Split (1)

train · 1.32k rows

id stringlengths 16 65	title stringlengths 19 278	text stringlengths 894 136k	keyphrases listlengths 1 34	prmu listlengths 1 34
Anal_Bioanal_Chem-3-1-2117333	Modification and re-validation of the ethyl acetate-based multi-residue method for pesticides in produce	The ethyl acetate-based multi-residue method for determination of pesticide residues in produce has been modified for gas chromatographic (GC) analysis by implementation of dispersive solid-phase extraction (using primary–secondary amine and graphitized carbon black) and large-volume (20 μL) injection. The same extract, before clean-up and after a change of solvent, was also analyzed by liquid chromatography with tandem mass spectrometry (LC–MS–MS). All aspects related to sample preparation were re-assessed with regard to ease and speed of the analysis. The principle of the extraction procedure (solvent, salt) was not changed, to avoid the possibility invalidating data acquired over past decades. The modifications were made with techniques currently commonly applied in routine laboratories, GC–MS and LC–MS–MS, in mind. The modified method enables processing (from homogenization until final extracts for both GC and LC) of 30 samples per eight hours per person. Limits of quantification (LOQs) of 0.01 mg kg−1 were achieved with both GC–MS (full-scan acquisition, 10 mg matrix equivalent injected) and LC–MS–MS (2 mg injected) for most of the pesticides. Validation data for 341 pesticides and degradation products are presented. A compilation of analytical quality-control data for pesticides routinely analyzed by GC–MS (135 compounds) and LC–MS–MS (136 compounds) in over 100 different matrices, obtained over a period of 15 months, are also presented and discussed. At the 0.05 mg kg−1 level acceptable recoveries were obtained for 93% (GC–MS) and 92% (LC–MS–MS) of pesticide–matrix combinations. Introduction For monitoring and control of pesticide residues, multi-residue methods are very cost-effective and are used in many laboratories. The pesticides are usually first extracted with an organic solvent of high or medium polarity. Typical solvents used for this purpose are acetone [1–4], ethyl acetate [5–26] (Table 1), and acetonitrile [26–31]. With all three options, pesticides are partitioned between an aqueous phase and an organic phase. With acetone and acetonitrile this is done in two successive steps, with ethyl acetate in one step. With regard to extraction efficiency, ethyl acetate has been shown to be equivalent to the water-miscible solvents for both polar and non-polar pesticides in vegetables, fruit, and dry products (after addition of water) [6, 7, 26, 32]. It is also suitable for products with a high fat content—because of the solubility of fat in ethyl acetate, pesticides are released and extracted efficiently. The extract obtained is compatible with gel-permeation chromatography (GPC), the clean-up procedure most suitable for this type of sample. Ethyl acetate is very suitable for GC analysis. It has good wettability in GC (pre)columns; this is of benefit for solvent trapping of the most volatile analytes, which is required for refocusing after injection. Its vapor pressure and expansion volume during evaporation also favor large-volume injection. Finally, it is compatible with all GC detectors. The same extract can also be used for LC analysis, after a solvent change into, e.g., methanol [11, 15–18, 26], as is done for acetone-based methods also [33]. Table 1Examples from literature. Conditions typically used in ethyl acetate-based multi-residue analysisSample (g)AdditionEtAc (mL)Na2SO4 (g)Extr.Phase separationRe-extr.Evap./reconst. (aliquot/to mL)Clean-upEvaporation (from/to mL)Final extr. g mL−1Inj. (μL)Analysis 50–10050B–5→1GPC None0.1910GC–NPD/ECD1987 [5]75–20040TF/Na2SO4–100→5GPCEluate→51.5?GC–NPD, FPD1991 [6](dilute)0.3?GC–ECD5–2010TLet settle–10→12.51–5GC–FPD/NPD1992 [7]15 mL H2O (wheat)0.550–10050TF/Na2SO4–0.52–8GC–MS/FPD/ECD1996 [4]50–250100BF/Na2SO4–All→100GPCEluate→111GC–NPD/ECD1998 [8]75–20040TF/Na2SO4–100→5SPE (ENV+)3 mL1.252GC–ITD/ NPD/ECD1999 [9]20–100TSee clean-up–Cartridge water abs. Polymer+ GCB/Na2SO450→dry→2 ace/hex12GC–MS, GC–NCI-MS2001 [10]GC–FPD LC–PCR-Flu82 g NaHCO35070TFYesAll→20 MeOH0.45–10LC–MS–MS2002 [11]25–10075TF (vac)–All→25 +25 cyclohexaneGPCEluate→111GC×GC–TOF-MS, GC–TOF-HRMS2003 [12]F/Na2SO4Rinse2004 [13]305–6 g NaHCO36030–40T (30 °C)F/cotton wool–1+0.1 IS→1–0.510GC–TOF-MS (DMI)2003 [14]25–5025TLet settle or centrifuge–1→1 H2O–0.520LC–MS–MS2003 [15]75NaOH if pH < 4.520040TF/Na2SO4–100→5–5→MeOH2.510LC–MS–MS2004 [16]151 mL 6.5 mol L−1 NaOH9013TF/Na2SO4RinseAll→15 MeOH–110LC–MS–MS2004 [17]151 mL 6.5 mol L−1 NaOH90TF/Na2SO4Yes 2×All→15 MeOH–150LC–TOF-MS2005 [18]10–50a10BF–All→5GPCEluate 35→2210GC–MS–MS2006 [19]6–503BCentr.Yes All→5GPCEluate 84→1b51GC–MS2006 [20]20–8050–100TFYesAll→ace/hexSPE SAX/PSA All→32.42GC–ECD2006 [21]50–10075TF/Na2SO4RinseAll→10–52GC–NPD/MS2006 [22]5–10TF/Na2SO4RinseAll→1–510GC–MS–MS2006 [23]2.5–52TF (syringe)––0.550GC–NPD2006 [24]305–6 g NaHCO36030–40T (30 °C)F–1→0.9 +0.1 IS–0.520GC–FPD2006 [25]510 mL H2O (barley)5015SF/Na2SO4–25→1GPCEluate→10 ACN 0.2525GC–TOF-MS, LC–MS–MS2006 [26]252 mL 4 mol L−1 phosphate buffer4025TCentrifuge–GC: -GCB/PSA disp0.520GC– MSThis workLC: 0.48→1.5 (MeOH/water)–0.210LC–MS–MSaEthyl acetate–cyclohexane, 1:1bAdditional SPE clean-up step with Florisil EtAc/Hex 1:1 5 mL evap. to 1 mLT, Turrax; B, blender; S, shaking; F, filtration; MeOH, methanol; ACN, acetonitrile; ace, acetone; hex, hexane Although multi-residue methods based on ethyl acetate extraction have been used for more than 20 years, and continue to be used in many laboratories (they are, for example, the official methods in Sweden and Spain and are also commonly used in the Netherlands, UK, Czech Republic, Japan, and China), the methods described in the literature frequently include steps that make them, in our opinion, unnecessary laborious. Such steps include repeated extraction, filtration, clean-up steps involving GPC for non-fatty matrices, column chromatography or solid phase extraction (SPE) manifolds and evaporative concentration. Typical examples are given in Table 1. It will be shown in this paper that most of the laborious steps can be replaced by more efficient alternatives—repeated extraction is not required, an aliquot is taken after settling or centrifugation rather than filtration, use of GCB instead of GPC for removal of chlorophyll, use of dispersive SPE instead of classical SPE for clean-up (analogous to an acetonitrile-based method [29]), and injection of larger volumes into the GC instead of manual evaporative concentration. The objective of the work discussed in this paper was to update and improve the ethyl acetate-based multi-residue method for pesticides in vegetables and fruit in respect of straightforwardness, robustness, and ease and speed of sample and extract handling. Aspects studied include dispersive clean-up using combined GCB/PSA, the possibility of preventing unacceptable adsorption of “planar” pesticides by GCB, by addition of toluene, and large-volume (20 μL) injection in GC. The method has been validated for 341 pesticides and degradation products which are analyzed by GC–MS or LC–MS–MS. For the latter the initial raw extract was used and injected after a solvent change to methanol–water. The suitability of the method as a multi-residue, multi-matrix method is evaluated by use of analytical quality-control data generated during 15 months for 271 pesticides and degradation products for over 100 different matrices, including less common and exotic crops. Results obtained for proficiency test samples during three years are also presented. Experimental Chemicals and reagents Pesticide reference standards were obtained from C.N. Schmidt (Amsterdam, The Netherlands). For GC–MS a mixed stock solution containing 135 pesticides (Table 7; concentration 50 mg L−1 for each pesticide) was obtained from Alltech–Grace (Breda, The Netherlands). The full chemical names of the metabolites of phenmedipham and pyridate are methyl N-(3-hydroxyphenyl)carbamate and 3-phenyl-4-hydroxy-6-chloropyridazine, respectively. Solvents were from J.T. Baker (ethyl acetate, Resi-analysed; Deventer, The Netherlands), Labscan (toluene, Pestiscan), and Rathburn (methanol). Anhydrous sodium sulfate, ammonium formate, potassium dihydrogen phosphate, disodium hydrogen phosphate, acetic acid, and diethylene glycol (all p.A. quality) were from Merck. Water was purified by use of a MilliQ reagent-water system (Millipore). Bondesil primary secondary amine (PSA, 40 μm) was obtained from Varian (Middelburg, The Netherlands) and GCB (graphitized carbon black) was purchased as Supelclean ENVI-carb (120–400 mesh, Supelco, Zwijndrecht, The Netherlands). For GC–MS, in addition to the mixed stock solution, individual stock solutions of other pesticides were prepared in ethyl acetate. From these, additional mixed solutions were prepared in ethyl acetate. For LC–MS–MS analysis, individual stock solutions were prepared in methanol. Mixed solutions were prepared from the individual stock solutions and diluted with methanol. The mixed solutions were used for fortification of samples and for preparation of matrix-matched standards. The extraction solvent was a solution of internal standard (0.05 mg L−1 antor (diethatyl-ethyl)) in ethyl acetate. Matrix-matched standards were prepared by addition of mixed solutions to control sample extracts. Dilution of the sample extract with mixed solution was never more than 10%. Instrumentation GC–MS analysis GC–MS analysis was performed with a model 8000 Top GC equipped with a Best PTV (programmed temperature vaporizer) injector, an AS800 autosampler, and a Voyager mass spectrometer (Interscience, Breda, The Netherlands). The instrument was controlled by Masslab software. The injector was equipped with a 1 mm i.d. liner with porous sintered glass on the inner surface. The GC was equipped with a 30 m × 0.25 mm i.d., 0.25 μm film, HP-5-MS column and a 2.5 m precolumn (same as the analytical column, connected by means of a press-fit connector). For PTV injection in solvent-vent mode 20 μL was injected at 5 μL s−1. The solvent was vented at 50°C in 0.67 min using a split flow of 100 mL min−1. The split valve was then closed and the analytes retained in the liner were transferred to the GC column by ramping the temperature at 10° s−1 to 300°C. Total transfer time was 2.5 min after which the split was re-opened. Helium was used as carrier gas at constant flow (1.5 mL min−1). The oven temperature was maintained at 90°C for 2 min after injection then programmed at 10° min−1 to 300°C which was maintained for 10 min. The transfer line to the MS was maintained at 305°C. Mass spectrometry was performed with electron-impact (EI) ionization (electron energy 70 eV) at a source temperature of 200°C. Data were acquired in full-scan mode (m/z 60–400), after a solvent delay of 5.5 min, until 30 min. Scan time and inter-scan delay were 0.3 and 0.1 s, respectively, resulting in 2.5 scans s−1. The detector potential was 450 V. Masslab software (Interscience, The Netherlands) and an Excel macro developed in-house were used for data handling and quantitative data evaluation. LC–MS–MS analysis LC was performed with an Agilent, model 1100 instrument comprising degas-unit, pump, autosampler, and column oven. A 4 mm × 2 mm i.d. C18 guard column (Phenomenex) and a 150 mm × 3 mm i.d. LC column (Aqua, 5 μm C18, Phenomenex) were coupled to a triple-quadrupole mass spectrometer (model API2000 or API3000, Applied Biosystems, Nieuwerkerk a/d Yssel, The Netherlands). Analyst 1.2 and, later, 1.4 were used for instrument control and data handling. Additional data processing was performed using an Excel macro developed in-house. Compounds were separated by elution with a gradient prepared from methanol–water–1 mol L−1 ammonium formate solution, 20:79.5:0.5 (component A) and methanol–water–1 mol L−1 ammonium formate solution, 90:9.5:0.5 (component B). The composition was changed from 100% A to 100% B in 8 min and was then isocratic until 24 min. The composition was then changed back to 100% A in 1 min and the column was re-equilibrated for 10 min before the next injection. The flow rate was 0.3 mL min−1 which was introduced into the MS without splitting. The injection volume was 20 μL and 10 μL for the API2000 and API3000, respectively. Data were acquired in multiple-reaction-monitoring (MRM) mode. Electrospray ionization (ESI) (called turbo ion spray for the instruments used) mass spectrometry was performed in positive-ion mode. For the API2000 the nebulizer gas, turbo gas, and curtain gas were 20, 50, and 40 arbitrary units (a.u.), respectively. The ion-spray potential was 5000 V. Nitrogen was used as collision gas (4 psi). For the API3000 the nebulizer gas and curtain gas were 12 and 10 a.u. and the turbo gas was 7.5 L min−1. The ion spray potential was 2000 V. Nitrogen was used as collision gas (4 psi). For both instruments, the pause time was 5 ms. The dwell times for the pesticide transitions varied between 10 and 25 ms. The precursor and product ions and the collision energy (data for API3000) for each pesticide or degradation product are listed in Table 8. In the acquisition method one transition for each pesticide was measured. All transitions were acquired in one time window. The total cycle time was 2.24 s resulting in 8–10 data points across the peak. To measure the second transition a second method was created and run if confirmation was needed. Sample preparation Vegetable and fruit samples were taken from batches of samples as received from the food industry and trade for routine multi-residue analysis. After removal of stalks, caps, stems, etc., as prescribed by 90/642/EEC Annex I [34], an amount corresponding, at least, to the minimum size of laboratory samples (usually 1–2 kg [35]) was homogenized in a large-scale Stephan food cutter. A subsample (25 g) was weighed into a centrifuge tube. Fortification was performed at this stage. Phosphate buffer (pH 7, 4 mol L−1, 2 mL) and extraction solution (ethyl acetate with internal standard, 40 mL) were then added. Just before Turrax extraction anhydrous sodium sulfate (25 g) was added. After Turrax extraction (1 min) the tubes were centrifuged (sets of four). For GC–MS analysis, Eppendorf cups were prefilled with 25 mg PSA and 25 mg GCB. To avoid a weighing step, scoops were made in-house for this purpose. Their accuracy was established to be 25 ± 2 mg (n = 10). For clean-up, 0.8 mL extract and 0.2 mL toluene were added to the cup with the SPE materials. The cups were then closed and the samples were vortex mixed for 30 s and centrifuged (up to 24 at one time). One aliquot was transferred to an autosampler vial with insert, and a second aliquot was transferred to an autosampler vial and stored under refrigeration as back-up extract. The calculated amount of initial sample in the final extract was 0.5 g mL−1. For LC–MS–MS analysis the initial extract (3.2 mL for the API2000 and 0.48 mL for the API3000) was transferred to a disposable glass tube. After addition of a solution of diethylene glycol in methanol (10%, 200 μL) the extract was evaporated to “dryness” under a gentle flow of nitrogen gas at 35°C (up to 36 tubes in a heater block). The residue was reconstituted in methanol (1 mL and 0.75 mL for the API2000 and API3000, respectively), by use of vortex mixing and ultrasonication (5 min). The extract was then diluted 1:1 with component A. After centrifugation one aliquot was transferred to an autosampler vial with insert, and a second aliquot was transferred into an autosampler vial and stored under refrigeration as back-up extract. The final extract concentration was 1 g mL−1 and 0.2 g mL−1 for the API2000 and API3000, respectively. For dry products (e.g. cereals) 5 g was weighed and 20 mL water was added. After soaking for 2 h samples were processed as described above. A larger amount of extract was taken for evaporation to compensate for the reduced amount of sample processed and to bring the final extract concentration to 0.2 g mL−1. With the final method, one person can process 30 samples in eight hours. Here processing includes specific preparation before homogenization (i.e. removal of caps from strawberries, etc.), homogenization of the samples, extraction, cleaning the Turrax between samples, clean-up for GC–MS, and solvent switch for LC–MS–MS, i.e. from laboratory sample to ready-to-inject solutions in autosampler vials. Quantification GC–MS For each pesticide the concentrations were calculated for two diagnostic ions. In previous validation work (not published) using the same software it was found that for most pesticides automatic integration and repeatability of response were better when peak height, rather than area, was used. Peak height was therefore used, with few exceptions (e.g. pesticides prone to tailing, for example 2-phenylphenol). All responses were normalized to the response of the internal standard (antor). One-point calibration was performed using a fixed matrix-matched standard (tomato, see Results and discussion section) at a level corresponding to five times the LOQ. The linearity of the plot of MS response against concentration was verified periodically over the range 0.01 to 1–5 mg kg−1. For most pesticides linearity was adequate (relative response within 20% of the calibration standard) up to at least 1 mg kg−1. LC–MS–MS The internal standard (antor) was evaluated qualitatively only to confirm injection of the sample extract. Because of unpredictable and varying matrix effects for several of the matrices included in this work, normalization against the internal standard was not considered feasible. For each sample matrix that was fortified, a matrix-matched standard was also prepared by spiking the final extract of the corresponding control sample. Peak area was used for quantification. One-point calibration was performed using the matrix-matched standard at a level corresponding to five times the LOQ. Linearity of the MS response against concentration was verified periodically over the range 0.01 to 1 mg kg−1. For most pesticides, the relationship was linear (relative response within 20% of the calibration standard) up to at least 0.5 mg kg−1. Validation Initial method validation was performed in accordance with EU guidelines [36, 37]. Two times five portions of the homogenized sample were spiked with a mixture of pesticides at a low level (0.01 mg kg−1 or lower) and at a level ten times higher. Together with two unfortified control portions of the sample, they were processed and analyzed as outlined above. Additional method-performance data were acquired by analyzing fortified samples concurrently with each batch of samples. The spike level (0.05 mg kg−1 for most pesticides) was five times the LOQ. With each batch different products were selected as much as possible. In the compilation the emphasis was on products which are less frequently reported in the literature to challenge the applicability of the method as a “multi-matrix method”. For this purpose samples were not pre-screened for absence of pesticides and, consequently, occasionally recoveries could not be determined, because of the relatively high levels incurred. Such results were eliminated from the data set. Spectrophotometric measurement of removal of chlorophyll For evaluation of the removal of chlorophyll by GCB and comparison with GPC, a lettuce extract was prepared by extracting 25 g lettuce with 40 mL ethyl acetate after addition of 25 g anhydrous sodium sulfate. As a reference, 0.8 mL ethyl acetate was added to 3.2 mL of this extract to bring the extract concentration to 0.5 g mL−1. For dispersive SPE, 100 mg GCB was added to sets of duplicate tubes and 3.2 mL extract was added to all tubes. Solvent was then added to four sets of tubes: set one 0.8 mL ethyl acetate, set two 0.4 mL ethyl acetate and 0.4 mL toluene (i.e. 10% toluene), set three 0.8 mL toluene (20% toluene), and set four 0.8 mL xylene (20% xylene). The extracts were vortex mixed and centrifuged. For GPC clean-up, 2.5 mL lettuce extract was injected on to a 40 cm × 28 mm i.d. Biobeads SX3 column with 1:1 ethyl acetate–cyclohexane as eluent. The fraction collected was such that at least 50% of the pyrethroids were recovered (fraction from 105–200 mL). The eluate was first concentrated, by rotary evaporation at 40°C, to approximately 5 mL, then transferred to a tube for further concentration, under nitrogen gas, to 2.5 mL. Final extract concentration before and after clean-up was always 0.5 g mL−1. Aliquots of the extracts were transferred to a cuvet for spectrophotometric analysis at 450 nm. If required, the extracts were diluted with ethyl acetate to bring absorption within the linear range. The amount of chlorophyll in the uncleaned extract was defined as 100%. For calibration purposes the uncleaned extract was diluted 10, 20, 40, 50 and 100 times with ethyl acetate and a calibration plot was constructed. Chlorophyll remaining after clean-up was determined from the decrease in absorption at 450 nm compared with the absorption of the uncleaned lettuce extract. Results and discussion Monitoring of residues in fresh produce for the food industry, especially trade and retail, calls for rapid turnaround, preferably within one or two days. This means sample preparation must be rapid and straightforward. With regard to cost and waste, consumption of solvents and reagents should be low. At the same time, EU directives with regard to sample definition (90/642/EEC, [34]) and laboratory sample size (2002/63/EC [35]) for residue analysis should be respected. This means, for example, that that a total of 2 kg grapes (after removal of stalks), five whole melons, or 1 kg strawberries (after removal of caps) must be processed. The actual analysis is performed on a subsample of the laboratory sample, after appropriate comminution. The more thorough the comminution, the smaller the subsample can be and the lower the amount of solvent needed for extraction. It has, furthermore, been reported that for well homogenized samples extraction by vortex mixing or shaking, instead of high-speed blending (Turrax) suffices for effective extraction [29], although there is still some debate on this matter [38]. Homogenization For homogenization there are several possibilities. Food choppers or kitchen blenders are often used. Very thorough homogenization can be achieved with the latter, but it is not possible to process the entire laboratory sample at once. For this reason, large-scale food choppers are more suited. With such devices, homogeneity is not always optimum, as can be observed with, e.g., tomatoes, for which small pieces of skin drift in the “soup” obtained after homogenization. Subsampling of very small amounts is, therefore, not acceptable after this procedure, because the subsample would be insufficiently representative of the original sample. More thorough homogenization can be achieved after addition of dry-ice or liquid nitrogen (cryogenic homogenization). This procedure is recommended when reducing the subsample for analysis to 10 g. This procedure is more laborious, however, because it involves cutting the sample into pieces, freezing the sample (usually overnight), cryogenic comminution, then dissipation of the dry-ice or liquid nitrogen before further processing or storage. It also puts higher demands on the cutter (blades) and requires additional precautions for the operators (protection against low temperatures and noise). Cryogenic comminution has been recommended for some pesticides because it reduces their degradation during this step [39]. In recent years the food trade and retail have been intensifying their residue-monitoring programs and require analytical data before harvest, before accepting an assignment, or before releasing their products from distribution centers to supermarkets. For fresh produce this means there is a much pressure on laboratories for rapid turnaround (24–48 h). This is difficult to achieve when the analysis involves overnight freezing for cryogenic comminution. Thus, for reasons of ease and speed, it was decided to retain the current procedure—ambient homogenization of the entire laboratory sample by use of a large scale food cutter (thus accepting the consequence that for a limited number of pesticides the concentration found might be an underestimate). Because of non-optimum homogenization with the food cutter, subsamples should not be too small, and further comminution is required for efficient extraction of systemic pesticides. This can be achieved during extraction by use of an Ultra Turrax. We have previously established the minimum size of subsample that did not negatively affect the repeatability of the analysis. This was done with samples which contained residues. For subsamples (n = 7) of 50 and 25 g, the relative standard deviation (RSD%) was below 8% for several pesticide–matrix combinations. For pear leaves (regarded as a difficult matrix to homogenize) containing bromopropylate, phosalone, and tolylfluanide it was observed that the RSD increased from <8% to 14–18% when the amount of subsample was reduced from 25 g to 12.5 g. From this it was concluded that, with our procedure, 25 g was the minimum required amount of subsample. pH adjustment In the ethyl acetate-extraction procedure analytes are extracted and partitioned between water (from the matrix itself, or added water for dry crops) and ethyl acetate in one step. For basic and acidic compounds the partitioning can be affected by pH, which can vary substantially with the matrix. Because the same extract is to be used not only for GC–MS but also for LC–MS–MS (after changing the solvent to methanol) which, preferably, should also include analysis of basic and acidic pesticides, control of pH was regarded as necessary. A pH of approximately 6 was chosen as compromise for efficient extraction of basic and acidic compounds. Although acidic pesticides were not included in this work, data in the literature (for barley without pH adjustment, i.e. non-acidic conditions [26]) indicate they are extracted into ethyl acetate. For pH adjustment others have used sodium hydroxide [16–18] or sodium hydrogen carbonate [11, 14, 25] (Table 1). A disadvantage of this is that the amount of salt needed depends on the acidity of the sample. Addition of too much will result in a high pH and possible degradation of base-sensitive pesticides. To keep the method as straightforward as possible the pH was adjusted using a solution of concentrated phosphate buffer (4 mol L−1, 2 mL). A solution was preferred over addition of solid salts because this enabled use of a dispenser and eliminated additional weighing of the salts. The buffer resulted in appropriate pH adjustment for most matrices, although there were exceptions, for example lemon and lime. Extraction The two conditions most relevant to extraction efficiency are the sample-to-solvent ratio and addition of salt, which in ethyl acetate-based multi-residue methods has always been sodium sulfate. The amount of ethyl acetate (in mL) relative to the amount of sample (in g) is, typically, at least 2:1. This ratio has been used for many years (Table 1). It results in good extraction efficiency and is practical with regard to achieving phase separation and avoidance of emulsions. To avoid sacrificing decades of method history no attempts were made to reduce the ratio; to do so might also adversely affect recovery and/or complicate phase separation. Larger amounts (as used by several other laboratories; Table 1) result in greater solvent consumption and more dilute extracts. In previous work [15] it has been shown that the efficiency of extraction of polar pesticides improves with the amount of salt added. When 50 mL ethyl acetate and 25 g sample were used, 25 g sodium sulfate was sufficient to obtain recoveries of 80% or better, even for very polar and highly water-soluble compounds, for example acephate and methamidophos. Because these recoveries were obtained with a single extraction it was found unnecessary to perform repeated extraction, as some laboratories are doing [11, 18, 20, 21]. For addition of the sodium sulfate an automatic salt-dispenser coupled to a balance, as is used in our laboratory, or a scoop, was found to be very convenient. The extraction procedure involves successive addition of buffer, extraction solution (ethyl acetate with internal standard), and sodium sulfate to the centrifuge tube containing the sample, after which the pesticides are extracted and partitioned in one step using a Turrax. During this step the subsample is further comminuted for efficient extraction of the pesticides from the matrix. Vortex mixing, shaking or sonication were regarded as less efficient for subsamples that were homogenized in a large-scale food cutter under ambient conditions, but this was not investigated, partly because a variety of samples containing residues would be required to do so in an appropriate manner. It was noted from the literature that filtration is often performed to separate the solid pellet from the liquid. Again, there is no real need for this step, which involves additional glassware and, occasionally, rinsing (diluting) of the extract. For many samples a clear ethyl acetate extract is obtained after settling; if not the tubes can be centrifuged. This is no more laborious than filtration and does not involve additional glassware. Because the same Turrax is used for several samples, carry-over is an aspect to be considered. Between samples the Turrax is cleaned first by rinsing with water, by means of a flow-through beaker, then by brief immersing in two beakers containing ethyl acetate. Using this procedure, carry-over was tested by analyzing a blank after a sample that had been fortified at 5 mg kg−1. Carry-over was less then 0.1%, indicating that the straightforward cleaning procedure was sufficient to avoid cross-contamination up to 5 mg kg−1 when setting reporting limits not lower than 0.01 mg kg−1. GC–MS analysis Clean-up In ethyl acetate-based multiresidue methods either no clean-up or GPC clean-up is performed. This has hardly changed over the years (Table 1). In contrast with acetone and acetonitrile-based methods, in which SPE is commonly employed, this has been reported only occasionally for ethyl acetate-based methods. Obana et al. [10] used a cartridge packed with layers of water-absorbing polymer and GCB. Sharif et al. [21] described a clean-up using SAX/PSA but the scope of the method was restricted to organochlorine and organophosphorus pesticides. Zhang et al. [20] used a clean-up based on Florisil and achieved adequate recovery of many pesticides but not the more polar organophosphorus pesticides. It has been stated that in GC analysis with use of highly selective detectors, for example MS–MS no clean-up is required, even when injecting 15 mg equivalent of matrix (green bean, tomato, pepper, cucumber, marrow, egg plant, and water melon [40]). Other laboratories experienced problems with contamination of the GC inlet and tried to solve this by automatic exchange of liner inserts [14, 41]. This is in agreement with our experience that injection of 10 mg matrix equivalent, especially for leafy vegetables, does result in rapid deterioration of system performance because of accumulation of non-volatile material in the inlet. This makes the system less robust, and frequent exchange of the liner (daily) and GC–pre column (weekly) is required. Another problem encountered with injection of the uncleaned extracts was a shift in the retention times of pesticides relative to that of the calibration standard for some sample extracts. This shift was insufficiently corrected by automatic adjustment of retention times relative to that of the internal standard. Typically, shifts were in the range 0.05–0.20 min and were most abundant for the “azole” pesticides. Such shifts can complicate automatic peak assignment during data-handling. When data acquisition is performed in a non-continuous mode (e.g. selected-ion monitoring or MS–MS) such shifts also increase the risk of pesticides shifting from their acquisition window. For injection of relatively large amounts of matrix (e.g. 10 mg) in GC analysis clean-up for removal of bulk co-extractants is therefore regarded as a prerequisite for robust analysis of a wide variety of vegetable and fruit matrices. For vegetables and fruit matrices, chlorophyll (MW ∼900) and other pigments, for example carotenoids (e.g. β-carotene, MW 537) are typical bulk co-extractants. Most of these compounds are of low volatility and are not apparent as interferences in the chromatograms; they do, however, accumulate in the liner of the GC and eventually have an adverse effect on transfer of analytes to the column and/or on peak shape. Because of its high molecular weight, chlorophyll can be removed by GPC. A disadvantage is that the extract is strongly diluted and reconcentration by rotary evaporation is almost inevitable when LODs of 0.01 mg kg−1 are required. Such a step would contribute substantially to overall sample-preparation time. Although a very efficient on-line combination of GPC and GC–MS was described recently [42], avoiding GPC whenever possible would be even more straightforward. Solid-phase extraction is an alternative clean-up procedure which involves less dilution and is less laborious. Even more efficient is SPE in the so-called dispersive mode, as described by Anastassiades et al. [29]. Here the solid phase is simply added to the extract, thereby avoiding typical SPE procedures such as conditioning, sample transfer, elution, and evaporative reconcentration. The pesticides partition between the solid phase and the solvent and after vortex mixing and centrifugation the supernatant is ready for analysis. Two stationary phases, graphitized carbon black (GCB) and phases with amino functionality, have been shown to be particularly effective for removing co-extracted material from the raw extract while not removing most of the pesticides; this makes them very suitable for wide-scope methods [28, 29, 31, 38, 43–45]. Initially, a method was envisaged using SPE column clean-up with GCB, because for leafy vegetables this was found to be the only sufficiently effective alternative to GPC. After the publication on dispersive SPE [29] it was decided to investigate this approach, thus sacrificing some clean-up potential (as has been reported in the literature [31]) for ease and speed. GCB is well known to adsorb planar molecules, including chlorophyll and other pigments but also pesticides with planar functionality. In acetonitrile-based methods, toluene (typically 25%) is often added to the eluent to desorb these pesticides also from the SPE column [28, 38, 43, 45]. One of the objectives of this work was to investigate the possibility of using GCB in a dispersive clean-up step without unacceptable losses of planar pesticides. First we investigated which pesticides, dissolved in ethyl acetate, are adsorbed by GCB. A somewhat arbitrary, 25 mg mL−1 GCB phase was added to standard solutions. After vortex mixing and centrifugation the solution was analyzed by GC–MS (165 pesticides) and, after changing the solvent to methanol, by LC–MS–MS (another 70 pesticides), and the responses were compared with those from untreated standard solutions. For 35 pesticides (15%) adsorption was observed (Table 2). In addition to the pesticides included in this test, it is known from the literature [44] that chinomethionate, furametpyr, and pyraclofos are also adsorbed by GCB (from acetone–cyclohexane, 1:4). Table 2Pesticides adsorbed by GCBaStrong adsorption (rec. 0–50%)Medium adsorption (rec. 50–70%)Not consistentMeasured by GC–MS ChlorothalonilAzinphos-ethylPhosmet CyprodinilAzinphos-methylProchloraz FenazaquinChlorpyrifos-methylPyrazophos HexachlorobenzeneDicloranTrifluralin MepanipyrimEPN PentachloroanilineFenamiphos PhosalonePhorate PyrimethanilQuintozene QuinoxyfenMeasured by LC–MS–MS CarbendazimFenpyroximate ClofentezineFlufenoxuron DesmediphamTricyclazole DiflubenzuronTriflumuron FlucycloxuronThiophanate-methyl Hexaflumuron Phenmedipham Pymetrozine ThiabendazoleaPesticides in ethyl acetate, 25 mg GCB mL−1 solventrec., recovered To investigate how much toluene is required to prevent adsorption of planar pesticides by GCB in dispersive SPE, the partitioning experiment was repeated with standard solutions of 10, 20, or 30% toluene in ethyl acetate. This was done for the GC–MS pesticide mixture only. As is apparent from Fig. 1, even 10% toluene dramatically improved recovery. With 20% toluene recovery of all pesticides was higher than 65%. It should be noted that this experiment with standard solutions is the worst case. For real samples chlorophyll and carotenoids will also affect the distribution in favor of the pesticides in solution. Use of 30% of toluene further improved recovery only slightly. Twenty percent was regarded as optimum with regard to distribution and ease of solvent elimination in large-volume injection (see below). In addition to toluene, two alternative analogues, benzene and xylene, were also considered. Benzene, was not tested because it could not be used in routine practice because of its carcinogenic properties (although it would have been favorable with regard to solvent elimination). Xylene was tested in a similar way as toluene. Results obtained for hexachlorobenzene and chlorothalonil by use of the two solvents are compared in Fig. 2. Slight but consistently better recovery was obtained with xylene—>70% recovery could now be obtained for all pesticides. Because of its greater volatility, however, toluene was finally selected. Fig. 1Effect of the amount (%) of toluene in ethyl acetate on recovery of pesticides adsorbed by GCB (25 mg mL−1). hcb, hexachlorobenzene; pca, pentachloroaniline; ctn, chlorothalonil; mep, mepanipyrim; cypr, cyprodinil; pyri, pyrimethanil; fena, fenazaquin; quin, quinoxyfen; pyra, pyrazophos; epn, EPNFig. 2Comparison of toluene and xylene as additives for preventing adsorption of planar pesticides by GCB in dispersive SPE Obviously, toluene is also likely to affect adsorption of chlorophyll and/or carotenoids and might reduce the effectiveness of clean-up. To investigate this, a lettuce extract was prepared, the dispersive clean-up experiments were performed with different amounts of toluene, and removal of chlorophyll was verified. Visually it was clearly apparent that, despite addition of toluene, the intense green color turned light yellow, indicating that chlorophyll was removed to a large extent. To enable more quantitative evaluation, the extracts were also measured with a spectrophotometer at 450 nm. For comparison, the same extracts were also cleaned by GPC. The results are presented in Table 3. Without toluene, chlorophyll was very effectively removed. Absorption at 450 nm was reduced by 94%. Toluene, as expected, reduced adsorption of chlorophyll, but removal was still 87% or 78%, after addition of 10% or 20% toluene in ethyl acetate, respectively. Similar to observations with the planar pesticides, adsorption was reduced slightly more by use of xylene than by use of toluene. With GPC, chlorophyll removal was 60%. It should be noted here that the elution window was relatively wide, to include pyrethroids within the scope of the method. The elution windows for chlorophyll (and carotenoids) partially overlap those for pyrethroids, as has also been reported by others [44]. From these experiments it can be concluded that chlorophyll has more affinity than the planar pesticides for GCB. In dispersive SPE toluene effectively prevents unacceptable adsorption of planar pesticides while to a large extent maintaining its cleaning properties in respect of chlorophyll. Dispersive GCB not only enables much faster chlorophyll removal, it is also more effective when including pyrethroids in the scope of the method. For non-fatty vegetable and/or fruit matrices, therefore, GPC is not required and dispersive GCB clean-up is a much faster alternative without sacrificing scope. Table 3Removal of chlorophyll by dispersive SPE (GCB) and GPCClean-up procedureChlorophyll removal (%)Dispersive SPE, 100% ethyl acetate94Dispersive SPE, 10% toluene in ethyl acetate87Dispersive SPE, 20% toluene in ethyl acetate78Dispersive SPE, 20% xylene in ethyl acetate71GPC (fraction incl. pyrethroids)60Sample extract: lettuce 0.5 g mL−1. Dispersive SPE: 25 mg GCB mL−1. GPC: wide scope elution window, i.e. including pyrethroids. The GCB clean-up enabled continuous injection of extracts of leafy vegetables without rapid system deterioration. With some matrices, however (e.g. plums, grapefruit), retention time shifts were still observed. In addition, depending on the matrix, quite intensive interferences could be observed in the GC–MS TIC chromatograms. Further clean-up by PSA, complementing the GCB clean-up by removing compounds such as organic acids and sugars by hydrogen bonding, was therefore investigated. To keep sample clean-up as straightforward and rapid as possible focus was on a combined dispersive GCB/PSA clean-up. After the outcome of the GCB experiments, partitioning of the pesticides and co-extractants will be between PSA and ethyl acetate–toluene, 8:2. Because no information was available about the distribution of pesticides between these two phases, this was obtained by analyzing pesticide standards in ethyl acetate–toluene, 8:2, with and without PSA. Preliminary experience with dispersive PSA clean-up revealed that with some matrices (e.g. cereals) 25 mg mL−1 did not result in complete elimination of interfering compounds (e.g. fatty acids) typically removed by PSA. Partitioning with a much larger amount of adsorbent (200 mg mL−1) was, therefore, also studied. With 25 mg mL−1 losses of 30–40% were observed for sixteen pesticides, most probably as a result of adsorption, although the possibility of degradation induced by the basic nature of the PSA material could not be fully excluded. The findings were confirmed by the experiment with 200 mg PSA mL−1 (Table 4). The pesticides for which interaction with PSA was observed all had a C=O or P=O group in common (except for chlorothalonil). Our findings are not in full agreement with those of Anastassiades et al. [29] who did not observe losses as a result of using PSA. For this there can be two explanations. In our experiment adsorption was tested with standard solution rather than matrix. Co-extractants in matrix are likely to compete with the pesticides during adsorption. Second, with our method the organic phase (ethyl acetate–toluene, 8:2) is less polar than the acetonitrile phase; this could result in a stronger interaction between the polar functionality of the pesticides and amino functionality of PSA. From our results it became clear that with regard to the amount of PSA “the more, the better” does not apply. Another observation was that a hump appeared in the TIC chromatogram after a 20-μL injection of solvent mixed with 200 mg PSA mL−1. This hump, which eluted between 6 and 12 min, consisted of many peaks and a variety of masses. Cleaning of the PSA by washing with ethyl acetate (3 × 20 mL for 1 g), then drying by rotary evaporation, eliminated this contamination without affecting the clean-up properties. To keep the method straightforward, 25 mg PSA mL−1 was used as default, and the material was not cleaned before use. Table 4Adsorption of pesticides by PSAPesticideRecovery (%)Acephate43aAcrinathrin41bAsulam0aCarbaryl56bChlorothalonil17bCycloxidim39aDichlorvos33bDimethoate62bHymexazol0aMevinphos62bPhosmet25bPhosphamidon63bProfenofos56bPyridate40aPyridate-metabolite7aSethoxydim48aaAfter partitioning with ethyl acetate, 25 mg mL−1 and LC–MS–MS analysisbAfter partitioning with ethyl acetate–toluene, 8:2, 200 mg PSA mL−1 and GC–MS analysis The clean-up proved effective at reducing retention time shifts. As an example, for a plum extract without clean-up, the retention times of 24 pesticides (out of 140) were shifted by more than 0.05 min compared with the calibration standard. After clean-up this occurred for three pesticides only. With other matrices also shifts were reduced, but for some matrices (herbs, e.g. parsley) deviations were still quite common. As an illustration of the removal of co-extractants from the ethyl acetate extract (or, in fact, from the ethyl acetate–toluene, 8:2, extract) by dispersive GCB/PSA clean-up, GC–MS total ion current chromatograms of extracts obtained with and without clean-up are shown in Fig. 3. The most apparent differences are indicated. Several abundant matrix peaks are removed or strongly reduced. For lettuce, the overall background level between 15 and 25 min was also reduced. This clearly visible clean-up was mainly caused by the PSA material. With GCB alone differences between cleaned and uncleaned were much less apparent. The main benefit of GCB was prevention of rapid build up of non-volatile material (chlorophyll) in the liner, which enables prolonged use of the system without maintenance. Experience with method for more than three years and analysis of over 15,000 vegetable and fruit samples shows that, on average, the liner must typically be replaced weekly (after 150–200 injections; iprodion, dimethipin, and chlorfenapyr are the first for which response is lost). Further GC–MS maintenance consists in replacement of pre-column once of twice a month. The GC column is replaced approximately twice a year. The source of the MS is cleaned once a month. Fig. 3GC–MS chromatograms. Overlay total ion chromatograms (TICs) obtained after 20 μL injection of an extract of mandarin (top) and lettuce (bottom) without (higher peaks) and with clean-up In a continuing search for even further simplification of sample preparation, the possibility of combined extraction and dispersive SPE clean-up in one step was investigated. For two matrices (lettuce and mandarin, fortified with 140 pesticides, triplicate experiments) the solid phase materials (GCB/PSA, relative amounts similar to previous experiments) were added directly to the centrifuge tube containing the sample, sodium sulfate, and the extraction solvent (to which 20% toluene had been added). After Turrax extraction and centrifugation, the extract was ready for injection into the GC. Recovery was compared with that obtained by use of dispersive clean-up after separation of the ethyl acetate extract from the sample mixture. As could be seen from the color of the extract (the lettuce extract was almost colorless) the GCB remained effective. Adsorption of chlorophyll is based on planarity (shape) rather than polarity and, therefore, this will occur from both the aqueous and the organic phases. As was to be expected, the same was not true for PSA. The presence of water prevented adsorption of co-extractants with a hydroxyl group, i.e. almost identical GC–MS total-ion chromatograms were obtained from extracts which were not cleaned and from those cleaned in the centrifuge tube. Pesticide recovery obtained after use of successive or simultaneous dispersive SPE clean-up was very similar, although recovery of some pesticides in the combined approach was too high, because of co-elution of interferences. The final method therefore used successive extraction and dispersive SPE clean-up. Large-volume injection GC–MS analysis of sample extracts was performed in full-scan mode. This enables detection of any GC–amenable pesticide. Because system LOQ for a quadrupole mass spectrometer in full-scan mode is limited, conservatively estimated at 100 pg, 10 mg matrix equivalent must be introduced into the GC to reach a target LOQ of 0.01 mg kg−1. With an extract concentration of 0.5 g mL−1, this means 20 μL must be introduced into the GC. Off-line tenfold evaporative concentration and then 2 μL injection could also be performed, but this would involve clean-up of larger volumes of extract, the risk of loss of the volatile pesticides (e.g. dichlorvos), and an additional step in sample preparation. Although large-volume injection in GC is a well established technique [47, 48], many routine laboratories are still reluctant to apply it; if they do, the volume is often restricted to 5–10 μL. Such volumes can be accommodated in liners with a frit or even in empty (baffled) liners when injection speed is carefully adjusted. For larger volumes there is a risk of flooding [46], i.e. that extract is lost as liquid through the split exit. To prevent this, liners can be packed with a variety of materials. Packing materials often have the disadvantage of a large surface area with active sites, however, resulting in degradation and/or adsorption of thermo labile and/or polar pesticides; problems can also be encountered with splitless transfer of higher boiling pesticides (e.g. deltamethrin) from the liner to the GC column. Other disadvantages can be a pressure drop over the liner (slows down solvent elimination) and liner-to-liner variability requiring re-optimization of the solvent-elimination process after liner replacement. A means of by-passing the disadvantages of packed liners while still achieving accommodation of 20–50 μL of liquid was described in 1993 by Staniewski and Rijks [49]. They developed a liner with a sintered porous glass bed on the inner surface wall of the liner. The liquid is retained in the porous glass bed. The potentially active glass surface area is relatively small compared with the materials in packed liners. The gas flow is not obstructed, because the centre of the liner is empty. This enables efficient solvent vapor removal during solvent elimination and efficient transfer of analytes to the analytical column during splitless injection after solvent elimination. Since the early 2000s such liners have been commercially available for PTV injectors from several suppliers, and since then our laboratory has implemented 20 μL as default injection volume for ethyl acetate. After the development of the dispersive GCB clean-up, the solvent to be introduced into the GC contained 20% toluene, which might effect the processes involved in large-volume injection differently from 100% ethyl acetate. Because toluene does not evaporate azeotropically with ethyl acetate and is less volatile, it will be the main solvent left at the end of the evaporation process. Injection of 20 μL 20% toluene in ethyl acetate means that 4 μL toluene is introduced. The PTV used in this work was equipped with a 1 mm i.d. porous glass bed liner that could hold approximately 30 μL within the zone that is appropriately heated during splitless transfer. Up to this volume there is no need for optimization of injection speed. To obtain information about splitless transfer of the last few microliters of toluene after solvent elimination, cold splitless injections of 1, 2, and 3 μL of standards in 100% toluene were performed. Even with 2-μL volumes peak distortion (fronting peak shape) was observed for pesticides of medium volatility. With 1 μL injections peak shape was good and for several pesticides even better than for ethyl acetate. On injection of 20 μL standard in ethyl acetate–toluene, 8:2, in the solvent-vent mode, no peak distortion was observed, indicating that less then 2 μL toluene remained in the injector after the solvent-vent step. As observed earlier with large-volume injection of ethyl acetate, the vent time (here set at 40 s using an initial PTV temperature of 50°C) was not at all critical, even for the most volatile pesticide (dichlorvos). Venting for 35 or 50 s did not dramatically affect responses or peak shape of the pesticides. In our experience, this phenomenon is typical for porous glass bed liners and contributes to the robustness of the method. Validation of GC–MS method In the past a method based on simple ethyl acetate extraction followed by direct GC–MS analysis of the raw extract [4] had been validated for concentrations in the range 0.05–0.5 mg kg−1. The modified method described here involved a dispersive clean-up step, large-volume injection, and injection of ten times more matrix into the GC. Re-validation was therefore required, and focused on method performance at low concentrations. This was done using lettuce as matrix. The validation set consisted of two control samples, five fortifications at 0.001–0.05 mg kg−1 and five fortifications at a level ten times higher. Over 200 pesticides were included in the validation procedure. The results are presented in Table 5. For the 0.01–0.5 mg kg−1 concentration range the EU criteria (recovery 70–110%, RSD 30%, 20%, or 15% for ≤0.01, >0.01–0.1, and >0.1–1 mg kg−1, respectively [37]) were met for 184 of the 201 pesticides included in the validation. At a level a factor of ten lower (fortification in the 0.001–0.01 mg kg−1 range for most pesticides) 147 pesticides could still be detected and for most (78%) of these recovery and RSDs were acceptable. For many pesticides S/N ratios were surprisingly good and background-corrected mass spectra often contained sufficient diagnostic ions (or were even recognizable mass spectra) to enable identity confirmation, as is illustrated in Fig. 4. The limits of detection, defined as S/N = 3 for one favorable diagnostic ion for each pesticide, were determined on the basis of the signals from the low fortification levels and the average noise observed in duplicate control samples. The LOD was at or below 0.001 mg kg−1 for 78 pesticides, between 0.001 and 0.005 mg kg−1 for 73 pesticides, between 0.005 and 0.01 mg kg−1 for 29 pesticides, between 0.01 and 0.05 mg kg−1 for 16 pesticides, and higher for four pesticides. Table 5GC–MS re-validation data for pesticides in lettuce PesticidetR (min) m/z (quant)Level (mg kg−1)Rec. (%)RSD (%)Level (mg kg−1)Rec. (%)RSD (%)LOD (mg kg−1)1Acephate10.451360.0263540.2575890.0062Acrinathrin22.062890.018118150.1789490.0033Aldrin16.582650.003139250.0319420.0024Atrazine14.172150.00291210.0189870.0025Azinphos-methyl21.641600.0111990.09811070.0096Azoxystrobin25.803440.018280.0999250.0037Benalaxyl19.821480.0058590.0479080.0028Benzoylurea (deg)a8.9014111350.02511069Bifenthrin20.911810.0078490.0688913≤0.00110Biphenyl9.811540.00697100.0631015≤0.00111Bitertanol22.891700.0038390.0319040.00212Bromophos17.023310.0039970.0321052≤0.00113Bromopropylate20.943430.003103130.0328950.00114Bromuconazole20.861730.002109120.024916≤0.00115Bupirimate18.722730.0036180.0329150.00116Buprofezin18.681720.00285140.0199280.00117Cadusafos13.461580.002117180.02192110.00118Carbaryl15.841150.0049390.049380.00219Carbofuran14.101640.0038870.0339330.00220Chlordane, alpha-17.813730.0010.0159240.00221Chlordane, gamma-18.123730.0028470.0159640.00122Chlorfenvinphos17.473230.0038460.039750.00123Chloroaniline, 3-7.491270.0020.02525460.00324Chlorobenzilate19.102510.0050.059540.01025Chlorothalonil15.052640.004146150.0421369≤0.00126Chlorpropham13.081710.0060.0599560.01527Chlorpyrifos16.673140.003102160.03410250.00228Chlorpyrifos-methyl15.702860.00110550.0151026≤0.00129Chlorthal-dimethyl16.773010.0059070.0519140.00130Cinerin-118.671500.0538430.5289360.04131Clofentezine22.453040.0140.14101140.05032Cyfluthrin I23.332260.0419170.4079360.02333Cyfluthrin II23.602260.04110080.4078880.01634Cyhalothrin-lambda21.911810.003110100.0299360.00235Cypermethrin-I23.651630.018107290.1849650.00836Cypermethrin-II23.831810.01894160.1849750.00637Cypermethrin-III24.071810.01896100.1849660.01338Cyproconazole18.972220.00672200.0598870.00139Cyprodinyl17.192240.005105250.0518510≤0.00140Cyromazine14.471660.0130.1382560.04041DDE, o,p′-17.902480.0020.0159230.00942DDE, p,p′-18.502480.001110110.0151005≤0.00143DDT, o,p′-19.322350.00110290.0159470.00144DDT, p,p′-20.282350.00286110.0169580.00145Deltamethrin25.442530.02211490.22310650.01446Demeton-S-methyl-sulfone16.111690.0371150.3029190.00447Desmethylpirimicarb15.421520.0030.0267670.00548Diazinon14.701370.00298140.0199430.00149Dichlofluanid16.412240.0047990.044988≤0.00150Dichlorvos8.001850.00210760.018927≤0.00151Dicloran13.962060.00396160.02910620.00352Dicofol (as DCBP)16.752500.0050.049126330.01053Dieldrin18.562630.0040.0419560.00554Diethofencarb16.532670.0059850.0469660.00155Difenoconazole-I25.123230.02994100.2889530.00656Difenoconazole-II25.363230.0299190.2889930.00357Diflubenzuron (deg)6.631530.00512490.0510720.00258Dimethoate13.971250.0090.0919140.01759Dimethomorph25.883010.0219570.2078750.00260Diniconazole19.542680.0020.01889120.00361Diphenylamine12.761690.00386100.0287215≤0.00162Disulfoton14.81880.00510150.059630.00263DMSA13.192000.0058790.0529270.00264DMST14.372140.0050.05373320.01965Dodemorph16.951540.00567260.0469170.00266Edifenfos18.073100.00596100.059480.00167Endosulfan-alpha18.08239+1970.0050.0479350.01068Endosulfan-beta19.19195+2410.0050.0468710.02069Endosulfan-sulfate19.98274+2370.00582100.0479740.00470Endrin20.942450.0050.0519080.00671EPN20.571690.01103230.0999470.00172Epoxiconazole20.551940.0070.0669210.01073Esfenvalerate24.771250.0040.0369850.00874Ethion19.362310.0030.039730.00775Ethoprofos12.861580.00388170.0269350.00176Etofenprox23.851640.005100110.0499350.00477Etridiazole10.742110.0149580.1389840.00178Etrimfos15.012920.0039640.025935≤0.00179Famoxadone25.903300.019790.19650.00380Fenamiphos18.233030.0159760.1549111≤0.00181Fenarimol22.131390.0040.03810140.00882Fenazaquin21.221600.003152120.02711480.00183Fenbuconazole23.301290.0030.039230.00684Fenhexamid20.101770.0030.0269070.00485Fenitrothion16.252600.0010.0159580.00386Fenoxycarb20.891160.01511780.1549440.00287Fenpiclonil20.782380.0078850.0719280.00388Fenpropathrin21.051810.00577130.0592130.00189Fenpropimorph16.631280.0010.019320.00290Fenthion16.632780.0029970.023995≤0.00191Fenvalerate24.541670.0040.03610380.00692Fipronil17.573670.0028160.024949≤0.00193Flucythrinate-I23.771990.01793110.1749210.00494Flucythrinate-II18.511990.0179460.1749340.00495Fludioxonil19.052480.003113130.0279730.00196Flufenoxuron (deg)14.793310.012104130.118118190.00597Flusilazole18.702330.0066880.055876≤0.00198Flutolanil18.303230.0038190.025868≤0.00199Fluvalinate, tau-24.802500.02595110.2459550.004100Folpet17.651470.01696160.15991150.009101Fonofos14.552460.0059460.0499270.001102Formetanate15.271220.050.498102620.188103Formothion15.271700.005102130.0498940.004104Fuberidazole15.791840.00583290.05155170.001105Furalaxyl17.592420.00595100.05110190.002106Heptachlor12.192720.0010.0149250.003107Heptachlorepoxide-I17.453530.0030.03397120.004108Heptachlorepoxide-II17.363530.00196130.015948≤0.001109Heptenophos12.241240.0039550.03933≤0.001110Hexachlorobenzene18.332840.00575280.04996150.001111Hexaconazole18.322160.0020.028770.003112Imazalil18.372150.00579500.0577140.002113Iprodione20.753160.01210870.129540.004114Isofenphos17.462130.0050.0519330.010115Jasmolin-I19.361230.0530.5287750.100116Kresoxim-methyl18.732060.0149560.1399190.005117Lindane14.411830.00286180.029960.001118Linuron16.352480.0050.0487990.010119Lufenuron (deg)11.481760.011123200.11476340.004120Malathion16.431730.0030.0349850.005121Mecarbam17.493290.0030.0299350.004122Mepanipyrim18.072220.0010.0139280.002123Mepronil19.542690.0020.02387100.005124Metalaxyl15.952060.00392100.0289750.002125Metaldehyde8.87890.0050.05111620.021126Methacrifos11.281800.00397170.029854≤0.001127Methamidophos7.751410.02636240.25847150.005128Methidathion17.821450.00381200.0310150.001129Methiocarb16.261680.002109590.0277460.001130Methoxychlor21.032280.0020.02590100.003131Metoprene17.56730.0110450.1039330.003132Mevinphos10.361920.003104160.03991≤0.001133Monocrotophos13.431920.0468480.4568870.021134Myclobutanil18.661500.0060.0559750.012135Nuarimol20.283140.0050.0498970.008136Omethoate12.391560.00557190.05453140.002137Oxadixyl19.381630.0120.1249240.038138Oxydemeton-methyl (deg)6.631100.0050.0527970.010139Paclobutrazole18.112380.007197280.07906≤0.001140Parathion16.692910.011106260.1069160.004141Parathion-methyl15.712630.0028870.021942≤0.001142Penconazole17.352480.00390100.03944≤0.001143Permethrin-cis22.651830.00510170.0499870.003144Permethrin-trans22.771830.0010.0119870.001145Phenothrin-I21.401830.0059780.059290.001146Phenothrin-II21.511230.0059360.0593100.004147Phenthoate17.532740.00510380.0489150.001148Phenylphenol, 2-11.561700.0059660.0529540.001149Phorate13.562600.0059860.059250.001150Phosalone21.611820.00111750.0091015≤0.001151Phosmet20.901600.005123160.0521004≤0.001152Phosphamidon-I14.751270.01193160.1059030.002153Phosphamidon-II15.491270.0118990.1059120.005154Piperonyl butoxide20.361760.0040.03789100.010155Pirimicarb15.251660.00210190.02955≤0.001156Pirimiphos-methyl16.262330.0020.0168720.004157Prochloraz22.971800.0040.03810160.007158Procymidone17.682850.003104150.0299170.001159Profenofos18.423370.0059780.05295100.001160Propargite20.313500.010.1029670.020161Propham10.731790.0059750.0499450.001162Propiconazole-I19.892590.0149250.1418990.003163Propiconazole-II20.022590.0149050.1418790.002164Propoxur12.621100.0029660.02927≤0.001165Propyzamide14.581750.00576390.0469920.001166Prothiofos18.372670.00385190.03210190.001167Pyrazophos22.172210.003137110.031454≤0.001168Pyrethrins19.621230.0530.52899130.087169Pyridaben22.821470.0059690.0519430.001170Pyridaphenthion20.801990.00599100.0489350.003171Pyrifenox-I17.392620.0118470.1069560.003172Pyrifenox-II14.682640.0110.1069060.170173Pyrimethanil14.651980.002135140.021234≤0.001174Pyriproxyfen21.651360.002119180.024916≤0.001175Quinalphos17.551460.0047090.0418780.002176Quinoxyfen19.902720.001113130.01410513≤0.001177Quintozene14.502370.005106100.04610820.003178Simazine16.172010.0049190.0399570.002179Spiroxamine15.671980.01899170.1768120.009180TDE, o,p′-18.672350.0039950.028954≤0.001181TDE, p,p′-19.362350.00186100.014907≤0.001182Tebuconazole20.282500.0090.0899190.031183Tebufenpyrad21.121710.00592170.0528770.001184Tecnazene12.562030.00510860.0489960.002185Teflubenzuron (deg)8.121970.003174250.025124250.002186Tefluthrin14.911970.0010.01489140.002187Terbufos14.462310.00510080.052953≤0.001188Tetraconazole16.853360.0039530.026886≤0.001189Tetradifon21.443560.0030.039480.010190Thiometon13.78880.0059350.0551003≤0.001191Tolclofos-methyl15.802650.0019160.011025≤0.001192Tolylfluanid17.422380.00385170.0319620.002193Triadimefon16.752080.00790140.0659760.005194Triadimenol17.851680.0050.0538520.029195Triazamate17.952420.0030.02890100.010196Triazophos19.622570.005109370.05489200.001197Trifloxystrobin19.921160.00691130.05588110.002198Triflumizole17.702780.007102150.06680150.001199Trifluralin13.333060.00292190.019948≤0.001200Vamidothion17.95870.019*0.18710050.045201Vinclozolin15.711980.00597160.0479370.003aBenzoylurea(deg) = 2,4-difluorobenzamideLOD: Amount for which S/N = 3, or in the event of an interfering peak, the average peak height for fortified sample (n = 5) should be 3.3 times the average peak height for control sample (n = 2)Fortification level below LOD as defined aboveUnderlined values are outside EU criteria for method validationFig. 4GC–MS extracted-ion chromatograms obtained from lettuce with (upper traces) and without fortification with pesticides, and the corresponding mass spectra (upper, reference spectra; lower, background-corrected spectra from the sample). a, b, 0.005 mg kg−1 disulfoton (m/z 88); c, d, 0.002 mg kg−1 fipronil (m/z 367); e, f, 0.006 mg kg−1 biphenyl (m/z 154) This initial validation clearly showed it is possible to introduce 10 mg of matrix equivalent of generic extracts obtained after ethyl acetate extraction of leafy vegetables. Adequate quantitative data are obtained for most of the pesticides at levels of 0.01 mg kg−1 or even below. Detection limits were usually well below 0.01 mg kg−1 after full-scan acquisition with a single-quadrupole MS. This means that for most pesticides at the target LOQ of 0.01 mg kg−1 (i.e. the lowest maximum residue limit set in the EU for vegetables and fruit), the signal-to-noise ratio is adequate for reliable automatic integration of peaks and that confirmation of identity of the pesticide is possible from its mass spectrum or at least one or two other diagnostic ions. Pesticides that did not meet the EU criteria for quantitative analysis, and/or for which relatively high LODs were obtained, included many compounds known to be troublesome in GC analysis because of to their high polarity or thermal lability. Typical examples are acephate, cyromazine, dicofol (screened for as its degradation product dichlorobenzophenone), dimethoate, imazalil, metaldehyde, methamidophos, methiocarb, omethoate, and the benzoylureas (measured as one common and one compound-specific degradation product). The relatively low recovery of the polar organophosphorus pesticides (acephate, methamidophos, and omethoate) can be attributed to the GC measurement and not to poor extraction efficiency, as was apparent from LC–MS–MS analysis of samples using the same extraction technique (see section LC–MS–MS analysis). For several other polar or labile pesticides adequate quantitative data were obtained during this initial validation, but from previous experience and the results obtained after implementation of the method it was clear that for such compounds LC–based analysis is more robust than GC–MS analysis. Typical examples include carbaryl, carbofuran, clofentezin, monocrotophos, and oxydemeton-methyl. Analytical quality-control data from routine GC–MS analysis The initial validation data are continuously being supplemented by performance data generated as part of the analytical quality-control during routine analysis of the samples, to gain insight into reproducibility, robustness, recovery, and selectivity with other matrices. For this, with each analytical batch, one of the samples submitted for routine analysis was spiked with 135 pesticides at five times the target LOQ level (i.e. samples were spiked with 0.05 mg kg−1 of most of the pesticides). A compilation was made of recovery data from a period of 15 months which included analysis of approximately 100 different vegetable and fruit commodities. Given the wide variety of commodities, matrix-matched calibration is quite tedious and would substantially increase the number of standard solutions to be analyzed in the GC sequence. It was therefore decided to select one relatively simple matrix (tomato) as default for matrix-matched calibration, i.e. recoveries for all commodities were calculated against the tomato-matrix standard. For each pesticide, calculations were performed for two diagnostic ions. All together this resulted in approximately 30,000 values. According to the current EU guideline on quality control in pesticide residue analysis [37], the recovery obtained during routine analysis should be within 60–140%. An overview of the percentage of recovery values within or outside the 60–140% criterion for a wide variety of matrices is presented in Table 6. With such large number of pesticides (or, actually, diagnostic ions) and matrices, one failing combination or more occurred for most matrices. There are several causes for this. Main reasons for recovery below 60% could be poor extraction efficiency or incomplete transfer of the pesticides to the GC column (e.g. adsorption and/or degradation in a contaminated inlet). Higher recovery may occur when a compound from the matrix generates the same diagnostic ion as a pesticide and co-elutes with that pesticide (i.e. detection was not selective). Another reason could be that the matrix effect induced in the GC inlet [50] for a pesticide in a particular matrix is more pronounced than that in the tomato-based calibration standard. Table 6Overview of percentage of recovery valuesa within or outside the EU 60–140% criterion [37] after GC–MS analysis MatrixPercentage of all recovery valuesa60–140%<60%>140%1Beetroot100002Cucumber (1/2)100003Mint (1/2)100004Sharonfruit (1/2)100005Witloof100006Asparagus99107Bean sprouts99018Corn syrup99009Fennel leaves990110Grape990111Kohlrabi (1/3)991012Lima bean990113Pak choi (1/2)990114Pear concentrate990115Pumpkin990116Salsify990017Sharonfruit (2/2)990118Strawberry990119Sugar pea991020Taro990121Bitter cucumber980222Cucumber (2/2)981123Egg plant980224Kidney bean981125Kohlrabi (2/3)981126Mushroom980227Pineapple981128Sweet pepper980229Tomato puree (processed)980230Turnip 981031Turnip tops (1/2)980232Alfalfa971233Cauliflower971234Cherry970335Chestnut972136Endive970337Fig970338Kangkung (1/2)971239Kangkung (2/2)972140Ladies’ fingers970341Mango970342Pear puree (processed)970343Sorrel973044Soybean sprouts970345Asparagus bean961346Orange962247Potato leaves962248Rhubarb962249Artichoke950550Tangelo952351Tarrragon953252Wine (red)951453Apricot940654Chives (1/3)943355Chives (2/3)944256Dill leaves944257Melon puree (processed)941558Mineola941659Pak choi (2/2)942460Sugar water946061Broad bean931662Celery leaves (1/4)933463Chervil935264Dates937065Sweetcorn (1/3)934366Carrot921767Haricot bean920868Oregano925369Parsnip922670Fennel910971Green pea (1/2)914572Passion fruit (1/2)912773Celery leaves (2/4)906474Green pea (2/2)901975Lemon puree908276Mint (2/2)905577Pomegranate901978Purslane901979Water cress902880Lettuce897481Chili pepper (1/2)886682Chinese cabbage8701383Passion fruit (2/2)8731084Bamboo shoots8601485Celery leaves (3/4)867786Honey8614087Potato puree (processed)8614088Sugar pea8501589Turnip tops (2/2)8501590Lime8441291Blueberry8321692Potato8315293Celery leaves (4/4)8231594Green pea8211795Apple pulp (processed)8161396Cassava8191097Chives (3/3)8171298Kohlrabi (3/3)7802299Parsley (1/2)78616100Thyme (1/3)78220101Kale77617102Chili pepper (2/2)76159103Coriander leaves76186104Sweetcorn (2/3)75187105Sweetcorn (3/3)74917106Parsley (2/2)73207107Thyme (2/3)73324108Rocket72325109Thyme (3/3)66295110Golden berry (physalis)65134aRecoveries at 0.05 mg kg−1 (0.10–0.30 mg kg−1 for 22 pesticides). Calculated for 135 pesticides, two diagnostic ions each, against a standard prepared in blank tomato extract. The pesticides included are listed in Table 7 Failing pesticide–matrix combinations were most abundant for herbs, kale, sweetcorn, and golden berry, for which up to 35% of recovery values (calculated using the two diagnostic ions for each pesticide) were outside the 60–140% range. These products contain larger amounts of co-extractants than most other vegetables and fruits, which may result in insufficient detection selectivity, enhanced response as a result of a matrix effect (more shielding of active sites in the inlet), and contamination of the inlet. For this type of product more selectivity, e.g. by use of MS–MS would be beneficial. Such detection is also more sensitive than single quadrupole full-scan detection and would enable reduction in the amount of matrix introduced, thus reducing build up of contamination. Overall, when data for all 110 QC samples were included, recovery was acceptable for 91% of the diagnostic ions measured. On the basis of the same data, an overview by pesticide is presented in Table 7. For each pesticide two diagnostic ions from the full-scan data were integrated and concentrations were calculated. In routine practice, however, the most convenient way of reviewing the data is by using one and the same diagnostic ion for each pesticide, irrespective the matrix. On the basis of the data set obtained (nearly 14,700 pesticide–matrix combinations) the most favorable of the two diagnostic ions, i.e. the ion for which the highest number of recoveries within 60–140% was obtained, was assigned as the Quan ion (default quantification ion). By using this ion, acceptable recoveries were obtained for 93% of pesticides–matrix combinations. This also means that 7% or, in absolute figures, 1008 of the pesticide–matrix combinations did not meet the criterion. 40% of these failing combinations could be accepted after use of the alternative ion, for which calculations were also performed automatically during data processing. Low recoveries (<60%) for both diagnostic ions were obtained for 2.7% of pesticide–matrix combinations. High recoveries (>140%) were obtained for 2% of the combinations. For this latter group manual evaluation of other ions, if available and sufficiently abundant, could further increase the number of acceptable recoveries. Because this is a time-consuming process, it was not done routinely. In the event of deviating recovery, assessment of the results to be reported was based on visual evaluation of the extracted ion chromatograms of the two diagnostic ions at least. On the basis of on the findings it was then concluded the pesticide could not be determined in that specific matrix, or only at higher levels. Table 7Recoveries over all matrices (GC–MS analysis)PesticideQuan. ion m/zQual. ion m/zFortification level (mg kg−1)# QCs matrices (see Table 6)Both diagn. ions 60–140%One of diagn. ions 60–140%Both diagn. ions >140%Both diagn. ions <60%Average recov. (%) Quan. ionRSD (%)Acrinathrin 2082890.10110107107309716Azaconazole1732170.05110107107219714Azoxystrobin3883440.0510897102089615Benalaxyl2061480.051101081090110013Bifenthrin1811660.051091091100010213Biphenyl1541530.051109394799820Boscalid1121400.1310998100289616Bromopropylate3413430.051101001019010914Bromuconazole2951730.051101001054110218Bupirimate2732080.02110108109019615Buprofezin1721050.051091051082010212Cadusafos1581590.051101051071210413Chlorfenapyr3643280.041101031062210216Chlorfenvinphos3232670.051101031037010316Chlorpropham2131270.051081011062210514Chlorpyrifos3142860.051091071090110114Chlorpyrifos-methyl2882860.051081011044210216Chlorthal-dimethyl3323010.051101101100010114Cinerin-11231500.111101041054110115Cyfluthrin2261990.201101021060410017Cyhalothrin, lambda-2081810.05108104109109916Cypermethrin1631810.15105991072010214Cyproconazole2222240.051101031051410216Cyprodinil2242250.05109101102088515DDE, p,p′-2463180.061101101100010113DDT, o,p′-2352370.051101061072110314DDT, p,p′-2372350.0511082909119820Deltamethrin2532550.101109198489517Diazinon1791370.051091081100010113Dichlorvos1851090.051109096869920Dicloran2061600.0510896102359915Dieldrin263790.051101091090110414Diethofencarb1682670.051101071081110015Difenoconazole3232650.10107101106049616Dimethipin118760.05110951045110416Dimethomorph3873010.10110981000108916Dimoxystrobin2051160.051101081090110012Diniconazole2702680.15645862119717Diphenylamine1691670.051101071070310116Dodemorph2381540.05110109109019615Endosulfan-alpha195+241239+1970.501109510010010712Endosulfan-beta195+241237+1600.101101071073010214Endosulfan-sulfate272+229274+2370.051091021072110416EPN1573230.051101031063110317Epoxiconazole1921380.05110106108119814Esfenvalerate1671250.151101021034310615Ethion2311530.051101061064010314Ethoprophos1582000.051101071081110413Etofenprox3761640.05110102104249715Etridiazole2111830.0510980822179721Fenarimol2191390.051101061081110316Fenazaquin1601450.05110105105148816Fenbuconazole1291980.05110105107129917Fenitrothion2772600.05108991027110616Fenoxycarb1861160.05110891018110517Fenpiclonil2381740.051101011063110217Fenpropathrin1811410.051091011046010313Fenpropimorph1281290.051101081091010114Fenvalerate1671250.25110102103259815Fipronil3673690.05110101100379918Flucythrinate1991570.051101021063110315Fludioxonil2481820.05109105107129817Flusilazole2332060.05110104107129715Flutolanil3232810.051101071091010013Flutriafol2191230.041101021045110314Fluvalinate, tau-2502520.151109799569915Furalaxyl242950.051101061073010113Heptenophos1241260.05109971046010218Hexaconazole2162140.051101061081110214Iprodione3163140.10103798881310020Jasmolin-11641230.0411092104429715Kresoxim-methyl1162060.051091061090110015Lindane1832190.05110107110009915Malathion1731270.051081031073010417Mecarbam3291310.051101091100010115Mepanipyrim2232220.0511088917128519Mepronil2691190.10110109110009715Metalaxyl2061600.051071051082010312Methidathion145850.05109858919210715Metrafenone3953930.05110104106229414Mevinphos1921270.05110889017310417Myclobutanil1791500.05110102107219815Nitrothal-isopropyl2362540.05110108108119913Nuarimol2352030.051101081100010115Oxadixyl1631320.15110106107129913Parathion2911090.051101051091010515Parathion-methyl2632470.05109861028010717Penconazole1592480.051091081100010015Pentachloroaniline2672650.1111096970138115Pentachlorothioanisole2962460.0511087890217716Permethrin-cis1831630.051101081100010114Permethrin-trans1831630.051101061073010013Phenylphenol, 2-1701410.05109102107309813Phosalone1821840.05110909213510119Phosmet1611600.05109769016410022Phosphamidon2641270.05110919413310319Picoxystrobin3351450.051101051091010312Piperonyl-butoxide1761770.051071061091010013Pirimiphos-methyl2763050.051101091091010213Procymidone2832850.051081061081110014Profenofos3372060.05108931028010417Propargite1731350.331091041091010316Propiconazole2592610.05109106107219914Propyzamide1731750.051101071082010212Prothiofos3092670.05110108109109913Pyrazophos2212320.051109999389118Pyrethrins1231600.36110871037010518Pyridaben1471480.05110107107129914Pyridaphenthion3401990.05110961017210217Pyrifenox2622640.051101081100010015Pyrimethanil1991980.05110107106139014Pyriproxyfen2261360.051101041072110316Quinalphos1571460.051101041054110414Quinoxyfen3072720.05110106106049214Quintozene2371420.05110107107129316Silafluofen1792860.05110106106049814Spirodiclofen3123140.251109596689619Spiromesifen2722540.05110105108119616Spiroxamine1001980.10110107109019613TDE, p,p′-2352370.05110971005510314Tebuconazole2502520.15676667019715Tebufenpyrad1713180.051101071081110013Tebupirimfos2343180.051101081091010114Tefluthrin1771970.051101061073010313Tetraconazole3363380.05110109109109914Tetradifon3562290.15109109110009914Thiometon881250.051101081100010415Tolclofos-methyl2652670.051081071072010113Tri-allate2682700.051101041054110413Triazamate2422270.051101071073010214Triazophos2852570.05109951008210418Trifloxystrobin1311160.051101081091010314Triflumizole2782870.03110105107039915Trifluralin2643060.051101071072110114Vinclozolin2121980.051071061091010311Total 146961368814057402300% of # QCs93.195.22.72.0 It should be noted that the above evaluation applies to a level five times the reporting level, which was set at 0.01 mg kg−1, or the LOQ if higher than 0.01 mg kg−1. At lower levels interferences may have a larger effect and, consequently, more frequent deviations from the 60–140% criterion (most probably >140%) may be observed. For higher levels, the opposite would be true. Pesticides for which low recoveries (<60%) were frequently obtained (10–21 of 110 QC samples) included iprodione and p,p′-DDT (degradation in inlet), dimethomorph (polar, relatively non-volatile, could be troublesome in splitless transfer), pentachloroanisole, pentachloroaniline, and mepanipyrim (no clear explanation, but probably related to the dispersive SPE clean-up). There were no indications for poor extraction efficiency. High recovery (>140%) frequently occurred for etridiazole, methidathion, mevinphos, phosmet, phosalone, phosphamidone, and endosulfan-alpha (10–21 times out of 110 QC samples, often in herbs and peas). This was attributed to matrix effects and interferences. Overall, the pesticides that failed most frequently (11–28 times out of 110) during routine analytical quality control were (in descending order) etridiazole, iprodione, methidathion, pentachlorothioanisole, mevinphos, phosmet, p,p′-DDT, mepanipyrim, phosalone, phosphamidon, biphenyl, dichlorvos, spirodiclofen, pentachloroaniline, deltamethrin, tau-fluvalinate, and pyrazophos. These would be the most relevant for inclusion in alternative methods, for example GC–MS–MS or LC–MS–MS. Average recovery and RSD were calculated for pesticide–matrix combinations that passed the acceptable recovery criterion. The results are included in Table 7. Average recovery was usually close to 100% and RSDs approximately 15%. For the pesticides known to be adsorbed by GCB systematically lower average recovery (77–90%) was obtained, which is in agreement with the results obtained during method development. These comprehensive data show that with a relatively inexpensive single-quadrupole MS detector in full-scan mode it is possible to obtain reliable quantitative data down to the 0.01 mg kg−1 level, or even lower, for a wide range of pesticides in a wide variety of matrices after generic rapid sample preparation based on extraction with ethyl acetate. Unified calibration based on a tomato-matrix standard is, furthermore, a feasible approach. One should, however, be aware there are also limitations and that some pesticide–matrix combinations cannot be determined in the 0.01–0.1 mg kg−1 range, and that for other pesticides calibration against the corresponding matrix instead of tomato is required to bring quantitative results within the AQC criteria, especially for MRL violations, when more stringent criteria apply. The data also reveal that the only way to gain full insight into analyte recovery and method selectivity with a wide variety of matrices is by performing analytical quality control on all pesticides which are reported, rather than on a subset, as is suggested in the EU guideline [37]. A subset will suffice for demonstration of adequate sample preparation and injection but will not reveal limitations in the selectivity of GC–MS. GC single-quadrupole MS remains an effective tool for routine GC analysis of pesticide residues. For many vegetable and fruit matrices there is no real need to change to more advanced (and expensive) MS techniques, for example MS–MS (which has limited scope) or accurate mass TOF-MS (which has a limited dynamic range). Use of such equipment would be justified for more complex matrices and when low μg kg−1 LOQs are required—for example analysis of some pesticides in baby food. LC–MS–MS analysis Clean-up The ethyl acetate extraction procedure is also appropriate for many pesticides not amenable to GC analysis [11, 15, 16, 18, 26]. Typically no clean-up is performed (Table 1). One reason for this is that with regard to chromatographic performance LC columns tend to be more tolerant of injection of bulk matrix than GC columns. In our experience, continual injection of 20 mg equivalent of vegetable and fruit extracts does not result in deterioration of chromatographic performance or unacceptable contamination of the ion source (the system used here was an API2000). In LC–MS co-extracted matrix does have an effect on the response, however, by interfering with the ionization process. This results in suppression (sometimes enhancement) of the response to a pesticide in a matrix compared with that in a solvent standard [51] and complicates quantification of pesticides in the samples. The possibility of reducing matrix effects by use of dispersive SPE clean-up was investigated in a similar way as for GC. First, the effectiveness of the clean-up step was investigated by addition of 25 mg GCB and 25 mg PSA to 1 mL raw extract of a mixed spinach–grape–onion sample (1:1:1, 1 g ml-1). Seventy pesticides (the ones in Table 8 with API2000 in the MS-MS column) were added after clean-up and analyzed by LC–MS–MS. The response was compared with that of solutions of equal concentration in the raw extract and a solvent standard. Clean-up increased the number of pesticides for which no pronounced matrix effect (less than 20% suppression or enhancement) was observed from 38 to 84%. Several of the pesticides (Tables 2 and 4) were adsorbed by the SPE material, however. Although adsorption by the GCB could have been avoided or reduced by addition of toluene (although less practical when changing from extraction solvent to methanol/water), it was concluded that PSA was not compatible with a generic method for pesticides amenable to LC–MS–MS. It was therefore decided not to include a clean-up step for LC–MS–MS analysis and to use the initial raw ethyl acetate extract. Another reason for not further pursuing clean-up in LC–MS–MS analysis was that the sensitivity of current triple-quadrupole instruments enables injection of only small amounts of matrix into the LC–MS–MS system (e.g. 2 mg) while still achieving the desired limits of quantification. Experiments showed that tenfold dilution of 1 g mL−1 extracts increased the number of pesticides for which no pronounced matrix effect occurred from 65 to 82% and from 10 to 65% for cucumber and cabbage, respectively. Table 8LC–MS–MS settings and performance-validation characteristicsPesticidetr (min)Precurs.Prod. ion 1DPFPCECXPProd. ion 2CECXPVegetablesn0.01 mg kg−10.1 mg kg−1Fruitsn0.01 mg kg−10.1 mg kg−1MS–MSMatrixRec. (%)RSD (%)Rec. (%)RSD (%)MatrixRec. (%)RSD (%)Rec. (%)RSD (%)Abamectinea,c21.78913054634033221454910Cuc/lett466186815Apple/grape41593915546API2000Acephate5.518414331150111295336Cuc/lett48021759Apple/grape48087610API2000Acetamiprid10.52231269127029101771114Cuc/lett4993967Apple/grape41174986API2000Aldicarba11.72081161611011889216Cuc/lett4103209112Apple/grape4991310913API2000Aldicarbsulfon7.922386322002112148136Cuc/lett41049834Apple/grape41205913API2000Aldicarbsulfoxide7.22071324630091089196Cuc/lett410912894Apple/grape41094863API2000Asulam3.623115641260151292336Cuc/lett435282938Apple/grape413231030API2000Azamethiphos12.1325183362202314112538Cuc/lett41016944Apple/grape4106119110API2000Azinfos-methyl13.53181324160236160156Lettuce59316884Orange56911799API3000Bendiocarb12.22241671610013101092518Lettuce510210966Orange58681086API3000Bifenazate13.93011981611013161702714Lettuce5359337Orange5937835API3000Bitertanol15.233826921120132099218Lettuce59610817Orange59310818API3000Butocarboximb11.621375413002141561712Lettuce5101239312Orange572168919API3000Butoxycarboxim7.7223106362501381661110Cuc/lett4116910415Apple/grape41184957API2000Carbaryl12.520214510137013121273710Cuc/lett41004954Apple/grape41111210010API2000Carbendazim11.41921604623023121324310Cuc/lett4104210210Apple/grape212211051API2000Carbofuran13.32221654629017121232910Cuc/lett41241211113Apple/grape410411934API2000Carbofuran, 3-OH10.4238220312109161631912Lettuce5918944Orange51007916API3000Carboxin12.62361431135021293512Lettuce5879802Orange5899846API3000Chlorbromuron14.0295206413502712182254Lettuce510019865Orange58330846API3000Chlorfluazuron18.15423854027029301582912Lettuce5797895Orange57421868API3000Clofentezin15.5303138512802110102618Cuc/lett493177610Apple/grape41272410116API2000Clomazone13.62401253119025889676Lettuce59741046Orange5905899API3000Clothianidin10.2250132367023101691710Lettuce599111002Orange511041003API3000Cycloxydim14.93262804626019221802914Cuc/lett418118829Apple/grape438457028API2000Cymoxanil11.1199128181201310111258Cuc/lett48313957Apple/grape4908992API2000Cyromazine7.116785402402661252510Cuc/lett496107811Apple/grape4967813API2000Demeton13.625989261801361981116Lettuce59714854Orange57615769API3000Demeton-S-methyl12.523189315021461374Lettuce5935864Orange5816818API3000Dem-S-meth-sulfone8.826316941350236109414Lettuce510412926Orange51002974API3000Desmedipham13.13011825134013141542512Cuc/lett48610883Apple/grape495228315API2000Diafenthiuron18.13853294126027222784518Lettuce50–0–Orange51049927API3000Dichlofluanidec14.1333224462701718123378Cuc/lett42111636116Apple/grape433825468API2000Dicrotophos9.5238112412701781931316Cuc/lett41105993Apple/grape4100129310API2000Diflubenzuron14.53111584627019121414710Cuc/lett47915841Apple/grape4101610212API2000Dimethirimol13.1210715129045498378Lettuce5997975Orange591101055API3000Dimethoate10.623019911350134125292Lettuce5987964Orange510917956API3000Diniconazole15.6326705631063141594516Lettuce57810936Orange59416965API3000Disulfotonc15.7275891190276614110Lettuce5536647Orange58516864API3000Disulfoton-sulfone12.830797311503981531714Lettuce5113101057Orange58161068API3000Disulfoton-sulfoxide12.82911852614017162131514Lettuce5111101156Orange59251015API3000Diuron13.3233723621037446356Lettuce511161017Orange5947946API3000DMSA11.620192261502561371310Lettuce510213974Orange58513877API3000DMST12.3215106261602181511310Lettuce5975956Orange58413855API3000Ethiofencarb12.822610736220218169914Cuc/lett48130945Apple/grape499179420API2000Ethiofencarbsulfon9.7258107362402162011116Cuc/lett4120101055Apple/grape410189711API2000Ethiofencarbsulfoxide9.9242107311802381851314Cuc/lett411413972Apple/grape4127101077API2000Ethirimol13.321014051370311298376Cuc/lett4963886Apple/grape486268126API2000Famoxadonea14.63923311113015222382518Lettuce59015801Orange5889801API3000Fenamiphos14.530421741350294234214Lettuce5878874Orange5937935API3000Fenamiphos-sulfone12.23363088136023222662920Lettuce51028945Orange58116868API3000Fenamiphos-sulfoxide12.13201715623027142333514Lettuce511410944Orange59781085API3000Fenhexamid14.2302975129035855598Lettuce58415824Orange5856845API3000Fenpyroximate19.34223666136021261354310Cuc/lett4988959Apple/grape4111910410API2000Fensulfothione13.03092814626021222532518Lettuce5967893Orange510123838API3000Fensulfothion-sulfone13.03252693612021181913312Lettuce510310988Orange58561006API3000Fenthion13.9279231261302116Lettuce511131818Orange53822748API3000Fenthion-sulfone12.5311125513202982792522Lettuce5956904Orange51011896API3000Fenthion-sulfoxide12.4295280462302520109458Lettuce5932946Orange5948876API3000Fipronil14.14373686637023262903716Lettuce570248811Orange592289012API3000Flucycloxuron17.34842896636015201324910Cuc/lett411341043Apple/grape41633812126API2000Flufenoxuron17.148915810136027121416510Cuc/lett410717908Apple/grape4172501028API2000Formetanate 12.2222165361901914120378Lettuce5100141036Orange5956957API3000Fosthiazate12.7284104312002362281522Lettuce59981026Orange5842986API3000Furathiocarb16.53831957637025162521918Cuc/lett455325538Apple/grape48717847API2000Hexaflumuronc15.24611585130027101416110Cuc/lett49124827Apple/grape41711511416API2000Hexythiazox17.43531684127035122282118Cuc/lett499198415Apple/grape4120268411API2000Hymexazolc5.8100546636021444292Cuc/lett476345049Apple/grape445152220API2000Imazalil15.02971594629033122012916Cuc/lett49047612Apple/grape411179013API2000Imidacloprid10.02561754124025142092118Cuc/lett49998111Apple/grape412112897API2000Indoxacarb15.15282494124023181503510Lettuce56032735Orange5846786API3000Iprovalicarb14.13211193116029102031318Lettuce510851047Orange59749010API3000Isoxaflutole12.93602514627019222205522Lettuce576189015Orange58618985API3000Linuron13.82491604629025121822114Cuc/lett4103168611Apple/grape490261016API2000Metamitron10.7203175512902314104316Cuc/lett480118717Apple/grape49717959API2000Methabenzthiazuron13.32221653120021121504512Lettuce51064986Orange58481079API3000Methamidofos4.61429441240216125198Cuc/lett483167919Apple/grape48611815API2000Methiocarb13.82261694630013141212510Cuc/lett4949954Apple/grape41015941API2000Methiocarbsulfon10.7258122563702582011316Cuc/lett4109129911Apple/grape4949876API2000Methiocarbsulfoxide10.12421854629019141703114Cuc/lett411651013Apple/grape412681042API2000Methomyl8.81638821130136106138Cuc/lett41532213619Apple/grape4125141037API2000Methoxyfenozide13.83693132420013241333410Lettuce5937914Orange59113913API3000Metobromuron13.12591704628025121482112Cuc/lett411219996Apple/grape49699912API2000Metoxuron11.6229723119037446352Lettuce510481004Orange59581024API3000Monocrotofos9.22241274124021101931116Cuc/lett41085904Apple/grape411189810API2000Monolinuron12.8215126412602381481912Cuc/lett41047986Apple/grape411171078API2000Omethoate6.52141253623029101831514Cuc/lett498138513Apple/grape41025862API2000Oxamyla8.0237722116023490116Cuc/lett410731907Apple/grape412814979API2000Oxamyl-oxim6.6163723623017490256Cuc/lett41006853Apple/grape411831019API2000Oxycarboxin10.92681752617019141473510Lettuce5986964Orange58522785API3000Oxydemeton-methyl8.5247169412301914109358Cuc/lett49811897Apple/grape41045964API2000Paclobutrazole13.829470363204541255110Lettuce5969878Orange57767696API3000Pencycuron15.43291255634035102182318Cuc/lett41005773Apple/grape41184929API2000Phenmedipham13.23011685129013141362910Cuc/lett4997965Apple/grape4108118411API2000Phenm.-metabolite10.0168136312001410108268Cuc/lett410791035Apple/grape4101149617API2000Phorate15.5261752615021447458Lettuce596279111Orange51042886API3000Phorate-sulfone12.92931712615017101153710Lettuce511410959Orange58361044API3000Phorate-sulfoxide12.82771994127017697454Lettuce5998963Orange5986914API3000Phosphamidon11.73001744125019141273310Lettuce510151075Orange59871007API3000Picolinafen16.43772385622041142562920Lettuce5818966Orange51038995API3000Pirimicarb13.023972263603141822312Lettuce5996964Orange5899923API3000Pirimicarb, desmethyl 11.62257221360334168216Lettuce51034983Orange531144215API3000Prochloraz15.4376308463101322704116Cuc/lett490157813Apple/grape484389464API2000Profoxydim16.24662806614027201803512Lettuce53325306Orange54934555API3000Propamocarb8.5189102311902561441912Lettuce5754724Orange52214188API3000Propoxur12.2210111312101981681114Cuc/lett411431005Apple/grape41183985API2000Prothiocarb7.4191146462402112Cuc/lett485266337Apple/grape410658310API2000Pymetrozine9.021810556370278201916Cuc/lett46526858Apple/grape4477717API2000Pyraclostrobin15.13881941350196163336Lettuce57213776Orange5874837API3000Pyridate metabolite10.42077756340456104318Cuc/lett410012874Apple/grape4899755API2000Rotenone14.739521310137031161923314Cuc/lett49313938Apple/grape494169430API2000Sethoxydim15.23281784626025142201918Cuc/lett46739883Apple/grape459349628API2000Spinosyn A22.073314296280431298836Lettuce5959936Orange5974922API3000Spinosyn D24.174714296110471298894Lettuce5863936Orange5997925API3000Tebuconazole14.83087061140516125538Lettuce5806933Orange5958964API3000Tebufenozide14.53531332618023102971322Cuc/lett4103168611Apple/grape4106427833API2000Temephos16.34671257132039104193532Lettuce56227816Orange5927959API3000Tepraloxydim12.73422503118019281662912Lettuce54419607Orange57315624API3000Terbufos16.728910311120131057378Lettuce57327758Orange580248112API3000Terbufos-sulfone13.5321171211301912115396Lettuce5108410111Orange59969310API3000Terbufos-sulfoxide13.53051876110171097598Lettuce510631035Orange5985979API3000Thiabendazole12.22021755637035121314510Cuc/lett487121013Apple/grape4982927API2000Thiacloprid11.025312641210278905316Lettuce59791023Orange510261167API3000Thiametoxam9.02922114627019241323310Lettuce5944974Orange51019996API3000Thiocyclamd12.6182137211602112732914Lettuce59611896Orange5100158211API3000Thiodicarb12.73558820130316108218Cuc/lett4371154298Apple/grape4834794API2000Thiofanox12.921957119019661154Lettuce5nd819321Orange5nd–8430API3000Thiofanox-sulfone10.2251571635026276214Lettuce5110161015Orange58525858API3000Thiofanox-sulfoxide9.82351043132017457272Lettuce511021053Orange510911886API3000Thiometonc13.0247891611023661458Lettuce596171009Orange587111002API3000Thiophanate-methyl12.13431513021025123111723Cuc/lett46687516Apple/grape441593798API2000Tolylfluanidea14.73642383121019181374110Cuc/lett43111642115Apple/grape475932481API2000Triadimefon14.02941973118023122251918Lettuce59210866Orange5897787API3000Triadimenol14.1296701613031499218Lettuce51017876Orange5897829API3000Triazoxide13.5248685632047495376Lettuce5991027619Orange5431076910API3000Trichlorfon10.6257109462602782211518Cuc/lett41161610422Apple/grape41148994API2000Tricyclazole11.51911365636039101633112Cuc/lett41055926Apple/grape49611833API2000Triflumuron14.93591563020023121394710Cuc/lett4949927Apple/grape4118121098API2000Triforine13.24353901210013302154015Cuc/lett498131016Apple/grape49710939API2000Vamidothion10.4288146463001912118318Cuc/lett411116963Apple/grape4119111047API2000Cuc, cucumberLett, lettuceaNH4 adductbNa adductcLOQ level 0.05 mg kg−1dLOQ level 0.02 mg kg−1 Routine experience with LC–MS–MS analysis for over four years, both with the API2000 (20 mg matrix) and the API3000 (2 mg matrix) has shown that injection of uncleaned extracts does not result in special maintenance requirements. The source is cleaned with a tissue daily. The LC column typically lasts for 6 months. Changing the solvent Because ethyl acetate is less suitable for direct injection in reversed phase LC, the solvent was changed. Because only small amounts of the raw extract need to be evaporated (less than 0.5 mL in the final method) and evaporation blocks enable simultaneous evaporation of many (typically 24–36) extracts, this step adds little to the overall sample-preparation time. Changing the solvent was even regarded as advantageous. It resulted in more freedom in selection of the final solvent to be injected into the LC, which can be critical for very polar compounds (e.g. in acetonitrile-based extraction methods, injection of 100% acetonitrile easily leads to band-broadening for methamidophos). It is also easier to compensate for the smaller amount of sample processed for dry crops (because of the need for addition of water) by evaporating a larger amount of the ethyl acetate extract. In previous work [15] a small amount of a diethylene glycol (added as solution in methanol) was added, because this was found to facilitate reconstitution, thereby improving recovery for some pesticide–matrix combinations. It was also shown that the evaporation step did not require special attention and that continuing the process for another half hour after completion of evaporation of the solvent did not affect recovery. The same procedure was therefore used here without re-evaluating the real need for it. Reconstitution was performed by first dissolving in methanol (ultrasonication) and then dilution with LC mobile phase component A. Validation of LC–MS–MS method The LC–MS–MS method was validated in three separate studies, one using the API2000 with injection of 20 mg matrix equivalent and the other two using the API3000 with injection of 2 mg matrix equivalent. A total of 140 pesticides and degradation products were included. In contrast with the full-scan acquisition in GC–MS, in LC–MS–MS data were acquired for a fixed, limited, set of pesticides. Although many pesticides from the GC–MS method can also be analyzed by LC–MS–MS, emphasis was on pesticides that were not, or less, amenable to GC analysis. Recovery, based on matrix-matched calibration, and repeatability were evaluated at the 0.01 and 0.1 mg kg−1 level for vegetable and fruit matrices; the results are listed in Table 8. Although acceptable performance data were obtained for most of the pesticides, low recovery and/or high variability were observed for some. Among these were compounds that were also reported as troublesome by other workers using alternative multi-residue methods, e.g. asulam [30]. Low recovery could be partly attributed to poor extraction efficiency (asulam, hymexazol, and, in orange, propamocarb) or degradation during sample preparation (cycloxydim, sethoxydim, profoxydim, tepraloxydim, dichlofluanide, tolylfluanide, thiodicarb, thiophanate-methyl, and, in lettuce, disulfoton and furathiocarb). The degradation seems to be related to the change of solvent, as is apparent from comparison of GC–MS and LC–MS–MS validation data for dichlofluanide, tolylfluanide, and disulfoton. Fortunately, for many of these the degradation products formed are also part of the residue definition and are included in the method. Indeed, elevated recovery was observed for the degradation products when determined in the same validation set as the parent compound. In the analysis, therefore, degradation is not necessarily a problem, because the results (expressed as defined in the residue definition) have to be summed. In routine analytical quality control (see below) the data were evaluated this way. Analytical quality-control data from routine LC–MS–MS analysis In the same way as for GC–MS analysis, the initial validation data are continually being supplemented by performance data generated as part of analytical quality control during routine analysis of samples. With each set of analytical samples at least one was fortified with the full quantitative suite (i.e. 136 pesticides and degradation products) at the 0.05 mg kg−1 level. A compilation was made from all the data generated over a period of 12 months, which included data for more than one hundred vegetable and fruit matrices. A limited number of dry matrices (flour, milk powder) were also included in the set. The data were evaluated for one transition for each pesticide, using the API3000 and injection of 2 mg equivalent of matrix (10 μL of a 0.2 g mL−1 extract). Examples of typical extracted ion chromatograms are shown in Fig. 5. Fig. 5Typical extracted ion chromatograms obtained by LC–MS–MS analysis of vegetable and fruit extracts (calibration standard in mango matrix, 10 pg μL, corresponding to 0.05 mg kg−1) For all fortified samples the matrix effect was also established by analyzing the corresponding matrix-matched standard, at the same level as in the extract of the fortified sample, against a solvent standard. Suppression (or enhancement) of up to 20% was regarded as acceptable for quantification. The number of compounds for which the response in matrix relative to that in solvent was between 80 and 120% is given in Table 9 for each matrix. Whereas for beetroot, asparagus, and kangkung little or no matrix effects exceeding 20% were observed, such effects were much more common for herbs and citrus fruits. Table 9Overview of matrix effects and recoverya within or outside the EU 60–140% criterion [37] after LC–MS–MS analysis NMatrix effectsnRecovery# Pesticides# PesticidesRel. resp. 80–120%>20% suppr.>20% enhanc.Calc. using solvent stdCalc. using matrix-matched std60–140%<60%>140%60–140%<60%>140%Corn syrup (2/2)1351340110497439941Beetroot135133111041013010130Corn syrup (1/2)13513221104984210022Kangkung13513221104911129482Green pea1351313110497529932Asparagus1351304110497709860Coco nut135130411046341059450Papaya1351303210496449842Cauliflower135129151041012110211Fennel135129421041003110130Cherry (2/3)135128701041004010040Cherry (1/3)1351277110492849824Ladies’ fingers1351278010497709770Mango (1/2)1351276210497349824Cherry (3/3)135126811041004010220Mango juice135126361041011210400Mushroom135126721041022010301Taro1351267210496449914Plum (3/3)13512582104957210040Fennel leaves (2/2)1351245610499239923Milk powder135124651045845159450Grape1351239310498339833Spinach13512312010494829653Tamarind135123841046737079250Cassava135122761048716178260Raspberry (1/3)1351221211048420092120Sweet pepper1341221021031001210012Apple puree1351215910499509770Corn flour13512111310495639572Courgette135121771041002210031Tomato puree1351211041041013010310Raspberry (2/3)135120150104985110031Broccoli135119142104901049383Flour (2/2)13511921410495279734Peach (1/2)135119160104995010040Mango (2/2)134117125103966110030Milk/flour mix1351171261044360155490Bitter cucumber13511617210499239914Melon puree135116181104995010310Tomato13511613610493839653Lettuce, crinkley 13411419110397339724Pear13411414610397609940Flour (1/2)1351131481047328385190Plum (1/3)13511313910493659824Celery leaves (1/3)135112221104901229734Purselane13511223010496629842Apricots135111231104901319761Artichoke135111177104911219581Cucumber1351101510104995010130Horseradish powder1351101510104881159752Tarrragon (2/2)13511081710496449464Avocado (1/2)1351092241048121290131Haricot bean1351092511048320190131Kiwi1351091016104976110022Peach (12/2)135108243104881429392Raspberry (3/3)1351072621048022290113Blackberry1331061710102911019192Diced pumpkins135106272104958110031Plum (2/3)1351062361048618085181Yam13510612810497619680Avocado (2/2)1341032921036834180221Dill leaves135103151710494739392Honey106103308282008200Chervil13510229410495909851Parsley13510229410495459914Nectarine13410129410392839841Bean sprouts106100518276607840Sweetcorn (1/2)10699618276517732Beetroot leaves13598325104851909950Chestnuts10698178276427930Pomegranate (1/2)135973711048420010040Pomegranate (2/2)135973711048420010040Pear syrup10695388279308020Alfalfa106941118275707840Fennel leaves (1/2)10692868274537624Chili pepper13591404104958110112Turnip tops106901518276247804Blueberry135894331026636091110Litchi13588452104782609941Salak13588425104822029941Pepper powder10687163825427170111Celery leaves (2/3)135854191049310110022Lemon13484473104782069734Physalis13583484104713309950Maize (feed)1358153110495639338Sweetcorn (2/2)13580505104792239860Coriander (1/2)13579560104683429563Mangostan1357640191044654469350Celery leaves (3/3)13475581103861619922Laos13573575104703319941Chives13571577104985110211Coriander (2/2)135656010104832109860Tea (black)136656921046043187143Lemon puree135538021046836010310Ginger13546863104683429833Grapefruit (1/2)13346870102435909813Grapefruit (2/2)13546881103614119733Oregano1354675141045250287161Kumquat13538952104475619464Lime13438942103485239643Tarrragon (1/2)135389521044163090131Italian herb mix135331011104544919581Total QC results1349710488256644310395861816131649533708154Percentage of total results78193831629271aRecovery at 0.05 mg kg−1 (higher for seven pesticides). The pesticides included are listed in Table 10N is the total number of individual compounds (pesticides and metabolites) added to the matrixn is the total number of pesticides added to the matrix. Compounds belonging to the same residue definition counted as one In contrast with GC, for which matrix effects are mainly caused by shielding of active sites in the inlet and were, to some extent predictable (in relation to the matrix load injected and the lability and/or polarity of analyte), in LC–MS–MS matrix effects are much less predictable. Although they do depend on the amount of matrix introduced into the system, and also tend to be more abundant in complex (“aromatic”) matrices, it cannot be readily predicted for which pesticides the effects occur. For this reason use of one matrix-matched standard as representative calibrant for a whole range of commodities, which worked reasonably well in GC–MS analysis, was not feasible in LC–MS–MS analysis. Consequently, critical evaluation of the matrix effect was required; if unacceptable suppression occurred there was no alternative to quantification by use of the appropriate matrix-matched calibration standard or, when not available, by standard addition. Recovery of the pesticides from the fortified samples was calculated relative to that from a solvent standard and a matrix-matched standard and tested against the 60–140% criterion for evaluation of routine analytical quality-control samples [37]. A total of more than 10,000 recovery values were evaluated. Without matrix-matched calibration, acceptable recovery was obtained for 83% of the pesticides. Deviating recoveries were usually too low, mainly because of ion suppression, as is apparent from the results obtained from determination of recovery using matrix-matched calibration, for which 92% met the criterion. Concentrating on performance at the pesticide level (Table 10) enables easy identification of troublesome pesticides. All compounds belonging to the same residue definition were summed (according to the residue definition) and counted as one, thereby compensating for possible conversion during sample pretreatment. This way the low recovery of dichlofluanide and the corresponding high recovery of DMSA were acceptable for most matrices because recovery for the sum met the criterion. Pesticides for which multi-matrix analysis under fixed conditions was less favorable included asulam, bifenazate, cyromazine, furathiocarb, propamocarb, pymetrozine, and thiocyclam (low recovery because of varying extraction efficiency and/or degradation). As already observed during validation, the method was also less suitable for cycloxydim, profoxydim, sethoxydim, and tepraloxydim. For these compounds recovery was too high, possibly because of degradation in the calibration standard used for preparation of the matrix-matched standards. Table 10Recovery over all matrices (LC–MS–MS) # ACQ samples# Recov. 60–140%# Recov. <60%# Recov. >140%Average recov. (%)aRSD (%)a1Abamectin1021002086172Acephate102939078133Acetamiprid10297509011Aldicarb102101019113Aldicarb-sulfone102102009212Aldicarb-sulfoxide102966084134Asulam1026932185175Azamethiphos1021020089126Azinfos-methyl102965187157Bendiocarb93930088128Bifenazate986037185189Bitertanol10298408415Butocarboxim102101108814Butoxycarboxim10210110911210Carbaryl102100118713Carbendazim10097219314Carbofuran102100119212Carbofuran,3-hydroxy-10210200931111Carboxin1029750841312Chlorbromuron1029840861413Chlorfluazuron1029381871514Clofentezine10289130801515Clomazone938931851216Clothianidin939120911217Cycloxydim1026811231041918Cymoxanil10210200911519Cyromazine10249530741220Demeton102102008914Demeton-S-methyl102100208714Demeton-S-methylsulfone10210110911221Desmedipham10296608314Dichlofluanid10236660801922Dicrotophos10210020891423Diflubenzuron1029840821524Dimethirimol9390308911Dimethoate10210110901225Diniconazole9384818616Disulfoton93672517513Disulfoton-sulfone9393008812Disulfoton-sulfoxide938904961626Diuron9392108714DMSA1024106110917DMST102961510416Ethiofencarb10299308614Ethiofencarb-sulfone102102009013Ethiofencarb-sulfoxide10210110921527Ethirimol1029840881228Famoxadone10295708314Fenamiphos102100208914Fenamiphos-sulfone102102009112Fenamiphos-sulfoxide939210901129Fenhexamid1029660851230Fenpyroximate102921008713Fensulfothion102102008811Fensulfothion-sulfone9391208512Fenthion10299308714Fenthion-sulfone10299218815Fenthion-sulfoxide10210200931431Flucycloxuron1029480881532Flufenoxuron1029390871433Fosthiazate939300901234Furathiocarb10279203841635Hexaflumuron10290102851836Hexythiazox10291110851537Imazalil1019290831438Imidacloprid1029930901439Indoxacarb1019650861640Iprovalicarb939210871341Isoxaflutole9383100821442Linuron1029741851243Metamitron1029750881544Methabenzthiazuron939300881345Methamidophos102901207512Methiocarb102100208513Methiocarb-sulfone102841807815Methiocarb-sulfoxide1029921881246Methomyl102890131011447Methoxyfenozide10210110851448Metobromuron1029741871249Metoxuron939300891250Monocrotophos10210110901251Monolinuron102101108614Omethoate10299308312Oxamyl102100208912Oxamyl-oxime10210110881252Oxycarboxin102102009112Oxydemeton-methyl1029750861353Paclobutrazole10210110871254Pencycuron10296608114Phenmedipham10294718314Phenmedipham-metabolite102100209315Phorate102683407419Phorate-sulfone9393008812Phorate-sulfoxide10210110901255Phosphamidon939300891056Picolinafen9386618415Pirimicarb102101018912Pirimicarb, desmethyl-10210011901257Prochloraz1019470831458Profoxydim99543213992159Propamocarb1019920701560Propoxur10210020881661Pymetrozine10273290892062Pyraclostrobin1029570851463Pyridate-metabolite1029291861564Rotenone1029390811565Sethoxydim102723271061966Spinosyn-A9388508217Spinosyn-D9382110831567Tebuconazole939030861668Tebufenozide1029930861469Temephos1029480871670Tepraloxydim1026204011414Terbufos93623017715Terbufos-sulfone9390308613Terbufos-sulfoxide939210881271Thiabendazole989251861372Thiacloprid939030881273Thiametoxam939120891374Thiocyclam93642907816Thiodicarb102624008216Thiofanox10298318514Thiofanox-sulfone102102009013Thiofanox-sulfoxide10210110921475Thiometon9388418716Thiophanate-methyl102831907712Tolylfluanid101366507622Triadimefon10299308513Triadimenol1029831871276Triazoxide1029093841677Trichlorfon10210101871278Tricyclazole1029660871279Triflumuron10189102841880Triforine1029732871581Vamidothion10210110891182Sum aldicarb10210110881183Sum butocarboxim10210110901184Sum carbendazim1019740831285Sum carbofuran10210200921086Sum dimethoate10210020861087Sum dichlofluanid102891121071788Sum disulfoton938940861389Sum ethiofencarb10210200891190Sum fenamiphos10210110901191Sum fensulfothion10210200861192Sum fenthion10210200891293Sum methiocarb10210020831294Sum methomyl10210020871295Sum oxamyl10210110881096Sum oxydemeton-methyl10210110881197Sum phenmedipham10210110881398Sum phorate1029750811299Sum pirimicarb102101109012100Sum terbufos9388508113101Sum thiofanox102102008911102Sum tolylfluanid10195608015103Sum triadimefon10299308613aAverage and RSD for recoveries within 60–140% rangeMatrix-matched calibration, API3000Level = 0.05 mg kg−1 for most pesticides/metabolitesBold indicates pesticides, including metabolites that are part of residue definition, if appropriate Averaging acceptable recoveries reveals there is some bias, because the values are mostly approximately 87% (in contrast with the GC–MS data, for which the average was approximately 100%). It was noted that for dry crops relatively low recovery (typically between 60–70%) was obtained for all pesticides. The cause is not clear. This bias can also be seen in tables in other papers (barley [26], soya grain [33]). Independent evaluation of method performance by proficiency testing From results obtained over the years from participation in proficiency tests, an additional and independent verification of method performance could be made. The data are summarized in Table 11 and clearly show that good quantitative data were consistently obtained from both GC–MS and LC–MS–MS, with method performance good (Z-score<2) 54 times, doubtful (2 < Z < 3) three times, and never poor. It also shows that the calibration approach (one-point calibration, tomato-matrix standard for GC and matrix-matched standard for LC) is fit-for-purpose. Table 11Results from the analysis of Fapas (series 19) proficiency test samples (2003–2005)SamplePesticideMRMSpike level added (μg kg−1)Inter-lab. result (μg kg−1)TNO result (μg kg−1)Z-score TNO#53 AppleFenpropathrinGC–MS5004055281.7Parathion-methylGC–MS705947−0.9TetradifonGC–MS14011591−0.9TriazofosGC–MS14011974−1.7VinchlozolinGC–MS6053530.0#52 CucumberIprodioneGC–MS1009489−0.3MethomylLC–MS–MS2825280.5ThiabendazoleLC–MS–MS50128113−0.5#51 PearCarbendazimLC–MS–MS15011660−2.2Dodinenot in MRM6059**ImazalilLC–MS–MS4002372730.8#49 MelonChlorprophamGC–MS109111.0ChlorpyrifosGC–MS887−0.7DimethoateLC–MS–MS151915−0.9PirimicarbLC–MS–MS201916−0.7#48 TomatoAzoxystrobinGC–MSNot given201166−0.9BifenthrinGC–MSNot given83990.9BuprofezinGC–MSNot given1081311Chlorpyrifos-methylGC–MSNot given319281−0.6ProcymidoneGC–MSNot given712668−0.4#47 GrapefruitDiazinonGC–MSNot given2622940.6HeptenophosGC–MSNot given1682341.9MalathionGC–MSNot given715690−0.2MethidathionGC–MSNot given567540−0.3#46 LettuceBromopropylateGC–MS806751−1.1DimethoateLC–MS–MS3002853160.6OxadixylGC–MS1201271340.3PenconazoleGC–MS1008251−1.7Tolclofos-methylGC–MS16013775−2.1#42 AppleChlorfenvinphosGC–MS907150−1.3ChlorpyrifosGC–MS400259241−0.3MethamidophosLC–MS–MS604431−1.3MonocrotophosLC–MS–MS805856−0.1OmethoateLC–MS–MS150108103−0.2TrifluralinGC–MS10059620.2#41 BasilKresoxim-methylGC–MS1509486−0.4ProcymidoneGC–MS1208778−0.5PropyzamideGC–MS1008159−1.2VinclozolinGC–MS604744−0.3#38 TomatoAzoxystrobinGC–MS150137132−0.2BupirimateGC–MS1008362−1.1Chlorpyrifos-methylGC–MS807253−1.2QuinalphosGC–MS140124105−0.7#37 LemonDiazinonGC–MS8042420.0FenitrothionGC–MS10078800.1MetalaxylGC–MS12094930MethidathionGC–MS1501091541.9#35 LettuceCarbendazimLC–MS–MS805331−1.9lambda CyhalothrinGC–MS806654−0.8MetalaxylGC–MS1209486−0.4#34 AppleDiphenylamineGC–MS503929−1.2Pirimiphos-methylGC–MS5041420.1PropargiteGC–MS2001621720.3TetradifonGC–MS1008338−2.5#29 Sweet pepperDichloranGC–MS2001792000.6MecarbamGC–MS100901201.5MethamidophosLC–MS–MS6051540.3 Conclusions The ethyl acetate-based multi-residue method has been modified to meet today’s demands in respect of ease and speed of sample preparation. For GC–MS analysis, combined GCB/PSA dispersive clean-up enables prolonged injection of vegetable and fruit extracts (10 mg matrix equivalent) without maintenance. Retention time shifts induced by some matrices compared with the calibration standard are reduced by the clean-up procedure. Interferences are partially removed, resulting in cleaner (extracted ion) chromatograms. The last two benefits aid correct automatic peak assignment and confirmation. Addition of toluene during dispersive clean-up prevented unacceptable adsorption of planar pesticides by GCB yet removal of chlorophyll and other pigments was still sufficient. Use of liners with a sintered porous glass bed on the inner wall makes 20 μL injection non-critical and robust. In GC, use of a universal matrix-matched standard (tomato) is a feasible means of compensating for the matrix effects of many other vegetable and fruit samples. For most pesticides, LOQs of 0.01 mg kg−1 can be obtained by GC–MS with full-scan acquisition. The same initial extract (i.e. without any clean-up) can be used for LC–MS–MS analysis, after changing the solvent to methanol–water. LC–MS–MS is relatively tolerant of injection of matrix—despite the absence of any clean-up no special maintenance was required. Matrix-induced suppression was observed for several matrices, however, especially herbs and citrus, and must be evaluated for all pesticide-matrix combinations. In contrast with the GC–based method, use of a universal matrix-matched standard to compensate for matrix effects was not feasible. Evaluation of analytical quality control data for 271 pesticides and degradation products in over one hundred matrices showed that, at the 0.05 mg kg−1 level, recovery was acceptable for 92% (LC–MS–MS) and 93% (GC–MS) of all pesticide–matrix combinations. It also revealed that the method fails in the other 7–8% because of lack of specificity (mostly in GC–MS) or because of poor extraction efficiency and/or degradation (LC–MS–MS). The only way to identify these limitations is by thorough and continual evaluation of the quantitative performance of the method for all the pesticides (rather then a “representative subset”) in all the matrices.	[ "pesticides", "multi-residue analysis", "foods/beverages", "gc-ms", "lc-ms/ms" ]	[ "P", "P", "M", "U", "U" ]
Biochim_Biophys_Acta-1-5-2212780	The role of mitochondria in protection of the heart by preconditioning	A prolonged period of ischaemia followed by reperfusion irreversibly damages the heart. Such reperfusion injury (RI) involves opening of the mitochondrial permeability transition pore (MPTP) under the conditions of calcium overload and oxidative stress that accompany reperfusion. Protection from MPTP opening and hence RI can be mediated by ischaemic preconditioning (IP) where the prolonged ischaemic period is preceded by one or more brief (2–5 min) cycles of ischaemia and reperfusion. Following a brief overview of the molecular characterisation and regulation of the MPTP, the proposed mechanisms by which IP reduces pore opening are reviewed including the potential roles for reactive oxygen species (ROS), protein kinase cascades, and mitochondrial potassium channels. It is proposed that IP-mediated inhibition of MPTP opening at reperfusion does not involve direct phosphorylation of mitochondrial proteins, but rather reflects diminished oxidative stress during prolonged ischaemia and reperfusion. This causes less oxidation of critical thiol groups on the MPTP that are known to sensitise pore opening to calcium. The mechanisms by which ROS levels are decreased in the IP hearts during prolonged ischaemia and reperfusion are not known, but appear to require activation of protein kinase Cε, either by receptor-mediated events or through transient increases in ROS during the IP protocol. Other signalling pathways may show cross-talk with this primary mechanism, but we suggest that a role for mitochondrial potassium channels is unlikely. The evidence for their activity in isolated mitochondria and cardiac myocytes is reviewed and the lack of specificity of the pharmacological agents used to implicate them in IP is noted. Some K+ channel openers uncouple mitochondria and others inhibit respiratory chain complexes, and their ability to produce ROS and precondition hearts is mimicked by bona fide uncouplers and respiratory chain inhibitors. IP may also provide continuing protection during reperfusion by preventing a cascade of MPTP-induced ROS production followed by further MPTP opening. This phase of protection may involve survival kinase pathways such as Akt and glycogen synthase kinase 3 (GSK3) either increasing ROS removal or reducing mitochondrial ROS production. 1 Introduction The heart is one of the most energy demanding tissues in the body and is totally dependent upon oxidative phosphorylation to supply the large amount of ATP required for beat-by-beat contraction and relaxation. If the blood flow to the heart is impaired (ischaemia), as occurs when a blood clot occludes a coronary artery (coronary thrombosis) or where cardiac surgery requires the heart to be disconnected from the blood supply, the source of oxygen is removed leading to the cessation of oxidative phosphorylation. This causes tissue ATP and creatine phosphate concentrations to decrease with a concomitant rise in ADP, AMP and Pi concentrations. Although glycolysis is activated, it is unable to meet the demand of the beating heart for ATP. Consequently, the heart rapidly ceases to beat as the contractile machinery is inhibited by elevated Pi and ADP, combined with the decreasing pH that accompanies the accumulation of glycolytic lactic acid [1–3]. The heart can usually survive a short period of ischaemia and then recover upon reperfusion. Although the performance of the heart may be impaired initially (stunning), given time recovery is complete. However, if the period of ischaemia is too long, the tissue becomes irreversibly damaged. Hence, if the heart is to be salvaged, it is important to restore the blood flow as soon as possible. Yet, paradoxically, such reperfusion can exacerbate the damage occurring during the ischaemic period. This is known as reperfusion injury and is accompanied by enzyme release and morphological changes characteristic of necrosis [1–3]. The extent of damage can be visualised as an area of necrotic tissue known as the infarct whose area can be determined to provide a quantitative measure of injury. Quantification of damage may also be provided by measuring the release of intracellular proteins such as lactate dehydrogenase or troponin I [4]. In addition to the necrotic cell death that represents the major damage to the reperfused heart there is also evidence that some myocytes around the periphery of the infarct die by apoptosis [5,6]. Understanding the causes of reperfusion injury and devising ways of preventing it is of major clinical importance in cardiac surgery and the treatment of coronary thrombosis. There is increasing evidence that mitochondrial dysfunction plays a central role in mediating both the necrotic and apoptotic components of reperfusion injury, and that one of the most effective ways of protecting hearts from such injury, known as ischaemic preconditioning (IP), acts to attenuate this [1,3,7–10]. This review will address the causes of reperfusion injury, emphasising the role of the mitochondrial permeability transition pore (MPTP) and prevention of its opening by IP. The proposed signalling mechanisms through which IP may exert its effects will be discussed, including the proposed role of mitochondrial potassium channels which has been a major but controversial area of research (see [11–13]. However, first it is necessary to explain the phenomenon of preconditioning. 2 The phenomenon of ischaemic preconditioning Ischaemic preconditioning (IP) involves exposing the heart to brief periods of ischaemia (typically 2–5 min) interspersed with periods of normal perfusion prior to the prolonged ischaemia initiated within an hour of the preconditioning protocol. This protocol was first shown to offer strong protection against reperfusion injury in the dog heart by Murry et al. in 1986 [14] and has since been confirmed in all species investigated, including humans [15–17]. However, it should be noted that significant reduction in infarct size only occurs when the optimal duration for the test period of ischemia is chosen and is lost if the ischaemic period is too long. Preconditioned hearts exhibit a smaller infarct size and intracellular enzyme release (indicators of necrotic cell death) and fewer arrhythmias [14,18], whilst contractile function is preserved [19]. Typically, 2–3 brief (2–5 min) cycles of ischaemia and reperfusion are used in IP, there being little advantage of additional cycles [15]. If ischaemia is initiated more than 1–2 h after the preconditioning protocol, protection is lost but re-emerges again after about 24 h and lasts for up to 3 days. This is termed the second window of preconditioning [15,20–22]. A wide variety of drugs that activate different signalling pathways thought to be involved in mediating IP (see Section 6) can also induce protection and thus preconditioning is often used as a generic term to encompass any protocol applied before prolonged ischaemia that protects the heart during reperfusion. Not to be confused with such preconditioning is a more recently described phenomenon known as post-conditioning. Here hearts are given several very brief ischaemic interludes (10 s) during the early phase of reperfusion which also improves the recovery of the heart and reduces infarct size [23–25]. This protocol has considerable clinical potential since intermittent reperfusion can be induced during angioplasty prior to permanent restoration of blood flow. It is beyond the scope of this review to consider the similarities and differences between preconditioning and postconditioning in any detail. However, many of the mechanisms and signalling pathways mediating protection are thought to be the same [25] and where effects on mitochondria have been demonstrated in postconditioning they will be noted. 3 Causes of reperfusion injury Increases in cellular [Ca2+] and reactive oxygen species (ROS), initiated in ischaemia and then amplified upon reperfusion, are thought to be the main causes of reperfusion injury. Mitochondria are involved both in the production of ROS and as targets for the damaging action of both ROS and calcium [1,3,9,26]. 3.1 Damage occurring in ischaemia (Reviewed in [1;3;26]) During ischaemia, the increased rate of glycolysis causes lactic acid to accumulate and the intracellular pH (pHi) to drop rapidly. This activates the Na+/H+ antiporter as the cell endeavours to restore pHi. However, the rapidly declining ATP concentrations cause inhibition of the Na/K ATPase and lead to a rise in intracellular [Na+]. This in turn reduces the ability of the cell to restore its pHi and increases intracellular [Ca2+] since the Na+/Ca2+ antiporter that usually pumps Ca2+ out of the cell, is inhibited or reversed. The conversion of ATP to ADP and AMP is rapid and reversible. AMP is slowly converted into adenosine and then inosine and xanthine through a purine degradation pathway. These nucleosides leak out of the cell (and may have vasodilator effects through purinergic receptors) and lead to a gradual depletion of adenine nucleotide which may contribute to the reduced cardiac performance (stunning) seen on reperfusion. Another feature of ischaemia that is thought to contribute to the damage seen during reperfusion is the production of reactive oxygen species (ROS). There is a small increase in ROS production immediately the heart becomes ischaemic which remains stable for 20–25 min but then increases dramatically [27]. This large increase is thought to play a major role in damaging the heart during ischaemia and sensitising it to reperfusion as described below. However the source of the ROS is unclear and might involve complex 1 and complex 3 of the respiratory chain or perhaps more likely xanthine oxidase acting on xanthine formed from the degradation of adenosine as noted above [26,28–31]. The depletion of ATP combined with elevated [Ca2+] and ROS leads to a gradual decline in cellular integrity as degradative enzyzmes are activated and ATP-dependent repair processes are unable to operate [1,3,26,32]. If the tissue remains ischaemic for only short periods and the mitochondria remain sufficiently intact to generate the ATP, tissue damage is slight and can be repaired by ATP-dependent processes upon reoxygenation. However, a critical point is reached when recovery is not possible and reperfusion actually causes further damage to the heart, causing extensive necrosis with associated enzyme release [1,3]. 3.2 Damage occurring during reperfusion Reperfusion is associated with a burst of reactive oxygen species (ROS) production [27], but here too the source of the ROS is debated. Although some may be produced by xanthine oxidase and NADPH oxidase, it is probable that most is formed by complex 1 and complex 3 of the respiratory chain [26,28–30]. When the respiratory chain is inhibited by lack of oxygen and then re-exposed to oxygen, ubiquinone can become partially reduced to ubisemiquinone. This can then react with the oxygen to produce superoxide that is reduced to hydrogen peroxide by superoxide dismutase. Hydrogen peroxide is removed by glutathione peroxidase or catalase, but if ferrous ions (or other transition metals such as copper) are present it will form the highly reactive hydroxyl radical through the Fenton reaction [28]. Mitochondrial proteins are especially susceptible to ROS induced damage and this is reflected in the impaired respiratory chain activity of mitochondria isolated from ischaemic hearts (see [26,31]). Thus ROS have direct effects on several respiratory chain components, most especially on complex 1 but also on complex 3, and other iron sulphur proteins such as aconitase. ROS can also cause thiol oxidation and inhibition of the ATPase and adenine nucleotide translocase. In addition, ROS cause oxidation of glutathione that may then form mixed disulphides with proteins. Such protein modification is thought to have inhibitory effects on ion pumps and therefore exacerbate the effects of ATP deprivation on ionic homeostasis [33–35]. ROS also cause peroxidation of the unsaturated fatty acid components of the phospholipids, and especially cardiolipin of the inner mitochondrial membrane, and this leads to further inhibition of respiratory chain activity [36,37]. Furthermore, lipid peroxidation causes the release of reactive aldehydes such as 4-hydroxynonenal that can modify membrane proteins [38]. Overall, it is thought that the combined effects of ROS and elevated [Ca2+] play a critical role in the transition from reversible to irreversible reperfusion injury, and that mitochondria are the major target of these agents. In particular, they lead to the opening of the mitochondrial permeability transition pore, that is now widely accepted to play a critical role in reperfusion injury [1,3,10,26,39]. 4 The mitochondrial permeability transition pore (MPTP) 4.1 Properties of the MPTP Under normal physiological conditions, the mitochondrial inner membrane is impermeable to all but a few selected metabolites and ions; this is essential to maintain the membrane potential and pH gradient that together drive ATP synthesis through oxidative phosphorylation. However, under conditions of high matrix calcium, especially when this is accompanied by oxidative stress, high phosphate and low adenine nucleotide concentrations, a non-specific pore opens in the inner mitochondrial membrane known as the mitochondrial permeability transition pore (MPTP). The properties and regulation of the MPTP are reviewed extensively elsewhere [3,40,41]. Once open, this pore allows free passage of any molecule of < 1.5 kDa and so disrupts the permeability barrier of the inner membrane. This has two major consequences. First, it allows unrestricted proton movement across the inner membrane, causing oxidative phosphorylation to be uncoupled. Not only does this prevent ATP synthesis but it also enables the proton-translocating ATPase to reverse direction and so actively hydrolyse ATP rather than synthesise it. Under such conditions, intracellular ATP concentrations rapidly decline, leading to the disruption of ionic and metabolic homeostasis and activation of degradative enzymes such as phospholipases, nucleases and proteases [3,9,26]. Unless pore closure occurs, these changes will cause irreversible damage to the cell resulting in necrotic death. It should be noted that opening of just a single pore in one mitochondrion is likely to cause its immediate depolarisation. This will then activate further pore opening in the same mitochondrion since MPTP opening of calcium loaded mitochondria is activated by depolarisation [42]. Thus mitochondria are either fully open or closed, and it is the fully open state that leads to the second consequence of MPTP opening; the mitochondria swell extensively [43]. This occurs as all small molecular weight solutes equilibrate across the inner membrane, leaving the high concentration of matrix proteins to exert a colloidal osmotic pressure that leads to the uptake of water and matrix swelling. Although unfolding of the cristae allows the matrix to expand without rupture of the inner membrane, the outer membrane will break and lead to the release of proteins in the intermembrane space such as cytochrome c and other factors that play a critical role in apoptotic cell death [44,45]. 4.2 The molecular identity of the MPTP The molecular identity of the mitochondrial permeability transition pore remains uncertain [40,41,46], but it is generally accepted that an inner membrane component undergoes a calcium-triggered change in conformation that is facilitated by cyclophilin D (CyP-D), a peptidyl-prolyl cis-trans isomerase [46,47]. The role of CyP-D was first suggested by the discovery that cyclosporin A (CsA) acts as a potent inhibitor of pore opening [48]. Further studies revealed that the potency of different CsA analogues to inhibit pore opening correlates with their ability to inhibit the peptidyl-prolyl cis-trans isomerase activity within the matrix [49,50] that was subsequently identified as CyP-D [51,52]. Extensive work from many laboratories confirmed the critical role of CyP-D and this was finally put beyond doubt by the demonstration that MPTP opening in liver mitochondria from CyP-D knockout mice is much less sensitive to calcium than normal mitochondria, and is no longer inhibited by CsA [53–55]. The identity of the membrane component of the MPTP is less certain. However, the most widely accepted view is that the adenine nucleotide translocase (ANT) normally fulfils this role and extensive circumstantial data supports this view (see [9,56]). Thus opening of the MPTP is inhibited by adenine nucleotides with a similar concentration dependence and specificity as they exhibit when acting as substrates for the ANT, and this inhibition is overcome by the specific inhibitor of the ANT, carboxyatractyloside (CAT) that traps the ANT in its “c” conformation. By contrast, another inhibitor of the ANT, bongkrekic acid, that causes the carrier to take up the alternative “m” conformation, inhibits pore opening [57]. The ANT can also account for the sensitisation of the MPTP to calcium by oxidative stress and the vicinal thiol reagent phenylarsine oxide (PAO) [57]. Thus cysteine residues 160 and 260 of rat ANT2 can be cross-linked by oxidative stress or PAO, with modification of Cys160 alone being sufficient to prevent the inhibition of MPTP opening by adenine nucleotides, so stimulating pore opening [58]. Further evidence for an important role for the ANT is the ability of the ANT to bind to CyP-D in a CsA-sensitive manner [58,59]. In addition, when the purified ANT is reconstituted into proteoliposomes high calcium concentrations can induce the formation of non-specific channels [60] and this process is sensitised to [Ca2+] by the addition of purified cyclophilin [61]. Nevertheless, despite the strong evidence in favour of the ANT being the critical membrane component of the MPTP, it is unlikely to be essential. Thus in an elegant study, that has yet to be confirmed by others, mitochondria from mouse livers in which ANT1 and ANT2 had been knocked out were found to exhibit MPTP opening that was inhibited by CsA [62]. However, pore opening in the ANT-knockout mitochondria required much higher calcium concentrations than did wild-type mitochondria, and was no longer sensitive to adenine nucleotides confirming that the ANT is at the very least playing a critical regulatory role. One possibility is that the ANT is the normal membrane component of the MPTP but that other less abundant members of the mitochondrial carrier family can fulfil this role in the absence of the ANT [63]. An alternative possibility, proposed by Lemasters, is that unfolded membrane proteins form the MPTP and that the ANT is the most abundant of these [64]. Many other proteins have been proposed to be components of the MPTP, including the peripheral benzenediazipine receptor, creatine kinase and the voltage dependent anion channel (VDAC) [3,40,41,65]. Of these perhaps the strongest candidate is VDAC since an interaction between VDAC and the ANT has been demonstrated and there is evidence that the MPTP may be associated with contact sites between the inner and outer mitochondrial membranes [3,40,65]. Furthermore, initial reports suggested that VDAC might be the locus for inhibition by ubiquinone analogues [66]. However, the recent demonstration that liver mitochondria from mice lacking VDAC1 exhibit normal MPTP opening and inhibition by ubiquinone analogues [67] make an essential role for VDAC1 unlikely. See note added in proof. 5 The MPTP opening plays a central role in reperfusion injury 5.1 The MPTP opens during reperfusion but not ischaemia We have developed a technique to demonstrate that the MPTP is kept firmly closed in the heart under normal physiological conditions, but opens upon reperfusion following a period of ischaemia. This technique involves measuring the extent of [3H]-2-deoxyglucose entrapment in mitochondria that have undergone the permeability transition [68]. Others have used fluorescence microscopy to measure MPTP opening in isolated cardiac myocytes and have also confirmed that the pore opens under conditions of simulated ischaemia and reperfusion [69]. It might be expected that the pore would also open after prolonged ischaemia and some have reported this to be the case [70,71]. However, these studies relied on cytochrome c release as a measure of MPTP opening, but it is known that this can occur independently of the MPTP as a result of Bax translocation to mitochondria during ischaemia [72]. By contrast, our own direct measurements [68] and those of others [73] do not support MPTP opening in ischaemia. Since the pore has been shown to be powerfully inhibited by low pH (< 7) [74] and the time course of opening during reperfusion correlates with the return of pH to normal [75], we have suggested that it is the low pH accompanying ischaemia that prevents pore opening despite the conditions of oxidative stress, low adenine nucleotide and elevated phosphate and calcium [75]. Indeed, as would be predicted, maintaining a low pH during the initial phase of reperfusion is known to protect hearts from reperfusion injury [76–78] and may play a role in the cardioprotective effects of inhibitors of the Na+/H+ exchanger such as amelioride and cariporide (see [3,79,80]). A similar mechanism has also been implicated recently in the protection offered by postconditioning [81,82]. It has also been proposed that the rapid energisation of mitochondria at reperfusion will lead to electrogenic calcium uptake into the mitochondria. Since the calcium has accumulated in the cytosol during ischaemia, this might be expected to induce mitochondrial calcium overload and hence MPTP opening [1,9]. However, measurements of mitochondrial [Ca2+] in isolated cardiac myocytes imply that it is the mitochondrial [Ca2+] at the end of ischaemia rather than during reperfusion that correlates better with cell injury [83]. Another factor that will reinforce MPTP opening at reperfusion is the surge of ROS production that is known to occur [27,84] and this may well be the most important factor in determining the outcome of reperfusion [73]. Indeed, there is increasing evidence that the extent of MPTP opening is a critical factor in the transition from reversible to irreversible reperfusion injury [1,3,9,65]. More accurately, it is the extent to which pores remain in an open state that correlates with damage, since we have demonstrated that some mitochondria in which the MPTP opens initially, subsequently undergo pore closer as reperfusion continues [75,85]. It is possible that myocytes in which the majority of mitochondria have undergone such reversible MPTP opening will have swollen sufficiently to release pro-apoptotic proteins such as cytochrome c and thus initiate apoptosis as noted in Section 4.1. These myocytes are likely to be at the periphery of the infarct where the ischaemic insult was incomplete and thus would account for the ring of apoptotic myocytes that have been observed to surround the necrotic core of the infarct [5,6]. It should be noted that mice in which CyP-D had been knocked out demonstrate normal development and cells from them respond to a range of apoptotic stimuli in the same manner as wild-type mice [54,55]. Thus MPTP opening cannot the mechanism by which apoptosis is mediated in the healthy animal, but this does not preclude apoptosis being induced in this way under pathophysiological situations such as those described above [46]. 5.2 Inhibition of MPTP opening protects hearts from reperfusion injury If the opening of the MPTP is a critical event in reperfusion injury then it would be predicted that inhibition of pore opening should protect hearts from injury. Many studies have confirmed that this is the case. Protection was first demonstrated with CsA which, when added prior to global ischaemia, was shown to improve haemodynamic function and ATP levels on reperfusion and to decrease necrotic damage as reflected in intracellular enzyme release [86]. Subsequent studies have shown that other CsA analogues, and the alternative CyP-D inhibitor sanglifehrin A (SfA) also provide protection under these conditions [68,87,88] whilst mitochondria isolated from these hearts showed less pore opening in response to a given calcium load [88,89]. In addition, CsA and SfA can reduce the infarct size of hearts in which a coronary artery is occluded and then re-opened to mimic the treatment of a coronary thrombosis. In this model of reperfusion injury, protection is observed even when the drug is added only at reperfusion [90,91]. Less direct approaches to inhibit MPTP opening are also effective at protecting the heart from reperfusion injury. Thus free radical scavengers are well known to be protective and in the case of one such agent, the anaesthetic propofol, we have used the 2-deoxyglucose entrapment technique to confirm inhibition of the MPTP directly [92]. In addition, we have demonstrated that propofol provides protection from reperfusion injury in a pig model of open heart surgery [93]. Inhibition of the sodium proton exchanger with cariporide is also known to protect the heart from reperfusion injury [94] and here too inhibition of MPTP opening correlates with this protection [80]. This protection probably reflects less sodium and hence calcium loading of the cardiac myocytes during ischaemia, coupled with a lower pH during the initial phase of reperfusion, both of which will reduce MPTP opening. The most powerful agent for inhibiting MPTP opening and protecting the heart is pyruvate [75,95]. The presence of 10 mM pyruvate during ischaemia and reperfusion can almost totally protect rat hearts from 30 min global ischaemia and this is accompanied by greatly reduced MPTP opening followed by total pore closure [75]. Pyruvate has three beneficial effects that may contribute towards its inhibition of MPTP opening: it is a free radical scavenger, it maintains a lower intracellular pH during the early phase of reperfusion and it is an excellent fuel for ATP synthesis that bypasses the ATP-requiring steps of glycolysis [75,96]. 5.3 Protection by preconditioning involves inhibition of MPTP opening In view of the critical role of the MPTP in reperfusion injury and the proven ability of inhibitors of the MPTP to protect from injury it might be predicted that ischaemic preconditioning (IP) would inhibit pore opening. Both direct and indirect techniques have been used to confirm this. Our own experiments have used the 2-deoxyglucose entrapment technique to demonstrate that less MPTP opening occurs at reperfusion in preconditioned hearts and that as reperfusion continues the majority of mitochondria that originally opened close again in the IP hearts but not in control hearts [85]. Others have used confocal microscopy to monitor mitochondrial membrane potential in an isolated cardiac myocyte model of preconditioning and shown that IP desensitises the mitochondria to pore opening induced by oxidative stress [69,97]. Recent experiments performed with two-photon microscopy in the perfused heart have confirmed these data [98]. Yet other studies have investigated the sensitivity of isolated mitochondria from control and IP hearts to pore opening in response to added calcium [85,99,100]. Both light scattering and mitochondrial calcium retention techniques demonstrated that when mitochondria were isolated immediately after the preconditioning stimulus, those from IP hearts were not any less sensitive to calcium-induced pore opening than those from control hearts. Indeed, if anything they were more sensitive [85,100]. When isolated at the end of ischaemia or during the first few minutes of reperfusion mitochondria were more sensitive to MPTP opening than those isolated prior to ischaemia but at these time points mitochondria from IP hearts were considerably less sensitive to calcium-activated pore opening than those from control hearts [99–101]. These data clearly show that the IP protocol alone is not having a direct effect on the MPTP but suggest that IP inhibits a process occurring during ischaemia and reperfusion that is responsible for sensitising the MPTP to calcium. This is an important fact to bear in mind when evaluating the extensive literature surrounding the mechanisms by which preconditioning exerts its protective effects as will be discussed further below (Section 6). Temperature preconditioning, where hearts are exposed to three brief hypothermic (26 °C) episodes prior to normothermic ischaemia is also thought to protect through inhibition of the MPTP [101] as is post-conditioning [102]. Thus, mitochondria isolated from such hearts during reperfusion are less sensitive to MPTP than those isolated from control hearts with temperature preconditioning being even more protective than IP [101]. 6 Signalling pathways linking preconditioning to inhibition of the MPTP There is a plethora of data implicating many different signalling pathways in preconditioning and the relevant role of each remains hotly debated (see [16,30,103–107]). Many of these studies have utilised pharmacological agonists and antagonists of components of the different signalling kinase cascades and their potential targets. The complex interactions that exist between different signalling pathways makes interpretation of these data complicated and this is made worse by the questionable specificity of some of the pharmacological interventions. The reader is referred elsewhere for a detailed consideration of the many potential pathways that have been proposed [30,103–107]. Here we will only provide a brief overview before focussing on how they may exert their effects on the mitochondria. 6.1 The role of protein kinase C There is extensive evidence that protein kinase C (PKC) plays a central role in preconditioning since inhibition of PKC has been shown to block the protection afforded by IP and pharmacological activators of PKC are cardioprotective (see [105]). There remains some controversy over which of the many PKC isoforms may be involved in IP, whether they translocate to the particulate fraction and how they exert their effects [103]. Nevertheless, there is a large body of evidence to implicate PKCε as an important player in IP [105]. Thus hearts from PKCε knockout mice do not exhibit IP [108] whereas transgenic mice with cardiac-specific over-expression of PKCε or expression of an activator of PKCε are protected from reperfusion injury [109–111]. Some studies have reported PKCε translocation to the particulate fraction, including mitochondria [112–114] and have suggested a direct inhibition of the MPTP by PKCε involving phosphorylation of components of the MPTP such as the voltage dependent anion channel (VDAC) [113–115]. However, in our own studies we were unable to detect PKCε translocation to the mitochondria of IP hearts (Fig. 1A) whilst others have observed that PKC translocation to the particulate fraction following IP is transient and lost during progressive brief cycles of ischaemia and reperfusion [116]. Although a role for PKCε in IP seems established, the mechanism(s) by which it exerts its effects are less clear as will become apparent below (see sections 7–10). It is also uncertain how ischaemic preconditioning activates PKCε, although several pathways may be involved. Factors released during the brief ischaemic periods such as adenosine, bradykinin, noradrenaline and opioids may bind to their G protein-coupled receptors to stimulate phospholipase C, producing diacylglycerol that activates PKC. Indeed, all of these factors can pharmacologically precondition the heart [104]. Accumulating evidence supports a role for the modest increase in ROS that occurs during IP protocol in the activation of PKC [27,117,118]. Thus ROS are known to activate PKC in the isolated heart [118–120] and IP can be prevented if free radical scavengers are present during the preconditioning phase [100,101,118,121] Oxidation of critical cysteine residues on PKC isoforms is known to cause their activation [35,122] and thus provides a mechanism by which ROS could activate PKC. 6.2 The role of nitric oxide and cyclic GMP dependent protein kinase Data from several laboratories have implicated activation of cyclic GMP dependent protein kinase (PKG), perhaps by nitric oxide, in the signalling pathway for IP. Thus it has been reported that nitric oxide donors and PKG activators can induce preconditioning whilst NO scavengers and PKG inhibitors prevent preconditioning [106,123,124]. Pharmacological studies from Garlid's laboratory have led them to conclude that PGK and PKCε work in concert to induce IP through an effect on the putative mitochondrial ATP-dependent potassium channel (mitoKATP — see Section 9 below). They propose that cGMP activates PKG localized at the cytosolic surface of the mitochondrial outer membrane and that this phosphorylates some target protein which in turn can somehow activate PKCε residing in the intermembrane space of mitochondria. This PKCε then would phosphorylate the mitoKATP channel to mediate preconditioning [123]. In a subsequent paper they proposed that activation of the mitoKATP channel increased ROS formation and that this ROS activated a second pool of PKCε that inhibited the MPTP [125]. Our own data do not support this (see Sections 7). It should be noted that an alternative or additional mechanism has been proposed by which nitric oxide may protect the heart when present during the initial stages of reperfusion (see [128]). Cytochrome oxidase is competitively inhibited by nitric oxide and this would lead to an inhibition of respiration and hence a reduction in the mitochondrial membrane potential. It is suggested that one consequence of this would be less mitochondrial calcium accumulation and hence less MPTP opening. However, another consequence of a lower mitochondrial membrane potential would be impaired ATP production which would not be predicted for hearts showing improved recovery. Similar arguments have been made for the proposed depolarisation caused by mitochondrial KATP channel opening and these will be considered in more detail in Section 10.2. 6.3 The role of pro-survival kinases Several laboratories have presented evidence to implicate activation of pro-survival kinases such as Akt (protein kinase B) in preconditioning [107,126]. It is proposed that tyrosine kinases are activated during IP through some ill-defined interaction with G-protein coupled receptors such as the adenosine and bradykinin receptors and that this causes activation of phosphatidyl inositol 3 phosphate kinase (PI-3-kinase). The resulting phosphatidylinositol 3,4-bisphosphate then activates phosphoinositide-dependent kinase 1 (PDK1) that in turn phosphorylates and activates Akt. However, it is also possible that activation of tyrosine kinases or inhibition of tyrosine phosphatases and the phosphatidyl inositol 3,4,5,-trisphosphate phosphatase PTEN (Phosphatase and TENsin homolog) may be mediated by the ROS produced during the preconditioning stimulus [107,127]. In support of this mechanism it has been demonstrated that phosphorylation of Akt is enhanced by IP, especially during reperfusion, and that pharmacological inhibitors of either Akt or PI-3-kinase prevent preconditioning [126,128–130]. It has further been proposed that following phosphorylation, activated Akt phosphorylates endothelial nitric oxide synthase to produces NO that acts via PKG as noted above [106,107]. Sollott et al. [97] have suggested that all these kinases may converge to phosphorylate and inhibit glycogen synthase kinase 3β (GSK3β) that they report translocates to the mitochondria. Inhibition of GSK3β by Akt-induced phosphorylation would then mediate protection. Although it is known that GSK3β is pro-apoptotic and that its phosphorylation inhibits its activity and stimulates cell survival [131], the mechanism of any mitochondrial effects remain to be elucidated. However, it may be significant that GSK3β has been reported to phosphorylate VDAC, and that this phosphorylation is associated with a sensitisation to cytotoxic drugs that stimulate pore opening [132]. VDAC was widely regarded as an integral component of the MPTP, but recent data showing that mitochondria from the VDAC1 knockout mouse have a normal MPTP cast doubt on this [53]. See note added in proof. 6.4 The role of AMP-activated protein kinase Yet another protein kinase, AMP-activated protein kinase (AMPK), is activated by a large variety of cellular stresses that deplete ATP and increase AMP, including glucose deprivation, hypoxia, ischaemia, oxidative stress and hyperosmotic stress [133]. It could be argued that the many pharmacological agents known to induce preconditioning, from KATP channel openers, to uncouplers and respiratory chain inhibitors (see Section 10), might all represent insults to the heart that increase AMP and so activate AMPK as the focus of a common signalling mechanism to protect the heart [133]. Indeed, transgenic mice expressing a kinase dead form of AMPK in the heart have been shown to exhibit greater myocardial necrosis and apoptosis after ischaemia/reperfusion [134] and cannot be preconditioned [135]. Furthermore, Nishino et al. [136] have shown that PKC inhibitors abolished AMPK activation by ischaemic episodes in the myocardium, suggesting that AMPK might be activated by IP in a PKC-dependent manner. However, our own data fail to show any abrogation of the IP-induced protection by the inhibitor of AMPK, compound C, under conditions where the inhibitor of PKC, chelerythrine, did overcome the protection [101]. Taken together, all these data suggest that several signalling pathways may interact or act in parallel to induce IP, but the ultimate target of their action remains unclear. Since IP involves inhibition of the MPTP the signalling pathway must ultimately inhibit MPTP opening, but this could be achieved either by direct phosphorylation of a component of the MPTP or indirectly by influencing factors that enhance pore opening such as by reducing oxidative stress or calcium overload. 7 Mechanism of inhibition of the MPTP by IP 7.1 Is there evidence for regulation by phosphorylation? As noted above, there are several reports that protein kinases may migrate to the mitochondria in response to IP suggesting that phosphorylation of a component of the MPTP may be responsible for its inhibition. Baines et al. have reported that the MPTP of isolated heart mitochondria was inhibited following incubation for 15 min with purified PKCε plus phorbol ester [115]. However, such experiments are hard to interpret since incubating isolated mitochondria in vitro can change the sensitivity of MPTP opening to calcium, especially if ATP is added since this is an inhibitor of MPTP opening in its own right [3,40]. Similar criticisms could be levelled against experiments in which incubation of isolated rat liver mitochondria with a cGMP analogue in the presence of a cytosolic extract and ATP caused inhibition of the calcium-induced mitochondrial permeability transition [137], although here reversal of the effect by the PKG inhibitor KT5823 makes the data more convincing. However, no data are available to show changes in phosphorylation of any mitochondrial protein following IP that might account for inhibition of the MPTP. Furthermore, when we and others measured MPTP opening in mitochondria isolated immediately after the preconditioning stimulus we observed no reduction in sensitivity to calcium as might have been predicted if a component of the MPTP had been phosphorylated [85,100]. Indeed, it is unclear how activated cytosolic protein kinases could cross the outer membrane to phosphorylate and inhibit components of the MPTP, let alone cross the inner membrane into the matrix. Although there are reports to suggest that this occurs [97,113,114], our own attempts to demonstrate the translocation of PKCε or GSK3β to mitochondria have consistently failed as illustrated in Fig. 1. A key aspect of our experiments was the use of Percoll gradient centrifugation to remove plasma membrane contamination which we have shown is significant in less pure mitochondrial fractions. This was demonstrated by the presence of monocarboxylate transporter 1 (MCT1) in the crude but not the purified mitochondria. Our inability to detect protein kinase translocation into the mitochondrial matrix is entirely consistent with extensive work from this laboratory more than 20 years ago in which high specific activity 32P labelling was used to investigate the phosphorylation of mitochondrial matrix proteins. Such studies demonstrated the presence of two proteins within the mitochondrial matrix that could be phosphorylated in Percoll-purified mitochondria from both heart and liver. These were pyruvate dehydrogenase (PDH) and branched chain 2-oxoacid dehydrogenase (BCDH) [138]. This was true whether the mitochondria were from control or glucagon-treated (raised cyclic AMP) rats [139], yet mitochondria from glucagon-treated rats were found to exhibit enhanced calcium retention characteristic of reduced sensitivity of the MPTP to calcium-induced opening [140,141]. Although these data do not support a role for protein kinases within the mitochondrial matrix in the inhibition of the MPTP following preconditioning, they do not exclude the possibility of regulation by phosphorylation of proteins exposed to cytosolic kinases, such as those in the outer mitochondrial membrane or bound to it using scaffolding proteins [142,143]. Indeed, work from this laboratory many years ago demonstrated that when mitochondria were isolated from hepatocytes incubated with high specific activity 32P, or isolated mitochondria were incubated with γ-[32P]-ATP, additional phosphorylated proteins were observed. Two of these proteins, both 30–35 kDa, demonstrated an increase in phosphorylation with glucagon treatment [139,144]. However, a weakness of the 32P labelling technique is that it will not detected proteins whose phosphorylation turnover is small and this may account for the larger number of phosphorylated proteins within mitochondria detected using a phosphoprotein-specific dye [145]. Nevertheless, a lack of turnover might suggest that such a phosphorylation does not play an important role in signalling. There have been an increasing number or reports in which proteomic and immunodetection methods have been used to detect a variety of other kinases associated with the mitochondria [142,146]. However, the extent to which these represent real mitochondrial kinases rather than non-specific mitochondrial contamination of the mitochondrial fraction is hard to assess. We have recently used a proteomics approach to identify any mitochondrial proteins whose phosphorylation might be altered by preconditioning. Mitochondria were isolated from control and IP hearts immediately after the preconditioning stimulus as well as the end of ischaemia and after three min reperfusion and phosphor-proteins detected using 2-D gel electrophoresis and staining with ProQ Diamond phospho-protein fluorescent dye. Although a significant number of phospho-proteins were detected in addition to the predominant spots representing PDH and BCDH, no consistent changes were detected in any of these in response to IP (unpublished data of S.J. Clarke, I. Khaliulin, Kate Heesom and A.P. Halestrap). Typical data are shown in Fig. 2. Overall, our data do not support a role for mitochondrial protein phosphorylation in mediating the inhibition of MPTP opening by IP. Rather, our data suggests that the inhibition of the MPTP by preconditioning may be secondary to either diminished ROS production or calcium overload as described below. 7.2 Effects of IP on ROS production and calcium loading during ischaemia and reperfusion as potential mediators of MPTP inhibition As noted above, opening of the MPTP is triggered by calcium overload especially when this is accompanied by oxidative stress and these two parameters appear critical determinants of the recovery of isolated heart cells from simulated ischaemia. Indeed, the extent of recovery has been reported to correlate inversely with mitochondrial matrix calcium concentrations at the end of ischaemia [83,147] and to be triggered by the rise in ROS during reperfusion [73]. It is well established that preconditioning reduces ROS production both at the end of ischaemia and during reperfusion [27,84,148,149] and decreases mitochondrial calcium overload [150–153]. Thus it is quite possible to account for the IP-mediated inhibition of MPTP opening at reperfusion merely through these indirect effects without having to invoke phosphorylation of any mitochondrial protein. Our own data are entirely consistent with this mechanism. Opening of the MPTP in mitochondria isolated immediately following the preconditioning stimulus showed the same sensitivity to calcium as in control mitochondria [85,100]. Mitochondria isolated at the end of ischaemia, or during reperfusion, were more sensitive to calcium-induced MPTP opening than those isolated prior to ischaemia. However, at these times the mitochondria from preconditioned hearts were less sensitive to MPTP opening than were mitochondria isolated from control hearts at the same time [99–101]. This difference in sensitivity to calcium correlated with the exposure of the mitochondria to oxidative stress as reflected in the carbonylation of mitochondrial proteins [100,101], a surrogate marker for oxidative stress [154]. These data confirm that the oxidative stress mitochondria experience at the end of ischaemia and during reperfusion is significantly attenuated in preconditioned hearts. As such they provide strong evidence in favour of IP exerting its primary effect on the reduction of ROS levels during ischaemia and reperfusion with the secondary consequence that MPTP opening is diminished. If this is the case, the key signalling pathways in preconditioning will be those that mediate the decrease in ROS levels at the end of ischaemia and during reperfusion. This could be the result of decreased ROS production or of improved ROS removal, but the ability of preconditioning to protect against oxidative stress mediated by exogenous hydrogen peroxide [155,156] argues in favour of the latter. However, an additional effect of IP to reduce ROS production cannot be ruled out. 8 The role of mitochondrial potassium channels Plasma membrane ATP-dependent potassium channels (KATP) are strongly expressed in the heart and have long been associated with cardioprotection [12,157]. It was originally proposed that opening of the plasma membrane KATP channel might hyperpolarize the cell leading to a shorter action potential duration (APD) and hence less calcium entry during ischaemia which would protect the hearts from calcium overload. In support of this, mouse hearts whose KATP channel Kir 6.2 had been knocked out exhibited greater calcium overload and ischaemia/reperfusion injury than control hearts [146]. A role for the sarcolemmal KATP channels in the protective mechanism exerted by preconditioning was first suggested by the observation that KATP channel blockers such as glibencamide prevent preconditioning (see [12,157]). Further support came from the demonstration that hearts from the Kir 6.2 knockout mice were insensitive to either ischaemic or pharmacological preconditioning [158–160]. However, despite this convincing evidence in favour of a role for the sarcolemmal KATP channel in preconditioning, other data argue against such a role. In particular conditions were found in which cardioprotection could be observed in the absence of APD shortening whilst a range of KATP channel openers were reported to show a poor correlation between their effect on APD and their protective effects (see [11,12,161–163]). These data led to a change of emphasis, away from the sarlcolemmal KATP channel towards a similar KATP channel proposed to reside in the mitochondrial inner membrane. 8.1 The identity of mitochondrial K+ channels and their physiological roles The presence of an electrogenic mechanism for K+ entry into mitochondria (K+ channel) together with a K+/H+ antiporter to pump K+ out again has been known for more than 30 years, and their relative activities are thought to play a key role in the regulation of matrix volume (see [43,164]). What is less clear is the identity of these channels and transporters since none have been purified sufficiently to allow sequencing. Nor has genomic analysis revealed any plasma membrane K+ channel isoform or spliced variant with a mitochondrial targeting sequence [12]. Rather the presence and properties of the K+ channels and transporters have been surmised from their functional characterisation (see [165,166]). 8.2 Techniques used to characterise mitochondrial K+ channels In isolated mitochondria, potassium ion movements can be measured directly using isotopes or K+-sensitive fluorescent dyes, or indirectly by determining matrix volume changes (see [167–171]). For mitoplasts patch clamping has been employed [172–174] whilst studies have been performed on proteoliposomes containing reconstituted inner mitochondrial membrane proteins using electrophysiological measurement of single channel current or fluorescent techniques to measure net K+ transport [168,175–179]. A major concern with the patch clamping and reconstitution techniques is the possibility that any channels detected in the mitochondrial membranes may actually represent a small number of K+ channels present in contaminating sarcolemmal membranes. It is known that most preparations of isolated mitochondria are extensively contaminated with plasma membranes unless further purified by density gradient centrifugation, and even then it is difficult to achieve complete removal of contaminating membranes [180,181]. However, in the majority of published work no data were presented to discount such contamination. Thus it is important to confirm that the activity of any channels identified by such techniques can be demonstrated in mitochondria, and this can be achieved by measurements of mitochondrial matrix volume. It is widely agreed that the matrix volume is regulated by the relative activity of the K+ channels mediating membrane potential driven potassium entry into mitochondria, and the K+/H+ antiporter catalysing proton-linked K+ extrusion activity [43,164,165,182]. Thus activation or inhibition of any K+ channel should lead to an increase or decrease in matrix volume respectively. This can be measured either directly, using an isotopic technique based around the permeability of 3H2O into the matrix and the exclusion of [14C]-sucrose, or by light scattering (see [165,182]). The latter provides a very sensitive real-time measurement of matrix volume since as the mitochondria swell, their refractive index decreases and they scatter less light. A wavelength of 520 nm is frequently employed since this represents the isosbestic wavelength at which changes in the redox state of the mitochondrial cytochromes cause no change in absorbance [182]. However, mitochondria can exhibit light-scattering changes that are independent of changes in matrix volume, reflecting rather a change in morphology or shape. It has been known for many years that such morphological changes, often referred to in terms of orthodox and condensed conformations of mitochondria, can be induced by ligands of the adenine nucleotide translocase (ANT) that switch the carrier from the “c” to the “m” conformation [44,49,170,183,184]. Thus externally added ATP, ADP or bongkrekic acid induce the “m” conformation of the ANT and cause a contraction of the mitochondria detected as an increase in light scattering whilst carboxyatractyloside, that induces the “c” conformation, exerts the opposite effect. Yet these light scattering responses all occur without changes in matrix volume measured isotopically [44,49,170,171]. Garlid and Paucek dismiss these data, stating that “this claim cannot be supported by any theoretical argument”, but the basis on which they make this assertion is unclear. First, the claim is backed up by experimental data from two different laboratories [44,49,170,171]. Second, the curvature of the mitochondria relative to the wavelength of light determines its refractive index and hence light scattering properties [182]. Thus a shape change without a change in volume (e.g. from a sphere to a cylinder) will affect the light scattering signal. Since the ANT represents some 25% or more of the inner membrane protein in heart mitochondria [185], it is not surprising that when it switches its conformation it can cause a change in mitochondrial morphology and hence light scattering. Third, Garlid and Paucek argue that changes in light scattering induced by adenine nucleotides must reflect mitochondrial volume changes because they did not observe them in potassium free medium. They go on to say “If Das et al. had carried out these simple control experiments they would have avoided spurious claims about conformational changes”. However, we [170] and Brustovetsky et al. [171] have reported other controls that confirm that the adenine nucleotide-induced light scattering changes do not reflect changes in matrix volume. Thus additions of very low concentrations of valinomycin (0.4 to 1 pmol per mg mitochondrial protein) induced changes in matrix volume of 10–25% that were readily detectable using radioisotopes, yet the magnitude of the light scattering changes were similar to those induced by ADP or ATP for which no detectable changes in matrix volume were detected (see Fig. 3 and [170,171]). We have also performed experiments in potassium free media as suggested by Garlid and Paucek, which we report in Fig. 3. These data clearly show that such adenine nucleotide induced changes in light scattering do occur in potassium free medium and similar results were obtained by Brustovetsky et al. [171]. We suggest that experiments directed towards investigating the activity of K+ channels in intact mitochondria that rely entirely upon light scattering changes for measurement of matrix volume should be treated with caution. Parallel isotopic measurements of matrix volume provide an essential confirmation that such light scattering changes do reflect changes in volume rather than morphology. 8.3 Evidence for the mitochondrial KATP channel Some 20 years ago we showed that adenine nucleotides could inhibit electrogenic mitochondrial K+ entry into mitochondria and that this was probably mediated by their binding to the ANT [49,186,187], in agreement with earlier proposals [188]. More recent data obtained with brain mitochondria have confirmed this proposal [171]. However, the existence of a mitochondrial KATP channel similar to that found in the plasma membrane was first formally proposed as a result of data obtained using patch clamping of giant fused mitoplasts from liver mitochondria [172] and light scattering experiments in isolated mitochondria [189]. Subsequently this channel has been studied extensively in several other laboratories (see [165,166]). Experiments using reconstituted proteoliposomes reported the characterisation of these channels with respect to the specificity for nucleotide inhibition and the activity of a range of pharmacological channel openers and blockers. In many cases the properties determined are similar to those of the plasma membrane KATP channel lending support to the possibility that it is actually these channels present in contaminating plasma membranes that are being characterised, rather than true mitochondrial channels (see Section 8.2). However, there are a number of differences in the reported properties of mitochondrial and sarcolemmal KATP channels that argue against this. For example, mitochondrial KATP channels have been reported to require the presence of Mg2+ for ATP to inhibit them, which is not true of the plasma membrane KATP channel. In addition, ADP acts like ATP to inhibit the mitochondrial channel but opens the sarcolemmal channel, whilst GTP has been reported to open the mitochondrial KATP channel but not the plasma membrane one [176,177]. Furthermore, their measured conductances were reported to be distinct [165] and different dose response curves for a range of pharmacological openers and blockers have been described [165,168,190]. This is especially true of the KATP channel opener diazoxide that has been reported to show a sensitivity towards the mitochondrial KATP channel at least three orders of magnitude greater than for the sarcolemmal KATP channel whilst 5-hydroxydecanoate (5HD) blocks the mitochondrial channel whilst having little effect on the sarcolemmal one [167,168]. However, there are other data claiming that diazoxide will open and 5HD block plasma membrane KATP channels at the concentrations frequently used to modulate mitochondrial KATP channels when investigating their possible role in cardioprotection [191–194]. Although no molecular identity has been ascribed to the mitochondrial KATP channel at the genomic level, similarities with the components of the plasma membrane KATP channels have been proposed. These contain one of two channel components, Kir6.1 (48 kDa) or Kir6.2 (44 kDa) and the sulphohylurea receptors, SUR1 (177 kDa) or SUR 2 (174 kDa) [12]. Two studies reported that antibodies against Kir6.1 and Kir6.2 detected proteins of the right size (about 45 kDa) in heart mitochondria, but no rigorous steps were taken to remove sarcolemmal contamination [166,195]. In heart mitochondria SUR2 antibodies detected a protein of about 25 kDa in one study [195] but no protein in the other study [166], whilst a protein of about 28 kDa was labelled by [125I]-glibencamide [196]. Overall, the available data provide no strong evidence for the presence of a KATP channel in mitochondria that is closely related to the well-characterised plasma membrane KATP channel. Garlid's laboratory found proteins of 55 kDa and 63 kDa in a partially purified preparation from brain mitochondrial inner membranes that showed KATP channel activity when reconstituted into proteoliposomes. The 63-kDa protein showed labelling by a fluorescent glibencamide derivative [197] leading them to surmise that the 55-kDa protein represented the channel forming component, but no sequence data were presented to identify either component. Furthermore, because only Coomassie blue staining of proteins was used, it was not established that there were not other minor components present in the partially purified preparation used and the high activity of channel proteins would only require a small contamination to produce significant channel activity. Indeed, it should be noted that mitochondrial preparations used in this study were unpurified on density gradient and therefore unavoidably would have contained significant contamination by fragments of plasma membrane and thus plasmalemmal KATP channels. This might explain why Garlid and co-workers found 6 to 7 times larger amounts of KATP channel in their preparations of isolated brain mitochondria than reported in liver or heart mitochondria. In view of the problems associated with the reconstitution and patch-clamping studies discussed above, and the absence of a molecular identity for the mitochondrial KATP channel, the best evidence for the activity of mitochondrial KATP channels should come from the measurement of changes in matrix volume of isolated mitochondria in response to channel openers and blockers. In this context, Garlid's laboratory has performed extensive experiments using light scattering to detect matrix volume changes in response to adenine nucleotides and a variety of KATP channel openers and blockers. However, as noted above (Section 8.2), we believe that without direct measurements of matrix volume these data must be treated with caution. Thus, although data from both this laboratory and that of Brustovetsky confirmed effects of adenine nucleotides on light scattering that are consistent with the presence of a KATP channel, neither laboratory was able to detect volume changes isotopically despite demonstrating with valinomycin that the technique was quite sensitive enough to do so should they occur [170,171]. Furthermore, neither we nor Brustovetsky et al. were able to detect any significant changes in light scattering with a variety of KATP channel openers (diazoxide, nicorandil, cromakalim, pinacidil, minoxidil) or blockers (5HD and glibencamide) [170,171]. Our own data have led us to conclude that if mitochondrial KATP channels exist, there is no strong evidence for them being active in isolated mitochondria. However, this does not preclude such channels being active within the intact cell, since we might be losing a factor during mitochondrial isolation that is required for their activity. We will consider if there is any evidence for this in the context of studies investigating the role of mitochondrial KATP channels in preconditioning (Section 9.1). 8.4 Evidence for a mitochondrial calcium-activated K+ channel Our own laboratory was the first to suggest that calcium might regulate potassium entry into mitochondria and hence matrix volume (see [182,198]). We demonstrated that K+ entry into liver mitochondria was stimulated by hormones that activated gluconeogenesis, such as glucagon, adrenaline and vasopressin, and that this led to an increase in matrix volume that was important for the stimulation of mitochondrial respiration and ATP production [182,198]. The mechanism was shown to involve a rise in matrix calcium that inhibits pyrophosphatase leading to a rise in matrix pyroposphate (PPi). Indeed, elevating matrix [PPi] independently of Ca2+, for example by provision of butyrate, also caused modest mitochondrial swelling [186]. Since PPi was known to bind to the ANT [199–201] and adenine nucleotide depletion had been reported to increase mitochondrial K+ uptake [202,203], we proposed that the ANT could act as a potassium channel when ATP was displaced by PPi [186]. In support of this we confirmed earlier data [188] that the permeability of the inner membrane to K+ was increased in the presence of atractyloside which inhibits the ANT by trapping it in the “c” conformation [186]. This mechanism has recently received support from studies on brain mitochondria where it was shown that carboxyatractyloside could enhance potassium loss from de-energised mitochondria [171]. Our own data were unable to demonstrate calcium mediated increase in either matrix [PPi] or volume in isolated rat heart mitochondria [180,204]. However, data from O'Rourke's laboratory [174] have suggested the presence of a calcium-activated potassium channel (KCa) in these mitochondria. These workers employed patch-clamped mitoplasts to characterise mitochondrial KCa channels that, like the sarcolemmal KCa channels were blocked with charybdotoxin. Furthermore, antibodies against sarcolemmal KCa channels detected immunoreactive proteins in the mitoplasts. One problem with these data is that, as noted above, it is well established that isolated mitochondria are contaminated with plasma membrane and thus it is possible that the immunological and patch clamping data were actually detecting the sarcolemmal KCa. 9 Is there good evidence for mitochondrial K+ channels being involved in the protective mechanism of preconditioning? 9.1 The use of pharmacological agents is hampered by lack of specificity There are numerous studies reporting protection of hearts from ischaemia reperfusion injury by drugs that are claimed to be specific openers of the mitochondrial putative mitochondrial KATP channel, most notably diazoxide, and the prevention of this effect by supposed specific blockers of the channel, in particular 5HD (see [11,12,161–163]). However, when haemodynamic performance of the heart is used to assess recovery after ischaemia/reperfusion, the effect of 5HD to reverse the protective effects of diazoxide is not universally observed [205,206]. O'Rourke et al. have suggested a role for the putative mitochondrial KCa channels in preconditioning, in addition to mitochondrial KATP channels. This proposal was based on the observation that the KCa agonist NS-1619 decreased the infarct size upon reperfusion after ischaemia, and this was blocked by paxilline, a KCa blocker [13,174]. However, as noted above (Section 8.3), a major problem associated with such studies is the assumption that the pharmacological agents used are only exerting their effects on the mitochondrial KATP or KCa channels. In the absence of independent verification of this, the conclusions drawn must be treated with caution. This is especially so in the light of increasing evidence that these agents have other non-specific effects on mitochondria including inhibition of the respiratory chain and uncoupling as is discussed below. Diazoxide has been shown by many laboratories including our own to inhibit succinate dehydrogenase at concentrations usually used (50 μM) to open the mitochondrial KATP channel [169,206–209]. Because inhibiting succinate dehydrogenase will also block the citric acid cycle in an intact cell, this may well perturb cellular energy metabolism leading to mild depolarisation of the mitochondria and flavoprotein oxidation. It has also been reported that both diazoxide and pinacidil can uncouple mitochondria although the concentrations (> 50 μM) required to produce a significant depolarisation are at the high end of the spectrum used in in vivo studies [210,211]. Interestingly these effects appear to require an interaction with the ANT [210] and modelling studies based around the published structure of the ANT have shown potential drug binding sites on this membrane protein [212]. The specificity of mitochondrial KATP blockers is also doubtful, especially for the most commonly used 5 HD which is a substituted fatty acid and has the potential to be metabolised. Indeed, our own work from and that of others has confirmed that 5 HD can be activated to its CoA derivative and then further metabolised through the β-oxidation pathway [206,209,213]. As such it can act as a poor respiratory substrate. In addition, the slow kinetics of 3,5-dihydroxydecanoyl-CoA oxidation by l-3-hydroxyacyl-CoA dehydrogenase, the penultimate step of the pathway, can produce a bottleneck that can inhibit normal fatty acid oxidation [206,214]. Interestingly, the other commonly used mitochondrial KATP channel blocker, glibencamide, also has the potential to decrease fatty acid oxidation by inhibition of palmitoyl carnitine transferase [215–217]. Overall, the specificity of the mitochondrial KATP channel openers and blockers is poor and the ability of several of the KATP channel openers used to act as uncouplers or inhibitors of respiration is especially worrying. This is because it is well established that bona fide respiratory chain inhibitors [31,218–221], succinate dehydrogenase inhibitors [222] and uncouplers [223,224] are all cardio-protective when applied before or during ischaemia even though they have no effects on mitochondrial KATP channels. Thus, in order to establish a role for mitochondrial potassium channels in ischaemic preconditioning more direct evidence for their opening during IP must be provided. 9.2 Direct measurements of mitochondrial KATP channel activation during preconditioning Two approaches can be taken to investigate whether mitochondrial K+ channel opening is occurring during preconditioning. The first is the determination of matrix volume in situ and the second is measurement of other effects that an increase in electrogenic potassium flux into the matrix might have on mitochondrial function within the cell. 9.3 Measuring matrix volume in situ Although measurement of mitochondrial volume in situ in the perfused heart is not possible we have determined matrix volume in mitochondria isolated very rapidly from the perfused heart [206]. Our data showed a significant increase in matrix volume induced by both IP (35%) and by 50 μM diazoxide (50%) which is consistent with the opening of a mitochondrial KATP channel. However, 100 μM 5HD, the putative mitochondrial KATP channel blocker, also induced an increase in matrix volume (50%) and failed to prevent the increase induced by IP. Indeed, we found no correlation between matrix volume (pre-ischaemic, end ischaemic or during reperfusion) and the haemodynamic recovery of the heart [206]. Sollott et al. [97] have used a sophisticated Fourier analysis of transmitted optics linescan imaging with a 633-nm laser in isolated cardiac myocytes to estimate changes in mitochondrial diameter from which they calculate changes in matrix volume. They report a 2.5–4% increase in matrix volume with 30 μM diazoxide and 50 μM pinacidil consistent with opening of mitochondrial KATP channels. However, their approach makes the assumption that changes in diameter necessarily mean a change in matrix volume, but as noted above (Section 8.2), this is not necessarily the case. Furthermore, Sollott et al. report that a wide variety of other agents known to protect hearts from reperfusion injury through quite distinct mechanisms, including cyclosporin A, the sarcolemmal sodium proton antiport inhibitor Hoe694 and ligands of G protein–coupled receptors such as bradykinin and Tyr-D-Ala-Gly-Phe-D-Leu also cause similar increases in matrix volume [97]. They propose that such an increase in matrix volume is critical for providing a memory that allows protection even after removal of the primary stimulus. Their explanation as to how this is achieved is that swelling increases the threshold at which ROS induce the permeability transition. However, why this should be the case is unclear since measurements on isolated mitochondria have demonstrated that MPTP opening is enhanced at higher matrix volume [225]. Overall, the published data on measurements of mitochondrial volume in situ provide no strong evidence either for or against the opening of mitochondrial K+ channels. 9.4 Measuring flavoprotein oxidation Another approach used to investigate the activity of mitochondrial K+ channels in cardiac myocytes is based on the reasoning that if a K+ channel is opened the entry of K+ would cause depolarisation of the mitochondrial membrane potential. Marban and colleagues reported that with diazoxide this can be detected indirectly in rabbit cardiac myocytes by an oxidation of mitochondrial flavoproteins that is reversed by 5HD [226]. Similar effects have been reported in guinea pig myocytes using NS1619 to open the putative mitochondrial KCa channel with reversal by paxilline [227]. However, other workers have failed to reproduce the effects of diazoxide and 5HD in guinea pig and rat cardiac myocytes leading to the suggestion that the effects observed may reflect substrate deprivation [209,228]. In addition, there are theoretical reasons to question whether opening of the mitochondrial KATP channel would cause a significant depolarisation. Indeed, Garlid has argued convincingly that the fluxes of K+ predicted to occur if the channel opens would allow minimal perturbation of the membrane potential even when giving a significant increase in matrix volume [169,229]. In fact, such an increase in matrix volume might even increase membrane potential as a result of stimulating the respiratory chain itself [182]. Direct measurements of membrane potential with the fluorescent dyes TMRE or JC-1 failed to detect either an increase or decrease [228,230]. There are, however, other mechanisms by which changes in mitochondrial membrane potential and thus flavoprotein oxidation could be induced by putative mitochondrial K+ channel openers and blockers that reflect other non-specific sites of their action. For example, in substrate deprived myocytes where diazoxide has been shown to cause flavoprotein oxidation, this could be caused by the drug inhibiting succinate dehydrogenase and thus the citric acid cycle whilst the ability of 5 HD to overcome this effect might be the result of the drug acting as a respiratory substrate as noted above (Section 9.1 and [12]). 10 Proposed mechanisms by which opening of mitochondrial K+ channels might protect hearts from ischaemia reperfusion injury? For mitochondrial K+ channels to play a role in preconditioning there would have to be a mechanism by which their opening might be linked to improved recovery. Several possible mechanisms have been proposed. 10.1 Enhanced mitochondrial ATP production It is known that an increase in matrix volume can stimulate respiration and oxidative phosphorylation [182,198] and so might improve ATP production during reperfusion leading to enhanced haemodynamic recovery [161,231]. Although we were able to measure an increase in mitochondrial volume and rates of ADP-stimulated respiration in mitochondria isolated from hearts immediately following ischaemic preconditioning or diazoxide treatment, the effects were largely lost during reperfusion when they are predicted to be most important [206]. Furthermore, 5HD-treatment was shown to exhibit a similar increase in matrix volume but was not cardioprotective [206]. Garlid and colleagues have proposed an alternative, but not unrelated mechanism in which it is the maintenance of the intermembrane space and the close association between the outer an inner membrane that is the critical factor resulting from KATP channel opening [11]. They argue that during ischaemia the loss of membrane potential would decrease electrogenic K+ entry into mitochondria and that this would cause contraction of the matrix. It is proposed that that this would disrupt the interaction between VDAC in the outer membrane, octomeric creatine kinase (CrK) in the intermembrane space and the ANT in the inner membrane leading to faster permeation of ATP through VDAC into the intermembrane space. Further translocation of this ATP into the matrix through the ANT would lead to its hydrolysis, that would be detrimental to the heart. It is suggested that opening of the mitochondrial KATP channel would prevent this matrix condensation and so decrease ATP breakdown during ischaemia leading to less damage. There is some evidence that ATP decline during ischaemia is slowed by ischaemic preconditioning or diazoxide treatment, although the mechanism originally proposed was through inhibition of the F1Fo ATPase by enhanced binding of the ATPase inhibitor protein [232,233]. Furthermore, other data report the opposite effect of preconditioning, with the decline in ATP and the development of contracture being faster than in control hearts [232–235]. Garlid and colleagues further propose that during reperfusion, keeping the mitochondrial KATP channel open maintains the VDAC, CrK, ANT complex which is vital to export ATP rapidly from the mitochondria to the cytosol where it is used to drive contraction. Although this may be true, our own data showed that the slight increases in matrix volume of diazoxide-treated and ischaemic preconditioned hearts compared to control hearts was not significant at reperfusion [206]. The more probable determinant of the efficiency of myocardial ATP production during reperfusion is the extent to which the mitochondria are damaged. 10.2 Mild uncoupling leads to less calcium overload and ROS production It has been proposed that opening of K+ channels would depolarise the mitochondria sufficiently during reperfusion to reduce ROS production and calcium accumulation and hence prevent MPTP opening [12,151,152,162,163,227,236]. However, as noted above, the amount of depolarisation predicted from K+ channel opening is unlikely to cause much if any depolarisation [169,229] and this is confirmed by direct measurements of mitochondrial membrane potential that failed to detect any depolarisation [228,230]. Furthermore, should significant depolarisation occur at reperfusion, ATP synthesis would be compromised leading to poorer rather than better recovery of the heart. Yet NMR spectroscopy measurements have shown that the bioenergetic state of the preconditioned heart improves during reperfusion, consistent with the better haemodynamic function [232,237,238]. Nevertheless, it is not possible to rule out the possibility that opening of mitoKATP channels might cause a minor depolarisation during reperfusion, or slightly reduce the repolarisation, sufficient to reduce ROS formation without having a significant effect on ATP production. Although significant uncoupling during the reperfusion phase seems unlikely to be protective, it is known that adding low doses of uncoupler prior to ischaemia can precondition hearts [223,224]. Indeed, any interference with oxidative phosphorylation during the preischaemic phase seems able to exert a similar protective effect whether it is brought about by a brief ischaemic episode as in IP, by addition of a respiratory chain inhibitor [218–221] or a succinate dehydrogenase inhibitor [222]. Since many of the putative mitochondrial K+ channel openers have also been shown to have direct uncoupling effects [209–211,239] or to inhibit components of the respiratory chain including succinate dehydrogenase [169,206–209,240], this provides a common mode of action for these agents. How this might be translated into a protective effect at reperfusion is not clear, although a signalling pathway involving AMP-activated protein kinase provides one possibility as discussed above (Section 6.4). Another potential mechanism would be through increased levels of ROS production causing PKC activation and respiratory chain inhibitors such as antimycin are known to increase ROS production which is critical in mediating its protective effects [218,241]. It has also been reported that low doses of uncoupler can increase ROS production in isolated myocytes [224], although how this might occur is unclear since in isolated mitochondria it is well established that even very modest uncoupling greatly decreases ROS production [242]. In the case of antimycin, PKCε translocation has been demonstrated and its protective effects shown to be abolished in PKCε knockout mice supporting a role for ROS and PKCε activation in its protective mechanism [241]. 10.3 Production of ROS as a signalling mechanism Some groups have shown that diazoxide and other putative mitochondrial potassium channel openers cause an increase in ROS production by isolated heart mitochondria, cardiac myocytes and perfused hearts and that this can be blocked by 5HD [240,243–248], but this has not been reproduced in other studies [249,250] or in our own unpublished experiments. Since the protective effects of diazoxide and nicorandil can be overcome by free radical scavengers such as N-(2-mercaptopropionyl) glycine [12,100,240,245] it has been concluded by many workers that opening of the KATP channel causes ROS production and that this signals preconditioning through the pathways discussed above. Attractive though this hypothesis may be, it is difficult to formulate a mechanism by which this might occur. As noted above, even if there were modest uncoupling, this would be predicted to decrease ROS production rather than increase it. Another possible mechanism would be via an increase in matrix volume causing a stimulation of electron flow from complex 1 into complex 3 [182,251]. Indeed, there are reports that low doses of valinomycin can increase ROS formation in isolated mitochondria [248] and myocytes [11,163,247], whilst pretreatment of the perfused rabbit heart with valinomycin substantially decreased the infarct size on reperfusion [252]. However, in our own unpublished experiments (H. Buckman and A.P. Halestrap, unpublished data) we have used Amplex Red to measured ROS production by isolated heart mitochondria incubated with a variety of respiratory substrates in State 4 and were unable to detect an increase in ROS production with any mitochondrial potassium channel opener tested, or when matrix volume was increased by decreasing the osmolarity of the incubation medium. Indeed, in the latter case we actually detected a decrease ROS production. This is consistent with the major locus of ROS production being at a highly reduced site on complex 1 that is oxidised by an increase in matrix volume as electron flow out of complex 1 and into complex 3 is stimulated [182,251]. Another consequence of opening a mitochondrial K+ channel would be an increase in the mitochondrial pH gradient coincident with a decrease in membrane potential [253]. Whereas the latter would be predicted to decrease ROS formation as outlined above, there is evidence to suggest than an increase pH gradient or matrix pH can increase ROS formation from complex 1 [254]. However, in view of the ability of many of the mitochondrial KATP channel openers to inhibit components of the respiratory chain [169,206–209,240], its is perhaps more likely that their ability to increase ROS is through a direct interaction with a redox centre in one of the respiratory complexes. 11 Other proposed mechanism for inhibiting the permeability transition pore in preconditioning 11.1 The role of connexin 43 Recently, it has been proposed that connexion 43 (Cx43) may play a critical role in linking preconditioning to the mitochondria [255–258]. Cx43 is the major protein of the gap junction with hexameric assemblies forming connexons on adjacent cardiomyocytes. Its traditional role is to provide a route for intercellular communication through the propagation of action potential, signalling molecules and metabolites [258–260]. The permeability of the connexons can be regulated through many mechanisms including changes in [Ca2+], pH and phosphorylation by several protein kinases [258,259], including PKCε and PKG, both of which have been implicated in IP as noted above (see Section 6). Cx43 is normally partially phosphorylated with low conductance, but progressive dephosphorylation occurs during ischemia causing increased conductance. It has been proposed that this is important in propagating injury from one cell to another, although hearts from Cx43-deficient mice showed no difference in their sensitivity to reperfusion injury [261]. Studies in both pig and rat hearts demonstrated that the dephosphorylation of Cx43 and electrical uncoupling during ischaemia were prevented by IP in a PKC-dependent manner and that IP increased the co-localization of several protein kinases with Cx43 during ischemia [262,263]. Furthermore, hearts from Cx43-deficient mice showed no protection from reperfusion injury by either IP [261] or diazoxide [247] and in the latter case the ability of diazoxide to increase ROS was also abolished. These data imply a link between Cx43 and preconditioning and thus that Cx43 can in some way influence the opening of the MPTP, perhaps by a signalling mechanism involving ROS. Recent work has suggested that this may be through a translocation of Cx43 to the mitochondria following preconditioning [255–257]. A small fraction of Cx43, primarily in its phosphorylated form, was detected in purified mitochondria from rat, mouse, pig and human hearts and this was increased following preconditioning. In these studies, considerable efforts were taken to show that this Cx43 was not just the result of sarcolemmal contamination of the mitochondrial fraction and further evidence for this was provided by immunofluorescence confocal microscopy and immuno-gold electron microscopy. Furthermore, sub-fractionation of mitochondrial compartments led the authors to conclude that Cx43 is translocated to the inner mitochondrial membrane, although some Cx43 may remain associated with the outer membrane [257]. Although we have been able to detect Cx43 in purified mitochondria, and in preliminary experiments show that this was increased following preconditioning, we have expressed reservations about the likelihood of Cx43 translocation to the inner mitochondrial membrane [256]. Furthermore, following more extensive experiments in which we determined the extent of sarcolemmal contamination of mitochondria by measuring the presence of MCT1, we were unable to confirm any consistent increase in the Cx43 associating with the mitochondrial fraction following IP (Fig. 4). Thus the relevance of Cx43 translocation to the mitochondria remains uncertain, although the loss of preconditioning in Cx43-deficient hearts from mice argues strongly for some pool of Cx43 playing a critical role [247,261]. Since myocytes from the Cx43-deficient mice did not show any diazoxide-induced ROS formation [247], this is likely to involve a ROS-mediated signalling pathway, although the mechanism linking Cx43 to ROS production is unclear. From their data, Garcia-Dorado and colleagues propose that recruitment of Cx43 to the mitochondria in IP might be required for opening of the mitochondrial KATP channel leading to ROS production that triggers preconditioning [255,257]. However, when they prevented this recruitment with geldanamycin, a blocker of heat shock protein 90-dependent protein translocation across the outer mitochondrial membrane, the cardioprotective effects of IP were not lost [257]. Furthermore the cardioprotective effects of both diazoxide and isoprenaline were not accompanied by translocation of Cx43 to the mitochondria [257]. Thus the balance of evidence suggests that Cx43 translocation to mitochondria is not essential for preconditioning [256]. 11.2 Transient MPTP opening Hausenloy and colleagues have reported that when the MPTP inhibitors CsA or SfA are present during IP or preconditioning induced by uncoupler and diazoxide, cardioprotection is abolished [121]. They conclude that transient opening of the MPTP may play a critical role in triggering preconditioning and that this might be caused by the well-documented rise in ROS that occurs during preconditioning (see Section 6.1). However, we have challenged these conclusions both on theoretical and experimental grounds [264]. First, our own data using the mitochondrial 2-deoxyglucose entrapment technique failed to detect any increase in MPTP opening immediately following the preconditioning stimulus [68,75]. Second, although the authors suggest that transient MPTP opening during the preconditioning phase prevents mitochondrial calcium overload and that this mediates protection, it is the calcium overload at the end of ischaemia and during reperfusion that mediates the MPTP opening responsible for reperfusion injury. Third, the authors argue that uncoupler also preconditions hearts by opening the MPTP, but uncouplers will depolarize mitochondria independently of MPTP opening. Thus any protection mediated by uncoupler that involves a decrease in mitochondrial calcium loading should be insensitive to both SfA and CsA, but this is not what the authors observed [121]. Hence it seems probable that CsA and SfA are overcoming IP independently of any effects on MPTP opening. This might involve roles for cyclophilins in the preconditioning signalling pathway similar to those responsible for the immunosuppressive actions of these drugs [264]. 12 Conclusions and future directions The ability of preconditioning (and post-conditioning) to inhibit the opening of the MPTP during reperfusion and so provide cardioprotection seems to be well established. What is less clear is the mechanism(s) by which this is achieved. We end this review by presenting our own views on how preconditioning is most likely to be mediated. These are summarised schematically in Fig. 5. The observation that mitochondria isolated immediately after the preconditioning protocol show no protection from MPTP opening [85,100] argues against a signalling pathway that causes phosphorylation of some regulatory component of the MPTP as an early event. Nor have we been able to observe any consistent migration of protein kinases to the mitochondria or changes in phosphorylation of a mitochondrial protein (see Sections 6 and 7.1). By contrast, mitochondria isolated at the end of ischaemia or during reperfusion do show a decreased sensitivity to calcium-mediated MPTP opening [99–101]. Here too we were unable to detect consistent migration of protein kinases to the mitochondria or changes in phosphorylation of a mitochondrial protein, arguing against this protection being mediated by phosphorylation (see Section 7.1). However, the changes in sensitivity of the MPTP do correlate with the extent of protein carbonylation, a surrogate indicator of the oxidative stress the mitochondria have experienced [100,101]. Since it is well established that oxidative stress sensitises the MPTP to calcium, this would seem to provide an adequate explanation as to how preconditioning inhibits MPTP opening at reperfusion (see Section 7.2). Furthermore, this is entirely consistent with the decrease in ROS production late in ischaemia and during reperfusion that has been observed directly in preconditioned hearts [27,265]. If this explanation is correct then the focus of research should shift towards investigating the mechanisms underlying the decreased levels of ROS seen in preconditioned hearts at the end of ischaemia and during reperfusion. This will involve consideration of both the sources of ROS, which may include mitochondria, NAD(P)H oxidase or xanthine oxidase, and the pathways that remove them including superoxide dismutase, glutathione peroxidase and catalase [26,28]. The many signalling pathways identified as potential mediators of preconditioning, including ROS, PKCε and PKG (see Section 6) may converge on one or more of these steps to decrease ROS levels and hence reduce activation of the MPTP by thiol oxidation. The pharmacological agents used to implicate these pathways may interact specifically with their proposed target, but they may also have less specific effects such as on mitochondrial respiratory chain components (see Section 9.1) to produce a modest increase in ROS levels that stimulate signalling pathways rather than cause damage. Indeed, it does seem that any stress to the myocyte's energy status, be it through respiratory chain inhibition, uncoupling or ischaemia, causes preconditioning. This may also implicate the energy sensing protein kinase, AMP-activated protein kinase (AMPK) [133,266], in the signalling pathway of preconditioning, and there are data supporting such a role for AMPK in both the heart [101,135] and the liver [267]. In addition to the protection seen early in reperfusion that could be explained through the mechanisms described above, there is good evidence for ongoing protection against MPTP opening being important as reperfusion progresses. Thus inhibition of MPTP opening with CsA or SfA provides effective cardioprotection, as measured by decreased infarct size, even if the drug is not present at the start of reperfusion, provided it is added within the first 15 min [90,91]. Furthermore, there is a considerable body of evidence to suggest that activation of survival kinase pathways during reperfusion is critical for protection during this phase of the insult, and that these may converge on Akt and GSK3 to mediate their effects [97,107,126]. It is possible to bring these observations together by invoking the established phenomenon of MPTP-induced ROS production [268,269]. Once MPTP opening has occurred during the initial phase of reperfusion, the resulting ROS production would lead to a progressive increase in pore opening in other mitochondria and thus an escalating number of necrotic cells as reflected in the ongoing LDH release and increased infarct size. CsA or SfA will prevent this ongoing opening, but so too might the survival kinase cascades by reducing the ROS production or increasing their removal as described below. It is well established that MPTP opening causes mitochondrial swelling and rupture of the outer mitochondrial membrane [44]. This leads to release of pro-apoptotic factors including cytochrome c, that activates caspase 9 and hence caspase 3 [45]. In addition, it has been shown that the pro-apoptotic factor Bax can translocate to the mitochondria during ischaemia [72] and this, in conjunction with cleaved Bid (tBid), might cause the cytochrome c release observed during prolonged ischaemia [70] despite there being no evidence of MPTP opening [68,98]. The resulting loss of cytochrome c will slow electron transfer out of complex 3 and thus potentially cause increased ROS production in either complex 3 or in complex 1. An additional mechanism that may lead to increased ROS production during ischaemia and reperfusion is through caspase 3 mediated cleavage of the p75 component of complex 1 [270,271]. Protection from the rise in ROS that accompanies such cytochrome c release could be mediated by survival kinase cascades in two ways. They might stimulate ROS removal as described above for the ischaemic phase or they could reduce the Bax-induced cytochrome c release. Indeed, it is well established that survival kinases can block apoptosis by inhibiting cytochrome c release [272]. This is brought about by Akt-mediated phosphorylation of the pro-apoptotic Bcl-2 family member Bad [273,274] and, via GSK3 phosphorylation and stabilisation of the anti-apoptotic Bcl-2 family member Mcl-1 [275]. Note added in proof Since submission of this article it has been reported that the properties of the mitochondrial permeability transition pore in mitochondria devoid of all VDAC isoforms are the same as in mitochondria from wild-type mitochondria (C.P. Baines, R.A. Kaiser, T. Sheiko, W.J. Craigen, J.D. Molkentin, Voltage-dependent anion channels are dispensable for mitochondrial-dependent cell death, Nat Cell Biol. 9 (2007) 550–555). This confirms that VDAC is not an essential component of the MPTP.	[ "ros", "ischaemia", "reperfusion", "mitochondrial permeability transition pore", "calcium", "pkc", "katp channel", "5hd, 5-hydroxydecanoate", "ampk, amp activated protein kinase", "ant, adenine nucleotide translocase", "apd, action potential duration", "bcdh, branched chain 2-oxoacid dehydrogenase", "crk, creatine kinase", "csa, cyclosporin a", "cyp, cyclophilin", "cx43, connexin43", "gsk3, glycogen synthase kinase 3", "ip, ischaemic preconditioning", "katp, atp-dependent potassium channels", "mitokatp, mitochondrial atp-dependent potassium channels", "mct1, monocarboxylate transporter 1", "mptp, mitochondrial permeability transition pore", "pdh, pyruvate dehydrogenase", "pdk1, phosphoinositide-dependent kinase 1", "pi-3-kinase, phosphatidyl inositol 3 phosphate kinase", "pkc, protein kinase c", "pkg, cyclic gmp-dependent protein kinase", "ppi, pyrophosphate", "ppiase, peptidyl-prolyl cis-trans isomerase", "pten, phosphatase and tensin homolog", "ros, reactive oxygen species", "sfa, sanglifehrin a", "sur, sulphohylurea receptor", "vdac, voltage activated anion channel" ]	[ "P", "P", "P", "P", "P", "P", "P", "R", "R", "R", "R", "R", "R", "R", "M", "M", "R", "R", "R", "R", "R", "R", "R", "R", "M", "R", "M", "M", "M", "R", "R", "R", "R", "R" ]
Psychopharmacologia-2-2-1705539	A comparison of nicotine dose estimates in smokers between filter analysis, salivary cotinine, and urinary excretion of nicotine metabolites	Rationale Nicotine uptake during smoking was estimated by either analyzing the metabolites of nicotine in various body fluids or by analyzing filters from smoked cigarettes. However, no comparison of the filter analysis method with body fluid analysis methods has been published. Introduction A number of studies were published over the last 30 years that attempted to determine the amount of tar and/or nicotine that smokers receive from their cigarettes (reviewed in Stephen et al. 1989; Pritchard and Robinson 1996; Scherer 1999). The methodology used falls into three broad categories: (1) the analysis of biomarkers in human body fluids or expired breath; (2) the measurement of smoking behavior (puff volume, duration, and frequency) followed by a smoking machine set to duplicate human puffing conditions; and (3) the analysis of spent cigarette filters and the calculation of smoke yields from the filter efficiency. Typically, biomarker measurements in blood/plasma and smoking behavior measurements require that sampling or measurements be made in a laboratory environment. There is a possibility that smoking behavior becomes atypical in this type of environment (Comer and Creighton 1978; Ossip-Klein et al. 1983). Urine, saliva, and spent cigarette filters can be collected in a smoker’s everyday environment. The analysis of 24-h urine samples for nicotine and major metabolites can provide quantitative data regarding uptake of smoke constituents as the product is used in a smoker’s everyday environment (see Byrd et al. 1998 and references therein for examples). Subject compliance can be an issue when trying to determine actual cigarette yields because in an unmonitored environment the subject must be relied upon to smoke only a given brand, not use any other nicotine-containing products, provide an exact accounting of every cigarette smoked during the collection period, and collect all urinary output. The filter analysis method is one of the least invasive of the methods mentioned above. The smokers can use their product in their normal environment and the only deviation from normal behavior is to save the filters. Compliance is not an issue when trying to determine the subject’s cigarette yield because the filters from the actual cigarettes smoked are analyzed. In most cases, the returned filter can be compared to those of the subject’s stated brand to assure brand compliance. The primary issue with this method has been that smoking behavior can produce changes in the filtration efficiency of the filter. The filtration efficiency can vary according to the velocity of the smoke passing through the filter and, to some extent, the length of the tobacco rod smoked (Overton 1973; Dwyer and Abel 1986; Norman et al. 1984). The method used for this study was developed to minimize the effects of smoking behavior on filter efficiency by measuring only the portion of the filter downstream of the ventilation holes (i.e., the mouth end) (St.Charles 2001; Shepperd et al. 2006). This results in relatively constant filtration efficiency over a wide range of smoking behavior. The objective of this study was to compare the nicotine yield of human-smoked cigarettes (mouth exposure) as measured by the filter analysis method and human smoke uptake as measured by biomonitoring under strictly controlled conditions. The biomonitoring measurements included salivary cotinine and 24-h urinary nicotine, cotinine, 3′-hydroxycotinine (3-HC), and their respective glucuronide conjugates. With a good correlation between the methods, future studies on smoker exposure can use the simpler filter analysis method rather than resorting to human biomonitoring techniques. Materials and methods Study design The clinical portion of this study was conducted by an independent contract research organization in 2003 (Covance Clinical Research Unit, Madison, WI, USA). The analysis of salivary cotinine and urinary nicotine metabolites was performed at Covance Laboratories (Harrogate, North Yorkshire, UK). Filter analysis was performed by the study sponsor (Research and Development, Brown & Williamson Tobacco Company, Macon, GA, USA). The study was approved by Covance’s Institutional Review Board and performed in accordance with applicable federal regulations. Subjects who participated in the study gave their informed consent, were told of the purpose of the study, and could withdraw at any time. Subject selection Habitual smokers were recruited by Covance. Enrollment criteria included males or nonpregnant, nonlactating females, between 21 and 65 years of age, within −20% to +30% of their ideal body weight, who smoked at least 15 cigarettes a day of the same cigarette brand during the previous year. Subjects were excluded if they were under 21 years of age, were pregnant or lactating, participated in any other clinical study within 30 days before study entry, had a history or showed signs of a significant medical or psychiatric condition, used prescription medications within 14 days before study entry, had a history of alcoholism or drug addiction within a year of study entry, or used alcohol or any nonprescription preparations within 72 h of study entry. A few subjects deviated from the enrollment criteria: underweight (1), overweight (5), medication or alcohol usage before study entry (4), low cigarette consumption (2), shortened brand loyalty duration (4), elevated clinical chemistry (2), abdominal/hernia surgery (7), and positive drug screen before study entry (2). Because these deviations were considered minor and not expected to interfere with the study objectives, the subjects were allowed to participate in the study. Subjects were assigned into one of four tar yield groups, which span the range of Federal Trade Commission (FTC) tar yields found in commercially available USA filtered cigarettes: 1–3 mg (ULL or ultralights/low), 5–6 mg (ULH or ultralights/high), 9–12 mg (LTS or lights), and 13–19 mg (FF or full flavor). The purpose was to cover a wide range of human nicotine exposure to allow robust correlations between the methodologies. The goal was to enroll 20 smokers per group; however, even with additional recruitment attempts, only 15 smokers enrolled in the ULL group (market share <2%). One subject in the FF group withdrew from the study due to illness. Table 1 summarizes the subjects’ demographics and their respective brand characteristics by tar band. Table 1Subjects’ demographics and cigarette brand characteristics by tar bandSubjectsULLULHLTSFFTotalNumber1520201974Sex (M/F)8/712/810/1014/544/30Height (cm)173 (159–204)174 (157–194)174 (161–187)177 (165–190)175 (157–204)Weight (kg)77 (57–104)75 (42–98)78 (56–98)77 (60–100)77 (42–104)Age35 (21–54)32 (21–47)34 (21–64)32 (21–53)33 (21–64)Menthol smokers335718Brands Menthol/nonmenthol2/53/64/84/613/25 Tar (mg/cig)1.4 (0.7)5.4 (0.5)10.1 (0.9)15.0 (1.7)8.3 (5.1) Nicotine (mg/cig)0.15 (0.05)0.46 (0.04)0.77 (0.06)1.07 (0.14)0.64 (0.34) CO (mg/cig)1.9 (0.9)6.8 (0.6)11.6 (1.4)13.9 (2.0)8.9 (4.6) Puffs/cig7.6 (0.7)8.0 (1.1)7.9 (1.0)8.2 (0.9)7.9 (1.0)Height, weight, and age are shown as the mean (range). Cigarette yields are shown as the mean (SD) for the FTC smoking method. Subjects were confined to the clinic for six calendar days to give five consecutive 24-h periods. Nine confinement periods were staggered and limited to ten subjects or less (generally of the same tar range group). During confinement period number 8, one LTS smoker and two FF smokers were allowed to participate with six ULL smokers to complete the LTS and FF cells. Subjects were fed a standardized bulk diet that excluded grilled, smoked, or barbecued food items. Consumption of water and other nonalcoholic beverages was unrestricted. Subjects refrained from strenuous exercise. During confinement, subjects smoked their usual cigarette brand ad libitum in a dedicated smoking room equipped with ventilation and air filtration. Use of any form of nicotine other than the subject’s declared own brand was prohibited. Cigarettes were purchased locally. Urine and saliva collection and analysis Twenty-four-hour urine samples were collected from each subject for five consecutive days. Collections started at approximately 0800 hours (first void excluded) and ended at approximately 0800 hours the following day (first void included). Urine was collected in 3-l amber plastic containers and kept refrigerated throughout the collection period. No chemical preservatives were used. After each 24-h sample collection, volume and pH measurements were recorded, a sample was taken for creatinine analysis, and 2 × 5-ml and 4 × 500-ml aliquots were taken and stored frozen at −70°C until shipped. Aliquots were shipped under dry ice to the analytical laboratory and stored at −70°C until analysis. Urinary nicotine, nicotine-N-β-glucuronide, cotinine, cotinine-N-β-glucuronide, trans-3′-hydroxycotinine, and trans-3′-hydroxycotinine-O-β-glucuronide were analyzed using solid phase extraction (SPE) and liquid chromatography with tandem mass spectrometric detection using a method developed and validated at the analytical laboratory (Analytical Procedure Covance no. 2002-010). Nicotine-d3, cotinine-d3 (Sigma-Aldrich, Poole, UK), and trans-3′-hydroxycotinine-d3 (Toronto Research Chemicals, North York, ON, Canada) were used as internal standards. Nicotine, cotinine (Sigma-Aldrich), trans-3′-hydroxycotinine, and the three corresponding glucuronides (Toronto Research Chemicals) were used as reference standards. Samples were incubated for 18–22 h at 37°C with β-glucuronidase, which enzymatically deconjugated the aglycones from their respective glucuronides. The free and cleaved aglycones and their respective deuterated internal standards were then extracted and analyzed. Saliva samples were collected in sterile Salivette tubes (Sarstedt, Newton, NC, USA) from each subject for five consecutive days at approximately 1830 hours. On day 4, two additional saliva samples were collected at approximately 0830 and 1330 hours. Clinical staff supervised and timed the process to assure compliance. Collected saliva samples were immediately stored at −20°C, shipped under dry ice to the analytical laboratory, and stored at −20°C until analysis. Salivary cotinine was analyzed by a method developed and validated at the analytical laboratory (Analytical Procedure Covance no. HB-02-061) based on a previously reported method (Bentley et al. 1999). Filter collection and analysis To control cigarette brand accessibility and document cigarette consumption, cigarettes were issued by clinical staff individually. The used filter had to be returned before a subject received their next cigarette. Used filters were processed under the supervision of clinical staff by having the subject remove any tobacco particles from the used filter before depositing it in an individually labeled glass container. Used filters were collected for five consecutive days with each day starting/ending at the same time as urine samples. The filters generated daily by each subject were shipped by overnight carrier at ambient temperature to the analytical laboratory. Immediately upon receipt, a 10.0-mm portion of the mouth end (tip) was cut from the filter using a specially designed jig, which kept the cut-length constant and kept a single edge razor blade perpendicular to the filter. Tip-to-tip length variation using the jig was within 3%. The tips were stored in labeled 30-ml glass jars with Teflon-lined lids at −20°C until analyzed. Nicotine yields from the human-smoked cigarettes were estimated by analyzing the tips for nicotine. Five separate machine smoking regimes were used to provide calibration curves for the filter tips of each brand style tested. Cigarettes from the same batches that the subjects smoked were used for calibration smoking. The smoking regimes were chosen to give a wide range of cigarette yields of approximately equal spacing. Table 2 shows the calibration puffing conditions for each tar band. These machine puffing regimes proved to cover the spread of human smoking results for all but 7% of the subject-days tested (5 ULH, 1 LTS, and 6 FF out of 370 total). Six of the 12 involved extrapolating by less than 0.1 mg of nicotine/cigarette and the maximum extrapolation was from 2.3 to 2.6 mg/cigarette. Table 2Calibration smoking conditions by tar bandULLULHLTSFF35/60/T + 3/O35/60/T + 3/O35/60/T + 3/O35/60/T + 3/O35/60/T + 3/B35/60/T + 3/B35/60/T + 3/B35/60/3 P/O70/60/T + 8/O35/60/4 P/O35/60/3 P/O35/60/6 P/O70/60/T + 8/B70/60/T + 8/B70/60/T + 3/O70/60/T + 3/O70/40/T + 3/B70/40/T + 3/B70/40/T + 3/B70/40/T + 3/OData are presented as puff volume (ml)/puff interval (s)/length smoked/ventilation holesO Open, B blockedT + # indicates overtipping plus distance (mm); # P indicates number of puffs. Calibration equations were calculated using a linear regression of nicotine yield as a function of tip nicotine. These equations were then used to estimate human-smoked cigarette yield from the measured tip nicotine. This method was validated in a separate study using duplicated human puffing profiles (Shepperd et al. 2006). Aging tests by the study sponsor have shown that tip nicotine values were constant when whole filters were stored in glass jars at ambient temperature for up to 31 days. For this study, tips were cut from the whole filters within 5 days. One to nine tips per sample were extracted using either 20 or 40 ml of methanol containing 0.038 mg/ml decanol internal standard. Tips were extracted for 40 min using a flatbed orbital shaker at 200 rpm. All available tips were divided into at least three separate batches, which were extracted and analyzed on different days to average analytical variation. For samples from days 2, 4, and 5, extract absorbance at 310 nm was also measured to audit the nicotine analysis using an independent method. Absorbance gives a measure of the tar deposited on the filter (Sloan and Curran 1981; Shepperd et al. 2006) and we have found that absorbance per tip correlates linearly with nicotine per tip. For days 1 and 3, replicate vials of the extract were stored at −20°C and subsequently analyzed for nicotine to audit the nicotine analysis further. The root mean square difference between the nicotine replicates was 5.5 μg/ml (7.3%). A backup set of replicate vials was kept for use when outlying data points were identified by either of the auditing methods. Samples were retested when they were outside of the 95% confidence interval of the absorbance per tip vs nicotine per tip correlation (days 2, 4, and 5) or when nicotine replicates differed by more than 20% (days 1 and 3). The extract was analyzed for nicotine by gas chromatography using an Agilent 5890 Series II with a flame ionization detector and a J&W Scientific 30 m Megabore® 0.53 mm ID, DB-Wax (1.0 μm film) fused silica capillary column (Agilent, Palo Alto, CA, USA). The UV absorbance of the extract was measured using an Ocean Optics (Dunedin, FL, USA) PC2000 spectrometer equipped with a fiber optic dip probe with a 2-mm path length. This gave absorbance values within the linear range of less than 2 without further dilution. Absorbance was measured on the same day the tips were extracted because it was found that extract absorbance declines with even overnight storage. Results Measured urinary concentrations for each metabolite were multiplied by their respective daily urine volumes and converted based on molecular weights to yield recovery results in nicotine equivalents. The sum of the nicotine equivalents for nicotine, cotinine, and 3-HC were calculated for each subject per day to give total daily urinary nicotine equivalents (UNE) in milligrams per day. Nicotine yield per cigarette was calculated from the mean nicotine per tip values for each subject-day and the appropriate calibration equation for the brand smoked. This was multiplied by the number of cigarettes smoked by the subject that day to give nicotine yield in milligrams per day. This study design allowed us to achieve a wide range of nicotine exposure suitable for the correlation of the three methods of nicotine uptake estimation. Nicotine yield from the human-smoked cigarettes (mouth exposure) as measured by filter analysis ranged from 3.7 to 67.1 mg/day, UNE ranged from 3.1 to 48.4 mg/day, and saliva cotinine ranged from 70 to 866 ng/ml. The mean (SD) proportions of urinary metabolites, including the glucuronides, were 22% (9%) for nicotine, 35% (6%) for cotinine, and 43% (13%) for 3-HC of the total UNE measured. When expressed as a percentage of nicotine entering the mouth, the proportions were 19% (9%) for nicotine, 31% (11%) for cotinine, 39% (17%) for 3-HC, and 89% (25%) for total nicotine equivalents. Statistical correlations Observed values from the three methodologies were correlated with each other. Figure 1 shows the three plots and correlations for all data points. Figure 1a is a graph of the UNE vs daily nicotine yield estimated from filter analysis; Fig. 1b shows UNE vs saliva cotinine; and Fig. 1c shows daily nicotine yield from filter analysis vs saliva cotinine. Linear regressions were significant for both the slope and intercept (p < 0.001) for all three correlations. The best correlation was obtained with UNE vs daily nicotine yield (Fig. 1a, R2 = 0.66). The slope indicates that 67% of the variation in nicotine mouth exposure calculated from filter analysis appeared as variation in the six urinary compounds measured. The standard error of the regression (SER) was 4.9 mg nicotine/day. Saliva cotinine did not correlate as well with either UNE (Fig. 1b, R2 = 0.49, SER = 5.9 mg/day) or nicotine yield (Fig. 1c, R2 = 0.45, SER = 7.4 mg/day). Because the amount of nicotine entering the mouth was greater than the sum of the UNE, the standard error of the nicotine yield would be greater than that of the UNE due to scaling even if the correlations were equivalent. To allow an equal comparison, the standard error of the percent of the difference between calculated (using the regression equation) and measured values was calculated for the three correlations. The standard error calculated in this manner was 32 and 42% for UNE vs nicotine yield and saliva cotinine, respectively, and 41% for nicotine yield vs saliva cotinine. Summary statistics for all correlations are in Table 3. Fig. 1Correlations of the three methodologies (individual measurements)Table 3Summary of the statistics for linear correlationsCorrelation (y : x) Intercept (95% CI)Slope (95% CI)R2Standard error (%)Urine : filterIndividual4.3 (1.2)0.67 (0.05)0.664.9 (32)mg/day : mg/dayAverage2.7 (2.0)0.74 (0.08)0.833.2 (21)Urine : salivaIndividual5.8 (1.6)0.042 (0.004)0.495.9 (42)mg/day : ng/mlAverage5.4 (3.2)0.043 (0.009)0.555.2 (36)Filter : salivaIndividual6.9 (2.0)0.048 (0.006)0.457.4 (41)mg/day : ng/mlAverage5.6 (4.0)0.052 (0.011)0.546.5 (33)Filter : urine (creatinine normalized)Individual8.7 (1.9)10.8 (1.3)0.427.6 (41)mg/day : mg/day/mmol creatinineAverage7.1 (3.9)12.1 (2.7)0.526.7 (32)Urine : FTCmg/cig : mg/cigAverage0.49 (0.13)0.66 (0.19)0.410.27 (42)Filter : FTCmg/cig : mg/cigAverage0.52 (0.17)0.85 (0.24)0.410.35 (38)Saliva : FTCng/ml : mg/cigAverage211 (58)184 (81)0.22119 (46)Urine Urinary nicotine equivalents, Filter daily nicotine yield estimated from filter analysis, Saliva saliva cotinine concentration, FTC FTC nicotine yield, CI confidence interval The means of the five daily data values per subject are shown in Fig. 2 using the same format as Fig. 1. Again, for all correlations, the slopes and intercepts were significant (p < 0.01). The correlation for UNE vs nicotine yield (Fig. 2a) improved significantly as shown in Table 3. The slope increased to 0.74 while the intercept moved closer to zero. The saliva cotinine correlations (Fig. 2b,c) also improved, but not as dramatically as the urinary nicotine. The correlation for UNE vs saliva cotinine (Fig. 2b) had an R2 value of 0.55 with a standard error of 5.3 mg/day (36%). The correlation for nicotine yield vs saliva cotinine (Fig. 2c) had an R2 value of 0.54 with a standard error of 6.5 mg/day (33%). Fig. 2Correlations of the three methodologies (5-day average per subject) Other findings Although the primary purpose of this study was to determine the statistical correlations between the three methods of estimating nicotine uptake, data collected during the experiment allowed for additional observations. Creatinine normalization Normalization by urinary creatinine is useful when only a single or partial daily urine collection is taken. To test the effect of creatinine normalization, observed daily UNE were divided by their respective millimole creatinine. A correlation was calculated with nicotine yield as the dependent variable and UNE/mmol creatinine as the independent variable. The results are summarized in Table 3. Both the intercept and slope are significant (p < 0.001) but normalization with creatinine clearly degrades the correlation compared to the correlation without normalization. The findings suggest that normalization with creatinine adds another factor of variability, which has a degrading effect on R2, and therefore indicates that the use of 24-h urine samples without normalization is expected to provide more accurate results than the analysis of partial (<24 h) urine samples with normalization. This is in agreement with Heavner et al. (2006), who showed mechanistically that creatinine normalization may be appropriate for some, but not all, of the urinary metabolites of nicotine and that other methods of normalization may be more appropriate for spot urine samples. FTC smoke yields Because this was a confined clinical study, subjects would not necessarily be expected to behave as if they were in their normal environment. However, because a wide range of FTC cigarette yields were tested, it was considered worthwhile to test the correlation between the measured FTC yields and the three methods of nicotine estimation. Correlation results are summarized in Table 3 for the 5-day mean values for the three methods vs the FTC values on a per subject basis. All correlations gave significant slopes and intercepts (p < 0.001). Differences between tar band groupings The results for mouth exposure to nicotine (from filter analysis), UNE, saliva cotinine, and cigarettes smoked per day were grouped by tar band and the mean, standard deviation, and one-way analysis of variance (ANOVA) were calculated on a per subject basis (Altman and Bland 1997). The results are shown in Table 4. Group means increased in rank order with increasing group yield except for saliva cotinine and cigarettes smoked per day. All measurements except cigarettes smoked per day gave a significant (p < 0.05) effect of group by ANOVA. For each measurement, results with different letter assignments are significantly different at the 95% confidence level (Fisher test). Normalizing urinary nicotine metabolites with creatinine resulted in an increase in p value for effect of group and no significant differences between the FF, LTS, and ULH groups. All three groups were significantly higher than the ULL group. This again shows that creatinine normalization degrades the discriminating power. Table 4Mean ± SD for measurements grouped by tar band. ULLULHLTSFFANOVA pNo. of subjects15202019–Mouth exposure (mg nicotine/cigarette)0.63 ± 0.18a0.94 ± 0.29b1.19 ± 0.33c1.37 ± 0.58c<0.0005Mouth exposure (mg nicotine/day)14.7 ± 5.6a22.0 ± 7.1b24.4 ± 6.4bc28.3 ± 12.6c<0.0005Urinary nicotine (mg/cigarette)0.59 ± 0.19a0.84 ± 0.27b0.96 ± 0.21b1.20 ± 0.41c<0.0005Urinary nicotine (mg/day)13.6 ± 5.7a 19.3 ± 6.8b19.8 ± 5.9b24.5 ± 8.9c<0.0005Creatinine normalized urinary nicotine (mg/day/mmol)0.93 ± 0.58a1.31 ± 0.51ab1.43 ± 0.58b1.45 ± 0.52b0.040Saliva cotinine (ng/ml)216 ± 107a347 ± 119bc314 ± 88b411 ± 149bc<0.0005Cigarettes/day22.8 ± 4.7a23.4 ± 3.6a21.2 ± 4.5a20.8 ± 4.2a0.190Cigarettes/day (self-reported minus measured)1.6 ± 5.1−1.0 ± 8.0−0.9 ± 5.4−0.1 ± 5.9Results with different letters are significantly different at the 95% confidence level Self-reported vs actual cigarette usage Table 4 also shows the difference in self-reported and measured cigarette usage. Self-reported cigarette usage was questioned and compliance with respect to truthfulness can be a concern (Byrd et al. 1998). In this study, subjects were asked to estimate their cigarette usage as part of the recruiting process. This gave a basis to compare with the measured usage during the study (overall mean = 22 cigarettes/day, SD = 4.2, range 15.8 to 32.0). The mean value for the (5-day average− self-reported) usage per subject was −0.2 cigarettes/day (SD = 5.9, range −19.4 to +11.7). Even when broken down to the smaller tar band groups of 15 to 20 subjects, the mean difference in measured and reported usage was less than 2 cigarettes/day. Within-subject variation Because the trial took place over five consecutive days, within-subject, day-to-day variation was calculated for each of the measured variables. In addition, on day 4, saliva samples were taken at approximately 0830, 1330, and 1830 hours for measurement of saliva cotinine to estimate within-day variation. Results are summarized in Table 5 and are expressed as a pooled (root-mean-squared) coefficient of variation. The pooled day-to-day variations for each variable were similar, ranging from 15.2% for cigarettes smoked per day to 18.1% for UNE per day. The within-day variation of 8.1% for saliva cotinine was approximately half the day-to-day variation of 15.7%. When a single factor ANOVA was calculated using saliva cotinine for the three time periods sampled on day 4, the variation between time periods was insignificant (p = 0.58). However, this could have been overwhelmed by the subject-to-subject variation in nicotine uptake. This factor was removed by dividing the individual measurement by the daily average per subject for all three time periods. A single factor ANOVA was significant for time of day (p < 0.001) with the mean (95% confidence interval) saliva cotinine values being 1.008 (0.015), 0.96 (0.012), and 1.03 (0.013) times the average daily value per subject for the 0830, 1330, and 1830 hour samples, respectively. Even though the time of day had a statistically significant effect, the practical differences were small. Table 5Within-subject variation expressed as a pooled coefficient of variation (CV)AnalyteTypePooled CV (range)%Filter (nicotine yield/day)Day-to-day16.7 (6–36)Urine (nicotine equiv./day)Day-to-day18.1 (4–54)Saliva cotinine concentrationDay-to-day15.7 (4–47)Saliva cotinine concentrationWithin-day8.1 (0.8–16)Cigarettes/dayDay-to-day15.2 (3–36) Discussion All three of the estimation methods correlated significantly with each other, but the best overall correlation was between the filter analysis method and UNE. The slopes for the correlations of UNE as a function of mouth exposure to nicotine were 0.67 and 0.74 for the individual and 5-day average regressions, respectively, implying that about 70% of the difference in the UNE measured in this study was due to a difference in mouth exposure to nicotine. However, the mean of the total UNE expressed as a percentage of the nicotine entering the mouth was 89%. The mean value falls within the reported range of 80% (Benowitz et al. 1994) to 90% (Curvall et al. 1991) but the slopes fall below this range. The difference between the two methods is due to the significant intercept calculated using the linear regression. The intercept using the 5-day average results was lower than the intercept using the individual results and the slope was greater. The intercept could represent compartmental nicotine and/or metabolite storage with subsequent carryover from storage before entering the study. This influence should diminish with a 5-day average compared to using single day results. An example of this is demonstrated in Fig. 1a where there is a single circled data point that appears to be an outlier. Urinary output of nicotine equivalents was approximately two times the mouth exposure of nicotine for that day. However, this was a day 1 measurement for a single subject. For subsequent days, the data points for this subject are buried within all the other data points. One possibility is that the subject had a much larger exposure to nicotine before participating in the study and the unusually high urinary output was due to clearance of the prior exposure. Averaging the results over 5 days would take out much of the metabolic influence and result in a much better correlation as demonstrated with these results. For the saliva cotinine correlations, the R2 values only improved slightly with averaging. Part of the reason for the improvement with all correlations can be attributed to averaging out measurement variation. This should be similar for all correlations. However, the additional improvement in the filter vs urine correlation can be explained by metabolism. Nicotine has a relatively short serum elimination half-life of 2.2–2.9 h (Scherer et al. 1988; Benowitz and Jacob 1993; Benowitz et al. 1999, 2002, 2004), which means that the urinary nicotine should be from the nicotine taken in that day. Cotinine has a serum elimination half-life typically reported at 16–18 h (Scherer et al. 1988; Benowitz et al. 1999, 2002, 2004; De Schepper et al. 1987) and a similar urine elimination half-life (Benowitz and Jacob 1993; De Schepper et al. 1987). 3-HC, which is downstream metabolically from cotinine, has a serum elimination half-life of 5.9–6.6 h with a similar urine elimination half-life (Scherer et al. 1988; Benowitz and Jacob 2001). Given the pharmacokinetic information, these urinary metabolites must have originated from nicotine exposure on multiple days as has been demonstrated directly using nicotine infusion (Scherer et al. 1988). Because measurements were taken over sequential days of input (filter analysis), the temporal source of urinary and salivary metabolites could be estimated by multiple regression of the metabolites vs the daily nicotine exposure. This analysis was performed using the nicotine exposure for the current day and the two previous days according to the equation: where (N), (N-1), and (N-2) equals the nicotine exposure for the current day, 1 day before, and 2 days before the urine or saliva sample, respectively. Metabolite measurements from days 3, 4, and 5 were used. None of the coefficients for nicotine exposure from 2 days before the urine collection were significant (p > 0.15). For saliva cotinine, the coefficient for 2 days before was not significant at the 95% confidence but the p value of 0.07 was small enough to warrant consideration. Given this, the regressions were recalculated using (N) and (N-1) and including the urinary metabolite measurements from day 2 as well. The results of this regression are shown in Table 6. Table 6Multiple regression results for source of urinary and salivary metabolites as a function of nicotine exposure from current and previous dayCoefficientSum of urinaryUrinary nicotineUrinary cotinineUrinary 3-HCSalivary cotinineR20.720.400.640.420.50Intercept3.51 (<0.001)0.77 (<0.01)0.85 (<0.01)1.89 (<0.001)97.9 (<0.01)Previous day0.24 (<0.001)−0.004 (0.89)0.08 (<0.01)0.16 (<0.001)5.7 (<0.01)Current day0.46 (<0.001)0.16 (<0.001)0.18 (<0.01)0.12 (<0.01)4.2 (<0.01)Results expressed as regression coefficient (p value) For the sum of all urinary metabolites, the coefficients were significant for the current day and prior day’s exposure. This implies that the total urinary metabolites originated from nicotine exposure over 2 days. The only coefficient that was significant for the urinary nicotine regression was for the current day’s exposure as would be expected. The coefficient of 0.16 implies that 16% of the nicotine entering the mouth appeared as urinary nicotine plus the glucuronide. This agrees closely with the estimates of 13.8% (Benowitz et al. 1994) and 10–15% (Curvall et al. 1991). For the urinary cotinine regression, coefficients for the current day and previous day were significant with the current day being about twice that of the previous day. The two coefficients imply that 26% of the nicotine entering the mouth appears as urinary cotinine plus the glucuronide. This also agrees closely with estimates of 25.6% (Benowitz et al. 1994) and 20–25% (Curvall et al. 1991). For the 3-HC regression, the coefficients for the current day and previous day were significant with the current day being slightly smaller than the previous day. The sum of the coefficients (0.28) was smaller than the reported 41 to 60% of the total urinary metabolites (Benowitz et al. 1994; Curvall et al. 1991). It is possible that because 3-HC is the third step in the metabolism of nicotine, there was too much of a smoothing effect for the change in daily mouth exposure to fully capture. For saliva cotinine, all coefficients were significant, suggesting that saliva cotinine results are an amalgam of at least 2 days of nicotine exposure. Saliva cotinine correlations suffer because results are expressed as concentration rather than an absolute value, and as such can be influenced by body size. In addition, other variables come into play that are not easily explained. An example of this is shown in Figs. 1 and 2 by the data points with boxes around them. These were from one subject that appeared to have unusually low saliva cotinine values for both UNE (Figs. 1b and 2b) and mouth exposure of nicotine (Figs. 1c and 2c). This subject had the highest mouth exposure of nicotine of all the subjects studied, yet the saliva cotinine values were only slightly above midrange. There was nothing unusual about this subject and average urinary output was within 1 SD (1,100 ml) of the average for all subjects (2,400 ml). It is unlikely that nicotine yields were overestimated because Fig. 1a shows that the data points are scattered about the regression line for UNE vs nicotine yield. It is also unlikely that the saliva cotinine concentrations were in error because the saliva cotinine for each day was analyzed in separate batches along with other samples. The subject did not rapidly metabolize cotinine to 3-HC because cotinine accounted for 47% of the urinary metabolites measured for this subject compared to an average of 35% for all subjects indicating the converse. Therefore, it must be concluded that this is an anomaly characteristic of this individual. As shown in Table 3, with many of the estimation methods, the correlations are only slightly better than simply correlating with FTC nicotine yield of the cigarette. The average nicotine mouth exposure, as measured by the filter method, correlated with the FTC nicotine yield with a standard error of 38% and an R2 value of 0.41. The correlation of nicotine mouth exposure with saliva cotinine gave a standard error of 33% and creatinine-normalized urinary metabolites gave a standard error of 32% with R2 values approximately 0.1 higher. Correlations of FTC nicotine with biomarkers in this study were stronger than that reported in other studies (Byrd et al. 1998; Jarvis et al. 2001; Ueda et al. 2002; Hecht et al. 2005; Bernert et al. 2005). We believe that there are valid reasons for this. One is that the correlations were performed using the 5-day averages per subject instead of a single sample per subject. In addition, exact compliance to brand and cigarette consumption was assured in the current study, whereas all but one (Bernert et al. 2005) of the referenced studies used self-reported brand identification and cigarette consumption. Self-reported brand information was reported to have about a 25% error rate when compared with packs returned (Peach et al. 1986) or in a test–retest comparison (Eisenhower et al. 1993). This and the potential for use of alternate brands during a study can further confound a correlation with biomarkers. Other confounding factors are the use of creatinine normalized spot urine samples instead of 24-h urine samples and analyzing for a subset of the nicotine metabolites used in the current study. Simple creatinine normalization is only biologically valid for xenobiotics that have the same excretion mechanism and urinary flow rate dependence as creatinine. Heavner et al. (2006) have shown that 1-HC (free and glucuronide) and cotinine glucuronide have urinary flow rate dependence similar to creatinine while nicotine (free and glucuronide), free cotinine, 1-hydroxypyrene, and the free and glucuronide forms of 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol do not. In conclusion, two methods stand out as superior. One is the filter method, which estimates mouth level exposure directly on a per cigarette basis. Filter collection need not be quantitative, the filter returned can be compared to the brand it is supposed to be and it can be readily determined if it was smoked or not. Thus, brand compliance and smoking status can be assured even if the subject happened to occasionally use a different brand or other form of nicotine during the study. With self-reported daily cigarette use from groups of at least 15–20 subjects, the exposure per cigarette can be converted accurately to daily exposure. In addition, mouth exposure to tar can also be estimated using the filter method (Shepperd et al. 2006). The other measurement, which is considered by many to be the “gold standard,” is the measurement of urinary nicotine and metabolites from 24-h urine samples without creatinine normalization. This method appears to reflect the mean daily nicotine uptake of the last 2 days.	[ "nicotine", "filter", "saliva", "urine", "metabolites", "smoking behavior", "cigarette smoking" ]	[ "P", "P", "P", "P", "P", "P", "P" ]
Purinergic_Signal-3-4-2072911	Lack of evidence for direct phosphorylation of recombinantly expressed P2X2 and P2X3 receptors by protein kinase C	P2X3 and P2X2+3 receptors are present on sensory neurons, where they contribute not only to transient nociceptive responses, but also to hypersensitivity underlying pathological pain states elicited by nerve injuries. Increased signalling through P2X3 and P2X2+3 receptors may arise from an increased routing to the plasma membrane and/or gain of function of pre-existing receptors. An obvious effector mechanism for functional modulation is protein kinase C (PKC)-mediated phosphorylation, since all P2X family members share a conserved consensus sequence for PKC, TXR/K, within the intracellularly located N-terminal domain. Contradictory reports have been published regarding the exact role of this motif. In the present study, we confirm that site-directed elimination of the potential phosphor-acceptor threonine or the basic residue in the P+2 position of the TXR/K sequence accelerates desensitization of P2X2 receptors and abolishes P2X3 receptor function. Moreover, the PKC activator phorbol 12-myristate 13-acetate increased P2X3 (but not P2X2) receptor-mediated currents. Biochemically, however, we were unable to demonstrate by various experimental approaches a direct phosphorylation of wild-type P2X2 and P2X3 receptors expressed in both Xenopus laevis oocytes and HEK293 cells. In conclusion, our data support the view that the TXR/K motif plays an important role in P2X function and that phorbol 12-myristate 13-acetate is capable of modulating some P2X receptor subtypes. The underlying mechanism, however, is unlikely to involve direct PKC-mediated P2X receptor phosphorylation. Introduction P2X receptors constitute an abundant class of ligand-gated ion channels, which respond to extracellular ATP and related nucleotides with the opening of an intrinsic pore permeable to Na+, K+ and Ca2+ [1, 2]. A family encompassing seven P2X subunit genes, designated P2X1–7, has been identified in rodents and mammals. Six of the seven subunit isoforms (P2X1-P2X5 and P2X7) are able to assemble into homotrimeric receptor channels [3–8] with distinct pharmacological and electrophysiological phenotypes [9]. According to their sensitivity to the synthetic ATP analogue α,β-methylene adenosine 5′-triphosphate (αβ-meATP) and the rate of current desensitization, P2X receptors are generally subgrouped into at least two categories: (1) rapidly desensitizing and αβ-meATP sensitive (P2X1 and P2X3) and (2) slowly or non-desensitizing and αβ-meATP insensitive (P2X2, P2X4 and P2X7) receptors. The term “non-desensitizing” means that the currents are maintained for at least a few seconds in the continuous presence of agonist. A peculiar phenotype is inherent to heteromeric P2X2/P2X3 (P2X2+3) receptors, which feature αβ-meATP sensitivity combined with a non-desensitizing current response. P2X receptors are found on the surface of a large variety of cells, where they are involved in numerous sensory processes including nociception under both physiological and pathological processes [10–15]. On nociceptive sensory neurons, extracellular ATP acts as a pain-producing neurotransmitter predominantly through homotrimeric P2X3 receptors or heterotrimeric P2X2+3 receptors. Conditions and mediators that facilitate P2X3 and P2X2+3 receptor-mediated signalling, apparently by potentiating ionic currents, include (1) substance P and bradykinin [16], (2) nerve injury [17] and (3) calcitonin gene-related peptide (CGRP), a potent vasodilator and proinflammatory agent [18]. How this increase in receptor function occurs has not been fully solved. One obvious candidate mechanism is protein kinase C (PKC)-mediated phosphorylation, since all P2X family members share a conserved consensus sequence for PKC-mediated phosphorylation (TXR/K) within the intracellularly located N-terminal domain (Fig. 1a, b). Indeed, results obtained with pharmacological activators and inhibitors are consistent with PKC being involved in P2X3 receptor potentiation by substance P and bradykinin [16], as well as the calcitonin gene-related peptide [18]. Direct PKC-mediated phosphorylation has so far been demonstrated only for the P2X2 receptor, with 18T of the 18TPK20 sequence as the phospho-acceptor site [19]. Mutational analysis has further shown that the TXR/K sequence of the P2X1, P2X2 and P2X3 receptors plays an important role in controlling the rate of receptor desensitization [16, 19, 20]. Changes of the rate of desensitization could profoundly influence the efficiency of synaptic transmission and thus contribute to pathological pain states. Overall, there is significant interest in a better understanding of the mechanisms involved in the short-term and long-term regulation of P2X receptors. Fig. 1Effect of site-directed modification of a putative PKC motif on currents mediated by ATP-activated P2X2 and P2X3 receptors. a Schematic model of the transmembrane topology of the rat P2X3 subunit illustrating the N-terminal position of the 12TTK14 sequence. b Alignment of intracellular N-terminal amino acid sequences of the seven P2X subunit isoforms reveals a highly conserved consensus motif, TXR/K. c Typical current traces elicited by applying 10-s pulses of 100 μM ATP to oocytes expressing the indicated wild-type or mutant P2X2 receptors. d Typical current traces elicited by applying 10-s pulses of 100 μM ATP to oocytes expressing the indicated wild-type or mutant P2X3 receptors. Gray areas indicate the duration of ATP application. e All the P2X2 and P2X3 receptors and receptor mutants were expressed efficiently at the cell surface. Intact, healthy oocytes expressing the indicated proteins for 2 days were surface-labelled with the membrane impermeant reactive Cy5 dye and then extracted with dodecylmaltoside. Recombinant proteins were isolated by Ni2+ chelate chromatography and resolved by reducing SDS-PAGE. Shown is a fluorescence scan of an SDS-PAGE gel The aim of this study was to assess biochemically the conditions in which P2X2 and P2X3 receptors become directly phosphorylated by PKC in a heterologous system. While this work was in progress, a similar study was published for the P2X3 receptor that arrived at essentially the same conclusion, namely that the P2X3 receptor is unlikely to be a direct PKC substrate [21]. Our results complement and extend this finding by demonstrating that the P2X2 receptor also does not serve as a substrate for direct PKC-mediated phosphorylation. Materials and methods Materials The sources of antibodies, enzymes or peptide substrates are specified at the appropriate places in the text. Chemicals not otherwise specified were purchased in the highest available quality from Sigma-Aldrich (Taufkirchen, Germany) or Merck (Darmstadt, Germany). Peptide-specific antibodies to P2X1, P2X2 and P2X3 subunits Peptides ATSSTLGLQENMRTS (residues 385-399 of the P2X1 subunit), QQDSTSTDPKGLAQL (residues 458–472 of the P2X2 subunit) and VEKQSTDSGAYSIGH (residues 383–397 of the P2X3 subunit) were selected based on their lack of similarity with known proteins and predicted hydrophilicity for raising polyclonal antibodies against the C-terminal end of the rat subunits P2X1, P2X2 and P2X3, respectively. Rabbit polyclonal antibodies were ordered through the Custom Antibody Production Services of Eurogentec (Seraing, Belgium). After 3 months and a total of four immunizations, rabbits were sacrificed by complete bleeding. Peptide-specific antibodies were purified from the serum by affinity columns coupled with the peptides used for immunization. Immunoblotting showed that the antibodies recognized only the receptor against which they were raised, thus demonstrating isoform specificity of these antibodies (results not shown). The antibodies were aliquoted at 0.12–0.16 mg/ml in phosphate-buffered saline (PBS) containing 0.1% bovine serum albumin (BSA) and 0.01% thimerosal and stored at −80°C. cDNA constructs Plasmids encoding wild-type and N-terminal hexahistidine-tagged (His-tagged) versions of rat P2X1 subunits (rP2X1, GenBank accession no. X80477), P2X2 subunits (rP2X2, GenBank accession no. U14414) or rat P2X3 subunits (rP2X3, GenBank accession no. X90651) in the oocyte expression vector pNKS2 [22] were available from previous studies [3, 8]. Insertion and replacement mutations were introduced by QuikChange site-directed mutagenesis (Stratagene, La Jolla, CA, USA). For protein expression in mammalian cells, the receptor cDNAs were subcloned from the oocyte expression vector pNKS2 into pcDNA3.1. The plasmid His-rP2X2-EGFP-pcDNA3.1 with a C-terminal enhanced green fluorescent protein (EGFP) fusion was generated by eliminating the stop codon in His-rP2X2 and adding the coding sequence of EGFP (Clontech, Palo Alto, CA, USA) in frame. All constructs were verified by restriction analysis and nucleotide sequencing. A pEGFP-C1 plasmid harbouring the N-terminally EGFP-tagged splicing factor SF3B1 was kindly provided by our colleague Dr. Walter Becker [23]. P2X receptor expression in Xenopus laevis oocytes Capped cRNAs were synthesized from plasmids linearized downstream to the 100 bp long polyA tail provided by the oocyte expression vector pNKS2 [22] and injected at 0.5 μg/μl in 50-nl aliquots into collagenase-defolliculated X. laevis oocytes using a Nanoliter 2000 injector (WPI, Sarasota, FL, USA), as detailed previously [24]. Oocytes were maintained at 19°C in sterile oocyte Ringer’s solution [ORi: 90 mM NaCl, 1 mM KCl, 1 mM CaCl2, 1 mM MgCl2 and 10 mM hydroxyethylpiperazine ethanesulfonic acid (HEPES), pH 7.4] supplemented with 50 μg/ml of gentamicin. Heterologous expression in HEK293 cells cDNAs encoding P2X subunits in pcDNA3.1 vector (Invitrogen, Karlsruhe, Germany) or the splicing factor SF3B1 in EGFP-C1 vector were transiently transfected into HEK293 cells using Lipofectamine LTX (Invitrogen, Karlsruhe, Germany). Cells were cultured at 37°C in Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen) supplemented with 10% (v/v) fetal calf serum (FCS) (PAA Laboratories, Linz, Austria), 100 U/ml of penicillin G and 100 μg/ml of streptomycin. Metabolic [35S]methionine labelling of oocytes For metabolic [35S]methionine labelling of oocyte-expressed recombinant proteins, cRNA-injected oocytes and non-injected controls were incubated overnight with L-[35S]methionine (>40 TBq/mmol; PerkinElmer, Boston, MA, USA) at about 100 MBq/ml (0.4 MBq per oocyte) in ORi at 19°C, and then chased for 24 h. Surface fluorescence labelling of oocytes Selective fluorescence labelling of plasma membrane-bound P2X receptors was achieved by using the amino-reactive fluorescent Cy5 dye (GE Healthcare, Freiburg, Germany), which is membrane impermeant due to its two sulfonic acid groups. Two days after cRNA injection, oocytes were washed with oocyte-PBS pH 8.5 (20 mM Na phosphate, 110 mM NaCl, 1 mM MgCl2), and then incubated for 30 min at ambient temperature with Cy5 dye, which was diluted 200-fold to a final concentration of 50 μg/ml from a dimethyl sulfoxide (DMSO) stock solution. The reaction was terminated by washing the cells with oocyte-PBS, followed by membrane protein extraction with dodecylmaltoside and receptor purification (see below). Isolation of recombinant P2X receptors from X. laevis oocytes or HEK293 cells His-tagged receptors were purified by Ni2+ nitrilotriacetic acid (NTA) agarose (Qiagen, Hilden, Germany) affinity chromatography from non-ionic detergent extracts of oocytes or HEK293 cells in the presence of iodoacetamide essentially as described [3, 8, 25]. Slight modifications were that buffers were supplemented with ethylenediaminetetraacetate (EDTA)-free Halt™ protease inhibitor cocktail (Pierce, Rockford, IL, USA), 50 mM NaF and 50 mM Na4P2O7 for phosphatase inhibition, and that digitonin was replaced by 0.2% and 0.05% dodecylmaltoside for membrane protein extraction and repetitive washing of resin-bound proteins, respectively. Proteins were released from the washed Ni2+ NTA agarose beads with non-denaturing elution buffer consisting of 250 mM imidazole/HCl (pH 7.4) and 0.02% dodecylmaltoside, and then kept at 0°C until analysed on the day of purification. In addition, immunoprecipitation assays were performed using either cRNA-injected Xenopus oocytes or transiently transfected HEK293 cells. Cells were lysed in ice-cold immunoprecipitation lysis buffer containing 150 mM NaCl, 25 mM Tris/HCl pH 7.4, 20 mM NaF, 20 mM Na4P2O7, 10 μl/ml Halt™ protease inhibitor cocktail without EDTA (Pierce, Rockford, IL, USA), 1% sodium dodecyl sulfate (SDS) (w/v) and 50 mM iodoacetamide. For efficient solubilization, lysates were incubated for 30 min on ice with occasional vortexing, and then cleared by centrifugation for 15 min at 16,100 g and 4°C. The cleared SDS extract was diluted 1:1 with immunoprecipitation binding buffer [150 mM NaCl, 20 mM Tris/HCl pH 7.4, 0.4 mM phenylmethylsulfonyl fluoride (PMSF), 0.5% Triton X-100 v/v and 25 mM iodoacetamide], supplemented 1:1,000 with the desired peptide-specific P2X2 or P2X3 antibodies or a polyclonal goat anti-EGFP antibody (Rockland, Gilbertsville, PA, USA) and incubated for 2 h on ice. Immunocomplexes were bound to the protein A Sepharose by overnight head-over-head rotation at 8°C, then washed 5 times with immunoprecipitation binding buffer containing 0.1% Triton X-100 and finally eluted by incubating the beads for 15 min at 56°C in reducing sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer. In vitro PKC phosphorylation assay In vitro phosphorylation assays were performed by using P2X receptors as substrates, purified as detailed above. In addition, a selective peptide PKC substrate (RRGRTGRGRRGIYR, Calbiochem-Merck, Darmstadt, Germany, final concentration 150 μM) corresponding to the amino acid sequence 1487-1500 of the hepatitis C virus polyprotein [26] was used as a positive control. The proteins or the peptide were incubated for 30 min at 30°C with 0.02 unit (0.02 μg) of purified catalytic PKC subunit from rat brain and 50 μM [γ-32P]ATP (10 mCi/ml, Amersham Biosciences GE Healthcare, Freiburg, Germany) in phosphorylation buffer containing 6 mM CaCl2, 10 mM MgCl2, 20 mM Tris/HCl pH 7.5, 1 mM EDTA, 1 mM mercaptoethanol, 0.05% Triton X-100 (v/v) and 0.1 mg/ml BSA. The reaction was terminated by boiling in reducing SDS-PAGE sample buffer for 30 s. Proteins were resolved by reducing SDS-PAGE and visualized by immunoblotting or phosphorimaging as detailed below. In vivo phosphorylation assay Transiently transfected HEK293 cells were metabolically labelled by a 3-h incubation with 500 μCi/ml of [32P]orthophosphate (Amersham Biosciences GE Healthcare, Freiburg, Germany) in phosphate-free oocyte Ringer solution or in phosphate-free DMEM (Invitrogen, Karlsruhe, Germany) as appropriate. Cells were treated with or without 100 nM phorbol 12-myristate 13-acetate (PMA; Sigma-Aldrich, Taufkirchen, Germany) or 100 nM calyculin A (Tocris Bioscience, Ellisville, MO, USA) for 10 min at ambient temperature. Non-incorporated radioactivity was removed by rinsing the cells twice with tracer-free medium. Affinity purification or immunoprecipitation was performed as described above followed by SDS-PAGE and immunoblotting or phosphorimaging. SDS-PAGE Samples for SDS-PAGE were denatured by incubating in reducing SDS-PAGE sample buffer for 15 min at 56°C and then electrophoresed on a Laemmli SDS-PAGE gel (10% acrylamide) in parallel with [14C]-labelled molecular mass markers (Rainbow™, Amersham Biosciences GE Healthcare, Freiburg, Germany). Cy5-labelled proteins were visualized by scanning the wet SDS-PAGE gel with a fluorescence scanner (Typhoon, GE Healthcare, Freiburg, Germany). For visualization of radiolabelled or non-labelled proteins, SDS-PAGE gels were either fixed, dried and exposed to a PhosphorImager screen, and scanned using a Storm 820 PhosphorImager (Amersham Biosciences GE Healthcare, Freiburg, Germany) or/and blotted to polyvinylidene fluoride (PVDF) membranes. Immunoblotting Proteins separated by SDS-PAGE were electroblotted onto PVDF membranes. Membranes were blocked for 1 h with 5% non-fat milk in PBS (or, alternatively, with 5% BSA if the phosphothreonine monoclonal antibody was used for immunoblotting), washed 3 × 5 min with PBST (0.1% Tween 20 in PBS) and incubated for 2 h at 4°C in PBS containing the appropriate primary antibody. Phosphorylated proteins were detected by immunoblotting with a phosphothreonine monoclonal antibody (Cell Signaling Technology, Beverly, MA, USA). Alternatively, the peptide-specific antibodies to P2X2 and P2X3 subunits were used to verify receptor expression levels. Following washing in PBST, an appropriate horseradish peroxidase-conjugated secondary antibody (goat anti-mouse or goat anti-rabbit, Pierce, Rockford, IL, USA) was applied for 1 h in PBS supplemented with 1% non-fat milk. After another 3 × 5 min washes with PBST, immunoreactive bands were visualized by chemiluminescence imaging using a Fujifilm LAS-3000 system (Tokyo, Japan) and the AIDA Image Analyser Software Version 4.08 (Raytest, Straubenhardt, Germany). Two-electrode voltage clamp current recordings and electrical capacitance measurements Two-microelectrode voltage clamp measurements of X. laevis oocytes were performed 2 days after cRNA injection using a Turbo TEC-05 amplifier (npi electronics, Tamm, Germany) interfaced by an INT-20X AD/DA converter (npi electronics, Tamm, Germany) to a personal computer running the CellWorks Lite 5.1 software (npi electronics, Tamm, Germany), as detailed previously [27–29]. Current traces were elicited by 100 μM ATP at 1- to 5-min intervals (Sigma–Aldrich, Taufkirchen, Germany) at a holding potential of −60 mV in nominally Ca2+-free bath medium (designated Mg-ORi) to avoid activation of endogenous Ca2+-activated Cl− channels. ATP was prepared in Mg−ORi at its final concentration and applied by gravity-fed perfusion at a flow rate of 10 ml/min to an oocyte held by the two microelectrodes in the bath chamber (volume ∼10 μl). Switching between different perfusion solutions was achieved with electromagnetic valves controlled by the CellWorks Lite 5.1 software. Electrophysiological recordings were analysed using the Origin 6.0 software (Microcal Software Inc., Northampton, MA, USA). Single exponentials were fitted to the decay phase of the macroscopic currents with a nonlinear least squares method. The time constant τ was taken as an estimate of the rate of receptor desensitization. To monitor PMA-induced changes in the oocyte surface area, depolarizing voltage steps of ΔV = 10 mV were applied from the holding potential to elicit a capacitive transient Icap, the integral of which yields the charge movement Qcap and thus the membrane capacitance Cm = Qcap/ΔV [30–32]. PMA was dissolved in DMSO to a stock solution of 10 mM and stored at −20°C. On the day of the experiment, the stock solution was pre-diluted to 100 μM in DMSO and then further diluted with Mg-ORi to the final concentration of 100 nM. Vehicle control experiments with a final concentration of 0.1% DMSO did not reveal any effects on P2X2 receptor-mediated currents measured with X. laevis oocytes (results not shown). All measurements were made at room temperature (21–23°C). Data analysis Data are shown as means±SEM. The paired Student’s t-test (two-tailed) was used to compare current amplitudes with statistical significance set to p < 0.05. Results Site-directed modification of a putative N-terminal PKC site affects the electrophysiological phenotype of P2X2 and P2X3 receptor channels First, we tested each of the non-mutated and mutant receptor constructs for expression of an ATP-gated inward current in cRNA-injected X. laevis oocytes. Application of 100 μM ATP induced typical non-desensitizing currents through oocyte-expressed P2X2 receptors (Fig. 1c, left current trace). The P2X2 subunit shares with other P2X subunit isoforms a conserved consensus site for PKC phosphorylation (TXR/K, where T is the phosphorylation site and X denotes any amino acid; see Fig. 1a, b). Disruption of the conserved N-terminal PKC site by mutating the phosphor-acceptor 18T to A resulted in a rapidly desensitizing channel (Fig. 1c, middle current trace). Also, replacement of 20K by T resulted in a desensitizing phenotype (Fig. 1c, right current trace). Since PKC preferentially phosphorylates threonine (or serine) residues that are close to a C-terminal basic residue, replacement of 20K in the P+2 position by a non-basic residue such as T will destroy the putative PKC site despite preserving the phosphor-acceptor 18T. Virtually identical findings had previously been reported by Boue-Grabot et al. [19]. The average half-times of desensitization (τ) of the 18A and 20T P2X2 mutants were quite similar, with mean τ values (±SEM) of 4.5 ± 0.4 s and 7.3 ± 0.5 s, respectively, of eight oocytes from two females per value. ATP applied to oocytes expressing the P2X3 receptor elicited a rapidly desensitizing inward current (Fig. 1d, left current trace) as is typical for this P2X receptor subtype. Incorporation of similar mutations, as in the P2X2 subunit in the analogous positions of the P2X3 receptor T12A or K14T, completely abolished P2X3 receptor function (Fig. 1d, middle and right current traces). Visualization of plasma membrane-bound fluorescence-labelled P2X receptors by affinity purification combined with SDS-PAGE and fluorescence scanning verified that all the receptors were efficiently exported to the oocyte surface (Fig. 1e). The PKC activator PMA augments P2X3 receptor-mediated currents, yet leaves P2X2-mediated currents unchanged Since the mutagenesis data for the P2X2 and P2X3 receptor were consistent with the possibility of a regulatory role for PKC in P2X receptor function, we examined whether the PKC activator PMA was capable of affecting the electrophysiological phenotype of P2X2 or P2X3 receptors. In planning the experiments, we considered that PMA treatment of X. laevis oocytes is known to induce a pronounced endocytotic reduction of the cell surface area, as evidenced microscopically and electrophysiologically by the disappearance of virtually all microvilli and a concomitant decrease of the electrical capacitance of the plasma membrane, respectively [31, 32]. Associated with oocyte surface reduction is a proportional internalization of membrane proteins, such as endogenous Na+-K+ pumps, suggesting that PMA treatment would stimulate internalization of recombinant P2X receptors as well. Re-examination of the time course of the PMA-induced reduction of the oocyte surface area showed that the electrical capacitance did not start to decline significantly before 10 min after PMA application (Fig. 2a). Accordingly, all current recordings were done within a time window of 10 min after PMA application to prevent any contribution of receptor internalization to the PMA effect on P2X receptor-dependent currents. Fig. 2Effect of PMA on ATP-induced currents mediated by expressed P2X2 and P2X3 receptors in X. laevis oocytes. a The electrical membrane capacitance was monitored as a measure of the oocyte surface area during sustained stimulation with PMA. Each bar represents the mean Cm±SEM calculated from the areas under the current transients elicited in 10-s intervals by five consecutive 10-mV depolarizing steps. b, c Representative current traces elicited by 10-s lasting pulses of 100 μM ATP applied in 10-min intervals to oocytes expressing the indicated P2X receptors or P2X receptor mutants. Where denoted by a black bar, oocytes were pre-incubated with 100 nM PMA before ATP was co-applied. PMA was without effect on the P2X2 receptor-mediated current (b), but induced a marked increase of the current amplitude mediated by P2X3 receptors (c). Gray areas indicate the duration of ATP application Figure 2b and c show typical current traces elicited by ATP from P2X2 or P2X3 receptor-expressing oocytes before and after a 10-min exposure to PMA. PMA neither affected the shape nor the amplitude of P2X2 receptor-mediated current (Fig. 2b). However, PMA produced a marked increase of the current amplitude mediated by the P2X3 receptor (Fig. 2c). In one experiment with four oocytes, a statistically significant increase of the P2X3-mediated current amplitude by PMA of 155 ± 42% (SEM) was determined. P2X2 and P2X3 receptors exist in a non-phosphorylated state in both X. laevis oocytes and HEK293 cells To explain the accelerated desensitization upon site-directed perturbation of the 18TXK20 motif, it has been suggested that the wild-type P2X2 receptor is constitutively phosphorylated at 18T and that this phosphorylation is responsible for the slow desensitization rate [19]. To address this issue, we next examined biochemically whether P2X3 receptors also exist in a constitutively phosphorylated state; the P2X2 receptor was also analysed to allow for a direct comparison of the results. Both parent P2X receptors and mutants were purified by Ni2+ chelate affinity chromatography from dodecylmaltoside extracts of [35S]methionine-labelled oocytes, resolved by SDS-PAGE, and blotted onto a PVDF membrane. However, probing with a phosphothreonine-specific monoclonal antibody revealed no discernible signal (Fig. 3a) despite the presence of significant amounts of [35S]methionine-labelled P2X2 and P2X3 protein in the SDS-PAGE gel (Fig. 3b). Fig. 3Immunoblots show no constitutive phosphorylation of oocyte-expressed P2X2 or P2X3 receptors. The indicated wild-type or mutant P2X subunits were purified by Ni2+ chelate chromatography from [35S]methionine-labelled X. laevis oocytes and resolved by denaturing SDS-PAGE. a Shown is a representative immunoblot probed with a phosphothreonine-specific monoclonal antibody and a peroxidase-conjugated secondary antibody. No phosphorylation signal could be detected at the SDS-PAGE migration position of P2X2 and P2X3 subunits, which are indicated by the adjacent PhosphorImager scan in panel b. b Direct PhosphorImager visualization of metabolically [35S]methionine-labelled parent and mutant P2X subunits resolved by SDS-PAGE of the same samples as in panel a. cLeft panel: representative immunoblot of P2X2 subunits tagged with either an N-terminal or C-terminal hexahistidine sequence (designated His-P2X2 or P2X2-His, respectively) and probed with a phosphothreonine-specific monoclonal antibody and a peroxidase-conjugated secondary antibody. No phosphorylation signal could be detected at the SDS-PAGE migration position of P2X2 subunits, which were visualized by immunoblotting with the P2X2 subunit polyclonal antibody (middle panel) or PhosphorImager scanning of incorporated [35S]methionine (right panel) To exclude the possibility that the apparent lack of receptor phosphorylation resulted from an interference between the N-terminal His tag and the PKC site of P2X3 or P2X2 receptor subunits, which are separated by only 12 or 20 residues, respectively, we performed analogous experiments with a C-terminally His-tagged P2X2 receptor, P2X2-His. As apparent from Fig. 3c, neither N-terminally nor C-terminally His-tagged P2X2 receptor became phosphorylated, as evidenced by immunoblotting with a phosphor-specific anti-threonine antibody (Fig. 3c). The presence of P2X2 receptor protein was verified by immunoblotting and metabolic labelling with [35S]methionine (Fig. 3c, middle and right panel). Boue-Grabot et al. [19] have demonstrated phosphorylation of P2X2 receptors expressed in HEK293 cells. To address the possibility that P2X receptor phosphorylation is host cell specific, we analysed HEK293 cells transiently transfected with P2X2 or P2X3 expression plasmids. P2X receptor proteins were isolated by immunoprecipitation and resolved by reducing SDS-PAGE. The efficient expression of the P2X3 receptor or the C-terminally EGFP-tagged P2X2 receptor in HEK293 cells was verified by immunoblotting with subunit bands migrating at 54 kDa or 83 kDa (27 kDa contributed to the P2X2 subunit by the EGFP tag), respectively (Fig. 4b). However, phosphorylated P2X2 or P2X3 subunits could not be detected by immunoblotting with a phosphothreonine-specific antibody (Fig. 4a). In contrast, the PKC-phosphorylated splicing factor SF3B1, which was expressed in HEK293 cells as a positive control substrate and immunoprecipitated with an EGFP antibody, was efficiently phosphorylated under identical conditions, assuring that the assay was able to detect phosphorylated proteins (Fig. 4a, lane 1). Taken together, these results suggest that both P2X2 and P2X3 receptors exist constitutively in a non-phosphorylated state in both X. laevis oocytes and HEK293 cells. Fig. 4Immunoblotting reveals no constitutive phosphorylation of HEK293 cell-expressed P2X2 and P2X3 receptors. The indicated proteins were isolated by immunoprecipitation using the peptide-specific P2X3 subunit polyclonal antibody or an EGFP-specific polyclonal antibody as appropriate and resolved by reducing SDS-PAGE. a Shown is an immunoblot probed by phosphothreonine-specific monoclonal antibody and a peroxidase-conjugated secondary antibody. Note that phosphorylation of the P2X3 receptor or the EGFP-tagged P2X2 receptor is not detectable, whereas a positive control—the EGFP-tagged splicing factor SF3B1 (migration mass 97 kDa [23])—was detected by the antibody. b The same samples as in panel a were immunoblotted using peptide-specific P2X2 and P2X3 subunit polyclonal antibodies to verify the expression of the transfected genes PMA-activated PKC does not drive the phosphorylation of P2X2 or P2X3 receptors A majority of protein kinases are quiescent unless their activity is stimulated by specific stimuli and effectors. To drive phosphorylation by active endogenous PKC, we incubated non-denatured recombinant His-tagged P2X1, P2X2 and P2X3 receptors isolated by Ni2+ NTA agarose affinity chromatography from dodecylmaltoside extracts of X. laevis oocytes with PMA-stimulated lysates of X. laevis oocytes (Fig. 5a) or HEK293 cells (Fig. 5b) in the presence of [γ-32P]ATP. As a positive control, a specific peptide PKC substrate RRGRTGRGRRGIYR was incubated with the lysates. This peptide was equally well phosphorylated by addition of either cell lysate or a commercial preparation of the rat brain PKC subunit (Fig. 5a, b). In contrast, however, no 32P incorporation could be detected in the SDS-PAGE migration position of the P2X1, P2X3 or P2X2 receptor subunit at 56 kDa, 54 kDa or 66 kDa, respectively. The presence of purified P2X1, P2X2 and P2X3 receptor subunit protein in this assay was verified by immunoblotting using P2X1-, P2X2- or P2X3-specific antibodies (Fig. 5c). The phosphorylation of the synthetic peptide clearly demonstrates that both X. laevis oocytes and HEK293 cells contain sufficient endogenous PKC to drive at least the phosphorylation of a synthetic peptide substrate. Fig. 5Lysates of X. laevis oocytes or HEK293 cells contain endogenous PKC, but do not support PMA-driven phosphorylation of P2X2 or P2X3 receptors. The indicated wild-type and mutant P2X receptors were purified by non-denaturing Ni2+ chelate chromatography from X. laevis oocytes, supplemented with [γ-32P]ATP, 100 nM PMA and oocyte lysate (a) or HEK293 cell lysate (b), and incubated for 30 min at ambient temperature. In parallel, a peptide PKC substrate (final concentration 0.5 μg/μl) was incubated with [γ-32P]ATP, PMA and either of the two cell lysates or the purified rat brain PKC catalytic subunit as indicated. The proteins were resolved by reducing SDS-PAGE and visualized by phosphorimaging. Note that the band pattern was similar irrespective of whether protein of mock-injected oocytes or P2X receptor-expressing oocytes was used as a substrate, arguing against specific labelling of either P2X receptor. Arrows indicate migration positions of P2X1, P2X2 and P2X3 subunits. Arrowhead in a, dominant background band that is present in all oocyte samples. The peptide PKC substrate was similarly labelled by lysates or purified rat brain PKC, confirming the presence of endogenous PKC in X. laevis oocytes and HEK293 cells. c The same samples as in panels a and b were resolved on a separate SDS-PAGE gel. The indicated P2X subunits were detected by immunoblotting using the appropriate P2X subunit-specific polyclonal antibodies. A sample of mock-transfected cells was simultaneously probed with the P2X2 and P2X3 subunit-specific polyclonal antibodies Discernible phosphorylation of P2X2 and P2X3 receptors might be obscured by a low level of endogenous PKC. Therefore, an analogous experiment was performed, in which the purified P2X2 and P2X3 receptors were incubated with purified rat brain PKC catalytic subunit rather than cell lysate in the presence of [γ-32P]ATP. Again, the specific peptide PKC substrate was efficiently phosphorylated as demonstrated both by direct phosphorimaging (Fig. 6a) and immunoblotting with the phosphothreonine monoclonal antibody (Fig. 6b). Phosphorylated P2X2 or P2X3 subunits could not be detected by either method (Fig. 6a, b) despite the presence of abundant amounts of the respective P2X subunits in the assays (Fig. 6c). Fig. 6PKC phosphorylates a specific PKC substrate in vitro, but does not phosphorylate P2X2 or P2X3 receptors. The immunoprecipitated recombinant P2X2 or P2X3 receptor from X. laevis oocytes or a synthetic peptide (1 or 10 μg) derived from the non-structural protein 3 of hepatitis C virus, which serves as a specific substrate for PKC, were incubated for 30 min with purified rat brain PKC catalytic subunit (0.02 μg corresponding to 0.02 U/μl) in the presence of 50 μM [γ-32P]ATP (10 mCi/ml), and then subjected to SDS-PAGE and phosphorimaging. The rat brain PKC catalytic subunit does not require Ca2+ and phosphatidylserine for activity [42]. 32P was incorporated into the PKC substrate (left panel) and was also detected by immunoblotting with an anti-phosphothreonine antibody (middle panel). However, there was no detectable 32P incorporation at a migration position of ∼66 kDa or ∼54 kDa of the P2X2 or P2X3 subunit, respectively, the presence of which could be verified by immunoblotting (right panel) PMA or calyculin A do not enhance phosphorylation of P2X3 receptors in intact HEK cells In a final set of experiments, we attempted to directly demonstrate phosphorylation in vivo by using metabolic [32P]orthophosphate labelling combined with immunoprecipitation analysis as a read out. Mock-transfected and P2X3 receptor-transfected HEK293 cells were incubated with [32P]orthophosphate in phosphate-free DMEM, followed by incubation with the PKC activator PMA and/or the phosphoserine/phosphothreonine phosphatase inhibitor calyculin A. The PhosphorImager scan revealed an identical pattern of phosphorylated bands almost irrespective of the experimental conditions, with no evidence for a significant 32P-labelling of the P2X3 subunit (Fig. 7, top panel). P2X3 receptor expression was verified by immunoblotting (Fig. 7, bottom panel). Fig. 7PMA and calyculin A do not induce phosphorylation of P2X3 receptors in intact HEK cells. Mock- and P2X3 receptor-transfected HEK293 cells were incubated with [32P]orthophosphate (500 μCi/ml) for 3 h in phosphate-free DMEM, followed by a 10-min incubation with 100 nM of each, the PKC activator PMA and/or the phosphatase inhibitor calyculin A. Triton X-100 extracts were then prepared from the cells and subjected to immunoprecipitation using the peptide-specific P2X3 subunit polyclonal antibody, followed by reducing SDS-PAGE and phosphorimaging. The migration position of the P2X3 subunit is marked by an arrow, indicating the absence of 32P incorporation into the P2X3 subunit. P2X3 receptor expression was verified by immunoblotting with the P2X3 subunit polyclonal antibody (bottom panel) Discussion Both P2X3 and P2X2+3 receptors are present on sensory neurons, where they play an important role in nociception. In addition to mediating transient nociceptive responses, these receptors contribute critically to the hyperexcitability that underlies abnormal pain states elicited by nerve injuries. One mechanism for the increased responsiveness to ATP is sensitization of P2X3 and P2X2+3 receptors arising from a significant enhancement in the trafficking of P2X3 subunit-containing receptors and/or an increase in function of pre-existing receptors [17]. An obvious candidate mechanism for functional modulation is phosphorylation, which is believed to occur universally in all ligand-gated ion channels as reviewed by Levitan [33]. Evidence for a control of P2X receptor channel function by PKC-mediated phosphorylation Site-directed elimination of the putative N-terminal PKC site has already previously been shown to severely affect the electrophysiological phenotype of several P2X receptor subtypes, including P2X1, P2X2 and P2X3 receptors from rat and/or human [16, 19, 20, 34] and P2X5 receptors from frog [35]. Replacing the putative phospho-acceptor residue 18T or the basic residue in the P+2 position (20K) by A or a non-basic residue, respectively, imposed rapidly desensitizing properties onto the non-desensitizing wt P2X2 receptor without grossly affecting the current amplitude. In contrast, when analogous mutations were introduced into the rapidly desensitizing receptor subtypes P2X1 or P2X3, peak currents declined to low or undetectable levels, respectively [16 ,20, 34]; the same observation was also made with the slower desensitizing frog P2X5 receptor [35]. Overall, the results obtained by site-directed elimination of the N-terminal TXR/K motif can be plausibly reconciled with the hypothesis that the rate of P2X receptor channel desensitization is controlled by a phosphorylation-dephosphorylation mechanism, with phosphorylation at this site being associated with a decreased rate of desensitization. Additional evidence for a possible involvement of PKC-dependent phosphorylation comes from experiments with the PKC activator PMA, which was consistently found to elicit a severalfold increase in currents mediated by the P2X1 receptor [36] or the P2X3 receptor [16, 18, 21] without changing the rate of desensitization. A PMA-induced change of the current desensitization was noted for C-terminally truncated P2X2 receptors, which are known to exhibit a desensitizing phenotype in contrast to the parent full-length P2X2 receptors [19, 37]. PMA treatment converted the rapidly desensitizing phenotype of truncated P2X2 receptors to slowly desensitizing, but did not affect the kinetic properties or current amplitudes of the full-length P2X2 receptors. The same lack of effect of PMA has been observed with wild-type P2X2 receptors stably expressed in HEK293 [38]. It has been inferred from these data that P2X2 receptors are constitutively phosphorylated and that this phosphorylation is responsible for their non-desensitizing phenotype. Consistent with this view, immunoblot analysis with a phosphothreonine-proline-specific monoclonal antibody efficiently detected wild-type P2X2 receptors with an intact PKC site, but not P2X2 receptor mutants with a disrupted PKC site [19]. Constitutive phosphorylation has been also demonstrated for the P2X1 receptor expressed in HEK293 cells [34]. Further studies demonstrated that exogenously added uridine triphosphate (UTP) potentiated the current mediated by endogenous and expressed P2X3 receptors in response to αβ-meATP as agonist. Substitution of phospho-acceptor residues of PKC consensus sites located in the ectodomain of the P2X3 receptor abolished this UTP-induced current potentiation, suggesting a role of ecto-protein kinase C in P2X3 receptor regulation [39, 40]. Evidence against a control of P2X receptor channel function by PKC-mediated phosphorylation In direct contrast to the results of Boue-Grabot et al. [19], we were unable to demonstrate a direct phosphorylation of wild-type P2X2 receptors expressed in X. laevis oocytes and HEK293 cells. The various experimental approaches for phosphorylation detection that we applied included immunoblot analysis of purified P2X2 receptors. In the experiments shown, we used a phosphothreonine-specific monoclonal antibody instead of a phosphothreonine-proline-specific antibody, because the TXK motif of the P2X3 receptor contains a threonine in the -X-position instead of the proline found in the amino acid sequence of the P2X1 and P2X2 subunit (cf. Fig. 1b). However, even when we used the same phosphothreonine-proline-specific antibody as Boue-Grabot et al. [19], we could not demonstrate phosphorylation of P2X2 receptors in our experiments (data not shown). Recently, it has been shown that a proline in the P+1 position strongly disfavours substrate recognition by PKC [41], suggesting that our negative results with P2X2 receptors may not result from technical limitations, but reflect the inherent inability of P2X2 receptors to serve as substrates for PKC. Notably, P2X3 receptors, despite harbouring a T in the P+1 position, could also not be demonstrated to become phosphorylated by PKC in our experiments, as well as in a recently published paper [21]. Further evidence against a direct PKC-mediated P2X receptor phosphorylation comes from co-expression studies of P2X1 and G protein-coupled receptors. Co-expression of the metabotropic glutamate receptor mGluR1α with P2X1 receptors produced a 2.5-fold increase of the P2X1 receptor-mediated current that was blocked by the PKC inhibitor staurosporine and could be mimicked by PMA treatment of singly expressed P2X1 receptors. Although the pharmacological data are basically consistent with PKC-mediated receptor phosphorylation being involved, mutant P2X1 receptors with a disrupted N-terminal PKC site (T18V) still showed PMA-mediated and mGluR1α-mediated potentiation. Also, PMA treatment did not lead to an increase in the (rather weak) basal phosphorylation of the P2X1 receptor [36]. Functional data supporting a role of PKC in P2X receptor regulation have been detailed above. There are, however, also functional data that are inconsistent with the view that the TXR/K motif represents a true PKC site: (1) replacement of the phospho-acceptor threonine by an acidic residue to mimic incorporation of a negatively charged phosphate group by constitutive phosphorylation was unable to restore the function of P2X1 receptors [34] and P2X3 receptors [16]; (2) the PMA-induced potentiation of hP2X1 receptor currents could be blocked by the PKC inhibitor staurosporine, but surprisingly, not by site-directed elimination of the PKC site [36]; thus it is much more likely that PMA exerts its effect on the P2X1 receptor (and the P2X3 receptor) through a yet to be identified accessory protein. Taken together, there is no doubt that the TXR/K motif plays an important role in the functioning of P2X receptors, and that PMA is capable of modulating P2X1- and P2X3-mediated receptor currents. The underlying mechanism, however, is unlikely to involve direct PKC-mediated phosphorylation of P2X receptors. The existence of a proline in the P+1 position in several P2X subunit isoforms (P2X1, P2X2, P2X4) might be one causative factor preventing direct phosphorylation by PKC. Additional steric constraints may explain why P2X3 receptors also do not serve as PKC substrates, despite lacking the proline in P+1. The exact role of the TXR/K motif awaits further clarification.	[ "phosphorylation", "p2x2", "p2x3", "protein kinase c", "p2x receptors" ]	[ "P", "P", "P", "P", "P" ]
Behav_Genet-3-1-1914288	Assessment and Etiology of Attention Deficit Hyperactivity Disorder and Oppositional Defiant Disorder in Boys and Girls	Attention deficit hyperactivity disorder (ADHD) and oppositional defiant disorder (ODD) are more common in boys than girls. In this paper, we investigated whether the prevalence differences are attributable to measurement bias. In addition, we examined sex differences in the genetic and environmental influences on variation in these behaviors. Teachers completed the Conners Teacher Rating Scale-Revised:Short version (CTRS-R:S) in a sample of 800 male and 851 female 7-year-old Dutch twins. No sex differences in the factor structure of the CTRS-R:S were found, implying the absence of measurement bias. The heritabilities for both ADHD and ODD were high and were the same in boys and girls. However, partly different genes are expressed in boys and girls. In both clinical and population samples, children diagnosed with attention deficit hyperactivity disorder (ADHD) and oppositional defiant disorder (ODD) are predominantly male (Gaub and Carlson 1997; Biederman et al. 2002; Loeber et al. 2000). These sex differences could either be the result of a higher liability for these disorders in boys than girls, or could be attributable to a sex effect in the actual measurement of the phenotype. For example, boys and girls with similar levels of problem behavior may receive systematically different rating scores if the items of the instrument do not reflect the problem behaviors in the same manner in boys and girls. This may conceptualized as bias with respect to sex. In this study, we will investigate this issue in the measurement of ADHD and ODD by means of the Conners Teacher Rating Scale-Revised:Short (CTRS-R:S) version. The presence of sex differences in the prevalence of ADHD and ODD also raises the question whether there are sex differences in the etiology of these disorders. Sex differences in etiology can only be interpreted if the measurement of a disorder is not biased with respect to sex. Lubke et al. (2004) discussed in detail the importance of establishing unbiasedness with respect to sex for the correct interpretation of any sex differences in the results of genetic modeling. The aim of this paper is, therefore, to investigate measurement bias in teacher ratings of ADHD and ODD with respect to sex. If unbiasedness can be shown to be tenable to reasonable approximation, we shall, as the second goal of this paper, estimate the genetic and environmental contributions to the phenotypic variance in ADHD and ODD in boys and girls. Mellenbergh (1989) defined unbiasedness, or equivalently, measurement invariance (MI), with respect to group to mean that the distribution of the observed test score, conditional on the latent construct that the test measures, is identical over groups (e.g., boys and girls). In more simple terms, this means that the instrument is measuring the same construct in boys and girls (Mellenbergh 1989; Meredith 1993). If this is the case, we expect the score of a given person to depend on that person’s score on the latent construct, but not on that person’s sex. If MI does not hold, a boy and a girl, who are characterized by the same degree of problem behavior, may obtain systematically (i.e., regardless of measurement error) different scores on the instrument. This is undesirable, because we wish our measurements to reflect accurate and interpretable differences between cases in different groups. Analyses aimed explicitly at establishing MI with respect to sex, according to the approach outlined by Meredith (1993), have yet to be conducted with respect to ADHD and ODD. Although MI has not been investigated, there have been some studies, which addressed sex differences in the factor structure in teacher ratings of ADHD. Fantuzzo et al. (2001) examined the factor structure of the 28-item version of the CTRS with exploratory factor analyses. They reported a three-factor solution, which accounted for 58% of the variance. The factors admitted interpretations in terms of conduct, hyperactivity, and passivity. The invariance of the factor structure was established by comparing results from random subgroups with the results from subgroups based on sex. The results supported the similarity of the factor structure across sex. A concern in this study is that the subjects in this study were 580 children from 33 classrooms located in low-income neighborhoods. It is, therefore, unclear how representative the sample is of the general population. Furthermore, because each teacher rated more than one child, and children were clustered into classes, the assumption of independent observations, which is important in statistical inference, might not hold. In this study, we conducted confirmatory factor analyses (CFA) of data from a large general population sample of 7-year-old twins, rated by their teachers on ADHD and ODD. Two questions are addressed. First is the measurement model that relates differences in the latent constructs of ADHD and ODD to the observed behavior problem scores identical in boys and girls, i.e., is MI tenable? Second, do the magnitudes of the genetic and environmental influences differ, or do different genes play a role in boys and girls? Methods Subjects and procedure This study is part of an ongoing twin study of development and psychopathology in the Netherlands. The subjects were all registered at the Netherlands Twin Registry (NTR; Boomsma et al. 2002, 2006). We assessed a sample of Dutch twins from the birth cohorts 1992–1996. These twins were assessed by their teachers when they were 7 years old. The twins at age 3 are representative of Dutch 3-year-old children with respect to their scores on measures such as the Child Behavior Checklist (CBCL; van den Oord et al. 1995). The socioeconomic status of the parents of the twins was somewhat higher than the level in the general Dutch population (Rietveld et al. 2004). When twins reached the age of 7 years, parents were asked to provide informed consent to approach the teacher. Consent was given by 80.1% of the parents. Teachers of these pairs received a questionnaire by mail, and were asked to return it to the NTR by mail. The response rate of the teachers was 80.0%. CTRS data were available for at least one twin in 1,651 twin-pairs (1,511 complete pairs). The maternal CBCL-AP scores at age 7 years were not significantly different between families in which parents provided permission to approach the teachers (mean = 2.95, SD = 2.93) and families in which parents did not do so (mean = 3.15, SD = 3.18; t(3,063) = 2.0, p = 0.133). However, mean maternal AP ratings were significantly higher in twins whose teachers did not return the questionnaire (mean = 3.34; SD = 3.13) than in twins whose teachers did return the questionnaire (mean = 2.78; SD = 2.81; F(1) = 16.82, p < 0.001). To avoid biased test results due to statistical dependency of the twin data, we randomly included the score from only one of the members of a twin-pair in the CFA. The resulting sample for CFA consists of 1,651 individual twins (800 boys and 851 girls). In the genetic analyses, we included all complete twin-pairs. Data were available for both members of a twin-pair in 248 MZ male, 251 DZ male, 294 MZ female, 234 DZ female, and 484 DZ opposite sex pairs. Some twins were rated by the same teacher (877 pairs, 58%) while the remaining twins were rated by different teachers (634 pairs, 42%). The genetic analyses accounted for potential differences between same and different teacher ratings by using the model developed by Simonoff et al. (1998). Incomplete twin-pairs were excluded from the genetic analyses. Measure The CTRS-R is a widely used instrument to assess behavior problems (Conners 2001; Conners et al. 1998). The CTRS-R was developed by factor analyzing a large set of items, and including items that load highly on interpretable common factors. In addition to the scales that were derived based on factor analysis, an ADHD index was included. This index comprises the best 12 items for distinguishing children with ADHD from children without ADHD as assessed by the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV; American Psychiatric Association 1994; Conners 2001). The short version of the CTRS-R, which was used in this study, contains 28 items from the following scales: oppositional (ODD, five items); cognitive problems-inattention (IN, five items); hyperactivity (HI, seven items); and the ADHD index (ADHD-I, 12 items). One of the items (item 27; excitable, impulsive) is included in both the hyperactivity scale and the ADHD index. The items are rated on a 4-point Likert scale for symptom severity (i.e., 0 = ”not true at all”; 1 = ”just a little true”; 2 = ”pretty much true”; 3 = ”very much true”). The internal consistency and stability of the CTRS-R:S version are good, as the Cronbach’s alpha coefficients range from 0.88 to 0.95, and 6–8-week test–retest correlations range from 0.72 to 0.92 (Conners 2001). Distribution of the items Because of the categorical nature of the item scores, Pearson product moment correlations underestimate the correlation of the underlying latent trait, and consequently the parameter estimates obtained in factor analyses or principal component analyses based on the Pearson correlation matrices are biased (Dolan 1994). We, therefore, adopted an approach that is suitable for factor analyzing discrete item scores. Polychoric correlations between items were obtained based on the liability threshold model (Lynch and Walsh 1998). In the case of a 4-point Likert scale, three thresholds divide the latent liability distribution into four categories. Criteria of MI The criteria of MI are empirically testable in the common factor model (Meredith 1993). MI criteria are: (1) equality of factor loadings over groups; (2) equality of item intercepts over groups (i.e., differences in item means are the result of differences in factor means), and (3) equality of residual variances (i.e., variance in the observed variables, not explained by the common factor) over groups. When satisfied, these restrictions ensure that any differences in the mean and variance of the observed variables are due to differences in the mean and variance of the common factor. For ordinal data, MI can be tested by constraining the thresholds and factor loadings, and residual variances to be equal in boys and girls. These constraints allow estimation of group differences in the means and variances of the common factor. To this end, the mean and variance of the common factor are fixed at 0 and 1, respectively, in an arbitrary reference group. We chose to estimate the mean and variance in girls, and to use boys as the reference group. In doing so, we are modeling the observed group differences as a function of the differences in the latent liability. Statistical analyses The polychoric correlation matrices of the items of the four subscales were calculated using Prelis (Jöreskog and Sörbom 1996). All CFA were performed on raw data using Mx (Neale et al. 2003). In principle, factor analysis of p ordinal items can be carried out using full information maximum likelihood (FIML) estimation. However, this requires repeated integration of the p-variate normal distribution, which can become computationally demanding even with as few as 12 items. We, therefore, estimated the model parameters using marginal maximum likelihood estimation (MML; Bock and Aitkin 1981). MML maximizes the likelihood of the data conditional on the latent trait, in contrast to FIML, which maximizes the unconditional likelihood. The advantage of MML compared to FIML is that it is computationally much less demanding. To test if the Conners scales are MI with respect to sex, the factor structure was constrained to be identical in boys and girls. The ODD, IN, and HI scales resulted from factor analyses, and a single factor was fit to these scales. In contrast, the ADHD-I contains items related to problems with inattention and hyperactivity, and thus a two-factor model was fitted. To detect prevalence differences in ADHD and ODD across sex, the means and variances of the latent factors were constrained to be equal in boys and girls. The fit of the models was compared by χ² tests, with a type I error probability set at α = 0.01. Browne et al. (2002) noted a complication of the χ² test. Specifically, they showed that χ² is influenced by the unique variances of the items. If a trait is measured reliably, the inter-correlations of the items are high, the unique variances are small, and the χ² test may suggest a poor fit even when the differences between the expected and observed covariance matrices are trivial. The standardized root mean square residual (SRMR; Bentler 1995) is a fit index that is not sensitive to the size of the correlations. To avoid the rejection of a simpler model due to high inter-item correlations, we only reject a model if a significant χ² test is accompanied by large residuals (SRMR > 0.08; Hu and Bentler 1999). After investigating MI with respect to sex, we look at sex differences in the genetic and environmental influences on the individual differences in the sum scores of the scales, given that MI is tenable (Lubke et al. 2004). The polychoric correlations of the four scales were calculated by sex and zygosity in PRELIS (Jöreskog and Sörbom 1996). The genetic analyses were performed on the raw data using Mx (Neale et al. 2003). With data from MZ and DZ twins, the variance in behavior can be attributed to genetic and environmental factors. In our sample, 58% of the twins were in the same classroom and 42% of the twins were in different classrooms. Correlations between twins may be higher when children are rated by the same teacher. Therefore, a correlated error model developed by Simonoff et al. (1998) was used to analyze the data. In this model, individual differences in behavior are explained by additive genetic factors (A), common environmental factors that are shared between two twins of a pair (C), and unique environmental factors (E). The unique environmental factors are allowed to correlate in twins who are placed into the same classroom, and do not correlate in twins who are placed into different classrooms. For the genetic analyses, the items of each subscale were summed, and the sum-score was recoded so that three thresholds divide the latent liability distribution into four categories. The thresholds were chosen in such a way that the categories contain more or less equal numbers of subjects. We preferred this procedure to the analysis of the raw sum scores, because these are skewed, and therefore cannot be analyzed with maximum likelihood approaches based on the assumption of normality (Derks et al. 2004). Sex differences in genetic and environmental influences were examined in two ways. First, we investigated if the estimates of the genetic and environmental variances are equal in boys and girls. Secondly, we investigated if the same genes influence phenotypic variation in boys and girls. These qualitative sex differences were evaluated by constraining the genetic correlation in opposite sex twins at 0.5 (similar as in same-sex DZ twins). If different genes play a role in boys and girls, the genetic correlation is expected to be lower than 0.5 in opposite sex twins. Results Measurement invariance We tested for MI by constraining the factor loadings, thresholds, and residual covariance matrices to be equal for boys and girls while allowing the factor means and variances to be different. The factor structure of ODD was MI with respect to sex (χ²(18) = 16.66, p > 0.10; SRMR = 0.01 and 0.06 in boys and girls, respectively). MI was also tenable for the ADHD-I (χ²(55) = 70.41, p > 0.05; SRMR = .03 and 0.05 in boys and girls, respectively). Both IN and HI showed statistically significant different factor structures in boys and girls (χ²(18) = 98.45, p < .001, and χ²(26) = 57.99, p < 0.001, respectively). However, the residuals between expected correlation matrices under the constrained and the non-constrained model were small (SRMR = 0.01 in girls and SRMR = 0.02 in boys, for both IN and HI). Apparently, the lack of fit was the result of the high inter-correlations between the items (Browne et al. 2002), and not of large residuals between the expected covariance matrices. Therefore, we tentatively accept MI with respect to sex. Table 1 provides the factor loadings and thresholds of the best-fitting models. We also included the factor loadings as reported by Conners (2001) to facilitate the comparison of our sample with the sample that was used to create the scales. Note that the factor loadings, as reported by Conners (2001), are generally much lower, as these estimates are based on the assumption of a continuous, normal distribution of the item scores, an assumption that is obviously violated in the instance of a four-category Likert scale. Table 1Promax rotated factor loadings and thresholds (T) of the best-fitting factor model for the four subscales of the CTRS-R:SItemsLoadings factor 1Loadings factor 2Loadings as reported by Conners (2001) (N = 1,897)T1T2T3Oppositional2. Defiant0.91—0.480.711.652.456. Defies0.88—0.501.022.002.7410. Spiteful0.69—0.301.492.323.1315. Argues0.84—0.510.821.802.5220. Explosive0.78—0.411.201.923.12Cognitive problems-Inattention4. Forgets things0.93—0.550.341.291.828. Poor spelling0.81—0.500.270.951.4413. Poor reading0.71—0.450.390.871.2118. Lacks interest0.70—0.471.011.652.2722. Poor arithmetic0.82—0.490.551.221.68Hyperactivity3. Restless0.73—0.620.100.941.517. Always on the go0.79—0.570.661.311.8711. Leaves seat0.82—0.580.621.371.8317. Difficulty waiting0.83—0.650.120.911.5821. Runs about0.79—0.441.211.892.5124. Difficulty playing0.87—0.520.341.172.0327. Excitable0.86—0.660.401.121.84ADHD indexAP1. Inattentive0.99−0.09—−0.360.761.4714. Attention span0.95−0.03—0.030.871.5516. Only pays attention0.450.23—0.361.201.8819. Distractibility0.94−0.01—0.591.542.2525. Fails to finish0.720.09—0.241.271.9926. Not follow instructions0.680.09—0.381.301.86HI5. Disturbs0.050.80—0.201.031.529. Remain still0.050.88—0.921.682.2412. Fidgets0.160.67—0.120.941.5323. Interrupts−0.050.85—0.541.372.0727. Excitable−0.090.91—0.461.151.8328. Restless−0.060.97—0.501.191.81AP attention problems; HI hyperactivityThe thresholds are constrained to be equal in boys and girls. The means and variances of the latent factors in boys are constrained at 0 and 1, respectively. The means and variances of the latent factors are freely estimated in girls. The variances in girls were not significantly different from 1. The means of the latent factors are estimated at −0.60 (oppositional), −0.86 (hyperactivity), −0.42 (ADHD-I AP), and −0.52 (ADHD-I-HI). The mean of the cognitive problems-inattention factor is not significantly different between boys and girls Genetic analyses Having established MI of the CTRS-R:S scales with respect to sex, we estimated the twin-correlations and carried out a genetic analysis of the data. Twin correlations are shown in Table 2, for same and different teachers, respectively. The genetic model fitting results of the four scales are reported in Table 3. All correlations are higher in MZ twins than in DZ twins, suggesting the presence of genetic influences on individual differences. The correlations are higher in twin pairs rated by the same teacher than for twin pairs rated by different teachers. This was taken into account by using a correlated error model (Simonoff et al. 1998). The lower correlations in opposite-sex DZ twins than in same-sex DZ twins suggest that different genes play a role in boys and girls. Table 2Polychoric twin-correlations of the CTRS rated by same teachers (ST) versus different teachers (DT)OppositionalCognitive problems-inattentionHyperactivityADHD indexSTDTSTDTSTDTSTDTMonozygotic males0.860.530.900.760.810.590.850.59Dizygotic males0.500.490.640.290.420.140.490.19Monozygotic females0.870.430.920.640.830.470.870.52Dizygotic females0.660.100.600.420.340.250.450.24Opposite-sex twins0.370.130.440.230.300.170.370.27Table 3Genetic model fitting results on CPRS-R:S ratingsModelParameters-2 LLWith modelΔdfΔχ²pOppositional behavior in 7-year-old children1. Fully saturated844779.36————2. Equal correlations same and different teachers784812.261632.91<0.0013. ACE model, estimates of ACE are allowed to be different between same and different teachers444817.9414038.580.5344. ACE, ACE same teacher = ACE different teacher404853.843435.90<0.0015. Correlated errors ACE model424819.8014240.440.5406. Correlated errors AE model404823.93524.130.1277. Correlated errors AE model, AE boys = AE girls384829.76625.830.0548. Correlated errors AE model, AE boys = AE girls, opposite-sex genetic correlation-free374821.86717.900.005Cognitive problems-inattention in 7-year-old children1. Fully saturated846785.90————2. Equal correlations same and different teachers786822.571636.67<0.0013. ACE model, estimates of ACE are allowed to be different between same and different teachers446815.9414030.040.8744. ACE, ACE same teacher = ACE different teacher406843.873427.93<0.0015. Correlated errors ACE model426815.8514229.950.9186. Correlated errors AE model406820.01524.160.1257. Correlated errors AE model, AE boys = AE girls386821.99621.980.3728. Correlated errors AE model, AE boys = AE girls, opposite-sex genetic correlation-free396817.50714.490.034Hyperactivity in 7-year-old children1. Fully saturated846880.07————2. Equal correlations same and different teachers786906.971626.90<0.0013. ACE model, estimates of ACE are allowed to be different between same and different teachers446914.5914034.420.7154. ACE, ACE same teacher = ACE different teacher406942.353427.76<0.0015. Correlated errors ACE model426924.1214244.050.3856. Correlated errors AE model406926.76522.640.2677. Correlated errors AE model, AE boys = AE girls386929.06622.300.3178. Correlated errors AE model, AE boys = AE girls, opposite sex genetic correlation-free376919.99719.070.003ADHD in 7-year-old children1. Fully saturated847420.21————2. Equal correlations same and different teachers787458.101637.89<0.0013. ACE model, estimates of ACE are allowed to be different between same and different teachers447472.5914052.380.0914. ACE, ACE same teacher = ACE different teacher407507.403434.81<0.0015. Correlated errors ACE model427477.7814257.570.0556. Correlated errors AE model407479.14521.360.5077. Correlated errors AE model, AE boys = AE girls387480.72621.580.4548. Correlated errors AE model, AE boys = AE girls, opposite-sex genetic correlation-free377475.93714.790.029 Model fitting analyses showed that variation in all four scales could be explained by additive genetic and unique environmental effects. The influences of the shared environment were not statistically significantly. The magnitude of the influences of genes and environment did not differ between boys and girls. The standardized estimates of genetic and environmental influences are shown in Table 4. Genetic effects explained 56–71% of the variation in the CTRS subscales. Unique environmental effects explained the remaining 29–44% of the variation. For all four scales, the genetic correlation was significantly lower than 0.5 in opposite-sex twins. This implies that different genes are expressed in males and females. The genetic correlation in opposite-sex twins was 0.16 for oppositional behavior, 0.35 for cognitive problems-inattention, 0.21 for hyperactivity-impulsivity, and 0.32 for the ADHD index. Table 4Standardized estimates of the genetic and environmental effects on problem behavior% Additive genetic effects (A)% Non-shared environmental effects (E)Genetic correlation opposite-sex twinsOppositional56440.16Cognitive problems-inattention71290.35Hyperactivity58420.21ADHD index61390.32 Discussion The purpose of this study was twofold. First, we investigated if teacher ratings on ADHD and ODD are measurement invariant with respect to sex. Secondly, genetic and environmental influences on variation in ADHD and ODD were compared between boys and girls. Measurement invariance Teacher ratings on ADHD and ODD were found to be measurement invariant with respect to sex. In other words, teacher assessments of these behavior problems relate to the same latent variables in boys and girls. Sex differences in observed scores on ADHD and ODD can, therefore, be interpreted as differences with respect to the latent construct. This supports the contention that the reported sex differences in ADHD and ODD (Gaub and Carlson 1997; Loeber et al. 2000; Maughan et al. 2004) are due to a higher liability for the disorder in boys than girls and not to measurement bias. Quantitative and qualitative differences in the heritability among boys and girls More than half the variance in ADHD and ODD in boys and girls is attributable to genetic influences. The remaining variance is attributable to unique environmental influences. The magnitude of the influences of genes and environment is the same in boys and girls. However, part of the variance in ADHD and ODD is attributable to different genes in boys and girls. We base this on the fact that the genetic correlation between DZ opposite-sex twins was significantly lower than 0.5, which is the theoretical value (in the absence of assortative mating), if the same genes influence behavior in boys and girls. We observed a genetic correlation lower than 0.5 in DZ opposite-sex twins for oppositional behavior, cognitive problems-inattention, hyperactivity, and the ADHD index. Few studies have addressed quantitative and qualitative sex differences in heritability estimates from teacher ratings. Saudino et al. (2005) reported qualitative sex differences in heritability of teacher ratings of hyperactive behavior in 7-year-old twins. They did not report any quantitative sex differences, which is in agreement with the current findings. Vierikko et al. (2004) report lower correlations in opposite-sex twins than in same-sex DZ twins of teacher ratings of hyperactivity-impulsivity in 12-year-old twins. However, both genetic and shared environmental effects were found to contribute to the phenotypic variance in these data. It was not possible to determine if the lower opposite-sex correlations were the result of sex-specific genetic influences or sex-specific shared environmental influences, although the presence of the latter appeared more likely. These findings disagree with the current results in the sense that we did not find any evidence for shared environmental influences. However, both studies suggest that teacher ratings are influenced by partly different etiological factors in boys and girls. The finding of different genetic influences in boys and girls in teacher ratings stands in contrast with results based on parental reports. In parent ratings, qualitative sex differences are not found for attention problems (Rietveld et al. 2004) or ODD (Hudziak et al. 2005). The different findings in parent and teacher ratings may be explained by the fact that the behavior of children depends on the context in which they are observed. Apparently, inattentive, hyperactive, and oppositional behavior of boys and girls are influenced by partly non-overlapping factors at school, while this is not true for these behaviors at home. The finding of sex-specific genetic variation has implications for gene-finding studies of ADHD and ODD. The fact that the genes which influence the behavior of boys and girls do not completely overlap indicates that some quantitative trait loci may explain variation in boys but not in girls and vice versa. Therefore, the data from boys and girls cannot be collapsed when studying genetic effects in teacher ratings. In the NTR, teacher data are collected at the ages 7, 10, and 12 years. The sample sizes at the ages 10 and 12 are currently relatively small. In the future, we plan to address the issue of qualitative sex differences in teacher ratings in a longitudinal framework. The results of such a study will reveal if the finding of sex differences in the specific genes that play a role is also present in older children. Another important issue that may be addressed is the MI of ADHD with respect to age. The results of this study should be interpreted in the light of the following limitations. First, we did not replicate the factor structure of the CTRS-R:S by means of exploratory factor analyses of the 28 items. To take the ordinal nature of the data into account, we used the liability threshold model (Lynch and Walsh 1998). We limited the number of common factors to keep the computational burden manageable. Therefore, we performed CFA, in which we assumed that the items are correctly assigned to the four scales and that cross-loadings are absent. Second, teacher ratings were shown to be measurement invariant with respect to sex, but this finding may not generalize to parent ratings. The correlations between Conners parent and teacher ratings are small to moderate with a range 0.18–0.52 (Conners 2001). It has been shown that parents and teachers rate partly different aspects of the child’s behavior (Derks et al. in press; Martin et al. 2002). Future studies will reveal if MI is also tenable in parent ratings. Assessment of ADHD and ODD symptoms, through teacher reports on the CTRS-R:S, provides a solid starting point for measuring sex differences in mean scores or in heritabilities. Variation in teacher ratings of children’s problem behavior is mainly influenced by genetic factors. The size of the genetic influences does not depend on the child’s sex, but partly different genes are expressed in boys and girls. Future studies should reveal if these findings generalize to children from different age groups.	[ "adhd", "odd", "sex differences", "measurement invariance", "conners teacher rating scale" ]	[ "P", "P", "P", "P", "R" ]
Bioinformation-1-5-1891680	A web server for transcription factor binding site prediction	Promoter prediction has gained increased attention in studies related to transcriptional regulation of gene expression.We developed a web server named PMSearch (Poly Matrix Search) which utilizes Position Frequency Matrices (PFMs) to predict transcription factor binding sites (TFBSs) in DNA sequences. PMSearch takes PFMs (either user-defined or retrieved from local dataset which currently contains 507 PFMs from Transfac Public 7.0 and JASPAR) and DNA sequences of interest as the input, then scans the DNA sequences with PFMs and reports the sites of high scores as the putative binding sites. The output of the server includes 1) A plot for the distribution of predicted TFBS along the DNA sequence, 2) A table listing location, score and motif for each putative binding site, and 3) Clusters of predicted binding sites. PMSearch also provides links for accessing clusters of PFMs that are similar to the input PFMs to facilitate complicated promoter analysis. Background Transcription factors play a pivotal role in the regulation of gene expression by sequence specific binding to the promoters of target genes. Prediction of putative transcription factor binding sites has become an important resource to explore genome organization and regulatory mechanisms. The binding specificity of transcription factors are usually represented by known sequence motifs (consensus sequences) or matrices (PFM (Position Frequency Matrix) or PWM (Position Weight Matrix)). [1] High throughput analyses using SELEX (Systematic Evolution of Ligands by Exponential Enrichment) and CHIP-Chip (Chromatin Immunoprecipitation-Microarray) along with computational sampling methods have generated thousands of PFMs. These data together with the related transcription factor information are curated in online databases such as Transfac [2], JASPAR [3], etc. Online applications, such as MatInspector [4,5], MATCH [6] and ConSite [7], have been built to utilize PFMs to predict transcription factor binding sites (TFBSs) embedded in promoter sequences. Many of the servers are comprehensive but lack the information on transcription factors whose binding specificities are similar to the input PFMs, which could effectively assist regulatory module finding. To provide such information and to meet the needs for efficient prediction tasks in the study of gene regulation network, we developed a server named PMSearch (Poly Matrix Search) for predicting TFBSs in DNA sequences with the novel functions of fetching similar PFMs and processing multi-forms of input motifs. PMSearch made the following improvements: 1) It has a more succinct and friendly user-interface. 2) It provides user with the convenience of customizing any set of PFMs from the local dataset. 3) It generates a resizable plot that shows the distribution and scores of predicted binding sites in variable scales. 4) It provides clusters of PFMs similar to the input PFMs. Methodology We have implemented a scoring scheme adapted from previous algorithms. [4,6] The uninformative nucleotides (ambiguous letters: N, B, D, H, and V) on either end of an input PFM are discarded before searching to enhance efficiency and specificity. We downloaded 507 PFMs from JASPAR and Transfac Public 7.0 to construct a local dataset from which the user may take any subset of PFMs, or along with userdefined PFMs, for a prediction task. In addition to PFMs, the server could also accept sequence motifs (modeled in the format of consensus sequences), which will then be converted into pseudo PFMs and applied in subsequent prediction. We have set a default global cutoff value of 0.85 previously adopted by TFSEARCH. [9] The user may adjust this cutoff if required. User interface The web interface of PMSearch helps in promoter analysis. The inputs are composed of PFMs and sequences under analysis. The user may submit PFMs, motifs or retrieve desired PFMs from our dataset using keywords (e.g., name of the transcription factor), its source database accession number (e.g. Transfac Matrix Accession number M00001) or our local accession number (e.g. X00001). Sequences in FASTA format or Genbank format are acceptable and will be auto-parsed. When the user initiates a predication task, each sequence will fork an independent task in which corresponding results can be retrieved respectively. In the output, PMSearch plots a figure that illustrates binding sites that are scored above the cutoff value (Figure 1). The user may specify a sub-region on the sequence that will be plotted for more detailed view, such as a specific promoter region in a sequence file. Other results including scores, sequence motifs and closely located putative binding sites (clusters) are also listed. In addition, the server provides a hyperlink for the user to fetch PFMs that are related to the input PFMs by implementing a gap-allowed alignment algorithm. A comprehensive instruction for users is available online. Conclusion PMSearch is an easy-to-use and efficient tool that utilizes PFMs to predict transcription factor binding sites in DNA sequences. It offers user the flexibility to search for putative TFBS with any set of PFMs. PMSearch outputs a plot demonstrating distribution of predicted binding sites and a table of the locations, scores, motifs and clusters of predicted binding sites. In addition, PMSearch provides the user access to PFMs that are relative to an input PFM for more sophisticated promoter analysis, as the predicted binding sites of one transcription factor could also be bond with other transcription factors that share the similar binding specificity. Such information may give hints to untangle the composite transcription regulatory network. We propose to update PFMs in our local dataset regularly. The source codes are available from the authors upon request. Caveats It has been proposed that many of the predicted TFBSs lack biological function in vivo. [10] This could result from 1) The predicted site is located in a context which is insufficient to facilitate transcription factor binding, 2) The low specificity of the input PFM gives rise to large portion of false positive predictions. We suggest the users prepare a list of certain transcription factors that are suspected to regulate the target gene to specify a prediction task.	[ "web server", "transcription factor binding site", "motif", "position frequency matrix" ]	[ "P", "P", "P", "R" ]
J_Digit_Imaging-3-1-2039858	Mastering DICOM with DVTk	The Digital Imaging and Communications in Medicine (DICOM) Validation Toolkit (DVTk) is an open-source framework with potential value for anyone working with the DICOM standard. DICOM’s flexibility requires hands-on experience in understanding ways in which the standard’s interpretation may vary among vendors. DVTk was developed as a clinical engineering tool to aid and accelerate DICOM integration at clinical sites. DVTk is used to provide an independent measurement of the accuracy of a product’s DICOM interface, according to both the DICOM standard and the product’s conformance statement. DVTk has stand-alone tools and a framework with which developers can create new tools. We provide an overview of the architecture of the toolkit, sample scenarios of its utility, and evidence of its relative ease of use. Our goal is to encourage involvement in this open-source project and attract developers to build off and further enrich this platform for DICOM integration testing. INTRODUCTION Digital Imaging and Communications in Medicine (DICOM)1 plays a major role in the health care information technology (IT) field as the standard for medical images and communication throughout the hospital. With the organization of the DICOM Standards Committee in 1996 and the support of major medical groups and imaging vendors worldwide, DICOM has become a dominant integration mechanism in the hospital enterprise. The DICOM standard version 3.0 contains 16 parts (more than 2,000 pages), and the standard itself is constantly evolving as new software and imaging technologies are developed. Mastering a technology like DICOM can be a daunting task, but the successful student will be on the right path with the DICOM Validation Toolkit (DVTk) from DVTk.org. Consider DVTk the lab kit that is handed out with the DICOM standard on the first day of DICOM 101 class. As with any other technology, true learning and understanding require hands-on experience.2 Although the standard itself does not define or identify testing or validation procedures to assess conformance, a number of third-party tools have been developed to fill that role. DVTk is a powerful tool for mastering the intricate DICOM file format and transfer syntax. The mantra of DVTk is to make DICOM easy. If a picture archive and communication system (PACS) adminsitrator, for example, is having a problem with imaging device integration, DVTk can be the first line of defense in tracking the problem. When vendors are pointing fingers at one another, DVTk can help them past recriminations and on to real solutions. For medical software developers, integrators, and testers, DVTk can help to more quickly produce robust systems. The DICOM Validation Tool (DVT) test framework is a flexible architecture and uses service–object pair (SOP) class definition files, making it adaptable as the DICOM standard evolves. It includes a graphical user interface (GUI) and a command line interface, DICOM media validation, service class user (SCU) and service class provider (SCP) emulators, and a rich scripting language. For more advanced testing, VBScript (Microsoft Visual Basic Scripting edition; Microsoft; Redmond, WA) or JScript (Microsoft) can be executed by DVT, and a set of .NET assemblies is available for developing stand-alone test tools with languages such as Visual Basic.NET (Microsoft) and C# (Microsoft). If a validation tool is to be taken seriously, it must be vendor neutral. With the release of DVT 2.1 in 2005, DVTk now exists as an open-source community project. Not only does this encourage more vendors to contribute (because of the incentive of reduced development and integration costs), but the adoption and proliferation of standards make it easier for individual developers to contribute to open-source projects such as DVTk. Independent software developers are sometimes discouraged from proceeding because of the chance that their work will provide a “one-off” solution only and not be widely used. DVTk, with the backing of the mature DICOM standard, is attracting talented developers with a desire to create something useful and lasting. This success assures that time devoted to learning this toolkit will not be wasted. This article focuses primarily on the DVT main application. We provide background on the toolkit and specific examples about the two intended roles of DVT: service and development. Current efforts within the DVTk project and future directions are also highlighted. DVTk HISTORY AND ARCHITECTURE In 2000, Agfa HealthCare (Mortsel, Belgium) and Philips Medical Systems (Eindhoven, The Netherlands) decided to coordinate activities around DICOM validation testing by bringing together efforts already started by both companies under a joint DVT project. The intention was to produce a DVT that could not only be used internally by both companies to test their own products but also made freely available to other original equipment manufacturers (OEMs) as a means of testing their products to the same level of detail. The ultimate aim was to reduce the time spent integrating proprietary systems by first exposing these systems’ equipment to tests run using DVT. DVT project has a steering committee with responsibility for guiding legal, technical, and commercial aspects. The steering committee meets every 6 months to discuss past progress, current issues, and future requirements. A project manager was elected by the steering committee to manage the DVT project on a daily basis and report back to the committee. Development tasks were divided up based on the available skills of developers who report to the project manager. In its first years, Agfa and Philips provided the personnel to staff the DVT project. In September 2005, DVT was made into an open-source project (http://www.dvtk.org, DICOM Validation Toolkit; last accessed May 2007) under SourceForge (http://sourceforge.net/projects/dvt, SOURCEFORGE.NET ; last accessed May 2007) as DVTk and is licensed under the GNU Lesser General Public License (LGPL). The steering committee decided that the time had come to begin promoting DVTk to a wider audience, with the aim of attracting other companies who might join and supply development resources. Around this time, ICT Healthcare (Eindhoven, The Netherlands), which had already been supplying development resources to the project, joined as a full member with representation on the steering committee. The goal was to make DVTk the independent gold standard for DICOM validation and thereby improve the interoperability of all vendor products using the DICOM interface. The dvtk.org Web site is now the location for the latest downloads, defect tracking details, and forums on using DVTk. Interested individuals and companies may join the project through the Web site. A weekly project telephone conference coordinates the activities of the development team. DVTk-based applications include DVT, the main application, which together with the core forms what is now referred to as the DICOM Validation Framework (Fig. 1). The core includes a DICOM testing data model and object-oriented class structure. The DVTk and DVTkData libraries provide access to this data model and class structure via the managed code adapter. The core includes the DICOMScript language that is supported by DVT. This language can be separated into basic programming, which includes the SEND and RECEIVE commands for simulating DICOM SCPs and SCUs, and advanced programming. The advanced programming language includes commands such as SYSTEM, for executing operating system native applications, and a number of commands for working with the Data Warehouse feature. The Data Warehouse is a run-time memory structure in which the user can store Association Control Service Element (ACSE) requests and responses, DICOM commands, and DICOM objects for reuse across test scripts and sessions. Fig 1.DICOM Validation Framework. For more advanced testing scenarios, DVTk includes the Script Support library and the High-Level Interface (HLI) library. The HLI library is a newer abstraction built on top of the core that makes it easy to write multithreaded tests. The application program interface (API) exposed by this library encapsulates many of the low-level core API classes and methods that make writing VBScripts similar to writing scripts in DVTk’s native DICOMScript language. VBScripts can be executed entirely within the DVT GUI application or command-line executable or debugged in Visual Studio .NET. Other DVTk-based applications built on the core and currently available on dvtk.org include: DICOM Network Analyzer, a network sniffer and DICOM protocol analyzer; DICOM Editor, for displaying and editing DICOM files; DICOM Compare Tool, for comparing the attributes and values of two DICOM files; DICOM Attribute Validator, for validating DICOM files, including Structured Report objects, against definition files; DICOM File Stripper, for removing all but mandatory attributes from a DICOM file; and DICOM File Anonymizer, for removing patient and physician information from a DICOM file. To date, the following contributions have been made to the DVTk project by outside parties, companies, and/or institutions: Medical Communications (UK) provided the underlying Transmission Control Protocol/Internet Protocol (TCP/IP) Capture File to the DICOM protocol data unit (PDU) conversion utilities used by the DICOM Network Analyzer application.The National Institute of Standards and Technology (NIST; Gaithersburg, MD) provided the Health Level Seven (HL7)3 validation Web services used in the latest DVTk HL7 validation components. GETTING STARTED WITH DVTk OUT OF THE BOX The latest version of DVT can be downloaded from http://www.dvtk.org . It comes packaged as a Microsoft Windows InstallShield application. The installation subfolder includes a user’s guide and Windows help files for the extensible .NET assemblies (available only if MS Visual Studio .NET is installed). A large number of examples on how to use the many features of DVT are included in the subfolder. Starting DVT presents the user with an empty workspace. DVT is designed to work on a single project at a time, although multiple views of the project may be opened from which multiple tests may be simultaneously executed. A DVT project is a container for one or more test sessions. A session is a container for the configuration of one or more tests to be performed against a system under test (SUT). Project and session configuration properties are stored in flat files. The DVT GUI exposes most of the configuration properties, although some of the more advanced settings, such as STRICT-VALIDATION, are available only by directly editing the session file. Some settings can also be modified in script files. For example, the STRICT-VALIDATION script command overrides the value of the STRICT-VALIDATION session property. Some settings, such as CALLED_AE_TITLE, also have built-in default values that are assumed if the property is not defined in a script or session file. Session files come in three types: emulator, script, and media. Emulator sessions are used when DVT should act as an SCU or SCP emulator to test the DICOM transfer syntax. The emulators have support for Verification SCU/SCP, Storage SCU/SCP, and Print SCP. Script sessions are used when DVT is used to execute a DICOMScript, DICOMSuperScript, or VBScript. A DICOMSuperScript is simply a script that calls one or more DICOMScripts. Script sessions have support for network SCP/SCU message exchange and DICOM media file creation. Media sessions are used to validate the DICOM file format as a DICOMDIR and/or DCM media files. When validating a DICOMDIR file, any referenced files are also validated. An easy way to gain familiarity with DVT is go through the examples in the subfolder using DVT as both the test tool and the SUT. Opening up the example project displays a tree view of all test sessions defined in the project and provides details on the session currently selected in the tree. Clicking on the Window menu and selecting the New Project View and Tile item displays two views of the project. The next example will walk through the Modality Worklist (WLM) SCP/SCU test script sessions. Selecting the WLM_SCP session in the top view and the WLM_SCU session in the bottom view presents a view from which two test sessions may be executed simultaneously (Fig. 2). Fig 2.DVT-Worklist Management SCP and SCU test sessions. At the top of the Session Information tab are general settings, including the session type, session ID (used in uniquely naming results files), and location of files used/created by the test session. The DVT Roles Settings section defines the Application Entity (AE) title and some connection settings that DVT assumes during the test session. The SUT Settings section defines the AE title and some connection settings, including the TCP/IP address, of the system the session is testing. Because this example uses DVT as both the testing system and the SUT, the DVT role settings in the top view are synchronized to the SUT settings in the bottom view, and the SUT settings in the top view are synchronized to the DVT role settings in the bottom view. Locations of the SOP class definition files required for the test session are selected under the Specify SOP Classes tab. The selected definition files are loaded into memory when the test session starts, and DVT uses these to validate the DICOM messages and objects it sends and receives. A definition file describes a single DICOM SOP class in terms of the combination of DICOM Message Service Element (DIMSE) commands and Information Object Definitions (IODs) that make up the SOP class. In this case, both test sessions specify the Modality Worklist Information Model–FIND SOP Class item. The standard definition files that come with DVT in the definitions subfolder are taken directly from the DICOM standard parts 3 and 4. Private definition files can be made that extend the validation capabilities of DVT by copying one of these standard files and modifying it with private user IDs (UIDs), modules, and attributes. A typical customization might specify that when a particular SUT is known to always send a value for the Patient Name (0010,0010) attribute, the corresponding definition file can be modified to define the Patient Name attribute as a type 1 (mandatory, nonzero-length) attribute. Opening the WLM_SCP tree node in the top view and selecting the DICOMScript, 1.ds, displays a Script tab in the right side of the view. This script defines the test steps for the SCP. Simple DICOMScripts such as this can be written using only the SEND and RECEIVE commands. A large number of additional commands are available for more advanced testing scenarios. The Script tab is read-only, but Windows Notepad can be launched from a context menu on the script name in the tree view. A handy DVT DICOM script reference (dvtDICOMscript.hlp) file is in the docs folder. This script first executes a RECEIVE command for an association request message that specifies a single presentation context, consisting of the Modality Worklist Information Model–FIND SOP Class and three possible transfer syntaxes. The SCU script in the bottom view SENDs an association request message. The only required parameter for the ASSOCIATE-RQ message is the presentation context. For the SCP’s RECEIVE command, the ASSOCIATE-RQ message is referred to as the reference or expected object. This will be compared to the received ASSOCIATE-RQ message. DVT will perform validation in two steps: The received object is first validated against any loaded definition files. This step ensures that the correct attributes are present in the object and that they are encoded correctly.This step is optional and applied only if a reference object is present in the script. Checks made here are that the expected number of attribute values has been received and that each attribute value matches the corresponding reference object attribute value. Often, it is not known with what values an SUT will respond, in which case no attributes can be specified in the reference object, so that only step one validation occurs (http://www.dvtk.org “DVT User Guide”, version 2.1, August 2005). The VALIDATION script command determines how the DICOM objects are validated by DVT. By default, validation is ENABLED, meaning full validation will occur. The value of the STRICT-VALIDATION session property determines how the result of the validation is handled by DVT. If STRICT-VALIDATION is enabled, the presence of attributes in the received message must match the definitions exactly. If they do not match, then DVT reports a FAILED validation and aborts further DICOMScript interpretation. If STRICT-VALIDATION is disabled and attributes do not match, then DVT reports a WARN message. By default, STRICT-VALIDATION is disabled. Following a successful association validation and negotiation, the SCU script SENDs a C-FIND-RQ message of type Modality Worklist–FIND. The SCP script RECEIVEs the C-FIND-RQ, validates it, and proceeds to SEND three hard-coded C-FIND-RSP messages. The SCU script was written to mirror the SCP script, so the three reference C-FIND-RSP messages it defines match exactly with those sent by the SCP. These scripts also demonstrate two DVT scripting features: Value Mapping and Value Representation (VR) Keywords. Value mapping allows the user to substitute a label for a value defined in a script, generated by DVT, or received from the SUT, and then refer to that value with the label further down in the script. The LABEL: keyword is used to map an attribute value defined in the script or received from the SUT. The NEW: keyword tells DVT to generate a new value and assign it to the given label. An example of this is in the SCP script, where the Study Instance UID attribute (0020,000D) is assigned a new value generated by DVT and named with the label StudyInstanceUid1. The generated UID value can be referred to subsequently in the script as StudyInstanceUid1 (although this was not done in this example). VR keywords, such as the AUTOSET keyword used in both the SCU and SCP scripts, tell DVT to generate a value and assign it to the corresponding attribute. In the example case, AUTOSET is used in the Scheduled Procedure Step Start Date attribute to ensure that values match between the two scripts. Both Value Mapping and VR Keywords are sensitive to the type of attribute to which they are applied. This test is begun by right clicking on the SCP script and selecting execute. When DVT is executing a test session, all tabs except the Activity Logging tab are hidden, the session tree in the corresponding view is disabled, and the test stop button in the toolbar is enabled. The SCP’s Activity Log will indicate that it is waiting for a connection on port 104. Executing the SCU script in the lower window will result in a number of messages in the Activity Logs, followed by the termination of the test sessions. Comparing and correlating the SCP and SCU activity log entries provides a good visual reference for the DICOM message flow between two AEs. By default, the AUTO-TYPE-2-ATTRIBUTES session property is set to true in the session file, which means that DVT will automatically add any zero-length type 2 (type 2 attributes must be included; however, they may be encoded with a zero-length value or no value) attributes from the definition file to the dataset before sending to the SUT. This behavior is recorded in the SCU’s activity column by the “Automatic Type 2 Attributes population...” statement. This feature means that only type 1 (required, nonzero-length) attributes must be explicitly stated in the scripts. After completion of the scripts, each view displays the Validation Results tab, where the results of the test sessions are displayed. DVT stores the results of each test in a _res.xml file in the test session’s configured results directory. The file name takes the form <Detail∣Summary>_nnn_<scriptName>_res.xml, where nnn is the session ID from the session properties. Varying the session ID from one test execution to the next allows storage of multiple sets of results in the results directory. Two session Boolean properties define whether summary or detailed results are generated: SUMMARY-VALIDATION-RESULTS and DETAILED-VALIDATION-RESULTS. If the test session’s STORAGE_MODE property is set to as-media or as-dataset, any media files received by DVT during the test are also stored in the results directory, again using the session ID to help uniquely name the files. The generated results files are listed in the session tree under the script name. Selecting one of the results files displays it in the Validation Results tab. The Validation Results tab is a Hypertext Markup Language viewer allowing navigation between the summary results, detailed results, and any generated media files using hypertext links. DVT does not include a DICOM file viewer. To automatically view a generated .dcm file by clicking on a link in the Validation Results tab, a DICOM image viewer must be installed and associated with the .dcm file type. The SCP’s Validation Results tab will show a test result of PASSED, and a .dcm media file is created that contains the Modality Worklist–FIND dataset received in the C-FIND-RQ message sent from the SCU. The SCU’s Validation Results tab also shows a result of FAILED, with nine errors reported. The Summary Results File lists the errors; in this case, three type 1 attributes are missing from each C-FIND-RSP message sent from the SCP. Clicking on a Link to Detailed Result link displays the errors in the context of the complete C-FIND-RSP message. The detailed results file contains the complete contents of each message sent and received in chronological order. Any comment lines in the DICOMScript that begin with ## are copied directly to the detailed results file. This is a good way to insert additional test information directly into the results. For additional troubleshooting output in the Activity Logging tab and detailed results file, the LOG-RELATION, LOG-DEBUG, LOG-DULP-STATE, and PDU-DUMP test session properties can be enabled. In the example described here, it is left as an exercise for the user to add the missing type 1 attributes to the SCP script so that the SCU test session result becomes PASSED. DICOMSuperScripts (script files with a .dss extension) enable the reuse of DICOMScripts in various test scenarios. The storage example script sessions included with DVT demonstrate the benefits of DICOMSuperScripts. DVTk FOR SERVICE TROUBLESHOOTING As demonstrated in the previous section, DVTk can be a useful tool when troubleshooting modality worklist problems. Another common use case would occur when adding new modalities to a PACS network. It can be frustrating to ensure that AE titles, ports, and IP addresses match between the modality and PACS configuration. This section will illustrate ways in which DVT can be used to troubleshoot a PACS-modality interface problem and describe another, more service-orientated tool available from DVTk.org. Because DVT can act as both an SCP and SCU, it is an excellent starting point for troubleshooting PACS-modality problems. One of the most common and most frustrating failures is when a modality does not send images to the PACS. Most modalities provide little or no error information when image storage problems occur with the PACS. The PACS system is often equally unhelpful. Without an error message of some kind, PACS administrators have no guidance on where to begin in addressing the problem. DVT can simulate the modality or the PACS to pinpoint the source of the problem while obtaining hard evidence that can be used to engage the involvement of additional service layers, such as the hospital networking group or modality vendor. The Emulator_1 test session in the example project that comes with DVT installation can be used to emulate the problem modality. The problem modality’s AE title and PACS connection properties must be copied to the Emulator_1 session’s properties. Although not absolutely necessary, the modality being emulated should be taken off the network to prevent any AE title conflicts during troubleshooting. The first test to perform is a simple Verification to ensure that there is DICOM network communication between the modality AE and the PACS. The test is started by opening the Emulator_1 session’s tree node, right clicking on the Storage SCU Emulator node, and selecting the Execute menu item. All tabs except the Activity Logging tab are hidden, and a dialog box is displayed. Clicking on the Echo button tells DVT to send a C-ECHO message to the SUT. The activity logging tab will display the operations performed. For the purposes of this article, the tests are run against the open-source DICOM PACS DCM4CHEE (http://www.dcm4che.org, dcm4chee-2.x DICOM Clinical Data Manager system; last accessed May 2007). If the verification test succeeds, then one can assume the patency of DICOM network connectivity between the modality and PACS. The next test is to attempt image storage to the PACS. This test requires a set of DICOM test images, preferably from the modality that is experiencing the problem. The images should contain no names, IDs, or UIDs that will conflict with real-world data. It is important, however, that the data used to perform the test matches the real-world data in structure as closely as possible, including any private data elements the modality creates. A copy of the modality’s DICOM conformance statement comes in handy here. This will help simulate the behavior of the modality as closely as possible. Under the Specify Transfer Syntax (TS) button on the Session Information tab, all transfer syntaxes supported by the modality as a storage SCU can be selected. For example, to test a computed tomography (CT) modality, one would select the CT Image Storage SOP Class (1.2.840.10008.5.1.4.1.1.2). For this test, the storage SOP class and transfer syntax that match the DICOM test images that will be stored to the PACS should be selected. Note that this step is not actually necessary; when emulating a storage SCU, DVT will automatically add to the association negotiation the transfer syntaxes and SOP Classes from the DICOM media files being stored. Selecting the correct transfer syntaxes and SOP Classes is necessary when emulating a storage SCP, however. Executing the Storage SCU Emulator opens the Storage SCU Emulator dialog box. In the dialog box, the Add button is used to add the test images to the list that will be sent to the SCP. Selecting the Validate before export option tells DVT to validate the media files against the corresponding definition file prior to sending them. The number of associations the modality uses when sending multiple images is also specified here (most modalities would send multiple images on a single association). Finally, clicking the Send button will tell DVT to execute the storage test and close the dialog. DVT’s detailed results will show the verification of each media file followed by each storage transaction. The Storage SCU emulator does not automatically do storage commitment as the SCP emulator does. To simulate the modality performing a storage commitment request, one could create a DICOMScript session similar to the Commit_SCU example session, modifying the SOP instance UIDs to match the images stored. This test should be performed immediately after the storage test. To demonstrate some of the more advanced DICOMScript features and the Data Warehouse in DVT, one can imagine a scenario where a CT modality has recently undergone a software upgrade. Since the upgrade, technicians have been reporting that although they can successfully postprocess images on the modality before sending them to the PACS, processing of images retrieved from the PACS back to the modality fails. No errors have been reported by the PACS or modality on storage or retrieval. One possible reason is that the PACS is doing something to the stored images that is preventing the modality from processing them. To evaluate, one can perform what is called a PACS transparency test. The idea is to make what the PACS is doing (if anything) to the images transparent by comparing the before and after images. This test will require a DICOM CT image file from the modality on the DVT workstation and a DVT project with two sessions. A script session playing the role of SCU is used to put the image in the DVT Data Warehouse, store it to the PACS, and initiate a C-MOVE from the PACS back to DVT. An Emulator SCP session is used to receive the image moved back from the PACS and for performing the validation of the received image against the original image in the Data Warehouse. The first commands the script performs are to reset the Data Warehouse and read the test image into the Data Warehouse (Fig. 3). The first READ command loads the DICOM image into the Data Warehouse and uses the value of attribute 0×00080018, the SOP Instance UID, as a reference value. This will allow the SCP Emulator that subsequently receives the image back from the PACS to automatically compare it to the image in the Data Warehouse based on the fact that the SOP Instance UID values are the same. The second READ command loads the same image again into the Data Warehouse but this time references it with CTIMAGE1. This reference will be used to export the image directly from the Data Warehouse to the PACS. As an alternative to READing the image into the Data Warehouse, one could use the CREATE and SET commands to build an image object directly in the Data Warehouse. DVT has the ability to generate pixel data patterns for VRs of type other byte (OB), other float (OF), and other word (OW). For the purposes of this example, it is best to use an image from the problem CT modality. The next operation the script performs is to establish a C-STORE CT Image association with the PACS. After establishing the association, the script creates a C-STORE-RQ command object in the Data Warehouse and exports the CT Image referenced by CTIMAGE1 using the C-STORE-RQ command referenced by CSTOREREQID, over the established association (Fig. 4). After closing the storage association, the script creates a new association on which to perform the C-MOVE, sends the request to move the image from the PACS to the DVT AE, and then closes the association (Fig. 5). Note that the Patient ID (0×00100020) and SOP Instance UID (0×00080018) were manually copied from the CT Image test file. Also, before executing this test, the SCP Emulator must already be started and waiting for the C-STORE from the PACS in response to the C-MOVE request. If the test executes successfully, the statement “Reference Dataset with identifier ‘1.3.46.670589.10.900123.19970114.35042000040’ found in Warehouse” should appear in the Activity Log indicating that DVT found the same CT image previously loaded into the Data Warehouse and will use it to perform a comparison of the dataset attributes and values. Fig 3.PACS transparency test using DICOMScript-reading image into the Data Warehouse.Fig 4.PACS transparency test using DICOMScript-exporting image from the Data Warehouse.Fig 5.PACS transparency test using DICOMScript-moving the image back to DVT. This test requires that the PatientRootQueryRetrieve-MOVE.def definition file is loaded in the script session and the CTImageStorage.def definition file is loaded in the Emulator SCP session. If, as suspected, the PACS is modifying the CT Image object before returning it to the modality, the Emulator SCP results data will identify where these modifications occurred. In this case, the hypothetical tester notices that DVT is indicating that a set of private attributes in the Data Warehouse copy of the image are missing from the image returned from the PACS. Upon further investigation, it turns out that the software upgrade performed on the modality included the addition of a set of new private attributes required by the modality to perform new postprocessing algorithms on the images. The PACS does not support these new private attributes for some reason and is stripping them from the image objects. Instead of using a combination of a DICOMScript test session and Emulator test session, this test could have been implemented with a single VBScript session using the HLI library. In this case, separate threads would be created for the SCU and the SCP parts. This approach is left up to the user as an exercise. Although DVT can analyze all aspects of DICOM interfacing, it tends to be overloaded for field service issues. DVTk.org has a number of service-orientated tools also based on the DVTk framework. One of these is the DICOMSniffer & Analyzer Tool. The tool is a GUI-based network sniffer and DICOM protocol analyzer. This tool uses the WinPcap open-source network library for packet capturing and filtering and includes a user guide. This tool can sniff a live DICOM network stream given two endpoints, such as a modality and PACS. The troubleshooting power of this tool should not be underestimated. It can perform message validation between two devices similar to DVT but does not require replacing an SUT device with the testing tool. Figure 6 shows the tool capturing a modality worklist query/response communication between two IP addresses. Fig 6.DICOM Sniffer & Analyzer tool in capturing mode. After the capture process, the user is presented with a per-association analysis of the DICOM control and datasets that were captured. The Association Overview tab shows the requested and accepted services and each field of the association PDUs. In the Service Elements Overview tab (Fig. 7), one can see all of the DIMSE messages that were transmitted during the association, save a copy, and view the PDUs. The summary and detailed validation results are similar to those produced by DVT, including message validation against SOP Class definition files. The tool also has the option of saving the captured data to a capture file for latter analysis or for sending on to additional service personnel. The DICOM Sniffer & Analyzer is a tool every PACS administrator should have on his or her workstation. Fig 7.DICOM Sniffer & Analyzer tool in analyzing mode. Two more tools built on the DVTk framework that should be in the imaging professional’s toolbox in working with DICOM on a regular basis are DICOM Compare and DICOM Editor. DICOM Compare can compare the attributes and values of two DICOM files. It can also filter out of the compare process any attributes and sequences to make it easier to identify offending elements. A known “good” DICOM object can be compared to one causing a problem that was captured from the network with DICOM Sniffer. With the DICOM Editor, one can add/delete/modify any attribute or sequence and sequence item and save the modified DICOM file to any location. This tool goes hand-in-hand with the DVT, DICOM Sniffer, and DICOM Compare tools. After those tools have identified the DICOM elements that may be causing a DICOM object storage problem, the editor can be used to manually fix the DICOM object and then resend it to the system where the failure is occurring. If the store succeeds, chances are the source of the problem has been identified. DVT AS A DEVELOPMENT TEST TOOL Another use of DVT is automated unit or system testing in a software development environment. Most modern source code management systems include automated build-and-test subsystems. Scripts can be created to test the DICOM interfaces of a vendor’s product, and the command line version of DVT can then be launched to execute those scripts against the product as part of the normal automated build-and-test processes. The DVT results and output files can then be saved along with the rest of the build/test artifacts. The command line version of DVT can be called on a DICOMScript as follows: dvtcmd Modality_System_Test.ses Modality_System_Test.ds. The summary and detailed results Extensible Markup Language files output by DVT make it easy for an automated test system to determine the results of the test, generate a report, and take any other appropriate action, such as e-mailing the development team. A tool as programmable and rich in output as DVT is valuable to the entire iterative development process. Software validation teams can link test session files, scripts, and result files to software defect issues, so that a developer can then use them to reproduce the defect. A validator can subsequently use those same scripts to test the fixed software. Maintaining a repository of DVT scripts is also an excellent way to run regression tests. If the DICOM software being tested is in the .NET language family, tighter testing integration with DVTk can be achieved via NUnit,4 the open-source unit testing framework for .NET. This integration uses a DVTkInvoker class to invoke the command line version of DVT with a session file, script file, and results file as arguments. One can use DVT Clients and Servers within NUnit to handle the various DICOM SOP Classes, starting and stopping DVT as necessary and obtaining validation results information. One future goal is to be able to start up and test an entire workflow scenario from NUnit using DVT. DVTk makes it quick and easy to perform repeatable, detailed testing of DICOM interfaces. In the following example, we posit a stress test to run against a storage SCP service. Because this is a stress test, the test application must hit the SCP simultaneously from multiple SCUs, so that multiple threads will be needed. Writing multithreaded Visual Basic (VB) test applications is a breeze with the new HLI library. In fact, it looks quite similar to a DICOMScript, with high-level abstractions for creating and releasing associations and for sending and receiving messages. This example using the HLI requires DVTk alpha version 2.1.007.006 or greater , which can be downloaded from dvtk.org after registering as a Plus User. Plus Users have access to early releases and draft documents, such as the draft version of the HLI API help file. In Visual Studio .NET, a VB console project is created and references to the Dvtk.dll, DvtkData.dll, and DvtkHighLevelInterface.dll libraries are added. This simple stress test application takes three arguments: the number of modalities to create (each modality runs on a separate thread), the number of images each modality will store to the SCP under test, and the location of the DVTk definition files. Figure 8 shows the imports required and the Sub Main(ByVal CmdArgs() As String) routine in module Module1. The first operation this application entry point method performs is to call Dvtk.Setup.Initialize(). This must be performed before any calls to the DVTk libraries, matched by the last call in the application, which should be Dvtk.Setup.Terminate(). Then it creates and initializes the MainDicomThread object, attaches it to the activity logging form, starts the main DICOM thread, and finally waits for the threads to complete. After the threads are complete, it logs selected performance data gathered during the stress test. Fig 8.Multithreaded stress test using VB.NET-main subroutine and imports. Figure 9 shows the MainDicomThread class that inherits from the DvtkHighLevelInterface.Dicom.Threads.DicomThread class. A single instance of this class is started from the entry point method. This class’s thread execute method starts all of the modalities and returns. As it creates each modality (instances of CtStorageScu), it generates unique identifying information for each SCU, such as AE title. Some of the options are analogous to test session properties. This test application could have loaded a script session configuration file and initialized some of the options from it, but in this case, all of the necessary session options are set directly in the code. After an SCU is created, it is immediately started by calling its thread Start() method. Fig 9.Multithreaded stress test using VB.NET-MainDicomThread class. Figure 10 shows a CtStorageScu class that also inherits from DvtkHighLevelInterface.DICOM.Threads.DICOMThread. The purpose of the execute method is to C-STORE a specified number of images to the SCP under test. It first creates unique patient, study, and series level identifiers for the images that will be stored. It then creates the C-STORE CT Image association; then loops to store the images, cloning a CT image dataset that was read from a DICOM CT image file in the constructor; creates unique image identifiers for each image; sends the C-STORE-RQ message; receives the response message; and checks the response status value. Finally, it logs its total storage time for later analysis and closes the association. This application could have been created and executed from the DVT GUI, but developing it in Visual Studio allows the application to be debugged and compiled to an executable file. This application is executed from the command line as: Storage_SCP_Stress_Test.exe 20 1000 “C:\Program Files\DVT\Definitions”. Fig 10.Multithreaded stress test using VB.NET-CtStorageScu class. Figure 11 is an excerpt from the Hl1Form that receives all of the logging from the DicomThread objects. Notice the multithreading capabilities of the HLI in this output. While modality 10 is validating an A-ASSOCIATE-AC message it received, modality 3 is handling a CSTORERSP message. This test created 20 CT modalities that are each simultaneously sending a 1,000-image series to the PACS. Fig 11.Multithreaded stress test-activity logging. RECENT EXTENSIONS AND FUTURE WORK DVTk has been extended recently to support HL7 validation in addition to DICOM validation. This is currently used by the DVTk Integrating the Healthcare Enterprise (IHE)5 Actors framework where it is possible to configure DVT to emulate the role of certain actors in the IHE integration profiles. The idea is that the IHE Actors necessary to allow the SUT to be tested are emulated by DVTk. DVTk will then validate the DICOM and HL7 transactions taking place between actors and, in addition, compare the values of certain attributes such as Patient ID, Patient Name, etc., between messages to ensure consistent use. In the future, the aim is to enhance the capabilities of the DVTk IHE Actors framework by supporting additional functionality in the existing actors and supporting other actors needed in the various IHE Integration Profiles. Support for other protocols is also envisioned. The DVTk development team is also involved in the new IHE Connectathon Toolkit (code-named “Gazelle”), which is being developed as a successor to the MESA Toolset. It is hoped that the DVTk DICOM validation engine can be wrapped as a Web service for use in Gazelle. Various new tools will be developed using the DVTk frameworks. A GUI for the DVT-IHE framework; various new emulators, such as a radiology information system emulator and modality emulator; and stand-alone validation applications are likely candidates. The current number of development resources limits what can be done—anyone wishing to contribute can do so via the Web site. The aim for future IHE extensions is to integrate any other validation services into the DVTk framework as Web services. This may include Cross-Enterprise Document Sharing validation, for example. It is hoped that the Gazelle cooperation will result in web services that can be reused for this purpose. CONCLUSION Originally developed by Agfa and Philips to test their products, DVTk has grown into a professionally managed, open-source, vendor neutral, DICOM Validation Framework. Flexible for the changing standard, programmable, and extensible, DVTk is a powerful tool for anyone working with the DICOM standard. The toolkit includes GUI and command line versions of the main validation application, DVT, and a collection of .NET libraries for creating new validation and test tools. The large collection of example validation sessions that come with the toolkit are a great place to start for understanding how the DICOM standard works in practice. Hospital IT staff can use DVTk for simple tasks, such as pinging the network for the existence of modality AE titles, or for more detailed troubleshooting, such as checking for the presence of specific DICOM attributes and values in messages and media files. Other DVTk-based tools are available for tasks such as editing or comparing DICOM files or for capturing and analyzing messages on a live DICOM stream. The XML-structured test results and media files created by the validation framework provide great evidence for IT staff when discussing a problem with vendors. Developers and testers can create scripts for testing their products DICOM conformance, or build on the framework by creating new standalone test tools. NUnit integration and the command line version of DVT make it possible to incorporate DVTk and the generated test results into automated build-and-test systems. Current and future efforts include support for IHE actor validation, including HL7 validation, and new device emulators, such as a Radiology Information System (RIS) emulator. As the digital hospital enterprise continues to grow, DVTk is well positioned to take on these new validation roles. We encourage the reader to continue working with DVTk as a means to master the changing DICOM environment. With the move to an open-source community approach, DVTk is well on its way to becoming the independent gold standard for DICOM interface testing.	[ "dicom", "ihe", "systems integration", "pacs dicom ihe conformance", "health level 7" ]	[ "P", "P", "R", "R", "R" ]
Bioprocess_Biosyst_Eng-2-2-1705497	Genetic algorithm for multi-objective experimental optimization	A new software tool making use of a genetic algorithm for multi-objective experimental optimization (GAME.opt) was developed based on a strength Pareto evolutionary algorithm. The software deals with high dimensional variable spaces and unknown interactions of design variables. This approach was evaluated by means of multi-objective test problems replacing the experimental results. A default parameter setting is proposed enabling users without expert knowledge to minimize the experimental effort (small population sizes and few generations). Introduction In many technical problems scientists face the problem of identifying optimal process conditions like pH, temperature, concentrations or other variables. A typical example is fermentation medium development. The composition of a fermentation medium consisting of carbon sources, nitrogen sources, mineral salts, trace elements, amino acids and/or peptides, vitamins and other growth factors determines the chemical and nutritional environment of cells in a bioreactor and is thus vital for the effective manufacturing of bioproducts. Due to the large number of process variables and the metabolic complexity of microorganisms or cells methods of experimental design need to be applied in order to identify values of the relevant variables resulting in an improved performance. Considering several objectives to be optimal there is per definition not one optimal solution but a set of efficient solutions. Mathematically such a multi-objective optimization problem (MOP) can be described in terms of Eq. 1. The functional relation between the m design variables and the objective functions φ does not necessarily need to be known, as y will be obtained from n experimental observations. An important criteria for MOPs is that of Pareto optimality, shortly saying that a member of the efficient set is not dominated by any other. In Fig. 1 the principle of dominance is exemplified in the objective space, where for the case of maximizing two objectives the efficient solutions are indicated. Fig. 1Efficient points (black) are not dominated by any other point. For a dominated point (white) there exist at least one other point that has greater values in both objectives y1 and y2 Classical statistical experimental design methods (Plackett–Burman Design, Response Surface Method) have drawbacks like screening for principal components and assuming an unimodal objective function, reviewed in [1]. More sophisticated stochastic search strategies like genetic algorithms (GAs) have grown in popularity since Rechenberg [2] and Holland [3] first published their work on this subject. GAs are based on evolutionary principles, encoding several sets of design variables on binary strings (individuals in a population) which are processed by GA operators (crossover and mutation) throughout several generations. The principle “survival of the fittest” assures a convergence towards optimal values in the design variables with proceeding generations. The first GA dealing with multiple objectives was the Vector Evaluated Genetic Algorithm (VEGA) proposed by Schaffer [4]. This multi-objective optimization strategy has already been applied successfully for experimental medium optimization in many cases [1]. The most recent published multi-objective GAs are the Non-Dominated Sorting Genetic Algorithm-II [5] and the Strength Pareto Evolutionary Algorithm (SPEA) [6]. The SPEAs main feature is processing two populations: besides a normal population an external population serves as a kind of archive, keeping track of the efficient set. Especially in the case of experimental design with a limited number of experiments the SPEA is supposed to have advantages compared to VEGA with respect to the experimental effort. The new software tool with a genetic algorithm for multi-objective experimental optimization making use of SPEA will be outlined. The performance of GAME.opt will be evaluated with the help of special test-functions to achieve appropriate parameter settings for experimental design. Methods GAME.opt was developed with LabVIEW 7 (National Instruments). The Application Builder was used to create a stand alone application running under Windows platforms. Data handling is based on spreadsheet files, which can be accessed via a text editor or Microsoft Excel. GAME.opt will be made available on request by the Technical University of Munich. As the core algorithm is based on the SPEA, fitness assignment, selection and clustering is described in detail by Zitzler and Thiele [6]. To abstract the algorithm: fitness assignment is either based on the strength of an individuals dominance or on the degree an individual is dominated by others. A mating pool equal to the population size is filled by selecting individuals using binary tournament without replacement [7]. Crossover is performed between two parental binary strings from the mating pool at the specified number of crossover points in order to create two off-springs. Each bit of these new strings is inverted with a probability which is defined by the mutation rate. Clustering is used in order to bound the external population to a maximal value Nextern. Each time the external population exceeds this value it is pruned to a size of Ncluster by means of average clustering. The first step when using GAME.opt is to create a new project where all information on the experimental design is defined. For this purpose the user is led through the steps explained together with some recommendations in the following. Design variables The design variables considered in the problem (x in Eq. 1) need to be defined initially. Alternatively to choosing the amount of bits encoding a variable GAME.opt gives the opportunity to choose the amount of levels, resulting from the variable’s bounds and its increment. The choice of levels can be necessary for practical purposes but causes problems in the coding of a decision variable as follows. A general decoding function linearly transforms an n-bit-long binary string (a1...an) into a bounded real number x ∈ [xL, xU]: where a1 is the most significant bit and an is the least significant bit [8]. This kind of transformation is just unique if the amount of levels equals (2n − 1). In other words, like its natural paradigm, the binary code is degenerated if there are more bit combinations than levels (the opposite case will not occur as the program computes the minimal amount of required bits). Considering the above mentioned the following points are important when defining the design variables:If the experimental setup prevents an adequate choice of levels the coding can be made visible in GAME.opt. In order to assure equal processing of all variables the amount of levels among the variables should not vary too much.From the amount of levels results the length of the binary string that is needed to encode the design variables, which is shown in GAME.opt in terms of bits for each variable. From a set of M binary strings with the length l each point in the search space (X in Eq. 1) can be reached with a probability p. The relation is given in Eq. 3 and GAME.opt will display p when selecting the number of individuals (M) [9].The value of p should be higher than 0.99.As each individual corresponds to one experiment, the experimental effort will be correlated to the chosen amount of levels.The increment should be higher than the possible precision in the experiment. Experimental results The experimental observations y have to be specified in GAME.opt just in terms of their names and units. The following items should be considered when planning the experiments:All parameters different from the design variables have to remain fixed in all experiments throughout the optimization procedure.The experimental error should be lower than the differences that are expected to result from different levels in the design variables. Objective functions The objective functions in GAME.opt can be any linear combination of the experimental results and the decision variables. GAME.opt will maximize all objective values. If for example one objective is minimizing a certain experimental result, the objective function is –1 times the according result. When defining the objective functions it is important to make sure that they are independent from each other. For example it is no use maximizing simultaneously the amount of a reaction’s product and minimizing the amount of non-converted reactants at the end of a batch process, because a high product concentration is correlated to low resting concentrations of reactants. A reasonable combination is for example maximization of an experimental result while minimizing the sum of decision variables, possibly reducing the costs of a process. Optimization procedure Once the MOP is defined by the above mentioned steps a random generated initial population is available. This set of design variables can be exported from GAME.opt and the according experiments are performed. After the experimental results are entered in GAME.opt the next population is generated and checked for individuals that have already appeared in order to prevent repeated experiments. This iterative process is performed until satisfying results are obtained. The external population contains the efficient set recovered so far and thus constitutes the result of the optimization process. The experimental results in the external population are already assigned because their members originate from former populations. Results A previous version of GAME.opt was used for experimental design for media optimization (13 medium components, 20 individuals/generation, 91 bits/binary string, eight generations) and gave comparable results to another GA making use of VEGA with half the number of experiments [10]. Further process optimizations considering media composition together with process parameters are currently performed in our laboratory. In order to evaluate the software for a proper processing and its general ability to solve MOPs it was tested by means of several multi-objective test problems in such a way that the experimental results are replaced by evaluation of mathematical functions. An exemplary MOP with three objectives and six design variables is given in Eq. 4. This MOP is based on Deb et al. [11], where the development of test problems is described. The MOP in Eq. 4 was solved with GAME.opt, therefore the six design variables have been defined in the program together with the bounds, being 0 and 1 for each variable. As mentioned above the choice of the increment and thus the amount of levels is crucial for a proper processing of the variables. An increment of 0.1 will give 11 Levels, namely [0, 0.1, 0.2, ..., 1] encoded with 4 bits giving 16 levels. According to Eq. 2 an adequate increment for a given number of bits n results from: Table 1 confronts a proper coding with 4 bits with a degenerated coding for a variable range from 0 to 1. Table 1Four bit binary coding of the design variables in Eq. 4 for different increments (incr)Binary codeDegenerated codingDecoded value (incr = 0.1)Proper codingDecoded value (incr = 1/15)0 0 0 0000 0 0 10.11/150 0 1 00.12/150 0 1 10.23/150 1 0 00.34/150 1 0 10.35/150 1 1 00.46/150 1 1 10.57/151 0 0 00.58/151 0 0 10.69/151 0 1 00.710/151 0 1 10.711/151 1 0 00.812/151 1 0 10.913/151 1 1 00.914/151 1 1 111 Using the proper coding shown in Table 1 the GA was run eight generations with 20 individuals in the normal population and 10 individuals in the external population. Two crossover points and a mutation probability of 1% was chosen what has been identified as a reasonable default setting in other problems. As a rule of thumb the size of the external population should be half of the normal population, for which a minimal number can be estimated by Eq. 3. Table 2 summarizes the default parameter setting. Table 2Default parameter setting for the genetic algorithmParameterValueCrossover points2Mutation rate1%Individuals in normal population (N)Eq. 3Individuals in external population (Nextern)N/2Individuals in external population after cluster analysis (Ncluster)N/2 The results are shown in Fig. 2, by means of the objective values f1, f2, f3 of the ten individuals in the external population after eight generations. Fig. 2The objective space of the test problem lies between the surfaces. Optimal solutions are located on the unit sphere in the first quadrant. The black dots indicate the efficient set after eight generations of GAME.opt (external population). The two views of the plot point out the closeness of the efficient set to the Pareto front (a) and its equal distribution (b) The two spherical surfaces in Fig. 2 constitute the objective space of Eq. 4. For the minimization problem the first quadrant of the unit sphere constitutes the Pareto optimal frontier where g(x) = 0 and it is already approximated well by the ten members of an external population after eight generations (Fig. 2a). Beside the closeness to the Pareto frontier the distribution of the solutions is an important performance criteria. Figure 2b reveals the equal distribution of the ten solutions resulting amongst other things from the cluster analysis. Optimal solutions are characterized by xi = 0.5 for i = 1, 2, 3, 4, whereas x5 and x6 can take any value in the given range. Splitting the design variables in such a way allows to test the algorithm for a crucial feature of GAs, the populations diversity. On the one hand the algorithm is expected to converge fast to the optimal solutions but on the other the populations diversity must be high enough to ensure exploration of the whole Pareto optimal frontier. This fact is reflected in the box plots of the initial and the eighth generation together with their corresponding external populations in Fig. 3. Fig. 3Box plot of the initial population (a), the first external generation (b), the eighth generation (c) and the eighth external generation (d). The boxes contain the middle of 50% of the data. The line in the box indicates the median. The bold dashed line separates variables with an optimal solution (x1, x2, x3, x4 = 0.5) from variables without optimal value (x5, x6) Initially the individuals are randomly distributed, what is reflected by the box plots of the six design variables in Fig. 3a. A trend towards 0.5 is already visible in Fig. 3b where box plots are shown for the first external population. In Fig. 3c, d, depicting the box plots of the corresponding populations after eight generations, a clear distinction between the variables with optimal value and those without can be made. After eight generations the diversity of x5 and x6 is still high in both populations. The remaining variables show the expected divergence towards 0.5 in the external population but remain in parts distributed over a wider range in the normal population. From this result it gets obvious that GAME.opt performs well with a small amount of individuals and a short evolution on a MOP considered a benchmark of evolutionary computing. Conclusions In order to keep the experimental effort low, GAME.opt will use small population sizes and fewer generations compared to other GA applications, where the amount of objective function evaluations is mostly limited by the computational costs. For this reason the external population is expected to be only a rough approximation of the true Pareto optimal frontier. Nevertheless in all test problems a satisfying performance of GAME.opt was achieved in spite of the relatively low number of experiments (160 compared to 166 possible design variable combinations with the test problem presented). In the near future, efficient parallel search strategies for experimental design are becoming more and more important in bioprocess optimization due to the availability of new parallel stirred-tank bioreactor technologies [12].	[ "genetic algorithm", "software tool", "experimental design", "multi-objective optimization" ]	[ "P", "P", "P", "P" ]
Apoptosis-4-1-2423418	Hypertonicity-induced cation channels rescue cells from staurosporine-elicited apoptosis	Cell shrinkage is one of the earliest events during apoptosis. Cell shrinkage also occurs upon hypertonic stress, and previous work has shown that hypertonicity-induced cation channels (HICCs) underlie a highly efficient mechanism of recovery from cell shrinkage, called the regulatory volume increase (RVI), in many cell types. Here, the effects of HICC activation on staurosporine-induced apoptotic volume decrease (AVD) and apoptosis were studied in HeLa cells by means of electronic cell sizing and whole-cell patch-clamp recording. It was found that hypertonic stress reduces staurosporine-induced AVD and cell death (associated with caspase-3/7 activation and DNA fragmentation), and that this effect was actually due to activation of the HICC. On the other hand, staurosporine was found to significantly reduce osmotic HICC activation. It is concluded that AVD and RVI reflect two fundamentally distinct functional modes in terms of the activity and role of the HICC, in a shrunken cell. Our results also demonstrate, for the first time, the ability of the HICC to rescue cells from the process of programmed cell death. Introduction It has become increasingly evident in recent years that the mechanisms of cell volume regulation are employed in a variety of physiological processes, in addition to just the maintenance of cell homeostasis. Among these processes are the coordination of transport across the apical and basolateral membranes in epithelia, the locomotion of cells, the regulation of metabolic processes in the liver and, most notably, the control of cell proliferation and apoptosis [1–3]. Since the rates of ion transport through channels are some 4–5 orders of magnitude higher than those achieved by transporters (carriers and pumps), any modulation of ion channel activity may serve as a rapid and efficient mechanism of osmolyte transport and cell volume regulation. Accordingly, hypertonicity-induced cation channels (HICCs) were found to be the main mediators of the “regulatory volume increase (RVI)” of a shrunken cell, whenever a system was analysed in a quantitative fashion [3–5]. Despite the low selectivity of HICCs for Na+ over K+, activation of these channels leads to an actual net uptake of cations, which is mainly due to the membrane voltage, which is negative intracellularly. Anions and osmotically obliged water follow, allowing the cell to complete the process of RVI [3–5]. Persistent isotonic shrinkage of a given cell, as opposed to hypertonicity-induced shrinkage, is one of the hallmarks of apoptosis. In many instances, this “apoptotic volume decrease (AVD)” employs the activation of K+ channels, which serve as the main routes of K+ exit [6–9]. Cl− channels also should open, with this parallel transport strongly depending, however, on whether or not the electro-chemical driving forces actually favour conductive Cl− export [1, 2, 10–12]. The question is thus raised of how hypertonic stress might interfere with AVD. Would osmotic shrinkage facilitate the process? Or would the activation of HICCs or RVI oppose the induction of apoptosis? Materials and methods Cell culture Human cervix HeLa cells were grown as monolayers in minimum essential medium supplemented with 10% fetal bovine serum, 100 μg/ml streptomycin and 40 IU/ml penicillin, under 95% air/5% CO2 at 37°C. For the experiments, cells were mechanically detached from the plastic substrate of the culture flasks with a jet of culture medium and then kept in suspension for 15–120 min. Determination of cell volume Changes of HeLa cell volume were quantified by means of electronic cell sizing on a Coulter-type analyzer (CDA-500; Sysmex, Kobe, Japan), as previously described [13]. The experimental solution contained (in mM): NaCl, 95; KCl, 4.5; MgCl2, 1; CaCl2, 1; NaHCO3, 10; HEPES, 5. Osmolality was adjusted to values from 300 to 600 mosmol/kg-H2O by adding mannitol under osmometric control (OM802; Vogel, Giessen, Germany). Measurements were performed at room temperature. Patch-clamp experiments Membrane currents were recorded in the fast whole-cell mode of the patch-clamp technique using 2 MΩ borosilicate pipettes. Currents were recorded with an Axopatch 200B amplifier (Molecular Devices, Union City, CA), filtered at 5 kHz with a four-pole Bessel filter and digitized at 20 kHz. pClamp 9.02 software (Molecular Devices) was used for control of the pulse protocol, as well as for data acquisition and analysis. Series resistance was <5 MΩ and was compensated (by 70–80%) to minimize voltage errors. In the experiments, voltage ramps from −80 to +20 mV and of 1 s duration were applied every 10 s; holding voltage was −30 mV. The bath solution (pH 7.5) contained (in mM): NaCl, 94; Na-gluconate, 6; MgCl2, 1; CaCl2, 2; TEA-Cl, 2; HEPES, 10; d-glucose, 10. Osmolality was adjusted to 310 (isotonic) and 400 mosmol/kg-H2O (hypertonic) by the addition of mannitol. The pipette solution (pH 7.3, 300 mosmol/kg-H2O) contained (in mmol/l): NaCl, 26; Na-gluconate, 69; MgCl2, 1; TEA-Cl, 2; Na2-ATP, 2; Na2-GTP, 0.5; HEPES, 10; EGTA, 1. With these ion gradients, the equilibrium potentials for Na+ and Cl− (as the only permeant ions) are set at 0 and −30 mV, respectively [13]. Experiments were performed at room temperature. Monitoring apoptotic cell death Cell viability and caspase-3/7 activity were determined by use of a calorimetric MTT assay (Cell Counting Kit-8; Dojindo, Kumamoto, Japan) and a fluorometric probe (Apo-ONE Homogeneous Caspase-3/7 Assay; Promega, Madison, WI), respectively. Fragmentation of DNA was determined by means of a photometric enzyme immunoassay (Cell Death Detection ELISAPLUS, Roche Applied Science, Diagnostics, Mannheim, Germany) with HeLa cells plated on 96-well plates. Following exposure to the various experimental conditions, cells were lysed and centrifuged at 200g. The cytoplasmic (DNA containing) fraction was then transferred to streptavidin-coated microtiter plates that had been incubated with a biotinylated monoclonal anti-histone antibody. Finally, the amount of the fragmented DNA on bound nucleosomes was determined with a peroxidase-conjugated monoclonal anti-DNA antibody and the substrate ABTS [2,2′-azino-di-(3-ethylbenzthiazoline sulfonate)], which was quantified photometrically at 405 nm. In all measurements, cells were incubated for 6 h at 37°C. Testing for necrosis and late apoptosis Necrosis and late apoptosis of cells were quantified fluorometrically on the basis of a propidium iodide staining of nuclei (excitation/emission at 530 nm/620 nm) performed with reference to an overall staining of cells by Hoechst 33342 (excitation/emission at 350 nm/460 nm). Chemicals All reagents were purchased from Sigma–Aldrich (Tokyo, Japan). Statistical analysis Data are presented as mean values ± SEM with n denoting the number of observations. Student’s t tests for paired and unpaired data were used as appropriate and P < 0.05 was considered significant. Results and discussion As was determined by means of electronic cell sizing, increasing osmolality from 300 mosmol/kg-H2O to 330, 400, 500 and 600 mosmol/kg-H2O led to an initial shrinkage of HeLa cells to 96.6 ± 0.4, 82.6 ± 1.9, 72.5 ± 0.2 and 66.2 ± 0.6% of the control level, respectively (n = 10; Fig. 1a). When plotted as a function of 1/osmolality (which is equivalent to a ‘Boyle-van’t Hoff graph’ [3, 14]), these data could readily be fitted by a straight line; its intercept with the y-axis yields the osmotically active versus inactive space of the cells, with the latter amounting to 30.6% of total cell volume (Fig. 1a, Insert). This osmometric behaviour of HeLa cells was then followed by an active volume recovery to 101.7 ± 0.4, 92.6 ± 0.5, 89.7 ± 0.8 and 88.4 ± 0.9%, in 120 min of hypertonic stress (Fig. 1a, b). The process of “regulatory volume increase (RVI)” occurred at essentially constant rates that equalled 0.4 ± 0.2, 0.9 ± 0.2, 1.6 ± 0.2 and 1.8 ± 0.2%/10 min, respectively, in the range of 5–120 min after application of hypertonicity. Fig. 1Hypertonicity-induced HeLa cell shrinkage and RVI determined by electronic cell sizing. (a) The volume response of cells to increases in osmolality (from 300 mosmol/kg-H2O) to 330, 400, 500 and 600 mosmol/kg-H2O at time zero (arrow) are depicted. The insert gives the initial (passive) amount of cell shrinkage as a function of 1/osmolality (equivalent to a ‘Boyle-van’t Hoff plot’ of these data; see text for details). 3.33 kg-H2O/osmol = 1/(0.300 osmol/kg-H2O) is the 100% value under isotonic conditions. (b) The volume of HeLa cells at 120 min of various levels of hypertonic stress, after partial volume recovery (RVI) had been achieved. n = 10 for each experimental condition It has become increasingly evident in recent years that persistent cell shrinkage is one of the hallmarks of apoptosis [6, 15]. Also, it has recently been found that after 2-h stimulation with staurosporine (STS), TNFα or Fas ligand, HeLa cells lose their ability to undergo RVI (12). To test whether HeLa cell shrinkage and/or RVI interferes with apoptotic volume decrease (AVD), 4 μM STS was used to induce apoptosis in HeLa cells [6], and a series of volume measurements was made. This was done under isotonic conditions (300 mosmol/kg-H2O) and under hypertonic conditions in which osmolality was increased to 330, 400, 500 and 600 mosmol/kg-H2O (n = 10). Within an experimental time frame of 120 min, STS was found to reduce cell volume to 65.8 ± 1.6% that of the control, under isotonic conditions (Fig. 2a: diamonds). 5 min after osmolality was increased to 330, 400, 500 and 600 mosmol/kg-H2O, cell volume equalled 89.6 ± 0.7, 81.4 ± 0.2, 71.4 ± 0.9 and 64.5 ± 0.8% of control, respectively. These values are virtually identical to the values of purely passive or physical responses to hypertonic stimulation in the absence of STS (cf. to Fig. 1a). Moreover, an osmotically inactive space of 35.4% could be calculated from these measurements; this value is very similar to that obtained in the control experiments (see above). Of note, no RVI followed hypertonicity-induced cell shrinkage when STS was present. This indicates that the RVI response was completely inhibited by STS (or surpassed by the STS-induced AVD process), a phenomenon which clearly holds for the full 2 h period after STS application. Using the respective values for volume at 5 min as passive reference values, AVD was computed to be 30.8 ± 0.7, 22.2 ± 0.7, 12.1 ± 0.5 and 7.2 ± 0.7% for 330, 400, 500 and 600 mosmol/kg-H2O, respectively (calculated as the volume at 5 min minus the volume at 120 min). So as summarized in Fig. 2b, a decrease of AVD was observed with increasing hypertonic stress. On the one hand, then, hypertonic stress appears to suppress STS-induced AVD; on the other hand, STS seems to abolish hypertonicity-induced RVI. Also of note, in this respect, is a recent study on HeLa cells in which tyrosine kinases, PKC, and p38 MAPK were identified as part of the signalling machinery employed in the activation of HICC and RVI [16]. Given the non-specificity of STS with respect to its inhibition of various protein kinases, it is not surprising that RVI is inhibited in the presence of the compound, while AVD is induced. Fig. 2Effects of hypertonic stress on apoptotic volume decrease (AVD). (a) At time zero (arrow), 4 μM staurosporine (STS) was added to induce apoptotic shrinkage of HeLa cells under isotonic conditions (◆) or together with an increase of osmolality to 330 (■), 400 (▲), 500 (●) and 600 (▼) mosmol/kg-H2O. Hypertonic cell shrinkage, at 5 min, was very similar in the presence and absence (see Fig. 1) of STS. (b) AVD under isotonic conditions (volume at 0 min minus volume at 120 min) and hypertonic conditions (volume at 5 min minus volume at 120 min). n = 10 for each experimental condition To study the interference of STS-induced AVD with hypertonicity-induced RVI in more detail, experiments were performed in which osmolality was increased from 300 to 500 mosmol/kg-H2O, at time zero (see Fig. 3a: arrow). In addition, 4 μM STS was applied either 2 h before (t = −120 min), 1 h before (t = −60 min), at the same time as (t = 0) or 1 h after (t = 60 min) hypertonic stimulation (see Fig. 3a: arrowheads). AVD was then computed from the overall changes in cell volume (i.e. the volume at 0, −60 or −120 min, respectively, minus the volume at 120 min) minus the amount of passive cell shrinkage (i.e. the volume at 0 min minus the volume at 5 min). Hypertonic stress (500 mosmol/kg-H2O) significantly reduced AVD from the control value of 34.2 ± 0.5% (Fig. 2b: 300; also marked as STS-ISO, in Fig. 3b) to 12.1 ± 0.5 or 12.5 ± 0.5% (Fig. 3b: STS-HYPER or HYPER/1hSTS) when given simultaneously with, or 1 h before, application of STS (n = 10, P < 0.001 each), respectively. Also apparent from these experiments is that no RVI occurred in the presence of STS. It is noteworthy that increasing the osmolality to 500 mosmol/kg-H2O after 1-h pretreatment with STS still led to a markedly reduced AVD of 25.9 ± 0.7% (P < 0.001; Fig. 3b: STS1h/HYPER). A hypertonic stimulation after 2-h pretreatment with STS, however, was not effective at suppressing AVD (Fig. 3b: STS2h/HYPER). These results suggest that hypertonic stimulation of HeLa cells is somehow interfering with the process of AVD when applied in the time range of 1 h before to 1 h after the application of STS. Fig. 3Interference of AVD with RVI. (a) Osmolality was increased from 300 to 500 mosmol/kg-H2O at time zero (arrow), with 4 μM STS applied (at arrowheads) either 2 h before (STS2h/HYPER: ■), 1 h before (STS1h/HYPER: ●), at the same time as (STS-HYPER:▼) or 1 h after (HYPER/1hSTS: ▲) the start of hypertonic stress. (b) Total AVD computed from the overall volume changes shown in (a) minus the initial amount of passive cell shrinkage in the first 5 min of hypertonic stress (i.e. volume at 0 min minus volume at 5 min). STS-ISO refers to the AVD observed under isotonic conditions. n = 10 for each experimental series Next, the effects of STS on activation of the HICC were tested. This was done by increasing osmolality at various times relative to STS application. As shown in Fig. 4 (STS-HYPER), STS had no significant short-term effect on activation of the HICC when given together with hypertonic stress, i.e. the peak activation of HICC current was virtually unaffected for 5–7 min by STS applied simultaneously with hypertonic stress. This finding fits with the observation that HICC activation appears to occur on a considerably faster time scale than the initiation of AVD does. In this respect, it is also of note that even with STS applied 1 h ahead of hypertonicity (STS1h/HYPER), the HICC was still activated to a level approximately half of the control value, i.e. to 1.59 ± 0.39 pA/pF. With STS given 2 h before hypertonic stress (STS2h/HYPER), HICC currents were almost completely inhibited (0.62 ± 0.20 pA/pF, n = 5–33). Fig. 4Effects of STS on HICC activation. (a) HICC currents in the absence (Control) or presence of STS, which was applied simultaneously with hypertonic stress (STS-HYPER), 1 h ahead of hypertonic stress (STS1h/HYPER) or 2 h ahead of hypertonic stress (STS2h/HYPER). Representative traces of 5 to 33 experiments. (b) Typical I/V-relationships under these conditions. (c) Summary of HICC current densities obtained under the four different conditions shown in (a) Whenever studied in a quantitative way, HICCs have been proven to be the main mechanism of the RVI of a given system [3–5]. This also holds for HeLa cells, in which the relative contribution of HICC activity and Na+/H+ antiport to RVI is approximately 2:1 (with no detectable contribution of Na+-K+-2Cl− symport at all) [5, 13]. This raises the question of whether the apparent interference of hypertonic stress with AVD may occur via activation of the HICC. In the experiments shown in Fig. 5a, osmolality was increased from 300 to 500 mosmol/kg-H2O, at time zero (arrow). Following 60 min of hypertonic stress, 300 μM flufenamic acid (FFA, an established blocker of the HICC in HeLa cells [4, 5, 13]), STS (4 μM) or FFA plus STS were then applied (arrowhead). As expected, FFA significantly blunted the progress of RVI, presumably by inhibition of the HICC. The volume change from 60 to 120 min (volume at 120 min minus volume at 60 min) was −2.0 ± 0.6% in the presence of FFA; in the control experiments, on the other hand, the 60-to-120 min volume change due to RVI equalled 10.2 ± 0.9% (n = 10, P < 0.001; Fig. 5b: Control). Of note, STS led to a pronounced decrease in cell volume of −12.5 ± 1.6% (Fig. 5b: STS), and the effects of FFA plus STS were additive, resulting in a decrease in cell volume of −19.0 ± 1.0% (P < 0.01; Fig. 5b: STS+FFA). Experiments with 1 mM 2-aminoethoxydiphenyl borate (2-APB) as an alternative blocker of the HICC [17] yielded virtually identical results. With 2-APB, the 60-to-120 min volume change equalled 2.9 ± 1.1% (Fig. 5b: 2-APB), and again, the effects of HICC blockage and STS (volume change of −16.3 ± 0.6%; P < 0.05) were additive (Fig. 5b: STS+2-APB). Fig. 5Effects of HICC channel blockers and STS on RVI and AVD. (a) Osmolality was increased from 300 to 500 mosmol/kg-H2O at time zero (arrow) for all groups. 300 μM flufenamic acid (FFA), STS or FFA plus STS were applied at t = 60 min of hypertonic stress (arrowhead) as indicated. (b) Summary of the volume changes from 60 to 120 min (volume at 120 min minus volume at 60 min) observed in the experiments shown in (a) and two additional experiments in which 1 mM 2-aminoethoxydiphenyl borate (2-APB) was used as a blocker of HICC channels, alone or in combination with STS. n = 10 for each experimental condition. *** Significantly different from Control with P < 0.001; #,## significantly different from STS with P < 0.05 and P < 0.01, respectively It may be argued that, due to the limited specificity of FFA and 2-APB, the effects of both compounds described above may be due to an inhibition of volume-sensitive outwardly rectifying (VSOR) Cl− channels (see [2, 18] for review), rather than blockage of the HICC. In an additional series of measurements, however, it was found that even with application of 10 μM DCPIB (4-(2-butyl-6,7-dichlor-2-cyclopentyl-indan-1-on-5-yl) oxybutyric acid), an effective blocker of the VSOR Cl− channel [19], an RVI value of 10.6 ± 0.6% (when added at t = 60 min; n = 10) was obtained; this is virtually identical to the value of 10.2 ± 0.9% observed under control conditions (see above). Furthermore, 10 μM DCPIB inhibited VSOR currents in HeLa cells by 67.6 ± 3.6% (n = 6) whereas it did not exhibit an effect on HICC currents at all (n = 8; Wehner and Numata, unpublished results). These data strongly imply that, even in early-stage apoptosis, HICC activation is an effective mechanism of RVI in HeLa cells. Furthermore, they provide additional strong evidence for the functional interplay of AVD and RVI. In this regard, the apparent lack of RVI in HeLa cells undergoing AVD (Fig. 2a) could be accounted for as follows: In early-stage apoptosis (say, within 1 h after STS application) the AVD process surpasses the RVI process, which involves HICC activity, whereas in the later stage of apoptosis, the RVI process is suppressed because of inhibition of the HICC by STS. Finally, we characterized the interplay between HICC activation and STS-induced cell death and apoptosis. As determined by an MTT assay (Fig. 6a), a caspase-3/7 activity assay (Fig. 6b), and an apoptosis assay measuring nucleosome-bound DNA (Fig. 6c), increasing osmolality from 300 mosmol/kg-H2O to 400, 500, and 600 mosmol/kg-H2O led to significant and parallel reductions of both cell death and apoptosis observed 6 h after STS application; the maximal effect was observed at 500 mosmol/kg-H2O and above. This is a surprising finding because, based on what has been reported so far concerning the interplay of cell volume and apoptosis, one might expect a synergistic effect of cell shrinkage on the induction of programmed cell death [15]. Furthermore, FFA and 2-APB almost completely blocked these effects, the fact providing strong further evidence that it is the activation of the HICC that is actually rescuing HeLa cells from apoptosis. Fig. 6Rescue from STS-induced apoptosis by hypertonic stimulation and inhibition of rescue by HICC blockers. (a) Cell death observed 6 h after STS treatment at various osmolalities, and at 500 mosmol/kg-H2O with 300 μM FFA or 1 mM 2-APB as indicated (n = 12–34). (b) Apoptosis monitored as caspase-3/7 activity induced by 6-h exposure to STS at various osmolalities, and when FFA or 2-APB was also applied, at 500 mosmol/kg-H2O (n = 12–26). (c) DNA fragmentation analyzed with a nucleosomal ELISA assay after a 6-h exposure to STS at various osmolalities, and when FFA or 2-APB was also applied, at 500 mosmol/kg-H2O (n = 6) To insure that the above measurements were not corrupted by the necrosis of cells, propidium iodide staining was performed. Propidium iodide is known to selectively bind to cell nuclei, and because it does not permeate intact cell membranes, it provides a reliable measure for the necrosis and late apoptosis of cells. Following a 6-h incubation, the percentages of necrotic cells were 0.1 ± 0.1% (for 300 mosmol/kg-H2O), 1.8 ± 0.3% (for 300 mosmol/kg-H2O—STS), 0.8 ± 0.1% (for 400 mosmol/kg-H2O—STS), 0.6 ± 0.1% (for 500 mosmol/kg-H2O—STS), and 0.5 ± 0.1% (for 600 mosmol/kg-H2O—STS; n = 8–10, for each condition). It is concluded from these experiments that STS, at 4 μM, is an effective and reliable tool for the induction of apoptosis in HeLa cells, as reported also in an earlier study from this laboratory [6]. It may be argued that hypertonic stress (and HICC activation) just delays, rather than prevents, STS-induced cell death in HeLa cells. As propidium iodide staining after 24 h of STS exposure revealed, however, the necrosis or late apoptosis of cells was 94.7 ± 0.7% for 300 mosmol/kg-H2O, and 76.1 ± 2.8% for 500 mosmol/kg-H2O. Clearly, even under these harsh experimental conditions (i.e. stimulation with STS for 1 day), hypertonicity rescued HeLa cells from death to a significant degree (P < 0.001; n = 8). In the last series of measurements, the timing of apoptosis induction by STS versus hypertonic stress was analysed. As depicted in Fig. 7, increasing osmolality from 300 to 500 mosmol/kg-H2O rescued HeLa cells from apoptosis induced by 6-h application of STS if this maneuver was performed 2 h before (HYPER2h/STS), 1 h before (HYPER1h/STS) or at the same time as the application of STS (STS-HYPER). Moreover, even with STS given 1 h ahead of hypertonic stress, apoptotic cell death was reduced to some 50% of the isotonic control value (Fig. 7). Of note, the overall profiles of the effects of the timing of STS application versus hypertonic stress on cell death and apoptosis, as shown in the figure, are undistinguishable from those observed on AVD (see Fig. 3) as well as on HICC activation (Fig. 4). Fig. 7Interference of HICC activation with STS-induced apoptosis. For the control, STS was applied for 6 h under isotonic conditions (STS-ISO). For the other groups, it was applied 2 h before (STS2h/HYPER), 1 h before (STS 1h/HYPER), at the same time as (STS-HYPER), 1 h after (HYPER1h/STS) or 2 h after (HYPER2h/STS) osmolality was increased from 300 to 500 mosmol/kg-H2O. Apoptosis was monitored as a decrease of cell viability (a, n = 12 to 42), caspase-3/7 activity expressed as relative fluorescence units (FU; b, n = 6 and 12), and DNA fragmentation measured in an enzyme-linked (ELISA) assay as absorbance values (OD, optical density; c, n = 6) It has long been known that AVD occurs in parallel with a loss of cellular K+ and a gain of cellular Na+ [20–22]. To some extent, these effects might reflect reduced activity of the Na+/K+-ATPase [23]. Nevertheless, if the activation of the HICC were triggered as a consequence of conductive K+ export and AVD, this would lead to very similar results. Also of interest in this respect is a recent study on human leukemia U937 cells reporting two distinct phases during STS-induced apoptosis: an initial phase in which cellular K+ and Cl− decreased in parallel, followed by a second phase in which there was a further decrease in cellular K+ and an enormous increase in cellular Na+ (accompanied by a rise in cellular Cl−) [24]. Of note, U937 cells were shown to express HICC currents [25]. Thus, the ionic scenario described could also reflect the interplay of HICC activation and AVD, as proposed in the present study. In rat hepatocytes, 405 mosmol/kg-H2O was insufficient to induce apoptosis but it sensitized cells to CD95-triggered cell death, whereas 505 mosmol/kg-H2O functioned as an effective apoptosis inducer per se [26–29]. Since rat hepatocytes express (amiloride-sensitive) HICC currents [4, 30–32], it would be interesting to examine the degree to which blockage of the HICC might interfere with the induction of cell death in this system. Conclusions Taken together, our results indicate that hypertonic stress caused a significant reduction in the staurosporine (STS)-induced apoptosis of HeLa cells. When analysing the functional interplay between apoptotic volume decrease (AVD) and regulatory volume increase (RVI) through the use of different activation protocols and effective blockers, the activation of hypertonicity-induced cation channels (HICCs) could be identified as the molecular mechanism by which HeLa cells are actually rescued from STS-induced apoptosis in hypertonic conditions. The precise mechanism by which HICC activation inhibits AVD and apoptosis remains to be elucidated. Nevertheless, the results reported herein demonstrate a novel role of the HICC in the induction of programmed cell death. They are also in perfect agreement with a recent study in which inhibition of the HICC was found to sensitize HeLa cells to shrinkage-induced apoptosis [33]. We have only begun to understand the role of the HICC in these processes [4, 5], and the general relevance of the channel in cell proliferation and apoptosis remains a question of considerable interest.	[ "cation channel", "apoptosis", "cell shrinkage", "hela cells", "volume regulation" ]	[ "P", "P", "P", "P", "P" ]
Knee_Surg_Sports_Traumatol_Arthrosc-3-1-1950586	Femoroacetabular impingement in 45 professional athletes: associated pathologies and return to sport following arthroscopic decompression	Femoroacetabular impingement (FAI) occurs when an osseous abnormality of the proximal femur (cam) or acetabulum (pincer) triggers damage to the acetabular labrum and articular cartilage in the hip. Although the precise etiology of FAI is not well understood, both types of FAI are common in athletes presenting with hip pain, loss of range-of-motion, and disability in athletics. An open surgical approach to decompressing FAI has shown good clinical outcomes; however, this highly invasive approach inherently may delay or preclude a high level athlete’s return to play. The purpose of this study was to define associated pathologies and determine if an arthroscopic approach to treating FAI can allow professional athletes to return to high-level sport. Hip arthroscopy for the treatment of FAI allows professional athletes to return to professional sport. Between October 2000 and September 2005, 45 professional athletes underwent hip arthroscopy for the decompression of FAI. Operative and return-to-play data were obtained from patient records. Average time to follow-up was 1.6 years (range: 6 months to 5.5 years). Forty two (93%) athletes returned to professional competition following arthroscopic decompression of FAI. Three athletes did not return to play; however, all had diffuse osteoarthritis at the time of arthroscopy. Thirty-five athletes (78%) remain active in professional sport at an average follow-up of 1.6 years. Arthroscopic treatment of FAI allows professional athletes to return to professional sport. Introduction In publications dating back several decades, irregularities in the morphology of the femoroacetabular articulation have been implicated as a possible source of hip pain in young, active athletes. A French article published in 1979 described hip pain associated with structural abnormalities of the proximal femoral neck in athletes participating in hockey, football, soccer, rugby, martial arts, and tennis [5]. More recently, developments pioneered by open hip surgeons have shown that morphologic abnormalities of both the femur and acetabulum underlie a large number of labral and chondral injuries in the hip [1, 29]. Wenger et al. [29] noted the presence of osseous abnormalities, including dysplasia and femoroacetabular impingement (FAI), in the majority of patients with labral tears. This suggests that isolated treatment of soft tissue pathologies may not be adequate without concomitantly addressing underlying structural abnormalities. The concept of FAI has been defined by Ganz and colleagues [1, 8, 27]. In this condition, a structural or spatial abnormality of the femur (cam) or acetabulum (pincer) damages the chondrolabral structures during normal joint movement. The most common situation is a mixed cam and pincer pathology, occurring along the anterior femoral neck and the anterior–superior acetabular rim. In high flexion and internal rotation movements, abutment and impingement of the labrum and cartilage occurs. The precise etiology of FAI is not well understood. Sub-clinical slipped capital femoral epiphyses [15, 25], mal-union of femoral neck fractures [6], and decreased femoral anteversion [26] are described causes of cam impingement. Relative posterior opening of the acetabulum (acetabular retroversion) or global (coxa profunda) overcoverage of the femur by the acetabulum are described causes of pincer impingement [23, 24]. Cam and pincer lesions lead to distinct patterns of labral and chondral damage [1, 10] and long-standing impingement is likely a significant cause of previously described idiopathic hip joint degeneration [1]. Ganz et al. developed an open surgical dislocation approach to decompress FAI [7, 13]. This technique has demonstrated good results of FAI decompression in a general population [2]. However, we believe that an arthroscopic approach involves less post-operative morbidity and allows patients, including professional athletes, to return to high-functioning lifestyles [28]. The purpose of this study was to define associated pathologies and determine if professional athletes could return to high-level athletics following arthroscopic decompression of FAI. Methods Patient selection A retrospective chart review was performed of all professional athletes presenting for arthroscopic treatment of FAI by the senior author between October 2000 and September 2005. All athletes presented with debilitating hip pain and an inability to participate in their sport. All were diagnosed with FAI based on physical and radiographic examination. Physical examination criteria included a positive impingement test [14] or flexion–abduction–external rotation (FABER) test. A positive impingement test was defined as groin pain with 90 degrees of hip flexion and maximal internal rotation. A positive FABER test was defined as asymmetry in the distance between the lateral knee and the exam table between the injured hip and the non-injured hip. Radiographic criteria included decreased anterior and superior femoral head–neck offset, acetabular retroversion (as defined by a positive cross-over sign), or coxa profunda (as defined by the medial acetabular border overlapping the ilioischial line). For these athletes with documented physical and radiographic evidence of FAI, conservative treatment was limited to only 6 weeks. Based on evidence described by Beck and colleagues [1], our thought was that conservative treatment could not address the underlying bony abnormalities and any further treatment delay in these highly active patients would likely precipitate further irreversible damage to the articular cartilage. Inclusion criteria for this study included professional athletes with at least one positive physical exam finding, at least one positive radiographic finding, and failure of at least 6 weeks of conservative therapy. Forty-five professional athletes from various sports met the inclusion criteria and were included in the study (Table 1). There were 42 male and 3 female patients. Table 1Pre-operative sports activities of the 45 professional athletesSportNo. of patientsHockey24Golf6Football5Soccer3Dance2Baseball2Martial arts1Tennis1Jockey1Total45 Arthroscopic procedure The modified supine approach to hip arthroscopy was used [3]. Two portals (anterior, anterolateral) were established and a third portal (distal lateral accessory) was established as needed [12]. At the time of arthroscopy, the central compartment of the hip was first inspected. Pincer lesions were typically identified in the superior acetabular quadrant. For the treatment of small pincer lesions (<2 to 3 mm), osteoplasty of the proximal femoral neck was occasionally sufficient to relieve the impingement [28]. Patients with moderate or large-sized lesions underwent acetabular rim trimming, using an arthroscopic osteotome or 5.5 mm round motorized burr [22]. Acetabular labral tears were arthroscopically identified, and based on previous research on the vascularity of the labrum [9, 11], detached and peripheral midsubstance tears were typically repaired [12, 22]. Care was taken to preserve as much healthy, viable labrum as possible, however, degenerative or frayed tissue was debrided to a stable remnant. Labral repair was performed with suture anchors to repair detached labral tears or to re-fix the labrum following iatrogenic detachment for complete resection of pincer lesions. The previously described technique involved placement of the anchors high on the acetabular rim in the area of detachment [12]. Typically, one bioabsorbable anchor (BioRaptor, Smith+Nephew, Andover, MA, USA) was placed at the 12:00 acetabular position, and re-inforcement was placed either anteriorly or posteriorly of this area as needed. One limb of suture (Ultrabraid, Smith+Nephew, Andover, MA, USA) was passed between the labral tissue and the rim, and was retrieved through the substance of the labral tissue. In cases of inadequate availability of labral substance or in patients with highly degenerative, friable tissue, the suture was passed around the labral tissue. Standard arthroscopic knots fixed the repair to the rim. For midsubstance labral tears, a suture passer looped an 0-Vicryl around the torn tissue to approximate the edges of the midsubstance split. The condition of the articular cartilage was evaluated according to the Outerbridge Classification System [18]. For this study, the Outerbridge Classification System was defined as follows: grade 0, normal cartilage; grade 1, cartilage with softening and swelling; grade 2, a partial thickness defect with fissures on the surface that do not reach the subchondral bone or exceed 1.5 cm in diameter; grade 3, fissuring to the level of the subchondral bone in an area with a diameter more than 1.5 cm; and, grade IV, exposed subchondral bone [18]. Chondral procedures performed in this group included chondroplasty and microfracture of lesions of the femoral head and acetabulum as needed. Following inspection of the central compartment, traction was released and the peripheral compartment was evaluated for cam lesions. A dynamic examination of the hip in all motion planes allowed for direct visualization of impingement at the femoroacetabular interface. For the treatment of cam lesions, an osteoplasty was performed with a 5.5 mm round motorized burr, restoring the anatomic concavity of the femoral head–neck junction [28]. Burring was performed from superior to inferior along the anterior femoral head–neck junction, to a depth of approximately 5–7 mm and a width of approximately 8–12 mm. Care was taken to avoid resecting too distally along the femoral neck to avoid the lateral epiphyseal vessels. A final dynamic examination was performed to confirm adequate decompression, and a lack of entrapment of the labrum during joint motion. In general, the above-described procedure took approximately 2–3 h, depending on the degree of labral and chondral pathology. Traction time was limited to less than 2 h to reduce the risk of neurovascular compromise. Post-operative management Post-operative medical prophylaxis included an aspirin or a low-molecular weight heparin, a non-steroidal anti-inflammatory, and an antibiotic. Each of the 45 athletes underwent post-operative rehabilitation according to the following generalized protocol that was devised based on personal experience. For basic procedures involving decompression of FAI and labral treatment, weightbearing was restricted to 20 pounds, flatfoot for 4 weeks. A brace was prescribed to be worn for 10 days to protect the hip and limit abduction and rotation. Continuous passive motion (CPM) was used to apply 0–90 degrees of hip flexion for up to 8 h per day for 4 weeks. Night boots were worn for ten post-operative days to limit rotation during sleep. If capsular modification procedures (thermal capsulorrhaphy or capsular plication) were performed, rotation precautions were extended for a total of 21 days to avoid capsular stress. If microfracture was performed, weightbearing restrictions and CPM use were extended for a total of 8 weeks. Physical therapy exercises were implemented within four post-operative hours. It is our belief, based on clinical observation, that this, in addition to an early emphasis on passive hip motion (particularly internal rotation) reduces the incidence of adhesion formation. Active hip flexion was limited for 4 weeks, based on clinical observation, to minimize the risk of hip flexor tendonitis. Athletes were typically allowed to return to full competitive activity between 12 and 16 post-operative weeks. Return-to-play statistics were determined by retrospective chart review and personal follow-up communication with the athletes. Results The average age of the athletes at the time of surgery was 31 years (range: 17–61). Eleven athletes (24%) previously underwent hip arthroscopy by multiple primary surgeons for the treatment of labral and chondral pathologies and experienced a recurrence of hip symptoms. The average time to follow-up was 1.6 years (range: 6 months to 5.5 years). Twenty-two athletes (49%) were treated for an isolated cam lesion and three athletes (7%) were treated for an isolated pincer lesion. Twenty-one athletes (47%) had a mixed pathology of both cam and pincer lesions. All of the athletes had acetabular labral tears. Twenty-five patients (56%) underwent either labral repair or re-fixation following rim trimming with suture anchors (average 1.3 anchors per patient, range: 1–3). Twelve patients (27%) had intra-substance labral repair. Five patients had labral debridement; one patient had a detached tear, one had frayed labral tissue, and three patients had complex tears consisting of detached, frayed, and flap components. Two patients were labral deficient from a previous arthroscopic debridement. They underwent labral grafting using an iliotibial band autograft. Twenty-one patients (47%) had a grade IV acetabular chondral defect; 14 underwent arthroscopic microfracture, 5 underwent thermal chondroplasty, and 2 patients had no treatment due to the diffuseness of their disease. Three patients (7%) had a grade IV femoral head chondral defect; one was treated with microfracture, one with chondroplasty, and one patient had no treatment due to the diffuseness of his disease. Seventeen (38%) had a grade I–III acetabular chondral defect (13 treated with chondroplasty and 1 treated with microfracture) and 11 (24%) had a grade I–III femoral head chondral defect (all 11 were treated with chondroplasty). Four patients (one baseball player, one football player, one hockey player and 1 golfer) had extensive diffuse OA at the time of arthroscopy, but opted to delay arthroplasty. Twenty-six patients (58%) had a partial tear of the ligamentum teres and 3 patients (7%) had a complete ligamentum teres avulsion. Thirteen (29%) patients underwent thermal capsulorrhaphy and 9 (20%) underwent plication for capsular redundancy. Seventeen patients (38%) had loose bodies in the joint and two patients required excision of myositis ossificans of the rectus femoris. Three patients (7%) underwent cheilectomy of a stenotic cotyloid fossa and 9 patients (20%) underwent arthroscopic release of a tight iliopsoas. Figure 1 illustrates restoration of femoral head–neck concavity in an athlete treated for cam impingement (Fig. 1) and Fig. 2 illustrates successful removal of the anterior acetabular overhang in an athlete treated for pincer impingement (Fig. 2). Fig. 1A 27-year-old professional hockey player was evaluated for right hip pain. a A pre-operative cross-table lateral radiograph demonstrated convexity at the anterior femoral head–neck junction. b The presence of a cam lesion was verified arthroscopically. The femoral head (FH) and acetabulum (Ac) are visualized in the peripheral hip compartment with the camera in the anterior portal. c A motorized burr was used to restore anterior concavity at the head–neck junction. This was visualized at the superior acetabular (Ac) position (12:00) with the camera in the anterior portal. d A post-operative film verified successful decompression of the cam impingement and restoration of the femoral head–neck junction concavityFig. 2A 22-year-old professional football player was evaluated for right hip pain. a A pre-operative AP radiograph demonstrated a cross-over sign of the right acetabulum, indicating acetabular retroversion. b The presence of a pincer lesion was verified arthroscopically with the camera in the anterolateral portal with excessive bone along the anterior–superior acetabulum (Ac). c A motorized burr in the anterior portal resected the lesion with the burr shield placed against the labrum (L) for protection. d The labrum was fixed back to the rim using two suture anchors. e A post-operative film verified the successful removal of pincer impingement and lack of retroversion in the superior region of the acetabulum Forty-two athletes (93%) returned to professional sport following hip arthroscopy. Three players (1 football player, 1 hockey player, and 1 baseball player) did not return to play following arthroscopy. Each of these patients had diffuse osteoarthritis at the time of arthroscopy. Five athletes (11%) required re-operation. Three underwent lysis of adhesions and two had symptomatic treatment of extensive osteoarthritis. All of the patients who underwent revision surgery for lysis of adhesions returned to professional play and the two with extensive osteoarthritis did not return to play. Thirty-five of the 45 athletes (78%) remained active at the professional level at an average of 1.6 years after hip arthroscopy. Discussion Intra-articular hip injuries are a recognized cause of pain, mechanical symptoms, and disability in athletes. Traumatic intra-articular injury results from acute injury, including hyperabduction injuries, direct hip contact, and joint dislocation or subluxation. Atraumatic hip injury, however, is less well understood. Various motions exerted during sport, particularly flexion combined with internal rotation (hockey-goalie stance), have been suggested as potential causes of overuse hip injury [19, 21]. In this position, a cam lesion extending anteriorly or superiorly from the proximal femoral neck would impinge at the anterosuperior acetabular rim. Combining this understanding with recent reports by Wenger et al. [29] and Ganz et al. [1, 2, 8, 13], morphological abnormalities of the femoroacetabular joint have been closely examined as possible underlying sources of atraumatic hip injuries in athletes. Ganz and colleagues described FAI resulting from structural abnormalities of the proximal femoral neck (cam), acetabular rim (pincer), or most commonly, a combination of the two pathologies (mixed). Damage to the articular cartilage and acetabular labrum results from this pathologic bony contact and FAI is a likely trigger of early hip joint degeneration. In our experience, FAI is a common problem that has only recently been recognized in athletes with a primary hip complaint. An open surgical dislocation approach was first described to treat hip impingement [7]. This technique showed good midterm results in a general population [2]. However, increasing recognition of FAI in a high-demand, competitive athletic population encouraged the development of a less invasive arthroscopic approach to facilitate prompt return to competitive play. Various studies have demonstrated that hip arthroscopy is a safe and effective intervention in athletes with intra-articular injuries [4, 17, 19, 20]. Philippon reported on a series of ten elite athletes who underwent hip arthroscopy for labral debridement with thermal capsulorrhaphy [20]. All of the athletes returned to high-level athletic activities. McCarthy et al. [17] reported 80% excellent results following hip arthroscopy in elite athletes (average follow-up of 23.6 months). Byrd et al. [4] reported a series of 44 hip arthroscopies in 42 recreational, high school, collegiate, elite, and professional athletes. Post-operative improvement, as quantified by the modified Harris Hip Score, was present in all classes of athletes and in athletes undergoing any arthroscopic procedure (removal of loose bodies, debridement of ligamentum teres, excision of osteophyte, labral excision, microfracture, chondroplasty). Byrd et al. reported better results in athletes who recalled a traumatic onset to their hip symptoms, when compared to those with an acute or insidious onset. The authors suggested that an unaddressed pre-disposition to injury might have had a negative impact on self-reported outcomes in athletes [4]. FAI may be the unaddressed pre-disposition to which Byrd and colleagues were alluding. While these above articles demonstrated successful outcomes following debridement procedures in athletes, there are a few caveats that must be addressed. First, it is likely that simple excision of damaged or diseased tissue, like with meniscal tears or loose bodies, can provide short-term abatement of pain and mechanical symptoms, and allow an athlete to return to professional athletics. However, recent studies have shown that simple hip debridement procedures failed to address the primary pathology in a vast majority of cases [1, 29]. A recent unpublished study performed by the senior author showed that a major cause of revision hip arthroscopy in all patients was treatment of labral and chondral re-injury and decompression of previously unaddressed FAI. Future studies examining the clinical and radiographic long-term outcomes of athletes treated with debridement procedures need to be compared to those athletes treated for FAI at an early stage of this disease. The purpose of this study was to simply state that despite the fact that FAI surgery is a more extensive procedure than debridement, professional athletes can still return to play following this procedure. An arthroscopic approach to treat FAI was recently reported [28]; however, early outcome studies are lacking. The results of this study show that professional athletes with FAI can return to high-level competitive sport following this procedure. Ninety-three percent of athletes returned to their previous level of sport. Three athletes did not return to sport (1 hockey player, 1 football player, and 1 baseball player), however all had diffuse hip osteoarthritis present at the time of arthroscopy. One patient (a senior PGA golfer) was able to return to sport following arthroscopy, despite his diffuse osteoarthritis. Based on this information, it seems that arthroscopic treatment of FAI in the presence of OA can allow patients to return to low-impact, however, not likely high-impact professional sports. A larger cohort of patients is needed to test this hypothesis. Avascular necrosis (AVN) and femoral neck fracture are recognized risks of osteoplasty. A recent study showed that up to 30% of the femoral neck may be resected before it is structurally compromised [16]. In this series of patients, no post-operative cases of AVN or femoral neck fracture were diagnosed. Additional potential complications include adhesion formation at either the capsulolabral margin or between the capsule and the site of bony resection. In this series of 45 athletes, 3 patients (7%) required revision arthroscopy for the lysis of adhesions (LOA). All athletes returned to competitive sport following revision arthroscopy for LOA. Two additional athletes had extensive OA at the time of primary arthroscopy and subsequently underwent revision arthroscopy for symptomatic treatment of OA, including osteophyte debridement. Both did not return to professional athletics due to the extent of their disease. There were no reported cases of heterotopic ossification or neurovascular compromise, including lateral femoral cutaneous, sciatic, and pudendal neuropraxia, reported, although this statistic may have been underestimated due to the nature of this professional athlete cohort. A limitation of this study was the inherent selection bias involved with the study of professional athletes. This patient sub-group was financially motivated to return to play and may have been less likely to report post-operative symptoms and complications than the rest of patients treated for FAI. Despite this, we believed that a return-to-play analysis was critical to assess the outcomes of FAI treatment in a high-demand population. This cohort of professional athletes was debilitated prior to hip arthroscopy, and following intervention, was able to perform in physically intense professional sport activities. Also, the athletes’ precise reason for retirement and whether it was related to their hip injuries was difficult to discern. For this reason, given the relatively short span of all professional athletes’ careers, we chose to include all athletes at a minimum of 6 months post-operative. This time period was selected as an ample time period for completion of post-operative rehabilitation and return to sport. Another limitation of this study was the lack of follow-up subjective and objective data. Current outcomes instruments (modified Harris Hip score and the Non-Arthritic hip score) have not been validated for use in high-level athletes. It is our belief that these scoring systems fail to address the activities that are most limiting to athletes, and hence, underestimate the degree of debilitation in professional athletes. Evaluating the applicability of current scoring systems in athletes, and developing new outcomes instruments for athletes will be the focus of future studies. Other limitations of this study included unequal distribution of gender (42 males and 3 females) and professional sports. Because different sports place different demands on the hip, we cannot draw conclusions among the various sports. The arthroscopic treatment of FAI represents the evolution of hip arthroscopy. In the past, soft tissue pathologies were treated in isolation, without addressing underlying impinging osseous abnormalities. Recent developments, particularly by Ganz and colleagues, have enlightened us to the pathology of FAI and the associated treatment options. In order to treat high-demand patients with FAI, an arthroscopic technique was developed. This study has demonstrated that full return to professional competitive sport is possible following arthroscopic treatment of FAI. Additional studies are needed to determine the effect on long-term joint degeneration of early surgical intervention to treat FAI.	[ "femoroacetabular impingement", "cam", "pincer", "hip arthroscopy", "osteoplasty" ]	[ "P", "P", "P", "P", "P" ]
Immunogenetics-2-2-1592254	A novel family of diversified immunoregulatory receptors in teleosts is homologous to both mammalian Fc receptors and molecules encoded within the leukocyte receptor complex	Three novel and closely related leukocyte immune-type receptors (IpLITR) have been identified in channel catfish (Ictalurus punctatus). These receptors belong to a large polymorphic and polygenic subset of the Ig superfamily with members located on at least three independently segregating loci. Like mammalian and avian innate immune regulatory receptors, IpLITRs have both putative inhibitory and stimulatory forms, with multiple types coexpressed in various lymphoid tissues and clonal leukocyte cell lines. IpLITRs have an unusual and novel relationship to mammalian and avian innate immune receptors: the membrane distal Ig domains of an individual IpLITR are related to fragment crystallizable receptors (FcRs) and FcR-like proteins, whereas the membrane proximal Ig domains are related to several leukocyte receptor complex encoded receptors. This unique composition of Ig domains within individual receptors supports the hypothesis that functionally and genomically distinct immune receptor families found in tetrapods may have evolved from such ancestral genes by duplication and recombination events. Furthermore, the discovery of a large heterogeneous family of immunoregulatory receptors in teleosts, reminiscent of amphibian, avian, and mammalian Ig-like receptors, suggests that complex innate immune receptor networks have been conserved during vertebrate evolution. Introduction Immune receptors that bind similar ligands often form clusters that are encoded on distinct chromosomes in the human genome. For example, receptors that bind to the Fc portion of Ig [fragment crystallizable receptors (FcRs)] are found on chromosome 1q21-23 and consist of the high-affinity IgG and IgE receptors (FcγRI and FcεRI) and the related low- or moderate-affinity Ig receptors (FcγRII, FcγRIII, and FcγRIV) (Hullet et al. 1991; Ravetch and Kinet 1991; Hullet and Hogarth 1994; Daeron 1997; Falk et al. 2005; Falk and Ravetch 2006). Alternatively, immune receptors involved in the recognition of major histocompatability class 1 (MHC I) antigens, such as killer cell Ig-like receptors (KIRs) and leukocyte Ig-like receptors (LILRs), are clustered together on chromosome 19q13.4 in a region known as the leukocyte receptor complex (LRC) (Kremer et al. 1992; Wagtmann et al. 1997; Wende et al. 1999; Martin et al. 2002). Recently, comparative studies have suggested that FcRs and LRC-encoded receptors, although genomically and functionally distinct, appeared to have evolved from a common ancestor. This contention is supported in part by recent attempts to find LRC regions in other species. Mouse chromosome 7 is syntenic to the human LRC and encodes the paired Ig-like receptors (PIRs) that were originally discovered in an attempt to find a murine IgA receptor ortholog (CD89; Kubagawa et al. 1997; Hayami et al. 1997). Subsequently, searches for PIR relatives resulted in the identification of the chicken Ig-like receptors (CHIRs) in birds (Dennis et al. 2000), representing a multigene family encoded within an avian LRC-like region (Nikolaidis et al. 2005a; Viertlboeck et al. 2005). More importantly, CHIR Ig domain sequences are not only related to the mammalian PIRs and KIRs, but also to various FcR Ig domains (Dennis et al. 2000). This previously unrecognized relationship was supported by phylogenetic and comparative structural analysis and was the first evidence suggesting that certain genomically and functionally diverse mammalian and avian immune receptor families probably evolved from a common ancestor (Dennis et al. 2000; Nikolaidis et al. 2005b). Compared to mammals and birds, bony fish are among the most phylogenetically distant vertebrates with characteristic innate and adaptive immune systems that have mammalian-like B and T cells (Yoder 2004). Furthermore, due to the establishment of clonal lymphoid cell lines, catfish are known to also possess natural killer (NK)-like cells capable of killing xenogeneic targets by an antibody-dependent cell-mediated cytotoxicity (ADCC) process (Shen et al. 2002–2004). This functional evidence suggested that catfish NK cells possess a putative FcμR (Shen et al. 2003). Additionally, catfish NK cells can kill allogeneic targets in an ADCC-independent manner, indicating that target recognition may also involve KIR and/or LILR-like receptors (Shen et al. 2002, 2004). However, to date, FcR and/or KIR/LILR homologs have not been described in ectothermic vertebrates, making it difficult to elucidate the evolutionary history of these important receptor families. Large families of putative activatory and inhibitory novel immune-type receptors (NITRs) also belonging to the Ig superfamily (IgSF) have been described in teleosts. Pufferfish (Spheroides nephelus), zebrafish (Danio rerio), catfish, and rainbow trout (Oncorhynchus mykiss) all express NITRs (Strong et al. 1999; Yoder et al. 2001, 2002; Hawke et al. 2001). Because some NITRs are also expressed by catfish NK cells (Hawke et al. 2001), it has been suggested that these receptors may represent the “functional orthologs” of the mammalian KIRs and/or LILRs (Yoder et al. 2001; Hawke et al. 2001). However, NITRs do not exhibit any sequence identity or phylogenetic relationship to such receptors (Yoder et al. 2001) and have only been identified in bony fishes. Herein is described a novel set of IgSF receptors present in the channel catfish, and possibly other teleosts, that are composed of Ig domains related to functionally diverse immunoreceptor families present in mammals. Detailed sequence analysis supports that individual catfish leukocyte immune-type receptors (IpLITRs) contain a unique composition of Ig C2-like domains sharing homology with both FcRs and receptors encoded within the LRC. Thus, IpLITRs provide further evidence supporting the common evolutionary origins of genomically and functionally distinct immunoregulatory receptor families present in mammals. Materials and methods Experimental animals and cell lines Catfish (1–2 kg) were obtained from a commercial source (ConAgra, Isola, MS, USA) and maintained in individual tanks as described previously (van Ginkel et al. 1992). Catfish peripheral blood leukocytes (PBL) were isolated from heparinized blood by centrifugation on a cushion of Ficoll-Hypaque (Lymphoprep, Accurate Chemical, Westbury, NY, USA) as described previously (Miller et al. 1994a). Catfish cells lines were grown at 27°C in AL-3 medium consisting of equal parts AIM-V and L-15 (Life Technologies, Gaithersburg, MD, USA) adjusted to catfish tonicity with 10% (v/v) deionized water and supplemented with 1 μg/ml NaHCO3, 50 U/ml penicillin, 50 μg/ml streptomycin, 20 μg/ml gentamicin, 50 uM 2-ME, and 3% heat-inactivated, pooled, normal catfish serum (Miller et al. 1994a,b). 3B11 is a cloned autonomous B cell line generated from an outbred catfish by mitogen stimulation (Wilson et al. 1997). G14D is an automonous T cell line derived from a gynogenetic catfish (Hogan et al. 1999), whereas 42TA is a macrophage cell line derived from a different outbred fish (Miller et al. 1994b). TS32.15 and TS32.17 are clonal nonautonomous antigen-dependent cytotoxic T cell lines developed from an alloantigen immunized outbred fish (Stuge et al. 2000). 1F3 is a clonal nonautonmous NK-cell line developed from an alloantigen-stimulated culture of PBL from a naive outbred fish (Shen et al. 2004). The nonautonomous cytotoxic T cell and NK cell lines require weekly restimulation with irradiated allogeneic cells for continuous proliferation. The channel catfish ovary fibroblast cell line (CCO; Browser and Plumb 1980) was provided by Dr. V.G. Chinchar (University of Mississippi Medical Center). CCO cells were maintained in Dulbecco’s modified Eagle’s medium (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% FCS. Channel catfish LITRs and sequence analysis In a search for IgSF receptors expressed by alloantigen stimulated catfish PBL and various catfish clonal cell lines, several expressed sequence tag (EST) libraries were analyzed. Subsequently, IpLITR1 and IpLITR3 cDNAs were identified by single pass sequencing of a cDNA library generated from a catfish 7-day-old mixed lymphocyte culture (MLC) enriched in NK-like cells, MLC52-1 (Shen et al. 2004; Stuge et al. 1997). IpLITR2 was obtained from the catfish 42TA macrophage cell line (Miller et al. 1994b) cDNA library. The cDNAs were sequenced on both strands using universal forward and reverse primers and gene-specific primers. Nucleotide and amino acid sequences were analyzed using DNASTAR software (Madison, WI, USA) and aligned using CLUSTALW (Thompson et al. 1997). Neighbor-joining (NJ) trees with pairwise gap deletions were drawn using MEGA v3.0 (Kumar et al. 2001). Maximum parsimony and maximum likelihood trees were drawn with essentially the same results (data not shown). Similarity searches were performed with the extracellular Ig domains of IpLITR1, IpLITR2, and IpLITR3 using position-specific iterative (PSI) and Basic Local Alignment Search Tool (BLAST) analysis (Altschul et al. 1990) against the National Center for Biotechnology Information (NCBI) nonredundant database. SOSUI (signal) beta version, Simple Modular Architecture Research Tool (SMART) (Letunic et al. 2004), Pfam databases (Bateman et al. 2004) and the three-dimensional position-specific scoring matrix server (Kelly et al. 2000) were used for predictions of signal peptides, Ig domains, transmembrane (TM) segments, and secondary structure. Sequence decorations were performed using GeneDoc (http://psc.edu/biomed/genedoc). Southern and Northern blot analyses Genomic DNA was prepared from erythrocytes of 16 catfish siblings. These siblings were representatives of a 78-member family obtained from the Catfish Genetics Research Unit, United States Department of Agriculture, Agricultural Research Service (Stoneville, MS, USA; designated family 1). The sibling catfish 1–16 used were A10, A5, A32, A22, A38, A36, A16, A29, A18, A49, A159, A15, A23, A25, A186, and A37, respectively (Quiniou et al. 2005). The DNA (10 μg) was digested to completion with Pst1 or EcoRI, separated on 1% agarose gels and transferred by capillary action onto Hybond-N+ membranes (Amersham Pharmacia Biotech, Arlington Heights, IL, USA) using standard techniques. Hybridizations were performed in Rapid-hyb buffer (Amersham Pharmacia Biotech) at 65°C and membranes were washed at high stringency (65°C with 0.1× saline-sodium citrate, 0.1% sodium dodecyl sulfate). Specific probes (Table 1) were amplified by PCR using IDPol DNA polymerase (ID Labs Biotechnology, London, Ontario, Canada) according to the manufacturer’s recommended protocol. Typical parameters were: 1 min 94°C, followed by 29 cycles of 94°C 30 s, 62°C 30 s, 72°C 1 min, then extension at 72°C for 10 min. Probes were random primed labeled with [32P] 2′-deoxycytidine 5′-triphosphate using a Megaprime labeling kit (Amersham Pharmacia Biotech). Table 1IpLITR PCR primers used for generating Ig domain-specific probes and gene-specific primers used in RT-PCR experimentsPrimer IpLITR1 D1Forward5′ GTCGGACAAGGTCAAGTTCTG 3′Reverse5′ GGCTTTTGGTCTCTCTATCAC 3′IpLITR1 D4Forward5′ AATCCTGATGAACAGGTGTACC 3′Reverse5′ GTGTTTACAGTGGTAGAAACC 3′IpLITR2 D3Forward5′ AGTCGTGAAGGAGCTGTACTGA 3′Reverse5′ TTCACTGCCAGAATGATGGTCAC 3′IpLITR3 D6Forward5′ CTGAGGGACATCCTCTGACCTT 3′Reverse5′ GTGTTTACAGTGGTAGAAACC 3′IpLITR1 GSPForward5′ GTCGGACAAGGTCAAGTTCTG 3′Reverse5′ CTGCAGACATGATGAACTTCT 3′IpLITR2 GSPForward5′ GTCGGACAAGCTCAAGAGTTT 3′Reverse5′ TGTGTAGTAGAGTGGGTTTCC 3′IpLITR3 GSPForward5′ GTCGGACAAGGTCAAGTTCTG 3′Reverse5′ GCTTTTGGGTGACTCTCCTCT 3′IpEF1α GSPForward5′ GACTGCCACACTGCTCACATTG 3′Reverse5′ TTAGTTACTCAGCAGCTTTCTTCCGSP gene-specific primers Total RNA from catfish PBL and various tissues was prepared using RNA-Bee (Tel-test, Friendswood, TX, USA); 10 μg of each sample were electrophoresed on 1.5% formaldehyde-agarose gels and transferred by capillary action onto Hybond-N+ (Amersham Pharmacia Biotech), hybridized, and washed as above. Reverse transcription (RT)-PCR Total RNA from PBL, MLC, various tissues, and catfish clonal cell lines were prepared as described above, and treated with DNase I (Invitrogen Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s recommended protocol. One μg RNA was reverse transcribed using an oligo-T primer and 200 units of Superscript III reverse transcription (RT) (Invitrogen Life Technologies). RT-PCR was performed using specific primers for IpLITR and the housekeeping gene elongation factor 1 alpha (IpEF1α; Table 1). Typical parameters were: 3 min 94°C, followed by 30 cycles of 94°C 45 s, 58°C 45 s, 72°C 2 min, then extension at 72°C for 10 min. PCR products were visualized on 1.2% Tris-acetate-EDTA agarose gels. PCR products were also cloned into pCR4-TOPO® (Invitrogen Life Technologies) and verified by sequencing. The Genbank accession number for IpEF1α is CB938718. Results Catfish LITR sequences In a search for IgSF receptors expressed by alloantigen-stimulated catfish PBL and various catfish clonal cell lines, several EST libraries were analyzed. Three IpLITR sequences were subsequently identified that encoded type I TM proteins with extracellular C2-like Ig domains, which were predicted using SMART (see Electronic Supplementary Material). Even though these sequences vary in length, they are highly similar (Fig. 1a,b). All three receptors have identical signal peptides and IpLITR1 and IpLITR3 have almost identically encoded Ig domains D1, D2, D3, and D4 (Fig. 1a). Comparatively, IpLITR2 D1 and D2 are 77.2 and 86.8% identical at the amino acid level to their IpLITR1 and IpLITR3 counterparts (Fig. 1c). IpLITR1 encodes a 346 amino acid extracellular region consisting of four Ig domains, a 23-amino acid TM segment and a 116-amino acid cytoplasmic tail (CYT). The IpLITR1 CYT contains two immunotyrosine-based inhibition motifs (ITIMs) centered at Y439 and Y461 and an ITIM-like motif (SEYTTE) centered at Y479 (Fig. 1b) (Daeron and Vivier 1999; Ravetch and Lanier 2000; Billadeau and Leibson 2002). An overlapping immunotyrosine-based switch motif (ITSM; TVYSQL) centered at Y465 is also present in the CYT of IpLITR1 (Shlapatska et al. 2001). In contrast, the smaller IpLITR2 and the longer IpLITR3 transcripts encode for molecules with identical 25-amino acid TMs containing a lysine residue and very similar charged CYTs (Fig. 1b). Overall, the individual IpLITRs are composed of a membrane distal to membrane proximal ordering of Ig domains, each containing similar D1s and D2s. However, the membrane proximal domains vary, i.e., IpLITR3 D5 and D6 are only 15.7–24.7% and 15.2–25% identical, respectively, to all other IpLITR Ig domains, and IpLITR2 D3 is only 17.9–39.3% identical to other IpLITR domains (Fig. 1c). Fig. 1Predicted amino acid sequence, domain comparisons, and schematic representation of IpLITR1, IpLITR2, and IpLITR3. a Alignment of the extracellular and b TM/CYT regions of IpLITR1, IpLITR2, and IpLITR3. Signal peptide (SP), and immunoglobulin (Ig) domains are labeled; cysteine residues predicted to be involved in intrachain disulfide bonds are marked with asterisks; gray shaded residues represent differences from IpLITR1 in a and differences from IpLITR2 in b. TMs are underlined, ITIM-like motifs are boxed, an overlapping ITSM within the CYT of IpLITR1 is indicated by a bracket, and TM charged residues are shaded black and marked (+). c Phylogenetic analysis of Ig domains in IpLITRs. NJ trees with pairwise gap deletions were drawn using MEGA v3.0 (Kumar et al. 2001) with 10,000 bootstrap replications, and bootstrap values >50% are shown. Branch lengths were measured in terms of amino acid substitutions and a scale bar are shown below the trees. The predicted SP, Ig domains, TM, and CYT are indicated. ITIM-like motifs are shown as boxes, N-linked glycosylation sites are marked as ballpoint lines. Individual IpLITR domains are shaded according to their relatedness between IpLITRs and percent amino acid identity with IpLITR1 Ig domains indicated to the left of IpLITR2 and IpLITR3. IpLITR2 D3 and IpLITR3 D5 and D6 are 15.2–39.3% identical to all IpLITR1 Ig domains IpLITRs belong to polymorphic and polygenic gene families Sibling restriction fragment length polymorphism (RFLP) patterns using domain-specific probes for the various IpLITRs show a large number of hybridizing bands with each probe used, as well as differences in the banding patterns among siblings (Fig. 2). These findings indicate that the IpLITR gene complex is both polygenic and polymorphic. However, it is presently unknown whether these polymorphic differences are due to variability within the exons encoding these receptors and/or within the introns. Some of the high-intensity bands found in each of the Southern blots may reflect multiple IpLITR copies, further increasing the possible number of genes. Importantly, the Southern blot segregation patterns show that there is more than one IpLITR locus. For example, among the 16 sibling fish used there are at least nine, perhaps more, different RFLP patterns for IpLITR D1 (Fig. 2b). If these genes were closely linked then only four RFLP patterns (barring cross-over events) would be observed, as shown for the MHC class IIβ RFLP patterns (Fig. 2a). When all Southern blots are compared, at least three distinct linkage groups (marked by ●, ▲, and ■ ) are observed, indicating the presence of multiple independently segregating loci, and none of these are linked to the MHC II RFLP (Fig. 2b–d). Similar results were also obtained using the IpLITR2 D3 probe (data not shown). Mining of the zebrafish genome for IpLITR relatives [whole genome shotgun database of the NCBI (Sanger Genomic Institute, Cambridge, UK)] further supports that these receptors are present on different chromosomes because five contigs were identified with IpLITR-like homologs on zebrafish chromosomes 3, 7, and 8 (Table 2). Fig. 2Southern blot analyses. Genomic DNA from 16 sibling catfish was digested, separated, transferred to nylon membranes, and hybridized with a MHCIIβ, b IpLITR1 D1, c IpLITR1 D4, and d IpLITR3 D6-specific probes. Letters a–i indicate different segregation patterns (b′, e′, and h′ represent variant segregation patterns that may be the result of recombination). Representative RFLP bands illustrating different linkage groups are indicated by ●, ▲, and ■. Kilobase size markers are indicated to the left of each blotTable 2Identification of IpLITR-like genes in the zebrafish genomeAccession numbersChromosomeBX0051568BX0005248BX5111763BX6496273CAAK01000490.17Search (04/05) of the zebrafish whole genome shotgun database of the NCBI (Sanger Genomic Institute) IpLITR expression Northern blot analyses using a cross hybridizing IpLITR1 D1 probe revealed that LITR message is highly expressed in the hematopoietic pronephros and mesonephros, with predominant transcript sizes of ∼6.6, ∼4.0, and ∼3.8 Kb (Fig. 3a). Low levels of IpLITR D1 expression were detected in the spleen, with little to no message detected in the heart, liver, gill, or muscle. The use of the more specific IpLITR2 D3 probe revealed message expression not only in the pronephros and mesonephros, but also in the spleen, gill, and perhaps heart with transcripts of ∼6.0, ∼3.7, ∼2.7, and ∼1.9 Kb (Fig. 3b). RNA load levels for each of the tissues tested were similar based upon hybridization with a catfish EF1α probe. RT-PCR, using primers specific for the putative inhibitory IpLITR1 and the putative activating IpLITR2 as representative IpLITRs, was also used to assess tissue expression. As shown in Fig. 3c, IpLITR1 and IpLITR2 are in most cases coordinately expressed in a variety of different tissues and confirm the Northern blot data. In addition, multiple PCR products were amplified using the IpLITR1 and IpLITR2 specific primers, which is reminiscent of the multiple transcripts observed following RNA blotting. Because the predicted sizes of PCR products for the prototype IpLITR1 and IpLITR2 are 1,461 and 780 bp, respectively, at least two additional IpLITR1 products (1,545 and 2,392 bp) and one additional IpLITR2 product (1,298 bp) were amplified (Fig. 3c). Fig. 3Northern blot and RT-PCR analyses of IpLITR tissue expression. Catfish LITR tissue expression was performed by Northern blot analysis using a an IpLITR1 D1-specific probe, and b an IpLITR2 D3-specific probe. Total RNA from pronephros (head kidney), mesonephros (trunk kidney), spleen, heart, liver, gill, and muscle were examined. Kilobase markers are on the left margin and arrows indicate the major hybridizing bands observed. RNA integrity and load levels were determined by hybridization using a catfish EF1α probe as a housekeeping gene. c RT-PCR analyses of IpLITR1 and IpLITR2 in various catfish tissues. The sizes of the IpLITR bands verified by sequencing are indicated at the right margin and base pair sizes are at the left margin To further examine the expression of IpLITR1 and IpLITR2, RT-PCR was performed using RNA from catfish clonal macrophage, B, T, and NK-like cell lines, as well as a MLC (Fig. 4a). RT-PCR products of 1,461, 1,545, and 2,392 bp were obtained using IpLITR1-specific primers, and these products were detected variably among the cell lines examined. It was also found that three RT-PCR products were also variably amplified from the different catfish leukocyte cell lines using IpLITR2-specific primers, the expected 780-bp product, a 1,298-bp product (previously observed in tissues; see Fig. 3c), and an additional 1,060-bp product (Fig. 4a). In contrast, no IpLITR message was detected in the fibroblast CCO cell line (data not shown), indicating that these genes are primarily expressed in cells of immune function. Fig. 4a RT-PCR analyses of IpLITR1 and IpLITR2 expression in various catfish clonal cell lines and a polyclonal MLC. Total RNA was obtained from the catfish 42TA macrophages, 3B11 B cells, TS32.15 and TS32.17 nonautonomous cytotoxic T cells, autonomous G14D T cells, a MLC, and 1F3 NK-like cells. RT-PCR was performed using primers specific for IpLITR1, IpLITR2, and IpEF1α. The sizes of the IpLITR bands verified by sequencing are indicated at the right margin and base pair sizes are at the left margin. b Schematic representation of IpLITR-types identified by sequencing of RT-PCR products obtained using IpLITR1 and IpLITR2-specific primers. The originally identified prototype IpLITRs are boxed, and arrows indicate the relative positions of the primer pairs used in RT-PCR reactions. Sizes in base pairs corresponding to the bands observed in a are indicated above each schematic. The predicted SP, Ig domains, TM, and CYT are indicated. ITIM-like motifs are shown as hatched boxes and noncanonical immunotyrosine-based activation motif-like motifs shown as white boxes within the CYT. Individual Ig domains are shaded according to relatedness as described for Fig. 1, and percentages above each receptor represent percentage of amino acid identities of the predicted extracellular region of the IpLITR-like sequences when compared with the prototype IpLITRs The sequencing of the various IpLITR1 and IpLITR2 RT-PCR products from all sources (both tissues and cell lines) resulted in the identification of multiple different IpLITR-related sequences (Stafford et al. manuscript in preparation). Schematic diagrams of representative IpLITR1 and IpLITR2 variant forms are shown in Fig. 4b. These highly related IpLITR1- and IpLITR2-like sequences differed in the number of Ig-like domains present, and in the case of some IpLITR1-variants, also differed in the length of the CYT. Homology between IpLITRs, LRC, and FcR genes When the extracellular region of IpLITR1 was used in BLASTP analysis, various adhesion molecules, FcRs, FcR-like proteins (FCRLs), LILRs, and KIR gene family members were identified with significant homology (i.e., E values <10−7), albeit with moderate amino acid identities ranging from ∼25–30% (Table 3). Examples include the adhesion molecules CD22 and carcinoembryonic antigen-related cell adhesion molecule 5, Xenopus leukocyte FCRLs, XFL, PIRA1, LILRA2, CHIRA2, KIR3DL1, FCRL5, and FcγRIIB. When the Ig domains of IpLITR2 were used for screening, most matches were members of the FcR gene family, e.g., FcγRI, FcγRII, FcγRIII, and FcεRIα, from humans, primates, and rodents, as well as various FCRL proteins (Table 3). Searches using the extracellular region of IpLITR3 were essentially the same as the IpLITR1 BLASTP results, and PSI-BLAST analysis further reinforced the evolutionary relationships between IpLITRs and mammalian FcRs, FCRLs, KIRs, and LILRs, as well as avian CHIRs (data not shown). Table 3Representative IpLITR BLASTP resultsReceptor AccessionScoreE-valueChromosomeIpLITR1CD22HumanNP_001762884e-1619q13.1XFL1.4XenopusAAQ56585845e-15?CEACAM5HumanCAA34474845e-1519q13.1PIRA1RatAAD50905731e-11?LILRA2 (ILT1)ChimpanzeeNP_001009044722e-1119FCRL5 (FcRH5)MouseNP_899045716e-113XFL1.3XenopusAAQ63874716e-11?LILRA2 (ILT1)HumanAAD50364701e-1019q13.4KIR3DL1HumanAAC83928693e-1019q13.4KIRChimpanzeeAAF86243684e-1019CHIRA2ChickenCAG33731661e-9?KIR3DL1OrangutanAAM78465661e-919FCRL5 (FcRH5)HumanCAH71428662e-91q21KIR3DS1HumanAAV32446662e-919q13.4FcγRIIBRatAAL29888631e-813q24XFL1.6XenopusAAQ63873631e-8?XFL1.2XenopusAAQ56587622e-8?CHIRB3ChickenCAG25957623e-8?LILRA3 (ILT6)HumanAAB87661608e-819q13.4LIRBovineXP_586587601e-7?PIRA2MouseAAC53213592e-77IpLITR2IRTA2BovineXP_595289761e-12?FcγRIIGuinea pigA34636613e-8?FCRL5 (FcRH5)MouseNP_899045606e-83FCRL3 (FcRH3)MouseAAS91578606e-83FcγRIMacaqueAAL92095591e-7?FCRL5 (FcRH5)HumanNP_112571574e-71q21FcγRIαHumanCAI12557575e-71q21FcγRIIBMouseAAH19758569e-71XFL1.4XenopusAAQ56585569e-7?IRTA1BovineXP_614632552e-6?FcεRIαHumanAAH05912538e-61q23FREB2HumanAAX11390538e-61q23.3FcγRIHumanAAA58414538e-61q21.2Similarity search (BLASTP) of the NCBI’s nonredundant protein database using the four Ig domains of IpLITR1 and the three Ig domains of IpLITR2CEACAM5 carcinoembryonic antigen-related cell adhesion molecule, XFLXenopus leukocyte FCRL Phylogenetic analyses and alignments The IpLITR D1 and D2 Ig domains are related to those of mammalian FcRs and FCRLs (Fig. 5a) in that these domains clustered with the FcR/FCRLs and away from catfish NITRs and from representative LRC members in a phylogenetic tree. In contrast to the D1/D2 sequence analyses, the membrane proximal D3 and D4 sequences of IpLITR1 and IpLITR3 resulted in a significant shift within the phylogenetic tree; IpLITRs now clustered with the LRC-encoded genes (Fig. 5b). The LRC relationship is supported by high bootstrap values when comparing IpLITR1 and IpLITR3 D3 and D4 sequences with D2 and D3 sequences of representative LILRs and KIRs, as well as with D1 and D2 sequences of NKp46 and CHIRs. The D5 and D6 sequences of IpLITR3 are related to the D1 and D2 domains of the FcR family (Fig. 5c). Thus, IpLITR3 is the best example of the unique domain structure of these teleost receptors, as it encodes for combinations of Ig domains that are phylogenetically related to the FcRs and FCRLs (D1 and D2; D5 and D6), as well as LRC encoded receptors (D3 and D4). Fig. 5Phylogenetic analyses reveal unique Ig domain composition within individual IpLITRs. The Ig domains compared are indicated by gray shading in the schematics to the left of each phylogenetic tree. a Comparisons of IpLITR, representative FcR and FCRL, and human LRC D1 and D2 sequences. b Comparisons of IpLITR1 D3 and D4 with representative FcR, FCRL, KIR, LILR, and PIRA1 D2 and D3 sequences, as well as NKp46 and CHIRA2 D1 and D2 sequences. c Comparison of IpLITR3 D5 and D6 with D1 and D2 of representative mammalian FcRs, FCRLs, and LILRs/KIRs. The accession numbers for the various Ig domain sequences used are: human (hu)FcγRI(CAI12557), huFcγRII (CAA35642), huFcγRIII (CAA36870), huFcεRI (AAH05912), huFCRL3 (AAH28933), huFCRL4 (AAK93970), huFCRL5 (NP_112571), huFcαR/CD89 (AAH27953), huLILRB1 (AAH15731), huLILRB2 (AAB87662), huKIR3DL1 (AAC83928), huKIR3DS1 (AAV32446), huSIRP (CAA71404), human NKp46 (AJ001383), mouse (mo)FcγRI (AAD34931), moFcγRII (AAA37608), moFcγRIII (NP_034318), moFcεRI (NP_034314), moFCRL3 (AAS91578), moNKp46 (AJ223765), rat (ra) FcγRIII (AAA42050), raPIRA1 (XP_341773), raNKp46 (AF082533), guinea pig (gp) FcγRII (D13692), macaque (mac)FcγRI (AAL92095), bovine (bo)FCRL5 (XP_595289), boKIR3DS1 (AAP33626), boSIRP (CAA71943) chimpanzee (cp)LILRA2 (NP_001009044), orangutan (or)KIR (AAM78473), monkey (mk)KIR3DL20 (AAU50562), CHIRA2 (CAG33731), catfish (Ip)NITR1 (AF397454), and IpNITR2 (AF397455). For all analyses, SIRPs and/or IpNITRs were included as outgroups to root the trees. NJ trees with pairwise gap deletions were drawn using MEGA v3.0 (Kumar et al. 2001) with 10,000 bootstrap replications, and bootstrap values >50% are shown. Branch lengths were measured in terms of amino acid substitutions, and scale bars are shown below the trees Subsequently, the D1 and D2 Ig domain sequences of IpLITR2 (as a representative) were aligned with various mouse and human FcRs and three representative human FCRLs. A majority of the predicted β-strands in IpLITR2 is present in similar positions as the β-strands found in FcR/FCRL family members (Fig. 6a), indicating a predicted degree of secondary structural conservation between IpLITRs and these mammalian immune receptors. In most cases, the predicted IpLITR2 β-strands were present in regions of high amino acid identity/similarity; however, none of the FcR residues known to contact the Fc portion of Ig were present (Fig. 6a). The D3 of IpLITR2, a domain specific to this IpLITR type, appears to be related to the membrane distal domain (D1) of several CHIRs (Fig. 6b), and alignment of IpLITR2 D3 with representative CHIR D1s indicated a degree of amino acid identity/conservation between these Ig domains (Fig. 6c). To reinforce the unique nature of domains present within this receptor, IpLITR2 D1 did not cluster with the CHIRs but was related to mammalian FcR D1 domains (Fig. 6b). Fig. 6IpLITR2 contains Ig domains related to both FcRs/FCRLs and CHIRs. a Amino acid alignment of IpLITR2 D1 and D2 sequences with representative mammalian FcRs and FCRLs. Accession numbers are as in Fig. 5. Gray shading: residues similar/identical to IpLITR2. Boxed residues: contacts for Ig Fc in huFcεRI. Black and gray arrows represent the predicted β-strands for IpLITR2 and huFcεRI, respectively. Hatched boxes indicate conserved cyteines and dashes represent gaps. b Phylogenetic analysis of IpLITR2 D1 and D3 sequences (gray shaded) compared with representative mammalian FcR sequences (as in Fig. 5) and D1 sequences of representative CHIRs. The accession numbers for the various Ig domain sequences used are: CHIRB2 (XP_422905), CHIRB1 (CAH55757), CHIRAB2 (CAG33733), CHIRA2 (CAG33731), CHIRB5 (AJ879908), CHIR A1 (AF306851), CHIRB4 (XP_428342), CHIRB6 (CAI53861), and CHIRAB3 (AJ879909). NJ trees with pairwise gap deletions were drawn using MEGA v3.0 (Kumar et al. 2001) with 10,000 bootstrap replications, and bootstrap values >50% are shown. Branch lengths were measured in terms of amino acid substitutions, and scale bars are shown below the trees. c Alignment of IpLITR2 D3 and CHIR D1. Gray shading indicates residues that are similar/identical to IpLITR2, black and gray arrows represent predicted β-strands for IpLITR2 and CHIRA2, respectively; “dotted” lines indicate predicted helices and hatched boxes indicate conserved cyteines. Gaps in alignment are indicated by dashes Attempts to model IpLITR Ig domains were unsuccessful as no sufficient modeling templates were identified using SWISS-Model. However, similarity searches against the NCBI’s protein data bank (PDB) indicated a significant match between D1 and D2 of IpLITRs and the Ig domains of human FcεRIα (E value <10−8). Conversely, the membrane proximal D3 and D4 Ig domains of IpLITR1 and IpLITR3 had significant alignment scores (E value <10−7) with the first two Ig domains of the human natural cytotoxicity receptor NKp46. Taken together, these PDB search results further support that IpLITRs D1 and D2 Ig domain sequences are related to mammalian FcRs and FCRLs, and that the D3 and D4 Ig domains are related to LRC encoded receptors. Mining of the zebrafish genome identified five contigs (Table 2) containing sequences encoding IpLITR-like proteins predicted to contain several Ig domains and, in some cases, signal peptides, TM segments, and CYTs with signaling motifs (Fig. 7). These findings indicate that IpLITR-like molecules are not unique to the catfish, but are likely common in other teleost species. Because IpLITR3 was identified as the closest match from the NCBI’s database to the zebrafish (Zf) LITR-like proteins, it was used for further comparisons. Phylogenetic analysis of the various Ig domain sequences indicated that the ZfLITR-like proteins were closely related to those of IpLITRs (Fig. 7a). In most cases, the membrane proximal domains of ZfLITRs are highly related to their IpLITR3 counterparts with >60% amino acid identity (Fig. 7b). However, one of the ZfLITRs (XP_692267) contains two Ig domains similar to IpLITR3 D6 in the membrane distal region. This zebrafish receptor also contains three Ig domains that are phylogentically related (Fig. 7a) with high amino acid identity (∼50%) to IpLITR3 D4 (Fig. 7b). The other ZfLITR-like proteins are also composed of combinations of Ig domains that are related to IpLITR3 Ig domains with varying degrees of amino acid identities (e.g., 25–66%). However, none of the predicted ZfLITR-like molecules contained membrane distal Ig domains that were related to IpLITR D1 and D2, which are shared among all the IpLITRs identified to date. This suggests either that LITR membrane distal domains may have coevolved with distinct ligands within individual fish species or that ZfLITRs with D1 and D2 domains similar to IpLITRs have yet to be identified. It should be noted, however, that these zebrafish domains were obtained from genomic sequence and that they do not accurately represent the precise organization of the Ig domains likely to be present in encoded ZfLITRs, which will only be determined from cDNA sequences. Nonetheless, the genomic segments are useful for domain-to-domain comparisons and reinforced the presence of IpLITR-like sequences in another teleost species. Fig. 7Annotation of zebrafish LITR-like proteins. a Phylogenetic analysis of ZfLITR and IpLITR3 Ig domains. Ig domains were predicted using SMART (Letunic et al. 2004) and the amino acid sequences aligned using CLUSTALW (Thompson et al. 1997). NJ trees with pairwise gap deletions were drawn using MEGA v3.0 (Kumar et al. 2001). Only branch values >50 are shown, and the major Ig domain groups are designated by brackets and shaded according to their relatedness with IpLTR3 Ig domain. b Schematic representation of ZfLITR-like proteins compared with IpLITR3. Domains are shaded according to their phylogenetic relationship with IpLITR3 domains. Hatched bars on IpLITR3 and Zf XP_692267 indicate signal peptides, TM segments are solid black lines, and ITIMs are shown as gray boxes for XP_694305. Amino acid identity vs corresponding IpLITR3 Ig domains are indicated as percentage value above each domain, and only IpLITR3 D3, D4, D5, and D6 were related to the Ig domains of the representative ZfLITR analyzed, as indicated by the dashed box. Predicted size in amino acids for each molecule is indicated on the left Discussion The novel IpLITRs described herein represent a polymorphic and polygenic gene complex that are primarily expressed in hematopoietic tissues. The polymorphic nature of this gene complex was indicated by RFLP analysis, which revealed differences in where various restriction enzymes cut the genomic DNA among the sibling catfish. However, it should be noted that only detailed sequence analyses of these IpLITR genes in a number of individuals would be required to determine if any polymorphic differences are also found within the coding regions. In addition, segregation analyses suggested that IpLITR genes are encoded within multiple independently segregating but homologous loci, similar to the situation seen with the ZfLITR homologs identified by genome mining. These paralagous relationships strongly suggest that IpLITRs are products of multiple gene duplication and translocation events from a common ancestoral gene. Whether or not the receptors encoded by the different IpLITR loci are functionally redundant remains to be determined. However, IpLITRs are related to functionally and genomically distinct mammalian FcRs and LRC encoded receptors, suggesting that such paralogous loci could have evolved to give rise to an array of immunoregulatory receptors with different functions. The coordinate expression of IpLITRs by myeloid and lymphoid cell lines in combination with their potential activating and inhibitory signaling capabilities suggests that they play an important role in immune cell function(s). This possible importance of IpLITRs in immune responses is highlighted by the RT-PCR expression studies where a large number of highly related, but unique, IpLITR sequences were identified. These variant IpLITR sequences differed in the number of Ig domains present, and in some cases in the length of the CYT. Therefore, in addition to the prototype IpLITR1, IpLITR2, and IpLITR3 described here, other IpLITR family members are coexpressed by catfish myeloid and lymphoid cells. Whether or not some of the variously sized transcripts represent messages from different IpLITR genes or are the result of alternative splicing is unknown. However, the large number of unique IpLITRs found expressed suggests that IpLITRs represent a complex immune receptor family that can potentially generate a large number of IpLITR types. The prototypical IpLITR1, IpLITR2, and IpLITR3 encode receptors with four, three, and six Ig domains, respectively. Pairwise comparisons and phylogenetic analysis of these domains suggest a high degree of conservation and a conserved membrane distal to membrane proximal ordering. Each of these three receptors has very similar D1 and D2 sequences, and IpLITR1 and IpLITR3 have almost identical D3 and D4s. However, variation does occur between predicted membrane proximal Ig domains. The IpLITR2 D3 sequence is unique to this receptor and IpLITR3 D5 and D6 are not similar to any other of the prototypical IpLITR Ig domain sequences. Identification of LITR-like sequences on different zebrafish chromosomes and comparisons with IpLITRs supports the notion that these polymorphic and polygenic immune receptors are not unique to catfish. Like IpLITRs, the predicted ZfLITRs are also composed of unique combinations of Ig domains. Such differences in the Ig domain compositions of IpLITRs and ZfLITRs may influence their ligand binding capabilities, formation of homo- or heterodimers and overall immunomodulatory functions. IpLITR1 has a long CYT with two consensus ITIMs and a third ITIM-like motif, whereas both IpLITR2 and IpLITR3 have shorter CYTs devoid of recognizable signaling motifs. However, these latter two receptors each contain a lysine residue within their TM segments, indicating that associations with an adaptor molecule(s) would be necessary for them to initiate intracellular signaling (Daeron and Vivier 1999; Ravetch and Lanier 2000; Billadeau and Leibson 2002). Within the CYT of IpLITRI was also an overlapping ITSM suggesting that multiple motifs (ITIMs and ITSM) within the CYT of this receptor may participate in downstream signaling cascades through recruitment of various adaptor molecules (Sidorenko and Clark 2003). Similarly, predicted ZfLITR family members were shown to either contain long CYTs with ITIM motifs or short CYTs with charged TM segments. Genes for immunotyrosine-based-activation-motif-bearing adaptor molecules (e.g., FcRγ chain, CD3ζ, and DAP) have been identified in several ectothermic vertebrates, including frogs and fish (Guselnikov et al. 2003a,b). Hence, it seems likely that IpLITR2 and IpLITR3 would each associate with such molecules and function in a similar manner as mammalian receptors with charged TM segments. Thus, the paired expression pattern of IpLITRs with activatory and inhibitory signaling potential is very similar to that shown for several mammalian and avian families of immune receptors, including PIRs, LILRs, KIRs, FcRs, and CHIRs. Similarity searches for IpLITR relatives revealed a striking dichotomy. The majority of matches were members of two distinct mammalian receptor families. While IpLITR2 appeared to be more closely related to FcRs and FCRLs, IpLITR1 and IpLITR3, exhibiting very similar D1–D4 sequences, were related to both FcR/FCRLs and LRC-encoded receptors. Thus, it is postulated that the D1 and D2 domains common to all IpLITRs are evolutionarily more closely related to domains encoded by the mammalian FcR and FCRLs, a contention supported by the retrieval of the human FcεRI alpha chain as the most similar structure from a PDB search. Phylogenetic analyses also supported this hypothesis because the IpLITR D1 and D2 clustered with members of the FcR and FCRL receptor families and away from representative LRC gene family members. Although the D1D2 of IpLITRs are phylogenetically related to FcR and FcRLs, there is no evidence to support that IpLITRs bind Ig. It appears more likely that the relatedness of these membrane distal Ig domains, as also demonstrated for FCRLs, PIRs, and CHIRs, are only evident at the phylogenetic level and that they are not functional homologs. Furthermore, during the preparation of this manuscript, a “bona fide” Ig-binding catfish FcR homolog, which is not a member of the IpLITR receptor family, was identified and characterized (Stafford et al. 2006). In contrast to the D1 and D2 domains, IpLITR1 and IpLITR3 D3 and D4 Ig domains are related to several members of the LRC (human chromosome 19q13.4). As above, these two domains clustered with one family of receptors (the LRC) and not the other (i.e., FcR/FcRLs). The identification of NKp46 as a structural relative from the PDB also supports this notion. Interestingly, D5 and D6 of IpLITR3 are related to FcR and FCRLs, but their membrane proximal location suggests that they are unlikely to be involved in ligand recognition (i.e., Ig). Finally, the third domain of IpLITR2 clustered with the Ig domains of several representative CHIRs. Taken together, IpLITR extracellular domain sequences appear reminiscent of both FcR/FCRLs and members of the mammalian and avian LRC, implying that catfish LITRs are distant relatives to each of these genomically and functionally diverse receptor gene families. At present, the functional significance of these novel teleost immune receptors is unknown. Recently, key insights into the evolution of the FcR and LRC gene families and their potential common origin has been presented: (1) human chromosomes 1 and 19 are believed to be in paralogous regions of the genome due to an entire genome duplication event (Kasahara 1999; Shiina et al. 2001); (2) attempts at cloning rodent FcαR (CD89) orthologs identified the PIR gene family located on a mouse chromosome, which is syntenic to the human LRC (Kubagawa et al. 1997; Hayami et al. 1997); (3) the subsequent search for distant PIR relatives resulted in the discovery of the CHIRs, providing the first molecular evidence for a common ancestor of LRC and FcR gene families (Dennis et al. 2000); (4) sequencing of genomic segments encoding the CHIR gene family identified the chicken equivalent of an LRC that appears to have also been expanded by gene duplication events (Nikolaidis et al. 2005a); and (5) the identification of Ig-like domains in chickens, Xenopus, and bony fish that are related to Ig domains of mammalian FcRs and members of the LRC (Nikolaidis et al. 2005b). Taken together, these findings suggest that immune receptors encoded by human chromosome 1q21 and 19q13 most likely evolved from a common ancestor that has since undergone multiple rounds of duplication and homologous recombination events. Although there is functional evidence for the existence of FcR- or KIR/LILR-like homologs in ectothermic vertebrates (Coosemans and Hadji-Azimi 1986; Shen et al. 2003, 2004), their genetic identification has not been previously reported and the inability to clone these receptors has made it difficult to elucidate their evolutionary history. The recent discovery of mammalian FCRL genes located on chromosome 1q21-22, however, has expanded the view of immune receptor evolution (Davis et al. 2001; Hatzivassiliou et al. 2001; Miller et al. 2002; Guselnikov et al. 2002). For example, FCRL3, FCRL4, and FCRL5 not only encode D1, D2, and D3 Ig domains homologous to the corresponding domains of FcγRI, but also contain variable numbers of unique membrane proximal Ig domains (Davis et al. 2001; Hatzivassiliou et al. 2001). Homology searches demonstrated that the FCRLs, also termed IFGP (for IgSF, FcR, and gp42) exhibited a “chimeric” Ig domain compositions and consequently these receptors were proposed to be a phylogenetic link between the FcRs, rat NK cell-specific gp42 antigen, and adhesion molecules (Guselnikov et al. 2002). This feature of similar Ig domains encountered in homologous but functionally diverse receptors is also a striking feature of the IpLITRs and supports the hypothesis that certain mammalian immune receptor families (i.e., FcRs and KIRs) have evolved from a common ancestor. Notably, although the IpLITRs D1 and D2 domains are phylogenetically and structurally similar to the mammalian FcRs and FCRLs, the IpLITR ligand(s) is unknown. This is similar to the situation with mammalian FCRL molecules, whose membrane distal Ig domains are clearly related to the classic FcRs, but are unlikely to be functional homologs because they do not bind Ig (Hatzivassiliou et al. 2001; Davis et al. 2005). Therefore, although certain Ig domains encoded by several receptor gene families identified in humans, rodents, birds, and fish (i.e., FCRLs, PIRs, CHIRs, and IpLITRs, respectively) are related to FcRs, they appear to have evolved to recognize different ligands. Recently, it was shown that rodent PIRs recognized MHC class 1 and β2M (Takai 2005), raising the possibility that related immunoregulatory receptors in other species (i.e., CHIRs and IpLITRs) could also be involved in MHC recognition. In summary, the discovery of individual catfish receptors that encode Ig domains related to both FcR and LRC gene families provides further support for the common origin of these important mammalian immune receptors. Identification of LITR-like sequences in the zebrafish genome also demonstrates that these receptors are likely found in all teleosts. Annotation of these genomic segments will be necessary to determine if zebrafish LITR homologs exhibit conserved genomic features and syntenic genomic location with mammalian 1q21-23 and/or 19q13. Thus, the discovery of IpLITRs is an important step towards further understanding the common origin and evolutionary history of the different families of mammalian immunoregulatory receptors. The investigation of the functional significance of these novel teleost receptors is underway and will be aided by the availability of channel catfish cell lines that coordinately express IpLITRs and the eventual production of IpLITR specific antibodies. Electronic supplementary material Below is the link to the electronic supplementary material. Electronic Supplementary Figure 1 52 kb	[ "teleosts", "innate immunity", "evolution", "molecular immunology", "inhibitory receptors" ]	[ "P", "P", "P", "M", "R" ]
Hum_Reprod-1-1-2387217	Predicting the FSH threshold dose in women with WHO Group II anovulatory infertility failing to ovulate or conceive on clomiphene citrate	BACKGROUND The objective of this investigation was to establish independent predictors of follicle-stimulating hormone (FSH) threshold dose in anovulatory women undergoing ovulation induction with FSH preparations. Introduction Approximately 20–30% of the women seeking fertility treatment present with anovulation (Healy et al., 1994). Clomiphene citrate has been used as the first line treatment for inducing ovulation in anovulatory women. In most women who failed to ovulate or conceive with clomiphene citrate, ovarian stimulation with gonadotrophins leads to ovulation (Coelingh-Benning et al., 1998; Yarali et al., 1999; Platteau et al., 2006; Balen et al., 2007). The use of low-dose step-up protocols has been shown to be successful in producing mono-follicular development in more than half of the women undergoing stimulation with gonadotrophins, and thereby minimizing the risk of multiple follicular development and the associated risks of ovarian hyperstimulation syndrome (OHSS) and multiple pregnancy (Homburg and Howles, 1999). A typical low-dose step-up protocol starts with 50–75 IU/day of follicle-stimulating hormone (FSH) or menotrophins for ∼7 to 14 days, and the dose is then increased by 37.5–50 IU/day if adequate response is not obtained. For some patients, the dose of gonadotrophin needed to provide adequate response is relatively high and the use of low-dose regimens will in these cases lead to stimulations of long duration. A long ovarian stimulation is associated with more monitoring visits and would be expected to result in increased costs and inconvenience for the patients. Selection of the most optimal starting dose of gonadotrophin for each individual patient should minimize these disadvantages, while ensuring adequate response. Imani et al. (1998) have identified clinical and laboratory parameters such as body mass index (BMI), menstrual cycle history, mean ovarian volume and free androgen index (FAI) which can predict ovulation in WHO type II anovulatory women undergoing clomiphene citrate treatment. A similar investigation in anovulatory women undergoing ovulation induction with gonadotrophins suggested BMI, clomiphene citrate resistance, serum insulin-like growth factor-1 (IGF-1) and serum FSH as significant predictors of FSH threshold dose (Imani et al., 2002). The clinical application of this latter result is complicated by the fact that the IGF-1 assay is not part of most clinics standard battery for laboratory screening of anovulatory patients. Therefore, further data are needed to identify simple parameters which can predict the FSH threshold dose in women who will undergo low-dose protocols with gonadotrophins. The purpose of the present prospective investigation was to evaluate the influence on the FSH threshold dose of clinical, sonographic and endocrine parameters available at the start of ovarian stimulation as well as type of FSH preparation to be used in a low-dose step-up protocol. The data from this investigation form the basis for constructing a nomogram which provides an indication of the expected FSH threshold dose for the individual patient, facilitating the selection of the most appropriate starting FSH dose for ovarian stimulation. Materials and Methods Study population and design This investigation included 151 women with anovulatory WHO Group II infertility who were included in a randomized, open label, assessor-blind, parallel group, multicentre, multinational, non-inferiority study comparing two different FSH preparations (highly purified (HP)-FSH, BRAVELLE, Ferring Pharmaceuticals A/S, Copenhagen, Denmark and recombinant FSH (rFSH), follitropin alfa, GONAL-F, Serono, Geneva, Switzerland) with respect to ovulation rate. Patients were recruited in 22 fertility centres (12 in Belgium, 7 in Denmark and 3 in the UK) that on average each included seven patients, ranging from 2 to 20. A detailed description of the study population, all inclusion and exclusion criteria, all study procedures, as well as the main outcome of the study, is described in detail in a separate publication (Balen et al., 2007). Patients were to be 18–39 years of age with BMI 19–35 kg/m2, have been diagnosed with infertility for ≥1 year, have chronic anovulation (amenorrhea, oligomenorrhea or anovulatory cycles based on progesterone levels in patients with cycle lengths of 21–35 days) and have failed to ovulate with clomiphene citrate doses of at least 100 mg/day for at least 5 days or failed to conceive after three cycles of ovulation induction with clomiphene citrate. Women with a history of ≥12 unsuccessful ovulation induction cycles, persistent ovarian cysts (≥15 mm) for >1 cycle or ovarian endometrioma on ultrasound and any significant systemic disease, endocrine or metabolic abnormalities were excluded. The study was carried out in accordance with the declaration of Helsinki on good clinical practice and ethical committee approval was obtained in all participating centres. All patients underwent screening assessments prior to start of ovarian stimulation. These included evaluation of demographics (age), physical characteristics (body weight, height, waist and hip circumference), obstetric history (previous pregnancies, previous live births), infertility history (duration of infertility, number of previous ovulation induction cycles with clomiphene citrate and in total, failure to ovulate with clomiphene citrate, failure to conceive with clomiphene citrate) and menstrual cycle history. Just prior to the first dose of gonadotrophin, patients underwent a transvaginal ultrasound and blood sampling for endocrine evaluation. The sonographic recordings comprised dimensions of right and left ovary, number of follicles (antral follicles >2 mm) in right and left ovary and endometrial thickness. The endocrine evaluation included the following parameters: luteinizing hormone (LH), FSH, estradiol (E2), prolactin, total testosterone, sex hormone binding globulin (SHBG), FAI, glucose and insulin. Chemiluminescent immunometric assays were used for LH, FSH, prolactin, total testosterone and SHBG, and radioimmunoassays were used for E2 and androstenedione, whereas a hexokinase assay was used for glucose and immulite technology was applied for analysis of insulin. All endocrine analyses were conducted by a central laboratory (Quest Diagnostics—Quest Diagnostics Limited, Heston, UK; Quest Diagnostics Limited, Van Nuys, CA, USA; Nichols Institute, San Juan Capistrano, CA, USA). FAI was calculated as (total testosterone/SHBG) × 100. All patients underwent ovarian stimulation following a low-dose step-up protocol. Stimulation was started 2–5 days after a spontaneous or progesterone-induced menstrual bleed. The starting dose of gonadotrophin was 75 IU daily, which was maintained for 7 days. After the first 7 days, the gonadotrophin dose was either maintained or increased by 37.5 IU increments according to individual response. Patients were maintained on any specific dose level for at least 7 days. The maximum allowed daily dose was 225 IU and patients were treated with gonadotrophin for a maximum of 6 weeks. Gonadotrophin stimulation was maintained until at least one of the following criteria for human chorionic gonadotrophin (hCG) administration was met: one follicle with a diameter of 17 mm or greater, or 2–3 follicles with a diameter of 15 mm or greater. Patients were not given hCG in either of the following situations: no follicular response after 6 weeks of gonadotrophin treatment, ≥4 follicles with diameters of ≥15 mm or serum E2 levels >2000 pg/ml. Patients who reached the hCG criteria received a single s.c. or i.m. injection of hCG (PROFASI, Serono) at a dose of 5000 IU to trigger ovulation. At least one blood sample was taken during the midluteal phase (6–9 days after hCG administration) and analysed for progesterone by a central laboratory using competitive immunoassay using direct chemiluminometric technology with a sensitivity of 0.48 nmol/l (Quest Diagnostics Limited, UK). Ovulation was defined as a mid-luteal serum progesterone concentration of ≥25 nmol/l (≥7.9 ng/ml). The FSH threshold dose was defined as the dose of FSH that was given when the criteria for hCG was reached. The analysis involved the influence of the type of gonadotrophin preparation (i.e. HP-FSH or rFSH) as well as of clinical, sonographic and basal endocrine parameters on the FSH threshold dose. All potential predictive variables were investigated: age, BMI, waist–hip ratio, previous pregnancy, duration of infertility, failure to ovulate on clomiphene citrate, menstrual cycle history (amenorrhea if the patient had absent menstruation, oligomenorrhea if menstruation intervals were longer than 35 days or cycle length of 21–35 days in which case anovulation had been documented by progesterone levels), total number of follicles, mean ovarian volume (based on length, width and depth of both right and left ovary), LH/FSH ratio, androstenedione, total testosterone, SHBG, FAI, prolactin, estradiol, glucose, insulin and insulin/glucose ratio. Statistical analysis The primary end-point of the study was ovulation rate after one cycle of gonadotrophin treatment, and these data have been reported elsewhere (Balen et al., 2007). The influence of the selected clinical, sonographic and endocrine parameters as well as type of gonadotrophin on FSH threshold dose were assessed in univariate and multivariate analyses using proportional odds polytomous logistic regression models (an extension of the model and method for binary response). Multivariate analyses were performed including characteristics that had statistically significant influence on the FSH threshold dose in the univariate analysis first by entering all selected characteristics into the polytomous logistic model in a forward selection approach and secondly by entering all selected characteristics into the polytomous logistic model using a backward elimination approach. P-values of <0.05 were regarded as statistically significant. Score tests were used to evaluate the proportional odds assumption and Wald tests were used to assess the influence of the baseline characteristics on the FSH threshold dose. For the present analysis, all parameters were analysed as continuous variables, except for type of gonadotrophin, previous pregnancy, failure to ovulate on clomiphene citrate and menstrual cycle history (glucose and insulin were categorical variables as normal/abnormal and insulin/glucose ratio was a continuous variable). In addition to the above, the model was evaluated on the subset of patients with mono-follicular development (one follicle ≥17 mm and no follicles of 15–16 mm). In this evaluation, the same set of variables was used as selected for the main model according to the above procedure. The predictive performance of the model was assessed by using the ‘leave-one-out’ cross-validation procedure. In addition, the residual standard error after a standard multiple regression analysis was evaluated. Finally, ROC curves were produced and AUC under the ROC curves were calculated. For the standard situation where ROC curves are used, the classification is binary, whereas the threshold dose had four possible outcomes (87.5, 112.5, 150 and 187.5) and the prediction model therefore also predicts into four classes. A consensus approach on how to handle multiclass ROC has not been reached (Landgrebe and Duin, 2007), but when the outcome classes are ranked, ROC curves can be made for each of the possible levels (Waegeman et al., 2006). Results The clinical, sonographic and endocrine characteristics of the 151 patients included in the study are shown in Table I. The mean age of the study population was 28.9 years, ranging from 19 to 39 years. The BMI spanned from 18 to 39 kg/m2, with 54% of the patients having a normal BMI, 32% being overweight and 14% being obese and with the mean BMI being 24.8 kg/m2. On average, patients had been infertile for ∼3 years and had undergone four previous ovulation induction cycles with clomiphene citrate, with 34% of the patients failing to ovulate on clomiphene citrate. Regarding menstrual cycle history, 56% of the patients were oligomenorrheic and 15% amenorrheic. An LH/FSH ratio above 1 was noted for 59% of the patients. Among the 151 patients who started stimulation with FSH (HP-FSH 73, rFSH 78), 132 patients met the hCG criteria (HP-FSH 65, rFSH 67) and contributed with data to the analysis of FSH threshold dose. There were 19 patients who initiated gonadotrophin therapy but did not contribute to this analysis; 10 discontinued from the study during ovarian stimulation (five because of excessive response), 2 had spontaneous ovulation and 7 received hCG despite either not meeting or exceeding the hCG criteria. Table I. Clinical, sonographic and endocrine parameters just prior to starting ovarian stimulation with FSH (Day 1 of stimulation). Total (n = 151) Clinical parameters Age (years) 28.9 ± 3.7 Body weight (kg) 68.6 ± 13.2 BMI (kg/m2) 24.8 ± 4.5 Primary infertility, n (%) 97 (64%) Duration of infertility (years) 2.8 ± 1.7 Previous cycles of ovulation induction (all) 4.9 ± 2.4 Previous cycles of ovulation induction (with clomiphene citrate) 4.1 ± 2.4 Clomiphene citrate non-responders, n (%) Failure to ovulate on clomiphene citratea 51 (34%) Failure to conceive on clomiphene citrateb 100 (66%) Menstrual cycle history, n (%) Amenorrhea 23 (15%) Oligomenorrhea 84 (56%) Anovulatory cycles with cycle length 21–35 days 44 (29%) Sonographic parameters Mean ovarian volume (cm3) 7.1 ± 3.5 Antral follicles >2 mm 14.8 ± 12.5 Endometrial thickness (mm) 4.1 ± 2.2 Endocrine parameters LH (IU/l) 7.1 ± 4.6 FSH (IU/l) 5.5 ± 2.4 LH/FSH ratio 1.4 ± 0.8 Prolactin (μg/l) 16 ± 13 Androstenedione (nmol/l) 6.74 ± 3.08 Total testosterone (nmol/L) 1.8 ± 0.6 Sex hormone binding globulin (nmol/l) 61 ± 40 Free androgen index 4.45 ± 3.47 Estradiol (pmol/l) 172 ± 98 Glucose (mmol/l) Fasting 5.1 ± 0.6 Non-fasting 5.2 ± 1.5 Insulin (pmol/l) Fasting 66.4 ± 43.1 Non-fasting 143.9 ± 192.6 All data are mean ± SD or number of patients (percentage of patients). aOn at least 5 days of 100 mg/day. bAfter three cycles. Table II provides the outcome of the univariate and multivariate analysis for the predictors of FSH threshold dose. Eight parameters were statistically significant predictors of FSH threshold dose (P < 0.05): age, BMI, failure to ovulate with clomiphene citrate, menstrual cycle history (amenorrhea, oligomenorrhea or anovulatory cycles with cycle length 21–35 days), mean ovarian volume, LH/FSH ratio, testosterone and FAI. The type of gonadotrophin preparation, HP-FSH or rFSH, did not predict FSH threshold dose. All eight significant parameters were included in a multivariate analysis which showed that only three of the parameters were independently statistically significant (P-values < 0.001) predictors of FSH threshold dose: BMI, mean ovarian volume and menstrual cycle history (Table III). The odds for needing >75 IU FSH were 1.18 [95% confidence interval (CI): 1.07–1.29] for BMI and 1.22 [95% CI: 1.09–1.37] for mean ovarian volume; thus the higher the BMI and ovarian volume prior to start of stimulation, the higher the FSH threshold dose. Regarding the menstrual cycle history, the odds for needing a FSH dose >75 IU were 11.9 [95% CI: 3.35–42.0] and 2.57 [95% CI: 0.97–6.79] for amenorrhea and oligomenorrhea, respectively, compared with anovulatory cycles with a length of 21–35 days. The observed data for these three independent predictors are presented in Table III. Table II. Statistically significant predictors of FSH threshold dose (univariate and multivariate analysis). Univariate analysis Multivariate analysis OR (95% CI) P-valuea OR (95% CI) P-valuea Age (years) 0.91 (0.83–1.00) 0.040 — — BMI (kg/m2) 1.10 (1.02–1.19) 0.010 1.17 (1.07–1.29) <0.001 Failure to ovulate on clomiphene citrate: yes versus no 2.27 (1.14–4.52) 0.020 — — Menstrual cycle history <0.001 <0.001 Amenorrhea versus cycle length 21–35 days 8.33 (2.81–24.7) 11.88 (3.35–42.1) Oligomenorrhea versus cycle length 21–35 days 3.48 (1.51–8.03) 2.57 (0.97–6.79) Mean ovarian volume (cm3) 1.18 (1.06–1.31) 0.002 1.22 (1.08–1.37) <0.001 Free androgen index (nmol/l) 1.25 (1.13–1.38) <0.001 — — Total testosterone (nmol/l) 2.14 (1.14–4.03) 0.018 — — LH/FSH ratio (IU/l) 1.59 (1.03–2.45) 0.038 — — aWald test for effect of factor/covariate. Table III. BMI, mean ovarian volume and menstrual cycle history by observed FSH threshold dose. BMI (kg/m2) Mean ovarian volume (cm3) Menstrual cycle history (%) Amenorrhea / oligomenorrhea / cycle length 21–35 days 75 IU (n = 67) 23.8 ± 4.2 6.3 ± 2.9 8% / 48% / 45% 112.5 IU (n = 46) 25.5 ± 4.8 7.3 ± 3.4 22% / 63% / 15% 150 IU (n = 15) 25.8 ± 4.0 9.5 ± 4.8 33% / 47% / 20% 187.5 IU (n = 4) 28.1 ± 4.6 10.2 ± 2.3 25% / 75% / 0% All data are mean ± SD or percentage of patients. The model predicted the FSH threshold dose to be the same as the dose observed in the study for 59% of the patients. The AUC under the ROC curve was 0.78 for 75 IU (Fig. 1a) and 0.79 for 112.5 IU (Fig. 1b). In a multiple regression analysis with the same three variables as in the logistic model, the FSH threshold dose was estimated on a continuous scale. From this estimation, the residual mean square error was 25 IU and the R2 = 25%. Figure 1: ROC curves for threshold dose (a) 75 IU versus <75 (AUC = 0.78) and (b) ≤112.5 IU versus >112.5 IU (AUC = 0.79). A simple nomogram can be constructed based on the two clinical and one sonographic parameters that significantly predict the FSH threshold dose. Figure 2 displays three nomograms—one for each type of menstrual cycle history—where the shaded areas indicate the predicted FSH threshold dose, based on the BMI displayed on the x-axis and mean ovarian volume on the y-axis. For example, the expected threshold FSH dose for a patient with amenorrhea (Fig. 2a), mean ovarian volume of 10.5 cm3 and a BMI of 30 kg/m2 will be 150 IU. In the case that the patient presents with oligomenorrhea (Fig. 2b), but similar mean ovarian volume and BMI, the expected threshold FSH dose would be 112.5 IU. As another example, a threshold FSH dose of 75 IU will be expected for a patient with anovulatory cycles of 21–35 days, mean ovarian volume of 8.5 cm3 and BMI 27 kg/m2, whereas it will be 112.5 IU for a patient with oligomenorrhea and similar mean ovarian volume and BMI. Figure 2: Nomograms for prediction of individual FSH threshold dose in anovulatory patients undergoing ovulation induction with FSH preparations, according to menstrual cycle history, BMI and mean ovarian volume. (a) Amenorrhea, (b) oligomenorrhea and (c) anovulatory cycles with cycle length 21–35 days. The accuracy of the model was 60% when applied to the subset of patients with mono-follicular development (80 patients), and thus similar to that achieved for the overall study population. Discussion The main finding in the present study was identification of the three variables: menstrual cycle history, BMI and mean ovarian volume, as independent predictors of the FSH threshold dose during FSH treatment of anovulatory infertility. For clinical use, such data could be implemented and used as screening predictors to determine the individual FSH threshold dose of any patient before start of gonadotrophin treatment, in order to select the most appropriate starting dose. For this purpose, we have used the three independent predictors to develop the FSH dose nomograms presented in Fig. 2. The proposed model had an accuracy of 59–60%, which is well above the 25% that would be expected by random allocation to one of the four dose levels evaluated in this study. Our findings are partly in line with those obtained by Imani et al. (2002) who studied 90 anovulatory women, also treated with either urinary or rFSH, and showed that BMI, clomiphene resistance, free IGF-1 and basal FSH were independent predictors of the FSH threshold dose. Imani et al. (2002) made an equation showing that the predicted dose was: [4 BMI (in kg/m2)] + [32 clomiphene citrate resistance (yes = 1 or no = 0] + [7 initial free IGF-I (in ng/ml)] + [6 initial serum FSH level (in IU/l)] − 51. Similar to our study, it was found that menstrual bleeding interval as well as clomiphene resistance were significant predictors in the univariate analysis, but among the possible clinical predictors only clomiphene resistance and BMI remained significant in the multivariable analysis. The population studied by Imani et al. (2002) seems very similar to the present study, even though the latter included patients if they failed to conceive after three clomiphene cycles, whereas the study by Imany et al. (2002) only included women with six failed ovulatory clomiphene citrate cycles, but it seems unlikely that this would explain the difference in predictive factors. A more plausible explanation could be that the study by Imani et al. (2002) included, in the forward step-wise multivariate analysis, some endocrine factors, such as free-IGF-I and its binding proteins, that may alter the predictors that remained in the final model during the analysis. It is worth briefly discussing a statistical methodological difference between the present model and that suggested by Imani et al. (2002). The polytomous logistic regression model used on the present data set predicts FSH threshold doses that correspond to actual dose levels applied in the clinical trial, in contrast to a multiple regression where the predicted dose is on the continuous scale, as the equation published by Imani et al. (2002). In our model, the residual mean square error was 25 IU and therefore the individual prediction is at least as good as in the model suggested by Imani et al. (2002) where the residual mean square error was 35 IU. The R2 value of 25% in the present model is explained by the relatively few patients in the two highest dose levels in our trial compared with a more even distribution of patients in the trial conducted by Imani et al. (2002) yielding an R2 of 39% in their model. Our finding that the menstrual status was important is in line with expectations. Patients with amenorrhea may need more FSH to reach the threshold than a patient with oligomenorrhea, who may also need more FSH compared with patients with anovulatory cycles, indicating that the severity of the ovarian dysfunction may well be reflected in the degree of menstrual cycle disturbance. The relative resistance to FSH in amenorrheic patients is supported by the finding that these patients also have a relative resistance to endogenously released FSH during clomiphene citrate treatment (Imani et al., 1998), as patients with clomiphene resistance do release a similar amount of endogenous FSH as the clomiphene sensitive women (Polson et al., 1989). This supports the concept that it is indeed the threshold of the ovarian sensitivity that is altered. A high BMI was associated with a higher FSH threshold dose; an observation that is supported by the findings that the total dose of gonadotrophins needed to induce ovulation is increased in parallel with body weight (Hamilton-Fairly et al., 1992; Balen et al., 2006). The cause of the association with BMI may be that the endocrine disturbance is related to the degree of body fat and hyperinsulinaemia, but it may also be due to a lower bioavailability of exogenous gonadotrophins in obese women (Chan et al., 2003; Steinkampf et al., 2003). The type of gonadotrophin used for stimulation was not a significant predictor of FSH threshold dose, and thus of limited relevance compared with the patient's clinical characteristics. In theory, the threshold is defined as the lowest dose that induces growth of the single most sensitive follicle. In this clinical trial, a dual criterion for hCG administration was used, covering not only the development of a single dominant follicle, defined as one follicle of 17 mm or greater, but also the presence of 2–3 follicles of 15 mm or greater. Therefore, our threshold is based on the stimulation in those 132 (87%) patients who reached either of these criteria, including those patients with up to 3 follicles ≥15 mm. The threshold dose in the study and subsequently the model could also be termed the ‘response dose’ and can be viewed as a surrogate for the real threshold. It should be considered that serum FSH may accumulate over time so the theoretically ideal threshold should be defined as the dose that caused growth of a single dominant follicle to around 9–10 mm, where dominance is normally established. The exact timing of dominance was not determined in the present study, but it is important to note that the model for threshold dose predictions also applied in the subset of patients with mono-follicular development defined as only one follicle ≥17 and no follicles of 15–16 mm. The average number of follicles ≥15 mm was 1.5 (Balen et al., 2007) indicating that some patients did indeed receive FSH doses slightly above the threshold for single dominant follicle development. However, the analysis restricted to patients with mono-follicular development showed that the accuracy of the model was the same for this subset of patients as for the overall study population. On the basis of the intention to treat population in the present study, 92% of the patients received hCG and the ongoing pregnancy rate was 19%, of which 79% accounted for singletons and 21% for multiple pregnancies (Balen et al., 2007). These figures are much in line with literature data on 7 day step-up protocols (Homburg and Howles, 1999), but it could be argued that the multiple pregnancy rate is somewhat high, compared with a number of recent studies using starting doses of FSH <75 IU/day (Balasch et al., 2000; Calaf Alsina et al., 2003). Indeed, multiple pregnancy rates as low as 6% have been reported in a large clinical trial using starting doses of 50 IU/day, even though up to 3 follicles >16 mm before hCG administration was accepted (Calaf Alsina et al., 2003). An alternative approach is to restrict hCG administration to those patients who have only one or two mature follicles, but this would increase the cancellation rates. As published (Balen et al., 2007), 3.3% (5/151) of the patients had their cycle cancelled before hCG, due to development of multiple follicles. Identification of these patients with a very low threshold would have necessitated a starting dose of <75 IU/l. Ovarian volume was another significant parameter predicting the FSH threshold dose. The larger the ovaries, the more FSH was needed to obtain adequate ovarian response. This finding is consistent with earlier clinical studies of ovulation induction with gonadotrophins, where it has been shown that patients with larger ovaries need longer duration of stimulation and higher doses of gonadotrophins (Lass et al., 2002). Van den Meer et al. (1994) studied three consecutive cycles in 16 patients during gradually increased doses of intravenously infused urinary FSH and elegantly documented how the inter-individual variability in FSH threshold dose was far greater than the variability from cycle to cycle. The key problem is therefore to define the appropriate dose for any individual patient and the present study provides a suggestion for nomograms that may be used clinically. A major concern, if patients were dosed according to the nomograms, would be that 13% would receive a higher starting dose that the observed threshold dose. The implications of an excessive starting dose are likely to be the same as when step-down protocols are used, i.e. risk of multiple follicular development. Indeed, van Santbrink et al. (2002) have used the prediction model developed by the same group of investigators (Imani et al., 2002), and retrospectively evaluated the doses used in a step-down regimen. In that study, even though the average excessive dose was 28.5 IU/day above the threshold, the step-down approach resulted in an overall multiple pregnancy rate of 17% and not a single case of OHSS (van Santbrink et al., 2002). The expected main advantage of dosing according to the nomogram would be a shortening of the stimulation phase. It is important to note that the suggested dosage nomograms could be easily used clinically because they only include simple characteristics that are routinely recorded in all anovulatory patients. Before using this model in clinical practice, a randomized controlled trial should be conducted to compare the presently used low-dose step-up protocol with a flexible protocol using this model. Funding The study was sponsored by Ferring Pharmaceuticals A/S, Copenhagen, Denmark. Funding to pay the Open Access publication charges for this article was provided by Ferring Pharmaceuticals A/S, Copenhagen, Denmark.	[ "threshold dose", "anovulation", "predictors", "follicle-stimulating hormone", "efficiency" ]	[ "P", "P", "P", "P", "U" ]
Eur_Radiol-4-1-2270358	Dual-source CT for chest pain assessment	Comprehensive CT angiography protocols offering a simultaneous evaluation of pulmonary embolism, coronary stenoses and aortic disease are gaining attractiveness with recent CT technology. The aim of this study was to assess the diagnostic accuracy of a specific dual-source CT protocol for chest pain assessment. One hundred nine patients suffering from acute chest pain were examined on a dual-source CT scanner with ECG gating at a temporal resolution of 83 ms using a body-weight-adapted contrast material injection regimen. The images were evaluated for the cause of chest pain, and the coronary findings were correlated to invasive coronary angiography in 29 patients (27%). The files of patients with negative CT examinations were reviewed for further diagnoses. Technical limitations were insufficient contrast opacification in six and artifacts from respiration in three patients. The most frequent diagnoses were coronary stenoses, valvular and myocardial disease, pulmonary embolism, aortic aneurysm and dissection. Overall sensitivity for the identification of the cause of chest pain was 98%. Correlation to invasive coronary angiography showed 100% sensitivity and negative predictive value for coronary stenoses. Dual-source CT offers a comprehensive, robust and fast chest pain assessment. Introduction Comprehensive CT angiography protocols for a complete assessment of the thoracic vessels, often referred to as “triple rule out” protocols, are used in the differential diagnosis of chest pain increasingly often [1–4]. These protocols aim to opacify pulmonary and coronary arteries as well as the aorta simultaneously to rule out pulmonary embolism, coronary artery disease and aortic aneurysm or dissection in a single exam. With ECG gating, the acquired images of the coronary arteries should have a similar diagnostic accuracy as a specific coronary CT angiography. Quite a few studies have meanwhile shown the feasibility of a simultaneous evaluation of these vascular territories in one single breathhold scan with a good sensitivity in the identification of the cause of chest pain [5]. Also, recent studies indicate that coronary CT angiography can be helpful for a fast and cost-effective triage of chest pain patients [6, 7]. So far, a major limitation of these studies especially in acutely ill patients is the restricted image quality of the coronary arteries in high heart rates [8]. The administration of beta-blockers to lower heart rates is general practice in 16- and 64-slice CT [9]. However, this approach is time consuming and limited by contraindications, and a sufficient reduction of the heart rate cannot be achieved in many acutely ill patients. Initial studies of dual-source CT (DSCT) cardiac imaging have shown a robust image quality and a very good diagnostic accuracy of coronary CT angiography even in high heart rates [10, 11]. Additionally, a more comprehensive cardiac assessment including wall motion and valve function is possible with DSCT [10]. The aim of this study was to assess the diagnostic accuracy of a specific dual-source CT protocol for chest pain assessment, regarding a 6-month follow-up for other findings and coronary angiography in patients in whom coronary artery disease could not be reliably excluded as standard of reference. Materials and methods Patients One hundred nine consecutive patients [31 women, 78 men; median age 64 (59–67) years] were prospectively enrolled in the trial. The study was approved by the institutional review board, and informed written consent was obtained from every patient prior to the examination. All patients with acute chest pain were eligible for the study if referred by a colleague after initial diagnostic workup including physical examination, ECG and serum levels of creatinine and TSH. Exclusion criteria were positive ECG changes or troponine test, severe ventricular arrhythmia, a history of severe allergoid reaction to iodinated contrast material, renal insufficiency and young age below 30 years. Also, severe dyspnea with inability to hold the breath for at least 15 s was regarded as exclusion criterion. The patients were asked to hold their breath for approximately 15 s prior to the examination. If the breathhold could not be maintained, the patient was excluded from the study, and seven otherwise eligible patients had to be excluded for this reason. Heart rates were 68 ± 14, ranging from 58 to 118 bpm. Beta-blockers were not administered in preparation of the scan. Eighteen patients had known coronary artery disease and 14 were on continuous oral beta blocker medication. One patient had a known chronic aortic dissection. Examination Examinations were acquired on a DSCT Somatom Definition (Siemens, Forchheim, Germany) in supine position. A coronal topogram was scanned to plan the spiral acquisition, which included the whole chest from the first ribs to the diaphragm. The level of the carina was defined as the trigger point where the second tube is switched on additionally. The volume and flow rate of the contrast material (Ultravist 370, Schering, Germany) were adapted to the patient’s body weight as described elsewhere [5] according to Table 1. A bolus tracking technique was used to assess the transit time of the contrast agent. For this purpose, a region of interest was defined in the ascending aorta to detect the arrival of the contrast material bolus in images scanned at 2-s intervals, and the scan was initiated with a delay of 4 s if the density increased more than 100 Hounsfield units (HU). The patient was then instructed to hold his breath at a mild inspiratory position, and the spiral scan was started. The parameters for the acquisition in craniocaudal direction were 0.33-s gantry rotation time, 120-kVp tube voltage, 560-mAs effective tube current with 0.6-mm collimation and double Z-sampling. The pitch and the tube current modulation were adapted automatically to the heart rate of the patient. Pitch values ranged between 0.2 and 0.48, whereas the full tube current was only applied at 70% of the cardiac cycle for heart rates below 70 bpm and between 30 and 70% for higher heart rates. Scan duration was 15 ± 3 s. Table 1Body-weight-adapted contrast material injection protocolWeight (kg)Volume (ml)Flow (ml/s)501063.5551113.7601163.9651214.0701254.2751304.3801344.5851384.6901424.7951464.91001505.01051545.11101575.21151615.41201645.5 Image reconstruction Continuous ECG-gated axial slices and coronal images of the whole chest were reconstructed with a temporal resolution of 330 ms using a soft kernel (B30f). Slice thickness of 3 mm and 2.5-mm increment were applied to avoid artificial inhomogenities in the lung parenchyma due to undersampling with gating. Additionally, axial slices of the heart were reconstructed for a limited range from the carina to the diaphragm with a 200-mm field of view, 0.75-mm slice thickness and 0.5-mm increment using a B26f convolution kernel. If stents had been implanted, an additional reconstruction was performed using a B46 kernel to reduce blooming and enhance the depiction of the lumen. For ECG gating, a single segment reconstruction algorithm utilizing quarter scan segments from both detectors was applied to achieve a temporal resolution of 83 ms, and delays were set to 300 ms and 70% of the cardiac cycle for the initial reconstructions. Also, axial slices with 1-mm thickness and increment were reconstructed with a B26f kernel for the whole cardiac cycle at 10% intervals for dynamic evaluation. Diagnostic evaluation The examinations were read by two radiologists in consensus immediately after the exam, including a review of axial and coronal slices for evaluation of pulmonary embolism and aortic disease (Figs. 3 and 4). Additionally, a complete cardiac workup (Figs. 1 and 2) was done using a 3D workstation (Multimodality Workplace, Siemens, Forchheim, Germany) to interactively browse multiplanar reconstructions and volume-rendered images if necessary. Multiphase reconstructions were reviewed as cine loops of the heart with interactive browsing of dynamic short- and long-axis sections. The coronary artery tree was extracted using region-growing algorithms, and curved multi-planar reformats were created for review (Fig. 1a,b). In case of motion artifacts, additional reconstructions with an optimized delay were available. If these still did not resolve motion artifacts sufficiently to make a sufficient diagnostic evaluation feasible, the depiction of the respective coronary artery segment was rated as non-diagnostic and excluded from further analysis. Additionally, the multiphase reconstructions were reviewed to assess wall motion abnormalities and valvular disease (Fig. 2a-d). Fig. 1High grade stenosis (arrows) of the left anterior descending coronary artery. a Curved multiplanar reconstruction along the centerline of the vessel. b Orthogonal curved multiplanar reconstruction. c Left anterior oblique-cranial projection at invasive angiography. d Respective right anterior oblique projectionFig. 2Subendocardial scar (arrows) from a myocardial infarction in the territory of the left anterior descending coronary artery. a Diastolic mid-ventricular short axis view. b Systolic short axis view. c Diastolic left two-chamber view. d Systolic two-chamber view. Note the dyskinesia bulging out of the anterior wall. e Multiplanar reconstruction of the left anterior descending coronary artery. Note the stent (white arrow) in the proximal vessel, which had been implanted after interventional re-perfusion of the occluded vessel Indication for invasive angiography was evaluated by a cardiologist based on CT and clinical findings. Invasive angiography was always initiated if significant coronary artery disease could not be reliably excluded in CT and there was no other evident pathology explaining the symptoms. Selective catheterization and coronary angiography were performed by an experienced cardiologist using a transfemoral arterial approach and standard Seldinger technique. At least three standard projections including a 45° left anterior oblique, a 30° right anterior oblique and a 45° left anterior and 30° cranial projection were acquired for both coronary arteries. For quantitative coronary assessment (QCA) analysis, a densitometrical analysis of two projection images was performed using the Quant-Cor QCA software (Siemens, Erlangen). Conventional coronary angiograms of 29 patients were available for correlation. A third reader correlated the findings in the reports to invasive coronary angiography and performed dose calculations, which were based on the dose length product given in the patient protocol of the CT scanner. Also, he reviewed the files of patients with initially negative results for further diagnoses in the subsequent 6 months. The readers of the initial CT images were unaware of any findings from conventional angiography, and most invasive angiographies were performed to further evaluate or treat lesions detected in CT. Thus, the cardiologists performing the coronary angiography were mostly aware of the CT findings. Statistical analysis Continuous variables are given as median and range. Ninety-five percent confidence intervals were calculated using MedCalc software (MedCalc Software, Mariakerke, Belgium). Diagnostic accuracy was calculated as sensitivity, specificity, positive and negative predictive values based on standard contingency tables. Results Technical success The scans were acquired from all 109 patients without severe adverse events. Although previously tested, three patients were unable to hold their breath for the duration of the scan, which resulted in artifacts in the lower part of the chest, rendering the coverage and depiction of the coronary arteries insufficient for diagnostic evaluation. All examinations showed an adequate contrast enhancement of the pulmonary arteries, the coronary arteries and the aorta except for six examinations. In two, the bolus was too short due to hemodynamic parameters, making the assessment of pulmonary embolism only possible in the central arteries. In three examinations an erroneous administration of a smaller volume of contrast material caused a weak opacification of the pulmonary arteries, and in one examination overall opacification was too weak (220–245 HU) for coronary artery assessment. Patient findings The diagnoses in the patient population were spread over the different organs and vascular territories quite evenly (Table 2). Ten patients had pulmonary embolism (Fig. 3) and three had morphological signs of chronic pulmonary hypertension. Twelve patients showed other, non-vascular pulmonary findings including pneumonic consolidation in four patients and edema from left heart failure in four cases. Thirty-seven patients had coronary atherosclerosis, and in 21 of these patients, stenoses were rated as hemodynamically significant (>50%) (Fig. 1). Three coronary arteries were occluded. Fifteen patients had had bypass surgery, and five bypass graft vessels were occluded. One bypass was rated as stenotic, but invasive angiography showed a patent lumen and a very slow runoff. There were 20 non-coronary cardiac findings including valvular disease in seven patients. Among those, there were four cases of aortic stenosis, one bicuspid aortic valve and one with vegetations from endocarditis. In one patient an insufficiency of tricuspid valve leaflets was diagnosed. Three patients had had aortic valve replacement, and the valve prostheses worked normally. Seven patients showed an abnormal global or regional wall motion with pulmonary congestion in four cases. Areas of myocardial infarction were noted in three patients (Fig. 2). Other non-coronary cardiac findings were left ventricular hypertrophy, scars from myocarditis and ventricular thrombi. Among the aortic pathologies there were seven aneurysms, five dissections (Fig. 4), one case of intramural hematoma and one acute rupture with active bleeding into the mediastinum. Incidental vascular findings were a lusorian artery and a right descending aorta with Kommerell’s diverticulum (Fig. 5). Another frequent finding was neoplastic lesions, partially with unknown primary cancer. Three patients had disseminated pulmonary metastases, one patient had a previously known lymphoma and one an esophageal cancer. In 26 patients there were no findings that were regarded as the main cause of chest pain. The review of the files of those patients in a 6-month follow-up showed additional diagnoses in two patients: One had intermittent arrhythmia that was identified as the cause of chest pain. The other patient developed pleural effusion and was diagnosed to have pleuritis. In the other 24 patients there was no evidence of the cause of chest pain in follow-up. Considering those results as correct-negative, the sensitivity of the protocol in the identification of the cause of chest pain amounts to 98%. Specificity can be estimated at 92% based on these data, although the exact significance of the findings for chest pain is frequently unclear and may therefore be regarded as arbitrary. Fig. 3Curved multiplanar reconstruction of the lower lobe pulmonary arteries showing emboli in the lobar and segmental arteriesFig. 4Dissection of the descending aorta. a Multiplanar reconstruction. Arrows indicate the depiction of the left coronary artery in the gated reconstruction (black arrow), the dissection membrane (white arrow) and the junction between the gated scan and the non-gated scan, which had been continued in this case to include the whole extent of the dissection (outlined arrow). b Angulated para-coronal reconstruction showing the topographyFig. 5Kommerell’s diverticulum (arrows). a Volume-rendered reconstruction showing the origin from the aortic arch and the course of the subclavian artery. Note the median-right course of the descending aorta ventral to the spine. b Curved multiplanar reconstruction showing the proximity to the tracheaTable 2Diagnoses found in the study populationTerritoryTypeDiagnosisnPulmonary19%Vascular62%Embolism77%10Pulmonary hypertension23%3Non-vascular38%Pneumonia50%4Edema50%4Cardiac40%Coronary84%Atherosclerosis100%37Relevant stenosis57%21Occlusion8%3Bypass grafts18%Patent75%15Occluded25%5Non-coronary13%Valvular50%7Myocardial50%7Aortic13%Aneurysm50%7Dissection43%6Rupture7%1Other7%Vascular variants25%2Neoplastic75%6Negative24%26Sum of patients100%109 Correlation to invasive angiography The 29 conventional coronary angiograms that were available for correlation of the coronary findings were evaluated in a segment-based and a patient-based analysis. Thirty-two coronary artery segments (7%) of these patients were excluded from analysis due to small size or insufficient visualization in CT. Of the 29 patients, 25 had evidence of coronary atherosclerosis in CT. Nineteen stenoses were rated as potentially hemodynamically relevant with more than 50% stenosis in CTA, but invasive angiography confirmed only 15 of those, whereas the other 4 were overrated. There were no false-negative results, i.e., all the stenoses had been cited in CTA, although it has to be taken into account that the indication for invasive angiography was frequently derived from CTA findings. Thus, sensitivity, specificity, positive and negative predictive value amounted to 100%, 99%, 79% and 100%, respectively, in the segment-based analysis. In the patient-based analysis, all 13 patients with significant stenoses were identified. There were two patients with completely negative CTA and symptoms very suggestive of coronary ischemia who had additional invasive angiography, and both of them remained negative. The overrated stenoses occurred in patients with other, significant stenoses. Thus, CTA achieved full diagnostic accuracy in correlation to invasive angiography for this patient-based analysis. The database of the coronary angiography laboratory was screened for those patients who had not initially been examined invasively, and there were no readmissions for angiography within 6 months after inclusion of the last patient. Radiation exposure As the patient protocols had not been archived in all patients, there were only 51 dose-length products available for calculation. Based on a conversion factor of 0.012 for chest exams, the estimated equivalent dose ranged between 9.2 mSv in low heart rates with strict dose modulation to 19.5 mSv in tall patients with heart rates around 70–80 bpm, which resulted in a wide pulsing interval of 30–70% with a still quite low pitch below 0.3. Overall, the median radiation exposure was 15.1 mSv (95% CI 9.9–18.8). Discussion CT angiography can be regarded as the modality of choice for the assessment of pulmonary embolism and aortic disease in acute care [12]. The fast volume coverage that has become available with recent CT technology has triggered efforts for a combined assessment of these pathologies, and the additional evaluation of the coronary arteries would make the protocol a universal diagnostic tool for chest pain assessment [2, 4, 12]. Quite a few studies showed the feasibility of this approach based on 64-slice CT technology with temporal resolution in the area of 165–200 ms, but also indicated that coronary CTA still represented the limitation of the method due to the high incidence of high heart rates that made coronary assessment unreliable [1, 3]. Recent studies have shown that diagnostic images of the coronary arteries can be obtained reliably with dual-source CT even in very high heart rates [10, 11]. Thus, dual-source CT was expected to greatly improve the diagnostic accuracy of the combined CT protocol, and our results indicate that this assumption was correct: There were no examinations rated as non-diagnostic because of blurring of the coronary arteries due to high heart rates. There were six exams with partially insufficient contrast opacification, which mainly affected pulmonary artery evaluation and three examinations with breathing artifacts that impeded coronary assessment, but otherwise the main coronary artery segments were reliably assessable. This is reflected in the detailed analysis of the coronary CT angiography in relation to invasive angiography, in which only 7% of coronary artery segments were excluded, mainly due to very small size in distal vessel segments and side branches. Regarding the diagnostic accuracy, the excellent performance in the patient-based analysis confirms that the method can be very helpful for fast patient triage. However, it has to be acknowledged that stenoses were quite frequently rated as ‘potentially relevant’ in CTA so that invasive angiography was not only necessary for intervention, but also for definite diagnosis. Also, some false-positive results would have to be expected in larger patient groups. Still, it has to be emphasized that a valid, relevant diagnosis could be made in 62% of the patients, and in 25% relevant diseases were ruled out. Regarding sensitivity for the cause of chest pain, the value of 98% in the present study is somewhat higher than the 93% observed in a study on ECG-gated 64-slice thoracic CTA [1] and compares well with an initial DSCT study that reported a sensitivity of 96% [13]. In 13% of patients there were findings such as coronary atherosclerosis or vascular variants that were not immediately relevant as a reason for chest pain, but that are still valuable information for further patient management. Comparing the results of coronary CTA correlation in this study, sensitivity, specificity, positive and negative predictive values of 100%, 99%, 79% and 100% are comparable to an initial study by Scheffel et al. [11]. They observed respective values of 96%, 98%, 86% and 99% for dual-source CT coronary CTA without heart rate control in a high pretest-probability group, although the rate of 1.4% non-assessable segments was remarkably low compared to 7% in our study. The somewhat higher negative and lower positive prediction may be attributed to a more sensitive diagnostic reading in our group of acutely ill patients. In a similar 64-slice CT study by Nikolaou et al. [9], respective values of 97%, 79%, 86% and 96% with 10% non-assessable segments were recorded for an intermediate likelihood population with strict heart rate control. These values suggest that a similar accuracy can be achieved with DSCT without beta blockers. Due to the variability in body size and habitus, the equivalent doses calculated from dose length products are only correct for less than half of the population and therefore can only serve as rough estimations. The median value of 15.1 mSv is about twice that of coronary artery CTA with dual-source CT, but only moderaltely higher than that for coronary CTA alone with single-source 64-slice CT [14, 15]. The increase is mainly related to the high heart rates in this acute patient group that require a wider pulsing interval for the reconstruction of systolic images. The longer coverage including the upper chest contributes rather little to the dose because it is acquired normally with only one tube switched on. Therefore, performing coronary CTA alone without coverage of the upper chest would save only little of the dose in this patient group. Thus, the additional dose from the longer coverage is well invested considering the diagnostic value. Also, comparing DSCT to single source CTA, the high diagnostic accuracy of DSCT can be regarded as well invested dose compared to single source scanners in which the dose can be estimated to be higher for this patient group because the pitch would have to be reduced to apply multi-segment reconstruction algorithms [16, 17], and pulsing would mostly have to be switched off to yield diagnostic examinations with optimized reconstructions [8]. Also, considering alternative diagnostic approaches for the evaluation of chest pain, for example, including pulmonary CTA, aortic CTA and invasive coronary angiography, this combination would result in a higher total patient dose [1]. Magnetic resonance imaging would be an alternative modality for the assessment of pulmonary embolism and aortic pathology, and an MR exam of myocardial wall motion and perfusion could serve as a method to rule out hemodynamically relevant coronary pathology without radiation exposure [18, 19]. However, the limited availability and the long examination times currently let such an approach seem less applicable in a clinical setting [20]. Several limitations of this study have to be acknowledged. One limitation is the exclusion of patients with the evidence of an acute coronary syndrome based on ECG or serum markers, which causes a bias in the assessment of the diagnostic accuracy of coronary CTA. Another limitation is the lacking correlation to other diagnostic approaches in the evaluation of pulmonary embolism and aortic disease. Of course, additional ventilation/perfusion scintigraphy and aortic MRI would be desirable for study purposes, but cannot be justified considering the necessary time and the encroachment of acutely ill patients and the involved radiation exposure. Also, additional invasive coronary angiography of the patients without coronary findings would be helpful to assess the actual negative predictive value, but the application of this invasive procedure involving radiation exposure to patients with other pathologies would of course not be acceptable. The follow-up of the coronary angiography database that showed no readmissions at least confirms the unlikelihood of relevant coronary artery stenoses in these patients. The fact that immediately therapy-relevant findings in all vascular districts were represented and quite evenly distributed in our patient population emphasizes that a combined protocol can be a valuable tool for diagnostic workup of acute, unclear chest pain. In conclusion, the specific dual-source CT protocol for chest pain assessment proved to be a very helpful tool offering a fast diagnostic workup and patient triage.	[ "chest pain", "ct angiography", "pulmonary", "embolism", "coronary artery disease", "aortic dissection" ]	[ "P", "P", "P", "P", "P", "P" ]
Pediatr_Nephrol-3-1-2100430	Cardiovascular complications of pediatric chronic kidney disease	Cardiovascular disease (CVD) mortality is a leading cause of death in adult chronic kidney disease (CKD), with exceptionally high rates in young adults, according to the Task Force on Cardiovascular Disease. Recent data indicate that cardiovascular complications are already present in children with CKD. This review summarizes the current literature on cardiac risk factors, mortality and morbidity in children with CKD. Learning objectives: To review recent data on the epidemiology of CVD in children with CKD To understand the mechanisms of cardiovascular abnormalities in pediatric CKD To review recent advances in the diagnosis and clinical presentation of cardiovascular complications in children with CKD To outline current understanding in the strategies to prevent progression of CVD in children with CKD. Cardiovascular mortality and morbidity The survival of children with CKD in the U.S. remains low: for children on dialysis the lifespan is 40–60 years less and for transplant patients, about 20–25 years less than that of an age- and- race-matched US population [1, 2]. The most likely cause of this is increased cardiovascular mortality due to the development of accelerated ischemic heart disease and premature dilated cardiomyopathy. The evidence comes from studies of young adults who developed renal failure during childhood. Oh et al. [3] analyzed the outcome of 283 young adults with childhood onset CKD between 1970 and 1997. Fifty percent of the deceased patients had died of cardiovascular or cerebrovascular events. Groothoff et al. [4] conducted a national retrospective and prospective cross-sectional study to evaluate the late physical, social and psychological effects of renal insufficiency (LERIC) in all Dutch children who started renal replacement therapy between 1972 and 1992. Of 381 patients, 85 had died. Cardiovascular disease was the most common cause of death and accounted for 41% of all deaths. Cerebrovascular accident, congestive heart failure, myocardial infarction and cardiac arrest (respectively) were the most common causes of cardiac death. Similar analysis of long-term survival from the Australia and New Zealand Dialysis and Transplant Registry [5] of all children and adolescents who were under 20 years of age when renal replacement therapy commenced (study period was 40 years) showed mortality rates 30 times higher than in the age-matched general population; CVD was the most common cause of death (45%). Cardiovascular death happens not only in later life, but also in childhood CKD. In the general pediatric population, the incidence of annual death due to cardiac disease is less than 3%. Yet annual reports from the United States Renal Data System (USRDS) indicate that over the last decade CVD has remained the second most common cause of death in children on chronic dialysis or after transplantation, accounting for approximately 20–25% of all deaths. Parekh et al. [6], using the USRDS database, performed a detailed cross-sectional analysis to evaluate the risk of a cardiac death in children and young adults (age 0–30 years) and to identify factors potentially associated with CVD mortality. A total of 1,380 deaths between 1990 and 1996 were analyzed. There were 311 cardiac deaths (22.5% of the total). Cardiac deaths in children and young adults in whom ESRD developed during childhood were approximately 1,000 times more frequent than in the general pediatric population. Of the specific categories of cardiovascular deaths, cardiac arrest was the most common cause in each of the age groups, followed by arrhythmia and cardiomyopathy. These causes of cardiac death are different from those of adults. In adults, coronary artery disease and chronic congestive heart failure are the leading causes of CVD mortality and, as shown by Parekh et al. [6], these causes are extremely rare in children and adults younger than 30 years of age. The incidence of cardiac arrest in the youngest age group (0–4 years) was five to ten times higher than in other age groups, perhaps, as noted by the authors, a reflection of the difficulty of ascertaining the true cause of death in young children. Some of these young children might have died from other co-morbid conditions such as congenital disorders that are not included in the USRDS database. The high rate of sudden death in children, especially in infants with ESRD, is poorly understood and warrants further investigation. In adults, sudden death is often a result of fatal arrhythmias due to acute ischemia of preexisting atherosclerotic disease. It is believed that arrhythmias are also the likely cause of most cases of sudden cardiac death (SCD) in children. However, the origin of acquired malignant arrhythmias in children is unlikely to be an atherosclerotic lesion. Dilated, especially hypertrophic, cardiomyopathies are a leading cause of SCD in children [7]. The macroscopic and microscopic structural abnormalities in cardiomyopathies involve fibrosis and cellular hypertrophy prone to produce an electrical instability with resultant arrhythmias. Ischemia of small coronary vessel disease secondary to medial hypertrophy might result in dispersion of repolarization properties and arrhythmia from re-entrant or autonomic mechanisms. As we discuss in more detail below in this review, children with CKD develop left ventricular hypertrophy (LVH), which is frequently severe, especially in children on prolonged dialysis therapy [8, 9]. It is currently unknown if LVH in young patients with CKD is characterized by structural abnormalities similar to familial or idiopathic hypertrophic cardiomyopathies associated with SCD. Whether LVH of children with ESRD can contribute to increased SCD is also not known. Another possibility for deadly arrhythmias in children with ESRD is acute changes in the cardiac extra- or intracellular ionic milieu, especially involving abnormalities of sodium- and potassium-based repolarization currents. Because the causes of cardiac death in children and adults are different, it is not surprising that none of the traditional or uremia-related risk factors for adult atherosclerotic CVD predicted cardiac death in the study by Parekh [6]. One of the examples presented in this paper is the effect of race on CVD death. In adults with ESRD, white males are at higher risk for CVD mortality. In contrast, black children as shown by Parekh et al. [6] appear to have an increased risk for cardiac death. Another important observation of this study is that transplant recipients had 78% lower risk of cardiac death than dialysis patients. However, the authors pointed out that the CVD mortality rate in transplanted patients was still significantly higher (approximately ten times) than in the general pediatric population. The analysis of cardiac morbidity in children on chronic dialysis performed by Chavers et al. [10] has confirmed that cardiac disease in children is different from adults. A total of 1,454 Medicare incident pediatric (0–19 years) dialysis patients were identified from 1991 to 1996. Among them, 452 (31.2%) developed a cardiac-related event. Arrhythmia was most common (19.6%), followed by valvular disease (11.7%), cardiomyopathy (9.6%) and cardiac arrest (3%). Ischemic heart disease was extremely rare in these children. Risk factors for cardiovascular disease The risk factors and pathogenic mechanisms of development of CVD in young adults who had onset of CKD in childhood are better understood than are those producing cardiac morbidity and mortality in children. The conventional thinking is that two groups of risk factors are responsible for accelerated CVD in adults with CKD (Table 1). Table 1Cardiovascular risk factors in chronic kidney disease in adultsTraditionalCKD-relatedOlder ageDecreased GFRWhite raceProteinuriaMale genderPeripheral renin-angiotensin-aldosterone activityHypertensionAbnormal calcium and phosphorus↑ LDL CholesterolDyslipidemia↓ HDL CholesterolHypoalbuminemiaDiabetes mellitusHemodynamic overloadTobacco useAnemiaPhysical inactivityThrombogenic factorsPsychosocial stressHyperhomocysteinemiaFamily history of CVDOxidative stressLVHInfection (Chlamydia pneumoniae)ObesityChronic inflammation First, as compared to the non-uremic population, there is an over-representation in uremic patients of classical risk factors, e.g., diabetes, hypertension and hyperlipidemia. A majority of the adults who develop ESRD do so as a complication of diabetes or generalized atherosclerosis. Often cardiac disease antedates the onset of CKD in these patients. Unfortunately, children with CKD share with adults a similar high prevalence of risk factors for adult-type atherosclerotic CVD. It is also troubling that the frequencies of these traditional risk factors have not changed over last decade. The North American Pediatric Renal Trials Collaborative Studies (NAPRTCS) data demonstrate that hypertension develops at early stages of CKD (48%) and persists (50–75%) in uremic children [11–14]. Dyslipidemia is found in 70 to 90% during chronic dialysis [15–18]. Successful renal transplantation leads to a dramatic improvement in renal function and elimination of many risk factors for atherosclerotic CVD that were present while on dialysis. However, transplant recipients are not free from multiple complications, and transplantation may amplify some of the traditional risk factors. Indeed, the prevalence of hypertension in pediatric renal allograft recipients is between 50–80% [19–21]. Also, hyperlipidemia may not disappear after renal transplantation; the reported prevalence is above 50% [22–24]. The NAPRTCS data also show that the rate of obesity, another CV risk, is increasing in children with ESRD at the time of transplantation (12.4% after 1995 vs. 8% prior to 1995) [25]. Single-center study data indicate that the number of obese patients can double at 1 year after transplantation [26]. Second, there is a multitude of uremia-related risk factors for atherosclerotic CVD. In adults with CKD, an increased homocysteine level appeared to be an independent risk factor for CVD morbidity and mortality [27–29]. Anemia has been linked to negative CVD outcome [30]. Hyperphosphatemia with increased calcium-phosphorus product constitutes a risk for cardiovascular calcification, cardiac ischemia and adverse cardiovascular outcomes [31, 32]. Elevated serum C-reactive protein (CRP), a marker of systemic inflammation, has been found to be a strong predictor of cardiac morbidity and cardiac death in CKD patients [33]. Another marker of inflammation, IL-6, has been associated with increased cardiac morbidity in adults with CKD [34, 35]. The role of systemic inflammation in CKD has been reviewed recently elsewhere [36]. Inflammation is directly linked to oxidative stress, which is now considered as a hallmark of uremia [37]. Asymmetric dimethylarginine (ADMA), an endogenous inhibitor of nitric oxide synthase and a marker of oxidative stress, has been extensively evaluated in patients with CKD by Zoccali et al. [38]. The authors demonstrated that elevated ADMA per se was responsible for a 52% higher risk of death and for a 34% higher risk of cardiovascular events in dialysis patients [39]. Another subset of novel CVD risk factors is adipokines. Adipokines, leptin and adiponectin, are the products of adipose tissue involving in regulation of lipid and glucose metabolism. Abnormal adipokines are strongly linked to insulin resistance, a known CV risk in CKD [40]. It has been determined that lower plasma adiponectin concentration independently predicts increased CVD morbidity and mortality in adults with ESRD [41]. Clinical Practice Guidelines for Cardiovascular Disease in Dialysis Patients extensively reviewed the current literature on CVD biomarkers, and the reader is referred to this publication for more information [42]. As in adults, in children with CKD many of these risk factors are very prevalent (Table 2). Table 2Prevalence of risk factors for CVD in children with CKDRisk factorsCRIa (%)Dialysisb (%)Transplantc (%)ReferencesHypertension4852–7563–8111a, 12–14b, 19–21cDyslipidemia25–5333–8755–8415a, 15–18a, b, 22–24cAnemia4840–6732–6443a, 44–47b, 48–51c,Hyperparathyroidism32.6–43.758–43Hyperhomocysteinemia87–9225–10052c, 53b, c, 54c, 55b↑ CRP761656b,57cHypoalbuminemia–40–60–43b, 44b, 58–59b Recent studies in adults on chronic hemodialysis place a malnutrition-inflammation complex at the center of a debate about the role of traditional and non-traditional risk factors for poor cardiovascular outcome. This issue emerged after publication of a series of articles, summarized in the review by Nurmohamed and Nube [60], describing the phenomenon of “reverse epidemiology.” The studies have shown that in adults on chronic hemodialysis low blood pressure, low body mass index (BMI), low serum cholesterol and low serum homocysteine are often correlated with an unfavorable clinical outcome. Thus, whereas traditional risk factors of CVD are correlated with an unfavorable outcome in the general population and patients with CKD not yet on dialysis, in hemodialyzed patients, mild hypertension, hypercholesterolemia and being overweight appear to be protective and associated with an improved survival. It has been speculated that a malnutrition-inflammation-atherosclerosis complex underlies, at least partly, the phenomenon of reverse epidemiology, since malnutrition causes a low BMI, hypocholesterolemia and low serum homocysteine levels. Mechanisms of cardiovascular disease in chronic kidney disease There are two parallel processes involved in the development of CVD in CKD patients (Fig. 1). Fig. 1Cardiovascular disease in chronic kidney disease The first is cardiac remodeling leading to hypertrophy of the left ventricle (LV) as a response to mechanical or hemodynamic overload. Two different patterns of LV remodeling can produce increase in LV mass (LVM) [61]. The concentric LV remodeling and hypertrophy may be the results of pressure overload as occurs with hypertension, whereas eccentric LVH may be related to volume and sodium retention, anemia and arteriovenous shunt. The patterns of sarcomere formation induced by pressure or volume overload are distinct. Pressure-induced concentric LVH is characterized by a parallel addition of sarcomeres resulting in the increase of cross-sectional area and diameter of the myocytes. Increase in LVM in this case is obtained by a marked increase in wall thickness with a less evident increase in the LV cavity that yields an elevated relative wall thickness and concentric LVH. From the physiological view, increased systolic blood pressure (BP) and pulse pressure, due to increased peripheral resistance and arterial stiffness, are the principal factors opposing LV ejection and leading to an increased LV workload and concentric LVH. An increase in LVM can also be obtained by an increase in the LV cavity with a symmetric increase in wall thickness to maintain the ratio between the wall thickness and LV transversal radius (relative wall thickness) normal, producing eccentric LVH. In this case, the addition of sarcomeres occurs mainly in series resulting in longitudinal cell growth. In the transition to maladaptive LVH, LV dilatation becomes disproportional to wall thickness, with myocytes elongated without an increase in diameter. Experimental models of cardiac hypertrophy support the theory that mechanical stress due to either pressure or volume overload is a trigger for activation of other multiple mechanisms leading to myocardial remodeling [62]. These factors include a local overexpression of the renin-angiotensin-aldosterone system (RAAS), adrenergic system, inflammatory cytokines and other autocrine and paracrine mechanisms. In patients with CKD, these mechanisms might be activated independently of hemodynamic overload since uremia per se is associated with an alteration in multiple humoral factors [63, 64]. With time, a maladaptive phase of LVH develops, characterized by decreased capillary density, decreased coronary reserve and subendocardial perfusion, a tendency to arrhythmia, and the development of myocardial fibrosis. All this leads to myocyte death and, finally, to diastolic and systolic dysfunction. The second process involves vascular injury. Exposure to CV risk factors results in vascular changes, including atherosclerotic and arteriosclerotic processes and vascular calcification. Atherosclerosis refers to the process of plaque formation or atheroma development. The process of atheroma formation begins with an accumulation of lipid-containing foam cells (macrophages) in the vascular intima and evolves into successive structures that penetrate the vascular wall and include lipids, smooth muscle cells and collagen fibers [65]. Calcification is an intrinsic part of the process and generally involves the intima. Atherosclerotic lesions have a patchy distribution along the length of the artery and cause local stenoses and occlusions. Recently, endothelial progenitor cells (EPCs) have been identified as contributing to angiogenesis [66]. The EPC number has been shown to be reduced in patients with CVD, leading to speculation that atherosclerosis may be caused by a consumptive loss of the endothelial repair capacity. Animal experiments have shown that EPCs reendothelialise injured vessels and that this reduces neointimal formation, confirming that EPCs have an atheroprotective effect. Arteriosclerosis is arterial stiffening involving the entire arterial tree, although it principally affects the elastic arteries. Unlike atherosclerosis, arteriosclerosis involves both intimal and medial thickening. In CKD, arteriosclerosis can occur in the absence of significant atherosclerotic disease [67–69]. Arteriosclerosis is associated with vascular remodeling characterized by increased wall thickness, lumen enlargement and increased length of arteries. This leads to an increase of systolic BP and pulse pressure and arterial stiffening. The triggers for vascular calcification are complex and include metabolic, mechanical, infectious and inflammatory injuries. Increased calcium x phosphate ion product or hyperphosphatemia may be the key promoter of vascular calcification [70]. The mechanisms include either stimulation of the uptake and precipitation of calcium and phosphate into the vessel or a decrease of the inhibitory process that prevents these ions from precipitation. Fetuin-A, the anti-inflammatory protein, is a highly potent inhibitor of serum calcium-phosphate complex formation [71, 72]. The serum concentrations of Fetuin-A are decreased in ESRD patients [73]. Another promoter of vascular calcification is 1,25(OH)2D3, which may have a direct effect on the calcium deposition in vascular smooth muscle cells [74]. In a comprehensive review, Vattikuti and Towler [75] defined four major histo-anatomic variants of vascular calcification, including classic atherosclerotic (fibrotic) calcification related to disorders of lipid metabolism, medial arterial calcification, vascular calcifilaxis and cardiac valve calcification. Spectrum of cardiovascular abnormalities in children with CKD Over the last decade, abnormalities of the LV such as LVH and LV dysfunction, abnormalities of the large arteries such as stiffness and increased intima-medial thickness (IMT) of the carotids, and calcification of the coronaries have been accepted as early markers of cardiomyopathy and atherosclerosis. They constitute strong independent predictors of cardiac morbidity and mortality both in the general population and in adults with CKD. In children and young adults with CKD, recent studies have proven that these abnormalities are also present and that risk factors for cardiac and vascular injury in children with CKD are similar to those for adults. Left ventricular hypertrophy LVH develops when renal insufficiency, is mild or moderate in children and progresses as renal function deteriorates. About one third of children with mild to moderate renal insufficiency have an increased left ventricular mass (LVM) index [76–78]. In a 2-year prospective longitudinal study of 31 pediatric subjects with CKD stage 2–4, Mitsnefes et al. [79] showed that a substantial proportion of children had a significant increase in the LVM index, with many of the children developing LVH. Indeed, 32% of the patients who initially had a normal LVM index have developed incident LVH. At initiation of maintenance dialysis, 69–82% of pediatric patients have evidence of LVH [80, 81]. LVH persists (40–75%) during long-term dialysis [8, 76, 82–85], with both concentric and eccentric geometric patterns of LVH present in these patients. Post-mortem studies confirmed high rate (more than 50%) of LVH in children with ESRD [86]. Small retrospective studies also suggest that with a better BP and volume control, LVH regression might be achieved in young patients on dialysis [80, 81]. On the other hand, a recent retrospective study demonstrated that LVH remains very prevalent and severe in a selected group of children who remained on maintenance dialysis for at least 2 years [9]. As in children prior to transplantation, most pediatric studies indicate that LVH remains common post transplant (48–82%) [87–91]. In contrast, a significantly lower frequency of LVH was found in a study by Englund et al. [92], who reported the results of a longitudinal analysis of children receiving renal transplants 10–20 years ago. Of 53 children who received a renal transplant between 1981 and 1991, 47 survived and were observed for 10 to 20 years. At the 10-year follow-up, echocardiography showed minor LVH in only two children with hypertension. No child without hypertension at 10 years post transplant had LVH. The factors associated with cardiac hypertrophy in children are similar to those in adults with CKD. As in adults, most pediatric studies of patients with pre-terminal, terminal renal failure and post transplant found significant relationships between low hemoglobin and an increased LVM index [8, 78, 79, 90]. However, recent adult studies in mild to moderate CKD or in chronic dialysis determined that correction of anemia was not associated with regression of LVH [93–96]. Authors suggested that relationships between anemia and LVH might not be causal. Of note, the above studies enrolled subjects with relatively mild degrees of baseline anemia and could not answer the question whether treatment of patients with initially significantly decreased hemoglobin levels might lead to a reduction of LVM. In contrast, Morris et al. [97] observed a significant reduction in the LVM index with the correction of severe anemia in seven children on chronic dialysis. There are several studies on the association between parathyroid hormone (PTH) levels and LVH in adults with CKD [98, 99]. In children, elevated PTH is associated with progression of LVH in stages 2–4 CKD [79]. The possible mechanisms of parathyroid-induced cardiac hypertrophy in CKD include a direct effect of PTH on cardiomyocytes and an indirect effect via elevated BP [99]. A support for a causal relationship comes from in vitro studies showing that PTH appears to have chronotropic, inotropic as well as hypertrophic effects on cardiomyocytes [100, 101]. In adults with CKD, hypertension is directly linked to the development of LVH [39]. The relationships between BP and LVH in pediatric CKD are unclear. Consistent correlations of LVM and BP are limited to children with ESRD [8, 85]. However, a detailed cross-sectional analysis of BP characteristics by ambulatory blood pressure monitoring (ABPM) in children from the ESCAPE trial did not demonstrate any relationship between office BP or ABPM parameters and LVM, suggesting only a minor role of hypertension in the pathogenesis of LVH in early CRI [78]. In contrast, analysis of longitudinal data suggests that ABPM might be an important tool to assess the risk of development of LVH in children with CKD [79]. In this study, authors determined that an increase in the nighttime systolic BP load (number of BP measurements above the 95th percentile BP value) were independently associated with the increase in the LVM index over time, arguing that persistent and chronic elevation of BP might be more important in the development of LVH [79]. Left ventricular function In contrast to adults, in whom systolic dysfunction is frequently associated with early cardiac failure and decreased survival, in children with CKD systolic LV function is usually preserved [76, 102–104]. On the other hand, diastolic dysfunction, often the initial abnormality of cardiac function, is already present in children with CKD. Doppler measurement of mitral inflow velocity has been the most widely used method to assess LV diastolic function. Using this method, Goren et al. [103] showed that LV relaxation (E/A ratio) was impaired in dialyzed children as compared to controls. Johnstone et al. [76] also found a reduction in the E/A ratio in children on chronic peritoneal dialysis and with pre-terminal renal failure, although none of these patients had an E/A ratio <1.0, which is considered to be abnormal. Unfortunately, the transmitral Doppler velocities and, therefore, the E/A ratio, are affected by several factors, including left atrial pressure and preload. This is particularly important for patients with advanced chronic renal failure, since many of them are hypervolemic. Recently, new indices were introduced to evaluate diastolic function using tissue Doppler imaging (TDI). In contrast to E/A, the TDI indices may be less load dependent and provide a more accurate measure of diastolic function. Recent studies employing TDI determined that children with CKD might have abnormal diastolic function [105, 106]. In these studies, children on chronic dialysis had significantly worse diastolic dysfunction than children with mild-to-moderate CRI or post transplant. Poor diastolic function in patients on dialysis was associated with anemia, hyperphosphatemia, increased calcium-phosphorus ion product and LVH. The clinical significance of diastolic dysfunction in pediatric patients with CKD is not known. Longitudinal studies are necessary to determine if abnormal diastolic function predicts the development of systolic dysfunction and congestive heart failure in these patients. Arterial structure and compliance Studies of young adults who developed ESRD during childhood found a high prevalence of abnormal carotid IMT, diminished arterial wall compliance and coronary artery calcification (CAC). These vascular abnormalities are accepted as markers of asymptomatic atherosclerosis and predictors of future symptomatic CVD in the general population and in adults with CKD [107, 108]. Groothoff et al. [109] found increased arterial stiffness and showed that systolic hypertension was the main determinant of abnormal arterial wall compliance. Goodman et al. [110] showed that among 23 patients on chronic dialysis who were younger than 20 years of age, none had evidence of CAC; in contrast, 14 of the 16 patients who were 20 to 30 years of age had evidence of CAC on CT scanning. Oh et al. [3] screened for coronary and carotid artery disease in 39 patients, aged 19 to 39 years, with childhood onset ESRD. Coronary artery calcification was present in 92% and carotid IMT was significantly increased compared to matched controls. Carotid IMT was correlated with cumulative dialysis and serum Ca x P product in their study. Milliner et al. [111], in an autopsy study of pediatric patients with ESRD who died in 1960–1983, showed a high prevalence of soft tissue and vascular calcinosis. In their study, CAC was present in 28%. Peak Ca x P product, peak serum P and cumulative dose of calcitriol were significantly associated with the severity of the calcinosis. Civilibal et al. [112] screened 53 children with ESRD for the presence and predisposing factors of CAC. Coronary artery calcification was present in 15% of patients (three hemodialysis patients, three peritoneal dialysis patients and two renal transplant recipients). The patients with CAC had a longer duration of total dialysis, had higher time-integrated serum phosphorus, calcium-phosphate product, intact parathyroid hormone, vitamin B12 levels and the amount of cumulative calcium-containing oral phosphate binders. In a study by Briese et al. [113] of 40 young adults (mean age 23.6 years) who developed ESRD at a mean age of 11.5 years, carotid artery IMT was similar to healthy controls and only 4 (10%) patients had evidence of coronary calcification. Authors noticed that a relatively low rate of cardiac calcification compared to other studies might be explained by a significantly lower amount of prescribed calcium-containing phosphate binders and vitamin D preparations in their patients. Of note, these patients had decreased vascular reactivity and a high rate of LVH (68.2%). As in adults, cardiac valve calcification was also described in children on maintenance dialysis [114]. Evidence of early atherosclerotic changes was found in the study by Nayir et al. [115], who reported on the histopathology of internal iliac artery samples obtained at the time of kidney transplantation in 12 children. Five arteries had fibrous or fibroelastic intimal thickening, medial mucoid ground substance and disruption of the internal elastic lamella. Two of these had microcalcification in the intimal layer; another two demonstrated atheromatous plaques. These abnormalities were associated with longer duration of renal failure. Vascular abnormalities in children develop in parallel with cardiac abnormalities early in the course of CKD and become more severe as end-stage disease is reached [116]. Mitsnefes et al. [54] showed that carotid arteriopathy is present in children after successful renal transplantation and is associated with hypertension. Litwin et al. [117], as a part of the ESCAPE trial, investigated vascular structure and function in children with chronic renal failure and after renal transplantation. The authors demonstrated vascular abnormalities in all patient groups with the most marked changes in the dialysis patients. The degree of arteriopathy in their study was correlated with conventional CVD risk factors such as hypertension and dyslipidemia in pre-dialytic CKD, while in children on dialysis and after transplantation, hyperphosphatemia, hyperparathyroidism and treatment with calcium-containing phosphate binders were determinants of arterial abnormalities. Another important observation in this study was a significantly lower carotid IMT in children post transplant, suggesting that the vascular abnormalities partially regress. Epidemiological studies should determine if vascular abnormalities detected during childhood CKD are associated with future accelerated coronary artery disease. Endothelial dysfunction as a marker of early atherosclerosis in children with CKD Endothelial function can be evaluated by the assessing of endothelial vasodilatation. Healthy vascular endothelium will respond to the release of nitric oxide by vasodilatation. With endothelial injury, the response to nitric oxide is diminished [118], and endothelial-mediated vasodilatation is impaired. Flow-mediated dilation (FMD) of the brachial artery is currently used to measure endothelial function. This test consists of applying obstruction to the flow of the brachial artery by placing the tourniquet or inflated BP cuff for few minutes. The artery dilates due to post-obstruction reactive hyperemia. FMD is calculated based on the difference in the diameter of the brachial artery before obstruction and after it released. Impaired (decreased) FMD predicts CVD morbidity and mortality in adults with ESRD [119]. It has been shown that endothelial dysfunction as determined by impaired FMD is also present in children with advanced renal failure, on chronic dialysis and after renal transplantation [120–122]. Evaluation and treatment recommendations The overall strategy in the prevention of cardiovascular complications in children with CKD is avoidance of long-term dialysis. The goal is to prevent the development and delay the progression of cardiomyopathy and atherosclerosis. The identification of modifiable risk factors and markers of CVD and early intervention should be initiated at the time of mild-to-moderate renal insufficiency, prior to the need for dialysis. Even though kidney transplantation poses a continuous CV risk (hypertension, hyperlipidemia and allograft dysfunction), it eliminates many uremia-related risks, reduces the risk of cardiac death by approximately 80 percent and prolongs the life span by 20–30 years. Thus, kidney transplantation should be the ultimate goal to minimize cardiovascular morbidity and mortality in patients with advanced CKD. For those patients who must have long-term dialysis, the strategy is directly linked to achievement of optimal dialysis outcomes, which include aggressive monitoring and management of hypertension, dyslipidemia, calcium-phosphorus metabolism, anemia, nutrition, systemic inflammation and other dialysis complications. Current recommendations on evaluation of CV abnormalities and treatment of modifiable risk factors in children are based mostly on clinical experience and adult data. Recent K/DOQI Clinical Practice Guidelines for Cardiovascular Disease in Dialysis Patients [42] recommend echocardiographic evaluation for the presence of cardiac disease in children (cardiomyopathy and valvular disease) at the time of initiation of dialysis therapy along with the screening for dyslipidemia, hypertension, anemia and increased Ca×P product. Management of modifiable risks in children with CKD should follow recommendations from the K/DOQI guidelines for treatment of anemia [123] and dyslipidemia [15] and from the Fourth Report on Blood Pressure in Children for Management of Hypertension [124]. Current K/DOQI guidelines for treatment of anemia recommend keeping the hemoglobin level above 11 g/dl by using an appropriate iron therapy and recombinant erythropoietin. In the opinion of the Working Group, there is insufficient evidence to recommend routinely maintaining hemoglobin levels above 13 g/dl or more. K/DOQI recommends evaluation of dyslipidemia in adolescents upon presentation with CKD stage 5 (GFR <15 ml/min/1.73 m2 or on dialysis), at 2–3 months after a change in treatment or other conditions known to cause dyslipidemia and at least annually thereafter. Reasons to repeat lipid measurements after 2–3 months include changes in the kidney replacement therapy modality, treatment with diet or lipid-lowering agents, immunosuppressive agents that affect lipids (e.g., prednisone, cyclosporine or sirolimus) or other changes that may affect plasma lipids. The assessment of dyslipidemia should include a complete fasting lipid profile with total cholesterol, LDL, HDL and triglycerides. The definition of dyslipidemia differs in children and adults. Hyperlipidemia in children is defined as lipid levels greater than the 95th percentile for age and gender. The normative data for lipids in children and adolescents currently used are from the Lipid Research Clinics Program from the NIH published in 1980 and can be found in the 2003 K/DOQI guidelines for management of dyslipidemia in chronic kidney disease [15]. For adolescents with stage 5 CKD and a level of LDL ≥130 mg/dl, K/DOQI recommends treatment to reduce LDL to <130 mg/dl. If LDL is <130 mg/dl, fasting triglycerides ≥200 mg/dl and non-HDL cholesterol (total cholesterol minus HDL) ≥160 mg/dl, treatment should be considered to reduce non-HDL cholesterol to <160 mg/dl. All children with dyslipidemia should follow the recommendations for therapeutic lifestyle changes (TLC), which include diet modification with a reduction in saturated fat intake and increase in fiber intake, and moderate physical activity. Adolescents should be counseled about avoiding smoking. Unfortunately, non-compliance with TLC is one of the major problems in the management of dyslipidemia in adolescents. Pediatric nephrologists must also recognize that appropriate caloric intake, including calories from fat, should be emphasized to avoid malnutrition and ensure normal growth and development, especially in young children. If LDL cholesterol is ≥160 mg/dl and non-HDL cholesterol ≥190 mg/dl, statin therapy is recommended in children older than 10 years. Target blood pressure in children should be lower than the 90th percentile for normal values adjusted for age, gender and height or less than 120/80 mm Hg, whichever is lower. Ambulatory blood pressure monitoring is recommended to assess the circadian rhythm. Angiotensin-converting enzyme inhibitors or angiotensin receptor blockers may be the preferred antihypertensive agents to slow the progression of CKD in children and possibly for regression of LVH. The Working Group on Cardiovascular Disease in Dialysis Patients recommends maintaining calcium and phosphorus levels within the normal range and the Ca×P product <55 mg2/dl2 in children on chronic dialysis [42]. There are no data available to make evidence-based recommendations on management of hyperhomocysteinemia, chronic inflammation or other potential CVD risk factors. Questions(Answers appear following the references) A 14-year-old African-American boy with ESRD secondary to FSGS has been treated with thrice-weekly hemodialysis for 1 year. His pre-dialysis BP is 132/85, hemoglobin is 9.8 g/dl, serum calcium is 10.6 g/dl, serum phosphorus is 8.2 g/dl and iPTH is 412 pg/ml. He is taking a calcium channel blocker to control his hypertension, erythropoietin and iron to control his anemia and calcium carbonate and a vitamin D IV preparation to control his renal bone disease.Which ONE of the following statements would be the BEST therapeutic intervention in this child to minimize the risk of future CV complications?Aggressive treatment of hypertensionMaximize dialysis treatment to achieve dry weightCorrection of anemiaKidney transplantationReduction of Ca-P productThe patient from the previous question was diagnosed with eccentric LVH.With regard to treatment of his LVH, which ONE of the following actions is MOST LIKELY to lead to reduction of LVH?Adding ACE inhibitor as a second antihypertensive agentIncrease the dose of erythropoietin and assure appropriate iron status to treat anemiaAchievement of dry weightSwitch from hemodialysis to peritoneal dialysisMaximize treatment of secondary hyperparathyroidismThe annual evaluation showed that this patient has LDL cholesterol of 170 mg/dl. You would like to start this patient on atorvastatin.With regard to the use of “statin” in this patient, which ONE of the following statements is correct?Statins are not approved by the United States Food and Drug Administration (USFDA) for use in children and adolescentsGrapefruit is contraindicatedThe dose of atorvastatin should be reduced by 50% compared to dosage in adult patients on dialysis to prevent adverse effects on growth and developmentThe target level for LDL cholesterol in children with CKD should be below 100 mg/dlAtorvastatin therapy will reduce mortality risk in this patientWhich ONE of the following statements concerning development of cardiovascular complications in a child with CKD is NOT correct?Cardiovascular disease is the cause of mortality in approximately 25% of chronically dialyzed childrenCardiac arrest is the most common cardiac cause of death in children with ESRDCardiac and vascular remodeling might develop in children during early stages of CKDCorrection of anemia reduces concentric LVHLVH is the most common cardiac abnormality in children on maintenance dialysisThe patient described in the previous questions is at risk for development of vascular calcification.Which ONE of the following choices does NOT correctly describe the mechanisms of vascular calcification in dialyzed patients?Histo-anatomic variants of calcification include classic atherosclerotic (fibrotic) calcification related to disorders of lipid metabolism, medial arterial calcification, vascular calcifilaxis and cardiac valve calcificationIncreased Ca-P product and hyperphosphatemia are the key drivers of vascular calcification in young patients with CKDInflammatory mechanisms are involved in mediating all stages of atherosclerosis including calcificationIncreased serum level of Fetuin-A promotes Ca-P ion product precipitation1, 25 (OH)2D3 increases calcium deposition in the vascular wall Answers dcbdd	[ "chronic kidney disease", "cardiovascular disease", "children", "esrd" ]	[ "P", "P", "P", "P" ]
Intensive_Care_Med-3-1-2039828	Management of invasive pulmonary aspergillosis in non-neutropenic critically ill patients	During recent years, a rising incidence of invasive pulmonary aspergillosis (IPA) in non-neutropenic critically ill patients has been reported. Critically ill patients are prone to develop disturbances in immunoregulation during their stay in the ICU, which render them more vulnerable for fungal infections. Risk factors such as chronic obstructive pulmonary disease (COPD), prolonged use of steroids, advanced liver disease, chronic renal replacement therapy, near-drowning and diabetes mellitus have been described. Diagnosis of IPA may be difficult and obtaining histo- or cytopathological demonstration of the fungus in order to meet the gold standard for IPA is not always feasible in these patients. Laboratory markers used as a non-invasive diagnostic tool, such as the galactomannan antigen test (GM), 1,3-β-glucan, and Aspergillus PCR, show varying results. Antifungal therapy might be considered in patients with persistent pulmonary infection who exhibit risk factors together with positive cultures or sequentially positive GM and Aspergillus PCR in serum, in whom voriconazole is the drug of choice. The benefit of combination antifungal therapy lacks sufficient evidence so far, but this treatment might be considered in patients with breakthrough infections or refractory disease. Introduction Invasive pulmonary aspergillosis (IPA) has emerged as an important cause of morbidity and mortality in patients receiving intensive chemotherapy, allogeneic stem cell transplantation, and solid organ transplantation. However, during recent years, several reports have described a rising incidence of IPA in critically ill patients admitted to the intensive care unit (ICU), even in the absence of an apparent predisposing immunodeficiency [1–6]. The incidence of IPA in the ICU ranges from 0.3% to as much as 5.8% [2, 3, 6] and it carries an overall mortality rate exceeding 80%, with an attributable mortality ofalmost 20% [4, 5]. This high mortality is at least partially related to difficulties in timely diagnosis, caused by insensitive and non-specific clinical signs and lack of unequivocal diagnostic criteria. In this review, we will describe the pathophysiological mechanisms and risk factors for IPA in non-neutropenic critically ill patients, limitations and advances in the diagnostic process, and alterations in treatment with antifungal therapy. A Medline/PubMed search was performed for all articles about IPA in critically ill patients in relation to risk factors, diagnosis and antifungal therapy. All publication types of human studies in the English language were searched and an extraction of relevant articles was made for the purpose of this narrative review. Pathophysiological mechanisms and risk factors Aspergillus spp. are opportunistic moulds that cause both allergic and invasive syndromes. The genus Aspergillus contains approximately 175 species, only a minority of them have been associated with human disease. Infections are caused mostly by Aspergillusfumigatus; next in line are Aspergillus flavus, Aspergillus terreus, Aspergillus niger andAspergillus nidulans [7]. Aspergillus is found in soil, water, food, and in the air and grows on a wide variety of organic material, such as decaying vegetation. The conidia (spores) are easily aerosolised. The route of transmission is by air. Although exposure is universal, invasive infection occurs almost exclusively in immunocompromised individuals. Infections have frequently been described in patients with haematological malignancies and solid organ transplant recipients, but also in patients undergoing chronic intermittent haemodialysis in whom these infections were associated with hospital construction and/or ventilation systems contaminated with Aspergillus spp. [8]. Even hospital water is a frequently overlooked source of nosocomial aspergillosis [9, 10]. Natural antifungal defence in humans is based on normal mucosal barriers and an intact macrophage and neutrophil function. Alveolar macrophages form the first line of defence against inhaled Aspergillus conidia that reach the alveoli. Macrophages normally are capable of killing the conidia and preventing germination, by releasing cytokines such as tumour necrosis factor (TNF)-α and macrophage inflammatory protein (MIP)-1α [11]. During neutropenia, TNF-α and MIP-1α synthesis is reduced and the conidia can germinate to form hyphae. T-cell mediated acquired immunity also has an important role in protecting against fungal infections, as Aspergillus antigens are able to induce T-helper (Th)-1 and Th-2 type reactivity [12, 13]. Th-1 reactivity is displayed by an increase of interferon-γ and interleukin (IL)-12 and has protective effects against infection. In contrast, Th-2 reactivity is characterised by production of IL-4 and IL-10 and leads to disease progression, at least in a murine model of IPA [12, 14, 15]. Critically ill patients in ICU exhibit a complex change in immune function characterised by deactivation of macrophages and an altered cellular response due to the severity of illness which is also termed “immunoparalysis” [16, 17]. This immunologic derangement might explain why Aspergillus infections are able to develop in critically ill patients who do not display the predisposing classical risk factors [17, 18]. Many other factors will negatively influence the immune function during critical illness, such as (acute) hyperglycaemia [19] and the use of corticosteroids [20–23]. Corticosteroids have profound effects on the distribution and function of neutrophils, monocytes, and lymphocytes and they directly stimulate the growth of Aspergillus fumigatus in vitro possibly via sterol binding proteins in the fungus [21]. In particular intravenous corticosteroids treatment in patients with chronic obstructive pulmonary disease (COPD), is associated with a rising incidence of IPA [1, 24–26]. Also broad-spectrum antibiotics, which affect the distribution of normal flora, have been described as a risk factor [27]. However, not every critically ill patient in the ICU is at risk for developing invasive fungal infections. Apparently, other specific – patient-related – predisposing conditions seem to be associated with the development of IPA, in which COPD and other chronic lung disease [1–4, 28–32], diabetes mellitus [2, 31, 33], acute liver failure/advanced liver cirrhosis [2–4, 34], chronic renal failure [8, 35], and near drowning [4, 36–38] have been described (Table 1). In these, mainly retrospective, studies a mean in-hospital mortality of 80% was found in patients with highly suspected or proven IPA in the presence of at least one of these underlying conditions, despite antifungal therapy. Remarkably, patients who were suspected to be colonised only with Aspergillus spp. (i. e. no signs of pulmonary infection) demonstrated even a high in-hospital mortality rate [3, 31], which might suggest that colonisation should be considered as a potentially important finding. Table 1Risk factors for IPA in non-neutropenic critically ill patients in the ICURisk factorReferenceCOPD in combination with prolonged corticosteroid use [1, 24–26]High-dose systemic corticosteroids > 3 weeks (e. g. prednisone equivalent > 20 mg/day) [2, 31, 33]Chronic renal failure with RRT [8, 35]Liver cirrhosis/acute hepatic failure [2–4, 34]Near-drowning [4, 36–38]Diabetes mellitus [2, 3, 31, 33]COPD, Chronic obstructive pulmonary disease; RRT, renal replacement therapy Diagnostic features A few years ago, a consensus for standard definitions and diagnosis for invasive fungal infections in immunocompromised patients with cancer and recipients of haematopoietic stem cell transplants was established by the European Organization for Research and Treatment of Cancer and the Mycoses Study Group (EORTC/MSG) in which three levels of probability for invasive fungal infections are proposed: “proven”, “probable” and “possible” [39]. However, these guidelines are intended for use in clinical trials and for epidemiologic research and, moreover, are focused on patients with classical risk factors such as neutropenia, malignancies and after haematopoietic stem cell transplantation, and it may not be possible to extrapolate them to the non-neutropenic critically ill patient. The diagnosis of IPA in non-neutropenic critically ill patients is difficult because signs and symptoms are non-specific, and the initiation of additional diagnostic examinations is often delayed because of a low clinical suspicion. For the timing of suspicion of IPA in these patients, the combination of persistent or rapid developing infiltrative abnormalities on thoracic imaging and/or a persistent pulmonary infection despite broad spectrum antibiotics accompanied by one or more predisposing conditions (Table 1) might be the moment for triggering further diagnostic exploration. Although histopathological evidence of IPA is defined as the gold standard due to a very high tropism for blood vessels [39], trans-bronchial biopsy or surgical lung biopsy via mini-thoracotomy may not justify the risk of this invasive procedure in critically ill patients on mechanical ventilation with, sometimes, severe bleeding diathesis. Therefore, the diagnostic process will contain thoracic imaging and microbiological examination by means of direct microscopy and culture of sputum or broncho-alveolar lavage fluid (BALF). Excluding the possibility of contamination during the pre-analytical phase of a sample, isolation of Aspergillus spp. in the respiratory tract may represent three clinical situations: (1) evidence of current disease, (2) true colonisation, or (3) a marker for the future development of invasive disease. In the immunocompetent host, cultures of Aspergillus in respiratory secretions are usually a result of colonisation [32]; in the immunocompromised host, however, they may indicate invasive disease [6, 7, 39]. The positive predictive value of cultures in general is as high as 80–90% [2, 6, 40], although in some groups of patients (e. g. after lung transplantation) specificity is much lower [41]. Sensitivity of cultures in the diagnosis of IPA is poor [2, 6, 40]. Direct microscopic examination of sputum or BAL, stained with specific fluorescent stain for chitin (a fungal cell wall component) is easy to perform, rapid to read, and improves the sensitivity of microbiological examination [40, 42]. Direct visualisation of the hyphae makes it possible to discriminate between septate (e. g. Aspergillus, Fusarium and Scedosporium) and non-septate (e. g. Mucorales) moulds. Ubiquitous moulds of the order Mucorales cause serious infection in immunocompromised patients and, in contrast to Aspergillus, Fusarium and Scedosporium spp., they are not susceptible to voriconazole, which is an important factor when choosing a pre-emptive antifungal drug. The extra value of a positive culture is that growing of the fungus enables identification and susceptibility testing to antifungal drugs. This is important in view of the antifungal resistance of – for example – Aspergillus fumigatus isolates to voriconazole [43]. Nonetheless, reliance on microscopy and/or culture alone results in substantial underdiagnosis due to the low sensitivity. Fibre-bronchoscopy with inspection of the tracheobronchial tree, sampling of deep airway secretions and BAL can be helpful; the macroscopic finding of ulcerative lesions and/or pseudomembranes together with a positive microscopy and/or culture is highly suggestive for Aspergillus-related tracheobronchitis. Chest computerised tomography (CT) has proved to be an important tool for the diagnosis of IPA in neutropenic, severely immunocompromised patients, even in the absence of evident lesions on a conventional chest X-ray. Radiological findings might include nodules with rapid growth and/or cavitations. A `halo sign' (a pulmonary mass surrounded by a zone of lower attenuation with ground-glass opacification produced by adjacent haemorrhage) and/or the `air crescent sign' (crescentic radiolucencies around a nodular area of consolidation) may be present [44–50]. The frequency of the halo sign in patients with IPA is relatively high in the early stages of the disease, but becomes progressively lower with the passage of time [51]. Combining the halo sign and the air crescent sign, the sensitivity for IPA is more than 80% as specificity reaches 60–98% [47]. However, thoracic imaging in mechanically ventilated ICU patients is less helpful due to many confounding factors such as atelectasis and, sometimes major, pleural effusions. A lower sensitivity (5–24%) of the halo sign and air crescent sign in non-neutropenic patients has been reported in the literature [3, 6, 52]. Due to the high tropism for blood vessels, IPA might be complicated by localisations in the central nervous system. Therefore, in patients with documented or highly suspected IPA, CT scanning or magnetic resonance imaging (MRI) of the brain should be considered to exclude dissemination to the brain. In the past decade, non-invasive diagnostic tests, serological and molecular, have focused on the detection of surrogate markers for Aspergillus spp., such as the galactomannan (GM) antigen, 1,3-β-glucan and the detection of Aspergillus DNA by PCR. GM is a major Aspergillus cell-wall component that is released during the growth phase of the fungus, and detection of GM would be indicative for invasive disease [53]. Many studies have been done in order to investigate the value of the commercial Platelia Aspergillus assay (BioRad™, Marnes-La-Coquette, France) as a diagnostic tool for IPA, but mainly in patients with haematological malignancies [54–57]. The specificity of the GM assay for diagnosing IPA is at least 85%, as demonstrated by these studies, but the sensitivity of the assay varied considerably between 29% and 100% depending on the cut-off value. The most important finding of these studies was that in around two-third of patients, circulating antigen could be detected at a mean of 8 days before a probable diagnosis was made by a combination of radiographic findings and Aspergillus isolation [53, 54]. In a recent meta-analysis, 27 studies were included regarding the value of the GM serum assay for surveillance of IPA in high-risk patients. The median sensitivity for proven cases was 71% (specificity 89%); for proven or probable cases, median sensitivity and specificity were 61% and 93%, respectively [58]. Specificity increased to 95% using a cut-off value of 1.5 in cases with proven or probable IPA. Because GM is a water-soluble carbohydrate, it can also be detected in BALF. Although the Platelia ELISA (enzyme-linked immunosorbent assay) is not validated for detection of GM in this fluid, there is an increasing tendency to use these samples for diagnosis of IPA. In small clinical studies among patients with haematological malignancies and in solid organ transplant recipients, the sensitivity of the GM EIA (enzyme immuno-assay) applied to BALF ranges from 85% to 100% with a high index cut-off (> 1.5) to define positivity [59–62]. However, there are several clinical circumstances that might influence the diagnostic performance of the GM antigen test in either serum or BALF. First, the false-positive reactivity which might be caused by gastro-intestinal translocation of fungal GM from contaminated food or drink, as demonstrated in small children [63], and the use of the intravenous antibiotics piperacillin–tazobactam and amoxicillin–clavulanic acid, which is associated with serum ELISA reactivity in patients without evidence of IPA [64–67]. Second, an important factor that affects the release of GM antigens is antifungal drug therapy. Different animal and human studies have shown decreased sensitivity of the GM assay when (prophylactic) antifungal drugs were used [63, 68, 69]. In several prospective studies that assessed the performance of antigen detection, patients received antifungal prophylaxis with itraconazole, which might have a significant effect on the sensitivity of the assay [53, 54]. Third, when the Platelia Aspergillus ELISA kit was launched in Europe about a decade ago, another cut-off serum ratio was recommended than at present. Over the past years, several studies suggested lower cut-off values, ranging from 0.5 to 1 [39, 53, 54, 70–72]. It is clear that alterations in cut-off level will change the performance of the assay. Finally, the studies show that monitoring GM levels is crucial in order to diagnose (and eventually monitor treatment outcome) correctly, which means that the assay has to be performed twice weekly, preferably on receipt of the specimen. In critically ill patients without classical risk factors for IPA, the diagnostic value of the GM assay has been investigated only in one retrospective study and demonstrated a sensitivity of only 53% in patients with proven or probable IPA (cut-off value 1.0) [3]. Thus, it has to be stressed that the available data from patients with (haematological) malignancies and after solid organ transplantation can not be extrapolated to the critically ill patient in general. In the meantime, due to lack of more reliable, non-invasive diagnostic tests, the GM assay could be used as an additive tool in the diagnostic work-up of IPA. The 1,3-β-glucan is a cell wall component of many filamentous fungi and yeasts, including Aspergillus spp. and Candida spp. Reproducible assay results, with high specificity and a high positive predictive value, demonstrated that use of an assay to detect serum 1,3-β-glucan derived from fungal cell walls is a useful diagnostic adjunct for invasive fungal infection [73]. In addition, false-positive tests have been found in patients after haemodialysis, cardiopulmonary bypass surgery, high-dose immunoglobulin treatment, and after exposure to glucan-containing gauze [33]. Furthermore, in a recent small prospective study among ICU patients, serum glucan levels did not appear to be specific for fungal infections, as serum glucan levels were also elevated in bacterial infections [74]. Hence, the usefulness of 1,3-β-glucan in the diagnosis of IPA has to be further evaluated. Amplification of nucleic acid by PCR technology for the diagnosis of IPA is being increasingly studied. It can be applied to serum and BAL specimens [75–79]. Experience is limited to patients with haematological malignancies. White et al. evaluated the performance of a real-time PCR in whole blood in a group of patients with haematological malignancies and showed sensitivity of 92.3% and specificity of 94.6% for the diagnosis of IPA with good agreement of the GM ELISA [80]. They concluded that a negative PCR obtained twice weekly allowed a wait-and-see approach concerning starting antifungal treatment. However, comparable to the GM antigen test, there are a number of factors that potentially have an impact upon the clinical sensitivity of PCR. The magnitude of the quantitative PCR signal falls with antifungal therapy, thereby causing false-negative PCR results [81] while the (transient) colonising presence of Aspergillus in the respiratory tract may suggest a low positive predictive value [78]. Furthermore, patients at risk for IPA are often prescribed a multitude of drugs and fluids, all of which may act as non-specific inhibitors of the PCR. For example, anticoagulants inhibit PCR, thereby limiting its sensitivity [82]. One might conclude that the use of GM antigen test, 1,3-β-glucan, and Aspergillus PCR as serological and molecular markers cannot be advocated for routine use in critically ill patients, and caution is warranted in the interpretation of positive test results in patients without a clinical suspicion of pulmonary infection as well as negative test results in patients with persisting pneumonia. However, the finding of sequentially positive GM tests in serum or BALF – using higher cut-off values – together with a positive Aspergillus PCR, in a patient with persisting pulmonary infection who carries one or more risk factors, is highly indicative for IPA and might justify treatment with antifungal therapy. Antifungal therapy There have been important developments in antifungal drugs in the past few years, although amphotericin B deoxycholate has been the standard therapy for IPA for decades. However, multiple studies have now established not only its lack of efficacy due to an increasing antifungal resistance but also demonstrated unacceptable toxicity of this compound, in particular nephrotoxicity [83, 84]. Continuous infusion of amphotericin B deoxycholate over 24 h may reduce its nephrotoxicity [85, 86], although the efficacy may be reduced due to lower peak serum levels [87]. The use of lipid formulations may also reduce toxicity and have been studied extensively for empirical use in febrile neutropenia with the same efficacy rate as conventional amphotericin B [88]. However, lipid formulations are more expensive, and the initial use of higher doses does not improve efficacy and is associated with greater toxicity than lower doses, which suggests that high doses may not be routinely warranted. Among the triazoles, itraconazole has activity against Aspergillus, but its clinical utility in critically ill patients with IPA has been limited by drug interactions and toxicity as well as erratic bioavailability of the oral suspensions [89]. Furthermore, strains of Aspergillusfumigatus resistant to itraconazole have already been described [90]. A large multicentre randomised trial established that voriconazole provides higher response rates and better survival than amphotericin B in the treatment of “probable or proven” IPA among patients with haematological diseases with fewer drug-related adverse events [91]. As a result of this study, voriconazole is increasingly recommended as initial therapy for IPA [92]. Voriconazole is available for intravenous and oral use. It is rapidly absorbed within 2 h after oral administration and the oral bioavailability is over 90% [93]. Clearance is hepatic via N-oxidation by the hepatic cytochrome P450 (CYP) isoenzymes, CYP2C19, CYP2C9 and CYP3A4 [93], which makes the potential for drug interactions considerable. For instance, voriconazole considerably reduces the clearance of intravenous midazolam [94], and fatal interactions with highly active antiretroviral therapy (HAART) have been described [95]. As voriconazole has limited aqueous solubility, the intravenous form includes the solvent vehicle sulfobutylether beta cyclodextrin sodium [96]. The clearance of sulfobutylether beta cyclodextrin sodium is linearly related to creatinine clearance and accumulation has been described in subjects with moderate to severe renal impairment [96], although in animal experiments the frequency of acute toxicity of sulfobutylether beta cyclodextrin sodium is low. Target organs for toxic effects are the kidney and liver, causing obstruction of renal tubules and necrosis in the liver respectively [96]. Because of this potential toxicity, it is recommended to treat patients with moderate to severe renal failure and who are on renal replacement therapy only with the oral form of voriconazole, if feasible. However, in critically ill patients safe oral administration of drugs is difficult to accomplish as, for instance, gastric reflux, gastro-intestinal bleeding and impaired function of the intestine are frequent co-morbidities, leading to potentially insufficient intestinal absorption of the drug. Furthermore, it is important to recognise that some patients may have inadequate levels of the oral drug, particularly if a standard dose of 200 mg twice daily is used rather than the recommended 4 mg/kg twice daily dose which has been studied for the intravenous formulation [97]. At present, no clinical data are available regarding the bioavailability of the oral form of voriconazole in critically ill patients. It might be considered that the fear of potential adverse effects from intravenous solutions does not justify the risk of insufficient treatment by oral solutions. Moreover, limited clinical data showed no obvious toxicity in patients undergoing intermittent haemodialysis who were treated with daily 400–800 mg of intravenous voriconazole during 2 weeks [96], and the pharmacokinetics of voriconazole appears not to be affected by continuous renal replacement therapy (CRRT). On the basis of pharmacokinetics, dose reduction is not recommended in patients receiving CRRT [98, 99]. Regarding the efficacy of voriconazole, there is a potential concern with (long-term) voriconazole therapy as occasional breakthrough infections with yeasts and moulds, with decreased susceptibility to voriconazole, have been reported [43, 100, 101]. Posaconazole is a promising new triazole with broad-spectrum antifungal profile and has shown activity for salvage treatment of IPA in patients who are refractory to or intolerant of conventional therapy. A recent multicentre, prospective study among haematological and non-neutropenic patients with refractory IPA demonstrated a 42% overall success rate for posaconazole recipients versus 26% for control subjects [102]. Posaconazole is, however, only available for oral administration, which makes it probably less applicable in critically ill patients who are susceptible to impaired drug absorption in the digestive tract. Echinocandins are a novel class of parenterally administered semi-synthetic lipopeptides with a pathogen-specific mechanism for non-competitive inhibition of biosynthesis of the fungus cell-wall enzyme complex 1,3-β-D-glucan [103]. The echinocandins have documented in vivo activity against Candida spp. and Aspergillus spp. At present, there are three approved echinocandins, caspofungin, anidulafungin and micafungin, of which particularly caspofungin has demonstrated efficacy for the treatment of IPA. The first clinical trial to document the efficacy of caspofungin was among patients with “proven or probable” IPA who had treatment failure with (liposomal) amphotericin B, itraconazole or voriconazole, or who were intolerant to these antifungal drugs [104]. Caspofungin seems to be as effective as and generally better tolerated than liposomal amphotericin B when given as empirical antifungal therapy in patients with persistent fever and neutropenia [105]. However, prospective randomised clinical trials aimed at the treatment of IPA in (non-neutropenic) critically ill patients are lacking. In general, there is no conclusive evidence that extended-spectrum triazoles are superior to echinocandins or polyenes, or vice versa, for monotherapy of IPA in ICU patients. We need prospective randomised controlled trials to solve this issue in critically ill patients. Table 2 gives an overview of treatment options with antifungal drugs for IPA. Table 2Treatment options with antifungal drugs for IPA in critically ill patients in the ICUSettingFirst choiceAlternativesPrimary therapy of IPAVoriconazole 6 mg/kg q 12 h i. v. on day 1, then 4 mg/kg q 12 h i. v.Liposomal amphotericin B 3-5 mg/kg/day i. v.ororVoriconazole 400 mg q 12 h oral on day 1, then 200 mg q 12 h oral aAmphotericin B deoxycholate 1 mg/kg/day i. v.orCaspofungin 70 mg i. v. on day 1, then 50 mg/day i. v. ba Oral administration is recommended only in patients with intact intestinal absorption; b In patients with moderate to severe hepatic failure, dose reduction is recommended to 35 mg/day i. v. Combination therapy Because the efficacy of antifungal therapy for IPA is poor, with more than 50% of all patients experiencing failure of first-line therapies [106, 107] empirical administration of combination antifungal regiments for proven or probable IPA may be an important strategy to improve outcome. Theoretically, there are several foreseeable advantages of combination therapy, such as a widened spectrum and potency of drug reactivity, more rapid antifungal effect, synergy, and a reduced risk of antifungal resistance [108]. The available antifungal drugs target four different cell functions: cell membrane integrity (polyenes), ergosterol biosynthesis (azoles, allylamines), DNA synthesis (pyrimidine analogues) and cell-wall integrity (echinocandins). Although antifungal drugs are targeted against specific cell functions, many drugs also have pleiotropic effects that may inhibit other elements of fungal homeostasis [108]. For instance, azoles inhibit many cytochrome-dependent enzymes of the fungal respiration chain and amphotericin B generates oxidative species that damage fungal mitochondrial function and enhance macrophage fungal killing [109]. These subtle effects could be enhanced when one antifungal is applied together with a second drug, resulting in synergy. Conversely, the combination could act antagonistically, e. g. when one antifungal agent affects the targets of the other one. Only a few clinical studies with small numbers of patients have tried to address the need for combination antifungal therapy [110–113] . These studies have important limitations because they lack a control group and because of many other uncontrolled factors including the choice of combination, the duration of therapy, concomitant antibacterial or antiviral treatment, the lack of adequate follow-up to estimate relapses, and no discrimination between primary and sequential therapy. Results from those studies should therefore be viewed with caution. Large, adequately powered, prospective clinical trials are needed but they might be difficult to perform because of inadequate enrolment, differences in providing benefits, difficulty in documenting fungal infection, controversy regarding endpoints, the lack of surrogate markers to correlate in-vitro evidence to outcome prognosis, and the associated costs [114]. Nevertheless, combination antifungal therapy might be considered for certain clinical conditions such as refractory disease or breakthrough infections [115, 116]. Conclusion Recent data indicate that IPA may be an underestimated opportunistic fungal infection in critically ill patients, even in the absence of severe pre-existent immunological disorders, and carries a high mortality rate. A decrease in immune function or dysregulation of the immune system due to the severity of illness, together with specific underlying risk factors such as COPD, diabetes mellitus, chronic renal replacement therapy, advanced liver disease, and long-term use of steroids, might explain the relatively high occurrence of IPA among these patients. The high mortality rate is partially related to difficulties in timely diagnosis because of non-specific signs and symptoms, low clinical suspicion, and time delay due to high risks for invasive procedures to obtain histopathological evidence for diagnosing IPA. The presence of a persistent pulmonary infection despite broad-spectrum antibiotics or abnormal thoracic imaging by CT scanning together with one of these risk factors should trigger further diagnostic exploration by collecting respiratory secretions and/or laboratory markers. Meeting the gold standard alone should not be the threshold for starting antifungal therapy, considering the high mortality rate. Invasive infection in patients with negative cultures might be supported by positive serological and molecular markers such as galactomannan antigen testing and Aspergillus PCR, which requires at least two sequentially positive samples. Antifungal therapy might be considered when persistent pneumonia with positive cultures for Aspergillus spp. or sequentially positive GM and Aspergillus PCR are present, accompanied with one of those risk factors for which voriconazole appears to be the first-line treatment. Combination therapy might be considered in breakthrough infections with moulds or yeasts or in refractory disease, although clear evidence is lacking.	[ "invasive pulmonary aspergillosis", "non-neutropenic critically ill", "antifungal therapy", "serological markers" ]	[ "P", "P", "P", "R" ]
Eur_J_Pediatr-4-1-2254659	What is new in iron overload?	Children with severe chronic hemolytic anemia or congenital erythroblastopenia are transfusion dependent. Long-term transfusion therapy prolongs life but results in a toxic accumulation of iron in the organs. The human body cannot actively eliminate excess iron. Therefore, the use of a chelating agent is required to promote excretion of iron. So far, iron chelation has been done by subcutaneous infusion of deferoxamine given over 10 h, 5–6 days per week. Compliance is poor and chelation often insufficient. Ferritin measurements and sometimes liver biopsies are used to evaluate the iron burden in the body. At the present time, new iron chelators that can be given orally are available. Furthermore, magnetic resonance imaging (MRI) assessment of tissue iron is a noninvasive and highly reproducible method, which is able to quantitate organ iron burden. In conclusion, iron overload can be measured more accurately with noninvasive methods such as MRI. Deferasirox is a once-daily oral therapy for treating transfusional iron overload, which improves patient compliance and quality of life. Introduction Children suffering from severe chronic hemolytic anemia such as β-thalassemia, or congenital erythroblastopenia are at risk of developing severe iron overload. Regular blood transfusions are potentially life saving. However, every single transfusion of one unit of blood brings 200 mg of iron into the body of the recipient. Patients receiving 3 U every 4 weeks (39 U per year) will accumulate 7.8 g of iron by the end of the year, in addition to the amount absorbed by the gut. The human body has no mechanism for active excretion of excess iron. Repeated transfusions lead to elevated iron levels that result in major organ damage. Excess iron is deposited largely in the liver, spleen, myocardium, and several endocrine organs. In young patients, growth and maturation is slowed, and in all age groups, liver disease, diabetes mellitus, and other complications develop. Once the body’s storage capacity is exceeded, free iron catalyses the formation of highly reactive hydroxyl radicals, which cause membrane damage and protein denaturation. This process leads to tissue damage and ultimately to significant morbidity and mortality [3]. Measuring iron levels (Table 1) Table 1Comparison of methods for evaluating iron overload InvasiveCostLimitationValidationFerritinNoLowInfection/inflammationYesLICYesMedianFibrosis/cirrhosisYesSQUIDNoHighNot easily availableYesMRINoHighDedicated personYesLIC liver iron content, SQUID superconducting quantum interference device, MRI magnetic resonance imaging There are several methods of assessing body iron load. As 90% of excess iron is deposited in the liver, most techniques focus on measuring liver iron levels, and it is widely accepted that liver iron content (LIC) provides an accurate measure of whole-body iron concentration. The invasive nature of the liver biopsy means that other markers such as serum ferritin levels are frequently employed. Other approaches using biomagnetic susceptometry and magnetic resonance imaging (MRI) are also being assessed in order to identify an accurate, low-risk, and convenient approach to assessing patient iron status. Ferritin Serum ferritin is easily measured using a commercially available kit. A ferritin constantly below 2,500 mg/l has been shown to reduce the risk of cardiac complications, but a target value of 1,000 mg/l is recommended. Factors such as inflammation, ascorbate status, and hepatitis can affect serum ferritin levels. Therefore, results should be interpreted with caution and a trend in the evolution of serial measurements is a better index than day-to-day variation. In patients with transfusion-dependent diseases, chelation should be initiated after 10–20 blood transfusions or when ferritin level rises above 1,000 mg/l. Liver biopsy LIC measurement is the most accurate method for assessing body iron (Fig. 1). It also provides information about the severity of the liver disease [8]. Liver samples can be obtained by percutaneous or transjugular access. LIC measurement from a needle biopsy has been associated with a coefficient of variation of less than 10% in livers with no advanced disease. However, it has been shown that variability can be much higher in fibrotic and cirrhotic livers. Values of LIC are expressed as milligrams of iron per gram of dry liver weight. Fig. 1Liver biopsy (Perls stain): iron deposits are in blue. Courtesy of B. Turlin, Pathology Department, University Hospital of Rennes Liver biopsy is an invasive technique that is associated with some pain and at risk of hemorrhage and infection. It is not indicated for routine assessment. Superconducting quantum interference device (SQUID) SQUID is capable of measuring very small changes in magnetic flux. Iron stored as ferritin and hemosiderin is the only relevant paramagnetic material in the human body. The magnitude of paramagnetic response is directly related to the amount of iron in a certain volume of tissue. This is a noninvasive method, with a linear correlation with LIC assessed by biopsy. However, equipment availability is extremely limited [6]. Magnetic resonance imaging (MRI) MRI measures tissue iron concentration indirectly by detecting the paramagnetic influences of storage iron (ferritin and hemosiderin) on proton resonance behavior. With MRI, the transmitted signal is a microwave, which excites water protons in the body to higher magnetic energy states. As these water protons relax back to the unexcited state, they emit microwaves that are received and interpreted by the scanner. Iron deposits act like little magnets when placed in a strong magnetic field, disrupting coherence among the protons and darkening the image more quickly (Fig. 2) [13]. LIC determined using MRI shows excellent correlation with that obtained from liver biopsy. Furthermore, MRI has the ability to evaluate the entire organ and gives more accurate measurement of LIC, particularly in patients with heterogeneous iron content (Fig. 3). It requires a dedicated imaging method and equipment operator [13]. Fig. 2Gradient echo images of liver collected by J. Wood [13] at four different echo times. The top four images were collected from a patient having a liver iron of 6 mg/g. The bottom four images were collected from a normal volunteer. All images darken as the echo time (TE) lengthens, but the iron-heavy tissue darkens faster. The half-life of this process is called T2* and the rate is called R2* (R2* = 1,000/T2)Fig. 3Gradient echo imaging illustrating discordant iron loading of the liver and heart. a Heavy liver iron loading (dark tissue) with heart sparing image. b Heavy cardiac iron loading with no liver deposition. From [13] Iron chelation therapy Within 1–2 years of initiation of regular blood transfusions, iron starts to accumulate in the body. The use of a chelating agent then becomes mandatory [6]. At the present time, three different therapies are available (Table 2). Table 2Comparison between iron chelators (adapted from [4])CharacteristicsDeferoxamine (Desferal)Deferiprone (Ferriprox)Deferasirox (Exjade)Route of administrationSubcutaneous or intravenousOralOralHalf-life20 min2–3 h8–16 hRoutes of iron excretionUrine/stoolUrinestoolDose range20–60 mg/kg per day50–60 mg/kg per day20–30 mg/kg/dayGuidelines for monitoring therapyAudiometry and eye exams annuallyComplete blood count weekly; ALT level monthly for first 3–6 months then every 6 monthsSerum creatinine level monthly; ALT level monthlySerum ferritinSerum ferritinSerum ferritinAssessment of liver iron annuallyAssessment of liver iron annuallyAssessment of liver iron annuallyAssessment of cardiac iron annually after 10 years of ageAssessment of cardiac iron annually after 10 years of ageAssessment of cardiac iron annually after 10 years of ageAdvantagesLong-term experienceOrally activeOrally activeEffective in maintaining normal or near-normal iron storesSafety profile well establishedOnce-daily administrationReversal of cardiac disease with intensive therapyEnhanced removal of cardiac ironDemonstrated equivalency to deferoxamine at higher dosesMay be combined with deferiproneMay be combined with deferoxamineTrials in several hematologic disordersDisadvantagesRequires parenteral infusionMay not achieve negative iron balance in all patients at 75 mg/kg per dayLimited long-term dataEar, eye, bone toxicityRisk of agranulocytosisNeed to monitor renal functionPoor complianceMay not achieve negative iron balance in all patients at highest dose Deferoxamine mesylate (Desferal) is the current standard for iron chelation therapy. The long-term efficacy of this approach has been demonstrated, limiting organ damage and preventing premature death. Deferoxamine has a very short half-life in plasma (5–10 min). Therefore, deferoxamine is administered parenterally, either subcutaneously 8–12 h a day at least 5 days a week, or intravenously, through a Port-a-Cath for continuous infusion. Low compliance with the prolonged subcutaneous administration is the main reason of ineffective treatment. Toxic side effects such as auditory, ocular, and neurotoxic abnormalities are also drawbacks in the therapy. Deferiprone (Ferriprox) is one of the hundreds of oral chelators that have been tested. It was shown to have comparable efficacy to subcutaneous deferoxamine. It has been used as a monotherapy or in combination with deferoxamine [7, 9]. It increases urinary excretion of iron and is more effective than deferoxamine in the removal of excess iron from the heart, as shown from MRI T2 studies [1, 9, 10, 11]. Deferiprone and its iron complex are cleared from the plasma with a half-life of 47–134 min. Therefore, it should be taken three times a day. The major toxic side effects are agranulocytosis, musculoskeletal and joint pains, gastric intolerance, and zinc deficiency. Those side effects are considered to be reversible. The combination of deferoxamine and deferiprone is most effective when used in the following way: deferiprone (80–110 mg/kg per day) during the day and deferoxamine (40–60 mg/kg per day) at least 3 nights a week. No new toxic effects have been reported for the combination. Deferasirox (Exjade) has been tested recently in clinical trials [2, 3, 12]. Its long half-life makes it suitable for once-daily oral administration. Efficacy at a dose of 20–30 mg/kg per day is at least equivalent to that of deferoxamine (≥35 mg/kg per day 5 days weekly) in a subgroup of patients with higher hepatic iron burdens. The most common toxic side effects are skin rash, increase in transaminases and creatinine, and gastrointestinal symptoms such as nausea, vomiting, diarrhea, and abdominal pain. The skin rash and gastrointestinal symptoms tend to decrease spontaneously with time and an increase in creatinine may require a dose adjustment. The increase in transaminases is usually linked to an insufficient dose of deferasirox and an increase in the hepatic iron overload. A higher dose of deferasirox will result in an improvement in transaminases. Discussion Chronic iron overload due to blood transfusions leads to significant morbidity and early mortality unless adequate chelation therapy is administered. Deferoxamine is the reference chelation therapy, but compliance to the treatment is often poor because it must be administered by prolonged subcutaneous or intravenous infusion. Oral chelators are now available: Deferasirox has the advantage of being administered once daily, with a low toxicity profile. Furthermore, its cost effectiveness has also been demonstrated [5]. MRI has emerged as the dominant technique by which to evaluate iron overload because of its sensitivity, reproducibility, and ability to image multiple organs in the body during a single imaging session.	[ "iron overload", "chelators", "diagnosis" ]	[ "P", "P", "U" ]
Ann_Biomed_Eng-4-1-2413127	Biomechanics of Traumatic Brain Injury: Influences of the Morphologic Heterogeneities of the Cerebral Cortex	Traumatic brain injury (TBI) can be caused by accidents and often leads to permanent health issues or even death. Brain injury criteria are used for assessing the probability of TBI, if a certain mechanical load is applied. The currently used injury criteria in the automotive industry are based on global head kinematics. New methods, based on finite element modeling, use brain injury criteria at lower scale levels, e.g., tissue-based injury criteria. However, most current computational head models lack the anatomical details of the cerebrum. To investigate the influence of the morphologic heterogeneities of the cerebral cortex, a numerical model of a representative part of the cerebral cortex with a detailed geometry has been developed. Several different geometries containing gyri and sulci have been developed for this model. Also, a homogeneous geometry has been made to analyze the relative importance of the heterogeneities. The loading conditions are based on a computational head model simulation. The results of this model indicate that the heterogeneities have an influence on the equivalent stress. The maximum equivalent stress in the heterogeneous models is increased by a factor of about 1.3–1.9 with respect to the homogeneous model, whereas the mean equivalent stress is increased by at most 10%. This implies that tissue-based injury criteria may not be accurately applied to most computational head models used nowadays, which do not account for sulci and gyri. Introduction The brain is often one of the most seriously injured parts of the human body in case of a road traffic crash situation.1,21,41 The incidence rate and mortality rate in Europe are estimated to be 235 and 15.4 per 100,000 of the population per year, respectively.41 Traumatic brain injury (TBI) is therefore considered as a widespread problem. Understanding the mechanisms inducing TBI is necessary for reducing the number of occurrences, e.g., by developing more appropriate protective systems and diagnostic tools. Brain injury criteria are used for the assessment of the probability of TBI for certain mechanical loading conditions. The most commonly used injury criterion in the automotive industry is the Head Injury Criterion (HIC).16,43 It is developed to predict TBI resulting from a translational acceleration of the head. One of the drawbacks of the HIC is that it is based on global kinematic data to predict TBI, whereas actual brain damage is caused at the cellular level as a consequence of tissue strains and stresses.35 Furthermore, it is based on experimental data, in which only anterior–posterior contact loading has been applied to human cadavers, not accounting for angular accelerations of the head. For a better approximation of the relation between TBI and a mechanical load, more advanced methods have been developed. For instance, three-dimensional finite element (FE) head models have been developed to predict brain injury.2,6,7,9,17,22,38,42,45,47 With these numerical head models, different injury mechanisms and loading conditions can be distinguished. However, in these models, the heterogeneous anatomy of the cerebrum is usually represented by a relatively homogeneous geometry. A comparison between the homogeneous geometry of a typical finite element head model and the complex structure of a real brain is given in Fig. 1. The main function of the heterogeneous morphology is to increase the cortical surface in order to obtain a more complex level of the brain functions.34 The most recent numerical head models include ventricles and the invaginations of the dura mater, but none include the convolutions of the cerebral cortex. Consequently, the stresses and strains that are predicted from these models likely do not represent actual tissue stresses and strains, at least in the cortex. Therefore, although tissue-based injury criteria may be used, their accuracy is expected to be limited. This might prohibit the direct use of tissue-based injury criteria. Such criteria predict injury at the tissue level and are based on in vitro and in vivo experiments.3,4,10,11,13,31,32 For a direct application of tissue-based injury criteria in a computational head model, a more detailed description of the biomechanical behavior of the cerebrum may be required, which can be achieved by including its morphologic heterogeneities in these models. A few two-dimensional FE models of the brain containing the convolutions of the cerebral cortex have been described in literature. Miller et al.30 compared different modeling techniques for the relative motion between the brain and the cranium. Nishimoto and Murakami33 developed a model to investigate the relation between brain injury and the HIC. However, these models have not been developed with the purpose of investigating the local biomechanics at the level of these convolutions. No conclusions have been drawn from these studies on the biomechanical influence of the heterogeneities of the cerebral cortex, due to the limited spatial resolution of the mesh.Figure 1(a) Numerical head model developed by Claessens.6,7 (b) Lateral view of the human brain. Adapted from Welker et al.44 Physical experiments have been conducted in several studies to investigate the biomechanical consequences of the heterogeneities of the cerebrum.14 In a study by Bradshaw et al.,5 a gel-filled chamber that represented the brain and skull in a coronal plane including the falx cerebri and the sulci of the cerebral cortex was subjected to a rotation with a peak acceleration of approximately 7800 rad s−2. An increase of the maximum principle strain in the cerebral cortex due to the sulci was found. The aim of this study is to investigate the biomechanical influences of the morphologic heterogeneities in the cerebral cortex. To achieve this, several two-dimensional FE models with detailed geometries of a part of the cerebral cortex have been developed. Also, an FE model with a homogeneous morphology of the cortex has been made. The loading conditions are based on simulations with a computational head model as used by Brands et al.6 The results of the simulations of the heterogeneous models will be compared to those of the homogeneous model. Methods In this study, plane strain models of small sections of the cerebrum are made using the FE code Abaqus 6.6-1 (HKS, Providence, USA). An explicit time integration is used, anticipating a dynamic load with a high magnitude and a short duration. The time increments are limited by the stability condition, which is determined in the global estimator function in Abaqus. Geometries To investigate the influence of the heterogeneities of the cerebral cortex, a homogeneous model and three heterogeneous models have been developed. The heterogeneous models, which are shown in Figs. 2a, 2c, and 2d have detailed geometries of a small part of the cerebrum, including also a part of the cerebrospinal fluid (CSF). The cranium is modeled by a boundary constraint, as will be detailed further on. Since the dura mater and the arachnoid are connected to the inside of the cranium in the region that is modeled,29,34 it is assumed that they can be ignored for this situation. The pia mater, which is a thin and delicate membrane covering the brain,29,34 is also not included, since it is expected to have no mechanical influence for the used loading conditions. The same assumption is used for the arachnoid trabeculae, which extend from the arachnoid to the pia mater and are less existent inside the sulci.34Figure 2(a) Heterogeneous geometry 1 and (b) its spatial discretization. (c) Heterogeneous geometry 2. (d) Heterogeneous geometry 3. (e) Homogeneous geometry The first geometry has one narrow sulcus on the right hand side and a small part of a sulcus on the left hand side. The second geometry contains two deeper and wider sulci than the other two geometries. The third geometry consists of one vertical sulcus and one partly horizontal sulcus, where horizontal and vertical refer to the x- and y-direction, respectively. These geometries, which represent typical stylized shapes of the cerebral cortex, are based on the topological studies by Mai et al.27 The left and right boundaries of the models are chosen to be periodic, i.e., the internal geometries near the opposite boundaries match. The periodicity of the boundary conditions will be explained further on. The models do not distinguish between gray (cerebral cortex) and white matter. In Fig. 2e, the homogeneous model is shown. Similar to the heterogeneous models, it also consists of CSF and brain tissue, but it does not contain any gyri and sulci. The outer dimensions of each model are 32 mm by 24 mm. The meshes consist of bi-linear, quadrilateral, reduced integration elements with hourglass control. The heterogeneous models also contain a small number of triangular elements. The total number of elements of the heterogeneous models ranges from 4243 to 4533 elements. The homogeneous geometry consists of 3072 elements. Material Properties For the material properties of the CSF, a nearly incompressible, low shear modulus elastic solid has been assumed, since the shear stress in the brain tissue due to the applied loading conditions is estimated to be about a factor 104 higher than that in the CSF. The material properties are listed in Table 1. The shear modulus of CSF is estimated from the loading conditions that are described further on by using in which G is the elastic shear modulus, η is the viscosity, γ is the estimated shear strain, and is the estimated shear rate. Because two different loading conditions have been used, also two different estimates for the CSF shear modulus have been used. However, with these shear moduli being much lower than that of the brain tissue, the exact values of these estimates do not affect the outcome of this study. The bulk modulus is obtained from literature.2,47Table 1Linear material parameters Bulk modulus (GPa)Shear modulus (Pa)Time constant (s)CSF2.20.036a∞0.12b∞Brain tissue2.5182.9∞98840.00013835.50.012231.20.3567.14.623.6112.12.7954.3aShear modulus in case of loading condition AbShear modulus in case of loading condition B The material properties of the brain tissue are described by a non-linear viscoelastic constitutive model that has been developed by Hrapko et al.19 This model was found to accurately describe the response of brain tissue for large deformations in both shear and compression. This model is extended here to account for compressibility. The constitutive model consists of an elastic part, denoted by the subscript ‘e’, and a (deviatoric) viscoelastic part, denoted by the subscript ‘ve’, with N viscoelastic modes. The total Cauchy stress tensor is written asin which the superscripts ‘h’ and ‘d’ denote the hydrostatic and the deviatoric part, respectively. For simplicity, the subscript i indicating the number of the viscoelastic mode will be omitted from this point on. The hydrostatic part of Eq. (1) is defined aswhere K is the bulk modulus and is the volume change ratio. The deviatoric elastic mode describes a non-linear response to the deformation gradient tensor F, which is given bywhere G∞ is the elastic shear modulus, is the isochoric part of the Finger tensor B, and and are the first and second invariant of the isochoric Finger tensor respectively. A, C, and b are fitting parameters describing the non-linearity of the elastic response. The third term on the right hand side of Eq. (1) consists of the summation of the viscoelastic modes. The deformation gradient tensor F is partitioned into an elastic deformation gradient tensor and a viscous deformation gradient tensor by assuming multiplicative decomposition25,36: The decomposition involves a fictitious intermediate state, which could exist after application of merely the viscous deformation gradient tensor This is the stress-free state, which after application of the elastic deformation tensor transforms into the final state. The third term on the right hand side of Eq. (1) describes the viscoelastic contribution to the stress as follows:with G the shear modulus, the isochoric part of the elastic Finger tensor and a a fitting parameter. The viscous deformation is assumed to be volume-invariant, i.e., and The viscous rate of deformation tensor is calculated from the flow rule aswhere the dynamic viscosity η is a function of the scalar equivalent stress measure for which the Ellis model is adopted:with subscripts 0 and ∞ denoting the initial and infinite values, respectively. The initial value for the viscosity is defined as η0 = Gλ, whereas the infinite viscosity is defined as η∞ = kη0. Here, λ refers to the time constant. Although differences between the material properties of the gray and white matter may exist, these differences are not well characterized. Therefore, no distinction between gray and white matter has been made in this study, except for the investigation of the influence of varying the material properties of gray matter with respect to those of white matter (see the Discussion and Conclusions). For simulating a head impact situation representative of road traffic accidents, an extra viscoelastic mode with a smaller time constant has been added to the behavior as characterized by Hrapko et al.19 The extra mode18 is based on the experimental data from Hrapko and co-workers in combination with the data by Shen et al.39 The linear material properties are listed in Table 1. The values of the non-linear viscoelastic parameters are shown in Table 2.Table 2Non-linear material parameters for brain tissueElasticViscousA = 0.73τ0 = 9.7 PaC = 15.6n = 1.65a = 1k = 0.39b = 1 Boundary Conditions The boundary conditions have been chosen such that they represent the biomechanical influences of the surroundings on the cerebral cortex model. Figure 3 shows the labeling of the corner nodes and the boundaries. The symbols x and y denote the components of position vector with respect to a Cartesian vector basis whereas u and v are the components of the displacement vector with respect to this basis.Figure 3Labeling of corner nodes and boundaries The Young’s modulus of the cranium is much higher than that of brain tissue.19,46 Still, in a contact loading situation of the head the deformation of the skull is important, because it initiates strain waves in the brain tissue. In this study, however, only inertial loading of the head is considered and therefore the cranium is assumed to be rigid. The cranium is incorporated in the boundary condition at Γ3. Because of the low shear modulus of the CSF, the influence of the rigid constraint associated with the cranium at boundary Γ3 in the x-direction can be neglected. Provided no rotation of the model occurs, the constraint equation for all nodes on boundary Γ3 iswith vs the vertical displacement of the skull. The boundaries Γ2 and Γ4 are subjected to periodic boundary conditions23: These constraints imply that throughout the deformation process the shapes of the opposite boundaries, Γ2 and Γ4, remain identical to each other, while the tractions on opposite boundaries are opposite to satisfy stress continuity, which can be written aswith the Cauchy stress tensor and the unit outward normal vector of boundary Γi. The lower boundary, Γ1, of the brain tissue in the model lies adjacent to brain tissue in neighboring regions. Therefore, boundary Γ1 has to be constrained accordingly. The applied constraint on Γ1 is obtained by tying all nodal displacements on Γ1 to a linear interpolation between the displacements of corner nodes C1 and C2. For any node on boundary Γ1, this results inwith the subscript 0 denoting the initial configuration. The displacements of corner nodes C1 and C2 are prescribed and calculated from the applied loading conditions. The loading conditions of the cerebral cortex model (micro-level in Fig. 4) are based on the loading conditions that have been used by Brands et al.6 for a three-dimensional numerical head model (macro-level in Fig. 4). In that model, an eccentric rotation has been applied to the skull to simulate an angular head acceleration around the neck-shoulder joint in the sagittal plane in the anterior–posterior direction. The eccentricity has been chosen to represent a typical neck length. The axis of rotation has been positioned at 155 mm below the anatomical origin, i.e., the ear hole projected to the sagittal plane. The rotation of the head model consists of two successive sine functions that describe the angular acceleration:Figure 4The loading conditions of the cerebral cortex model (micro-level) are derived from the region of interest in a parasagittal cross-section (15 mm offset from the midsagittal plane) of the head model (macro-level). Shown at the macro-level is the equivalent stress field of the head model at 10 ms In Eqs. (12) and (13), the angular acceleration is given in rad s−2. The loading conditions are applied to the cerebral cortex model by means of body forces. In all integration points of the elements in the model, a non-uniform body force is imposed that reversely simulates the inertial forces:in which represents the distributed load per unit of volume, ρ is mass density, t is time, and refers to the acceleration in the x-direction that is represented by these body forces. Note that for the head model the loading conditions contain an angular component, whereas the cerebral cortex model uses translational loading conditions. Because only a small part of the head is modeled and because of the small rotation of the head model with a maximum of 4°, the loading of the cerebral cortex model is assumed to be translational in x-direction only. The loading conditions of the cerebral cortex model, i.e., the representative accelerations are calculated from the head model (from the region indicated in Fig. 4) in two different approaches: In the first approach, the input accelerations of the head model are used to define the loading condition of the cerebral cortex model. This approach will be referred to as loading condition A.The translational acceleration can be calculated usingwith the angular acceleration, which is defined by Eqs. (12) and (13), and r the radius from the axis of rotation (neck-shoulder) in the head model to a point in the region of interest. The radius r is a function of the y-position in the cerebral cortex model. It varies between r(0) = 0.251 m at boundary Γ1 to r(0.024) = 0.275 m at Γ3. The accelerations at Γ1 and Γ3 are depicted in Fig. 5a. All other accelerations are interpolated linearly between these two boundaries, thereby creating a gradient across the height of the model. The acceleration gradient is important for the resulting shear stresses. Figure 5b shows the acceleration profile of the cerebral cortex model. The accelerations are used to calculate the body forces as a function of both time and y-position.The disadvantage of this loading condition is that a spatially constant acceleration gradient is assumed and therefore it does not account for the influence of the geometry of the cranium. To account for the geometry of the head, another loading condition has been developed that is described next.The second approach, loading condition B, uses output accelerations from a global head model simulation as the input of the cerebral cortex model. For this, a modified version of the head model, as used by Brands et al.,6 has been employed in the simulation code Madymo, in which the constitutive model for brain tissue by Hrapko et al.19 has been implemented. The accelerations obtained from the region inside the box in Fig. 4 from the head model are imposed on the cerebral cortex model. Hence, the influence of the geometry of the head is modeled indirectly by means of an acceleration profile that is obtained from the head model.The displacements of the brain tissue in the head model in the field of interest are almost entirely in the x-direction justifying the assumption of inertial loading (of the cerebral cortex model) in the x-direction only. In Fig. 6, the acceleration profiles as a function of the y-position are shown at 5, 10, and 20 ms. Similar to loading condition A, the accelerations are used to calculate the body forces as a function of both time and y-position.Figure 5Loading condition A. (a) Acceleration at the upper and lower boundary of the cerebral cortex model. (b) Acceleration profiles at different timesFigure 6Loading condition B: displacement (top) and acceleration (bottom) profiles derived from the output of the head model In order to quantify the influence of the morphologic heterogeneities, the equivalent stress is used, in which is the deviatoric part of the Cauchy stress tensor The equivalent stress is chosen, because the simulations are based on an angular acceleration of the head, in which deviatoric stresses are considered to be the most important.30 The maximum principal strain is considered important as well with respect to diffuse axonal brain injury.3,30 Therefore, also the maximum principal logarithmic strain is used to quantify the influence of the morphologic heterogeneities. Results Figure 7 depicts the development in time of the equivalent stress fields for the homogeneous model (top row) and the heterogeneous models from the simulation with loading condition A. Stress concentrations are present in the heterogeneous models at the surface of the brain tissue between two gyri at 5, 10, and 20 ms. Near boundary Γ1, all heterogeneous models have lower equivalent stresses compared to the homogeneous model at 20 ms.Figure 7The equivalent stress fields as a result of loading condition A In order to obtain a good comparison of the results for all geometries during the complete simulation time, the maximum and mean equivalent stress from the simulations with loading condition A are shown in Fig. 8 as a function of time. It shows the stresses in the brain tissue only. It can be noticed that the heterogeneous models have a higher maximum equivalent stress than the homogeneous model. Among the heterogeneous configurations, geometry 1 causes a noticeably lower maximum equivalent stress of 112 Pa compared to geometries 2 and 3, with a maximum equivalent stress of approximately 156 Pa. The large maximum equivalent stress in heterogeneous geometry 2 lasts longer than the stresses of the other geometries. The maximum equivalent stress of the homogeneous model reaches a value of 80 Pa. The mean equivalent stresses are nearly the same for all geometries.Figure 8Maximum and mean equivalent stress for the heterogeneous and homogeneous models as a result of loading condition A To investigate the influence of the heterogeneities, the equivalent stress of the cerebral cortex in the heterogeneous models is taken relative to that of the homogeneous model. For the maximum equivalent stress, this will be done by taking the maximum values, whereas for the mean equivalent stress, this will be done by taking the time averaged values. The maximum equivalent stress of the heterogeneous models 1, 2, and 3 is 1.31, 1.84, and 1.83 times higher than the homogeneous model, respectively. The mean equivalent stress of the heterogeneous models 1, 2, and 3 with respect to the homogeneous model is 1.09, 1.08, and 1.10, respectively. The equivalent stress fields obtained with loading condition B are displayed in Fig. 9. During the beginning of the simulation, the equivalent stress fields in the brain tissue are comparable for all models. When the field of higher equivalent stress moves downwards, the heterogeneities result in local peak stress concentrations, which can be seen at 10 ms for geometries 1 and 3. Later on, at 20 ms, the heterogeneous geometries 1 and 3 have less influence on the equivalent stress fields. The differences of geometry 2 with respect to geometries 1 and 3 are a consequence of the deeper sulci in geometry 2.Figure 9The equivalent stress fields as a result of loading condition B The maximum and mean equivalent stress of the cerebral cortex as a function of time obtained with loading condition B is shown in Fig. 10. The maximum equivalent stress is higher for the heterogeneous models than for the homogeneous models, but not for the complete duration of the simulation. After about 10–15 ms, the maximum equivalent stress of the heterogeneous models drops to approximately the same magnitude as the one obtained for the homogeneous model. For the heterogeneous models, the maximum equivalent stress reaches values of approximately 470, 565, and 624 Pa for geometries 1, 2, and 3, respectively. The homogeneous model has a maximum equivalent stress reaching 325 Pa. Also, the moment in time at which the maximum occurs differs from one geometry to the other. The mean equivalent stress values of all the geometries are similar.Figure 10Maximum and mean equivalent stress for the heterogeneous and homogeneous models as a result of loading condition B To quantify the influence of the heterogeneities, the equivalent stress of the brain tissue of the heterogeneous models is taken relative to the homogeneous model in the same manner as described previously for loading condition A. The maximum equivalent stress of the heterogeneous models 1, 2, and 3 has increased by 1.44, 1.74, and 1.92 with respect the homogeneous model, respectively. The mean equivalent stress of the heterogeneous models 1, 2, and 3 is 0.97, 0.99, and again 0.99 relative to the homogeneous model, respectively. The distribution of maximum principal strains for loading condition B at 10 ms is shown in Fig. 11. One can notice that a concentration of maximum principal strains at 10 ms occurs in the same location as the equivalent stress concentration at 10 ms (Fig. 9), both in case of loading condition B. The same method for the quantification of the influence of the heterogeneities is used, but with the maximum principal strain instead of the equivalent stress. For the simulations with loading condition A, the peak maximum principal strain in the brain tissue of the heterogeneous models 1, 2, and 3 has increased with respect to the homogeneous model by 1.22, 1.92, and 1.80, respectively. If loading condition B is used, the increases are 1.43, 1.84, and 1.90, respectively.Figure 11The maximum principal logarithmic strain field as a result of loading condition B at 10 ms Discussion and Conclusions In this study, the influences of the heterogeneities in the cerebral cortex were investigated. This was done with FE models of several different geometries from small detailed parts of the cortex. In a preliminary study, the boundary constraints were tested. The loading conditions were derived from a numerical head model. In order to determine which constraints on the boundaries would represent the surroundings best, a preliminary study was conducted in which several different constraints were applied to boundaries Γ1 and Γ3. The different conditions on boundary Γ3, i.e., rigid constraint or slip-condition in the x-direction, did not contribute to differences in the results of the brain tissue. This was probably caused by the low shear modulus of the CSF. For boundary Γ1, several different boundary conditions were tested and compared to the results of models with the same width, but with twice the height of the models in this study. By comparing the ‘normal’ and the ‘high’ models, the boundary condition at boundary Γ1 that represented adjacent brain tissue could be determined. The periodic boundary condition that was applied to boundaries Γ2 and Γ4, was used because of the assumed periodicity in the cerebral cortex. Since the loading conditions were dynamic they could not be applied to the model by directly imposing a deformation, which would induce boundary effects. For this reason, an indirect deformation was imposed by means of body forces. This approach worked well for the first 20 ms of the simulation, which had a total duration of 30 ms. After 20 ms the deformations of the brain tissue differed from the deformations derived from the head model that were indirectly applied to the model (Figs. 6 and 9). Therefore, only the first 20 ms of the simulation are considered to be realistic. Loading condition A was derived from the acceleration pulse that has been applied to a head model.6,12 The equivalent stress fields of the homogeneous models caused by loading condition A (Fig. 7) showed no similarities to the equivalent stress field of the head model (Fig. 4). Provided that the equivalent stress field of the head model is realistic, loading condition A can be considered unrealistic. Loading condition B was obtained from the resulting accelerations in the region of interest of the head model. The equivalent stress fields of the homogeneous models from simulations with loading condition B (Fig. 9) were approximately similar to the stress fields in the corresponding region of the head model (Fig. 4) during the first 20 ms. Nevertheless, differences in the equivalent stress fields existed at the surface of the cortex. This is due to the CSF layer in the cerebral cortex model being described elastically with a low shear modulus, as opposed to the head model, which contains a relatively stiff CSF/dura layer. In spite of all these obvious differences, by comparing these two loading conditions, it was shown that the different loading conditions have hardly any effect on the relative mean and maximum equivalent stress and the relative peak maximum principal logarithmic strain. Hence, the mechanical influences of the heterogeneities of the cerebral cortex seem to be independent of the loading conditions. The constitutive model for brain tissue was based on experiments on porcine white matter.19 This was used as a substitute for human brain tissue, as it was readily available and it allowed to conduct experiments with a shorter post-mortem time.12,26 Experiments conducted by Prange et al.37 on fresh human brain tissue indicated that its mechanical properties are approximately 30% stiffer than those of fresh porcine brain tissue. Although the constitutive model for brain tissue in the cerebral cortex model was not based on human brain tissue, the geometries of the models were based on the human cerebrum. This assumption is expected to have more effect on the absolute equivalent stress than on the relative equivalent stress, which is dominated by the heterogeneities of the model. Furthermore, no distinction between the mechanical properties of white and gray matter was made, so that the material properties of the cerebral cortex, which consists of gray matter, were based on experiments on white matter, as well. In their results, Prange et al.37 found differences between the material properties of the corpus callosum, i.e., the white matter that connects the two cerebral hemispheres, and the cerebral cortex, but not between the corona radiata, i.e., the white matter that lies inferior to the cerebral cortex, and the cerebral cortex. Nevertheless, in several other studies of head and brain models, different material properties were used for gray and white matter based on the assumption that white matter was more fibrous than gray matter.2,28,30,47 Therefore, simulations of the cerebral cortex model with shear moduli for gray matter ranging from 75% to 125% with respect to the shear moduli white matter were conducted as well. The results of these simulations indicated that the material properties had an effect on the magnitude of the equivalent stress, but not on the regions in which the peak equivalent stress was observed. The equivalent stress of the heterogeneous model relative to the homogeneous model was hardly affected by the different material properties. In this model, the meninges and the blood vessels were considered to have no mechanical influence on the cerebral cortex for the used loading conditions. Although Jin et al.20 suggested that the pia-arachnoid complex can have a mechanical influence on the brain tissue during an impact, it is still not clear if it would affect the influence due to the gyrification of the cerebral cortex. In a study performed by Ho and Kleiven,17 it was found that the vasculature of the brain can be neglected, as far as the mechanical influences are concerned in a head model not containing the gyrification of the cortex. In order to validate this model, a comparison with physical experiments is required. However, data of physical experiments at a typical length scale of this level is rather limited. Parallel to this study, the results of physical experiments, in which brain slices have been accelerated, showed increased equivalent strains near the sulci.24 Furthermore, some studies showed that angular accelerations of the head induce high stress concentrations in and near the cerebral cortex.8,15 Another method of validation can be performed by comparing the results of the model to clinically observed injury. It has been shown that small cortical infarcts exist in diffuse brain injury at the bottom of the sulci.40 This is in accordance with the locations of the high stress and strain regions in the cerebral cortex model. The two loading conditions and the different geometries resulted in different equivalent stress fields. The simulations with loading condition A resulted in a lower mean and maximum equivalent stress compared to the simulations with loading condition B. However, relative to the homogeneous model, it was observed that the equivalent stress was almost independent of the different loading conditions used in this study. The differences between the several heterogeneous geometries had more influence on the relative mean and maximum equivalent stress. The morphologic heterogeneities of the cerebral cortex led to an increase of the maximum equivalent stress by a factor of about 1.3–1.9, depending mostly on the geometry, whereas the relative mean equivalent stress values of all the geometries were 1.1 and 1.0 for loading condition A and B, respectively. Furthermore, the peak maximum principal logarithmic strain was increased by a factor of about 1.2–1.9 due to the morphologic heterogeneities of the cerebral cortex. This is a strong indication that predictions of brain injury obtained from head models with a homogeneous cerebrum should be interpreted with care. To obtain a more accurate assessment of injury, the influence of the morphologic heterogeneities in the cerebral cortex should be accounted for.	[ "traumatic brain injury", "cerebral cortex", "finite element model", "cerebrum", "brain tissue", "inhomogeneities" ]	[ "P", "P", "P", "P", "P", "U" ]
Ann_Biomed_Eng-2-2-1705524	Three-dimensional Numerical Modeling and Computational Fluid Dynamics Simulations to Analyze and Improve Oxygen Availability in the AMC Bioartificial Liver	A numerical model to investigate fluid flow and oxygen (O2) transport and consumption in the AMC-Bioartificial Liver (AMC-BAL) was developed and applied to two representative micro models of the AMC-BAL with two different gas capillary patterns, each combined with two proposed hepatocyte distributions. Parameter studies were performed on each configuration to gain insight in fluid flow, shear stress distribution and oxygen availability in the AMC-BAL. We assessed the function of the internal oxygenator, the effect of changes in hepatocyte oxygen consumption parameters in time and the effect of the change from an experimental to a clinical setting. In addition, different methodologies were studied to improve cellular oxygen availability, i.e. external oxygenation of culture medium, culture medium flow rate, culture gas oxygen content (pO2) and the number of oxygenation capillaries. Standard operating conditions did not adequately provide all hepatocytes in the AMC-BAL with sufficient oxygen to maintain O2 consumption at minimally 90% of maximal uptake rate. Cellular oxygen availability was optimized by increasing the number of gas capillaries and pO2 of the oxygenation gas by a factor two. Pressure drop over the AMC-BAL and maximal shear stresses were low and not considered to be harmful. This information can be used to increase cellular efficiency and may ultimately lead to a more productive AMC-BAL. Introduction Acute liver failure (ALF) is a severe disease with high mortality rates (60–90%). At present, the only efficient therapy is orthotopic liver transplantation (OLT).11 To bridge ALF patients to liver transplantation or to regeneration of the native liver, liver support systems are needed. The most promising bridging method for the treatment of ALF patients are bioartificial liver (BAL) support systems. These systems are extracorporeal devices that are generally comprised of a bioreactor in which living hepatocytes are seeded. In a clinical setting, toxic plasma of ALF patients is perfused through a BAL system and detoxified by viable hepatocytes. A major advantage of BAL systems, as compared to other non-biological liver support systems, is the capacity to provide a full range of metabolic functions to compensate for the complex metabolic disorders seen in ALF. Although many BAL systems have demonstrated their detoxifying capacity or liver-specific functions in in vitro, ex vivo and in phase I clinical studies, no BAL system has shown to significantly improve survival in ALF patients bridged to transplantation in a controlled trial (for review:40). Important issues for the development of an effective BAL system are local plasma flow and oxygen transport. Optimal cell function, i.e. detoxifying capacity, is only obtained when a BAL system is adequately perfused with plasma to enable efficient mass transfer. In this situation, sufficient nutrients (e.g. glucose, fatty acids) and oxygen are supplied, whereas unwanted metabolites (e.g. ammonia) are efficiently detoxified and metabolic end products (e.g. urea, CO2) subsequently removed. In principle, oxygen (O2) transfer is the main limitation in the efficiency of a BAL,3 because of the low solubility of oxygen in plasma and high demand for oxygen of functionally active hepatocytes. The AMC-Bioartificial Liver (AMC-BAL; Academic Medical Center, Amsterdam, The Netherlands, patent No: WO 97/12960) was developed with a special design for on-site oxygenation of hepatocytes and direct contact of plasma to the cells to enhance bidirectional solute exchange. The AMC-BAL has shown promising results in in vitro set-ups, small and large animal ALF models, and finally in a phase I clinical study with ALF patients.13,39 But to increase the efficiency of the AMC-BAL, further optimization of plasma perfusion and oxygenation of hepatocytes may be useful. Numerical techniques and Computational Fluid Dynamics (CFD) simulations are useful tools to gain insight in local flow field and oxygen transport in hollow fiber BALs19,27,30 as well as in other types of hepatocyte systems.16,23,25 These numerical techniques can also be used to determine the optimal working parameters and to further optimize the design of a BAL system. In this respect, we used CFD simulations as a methodological approach to analyze the configuration of the AMC-BAL. We constructed three-dimensional computer models of two representative unit volumes – micro models – of the laboratory-scale AMC-BAL. Fluid flow and oxygen transport and consumption were simulated in these micro models to gain insight in the flow field and local cellular oxygen availability. Numerical parameter studies were performed to assess possible improvements in local oxygen availability of hepatocytes. Several parameters were tested: (1) effect of the internal oxygenator system, (2) pO2 of the oxygenation gas, (3) culture medium flow rate, (4) incorporation of an external oxygenator in the extracorporeal circuit, (5) the number of oxygen capillaries, (6) the effect of plasma perfusion versus normal culture medium, and (7) the effect of changes of hepatocyte oxygen consumption characteristics in time. Finally, we discuss the implication of this information for the optimization of the AMC-BAL into a more efficient bioartificial liver. Materials and methods The internal geometry of the laboratory-scale AMC-BAL is schematically drawn in Fig. 1 in a longitudinal and transverse cross-section. The bioreactor is built from a cylindrical polycarbonate housing (Fig. 1l) in which two pieces of three-dimensional (3D) non-woven polyester matrix mat (Fig. 1e, g) are spirally wound around a massive inner core (Fig. 1i). A space between the two mat segments (Fig. 1f) is left open for an additional hepatocyte seeding port (not shown). In the hydrophilic matrix, high-density hepatocyte culture is possible. Between the matrix windings, hydrophobic gas capillaries (Fig. 1m) are positioned in parallel along the entire bioreactor as an internal oxygenator system to supply additional oxygen to the hepatocytes. Culture gas (95% air, 5% CO2) is perfused through these capillaries. Plasma enters the bioreactor through an inlet port (Fig. 1b) and flows through the inflow zone (Fig. 1j) further passes the hydrophilic matrix mat and through the void inter-capillary space (Fig. 1n). Plasma exits the bioreactor via the outflow zone (Fig. 1k) through the outlet port (Fig. 1d). Plasma and culture gas are counter-current flows. Figure 1.Longitudinal (left) and transverse (right) view of the internal geometry of the AMC-BAL with a – gas outlet; b – gas inlet; c – plasma inlet port; d – plasma outlet port; e – first mat segment; f – interspace; g – second mat segment; h – polyurethane potting to separate gas and fluid compartment; i – inner core; j – inflow zone; k – outflow zone; l – polycarbonate housing; m – gas capillaries; n – inter-capillary space through which plasma flows. The inline and triangular micro models are designated. Computer Model Since the entire BAL geometry is too large and complex to model, three-dimensional unit volumes – micro models – were isolated from the geometry (Fig. 1). Each three-dimensional micro model (Fig. 2) consists of 2 pieces of mat (thickness = 400 μm, height = 1563 μm, length = 50 mm; Fig. 1e, g) separated by a void interspace section (length = 6 mm; Fig. 1f). At both sides of the mat, gas capillaries (outer diameter = 380 μm) and an inter-capillary space are present. The two micro models differ only in their gas capillaries’ position. Either the capillaries are in a rectangular (inline micro model – Fig. 2) or in a triangular (triangular micro model – Fig. 2) pattern. These two configurations are the most distinct that could be isolated as the position of the capillaries relative to each other changes continuously along the course of the spiral mat (Fig. 1). These two micro models were therefore considered to be representative for the entire AMC-BAL. Simulations on both models will allow us to assess the influence of the capillary arrangement. Figure 2.AMC-BAL micro models. Upper left, inline micro model; upper right, triangular micro model (c – capillary wall, M – non-woven matrix mat, f – inter-capillary space); lower left, inline micro model with double number of capillaries; lower right, triangular micro model with double number of capillaries. Both models were created in the modeling software Gambit 2 (Fluent Inc., Sheffield, UK). All dimensions were derived from a laboratory-scale AMC-BAL or were supplied by the manufacturer. The curvature of the mat was neglected and the inflow and outflow zones before and after the mat segments in the AMC-BAL were not included in the micro models. Each standard micro model contains the equivalent capillary wall surface of 1 whole capillary. Since the entire laboratory-scale AMC-BAL contains 300 capillaries, each micro model can thus be considered as 1/300th part of AMC-BAL. Consequently, one AMC-BAL can be regarded as a combination of 300 separate micro models in parallel. Modeling Fluid Flow Theoretical Model To simulate fluid flow, the commercial CFD package Fluent 6.2 (Fluent Inc., Sheffield, UK) was used to numerically solve the steady-state Navier–Stokes equations. Standard pressure discretization scheme and second order upwind momentum discretization scheme were used. Pressure–velocity coupling scheme was SIMPLE. Fluid properties were set to those of culture medium to allow future validation with in vitro experiments. Culture medium was modeled as an incompressible, isothermal, Newtonian fluid composed of 90% water (ρ = 998 kg/m3) and 10% serum (ρ = 1030 kg/m3) with a resulting density of 1001.2 kg/m3. Dynamic viscosity μ of culture medium with 5 g/l bovine serum albumin (BSA) at 37°C was set to 0.691 mPa s.26 Plasma viscosity at 37°C was set to 1.3 mPa s.31 Resistance to Flow of the Non-woven Polyester Mat The non-woven polyester matrix was modeled as an isotropic homogeneous porous zone. A measure for the viscous resistance of the non-woven polyester matrix used in Fluent 6.2 was determined in an experimental set-up by establishing the pressure drop – flow rate relationship. A sample of this matrix fabric sheet was clamped between two connecting tubes. These tubes were placed in a closed circuit with an overflow reservoir, which provided a constant static pressure load over the matrix sample. Pressure difference over the matrix sheet was measured using a calibrated differential pressure transducer (Fuji Electrics FCX, Japan). Flow rate was measured gravimetrically. De-ionized water was used. Static pressure load was set so average velocities in the experiment were in the same order of magnitude as could be expected in the AMC-BAL (1–7 mm/s). A viscous resistance factor of 2.7 ± 0.2 × 1010 m−2 was calculated as an average of 10 measurements. Resistance to flow of hepatocyte cell layers (section “Modeling oxygen transport and consumption”) was modeled using the same viscous resistance factor. Boundary Conditions We assumed that total flow rate (15 ml/min) inside the research scale AMC-BAL is homogenously distributed. In this way, each micro model has the same flow rate of 0.05 ml/min, i.e. 1/300th of the total flow rate. Pressure inlet boundary conditions were used; outlet boundary condition was a zero pressure outflow; the capillary walls were ‘no-slip’ walls and symmetry boundary conditions were used at the side walls of the model. Modeling Oxygen Transport and Consumption Theoretical Model To simulate oxygen transport and consumption, Fluent 6.2 solves the steady-state convection–diffusion-reaction equation (Eq. 1). Discretization scheme was set to QUICK. The transported scalar φ is the local oxygen concentration (vol.%), which is also the product of the oxygen solubility α and the local oxygen partial pressure (pO2) according to Henry’s Law (φ = αpO2). Oxygen solubility α is 3.1385 × 10−5 ml O2/mmHgml fluid in culture medium and 2.855 × 10−5 ml O2/mmHgml fluid in plasma. D is the oxygen diffusion coefficient, which is 2.92 × 10−9 m2/s in culture medium and 2.18 × 10−9 m2/s in plasma.2,17,33,44 Oxygen consumption by the primary porcine hepatocytes in the AMC-BAL was modeled by an additional source term Sφ, which was only implemented in regions that were designated to contain hepatocytes. This source term (Eq. 2) was based on the Michaelis–Menten kinetics of O2 consumption by hepatocytes. As such, O2 uptake is dependent on the local hepatocyte density ρcell and in a non-linear way dependent on the local O2 availability. Michaelis–Menten parameters of primary porcine hepatocytes were obtained from Balis et al.1 Values of culture day 2 were taken as they are characterized by a highly demanding O2 uptake (VM = 0.7286 nmol/s106 cells; KM = 2 mmHg). These values were measured in 2D culture, whereas hepatocyte culture in the AMC-BAL is three-dimensional and at high density. For this reason, a lower O2 consumption may be expected as the specific O2 uptake rate per cell is reported to decrease at increasing cell densities.30 On the other hand, O2 consumption may rise significantly when a metabolic load is applied to the cells.4 As such, the proposed consumption characteristics are justified by assuming these values to represent the most limiting case for oxygen availability. Consequently, attention must be paid when comparing simulation results to other numerical studies.19,23,25 Hepatocyte Distribution in the Micro Models In an in vitro setting, the laboratory scale AMC-BAL is seeded with 1 billion porcine hepatocytes via three different loading ports, while gently rotating the bioreactor to assure a homogeneous cell distribution. As we assumed that one micro model is one 300th part of an entire laboratory scale AMC-BAL, one micro model contained 3.33 million cells. Since no detailed information on the internal hepatocyte distribution in the AMC-BAL is available, two cell distributions are hypothesized. In cell distribution 1, all hepatocytes are homogeneously distributed in the non-woven matrix (resulting hepatocyte density ρcell of 53.7 × 106 cells/ml). In previous macroscopic and microscopic studies of an early prototype AMC-BAL, the majority of the hepatocytes were immobilized within the three-dimensional matrix.14 However, given the undifferentiated method of hepatocyte seeding, it is possible that hepatocytes also adhere to the surface of gas capillaries. Therefore, in cell distribution 2, 50% of the hepatocytes are located in the matrix (ρcell = 31.7 × 106 cells/ml) and 50% in a hepatocyte layer around the capillaries. The hepatocyte layer was set to be 122 μm, with a ρcell of 81.7 × 106 cells/ml. This cell layer thickness is in the range of what is used in other numerical studies19,20 and is consistent with the size of primary porcine hepatocyte 3D cell clusters found in vitro.15 All proposed cell densities were also considered to be realistic.8,21,37 The complete AMC-BAL cannot be modeled by using only one cell distribution or only one micro model. Therefore, each cell distribution was applied to both micro models, leading to four basic configurations in total. All four micro model configurations were assessed independently. With a combination of these micro models, the entire AMC-BAL can be modeled in future studies. O2 Diffusion Constant Through Non-woven Matrix Free O2 diffusion through the mat zone is hindered by the non-woven arrangement of hydrophilic polyester fibers, which are considered to be impermeable to O2. To compensate for this hindered diffusion, a correction factor for the O2 diffusion coefficient in the mat volume was determined. On a microscopic view of the non-woven matrix, the mat porosity was determined to be 91%. A modified two-dimensional inline micro model was constructed with 250 polyester fiber circles randomly located within the boundaries of the mat volume. On this modified inline micro model, the pO2 distribution was simulated using hepatocyte distribution 1 (Fig. 3A). Standard boundary conditions were used and culture flow rate was zero. O2 diffusion coefficient in the mat zone of the modified micro model was the same as for free culture medium, since culture medium occupies the void spaces between the polyester fibers. Subsequently, an analogous simulation was performed on a regular inline micro model, i.e. with the mat modeled as homogeneous medium, with an adjusted O2 diffusion coefficient in the mat zone to match the results of the ‘modified’ model (Fig. 3B). A correction factor of 0.85 (Dmat = 2.48 × 109 m2/s) for O2 diffusion was necessary to take into account the presence of the non-woven polyester matrix mat. This correction factor is in accordance with the equation of Rayleigh (correction factor ∼ ), which has been experimentally verified for various sheets of porous media.29,32,34Figure 3.Determination of the correction factor for O2 diffusion through the non-woven polyester mat. (A) pO2 distribution [mmHg] on modified inline micro model (polyester fibers drawn) under standard boundary conditions, no medium flow; (B) identical pO2 distribution [mmHg] obtained with identical simulation settings, but with mat as homogeneous medium with an adjusted diffusion constant (with a factor of 0.85) to account for hindered diffusion. O2 Diffusion Constant Through Zones Containing Hepatocytes Oxygen can diffuse through both continuous fluid spaces as well as through hepatocytes. In regions seeded with hepatocytes, the resulting “effective” diffusion coefficient Deff will be lower than in free medium, since O2 diffuses more slowly through hepatocytes. The extent to which the diffusivity is reduced will logically depend on the local hepatocyte density. Since experimental data on O2 diffusion constants through hepatocyte cell layers with different cell densities is not readily available, a theoretical approach was chosen. Riley et al.35 developed an empirical relation (Eq. 3) that relates the effective diffusion coefficient Deff to the local free diffusion coefficient D0, the intracellular diffusion coefficient Dcell and the cell volume fraction φ. This equation, based on Monte Carlo simulations, shows a good agreement with available data throughout a wide range of cell volume fractions (0.04 < φ < 0.95). Dcell was set to 0.25 × 10−9 m2/s, which is an average of intracellular O2 diffusion constants in hepatocytes reported by Jones.22 Cell volume fraction φ is the local volume percentage of space that is occupied by hepatocytes and was calculated as the product of the local hepatocyte cell density and the average hepatocytes cell volume.9 Local free diffusion coefficient D0 is dependent on the perfusion fluid, e.g. culture medium or plasma, and on the location of hepatocytes in the model; either in the inter-capillary space or in the mat zone. O2 diffusivities in different regions of both micro models are presented in Table 1. TABLE 1.O2 diffusion constants in different zones of the micro models.Region in the micro modelD0 [10−9 m2/s]ρcell [106 cells/ml]φDeff [10−9 m2/s]Cult. med. PlasmaCult. med.PlasmaNo hepatocytes Inter-capillary space2.922.180012.922.18 In mat2.481.850012.481.85Cell distribution 1 In mat2.481.8553.70.1720.751.871.41Cell distribution 2 In mat2.481.8531.70.1020.852.111.59 In hepatocyte cell layer around capillaries – part in mat zone2.481.8581.70.2620.641.591.22 In hepatocyte cell layer around capillaries – part in inter-capillary space2.92 2.1881.70.2620.641.861.41The ratio of the effective diffusion constant to the free diffusion constant Deff/D0 is calculated from the local free oxygen diffusion constant D0 and the local cell volume fraction φ using Eq. 3. The local effective diffusion coefficients Deff in culture medium and plasma per region are presented in the last column. The obtained local diffusion coefficients are in the range of reported values for hepatocytes and other cell types.6,16,18,41,43 Boundary Conditions Oxygen partial pressure (pO2) of incoming culture medium was set to 146.5 mmHg, which was an average of three culture medium pO2 measurements using an ABL505 blood-gas analyzer (Radiometer Copenhagen). Axial and radial O2 transport resistances in gas capillaries were neglected, since the culture gas flow rate through the gas capillaries and their O2 permeability are high.5 As such, O2 supply through gas capillaries was modeled by imposing a constant 150 mmHg pO2 on the capillaries’ outer walls, which corresponded to the pO2 of culture gas used in vitro (95% air, 5% CO2). All other boundary faces have no-flux boundary conditions. Grid Dependency One micro model mesh contained approximately 3.75 million finite volume mesh elements. Further increase in the number of cells rendered identical simulation results. Simulations Overview and Assessment As previously stated, four micro model configurations were used: (a) the inline micro model with cell distribution 1, (b) the inline micro model with cell distribution 2, (c) the triangular micro model with cell distribution 1, and finally (d) the triangular micro model with cell distribution 2. Fifteen case studies (Table 2) were performed on each configuration. TABLE 2.Overview of simulations.CaseFluidpO2 gas (mmHg)pO2 medium (mmHg)Qmedium (ml/min)# capillariesKM (mmHg)Reference case (1) Standard boundary conditionsCulture medium150146.50.0512Effect of the internal oxygenator (2) pO2 gas = 0Culture medium0146.50.0512 (3) pO2 gas = 0; pO2 medium × 2Culture medium02930.0512Increasing oxygen availability (4) pO2 gas × 2Culture medium300146.50.0512 (5) pO2 gas = carbogenCulture medium722146.50.0512 (6) pO2 medium × 2Culture medium1502930.0512 (7) pO2 medium = carbogenCulture medium1507220.0512 (8) Fluid flow rate Qmedium × 2Culture medium150146.50.1012 (9) Fluid flow rate Qmedium × 10Culture medium150146.50.5012 (10) No. capill. × 2Culture medium150146.50.0522 (11) No. capill. × 2; pO2 gas × 2Culture medium300146.50.0522Clinical versus experimental setting (12) PlasmaPlasma150146.50.0512 (13) Plasma; pO2 gas × 2Plasma300146.50.0512Changes of hepatocyte O2 consumption in time (14) KM day 4Culture medium150146.50.0514.75 (15) KM day 5Culture medium150146.50.0517.5Each case is applied to the four basic micro model configurations (inline and triangular micro model each with cell distribution 1 or 2). The first case study was simulated using standard boundary conditions as described in previous paragraphs. This ‘reference case’ study (1) is the internal control for all other case studies with the same micro model configuration. The effect of the internal oxygenation system was assessed by disabling culture gas flow (2), and by assessing the effect of an external oxygenator that oxygenates incoming medium to a doubled pO2 as an alternative for internal oxygenation (3). To increase oxygen availability to the hepatocytes, different strategies were assessed. First, by using a more oxygen rich culture gas, i.e. a doubled pO2 gas (4) or carbogen gas (95% O2) (5). Second, by incorporating an external oxygenator to increase incoming medium pO2 to a doubled (6) or carbogen (7) level. Third, by changing culture medium flow rate to a doubled (8) or 10-fold (9) flow rate. Finally, by doubling the number of capillaries (10) or by combining the double number of capillaries with a doubled culture gas pO2 (11). In cases with a doubled number of capillaries, cell density in the hepatocyte layer around the capillaries was kept constant. This lead to a cell layer thickness of 68 μm, keeping the number of hepatocytes around all capillaries equal to 50% of total, as was initially proposed. In micro models with double number of capillaries the distance between the capillaries is halved (see Fig. 2). In the next case studies, we studied the effect of plasma versus culture medium, since this is the main difference in perfusion in a clinical setting as compared to an experimental setting. Standard boundary conditions (12) and a doubled pO2 gas (13) were used and compared with cases (1) and (4). Finally, time-related changes in oxygen consumption characteristics of hepatocytes were investigated by varying KM values. KM values of culture day 4 (14) and 5 (15) were compared to the reference case (1). According to Balis et al.,1KM values change remarkably throughout the first 5 days, whereas VM values remain constant. We therefore kept VM constant in these simulations. The proposed simulation cases on the four different micro model configurations were assessed by examining pO2 distributions. However, not every simulated pO2 profile can be shown due to brevity reasons. To evaluate the effect of the parameters discussed in the parameter study, the local effective hepatocyte utilization ratioVratio (Eq. 4) was studied.30 Vratio is the ratio of the observed oxygen consumption rate to the maximal hepatocyte oxygen consumption rate, i.e. VM, and ranges between 0 and 1 (asymptotically). A threshold for Vratio of 0.9 was chosen as introduced by Patzer30 and corresponds with a minimal pO2 level of 18 mmHg when KM = 2 mmHg. Consequently, oxygen availability in the AMC-BAL was quantified by determining the percentage of hepatocytes with Vratio > 0.9. Results Fluid Flow and Shear Stress Distribution Figure 4 represents an example of colorimetric contour plots of fluid flow and shear stress distributions in two different micro models with different hepatocyte distributions. In the upper part, velocity magnitudes (m/s, left legend) in a transverse plane midway through the first mat segment are shown for the reference case (case 1) of an inline micro model with hepatocyte distribution 1 (Fig. 4A1) and for a triangular micro model with hepatocyte distribution 2 and with a double number of capillaries (case 10–11) (Fig. 4B1). The contour plots can be mirrored with respect to the horizontal axis as micro models are symmetrical. Figure 4.Colorimetric contour plot of velocity magnitudes (m/s, left legend, upper part 1) and shear stress levels (Pa, right legend, lower part 2) in a transverse plane midway through the first mat segment, in the reference case (case 1) of an inline micro model with hepatocyte distribution 1 (A1 and A2 resp.) and in a triangular micro model with hepatocyte distribution 2 and with double number of capillaries (case 10–11) (B1 and B2 resp.).Figure 5.Colorimetric contour plot of pO2 (mmHg, left legend, upper part 1) and effective hepatocyte utilization ratio Vratio (dimensionless, right legend, lower part 2) in a transverse plane midway through the first mat segment, in the reference case (case 1) of an inline micro model with hepatocyte distribution 1 (A1 and A2 resp.) and in a triangular micro model with hepatocyte distribution 2 and with double number of capillaries and doubled culture gas pO2 (case 11) (B1 and B2 resp.) (Note: B1 different scale compared to A1). Fluid flows in the non-woven matrix mat zone and in the hepatocyte cell layers were orientated axially and are uniform in size (approx. 8 μm/s in Fig. 4A1 and 37 μm/s in Fig. 4B1). In the inter-capillary space, flow velocities have a poiseuille-like (parabolic) profile with maximal velocities of approx. 3.6 mm/s in Fig. 4A1 and 7.8 mm/s in Fig. 4B1. In the interspace between two mat segments, flow lines expand radially due to the sudden increase in local cross-sectional area. However, these flow lines reconverge when entering the second mat segment, where the velocity profile was identical compared to the first mat segment (not shown). Static pressure drop over the entire micro model was 15.7 Pa in the reference case of the inline micro model with hepatocyte distribution 1 (Fig. 4A) and 69.3 Pa for the triangular micro model with hepatocyte distribution 2 and double number of capillaries (Fig. 4B). The lower parts of Fig. 4 (A2 and B2) show corresponding shear stress distributions (Pa, right legend). Maximal shear stresses were generally located near the mat side surfaces and also – in case of cell distribution 2 – at the boundary between the hepatocyte cell layer and inter-capillary space. In Table 3, an overview is given of maximum velocity magnitudes in the inter-capillary space, uniform velocity magnitudes in the mat/hepatocyte layer zone, as well as static pressure losses over the entire micro model and maximal local shear stresses for all simulation cases. Maximum velocities in the inter-capillary space ranged between 3.6 and 45 mm/s, whereas uniform velocities in the mat/hepatocyte layer ranged between 8 and 110 μm/s. Simulation results showed that velocities in the mat zone and in hepatocyte layers were consistently two orders of magnitude smaller than in the inter-capillary space. Velocity profiles and shear stress profiles were similar for the different simulation cases, but differed in magnitude between cases when flow rate was altered or when the internal geometry had been changed, e.g. due to additional gas capillaries and/or hepatocyte layers around the capillaries. Static pressure loss over the entire micro model and local shear stress levels were additionally influenced by the type of fluid, e.g. plasma or culture medium. Static pressure loss over the entire micro model ranged from approx. 16 to 203 Pa. Maximum shear stress ranged between approx. 0.03 and 0.40 Pa. Largest maximum values were reached in hepatocyte distribution 2 with a 10-fold flow rate. TABLE 3.Overview of the maximum velocity in the inter-capillary space, the uniform velocity in the mat/hepatocyte layer zone, the static pressure loss over the entire model and the maximal shear stress for all simulation cases and for different hepatocyte distributions.CasesHepatocyte distr. 1Hepatocyte distr. 2Velocity in mat/hepatocyte cell layer (mm/s) – Maximum velocity in inter-capillary space (mm/s) (1,2,3,4,5,6,7,14,15) Standard flow rate0.0085 – 3.59 0.011 – 4.60 (8) Qmedium × 20.017 – 7.170.022 – 9.20 (9) Qmedium × 100.084 – 35.70.11 – 45.0 (10,11) No capillaries × 20.0195 – 5.690.037 – 7.75 (12,13) Plasma0.0085 – 3.590.011 – 4.60Static pressure loss over micro model (Pa) (1,2,3,4,5,6,7,14,15) Standard flow rate15.720.6 (8) Qmedium × 231.441.1 (9) Qmedium × 10155.7203.3 (10,11) No capillaries × 236.169.3 (12,13) Plasma30.039.3Maximum shear stress (Pa) (1,2,3,4,5,6,7,14,15) Standard flow rate0.0320.041 (8) Qmedium × 20.0640.083 (9) Qmedium × 100.310.40 (10,11) No capillaries × 20.0560.083 (12,13) Plasma0.0570.073Values for the inline and triangular micro model are identical. Importantly, both the inline and triangular micro model had the same flow field, static pressure loss and shear stress distributions within a simulation case with certain boundary conditions and hepatocyte distribution (comparison not shown). Oxygen Transport and Consumption Results of oxygen transport and consumption simulations are presented by means of two contour plots of pO2 and Vratio distributions (Fig. 5). Also, Table 4 presents detailed information for each simulated case and micro model configuration on the percentage of hepatocytes that consume oxygen at a minimal Vratio level of 0.9. TABLE 4.Percentages of total hepatocyte cell amount that attain a Vratio > 0.9.% Hepatocytes with Vratio > 0.9CaseHepatocyte distribution 1Hepatocyte distribution 2Inline (%)Triangular (%)Inline (%)Triangular (%)Mat (%)Capill. (%)Mat (%)Capill. (%)Reference case(1) Standard boundary conditions15.715.828.828.63.354.13.353.5The effect of the internal oxygenator (2) pO2 gas = 01.71.71.71.73.10.43.10.4 (3) pO2 gas = 0; pO2 medium × 26.26.26.35.911.01.710.31.7Increasing oxygen availability (4) pO2 gas × 230.430.250.350.35.195.15.094.9 (5) pO2 gas = carbogen62.464.580.788.361.2100.076.3100.0 (6) pO2 medium × 220.120.234.533.911.857.011.056.5 (7) pO2 medium = carbogen35.535.154.253.440.667.737.569.1 (8) Qmedium × 217.918.131.631.46.656.46.556.0 (9) Qmedium × 1028.428.645.745.622.268.922.168.9 (10) No capillaries × 231.831.856.756.814.1100.014.3100.0 (11) No capillaries × 2; pO2 gas × 261.059.987.584.875.3100.069.9100.0Clinical versus experimental setting (12) Plasma12.012.022.923.02.643.02.643.0 (13) Plasma; pO2 gas × 222.822.940.840.92.978.42.978.3Changes in hepatocyte O2 consumption in time (14) KM day 48.78.718.918.91.136.41.136.4 (15) KM day 55.05.112.913.00.325.30.325.5In the case of hepatocyte distribution 2, distinction is made between the percentage of the total number of hepatocytes in the mat and the percentage of total number of hepatocytes in the cell layers around the capillaries. Reference Case Simulations (Case 1) Figure 5A1 illustrates the pO2 distribution in a transverse plane midway through the first mat segment in the reference case (case 1) for the inline micro model with hepatocyte distribution 1. Highest pO2 (150 mmHg) was found close to the capillary wall. A large pO2 gradient extended radially from the capillary and was steeper at the side of the capillary in the mat zone. At a radial distance of 86 μm from the capillary, pO2 in the mat zone dropped already below 20 mmHg. A region where oxygen was depleted (0 mmHg) is present in the center of the hepatocyte-seeded mat throughout both mat zones. Analogous pO2 distributions were found in cross-sections of the micro model further downstream to the plane of Fig. 5A1, but with overall decreasing average oxygen level. Oxygen content of the perfused culture medium in the inter-capillary space dropped from its initial value of 146.5 mmHg to approx. 60 mmHg midway the first mat segment and further decreased downstream to 28 mmHg at the end of the first mat segment. At the start of the second mat segment, culture medium pO2 in the inter-capillary space was increased to 40 mmHg and again decreased to 22 mmHg at the end of the mat segment. As a standard for cellular oxygen availability, the effective hepatocyte utilization ratio Vratio was calculated (Fig. 5A2). Complying with pO2 distribution, hepatocytes in the center of the mat zone cannot consume oxygen at all (Vratio ∼ 0), whereas hepatocytes closest to the capillaries have enough oxygen available to consume oxygen at near maximum capacity (Vratio ∼ 1). Approximately 16% of all hepatocytes in the model consumed oxygen with Vratio > 0.9 (Table 4). A change to hepatocyte distribution 2 was characterized by a larger radial pO2 gradient around the capillaries and a more extended zone of zero oxygen content in the mat segments. Due to the additional presence of a dense hepatocyte layer around the capillary, pO2 dropped below 20 mmHg from a radial distance of 65 μm from a capillary (compare to 86 μm in hepatocyte distribution 1). Nevertheless, 29% of all hepatocytes attained a Vratio of more than 0.9. Fifty four percent of the hepatocytes in cell layers around the capillaries reached the threshold of Vratio > 0.9, whereas 3% of the hepatocytes in the mat zone. pO2 and Vratio distributions for the reference case of the other three micro model configurations are not shown for brevity reasons. Simulation results also showed that inline and triangular micro models, using the same hepatocyte distribution, showed qualitative and quantitative identical pO2 distributions. The Effect of the Internal Oxygenator (Case 2,3) Disabling culture gas flow (∼pO2 gas = 0 – case 2) led to a reduction of cells with Vratio > 0.9 to approximately 2% for all micro model configurations. Replacing the internal oxygenator by an external oxygenator, which doubles culture medium pO2 (case 3), led to a “Vratio > 0.9” fraction of approx. 6% for all micro model configurations. More specifically, in cell distribution 2, this was true for only 1% and 2% of the hepatocytes around the capillaries, and for 3% and11% of the hepatocytes in the mat zone, for case 2 and 3, respectively. Increasing Oxygen Availability (Case 4–11) Doubling culture gas pO2 (case 4) almost doubled “Vratio > 0.9” percentages (Table 4) from approximately 16% to 30% in micro models with hepatocyte distribution 1; an increase of 93%. A 75% increase (29–50%) was obtained in micro models with cell distribution 2. This increase mainly occurred in the hepatocyte layer around capillaries, i.e. 55% in case 1 to 95% in case 4. A further pO2 gas increase to carbogen level (case 5) led to an approximate quadruple and triple “Vratio > 0.9” fraction compared to the reference case for distribution 1 and 2, respectively. In the latter, hepatocytes in both the mat zone as well as in the cell layers have increased threshold percentages. Furthermore, about 7% more hepatocytes attained the threshold in the triangular capillary arrangement compared to inline when using hepatocyte distribution 2. In distribution 1, inline and triangular configurations rendered quasi-identical results. Doubling culture medium pO2 (case 6) led to a 28% and 20% increase in Vratio percentages for cell distribution 1 and 2, respectively. In the latter, the relative increase was higher in the mat zone as compared to the hepatocyte layers. Further increase of pO2 medium to carbogen level (case 7) led to a 125% and 88% increase in “Vratio > 0.9” percentages compared to the reference case for distribution 1 and 2, respectively. Inline and triangular models rendered identical results under the same hepatocyte distribution. Doubling culture medium flow rate (case 8) resulted in a 14% and 10% increase in “Vratio > 0.9” fraction for distribution 1 and 2, respectively. Further increase of culture medium flow rate up to a 10-fold (case 9) resulted in an 81% and 59% increase for distribution 1 and 2, respectively. Typically in cases with increased flow rate in cell distribution 2, Vratio percentages in the mat zone changed remarkably, whereas in the hepatocyte layers there was only a minor increase. Again, inline and triangular models rendered identical results under the same hepatocyte distribution. A 2-fold increase in the number of capillaries (case 10) resulted in a doubled percentage of cells that attain the Vratio > 0.9 threshold in cell distribution 1 (+102%) and 2 (+98%). In the latter, this increase occurred in the mat zone as well as in the cell layer around the capillaries. The combination of a double number of gas capillaries and a doubled pO2 gas (case 11) increased the threshold percentages to almost quadruple (+280%) for cell distribution 1 and triple (+200%) for cell distribution 2. Again, only minor differences existed between inline and triangular micro models. Figure 5B1 shows the pO2 distribution in a transverse plane midway through the first mat segment in the triangular micro model with cell distribution 2 (case 11). In this case, highest pO2 levels (∼300 mmHg) were only found close to the capillary walls. A very steep radial gradient was noted in the hepatocyte cell layer around the capillaries. pO2 in the center of the mat zone was 7 mmHg. As in the reference case, pO2 distributions of cross-sections further downstream showed analogous results as compared to Fig. 5A1, but with overall decreasing average oxygen level. The distribution of Vratio corresponding to the pO2 distribution of case 11 is depicted in Fig. 5B2. The entire hepatocyte layer (100%) around the capillaries and large parts of the mat zone (70%) had a Vratio of at least 0.9. Consequently in case 11, 85% of all hepatocytes had a Vratio higher than 0.9 (Table 4). Clinical Versus Experimental Setting (Case 12,13) Changing fluid properties from culture medium to plasma resulted in a 24% decrease in Vratio > 0.9 percentages in micro models with cell distribution 1 and a 20% decrease in cell distribution 2 (case 12). Doubling pO2 gas in the clinical setting (case 13) showed a relative rise of 91% and 78% compared to the clinical reference case (case 12) in cell distribution 1 and 2, respectively. Compared to the analog case in in vitro settings (case 4), the change of fluid properties also corresponds with a respective 25% and 19% decrease of the “Vratio > 0.9” fraction. Changes in Hepatocyte O2 Consumption in Time (case 14,15) Changing KM values to the situation on day 4 (case 14) or 5 (case 15) causes slight increases in absolute pO2 level. Also, pO2 gradients are less steep and the oxygen depleted zone in the center of the matrix is slightly smaller (not shown). However, “Vratio > 0.9” fractions decreased 45% and 68% for cell distribution 1 and 34% and 55% for distribution 2 compared to day 2 (case 1) for day 4 and 5, respectively. Discussion Fluid Flow and Shear Stress Distribution Perfusion of a micro model was largely influenced by the presence of the non-woven mat. Although the cross-sectional area of the inter-capillary space and the non-woven mat are roughly the same size, fluid flow in the mat zone was generally two orders of magnitude smaller as compared to the flow in the inter-capillary space. This effect is caused by the higher resistance to flow of the non-woven mat, forcing the majority of fluid flow through the inter-capillary spaces, which have a negligible flow resistance. This also causes the overall pressure loss to be minimal. Since the same hydraulic permeability is used for the hepatocyte layers, fluid velocities there are also in the range of micrometer per second. Consequently, apart from the incorporation of additional gas capillaries (case 10–11), also the presence of hepatocyte layers around the capillaries (cell distribution 2) increased flow velocities and pressure loss in the model as the free cross-sectional area of the inter-capillary space is reduced. Fluid flow simulations for the inline and triangular variant of the different case studies render identical results for velocity profiles, pressure loss and shear stress distributions. This was expected as the micro models consist of the identical geometrical entities, which are only changed in location relative to each other. From a fluid dynamical point of view, we conclude that the change in capillary arrangement along the course of the spiral mat in the AMC-BAL does not influence fluid flow, pressure drop or shear stress distribution. The static pressure loss over one micro model can be regarded as the total pressure drop over the entire AMC-BAL without the inflow and outflow zone (Fig. 1e–g), since the AMC-BAL can be represented by 300 micro models in parallel. This pressure drop, i.e. max. 200 Pa ∼ 1.5 mmHg, is negligible when compared to the pressure losses in the extracorporeal circuit. These low pressure gradients are in accordance with the in vivo situation in the liver lobule, in which pressure drops of approximately 3 mmHg over the sinusoids are normal.7,28 This is considered advantageous for cell culture inside the AMC-BAL. Shear stress was also assessed as it is a possible determinant of cellular damage and reduced metabolic function. Shear stress is directly proportional to the local velocity gradient and the fluid viscosity. Consequently, shear stresses are generally more elevated in cases with higher velocity magnitudes in the inter-capillary space (e.g. in case of increased flow rate, doubled number of capillaries, cell distribution 2) and where fluid viscosity is increased (e.g. when plasma was used instead of culture medium – ‘clinical setting’ – case 12–13). Results show that only hepatocytes located at the side of the mat and at the border of the hepatocyte layers with the inter-capillary space are subjected to a certain level of shear stress. In vivo values of hepatocyte wall shear stress are difficult to obtain, and comparison is therefore difficult. In human, liver sinusoid wall shear stress τ can be calculated as ∼ 1Pa (with static pressure drop Δp ∼ 3 mmHg; sinusoid diameter D ∼ 9 μm; length L ∼ 1 mm;7,28 and considering laminar flow and a cylindrical shape of the sinusoid). In mice sinusoids, shear stress was calculated to be 0.55 Pa.24,27 However, in vivo, hepatocytes are not directly subjected to this shear stress as they are shielded by the sinusoid endothelial lining. As a result, the maximal tolerable level of shear stress is lower than 1 Pa. Tilles et al.38 showed in in vitro studies on rat hepatocytes that hepatocyte function, measured as albumin and urea synthesis rates, was significantly decreased (resp. 2.6- and 1.9-fold) when hepatocytes were directly exposed to shear stresses higher than 0.5 Pa compared to low shear stresses (<0.033 Pa). Under standard conditions, maximum shear stress values in the AMC-BAL do not exceed 0.04 Pa. Additionally, the majority of the hepatocytes are not subjected to significant shear stresses. Small zones with increased shear stress exist but maximal values are considered to be acceptable and not detrimental for the viability and function of hepatocytes. However, shear stresses may increase up to 0.4 Pa when flow rate is increased 10-fold (case 9). Moreover, should plasma be used in this case, instead of culture medium, maximal shear stress values may then reach up to 0.71 Pa (not simulated; obtained by data extrapolation), which is in the critical range (>0.5 Pa38). Additionally, very high flow rates, e.g. 10-times the normal culture flow rates, may cause detachment of hepatocytes and therefore should not be used in spite of any possible increase in local O2 availability. Subsequently, case 9 (culture medium flow rate × 10) can already be discarded in the search for optimal O2 availability, based on fluid dynamical grounds. Oxygen Transport and Consumption We assessed the oxygen availability in different case studies and in different micro model configurations. In this paragraph, the results are discussed per case. Reference Case Simulations (Case 1) Highest oxygen concentrations in the mat zone are located near the gas capillaries and near the border of the mat and the inter-capillary space. In zones with limited flow velocities, cellular oxygen supply is clearly dependent on diffusion, which causes large radial pO2 gradients. On the other hand, diffusive O2 transport from the culture medium towards the hepatocytes also causes an axial gradient in culture medium oxygen content. As a consequence, culture medium only oxygenates the mat near the very start of the first mat segment and near the side of the mat. The depth of O2 penetration into the mat decreases rapidly further downstream, since as the majority of oxygen content in the convective flow is already consumed in the first part of the micro model. Nevertheless, oxygen level in the culture medium is not totally depleted at the end of the micro model as the fluid transit time is not sufficiently long for all oxygen to diffuse to the hepatocytes. In contrast, the O2 concentration in the gas capillaries remains constant along the micro model axis, so the zone around the capillaries, which is oxygenated has a constant radius throughout the entire micro model. This pattern of oxygenation leads to about 80% higher Vratio percentages in hepatocyte distribution 2 as compared to 1, since more hepatocytes are close to the gas capillaries in cell distribution 2. So despite steeper pO2 gradients near the gas capillaries and consequently a larger zone of depleted oxygen in the mat zone in case of hepatocyte distribution 2, this cell distribution leads to significantly higher Vratio percentages. However, this effect is related to the maximal oxygen diffusion distance, which is limited by the Michaelis–Menten oxygen consumption parameters, local hepatocyte density and local diffusion coefficient in general and by the specific local gas capillary and culture medium pO2 in this reference case. Since regions which are oxygenated by one gas capillary in particular do not overlap or influence each other in the reference case simulations, diffusive oxygen supply by the capillaries is independent of relative capillary location. Convective oxygen transport is also identical as flow distribution is irrespective of capillary placement (section “Fluid flow and shear stress distribution”). Consequently, inline and triangular capillary pattern give identical results concerning oxygen transport and consumption in these reference cases when the same hepatocyte distribution is used, as is confirmed by simulation results. Concerning oxygen transport, we can conclude that the oxygen supply in the AMC-BAL is not fully adequate to provide all hepatocytes with sufficient oxygen to maintain O2 consumption at minimally 90% of the maximal uptake rate (Vratio > 0.9). Given that the AMC-BAL can be modeled as a combination of micro model configurations, 16–28% of the hepatocytes in the AMC-BAL are adequately oxygenated. In other words, between 72% and 84% of the hepatocytes are subjected to a pO2 of less than 18 mmHg under the current conditions. Normal physiological pO2 values in the liver sinusoids are between 70 mmHg in the periportal area and 20 mmHg in the pericentral area.42 But it is difficult to determine the effect on viability and function of hepatocytes below this threshold. Studies indicate that hepatocyte respiration becomes impaired below 1–2 mmHg pO2.10 With respect to these values, further investigation of simulation results showed that under current conditions 67% and 52% of the hepatocytes in cell distribution 1 and 2 have a pO2 below 2 mmHg and will suffer from hypoxia. The remaining 17–20% of the hepatocytes are likely to be characterized by shifts to more anaerobic metabolic processes.36 However, it should be noted that the low “Vratio > 0.9” percentages are largely influenced by the very stringent hepatocyte oxygen consumption ratios used in this numerical model. Also, hepatocytes which are subjected to low oxygen levels for a prolonged period of time might show changes in oxygen consumption characteristics (VM, KM) to compensate for the lack of oxygen. This could account for account for higher oxygen availability in vitro. The use of different (VM, KM)-values throughout the model to address this behavior can be implemented in future work. Nevertheless, further increase in O2 availability may prove useful to increase the cell viability and subsequently the efficiency of the AMC-BAL. The Effect of the Internal Oxygenator (Case 2,3) Disabling the internal oxygenator leads to detrimental results and reduces the percentage of adequately oxygenated hepatocytes to virtually zero. This confirms the importance of the internal oxygenator and illustrates the small contribution of the incoming convective oxygen flux by culture medium. Incoming culture medium flow adds only 0.17 nmol O2 per second to the micro model, while maximal oxygen uptake VM equals 0.7286 nmol/s*106 cells. As such, theoretically no more than 0.2 × 106 cells, i.e. 6% of the total amount of hepatocytes per micro model, would be able to consume at maximal uptake rate. In reality, only 1.7% of the hepatocytes can consume at a Vratio > 0.9 due to the presence of pO2 gradients. Replacing the internal oxygenator by an external oxygenator, which doubles the flow pO2 and thus the convective oxygen flux, will only increase the “Vratio > 0.9” percentage to about 6%. Since this is still considerably less than in the reference cases, external oxygenation – even at elevated pO2 levels – is a not suitable replacement for the internal oxygenation system as the gas capillaries constitute the main oxygen supply. As such, the internal oxygenator is an essential part of the AMC-BAL. Increasing Oxygen Availability (Case 4–11) In the reference case models, merely 16–28% of the hepatocytes are adequately supplied with oxygen to consume at a Vratio > 0.9, depending on hepatocyte distribution. We therefore assessed several strategies to increase overall oxygen availability. Culture medium flow rate (case 8,9) does not have a considerable effect on cellular oxygen availability. A 2-fold increase in flow rate resulted in an increase in regions with Vratio > 0.9 of only 14%. This was expected as the convective oxygen supply is only a minor contributor to the overall oxygen supply. Consequently, doubling the medium pO2 (case 6) does not greatly increases the amount of regions with sufficient oxygen either. As already discussed, changes in culture medium properties are reflected in a small increase of regions with Vratio in the mat zone, and more specifically in the first mat segment as oxygen content of the fluid is rapidly decreased downstream. It is interesting to note that doubling the flow rate and doubling the medium pO2 both supply exactly the same amount of oxygen to the micro model: in the latter, oxygen is supplied at increased pO2, whereas in the former oxygen is supplied at standard level but with a higher flow rate. Nevertheless, doubling pO2 medium is preferred compared to increasing flow rate when aiming at a better effective hepatocyte utilization ratio since the doubled flow rate does not give the oxygen enough time to diffuse to the hepatocytes as the residence time before oxygen exits the micro model is cut in half. However, a disadvantage of increasing pO2 in the medium is the requirement of an external oxygenator, which makes the extracorporeal circuit more complex and costly and increases the extracorporeal plasma volume in a clinical setting. Doubling the pO2 of the oxygenation gas (case 4) is much easier to achieve in practice. Importantly, the “Vratio > 0.9” percentage increases significantly to almost a doubled level as compared to the reference case. In comparison, culture medium flow rate has to be increased more than 10-fold (case 9) to achieve similar results as with a 2-fold increase in pO2 of the oxygenation gas. This illustrates how oxygen supply is only minor dependent on fluid flow rate. However, saturation of “Vratio > 0.9” percentages with increased flow rate is not yet achieved at 10-fold flow rate. Nevertheless, a 10-fold flow rate is not preferred because of possible detachment of the hepatocytes and an increased risk of possible shear stress damage to the hepatocytes. The effective hepatocyte utilization ratio can even be further improved. Culture medium pO2 can be set to carbogen level (case 7) or the number of capillaries can be doubled (case 10). On average and for both hepatocyte distributions, these two methods lead to a 2-fold increase in the amount of hepatocytes consuming oxygen at Vratio > 0.9 as compared to the reference case. When both approaches are compared for hepatocyte distribution 1, the results are slightly better when using the carbogen medium pO2. But for hepatocyte distribution 2, the double number of capillaries leads to better results. An increase in the number of capillaries is preferred over carbogen medium perfusion, since carbogen medium perfusion leads to extremely high and possible toxic12pO2 values of almost 720 mmHg. Carbogen medium perfusion also requires an additional external oxygenator. Therefore, increasing the number of capillaries by 2-fold is a more easy and safe method to increase the effective hepatocyte utilization ratio. The results obtained with double capillary numbers are also better when compared to the doubling of the oxygenation gas pO2. In cell distribution 2, this is mainly caused by the redistribution of the hepatocytes located around the capillaries. By increasing the number of capillaries, the hepatocyte layer thickness decreased (section “Modeling oxygen transport and consumption”). As a consequence, the region that a gas capillary can sufficiently oxygenate now spans the entire hepatocyte layer so 100% of the cell layer (Table 4) is sufficiently oxygenated. In cell distribution 1, the total region with Vratio > 0.9 has also increased by two, as each additional capillary supplies an additional (constant) volume of hepatocytes with Vratio > 0.9. In contrast, doubling the oxygenation gas pO2 increases the region surrounding a single capillary with Vratio > 0.9 with a factor of less than two. The ultimate results in improving cellular oxygen availability and effective hepatocyte utilization ratio are found when oxygenation gas is set to carbogen level (case 5) or when a double number of capillaries is used in combination with a doubled oxygenation gas pO2 (case 11). Vratio > 0.9 percentages have quadrupled for cell distribution 1 and tripled for cell distribution 2 as compared to the reference case for both approaches. Simulation results show that in case 11, the respective gains in Vratio > 0.9 regions are cumulated when the case of a double number of capillaries and a doubled pO2 gas are combined. It can also be noted that there is a small but significant difference between the inline and triangular capillary pattern in case 5. When carbogen oxygenation gas is used, the region with Vratio > 0.9 surrounding a capillary has grown to such an extent that it could overlap with the region of surrounding capillaries. If a capillary is present straight opposed to another (inline pattern), this overlap is present. Consequently, the intersection of the two regions only contributes once to the amount of region with Vratio > 0.9. In contrast, when capillaries are in a triangular pattern, the region surrounding the capillary can extent to its fullest, thus increasing the percentage Vratio > 0.9 more compared to the inline micro model. This effect is more pronounced in the cell distribution 2 compared to distribution 1. Given the possible toxicity of the hyperoxic oxygen levels when carbogen oxygenation gas is used (case 5), the preferred method to optimally and maximally increase oxygen availability in the AMC-BAL is to double the number of gas capillaries together with a 2-fold increase in culture gas pO2. We note, however, that the calculated improvements in O2 availability and Vratio > 0.9 percentages do not necessarily translate into identical increases in cell viability or functional activity. Nevertheless, similar trends between simulation results and the in vitro situation are expected, but have to be validated. Clinical Versus Experimental Setting (Case 12,13) An important 20–25% decrease in the amount of hepatocytes with Vratio > 0.9 is caused by the lower diffusion coefficient (−9%) and oxygen solubility (−25%) in plasma as compared to culture medium. As in reference case 1, results are independent of capillary pattern and are higher in cell distribution 2 as compared to cell distribution 1. Doubling the oxygenation gas pO2 in the clinical setting leads to approximately the same relative increase in Vratio percentages as it does in the experimental setting. As such, we conclude that any relative increase in oxygen availability found in the experimental setting by any possible strategy can also be applied to the clinical setting, taking into account that the absolute Vratio > 0.9 fractions are decreased with a constant percentage depending on hepatocyte distribution. Changes in Hepatocyte O2 Consumption in Time (Case 14,15) Higher KM values of day 4 and 5 cause the decrease in oxygen consumption to start at higher oxygen levels compared to lower KM values of day 2. Consequently, pO2 gradients are less steep and oxygen penetrates further into the mat zone or into hepatocyte layers because of the reduced local O2 uptake. Paradoxically, “Vratio > 0.9” fractions have decreased. This is understood when converting the Vratio threshold to the minimal cellular pO2 to which the hepatocytes must be subjected to, using Eq. 4. Whereas in the reference case, hepatocytes are considered to be sufficiently oxygenated when cellular pO2 reaches at least 18 mmHg, threshold values are now approximately 43 and 68 mmHg for day 4 and 5, respectively. Since the overall pO2 level in the latter cases is not equally elevated, Vratio > 0.9 fractions are strongly reduced. Conversely, when applying an identical pO2 threshold instead of Vratio, hepatocyte percentages in case 14 and 15 are increased with 3% and up to 20% compared to day 2. As such, an increase in KM causes average cellular pO2 to increase, but whether there is also an increase in the number of hepatocytes that are sufficiently oxygenated is difficult to determine based on computer simulations, as this depends on the criteria, i.e. minimal Vratio versus minimal pO2, used. Conclusions A numerical model to investigate fluid flow and oxygen transport and consumption in the AMC-BAL was developed and applied to two representative micro models of the AMC-BAL. Two different gas capillary patterns, i.e. ‘inline’ and ‘triangular’, were used and combined with two proposed hepatocyte distributions, leading to four basic configurations in total. Fifteen case studies were performed on each of the configurations in order to gain insight in the fluid flow, shear stress distribution, oxygen availability and effective hepatocyte utilization ratio Vratio of the AMC-BAL and to assess possible strategies to further improve cellular oxygen availability. We found that the AMC-BAL does not provide sufficient oxygen to all hepatocytes to allow them to consume oxygen at 90% of their maximal uptake rate under standard operating conditions. The internal oxygenator is an essential part of the AMC-BAL. Doubling the number of gas capillaries together with a 2-fold increase in the oxygenation gas pO2 was found to be the optimal method to maximally increase O2 availability. Additionally, pressure drop over the AMC-BAL and cellular shear stress levels were found to be low and advantageous to cell culture. The developed model also allowed us to assess the effect of the transition from the in vitro to the clinical setting and the effect of the change of hepatocyte O2 consumption characteristics in time. Since large variations in simulation results between hepatocyte distributions are shown, an assessment of the in vitro hepatocyte distribution in the AMC-BAL is useful. Subsequently, an attempt to validate the numerical model with in vitro experiments should be made. Eventually, adoption of this information may lead to a more efficient and productive AMC-BAL in the near future.	[ "computational fluid dynamics", "fluid flow", "shear stress", "internal oxygenation", "external oxygenator", "acute liver failure", "oxygen partial pressure", "effective hepatocyte utilization ratio", "michaelis–menten oxygen consumption", "hepatocyte bioreactor" ]	[ "P", "P", "P", "P", "P", "P", "P", "P", "P", "R" ]
Dig_Dis_Sci-3-1-1914226	Rectal GIST Presenting as a Submucosal Calculus	This case report presents an incidental finding of a rectal GIST (gastrointestinal stromal tumor) presenting as a submucosal calculus, not previously reported. A 53-year-old man without a significant medical history presented with abdominal pain in the left lower quadrant, and with constipation. Upon rectal examination, a hard submucosal swelling was palpated 4 cm from the anus, at 3 o’clock, in the left rectum wall. X-ray photos, computerized tomography (CT)-scan and a magnetic resonance imaging (MRI) scan clearly showed a calculus. Excision revealed a turnip-like lesion, 3.1×2.3×1.8 cm. Analysis showed it was a rectal GIST, a rare mesenchymal tumor of the gastrointestinal tract, which expressed CD117 (or c-kit, a marker of kit-receptor tyrosine kinase) and CD34. Calcification is not a usual clinicopathological feature of GISTs [1–3], and although a number of rectal GISTs have been reported [4–9], we have found no cases so far of rectal GIST presenting as a submucosal calculus. Case Report A 53-year-old man without significant medical history presented with abdominal pain in the left lower quadrant and constipation. He got up that morning with a cramping abdominal ache that lasted over the day. He was nauseous, had vomited once, and ructus was present. There had been a small amount of hard stool and discomfort during defecation initially, but after the general practitioner had administered laxatives, he had liquid diarrhea. There was no complaint of constipation on a regular basis; this was a new finding. Tenesmus was absent. There was no bleeding or mucous discharge. He was feverish, with a temperature of 38.2°C. On examination, the abdomen was slightly distended. Auscultation showed diminished intestinal peristalsis, without borborygmi, and percussion was tympanic. Upon palpation, the abdomen was supple: there was no reflex rigidity nor guarding, and no rebound tenderness. Upon rectal examination, a hard submucosal swelling was palpated 4 cm from the anus, at 3 o’clock, in the left rectum wall. At the time, this was considered an accidental finding that was to be investigated later. There were no specific deviations in the lab results. Ultrasound investigation displayed some hydronephrosis of the left kidney and a stone (size: 0.13×0.7 cm) of the left ureter. X-ray photos displayed a spherical lesion, projecting in the pelvis and situated in the rectum, and some phleboliths projecting on the left colon, but not the ureter stone (Fig. 1). Proctoscopy was not conclusive, so further investigation was performed. A computerized tomography (CT) scan and a magnetic resonance imaging (MRI) scan also clearly showed the submucosal calculus (Figs. 2 and 3). Fig. 1X-ray photo showing the rectal calculus projecting in the pelvis (thick arrow) above the pubic symphysis and the phleboliths projecting on the left colon (thin arrow)Fig. 2Computerized tomography (CT) scan showing the rectal calculus in the center (arrow)Fig. 3Magnetic resonance imaging (MRI) scan. The arrow shows the rectal calculus in the rectum wallFig. 4Excision of the rectal calculus in the operating roomFig. 5Excision revealed a turnip-like lesion, dimensions 3.1× 2.3×1.8 cm Our patient was diagnosed with abdominal pain and constipation based on a ureter stone and slight hydronephrosis of the left kidney, with coincidental finding of a rectal tumor. He was sent home with adequate analgesia and spontaneously released the ureter stone. An appointment was made for elective excision of the rectal calculus. Transanal excision was performed in the operating room (Fig. 4) and revealed a turnip-like lesion, dimensions 3.1×2.3×1.8 cm (Fig. 5). No normal tissue margins were excised because the tumor was not connected to the surrounding tissues. After surgery, the patient went home and recovered well, without any further complaints or complications. The calculus was analyzed and consisted of fibroid and bony tissue, together with bundles of spindle-shaped cells with cigar-like nuclei. A partially calcified leiomyoma was considered at first, but supplementary immunoperoxidase analysis showed that the spindle-shaped cells were positive for CD117 (c-kit) and CD34. Unfortunately, histopathologic slides showing CD117 staining are not available to us. There was some focally positive smooth muscle actin, and MIB1 expression appeared negative. The mitotic count was low: fewer than five mitoses per 50 HPF. Only then did this analysis lead to the conclusion of rectal GIST, with extended calcification and ossification. It was classified as low-risk GIST because of the size (between 2 and 5 cm) and low mitotic count (Table 1). Table IPrognosis of primary GISTRiskSize (cm)Mitotic count (per 50 HPF)Very low risk<2<5Low risk2–5<5Intermediate risk<56–105–10<5High risk>5>5>10>Any mitotic rateAny tumor>10Note. from Fletcher et al (13). Abbreviations: HPF, high-power field. Conclusion Calcification is not a usual clinicopathologic feature of GISTs [1–3], and although a number of rectal GISTs have been reported, we have found no cases so far of rectal GIST presenting as a submucosal calculus. Furthermore, in the literature we found no cases of calcified GIST in other sites of the body. Expression of CD117 (c-kit) and CD34 proved that our calculus was in fact a rectal GIST. Calcification might be explained by internal bleeding of the tumor. The abdominal pain of our patient could be explained entirely by the ureteral stone. The rectal tumor might only explain the constipation. Because of the unpredictable biological behavior of gastrointestinal stromal tumors, the prognosis for our patient is uncertain, in spite of the low-risk character. For this reason, at multidisciplinary oncological deliberation, an expectant policy with regular follow-up was arranged.	[ "calculus", "gastrointestinal stromal tumors", "rectal tumor" ]	[ "P", "P", "P" ]
Lipids-3-1-2039812	Arachidonic Acid but not Eicosapentaenoic Acid (EPA) and Oleic Acid Activates NF-κB and Elevates ICAM-1 Expression in Caco-2 Cells	In patients with inflammatory bowel disease (IBD), intestinal activation of the transcription factor NF-κB as well as intercellular adhesion molecule (ICAM)-1 expression, which is involved in recruiting leukocytes to the side of inflammation is increased. Moreover, colonic arachidonic acid (ARA) proportions are increased and oleic acid (OA) proportions are decreased. Fish oils are protective in IBD patients however, a side-by-side comparison between effects of fish oils, ARA and OA has not been made. We therefore, compared effects of eicosapentaenoic acid (EPA) versus ARA and OA on ICAM-1 expression in Caco-2 enterocytes. To validate our model we showed that dexamethasone, sulfasalazine and PPARα (GW7647) or PPARγ (troglitazone) agonists significantly lowered ICAM-1 expression. ICAM-1 expression of non-stimulated and cytokine stimulated Caco-2 cells cultured for 22 days with ARA was significant higher as compared to EPA and OA. Furthermore, ARA increased NF-κB activation in a reporter cell-line as compared to EPA. Antibody array analysis of multiple inflammatory proteins particularly showed an increased monocyte chemotactic protein (MCP)-1 and angiogenin production and a decreased interleukin (IL)-6 and IL-10 production by ARA as compared to EPA. Our results showed that ARA but not EPA and OA activates NF-κB and elevates ICAM-1 expression in Caco-2 enterocytes. It suggests that replacement of ARA by EPA or OA in the colon mucosa might have beneficial effects for IBD patients. Finally, we suggest that the pro-inflammatory effects of ARA versus EPA and OA are not related to PPARγ activation and/or eicosanoid formation. Introduction Epidemiological studies have shown a low incidence of inflammatory bowel disease (IBD) in Eskimo’s as compared to West-European populations [1] and increasing incidences of IBD in Japan [2]. These findings suggest that an increased dietary intake of n-6 polyunsaturated fatty acids (PUFA) and a lower intake of n-3 PUFAs contribute to the development of IBD. Thus, n-3 fish oil PUFAs may have anti-inflammatory effects as compared to n-6 PUFA [3]. Indeed, elevated proportions of the n-6 PUFA arachidonic acid (ARA) in colon mucosa of both ulcerative colitis (UC) and Crohn's disease (CD) patients as compared to those of control subjects have been shown [4–7]. Although the proportion of the n-3 PUFA eicosapentaenoic acid (EPA) in the mucosa was in most studies not statistically different between IBD patients and healthy controls, one study reported a tendency towards lower EPA proportions in IBD patients [4]. In contrast to EPA, the proportion of the fish oil PUFA docosahexaenoic acid (DHA) was elevated in colonic mucosa of IBD patients as compared to control subjects [4, 5, 7]. Interestingly, the proportion of the n-9 monounsaturated fatty acid (MUFA) oleic acid (OA) was lower in colon mucosa of IBD patients [4, 7]. Because of these observations, and because EPA is a more important precursor of eicosanoids than DHA [8], we decided to compare the effects of EPA versus ARA and OA in an in vitro model of intestinal inflammation. OA is already the most abundant fatty acid present both in our diet [9] as well as in colon mucosa [4]. Therefore, and because of differences in ability to compete with ARA for incorporation in tissue phospholipids [10, 11], it is probably easier to lower mucosal ARA levels by increasing fish oil intake than by increasing OA intake. Indeed, by increasing their intake of fish oil, ARA in the colon mucosa of IBD patients was replaced by EPA and DHA [12], which was associated with significantly reduced corticosteroid requirements [12] and lower relapse rates [13]. It should be noted however, that not all intervention studies using fish oils were that positive, although the overall conclusion is that fish oil supplementation shows at least minor protective effects [14]. The intercellular adhesion molecule (ICAM)-1 plays an important role in the pathology of IBD. In IBD patients intestinal ICAM-1 expression [15] and plasma levels of soluble ICAM-1 (sICAM) are increased [16], and IBD is associated with polymorphisms in the gene encoding for ICAM-1 [17]. Moreover, animal models [18, 19] and a human intervention study [20] have shown that ICAM-1 blocking inhibited intestinal inflammation. The transcription factor NF-κB is a key regulator of the inflammatory response and activation of NF-κB seems to play a critical role in the initiation and perpetuation of intestinal inflammation in IBD [21, 22]. NF-κB activity in the colon is increased during active episodes in IBD patients and certain anti-inflammatory drugs commonly used for IBD appear to inhibit NF-κB [23–27]. In animal models, NF-κB blockade abolished experimental colitis [28, 29]. A side-by-side comparison of the n-3 PUFA EPA, the n-6 PUFA ARA and the n-9 MUFA OA on ICAM-1 expression and NF-κB activation of intestinal cells has, as far as we are aware of, never been performed. Therefore the aim of the present study was to compare the effects of EPA, ARA and OA on ICAM-1 expression and NF-κB activation in the human intestinal epithelial Caco-2 cell line in vitro. Materials and Methods Reagents Bovine serum albumin (BSA; endotoxin and fatty acid-free), sulfasalazine, dexamethasone, GW7647, oleic acid (OA), arachidonic acid (ARA), eicosapentaenoic acid (EPA) and indomethacin were obtained from Sigma Chemical Company (St Louis, MO). Troglitazone was purchased from Biomol (Plymouth Meeting, PA). Recombinant human IL-1β and interferon (IFN)γ were purchased from Roche Molecular Biochemicals (Mannheim, Germany). DMEM, trypsin, penicillin streptomycin (PS), sodium pyruvate (SP) and non-essential amino acids (NEAA) were obtained from Invitrogen Corporation (Paisley, UK). Fetal calf serum (FCS; South-American) was obtained from Greiner Bio-one (Frickenhausen, Germany). Intestinal Cell Cultures The human cell line Caco-2 was purchased from the American Tissue Type Collection (ATTC). Caco-2 cells were cultured in DMEM supplemented with 10% heat-inactivated FCS and 1% penicillin streptomycin (PS), 1% sodium pyruvate (SP) and 1% non-essential amino acids (NEAA). Cells were cultured at 37 °C in a 5% CO2 humidified atmosphere, refreshed every second day and separated by trypsin–0.03% EDTA, when they had reached 70–90% confluence. To evaluate the immune-modulating effects of different interventions, Caco-2 cells were plated in six well tissue culture plates at an initial density of 0.5 × 106 cells/mL in a total volume of 1.5 mL. Medium was replaced every other day for 24 days. After 24 days Caco-2 wells were fully differentiated into small intestinal enterocytes [30]. First, effects of immune-suppressive pharmacological compounds (sulfasalazine, dexamethasone and troglitazone) were tested to validate the model. Although these pharmacologic compounds have known immune-suppressive effects, effects on ICAM-1 expression in Caco-2 cells have—as far as we know—not been reported before. Therefore, after Caco-2 cells were fully differentiated, medium was replaced by medium containing the compound of interest in combination with an inflammation-inducing cocktail consisting of the cytokines IFNγ (100 U/mL) and IL-1β (50 U/mL). The compounds of interest were pre-incubated 30 min (sulfasalazine) or 2 h (dexamethasone, troglitazone or GW7647) before stimulation with the cytokine cocktail. After 16 h of cytokine stimulation, cells were used to determine cell surface ICAM-1 protein expression. Fatty Acid Experiments The effects of various fatty acids were evaluated using the same Caco-2 cell model. For this, various fatty acids were added at indicated concentrations 2 days after plating the cells and again for the following 22 days each time when the medium was refreshed. We used OA as a control n-9 MUFA because this is the most abundant fatty acid in the diet [9]. We further compared the effects of the n-6 PUFA ARA versus the n-3 PUFA EPA. We have explicitly chosen to use EPA instead of DHA since the proportion DHA in the colon mucosa of IBD patients was already higher as compared to control subjects [4, 5, 7]. In addition, EPA is most likely a more important eicosanoid precursor [8]. The fatty acids were dissolved in ethanol up to a final ethanol concentration in the medium of maximal 0.5% (v/v). To prevent cytotoxicity of the fatty acids the FA were bound to albumin, by pre-incubating the fatty acids dissolved in ethanol for 30 min at 37 °C in full culture medium together with 10% FCS, which also contained 0.1% BSA. Caco-2 cells were cultured with respectively 160 μM OA [C18:1(n-9)] versus 130 μM ARA [C20:4(n-6)] plus 30 μM OA (in total 160 μM fatty acids) or 6 μM EPA [C20:5(n-3)] plus 154 μM OA (in total also 160 μM fatty acids). By this approach the total molarity of fatty acids supplied was similar in all experiments, while supplying different amounts of the fatty acid of interest (i.e., 130 μM ARA or 6 μM EPA). These relatively low concentrations of OA (160 μM), ARA (130 μM) and EPA (6 μM) were chosen because they are four times higher than normally present in culture medium of Caco-2 cells containing 10% FCS. We have deliberately chosen for this low EPA concentration since the EPA concentration is very low in FCS. However, the EPA concentration used is—as for all fatty acids used in these experiments—already four times higher than normally present in culture medium. We cultured the cells for 22 days with these relatively low concentrations of fatty acids to simulate a realistic long-term in vivo change in dietary fatty acid intake. After 22 days culture with fatty acids, medium was replaced by medium enriched with the different fatty acids plus the cytokine cocktail [IFNγ (100 U/mL) and IL-1β (50 U/mL)]. After 16 h stimulation, ICAM-1 expression on living cells was measured and culture medium was collected to determine inflammatory protein expression profiles. To evaluate the effects of the fatty acids on NF-κB activity and the role of cyclooxygenase (COX)-enzymes, the experiments with the fatty acids ARA and EPA were repeated but now in our NF-κB reporter Caco-2 cell line with and without indomethacin (20 μM) added 2 h before and during cytokine stimulation. Prostaglandin PGE2 levels in the supernatant were quantitated using a PGE2 Biotrak enzyme-immunoassay (EIA) system (Amersham Biosciences Ltd, Buckinghamshire, UK) according to the high sensitivity enzyme immunoassay protocol 2. Flow Cytometry Analysis of ICAM-1 In order to quantify cell surface ICAM-1 protein expression on living Caco-2 cells, we developed a flow cytometry assay. After 16 h of stimulation, the cells were washed three times with PBS and detached with trypsin–0.03% EDTA. Next, medium was added and cell suspensions were centrifuged for 5 min at 1,200 rpm at room temperature, followed by resuspending the pellets in 500 μL PBS-1% BSA. Cells were counted and diluted to 106 cells/mL in PBS-1% BSA. Recombinant-phycoerythrin (R-PE)-conjugated mouse-anti-human CD-54 monoclonal antibody (anti-ICAM-1) or isotype-matched control antibody (Becton Dickinson Biosciences, San Diego, CA; 20 μL/106 cells) was added and incubated for 30 min on ice in the dark. Next, cell suspensions were centrifuged for 5 min at 1,500 rpm and pellets were resuspended in 500 μL PBS-1% BSA. The amount of fluorescence of 10,000 living cells was counted and analyzed with the FACSort and CellQuest analysis software (Becton Dickinson, Franklin Lakes, NJ). Stable Transfection of NF-κB in Caco-2 Cells For evaluating the effects of the various interventions on transcriptional activity of NF-κB, a stable NF-κB reporter Caco-2 cell line was created. The 6κB-TK-luciferase (NF-κB reporter) plasmid and neomycin resistance plasmid were both kindly provided by Dr. R.C. Langen (Department of Pulmonology, Maastricht University, The Netherlands). Cells were transfected using Lipofectamine 2000 (Invitrogen Corporation, Paisley, UK) according the manufacturers’ instructions. Positive clones were selected by culturing with geneticin (1 mg/mL). To determine luciferase activity, non-stimulated and 3 h cytokine (100 U/mL IFNγ and 50 U/mL IL-1β) stimulated cells were lysed in luciferase lysis buffer (Promega, Madison, WI) and stored at −80 °C. Luciferase (Promega) activity was measured according to the manufacturers’ instructions and expressed relative to total protein (Bio-rad assay; Bio-rad, Hercules, CA). Peroxisome Proliferator-Activated Receptor (PPAR)γ and PPARα mRNA Expression of Differentiated Caco-2 Cells Total RNA was extracted from differentiated Caco-2 cells with Trizol according to the manufacturers’ instructions (Gibco BRL, Gaithersburg, MD). Next, cDNA was made as described [31], and mRNA expression of PPARγ and PPARα was determined using commercially available Taqman gene expression assays (Applied Biosystems, Foster city, CA). Data were normalized against β-actin as housekeeping gene. Fatty Acid Composition of Caco-2 Cells Fatty acid incorporation into the Caco-2 cells was evaluated using extraction and analysis procedures as previously described [32]. Briefly, total lipids were extracted from 500 μL cell suspension in PBS-1% BSA according to the method of Bligh and Dyer [33]. Aminopropyl-bonded silica columns (Varian, Harbor City, CA) were used to separate phospholipids from the total lipid extract [34]. The phospholipids were then saponified, and the resultant fatty acids were methylated into their corresponding fatty acid methyl esters (FAMEs) [35]. Fatty acids were separated on an Autosystem (Perkin-Elmer, Norwalk, CT) gas chromatograph that was fitted with a silica-gel column (Cp-sil 88 for FAME, 50 m × 0.25 mm, 0.2-μm film thickness; Chrompack, Middelburg, The Netherlands) with helium gas (130 kPa) as the carrier gas. Both the injection and detection temperatures were set at 300 °C. The starting temperature of the column was 160 °C. Ten min after injection, the temperature was increased up to 190 °C at a rate of 2.5 °C/min. After 20 min at 190 °C, the temperature was increased up to 230 °C at a rate of 4 °C/min. The final temperature of 230 °C was maintained for 10 min. Data were analyzed by using CHROMCARD software (version 1.21; CE Instruments, Milan, Italy). The fatty acid compositions of the Caco-2 cells are expressed in relative amounts (% of total fatty acids identified; w/w). Inflammatory Protein Expression Profiles Using an Antibody Array Protein expression patterns of multiple cytokines, chemokines and growth factors, were detected simultaneously in Caco-2 cell culture media with the human cytokine antibody array III (Ray Biotech Inc., Norcross, GA) according to the manufacturers’ instructions. First, duplicates of cell culture media of Caco-2 cells cultured with ARA and EPA after cytokine stimulation were pooled. One millilitre of the pooled samples was added to the array membranes. After incubating and washing, the protein-bound membrane was incubated with a cocktail of biotin-labeled antibodies, followed by the addition of horseradish peroxidase-conjugated streptavidin. Array spot intensity was detected by using a LAS-3000 Lite Image reader (Raytest GmbH, Straubenhart, Germany) based on chemiluminecence imaging. Intensity of the spots was quantified in arbitrary units (a.u.) by densitometry using Aida software version 3.50 (Raytest GmbH), thereby correcting for background staining of the gel. Comparison of protein expression profiles was possible after normalization of each spot on an array using the positive controls, provided by the manufacturer. The sensitivity of the array is not the same for the various proteins. Differences in heights of bars from different proteins do therefore not necessarily represent differences in concentrations. The cytokines used for stimulation (IFNγ and IL-1β) were excluded from analysis. Detection of ICAM-1 on Caco-2 Frozen Sections To determine the localization of ICAM-1 in our in vitro Caco-2 cell model, Caco-2 cells were cultured and differentiated into small intestinal enterocyte on collagen-coated polyfluoroethylene transwell membrane inserts with a 0.4 μm membrane pore size (Corning Costar, Cambridge, MA). Differentiated Caco-2 cells were stimulated with IFNγ (100 U/mL) and IL-1β (50 U/mL) for 16 h, embedded in Tissue-Tek (Sakura Finetek, Zoeterwoude, The Netherlands) and rapidly frozen in 2-propanol (Fluka, Zwijndrecht, The Netherlands), dry-ice-cooled and stored at −80 °C. Serial cryosections (10 μm) were obtained using a Leica CM3050 cryostat (Leica Microsystems GmbH, Wetzlar, Germany) and thaw mounted on uncoated glass slides. Before processing or storage at −80 °C, the samples were air dried overnight. To detect ICAM-1 the sections were incubated 30 min in the dark at room temperature with recombinant-phycoerythrin (R-PE)-conjugated mouse–anti-human CD-54 monoclonal antibody or isotype-matched control antibody (Becton Dickinson Biosciences, San Diego, CA) 1:50 diluted in PBS-1% BSA. To detect cytokeratin (CK)-19 the sections were simultaneously incubated with a monoclonal antibody directed to CK-19, kindly obtained from Dr. E.B. Lane (University Dundee, Dundee, UK) 1:10 diluted. Then the sections were washed three times for 5 min in PBS. After that the secondary antibody goat anti-mouse IgG1 (ALEXA555) (Molecular Probes Europe, Leiden, The Netherlands) (1:500) against anti-CD54 (to evade fast quenching of the PE-label) and goat anti-mouse IgG2b (FITC) (Southern Biotech, Sanbio BV, Uden, The Netherlands) (1:50) against CK-19 diluted in PBS-1% BSA was added to the sections and incubated for 30 min. Again the sections were washed three times for 5 min with PBS. Finally, sections were mounted in Mowiol-TRIS pH 8.5 (Calbiochem, Omnilabo International, Etten-Leur, The Netherlands) containing 0.5 g/mL 4–6-diamino-2-phenylindole (DAPI; Molecular Probes Europe) to stain the nuclei. All sections were examined using a Nikon E800 fluorescence microscope (Uvikon, Bunnik, The Netherlands) coupled to a Basler A101C progressive scan colour CCD camera. By just a simple shift in filters, images were grabbed in fluorescence using the ALEXA excitation filter (540–580 nm), the FITC excitation filter (465–495 nm) and DAPI UV excitation filter (340–380 nm) in the red, green and blue channel, respectively. The images acquired were merged to examine the cellular localisation and level of expression of ICAM-1. Statistical Analysis Data were expressed as means and standard deviations (SD) for the non-stimulated condition, the cytokine stimulated condition and cytokine stimulated–non-stimulated (called net stimulated) condition. To determine statistical significance unpaired t-tests (comparison between two interventions) or ANOVA (comparison between three interventions) with a Bonferonni post-hoc test when differences between interventions were significantly different, were performed. First non-stimulated values between interventions were tested to examine differences in basal values. Then, net stimulated values between interventions were tested to examine differences in cytokine-induced changes. Furthermore cytokines stimulated values were tested to examine differences in “end” values. All statistical analyses were performed using SPSS 11 for Mac Os X (SPSS, Chicago, IL). P-values of less than 0.05 were considered statistically significant. Results Model Validation Stimulation with IL-1β and IFNγ increased cell-surface ICAM-1 protein expression on Caco-2 cells, which resulted in a net stimulated ICAM-1 expression of about 50 a.u. (Fig. 1, panels a and b). After pre-treatment of the cells for 2 h with dexamethasone, a corticosteroid with known therapeutic effects in IBD patients, the cytokine stimulated ICAM-1 expression was significantly decreased as compared to control (P = 0.044). Also the net stimulated ICAM-1 expression (25 a.u.) was significantly reduced as compared to control (P = 0.019). Next, we examined the effect of another frequently used therapeutic drug for IBD patients, sulfasalazine. Effects of sulfasalazine were comparable to those observed for dexamethasone. However, also ICAM-1 expression without cytokine stimulation (P = 0.004) (Fig. 1, panel a) was lowered. Altogether, these results show that in our cell model ICAM-1 expression is related to the clinical outcomes of drugs proven to treat IBD and could therefore be used as the main outcome parameter in the following experiments. To further validate characteristics of our in vitro model, we also localized the site of ICAM-1 expression on the Caco-2 cells by means of immunohistochemistry on frozen sections. As shown in Fig. 1 (panels c–f), ICAM-1 was expressed on the apical (lumen) side of the polarized Caco-2 cells. Besides a low constitutive expression (panel d), there was a clear increase after stimulation with the cytokine cocktail (panel e). This localisation is in line with the apical ICAM-1 expression, as found in intestinal biopsies from IBD patients [36]. Fig. 1Model validation. a ICAM-1 expression (in arbitrary units a.u.) on living control Caco-2 cells and after 2 h pre-treatment with 1 μM dexamethasone or 30 min 5 mM sulfasalazine with and without cytokine stimulation for 16 h (100 U/mL IFNγ and 50 U/mL IL-1β) and b net stimulated (stimulated–non-stimulated) ICAM-1 expression. Results represent means ± SD; n = 2. Representative representation of two independent experiments. aP < 0.05 versus control non-stimulated, bP < 0.05 versus control cytokine stimulated, cP < 0.05 versus control net stimulated. (c–f) Immunohistochemistry of ICAM-1 (red) on frozen sections of Caco-2 cells on a transwell showed increased apical expression after cytokine stimulation (100 U/mL IFNγ and 50 U/mL IL-1β). c Hematoxylin staining d Non-stimulated Caco-2 cells e Apical cytokine stimulated Caco-2 cells f Isotype control staining of ICAM-1. Red staining ICAM-1, green staining cytoskeleton (cytokeratin-19), blue staining nucleus. Magnification 40X Effects of PPAR Agonists on ICAM-1 Expression and NF-κB Activation Since PPARs are known modulators of inflammation and fatty acids are natural ligands for PPARs, we first examined PPAR expression in our differentiated Caco-2 cells. We found that PPARα and PPARγ mRNA are expressed in equal amounts (data not shown). Next we examined the effects of the PPARγ agonist troglitazone [a thiazolidinedione (TZD)] and the PPARα agonist GW7647 on ICAM-1 expression. Although both troglitazone and GW7647 significantly lowered the cytokine stimulated ICAM-1 expression as compared to control (P < 0.001 and P = 0.009, respectively), only troglitazone significantly (P = 0.008) reduced the net stimulated ICAM-1 expression (Fig. 2, panels a and b). Troglitazone and GW7647 were used at a concentration where they are selective for their receptor subtypes [37, 38]. It is known that PPARs suppress inflammation by inhibiting the transcription factor NF-κB, which is a key regulator of inflammation [39]. Therefore, we also examined effects of troglitazone on NF-κB transactivation in our Caco-2 NF-κB reporter cell line. These results showed that the cytokine cocktail induced after already 3 h a more than 200% increase in NF-κB transactivation. This increase of NF-κB transactivation was significantly (P = 0.008) lower after 2 h pre-treatment with the PPARγ agonist troglitazone (Fig. 2, panels c and d). Moreover, the inhibitory effect of pre-treatment with the PPARγ agonist troglitazone on cytokine induced NF-κB transactivation was significantly (P = 0.001) larger as compared to pre-treatment with the PPARα agonist GW7647 (data not shown). Fig. 2a, b ICAM-1 expression (in arbitrary units a.u.) on living control Caco-2 cells and after pre-treatment with PPAR agonists (2 h; PPARγ troglitazone 100 μM and PPARα GW7647 1 μM) and c, d NF-κB transactivation measured by luciferase activity (in RLU/mg) in a NF-κB reporter Caco-2 cell-line of control cells and after troglitazone pre-treatment with and without cytokine stimulation for 3 h (100 U/mL IFNγ and 50 U/mL IL-1β) and net stimulated (stimulated–non-stimulated). Results represent means ± SD; n = 2. Representative representation of two independent experiments. bP < 0.05 versus control cytokine stimulated, cP < 0.05 versus control net stimulated Fatty Acid Incorporation in Phospholipids Table 1 shows that the fatty acids kept for 22 days in the culture medium were incorporated into the phospholipid fraction (% of total fatty acids) of the Caco-2 cells. Fatty acid incorporation into phospholipids showed the same pattern as the fatty acid composition of total lipids (data not shown), although changes in the total lipids were more pronounced. Moreover, the fatty acid profiles of phospholipids (Table 1) and total lipids (data not shown) were similar in non-stimulated and stimulated Caco-2 cells. Cells cultured with ARA showed in particular an increase in the proportion of long chain n-6 PUFAs (mainly ARA) and a decrease in n-9 MUFAs (mainly OA) as compared with cells cultured with only OA. The cells cultured with EPA showed an increase in the proportion of n-3 fatty acids (mainly EPA) whereas the proportions of n-9 MUFAs and n-6 PUFAs did not change much as compared with cells cultured with only OA. Table 1Fatty acid composition in phospholipids of Caco-2 cells without (−) and with (+) cytokine stimulation (IL-1β and IFNγ) supplemented with different fatty acids (% of total fatty acids)Fatty acidsOA (160 μM)ARA + OA (130 + 30 μM)EPA + OA (6 + 154 μM)−+−+−+16:0 (PA)16.016.118.719.516.315.818:0 (SA)5.96.19.610.56.76.518:1 trans 3.03.16.15.72.72.718:1 (n-7)3.43.42.62.53.33.318:1 (n-9) (OA)47.146.314.413.845.245.018:2 (n-6) (LA)1.71.71.11.11.71.620:1 (n-9)2.12.10.30.31.71.720:4 (n-6) (ARA)9.39.429.328.98.98.920:5 (n-3) (EPA)0.40.40.00.02.12.222:1 (n-9)0.80.81.51.30.80.824:1 (n-9)2.02.10.91.02.02.122:4 (n-6)0.40.47.77.80.30.322:5 (n-3)0.50.50.40.41.31.322:6 (n-3) (DHA)1.11.10.60.60.70.6∑93.793.593.293.493.792.8∑ SAFA21.922.228.330.023.022.3∑ MUFA58.457.825.824.655.755.6∑ PUFA12.312.438.538.214.314.3∑ n-32.02.01.01.04.14.1∑ n-611.411.538.137.810.910.8∑ n-952.051.317.116.449.749.6Data are representative for two independent experiments and the data are derived from pooled samplesARA arachidonic acid, EPA eicosapentaenoic acid, LA linoleic acid, MUFA monounsaturated fatty acids, OA oleic acid, PA palmitic acid, PUFA polyunsaturated fatty acids, SA stearic acid, SAFA saturated fatty acids, Σ sum, − non-stimulated, + cytokine stimulated Effects of Fatty Acids on ICAM-1 Expression and NF-κB Activation ICAM-1 expression was analyzed on living cells. FACS analysis did not show significant changes in cell populations cultured with the various fatty acids, suggesting no significant cell death. As shown in Fig. 3 (panel a), ARA significantly elevated ICAM-1 expression as compared to OA or EPA (non-stimulated P < 0.001 ARA versus OA and ARA versus EPA, and stimulated P = 0.010 ARA versus OA, and P = 0.013 ARA versus EPA). Interestingly, stimulated and non-stimulated ICAM-1 expression on cells cultured with EPA did not significantly differ from those cultured with OA (Fig. 3, panel a). Despite these clear differences in non-stimulated and stimulated ICAM-1 expression between ARA and EPA or OA, the net stimulated ICAM-1 expression did not differ between the groups (ANOVA P = 0.134). Since dietary interventions aiming at lowering mucosal ARA content are most likely more successful by dietary EPA enrichments than by OA enrichment–because OA is the most abundant fatty acid present in our diet [9], making higher intakes not realistic–in further experiments we have focused on effects of ARA versus EPA. First, we examined the effects on NF-κB activation in the stable NF-κB reporter Caco-2 cell line (Fig. 4). NF-κB transactivation in stimulated Caco-2 cells cultured with ARA was significantly increased as compared to that in stimulated cells cultured with EPA (Fig. 4, panels a and b) (stimulated P = 0.017 and net stimulated P = 0.001). To evaluate whether the effects were related to the production of different families of eicosanoids we measured PGE2 production and examined the effects of the different fatty acids in the presence of the cyclooxygenase (COX)-inhibitor indomethacin. As expected, PGE2 production by Caco-2 cells cultured with ARA was significantly higher than by cells cultured with EPA (data not shown). Stimulation of the cells with the cytokine cocktail did however, not influence PGE2 production. Since cytokine stimulation enhanced NF-κB activation, this suggests that PGE2 is not directly involved in NF-κB activation. Indomethacin decreased PGE2 production of Caco-2 cells by approximately 70% (data not shown), but had no effect on NF-κB transactivation (Fig. 4, panels c and d) as compared to untreated Caco-2 cells. Also ICAM-1 expression was not affected (data not shown). However, as shown in Fig. 4, the effects of ARA versus EPA in the presence of indomethacin treatment (panel e) showed an identical pattern as observed without indomethacin (panel a). Also in the presence of indomethacin the net cytokine stimulated NF-κB transactivation was significantly higher after ARA as compared to EPA (P = 0.047) (Fig. 4, panel f). This indicates that the increased basal PGE2 production by ARA cannot explain the different effects of ARA and EPA on NF-κB activation. Fig. 3a–b ICAM-1 expression (in arbitrary units a.u.) on living Caco-2 cells cultured for 22 days with 160 μM oleic acid (OA) or 130 μM arachidonic acid (ARA) plus 30 μM OA or 6 μM eicosapentaenoic acid (EPA) plus 154 μM OA with and without cytokine stimulation for 16 h (100 U/mL IFNγ and 50 U/mL IL-1β) and net stimulated (stimulated–non-stimulated). Results represent means ± SD; n = 2. Representative representation of two independent experiments. ANOVA between groups, non-stimulated P < 0.001, stimulated P = 0.006 and net stimulated P = 0.134. aBonferroni P < 0.05 versus OA and EPA non-stimulated, bBonferroni P < 0.05 versus OA and EPA cytokine stimulatedFig. 4NF-κB transactivation with and without cytokine stimulation for 3 h (100 U/mL IFNγ and 50 U/mL IL-1β) (a, c, e) and net stimulated (stimulated–non-stimulated) (b, d, f) in a NF-κB reporter Caco-2 cell line measured by luciferase activity (in RLU/mg). a–b Reporter Caco-2 cells cultured for 22 days with 130 μM arachidonic acid (ARA) plus 30 μM oleic acid (OA) or 6 μM eicosapentaenoic acid (EPA) plus 154 μM OA. c–d Control reporter Caco-2 cells and after 2 h pre-treatment with the COX-inhibitor indomethacin (20 μM). e–f Reporter Caco-2 cells cultured for 22 days with 130 μM arachidonic acid (ARA) plus 30 μM OA or 6 μM eicosapentaenoic acid (EPA) plus 154 μM OA and 2 h pre-treatment with the COX-inhibitor indomethacin (20 μM). Results represent means ± SD; n = 2. bP < 0.05 versus EPA cytokine stimulated, cP < 0.05 versus EPA or EPA + indomethacin net stimulated Effects of Fatty Acids on Inflammatory Proteins Expression Profiles Finally, to explore the effects of ARA and EPA on a broader scale, we evaluated protein expression profiles consisting of various inflammatory mediators by using antibody arrays. Figure 5 shows that ARA treatment particularly increased expression of monocyte chemotactic protein (MCP)-1 and angiogenin, while EPA treatment increased IL-10, IL-6, macrophage inflammatory protein (MIP)-1δ, and growth regulated protein (GRO) expression. Fig. 5Protein expression profile in the culture medium of Caco-2 cells cultured for 22 days with 130 μM arachidonic acid (ARA) plus 30 μM oleic acid (OA) or 6 μM eicosapentaenoic acid (EPA) plus 154 μM OA after 16 h cytokine stimulation (100 U/mL IFNγ and 50 U/mL IL-1β) measured with an antibody array. Results are represented as mean ± SD; n = 2. CSF colony stimulating factor, ENA epithelial-derived neutrophil activating protein, MCP monocyte chemotactic protein, MDC, macrophage derived chemokine, MIG monokine induced by gamma interferon, MIP macrophage inflammatory protein, SDF stromal cell-derived factor, GRO growth regulated protein, MCSF Macrophage colony stimulating factor, Tpo thrombopoietin, EGF epidermal growth factor, OSM oncostatin M, VEGF Vascular endothelial growth factor, SCF stem cell factor, IL interleukin, TNF tumor necrosis factor, Ang angiogenin Discussion ICAM-1 seems important in the pathology of IBD [15]. Whether fish oils, which have been shown to protect against relapses in IBD patients on remission [13], have effects on ICAM-1 expression, is however, unknown. In addition, a direct side-by-side comparison of fish oils with ARA, which is postulated to have pro-inflammatory effects [40] and is elevated in the colon mucosa of IBD patients [4, 7], has never been made. The transcription factor NF-κB is important in regulating intestinal inflammation and is elevated in IBD patients [21, 22, 26]. We have now shown that the n-3 PUFA EPA, as compared to the n-6 PUFA ARA, clearly reduced cytokine stimulated NF-κB activation and ICAM-1 expression in enterocytes in vitro. Moreover, effects of OA on ICAM-1 expression were comparable to those of EPA. Because EPA and OA resulted in comparable changes in the proportions of ARA in the phospholipids of the enterocytes, the reported effects may be ascribed to the increased ARA proportion in the ARA cultured cells. Thus, decreasing cellular ARA levels seems to be a crucial step. Since OA is already the most abundant fatty acid present in our diet [9] and in the colon mucosa [4], decreasing mucosal ARA levels may be easier by increasing fish oil intake than by increasing OA intake. Moreover, EPA does compete with ARA for incorporation into tissue phospholipids [10, 11]. Our data showed that replacing ARA for EPA or OA decreased ICAM-1 expression and NF-κB activation in Caco-2 enterocytes. In line with our observations in enterocytes, n-6 PUFAs also increased NF-κB activation as compared to n-3 PUFAs in monocytes [41] and macrophages [42]. Also, earlier in vitro studies have demonstrated that fish oils reduced cytokine stimulated ICAM-1 expression in endothelial cells [43] and monocytes [44] as compared to conditions without addition of fatty acids. Moreover, in vivo ICAM-1 expression (surface and mRNA) on peritoneal macrophages was reduced in mice fed fish oils compared to that in mice fed coconut oil [45]. In humans, dietary fish oil supplementation lowered expression of ICAM-1 on ex vivo stimulated monocytes as compared to no supplementation [46]. However, our study is the first that examined effects of EPA versus ARA on ICAM-1 expression and NF-κB activation in enterocytes. We however, realize that, although enterocytes play an important role in intestinal inflammation, immune modulating effects of fatty acids in CD patients will be influenced not only by enterocytes, but also the interaction with other intestinal immune and non-immune cells is important. Therefore, in future experiments effects on other cell types, e.g. isolated from mucosal biopsies of CD patients should be evaluated. Moreover, to validate if the effects of EPA versus ARA are also applicable in the pathogenesis or even the treatment of CD patients, these effects should be confirmed in appropriate animal models of IBD. We used an approach of supplying different amounts of the fatty acids of interest (i.e., 130 μM ARA or 6 μM EPA), while the total molarity of fatty acids supplied was similar in all experiments by adding OA. These different concentrations of ARA and EPA were deliberately chosen because they were both four times the amount, in which the cells grow normally (i.e., the fatty acid composition of culture medium with 10% FCS). Since the EPA concentration is very low in FCS, we have deliberately chosen for a low EPA concentration. Higher concentrations of EPA would be interesting to examine, however, are difficult to be achieved with dietary interventions in vivo. Using iso-molaric total concentrations of fatty acids is essential, because an increase in total fat can be immune suppressive [47]. Therefore, we used OA as a reference fatty acid to make total fatty acid concentrations between experimental fatty acid conditions iso-molaric, i.e. OA was exchanged for ARA or EPA. In addition, we evaluated the condition of OA only. The results of this latter condition showed that decreasing ARA levels in the mucosa seems to be more important than increasing EPA levels. Regarding the pathways underlying the anti-inflammatory effect of fish oils, several suggestions have been made. As compared to n-6 PUFAs, n-3 PUFAs may have different effects on (I) signal transduction pathways, and (II) the types and levels of eicosanoids synthesized [40]. Regarding the first mechanism, our finding that EPA lowered NF-κB activation and ICAM-1 expression as compared to ARA indeed showed that EPA and ARA differently affected the NF-κB signal transduction pathway. In this respect, two peroxisome proliferators-activated receptors (PPARs), PPARα and PPARγ, seem relevant, since fatty acids are natural ligands for these PPARs. Both PPARs seem to have anti-inflammatory effects by inhibiting NF-κB activation [48, 49]. In our in vitro model, the PPARγ agonist troglitazone inhibited NF-κB activation, and cytokine and net stimulated ICAM-1 expression, while the PPARα agonist GW7647 only inhibited cytokine-stimulated ICAM-1 expression. This suggests that both PPARγ and PPARα have anti-inflammatory effects on enterocytes, in which effects of PPARγ seem more pronounced. This is in agreement with the finding that PPARγ agonists but not a PPARα agonist inhibited IL-8 expression of Caco-2 cells and HT-29 cells [50]. Differences between fatty acids in activation of PPARs might explain the activation of NF-κB by ARA and not by EPA in our Caco-2 cells. However, our findings are not supported by the PPARγ binding affinities of EPA and ARA, which are about the same [51], while we used much lower EPA concentrations than ARA concentrations. Secondly, not only EPA but also OA lowered intestinal ICAM-1 expression while OA is a poor PPARγ ligand [51]. Therefore, although PPARγ activation certainly protects against inflammation, our results do not suggest that the protective effects of EPA and OA as compared to ARA on intestinal inflammation are PPARγ-mediated. Therefore, there should be another explanation why effects of OA and EPA were comparable but different from those of ARA. The mechanism underlying a second possible explanation relates to the incorporation of fatty acids into cell membrane phospholipids and eicosanoids synthesis. Eicosanoids can modulate the intensity and duration of the inflammatory response. [8, 40] Replacement of ARA by EPA in the culture medium of the cells resulted in a decrease in the proportion of ARA in cell membranes and an increase in the EPA proportion and as expected the PGE2 production of Caco-2 cells cultured with EPA was also lower than PGE2 production of cells cultured with ARA. However, PGE2 production was not different between non-stimulated and stimulated cells, while stimulation resulted in an abundant elevation of NF-κB activation. To our opinion this rules out the role of PGE2 in the NF-κB activation of ARA and EPA. Moreover, although in our experiments indomethacin–a blocker of COX–indeed inhibited PGE2 production, it did not lower NF-κB activation. This is in line with findings of De Catherina et al. [43] showing that in endothelial cells the effects of DHA on VCAM-1 expression could not be inhibited by indomethacin, although prostacyclin production was completely suppressed. Therefore, we fully agree that in an in vivo situation pro-inflammatory effects of ARA are probably mediated by eicosanoids synthesized from ARA, however, these effects are not derived from a direct eicosanoid mediated activation of NF-κB as shown in our model. Finally, we evaluated inflammatory protein signatures of Caco-2 cell culture medium after 22 days treatment with EPA versus ARA. Since these signatures were analyzed in pooled material we cannot draw conclusions based on statistical analysis. However, this array was intended to generate hypotheses about differences in immune modulating effect of ARA versus EPA based on the protein expression profiles. These profiles indicate, that both fatty acids induced specific changes in various inflammation mediators. Most of these proteins, but not all, are regulated by NF-κB. For example, IL-6 protein expression was higher after EPA as compared to ARA, whereas NF-κB activation was higher after ARA. It should however, be considered that although highly important, NF-κB is not the only transcription factor involved in the regulation of IL-6 and other inflammatory proteins. Thus, expression of individual proteins cannot be predicted by the level of NF-κB activation solely. In addition, we found clear differences between EPA and ARA, i.e., that some proteins were expressed higher after EPA (IL-6, IL-10, MIP-1δ and GRO), while others were higher expressed after ARA (MCP-1, and angiogenin). The precise role of these individual proteins in the process of intestinal inflammation is not clear. However, the increased MCP-1 expression after ARA treatment is in line with the finding that MCP-1 expression is upregulated in the mucosa of IBD patients and correlates with disease activity [52]. The increased expression of the Th2 cytokine IL-10 after EPA treatment fits with the anti-inflammatory role for IL-10 in IBD pathology. In this respect, IL-10 knockout mice develop chronic intestinal inflammation and delivery of recombinant IL-10 to the intestinal mucosa by the bacterium L. lactis attenuated mucosal inflammation in two mouse models [53]. In humans, a pilot study using these bacteria showed also promising effects [54]. In conclusion, we have shown that ARA but not EPA and OA activates NF-κB and elevates ICAM-1 expression in Caco-2 enterocytes and we hypothesize that the effects are not related to PPARγ activation or eicosanoid formation.	[ "caco-2 cells", "intestinal inflammation", "prostaglandins", "peroxisome proliferator-activated receptor (ppar)" ]	[ "P", "P", "P", "R" ]
Eur_J_Pediatr-3-1-2151778	Hepatitis C virus infection acquired in childhood	Hepatitis C virus (HCV) infection occurs less frequently in children than in adult patients, and the natural history, prognosis, and clinical significance of HCV infection in children are poorly defined. We report here a descriptive follow-up of the clinical course, biochemical data, and viral markers observed in 37 children with anti-HCV. Ten patients included in the study tested persistently negative for serum HCV-RNA (group 1) and 27 patients tested persistently positive (group 2). In group 1, serum alanine aminotransferase (ALT) was normal in all patients, while two patients had non-organ-specific autoantibodies. In group 2, serum ALT was elevated in 13 of 27 patients, and five patients had non-organ-specific autoantibodies. HCV genotype 1a and 1b were the most prevalent among HCV-RNA-positive patients. Twenty liver biopsies were carried out on 17 patients in our series (mean evolution time, 11.2 years; range, 3–21 years). The liver specimens showed mild necroinflammatory changes in most patients, and fibrosis was absent or low grade. Two HCV-RNA-positive patients became persistently HCV-RNA negative. Of the 26 children investigated, 7 (one in group 1, six in group 2) had a co-infection with hepatitis G virus. Conclusion Most children chronically infected with HCV were asymptomatic and presented only mild biochemical evidence of hepatic injury. Autoimmunity in the form of non-organ-specific autoantibodies was common. HCV in children induced mild changes in the liver with a low level of fibrosis and at a low rate of progression. Introduction HCV (hepatitis C virus) infection occurs less frequently in children than in adult patients. The natural history, prognosis and clinical significance of HCV infection are poorly defined in childhood [11]; in contrast, HCV infection in adults presents a high degree of chronicity, with up to 50% of all HCV-infected adults developing progressive liver disease [21]. Results from prospective studies show that 20–30% of chronically infected adults develop compensated and eventually decompensated cirrhosis or hepatocellular carcinoma, or both, within 20 years of the initial infection [1]. The lower prevalence of HCV infection in children and the fact that most patients undergo antiviral drugs treatment result in very limited knowledge of the natural outcome of chronic HCV infection acquired at early ages of life. Thus, immediate goals in the investigation of HCV infection should be to characterize current epidemiology and to describe the pathogenesis and course of hepatitis C in children and adults [20]. To understand the evolution of HCV infection could be itself an important surrogate end-point for the evaluation of infected young patients with prognostic implications and therapeutical consequences. The aim of this study was to determine the clinical features and long-term evolution of HCV infection in a group of children who had never received treatment with antiviral drugs. Patients and methods Thirty-seven children (16 females, 21 males) with positive antibodies to hepatitis C (anti-HCV) were investigated retrospectively. These patients were followed-up for a period of 5 years. None had received treatment with antiviral drugs for viral hepatitis or had a history of intravenous drug abuse. All subjects made regular visits to our outpatient clinic, and the serum levels of alanine aminotransferase (ALT), albumin, prothrombin time, antinuclear antibodies (ANA), anti-mitochondrial antibodies (AMA), anti-smooth muscle antibodies (SMA), liver-kidney anti-microsomal antibodies type I (LKM), anti-gastric parietal cells antibodies (GPCA), rheumatoid factor, thyroxine (T4), thyroid stimulating hormone (TSH), anti-thyroid antibodies, anti-HCV and HCV-RNA were determined at least on five consecutive occasions at 1-year intervals. Hepatitis B virus (HBV) surface antigen (HBsAg), antibodies to HBV surface antigen (anti-HBs), antibodies to HBV core antigen (Anti-HBc), human immunodeficiency virus (HIV) and antibodies to hepatitis E (anti-HEV) were investigated in all patients during one visit. Hepatitis G virus-RNA (HGV-RNA) and antibodies to HGV (anti-HGV) were determined in 26 patients. Genotypes of HCV were performed in 21 viremic children. The duration of infection was calculated as the interval between the presumed date of infection and the date of the last visit to the clinic. Liver biopsies were obtained in 17 patients. Repeated biopsies were performed in three patients. Viral markers for HBV, HCV and HIV were tested by third generation ELISA (Axsym; Abbott Diagnostics, Chicago, Ill.). Anti-HEV was detected by a commercially available ELISA (Bioelisa HEV IgG; Biokit, Barcelona, Spain). HGV-RNA and anti-HEV antibodies were detected using commercial tests (Roche Diagnostics, Mannheim, Germany). HCV-RNA was detected by PCR (Amplicor HCV PCR test, Roche Diagnostics), and HCV genotyping was performed by a second generation line probe assay (INNO-LIPA HCV; Innogenetics, Ghent, Belgium). Liver biopsies were performed percutaneously and under the ultrasound guidance of an experienced operator. All liver biopsy specimens (>12 mm long) were fixed in formalin, embedded in paraffin stained with hematoxylin-eosin and Masson’s trichrome and then observed by a single experienced pathologist (AM). Histological necroinflammatory activity and fibrosis were scored separately. The degree of histological necrosis with inflammatory activity was scored using the three parameters of the histological activity index (HAI) developed by Knodell et al. [12]; these consist of: (1) piecemeal necrosis and bridging necrosis (score: 0–10), (2) lobular cytolysis (score: 0–4) and (3) portal inflammation (score: 0–4). Fibrosis was scored independently using the following system: F0, absence of fibrosis; F1, fibrous expansion of portal areas; F2, portal to portal bridging fibrous tracts; F3, portal-central bridging fibrous septa; F4, cirrhosis (bridging fibrous septa with parenchymal nodules). This score is a modification of the fibrosis scoring system of Knodell’s method and introduces one additional degree of fibrosis. Macrovesicular steatosis was graded semiquantitatively using a modified score from Brunt et al. [5]. Alternative and additional diagnoses identified in the liver biopsy were also recorded. Results Mean chronological age for the 37 patients (21 males and 16 females) at the last follow-up evaluation was 20.1 years (range: 9–30 years). Most of the subjects were diagnosed among patients who were screened for HCV infection after receiving transfusions of blood products for heart surgery, hematological diseases or casual injuries before 1992. Thus, most of the patients included in the study presented with other diseases that were unrelated to HCV infection: 16 patients had a congenital heart disease, seven had various hematological disorders, three had neurological diseases, one had vesicoureteral reflux and one had Turner syndrome associated to IgA deficiency. Ten patients were anti-HCV-positive and HCV-RNA-negative (Group 1). The presence of HCV-RNA was demonstrated in the other 27 patients (Group 2). Demographic characteristics and the main results obtained in these children are shown in Table 1. Table 1Characteristics and biologic data of children with HCV infection Patients negative for HCV-RNA (n = 10)Patients positive for HCV-RNA (n = 27)Age (years) at the follow-up evaluationa18 (9–26)20 (9–30)Duration of the HCV infection (years) in 36 patients at the follow-up evaluationa14 (6–23)16 (5–30)Route of infection Parenteral1024 Vertical02 Unknown01Alanine aminotransferase (U/l)First visitLast visit Normal (n )10714 ≤2 × UNL (n )b067 2–4 × UNL (n )055 >4 × UNL (n )091Non-organ specific antibodies Antinuclear antibody02 Anti-mitochondrial antibodies10 Anti-smooth muscle antibodies01 Liver-kidney anti-microsomal antibody type 111 Gastric parietal cell antibodies01HCV genotype 1a8 1b12 2a1 Unknown6HBV00HIV10HGV-RNA/Anti-HGV (n = 26)1/04/2Anti-HEV00aValues are given as the mean, with the range presented in parenthesisbUNL, Upper normal level Group 1 The once-yearly PCR analysis revealed that ten patients had positive anti-HCV and negative viremia, which reflected an ongoing or past infection with HCV. These patients were asymptomatic, and the liver function tests remained normal in all but two who had been diagnosed with congestive heart failure; the liver abnormalities of these two patients were normalized after their cardiac condition had improved. One female patient presented persistent positive AMA for 3 years, which then normalized and remained negative during the subsequent 9-year follow-up period. Throughout this period the patient was asymptomatic, and liver tests were normal. Two years after the laboratory tests were negative for antibodies, the patient presented with psoriasis. Another patient had persistent high titers (1:620) of anti-LKM. Fifteen years after acquiring the infection this patient lost the anti-HCV but the anti-LKM remained positive for two more years. During the follow-up this patient did not present liver abnormalities or other diseases. During the follow-up span of this study, another female patient became persistently anti-HCV negative 12 years after she was infected. At the time of writing, eight patients remain anti-HCV-positive and HCV-RNA-negative, and they show no evidence of liver disorders. Group 2 HCV-RNA was detectable in the blood of 27 patients, two of whom became HCV-RNA-negative. The age of exposure to the infection and the type of exposure could not be established in one patient; following the detection of HCV in the blood serum, this patient was followed for 11 years and was positive for HCV up to and including the last follow-up visit. All patients, with one exception, were asymptomatic. The exception had been diagnosed with leukemia in 1987 and received transfusions of blood products on several occasions. She underwent chemotherapy and had a cholestatic hepatitis; liver tests revealed persistent abnormalities. Anti-HCV and HCV-RNA were detected in 1991. The patient presented portal hypertension and progressive liver failure, and when she was 8 years old, in 1993, she received a liver transplant. The evolution of serum ALT levels is shown in Table 1. Serum levels of albumin and prothrombin time were abnormal only in the patient who developed hepatic failure. The presence of non-organ-specific autoantibodies (NOSAs) was detected in five patients. One patient (genotype1a) diagnosed with a single ventricle and Fontan surgery developed persistent anti-LKM (titers 1:160–1:620) and the rheumatoid factor throughout the evolution of the disease. Liver biopsy could not be obtained in this patient. Another male patient (genotype 1b) had positive ANA for 2 years, which subsequently normalized. This patient had increased ALT, and the liver biopsy showed minimal necrosis with inflammatory changes (HAI: 2) and no fibrosis. A third patient (genotype 1b) had persistent PGCA during her follow-up period (10 years), with slightly elevated ALT levels, mild necrosis with inflammatory changes (HAI: 4) and minimal fibrosis (stage 1). A fourth patient (genotype1b) with AST levels greater than fourfold the upper normal level presented SMA during the last 2 years of her follow-up (HAI: 6). The patient diagnosed with Turner syndrome and IgA deficiency (genotype 1b) developed throughout the years a seronegative poli-articular rheumatoid arthritis, psoriasis and celiac disease. She had intermittently positive ANA. The levels of T4 and TSH were normal in all patients, and no patient developed antithyroid antibodies. Abdominal ultrasounds did not show significant abnormalities. Only the patient who presented cirrhosis had changes associated to portal hypertension. Liver biopsies were performed in viremic patients who presented altered hepatic function tests and for whom parental consent had been obtained. Seventeen patients underwent 20 liver biopsies. Mean evolution time of the infection at the time of biopsy was established in 16 patients (mean: 11.2 years; range: 3–21 years). The results on necroinflammatory activity and fibrosis are shown in Table 2. Table 2Histopathological features of patients positive for HCV-RNALiver biopsyPatients (n)Necroinflammatory activity 1–26 3–47 5–64 6–180Fibrosis F08 F16 F22 F30 F4 (Cirrhosis)1Steatosis0 A second hepatic biopsy was performed in three patients after 5, 12 and 14 years, respectively. The minimal necroinflammatory activity observed in the first biopsy specimens of these three children remained unchanged in the second biopsy. One of the patients progressed from F0 to F1 fibrosis. A second patient never presented with fibrosis, and both samples of the third patient had minimal fibrosis (F1). During the follow-up period, 2/27 patients became persistently HCV-RNA-negative (for 9 and 10 years, respectively). Both patients showed minimal necrosis and inflammatory activity with minimal fibrosis (F1) in the biopsy performed before HCV-RNA clearance. One patient of Group 1 was HIV-positive, while one patient of Group 1 and six patients of Group 2 were HGV-positive. Viral coinfections observed in both groups are expressed in Table 1. Discussion In this study, chronic hepatitis C infection was silent in most of the children, and there was little biochemical evidence of liver disease in these patients. Only one girl with portal hypertension and liver failure developed symptoms related to the infection. Consequently, our data suggest that HCV infection may be currently underdiagnosed in children and, moreover, that young patients could become a potential source of infection. We therefore strongly recommend that children falling in risk groups be screened for HCV infection. Of the 37 pediatric patients in this series, seven (19%) had NOSAs. These antibodies are highly prevalent in subjects exposed to the HCV, and a positive test result for SMA, GPCA and ANA is part of the natural course of chronic HCV infection in adults and children [7]. Lenzi et al. [13] reported a higher prevalence of NOSAs in anti-HCV-positive adult patients than in normal controls (25 vs. 6%). Stroffolini et al. [22] found positive NOSAs in 36.9% of 502 subjects with HCV-RNA-positive chronic hepatitis. Muratori et al. [16] detected positive NOSAs in 16 of 47 children (34%) with chronic hepatitis C, and anti-LKM was present in 11% of the children. Of the 37 children participating in our study who manifested past or active HCV infection, two had serum anti-LKM. This incidence is lower than that observed by Muratori et al. (11%) [16] and Bortolotti et al. (10.3%) [4] in children with HCV infection. A possible explanation for this finding is the higher age and longer mean duration of HCV infection in the patients included in our series in comparison to the children in the other two studies. The prevalence reported by Lenzi et al. (1.3% [13]), Stroffolini et al. (2.25% [22]) and Reddy et al. (0% [18]) in adults support this hypothesis. Anti-LKM is an immunomarker of type 2 autoimmune hepatitis [2] that tends to present at younger ages and to affect mainly children. Our data suggest that age may be a factor contributing to the presence of anti-LKM1 in children infected with HCV. Moreover, we observed that one of our non-viremic patients had persistent antibodies after the test for anti-HCV became negative. Liver specimens from the children with hepatitis C showed mild necroinflammatory changes and a low level of fibrosis. In most patients, fibrosis was absent or low grade. Only one patient, as mentioned above, developed cirrhosis and liver failure over a period of 6 years. This leukemic patient had other additional risk factors for liver damage that may have affected this outcome. In this respect, Hoshiyama et al. [9] observed that chronic hepatitis C is more frequent among children with hepatitis C infection following blood transfusions for malignant disease than in patients with hepatitis C following blood transfusions for open heart surgery. We have compared our results with those obtained from other pediatric series [6, 8, 9, 23]. Table 3 summarizes the data on fibrosis obtained from liver biopsies of 210 pediatric patients (above-mentioned studies). Fibrosis was not detected or was low grade in 200 patients (95%), and severe fibrosis or cirrhosis was detected in ten patients (5%). Minola et al. [15] and Poynard et al. [17] demonstrated an inverse correlation between age at HCV infection and progression to cirrhosis that suggests that chronic HCV infection in childhood induces mild changes in the liver with a low level of fibrosis and a low rate of progression. However, it has been observed that mild chronic hepatitis C can be a progressive disease in adult patients [3, 19] and that the long evolution of infection acquired in childhood could increase the risk of fibrosis. Thus, some children infected early in life, in whom chronic disease has a mild liver expression, might develop fibrosis in adulthood. Table 3Hepatic fibrosis in biopsies from children with HCV infection Vogt et al. [23]Hoshiyama et al. [9]Guido et al. [8]El-Raziky et al. [6]This reportNumber of patients173811226a17Mean duration of the HCV infection (year)21.2 ± 4.67.1 ± 2.8 (n = 23)12.9 ± 3.1 (n = 15)8.04 ± 5.3Non-determined11.2 ± 5,6FibrosisNo/low grade14/2b231525/8120/47/9Severe/cirrhosis0/1c005/11/1d0/1eaSome patients were co-infected with HBV and /or had secondary iron overloadbTwo patients with congestive heart failurecPatient with Anti-HBs, anti-HBc and anti-HEV antibodiesdThis 12-year-old cirrhotic patient was not thalassaemic and HBV markers were negativeeThis leukemic patient was infected with HCV and was also receiving chemotherapy concurrently It is noteworthy that 6/27 of the children with HCV infection in our series had a co-infection with HGV (HGV-RNA-positive) or had a past HGV infection (anti-HGV-positive). This prevalence of HGV infection is higher than that found in healthy children (6%) in Spain [10]. These data suggest that HCV-infected children must be considered to be a risk group for HGV infection since HCV and HGV are both parenterally transmitted. Moreover, no differences were observed in biochemical and histological findings in HCV-infected children with and without a concomitant infection with HGV, which suggests that superimposed HGV infection does not influence the course of HCV infection. Conversely, our data do not support that HEV, which can be transmitted by a parenteral route [24], is prevalent in children who had undergone blood transfusions (prevalence rate: 0%). However, this low value may result from the low prevalence of HEV infection in Spain (2.8%) [14]. Conclusion Most children chronically infected with HCV are asymptomatic and present only mild biochemical evidence of hepatic injury. Spontaneous clearing of the virus occurs occasionally. Autoantibodies are common in HCV patients. The natural history of chronic hepatitis C in children differs from that in adults since HCV infection is relatively benign, induces mild changes in the liver with a low level of fibrosis and a low rate of progression and is rarely associated with severe or decompensate liver disease.	[ "hepatitis c virus", "anti-liver/kidney microsomal antibody", "infectious hepatitis", "liver fibrosis", "nonorgan-specific autoantibodies" ]	[ "P", "M", "M", "R", "M" ]
Clin_Oral_Investig-2-2-1544376	Bone apposition to titanium implants biocoated with recombinant human bone morphogenetic protein-2 (rhBMP-2). A pilot study in dogs	The aim of the present study was to investigate bone formation to recombinant human bone morphogenetic protein-2 (rhBMP-2)-biocoated and rhBMP-2-nonbiocoated titanium implants after implantation in dogs. Implantation of sand-blasted and acid-etched (C), chromosulfuric acid surface-enhanced (CSA), and rhBMP-2-biocoated CSA [BMP-A: noncovalently immobilized rhBMP-2 (596 ng/cm2), BMP-B: covalently immobilized rhBMP-2 (819 ng/cm2)] implants was performed in both the mandible and tibia of dogs. After 4 weeks of healing, the percentage of direct bone to implant contact (BIC) and the induced bone density (BD) at a distance of less than and greater than 1 mm adjacent to each implant was assessed. Histomorphometric analysis of implants inserted in the mandible and tibia revealed that BIC values appeared to be highest in the BMP-B group, followed by BMP-A, CSA, and C. BD as measured at a distance of <1 mm revealed obvious differences between groups: BMP-B>BMP-A>CSA>C. However, no differences between groups were observed at a distance of >1 mm. Within the limits of the present study, it may be concluded that rhBMP-2 immobilized by covalent and noncovalent methods on CSA-treated implant surfaces seemed to be stable and promoted direct bone apposition in a concentration-dependant manner. Introduction The adhesion of plasma proteins on the surface of titanium implants has been reported to play an essential role in the process of osseointegration [2, 4, 8, 9]. Each surface of a material is characterized by a unique composition of adsorbed proteins, which influences the type of cells that may adhere. Subsequently, the specific pattern of adsorbed proteins determines the type of tissue that will develop at the interface between the implanted material and the host [32, 41, 42]. In recent years, several modifications of specific surface properties such as structure, chemistry, surface charge, and wettability have been investigated to improve osseointegration of titanium implants [1]. Additionally, several growth factors and cytokines have also been suggested to stimulate a deposition of cells with the capacity of regenerating the desired tissue [27, 39, 46]. In case of endosseous titanium implants, an enhanced proliferation and differentiation of undifferentiated mesenchymal cells, osteoprogenitor cells, and preosteoblasts into osteoblasts may improve bone response and subsequently osseointegration [5]. One particular growth factor, bone morphogenetic protein (BMP), has shown considerable potential to stimulate bone formation both in extraskeletal sites [47, 48] and in defect models in different species [38, 50]. BMPs originate from the transforming growth factor-β family, including at least 18 proteins [29]. BMP-2, which has been described as an anthelix structure, seems to possess the highest osteoinductive potential among the BMPs [25]. In recent years, the regenerative potential of recombinant human BMP-2 (rhBMP-2) has been demonstrated in various experimental animal studies, including sinus floor augmentation, alveolar ridge preservation, bone augmentation procedures, and periodontal repair [3, 11–15, 31, 33, 49]. Most recently, the effects of rhBMP-2 on the osseointegration of titanium implants have also been investigated in experimental animal studies [14, 26, 33–37, 45]. Experimental titanium plasma-sprayed hollow cylinder implants were filled with a solution of rhBMP-2 soaked on an absorbable type-I collagen sponge before insertion. The histomorphometric analysis revealed a significantly increased bone regeneration in rhBMP-2-treated defects compared to controls. Furthermore, the level of osseointegration, as measured by direct bone-implant contact, was significantly higher for the rhBMP-2 implants compared to controls [34]. Recently, biologically active rhBMP-2 has also been covalently immobilized on metal surfaces [16, 19–21, 44]. Osseointegration of rhBMP-2-biocoated plasma spray-coated titanium–alloy cylinders, as evaluated histomorphometrically 4 weeks after implantation in the distal femur condylus in a gap healing model in sheeps, was predominantly characterized by circumferential bone formation and integration with minimal residual gaps. In contrast, control specimens generally exhibited a wide gap surrounding the implant cylinder [18]. The rationale for BMP immobilization was to avoid ectopic bone formation by a limited and targeted release of rhBMP-2 from the implant surface. Recently, treatment of titanium with chromosulfuric acid (CSA) has been reported to result in ultrahydrophilic bioadhesive surfaces, which in turn improves biocoating with rhBMP-2 [19]. However, there are currently no histological data evaluating osseointegration of CSA-modified and rhBMP-2-biocoated titanium implants. Therefore, the aim of the present study was to investigate histomorphometrically bone formation to CSA-enhanced and rhBMP-2-biocoated titanium implants after implantation in the mandibula and tibia of dogs in comparison to control titanium implants. Materials and methods Animals Two 3-year-old male mongrel dogs (approximate weight 25 kg) were used in the study. Both animals exhibited a fully erupted, healthy, permanent dentition. During the experiment, the dogs were fed ad libitum with soft-food diet and water. Animal selection, management, and surgery protocol were approved by the Animal Care and Use Committee of Belgrade University (ref. no. 2179). The experimental segment of the study started after an adaptation period of 4 weeks. Study design The study was performed in two surgical phases. In the first phase, extraction of the mandibular second, third, and fourth premolar and first molar (P2–M1) was performed bilaterally. After 4 months of healing, surgical implantation of rhBMP-2-biocoated and rhBMP-2-noncoated screw-typed implants was performed in a submerged healing procedure during the second phase. Throughout the study period, oral hygiene procedures were performed three times a week including tooth and implant brushing. Radiographs were obtained before and immediately after tooth extraction as well as immediately after implant installation. Both animals were killed after a healing period of 4 weeks. Implant preparation RhBMP-2 was prepared as previously described [21]. The biological activity of soluble rhBMP-2 was assessed with MC3T3-E1 cells by the induction of the de novo synthesis of alkaline phosphatase (AP) [43]. The half-activation constants (K0.5) were in the range of 20–75 nM [18]. Twenty-four screw-type implants (Camlog Screw Line, Wimsheim, Germany) were manufactured from commercially pure titanium. The core diameter of the implants was 3.3 mm and the total length was 11 mm. A total of six (n=6) implants were sand-blasted and acid-etched according to a standardized procedure (Promote, Altatec, Wimsheim, Germany) (C), while a total of 18 (n=18) implants were surface-enhanced by a novel procedure with CSA [21]. The treatment of metals with CSA (CSA–Ti–alloy) [21] leads to ultrahydrophilic (contact angles 0–10°, no hysteresis) bioadhesive surfaces [17]. A total of 12 (n=12) surface-enhanced implants were divided into two subgroups (A and B) and biocoated with rhBMP-2 [BMP-A: noncovalently immobilized rhBMP-2 (596 ng/cm2), BMP-B: covalently immobilized rhBMP-2 (819 ng/cm2)] [40]. RhBMP-2 was immobilized by covalent and noncovalent methods on these CSA-treated surfaces [18, 21, 40]. In brief, the implants were assigned to the following test and control groups: BMP-A (n=6), BMP-B (n=6), CSA (n=6), and C (n=6). To control the produced surface, the following “sibling method” was employed: Parallel to the preparation of the above dental implants for in vivo experiments, miniplates (10×5×1 mm) with identical Promote surfaces were surface-enhanced with CSA and coated with 125I-rhBMP-2 under identical conditions as the dental implants. In this way, the corresponding contact angles, the amount of immobilized rhBMP-2, and the in vitro biological activity [6] could be tested before the implants were placed into the animals. Only those dental implants were released for implantation, whose sibling miniplates reached the standard as mentioned above and whose surfaces showed an intense in vitro bioactivity by fluorescence microscopy [6]. Surgical procedure for both phases The dogs were anesthetized with 1 mg/kg sodium pentobarbital. To maintain hydration, both animals received a constant rate infusion of lactated Ringer’s solution while being anesthetized. Prophylactic antibiotics were administrated intraoperatively with a combination of 20,000 IU penicillin and 1.0 g streptomycin/10 kg body weight. In the first surgery, P2–M1 were carefully removed after reflection of full thickness mucoperiosteal flaps and tooth separation. After wound closure by means of mattress sutures, the sites were allowed to heal for 4 months. In the second surgery, the test and control implants were randomly allocated to both sides of the mandible (left and right sides, one implant each) and one implant each to either the posterior left or right tibia. In brief, bilateral vestibular incisions were made, and full thickness mucoperiosteal flaps were elevated to expose the respective sites for implant placement in the mandible. Surgical implant sites were prepared bilaterally, at a distance of 10 mm apart, according to the protocol suggested by the manufacturer. All implants in the mandible were sealed with cover screws (Camlog, Wimsheim, Germany). After irrigation, mucoperiosteal flaps were repositioned, and primary wound closure was achieved with consecutive polyglycolic acid 5.0 Polyester sutures (Resorba, Nürnberg, Germany). An area of approximately 15 cm in length and 4 cm in width was depilated on the respective side of the tibia using an electric shaver and a razor blade. After disinfection with polyvidone iodine (Betaisodona, Mundipharma, Limburg/Lahn, Germany), a skin incision was made and a flap was elevated to expose the respective sites of the tibia for implant placement. On the inner-posterior side of the tibias, implants were inserted at a distance of 30 mm apart. All titanium implants were inserted with good primary stability according to a low-trauma surgical technique under copious irrigation with sterile 0.9% physiological saline. After implant placement, the periosteum and fascia were sutured using 3.0 Polyester sutures (Resorba, Nürnberg, Germany). Animal killing and retrieval of specimens The animals were killed (overdose of sodium pentobarbital, 200 mg/kg i.v.) after 4 weeks. The jaws as well as the posterior tibias were dissected, and blocks containing the experimental specimens were obtained. Block sections of the anterior tibia served as additional control (UC). All specimens were fixed in 10% neutral buffered formalin solution for 4–7 days. The specimens were dehydrated using ascending grades of alcohol and xylol, and infiltrated and embedded in methylmethacrylate (MMA, Technovit 7200, Heraeus Kulzer, Wehrheim, Germany) for nondecalcified sectioning. After 18 to 24 h, the specimens were completely polymerized. Each implant site was cut in the mesiodistal direction and along with the long axis of the implant using a diamond wire saw (Exakt, Apparatebau, Norderstedt, Germany), resulting in four sections of approximately 500 μm in thickness [7]. Subsequently, all specimens were glued with acrylic cement (Technovit 7210 VLC, Heraeus Kulzer, Wehrheim, Germany) to opaque Plexiglas and ground to a final thickness of approximately 40 μm. All sections were stained with toluidine blue. Histological and histomorphometric analysis Histomorphometrical analyses as well as microscopic observations were performed by one experienced investigator masked to the specific experimental conditions. For histomorphometrical measurements, images were obtained using a light microscope (BX50, Olympus, Hamburg, Germany) at a magnification of 100×, associated with a video camera (SIS Color View3, Soft imaging System GmbH, Münster, Germany). Digital images were evaluated using a software program (SIS analySIS Auto Software 3.2, Soft imaging System GmbH, Münster, Germany). The percentage of direct contact between mineralized bone and the titanium surface (bone to implant contact—BIC) was measured at every thread on both sides of the implant [23]. BIC was subdivided into crestal and apical values (upper and lower halves of the intraosseous implant length, respectively). Furthermore, in the tibias, induced bone density (BD) adjacent to the titanium surfaces was measured at distances of <1 and 1 mm by assessing the ratio of mineralized bone vs bone marrow within the respective areas [28]. Results Clinical observations The postoperative healing was uneventful in both dogs. No complications such as allergic reactions, abscesses, or infections were observed throughout the study period. Histological and histomorphometric analysis BIC and BD values for each group and respective anatomical sites (mandible and tibia) are presented in Fig. 1a–c. In particular, all test and control implants inserted in the mandible generally exhibited new bone formation in direct contact with the implant interface. The formation of organized trabeculas of woven bone, recognized by osteon formation, could be observed in all groups. However, new bone formation in direct contact with the implant interface appeared to be higher in the BMP groups (Fig. 2a–d). In particular, BIC values appeared to be highest in the BMP-B group, followed by BMP-A, CSA, and C groups. No differences in BIC with respect to crestal and apical values were observed within or between groups (Fig. 1a). Fig. 1Boxplots with outliners for the medians and Q1–Q3 quartiles of BIC and BD (%) in different groups after 4 weeks of healing with respective values in the crestal and apical portion of the implant (BIC) and at a distance of less than and greater than 1 mm adjacent to the implant surface (BD). Lines below and above box plots min, max. control c. a BIC mandibula. b BIC tibia. c BD tibiaFig. 2Histology of representative titanium implants after 4 weeks of healing in the mandible (original magnification ×40). BIC appeared to be highest for BMP-B, followed by BMP-A, CSA, and C. a C. b CSA. c BMP-A. d BMP-B In comparison to the organized trabeculas of woven bone noted in the mandible, bone formation in the tibia seemed to be of a cancellous type, mainly characterized by tiny trabeculas (Fig. 3a–d). BIC values appeared to be highest in the BMP-B group, followed by BMP-A, CSA, and C groups. Again, with respect to crestal and apical BIC values, no differences were observed within groups (Fig. 1b). Histomorphometric analysis of BD in the tibia at a distance of <1 mm revealed obvious differences between groups. In particular, highest values were noted for both BMP groups (BMP-A=BMP-B). This was followed by CSA, also exhibiting higher BD values than C. No differences between groups were found at a distance of >1 mm (Fig. 2c). Fig. 3Histology of representative titanium implants after 4 weeks of healing in the tibia (original magnification ×10). BIC and BD appeared to be highest for BMP-B, followed by BMP-A, CSA, and C. a C. b CSA. c BMP-A. d BMP-B Discussion The present histological study was designed to evaluate bone formation and direct bone apposition to rhBMP-2-biocoated, CSA surface-enhanced, and C titanium implants after implantation in the mandible and tibia of dogs. In particular, rhBMP-2 was immobilized by covalent and noncovalent methods on CSA surface-enhanced titanium implants [18, 21, 40]. Within its limits, histomorphometrical analysis of implants inserted in both mandible and tibia after 4 weeks revealed that BIC values appeared to be highest in the BMP groups, followed by CSA and C groups. Furthermore, it was observed that BD as measured at a distance of <1 mm in the tibia was obviously highest in the BMP-B group, followed by BMP-A, CSA, and C groups. However, no differences between groups were observed at a distance of >1 mm. In this context, it is important to realize that the present pilot study does not have the statistical power to rule out the possibility of a difference between groups. Further experimental studies of higher power are needed to support equivalence or superiority [10]. On the other hand, it needs also to be pointed out that these are the first histological data evaluating bone formation and apposition on rhBMP-2-biocoated and CSA surface-enhanced titanium implants in the mandible and tibia. However, the present findings corroborate, to a certain extent, previous results observed in an ectopic bone formation model [40]. Electropolished titanium miniplates were surface-enhanced by CSA and coated with a total amount of 150–200 ng rhBMP-2. Periosteal flaps were prepared from the anterior surface of the tibias of adult rabbits and wrapped around the titanium specimens. Additionally, some titanium miniplates were inserted to which nonimmobilized soluble rhBMP-2 was added. After 28 days of healing, noncoated specimens revealed bone formation in 2/12 implants, rhBMP-2-coated implants in 6/8, and implants with free rhBMP-2 in 8/8 cases. However, in the case of rhBMP-2-coated implants, the induced bone had direct contact to the implant in all cases. In contrast, titanium miniplates inserted with free administered rhBMP-2 revealed direct BIC in just six cases, whereas in two cases, the titanium surface was separated by a fibrous capsule [40]. The finding that rhBMP-2 may promote periimplant bone regeneration and osseointegration of titanium implants is in accordance with the previous studies [14, 26, 33–37, 45]. In all of these studies, however, rhBMP-2 was admixed with a carrier (i.e., collagen, calcium-phosphate cement carrier), acting as a slow delivery system, for instillation with the implant due to a rapid diffusion of BMP after implantation in vivo. In particular, Sigurdsson et al. [33] evaluated rhBMP-2- (2×0.43 mg/ml in a type-I bovine collagen carrier) induced bone regeneration and osseointegration in a supraalveolar periimplant defect model in dogs. At 16 weeks after healing, bone regeneration (height) was significantly larger for rhBMP-2 than control defects. However, the large amount of BMP-induced bone was poorly adapted to the implant surface. In contrast, the smaller amount of new bone in the control group seemed to be well adapted [33]. In contrast, Wikesjö et al. [45] reported similar BIC values 16 weeks after implantation of rhBMP-2 (0.4 and 0.75 mg/ml in a calcium-phosphate cement carrier) or carrier alone subsequent to a vertical alveolar ridge augmentation procedure and simultaneous implant installation in dogs. Furthermore, Howell et al. [14] applied rhBMP-2 using a collagen sponge carrier to stimulate bone formation in artificially created defects around endosseous implants in the canine mandible. Nonresorbable expanded tetrafluoroethylene (e-PTFE) membranes served as controls. Histological analysis revealed that the addition of rhBMP-2 resulted in a significantly greater amount of new bone area and BIC after 4 and 12 weeks of healing than e-PTFE. Although membrane-treated sites were reported to have less new bone formation after 4 weeks of healing, this difference seemed to be equalized after 12 weeks [14]. Similar results were also reported by Sykaras et al. [34]. RhBMP-2-induced bone regeneration and osseointegration was evaluated in mandibular bony defects created within the hollow chamber of endosseous dental implants in dogs. Before insertion, hollow chambers were filled with 20 μg of rhBMP-2 soaked on an absorbable type-I collagen sponge (0.4 mg/ml). Histological observation revealed statistically significant higher BIC values in the rhBMP-2 group at 8 and 12 weeks after implantation [34]. There might be several explanations for the discrepancies noted in these studies. First of all, it must be emphasized that little information is available on the interaction between rhBMP-2 and the individual carriers. Furthermore, the high concentrations of rhBMP-2 used in these studies (in the milligram per milliliter range) strongly indicate that an optimal method is still lacking. In this context, it is important to point out that BMP-2 has the potency to induce or modulate apoptosis [24], and that in vivo application of high doses of BMP-2 may inhibit bone formation [22, 30]. Furthermore, the results of a recent cell culture study have shown that the dose-dependent effect of rhBMP-2 on AP induction in MC3T3-E1 cells plateaus out into a maximal response at 300–1,000 nM BMP-2 (i.e., 8–25 μg/ml) [43]. Indeed, the results of the present study have shown that BIC and BD values seemed to be ameliorated after application of rhBMP-2 at far lower concentrations. Furthermore, it must also be noted that the effects of rhBMP-2 were limited to a range of 1 mm, outlining that both covalent and noncovalent methods of immobilization seemed to be stable. This finding may also be supported by the observation that no differences with respect to BIC values were observed in crestal and apical areas of the implant surface in all groups. In accordance, it might be hypothesized that from a clinical point of view, both methods of immobilization are suitable to avoid ectopic bone formation due to a limited and targeted release of rhBMP-2 from the implant surface. However, further studies are needed to clarify this issue. Within the limits of the present study, it may be concluded that rhBMP-2 immobilized by covalent and noncovalent methods on CSA-treated implant surfaces seemed to be stable and promoted direct bone apposition in a concentration-dependent manner.	[ "titanium implant", "rhbmp-2", "animal study", "bioactive surface", "ultrahydrophilic surface" ]	[ "P", "P", "P", "R", "R" ]
Skeletal_Radiol-4-1-2271083	Use of FDG-PET in differentiating benign from malignant compression fractures	Objective The objective was to evaluate the use of fluorodeoxyglucose positron emission tomography (FDG-PET) in differentiating benign from malignant compression fractures. Introduction Differentiating between malignant and benign vertebral compression fractures can represent a diagnostic challenge and is particularly difficult in elderly patients, who frequently have a history of malignancy and are also predisposed to benign compression fractures from osteoporosis or treatment-related changes [1–3]. Differentiating between benign and malignant compression fractures has important therapeutic and prognostic implications. Computed tomography (CT) and magnetic resonance imaging (MRI) are routinely used in the evaluation of compression fractures; however, these imaging modalities do not always permit definite diagnosis [2–5]. In contrast to MRI or CT, fluorodeoxyglucose positron emission tomography (FDG-PET) can yield metabolic information that is based on increased glucose metabolism of malignant and inflammatory lesions. Tumor cells typically accumulate FDG, while traumatic fractures are not expected to significantly accumulate FDG. Therefore, FDG-PET may allow differentiation between malignant and benign compression fractures [6–9]. The purpose of our study was to evaluate the use of FDG-PET in differentiating benign from malignant compression fractures. Materials and methods This study was approved by the institutional review board of our institution, which waived the need for informed consent. The study was compliant with the Health Insurance Portability and Accountability Act. A retrospective search was performed using Boolean operators (Folio; Open Market, Proto, UT, USA) to identify all patients who had undergone whole-body FDG-PET at our institution from 2003 to 2006 and had the finding of a compression fracture mentioned in the PET report. Imaging studies and reports, medical records, and pathology reports of selected cases were reviewed. Patients We identified 33 patients with compression fractures who underwent whole-body FDG-PET. There were 23 women and 10 men, aged 48–93 years, with a mean age of 72 years. Twenty-nine patients had a history of malignancy (1 had leukemia, 5 lung cancer, 1 ovarian cancer, 2 breast cancer , 5 colon cancer, 9 lymphoma, 1 sarcoma, 1 pancreatic cancer, 1 Klatskin tumor, 1 gastrointestinal stromal tumor, 1 laryngeal cancer, and 1 had esophageal cancer). The 4 patients without history of malignancy underwent whole body PET for work-up of indeterminate lung nodules detected on prior CT or radiographs. Sixteen patients underwent FDG-PET and 17 patients underwent FDG-PET/CT at time of evaluation. Nine patients underwent CT and 14 patients underwent MRI of the spine within 4 weeks of FDG-PET. Seventeen patients were followed up with serial FDG-PET studies. Nine patients underwent biopsy and the pathologic results were used as a standard of reference. Twenty-four patients who did not undergo biopsy were followed up clinically and with repeat imaging, with MRI, CT, or FDG-PET for a period of 1–3 years and findings at clinical follow-up and imaging studies were used for lesion verification. In these patients, the development of new lesions and/or progression of existing lesions on imaging were used as criteria for malignant compression fractures. Image acquisition Whole-body PET was performed using an ECAT HR+ scanner (CTI Molecular Imaging, Knoxville, TN, USA). All patients fasted for at least 6 h prior to image acquisition, and blood glucose levels were measured prior to the injection of FDG. A dose of 15–20 mCi (555–740 MBq) of FDG was administered intravenously 45 min to 1 h prior to scanning. Patients were positioned supine on the scanner and emission images were acquired in 6–7 bed positions from the mandible to the mid-thigh or to the level of the ankles in the case of lower extremity lesions. Transmission images obtained with a rotating germanium 68-rod source were used for attenuation correction. Images were reconstructed using the ordered-subset expectation maximization (OSEM) algorithm. Combined PET/CT studies were performed with a 16-section hybrid PET/CT gantry (Biograph Sensation 16; Siemens, Erlangen, Germany), which comprises a 16-section high-performance multi-detector row CT scanner with a lutetium oxyorthosilicate-based PET scanner. The PET image spatial resolution was 5.0 mm full width at half maximum, with a section thickness of 3.5 mm. Patients fasted for at least 6 h prior to image acquisition, and blood glucose levels were measured prior to the injection of FDG. Two 10-oz cups of water were administered as negative contrast material 1 h prior to scanning. A dose of 15–20 mCi (555–740 MBq) of FDG was administered intravenously 45 min to 1 h prior to scanning. Patients were positioned supine on the scanner and emission images were acquired in 6–7 bed positions from the mandible to mid-thigh or to the level of the ankles in the case of lower extremity lesions. Images were reconstructed with Fourier rebinning and attenuation-weighted ordered-subsets expectation maximization. A low-dose CT scan was performed prior to PET imaging primarily for attenuation correction, with patients holding their breath mid-expiration, and included an area from the mandible to the mid-thigh or to the level of the ankles in the case of lower extremity lesions. Slice thickness was 5 mm. A diagnostic contrast-enhanced CT was performed subsequent to the PET/CT following the administration of 100 mL intravenous contrast material (Isovue 300; Bracco Diagnostics, Princeton, NJ, USA) at an injection rate of 2 mL/s using 2.5-mm sections. Image analysis Semiquantitative and qualitative evaluation of PET images was performed on a high-resolution workstation (Reveal-MVS; Mirada Solutions, Oxford, UK) by one investigator, who was blinded to the clinical and pathological results (MAB). The images were displayed in rotating maximum intensity projections and in axial, coronal, and sagittal planes. Semiquantitative analysis of FDG uptake was performed by creating a region of interest over the area of maximal radiotracer activity. Maximum standardized uptake values (SUVs) were automatically generated according to the following equation: SUVmax (bw) = Ctis/Dinj/bw, where SUVmax (bw) is maximum SUV normalized for body weight; Ctis, tissue concentration expressed as megabecquerels per milliliter; Dinj, injected dose expressed as megabecquerels; and bw, body weight expressed as kilograms. Lesions with SUV greater than 3.0 were considered malignant and less then 3.0 were considered benign. However, lesions with SUV between 2 and 3 are often indeterminate and do not always allow definite diagnosis, and also the pattern of uptake has to be considered. Qualitative assessment was made with the specific aim of establishing whether the lesion was benign or malignant. The radiotracer uptake by the lesion was compared with the liver, and those lesions with uptake greater than the liver were classified as malignant. Statistical analysis The recorded data were analyzed using JMP statistical database software (SAS Institute, Cary, NC, USA). The findings on the PET images as well as SUV of benign and malignant compression fractures were correlated with the final diagnosis of the lesion and determined either malignant or benign. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were calculated. The Student’s t test was used to determine whether there was a statistically significant difference between the SUV for benign and malignant compression fractures. A difference with p < 0.05 was considered to be statistically significant. Results Forty-three compression fractures were identified in 33 patients. Twenty-two fractures involved the thoracic and 21 fractures the lumbar spine. Nine patients underwent biopsy and 24 patients were followed up clinically and with repeat imaging. In the 9 patients who underwent biopsy, there were 3 benign and 6 malignant compression fractures. Of the 24 patients who were followed clinically and with repeat imaging, 6 patients were thought to have malignant and 18 patients were thought to have benign compression fractures. Overall, there were 21 patients with benign and 12 patients with malignant compression fractures. Out of the 43 compression fractures, there were 29 benign and 14 malignant fractures. Based on clinical history (acute onset of back pain often after minor trauma) and imaging characteristics on MRI (bone marrow edema) there were 5 acute benign compression fractures. Three patients were on bone marrow-stimulating therapy at the time of FDG-PET. Five patients were unable to undergo MRI because of pacemakers (3) or severe pain. Qualitative PET analysis Forty-three compression fractures were identified in 33 patients. Visual inspection of suspected lesions on FDG-PET characterized 26 compression fractures as benign and 17 fractures as malignant. The malignant lesions demonstrated intense radiotracer uptake (Figs. 1, 2), while benign compression fractures showed only mildly increased or no increased uptake on FDG-PET (Figs. 3, 4). Histological and clinical follow-up data were analyzed. By FDG-PET, 12 lesions were correctly classified as malignant (true-positive) and 24 lesions were correctly classified as benign (true-negative). Two malignant tumors were incorrectly classified as benign (false-negative) and 5 benign tumors were incorrectly classified as malignant (false-positive). In the 2 false-negative cases there was moderately increased uptake of the compression fractures, which was thought to represent acute benign fractures prospectively, but was found to be metastatic disease on subsequent biopsy (Fig. 5). SUV in these patients were 2.5 to 2.8 respectively, and the primaries in these cases were esophageal and lung cancer. Fig. 1a Malignant compression fracture in a patient with lymphoma. Sagittal FDG-PET image demonstrates intense radiotracer uptake in the L1 vertebral body (SUV = 7.1; arrow). b Axial FDG-PET image demonstrates tumor infiltration of the L1 vertebral body with associated extra-osseous component (arrows). c Sagittal T1-weighted MRI demonstrates diffuse low signal intensity of the bone marrow of the lumbar spine with compression fracture of L1 (arrow). d Sagittal fat-suppressed T1-weighted MRI following the administration of gadolinium demonstrates diffuse enhancement of the L1 compression fracture (arrow). e Axial image from a CT-guided biopsy demonstrates large extra-osseous soft tissue component (white arrow) and infiltration of the L1 vertebral body (black arrows)Fig. 2a Malignant compression fracture from metastatic colon cancer. Sagittal fused FDG-PET/CT image demonstrates intense radiotracer uptake in the T7 compression fracture (SUV = 6.3; arrow). b Axial fused FDG-PET/CT image demonstrates focal increased radiotracer uptake in the T7 vertebral body (arrow). c Sagittal T1-weighted MRI demonstrates compression fracture of T7 with low signal marrow infiltration (arrow). d Sagittal STIR image demonstrates mild hyperintensity of the T7 vertebral compression fracture (arrow). e Sagittal fat-suppressed T1-weighted MRI following the administration of gadolinium demonstrates diffuse enhancement of the T7 compression fracture (arrow). Malignancy was indeterminate on MRIFig. 3Benign compression fracture in a patient with lymphoma. Sagittal fused FDG-PET/CT image demonstrates compression fracture of T9 (arrow) without significant radiotracer uptake (SUV = 0.7)Fig. 4a Benign subacute compression fractures in a patient without history of malignancy. Sagittal FDG-PET image demonstrates mildly increased radiotracer uptake in the upper thoracic spine (SUV = 1.8; arrow). b Axial FDG-PET image at the level of T4 demonstrates mildly increased radiotracer uptake (arrow), consistent with subacute compression fracture. c Sagittal CT image demonstrates compression fracture of T4 (arrow) and multiple chronic osteoporotic compression fractures throughout the thoracic spineFig. 5a False-negative compression fracture on FDG-PET in a patient with metastatic esophageal cancer. Sagittal fused FDG-PET/CT image demonstrates mild to moderately increased radiotracer uptake in the L2 compression fracture (arrow; SUV = 2.5) that was thought to represent a benign compression fracture but was found to be malignant on subsequent biopsy. Note the mildly increased radiotracer uptake in the T11 compression fracture (arrowhead), which was thought to represent a benign fracture. b Axial fused FDG-PET/CT image demonstrates increased radiotracer uptake in the L2 vertebral body (arrow). c Sagittal CT image demonstrates L2 and T11 compression fractures (arrows), for which malignancy was indeterminate Three of the false-positive patients were on bone marrow-stimulating agents (Fig. 6). One of the false-positive patients had an acute compression fracture (Fig. 7). Except for that patient, there was no significant difference in uptake pattern of benign acute and benign chronic compression fractures (Fig. 8). Fig. 6a False-positive compression fracture in a patient on bone marrow-stimulating agents. Sagittal FDG-PET image demonstrates diffuse increased radiotracer activity throughout the spine. Note the compression fractures with focal kyphotic angulation at L1–L2 (arrow). The patient was thought to have diffuse metastatic disease with malignant compression fracture. b Sagittal FDG-PET image performed 2 months later demonstrates resolution of diffuse radiotracer uptake. Kyphotic angulation from compression fractures remains (arrow). The patient was off bone marrow-stimulating agents for 6 weeks at the time of the studyFig. 7a False-positive compression fracture in a patient with laryngeal cancer. Sagittal fused FDG-PET/CT image demonstrates increased radiotracer uptake in the T12 compression fracture (SUV = 4.9; arrow). b Sagittal T1-weighted MRI demonstrates compression fracture of T12 with linear low signal of the superior endplate and preservation of the normal marrow signal (arrow). Note compression fracture of L1 with relative preservation of normal marrow signal (arrowhead). c Sagittal STIR image demonstrates hyperintensity of the superior endplates of the T12 and L1 compression fractures, suggesting acute fractures (arrows). d Sagittal fat-suppressed T1-weighted MRI following the administration of gadolinium demonstrates enhancement of the T12 compression fracture (arrow). Mild enhancement is noted, involving the L1 compression fracture (arrowhead). Findings were thought to be benign on MRI and on follow-up imaging, there was resolution of bone marrow edema and enhancementFig. 8a Acute and chronic benign compression fractures in a patient without history of malignancy. Sagittal fused FDG-PET/CT image demonstrates multiple compression fractures throughout the thoracic and lumbar spine with mild increased radiotracer uptake at T11 (SUV = 2.1; arrow). b Sagittal T1-weighted MRI demonstrates multiple chronic compression fractures (arrowheads). Mild decreased signal intensity of the T11 compression fracture suggests edema (arrow). c Sagittal STIR image demonstrates mild edema in the T11 compression fracture suggestive of an acute fracture (arrow). There is no evidence of marrow edema in the remaining compression fractures Quantitative PET analysis Standardized uptake values were measured in all 43 compression fractures. SUV for benign compression fractures ranged from 0.7 to 4.9 with a mean of 1.94 ± 0.97 standard deviation (SD) on FDG-PET. Malignant compression fractures showed SUV from 2.2 to 7.1 with a mean of 3.99 ± 1.52 SD on FDG-PET. The difference between the SUV values of benign and malignant compression fractures was statistically significant (p < 0.001, Student’s t test; Fig. 9). Fig. 9Mean standardized uptake value (SUV) of benign and malignant compression fractures. The top of the boxes represent the mean and error bars represent the range of SUV. There is a statistically significant difference between the SUV values of benign and those of malignant compression fractures (p < 0.001, Student’s t test) Statistical analysis Sensitivity, specificity, positive and negative predictive values, and accuracy of FDG-PET in differentiating benign from malignant compression fractures were 86%, 83%, 84%, 71%, and 92% respectively. Discussion The correct diagnosis of benign and malignant compression fractures can be problematic, but has important prognostic and therapeutic implications. MRI, CT, or bone scintigraphy are commonly used in the diagnostic work-up of patients with compression fractures. However, in some cases, these imaging techniques do not permit definite diagnosis of the cause of the compression fracture. Diffusion-weighted MRI has been used successfully to differentiate benign from malignant fractures [10–13], but benign and malignant compression fractures can show significant overlap on quantitative assessment with apparent diffusion coefficient maps [12]. Also, some patients are unable to undergo MRI due to pacemakers, claustrophobia, or severe pain. This is especially true in the elderly population. Five of our 33 patients were unable to undergo MRI. In this context, metabolic imaging modalities such as FDG-PET might be used as a management problem solver. FDG-PET has been successfully used to differentiate benign from malignant neoplasms and in the evaluation of metastatic disease [14–16]. Preliminary studies and case reports have shown that FDG-PET might be helpful in differentiating benign from malignant compression fractures [8, 9]. Since elderly patients often have a history of malignancy and are also predisposed to benign compression fractures due to osteoporosis, we thought that FDG-PET might be used as a diagnostic problem solver in this patient population in cases of equivocal MRI or CT findings or if the patient was unable to undergo MRI. In our study, malignant compression fractures demonstrated significantly increased FDG uptake compared with benign fractures. Mean SUV of malignant and benign fractures were 3.99 ± 1.52 SD for malignant lesions and 1.94 ± 0.97 SD for benign lesions. There were 2 false-negative results. In these cases there was moderately increased uptake of the compression fractures, which were thought to be acute benign fractures prospectively, but were found to represent metastatic disease on subsequent biopsy. SUV in these patients were 2.5 to 2.8 respectively. There were 5 false-positive results; 3 of those patients were on bone marrow-stimulating agents, which mimicked tumor involvement. This effect on bone marrow FDG uptake has been described in the literature [17]. One to 2 months after cessation of bone marrow-stimulating therapy, FDG uptake returned to normal in our patients. The false-positive results occurred because the investigator was blinded to the clinical history. This is less likely to occur in a clinical setting where this history should be actively sought when interpreting positive findings, especially in cases of diffuse osseous uptake, mimicking a diffuse marrow infiltrative process as was seen in our cases. In fact, when we reviewed the original reports, these cases were reported to be consistent with the known use of bone marrow-stimulating agents. Only mildly increased or no increased uptake was seen in chronic benign fractures and mild to moderately increased uptake was seen in acute benign fractures, which was less than in malignant fractures. This is in contrast to bone scintigraphy, in which increased uptake persists for many months [18]. Positron emission tomography and PET/CT were equally sensitive in differentiating benign from malignant compression fractures. However, the CT portion of the PET/CT improved the exact fracture localization and was able to provide additional information on fracture morphology, which can be helpful in diagnosing benign vs. malignant fractures. Our study had several limitations. The first is the retrospective nature of the study. Second is the lack of histologic correlation in all cases. Only 9 patients underwent biopsy of the spine. However, we obtained clinical follow-up and serial imaging studies, including serial FDG-PET for a period of 1–3 years to evaluate for benign or malignant fractures. Also, there were no patients with osteomyelitis/discitis, which can mimic malignancy [8, 19, 20]. Another limitation is the lack of inter-observer variability, since only one observer interpreted the images. However, we compared our results with the reports of the original interpreter, who had access to all data. On the original reports, the 3 cases of patients on bone marrow-stimulating agents were interpreted as being consistent with the known use of bone marrow-stimulating agents. Malignancy was said to be indeterminate in the two false-negative reports. However, in our review, we categorized each fracture as either benign or malignant without the option of an indeterminate lesion. In summary, FDG-PET is a useful method of differentiating between benign and malignant compression fractures and can serve as a problem solver in cases of equivocal MRI or CT findings, and in patients who are unable to undergo MRI. We do not recommend FDG-PET as a screening test, but rather as an additional imaging modality in problem cases, particularly in elderly patients with osteoporosis and a history of malignancy. In these patients, FDG-PET has the additional advantage of being able to evaluate the entire skeletal system and screen for metastatic disease.	[ "fdg-pet", "benign", "malignant", "compression fractures", "spine" ]	[ "P", "P", "P", "P", "P" ]
Int_J_Legal_Med-3-1-2039830	Variables affecting the probability of complete fusion of the medial clavicular epiphysis	In this study, we have combined data on clavicle fusion from different studies and applied a binomial logistic regression analysis. As such, we aimed to assess whether or not variables such as sex, socioeconomic status, and ethnicity influence the probability of having mature, i.e., completely fused clavicles at a given age. We further explored whether the method of clavicle examination, i.e., diagnosis from either a dry bone specimen, an examination of X-rays, or an examination of computed tomography scans, affects the probability of being diagnosed with mature clavicles. It appeared that only ethnicity did not significantly affect this probability. Finally, we illustrated how the logit model may be used to predict the probability of being diagnosed with mature clavicles. Introduction The medial (or sternal) clavicular epiphysis matures relatively slowly. In the human skeleton, it is generally the last long bone epiphysis to fuse. Its developmental stage is therefore useful when estimating age at death in the post-pubertal period [1] or the age of living persons involved in criminal proceedings [13, 15]. In the Netherlands, a radiological clavicle examination may be applied to verify the age of young asylum seekers without valid identification documents. Various authors have studied clavicle development to determine the age interval in which fusion of the clavicular epiphyses occurs. Most of the published studies were based on the examination of dry bone specimens [3, 6, 9, 10, 17, 18]. When dealing with living individuals, examiners analyzed either X-rays [4, 13] or computed tomography (CT) scans [5, 15, 16]. In a number of studies, the results for different ethnic groups were evaluated or the results for males were compared with those for females. In most studies, such comparisons were, however, hampered by a relatively small sample size. In this study, we aim to assess from a large sample whether or not variables such as sex, socioeconomic status, and ethnicity affect the probability of having completely fused clavicles at a certain age. We further aim to compare results from X-ray and CT scan examinations with those from studies based on dry bone specimens. To do so, we have combined data from different studies and performed a binomial logistic regression analysis. We then illustrate how the resulting model can be applied to estimate the probability of completely fused clavicles at a given age. Data Published and unpublished data on the individual subjects, taken from various studies on clavicle development [3–6, 8, 10, 13, 15, 18] were combined. For each individual, information on the calendar age in years was combined with data on the fusion status of the clavicle, i.e., exhibiting a complete bony fusion of epiphysis and diaphysis (i.e., mature) or not. When data for both left and right clavicles were published, we used the side for which most records were provided. Data on the sex and ethnicity of the individuals were copied when available. In the context of this study, we defined ethnicity as being either of European (Caucasian), Asian, or Afro-American descent. It was further recorded which method was used for determining the fusion status, i.e., using dry bone specimens, X-rays, or CT scans. The socioeconomic status of individuals was estimated from the human development index (HDI) associated with the country in which the study took place. The HDI is a summary composite index that measures a country’s average achievements in three basic aspects of human development: longevity, knowledge, and the standard of living [19]. The HDI is used by the United Nations Development Programme, and data are available from 1975 onward. To accommodate data from two American studies that took place before this year [6, 18], we extrapolated the HDI by first calculating the linear regression of the HDI on time. In total, data of 3,552 individuals aged 15–33 years were combined. Figure 1 shows the age distribution within the sample. Table 1 provides further information on the number of individuals per category. Fig. 1Age distribution of individuals in the sampleTable 1Number of individuals per categoryCategoryValuesTotal number of individuals3,552SexMales2,133Females1,123Unspecified296EthnicityaEuropean descentb1,316Afro-American descent287Asian descent695Unspecified1,254MethodDry bone specimen1,374X-ray1,326CT scan852aA small number of individuals in McKern and Stewart [6] were of Afro-American descent. As no individual data were provided, all 374 individuals are classified “of European descent”.b“Of European descent” includes Caucasian and Latino individuals from North America. Statistical analysis Analysis of the data was performed using the SPSS (version 10.0) software package for statistical analysis. We carried out a binomial logistic regression analysis to assess the effect of various variables on the probability of mature clavicles. In logistic regression, the dependent variable is transformed into a logit variable, i.e., the natural log of the odds of the dependent occurring or not. This transformation ensures that the estimated probabilities are between 0 and 1. A logit model is a form of the generalized linear model. Age, sex, ethnicity, and socioeconomic status were considered as potential predictor variables for the probability of mature clavicles. The method of clavicle examination was further considered as a factor influencing the probability of being diagnosed with mature clavicles. Age and socioeconomic status (i.e., HDI) were analyzed as covariates. Sex, ethnicity, and the method of clavicle examination were analyzed as fixed factors. Possible interaction between these fixed factors and the covariate age was also investigated. The initial (exploratory) logit model may be described by the following equation: where p denotes the probability of the dependent, α the intercept, β the estimated effect of a covariate or factor-by-covariate interaction term on the logit of the probability, and γ the estimated effect of a fixed factor. Variables to be kept in the model were selected by means of backward elimination using the Wald test. Insignificant interaction terms were successively dropped. After each alteration to the exploratory model, possible significance of the remaining variables was reassessed. A p value lower than 0.05 was considered to be statistically significant. Not all studies that were used as a source for our data provided individual data on sex or ethnicity. Consequently, the described model (M1) could be tested on the data of a subgroup of 2,002 individuals only. No data from CT scans were available for this subgroup. To be able to include data from all available studies, we therefore also tested simplified models. For the first of these simplified models (M2), we omitted the variable ethnicity and the associated interaction term from the initial model. This model could be fitted to the data of a subgroup of 3,256 individuals. For this subgroup, data resulting from the evaluation of CT scans were drawn from a single study [15], as the second available study [5] did not provide details on the sex of individuals. As a next step, we therefore further omitted the variable sex and the associated interaction term from the initial model. The resulting model (M3) could be fitted to the data of all 3,552 individuals. As a final step, we assessed the significance of the change in −2 log likelihood between the resulting model and a nested model including the main terms only. We further compare the Nagelkerke pseudo R2. As such, we compare the difference in model fit. Results Table 2 provides the results for the three different models. Model M1 resulted in a p value higher than 0.05 for ethnicity after all non-significant interaction terms had been removed. Tested for a subgroup of 2,002 individuals, ethnicity therefore did not appear to affect the probability of having mature clavicles. Including this variable in the model does not improve the model fit. Exploring model M1 therefore effectively resulted in model M2, which could be fitted to the data of 3,256 individuals. Table 2Estimated parameters of logistic regression models for the logit of the probability of mature claviclesModelIndependent variablesBStandard errorp valueM1 N = 2,002Ethnicitya0.355M2 N = 3,256Age0.6290.0340.000Sex (ref: male)0.026 Female0.3680.166Human development index2.7460.5900.000Method (ref: dry bone specimen)0.000 X-ray−21.0463.055 CT scan−7.6242.346Method × age0.000 X-ray1.0290.135 CT scan0.4230.104Constant−18.6381.0020.000M3 N = 3552Age0.6300.0340.000Human development index3.0120.5840.000Method (ref: dry bone specimen)0.000 X-ray−20.7293.060 CT scan−7.5261.928Method × age0.000 X-ray1.0230.135 CT scan0.4010.083Constant−18.8030.9980.000aThe remaining parameters of model 1 (M1) are omitted from this section in the table, as final results for a model without the variable ethnicity are provided for model 2 (M2). From model M2, it appeared that age, the HDI (i.e., socioeconomic status), sex, and the method of clavicle examination all significantly affected the probability of having completely fused clavicles. Exploring possible interaction between sex and age did not indicate a significant effect. The interaction term method × age, however, yielded a p value of 0.000. The subsequent assessment of the significance of the difference in −2 log likelihood between the resulting model and a nested model omitting the interaction term yielded a p value of 0.000. The Nagelkerke pseudo R2 reduced from 0.815 to 0.795 after omission of the interaction term. All indicates a better model fit when the interaction term is included. Model M3, which could be fitted to all available data on clavicle fusion, resulted in comparable model parameters for the dummy variable for CT scans. Confounding by an observer effect is therefore reduced. Application of the model The most appropriate model to predict the probability of being diagnosed with mature clavicles, i.e., providing the best model fit to the currently available data, appeared to be model M2. The resulting model—i.e., after the removal of non-significant main and interaction terms and using the parameters for covariates and the intercept—may be described by: Parameters for the appropriate dummies for the categorical variables may be found in Table 2. From this equation, it follows that the predicted probability may be calculated using the following equation: Based on the currently available data on clavicle fusion, we can now predict the probability of having mature clavicles for different individuals. Here, we provide a few examples. The predicted probability of mature clavicles is 0.016 (1.6%) for a contemporary 19-year-old German male (HDI is 0.932). For a 20- or 21-year-old German male, this probability increases to 0.029 (2.9%) and 0.054 (5.4%), respectively. For females, the predicted probability values are increased to 0.023 (2.3%), 0.042 (4.2%), and 0.076 (7.6%), respectively. However, if the individuals have a lower socioeconomic status, these values are decreased. For contemporary 19-, 20-, and 21-year-old males from India (HDI is 0.611), for instance, they are 0.007 (0.7%), 0.012 (1.2%), and 0.023 (2.3%), respectively. When diagnosed from X-ray or CT scan, the values increase. For the 21-year-old Indian male, it would be 0.040 (4%) if diagnosed from an X-ray and 0.076 (7.6%) if diagnosed from a CT scan (composed using a slice thickness that is far from optimal). In Figs. 2 and 3, we show the differences in predicted probabilities at different ages between the various categories. In Fig. 2, we compare the predicted probability values for a German male with that for a German female and an Indian male. Figure 3 illustrates the difference in predicted probability for a German male depending on whether dry bone specimens, X-rays, or CT scans are examined. Fig. 2Predicted probability of having completely fused clavicles: a comparison between German males, German females and Indian malesFig. 3Predicted probability of being diagnosed with mature clavicles: a comparison between examination by means of dry bone specimens, X-rays and CT scans Discussion Ethnicity did not appear to significantly influence the probability of having completely fused clavicles, which is in line with the results of Schmeling et al. [12]. We should, however, keep in mind that the current database is not well balanced. In addition, a small number of individuals taken from McKern and Stewart [6] were misclassified in our analysis (see footnote Table 1). Additional data from future studies may contribute to the knowledge on the possible effect of ethnicity, if not strengthen our conclusion that there is no significant effect. The effect of sex on the probability of having completely fused clavicles appeared significant. The estimated model parameter for the dummy variable for females being greater than zero indicates a positive effect on the logit of the probability. The predicted probability of mature clavicles is therefore greater for females than for males of similar ages. This would point to a generally earlier onset of fusion in females. A similar disparity has been reported for epiphysial closure of all other long bones [11]. There was no indication of significant interaction between sex and age. We may therefore assume similarly shaped probability curves for males and females, indicating an equal number of years during which complete fusion may generally be achieved. A higher HDI, and therefore a higher socioeconomic status, appears to increase the probability of having mature clavicles. This would mean that under less preferable socioeconomic conditions, the process of fusion is generally delayed. This is in concordance with conclusions drawn by Schmeling et al. [12, 14], and was summarized before by Eveleth and Tanner [2]. The method of clavicle examination appeared to also be significant. The significant interaction term method × age indicates that we should assume a difference in the estimated effect of age depending on the method applied. For the age interval in which mature clavicles have been observed, the predicted probability of being diagnosed with mature clavicles is greater when X-rays or CT scans are used instead of dry bone specimens. It is noteworthy that the difference in the predicted probability is initially very small, but becomes increasingly greater with increasing age during a limited age interval. When keeping in mind that the curve for the predicted probability will be s-shaped, one may image that the difference is again only minimal for the older individuals. A number of theories can be postulated to explain a difference in the predicted probabilities between methods of examination. Inappropriate slice thickness Regarding the results from CT scans, one could postulate that the slice thickness affected the diagnosis. Mühler et al. [7] concluded that reconstructions composed with a slice thickness of 1 mm should be used. The vast majority of the analyzed data were, however, obtained from reconstructions with a greater slice thickness (up to 8 mm) [5, 15]. One can foresee that an almost completely fused clavicle may be diagnosed as being mature when details are lost due to the greater slice thickness. Schulze et al. [16]—from whom no data were included in our analysis, as no individual data were provided—observed two 19-year-olds with completely fused clavicles. Both observations were done from reconstructions with a slice thickness of 7 mm (Schulze, personal communication). We therefore think there is a realistic possibility of these cases having been misdiagnosed. Persistent small grooves or notches between diaphysis and epiphysis When comparing the results from their study based on CT scans with two studies based on dry bone specimens [6, 9], Kreitner et al. [5] hypothesized that persistent small grooves or notches seen in dry bone specimens may be classified as partial fusions of the epiphysis with the clavicular shaft. This would reduce the probability of dry bone specimens being classified as completely fused. Reduced visibility of the epiphyseal disc on X-rays X-rays used in the published studies have not been made with the purpose of examining the clavicles. Possibly, this has hindered a correct classification of the clavicle at times. Interobserver variability As the sample included results from two studies based on X-rays and two studies based on CT scans, we hold it unlikely that what we perceive to be an effect of the applied method is in fact an observer effect. Expectation bias In some X-rays, the (nearly fused) epiphyseal disc will have been more difficult to judge than in others. In such cases, knowing the actual age of the individuals could have affected the interpretation of the observation. Clavicles of relatively young individuals would then more often be classified as “not completely fused”, while those of relatively old individuals would be classified as “completely fused” more often. We know, however, that in at least two out of four studies, one based on X-rays [13] and one on CT scans [15], the examiner was not aware of the age of the subjects when classifying the images. We therefore do not believe that expectation bias could have played a role of any great importance. Statistical artifact In theory, it may be possible that due to an unfortunate coincidence, there were no relatively late maturing individuals examined by X-ray or CT scan.In theory, all of these aspects could have contributed to the significant effect of the variable method. However, if significance of the interaction between method and age is not explained as a statistical artifact, only expectation bias would have resulted in a significant change in the shape of the probability curve. In case of misclassifications of nearly fused clavicles or differences in the classification system among studies, we would have expected a higher number of individuals with completely fused clavicles in the relatively young subjects, even below the age of 20.If we, however, were to conclude that significance of the interaction term is in fact a statistical artifact, the term should be removed from the model, and model parameters should be recalculated. It is our ambition to investigate the matter further by comparing high quality CT scans, expected to allow a correct diagnosis, with X-rays taken from the same person. Conclusions The probability of having mature clavicles at a certain age is affected by the sex and socioeconomic status of the individual. The probability of being diagnosed with mature clavicles further appears affected by the method of clavicle examination. Predicted probabilities from the model may be used to illustrate acceptance or rejection of individual age claims. Please note that the odds on having mature clavicles given a certain age should not be confused with the (posterior) odds of having reached a certain age given that the clavicles have matured. These probabilities are only equal if we assume a prior odds of 1.	[ "clavicle", "age estimation", "medial epiphysis", "epiphyseal fusion" ]	[ "P", "R", "R", "R" ]
Arch_Dermatol_Res-2-2-1705535	The incidence of arthropathy adverse events in efalizumab-treated patients is low and similar to placebo and does not increase with long-term treatment: pooled analysis of data from Phase III clinical trials of efalizumab	A large-scale, pooled analysis of safety data from five Phase III clinical trials (including open-label extensions of two of these studies) and two Phase III open-label clinical trials of efalizumab was conducted to explore whether arthropathy adverse events (AEs) were associated with efalizumab treatment in patients with moderate-to-severe chronic plaque psoriasis. Data from patients who received subcutaneous injections of efalizumab or placebo were stratified for analysis into phases according to the nature and duration of treatment. These included: the ‘first treatment’ phase (0–12-week data from patients who received either efalizumab, 1 mg/kg once weekly, or placebo in the five placebo-controlled studies); the ‘extended treatment’ phase (13–24-week data from seven trials for all efalizumab-treated patients); and the ‘long-term treatment’ phase (data from efalizumab-treated patients who received treatment for up to 36 months in two long-term trials). Descriptive statistics were performed and the incidence of arthropathy AEs per patient-year was calculated using 95% confidence intervals (CIs). During the first treatment phase, a similar proportion of patients had an arthropathy AE in the efalizumab group (3.3%; 58/1740 patients) compared with the placebo group (3.5%; 34/979 patients); the incidence of arthropathy AEs per patient-year was 0.15 in the efalizumab group (95% CI 0.11–0.19) and 0.16 in the placebo group (95% CI 0.11–0.22). Analysis of first treatment phase data from one study (n = 793) showed that the incidence of psoriatic arthropathy per patient-year was lower in efalizumab-treated patients (0.10; 95% CI 0.05–0.18) than in those given placebo (0.17; 95% CI 0.08–0.30). During the extended treatment phase, the incidence of arthropathy remained low (0.17; 95% CI 0.14–0.22). Data from two long-term studies showed that there was no increase in the incidence of arthropathy AEs over time in patients treated with efalizumab for up to 36 months. Patients who had an arthropathy AE during treatment with efalizumab appeared to be more likely to have a history of arthropathy prior to treatment. Efalizumab does not appear to increase the risk of arthropathy AEs compared with placebo. Introduction Psoriasis is a chronic, immune-mediated, inflammatory skin disorder that is currently incurable. Consequently, the majority of people with psoriasis require long-term treatment to maintain disease control. Traditional immunosuppressive systemic treatments, such as acitretin, methotrexate, cyclosporine, hydroxyurea, and thioguanine, may be effective in controlling psoriasis in some patients but significant toxicity and the need to closely monitor patients limit the viability of these treatments for long-term, continuous use [23]. Recently developed systemic therapies that selectively target specific pathways in the inflammatory cascade of psoriasis generally have a much improved safety profile compared with traditional therapies [26]. Efalizumab (anti-CD11a; Raptiva®) is a recombinant humanized monoclonal IgG1 antibody that has been approved for the treatment of moderate-to-severe chronic plaque psoriasis. It interferes with the pathogenesis of psoriasis via multiple mechanisms, including inhibition of T-lymphocyte trafficking and T-lymphocyte activation and reactivation [1, 10, 11, 21, 25]. The safety and efficacy profile of efalizumab has been established in numerous clinical trials, in which more than 3,500 patients were enrolled and treatment was assessed for up to 3 years [4–6, 12–17, 22]. Although psoriasis can be associated with the co-morbidity of psoriatic arthritis, a minority of patients with psoriasis (7–30%) will develop this joint disease [27]. Nevertheless, psoriatic arthritis constitutes a major consideration in patients who are receiving long-term treatment for their psoriasis. A Nordic study of more than 5,000 patients with psoriasis showed that patients with arthritis exhibited greater impairment of psoriasis-related quality of life (QoL), longer disease duration, and greater self-reported disease severity, compared with patients who had psoriasis but no co-morbid arthritis [27]. A low incidence of arthropathy adverse events (AEs; any form of joint disease) associated with efalizumab treatment has been reported in both clinical studies and routine clinical practice [8, 12]. However, anecdotal reports of arthropathy in routine clinical practice have expressed concern that efalizumab may be associated with exacerbation of arthropathy [8]. To address this concern, we conducted a large-scale pooled analysis of safety data from five Phase III clinical trials (including open-label extensions of two of these studies) and two Phase III open-label clinical trials of efalizumab to explore whether arthropathy AEs were associated with efalizumab treatment in patients with psoriasis. Methods The primary objective of this pooled safety analysis was to assess the incidence of arthropathy AEs in patients who had received either efalizumab or placebo. Safety data were pooled from five randomized, double-blind, placebo-controlled clinical trials (including data from two open-label extension studies of two of these trials) and two open-label clinical trials of efalizumab [4–6, 12–17, 22]. Patients included in these Phase III studies were aged ≥18 years and had moderate-to-severe chronic plaque psoriasis, a psoriasis area and severity index (PASI) score of ≥12 at screening, and plaque psoriasis covering ≥10% of body surface area. All patients were candidates for either systemic anti-psoriatic therapy or had received systemic anti-psoriatic therapy. Patients included in these trials received subcutaneous injections with efalizumab, 1–4 mg/kg once weekly or 2 mg/kg once-every-other week, or placebo. Details of individual study methodologies are described in other publications [4–6, 12–17, 22]. Arthropathy AEs were defined according to the Coding Symbols for Thesaurus of Adverse Reaction Terms (COSTART) [3] preferred terms ‘arthritis’ and ‘arthrosis’, or the Medical Dictionary for Regulatory Activities (MedDRA, http://www.meddramsso.com/NewWeb2003/index.htm) preferred terms ‘arthritis not otherwise specified (NOS)’, ‘psoriatic arthropathy’, ‘arthropathy NOS’, ‘monoarthritis’, ‘polyarthritis’, and ‘osteoarthritis NOS’. Treatment groups analyzed Due to the variety of study designs, five analyses were considered: ‘first-treatment phase’, ‘first exposure phase’, ‘extended treatment phase’, ‘re-treatment phase’, and ‘long-term treatment’ (see Table 1). Table 1Summary of the Phase III data from five placebo-controlled clinical trials (including data from two open-label extension studies of two of these trials) and two open-label clinical trials of efalizumab included in the pooled safety analysisPublication (protocol number)Study designNumber of patients in each analysisFirst treatment (0–12 weeks)Efalizumab sc 1–4 mg/kg qw or 2 mg/kg qowFirst exposure Extended treatment (13–24 weeks)Long-term treatmenta (≤36 months)Re-treatmentPlaceboEfalizumab 1 mg/kgEfalizumab 2 mg/kgLeonardi [13] (ACD2058g)Randomized, double-blind, parallel-group, placebo-controlled170162166462123–55Lebwohl [12] (ACD2059g)Randomized, double-blind, parallel-group, placebo-controlled 122232243579289––Gordon [4] (ACD2390g)Randomized, double-blind, parallel-group, placebo-controlled 187368–368–––Papp [17] (ACD2600g)Randomized, double-blind, parallel-group, placebo-controlled 236449–449–449–Sterry [22] (IMP24011)Randomized, double-blind, parallel-group, placebo-controlled264529–772308–145Papp [16] (ACD2062g)Open-label–––34137–365Gottlieb [5, 6] (ACD2243g)Open-label–––339b290339–Menter [14] (ACD2391g)Open-label extensionc of study ACD2390g [4]–––174342––Menter [15] (ACD2601g)Open-label extensionc of study ACD2600 g [17]–––217622635d–Pooled analysis 97917404093,3942,111n.a.e565qow once-every-other week, qw once weekly, sc subcutaneousaNumber of patients at startbPatients received combined therapy with fluocinolone acetate (n = 169) or petrolatum (n = 170) for weeks 9–12; for months 3–15, the dose of efalizumab could be escalated to 4 mg/kg per week for up to 4 weeks if clinically indicatedcSome patients are included in the analyses more than once because patients in the open-label extension studies are also included in analyses of the parent studiesdIncluded patients who received either efalizumab or placebo in the parent study [17]eNot applicable because data are analyzed and reported separately for the study by Gottlieb et al. [6] and the study published by Papp [17] and Menter [15] It is worth noting that most of the studies included in this pooled analysis were designed and conducted before efalizumab had received regulatory approval and before it was known that doses of more than 1 mg/kg once weekly (the approved dose) did not confer additional treatment benefit (EMEA, Raptiva Summary of Product Characteristics; FDA US, FDA Prescribing Information for Raptiva). For this reason, only the efalizumab 1 mg/kg once-weekly dose data are reported for the ‘first treatment phase’ of the analysis. Due to the wide variety of study designs included in the pooled analysis, data for patients receiving any dose of efalizumab are combined for all other treatment phases analyzed. The ‘first treatment phase’ analysis included 0–12-week data from patients in the five placebo-controlled studies who received either efalizumab 1 mg/kg once weekly or placebo. This analysis allows a comparison between the efalizumab and placebo treatment groups. The ‘first exposure phase’ included 12-week data from all studies in patients who had their first exposure to any dose of efalizumab and, thus, did not include placebo data. This analysis was conducted to include the maximum number of patients who received efalizumab for their first 12 weeks of treatment (i.e., it included those patients who first received efalizumab treatment after crossing over from a placebo group, as well as the patients who first received efalizumab during weeks 0–12). The ‘extended treatment phase’ analysis included 13–24-week data in patients given any dose of efalizumab who had already received efalizumab during the first treatment phase. The ‘long-term treatment phase’ analysis included all patients who received continuous long-term treatment (up to 36 months) with any dose of efalizumab. Data were analyzed in 12-week segments to assess change in the incidence of arthropathy AEs over time. This analysis included data from two long-term studies [6, 15, 17], which were analyzed separately due to differences in study design. The ‘re-treatment phase’ analysis included all patients who re-started treatment with efalizumab following a treatment-free observation period. Statistical analyses Descriptive statistics were used to explore the association between efalizumab and the occurrence of arthropathy AEs. Results are expressed as point-estimates of the incidence rates (ratio of the number of patients with an arthropathy AE to the total number of patient-years at risk of an arthropathy AE) with their 95% confidence intervals (CIs). Descriptive comparisons are provided; no formal statistical tests were performed. Analyses were also conducted to explore the relationship between onset of arthropathy AEs during efalizumab treatment and a previous history of arthropathy (reported as a narrative by patients at the baseline visit) and the incidence of arthropathy AEs and clinical response to efalizumab treatment [measured at 12 weeks using the Physician Global Assessment (PGA) and PASI scales]. Differences in patient and psoriasis characteristics at baseline were also compared between patients who had arthropathy AEs and those who did not. An additional analysis of data from patients included in the first treatment phase of the study by Sterry et al. [22] (Table 1) was conducted to assess the incidence of psoriatic arthropathy in these patients. This was the only study to define arthropathy AEs according to MedDRA; other studies used the COSTART, which did not include ‘psoriatic arthropathy’ specifically as a preferred term. Baseline demographics and psoriasis characteristics and the proportion of patients with a previous history of arthropathy (as reported by patients at the baseline visit) were tabulated by presence/absence of an arthropathy event. Results The number of patients included in each of the pooled safety analyses from each of the seven trials and two open-label extensions is summarized in Table 1. Up to 3,394 patients received at least one dose of efalizumab. A total of 2,719 patients were included in the first treatment phase analysis, of whom the majority (64%; 1,740 patients) received efalizumab 1 mg/kg per week; 979 patients (36%) received placebo. Efalizumab 2 mg/kg per week regimen was given to 409 patients (15%) in two of the five studies [12, 13]; consequently, these patients were not included in the first treatment analysis. Patient demographics and baseline psoriasis characteristics were similar between treatment groups in the first treatment phase (Table 2). Table 2Baseline demographic and disease characteristics for patients in the placebo-controlled first treatment phaseCharacteristicsPlacebo (n = 979)Efalizumab 1 mg/kg per week (n = 1,740)Efalizumab 2 mg/kg per week (n = 409)Mean age (years), mean (SD)45 (12)45 (12)45 (13)Weight (kg), mean (SD)90.0 (20.0) 89.4 (19.6)93.6 (20.5)Mean BMIa (kg/m2)30.4 (6.4)30.2 (6.3)31.4 (6.6)Race, n (%) Caucasian891 (91)1,569 (90)356 (87) Other88 (9)171 (10)53 (13)Duration of psoriasis, mean number of years (SD)19.2 (11.4)19.1 (11.4)17.6 (11.7)History of arthritis, n (%)286 (29.2)529 (30.4)141 (34.5)BMI body mass indexaDue to missing height data, BMI was calculated for 971 patients in the placebo group, 1,719 patients in the efalizumab 1 mg//kg per week group and 404 patients in the efalizumab 2 mg/kg per week group First treatment phase (weeks 0–12) During the first 12 weeks of treatment a similar proportion of patients had an arthropathy AE in the efalizumab 1 mg/kg group (3.3%) and the placebo group (3.5%; Fig. 1a). Correspondingly, the incidence of arthropathy AEs per patient-year was 0.15 in the efalizumab 1 mg/kg group (95% CI 0.11–0.19) and 0.16 in the placebo group (95% CI 0.11–0.22; Fig. 1b). The majority of the arthropathy AEs was mild-to-moderate in intensity in both the efalizumab (41/58 events; 71%; 95% CI 57–82%) and placebo groups (31/34 events; 91%; 95% CI 76–98%). Fig. 1a Proportion of patients who had arthropathy adverse events (AEs) during each phase of the safety analysis and b incidence of arthropathy AEs per patient-year for each phase The additional analysis of data from the study by Sterry et al. [22] demonstrated that the incidence of psoriatic arthropathy per patient-year was lower in the group treated with efalizumab 1 mg/kg per week (0.10; 95% CI 0.05–0.18) than in the placebo group (0.17; 95% CI 0.08–0.30); the proportion of patients with psoriatic arthropathy was 2.3% (12/529 patients) in the efalizumab group and 3.8% (10/264 patients) in the placebo group. First exposure phase In total, 3,394 efalizumab-treated patients were included in this analysis. A small proportion of patients had an arthropathy AE (3.6%; Fig. 1a) and the incidence of arthropathy AEs per patient-year was also low (0.16; 95% CI 0.14–0.19; Fig. 1b). The incidence of arthropathy AEs in this group of patients was similar to that in the placebo group in the first treatment phase, as indicated by the overlap in CIs. Extended treatment phase (weeks 13–24) In total, 2,111 patients were included in the extended treatment phase analysis. During this phase, a low proportion of patients had an arthropathy AE (3.8%; Fig. 1a) and the incidence of arthropathy AEs per patient-year was also low (0.17; 95% CI 0.14–0.22; Fig. 1b). Overlap in the CIs indicates that the incidence of arthropathy AEs in this group of patients was also similar to that in the placebo group in the first treatment phase. Long-term treatment phase The results of two long-term studies were analyzed separately to assess the incidence of arthropathy AEs in patients treated with efalizumab. In both of these studies (Fig. 2), there was no overall increase in the incidence of arthropathy AEs over time. Furthermore, the incidence of arthropathy remained similar to that of the placebo group in the first treatment phase and stable between 12-week periods. Fig. 2Incidence of arthropathy AEs in long-term studies of patients treated with efalizumab a for up to 36 months and compared indirectly with pooled placebo data from the first treatment (FT) phase [5, 6] and b for up to 15 months and compared with the study’s placebo group during month 0–12 [15, 17]. *Following the first 3-month double-blind, placebo-controlled phase of this study, patients in the placebo group who continued were switched to open-label treatment with efalizumab. Consequently, the month 6, 9, 12 and 15 results included patients who had received placebo during the initial 3 months of the study In total, 339 patients were included in the analysis of the study by Gottlieb et al. [6]. These patients received continuous treatment with efalizumab 2 mg/kg once weekly for weeks 1–12 (fluocinolone acetate or petrolatum was co-administered during weeks 9–12), followed by continuous maintenance treatment with efalizumab 1 mg/kg once weekly for up to 36 months in patients who had a ≥ 50% improvement in PASI score. For months 3–15, the dose of efalizumab could be escalated to 4 mg/kg per week for up to 4 weeks if clinically indicated, then maintained at 2 mg/kg per week. During the entire study period, there was little variation in the incidence of arthropathy AEs (range 0.06–0.19; Fig. 2a). Reasons for discontinuation were diverse and were representative of the overall population; refer to Gottlieb et al. [6, 7] for details of discontinuations. For the other long-term study, the analysis included 3-month data from 449 efalizumab-treated patients in the placebo-controlled first treatment phase of the study [17] and data from 635 patients who entered the open-label extension phase and received efalizumab treatment [15]; 218 of the 635 patients included in the open-label extension had switched from placebo to efalizumab after completing the first treatment phase. Patients who entered the open-label extension phase continued to receive, or initiated treatment with, efalizumab 1 mg/kg once weekly for up to 15 months continuously. As in the long-term study by Gottlieb et al. [6], there was little variation in the incidence of arthropathy AEs during the entire study period (range 0.06–0.12; Fig. 2b). Re-treatment phase In total, 565 efalizumab-treated patients were included in the re-treatment phase of the analysis. In this phase, a lower proportion of patients had an arthropathy AE (2.7%; Fig. 1a) compared with the first treatment phase, and the incidence of arthropathy AEs per patient-year was also lower (0.12; 95% CI 0.07–0.19; Fig. 1b). The incidence of arthropathy AEs in this group of patients was lower than in the placebo group in the first treatment phase. Baseline characteristics and previous history of arthropathy There were no differences in baseline demographics or disease characteristics between the patients who had arthropathy AEs and those who did not. Patients who experienced an arthropathy AE during treatment with efalizumab appeared to be more likely to have a history of arthropathy prior to treatment. Of the patients who never developed an arthropathy AE during efalizumab treatment, 27% reported a previous history of arthropathy compared with 59% in patients who did have an arthropathy AE. During the first treatment phase, 88% (n = 34) and 76% (n = 79) of patients who developed an arthropathy AE had a history of arthropathy prior to receiving placebo or efalizumab 1 mg/kg once weekly, respectively. Arthropathy AEs and clinical response to efalizumab Arthropathy AEs appeared to be less likely to occur in patients who had a good clinical response to treatment (≥75% improvement in PASI score; 2.3% of patients had events) than in patients who had a partial response (50–74% improvement in PASI score; 3.5% of patients had events; Fig. 3) and non-responders (<50% improvement in PASI score; 4.5% of patients had events). The corresponding incidences of arthropathy AEs per patient-year were 0.10 in patients with a good clinical response (95% CI 0.05–0.18), 0.17 in patients with a partial clinical response (95% CI 0.11–0.25), and 0.21 in patients who did not respond (95% CI 0.15–0.28). Fig. 3a Proportion of patients with an arthropathy AE by response category on the psoriasis area and severity index (PASI) and physician global assessment (PGA) scales and b incidence of arthropathy AEs per patient-year by response category on the PASI and PGA scales When assessed using the PGA scale, arthropathy AEs also appeared to be less likely to occur in patients who had better clinical responses to treatment with efalizumab (Fig. 3). During the extended treatment phase, the incidence of arthropathy AEs per patient-year was 0.17 in patients with responses categorized as ‘cleared’, ‘excellent’ or ‘good’ on the PGA scale (95% CI 0.10–0.25) and 0.25 in patients with responses categorized as ‘fair’, ‘slight’, ‘unchanged’ or ‘worse’ on the PGA scale (95% CI 0.17–0.35). Discussion The placebo-controlled results of this large-scale pooled analysis of arthropathy data from seven clinical trials show that efalizumab does not appear to increase the risk of developing arthropathy AEs compared with placebo during the first 12 weeks of treatment. In addition, for patients treated with efalizumab, the incidence of arthropathy AEs did not appear to increase over time. The proportion of patients who had an arthropathy AE within any 12-week treatment period was low (<4.1%) through all treatment phases (first treatment, first exposure, extended treatment, re-treatment, long-term treatment). Joint disease has also been reported as a side-effect of other approved biological treatments for psoriasis, namely infliximab (EMEA public statement on infliximab, http://www.emea.eu.int/pdfs/human/press/pus/444500en.pdf) [2, 18, 19], alefacept (Biogen safety presentation on Alefacept to the FDA, http://www.fda.gov/ohrms/dockets/ac/02/slides/3865S1_04_Biogen-Safety/sld007.htm) [20, 24], and etanercept (EMEA Scientific discussion for the approval of Enbrel, http://www.emea.eu.int/humandocs/PDFs/EPAR/Enbrel/014600en6.pdf). Indeed, placebo-controlled studies of infliximab and alefacept indicate that in patients with moderate-to-severe psoriasis the incidence of arthralgia (joint pain) is 7 and 5%, respectively (Biogen safety presentation on Alefacept to the FDA, http://www.fda.gov/ohrms/dockets/ac/02/slides/3865S1_04_Biogen-Safety/sld007.htm) [18, 24]. Psoriatic arthritis has been reported as serious treatment-related AE in three placebo-controlled studies of etanercept in the treatment of chronic plaque psoriasis (incidence data have not been published) (EMEA Scientific discussion for the approval of Enbrel, http://www.emea.eu.int/humandocs/PDFs/EPAR/Enbrel/014600en6.pdf). The incidence of arthropathy AEs in the current analysis of efalizumab appears to be similar to that for arthralgia in studies of infliximab and alefacept. Moreover, the term ‘arthropathy’, used in the current study encompasses a variety of joint diseases, not just a single joint condition such as arthralgia or psoriatic arthritis, and therefore has greater potential to include more patients. However, this between-study comparison is indirect and thus should be treated with caution. Moreover, no arthropathy event (defined by any of the MedDRA or COSTART preferred terms) was excluded from the analysis. Also, data from first treatment phase of the study by Sterry et al. [22] indicate that the proportion of patients with psoriatic arthropathy specifically was low (2.3%) in patients treated with efalizumab 1 mg/kg per week—in fact, lower than in the placebo group (3.8%). It should be noted, however, that psoriatic arthropathy events were not confirmed by a rheumatologist—this is a potential limitation of the study. However, the umbrella term ‘arthropathy’ was designed to capture all joint diseases, including ‘psoriatic arthropathy’. Also, the incidence of psoriatic arthropathy in the study by Sterry et al. was in line with the incidence of ‘arthropathy’ in the overall pooled analysis. To put the results of this pooled analysis, which by its very nature included select patient populations (determined by the inclusion/exclusion criteria and study designs), in the context of routine clinical practice, post-marketing surveillance data were assessed. During post-marketing surveillance of efalizumab, which accounts for approximately 17,500 patient-years to date, serious arthropathies requiring hospitalization were reported with a frequency of about 4.8 per 1,000 patient-years in patients receiving efalizumab. It should be noted, however, that underreporting of AEs in routine clinical practice setting may lead to an underestimate of the true incidence of arthropathy. For both the 12-week first treatment and first exposure phases of the current analyses, the proportions of patients reporting an arthropathy AE appeared to be lower in the efalizumab groups than in the placebo group in the first treatment phase. Correspondingly, the incidences of AEs per patient-year in these treatment phases were also lower in the efalizumab groups than that observed in the placebo group in the first treatment phase. However, the proportion of AEs that were moderate or severe was greater in the efalizumab groups than in the placebo groups; too few patients had events to draw any meaningful conclusions. During the extension phase (weeks 13–24), the incidence of arthropathy AEs in efalizumab-treated patients remained similar to the placebo group in the first treatment phase. Previous history of arthropathy and poor clinical response may potentially indicate a risk for occurrence of new arthropathy AEs during treatment. Indeed, arthropathy AEs were most frequent in patients who did not respond to therapy with efalizumab or in patients with a history of arthropathy. Importantly, the data from the two long-term studies of efalizumab indicate that the incidence of arthropathy AEs remains stable and low for up to 3 years of continuous treatment. These results, coupled with efficacy data showing that the clinical improvements of the skin after 3 months of efalizumab therapy are maintained throughout 36 months of continuous dosing [5], support the suitability of efalizumab for the chronic, continuous treatment of patients with psoriasis. Reasons for patients’ discontinuations in the 36-month study by Gottlieb et al. [6, 7] were diverse and representative of the overall population included in this analysis and have been described previously. When considering the long-term analysis of the 36-month study (Fig. 2a), it should be noted that the number of patients who remained in the study decreased over time. This discontinuation rate is not unexpected for a study that is 3 years in duration but, by month 36, there is a small number of patients on which to base comparisons with the first treatment phase. Another factor that may confound between-phase analysis comparisons was the possible use of concomitant medications for psoriasis after the first treatment phase in the study by Gottlieb et al. [6], which permitted the use of topical corticosteroids and ultraviolet B phototherapy. Accordingly, it should be noted that comparisons of the results between any of the treatment phases of this analysis are observational (i.e., not direct) but do confirm the results of the long-term treatment phase and the placebo-controlled 12-week first treatment phase studies, suggesting that the risk of joint disease is not increased with continued efalizumab treatment and that the incidence of arthropathy is low and similar to placebo. Further investigation is needed to confirm the results of this preliminary analysis of arthropathy events during long-term treatment with efalizumab. In patients who re-started treatment for a further 12 weeks following an intervention-free period, the proportion of patients who had an arthropathy AE was lower than during the first treatment phase; the same was true for the incidence of arthropathy AEs per patient-year in re-treated patients. Although this scenario is likely to occur infrequently in clinical practice, these data show that if a patient needs to stop (e.g., during pregnancy) and then restart treatment, there appears to be no increased risk of arthropathy AEs. Although arthritis in patients with psoriasis has a significant impact on QoL [9, 27], it can, in most cases, be managed effectively [8]. In the small minority of patients who develop arthropathy during treatment, the symptoms can be managed successfully with non-steroidal anti-inflammatory drugs [8]. In conclusion, the results of this pooled analysis show that efalizumab does not appear to increase the risk of developing arthropathy AEs compared with placebo. Long-term studies of efalizumab indicate that the incidence of arthropathy AEs remains stable and low for up to 3 years of continuous treatment.	[ "arthropathy", "efalizumab", "psoriasis", "psoriatic arthritis", "arthritis" ]	[ "P", "P", "P", "P", "P" ]
Naturwissenschaften-4-1-2270361	Mechanical analysis of infant carrying in hominoids	In all higher nonhuman primates, species survival depends upon safe carrying of infants clinging to body hair of adults. In this work, measurements of mechanical properties of ape hair (gibbon, orangutan, and gorilla) are presented, focusing on constraints for safe infant carrying. Results of hair tensile properties are shown to be species-dependent. Analysis of the mechanics of the mounting position, typical of heavier infant carrying among African apes, shows that both clinging and friction are necessary to carry heavy infants. As a consequence, a required relationship between infant weight, hair–hair friction coefficient, and body angle exists. The hair–hair friction coefficient is measured using natural ape skin samples, and dependence on load and humidity is analyzed. Numerical evaluation of the equilibrium constraint is in agreement with the knuckle-walking quadruped position of African apes. Bipedality is clearly incompatible with the usual clinging and mounting pattern of infant carrying, requiring a revision of models of hominization in relation to the divergence between apes and hominins. These results suggest that safe carrying of heavy infants justify the emergence of biped form of locomotion. Ways to test this possibility are foreseen here. Introduction All higher primates (except humans) carry their young clinging to their fur from birth (Jolly 1972). The correlation between infant carrying and the form of locomotion of adult primates is clear, but no detailed study has focused on the mechanics of the problem. Among nonhuman primates, there is a change in the carrying pattern of infants by adults (mostly by the mother) as the infant grows (Jolly 1972; de Vore 1965; Goodall 1967; MacKinnon 1974; Fossey 1979, 1983; Tuttle and Watts 1985). Newborns are carried clinging in close ventro-ventral contact, often with additional support from the mother (Hoff et al. 1983). Change to infant support over the adult body (dorsal or lumbar clinging) occurs some months later for all nonhuman higher primates and extends for years in apes. It seems clear that safety in infant carrying imposes limits on the weight of infants. This work deals with the problem of infant carrying in Hominoids, characterized by increasing body size. Accepted hominoid phylogeny places the branch to the lesser ape gibbon as the oldest. It is followed by the great apes, with the older branch to the arboreal orangutan and the branch originating terrestrial gorilla and chimpanzee coming later, followed by biped hominids (Yunis and Prakash 1982; Ruvolo et al. 1994; Lockwood et al. 2004). Mechanical analyses of constraints for safe infant carrying in the usual primate pattern are reported. Measurements of tensile and friction properties of hominoid hairs are presented and conditions for mechanical equilibrium determined. The usual pattern of primate carrying of heavy infants is shown to be incompatible with bipedalism. This brings a new perspective to the hominization process and to the basic changes at the divergence between apes and hominins. Materials and methods Samples For the experimental study of the mechanical properties of ape hairs, three pieces of skin with hairs were obtained from frozen limbs of dead animals at the Laboratoire d’Anatomie Comparée, Muséum National d’Histoire Naturelle, Paris, France (one gibbon, one orangutan, and one gorilla—chimpanzee was not available). Due to difficulty in obtaining animal skins, the samples came from three individual animals, one piece of about 200 cm2 from each. The three pieces of skin were transported at room temperature immersed in formaldehyde and were treated afterwards by a taxidermist. After the skin was fixed, it became leather, and chemicals were washed out from the hairs, which looked natural. The effect of formaldehyde on hair has been tested in human head hair; results are presented in the Appendix, and evidence is that no significant change occurs in the relevant tensile parameters. The skin and hair from the different animals studied here are easily recognized by appearance and texture and are species-characteristic. Gibbon hairs are smooth and silky, while orangutan hairs are harsh, hard and very long; gorilla hair has an intermediate appearance. Estimated values of skin thickness and hair length are given in Table 1, with other parameters obtained. Table 1Parameter values for apes GorillaGibbonOrangutanSkin thickness (mm)~1.5~0.6~2.3Hair Length (cm)~6~4~10∅ (μm)66 ± 2 (21%)52 ± 2 (19%)120 ± 4 (24%)α (N)17 ± 2 (25%)6.0 ± 0.6 (33%)32 ± 1 (13%)FE (N)0.38 ± 0.03 (30%)0.18 ± 0.02 (36%)0.49 ± 0.02 (12%)Fr (N)0.67 ± 0.09 (39%)0.29 ± 0.04 (47%)0.79 ± 0.04 (17%)Sr0.26 ± 0.05 (56%)0.28 ± 0.05 (52%)0.07 ± 0.01 (51%)E (GPa)5.0 ± 0.6 (42%)2.8 ± 0.3 (43%)2.8 ± 0.2 (36%)Skin thickness and hair length (estimated), hair diameter ∅, and results obtained from the n = 10 tensile curves for each ape hair. The regression coefficient α and the elastic limit FE are measured in the initial linear Hookean part of the curve, while the force Fr and strain Sr at rupture are measured at the end of the curve. Mean values are given with their standard errors. The Young modulus E is given by α divided by the hair cross-sectional area (with error propagation). The variation coefficient (CV = 100 × standard deviation/average value) is given in parenthesis. The three skins provided several samples of different areas with the skin side fixed on pieces of wood, leaving the hairs exposed. The use of taxidermized specimens is the only possibility for simulating the natural situation for friction coefficient measurements. The basic data by Schultz (1931) on hair density of primates were also obtained from embalmed animals. Methods Several sectioned ape hair cross-sections were examined under a compound microscope (magnification up to 360), and external appearance as well as hair diameters were measured with a Zeiss Axiovert 200 microscope (several positions along the hair length). Ten single hairs of each of the three animals were cut and analyzed on an automated Instron Tensile Tester, at the standard 65 ± 2% relative humidity of the air (RH), 20 ± 2°C and speed 2.6 mm/min. The hair was fixed on one end to a fixed grip and on the other extremity to a movable grip, connected to a pulling force. The defined distance between grips (Lo) is between 20 and 50 mm. Curves of force (given in Newton, MKS unit) versus longitudinal elongation L (mm) were obtained; the strain, or relative elongation, is given by (L − Lo)/Lo. Single human head hair (with and without immersion in formaldehyde) was also analyzed as a standard, as a comparison was possible with previous data for human hairs (Robbins 1994; Nikiforidis et al. 1993; Franbourg et al. 2003). Twenty single hairs from each of the three animals, still on the skins, were also analyzed and pulled until detachment from the skin occurred. For friction measurements, the wooden pieces with the larger areas from each animal were fixed to a movable inclined plane. The other pieces (with areas in the interval 16–76 cm2) were used in measurements of the static friction coefficient μ (hairs against hairs) using variable weights attached to the wood face. The angle of the inclined plane was carefully changed until slipping started at its critical value, and the process was repeated several times to check reproducibility. The RH and the room temperature (20–26°C) were measured and controlled with a dehumidifier and an air-conditioning system. Results Tensile measurements Mechanical studies of animal fibers have been extensively performed over the last decades because of interest in textile production (mainly wool, see Feughelman 1997) as well as cosmetics (human head hair, see Robbins 1994). Animal hair is composed of three parts: an external thin cuticle (protective covering with a scale structure from root to tip), a thicker cortex with fibrous proteins, and a central porous medulla, which may be absent in finer hair. Ape hair viewed under an optical microscope are similar in structure to wool and human head hair, as in Fig. 1, where the external appearance with the cuticle scale structure is shown in Fig. 1a and the cross-section inner structure in Fig. 1b. Fig. 1Micrographs showing orangutan hair, whose light color allows better visualization; a external appearance with typical cuticle scale structure, b transversal cross section with external cuticle, thicker cortex, and inner medulla Mechanical properties of natural fibers are defined by the cortex of the fiber and are due to the molecular structure of keratin, which constitutes the cortex, as determined many years ago (Astbury and Street 1931; Astbury and Woods 1933). Extensive work during decades evidenced the dependence of mechanical properties essentially on the cross-sectional area of the fiber, thus enabling basic research to be conducted on relatively few single fibers, eliminating the need for statistical methods on a large number of samples (Feughelman 1997). Tensile measurements give curves of force against strain (relative elongation) for a single hair fiber. Force has an important meaning for this study; stress is obtained dividing the force by the cross-sectional area of the fiber. A typical result with standard human head hair is shown in Fig. 2a and is similar to earlier data on human hair (Robbins 1994; Nikiforidis et al. 1993; Franbourg et al. 2003). The curve displays an initial elastic (Hookean) linear region, followed by a plastic yield region and a postyield region before breakage. Transition from the elastic to the plastic region corresponds to a first-order transition of the alpha-keratin molecule from the coiled helices to the stretched beta-keratin form (Bendit 1957), and the plastic region is strongly influenced by the water content (Feughelman 1997). Fig. 2Measured force × strain (relative elongation) curves for hominoid a human, b gorilla, c gibbon, and d orangutan hairs. In a are stressed the initial Hookean linear elastic region (E), the plastic plateau (P), the post-yield region (PY), and the breaking threshold (B). The different curves in b, c, and d correspond to ten hairs of each individual ape. The inset in a is a magnification of the elastic linear region, showing the slope α and the force at the elastic limit FE The linear region extends up to 1% strain, a result well established for helical alpha-keratins of all animal hairs: the cortex composite has mechanical properties of a single material (Feughelman 1997). The value of the force at the Hookean elastic limit (called FE in this paper) corresponds to the limiting force for full recovery of the fiber after being relaxed. The regression coefficient of the linear region (called α in this paper) of the force versus strain curve divided by the cross-sectional area of the fiber gives the Young modulus of elasticity (E) of the fiber (dimension of pressure unit, GPa = 109 N/m2). The values α and FE are determined expanding the initial elastic part of the curve (which does not, in general, start precisely at zero), as seen in the insert in Fig. 2a. Further analysis of human head hair is given in the Appendix. Apparently, no tensile studies on ape hair have been reported. Curves obtained with n = 10 ape hairs of each of the three apes (one gorilla, one gibbon, and one orangutan) are shown in Fig. 2b–d. The curves display all the classical regions of animal hair. Human hair has a larger postyield region and a larger strain at breakage, related to the usual human hair treatment (frequent washing and combing). Maximum stretching is very different in the wet (50% strain) and dry (10% strain) states and may be influenced also by hair conditioners (Robbins 1994; Colombera and Joekes 2004). The initial linear region is well defined in all hair curves, and its variability among different hairs of the same ape is due to differences in hair cross-sectional area. The physical parameters α and FE of each of the ape curves were obtained with the procedure shown in the insert of Fig. 2a. Values of force (Fr) and strain (Sr) at rupture were measured directly at the end of the curves. Table 1 gives statistical averages of parameters obtained from tensile curves (α, FE, Fr, Sr) and measured hair diameter ∅ for each species, together with calculated Young modulus of elasticity E, with respective errors. The average values for FE are 60% (±2%) of Fr for the hairs of the three ape species. Hair diameters ∅ in Table 1 are averages over several positions measured along ten single hairs of each ape (∼50 measurements for each ape). One-way analysis of variance (ANOVA) test for these ∅ measurements evidenced significant difference between apes (F2,155 = 177; p < 0.0001). All tensile parameter values for gorilla hairs, with the only exception being the Young modulus E, are intermediate between gibbon and orangutan hairs. Table 2 gives statistical tests for these four parameters: significant differences result from two-population t tests and one-way ANOVA test. Table 2Statistics for tensile parameters of ape hairs FEαFrSrGibbon × Orangutant1011188.44.4p<0.0001<0.0001<0.00010.0003Gorilla × Gibbont105.36.24.10.32p<0.0001<0.00010.00080.75Gorilla × Orangutant103.17.61.34.0p0.007<0.00010.20.0009ANOVAF2,2744119209.3p<0.0001<0.0001<0.00010.0009Results from statistical tests for the n = 10 measured values of the four parameters obtained from the force × strain curves of hairs of the three apes. Two-population t-test (t10 values and corresponding probability p) are given for each parameter and each compared population in the upper lines. ANOVA test of the four parameters in the three ape populations and corresponding probability p in the lower line Hominoid species differ significantly regarding hair strength, which increases from gibbon to gorilla and to orangutan. However, the Young modulus has a different behavior: E values for gibbon and orangutan are equal (2.8 Gpa), but gorilla has a higher value (5.0 Gpa). The value for human head hair from literature (Robbins 1994; Nikiforidis et al. 1993) is 3.8 Gpa, intermediate between gorilla and orangutan hair. Human head hairs measured in this paper are within reported values, as detailed in the Appendix. Results obtained indicate that the force strength and average thickness of hominoid hairs are species-dependent and may be evolutionary meaningful. Data for human hair in the Appendix give further indication of species dependence. Such direction of research, together with definition of the Young modulus in terms of the inner hair structure, have clear interest but outside the scope of this work. For infant carrying, the really decisive parameter is FE, representative of the upper limit of the force that a single hair can safely withstand, because it is the limit for recovery after relaxing. It is convenient to express FE in kilogram-force (and gram-force), corresponding to the force weight of a given mass at the earth surface. This unit is often still used for load (instead of N = 0.10198 kgf) because it makes the mass value of the weight clear. The important values FE are then 18 ± 2 gf for gibbon, 39 ± 3 gf for gorilla, and 50 ± 2 gf for orangutan. The value for human head hair is 29 ± 2 gf (see Appendix). The force to pull hairs from the skin samples of the three ape species has also been measured and may be considered as an upper limit, in view of changes in skin due to taxidermy. They are typically near the elastic limit and lower than rupture value, as occurs with human head hair in the scalp (Robbins 1994), giving a further reason to focus on the FE values as upper limits for safe load carrying. A robust conclusion can be thus extracted. It is clear that to withstand the clinging infant weight, a large number of hairs held together is required. Results obtained evidence that bunches of about 100 hairs are necessary to carry infants weighing a few kilogram-force. Infant hands (and feet) can possibly grasp a bunch of hairs available in some square centimeters of skin so that safety in the usual pattern of nonhuman primate infant carrying critically depends on the density of hairs. Infant carrying Values of hair density for nonhuman primates have been obtained in the extensive work of Schultz (1931, 1968); values are not gender-dependent but vary with body region: density is highest in the vertex and smaller on the ventral than on the dorsal side of the trunk. Densities are generally compatible with the carrying requirement defined above, but the great apes have the lowest hair density among primates. Hair density values on the back of primates vary, from over 1,000 hairs/cm2 for smaller species (monkeys and gibbons), down to about 100 hairs/cm2 for great apes. The problem of safely carrying heavy infants requires thus particular consideration among the great apes. Analysis of hair density data (Schultz 1931) in many anthropoid primate taxa yield (Schwartz and Rosenblum 1981) a negative allometry of relative hair density (rhd = hair density/total surface area). This decrease in rhd with increase in primate body mass has been correlated to thermal constraints imposed by the decreasing ratios of surface area to body volume. However, chimpanzees apparently have a lower rhd than would be expected for their body volume (see Fig. 1 of Schwartz and Rosenblum 1981), with rhd similar to that of larger gorillas. The bunch of hairs available for clinging depends also on hair length and infant hand size. Accepting that about 3 cm are necessary to roll hairs on the infant fingers, the excess in length defines the area available for clinging. Clinging area per limb may be thus estimated to be 3, 10, and 50 cm2 for, respectively, gibbon, gorilla, and orangutan. In a dynamic situation, the load should be supported instantaneously by a single limb. The infant-carrying method of the arboreal Asian apes (gibbon and orangutan) is different from that of the terrestrial African apes (gorilla and chimpanzee). The arboreal apes carry their young over one side of the mothers’ pelvis, not on the back (de Vore 1965). In accordance with the small size (7 kgf when adults), the lesser ape gibbon has a high hair density, over 1,000 hairs/cm2, more than ten times the density of great apes (Schultz 1931, 1968). This ensures safety for carrying their light young in their acrobatic arboreal life. The arboreal but heavier orangutans (adult females weigh around 50 kgf and males 90 kgf) compensate the lower hair density by longer, thicker, and stronger hairs and thicker skin. Orangutan infants live with their mothers for about 7 years. The gorilla is about 50% heavier than orangutan, but instead of even greater hair thickness and length, the solution for terrestrial apes is the mounting position for heavier infant support. Orangutan hair and skin are, in this paper, estimated to be able to withstand about seven times more weight than gorilla hair and skin. Gorillas are the heaviest primates and have very few chest hairs (Schultz 1931, 1968). Adult gorilla is much heavier than adult human, but their newborns have only half the weight of human babies. In the first 1 or 2 months, the infant gorilla is supported manually by its mother as she walks tripedally or bipedally (Jolly 1972; de Vore 1965; Tuttle and Watts 1985; Doran 1997). The same occurs for chimpanzee babies, unable to support their own weight by clinging prior to 2 months of age (Plooij 1984). Change from quadruped to triped or biped motion occurs systematically among all great apes when infant safety requires manual support (Jolly 1972; de Vore 1965; Tuttle and Watts 1985). Slow and careful locomotion of female chimpanzees while carrying young infants has been reported (Goodall 1967). Two primary causes of mortality have been found among infant chimpanzees: inadequacy of the mother–infant bond and injuries caused by falling from the mother (Goodall 1967). It is clear that infant carrying is crucial and depends on not trivial behavior among African great apes. Weight limit for clinging in African apes The gorilla FE value, combined with hair density and length, indicates that load up to 40 kgf could be supported by clinging. However, the weight limit for clinging depends also on several other variables. Increase in hair pull-out or break in the dynamic situation, capacity of the infant to withstand its own weight and to grasp hairs without sliding, and skin capacity to withstand the load without rupture are all variables that may decrease the infant weight limit for clinging. A bunch of hairs on about 1 cm2 (±10%) of gorilla skin under load was observed, still on the skin, to simulate the real load effect in the actual clinging situation. Counting hairs in this bunch gave 140 hairs (±10%), in agreement with the value reported by Schultz (1968) for the back side of the trunk of gorillas (145 hairs/cm2), evidencing that there are no marked differences between the limb sample and the back-trunk hairs. This bunch of hairs supported 1 kg without problems, even with some pendulum movement. With 2 kg, several problems became evident. Besides eventual detachment of single hairs from the skin during movement, the hair bunch started to slip from the double-face adhesive tape used to attach it to the load support. It was necessary to make a new attachment system; the hair bunch was rolled several times around the weight support. Even so, loads of 2 kg and more could be observed only with the rolled hairs attached by super glue (ester cyanoacrylate) to the weight support. This indicates a possible problem also with the clinging capacity of infant fingers, and slippage may occur. Also, the skin itself began to deform and detach from the wood support, requiring a much stronger attachment of the skin to the wood base. The bunch of hairs broke at the same hair length for all under a static load of 7 kg, compatible with the values of force at rupture (Fr) obtained from the tensile measurements of single hairs (Table 1). An upper limit of 1 kgf/cm2 for clinging without problems may be deduced from this simple experiment, but the actual limit depends on the clinging capacity of infant fingers and on the mother’s skin resistance. The effective weight limit for clinging must be obtained from careful observation of live apes. Fossey (1979) describes physical and behavioral development of gorilla infants: at 4–6 months (weight about 5 kgf), the gorilla infant travels 60% of the time in ventral position while with 6–12 months (weight about 8 kgf), it travels in the dorsal position 80% of the time. A detailed investigation of the ontogeny of locomotion in the African ape (Doran 1997) revealed that up to 5 months of age, chimpanzees are slightly more precocious than gorillas. However, at 6 months, gorilla locomotor development becomes faster than that of chimpanzees, and surprisingly, much of the interspecific variation in behavior is explained by differences in body size (Doran 1997). Afterward, when the infants are of similar weights (although of widely disparate ages) gorillas and chimpanzees perform very similar locomotion activities (Doran 1997). It is, in this paper, suggested that the weight level of 5 kgf may be considered the limiting value for safe support over long periods by only ventral clinging, defined probably by a combination of security factors1 related to hair strength, infant clinging capacity, and skin resistance to pressure. The limiting clinging pressure on the mother’s skin for the African great apes may be thus estimated to be 0.5 kgf/cm2 (a factor of 2 in relation to the simple test made with the skin sample). The skin is also a visco-elastic medium, reflecting properties of a strain-induced aligned-collagen network that can stretch up to 100% before permanent damage (Silver 1987). The shear modulus for the human dermis ranges from 1.2 to 3.1 MPa, while for the human hypodermis, it ranges from 3.1 to 9.7 kPa (Gennisson et al. 2004). The pressure limit 0.5 kgf/cm2 (~50 kPa) could be a limit also for danger to the skin of the mother ape. It is known that one of the important differences between the skin of humans and of African apes is the higher elasticity of the human skin (Montagna 1982). The locomotor development of the infants must therefore adapt to such an effective weight limit. This does not mean that infants over 5 kg will fall. Larger infants may travel briefly in ventral position during stressful situations (Fossey 1979) or when the chimpanzee mother swings through trees (Goodall 1967). Usual locomotion of adult African apes is terrestrial quadrupedalism (~96% of the time in gorillas and ~86% in chimpanzees, from Doran 1997), and safe infant support over longer periods clearly requires the necessity of change to a mounting position, with the infant weight supported by the adult body. Dorsal clinging position In order to analyze the mechanics of the carrying system in the African great apes, it is also necessary to consider, besides the tensile properties of hairs, the hair–hair friction that prevents slipping, particularly for heavier infants in dorsal position. Fig. 3 is a sketch of the mounting position in African apes, showing the angle θ of the inclined plane where the infant stands. For simplicity in the sketch, the angle θ coincides with the angle defined by the usual knuckle-walking position of African great apes, but that is not necessary. The important parameter is the inclination of the base where the infant stands clinging; its relation to the average body inclination angle does not need to be taken into account in analysis of the friction effect. Fig. 3Sketch of ape carrying infant. For simplicity, in the sketch, the angle θ of the inclined plane where the infant stands (to which Eq. 1 refers) coincides with the angle defined by the knuckle-walking position of African great apes, which is not necessarily so, as discussed in the text The total infant weight, in the vertical direction, has two components, one in the inclined plane, favoring slipping, and one normal to the inclined plane, responsible for the friction force opposing slipping. Calling Wt the total infant weight carried in mounting position, Wc the effective weight limit for clinging, and μ the static friction coefficient (Bowden and Tabor 1956), the equilibrium of forces on the infant requires that force components on the inclined plane direction satisfy the condition: In the absence of clinging (Wc = 0), slipping on the body surface of contact starts for a critical value θc, given by tan θc = μ. Both friction and clinging are thus essential to hold heavy infants in the mounting position. The requirements for dynamic equilibrium may be greater, so that in fact the mounting position in dynamic movement may be unsafe even when Eq. 1 is satisfied. It should be stressed that Eq. 1 is a necessary condition for infant survival and therefore also for species survival. It is a very robust requirement of basic mechanics for static equilibrium. The ape hair–hair friction coefficient μ must be known to analyze the actual ape situation. Friction measurements Apparently, no results for hair–hair friction coefficients exist in the literature; only friction between human hair and other materials are reported (Robbins 1994). The skin samples were used to measure μ as a function of the supported load W in known conditions of relative air humidity and temperature, in a situation analogous to that depicted in Fig. 3. It was verified that μ depends on the hair direction, being larger for hairs in the parallel position (as in nature, with the two animals heading in the same direction) than in the antiparallel condition. Systematic measurements have been thus made in the parallel condition. Figure 4a shows a typical result of μ as a function of W. For hard solid surfaces, μ is constant and independent of the load W and contact area A. This is clearly not the case for hairs, especially at low loads. Long ago, it was shown that a variation of μ with W occurs for textiles and fibers (Gralen 1952; Makinson 1952), given by the empirical equation: Fig. 4Results for friction coefficient at 80% RH. a friction coefficient μ as a function of W for a gorilla skin with smaller area (16 cm2). The line shows a fit to the data using Eq. 2. bμ × W for (circles) gorilla with larger area (76 cm2); (squares) gorilla with smaller area (16 cm2), and (triangles) gibbon with intermediate area (42 cm2). Linear fits to the data are shown in lines The parameter a gives the friction coefficient in the limit of high loads. It has been proposed (Gralen 1952) that the parameter b is related to the pressure inside an oil lubricating film on the fiber surface. The thickness of this film decreases due to pressure, which causes changes in the friction mechanism. Hair is a natural fiber (Howell et al. 1959) and behaves in a similar way. In fact, the data could be well-fit with Eq. 2, shown as a solid line in Fig. 4a. Many curves were analyzed for different values of the contact area A and RH. A strong dependence of μ on A was observed, probably due to shear among the several hairs in contact. The fits to the linear equation μW = aW + b are more convenient for yielding the parameters a and b, as shown in Fig. 4b. Both a and b increase with A, but a is less sensitive to A, while b is roughly proportional to A, so that the term b/W in Eq. 1 is inversely proportional to the pressure W/A exerted on the surfaces in contact. Results indicate that the friction coefficient might be considered constant for pressures higher than 20 gf/cm2. This pressure corresponds to infant weight around 1 kgf, from the relation between body surface area and weight (Schwartz and Rosenblum 1981; Kleiber 1975) and estimating the contact area as one third of the infant surface area. The average value is μ = 0.27 ± 0.03 for gorilla hairs and μ = 0.20 ± 0.02 for gibbon hairs at 60% RH, in agreement with the expected null or small difference on friction depending on the fiber diameter (Robbins 1994). Such values are in good agreement with known values for single fibers (Bowden and Tabor 1956; Howell et al. 1959). The orangutan skin was not as perfect and homogeneous as the gibbon and gorilla skins, showing some parts without hairs, and provided only two samples, so dependence on area could not be analyzed. However, experiments of μ as a function of load gave similar values to those for the other apes. The hair–hair friction coefficient of apes obtained is intermediate between values for friction between human hair and hard rubber with dry (μ = 0.19) and wet (μ = 0.38) surface conditions (Robbins 1994), indicating that results obtained can be considered of general validity. Note that Robbins (1994) mentions the difficulty of testing hair–hair friction, important in hair combing, and that it should be similar to hair–rubber friction. Equilibrium condition in African apes It is possible now to focus the problem on our nearest relatives, the great apes gorilla and chimpanzee, applying Eq. 1 to their mounting condition, in which clinging and friction act together for safe infant carrying. The effective weight limit for constant ventral clinging, already discussed, is estimated to be 5 kgf. As the hair density in the back is higher (30% for chimpanzees), the estimate is Wc ~ 5–7 kgf. The maximum value Wt can be estimated from the weight of infants when they are no longer carried by mothers. This occurs when gorillas are about 2.5 years old and chimpanzees 5 years old, both weighing around 20 kgf (Fossey 1979; Doran 1997). By inserting such values in Eq. 1, and using the estimated constant friction coefficient μ, it is possible to obtain the maximum inclination of the plane supporting the infant for the safe carrying system in the African great apes. The condition θ < 29 to 34° is obtained. This basic equilibrium constraint related to friction in the mother–infant contact is satisfied in ape knuckle-walking quadruped position2. It should be stated that it is frequently reported that in the dorsal position, the African ape infant rides high on the mother’s neck and lies stretched and prostate (de Vore 1965; Fossey 1979). This position increases the fixed base of the load and places it in a body region with a promontory and a smaller inclination angle (θ ~ 20° as estimated2 from the same Fig. 7–20 in Schultz 1968). Older infants ride in a jockey position; they do not normally grip hairs with their feet, but press the sides of their ankles against the mother’s flank (Goodall 1967). Clearly, among terrestrial great apes, infant safety does not rely only in grasping hairs. It may be concluded that slipping imposes clear limits on the maximum body angle attained by heavy ape species carrying infants, representing a hindrance against evolution towards bipedality. This might explain the persistence of knuckle-walking among the great African apes. Effect of humidity on friction The two most relevant variables to friction in hairs (Robbins 1994) are the load pressing the two surfaces together, as analyzed previously, and the relative humidity RH. Dependence of μ on RH was reported for fibers (Howell et al. 1959) and human hair (Robbins 1994), and a similar effect was verified for ape hairs. Reduction in RH, from 80 to 38%, led to a decrease of gorilla hair μ by about 30%. The friction coefficient between human hair and hard rubber is reduced by 50% between the wet and dry states (Robbins 1994). Friction of gorilla hairs has been also measured in wet conditions (immersion in water, followed by slight shaking of the sample). The result is μ = 0.35 ± 0.03, in good agreement with reported values for human hair and hard rubber in the wet condition. The larger μ value in wet condition is possibly due to network of hydrogen bonds connecting hairs in contact. The thin outer epicuticle membrane of hair is hydrophobic, but hydrogen bonds may occur through the thicker exocuticle and endocuticle, especially the latter, which swells much more in water (Feughelman 1997, Robbins 1994). Such structures may be accessible to external water through the ratchet scale structure of hair. It should be concluded that reduction in humidity destroys the delicate balance between friction and clinging (Eq. 1) necessary for survival of infants of heavy ape species. Figure 5 shows the maximum carrying weight Wt obtained from Eq. 1 as a function of the inclination angle θ for μ varying from zero up to 0.40, with Wc = 5 kgf. Fig. 5Curves of total weight Wt supported in dorsal position as function of the inclination angle θ of the infant base on the mother’s trunk, obtained from Eq. 1, for several values of the friction coefficient μ (0–0.4, as indicated on top), with Wc = 5 kgf (limit value shown with dashed line). The line with cross is for Wc = 10 kgf and μ = 0.4. Vertical position corresponds to θ = 90° Discussion In this section, the relevance of the above results for models of hominization is discussed. Safe infant carrying is clearly very important for species survival and has to be considered in the discussion of the transition from apes to hominins. The mechanical analysis made in this paper evidenced that habitual bipedalism, with a large θ angle, is incompatible with the usual primate type of infant carrying. Even if one considers Wc = 10 kgf and μ = 0.40, the maximum attainable angle is θ ~ 49°, as seen in Fig. 5. It seems clear that together with bipedality probably also came the characteristic human type of infant carrying in the arms and hands of their mothers, making their locomotion and foraging particularly difficult. Females would not adopt bipedality if they could avoid it, and the species would not adopt bipedalism if females with infants would not engage in it. The transition to bipedality was not trivial. Etkin (1954) considered the problem of the burden on females “almost continuously carrying a child,” just to conclude that the female could not be an effective hunter, turning to a theory of the monogamous family unit. It is clear now that such a scenario could be considered only for Homo at a much later period. Iwamoto (1985) recalled examples of facultative bipedalism among monkeys and speculated that the “decisive factor (for habitual bipedalism) may have been some everyday necessity to carry something in both hands”. After criticizing proposals in which the “something” is food, Iwamoto suggested that the “something” could be their helpless babies. However, this proposal has been dismissed under the hypothesis that babies became helpless only with increase of brain in Homo. Now models for bipedalism origin based on “food acquisition and foraging strategies” are considered “likely” but are in fact based on the premises that bipedal locomotion frees the arms and hands and does not have high costs. However, such premises just forget the fundamental fact that bipedalism frees the arms and hands only of males and juveniles, but females, on the contrary, have arms and hands occupied by infant carrying, which represents a high cost for bipedalism, not considered up to now. A review of bipedalism mentions: “infants (from early hominin biped) probably clung to their mothers’ hair, as is the case for most other primates” (Richmond et al. 2001). It seems that the not trivial mechanical constraint of safe carrying of heavy infants has never been considered before. Current evidence suggests (Richmond et al. 2001; Kingston et al. 1994; Potts 1998) that the key adaptation to bipedalism originated in wooded environments during a drier period in East Africa. The equilibrium condition (Eq. 1) shows that reduction of friction in the dorsal clinging position would most probably lead to a decrease in infant weight and decrease in the body inclination angle θ, corresponding to the solution adopted by quadruped savanna monkeys in the arid savanna. It seems clear that habitual bipedalism would not be reached by continuous increase of the body angle. However, among hominoids, reduction of friction between infant fingers and mother hairs could reduce infant grasping capacity, increasing the period of mother’s manual support of newborns and eventually inducing habitual bipedalism. Concluding remarks and future perspectives An important conclusion from this work refers to interest in systematic study of tensile properties of hairs from an evolutionary point of view. The other clear conclusion is that models on locomotion and bipedalism evolution must focus on females carrying infants, who are the target of the strongest selective pressures, and this conclusion is independent of the form of locomotion of hominid ancestors (knuckle-walkers, terrestrial or arboreal quadrupeds, or full-time arboreal climbers). To disentangle the divergence between apes and hominins, several directions of research become clear from this work. Studies of living apes focused on the mechanical problem of infant carrying, as well as determination of the elastic limits of the ape skin and the limits for grasping capacity of ape infants, which are all essential to identify the critical factor responsible for emergence of bipedalism. It should be stressed that in a scenario with bipedalism emerging for safe infant carrying, the selective pressure would act particularly on females. This is not a problem, as only genes in the nonrecombinant part of the Y chromosome are not transmitted to both sexes and therefore cannot be associated with species locomotion. The process correlates with increase in the mother–infant bond, characteristic of the human lineage. Other hypotheses for emergence of bipedality (see Richmond et al. 2001 for a review) do not, in fact, explain why other primates did not follow such line. The many scenarios invoked previously for bipedalism evolution may sound plausible, but are not necessary; neither take into account the problem of infant carrying and can be considered complementary and a consequence of the selective pressure on females carrying infants. The difficulties in locomotion and food gathering for biped females carrying infants may well be at the origin of the necessity of group cooperation, which could initially have been among females, with males in their usual role of primate group protection. A fundamental question is the possible correlation between bipedalism and reduction of body hairs, the two basic biological modifications of hominins. It is evident that decrease in body hair as the initial modification would bring on bipedality as a necessary consequence (Amaral 1989), through the strong selective pressure of safe infant carrying, when infants could no longer cling to body hairs. The timing of reduction of body hairs is very controversial, since Darwin’s original (1871) emphasis on the issue. Several works indicated a very early beginning of body hair reduction, still in forested environments, based on thermoregulatory requirements (Newman 1970), rhd of Australopithecines (Schwartz and Rosenblum 1981), and comparison of human and ape skin (Montagna 1982, 1985). Bipedality preadaptive to nakedness was suggested later from water-consumption requirements under reduced humidity (Wheeler 1992). However, detailed analysis of thermal loads and water consumption for quadrupeds and bipeds in the furred and naked conditions suggested again that reduction of body hair started in a forested environment becoming drier (Amaral 1996). Savanna monkeys tolerate heat even at high levels of exercise, evidencing that the biological avenue to cope with heat stress in an open environment is to keep a hair covering, increase sweating capacity, and have a variable conductance (patas monkeys, Mahoney 1980, and baboons, Rogers et al. 1992). The only situation in which nakedness is favorable concerns dissipation of heat loads that do not come from sun absorption but from activity in a more closed forest condition at temperatures below that of the body (Amaral 1996). Furthermore, regarding water requirements, advantages for nakedness exist in dry ambients with temperatures below body temperature (Amaral 1996). Recent genetic work has focused on evolution of black skin as a result of naked unprotected skin under solar stress, estimated to have occurred at least 1.2 mya (Rogers et al. 2004), while clothing is a recent innovation of only about 70,000 years ago (Kittler et al. 2003). However, up to now, there is no date for the beginning of reduction of body hairs. Therefore, localization of the genetic changes responsible for reduction of body hairs in humans is necessary to settle the issue. The new perspective developed in this paper evidences continuity between physical and behavioral aspects of early hominins and their ancestors. The more differentiated aspects of humans are left to the emergence of Homo, at a much later period.	[ "infant carrying", "clinging", "friction", "bipedalism", "hair strength", "reduction of body hairs" ]	[ "P", "P", "P", "P", "P", "P" ]
Arch_Orthop_Trauma_Surg-4-1-2228384	Viscosupplementation in the hip: evaluation of hyaluronic acid formulations	This study compares three different hyaluronate formulations and evaluates functionality, time of satisfactory pain relief and also the delay in performing a total hip arthroplasty. One hundred and twenty patients (126 hips) received viscosupplementation with one of the three hyaluronate formulations. All patients were candidate for surgical treatment with a total hip arthroplasty. Three different products were consecutively used: Adant®, Synocrom® or Synvisc®. Patients were assessed 6 weeks after each infiltration using Visual Analogue Scale and Harris Hip Score. The Harris Hip Score increased significantly in two of the three groups compared to baseline, but no statistical significant difference was noted between the groups. Viscosupplementation provides significant pain reduction in two of the three groups. There is no significant difference in duration of the effect of the first infiltration between the three groups. The positive effect was still ongoing at the end point of the study in 46 hips: 51% of the patients did not undergo total hip arthroplasty, 3 years after viscosupplementation. Introduction Severe osteoarthritis (OA) of the hip is characterized by pain and reduced mobility. It affects primarily adults aged over 50 years, but the age of onset of hip OA linked complaints seems to decrease over the last decennia [12]. First-line treatment consists of pain reduction with analgesics and non steroidal anti inflammatory drugs (NSAID). Rehabilitation treatment, education and the use of walking aids can help providing pain relief and improving mobility. When those therapeutic options fail or the analgesic treatment causes intolerable side effects, joint lavage, hip osteotomy or total hip arthroplasty (THA) can be considered. This arthroplasty technique has greatly improved and patients experience pain reduction and improvement of the quality of life. However, the intervention still requires prolonged anaesthesia, and is followed by a long-term rehabilitation program. Though complications are rare and protheses have an increasing duration of life, reinterventions are still needed [4]. Viscosupplementation (VS) is the administration of hyaluronan and/or hyaluronic acid preparations to joint synovial fluid for the treatment of OA in order to restore the biologic properties of normal hyaluronic acid (HA). The use of VS with HA was first described to provide pain relief and to increase mobility of the knee joint. The VS is an effective treatment for OA of the knee with beneficial effects on pain, function and patient global assessment [14]. HA products have more prolonged effects than intra-articular corticosteroids [14]. Since 1984, this technique is also used for the management of OA of the hip joint [2]. In the available literature there is no consensus on the number of injections, the dosage per injection, the most appropriate formulation of HA, and the optimal method for controlling the needle positioning in the hip joint. Moreover, the patient selection criteria differ from one study to another. To our knowledge there is no documentation available on the possible time gain for patients treated with VS of the hip joint before they have to undergo THA. The results of a trial comparing three formulations of HA were presented. All patients were candidates for THA. The time of satisfactory pain relief, improved functionality obtained with each infiltration of HA and the delay in surgery were evaluated. Patients and methods Patient selection The use of VS in patients with severe OA is common practise in our hospital, thus not requiring approval of the ethical committee. Patients gave their informed consent for this prospective study. Between March 2001 and February 2005, 120 patients (126 hips), 49 males and 71 females, responding to the eligibility criteria listed below, received viscosupplementation. Three different products were consecutively used: Adant® (Tedec-Meiji Failma, Madrid, Spain) (Group 1) (Synthetic hyaluronic acid with an average molecular weight of 0.6–1.2 million Da), Synocrom® (Croma Pharma, Leobendorf, Austria) (Group 2) (Sodium hyaluronate with a average molecular weight of 1.6 million Da) or Synvisc® (Altana Pharma, Konstanz, Germany) (Group 3) (Hylan GF 20 with an average molecular weight of 6.0 million Da) (Table 1). Table 1Characteristics of infiltrated populationPeriod of treatmentPatientsGenderAgeAdant27-11-02 till 23-02-059135 ♂ and 65 ♀61.8 ± 12.8Synocrom19-01-04 till 04-10-04207 ♂ and 13 ♀62.1 ± 14.5Synvisc15-03-01 till 24-04-03157 ♂ and 8 ♀61.9 ± 15.3 Eligibility criteria Age between 30 and 70 years and suffering idiopathic radiologically confirmed hip OA. Visual Analogue Scale (VAS) score for pain greater then 30 (on a 100-point scale; 0 no pain and 100 “the worst pain imaginable”) Have persistent pain for longer than 1 month despite use of analgesics or NSAID’s. Be candidate for surgical treatment with a THA, according to the following criteria: continuous hip pain, also during the night, requiring daily intake of NSAID’s or pain medicationdisabled gait pattern and need of walking aid Be able to understand the information relative to viscosupplementation and to give informed consent. Exclusion criteria Pregnancy Contraindications to intra-articular hyaluronic-acid preparations Major hip dysplasia or congenital abnormality of the hip Patients with systemic corticosteroids or intra-articular corticosteroid injections in the last 6 months Contra-lateral THA or hip arthroscopy in the last 6 months Oral or parenteral anticoagulant therapy Previous hyaluronic acid hip infiltrations Skin diseases or infections Signs of haemarthrosis History of allergy or hypersensitivity to iodated contrast Treatment Patients received an intra articular infiltration with one of the three products. The manufacturer’s treatment recommendations were followed. Patients having initially experienced a satisfactory pain relief are offered a second and third infiltration or THA when the condition deteriorates. Injection of the viscosupplementation was performed under sterile conditions by the same experienced orthopaedic surgeon (MM) in all patients. After skin cleaning a lumbar puncture needle was inserted in a lateral approach. Layer by layer local anaesthesia was performed using lidocaine 1%. Iodinated contrast agent Ultravist® (Schering, Berlin, Germany) was injected. The needle positioning into the joint cavity was fluoroscopically controlled (Fig. 1). Arthrocentesis was carefully performed prior to each injection to remove any effusion. Fig. 1a Needle placement for injection of the viscosupplementation. b Fluoroscopic image of the needle insertion in the hip joint space After resting for 2 h, the patient was allowed to walk and to return home. The patient was advised to rest at home until the next morning. Oral symptomatic slow acting drugs for osteoarthritis were authorized if they were taken at a stable dose for more than 3 months prior to inclusion in the study. These analgesics were continued at a stable dose during the VS treatment. Evaluation All patients were assessed at baseline and 6 weeks after each infiltration. During this consultation in the outpatient clinic, the pain and functionality were evaluated using the VAS pain during walking score (100-point scale) and the Harris Hip Score (HHS). The latter is a clinical scoring system on a total of 100 points whereby the following subscales are rated: function (47 points), pain (44 points), range of motion of the hip (5 points) and absence of muscle contractures and length discrepancy (4 points). All side effects and complications of viscosupplementation were noted. In April 2005, all patients were contacted for follow-up assessment over the phone VAS and HHS. Statistical analysis Software (MS Excel)® was used to collect the data. Statistical analyses were of two kinds, first the differences in HHS and VAS pre-infiltration and post-infiltration (in the three groups) were compared using the paired t test and with the Wilcoxon paired test in the Synvisc group because this group is too small to analyse with the paired t test. The second comparison was made between the three groups (Synvisc, Synocrom and Adant). The differences in HHS and VAS evolution between the three groups were checked by the analysis of variance statistics with Tukey extension, which also was used for controlling if the groups were initially homogeneous. The duration of the viscosupplementation effect of the first infiltration was analysed using the Kaplan-Meier survival curves. The delay for the need of hip surgery was also analysed using a Kaplan-Meier survival curve. The statistical significance was set on P ≤ 0.05. Results The mean pre-infiltration HHS was comparable for the three groups and varied from 64.8 points in the Adant group to 66.8 points in the Synocrom-group. The post-infiltration HHS increased with 6.3 points in the Adant group (P < 0.001), with 10.6 points in the Synocrom group (P < 0.05) and with 6.1 points in the Synvisc-group (P > 0.05; Table 2). There was no statistical significant difference in the evolution of the HHS between the three groups (Table 2). Table 2Evolution in average HHS scoreHHS preHHS postDifferencePAdant64.8 ± 13.871.1 ± 15.7+ 6.3< 0.001Synocrom66.8 ± 13.877.4 ± 14.7+ 10.6< 0.05Synvisc66.3 ± 13.572.4 ± 14.5+ 6.1> 0.05 Viscosupplementation provided a highly significant pain reduction in the Adant-group (P < 0.0001), a significant pain reduction in the Synocrom-group (P < 0.05) and a pain reduction that did not reach significance in the Synvisc-group (P > 0.05). There was no significant difference in pain relief between the three treatment products (Fig. 2, Table 3). Fig. 2Pain relief in the three treatment groups a Adant, b Synocrom and c Synvisc The duration of the effect of the first infiltration in the three groups is shown in a Kaplan-Meier curve (Fig. 3). The first infiltration was the starting point. Endpoints were the second infiltration or operation of the afflicted hip, or when these were not applicable the latest patient contact, which can be considered as ongoing effect. Fig. 3Kaplan-Meier survival curve of the duration of effect of the first infiltration in days for the three different treatment groups There is no significant difference between the three groups (χ2 = 0.988 and P = 0.61). Figure 4 illustrates the duration of effect of the three groups globally. Fig. 4Kaplan-Meier survival curve of the duration of effect of the first infiltration in days (three treatment groups together) The positive effect was still ongoing in 46 hips, while in 80 hips patients had either received a second infiltration or THA at, the end of the study (Table 4). Table 3Evolution in average VAS pain during walking scoreVAS preVAS postDifference PAdant51 ± 2339 ± 27− 12< 0.0001Synocrom43 ± 2229 ± 23− 14< 0.02Synvisc47 ± 2630 ± 29− 17> 0.05Pre: at first infiltrationPost: at follow-up The delay in performing a hip operation is analysed using a Kaplan-Meier survival curve (Fig. 4). After 3 years, 51% of the patients have not undergone surgery. Because of the comparable outcome achieved with the three HA formulations, we did not differentiate between the three products in this survival analysis (Fig. 5). Fig. 5Kaplan-Meier survival curve for the delay to surgery in days for the three groups confounded Discussion The patients attending our outpatient clinic, who were judged eligible for THA, were offered the possibility of a VS therapy. Hundred twenty patients choose this treatment option and 126 hips were treated with Adant, Synocrom or Synvisc. All studies, published up till now, studying the effect of VS in the treatment of hip OA were performed in small patient groups (12–57 patients) [6, 9]. Table 4Duration of treatment effect after first infiltrationPatientsAverage duration (days)Effect of 1st infiltration still going on46 352 ± 258 (range 6–886)Effect of 1st infiltration terminated80174 ± 163 (range 13–724) The three preparations provided a significant pain relief and improvement of the HHS. The isolated Synvisc group never reached statistical significance in HHS score evolution and VAS during walk test after VS treatment; possibly due to the small number of patients (N = 15) in this group. As in the study of Tikiz [13], we found no significant difference between lower and higher molecular weight hyaluronic acids. We could identify two groups of responders, the ongoing responders who last on average for 352 days after the first infiltration. The second group, where patients either received a second infiltration or THA had an average duration of effect of 174 days. The VS method is widely used for OA of the knee joint, but there are only a few studies about its use in OA of the hip. Most authors agree that there should be a role for viscosupplementation in the treatment of hip OA. [1–3, 6, 7, 9–11] Our findings confirm the effect of VS in patients suffering OA of the hip. To our knowledge, this is the largest series of patients with hip osteoarthritis treated with viscosupplementation. There are no guidelines regarding the optimal number of injections needed to obtain optimal clinical response, which seems to be different [1, 2, 8, 10, 11] between products and dependent on the molecular weight (MW) [7]. We evaluated the time to relapse of one infiltration. Other authors advocated the use of ultrasound for controlling the needle position in the hip joint. [9, 3] In our series, we performed the infiltrations under fluoroscopy, in experienced hands this approach has proven to be accurate and safe. We use a very low amount of contrast liquid to avoid dilution of the product. Of all the reports published, we have the longest follow-up, up to approximately 3 years. The analysis by means of the Kaplan-Meier survival curve learns that there is a 50% probability of long-term effect (more than 2 years). We saw no infectious adverse events and no serious systemic reactions, but all the interventions are performed in the operating theatre under strict aseptic conditions. The adverse events rates ranged from 10 to 30% which is slightly higher than the rates reported in VS treatment of knee OA [7]. Repeated injections did not increase the risk of adverse events [7]. Some patients experienced transient hip pain after the infiltration but made a full recovery in the next days. In the study of Conrozier [6], transient hip pain was reported after 10.1% of the injections. The products used in this study were all well tolerated. Complications are rare but a single case of septic arthritis was reported after multiple injections of hyaluronate and glucocorticoid [5]. Gout, pseudogout and chondrocalcinosis have not been reported after hip infiltrations. Intra articular injection of hyaluronic acid can provide long-term pain relief and improvement of joint function even in patients eligible for THA. Despite the use of fluoroscopy, this technique can be performed in an outpatient clinic, allowing patients to return home on the day of the intervention. At present, viscosupplementation therapy for OA of the hip is only palliative. It can be an alternative for young THA candidates, patients with surgical contra-indications and patients in whom NSAID use is not appropriate. There is only 1 randomised, double blind, placebo controlled trial of patients with hip OA [15]. The study was designed with a three-armed parallel-group (Hyalgan vs. methylprednisolone vs. lidocaine). One hundred and one were treated. There was no statistically significant difference between the three products on any outcome measure including the primary outcome measure, i.e., ‘Pain on walking’, at three 3 months of follow-up. The results of our study should be considered in the light of the limitations of the design of this study. It is a non placebo controlled non randomised prospective study. It is known from experience with knee OA that the placebo effect of VS tends to be substantial [8]. The dimension of the groups treated with the three different products differs from 15 to 91 patients. Viscosupplementation with hyaluronic acid seems to be a valuable technique for the management of painful OA of the hip that may delay the need for surgical intervention. Further prospective randomised placebo controlled studies are necessary to draw definite conclusions.	[ "viscosupplementation", "hyaluronic acid", "hyaluronan", "intra-articular", "injections", "hip osteoarthritis", "non-operative treatment" ]	[ "P", "P", "P", "P", "P", "P", "M" ]
Exp_Brain_Res-4-1-2257995	The effect of age on task-related modulation of interhemispheric balance	Normal aging is associated with less lateralised task-related activation of the primary motor cortices. It has been hypothesized, but not tested, that this phenomenon is mediated transcallosaly. We have used Transcranial Magnetic Stimulation to look for age-related changes in interhemispheric inhibition (IHI). Thirty healthy individuals (aged 19–78 years) were studied using a paired-pulse protocol at rest and during a low-strength isometric contraction with the right hand. The IHI targeting the right motor cortex was assessed at two intervals, 10 ms (IHI10) and 40 ms (IHI40). The corticospinal excitability of the left hemisphere was assessed by means of input–output curves constructed during voluntary construction. Age was not correlated with IHI10 or IHI40 at rest. During muscle contraction IHI tended to increase at both intervals. However, this increase in IHI during the active condition (changeIHI) was less evident with advancing age for the 40 ms interval (r = 0.444, P = 0.02); in fact a degree of disinhibition was often present. There was no correlation between age and changeIHI10. Age was negatively correlated with the area under the recruitment curve (r = −0.585, P = 0.001) and the size of the maximum MEP collected (r = −0.485, P = 0.007). ChangeIHI and measures of corticospinal excitability were not intercorrelated. In conclusion, task-related increases in interhemispheric inhibition seem to diminish with advancing age. This phenomenon is specific for long-latency IHI and may underlie the age-related bihemispheric activation seen in functional imaging studies. The mechanism underlying changes in IHI with advancing age and the association with changes in corticospinal excitability need further investigation. Introduction Physical performance declines with increasing age (Hackel et al. 1992; Smith et al. 1999). Although age-related changes in muscular (Vandervoort 2002) and neural architecture (Dorfman and Bosley 1979; Haug and Eggers 1991; Madden et al. 2004) are an important cause of this decline, there is increasing interest in the role of functional alterations occurring throughout the motor system. Excitation–contraction uncoupling (Delbono et al. 1995), motor unit remodelling (Wang et al. 1999), changes in the agonist–antagonist activation pattern (Hortobagyi and Devita 2006) and reorganization of the central motor pathways (Kido et al. 2004; Ward and Frackowiak 2003; Minati et al. 2007) are all areas currently under study. Understanding the mechanisms underlying these phenomena will aid differentiation between inevitable functional failure and potential compensatory strategies which in turn may allow the development strategies to combat this age-related decline (Delbono 2003). Functional imaging has been used to assess age-dependent changes in the cerebral motor system in humans (for review see Ward 2006). In general, motor task-related brain activation is seen in a wider network with advancing age. The exact pattern of age-related change depends on the task being performed. However, it has been repeatedly shown that in older individuals there is increased bihemispheric activation during the execution of motor tasks that in younger individuals involve mainly lateralized processing. In particular, with advancing age there is less de-activation of the primary motor cortex (M1) ipsilateral to the moving hand (Ward and Frackowiak 2003; Naccarato et al. 2006; Ward et al. 2007). It has been suggested, but not tested, that the underlying mechanism for this phenomenon involves changes in interhemispheric connections between the motor cortices. Transcranial magnetic stimulation (TMS) can be used to study interhemispheric connectivity between the motor cortices. At rest, this connectivity consists primarily of an inhibitory effect (interhemispheric inhibition, IHI) with a latency of 6–50 ms (Ferbert et al. 1992; Di Lazzaro et al. 1999; Daskalakis et al. 2002; Chen 2004). This tonic IHI is modulated and follows a specific time course during preparation and execution of voluntary movement (Murase et al. 2004; Duque et al. 2005, 2007). During the execution of an isometric muscle contraction, the IHI (at 10 ms latency) targeting the M1 ipsilateral to the moving hand has been shown to increase (Ferbert et al. 1992). This “extra” inhibition is in keeping with the primarily unilateral activation pattern seen in imaging studies of movement and has been interpreted as a means of minimizing mirror activity (Leocani et al. 2000; Duque et al. 2005). Task-related changes in longer latency IHI, i.e. 40 ms, have not been studied in detail but there is some evidence that they may be distinct to the ones seen in short latency IHI (Chen et al. 2003). Previous TMS studies have shown that there is reduction of activity in the intracortical inhibitory circuits in the elderly (Peinemann et al. 2001; Hortobagyi et al. 2006). It has been suggested that reduced inhibition represents a mechanism to compensate for the deleterious effects of aging on several levels of the motor system including the primary motor cortex (Pitcher et al. 2003; Oliviero et al. 2006) but these notions remain mainly hypothetical (Wassermann 2002). The main purpose of this study was to investigate whether normal aging has an effect on the task-related modulation of IHI measured at two different latencies, 10 and 40 ms. We hypothesized that with increasing age there would be less extra activation in the inhibitory circuits targeting the M1 ipsilateral to the moving hand. In order to gain some insight of the mechanisms and the physiological meaning of these changes we also performed measures of corticospinal excitability on the side contralateral to the moving hand and looked for correlations with age and IHI measures. Methods Subjects Thirty healthy right-handed volunteers (mean age 42.9 years, range 19–78; 30% female) participated in the study after giving informed consent. They reported no history of neurological illness, psychiatric history, vascular disease or hypertension and they were not taking regular medication. The study was approved by local Ethics Committee. Transcranial magnetic stimulation Subjects were seated in an armchair with their eyes open. EMGs were recorded via Ag/AgCl electrodes placed over the First Dorsal Intersosseus (FDI) bilaterally, using a belly-tendon montage. Signals were filtered (30 Hz to 10 kHz), amplified using a Digitimer 360 (Digitimer Ltd, Welwyn Garden City, Herts., UK) and stored on computer via a Power 1401 data acquisition interface (Cambridge Electronic Design Ltd, Cambridge, UK). Analysis of data was carried out using Signal Software (Cambridge Electronic Design). Two figure-of-eight coils connected to two monophasic Magstim 200 stimulators were used for the experiments (all Magstim Co., UK). A 70-mm coil was used for motor hot spot identification and threshold measurements on both sides of the brain. The motor hotspot was defined as the scalp location where TMS consistently resulted in the largest MEP. The resting motor threshold (RMT) was defined as the lowest intensity needed to evoke an EMG response of 50 μV in 50% of the trials with the FDI relaxed; the active motor threshold (AMT) was defined as the intensity which evoked a 200 μV EMG response in 50% of the trials with a background FDI contraction of 10–15% of the maximum voluntary contraction (MVC). For all studies requiring activation of the FDI, visual feedback was provided using an oscilloscope. A recruitment curve (RC) for the active MEP amplitude elicited in the left FDI was obtained using the 70 mm coil while the subjects maintained a background FDI contraction of 15–20% MVC. Ten MEPs were collected and averaged at the following stimulus intensities: 90, 100, 110, 120, 140, 150, 160 and 170% AMT. The contraction level was tested by measuring the mean value of the rectified EMG in the 80 ms preceding the TMS pulse. The averaged peak-to-peak amplitude of the unrectified MEPs was then expressed as a ratio of the maximum peak-to-peak amplitude of the unrectified compound action potential (CMAP) evoked by supramaximal electrical stimulation of the ulnar nerve at the wrist using a Digitimer pulse stimulator (model DS7). Interhemispheric inhibition (IHI) was measured using a standardised paired-pulse paradigm (Ferbert et al. 1992). IHI is expressed as the reduction in the response to a suprathreshold TMS pulse (test) delivered to the M1 when another suprathreshold pulse is delivered to the contralateral M1 6–50 ms earlier (conditioning). Here, we concentrated on the IHI targeting the right M1. Hence the conditioning pulse was given over the left M1 using the same 70-mm coil. For the test pulse a smaller 50 mm figure-of-eight coil was held over the right M1; this allowed us to accommodate two coils on the head of all the subjects without compromising the exact positioning of the coils over the motor hot-spot (Ferbert 1992). MEPs were measured from the left FDI. Ten single (test) and ten paired-pulse (conditioning + test) trials were randomly intermingled and averaged. IHI was defined as the conditioned/test MEP amplitude ratio, smaller values reflecting stronger IHI. Interstimulus intervals (ISIs) of 10 and 40 ms were studied in different blocks. IHI was initially measured at rest (restIHI10 and restIHI40). The stimulation intensity for both the conditioning and the test stimuli was adjusted to evoke an MEP of 1–1.5 mV in the contralateral FDI muscle. For the active condition (activeIHI10 and activeIHI40) the subjects were instructed to contract the right FDI to 15–20% of their MVC in response to an auditory cue preceding the conditioning stimulus by 600 ms. In this way, we made sure that the subject had reached the target level of background contraction for at least 200 ms before the conditioning pulse. This was to avoid measuring changes in IHI that are associated with the preparation of the movement rather than a steady isometric contraction (Duque et al. 2005, 2007). EMG activity in the left FDI was also recorded and contaminated trials were rejected to avoid the confounding effect of mirror activity. The stimulation intensity both for the conditioning and test stimulus was the same as in the resting state, as in previous studies (Ferbert et al. 1992). The absolute values of activeIHI were then expressed as a ratio to the values at rest for the respective ISI (changeIHI10 and changeIHI40). ChangeIHI therefore reflects the change seen in the IHI targeting the right motor cortex when the right hand is active. Values <1 reflect stronger inhibition, while values >1 reflect less inhibition. Statistical analysis Analysis was performed using SPSS v.14 (SPSS Inc., USA). Exploratory plotting and statistical tests (Shapiro-Wilk) confirmed normal distribution for most variables. Some of the variables were skewed by a single outlier and this could not be corrected by logarithmic transformations; the outlier was therefore discarded from the dataset to allow use of parametric tests. Data obtained during the RC paradigm were plotted against the stimulation intensity. In most subjects the resulting curves were sigmoidal in shape. Two representative examples, one from a young and one from an old subject, are shown in Fig. 1. The area under the RC (AUC) was calculated using the method of trapezoid integration to provide a summary measure of the corticospinal output across all stimulation intensities. Data points were then fitted in the Boltzmann sigmoidal model based on the modified Levenberg–Marquardt nonlinear least-mean-squares algorithm (Press et al. 1986). This function is often used to provide parameter estimates for RCs (Capaday et al. 1999; Devanne et al. 1997). The data were best fitted by a 4-parameter equation in which the amplitude of the MEP at a given stimulation intensity (I) is estimated as: MEPmax and MEPmin represent the maximum and minimum MEP amplitude, respectively; I50 represents the stimulation intensity required to get a response 50% of the maximum. The inverse of the slope parameter (1/slope) is directly proportional to the maximal steepness of the curve, which occurs at I50 (Devanne et al. 1997). The R2 co-efficient of determination was greater than 0.89 in all subjects. Fig. 1Representative examples of the recruitment curves plotting the amplitude of the active MEPs against the stimulation intensity. The circles represent the collected data. The line represents the relation predicted by the Boltzmann model. The area under the curve (grey area) was calculated with the method of trapezoid integration using the actual data collected during the construction of the curves. CMAP, compound motor action potential, AMT, active motor threshold Pair-wise comparisons between different TMS measures were performed using paired t tests. Correlations with age and between TMS measures were assessed by computing Pearson’s product-moment correlation coefficient. We have also looked for correlations between these parameters and gender. Partial and part correlations were employed as appropriate. Significance level was set at P < 0.05. Results The mean and range of the values for all TMS measures are summarized in the Table 1. Table 1TMS measures of corticospinal excitability and interhemispheric inhibitionCorticospinal excitabilityLeft hemisphere Right hemisphereMotor thresholds (% stimulator’s intensity)Resting (RMT)37.6 (28–65)36.4 (29–50)Active (AMT)28.4 (19–45)28 (22–39)RC of active MEP amplitude(CMAP corrected)Mean MEP amplitude recorded 90% AMT0.02 (0.01–0.06)– 100% AMT0.03 (0.01–0.07)– 110% AMT0.06 (0.02–0.18)– 120% AMT0.09 (0.03–0.35)– 140% AMT0.25 (0.06–0.51)– 150% AMT0.30 (0.10–0.57)– 160% AMT0.34 (0.13–0.56)– 170% AMT0.22 (0.39–0.62)–Maximum MEP amplitude recorded0.38 (0.13–0.62)–AUC12.8 (1.8–23.4)–Parameter estimates (Boltzman model) MEPmax/CMAP0.4 (0.13–0.72)– I50 (%AMT)137 (115–168)– 1/slope0.14 (0.05–0.37)–IHI (targeting the right M1)10 ms40 msrest IHI0.61 (0.32–0.9)0.6 (0.28–1.1)activeIHI0.54 (0.1–0.89)0.57 (0.2–1)changeIHI (activeIHI/restIHI)0.96 (0.42–2.47)0.97 (0.5–1.49)CMAP, compound motor action potential; AUC, area under the curve; MEPmax, maximum MEP amplitude estimated; I50, stimulation intensity needed to get a response of 50% of the maximal MEP; IHI, interhemispheric inhibition; restIHI, IHI measured with both hands relaxed; activeIHI, IHI measured during a tonic contraction of the dominant hand at 15–20% MVC; changeIHI, values <1 indicate stronger IHI at the active condition* Significant correlation with advancing age (P < 0.05) Corticospinal excitability Threshold values were not different between the left and right M1. Age was not correlated with RMT or AMT on either side or the amplitude of the CMAP. The mean amplitude (CMAP corrected) of the active MEPs recorded during the construction of the recruitment curves was negatively correlated with age at all stimulation intensities between 100–160%AMT (100%AMT: r = −0.431, P = 0.02; 110%AMT: r = −0.397, P = 0.03; 120%AMT: r = −0.481, P = 0.007; 140%AMT: r = −0.556, P = 0.001; 150%AMT: r = −0.457, P = 0.11; 160%AMT: r = −0.463, P = 0.01). The same was true for the amplitude of the maximum MEP recorded (Pearson’s r = −0.485, P = 0.007) (Fig. 2a). Age was negatively correlated with the total area under the RC (Pearson’s r = −0.585, P = 0.001) (Fig. 2b). From the parameters estimated using the Boltzmann model, MEPmax showed a weak correlation with age (r = −0.397, P = 0.04); age was not correlated with the maximum slope or I50. Fig. 2Correlations between age and TMS parameters. There was a significant negative correlation between the amplitude of the maximum MEP recorded during the construction of the RC (a) and the total area under the RC (b). Age was negatively correlated with changeIHI at 40 ms (d) but not at 10 ms (c). CMAP, compound motor action potential, RC, recruitment curve constructed during tonic contraction of the target muscle, IHI, interhemispheric inhibition, restIHI, IHI measured with both hands relaxed, activeIHI, IHI measured during a tonic contraction of the dominant hand at 15–20% MVC, changeIHI, values <1 indicate stronger IHI at the active condition Gender was not correlated with any of the TMS measures of corticospinal excitability. Interhemispheric inhibition Most subjects showed some amount of IHI at rest; only in one subject could restIHI40 not be elicited. Paired-samples t test did not show any difference between restIHI10 and restIHI40 (P = 0.44). Because the amount of IHI depends on the intensity used for both the conditioning and the test stimulus (Ferbert et al. 1992; Chen et al. 2003) we looked for correlations between these values and age. Age was not correlated with the intensities used to stimulate either the left M1 (70-mm coil; 45.7 ± 7.1 for IHI10 and 45.8 ± 7.2 for IHI40) or the right M1 (50-mm coil; 46.2 ± 8.6 for IHI10 and 46.9 ± 8.1 for IHI40). The amplitude of the MEPs elicited by the test stimulus and by the conditioning stimulus was slightly higher for the restIHI40 but the values fell within the target range of 1–1.5 mV (see below). During activation of the right hand the MEP elicited by the test stimulus in the resting left FDI tended to increase (from 1.23 ± 0.53 to 1.44 ± 0.67 for IHI10 and from 1.48 ± 0.58 to 1.64 ± 0.70 for IHI40), but this was not significant across the group, as reported earlier for low-strength isometric contractions (Liepert et al. 2001). As expected, the amplitude of the MEP elicited by the conditioning pulse in the right FDI increased significantly during activation of the muscle (from 1.27 ± 0.40 to 5.95 ± 1.69, P < 0.001 for IHI10 and from 1.5 ± 0.47 to 6.17 ± 2.57, P < 0.001 for IHI40). There was an overall tendency for stronger IHI during activation of the right FDI at both ISIs, but this did not reach significance. In fact, changeIHI was quite variable ranging from more inhibition to less inhibition compared to the resting condition. Because changeIHI might be biased by the amount of IHI at rest we looked for correlations between changeIHI and restIHI; there was none for either ISI. There was no correlation between age and changeIHI10 (Fig. 2c). However, we found a significant positive correlation between age and changeIHI40 (r = 0.444, P = 0.02) (Fig. 2d). In other words, when younger individuals activated their right hand the amount of inhibition targeting the ipsilateral (right) motor cortex increased; with increasing age this phenomenon was gradually attenuated and often reversed. Because age showed a significant correlation with the amplitude of the active MEPs we performed semi-partial correlations correcting for the change seen in the amplitude of the conditioning MEP between the resting and the active condition (changeMEP = activeMEP/restingMEP). Again no correlation was found between age and changeIHI10; the correlation between age and changeIHI40 remained and was in fact slightly stronger (r = 0.553, P = 0.006). Gender was not significantly correlated with any of the IHI measures; in line with a previous report (De Gennaro et al. 2004) there was a tendency for a positive correlation between female gender and the amount of restIHI10 targeting the right hemisphere, but this finding did not reach significance in our group of subjects (r = 345, P = 0.07). Finally, changeIHI was not correlated with any of the measures of corticospinal excitability elicited from the stimulation of the left M1. Discussion The purpose of this study was to investigate the effect of age on the interhemispheric balance between the two primary motor cortices and in particular the modulation of inhibition targeting the right M1 during activation of the right hand (changeIHI). We found that the amount of tonic inhibition (restIHI) was very variable across the group but age could not explain this variability. However, age was significantly correlated with changeIHI. As reported previously, when young individuals perform a low strength grip, the inhibition targeting the ipsilateral M1 tends to increase (Ferbert et al. 1992). Our data show that with increasing age this extra inhibition is not as strong and some degree of disinhibition may be present instead. Interestingly, the effect of age was significant for the 40 ms ISI (changeIHI40) but no correlation was found between age and changeIHI10. Our results on changeIHI are in keeping with the findings from functional imaging studies. A number of studies have consistently shown that task-related activation is less lateralised in older individuals (for review see Ward 2006). This bilateral recruitment includes primary and non-primary motor areas; activation of the latter seems to characterise more complex tasks. Regarding the primary motor cortices, reduced deactivation of the M1 ipsilateral to the moving hand has been a common finding among studies using different scanning paradigms (Hutchinson et al. 2002; Ward and Frackowiak 2003; Naccarato et al. 2006; Ward et al. 2007). It has been hypothesized that this phenomenon could be mediated transcallosally. In support of this hypothesis, we have directly demonstrated that there is task-related reduction in interhemispheric inhibition with advancing age at least when the right hand is activated. The differential effect of age on changeIHI10 and changeIHI40 is an interesting finding that adds to the current knowledge of the cortical circuits that these measures reflect. So far, it is thought that both short latency IHI (8–12 ms) and long latency IHI (40–50 ms) are mediated through excitatory callosal projecting neurons that are distinct from corticospinal neurones (Lee et al. 2007). In accordance with this view, corticospinal output and IHI measures in this study did not correlate. The callosal neurons are thought to activate inhibitory interneurons in the contralateral M1 (Daskalakis et al. 2002; Chen 2004), but these connections may differ for short and long latency IHI (Chen et al. 2003; Kukaswadia et al. 2005). The neurotransmitters involved are not entirely understood but there is evidence of significant involvement of GABAB-ergic activity. However, Irlbacher et al. (2007), have recently demonstrated that exogenous enhancement of GABAB activity strengthens long IHI but has little effect on short IHI at rest. Long IHI was also sensitive to GABAA agonists (Irlbacher et al. 2007). Finally, Chen et al. (2004), has shown that activation of the target muscle results in a reduction in the IHI targeting the active M1 (the opposite direction than the one studied here) measured at 8 ms but has little effect on the IHI40. We have shown that advancing age has an effect on the modulation of IHI during voluntary muscle contraction at 40 ms but not at 10 ms. We have thus provided additional evidence that IHI10 and IHI40 may respond differently to dynamic changes within the motor system and thus may represent physiologically distinct phenomena. This evidence could be strengthened in the future by testing the effect of age on IHI in both directions, i.e. the IHI targeting the motor cortices ipsilateral and contralateral to the moving hand. Why is IHI40 not enhanced during voluntary contraction in older individuals? One possible explanation is age-related failure of the involved pathways. MRI studies have reported age-related reduced fractional anisotropy within the corpus callosum suggesting callosal fibre degeneration (for review see Minati et al. 2007). However, for callosal degeneration to be the cause of our finding, one would expect some effect on IHI10 as well. Another possibility is that the inhibitory circuits with which the callosal fibres synapse are less excitable. Previous studies have provided some evidence of age-related reductions in cortical inhibition. For example, GABAA-ergic short-interval intracortical inhibition (SICI) assessed with a biphasic stimulator (Peinemann et al. 2001) and cortical reciprocal inhibition (Hortobagyi et al. 2006) are reduced in older individuals. However, Wasserman et al. (2002) failed to show a correlation between SICI and advancing age in a large series of normal subjects. Others have found increased SICI in middle-aged adults (Kossev et al. 2002). Further experiments, which include active conditions are needed to study these associations in the aging brain. Finally, it should be noted that although TMS-measured IHI is thought to reflect mainly transcallosal connections, Gerloff et al. (1998) has demonstrated a role for other pathways involving subcortical and/or spinal structures, the exact nature of which is unclear. Age related changes in these pathways may therefore also contribute to the correlations that we have observed. Another question is whether the age-related reduced activity reported in multiple inhibitory systems, including IHI is an inevitable consequence of neurodegeneration or in fact represents a compensation strategy of the aging brain. Similar changes have been described in response to brain injury and have been interpreted as markers of cortical reorganization (Ward and Cohen 2004; Talelli et al. 2006). The relative preservation of changeIHI10 could suggest that the release from inhibition is not uncontrolled or at least it occurs in a hierarchical manner. For example, IHI10 may be important for direct suppression of mirror activity in the contralateral M1 (Duque et al. 2005). Longer latency IHI40, on the other hand, may reflect polysynaptic pathways extending even beyond the primary motor cortices. In that case relatively less inhibition in an active state could allow bilateral recruitment if that was necessary. In this study, there was no correlation between changeIHI40 (or any other IHI measure) and any of the measures of corticospinal excitability despite the fact that the latter appeared to be compromised in older individuals. This finding could mean that changeIHI40 does not reflect an adaptation to the failing corticospinal pathways. However, it should be kept in mind that if compensatory mechanisms are indeed operating with increasing age (or decreasing corticospinal output) TMS may not measure the real deficit but the compensated deficit. The next reasonable step could be to study a series of elderly individuals known to have different levels of manual dexterity and look for correlations between electrophysiological and behavioural measures, including mirror movements during performance of complex motor tasks. We found variable effects of age on parameters relating to corticospinal excitability. Motor thresholds were not affected in keeping with previous reports (Wassermann 2002; Oliviero et al. 2006; Hortobagyi et al. 2006). We have found negative correlations between age and the amplitude of the active MEP at most stimulation intensities, between age and the amplitude of maximum MEP and finally between age and the total area under the recruitment curve (AUC). These results are in agreement with previous findings of reduced amplitude of the MEPs elicited during an isometric contraction (Oliviero et al. 2006; Sale and Semmler 2005). The values derived from the Boltzmann model did not add extra information, since there was no correlation between age and the maximal slope of the curve or the stimulation intensity required to obtain 50% MEPmax (I50). Pitcher et al. (2003) have previously reported that I50 significantly increased with increasing age. In other words, the recruitment of the MEPs was slower at low stimulation intensities. The authors discussed that this finding may reflect reduced excitability or asynchronous activation of spinal motorneurons. In this study we did not perform measures of spinal excitability mainly due to time limitations. We have, however, constructed our RCs during background activation of the target muscle; that means that some a-motorneurons are already activated and the resulting measures are thus less subject to changes in spinal cord excitability (Day et al. 1989). We propose that increasing age does have a detrimental effect on the total output of primary motor cortex and that the mechanisms underlying (or compensating for) this effect are variable; thus a summary variable such as the AUC might be more suitable as a surrogate marker of overall corticospinal output. In conclusion, this study has shown for the first time that there is an age-related reduction in the extra inhibition targeting the right hemisphere during an isometric handgrip with the right hand. These changes may underlie the bihemispheric pattern of activation seen in functional imaging studies of older individuals performing a unimanual hand task. Should our findings be confirmed, studies of interhemispheric balance may prove to be a useful marker of reorganization in the aging brain.	[ "aging", "transcranial magnetic stimulation", "interhemispheric inhibition", "motor system" ]	[ "P", "P", "P", "P" ]
Anal_Bioanal_Chem-4-1-2226000	Characterization of cellular chemical dynamics using combined microfluidic and Raman techniques	The integration of a range of technologies including microfluidics, surface-enhanced Raman scattering and confocal microspectroscopy has been successfully used to characterize in situ single living CHO (Chinese hamster ovary) cells with a high degree of spatial (in three dimensions) and temporal (1 s per spectrum) resolution. Following the introduction of a continuous flow of ionomycin, the real time spectral response from the cell was monitored during the agonist-evoked Ca2+ flux process. The methodology described has the potential to be used for the study of the cellular dynamics of a range of signalling processes. Introduction The advance of miniaturized microfluidic systems for chemical and/or biochemical applications based on so-called Lab-on-a-Chip technology has demonstrated that such micro systems represent the ability to “shrink” conventional bench systems to the size of a few square centimetres with major advantages of speed, performance, integration, portability, reduced sample/solvent quantity, automation, hazard control and lower cost [1–6]. These merits are important for a variety of applications in analytical chemistry, biochemistry, clinical diagnosis, medical chemistry and industrial chemistry [1, 2]. Consequently, numerous micro total analysis systems (μ-TAS) and micro reactor systems have been developed, and many more are currently under investigation [2]. For the study of cellular and subcellular systems, a wide range of analytical methods have been used with fluorescence techniques being the most common. In addition, fluorescence-based imaging is a highly attractive methodology for the study of organelle dynamics, identifying subcellular compartments and monitoring biological kinetics [7]. Over the last two decades, Raman spectroscopy has become an increasingly important technology with ability to study the biophysics and biochemical processes involving cells [4, 8–12]. Since Raman spectroscopy is based on vibrational transitions where frequency shifts are associated with specific molecular vibrations within the sample of interest, it enables the identification of polarizable bio/chemical species, the elucidation of molecular structure and the investigation of interface reactions, all in a non-destructive manner. In addition, unlike fluorescence-based techniques, Raman spectroscopy does not require labelling dyes and since water is almost Raman “transparent”, the technique is ideally suited for analysing cell-based biological systems. Coupling a Raman spectrometer with a confocal microscope enables the acquisition of full spectral information with a high spatial (<1 μm) resolution in three dimensions. Raman spectroscopy, however, suffers from an inherent poor level of sensitivity compared for example to fluorescence, which can be 12–14 orders of magnitude more sensitive. As a consequence, despite advances in detector technologies, the technique can be less than ideal for the direct detection of intracellular components present at low concentrations, which may take from a few tens of seconds to a few minutes for spectral acquisition [12]. With the unexpected discovery of surface-enhanced Raman scattering (SERS) by Fleischman et al. [13], Raman intensity can now be dramatically increased (a factor of up to 105–1010) with the inclusion of metallic nano structures either on a substrate surface or in a colloidal solution [8]. For cellular and subcellular analysis with SERS, colloid nanoparticles (e.g. silver or gold) are normally loaded into cells by different means such as general incubation (fluid-phase uptake) or ultrasonication-assisted uptake [14, 15]. Due to its chemical inactivity, gold nanoparticles are generally regarded to be more suitable for incorporation within living cells [14]. To analyse living cells using Raman spectroscopy, the cells are usually fixed at a specific location. The most commonly used method for cell fixation is to seed or grow cells on a substrate, e.g. a microscope cover slide. Other approaches such as optical tweezers have also been reported [9]. However, the conventional batch operation in a static system limits the in situ Raman analysis, especially when a series of reagent treatments are required. With those limitations in mind, the development of microfluidic and associated Lab-on-a-Chip technologies provides unique opportunities for delivering and immobilizing cells on a microchannel surface, prior to introducing different reagents with a continuous flow in a given sequence [3, 4] whilst under spectroscopic investigation. In this study, we report the successful integration of a range of techniques including microfluidics, surface-enhanced Raman scattering (SERS) and confocal microspectroscopy which enables in situ characterization of single living CHO cells with high spatial and temporal resolution. By using the microfluidic methodology, cells and reagents were introduced into the chip in a continuous flow as a series of plugs in a given sequence where agonist (ionomycin) evoked intracellular Ca2+ fluxes and the cell’s real time spectral response was recorded. Materials and methods Confocal Raman microspectroscopy All Raman spectra were acquired with a LabRam inverted microscope spectrometer, manufactured by Jobin Yvon Ltd. Figure 1 shows the schematic of the experimental setup. The spectrometer was equipped with dual laser sources at wavelengths of 780 nm (diode laser, 70 mW) and 633 nm (He–Ne laser, 20 mW), confocal optics, a holographic transmission grating, and a charge coupled device (CCD) detector with 1,024 × 256 pixels. The instrument included a precision motorized X–Y sample stage for automated mapping at spatial resolution down to less than 1 μm and extensive software support (LabSpec 4.18) for data processing. In this study, an objective lens of ×50 magnification, 17-mm working distance and numerical aperture (NA) of 0.45 was used (L Plan SLWD 50, Nikon, Japan). This objective lens was mounted on a PI-721.10 piezo actuator (Physik Instrumente, Germany) for automatic focussing of the microscope objective at different depths in the Z direction enabling 3D mapping. A grating with 1,800 grooves mm−1, a confocal aperture of 300 μm and an entrance slit of 150 μm were selected for the experiments. The Raman spectrometer wavelength range was calibrated using the centre frequency of the silicon band from a silicon sample (520.2 cm−1). Using these conditions, a typical acquisition time of 1 s was used to collect SERS spectra from cells within the microchannel. Fig. 1Schematic of experimental setup Microfluidic device fabrication The microfluidic device was constructed using a manifold clamping method according to published procedures with some adaptations [16–18]. The assembly of the device is illustrated in Fig. 2a. Briefly, the microchip consisted of a PARAFILM® sheet (thickness 130 μm, American National Can Company, US) with a channel network and two glass plates which sandwiched the polymer film. The Y-shaped channel network (Fig. 2b) cut through the film was 500-μm wide. The top glass plate (B-270, 25 × 25 × 3 mm) had three holes (diameter 1.5 mm) drilled through at appropriated positions in order to link the ends of the channels with inlet/outlet tubing. The bottom glass plate was a thin quartz coverslip (22 × 22 mm, Agar Scientific Ltd, UK) which had a thickness of 250 μm in order to minimise the glass background during Raman measurements. This sandwich chip was then clamped using two aluminium frames with screws. The windows on the frames were designed for tubing connections (through top frame) and for optical passage (through bottom frame). Fig. 2(a) Assembly of microfluidic device and (b) Y-shaped channel network (channel depth 100 μm, width 500 μm) with cells loaded (c) for examination Two KDS 200 syringe pumps (KD Scientific Inc., USA) were used to deliver cells in suspension and test solutions into the microchip channel (Fig. 2c). Ethylene tetrafluoroethylene (ETFE) polymer tubing with an inner diameter of 250 μm, on–off valves, and appropriate fittings and connectors, all obtained from Upchurch (Upchurch Scientific Inc., USA), were used for plumbing to link the chip and the syringes. Cell culture and assay reagents CHO-K1 (Chinese hamster ovary, Cricetulus griseus) cells were supplied by ATCC/LGC Promochem (ATCC® No. CCL-61™, LGC Promochem, UK). The cells were cultured routinely in DMEM/F-12 medium without L-glutamine (Invitrogen Ltd, UK) which was supplemented with fetal bovine serum (Invitrogen Ltd) to a final concentration of 10%, and L-glutamine (Invitrogen Ltd) to a final concentration of 4 mM. An incubator was used at 37 °C supplying 5% CO2. The concentration of cells used in loading the chips during this experiment was in the range of 7.5 × 106 cells mL−1. A wash solution consisting of a modified Tyrodes buffer was used to wash the cells and to prepare the test solution. The Tyrodes buffer was composed of 145 mM NaCl, 2.5 mM KCl, 10 mM HEPES, 10 mM D-glucose and 1.2 mM MgCl2. CaCl2 (99.5%, BDH AnalaR, 150 mM dissolved in Tyrodes buffer) and probenecid (98%, Sigma, 0.834 M dissolved in 1 M NaOH aqueous solution) were then added to the Tyrodes buffer giving final concentrations of 1.5 mM and 2.5 mM for CaCl2 and probenecid, respectively. Ionomycin test solutions were made by adding ionomycin stock (1 mM in DMSO) into wash solutions for a concentration of 50 μM. Ionomycin was obtained from Calbiochem (Calbiochem of EMD Biosciences, Inc., USA), and DMSO from Sigma–Aldrich (99%, D2650). The flow rate for the introduction of ionomycin test solution into the microchip was 2 μL min−1. Gold colloid with a particle size of 50 nm (EM.GC50), suspended in water, was supplied by BBinternational Ltd, UK. The gold nanoparticles were introduced into cells by a passive uptake method where the cells were incubated with gold colloid solution at desired concentration (20% colloid solution, v/v) and room temperature for 50 mins. Prior to loading into the microfluidic chip, the cells were washed with wash solution to remove the culture media and gold nanoparticles outside cells. The cells were re-suspended in the wash solution for measurements. Results and discussion SERS effects After loading cells into the microfluidic channel, Raman spectra were taken from a selected single cell (Fig. 2c). By using the near-infrared laser excitation (780 nm) the optical and thermal effects of laser illumination on the living cells were minimised [19, 20]. In addition, using a near-infrared laser can significantly reduce the fluorescence interference background and light scattering from the quartz base plate. It has been reported that silver or gold nanoparticles with an individual size in the range of 20–60 nm can yield significant enhancement of Raman scattering for cellular analysis when nanoparticles are introduced into cells [14]. This has been confirmed in this study by using gold nanoparticles with a size of 50 nm. Figure 3 shows a typical spectrum (upper) taken from a cell incubated with gold colloid solution. The spectrum was recorded in the range from 300 to 2,800 cm−1 with an acquisition time of 1 s. We observed that with the introduction of gold nanoparticles the Raman spectra were clearly detectable under those conditions. For comparison, a spectrum obtained from a cell incubated in medium without gold colloid solution is also shown in Fig. 3 (lower). It can be seen that even though the spectra acquisition time was increased to 10 s it was still unable to obtain a good quality spectrum without gold nanoparticles present. This observation confirmed the significant enhancement of Raman scattering with the presence of gold nanoparticles which also enabled a fast spectra acquisition although the exact mechanism of the SERS enhancement of Raman signals is not fully understood [21]. The introduction of gold nanoparticles therefore significantly shortened signal acquisition times mainly owing to the increase in sensitivity which in turn allowed an array of spectra to be obtained in a relatively short time period. Fig. 3Raman spectra from cells following incubation in medium with gold colloid (upper spectrum) and without gold colloid (lower spectrum) Characterization of single living cells by 3D mapping Using the SERS technique described above it was possible to obtain spectra from a single living cell which represented a “fingerprint” from which various chemical constituents in the cell can be assigned. Combined with an automated microscope stage which offers spatial resolution (i.e. 1 μm), mapping of an entire cell area at a specific plane was carried out. It has been suggested from previous studies that most of the spectral bands associated with living cells occur in the range 800–1,700 cm−1 [11, 22, 23], hence this range was used in this study for mapping. Figure 4a shows an “image” of Raman spectra obtained from the middle layer across the single cell by mapping an area of 21 × 21 μm2 with a spatial step of 1 μm in both X and Y directions. The mapping provided information on the distribution of selected bands, as seen in Fig. 4b in the range from 1,290 to 1,370 cm−1, which represent most of the significant bands associated with DNA and proteins within a cell’s nucleus and cytoplasm [11, 23]. Fig. 4a Spectral mapping of a single CHO cell on an X–Y plane and b corresponding spectra from three positions in the area of nucleus, cytoplasm and membrane, respectively In general, the Raman spectra of single CHO cells showed contributions from all its cellular components including nucleic acids, proteins, lipids and carbohydrates. Table 1 summarises the band assignment for the Raman spectra taken from CHO cells based on the published data [8, 9, 12, 23–27]. Comparison of the spectra taken from different positions across the cell on an X–Y plane (Fig. 4) indicated that strong peaks from the nucleus spectrum corresponding to DNA sugar–phosphate backbone (895 and 1,142 cm−1), and bases G (1,320 and 1,487 cm−1), A (1,420 and 1,578 cm−1), T (1,176 and 1,376 cm−1) and C (1,420 cm−1) were noticeably reduced in the cytoplasm and membrane spectra (Table 1). This change was expected as the nucleus contains high densities of DNA, whilst the cytoplasm also had significant quantities of RNA contributing to the corresponding peaks. As expected, the spectrum taken from membrane area showed significant peaks corresponding to lipids (1,068 and 1,453 cm−1). Table 1Band assignment for Raman spectra of CHO cellsBands (cm−1)AssignmentsDNA/RNAProteinsLipids830Phosphodiester BkB [12, 23, 24]Tyr [12, 23, 24]895Phosphodiester BkB, deoxyribose [12, 23, 24]940ν (C–C), α-helix [9, 12, 23]1,004Phenylalanine [9, 12, 23]1,065ν (C–O) [12, 23]ν (C–C) chian [25]1,126ν (C–N) BkB [9, 12, 23]ν (C–C) chain [25]1,144Ribose–phosphate [12, 23]1,157Ribose–phosphate [20, 26]1,176T, C, G [9, 12, 23, 24]Phenylalanine [12, 23]1,230C [12, 23, 24]1,266Amide III [25]δ (C=CH2) [25]1,295δ (CH2) [25]1,320G [24]1,342A [12, 23, 24]1,376T, A, G [9, 12, 23, 24]1,448δ (CH) [9, 12, 23]δ (CH) [9, 12, 23]1,482A, G [12, 23, 24]1,566Hemoglobin [25]1,578A, G, purine [12, 23, 24]1,603A, C [24]1,643Amide I [27]Abbreviations: BkB DNA sugar–phosphate backbone, Tyr tyrosine, A adenine, T thymine, G guanine, C cytosine, ν stretching vibrations, δ deformation vibrations Using the confocal optics of the microscope system, the chemical concentration distribution in the Z axis at three levels in the nucleus was examined (Fig. 5) and indicated that the main peaks positions were generally identical but the peak heights were markedly different, indicating a concentration difference. From Fig. 5a significant a peak was observed at 1,320 cm−1 on the spectrum taken from Z = 3 μm (near to the bottom of the cell) which was assigned to DNA bases G [23], whilst the peak at 1,450 cm−1 from Z = 6 μm indicated a strong deformation from a C–H stretch in proteins [8, 12, 23]. Thus, by collecting spectra from different spatial positions, it provided an approach to build a 3D mapping of the distribution of chemicals within a single cell. Fig. 5Raman spectra taken at three levels along the Z direction in the nucleus area It should be noted that the spectra taken from different positions are distinct in terms of both Raman band amplitudes and band positions. This spectral variation can be attributed to the native chemical inhomogeneity within a cell. However, the possibility of non-uniform distribution of gold nanoparticles within the cell cannot be excluded [15]. In addition, this non-uniform distribution can also take place during the time-resolved monitoring because dynamic processes inside the cell can lead to local fluctuations of the particle densities and hence changes of the spectra. To further develop this spectroscopic technique for both qualitative and quantitative analysis, technologies need to be explored in order to deliver and position nanoparticles within cells in a controllable format for a uniform distribution. Nevertheless, it is still possible using this promising methodology to examine a specific point within a cell in a dynamic way with time, especially when applying stimulation, e.g. an agonist, under microfluidic control conditions. In situ monitoring of cellular chemical dynamics Finally, we monitored the chemical dynamics of the cell when exposed to the agonist ionomycin [3, 28]. After loading cells into the chip channel, a 40-min period of settlement was allocated before introducing the ionomycin solution at a constant flow rate of 2 μL min−1. Whilst focussed on a region close to the cell nucleus, a series of spectra were taken at an interval of 1 s in the range from 800 to 1,700 cm−1 (Fig. 6a). It can be seen from the spectra (Fig. 6b) that most of the peaks corresponding to nucleic acids, proteins and lipids remain visible with time but the peak heights vary noticeably, indicating the concentration change of these compounds. One of the most significant changes is the appearance of a peak at 1,643 cm−1 which can be assigned to amide I [27]. The time profile of this peak (Fig. 6c) reveals the concentration change of amide I within the cell, showing a similar trend to that of the Ca2+ flux evoked by the agonist ionomycin [3] which is commonly used in biomedical research to stimulate the intracellular production of proteins such as interferon [29]. Since amide I contributes spectrally to the Raman spectrum of interferon, it follows that the in situ monitoring of such molecules could in future be used to produce characterization of protein expression dynamics at subcellular levels. Clearly, more studies are required in order to understand the mechanism and cause of the concentration change of amide I within the cell when exposed to agonist ionomycin. Fig. 6In situ monitoring of cellular dynamics by recording a series of Raman spectra (a) from a specific area at an interval of 1 s. The first spectrum, at time 0, was taken when no agonist was introduced (b); (c) shows the time profile of the peak at 1,643 cm−1 corresponding to the change of amide I Conclusions The SERS technique has been used for characterisation of single CHO (Chinese hamster ovary) cells with a microfluidic device where gold nanoparticles were introduced into cells for Raman enhancement. The use of an inverted microscope optical systems in combination with a charge coupled device (CCD) detector allows the measurement of Raman spectra for simultaneous analysis of bio/chemical species within cells. In addition to a precision automated X–Y sample stage, the confocal optics provides discrimination between points of different depths within the cell, enabling chemical mapping in three dimensions. By using microfluidic methodology, the cell manipulation and reagent delivery were performed in a controllable manner. By introducing cells and test reagent into the chip in a continuous flow as a series of plugs in a given sequence, it enabled the in situ chemical characterization of single CHO cells with a high degree of spatial and temporal resolution. This allows the real time monitoring of the dynamics of the agonist-evoked Ca2+ flux response. The approach described has the potential to be used for the study of the spatial dynamics of a range of intercellular processes.	[ "microfluidics", "confocal microspectroscopy", "surface-enhanced raman scattering (sers)", "chinese hamster ovary (cho) cells", "dynamic monitoring" ]	[ "P", "P", "P", "R", "R" ]
J_Med_Internet_Res-6-2-1550592	Setting the Public Agenda for Online Health Search: A White Paper and Action Agenda	Background Searches for health information are among the most common reasons that consumers use the Internet. Both consumers and quality experts have raised concerns about the quality of information on the Web and the ability of consumers to find accurate information that meets their needs. Introduction Searches for health information are among the most common reasons that consumers use the Internet. The Pew Internet & American Life Project (Pew) reported in 2003 that 80% of Americans with Internet access have used the Web to get health or medical information [1]. The Internet has transformed the ability of consumers to find health information and to connect with other individuals with similar interests. The Internet has been recognized as an important source of health information by the federal government, which established a series of goals relating to access and quality of information on the Internet in the Healthy People 2010 action plan [2]. Health information on the Internet can dramatically improve consumers' health-care and lifestyle choices. However, increased access to Web-based information has also raised concerns about the quality of information consumers are using, and the impact of this information [3]. Disparities in access to information have also become apparent. These factors suggest the need to better understand how consumers find health information on the Web, how to evaluate the quality of information retrieved, and how to help consumers to critically evaluate and manage information. These factors suggest the need to better understand how consumers find health information on the Web, how they evaluate the quality of information retrieved, and how to they could be helped to critically evaluate and manage information. Research on health Web sites raised concerns about the quality of information on the Web [4]. A 2001 study by RAND for the California Healthcare Foundation showed that information on health Web sites is often incomplete or out of date [5]. This might be of little concern if consumers routinely consulted health-care professionals about the information. However, Pew found that 69% of consumers did not discuss the information they found with a doctor or nurse. Many people use search engines to find the information they use to help make personal health decisions. Search engines and the Internet have vastly improved access to health information for many consumers. However, search processes and results vary considerably among search engines, and are not transparent to consumers. The criteria used to identify and rank health-related Web sites vary among search engines, and often is not apparent to consumers. Search results may be affected by the structure of content on health Web sites, consumer search terminology, and the use of paid placements by the search engine. In short, research on health searches suggests that the process by which consumers locate health information on the Internet, and the evaluations they make regarding which Web sites to review are important variables in the quality of information they ultimately view and use. Improved understanding of factors influencing online searches will facilitate technical and educational approaches for maximizing quality and benefit of health searches. Methods In 2003, URAC and Consumer WebWatch (CWW), a project of Consumers Union, carried out a project funded by the Robert Wood Johnson Foundation to examine factors influencing the results of online health searches and to develop an agenda for future research and development that would improve the results of health searches. We reviewed published literature and industry reports, and convened two stakeholder groups consisting of consumers, quality experts, search engine experts, researchers, health-care providers, informatics specialists and others. Literature Review Method Our literature review was not exhaustive: its purpose was to provide a baseline understanding of consumer, Web site, quality measurement, and search engine factors that influence the results of searches for health information. We conducted a search of key terms in the Cumulative Index of Nursing and Allied Health Literature (CINAHL), Medline, PubMed, Expanded Academic ASAP, Lexis-Nexis, Proquest, Ingenta, and related databases in health care, information science, and computer science. The initial searches took place in early 2003, but citations were added as they were identified. Where initial searches revealed poor topic coverage, associated reference lists, books and other media that were considered to inform the topic were included. The following search terms were included: Web-based, Web site, information quality, Web search, consumer health, eHealth, health information, search engine, information retrieval, information seeking. We also examined bibliographies of articles retrieved by electronic searches and solicited recommendations from members of the project advisory committee. We discontinued searching in specific topic areas when project staff believed we had adequately described current understanding of key issue areas. Methods for Convening Stakeholder Meetings An open announcement about the project and recommendations from industry leaders helped identify interested stakeholders, and participants were selected by URAC and CWW with guidance from a project advisory committee. Not everyone invited was available to attend. We attempted to achieve a balance of different stakeholders at each meeting. Meetings were held in California and Washington, DC to facilitate participation. The purpose of each stakeholder meeting was to review existing knowledge about results of consumer searches for health information, and to develop recommendations for additional research, technical improvements, and educational approaches needed to improve the results of online consumer searches for health information. Participants reviewed the summary recommendations presented in this article after the meeting and had the opportunity to comment, but were not asked to vote on or endorse the recommendations. For the purposes of this project, we assumed that most searchers would prefer information that is accurate and reliable. These attributes are also components of effective health communication [2]. This was our working definition for quality Web sites. The perception of other elements that might be used to define quality, such as the site's reading level and comprehensiveness, will vary depending on the user and the user's needs at a given time. Results Results of Literature Review How Consumers Use the Internet to Locate Health Information An April 2003 report from the Pew Internet & American Life (Pew) report provided an overview of the US Internet consumer population [6]. The study found that Internet access has grown across-the-board, but that demographic gaps remain. A variety of factors continue to separate Internet users from non-users. Internet users tend to be younger and more affluent, and are more likely to be employed, white, well-educated, and to be suburban or urban residents. Pew noted that consumers often overestimate their knowledge of the Internet and their ability to locate information. A 2002 analysis by Houston et al using Pew data noted a need to educate patients about searching for health information online and for tools to help them identify high quality information [7]. They also found that chronically ill Internet users were often relatively new to the Internet, but noted that they were more likely than those in good health to discuss findings with their physicians. Consumer Search Strategies A 2002 Pew Internet & American Life Project poll found that the typical health-information seeker usually starts searching for medical information at a general search site, not a medical site. Eighty-one percent of online health seekers start at a search engine or use the search function of a general portal such as the Yahoo home page, MSN, or AOL. Consumers visit two to five sites during an average visit and typically spend at least thirty minutes on a search [8]. Several studies have investigated behaviors consumers exhibit in retrieving and health information on the Internet and in assessing its quality. Eysenbach and Köhler, examining Web searchers in Germany, found that although search technique was often suboptimal, Internet users found the health information they were looking for relatively quickly [9]. A search optimization firm, iProspect, reports that users generally use the same search engine for all types of search requests. Users look at up to three pages of search results to determine relevance, and abandon a search if they do not find appropriate results in the first three pages. Users usually modify their query after abandoning an initial search, and may at that point change search engines [10]. These findings illustrate the importance of search engines to the process of retrieving health information. They imply a business rationale for search engines to ensure that health searchers locate what they want, since they may otherwise lose search traffic. Comprehension, Literacy, and Access Issues Searches are heavily influenced by the search terms used, even when the terms used are considered to be synonyms. Use of lay terminology for a health subject can result in unrelated or misleading information [11]. Berland et al concluded that accessing health information using search engines and simple search terms was not efficient because Web sites are inconsistent in their provision of key information, and because high reading levels are required to comprehend Web-based health information [4]. Also, the relevance of information located was often of limited value, which may have been due to terminology used in the original search phrase. Non-English speakers face challenges finding and reviewing information on the Internet. One Internet accessibility study for people with disabilities found that there are significant access barriers. Governmental and educational health-information Web sites were more accessible than other categories, such as Web portals and community sites [12]. Physician Responses to Internet Information A study of physician views on online information found that physicians increasingly encounter patients who have conducted health searches. Use of the Internet by patients does appear to affect treatment processes: for example, many physicians reported having changed the treatment protocols they had initially planned as a result of consumer requests. Although most physicians believe the information their patients find is accurate, many believe that having to discuss this information with their patients decreases their efficiency and challenges their authority. Some are also concerned that the information may be inaccurate. The study concluded that quality of information on the Internet is critical, as it does influence both patient requests and physician treatment choices [13]. In an effort to steer patients to credible Web sites, some health-care organizations have begun to suggest ( "prescribe") credible Web sites to their patients in the course of their consultations [14]. Consumer Evaluation of Web Site Credibility Experts and consumers use different criteria for evaluating the quality of Web sites. Eysenbach found that consumers assessing the credibility of a Web site primarily looked for the source, a professional design, a scientific or official touch, language, and ease of use. Study participants never checked any "about us" sections of Web sites, disclaimers, or disclosure statements. Very few participants noticed and remembered which Web sites visited [9]. A Consumer WebWatch (CWW) study of consumers reported findings similar to Eysenbach's: once people get to a site, they do not use rigorous criteria to assess the site's credibility. For example, they almost never referred to a site's privacy policy. The average consumer paid far more attention to the superficial aspects of a site, such as visual cues, than to its content. Nearly half of all consumers in the CWW study assessed the credibility of sites based in part on the appeal of the overall visual design, including layout, typography, font size, and color schemes. In comparison, a parallel group of health and finance experts were far less concerned about the surface aspects of these industry-specific types of sites and more concerned about the breadth, depth, and quality of a site's information [15]. How Web Sites Influence Availability of Quality Health Information Techniques for Conveying Information about Web Site Content The structure of a Web site influences how information can be retrieved from the site by a search engine, as well as the usability of the site for consumers. Coding and structure of Web sites can facilitate retrieval by search engines or can pose a barrier to information retrieval. Coded information on a Web site is processed through the search engine algorithm, and determines whether and how the site is ranked in search returns. The same tags and codes that can be used to highlight information on a legitimate Web site may also be used by "spoofers" who try to lure traffic onto the site. In general, Web sites can support retrieval of information on their pages by using metadata, metatags and keywords to guide search crawlers to important content. These codes provide a means for relaying information directly to the search engine. Keywords are recognized indicators of specific services or products that can be used to increase specificity of searches and help Web sites attract "qualified" traffic. One strategy for enhancing search rankings of quality Web sites is to code certain types of information for consistent retrieval by the search engine. Efforts are under way to implement metadata codes to support a "semantic Web." The semantic Web uses code to establish relationships between words to enable search engines to effectively understand intent, rather than simply identifying the presence of a search term [16]. Quality Indicators for Web Site Content Eysenbach et al found wide-ranging differences in studies of the quality of health Web sites. There are significant variations in study methods and rigor, quality criteria, study population, and topic chosen. Operational definitions of quality are often inconsistent. As a result, the conclusions on quality of health-related Web sites vary widely. Eysenbach found that the most frequently used quality criteria include accuracy, completeness, readability, design, disclosures, and references provided [16]. Griffiths and Christensen evaluated the quality of Web-based information on treatment of depression to identify potential indicators of content quality, and to establish whether accountability criteria are indicators of quality [17]. They found that although the sites examined contained useful information, their overall quality was poor. Sites typically did not cite scientific evidence in support of their conclusions. Researchers have also studied the correlation between popularity of Web sites and quality of content. Meric et al found that more popular breast cancer-related Web sites were more likely than less popular ones to contain information on ongoing clinical trials, results of trials, and opportunities for psychosocial adjustment. These characteristics were also associated with a higher number of links. More popular sites were more likely to provide updates on other breast cancer research, information on legislation and advocacy, and a message board service. Measures of quality such as display of authorship, attribution or references, currency of information, and disclosure did not differ between popular and less popular sites [18]. In similar findings, Kunst et al found that while there is a correlation between credibility features and accuracy of information, the association is relatively weak [19]. These findings suggest that additional research is needed to identify indicators of content quality, and to correlate consumer preferences to quality indicators. Sites that include content correlated with popularity may best meet the public's desire for health information. Current search algorithms may not be in agreement with quality clinical indicators and performance measures currently used throughout the health-care industry. Codes of Conduct A wide range of tools has been developed to assist site developers to produce good quality sites and for consumers to assess the quality of sites. Adherence to accepted codes could theoretically be used as a factor in searches. Ratings instruments include codes of conduct, quality labels, user guides, filters, and third party certification. The value of these tools is unclear: studies have demonstrated that consumers do not routinely seek out information on certifications or adherence to voluntary codes. However, it is assumed by many that such codes benefit consumers indirectly by influencing Web site behaviors and practices. For example, most standards require sites to implement privacy protections and disclosure of site information as consumer protections. No research has been done on the effect of compliance to a code of conduct on Web sites. A number of organizations have developed quality criteria for health-related Web sites, some with verification and some completely voluntary. Voluntary, self-certifying standards have been developed by the eHealth Code of Ethics of the Internet Health Coalition [20], the American Medical Association [21] and the Health On the Net (HON) Foundation [22]. URAC has developed a health Web site accreditation program that involves independent verification of compliance with its standards. URAC accreditation includes review of the Web site by an external auditor [23]. Web Site Rating Instruments Web site ratings could be potentially be used to inform searchers and search engines as well, if ratings could be clearly correlated to quality. Two common approaches to rating Web sites include expert ratings, and user (consumer) ratings. Gagliardi and Jadad conducted two evaluations of Web site rating tools, published in 1998 and 2002 respectively [24,25]. They concluded that ratings instruments tend to proliferate and disappear, and that few have been validated for direct correlation between standards and quality. Few provide details on how they were developed, or provide instructions for use, or information about the inter-observer reliability and construct validity of the measurements. Kim et al reviewed published criteria for evaluating health-related information on the Web, and identified areas of criteria-based consensus [26]. They identified 29 published rating tools and journal articles that had explicit criteria for assessing health-related Web sites. The most frequently cited criteria were those dealing with content, design and aesthetics of site, disclosure of authors, sponsors or developers, currency of information (includes frequency of update, freshness, maintenance of site), authority of source, ease of use, and accessibility and availability. A number of tools are available to guide users in evaluating information on the Web. Interactive user guidance systems can be used to assess characteristics of Web sites. Tools such as DISCERN and QUICK allow Web site users to assess Web site credibility by responding to a series of questions [27]. Other organizations such as the National Library of Medicine, which operates MEDLINEplus, and the Medical Library Association, have developed guidelines and tips for consumers to evaluate health Web site content [28]. The European Union sponsored a collaborative project called MedCERTAIN to develop a rating system to enable consumers and professionals to rate quality information on the Web. The MedCERTAIN project evolved in a project called MedCIRCLE, which has developed a metadata coding language to mark quality indicators on health Web sites [29]. Discussion Search Engines and Mediators of Health Information Electronic and Human Mediation Search engines serve an essential function in enabling users to find relevant information on the Internet. Recognizing the challenges of sorting the enormous amount of information on the Internet, many organizations are augmenting or mediating the results of electronic searches. Mediation can be either electronic or human-augmented techniques for reviewing information and making a pre-selected set of information available to consumers. One challenge to search engines and human mediators is making access to personalized information as effortless as possible, as consumers rarely use even the advanced search features currently available to them [30]. How Search Engines Work Search engines and Web directories play a central role in facilitating access to health information. Web directories are organized Web site listings put together by human reviewers. Search engine listings are put together by automated systems and lack a navigable structure. Directories usually concentrate on indexing Web sites, while search engines typically index individual Web pages. Consumer searches for keywords will result in a valid match only if the keyword appears in the Web site's description. Hybrid models of search engines and directories are common. Search Engine Indexing and Retrieval Methods Virtually all commercial search engines rely on large powerful databases that utilize automated search agents called robots ("bots"), crawlers, or spiders. Search agents crawl the Web continuously to index information on Web sites. Crawlers capture metadata, page titles and textual content, and add them to the search engine's index or main database. The search engine's algorithm compares indexed data to the user term to process a search. Search engine algorithms are quite complex and scientific. They make frequent use of complementary directories aimed at optimizing and positioning Web sites in the right categories. Search algorithms are closely guarded as proprietary corporate information [31]. Current metrics for evaluating search engines include initial page retrieval capacity and the ability to revisit Web sites to update information. Currency of information, as demonstrated by elimination of non-working links to Web sites is also a performance metric. These criteria are features of business performance, not necessarily the content relevance or quality of the sites returned by a search. Content and format of Web sites determine how they are indexed by search engines. Some search engines use keyword location, frequency, phrasing, and density as indexing and ranking factors. Type and number of links associated with a Web site are common indexing factors. Web sites also use tags to identify certain types of information. Search engine databases include only Web sites that have been registered with or indexed by the search engine-hence the importance of Web site developers making their sites accessible to automated agents, or becoming known to directory developers. Ranking and Ratings Ranking of sites in the final display of search results is of great importance to Web sites, users, and search engines. Ranking effectively drives the likelihood of particular sites being recognized and visited because, as noted, consumers rarely look at more than three pages of results. A poorly designed or executed search may produce an unwieldy list of Web site results that is difficult to navigate. Alternatively, searches that are too narrowly drawn may omit important sites. Paid preference and placement by search engines also affects which sites are retrieved in a given search [32]. A study by CWW demonstrated that consumers experience considerable confusion about paid listings, and may not distinguish them from other returned listings [33]. The Federal Trade Commission has also expressed concern about how paid placement is disclosed to consumers, and has warned search engines to clearly distinguish advertising from search returns. Search engines may operate their own paid placement programs or obtain search results from third parties, who in turn operate paid placement programs. Mediated Searches Mediated searches may be as simple as having a librarian assist with a search, or they may be based on much more complex algorithms. Participants in the URAC/CWW stakeholder group noted that medical and general librarians play an important role in helping large segments of the population retrieve online information and learn effective search strategies. More complex mediated search strategies employ both human mediation and electronic queries to interface with users and focus a search. Many search engines offer filters that allow users to exclude unwanted search results, most typically pornographic sites. Users, including libraries, can also install blocking software to prevent unauthorized use. However, this electronic mediation may unintentionally block desired health information and create an access barrier. For example, because pornography-blocking software and filters cannot perfectly distinguish between pornographic and non-pornographic Web sites, such products may block access to legitimate health-information sites, particularly those related to sexuality [34]. Gateways employ filters, either electronic or human, to accept or reject types of sites of information based on preset criteria. Gateways are used to organize information on the Internet through selection of resources based on quality and relevance of information to a particular audience. Internet resources are reviewed, classified, and stored with descriptive information. In the US, healthfinder.gov®, is a widely used gateway to selected consumer health and human services information resources provided by US government agencies and other organizations serving the public interest [35]. Participants in the stakeholder meetings noted that domain name extensions such as .com or .org could be used as a distinguishing feature of Web sites for the purpose of focusing search efforts. The World Health Organization is considering the feasibility of requesting a "dot health (.health)" extension for a pre-selected set of trusted Web sites [36]. In informal proposals describing the .health domain name, the extension is proposed for health information, services and organizations under a framework promoting minimum standards of conduct. Oversight of Web sites would be delegated to independent verifying organizations. The advantage to sites for adhering to standards of content quality would be more ready identification of sites by search engines as a result of the .health domain name. Stakeholder Discussion of Literature Review Research Needs to Address Consumer Evaluation of Web Quality There is great variation in how consumers seek information via the Internet, and in how successful they are in searching for health information. Since there is significant consumer-level variation in how consumers search for health information, search algorithms that support variation and still return expected results will meet consumer needs most effectively. Additional research is needed on information needs of different consumer segments and how to effectively educate differing consumer segments to improve the results of their health searches. Research is needed on how to efficiently validate the quality of Web sites and communicate this information to consumers. Research Needs for Web Site Quality Indicators There is a need for tools to enhance recognition of quality Web sites by consumers and search engines. Such tools may be implemented by Web sites themselves, for example through increasingly sophisticated coding to highlight quality indicators. The MedCERTAIN project has been created precisely to address this issue, and has developed the HIDDEL vocabulary to mark features of Web sites [29]. Technical tools can be used to direct consumers more effectively to relevant, high quality information. In addition, since there are currently multiple tools for either self-evaluation or third party evaluation of Web sites, future research should be undertaken to validate these tools. As noted, gateways filter information to increase its relevance to consumers and provide expert assessment regarding validity of sources is available. It may also be useful to develop more sophisticated search models for providing useful and relevant information to consumers via customization approaches. Such approaches could potentially be embedded in search algorithms. In addition, more research is needed on the impact of Internet-based health information on outcomes. The benefits and risks of health information, both from a health outcome and a system outcome (quality, cost), are poorly understood and should be examined further. Research Needs for Search Factors Influencing Search Results Search engines are increasingly important as a tool for locating and organizing information from the vast Internet resource. The volume of information on the Web is so significant that consumers may need different types of mediators, such as search engines or librarians, to help manage the volume of information. Human assistance is also helpful to counteract electronic spoofing and to help consumers overcome limitations in their search strategies. To effectively improve health searches, more information is needed about search algorithms and how quality factors are identified in the algorithms. Search engines are also developing technology to search for synonyms, which may enhance health searches conducted by laypersons. It may also be helpful for search engines to develop methods to distinguish health related searches from other types of searches, rather than using a simple word match. Search technology to intuit consumer needs more effectively and learn from repeated searches could help search engines steer consumers to quality results. New technologies may ultimately be more effective than electronic filtering, requiring consumers to apply filters, or modifying their search strategies. With technology advances, search engines may be able to identify quality proxies that could improve page rankings of high-quality Web sites. Search engines could, for example, give higher ranking to "official sites" for diseases. They could also piggyback onto credibility assessments provided by groups such as healthfinder.gov, or give higher ranking to sites listed in directories from trusted independent sources. Ultimately, adoption of technological solutions depends on the ability of researchers to understand the relationship between electronic proxies for quality and actual quality of content. Discussion of Stakeholder Recommendations for Next Steps The URAC and CWW expert panels discussed consumer, Web site, and search engine factors that influence the outcomes of health searches. In the course of discussion, they developed a number of recommendations for future research and development (Textbox 1). Their recommendations fell into several categories: needs for health services research, consumer and provider education, technological improvements, and development of tools and information to improve the results of health searches. For some recommendations, the evidence base for implementation is strong; for others, not. Implementation of some recommendations will be enhanced by creation of a national research agenda for health information and targeted funding to study and improve consumers' ability to locate and retrieve quality health information on the Internet. Other recommendations could be embraced at any time by researchers, educators or technology organizations as a business need becomes increasingly evident. Recommendations of the Group Leadership for Health Search Improvement Organizations concerned about the quality and accessibility of health information online should continue to collaborate to promote "health search literacy."Collaborators should convene a leadership summit on health search literacy to discuss feasibility and implementation of many of the recommendations herein.Collaborating organizations shouldwork with funding organizations to develop a comprehensive long-term research agenda to improve health searches and increase access to quality health information;develop enhanced research methodologies to evaluate the quality, impact, and effectiveness of online health information. Consumer-directed Tools Create tools to support consumer health-information needs, including preset, prescreened health bookmarks and more guidance on how to reach health gateways and portals containing trusted health content.Develop and circulate a public domain brochure on health search strategies that could be branded and distributed by physicians, employers, health plans, and others to educate consumers.Develop public domain interactive, validated search strategy content pages that could be branded and used by health Web sites. Research Needs Identify the search needs and capabilities of diverse populations of searchers, including culturally diverse users and searchers with health needs of differing intensity and severity.Develop more understanding about how consumers interpret online health information, assess its credibility, and make health-related decisions.Research the relationship between consumer search strategies and consumer expectations for results to determine effective approaches for conveying information on the Internet.Research factors affecting physician assessments of Web-based information and how quality content affects physician recommendations to patients about online health-information resources.Assess the relationship between expert accreditation, quality seals, ratings and content quality, as well as the impact of such endorsements on both consumer behavior and Web site behavior.Research the correlation between Web site traffic volume and consumer satisfaction, particularly for health Web sites where there is variation in dimensions of quality such as accuracy, comprehensiveness, ease of navigation, and reading level.Evaluate content quality of Web sites in different domains, (eg, .gov, .edu, .com, and .org) to identify similarities and differences related to quality within and across categories of Internet domain namesEvaluate the impact of Internet-based health information on health outcomes: utilization, behavior change, knowledge, burden of illness and disease, or other measures.Research the relative effect of each component of a search algorithm (word frequency and placement, links, etc) for finding health information.Validate elements of some search algorithms, such as link frequency, as indicators of value/quality.Conduct periodic studies to monitor changes in accuracy and quality of content over time, including updating findings from the California HealthCare Foundation /RAND study [5]. Education Agenda Develop models for offering health search education at teachable moments and in diverse consumer settings.Promote dissemination of existing educational tools and resources to assist consumers in evaluating health information on the Web more effectively.Develop user-appropriate tools and approaches to assist Internet users with special needs. High priority user groups may include disability, low literacy, and non-English speaking groups.Urge provider organizations to educate provider members on the value of offering Internet information and interactive learning recommendations as part of the therapeutic intervention.Educate health Web site developers on how to make information easy to find and how to meet the content-level of their intended users.Urge education organizations, in collaboration with health organizations, to develop a school-based or publicly available health search curriculum. Technology Improvement Agenda Continue to develop interactive features on search engines and sites to customize and personalize health searches.Develop more functionality for search engines to enhance selected health queries by offering additional relevant information.Develop technological markers or indicators that could be uniformly applied by Web site developers to indicate accuracy and comprehensiveness of health Web sites.Develop codes to indicate when information on a Web site supercedes previous information.Develop collaborations between health quality and search engines experts to develop codes for validated quality proxies.Develop search technology similar to that used in the commercial sector to direct consumers to related, relevant information based on both searching and viewing behaviors.Enhance personalized searches by building search engine capability to "learn" from repeated searches and user behavior. Expanding the Market for Quality Develop a health equivalent of "BizRate" or "eBay" surveys that can be used by consumers to evaluate Web sites after viewing. Existing models for such a survey could be adapted and disseminated.Sponsor a competition for individuals or organizations to design a search algorithm that returns the most credible health results as evaluated by experts. Design a separate contest for the most effective business plan to make the business case for building quality factors into health searches. Conclusion The Internet has opened a vast library of information to consumers of health information and made that information more accessible than ever before. This represents a significant step forward for consumers. However, the volume of information and the variable quality of information has created new interpretive challenges. Now, one great challenge is helping consumers find the information they want that is also accurate, reliable, and presented in an accessible format. Searches for health information rely on a complex interplay of search algorithms, Web site content and coding, and consumer behaviors. The recommendations presented here address each of those factors with ideas for further research as well as more immediate recommendations for action. This agenda is a start at maximizing the potential of the Internet to deliver high-quality health information for diverse users.	[ "internet", "ehealth", "health services research", "information management", "quality of health care", "consumer participation", "patient education" ]	[ "P", "P", "P", "R", "R", "R", "R" ]
Neuropsychologia-2-1-2265793	Free recall in autism spectrum disorder: The role of relational and item-specific encoding	Autism spectrum disorders (ASDs) are characterised by a relatively specific pattern of typical and atypical memory functioning. Convergent behavioural and neuroscientific evidence indicates that this pattern of functioning may be the result of specific impairments in hippocampally mediated relational memory processes, whilst brain-mechanisms mediating item-specific memory processes remain intact. In the current paper we draw on a behavioural paradigm developed by Hunt and Seta [Hunt, R. R., & Seta, C. E. (1984). Category size effects in recall—The roles of relational and individual item information. Journal of Experimental Psychology: Learning, Memory and Cognition, 10, 454–464], which not only allowed us to determine whether individuals with ASD did indeed experience selective difficulties in relational processes, but in addition enabled us to gain insights into the severity of this impairment. Our results suggest that whilst individuals with ASD employ relational memory processes atypically, this impairment seems restricted to situations in which such processes need to be deployed spontaneously to facilitate memory. Under situations that provide environmental support for the processing of relational information, individuals with ASD did demonstrate the ability to employ such processes relatively effectively. These findings provide further support for the ‘Task Support Hypothesis’ and suggest that relational memory processes may in principle be functionally intact despite not being triggered by the same environmental situations as in typical development. 1 Introduction Autism spectrum disorders (ASDs) are clinically defined by difficulties in reciprocal social behaviour and communication and the presence of stereotyped patterns of behaviour and restricted interests (ICD-10: World Health Organisation, 1992; DSM IV-TR: American Psychiatric Association, 2000). In addition to this unique combination of symptomatology, the condition is also characterised by a relatively specific combination of typical and atypical functioning within the domain of memory. Since this patterning of memory functioning cannot be accounted for by the varying degree of language or general intellectual disability that often accompanies the core clinical features of ASD, it is thought to reflect a facet of the broader phenotype characterising the disorder. We propose that a cognitive framework that distinguishes between item-specific and relational memory processes may not only provide a suitable explanation for available behavioural evidence, an idea that we test in the present study, but may also prove useful in guiding future neuroscientific work relating to medial temporal lobe (MTL) functioning in ASD. On the basis of currently available evidence the patterning of memory functioning in ASD may be summarised as follows. Procedures such as recognition, priming and cued recall generally tend to yield typical levels of performance in ASD (Boucher & Lewis, 1989; Bowler, Gardiner, & Grice, 2000a; Bowler, Matthews, & Gardiner, 1997; Gardiner, Bowler, & Grice, 2003; Minshew, Goldstein, Muenz, & Payton, 1992; Minshew, Goldstein, Taylor, & Siegel, 1994; Tager-Flusberg, 1991). By contrast, free recall paradigms generally lead to diminished performance in this population especially when semantic, syntactic or phonological information is available to aid recall (Bowler, Gardiner, Grice, & Saavalainen, 2000b; Hermelin & O’Connor, 1967; Smith, Gardiner, & Bowler, 2007; Tager-Flusberg, 1991; but see López & Leekam, 2003 for contrary evidence). These free recall difficulties parallel findings from the typical aging literature (e.g. Craik & Anderson, 1999; Craik, Morris, Morris, & Loewen, 1990) and led Bowler et al. (1997) to posit a ‘Task Support Hypothesis’ according to which procedures that provide cues to the remembered material at test attenuate the memory difficulties experienced by individuals with ASD. Bowler, Gardiner, and Berthollier (2004) demonstrated that this framework could account for conflicting results regarding source memory capacities in individuals with ASD where previous studies had observed impaired performance on tests of source recall but undiminished performance on tests of source recognition (e.g. Bennetto, Pennington, & Rogers, 1996; Farrant, Blades, & Boucher, 1998). Although the task support hypothesis can account for the patterning of performance by individuals with ASD across a variety of memory paradigms, the causes for this greater reliance on support for the retrieval of previous experiences remain to date unknown. Earlier attempts to account for the pattern of intact and impaired memory processes in ASD have often invoked encoding as the source of difficulty. The most influential of these accounts is based on the seminal work of Hermelin and O’Connor (1970) who demonstrated that compared to non-ASD children who demonstrate superior recall for semantically and syntactically organised word sequences, children with autism do not tend to draw on such semantic and syntactic features to aid recall. On the basis of this evidence Hermelin and O’Connor (1970) argued that individuals with ASD do not encode stimuli meaningfully. Although several investigations have supported this hypothesis (e.g. Bowler et al., 2000b; Tager-Flusberg, 1991), three strands of more recent evidence indicates that the encoding difficulties seen in ASD may be more subtle than general problems with processing semantic information per se. We will briefly consider each of these in turn. First, individuals with ASD have been found to be subject to semantically induced memory illusions when Roediger and McDermott's (1995) procedure is used. In such paradigms individuals are asked to try to remember a series of words that include the strongest semantic associates of one non-presented target word (e.g. bed, dream, night, etc. for the target word ‘sleep’). Bowler et al. (2000b) and Beversdorf, Smith, Crucian, Anderson, and Keillor (2000) showed that individuals with ASD like typical individuals are more likely to falsely remember the semantically related target words than semantically unrelated words. Although the findings by Beversdorf and colleagues suggested that individuals with ASD may be better at discriminating the illusory target words from actually studied items, the finding that individuals with ASD did experience illusory memories shows that they are sensitive to the semantic associations of the studied words at least to some extent. The second strand of evidence concerns the observation that individuals with ASD exhibit relatively typical levels of performance following deep levels of encoding (Bowler et al., 1997; Gardiner et al., 2003; Mottron, Morasse, & Belleville, 2001; Toichi & Kamio, 2002). Deep levels of encoding generally involve the processing of semantic aspects of material (e.g. thinking about category membership of words), which typically leads to enhanced memory in comparison to shallower levels of encoding that involve the processing of non-semantic features of material (e.g. counting the number of syllables of words) (Craik & Lockhart, 1972). The finding of typical levels of performance following deep levels of encoding in ASD thus again suggests that under some circumstances such individuals encode semantic aspects of stimuli relatively effectively. Interestingly, studies employing levels of processing paradigms have also tended to note superior performance of individuals with ASD following shallow levels of encoding (e.g. Toichi & Kamio, 2002). This pattern has led Mottron et al. (2001) to suggest that rather than being deficient in processing semantic or ‘higher-level’ conceptual information, individuals with ASD may be superior at processing ‘low-level’ perceptual information and that this processing style may interfere with higher-level conceptual processes in some circumstances. We will return to this argument again later. The third and final strand of evidence regards a recent set of studies from our laboratory. In this set of experiments we asked participants to study a list of words, each of which was accompanied by a semantically related or semantically unrelated context word (e.g. ‘Wood’ in the context of ‘Tree’ vs. ‘Stone’ in the context of ‘Motor’). Whilst individuals with ASD failed to benefit from the semantic relatedness of to-be-remembered words and simultaneously presented context items on a test of free recall, their performance on a test of recognition was enhanced by such semantic relationships to a similar extent as found in typically developed comparison participants (Bowler, Gaigg, & Gardiner, 2008b). Again this finding suggests that individuals with ASD are sensitive to semantic features of stimuli, at least when test procedures support retrieval. The apparent contradiction between diminished use of semantic relations to aid free recall and relatively typical use of semantic features of stimuli under certain circumstances may be resolved by means of a closer analysis of what each of the paradigms described above requires of the participant. Experiencing an illusory memory on the basis of studying strong associates of a non-studied word implicates the relation between each studied word and the participant's existing knowledge base (e.g. Item A is associated with Concept X, Item B is associated with Concept X, etc.) and does not rely heavily on processing the relations amongst the studied items (Roediger, Watson, McDermott, & Gallo, 2001). Deeper levels of encoding equally do not necessitate relating studied items to one another but rather require enhanced attention to the semantic properties of each studied item. Finally, performance on tests of recognition has been found to rely more heavily on the ability to draw on information specific to individual items, including their semantic properties, rather than relationships among items (e.g. Anderson & Bower, 1972). In contrast, making efficient use of semantic features of stimuli during free recall tasks relies not only on the ability to process the semantic properties of each item but in addition on the ability to make use of these semantic features to establish associations amongst the items (i.e. Item A is associated to Item B because they are both associated with Concept X). The foregoing analysis leads us to speculate that individuals with ASD may have specific difficulties in using semantic information that emerge as a result of the relationships between items, whilst their capacity to draw on semantic information that is specific to individual items appears to be intact. This distinction between relational and item-specific processing has been widely applied to account for a variety of memory phenomena (Anderson & Bower, 1972; Hunt & McDaniel, 1993) and specific difficulties in processing relational information would explain why individuals with ASD rely on greater task support during retrieval. Such difficulties would also explain why individuals with ASD experience fewer episodically defined recollective experiences but somewhat more familiarity based experiences on tests of recognition that employ the ‘Remember/Know’ procedure (Bowler et al., 2000a; Bowler, Gardiner, & Gaigg, 2007; see Gardiner, 2001 for further details on the ‘Remember/Know’ procedure). Recollective experiences require that information be encoded and stored in relation to spatial and temporal contextual information whilst familiarity based recognition judgments can be mediated on the basis of available item-specific information alone (see Gardiner, 2001; Tulving, 1985, 2002 for further details). Furthermore inefficient use of organisational strategies such as semantic clustering (e.g. Minshew & Goldstein, 2001) or subjective organisation (Bowler, Gaigg, & Gardiner, 2008a) to facilitate memory in ASD also indicate that this population experiences difficulties in using relationships amongst items to organise their retrieval in free recall. The suggestion that ASD may be characterised by relatively specific difficulties in relational memory processes has recently also emerged on the basis of neuroscientific evidence (Nicolson, DeVito, Vidal, Sui, & Hayashi, 2006). Since the first direct examinations of the brains of individuals with ASD (Bauman & Kemper, 1985), atypicalities in areas associated with memory processes have repeatedly been documented (see Bachevalier, 1994; Kemper & Bauman, 1998; Palmen, van Engeland, Hof, & Schmitz, 2004 for reviews). Although the findings remain somewhat inconsistent, morphological abnormalities of the hippocampus are relatively well documented in ASD (see Nicolson et al., 2006). Areas surrounding the hippocampus, such as perirhinal, entorhinal and parahippocampal areas have less often been the focus of investigation but the observations by Bauman and Kemper (1985) suggest that at least the entorhinal cortex seems to be less affected than the hippocampus in ASD individuals.1 Until recently it has been difficult to relate these pathological findings to the memory difficulties experienced by individuals with ASD because the precise role of distinct medial temporal lobe areas in mediating memory processes was only vaguely understood. Accumulating evidence, however, now demonstrates that relational and item-specific processes are mediated by distinct sub-systems of the medial temporal lobe (MTL) memory system. More specifically, the hippocampus has been identified as the site of domain-general relational memory processes where individual features of an episode are integrated and organised (e.g. Eichenbaum, 2004; Holdstock, Mayes, Gong, Roberts, & Kapur, 2005; Squire, 1992). Areas outside the hippocampus, such as perirhinal, entorhinal and parahippocampal areas, on the other hand, seem to mediate more domain-specific item and contextual processes (e.g. see Davachi, 2006; Mayes, Montaldi, & Migo, 2007 for comprehensive reviews). Of particular interest in relation to ASD is the finding that episodically based recognition judgements that involve the recollection of contextual information (and are impaired in ASD) are primarily mediated by hippocampal processes whilst familiarity based recognition judgements (which are intact in ASD) are mediated by perirhinal processes (Brown & Aggleton, 2001; Davachi, Mitchell, & Wagner, 2003; Davachi & Wagner, 2002; Holdstock et al., 2005). This dissociation, together with the wider memory and neuropathological literature in ASD suggests that the item-specific/relational distinction may provide a useful heuristic device to guide further neuroscientific investigations of MTL pathology in ASD. For such an endeavour to be successful, however, it is necessary to test whether this framework provides an adequate explanation for the behavioural manifestations of memory difficulties in ASD. In the current paper we test the hypothesis that individuals with ASD are characterised by specific behavioural difficulties in relational memory processes whilst item-specific memory processes are spared. The current paradigm is based on a study by Hunt and Seta (1984), who argued that the efficiency of recalling items from a list of categorised words depended on the availability of both item-specific and relational information. Item-specific information, they suggest, is important in order to effectively distinguish amongst items from within a given category whilst relational information is important in order to recall the category per se. In order to test this hypothesis, they asked participants to study a list of words that included varying instances of items belonging to different categories (e.g. 2 Items of Fruit, 4 Professions, 8 Countries, 12 Animals, 16 Furniture). Hunt and Seta (1984) argued that because the relational nature of the items from the relatively small categories in such a list is relatively unobvious, effective recall of these categories depends disproportionately on the availability of relational information. By contrast, effective recall of items from the relatively large categories depends disproportionately on the availability of item-specific information because such information facilitates the differentiation of items within these categories. In support of their hypotheses, they showed that participants who encoded words through a relational orienting task (i.e. sorting words into categories) recalled items from the less obvious categories that were represented relatively infrequently in the study list significantly better than participants who encoded the words through an item-specific orienting task (i.e. rating words on pleasantness). In addition, the relational orienting task facilitated the recall of at least one item from each of the categories (particularly the relatively small categories) supporting the view that relational information is important for the recall of the categories per se. By contrast, participants who encoded words through the item-specific orienting task exhibited superior recall of items from the categories that were represented relatively frequently in the study list. In short, whilst the encoding of relational information disproportionately benefits the recall of words from relatively small (relative to other categories in the list) and therefore not very obvious categories, the encoding of item-specific information is disproportionately beneficial for the recall of items from relatively large categories. In the current experiment we asked participants to study two lists of words that, following Hunt and Seta (1984), consisted of varying instances of members from different categories. For the first list, individuals were simply asked to try to remember as many words as possible for an upcoming free recall test. Following this baseline condition, participants studied a second list whilst carrying out either the item-specific or relational encoding tasks employed by Hunt and Seta (1984). On the basis of evidence showing that typical individuals consistently benefit from semantic and categorical relationships to facilitate recall (e.g. Bousfield, 1953; Bower, Clark, Lesgold, & Winzenz, 1969) we predicted that during the baseline condition, the typical group would tend to rely on relational memory processes. In contrast, and on the basis of the evidence outlined above, we hypothesised that ASD individuals would rely more heavily on item-specific memory processes. Since relational information is particularly important for effectively recalling relatively small categories, we therefore predicted that during the baseline condition, the ASD group would exhibit disproportionate recall difficulties for items from relatively small categories whilst their recall of items from relatively large categories would not be as seriously compromised. In relation to performance following the relational orienting task, our prediction is less specific. If relational memory processes in ASD are impaired to such an extent that they cannot be deployed even when environmental support would facilitate such processes, the disproportionate recall impairment for smaller categories would persist. If, on the other hand, the impairment in relational memory processes is restricted to circumstances in which such processes would need to be deployed spontaneously, the task support hypothesis would predict that a relational orienting task would alleviate the recall difficulties in ASD thereby resulting in a relatively typical level of performance across category sizes. Finally, based on the evidence that individuals with ASD employ item-specific memory processes effectively, we predicted no recall impairment of this group following the item-specific orienting task. To the contrary individuals with ASD may outperform typical individuals in this condition because they may have developed superior skills in item-specific processing in order to compensate for their difficulties in relational processes. The finding that individuals with ASD tend to outperform typical individuals following shallow encoding tasks (e.g. Mottron et al., 2001) would be in line with this suggestion. 2 Method 2.1 Participants Twenty individuals with autism spectrum disorder (7 female, 13 male) and 20 typical individuals (7 female, 13 male) took part in this experiment. Participants were individually matched on Verbal IQ as measured by the WAIS-R or WAIS-IIIUK (The Psychological Corporation, 2002) and groups did not differ on Performance IQ, Full scale IQ or age. Ten participants from each group were randomly allocated to each of the two orienting task conditions (described below) with the constraint that IQ scores and age were similarly distributed across the two conditions. Table 1 summarises these data. All individuals with ASD were diagnosed by local health authorities and/or experienced clinicians, and met DSM-IV-TR (American Psychiatric Association, 2000) criteria for Asperger's disorder or Autistic disorder. The Comparison group was recruited via local newspaper advertisements. Brief interviews ensured that no participant had a history of neurological or psychiatric illness. Individuals gave their informed consent to take part in the study and were paid standard University fees for their participation. All but four individuals with ASD (two from each orienting task condition), who had been prescribed low doses of antidepressant medication, were free of psychotropic medication. Since the exclusion of these participants and their matched typical individuals did not alter the results significantly, all participants were included in the analysis. 2.2 Design and materials On the basis of Hunt and Seta's (1984) first experiment study lists were constructed from a master pool of words that consisted of 16 words from each of 10 categories selected from the Battig and Montague (1969) category norms. The frequency of the words ranged from 1 to 25 and the average category rank of words was 10 (see Battig & Montague, 1969 for details). The categories of Sports, Clothing, Weapons, Countries and Animals served as set A and the categories of Birds, Kitchen Utensils, Parts of the Body, Fruits and Vehicles served as set B. From each set, 5 study lists were constructed consisting of a total of 42 target items and 8 buffer items to counter primacy and recency effects (4 at the beginning and 4 at the end). Within each list relative category size was manipulated by selecting 2, 4, 8, 12 or all 16 items from the 5 categories (e.g. 2 Sports, 4 Clothing, 8 Weapons, 12 Countries and all 16 Animals). Across the 5 lists each category appeared at each category size once. Buffer items were selected from the categories of Professions and Parts of a Building for set A and Earth Formations and Alcoholic Drinks for set B. Words were printed in bold, arial font (size 36; Microsoft Word for Windows) in the centre of 8.2 cm × 7.6 cm, laminated cards. The 42 cards constituting the target items were ordered pseudo randomly with the constraint that the average lag between items from the same category be as close to 2 as possible (ranging from 0 to 5).2 The buffer items in the beginning and end of the target list were also randomised so that no more than two consecutive words were from the same category. The orders of items in the 5 lists from sets A and B were equivalent in terms of the list position of words from the differently sized categories. 2.3 Procedure Unlike Hunt and Seta (1984) the current experiment included a baseline condition during which participants were presented with one of the study lists (in the form of a deck of cards) from either set A or B and simply asked to try to remember as many words as possible. Participants were allowed to go through the cards at their own pace and the total amount of time they required to do so was recorded. Participants were instructed to put each card face down in front of them after they had tried to remember it and to not look at a card again once it was placed on the table. Immediately after the last word, oral free recall was requested. Following a 5–10-min break, individuals were given the respective deck of cards from the set not used during baseline, and given instructions for either the item-specific or the relational orienting task employed by Hunt and Seta (1984). For the relational orienting task printed category labels were placed on the table and participants were asked to sort the word-cards into their respective categories. In the case of uncertainty participants were asked to guess what category a word belonged to. For the item-specific orienting task, labels representing a 5-point pleasantness rating scale (very pleasant, a little pleasant, neutral, a little unpleasant, very unpleasant) were placed on the table, and individuals were asked to rate each of the words on this scale orally and not sort the cards underneath the labels. Regardless of orienting task individuals were asked to try to remember as many words as possible and following the last word all materials were cleared from the table and oral free recall was again requested. 3 Results A 2 (ASD vs. Comparison) × 2 (Baseline vs. Orienting Tasks condition) mixed ANOVA of the time participants spent looking through the decks of cards revealed a significant (F(1, 36) = 7.96, p < .01) effect of condition with participants spending an average of 323 s (S.D. = 404) looking through the cards whilst carrying out the orienting tasks compared to 244 s (S.D. = 254) during the baseline condition. Neither the main effect of group nor the interaction between group and condition were significant. An analysis of the time participants spent looking through the cards during the two orienting tasks revealed no main effects or interaction of the factors group (ASD vs. Comparison) and orienting task (Rate vs. Sort). Since the time participants spent looking through the cards correlated highly (r > .65) with overall recall levels for both groups in all conditions, encoding time was entered as a covariate in all subsequent analysis of the recall data (Miller & Chapman, 2001). 3.1 Baseline condition The free recall data for the baseline condition are illustrated in Fig. 1, which gives the average proportion of items recalled from the smaller (i.e. size 2, 4 and 8) and larger (i.e. size 12 and 16) categories for the ASD and Comparison groups. Overall, the ASD group recalled fewer words than the Comparison group (F(1, 37) = 8.08, p < .01) which was most marked with smaller categories resulting in a significant group by category size interaction (F(1, 37) = 6.89, p < .05). Thus, in line with our prediction that individuals with ASD would exhibit a recall decrement that would be indicative of specific difficulties in relational memory processes, the ASD group recalled significantly fewer words from the small categories (t(38) = 3.37, p < .01; Cohen's d = 0.96) but not the large categories (t(38) = 1.09, ns; Cohen's d = 0.30). We note that we have collapsed the recall data into ‘small’ and ‘large’ categories for simplicity and in order to facilitate the calculation of effect sizes. An analysis of the data across the five levels of category size yielded the same significant main effects and interactions (or lack thereof) as those reported above and in the analysis of the data from the orienting tasks below. As indicated earlier, Hunt and Seta (1984) suggest that the recall of at least one item from any given category (i.e. category availability—CA) represents the availability of relational information during recall, as does the amount of category clustering individuals employ during retrieval. Category clustering, as indexed by the Modified Ratio of Repetition is a simple ratio of the number of category repetitions (i.e. two consecutive items are recalled from the same category) to the total number of items recalled across all categories. As Hunt and Seta (1984) point out, more sophisticated measures of clustering are unsuitable for obtaining measures of organisation for each category size because they are mathematically undefined for a single category. In contrast to these indices of relational information, the number of items participants recall within a particular category (i.e. items per category—IPC) depends on the availability of item-specific information since such information aids the differentiation of individual instances of a particular category. In order to provide further insights into the use of item-specific and relational information to facilitate memory in ASD we computed these measures which are set out in Table 2.3 A 5 (category size) by 2 (group) mixed ANCOVA on the number of categories recalled (category availability) revealed a significant main effect for category size (F(4, 34) = 15.22, p < .001) with larger categories being nearly perfectly recalled whereas the smallest 2-item category was only recalled by 35% of participants. As expected, the ASD group recalled significantly fewer categories (F(1, 37) = 10.69, p < .01) and this effect was again characterised by a significant interaction between category size and group (F(4, 34) = 12.25, p < .001). Post hoc nonparametric comparisons showed that the ASD group recalled the small 2 and 4 item categories less often than typical participants (z = 2.82, p < .01, one-tailed) whilst both groups recalled the larger 12 and 16 item categories nearly perfectly. Although this result needs to be interpreted with caution due to the ceiling performance on larger categories, further evidence for the attenuated use of relational information to facilitate recall in ASD stems from the analysis of the category clustering data. Again this measure increased with category size (F(4, 34) = 10.13, p < .001) and again individuals with ASD clustered words into their respective categories less than the comparison group (F(1, 37) = 5.66, p < .05). Again the interaction between group and category size needs to be interpreted with some caution due to the floor performance on smaller categories. However, as the data set out in Table 2 indicate, clustering scores increased linearly with category size for typical participants whilst for the ASD group clustering only increased notably with a category size of 12. This quadratic trend is significant (F(1, 37) = 5.36, p < .05). In contrast, an analysis of the IPC data revealed no significant main effects of group (F(1, 37) = 0.04, ns) or category size (F(1, 37) = 1.25, ns) and no interaction between these factors (F(4, 34) = 0.23, ns). Thus, our findings from the baseline condition confirm that without any support, participants with ASD use relational information to aid recall to a lesser extent than typical individuals, whereas their use of item-specific information to help their recall appears similar to that of the Comparison group. 3.2 Orienting tasks Prior to analysing the recall performance following the orienting tasks, we assessed whether groups may have completed these tasks differently. During the category sorting task, participants in both groups performed at ceiling with only 3 ASD and 2 Comparison individuals committing either 1 or 2 errors. During the rating condition, ASD participants provided average ratings of 3.04 (S.D. = 0.41), which did not differ significantly from the average rating of 2.88 (S.D. = 0.27) given by the Comparison group. Similarly, an inspection of the frequency distributions of the ratings given by individuals revealed no differences between the groups. Taken together with the observation that groups did not differ significantly in terms of the time they spent looking through the deck of cards whilst they completed the orienting tasks these findings suggest no group differences in fulfilling the requirements of the orienting task instructions. Our analysis of the recall data following the orienting tasks (illustrated in Fig. 2) paralleled that of the baseline condition and encoding time was again entered as a covariate. A 2 (category size) by 2 (orienting task) by 2 (group) mixed ANCOVA of the recall data revealed a main effect of orienting task (F(1, 35) = 6.88, p < .05) indicating that recall following the relational encoding task (i.e. sorting words into categories) was superior to recall following the item-specific encoding task (i.e. rating words on pleasantness). The only other significant effect was an interaction between category size and orienting task (F(1, 35) = 12.79, p < .01), which replicates the findings reported by Hunt and Seta (1984). Post hoc comparisons showed that recall of items from the small categories was superior following the relational compared to the item-specific encoding task (t(38) = 4.15, p < .001; equal variance not assumed) whereas recall of items from the large categories was similar following either type of encoding task (t(38) = 1.51, p = .13). The lack of any interactions involving the group factor (Fs < 1) and the absence of a main effect of group (F(1, 35) = 0.67, ns) suggests that the provision of support in the form of orienting tasks attenuated the free recall difficulties seen in ASD. One may criticise this latter conclusion on the grounds that the reduced group sizes during the two encoding conditions decreased the statistical power of the analysis of these data in comparison to the baseline condition. In relation to this issue three aspects of our data are worth further comment. Most important amongst these is the observation that unlike performance during the baseline condition, recall following the orienting task conditions was not characterised by interactions between group and category size for either the relational orienting task (F(1, 17) = 0.36, ns) or the item-specific orienting task (F(1, 17) = 0.13, ns). Thus the patterning of performance as a function of category size did no longer differ as a function of group. Second, Cohen's d effect sizes for the between group differences in recall of items from small categories were reduced from 0.96 during the baseline condition to 0.57 following the relational orienting task and 0.08 following the item-specific orienting task (respective effect sizes for larger categories were 0.44 and 0.21). Third, although order confounds and differences in encoding time (i.e. time spent looking through deck of word cards) make analyses across baseline and orienting task conditions problematic, inspection of the data set out in Figs. 1 and 2 show that performance of ASD individuals following the relational orienting task was nearly identical to the comparison groups’ performance during the baseline condition. Interestingly the item-specific orienting task reduced performance of comparison participants to the level of ASD individuals’ baseline performance. We will return to the implications of these results in more detail in our discussion. Table 3 summarises the category availability, clustering and IPC data as a function of orienting task. As the category availability data suggest, overall recall of categories is generally better for larger categories (F(4, 32) = 24.97, p < .001) and following the relational orienting task (F(1, 35) = 12.04, p < .01). Furthermore, a significant interaction between category size and orienting task (F(4, 32) = 4.47, p < .01) indicates that the main effect of orienting task is mostly due to the increased availability of smaller categories following relational as compared to item-specific processing. Again the lack of a main effect of group or interactions involving the group factor (Fs < 1) suggests that the effect of item-specific and relational orienting tasks on the recall of categories was similar for the two participant groups. An analysis of the clustering data revealed a main effect of category size (F(4, 32) = 8.54, p < .001) and a marginally significant orienting task by category size interaction (F(4, 32) = 2.58, p = .056), which follows Hunt and Seta's observation of larger differences in clustering between the item-specific and relational encoding conditions for the smaller as compared to the larger categories. Again the group factor did not yield a main effect (F(1, 35) = 1.39, p = .25) or interactions with the other factors (Fs < 2). An analysis of the IPC data as a function of category size, group and orienting task, did not reveal any significant main effects or interactions (Fs < 1.1), thus not replicating Hunt and Seta (1984) who reported higher IPC scores following the item-specific orienting task, especially for larger categories. In summary, these analyses are in line with the suggestion that recall performance in ASD is no longer characterised by disproportionate difficulties in drawing on relational information when orienting tasks constrain the processes by which information is encoded. 4 Discussion In the current experiment we drew on a procedure developed by Hunt and Seta (1984) in order to evaluate the hypothesis that individuals with ASD are characterised by specific difficulties in relational memory processes. Furthermore we hoped to gain insights into the severity of such difficulties by assessing whether environmental support in the form of a relational orienting task could help individuals with ASD to employ such relational processes. Our results from the baseline condition support previous demonstrations (e.g. Bowler et al., 1997; Smith et al., 2007; Tager-Flusberg, 1991) of reduced recall in individuals with ASD when categorical information is available to aid recall. The finding that the ASD group showed selectively reduced recall of smaller but not larger categories confirms our prediction that ASD is characterised by relatively specific difficulties in relational but not item-specific memory processes. Further support for this view stems from the finding that the ASD participants recalled overall fewer categories and were less likely than typical participants to cluster items into their respective categories during recall. In contrast, the ASD participants recalled as many items per category (IPC) as the Comparison group indicating that they make as much use of item-specific information to facilitate memory as typical individuals. Together these results strongly suggest that, in the absence of any support, individuals with ASD employ relational memory processes to facilitate recall to a lesser degree than typical individuals whilst their ability to draw on item-specific information to aid recall seems relatively intact. Solely on the basis of the results from the baseline condition it is difficult to determine the severity of the relational memory difficulty evident in individuals with ASD. Our results from the supported encoding conditions shed some light on this issue. These results revealed that following item-specific and relational orienting tasks, overall recall performance between ASD and comparison participants were comparable. As noted in our results, we concede that this conclusion may be criticised on the basis of the reduced group sizes for each of the orienting task conditions, particularly because the ASD group's performance was numerically (if not significantly) worse than the typical group following the relational orienting task. What is crucial to note, however, is that unlike performance during the baseline condition the patterning of recall as a function of category size following the orienting tasks was very similar for the two groups as were the indices of relational and item-specific encoding. In other words, individuals with ASD no longer exhibited the disproportional difficulties with relational memory processes that characterised their performance during the baseline condition. In this context it is particularly noteworthy that the overall level of recall and the pattern of recall across category sizes of individuals with ASD following the relational orienting condition were almost identical to that of typical individuals during the baseline condition. Conversely, the comparison groups’ performance following the item-specific orienting task was nearly identical to that of the ASD group during the baseline condition. Thus, whilst the relational orienting task allowed individuals with ASD to achieve a level of performance comparable to that of typical individuals’ unsupported performance, the item-specific orienting task seems to have created a learning situation for typical individuals that mimics that experienced by individuals with ASD under normal circumstances. A possible limitation of our observations from the orienting task conditions is the fact that all participants first completed the baseline condition. On the basis of this order confound it may be argued that individuals with ASD simply required more practice in order to employ relational memory processes successfully. Although problematic to some extent, our conclusions would not be altered even if the improvement in performance by individuals with ASD is to some extent attributable to disproportionate practice effects. In relation to the task support hypothesis (Bowler et al., 1997), the findings from the orienting task suggest that support in the form of an orienting task (and perhaps increased practice) helps individuals with ASD to overcome difficulties in deploying relational memory processes effectively. Thus our main conclusion is that rather than lacking the capacity to process relational information sufficiently to aid recall, individuals with ASD experience difficulties in spontaneously deploying them in a way that fosters effective learning and memory in novel and unsupported situations. This conclusion is in line with an argument developed by Mottron and colleagues (Mottron, 2004; Mottron, Dawson, Soulières, Hubert, & Burack, 2006) on apparent conceptual difficulties in ASD. Rather than accepting the view that higher-level conceptual processes are impaired in this population, these authors contend that enhanced low-level perceptual processes compete with higher-level integrative functions. In the domain of memory this competition may occur between item-specific and relational encoding processes. As we have highlighted in our introduction, the distinction between item-specific and relational memory processes may not only prove useful in terms of understanding the behavioural pattern of memory functioning in individuals with ASD but it may also provide a fruitful heuristic framework for more direct investigations regarding the neuropathological correlates underlying memory functioning in this group. Since our observations are purely behavioural, we can only speculate about the neural underpinnings of the specific difficulties in relational memory processes that characterised performance of individuals with ASD in the current study. Given the evidence regarding morphological abnormalities of the hippocampus in ASD (e.g. Kemper & Bauman, 1998) and the growing evidence implicating this structure in relational memory processes, an appealing possibility is that the memory difficulties experienced by individuals with ASD stem from relatively specific functional atypicalities of hippocampally mediated memory processes (see also Nicolson et al., 2006). Although more direct neuroscientific investigations will be needed in order to specify the nature of this functional abnormality further, we would argue that at least two hypotheses may be generated on the basis of the current literature. Based on evidence suggesting that areas surrounding the hippocampus may under some circumstances mediate relational memory processes (see Eichenbaum, 2004 for a review), one possibility is that in ASD these adjacent areas are able to compensate for deficits in hippocampally mediated relational processes if environmental circumstances invite this level of processing. If environmental support is absent on the other hand, cortical areas adjacent to the hippocampus may simply perform their ‘default’ operations and mediate item-specific processes. Another possibility is that hippocampally mediated relational memory processes are principally intact but limited to such an extent that they are ineffectively deployed under spontaneous learning conditions. When environmental circumstances emphasise relational processes, however, this functional limitation may be sufficiently supported to permit a relatively typical behavioural expression of relational memory capacities. These two hypotheses are most likely not the only ones that may be put forward but we include them here to reinforce the point that the framework of item-specific versus relational memory processes provides a useful heuristic to generate future research to further specify the neural underpinnings of memory difficulties in ASD. In summary, our observations provide strong support for the view that individuals with ASD exhibit relatively specific difficulties in the spontaneous deployment of relational memory processes. We stress the term spontaneous because we think it important to distinguish between an impairment in the ability to engage in otherwise normally functioning processes and processes that are so impaired that they cannot function normally under any circumstances. Our finding that individuals with ASD no longer exhibit disproportionately attenuated memory for smaller categories following a relational orienting task illustrates that supporting learning environments promote relational processes in this group. Future studies will be needed in order to determine whether the relational processes employed by individuals with ASD under supported conditions are mediated by the same hippocampal-based neural mechanisms as in typical individuals or whether adjacent brain areas which typically mediate item-specific memory processes compensate for atypical hippocampal functioning.	[ "autism spectrum disorder", "task support hypothesis", "hippocampus", "medial temporal lobe memory system" ]	[ "P", "P", "P", "R" ]
Doc_Ophthalmol-4-1-2244701	Functional characteristics of patients with retinal dystrophy that manifest abnormal parafoveal annuli of high density fundus autofluorescence; a review and update	Purpose To examine the presence and functional significance of annular fundus autofluorescence abnormalities in patients with different retinal dystrophies. Methods Eighty one patients were ascertained who had a parafoveal ring of high density on fundus autofluorescence imaging. Sixty two had had a clinical diagnosis of retinitis pigmentosa (RP) or Usher syndrome with normal visual acuity. Others included a case of Leber congenital amaurosis and genetically confirmed cases of cone or cone-rod dystrophy (GUCA1A, RPGR, RIMS1), “cone dystrophy with supernormal rod ERG” (KCNV2) and X-linked retinoschisis (RS1). International-standard full-field and pattern electroretinography (ERG; PERG) were performed. Some patients with rod-cone or cone-rod dystrophy underwent multifocal ERG (mfERG) testing and photopic and scotopic fine matrix mapping (FMM). Results In patients with RP, the radius of the parafoveal ring of high density correlated with PERG P50 (R = 0.83, P < 0.0005, N = 62) and encircled areas of preserved photopic function. In the other patients, AF rings either resembled those seen in RP or encircled an area of central atrophy. Ring radius was inversely related to the PERG P50 component in 4 of 18 cases with a detectable response. FMM showed that arcs of high density were associated with a gradient of sensitivity change. Conclusions Parafoveal rings of high density autofluorescence are a non-specific manifestation of retinal dysfunction that can occur in different retinal dystrophies. Electrophysiology remains essential for accurate diagnosis. The high correlation of autofluorescence with PERG, mfERG and FMM demonstrates that AF abnormalities have functional significance and may help identify suitable patients and retinal areas amenable to future therapeutic intervention. Introduction Lipofuscin is derived from the shed and degraded photoreceptor outer segments and normally accumulates in the retinal pigment epithelium (RPE) with age [1–4]. An abnormal increase or depletion of lipofuscin manifests as high or low density areas respectively in fundus autofluorescence (AF) images [5] and such changes may aid in the detection and characterisation of a wide range of inherited retinal disorders, either by accentuating the visibility of fundus abnormalities or by revealing changes not accessible by routine examination or fluorescein angiography [5–13]. Lipofuscin has been shown to fragment when exposed to light [14] and its presence suggests continuing metabolic demand [5]. Absence of autofluorescence suggests either blocking of the incident radiation, photoreceptor cell death and RPE atrophy [5, 6] or disruption of the vitamin A cycle [15, 16]. Some patients with genetically different forms of retinitis pigmentosa manifest a ring of high density AF representing abnormal parafoveal accumulation of lipofuscin which usually encircles preserved foveal areas [17–20]. Similar parafoveal rings have recently been documented in other retinal dystrophies including some patients with Leber congenital amaurosis [21], Best disease [13], X-linked retinoschisis [22] and cone-rod dystrophy consequent upon mutation in GUCA1A [23], GUCY2D [24], RIMS1 [25, 26], or RPGRORF15 [26, 27]. Abnormal annular AF has also been described in some cases of “cone dystrophy with supernormal rod ERG” consequent upon mutation in KCNV2 [28, 29]. Unlike RP, these disorders often result in atrophic macular changes, manifesting as low density AF within the ring. The main aims of the current study were to review the electrophysiological phenotypes associated with annular AF abnormalities, and to examine two heterogeneous groups of patients in more detail by comparing indices of macular function associated with abnormal macular AF in rod-cone and cone-rod dystrophies. Materials and methods Autofluorescence imaging was performed using a scanning laser ophthalmoscope according to previously described techniques [5, 30]. Eighty one patients with abnormal AF of the posterior pole in the form of a high density parafoveal ring were reviewed. Sixty two had a clinical diagnosis of retinitis pigmentosa or Usher syndrome with a visual acuity of 6/9 or better and included 30 cases described previously [17]. Nineteen other individuals were ascertained from previous studies [21–23, 25–29] including 14 with cone-rod or cone dystrophy consequent upon mutation in RPGR ORF15 (4 cases), RIMS1 (8 cases) or GUCA1A (2 cases). Two subjects had “cone dystrophy with supernormal rod ERG” consequent upon mutation in KCNV2, two had RS1mutations and manifested rings in one or both eyes. One patient had Leber congenital amaurosis. Full-field ERGs were performed according to extended testing protocols incorporating the ISCEV minimum standard [31] in order to assess generalised retinal function. A stimulus 0.6 log units greater than the ISCEV maximum was also used, to demonstrate better the a-wave under conditions of dark adaptation. Pattern ERGs evoked by high contrast checkerboard reversal were recorded according to ISCEV recommendations [32] using standard parameters; checkerboard size 12 × 15 degrees, check size 45′, Michelson contrast 0.98. The PERG P50 component was used as an index of macular function. Additional PERG testing was performed using a range of circular checkerboard fields ranging from 3 to 18 degrees in diameter presented in a random order [17]. Check size was constant at 45′. International-standard mfERGs [33, 34] and fine matrix mapping were performed in some cases. Fine matrix mapping measures rod and cone system sensitivity and has been described previously [18, 35, 36]. In brief, detection thresholds for a spot of light were determined at 1-degree intervals over selected 9 × 9-degree retinal areas. Data were presented both as sensitivity contours, illustrating the position and orientation of tested retinal locations and as three-dimensional threshold profiles, plotted using interpolated values at 0.25-degree intervals, obtained by Gaussian filtering. Results All patients were selected according to fundus autofluorescence that was characterised by an abnormal high density parafoveal ring (Figs. 1, 2, 4–8). Sixty two with a clinical diagnosis of retinitis pigmentosa or Usher syndrome and a visual acuity of 6/9 or better had evidence of preserved RPE within the ring. Nineteen others had either cone or cone-rod dystrophy, “cone-dystrophy with supernormal rod ERG”, X-linked retinoschisis or Leber congenital amaurosis. Many of the older non-RP cases had evidence of macular RPE atrophy within the ring (Figs. 2b–d, 7c, 8b–c). Conversely, AF imaging in some RP patients showed mild to moderate atrophic RPE changes within the vascular arcades but external to the ring (Figs. 4d and 5e). Fig. 1Full-field ERGs, PERGs and AF in 3 patients with rod-cone dystrophy (a–c) including a patient with RP18 (a) and Usher syndrome (b). Row D shows full-field ERGs and AF in a case of Leber congenital amaurosis; ERGs and the AF image were obtained in the presence of nystagmus and are consequently noisy. Normal examples are shown for comparison (e). LU indicates log units greater (+) or less (−) than the ISCEV standard flashFig. 2Full-field ERGs, PERGs and AF in cone rod-dystrophy consequent upon mutation in RPGR (a), RIMS1 (b), “cone dystrophy with supernormal rod ERG” (c) and in a case of X-linked retinoschisis (RS1; d). Normal examples are shown for comparison (e). LU indicates log units greater (+) or less (−) than the ISCEV standard flash Figures 1 and 2 show representative full-field ERGs and PERGs in eight individuals with different retinal dystrophies. Figure 1 shows data from 3 subjects with rod-cone dystrophy and normal visual acuity including a case of autosomal dominant RP18 (Fig. 1a) and Usher syndrome (Fig. 1b). Figure 1d shows undetectable ERGs in a patient with Leber congenital amaurosis; ERGs are contaminated by the effects of nystagmus but no ERG is detectable, in keeping with severe generalised photoreceptor dysfunction. Figure 2 shows representative examples of cone-rod dystrophy consequent upon mutations in RPGR (Fig. 2a), RIMS1 (Fig. 2b) and KCNV2 (“cone dystrophy with supernormal rod ERG”; Fig. 2c). Figure 2d shows the ERG and AF findings in a patient with X-linked retinoschisis. In the 62 RP cases PERG P50 components varied between normal (>2 uV) and undetectable, consistent with varying degrees of macular involvement. There was no significant correlation with the bright flash ERG a-wave amplitude or 30 Hz flicker ERG amplitude. Figure 3 shows high positive correlation between PERG P50 and mean ring radius (r = 0.83, P < 0.0005). Data are shown from one RP patient that was tested comprehensively (Fig. 4); PERGs were normal to the smallest diameter checkerboard but minimal enlargement was seen as the stimulus field size was increased (Fig. 4b). Multifocal ERGs showed widespread reduction with relative preservation of the response associated with the central stimulus element (Fig. 4a), consistent with a central island of visual field preservation (Fig. 4c). The internal edge of visual field constriction corresponded closely with the ring of high density, as shown by photopic FMM (Fig. 4d). Scotopic fine matrix mapping revealed rod sensitivity losses that were severe and that encroached upon the central macula within the ring. Additional examples of mfERGs, visual fields and fine matrix mapping in RP patients are shown in Figs. 5 and 6. Fig. 3Comparison of mean ring radius with PERG P50 in 62 patients with rod-cone dystrophy (RP; broken linear regression line) and normal visual acuity and in 19 patients with other retinal dystrophies, including 4 cone or cone-rod dystrophy cases in which there was a detectable PERG (solid regression line)Fig. 4Multifocal ERGs (a), small field PERGs (b), Humphrey visual field (c) and photopic (d) and scotopic (e) fine matrix mapping in a patient with a clinical diagnosis of RP. Diamonds and error bars in (b) show mean values and standard deviations for 8 normal subjects; triangles and squares show patient data from right and left eyes. Contour plots (d and e) show sensitivity gradients over tested retinal locations; corresponding 3-D plots show retinal location (abscissa, degrees) and thresholds (ordinate, log units). Labelling (x) shows correspondence between the orientation of contour and threshold plots. Normal photopic and scotopic values are plotted in Fig. 5a and dFig. 5Contour sensitivity plots (rows 1 and 3) and 3-D threshold plots (rows 2 and 4) obtained in representative normal subjects (a, d) and in 3 RP patients (b, e and c, f). Subjects were tested under photopic (a–c) and/or scotopic conditions (d–f). Labelling (x) shows correspondence between the orientation of contour and threshold plots. Abscissa shows retinal location (degrees), ordinate axes show threshold (log units). Corresponding photopic FMM in individual (e) has been published elsewhere [18]. Normal 3-D plots show averaged data from 14 (a) or 12 (d) normal subjectsFig. 6AF images, mfERGs and corresponding Humphrey visual fields in 3 patients with RP and normal visual acuity Pattern ERGs were detectable in only 4 patients with cone or cone-rod dystrophy (GUCA1A, RPGR or RIMS1, Fig. 3). Pattern ERG P50 was inversely related to ring size in these subjects (Fig. 3). Data from one subject with cone-rod dystrophy (RIMS1) are shown in Fig. 7. Multifocal ERG showed widespread reduction with only relative preservation of the central response (Fig. 7a). Standard Humphrey visual fields showed a central scotoma and some superior field loss (Fig. 7b). Fine matrix mapping revealed severe threshold elevation across the macula but with relatively preserved photopic sensitivity over a central island of preserved RPE AF (Fig. 7c). Threshold values inside the ring are maximally elevated over a concentric atrophic area and show a gradient of increasing sensitivity over the arc of high density (Fig. 7c and e). The PERGs in this patient were undetectable (data not shown). Additional examples of fine matrix mapping in RIMS1 and RPGR patients are illustrated in Fig. 8. Central RPE atrophy was not always present (Figs. 2a, 8a). Fig. 7Multifocal ERGs (a), Humphrey visual field (b) and photopic (c, d) and scotopic (e, f) fine matrix mapping in a patient with cone-rod dystrophy consequent upon RIMS1 mutation. Labelling (x) shows correspondence between the orientation of contour and 3-D threshold plots. Abscissa shows retinal location (degrees), ordinate axes show threshold (log units). Threshold values for half the tested area have been removed from c and e, to expose the foveal values that would otherwise be obscuredFig. 8Contour sensitivity plots (rows 1 and 3) and 3-D threshold plots (rows 2 and 4) obtained in 3 patients with cone-rod dystrophy consequent upon RPGR (column 1) or RIMS1 mutations (columns 2 and 3). Subjects were tested under photopic (a–c) and scotopic conditions (d–f). Labelling (x) shows correspondence between the orientation of contour and threshold plots. Abscissa shows retinal location (degrees), ordinate axes show threshold (log units) Discussion This study reviews a heterogeneous group of 81 patients with genetically-determined retinal diseases that manifest a common feature on fundus autofluorescence imaging in the form of a parafoveal ring of high density. Patients with RP and normal visual acuity had rings that encircled preserved central AF. Mild to moderate atrophic changes were occasionally detected within the vascular arcades but eccentric to the ring (Figs. 4d, 5e). In the other retinal dystrophies that were examined in this and in previous studies [25–27], the ring could also encircle preserved central AF but in older subjects there was often central RPE atrophy within the annulus and preserved AF at more eccentric locations. As patients with RP or different forms of retinal dystrophy can have indistinguishable AF abnormalities, the AF appearance cannot be used to establish a diagnosis in such cases. Non-specific annular increases in AF may be associated with a wide variety of distinctive or pathognomonic full-field ERG changes that are essential for accurate diagnosis and functional phenotyping (see below). The data from 62 patients with RP and normal visual acuity show a high positive correlation between ring size and the PERG P50 component extending and confirming findings in a cohort of 30 of these cases [17], demonstrating the robust nature of this relationship. The findings are corroborated by high spatial resolution fine matrix mapping, visual field data and mfERG testing that shows preserved photopic sensitivity within central areas bordered by the ring, consistent with previous reports [18–20, 37]. Scotopic sensitivity losses encroach upon central macular areas suggesting that rod-system dysfunction precedes abnormal parafoveal accumulation of lipofuscin and progressive visual field loss [18]. Serial data indicate that the rings may vary greatly in terms of their stability; to date only 3 cases have been reported in which progressive ring constriction has occurred [37]. The rate of AF ring constriction may prove to be of prognostic value in predicting retention of visual acuity and visual field preservation, but further monitoring is required. In addition to RP, the parafoveal ring of high density may occur in cone or cone-rod dystrophy consequent upon mutation in GUCA1A [23], GUCY2D [24], RPGR [25, 26], RIMS1 [26, 27], in “cone dystrophy with supernormal rod ERG” (KCNV2) [28], in X-linked retinoschisis (RS1) [22] and in Leber congenital amaurosis [21]. Similar AF findings have also been documented in Best Disease [13] and other maculopathies [10]. In young patients with cone-rod dystrophy, small rings may have preserved central AF and are similar to those seen in RP cases [26]. Older individuals tend to manifest central atrophic changes that are encircled by the ring and there may be a central island of RPE preservation. Fine matrix mapping suggests that the abnormal accumulation of lipofuscin in cases of cone-rod dystrophy represents a transitional stage between relatively preserved parafoveal function and severe central dysfunction that is likely to precede central atrophy [26, 27]. In patients with cone-rod dystrophy consequent upon RPGR or RIMS1 mutations, serial studies have recently demonstrated ring expansion [26], suggesting an expanding front of macular photoreceptor dysfunction. This contrasts with RP where the opposite occurs [37]. It is possible that rings associated with other causes of maculopathy may also expand with time as lesions become larger with age. It is noted that RPGR mutations are more commonly associated with X-linked retinitis pigmentosa [38] with visual acuity reduction [39]; in the current study not all patients underwent genetic screening but none were known to have X-linked RP. Lipofuscin accumulation in the RPE is likely to reflect metabolic activity which is largely determined by outer segment renewal. There is evidence of light-induced degradation of lipofuscin [14], so that its presence would depend upon continuing metabolic demand. Evidence from RCS rats indicates that failure of RPE cells to phagocytose outer segments results in reduced lipofuscin levels [40]. Lipofuscin formation is almost completely dependent on a normal visual cycle and on the availability of dietary vitamin A [41]. In the RPE65 knockout mouse, the visual cycle is impaired and fluorescence is reduced [42]. Similarly, patients with RPE65 mutations can have reduced autofluorescence in the presence of near-normal fundi and OCT findings that do not suggest reduction in the photoreceptor layer thickness [15, 16]. It is widely accepted that photoreceptor degeneration results in atrophy and reduced rather than stable levels of autofluorescence and it is likely that autofluorescence is lost within a few weeks of photoreceptor loss. The presence of foveal or parafoveal AF in the photoreceptor dystrophies outlined in the current study, suggest that photoreceptors are likely to be intact and thus may be amenable to functional rescue. Recent developments, for example in the field of gene therapy, make early identification of candidate patients increasingly important. Patients with RP classically present with impaired night vision and visual field constriction, consistent with generalised retinal dysfunction involving rod more than cone photoreceptors. In the early stages the fundi may be near normal and rod-cone dystrophy is established by full-field ERG testing (Fig. 1). In patients with normal visual acuity, the degree of macular sparing cannot be predicted from the severity of peripheral dysfunction. The cone and cone-rod dystrophies are typically characterised by progressive worsening of visual acuity, dyschromatopsia, and photophobia with eventual central scotomata and peripheral field abnormalities. Ophthalmoscopic abnormalities, when present, are generally confined to the macula. Photopic full-field ERGs are typically delayed and subnormal with milder scotopic ERG abnormalities (Fig. 2a and b). A notable exception is the autosomal dominant cone dystrophy consequent upon mutation in GUCA1A; cone-mediated ERGs show amplitude reduction without significant implicit time delay [23] and although not diagnostic, may enable focussed mutational screening. Patients with “cone dystrophy with supernormal rod ERG” exhibit a wide range of fundus AF abnormalities including ring-like or bull’s eye changes, central atrophy or increased foveal AF [28]. Full field ERGs are pathognomonic in this disorder (Fig. 2c), have recently been shown to be consequent upon mutation in KCNV2 and have been described in detail [28, 29]. X-linked retinoschisis is usually associated with an electronegative bright flash ERG in keeping with inner retinal dysfunction (Fig. 2d); two cases (aged 47 years and 49 years) with annular AF abnormalities have been documented within a series of seven atypical but genetically confirmed cases [22]. The rings of high density in these individuals surround areas of central atrophy and may represent an outer boundary of macular photoreceptor dysfunction in addition to generalised inner rather than outer retinal disease. Commonly, younger individuals with X-linked retinoschisis manifest stellate macular lesions that are also visible in AF images [43]. Conclusions A parafoveal ring of high density autofluorescence is a non-specific manifestation seen in different retinal dystrophies. Electrophysiology remains essential for accurate diagnosis. The high correlation of autofluorescence with PERG, mfERG and FMM demonstrates that AF abnormalities have functional significance and may be an important parameter in the monitoring of these patients. Autofluorescence may be of prognostic value and may help identify suitable patients and retinal areas amenable to future therapeutic intervention.	[ "autofluorescence imaging", "retinitis pigmentosa", "usher syndrome", "cone-rod dystrophy", "electroretinography", "genotype-phenotype correlation" ]	[ "P", "P", "P", "P", "P", "M" ]
J_Med_Internet_Res-8-2-1550702	Who’s Using PDAs? Estimates of PDA Use by Health Care Providers: A Systematic Review of Surveys	Background Personal digital assistants (PDAs) find many uses in health care. Knowing rates of collective PDA use among health care providers is an important guiding step to further understanding those health care contexts that are most suited to PDA use and whether PDAs provide improved health outcomes. Introduction A handheld computing device, also commonly known as a personal digital assistant (PDA), is a mobile computer about the size of the palm of the hand. More modern devices can access external networks or the Internet through a wireless connection. Since 1993, when Apple launched the first PDA (Newton MessagePad), use of PDAs has increased worldwide, with global PDA sales projected to surpass 17 million in 2008. This represents a compounded annual growth rate of 17.8% between 2002 and 2008 [1]. Health care has not been immune to this technological advance in handheld computing. In fact, PDAs find many applications in health care. Family physicians and specialists have been using PDAs for general medical reference, such as drug interactions, pharmacopeias, and cardiac risk [2-4]. Other important applications of PDAs are those involving data collection and management, as in patient tracking, electronic Case Report Forms in clinical trials, patient diaries, and infection surveillance [4-9]. However, the suitability of PDAs across all health care contexts and whether they benefit health outcomes remain open questions. Many of us would agree that it is necessary to evaluate a technology before its adoption to allow health care providers to make informed decisions. However, given that technology is a moving target, a common problem with evaluation is that practice frequently precedes research. By the time researchers have obtained funding, completed a study, and published it, the technology is either in widespread use or has been abandoned [10]. As well, the appropriate type of evaluation is not independent of the stage of adoption of the technology. For example, if 90% of the target users have already adopted a technology, then studies evaluating its general utility will no longer inform the adoption decision. In this case, research should focus on optimization of the technology in use. This is a familiar scenario in information technology research, and it underscores the importance of understanding the rates of adoption in helping direct approaches to research [10]. In a general overview article, Fischer et al (2003) summarized the current literature covering the use of handheld devices in medicine, primarily related to PDA functionality [4]. While implementation issues were discussed, rates of adoption were not addressed. Further, a recent review of PDA use in health care by Baumgart (2005) examined operating systems, basic functionality, security and safety, and limitations of PDA use [11]. It is a thorough overview of studies published since 2000 that addresses applications of handheld computers for health care professionals, but it touches only briefly on the prevalence of handheld use. Therefore, to our knowledge, there has not been any structured review conducted to date that specifically addresses the extent of use of handheld devices and estimated adoption rates. As such, this paper aims to systematically summarize all available survey data on health care providers’ use of PDAs with the view of presenting the best available estimates of current PDA use. This paper also aims to project expected future adoption based on established technology diffusion models. From this information we draw implications for research and practice. Methods For the purposes of this systematic review of surveys, the term PDA is used synonymously to refer to any handheld device. Some examples include the following: Blackberry; Palm operating system devices, which include Palm Tungstens, Handspring Visor, and Sony Clie; and Pocket PC devices, which include the Compaq iPAQ and HP Jordana. Data Sources Surveys were identified as a subset selected from a broader systematic review examining all studies related to handheld devices in health care settings. Thus, initial search strategies and retrieved articles reflected this more extensive focus. This comprehensive literature search was conducted in consultation with an information specialist. The searched bibliographic databases covered both medical and engineering disciplines, including the following eight databases: Medline, Current Contents, Inspec, BA/RRM, Biotechnology, Biological Abstracts, EI Compendex, and EMBASE. The search was restricted to English-language literature published January 1993 (corresponding to the development of the first palm device) to February 2005. An updated search of Medline (PubMed) and EI Compendex (EI Village 2) was run near the project’s completion (January 30, 2006). Furthermore, the reference lists from included studies were examined in an effort to identify additional surveys not captured in the reference databases. In addition, surveys identified from Google searches and those known to the authors to have been conducted by private market research firms as well as physician groups were nominated for inclusion in our screening. Electronic Search Strategy The intent of searching the biomedical databases was to retrieve all studies related to handheld devices in health care. It is for this reason that the word survey was not included as a specific term in the original search strategy. The search did include the sample search terms detailed in Appendix 1. The search strategy for engineering databases limited retrievals to those articles relating to both handheld computing and health. All bibliographic databases were searched using subject headings tailored to each database and free-text terms in the titles and abstracts. Eligibility Criteria Surveys were included for this present review if they met the following initial criteria: related to an application in human health care and involved the use of a PDA device; contained original data; written in English (not including abstract or conference proceedings); published after 1993; and specifically reported handheld usage rates (prevalence of PDA use as a metric) in populations of health care professionals who were surveyed about the extent of their PDA use. Although conference proceedings were excluded, if deemed potentially relevant, a cross-check was conducted to see if there was an ensuing journal publication. A survey was not included if the handheld device being evaluated had undergone extensive custom modifications. A final set of unique references was identified and posted to the proprietary Web-based screening system SRS (Systematic Review Software). Selection Process The selection process for this present survey review consisted of two phases. First, it began with a screen of full-text articles that had already been retained because their title, abstract, or keywords suggested they contained relevant information on PDA use in health care settings. Therefore, for assessment of relevance, surveys were included if they appeared to contain pertinent study information and if there was no unequivocal reason for exclusion. Second, upon updating the searches, authors returned to the screening of the title, abstract, and keywords for each citation strictly to identify potentially relevant and most recent PDA usage surveys. Eligibility criteria were applied to the full-text surveys, which were reviewed independently by two reviewers (CG and KE). Disagreements were resolved by consensus. Figure 1 provides a modified QUOROM flow chart outlining the process for selecting identified PDA usage surveys. Figure 1 Modified QUOROM Flow Chart for Identified PDA Usage Surveys Data Abstraction The contents of each included survey were abstracted by one reviewer (CG), with an additional research assistant providing verification (TR). Analysis The data from all included surveys were extracted in a predefined, standardized fashion with abstraction verified by a second person and assessed descriptively (Appendix 2). Quality assessment methods for descriptive study designs such as surveys have not been established. Although some assessment frameworks exist for assessing survey research [12,13], none of them have been validated or empirically shown to include criteria that are associated with the reduction of bias in empirical surveys. Therefore, survey quality was not formally assessed. Table 1 Included surveys Year of Survey/Publication Author Prevalence of PDA Use Health Care Professional Group 1 1999/2000 Hucko [18] 15% (use in clinical work) Physicians 2 NS/2001 ACP-ASIM [19] 47% (use in clinical work) Specialists (Internists) 3 2001/2001 Versel* [20] 60% (use in practice) Physician Executives (organizational survey) 4 2001/2001 Martin [21] 19.3% (use in clinical practice) Physicians & Specialists 5 2001/2001 Taylor [22] 26% (use in practice) Physicians 6 2001-2002/2002 AAP [23] 38% (NS) Specialists (Pediatricians) 7 2000-2001/2002 Criswell* [24] 67% (use in practice) Residents (Family Medicine) (organizational survey) 8 2001/2004 Miller [25] 26.2% (office-based use) Physicians 9 2001/2004 Balen [26] 33% (use at work or home) Pharmacists 10 2001-2002/2004 Barrett [27] 75% (use in practice) Medical Residents 11 2002/2002 Martin [2] 27.9% (use in clinical practice) Physicians & Specialists 12 2002/2002 Versel* [28] 33% (use in physician offices) Physician Executives (organizational survey) 13 2002/2003 McCleod [29] 46% (use at medical institutions) Specialists, Medical Residents, & Fellows (Internists) 14 2002/2004 Carroll [30] 35% (use at work) Specialists (Pediatricians) 15 2002/2004 DeGroote [31] 61% (use on an academic health science campus) Health Sciences Faculty & Medical Residents 16 2003/2003 Martin [32] 32.9% (use in clinical practice) Physicians & Specialists 17 NS/2003 Vincent [33] 36% (use alone or in conjunction with log-card procedure in documenting) Medical Residents (Family Practice) 18 2003/2003 Versel* [34] 75% (carry & use PDAs) Physician Excutives (organizational survey) 19 2004/2005 AMA/Forrester [14] 57% (use regularly in a work week) Physicians, Specialists (Surgeons), & Medical Residents 20 2004/2005 Wilden [35] 91% own; 85% use on daily basis; 9% weekly; 215% monthly Specialists (Anestheologists) 21 2001/2005 Stromski* [36] 64% of programs report “most or all” residents use for clinical purposes Medical Resident Programs (Emergency Medicine) (organizational survey) 22 NS/2005 Stroud [37] 67% (NS) Nurse Practitioners & Students 23 NS/2005 Boonn [38] 45.1% (own or use daily) Specialists (Radiologists) NS/2004 Joy† [17] Difficult to interpret the prevalence numbers among the resident respondents Medical Residents (Obstetrics & Gynecology) 2004/2005 National Physician Survey (Canada)† [15] Unable to establish overall prevalence due to way data have been presented;48.6% of medical students have a PDA (although unable to infer use) Physicians, Specialists (various), & Medical Students Note: An excerpt from the “Taking the Pulse” study published in October 2004 by Manhattan Research [16] reports that 40% of all US physicians currently use a PDA, increasing from 35% in 2003. However, for this present review, the authors were unable to obtain a full copy of the report in spite of having contacted Manhattan Research on two separate occasions (February 2006). NS = not specified *Survey conducted at organizational level (vs individual level responses) †Survey of PDA use but prevalence data could not be established (referred to descriptively only) Results From a total of 816 full-text articles that underwent relevance assessment for a systematic review of the literature examining broad-ranging PDA use in health care, a subset of 18 surveys reporting PDA prevalence rates were identified (see Figure 1). Additionally, upon updating the search, an additional 959 records were retrieved and screened, from which 5 additional unique surveys were included. Furthermore, a total of 8 surveys were reviewer nominated, 3 of which were identified upon updating. Unfortunately, the authors were not able to obtain access to one Internet market research report. Prevalence numbers from 2 surveys were found too difficult to interpret, and, therefore, these data could not be utilized further in our results; however, we refer to both studies descriptively. It is from this pool of literature that a total of 23 unique surveys were identified (Table 1):15 were published articles in scientific journals, and 8 were nonacademic, reviewer-nominated citations that were either reports available for purchase, press releases, or trade magazine articles and thus not subject to formal peer review. Of these 8 surveys, 5 were conducted by Internet market research firms, 2 were conducted by physician groups, and 1 was conducted by a market research firm in conjunction with a physician group (American Medical Association). Survey Characteristics The included surveys were published between 2000 and 2005, with survey data collected between 1999 and 2004. One survey had a four-year lag between data collection and publication, three surveys had a lag of three years, and three surveys had a lag of two years. We were unable to determine publication lag in four surveys as no data collection dates were provided. Surveys were from the United States (16), Canada (4), Australia (1), both the United States and Puerto Rico (1), and both the United States and Canada (1). Survey methodology reflected the following: self-administered questionnaires distributed solely by mail (11); telephone interviews (2); Web-based online surveys (4); and combined distribution by electronic or postal mail as determined by the recipient (4). Two studies did not report the methodology used. Response rates ranged from 5.7% to 92.6% across 13 of the included surveys; 10 surveys did not report such rates. PDA Use In presenting the results, we group the PDA users by type of health care provider and personal characteristics (eg, age). In terms of PDA use, physician specialists were surveyed exclusively in five surveys. Three surveys examined practicing physicians, three included physicians and specialists combined, two included medical residents exclusively, while two surveyed an amalgam of physicians, specialists, medical residents, and/or students. Three surveys targeted physician executives and organizational practice leaders. One survey was directed at directors of family practice residency programs, while a further survey targeting individual PDA use in emergency medicine resident programs was completed at the organizational level. In addition to physicians as users of technology, one survey targeted practicing hospital pharmacists and another targeted a national sample of nurse practitioner students and faculty. One survey included faculty and residents across several health science disciplines, including medicine, dentistry, nursing, public health, pharmacy, and applied health science. To more accurately reflect handheld use across time, reported surveys were examined, when possible, from the timepoint when survey data were collected versus when published. When not possible, the publication date was the reported timepoint used. Collectively, the included surveys do indicate that PDA use is high, albeit somewhat variable, across studies. The reported prevalence rates of PDA use lend themselves well to an estimation of trend over time (Figure 2), and, as such, since 1999, there is evidence of an increase in PDA usage. Results do not include surveys completed at the organizational level. Surveys are presented according to data collection dates, with the exception of the American College of Physicians study (2001) [19], Stroud (2005) [37], and Boonn (1995) [38], which report publication dates only. The noted drop in 2003 is due to the paucity of surveys conducted in that year. Based on the most recent survey statistics (2004/2005), the current overall adoption rate varies between 45% and 85%, as derived from individual level survey data. In addition, of the five surveys completed at the organization level (eg, physician executives or medical program directors speaking on behalf of their individual members), the PDA use of their group members was estimated to be 60% (2001) [20], 67% (2001) [24], 64% (2001) [36], 33% (2002) [28], and 75% (2003) [34]. Figure 2Range plots of PDA usage by health care providers (n = 17); middle points represent range medians To elaborate on the percentage of overall adoptions rates, a US survey of 769 practicing physicians conducted in 1999 found that only 15% of physicians use a PDA in practice [18]. In a 2000/2001 survey of directors of family practice residency programs in the United States and Puerto Rico, use of handheld computers by either an individual or group was reported in 67% of the residency programs [24]. In 2001, 47% of 489 US-based internists surveyed were using a PDA [19]. A subsequent 2001 survey of 834 practicing physicians found that the proportion using PDAs had increased to 26% [22]. If we only look at professional use, then the increase is from 10% in 1999 to 18% in 2001 [22]. Among a national sample of practicing physicians surveyed in 2001, 26% reported using PDAs for office-based work [25]. In 2001/2002, 38% of 696 office-based physicians indicated that they used a PDA in their practice [23]. Of practicing hospital pharmacists surveyed in 2001, 33% reported using a PDA at work or home, with 28% using one on a daily basis [26]. These numbers reflect both types of use: personal and professional (ie, as an integral part of everyday practice). In 2001, 75% of residents in a teaching hospital reported using their PDA on a daily basis [27]. In 2002, 35% of US pediatricians were using a PDA at work, and 40% had one for personal use [30], and 46% of internal medicine physicians and residents were reporting PDA use [29]. In Canada, similar PDA use data have been collected since 2001 as part of the annual Physician Resource Questionnaire conducted by the Canadian Medical Association. PDA use among physicians increased from 19% in 2001 [21] to 28% in 2002 [2] and to a third in 2003 [32]. These data conclude that, in 2003, a third of Canadian physicians were using PDAs, which marked a 73% increase from 2001. Further, more than 50% of Canadian medical doctors under 35 years of age reported that they were using a PDA or wireless device in clinical practice [32]. The data did not differentiate type of professional use. In a PriceWaterhouseCoopers survey in 2001, 60% of the physician executives who responded indicated that their organization had at least one physician with a PDA [20]. Reportedly, this represented an upward trend from 26% in a similar 2000 survey. Further, in 2003, the trend continued, and 75% of respondents reported that their organization’s physicians were using PDAs. This increase in PDA use came after a steep decline to 33% in 2002 [28,34]. A sample of health science faculty and medical residents was surveyed in 2002 about their PDA use. Combined results from the various faculties and residents indicated that 61% used a PDA [31]. In 2004, 57% of a sample of US physicians indicated that they regularly used a handheld computer in a typical work week [14]. Results obtained in 2004 from a survey of members of the Austalian Society of Anaesthetists indicated that 91% of respondents owned a PDA; 85% reported using it on a daily basis, and 66% were reportedly “dependent” upon the handheld device, although the term dependent was not defined [35]. In 2005, when physician members of the Radiologicial Society of North America were surveyed, 45.1% reported owning or using a PDA on a daily basis [38]. However, the survey authors suggested use among this group of specialists appeared to be lower than for other physicians because a radiologist often works in front of a full workstation in clinic and therefore relies less on a mobile device. Further, PDAs are not yet well equipped to handle the tasks radiologists need to perform. In 2005, Stroud et al became the first group of researchers to address the use of PDAs in the field of nursing. Survey results concluded that the majority (67%) of participants used this technology [37]. While PDA use has clearly increased since 1999, it appears as though only a handful of studies have examined the prevalence and usage patterns of such technology outside of physician groups. Furthermore, when comparing the included surveys in depth, distribution of use is not uniform across selected characteristics of surveyed health care professionals. Therefore, further subgroup analyses from the included surveys are provided below. Patterns of handheld use are also briefly examined. Patterns of PDA Usage Age Based on a survey of 250 family physicians, as far back as 1995, younger physicians (less then 40 years of age) were more likely to consider carrying a handheld computer than older physicians (94% vs 84.5%) [39]. More recent data from this present review also suggest an age differential in usage patterns. A 2001 survey of 834 US practicing physicians found that use of handheld devices was higher among doctors under age 45 (33%) than among older doctors (21%) [22]. Another study found that pediatricians graduating from medical school in the last five years were more likely to use a PDA in practice than those who graduated more than five years ago [30]. According to a survey conducted by the American Academy of Pediatricians in 2001, PDA use was highest among those members under 30 years of age, with a reported usage rate of 75% [23]. Another study found that 60% of US internists below 40 years of age used a PDA, while only 34% older than 51 years did [19]. McLeod et al (2003) also found that PDA usage captured in 2002 among a sample of internal medicine physicians and residents under 30 years was much higher (68%) versus those over 40 years of age (37%) [29]. In Canada, 2003 usage was highest among younger physicians, with more than half of those under the age of 35 years (53%) using a PDA, compared with 15% of physicians aged 65 or older [32]. According to the American Medical Association/Forrester Research 2005 Physician and Technology Study, more doctors under the age of 40 years were reportedly using PDAs (55%) than those over 40 years (45%) [14]. In 2005, the mean age of nurse practitioners and students who reported using a PDA was 42 years [37]. Students and Medical Residents Residents tend to be younger, therefore it follows that they are more likely to use PDAs. This is also substantiated by direct evidence. A survey of directors of family practice in the United States and Puerto Rico conducted in November 2000 (306 responses) found that use of handhelds in residency programs, either by an individual or group, was 67% [24]. A 2001 survey of residents in a teaching hospital reported that more than 75% used their PDA on a daily basis [27]. Stromski et al (2005) surveyed emergency medicine residency programs in 2001 to identify the methods of procedure documentation to examine the number of programs transitioning to more advanced information technology systems (eg, PDA use). Their results indicated that 13% of the residency programs required the use of PDAs, 15% of programs purchased PDAs for their residents, and a similar proportion reported that PDAs were used by “most or all” of their residents to document procedures. Further, 64% of programs reported that “most or all” of their residents utilized PDAs for clinical purposes. DeGroote et al found that, in 2002, 71% of medical residents reported using PDAs versus 56% of faculty members [31]. In a 2002 survey, McLeod et al noted that the percent of frequent PDA users among internal medicine residents and fellows in training exceeded 70%, compared to only 50% of attending physicians [29]. From a survey of the experiences of family resident graduates in obtaining hospital privileges and in documenting procedures and deliveries, Vincent et al (2003) concluded that 36% of the respondents used a PDA alone or in conjunction with a log-card, paper-based system. Unfortunately, this study did not present any other prevalence data on PDA use [33]. However, from survey data captured in 2004, the handheld technology gap between residents and physicians began to close: a US study concluded that 73% of residents regularly used a handheld computer in a typical work week, followed closely by 71% of family/general practitioners [14]. In a survey of PDA use by nurse practitioner students and faculty, Stroud et al found that of the total respondents who reported PDA use, 73% were nursing students [37]. One survey by Joy et al (2004) met our initial criteria but could not be incorporated into the results analysis. Although this study did examine PDA use in obstetrics and gynecology residency programs, it was difficult to interpret the prevalence numbers among the resident respondents. Likewise, the National Physician Survey (2004) did not present overall PDA prevalence rates but did ask Canadian medical students if they had a PDA or wireless device [15]. Of the 2721 respondents, 24% in first year, 40.6% in second year, 70.6% in third year, and 71.6% in fourth year reported having a PDA, representing an overall average of 48.6% among students [15]. Unfortunately, these 2004 figures provide no information on how medical students were using this technology and in what contexts. Gender PDA usage among men and women was equal in a 2001 survey of internists [19]. Similarly, McLeod et al (2002) found no significant gender difference in PDA users among a 2002 sample of internal medicine physicians and residents [29]. However, pediatrician PDA users were most likely male, as reported in 2002 [30]. As well, the 2003 Physician Resource Questionnaire analysis concluded that male physicians were somewhat more likely to use a PDA in their practice than were females (35% vs 30%) [32]. More recent data from a 2004 survey of PDA use among US physicians, specialists, and medical residents suggested that male clinicians were slightly more likely than their female counterparts to regularly use handhelds (53% vs 47%) [14]. On the other hand, nurse practitioner data from 2005 show that men (82%) were notably more likely than women (64%) to use a PDA (P < 0.05) [37]. However, the authors cautioned that they were unable to determine the significance of this finding given that the actual survey sample of men (n = 38) as opposed to women (n = 188) was small. The authors suggested that if ease with PDA technology is less common in women, then the nursing profession, dominated by females, may need elevated momentum to adopt PDA technology across nursing practice [37]. Family Physicians versus Specialists The most recent Physician Resource Questionnaire (2003) analysis concluded that Canadian family physicians were just as likely to use a PDA (33%) when compared to medical (34%) and surgical (32%) specialists [32]. This was the third consecutive year these figures rose consistently across all physician groups in Canada [2,21,32]. However, according to a US survey of physicians published in 2005, the biggest adopters of PDAs in professional practice were family and general practitioners (71%) when compared to surgical specialists (54%) [14]. The above mentioned studies are the only survey data available directly comparing general physician use to that of specialists. Large and Hospital-Based Practices A US survey of practicing physicians found that use was higher among those who were wholly or partly hospital-based (33% and 29%, respectively) than among those who were office-based (23%) [22]. Usage was also higher among physicians in large practices (33%) than in solo practice (16%) [22]. Carroll et al (2004) also found that PDA users tended to not be in private practice [30]. Additional survey data from 2004 indicated that of US physicians practicing in primary practice offices with fewer than 10 physicians, 49% reported regular use of a handheld computer [14]. Miller et al (2004), reporting on a national sample of practicing physicians, found that in a group practice consisting of an average of nine physicians, handheld use was approximately 56% [25]. Urban versus Rural Physicians From a random sample of US pediatricians in 2002, PDA users were most likely from urban communities [30]. Similarly, results from Canada’s Physician Resource Questionnaire in 2001 indicated PDA use to be higher among physicians practicing in urban centers (19.9%) than in rural centres (13.4%) [21]. However, by 2002, rural use (29.6%) surpassed urban use (27.7%) among physicians [2]. In Canada, this trend continued in 2003, with 36.9% of rural respondents indicating PDA use versus 32.5% of urban respondents [32]. Professional Use Five surveys considered PDA use in both a professional and personal context; 17 studies exclusively captured professional use. One study reported general prevalence rates for PDA use among pediatricians; however, it did not specify if use was in clinical practice or outside of work. In order to discern professional use more closely, we explored administrative PDA uses versus direct use in clinical patient care. We found that of the surveys that concern PDA use within a health care setting, 17 of 23 studies (74%) reported use pertaining to administrative or organizational tasks, while 14 of 23 studies (61%) addressed PDA use in patient care. Billing and coding were the most frequently performed administrative PDA functions in 50% of the surveys reporting administrative uses. This was followed by 44% reporting calendar scheduling, 31% reporting Web and email access, 25% reporting address book use, and 25% stating use in charting patient details into an electronic health record. Other reported administrative tasks included the following: word processing, calculator, charge capture, procedure documentation, outpatient tracking, resident hours, telephone message tracking, general time management/personal organizer, patient referrals, procurement of supplies, patient census, order entry, dictation, and passwords and pins. In terms of patient care, access to drug information was reported in 93% of the surveys reporting clinical PDA use, while 50% reported prescribing, 43% stated accessing patient records, 43% described medical calculator use, and 36% indicated use in reference to laboratory values. Other reported clinical PDA uses included access to medical references, patient tracking and patient reminders, clinical decision pathways and managed care applications, telemedicine, and diagnostic imaging or radiology applications. Only one survey reported PDA use for patient education, and one referred to PDA use for research purposes. Discussion This paper summarizes the results from surveys examining adoption of PDA use. These survey data are in reasonably good agreement and suggest a sizable proportion of physicians use handheld devices. However, most of the sources of survey data did not distinguish well between types of applications being used most often and whether the PDAs were being used professionally for administrative purposes or for direct clinical work. It is encouraging to note that our findings are similar to those of an analysis of online registrations and downloads of a PDA drug reference guide, which concluded that approximately one fifth of US physicians (150000) and half of medical students in the United States (33000) were PDA users [40]. Our grouped survey data suggest that there is little information on the PDA usage rates among nonphysician health care providers. However, collectively, these data suggest that use of handheld devices has become a subject that health care professionals need to know about. By systematically gathering this usage information, it is difficult to deny the prevalence of PDAs in health care. With this basic understanding of current handheld usage patterns, we need to consider the impact of this development of mobile handheld technology on both practice and research. According to a commonly accepted descriptive model of the diffusion of innovations developed by Rogers, when the cumulative rate of users of a new invention is plotted versus time, the result is an S-shaped curve [41]. Interestingly, this appears to be true of most technological innovations, irrespective of the technology. For example, Hall and Khan (2003) reviewed the S-shape adoption patterns of a variety of 20th century consumer products (eg, washing machines, video cassette recorders) [42], while Teng et al (2002) developed historical diffusion curves for information technologies (eg, personal computers, email) [43]. Variations in diffusion slopes do exist given that some technologies will diffuse more rapidly than others. Health care information technologies have also been examined within this diffusion framework. England et al (2000) studied organizational and technological factors determining the rate at which innovations diffuse in the health industry [44]. In 2005, RAND Health completed a report characterizing the diffusion of electronic health records along an S-shaped adoption curve [45]. Technologies typically go through multiple phases during their adoption life cycle, which may last for many years [41,46]. The characteristics of the adopters change over time and so does the nature of suitable evidence to inform their adoption decisions. For example, innovators (the first 2.5% who adopt a new technology) do not need evidence to make an adoption decision. Early adopters (the next 13.5%) are satisfied with case studies and examples of successful adoption and benefits [41]. Examining the typical technology adoption curve for handheld devices (Figure 3) based on the adoption percentage of PDAs thus far from the most recent available data (2004/2005), it can be concluded that we are now at the steepest stage in the adoption S-curve, with a transition from the early majority to the late majority. Figure 3 The S-shaped diffusion of technology curve [41] The increase in PDA adoption means a potential reduction in hardware and training costs when using handheld devices in the provision of care and in research. Because of the high probability that target health care professionals may already have a handheld device and will already know how to use one, the overall hardware purchase costs could be reduced, and the end user will not necessarily have to be trained from scratch. To date, use of PDAs in health care appears to have preceded extensive evaluative research. PDA adoption rates, already high, continue to be a moving mark with projections for rapid growth in the short term. By comparing handheld device diffusion to other health information innovations, and by placing PDA use within existing diffusion models, we are able to better predict the future of handheld growth in health care and therefore develop more timely and appropriate evaluative research to accompany such growth. Unfortunately, we were unable to include information from two national physician surveys. The first report entitled “Taking the Pulse” was published in October 2004 by Manhattan Research [16]. Information gleaned from a report excerpt stated that 40% of all US physicians surveyed in 2004 were using a PDA, marking an increase from 35% in 2003. Reported top activities performed on a PDA by all US physicians (in order) were personal scheduling, professional scheduling, accessing a drug reference database, accessing online information, writing/entering clinical notes, and mobile email access [47]. These report findings are similar to our overall findings in this present review. The second national physician survey not incorporated into our analysis was the Canadian National Physician Survey (NPS) (2004), which provides valuable insight into what information technology, including PDAs, physicians and specialists have in their main patient care settings [15]. However, overall prevalence rates could not be determined from the data provided given the manner in which they were presented. Nonetheless, in reviewing the national data, we can descriptively draw some conclusions. First, it appears as though male physician PDA use is higher than that of females. This appears to be consistent across all tasks involving PDA use although differences do appear to be small. This is consistent with our general findings in which males are only marginally more likely to use a PDA than are females. Interestingly, when examining age-related data from the NPS, it appears as though the age factor may in fact be PDA task-specific. For example, electronic health record usage appears to decrease as the age of physician users decreases. However, PDA use for drug interaction information increases when the age of the physician user decreases. This appears contrary to most other surveys that show younger age is associated with higher general PDA use. Perhaps what this information tells us is that handheld use may be more complex when broken into task-specific strata. It is worthy to note that, with the exception of one survey focusing on nurse practitioner students, little mention was made in the surveys of PDA use by students across health care disciplines, including medicine. Several universities in Canada and the United States now mandate use of PDAs for medical undergraduate students and residency programs; therefore, it is assumed this could potentially affect prevalence rates. However, because none of the included surveys examined mandated use, we are unable to infer if this is responsible for recent increases. However, this raises an important issue to be considered in future studies related to students and rates of handheld adoption. To better understand the prevalence rates among the included surveys, it became important to categorize the drivers for PDA use as either professional or personal. We therefore attempted to discern what specific PDA tasks the respective health care professionals were performing. This was done by classifying, whenever possible, the use as administrative versus care. On the surface, it would appear that administrative and organizational tasks on a PDA exceed those related to patient care, perhaps signaling where the growth in adoption is most likely to occur. In this present review, we can only speak broadly to rates of adoption and patterns of use. Drawing inferences from the survey data was often limited by lack of, or differences in, operational definitions in aspects of handheld use being measured. For example, the term use was often not defined by frequency (eg, specific units of time—day, week, month). Taking these issues into consideration would be a useful exercise for future surveys as well as information technology prevalence studies in health care. In conclusion, physicians are increasingly accustomed to using a PDA, and, therefore, technology expertise will not likely be a barrier to deploying handheld applications. There is an urgent need to evaluate the effectiveness and efficiency of specific tasks using PDA technology (eg, implementation, searching, reference, data entry, reporting) to inform those persons developing and those using handheld applications. Furthermore, it is not clear why there is a paucity of evidence on the extent of adoption of PDAs by other health care providers: is it that they lag in the use of this technology or is it simply that they have yet to be studied? Limitations This review has a number of limitations. Issues around response bias and inability to draw causal inferences weaken survey methodology. It may be the case that those surveyed feel a stronger affinity to the survey sponsor, who has a greater interest in the questions asked, or are in complete disagreement with the topic at hand. This can skew results in difficult-to-measure ways. Quite possibly, the nonrespondents are the least committed (ie, nonusers of PDAs). As a result, the critical objective of drawing a true random sample of the populations that are the focus of the survey is compromised and the findings somewhat impure. The reported methodologies across these surveys appear to be heterogeneous, which limits their comparability. As noted, the quality of the included surveys could not be determined given the absence of validated quality assessment instruments, and, therefore, there was no adequate way to assess the influence of bias. A related issue is that some of the included surveys did not go through a rigorous peer-review process. These combined issues made judging the strength of the evidence not possible. One would assume surveys identified from scientific journals would be a source of less biased information. However, in defense of the nonacademic surveys, there is a consistency in results between those peer-reviewed versus those that were not. This may suggest that our main conclusions regarding adoption rates are fairly robust and not disconnected even with the inclusion of non–peer-reviewed evidence. Conclusions The objective of this study was to determine the adoption rates of PDAs in health care settings, and to project expected adoption in the future based on established technology diffusion models. Our findings from a systematic review indicate the current overall adoption rate for professional use of PDAs among health care providers, namely physicians, is 45% to 85%. Younger physicians, residents, and those working in large and hospital-based practices are more likely to use a PDA. Professional use in health care settings appears to be more focused on administrative tasks when compared to those related to patient care, although this requires further study. The adoption rate is now at its highest rate of increase according to a commonly accepted diffusion of innovations model. Additionally, the impact of PDA use on practice appears to be immediate in terms of costs and training. Familiarity will not likely be a barrier to deploying handheld applications in health care. However, there is a critical need to evaluate the effectiveness and efficiency of specific tasks using handheld technology within the health care system and across health care provider PDA user groups.	[ "health care", "systematic review", "survey", "personal digital assistant", "health technology adoption" ]	[ "P", "P", "P", "P", "R" ]
Int_Arch_Occup_Environ_Health-4-1-2254471	Sickness absence due to depressive symptoms	Objective There is no information on the duration of absence of depressed Dutch workers. The aim of this study was to determine the duration of sickness absence due to depressive symptoms in the working population. Introduction Mental illness is an important cause of disability and its impact on job performance and productivity is substantial. Depression is one of the most common mental disorders. According to the World Health Organization (2002), depression will be the leading cause of morbidity by the year 2010. Depression was found to be related to long absence from work (Harrison and Martocchio 1998; Nieuwenhuijsen et al. 2006) and to be a significant predictor of work disability (Kouzis and Eaton 1994; Skodol et al. 1994; Kessler et al. 1999). Depressed employees had a 28 times higher risk of absence as compared to employees who did not suffer from depression (Kouzis and Eaton 1994). Respondents with depression reported a mean of 35 days (95% CI 27–44) in the past year when they were totally unable to work or carry out their normal activities because of their depression (Kessler et al. 2003). When depressed employees stayed at work, they had a lower than normal productivity (Kessler and Frank 1997). Depressive episodes were reported to vary widely in their duration, with a median ranging from 2 to 12 months and with rates of chronicity (i.e. a duration of 12 months or more) between 15% and 50% (Keller et al. 1982; Coryell et al. 1994; Mueller et al. 1996; Eaton et al. 1997; Solomon et al. 1997; Spijker et al. 2002; Furukawa et al. 2005). Estimates of the mean duration of depressive episodes varied from 13 to 27 weeks (Blazer et al. 1994; Eaton et al. 1997; Kendler et al. 1997; Kessler et al. 2003). In international research the duration of sickness absence due to depressive disorder was 142 days in 213 adult psychiatric outpatients (Sorvaniemi et al. 2003). In Norway, the mean sickness absence duration in psychiatric patients amounted to 110 days in men and 103 days in women (median 57 and 52 days, respectively), with depression as the most common diagnosis (Hensing et al. 2000). In a Swedish sick-leave database a mean sickness absence duration due to combined depression and anxiety disorders was 73 days in men and 75 days in women (Hensing et al. 1996). In a reference population consisting of working people, the mean duration of a major depression was 52 days for a single episode (with 4.1% lasting longer than 6 months) and 74 days for a recurrent episode (with 9.2% lasting longer than 6 months) according to the Medical Disability Advisor (2006). There is no information on the duration of absence of depressed workers in The Netherlands (Raddjoe and van der Hoek 2005). The results of international studies are not comparable to the Dutch situation, because of differences in social legislation and sickness absence regulations. As it is very likely that the results of international studies are not comparable to the Dutch situation, the aims of this study were: (a) to investigate the mean (and median) duration of absenteeism due to depressive symptoms in the Dutch working population by sector and company size, and (b) to investigate gender and age influences on the mean (median) duration of absenteeism due to depressive symptoms. Studies on sickness absence because of depression revealed gender differences, with men having a higher risk of sickness absence when depressed (Laitinen-Krispijn and Bijl 2000). Women had a higher incidence of sickness absence due to mental illness (North et al. 1993; Hensing et al. 1996, 2000), but the duration of sick-leave spells due to mental illness were longer for men (Hensing et al. 1996, 2000). Based on Dutch national statistics (Central Statistical Office of the Netherlands 2006), in which women have longer absence durations and higher disability figures than men, we hypothesize that depressed women have longer absence durations than men (H1). We also hypothesize that older employees have longer absence durations than younger employees (H2) based on Dutch national statistics and studies of Hensing et al. (1996, 2000). Furthermore, we hypothesize that employees in small companies have longer absence durations as compared to employees in large companies (H3), because of less reintegration opportunities in small companies. Finally, we hypothesize that employees in educational and health care services have longer absence durations as compared to employees in other sectors (H4). Employees in the education and care sectors have a substantially increased risk of being at work when sick, because of difficulties in replacement (Aronsson et al. 2000) and therefore, when they take absence the duration may be longer. Methods Data The prevention, supervision and medical examination of sickness absence is a task of the employer for which he can choose to engage an occupational health department. ArboNed (the second largest occupational health department in the Netherlands) employs an absenteeism registration system for their affiliated companies. We register sickness absence and its causes in about 15% of the total Dutch working population. In our population, commercial services are over-represented (57% vs. 41%) and non-commercial services (e.g. health care, civil servants) are under-represented (17% vs. 34%) as compared to the total working population in the Netherlands. All absence periods that started between April 2002 and November 2005, and were diagnosed as depression were selected from the registration system. Absence periods were encoded as depression by the occupational physician when symptoms of depressed mood (such as feelings of profound sadness or emptiness) and/or reduced interest in activities that used to be enjoyed lasted for at least two weeks in combination with 3 or more of the following symptoms: decreased energy, decreased motivation, appetite changes, disturbed sleep pattern, agitation or inhibition, guilt feelings, low sense of self-worth, impaired concentration, or self-destructive thoughts. The symptoms must be accompanied by evident suffering and adverse effects on personal as well as social functioning. The age of the employees was registered at the moment of sick-leave. Absent men had a mean age of 41.6 years (SD = 9.9) and women of 38.2 years (SD = 9.9) (T = 17.2; df = 9,864; P < 0.01). The company size in which the employee worked was divided into four categories: <75 employees, 75–500 employees, 500–5,000 employees and >5,000 employees. In 95% of cases the company size was known. The companies were categorized into the following sectors (mean age ± standard deviation): construction industry (40.2 ± 10.2), health care (39.6 ± 10.1), trade (38.1 ± 10.3), catering industry (37.5 ± 10.0), industry (40.9 ± 9.8), education and public sector (43.9 ± 9.9), transportation and communication (41.0 ± 9.7), commercial services (38.6 ± 9.9), Other/unknown (39.9 ± 9.9). In 72% of cases the sector was known. Absence duration Between April 2002 and November 2005 a total of 9,910 newly originating episodes of absence due to depressive symptoms were registered. The period (in calendar days) between the first day of sick-leave and the date of return to work or disability pension was computed. In The Netherlands an employee can benefit a disability pension after one year of work incapacity1. Absence periods of 365 or more calendar days were considered to reach this limit, and therefore the maximal duration of an absence episode was 365 days. Absence duration was not corrected for part-time return to work, unless mentioned otherwise. When the number of absence days is corrected for part-time return to work, this is done by dividing the number of absence days by 1/reintegration percentage. For example, when the employee returns to work for 50% the number of absence days is divided by 2. When estimating the duration of absence spells, it is important to censor absences which have not ended by the end of the observation period or at the dismissal date (Blossfeld and Rohwer 2002). Therefore, we censored absences that did not end before the retrieval date (November 1, 2005). Absences which ended, because the employee resigned, were also censored. Statistical methods The data were analyzed using SPSS for Windows, version 13. Kaplan–Meier survival curves were computed in order to obtain mean and median absence duration. The Kaplan–Meier method (Kaplan and Meier 1958) calculates the risk set at every point in time where at least one recovery occurred. The risk set also includes episodes that are censored at this point in time. A censored episode contains the information that there was no return to work or reaching the one-year absence limit at the end of the study period or the dismissal date. The survival functions were plotted using life tables. The mean number of absence days corrected for part-time return to work in the different groups was compared using analysis of variance. If the main effect was significant, the Tukey post hoc test was performed in order to determine which groups differed from each other. Results In the period April 2002 to November 2005, 9,540 employees were absent with depressive symptoms. This is about 1% of the population covered by the occupational health department. Of these employees 9,196 (96.4%) had one absence episode, 321 (3.4%) had two absence episodes and 23 (0.2%) had three or more episodes of absence due to depressive symptoms. In Table 1, the course of sick-leaves with depressive symptoms is presented. Men more often returned to work and women more often reached disability pension (χ2 = 13.6; df = 3; P < 0.01). Table 1Number and course of depressive episodesMenWomenTotaln%n%n%Number of employees4,5225,0189,540Number of absences4,735100.05,175100.09,910100.0Returned to work within a year3,20567.73,35564.86,56066.2One year of incapacity for work71915.290317.41,62216.4Left the employment during the illness3076.53807.36876.9Not returned to work at end of observation period50410.653710.41,04110.5 Men and women of age groups <35 years, 35–44 years, 45–54 years and ≥55 years were distinguished. Figure 1 shows the life table survival function of episodes with depressive symptoms in each of these subgroups. Fig. 1Survival function of the duration of absence due to depression in men and women by age Women with depressive symptoms had a lower return to work rate than men. After 26 weeks 53.0% were still absent as compared to 48.8% in men. More women (25.5%) than men (22.3%) reached the disability pension date. In Table 2 the results of the Kaplan–Meier estimation are presented by age, company size and sector. The estimation of the mean duration of absence due to depressive symptoms (with a maximum of one year) was 200 (median 179) days in men and 213 (median 201) days in women. Table 2Mean (95% CI) and median (95% CI) duration of depressive episodes by age, company size and sector in men and womenNumber of absencesMean (95% CI)Median (95% CI)MenWomenMenWomenMenWomenAge <35 years1,3012,105184 (177–191)204 (198–209)153 (143–163)182 (171–193) 35–44 years1,6101,729200 (194–206)217 (211–223)181 (170–192)206 (193–219) 45–54 years1,3571,081212 (205–219)223 (215–230)199 (184–214)223 (205–241) ≥55 years456248205 (193–217)222 (205–238)182 (163–201)207 (163–251)Company size <75 employees1,3491,214214 (207–221)226 (218–233)198 (182–214)229 (210–248) 75–500 employees1,1741,225200 (192–207)210 (203–217)181 (167–195)201 (187–215) 500–5,000 employees1,4212,032196 (189–203)213 (207–218)174 (163–185)200 (188–212) >5,000 employees531423188 (177–198)208 (195–220)170 (157–183)179 (145–213)Sector Construction industry33839206 (192–220)206 (174–239)182 (151–213)189 (133–245) Health care1961,218212 (194–230)214 (207–221)202 (163–241)200 (184–216) Trade524406205 (194–216)212 (199–224)183 (163–203)200 (171–229) Catering industry176345197 (177–216)203 (189–217)168 (136–200)174 (144–204) Industry1,077341189 (182–197)205 (192–219)163 (149–177)192 (167–217) Education and public sector165196232 (213–251)242 (224–260)236 (199–273)272 (223–321) Transportation and communication417332196 (183–208)208 (194–221)182 (159–205)177 (151–203) Commercial services645682213 (203–223)219 (209–229)196 (176–216)212 (187–237) Other/unknown1,1971,616194 (186–201)212 (206–218)171 (160–182)201 (188–214)Total4,7355,175200 (196–204)213 (210–217)179 (172–186)201 (193–209) In all age categories, women with depressive symptoms were absent longer than men. Elderly employees had a longer duration of absence and a higher risk of reaching disability. Employees in companies with less than 75 employees were absent for a longer period than employees working in large sized companies. Employees in the educational and public sector had the longest absence duration, followed by employees in commercial services and health care. Men working in the industrial sector had the shortest duration of absence due to depressive symptoms. Table 3 shows the mean number of absence days corrected for part-time return to work. Men with depressive symptoms had significantly less absence days than women: 154 and 165 absence days, respectively (T = −5.3; df = 9,844; P < 0.01). Because the mean percentage of part-time return to work was almost equal in men and women, this was due to a shorter duration of absence episodes in men. Elderly employees with depressive symptoms had more absence days than younger employees (F(3, 9,862) = 16.5; P < 0.01), because of a longer duration of the absence episodes and lower part-time return to work percentages. The post hoc Tukey test revealed that the difference was significant in employees aged 35 years and more as compared to younger employees. Table 3Mean (95% CI) number of absence days corrected for partial work resumption of depressive episodes by age, company size and sector in men and womenMenWomenTotalMean (95% CI)Mean (95% CI)Mean (95% CI)Age <35 years140 (134–146)156 (151–160)150 (146–153) 35–44 years153 (148–158)169 (164–174)161 (158–165) 45–54 years164 (158–170)176 (169–183)169 (165–174) ≥55 years163 (152–173)178 (163–192)168 (160–176)Company size <75 employees166 (160–172)176 (169–182)171 (166–175) 75–500 employees154 (148–161)166 (159–172)160 (156–165) 500–5,000 employees150 (144–156)163 (158–168)158 (154–161) >5,000 Employees134 (125–142)157 (146–168)144 (137–151)Sector Construction industry164 (152–176)160 (132–189)164 (152–175) Health care168 (152–185)167 (161–174)168 (162–173) Trade158 (149–168)166 (155–177)162 (154–169) Catering industry153 (137–170)163 (151–176)160 (150–170) Industry143 (137–150)157 (145–169)147 (141–152) Education and public sector181 (162–199)196 (179–213)189 (177–201) Transportation and communication141 (130–151)151 (139–163)145 (138–153) Commercial services162 (154–171)164 (155–172)163 (157–169) Other/unknown151 (145–158)166 (161–172)160 (156–164)Total154 (150–157)165 (162–169)160 (158–162) Among employees working in companies with more than 5,000, we found more part-time return to work than in smaller companies, resulting in a relatively low mean number of corrected absence days (F(3, 9,344) = 14.7; P < 0.01). The post hoc Tukey test revealed that three subsets based on decreasing number of absence days could be distinguished: companies with less that 75 employees, companies with 75–5,000 employees and companies with more than 5,000 employees. The difference in mean number of absence days between the sectors was significant (F(8, 9,880) = 8.2; P < 0.01). According to the post hoc Tukey test, employees in the educational and public sector had the highest number of absence days corrected for part-time return to work. Employees in the transportation and communication sector had a significantly lower number of absence days as compared to the other sectors. Discussion Our study showed that depressed women were absent longer from work than men, which confirmed our first hypothesis (H1). The mean duration of absenteeism due to depressive symptoms was 213 (median 201) days in women and 200 (median 179) days in men. This could be due to gender differences in the severity of the depressive symptoms. However, no gender differences were found in the duration of depressive episodes in a population sample (Eaton et al. 1997) and in the severity or symptomatology of depressive disorders in general practice (Hildebrandt et al. 2003). In studies on absence and depression, reverse gender differences are found. For sickness absence in general, women have more sickness absence days than men (Harrison and Martocchio 1998; Kivimäki et al. 2003; Central Statistical Office of the Netherlands 2006). According to Laitinen-Krispijn and Bijl (2000), depression was a stronger risk factor for absence in men than in women. However, Hensing et al. (1996, 2000) found a higher incidence of absence due to psychiatric disorders in women as compared to men. Hensing et al. (1996, 2000) even found longer absence durations for men with psychiatric disorders as compared to women. They reported about neurosis, and we investigated depressive symptoms, which might explain part of the difference. Hensing et al. (1996, 2000) reported that the longer duration of absence among men could be explained by gender differences in co-morbidity with alcohol problems and other psychiatric disorders. They also mentioned gender differences in seeking health care, diagnostic procedures and treatment of psychiatric disorders. The social consequences of having a psychiatric disorder might be worse for men than for women, leading to under-reporting among men. Men might postpone the use of health care and get into a poorer mental condition. As a consequence, they need a longer period of recovery and are absent longer from work. Gender differences in the (assessment of) severity of psychiatric problems might also explain the different duration of sick-leave. However, if these mechanisms do play a role in our study, they are more unfavourable to depressed women than to depressed men. Dewa et al. (2003) reported that depressed women returned to work more often than men, rather than leaving employment. They defined the end of the absence period as returning to work either part-time or fulltime. In our study men returned more often to work than women. We regarded fulltime return to work as the end of an absence period, because we did not have the exact dates of part-time return to work. It is possible that women returned more often to work part-time than men. Depressed elderly workers are absent longer than young employees with depressive symptoms, confirming our second hypothesis (H2). Sickness absence increases with age, regardless of diagnosis (North et al. 1993; Marmot et al. 1995). Hensing et al. (2000) found higher rates of absence with psychiatric diagnoses in older ages. Like in our study, the increase they found is less linear than in general absence studies. This can be explained by the fact that sickness absence with a psychiatric diagnosis differs from sickness absence due to other diagnoses, since the duration of absence is long in the youngest age groups as well. Age differences in the severity of psychiatric disorder, in help-seeking behavior, recognition and/or treatment may influence the duration of sick-leave due to depression. An alternative factor, which may contribute to the longer duration, is that elderly employees prefer to retire from the labour force. In contrast with Hensing et al. (1996) we found that men in all age groups, especially in the youngest age groups, had fewer days of sick-leave than women in the corresponding age group. Further study is necessary to investigate these gender differences by age. In accordance with our third hypothesis (H3), the periods of absence due to depressive symptoms were shorter in large-scale companies than in smaller ones. The reintegration percentage is higher in large companies, resulting in a lower number of absence days. Probably, large companies have more opportunities for part-time return to work. Moreover, large companies have structured protocols as to how to deal with long-term absence from work. In the transportation and communications sectors, part-time return to work was most often observed. In our sample, these are mostly large companies in the postal and telecommunications sector, offering more reintegration possibilities. Although blue-collar workers generally have higher sickness absence than white-collar workers, depressed employees in industrial settings had shorter absence durations. A possible explanation might be that employees in the industrial sector are more susceptible to stigmatization of having mental problems, and return to work earlier. Also, in blue-collar workers, depressive symptoms may more often be masked by or diagnosed as somatic diseases. Employees in the educational and public sectors had longer duration of depressive episodes, resulting in more absence days, confirming our fourth hypothesis (H4). An explanation is that these employees find it harder to report themselves sick because of the difficulty to substitute their work and the personal ties with their clients and pupils (Aronsson et al. 2000). Probably they have more severe depressive symptoms at the moment of sick-leave. Another explanation for their longer absence might be the lesser reintegration opportunities in these sectors. Finally, the mean age of employees in the educational and public sectors was higher than average, and it is known from Dutch national sickness absence surveys that elderly employees are absent longer than younger ones. More research is needed to examine these possible explanations. Duration of absence due to depression compared to other absenteeism studies The duration of absence due to depressive symptoms is longer than reported in international research (Hensing et al. 1996, 2000; Sorvaniemi et al. 2003) and the figures according to the Medical Disability Advisor (2006). Differences in definitions and methods may account for this finding. Hensing et al. (1996) reported on episodes of combined depression and anxiety disorder, exceeding 7 days and without maximum duration. In contrast to our study, they did not censor for the end of the observation period. Hensing et al. (2000) investigated sickness absence due to neurosis with a minimum duration of 15 days until the end of the absence. We investigated absence due to depressive symptoms with a minimum duration of 1 day and maximized at 365 calendar days. In our study, 98% of cases had a minimum absence duration of 15 days. The code depressive symptoms was given in the second or third week of absence, when the absent employee consults the occupational physician. Sorvaniemi et al. (2003) studied sick-leaves with depression according to DSM-III-R in psychiatric outpatients. They used exact beginning and ending dates of sick-leaves, resulting in durations ranging from 2 to 374 days and a mean duration of 142 days. At this point, it should be reminded that depressive symptoms were not strictly assessed according to the DSM criteria, limiting the comparability of results. Besides, the occupational physician can only give one diagnosis per absence episode. Comorbidity could not be coded. Approximately two-thirds (64%) of respondents with 12-month major depressive disorder meet the criteria for at least one other mental disorder, with anxiety disorders (58%) more common than substance abuse (9%) or impulse control (17%) (Kessler et al. 2003). Comorbidity may substantially prolong the duration of an absence episode. The long duration of absence due to depressive symptoms might be explained by comorbid features. Alternatively, it is possible that depressive symptoms were reactive to somatic disease. In case the symptoms of somatic disease appeared first, the occupational physician encoded the somatic disease and not the reactive depressive symptoms. During sick-leave, the occupational physician evaluates the medical situation and recovery every 6 weeks. Bias can only be expected when the somatic disease was not known in the sick-leave period of the depressed worker. Further investigation of the influence of comorbidity by more detailed research of the medical files is necessary, as Chisholm et al. (2003) reported that the economic consequences of depression (e.g. number of lost work days) were influenced to a greater extent by the presence of medical comorbidity than by symptom severity alone. Duration of a depressive episode compared to the duration of absence The median duration of absence due to depression in our study exceeds the estimations of duration of depressive episodes in the general Dutch population (Spijker et al. 2002). In a population survey, Kessler et al. (2003) found a mean duration of a depressive episode of 16 weeks using the composite international diagnostic interview (CIDI). This is about half of our estimated duration of absence. Population studies may underestimate the duration of a depressive disorder, because adults with severe forms of depression are less likely to participate in such studies (Centers for Disease Control and Prevention 2004). Studies on the duration of a depressive episode using the additional life chart interview showed a mean duration of 26 weeks in men and 27 weeks in women (Eaton et al. 1997). This corresponds to the mean absence durations we found: 28.6 weeks in depressed men and 30.4 weeks in depressed women. Blazer et al. (1994) found an overall mean duration of 22.6 weeks by means of a diagnostic interview. For the purpose of their analysis, however, 3% of the participants who had had very long episodes were excluded. When chronic cases are considered, our estimates were comparable to the study of Spijker et al. (2002) in the general population. They found a chronic course (duration 12 months or more) in 24% of cases. Our survival analysis also indicated a chronic course in 24% of cases. It is important to investigate which employees are at risk for a chronic course. The duration of a depressive episode in The Netherlands is comparable to other countries, whereas the duration of absenteeism is longer. It could be hypothesized that Dutch employees return to work when they are completely recovered, while in other countries employees start working (part-time) during recovery. We think that differences in social legislation and benefits contribute to these differences. In most other European countries compensation schemes apply to work-related diseases (“risque professional”). In The Netherlands, however, the social insurance system does not take into account the cause of the disease (“risque social”). Moreover, social benefits could have contributed to the relatively long absence duration. Dutch employers are obliged to pay the employee on sick-leave at least 70% of his/her salary. In almost all cases the full salary is paid during the first year of sick-leave, which does not motivate employees to return to their work. Limitations of the study Our population is not a random sample from the total Dutch working population, because employers voluntarily engage our occupational health department. Therefore, results are not representative of the whole Dutch working population. However, the considerable sample size and the use of registered, rather than self-rated information on sickness absence, are strong points of this study. A depressive episode was diagnosed in employees who visited the occupational physician. Persons with lower grade depressive moods might not have reported complaints, and stayed at work. If absent, they might have returned to work before they consulted the occupational physician. Alternatively, they might have called off a visit to the occupational physician, because they intended to resume their work soon. Therefore, more severe cases of depression were included in our sample. The duration of absence due to depressive symptoms is related to the severity of depression (Dewa et al. 2003). We investigated the interaction between depression and work sectors, but could not examine the influence of work content and working conditions. Rugulies et al. (2006) reported work environment to influence the risk of developing severe depressive symptoms. Work factors also influence the probability of returning to work once absent (North et al. 1996; Väänänen et al. 2003; Nielsen et al. 2004). Clinical implications Because of the risk of chronicity, it is important that occupational physicians recognize depressive symptoms in an early stage of the absence episode It is recommended to develop and apply tools for recognizing employees at risk for chronic depression Special attention should be given to employees in educational and public services, commercial services and health care as they are at risk for longer absence duration in case of depressive symptoms. Study limitations Diagnosis of depressive symptoms in absent employees may be biased towards absences with a longer duration. Mildly depressed employees, who did not report ill or returned to work before consulting the occupational physician, were not included Depressive symptoms were not strictly assessed according to DSM criteria, limiting the comparability to other studies Our absence registration system does not account for comorbidity.	[ "depression", "sector", "company size", "duration of sick-leave", "mental health", "gender and age differences" ]	[ "P", "P", "P", "P", "R", "R" ]
Pediatr_Radiol-3-1-1950216	Magnetic resonance imaging protocols for paediatric neuroradiology	Increasingly, radiologists are encouraged to have protocols for all imaging studies and to include imaging guidelines in care pathways set up by the referring clinicians. This is particularly advantageous in MRI where magnet time is limited and a radiologist’s review of each patient’s images often results in additional sequences and longer scanning times without the advantage of improvement in diagnostic ability. The difficulties of imaging small children and the challenges presented to the radiologist as the brain develops are discussed. We present our protocols for imaging the brain and spine of children based on 20 years experience of paediatric neurological MRI. The protocols are adapted to suit children under the age of 2 years, small body parts and paediatric clinical scenarios. Introduction The use of imaging protocols allows the standardization of procedures and workflow as well as consistency of image quality. With appropriate training, radiographers can select and implement imaging protocols with relatively little requirement for radiological involvement. This use of skill mix provides radiographers with an expanded role and greater responsibility that increases job satisfaction and staff retention. Imaging protocols can also provide guidance for radiologists and radiographers for patients in shared care between secondary and tertiary care institutions. The use of protocols set up by the referring centre will prevent repeat MR scanning with the additional burden on MR scanner time. These advantages are offset by the minor disadvantage that radiologists’ personal preferences and the intricacies of individual scanners are not taken into consideration. At Great Ormond Street Hospital (GOSH), MRI protocols have been implemented for nearly 20 years and have been modified as scanners are replaced and new sequences developed. Our success is measured by the minimal number of patients recalled (less than 1%). Increasingly, we are asked about our MR protocols by departments throughout the United Kingdom and have set them out here to make them freely available to those who may find them beneficial to patient care. The protocols provided are confined to conventional imaging techniques and diffusion-weighted imaging (DWI), and do not include the advanced imaging techniques of perfusion imaging, diffusion tensor imaging and MR spectroscopy that remain largely within the research domain. Practical issues of MRI in children One of the biggest challenges of paediatric neuroimaging is the acquisition of high-quality diagnostic images, as it requires the infant or child to keep still for a long period of time (sometimes nearly 1 h). Children over the age of 7 years are scanned, where possible, without sedation and with recourse to play specialists and videos. Neonates under 2 months old (corrected age, if premature) are scanned during natural sleep induced by food, comfort and warmth (feed and wrap), often best after a period of sleep deprivation. If this fails then light oral sedation, e.g. chloral hydrate, may be used. The majority of our patients (approximately 70%) require sedation or a general anaesthetic. When sedation is required, it is usually heavy and heavier than that required for CT or nuclear medicine studies. Sedation is applied by specially trained nurses who are trained to decide when and how much to administer and monitor its effects, and who are skilled in the recovery of the child [1]. Details can be found in a review of the management of painless imaging in children [2]. All MR sequences may disturb the sleeping infant or child and ear protection such as earplugs and baby earmuffs should be used. Some motion can be avoided by swaddling infants, keeping them warm, and by placing moulded foam or airbags around the baby’s or child’s head. Cardiorespiratory monitoring with MR-compatible equipment is essential in all infants, whether sedated or not, and all sedated or anaesthetized children [2]. Commercial MR machines are designed for adult practice and few manufacturers make provision for the issues encountered in paediatric practice. Currently, most children are imaged in units that have a wider adult practice with only a handful of institutions throughout the world having dedicated paediatric MR facilities. The structures we need to examine in children are generally small and we aim for a maximum slice thickness of 5 mm in the brain and 3 mm in the spine. The slice thickness is reduced to 3 mm when acquiring images of the pituitary gland and orbits. We do not recommend ultrafast ‘breath-hold’-type T2-weighted (T2-W) sequences that have reduced contrast to noise as an alternative to sedating a child, as achieving high contrast to noise is of utmost importance given the small size of the paediatric head. Advances in coil design have improved signal-to-noise and we now use multichannel head and spine coils. Practical consideration of the changing brain with age Changes in normal values of measured parameters in very young children occur as the brain undergoes very rapid development during the first 2–3 years of life. This necessitates the creative modification of pulse sequence parameters in order to optimize the soft-tissue contrast between anatomical and pathological structures. The two most important physiological changes to take place within the first 2–3 years are reduction in the total brain water content and T1 and T2 shortening largely as a consequence of the maturation of myelin [3]. From birth to 6 months of age the maturation of myelin is best assessed using T1-weighted (T1-W) images. From 6 months until the completion of myelination at around 2 years of age, T2-W images are more useful. During the process of myelination on T1-W images the signal from white matter goes from dark to bright. For a variable period the subcortical white matter becomes isointense with grey matter. This loss of grey/white matter differentiation obscures structural detail and makes identification of subtle abnormalities such as polymicrogyria difficult to detect within the first 6 months of life. The white matter shows maturity later, between 6 months and 2 years, on T2-W images and at this stage T1-W images are essential to evaluate structural abnormalities [4]. To overcome the difficulty of increased brain water in those under 2 years of age, we use a dual-echo short-tau inversion recovery (STIR) sequence that has improved contrast resolution in this age group (Fig. 1) [5]. Typical values for a dual-echo STIR sequence are TE 30/128 ms, TR 5,400 ms and TI 130 ms. We prefer this sequence to the more conventional T2-W fast (or turbo) spin-echo sequences, for reasons of increased lesion conspicuity. However, we recognize that similar tissue contrast may be achieved if such fast spin-echo sequences are modified using increased TR and TE to compensate for the increased water content in this age group. The additional brain water is difficult to suppress and we do not find the fluid-attenuated inversion recovery (FLAIR) technique useful in this age group as lesion conspicuity is poor. Fig. 1Axial images acquired through the bodies of the lateral ventricles of a 6-month-old child. First (a) and second (b) echo (TE 30/120 ms, TR 5,500 ms, TI 130 ms) of a dual-echo short-tau sequence shows the increased contrast between grey and white matter on the second echo compared to (c) the T2-W (TE 90 ms, TR 3,500 ms) sequence that would be used in those over 2 years of age. The slice thickness (5 mm) and matrix size (512 × 192) were the same for both sequences Diffusion measurements in the brain of newborns and infants show that there is more movement of water (high apparent diffusion coefficient, ADC) with less directionality (low anisotropy) within the immature and unmyelinated brain than in the mature and myelinated adult brain [6, 7]. The most dramatic decrease in ADC values occurs within the first few months of life and the greatest changes are seen in the frontal and parietal white matter [8]. Failure to appreciate the normal changes with age of the ADC will result in the erroneous reporting of white matter abnormality particularly in the neonatal period. Factors influencing sequence choice For any imaging protocol, it is the specific combination of sequences that determines the diagnostic efficacy of the examination. It is clear that imaging protocols vary between institutions. This is largely because rapid advances in imaging technology and variations between manufacturers, applied in the context of investigating relatively rare disorders, precludes an effective evidence-based approach to sequence choice. The main advantages that MRI offers over alternative imaging modalities is the ability to demonstrate different tissue contrasts (principally T1-W, T2-W and spin density, but also flow and diffusion) in multiple imaging planes (principally sagittal, coronal and axial). The disadvantage of MRI is the artefacts that it generates in almost every image. The choice of sequence combination should reflect the multicontrast and multiplanar capabilities of MRI. We have found that applying the generic principle of combining T2-W images in two planes, supported by T1-W imaging in two planes, as the basis of our imaging protocols, serves to optimize the benefits of MRI whilst minimizing the impact of artefacts. Standard MRI studies Brain Our standard brain sequence (>2 years of age) is composed of an axial T2-W, coronal FLAIR, and coronal and sagittal T1-W images. Sagittal T1-W images allow the assessment of the midline structures, particularly the corpus callosum and cerebellum, which are frequently abnormal in congenital abnormalities. Coronal T1-W images are valuable in the diagnosis of abnormalities of the anterior visual pathway, schizencephaly and holoprosencephaly. Axial T2-W and coronal FLAIR images provide T2-W images in two different planes and have been shown to be complementary in children [9]. DWI is acquired in all children unless artefacts from, for example, dental braces or a ventriculoperitoneal shunt, preclude it, and an ADC is calculated using automated computer software and provided for reporting. In those under 2 years of age, the T2-W sequence is replaced by a dual-echo axial STIR sequence (see above). In some cases, a T2*-W gradient-echo (GE) sequence (“susceptibility-weighted” sequence), sensitive to changes in local field inhomogeneity caused by the breakdown products of haemoglobin, is added. The sequence is particularly useful in trauma and vascular malformations such as multiple cavernomas. Spine MRI is the modality of choice for imaging the intraspinal components of the paediatric spine. Prior to ossification of the posterior elements of the spinal column, US has been shown to be a valuable screening tool; however, infants with an abnormal sonogram or who have a neurological abnormality in the context of a normal sonogram, still require MR imaging [10]. Our standard spine imaging includes sagittal, fast spin-echo T1- and T2-W sequences (3-mm-thick slices). Both axial T1-W and T2-W images are acquired through any abnormality. Unlike most adult spine imaging protocols, groups of axial images through disc levels are not applied because degenerative disc disease is rare. Children with scoliosis and/or suspected spinal dysraphism routinely have axial T1-W images through the conus and filum terminale to detect lipomas of the filum terminale that may not be visible on sagittal imaging. Contrast medium The use of contrast medium is a personal choice and in the majority of units is requested on a case-by-case basis. At GOSH the radiographers or sedation nurses administer MR contrast medium according to guidelines set up by the neuroradiologists (Table 1). In the event of an unexpected finding the radiographers are encouraged to make a decision on contrast administration without discussion with a radiologist. Table 1Indications for contrast medium administrationVarious indicationsAcute inflammation Acute disseminated encephalomyelitis (ADEM) Optic neuritisAcute infection Abscess Cerebritis Discitis Empyema Encephalitis Meningitis Transverse myelitisNeurocutaneous disorders Congenital melanocytic naevus Neurofibromatosis type IITumours Benign and malignant Intracranial IntraspinalWhite matter disordersaVascular anomalies Cavernomas Developmental venous anomaliesVascular disorders Intraparenchymal haemorrhage Sturge-Weber syndrome VasculitisaAt presentation and at follow-up if necessary MRI protocols for specific areas The strategies applied to the brain are also applied to small parts such as the orbits and pituitary gland. The orbit is scanned using a STIR sequence to benefit from the fat saturation properties of the sequence and improve the conspicuity of lesions within the orbit. The slice thickness is reduced to 3 mm and, to increase the signal-to-noise, the matrix size is reduced, thereby slightly reducing the in-plane resolution. A heavily T2-W 3-D volume sequence of the petrous temporal bones is used to image the membranous labyrinth and is reconstructed in the coronal plane of the petrous temporal bone, and the axial and sagittal plane of the internal auditory meati (IAM). The sagittal plane is used to view the vestibulocochlear nerve in cross section. A constructive interference steady state sequence (CISS), available on Siemens scanners, is acquired and reconstructed at a slice thickness of 0.7 mm. The GE equivalent is the FIESTA sequence and the Philips equivalent is DRIVE. MRI protocols for specific neurological or neurosurgical presentations The majority of children undergoing an MRI brain scan will have epilepsy, stroke or a brain tumour and are often referred to specialists with an interest in these areas. Tumours of the brain and spine Brain Imaging of brain tumours is performed using our MRI brain with contrast enhancement protocol. We add contrast-enhanced T1-W imaging of the spine to establish the presence of any distant metastases in accordance with the United Kingdom Children’s Cancer Study Group (UKCCSG) guidelines for the imaging of brain tumours [11]. One of the commonest reasons for repeat scanning of a child with a brain tumour is the acquisition of contrast-enhanced imaging of the spine in a child who has had only a brain scan at the referring centre. The purpose of preoperative imaging of brain tumours is to assess tumour location and type, establish whether there are single or multiple lesions, define its relationship with vital structures and look for any complications such as hydrocephalus. Contrast-enhanced T1-W volumetric acquisitions are also acquired for image-guided surgery. Imaging of the spine is performed in all children with intracranial tumours (not only posterior fossa tumours) and although this requires a change of coil and an additional sequence, we find it beneficial as tumour histology is often not certain at the time of the initial imaging. Some tumours unexpectedly metastasize to the spine and assessment of the postoperative spine is made more difficult in the presence of blood products. Postoperative MRI is acquired using our MRI brain with contrast enhancement protocol to assess the degree of tumour clearance. We aim to do this within 72 h of surgery to reduce enhancement due to postoperative changes that can complicate the assessment of tumour clearance, in accordance with guidelines proposed by the UKCCSG [11]. If preoperative spinal imaging has not been acquired, the spine is imaged with both unenhanced and contrast-enhanced T1-W imaging to enable differentiation between haematoma and tumour metastases. Surveillance imaging (SI) during adjuvant treatment and follow-up is a major part of the management of childhood tumours and scanning frequency has important implications to anaesthetic and radiology service provision. In children with ependymoma and medulloblastoma, studies at GOSH have shown that SI reveals a substantial number of asymptomatic recurrences and survival appears to be improved in these patients compared with those patients in whom the recurrence is identified by symptoms [12, 13]. Benefits of SI following incomplete resection of posterior fossa low-grade pilocytic astrocytoma have also been shown [14, 15]. SI schedules have been written based on this work (Table 2). Table 2Proposed protocols for surveillance imaging in children with ependymoma, medulloblastoma and pilocytic astrocytomaTumourTiming of postoperative imagingFrequency of studyMacroscopically complete excisionIncomplete excisionCranial studySpinal studyCranial studySpinal studyEpendymoma24–48 hFirstNoneFirstNone1st year3–6 months6 months3 months3–6 months2nd–5th years6 months6–12 months6 months6–12 monthsMedulloblastomaa24–48 hFirstNoneFirstNone1st year3–4 months3–4 months3–4 months3–4 months2nd–6th year6–8 months6–8 months6–8 months6–8 monthsCerebellar low-grade astrocytomab24–48 hFirstNoneFirstNone1st year6 monthsNone6 monthsNone2nd yearAt 24 monthsNone6 monthsNone3rd yearAt 3.5 yearsNone6 monthsNone4–5th yearAt 5 yearsNone1 yearNone6th year onwardsNoneNone2 yearsNoneaAll imaging studies in children with medulloblastoma should include the entire neuroaxis.bThe surveillance imaging protocol following complete resection should be applied to children who have radiotherapy with the postoperative scan omitted Spine Imaging of spinal tumours is performed using our MRI spine with contrast enhancement protocol (Table 3) with dedicated axial imaging through the tumour. We add a standard brain scan with contrast enhancement to the initial or immediate postoperative scan to stage the tumour. If the brain scan is normal, further brain imaging is not acquired as part of SI and acquired only when clinically indicated (Table 4). Table 3Protocols for specific areasSpecific areasProtocolsOrbits (3-mm slices) Coronal and axial dual-echo STIR Coronal and axial T1-W spin-echoOrbits with contrast enhancement (3-mm slices) Coronal dual-echo STIR Coronal and axial T1-W spin-echo Contrast-enhanced coronal and axial T1-W images with fat saturationPituitary (3-mm slices) Sagittal and coronal T1-W spin-echo Coronal T2-W spin-echoPituitary with contrast enhancement (3-mm slices) Pituitary protocol Contrast-enhanced coronal and sagittal T1-W spin-echoInternal auditory meati 3-D volume axial CISS Brain MRIFace and neck MRI Coronal and axial dual-echo STIR Coronal and axial T1-W spin-echo Fat-saturated contrast-enhanced coronal and axial T1-W spin-echoMidline facial lesions Axial dual-echo STIR from floor of anterior cranial fossa to hard palate Sagittal T1-W and T2-W spin-echo (3-mm slices) Coronal T1-W spin-echo from nose to brainstemTable 4Standard MRI brain and spine protocolsTypes of brain and spine MRIProtocolsMRI brain (under 2 years old) Axial and coronal dual echo STIR Coronal and sagittal T1-W spin-echo DWI in three planes and calculated ADC map Axial “susceptibility-weighted” GE sequenceaMRI brain (over 2 years old) Axial T2-W fast spin-echo Coronal FLAIR Coronal and sagittal T1-W spin-echo DWI in three planes and calculated ADC map Axial “susceptibility-weighted” GE sequenceaMRI brain with contrast enhancement (under 2 years old) Axial and coronal dual-echo STIR Coronal T1-W spin-echo DWI in three planes and calculated ADC map Contrast-enhanced axial, coronal and sagittal T1-W spin-echo with magnetization transfer Axial “susceptibility-weighted” GE sequenceaMRI brain with contrast enhancement (over 2 years old) Axial T2-W fast spin-echo Coronal FLAIR Coronal T1-W spin-echo DWI in three planes and calculated ADC map Contrast-enhanced axial, coronal and sagittal T1-W spin-echo with magnetization transfer Axial “susceptibility-weighted” GE sequenceaMRI spine Sagittal T1-W and T2-W fast spin-echo Axial T1-W and T2-W fast spin-echo through target area and conus (Coronal T1-W spin-echo for scoliosis, if patient compliant)MRI spine with contrast enhancement Sagittal T1-W and T2-W fast spin-echo Contrast-enhanced sagittal T1-W fast spin-echo Axial T1-W spin-echo through target areaaOptional sequence Epilepsy UK guidelines for the imaging of epilepsy include children who have either multiple or focal seizures (NICE guidelines) [16]. In children with epilepsy we are aiming to detect focal cortical abnormalities; the majority are extratemporal and in a smaller proportion they are mesial temporal in origin. The majority of cortical abnormalities are easily diagnosed using conventional brain imaging and our standard brain protocol is used to investigate these children. The acquisition of a coronal T2-W sequence (either STIR, T2-W or FLAIR) allows the visualization of the mesial temporal lobe, and is a sequence often not acquired in referring centres. There are a group of children with intractable epilepsy, i.e. seizures not controlled by three drugs, who are considered for epilepsy surgery. In this subset of children a more rigorous epilepsy protocol is carried out that includes a 3-D volume T1-W acquisition and hippocampal T2 relaxometry (Table 5). To identify mesial temporal abnormalities the coronal studies are planned on a parasagittal scout image to lie perpendicular to the hippocampi. The 3-D, T1-W gradient-echo dataset is acquired isometrically in the sagittal plane (this saves time) and reconstructed in all three planes with the axial and coronal images tilted to follow the plane of the hippocampus. More recently a 3-D FLAIR sequence reconstructed along the same lines has been added. A sequence is acquired that measures the true T2 values of the hippocampi (T2 relaxometry), which is only of value in suspected mesial temporal sclerosis (MTS). It has been shown that T2 values are more sensitive to the presence of MTS than visual inspection and T2 relaxometry is of particular value in bilateral disease [17]. Normal values will vary between scanners, so normative data should be acquired. The radiographers carry out the postprocessing of the 3-D dataset and the T2 relaxometry. Table 5Protocols for particular clinical indicationsTypes of clinical indicationProtocolsBrain tumours MRI brain with contrast enhancement Contrast-enhanced sagittal T1-W images of whole spine Contrast-enhanced image-guided images when requiredStrokeAcute Axial T2-W fast spin-echoa Coronal FLAIRb Sagittal T1-W spin-echoDWI in three planes and calculated ADC mapIntracerebral 3-D TOF MRAAxial dual-echo STIR and T1-W spin-echo through the neckExtracerebral 2-D TOF MRA of the neck down to the aortic rootNon-acuteAcute stroke protocol without imaging of the neckEpilepsy Axial T2-W fast spin-echoa 3-D T1-W volume acquisition reconstructed in three planes Coronal T2-W fast spin-echob Coronal FLAIR (or 3-D FLAIR if available) Hippocampal T2-relaxometry (see text)Intraparenchymal haemorrhage MRI brain with contrast enhancement Intracerebral 3-D TOF MRA MRVNon-accidental head injury Standard MRI brain Axial GE “susceptibility-weighted” sequence Sagittal T2-W spin-echo and GE “susceptibility-weighted” sequence of the cervical spineMPRAGE magnetization prepared rapid acquisition gradient echo, MRV magnetic resonance venographyaReplaced by dual echo STIR sequence in the under 2 age groupbReplaced by a dual echo STIR coronal in the under 2 year age group. Non-traumatic intraparenchymal haemorrhage The aim of imaging children with intraparenchymal haemorrhage (IPH) is to detect the underlying cause and we advocate the use of CT angiography as a matter of urgency in those children who have a depressed level of consciousness or whose clinical condition is deteriorating. MRI is used to investigate clinically stable children, and we perform brain MRI with contrast enhancement and a 3-D, time-of-flight (TOF) intracranial MR angiography (MRA). In our experience nearly two-thirds of children presenting with nontraumatic IPH have an underlying cause diagnosed by MRI and MRA, with aetiologies including tumour, arteriovenous malformation (AVM), venous sinus thrombosis and cavernous angioma [18]. An attempt is made to perform MRI within 72 h of the ictus to overcome the problem of the visibility of blood products (T1 shortening due to methaemoglobin) on 3-D TOF sequences that may obscure the underlying pathology. The value of intravenous contrast agents in identifying underlying vascular causes has not been proven. We have a largely anecdotal view that it is helpful in identifying small AVMs and venous anomalies. Stroke The aims of conventional MRI are not only to detect the infarct, but also to provide information to establish the cause of the stroke and to exclude other causes (such as tumour or infection). The majority of children with stroke have a combination of risk factors including sickle cell disease, congenital heart disease, anaemia, prothrombotic disorders and infections such as varicella-zoster [19]. The rationale for the inclusion of intracranial MRA in all children with stroke is that cerebral arteriopathy is found in up to 80% of these children and most commonly affects focal areas of large intracranial arteries [19]. Specific entities such as moyamoya disease may also be diagnosed. The commonest abnormality identified is occlusion or stenosis, of unknown aetiology, affecting the terminal internal carotid artery (ICA) or proximal middle cerebral artery (MCA). Extracranial MRA and, either a dual-echo STIR or fat-saturated axial T1-W imaging, through the neck can detect arterial dissection, particularly in children with posterior circulation infarcts [20]. The inversion time of the STIR sequence is selected to suppress the fat within the neck. The fat saturation provided by both these sequences improves the conspicuity of the haematoma within the extracranial vessel wall. A 3-D TOF MRA sequence is used for the imaging of the intracranial vessels and a 2-D TOF MRA sequence for the extracranial vessels. The TOF scan times are shorter than phase-contrast (PC) MRA and there is lack of dependence on the choice of correct velocity encoding with obvious advantages when scanning ill children. Intracranial 3-D TOF MRA is also included in the investigation of children with IPH, although its sensitivity to T1 shortening may obscure the underlying abnormality (see above). As CT is often the first-line investigation in children with stroke, the potential of DWI to detect hyperacute cerebral infarction prior to changes on T2-W MRI is not realized. However, inpatients (e.g. cardiac patients, patients with recent-onset stroke) can be imaged early, and in these children DWI can be used to detect infarcts of different ages. Non-accidental head injury Although CT is the primary imaging tool in the evaluation of nonaccidental head injury (NAHI), we perform MRI in children with an abnormal CT or neurological symptoms and signs on day 3–4 in accordance with the protocol proposed by Jaspan et al. [21]. The rationale for performing MRI is that it is more sensitive than CT to both hypoxic–ischaemic injury, particularly when DWI is used, and subdural haematomas adjacent to the calvarium or in the posterior fossa. MRI is repeated at 2–3 months in those children with parenchymal brain injury or a persistent neurological abnormality. Children with suspected NAHI are imaged using our standard brain protocol. In addition axial GE (susceptibility-weighted) sequences are acquired, as they are more sensitive to blood products than standard sequences. Sagittal T2-W and GE images are acquired of the cervical cord, including the brainstem, particularly to identify shear injuries within the brainstem that have been proposed as a mechanism of hypoxic–ischaemic injury in children who have been shaken as the mechanism of injury [22, 23]. The aim of the sagittal T2-W sequence of the cervical cord is to look for any evidence of trauma as a result of violent shaking in an attempt to establish the true, albeit rare, incidence of pathology in the upper cervical spine [21]. Neonatal imaging Conventional imaging can detect patterns of regional brain injury in the neonatal period that can help time the injury, determine underlying mechanisms and ultimately provide some prognostic information. MRI is useful in detecting hypoxic–ischaemic injury, and germinal matrix and intraventricular haemorrhage, and can be useful to distinguish other pathologies that may mimic hypoxic–ischaemic encephalopathy in the neonatal period such as venous infarction, metabolic disease, infection, and congenital developmental abnormalities. Our neonatal imaging protocol is the same as the under-2-year brain protocol. Achieving high signal-to-noise is of utmost importance given the small size of the infant head and is improved by decreasing the slice thickness to 4 mm whilst decreasing the matrix size from 512 × 512 to 256 × 256 in a field-of-view of 180 mm. Although specific neonatal head coils have been developed, most nonpaediatric centres are unlikely to have access to them and improved results can be obtained by using an adult knee coil. Developmental delay Developmental delay, without other clinical features such as epilepsy or dysmorphic features, is not considered an indication for an MRI scan at our institution. This decision was made in conjunction with our neurologists as it was recognized that there is an extremely low yield of clinically relevant abnormalities seen in patients with developmental delay alone. The future The future of paediatric neuroimaging lies in the incorporation of research techniques, such as perfusion imaging and diffusion tensor imaging, into standard imaging protocols once they have been demonstrated to be of value within the clinical arena. The use of higher field strength magnets will have the advantage of increasing signal-to-noise and reducing scanning time, which will be of benefit to the child who is both small and liable to move.	[ "protocols", "mri", "children" ]	[ "P", "P", "P" ]
Eur_Spine_J-2-2-1602198	Vascular complications of prosthetic inter-vertebral discs	Five consecutive cases of prosthetic inter-vertebral disc displacement with severe vascular complications on revisional surgery are described. The objective of this case report is to warn spinal surgeons that major vascular complications are likely with anterior displacement of inter-vertebral discs. We have not been able to find a previous report on vascular complications associated with anterior displacement of prosthetic inter-vertebral discs. In all five patients the prosthetic disc had eroded into the bifurcation of the inferior vena cava and the left common iliac vein. In three cases the aortic bifurcation was also involved. The fibrosis was so severe that dissecting out the arteries and veins to provide access to the relevant disc proved impossible. Formal division of the left common iliac vein and artery with subsequent repair was our solution. Anterior inter-vertebral disc displacement was associated with severe vascular injury. Preventing anterior disc displacement is essential in disc design. In the event of anterior displacement, disc removal should be planned with a Vascular Surgeon. Key points Anterior prosthetic disc displacement frequently involves major arteries and veins. Removal of an L45 disc prosthesis is associated with vascular injury. A Vascular Surgeon should be involved in disc removal. Introduction Replacement of the inter-vertebral disc is designed to relieve symptomatic back pain secondary to disc degeneration by preserving segmental function. A successful artificial disc replacement would avoid the significant long-term problems associated with spinal fusion [9]. We report five consecutive cases of anterior displacement of prosthetic inter-vertebral discs resulting in erosion into the bifurcation of the vena cava and aorta, from a series of only 13 patients. One patient suffered recurrent back pain. These patients had no preoperative vascular symptoms to warn of this vascular involvement and all of these discs were removed in order to prevent further vascular injury. The complications associated with removal are described together with our suggestions on an approach to this challenging surgical problem. Prosthetic disc replacements (AcroFlex disc, DePuy Acromed, Raynham, MA, USA) were carried out as part of a clinical trial with ethical approval, regulatory approval and with informed patient consent. All patients had incapacitating back pain. Leg pain was not an exclusion criteria unless directly attributable to a prolapsed inter-vertebral disc. All patients in this study had magnetic resonance imaging evidence of degenerative disc disease. All patients had an Oswestry disability index (ODI) greater than 30% and all had discographically established concordant pain at the target level/s. Inclusion criteria, the follow-up protocol and disc design for this study have been described previously [5]. Case 1 A 33-year-old woman presented with a 5-year history of low back and right leg pain. A two level artificial inter-vertebral disc replacement was performed at L45 and L5S1 without complication through an anterior transperitoneal approach (L45 31.5 × 43 × 9 × 5° AcroFlex disc and L5S1 34 × 46 × 9 × 10° AcroFlex disc) (Fig. 1a–c). Fig. 1Postoperative X-ray with satisfactory placement of L45 and L5S1 AcroFlex disc, lateral view (a) AP view (b and c) with anterior disc displacement at 6 months (d) The patient made an uneventful recovery with improved symptoms at 6 weeks. However on X-ray the prosthesis at L45 had moved 5 mm anteriorly. At 6 months there was further L45 prosthesis displacement; however, mobility was preserved with full lumber flexion and extension at video fluoroscopy (Fig. 1d). The L45 artificial disc was replaced at 15 months with a size 2 SB Charité III disc containing an 8.5-mm spacer (Waldemar Link GmbH & Co., Hamburg, Germany). At surgery the disc had stretched and eroded into the left common iliac vein. The distal vein was occluded with thrombus. An injury to the medial wall of the left common iliac artery was repaired with 5/0 Prolene and the occluded vein was oversewn with 4/0 Prolene. This provided access and permitted removal of the disc. Postoperative recovery was complicated by a paralytic ileus only. The patient has an L5S1 prosthesis which remains in an acceptable position. She has subsequently become pregnant and had a normal delivery without the need for Caesarian section. Case 2 A 40-year-old woman presented with a 3-year history of lumber back pain. She underwent L45 disc replacement with a 29 × 40 × 9 × 10° AcroFlex disc. This surgery was complicated only by a temporary loss of bladder filling sensation. At 6 week follow-up her back and leg symptoms were significantly improved. Unfortunately at 6 months her back pain had returned and she was now suffering pain and tingling in the left L5 dermatome. On X-ray the prosthesis had moved anteriorly. At 1 year the patient agreed to have the prosthesis removed. At operation the prosthesis was eroding into the left common iliac artery and vein. Both vessels were mobilised with difficulty due to dense scar tissue, the inevitable venous and arterial injuries were repaired with 4/0 Prolene. The disc was replaced with a Stabilis cage (Stryker, UK) containing artificial bone graft which was secured with a 3.2 × 34 mm screw and washer. Postoperatively the patient wore class II compression hosiery and was prescribed 40 mg Clexane twice daily. Unfortunately, despite this prophylaxis, she suffered left common iliac artery and iliofemoral deep vein thrombosis. Full anticoagulation was achieved with Tinzaparin and subsequently Warfarin. Swelling is controlled by class II compression hosiery. Transluminal angioplasty has been performed for a stenosis of the left common iliac artery at the site of repair. Her claudication has improved and is now thought to be venous. Case 3 A 51-year-old man presented with bilateral leg pain and no motor or sensory loss. He had undergone laminectomy 12 years previously for a disc prolapse. A prosthetic inter-vertebral disc (29 × 40 × 9 × 10° AcroFlex disc) was placed in the L45 inter-vertebral space, without complication, using an anterior transperitoneal approach. The postoperative X-ray showed the prosthesis to be in an acceptable position, i.e. centrally placed in the disc space on antero-posterior projection and posteriorly placed on the lateral view (Fig. 2a, b). The patient’s symptoms were improved but the prosthesis had migrated 2 mm anteriorly at 6 weeks and 5 mm by 12 months (Fig. 2c). Fig. 2Satisfactory placement of L45 AcroFlex disc, lateral view (a) and AP view (b) with anterior displacement at 12 months (c) and subsequent spinal stabilisation with plate and pedicle screws (d) At operation the disc had migrated into the left common iliac vein posterior to the aortic bifurcation. A tear was made in the left iliac vein while mobilising these vessels. The left common iliac vein was divided transversely to facilitate removal of the disc. A 14-mm carbon fibre cage filled with iliac crest bone graft was inserted and the iliac vein re-anastomosed with 4/0 Prolene. Postoperative duplex scanning showed the left common iliac vein and artery to be patent. The spine was stabilised with a flat plate and pedicle screws 12 days later (Fig. 2d). Case 4 A 47-year-old man presented with lumbar back pain radiating to his thigh and buttock. He underwent L45 and L5S1 disc replacement with a 31.5 × 43 × 11° AcroFlex disc and a 31.5 × 43 × 9° AcroFlex disc respectively without postoperative complication. At 1 month follow-up he was well with no pain. Anterior displacement of the L45 disc was first seen on plain X-ray at 2 years and the disc was removed 11 months later. Having learned from previous bitter experiences, the vascular surgeon formally divided the left common iliac artery and vein so that there was good access to remove the disc and replace it with a Stabilis cage. The left common iliac artery and vein were then re-anastomosed. Despite this approach the right common iliac vein was injured and required repair. Full deep vein thrombosis prophylaxis was undertaken with pneumatic compression in theatre, peri-operative low molecular weight Heparin, postoperative compression stockings in addition to anticoagulation with Warfarin. Both iliac vessels were widely patent with no thrombus and normal blood flow velocity on duplex imaging at 1 month. Postoperatively he has continued to suffer pain after prolonged standing with occasional sleep disturbance. The L5S1 disc was stable on X-ray at 6 months with fusion progressing between L45. Case 5 A 44-year-old man who had long standing low back pain had a single level degenerate disc at L5S1. The L5S1 disc was replaced with a 29 × 40 × 10° 11 mm AcroFlex disc. Immediately postoperatively he was well with an improvement in his symptoms. Gradual displacement of the prosthesis occurred from 3 months. The prosthesis was removed at 19 months and fusion performed using a Stabilis cage and cancellous bone graft. During surgery a tear of the left common iliac vein occurred during mobilisation this was repaired with 2.0 Prolene. Formal division of the iliac vein and artery was not necessary as the disc was removed from L5S1. Preoperatively the ODI was 48% after fusion it had fallen to 10%. Discussion The five patients reported here gained early symptomatic relief from low back and leg pain with an artificial inter-vertebral disc. However, in each case, revision following surgery for benign disc degeneration was associated with a major vascular injury. Anterior displacement of the L45 disc caused pressure on the left common iliac vein and artery with stretching of the vessels and erosion into the vessel wall in three cases. It was impossible to mobilise the stretched vessels from the firmly adherent surrounding dense scar tissue without vascular injury. In one case the left common iliac vein was thrombosed while in the others repairs were necessary after almost impossible mobilisation of the vessels to expose the prosthesis. Damage to the left common iliac vein and artery in one case led to both venous and arterial thrombosis. It is likely that this complication will lead to longstanding symptoms of venous insufficiency. The only safe solution for adequate exposure of an anteriorly displaced visco-elastic artificial L45 disc is the formal division and subsequent re-anastomosis of the left common iliac vein and if necessary the left iliac artery; this strategy was successfully employed in cases 3 and 4. An anterior transperitoneal approach is essential to allow proximal and distal vessel control before division. A retroperitoneal approach in this type of revision surgery would be extremely hazardous and potentially life threatening due to an inability to control the right common iliac vein or vena cava when there is bleeding from the left common iliac vein. The aortic bifurcation would also be difficult to control. The Link SB Charité disc is currently the most widely implanted inter-vertebral disc. This articulating disc consists of an ultra high molecular weight polyethylene spacer between two separate cobalt–chromium–molybdenum alloy plates. Griffith et al. reported displacement or dislocation of the prostheses in 6 of 139 implants with an average follow-up of 11 months. There were six reported vein injuries although it is not clear whether these injuries were associated with displacement or revision surgery or at what level the implant was placed [6]. Cinotti et al. reported only one anterior displacement of 46 SB Charité III prostheses and no vascular injury, with a minimum follow-up of 2 years [3]. In a series of 105 patients with 154 SB Charité III discs five vascular problems were reported (two phlebitis, two pulmonary embolism and one acute leg ischaemia) although no information on the rate of revision or anterior displacement was presented in this series [10]. Zeegers et al. reported a prospective study of 50 patients where 75 SB Charité III prosthesis were inserted. None of 38 patients with full 2 year radiographic follow-up had migration of the prosthesis. One patient lost to follow-up had malposition and some slip but was lost to follow-up. Of 12 patients who underwent reoperation only one aortic injury was reported during failed disc re-positioning and subsequent fusion [12]. These reports are all non-randomised series without any control and in one case retrospective. The only published randomised study of inter-vertebral disc replacement compares the Charité disc with spinal fusion. This study showed that pain and disability scores are equivalent at 24 months follow-up. There was a 4.4% incidence of venous injury in the arthroplasty group and device failure in 5.1% compared to 9.1 and 2% respectively in the spinal fusion group [2]. This first design of the AcroFlex disc contained a hexane based polyolefin rubber core vulcanised between two titanium plates. A retrospective review of six patients who had AcroFlex discs, with a mean follow-up was 3.5 years, reported a poor outcome in two patients with fracture of the rubber component and persistent symptoms in a second. There were no reported cases of disc displacement or of vascular injury during one revision although the prosthesis was at L5S1 [4]. Fraser et al. reported a series of 28 patients who had the third generation AcroFlex disc implanted, 17 patients had the same disc design used in this study. Eight of 28 required revision surgery for deterioration of symptoms and failure of the prosthesis. Two of the four patients who had removal of the prosthesis suffered left iliac vein injury during surgery. Fraser et al. inserted the discs through a retroperitoneal approach and suffered one anterior displacement in this case series [5]. It should be noted that there were two pilot studies. In the first series all patients were implanted in Adelaide. Displacement occurred in series 2 at both test centres but only in Manchester was displacement severe enough to warrant removal for the reasons stated. Disc displacement due to placement of the disc too far anteriorly has been reported with the PRODISC [1]. The high incidence of disc displacement in this series (5 of 13) is thought to be associated with an anterior transperitoneal approach and more extensive removal of the annulus rather than malposition. Design is also an important factor in preventing displacement since it occurred in both centres. On revision surgery no osteointegration of these implants was apparent and displacement occurred despite re-design of the endplates, as a result of the series 1 failure, to a domed surface with 4–6 tapered fins on each surface. This redesign was also inadequate. Van Ooij has described a series of 27 patients with anterior displacement of four Charité discs one of which led to compression of the iliac artery causing symptoms of ischaemia [11]. A systematic review of total disc replacement examined the results from 564 arthroplasties in 411 patients. The majority of the cases considered were the Charité disc and only eight cases of vascular injury were identified [8]. The authors of both papers concluded that removal of artificial discs is potentially dangerous and no reliable strategy for revision surgery exists for a procedure without proven long-term reliability. The five cases described in this report highlight failure of three design criteria for a disc prosthesis: (i) the rapid fixation to bone, (ii) fail-safe design to prevent damage to surrounding structure in case of failure and (iii) medium-term osteointegration [7]. If a disc does fail and needs to be removed, surgery can be hazardous with associated vascular injury [8]. Consequently, we advise that a Vascular Surgeon should be available to try and mobilise or divide and re-anastomose the left common iliac vessels when needed for adequate exposure of a L45 disc. Furthermore, this revision disc surgery is usually difficult and should be planned to include autologous blood transfusion.	[ "vascular injury", "disc design", "artificial inter-vertebral disc", "inter-vertebral disc replacement", "inter-vertebral arthroplasty" ]	[ "P", "P", "P", "P", "R" ]
Evid_Based_Complement_Alternat_Med-4-1-1810360	Randomized Controlled Trials of Pediatric Massage: A Review	The existing reviews of massage therapy (MT) research are either limited to infants, adults, or were conducted prior to the publication of the most recent studies using pediatric samples. Randomized controlled trials (RCTs) of pediatric MT are reviewed. A literature search yielded 24 RCTs of pediatric MT, defined as the manual manipulation of soft tissue intended to promote health and well-being in recipients between 2 and 19 years of age. Because RCTs of pediatric MT varied considerably in the amount and types of data reported, quantitative and narrative review methods were both used. Single-dose and multiple-dose effects were examined separately. Among single-dose effects, significant reductions of state anxiety were observed at the first session (g = 0.59, P < 0.05) and the last session (g = 1.10, P < 0.01) of a course of treatment. Effects for salivary cortisol (g = 0.28), negative mood (g = 0.52) and behavior (g = 0.37) were non-significant. Three of eleven multiple-dose effects were statistically significant. These were trait anxiety (g = 0.94, P < 0.05), muscle tone (g = 0.90, P < 0.01) and arthritis pain (g = 1.33, P < 0.01). Results of studies not permitting effect size calculation were judged to be generally consistent with quantitative results. MT benefits pediatric recipients, though not as universally as sometimes reported. Numerous weaknesses endemic to MT research (e.g. low statistical power, frequent failure to report basic descriptive statistics) are identified, and recommendations for future pediatric MT research are discussed. Introduction Background Massage therapy (MT) is one of the most widely used complementary and alternative medicine (CAM) therapies in the United States according to National Health Interview Survey data. It is estimated that consumers spend between 2 and 4 billion dollars on 75 million visits to massage therapists annually (1). Studies indicate that parents are making increasing use of CAM therapies, including MT, for their children. One study found that 33% of parents reported using CAM for their child within the past year, with MT being one of the most popular therapies (2). Another study shows that families of children with special health care needs are almost twice as likely to have used CAM for their child (3). CAM use is now prevalent, even in many traditional medical settings [e.g. 49% of university-affiliated pain management centers in the US and Canada offer MT (4)], yet pediatricians and other health care professionals are often not informed about the CAM therapies that are being used by their patients (2). This may be especially true for pediatric CAM, where survey results indicate that 81% of parents currently using CAM for their child wanted to discuss it with their pediatrician, but only 36% did (5). For CAM to truly be integrated into the health care system, it needs to be openly discussed and recognized for its value, particularly in the area of palliative care (6). A scientific understanding of CAM therapies, such as pediatric MT, will permit a greater understanding of the value of this type of therapy. Numerous studies of MT for children and young adults have been conducted. These include randomized controlled trials (RCTs), evaluation studies, descriptive case studies and reviews. However, the existing reviews of MT research are either limited to adult participants, limited to infants, were conducted prior to the publication of the most recent studies using pediatric samples, or neglected to quantify results. Ottenbacher and colleagues (7) quantitative review of tactile stimulation for infants and young children, published nearly 20 years ago, found statistically significant beneficial outcomes for five of the six categories examined (these were motor/reflex, cognitive/language, social/personal, physiological and overall development; visual/auditory was non-significant). Field (8) conducted a narrative review, including research on persons of all ages, and concluded that MT had a multitude of condition-specific effects in addition to consistently providing reductions of anxiety, depression and levels of stress hormones (most notably cortisol) that were observed across studies. The two most recent reviews are quantitative reviews that focus specifically on either infants or adults. A Cochrane review systematically examines MT for preterm and low birth-weight infants, and reaches the negative conclusion that ‘there is insufficient evidence of effectiveness to warrant wider use of preterm infant massage’ (9). The quantitative review of MT performed on adults confirmed (and quantified) some of Field's conclusions while disputing others (10). Substantial reductions resulting from multiple sessions of MT were noted for depression (g = 0.62, P < 0.01) and trait anxiety (g = 0.75, P < 0.01), while MT's effect on cortisol levels was not statistically significant (g = 0.14). While this review supported the value of MT by quantifying its ability to substantially reduce symptoms of psychological distress in adults, taken together this set of findings called into question some of the theories most often invoked to explain the benefits of MT, especially the theory that MT benefits recipients primarily by activating the parasympathetic nervous system (10). Current Review While the latest quantitative reviews increase our understanding of MT and indicate new directions for research, they neglect MT performed on pediatric samples. The current review examines MT's effects in pediatric samples (defined here as studies where the mean age of participants was between 2 and 19 years), with a focus on RCTs. While some have argued against using RCTs to understand CAM modalities such as MT (11), RCTs represent the state-of-the-art for establishing cause and effect relationships in treatment research because they are the most effective study design for ruling out alternate (i.e. non-treatment) explanations for observed effects, including spontaneous recovery, placebo effects and statistical regression (12). Though there have been some new pediatric MT studies published since the reviews by Ottenbacher et al. (7) and Field (8), the number of RCTs that examine pediatric MT is still quite small. In addition, a weakness of existing MT research is that, more often than not, studies in this area do not include the minimal statistical detail necessary to calculate effect sizes. For these reasons, a full meta-analytic treatment of MT research with pediatric samples is not possible. However, consistent with Rosenthal's observation that a narrative literature review can only be improved by the addition of a simple, descriptive quantitative analysis (13), we quantified results where possible, and include these results in combination with a narrative review of other RCTs. In some cases, the present findings are being compared with analogous findings in the adult literature. When these comparisons are made, we are referring to the adult meta-analysis conducted by Moyer et al., unless otherwise stated (10). Operational Definition A notable challenge in reviewing MT studies is that there are many forms of MT in practice. The American Massage Therapy Association (AMTA) defines massage as ‘manual soft tissue manipulation, [including] holding, causing movement, and/or applying pressure to the body’ (14). As written, this very broad definition includes numerous MT approaches commonly used in clinical practice that are relevant to the current review, but could also include rare forms of medical massage (e.g. optic nerve massage (15), light compressive massage for congenital dacryocystocele (16), cardiac massage (17)), that are outside the intended scope of this review. For this reason, we focus on forms of MT that are consistent with traditional Swedish styles of massage. Swedish massage uses five main strokes to stimulate the circulation of blood through the body; petrissage (kneading), effleurage (stroking), friction, tapotement (tapping) and vibration. For the purposes of this review, MT is typified by the manual manipulation of soft tissue, performed by a person other than the recipient, intended to promote health and well-being. This operational definition allows a range of MT styles to be included in this review. Studies vary on many details, including the amount of clothing worn by recipients, whether a massage chair or massage table was used, whether MT took place in a clinical setting or at home, and whether MT was performed by a person with full, partial, or no training as a massage therapist. Studies also vary in which anatomical regions are massaged. Despite all these variations, it is reasonable to expect that there will be some consistent outcomes that result from MT. Eventually, as a scientific understanding of MT grows, studies that examine the importance of these variations will be advisable, but currently the questions of greatest interest are at a more fundamental level. Types of Effects MT effects can be divided into single-dose and multiple-dose. Single-dose effects include MT's influence on psychological or physiological states that are transient in nature and that might reasonably be expected to be influenced by a single session of MT. Multiple-dose effects are restricted to MT's influence on variables that are considered to be more enduring, or that would likely be influenced only by a series of MT sessions performed over a period of time, as opposed to a single dose. Frequently, both single- and multiple-dose effects are examined in the same study. One example is a study of MT for autistic children that examined the single-dose effect of MT on salivary cortisol (immediately prior to, and immediately following, an individual session of MT) and the multiple-dose effect of MT on depression (at the beginning of, and at the conclusion of, a sequence of MT sessions over time) (18). A second example is a study that evaluated children's distress during burn treatment, which included the single-dose effect of MT for state anxiety and the multiple-dose effect of MT for depression (19). Typically, studies include the terms ‘short-term effect’ and ‘long-term effect’ to indicate single- and multiple-dose effects, respectively. Our decision to use the single-dose and multiple-dose terminology is motivated by the desire to prevent any confusion that may arise related to how long an effect may last following the termination of treatment. None of the studies in the current review examine whether any MT effects last beyond the final day on which a participant receives treatment, making the use of the term ‘long-term effect’ potentially misleading. The potential benefits of MT can be further classified according to whether they are primarily affective, physiological or behavioral in nature. Affective refers to effects most closely associated with the recipients' feelings and emotions. Physiological effects are those concerned with recipients' vital organismic processes. Behavioral effects are those related to the recipients' observable responses to their environment. Study results reviewed here will first be separated by the single-dose versus multiple-dose distinction, then further categorized into affective, physiological and behavioral dimensions. Methods Literature Search A literature search was conducted by the first author (S.B.) using the keywords massage, child and pediatric to search the MEDLINE, LexisNexis, CINAHL and PsycInfo databases. We checked the NIH CRISP Database to search for other publicly funded studies currently in progress. MT researchers were also contacted to obtain studies that were unpublished, in press or otherwise not found by means of database searches. With the introduction of the first Massage Therapy Research Conference in Albuquerque, New Mexico (43), leaders in the field of massage research were readily contacted. We used this opportunity to access current unpublished pediatric massage studies. Studies obtained by these methods were inspected to ensure that they examined a form of MT consistent with the present study's operational definition of MT performed on a pediatric sample. Application of these criteria yielded 24 RCTs of MT with a pediatric sample. These studies, along with important details, are listed in Table 1. Statistical Analysis For studies that provided sufficient data, between-groups comparisons on variables of interest were converted to Hedges' g effect size by the second author (C.A.M.). Hedges' g, calculated as (Group mean 1 − Group mean 2)/Pooled SD, estimates the number of standard deviations that the average member of a treatment group differs from the average member of a comparison group for a given outcome. Hedges' g was selected over Cohen's d, a similar standardized mean difference effect size. This choice was made for two reasons. First, in some cases the original study data could only be converted to g. Second, using g makes the results of the current review as consistent as possible with the existing meta-analysis of MT effects for adults. In cases where a study employed more than one measure to examine the same outcome variable, results of multiple measures completed by participants or by blinded observers (but not those completed by non-blind observers) were standardized and then averaged, yielding one effect size per variable for each study. Similarly, if a study examined the immediate effects of more than one application of treatment, the results of the multiple applications or assessments were standardized and, when similar in magnitude, averaged in order to calculate a single effect size for that study. In the single case where these assessments clearly differed in a systematic way (i.e. state anxiety, where the effects of a final session were always larger than the effects of the initial session; paired sample t(3) = 4.46, one-tailed P < 0.02), separate effects were calculated for each timepoint. Effect sizes were coded such that positive values, for any variable, indicate a more desirable outcome (e.g. a reduction in anxiety) for the participants who received MT. Individual study effect sizes were subjected to a correction for small sample bias, then weighted by their inverse variance and averaged to generate a mean effect size for each outcome variable (20). All effect sizes were calculated according to a random effects model of error estimation. Statistical significance of the mean effect sizes was assessed by calculating the 95% confidence interval for the population parameter. A significance level of 0.05 or better is inferred when zero is not contained within the confidence interval. Results Of the 24 RCTs, only 9—accounting for a total of 200 participants—provided sufficient data for their results to be systematically quantified. Table 2 lists mean effect sizes for 16 outcome variables (g), as well as the number of studies contributing to each effect size (k), the total number of participants contributing to each effect size (N) and 95% confidence intervals. Five of these sixteen effect sizes were statistically significant. For the single-dose effects category, these included state anxiety at the first session (g = 0.59, P < 0.05) and at the last session (g = 1.10, P < 0.01) of a course of treatment. Effects for salivary cortisol (g = 0.28), negative mood (g = 0.52) and behavior (g = 0.37) were non-significant. Only 3 of the 11 multiple-dose effect sizes were statistically significant. These were trait anxiety (g = 0.94, P < 0.05), arthritis pain (g = 1.33, P < 0.01) and muscle tone (g = 0.90, P < 0.01). Of the 24 RCTs, 15 (accounting for 458 research participants) do not report sufficient data to permit effect size calculation, a frequent problem in MT research that makes objective interpretation of results difficult. Nevertheless, by judiciously comparing the scant data presented in this subset of studies with the objective data previously summarized, it should be possible to see if there are any dramatic contrasts among the findings. These interpretations, within the context of the more objective findings, appear in the results categories that follow. Single-Dose Effects Affective Dimension State anxiety. Field et al. have conducted several pediatric MT studies where anxiety is an outcome measure. These studies date back to 1992 when MT was applied to a group of pediatric psychiatric patients (21). Four studies with reportable effect sizes (No.'s 6, 14, 17 and 21 in Table 1), using a total of 81 participants, compared MT with either relaxation therapy or a reading comparison group with state anxiety—a momentary emotional reaction consisting of apprehension, tension, worry and heightened autonomic nervous system activity (22)—as a dependent variable. Three of the studies (No.'s 14, 17 and 21) used the State Trait Anxiety Inventory (two specify the child version). One study (No. 6) of MT for children with juvenile rheumatoid arthritis used a behavior observation of the child's anxiety level performed by a blinded rater. MT consistently reduced state anxiety in these studies. This single-dose effect was unique in that it is significantly larger at the second timepoint, so we examined these separately. Results of a first session of MT yield a statistically significant effect, g = 0.59 (95% CI = 0.15, 1.04). The effect at the last session is even greater, g = 1.10 (95% CI = 0.64, 1.57). These results are depicted graphically in Fig. 1. Possibly, the substantially larger effect occurring at the last session of treatment may be the result of participants' increasing comfort with MT (or with the massage therapist) over the course of time, or the effect may be related to MT's potential to reduce trait anxiety over a course of treatment (8). These pediatric results are consistent with the same effect found for adults, where this effect has been estimated as g = 0.37 (95% CI = 0.14, 0.59). Possibly, the larger effect for the pediatric samples examined here reflects a greater treatment aptitude for this population; however, with such wide confidence intervals, this is only speculation. Studies where state anxiety was a dependent variable but effect sizes could not be calculated include samples of children who experienced Hurricane Andrew (23), depressed adolescent mothers (24), and children and adolescents with a range of illnesses including diabetes (25), atopic dermatitis (26), asthma (27), bulimia (28) and leukemia (29). The most commonly used measures across these studies were the State Trait Anxiety Inventory for Children (STAIC) and the State Trait Anxiety Inventory (STAI) for adolescents. A study of children who experienced Hurricane Andrew assessed both the single-dose effects of state anxiety and the multiple-dose effects of trait anxiety. These studies are consistent with the previously reported effect sizes, in that MT appears to decrease state anxiety. The MT in all of these studies was conducted either daily or twice weekly over a treatment period of 30–45 days. MT sessions generally lasted between 20 and 30 min, the one exception being the leukemia study where MT sessions were 15 min in duration. Four of these studies trained parents to provide MT directly to their child and the remaining three either used trained massage therapists or massage students. Mood. Mood, which may be defined as ‘transient episodes of feeling or affect’ (30), has frequently been an outcome measure in pediatric MT studies. Study populations have included depressed adolescent mothers, and children and adolescents with cystic fibrosis, leukemia and bulimia. Two studies (No.'s 13 and 14 in Table 1), using a total of 50 participants, compared MT with either a reading or wait-list control to examine MT's effect on mood. One of these studies (No. 14) used the Profile of Mood States (POMS) (31) depressed mood subscale. The second study (No. 13) used a faces scale and a modification of the Children's Pain/Fear Thermometer Rating Scale. Taken together, these studies yield a non-significant effect, g = 0.52 (95% CI = −0.05, 1.10) that is consistent with the results found for adult recipients (g = 0.34, 95% CI = −0.08, 0.76). These positive but non-significant results suggest at least four possibilities. One is that the tools that have been used to measure mood in these studies are not precisely capturing overall mood, but are being affected by correlates of mood such as depression, pain or fear. A second possibility is that MT's modest effect on mood has not been examined with sufficient statistical power, resulting in wide confidence intervals. The third is that MT's positive effect on mood is affected by a moderator variable that has not yet been examined (e.g. recipient's comfort with MT; the existence of a therapeutic bond between the recipient and provider, etc.). Finally, the possibility that MT does not specifically have a positive effect on mood cannot be definitively ruled out. Physiological Dimension Salivary cortisol. Two studies (No.'s 6 and 13 in Table 1), using a total of 50 participants, compared MT with either relaxation training or a wait-list control to examine MT's single-dose effect on cortisol—a stress hormone associated with activation of the sympathetic nervous system in response to certain kinds of stressors. In both studies, salivary cortisol [but not urinary cortisol, which less accurately captures short-term stress responses (32) most likely to respond to single-dose MT] was sampled pre- and post-MT session, with a delay of 20–30 min after the session, because salivary cortisol samples reflect responses to stimulation occurring ∼20 min prior to collection. Both studies also took into account the diurnal cortisol cycle, characterized by an increase in secretory activity following awakening and a declining trend over the course of the day (33). Combined, these studies yield a non-significant reduction, g = 0.28 (95% CI = −0.27, 0.84) of salivary cortisol for the participants receiving MT in comparison to controls. This finding contrasts Field's assertion that reductions in cortisol level are one of MT's most reliable effects, but is consistent with meta-analytic findings based on adult samples, where cortisol effects were small and non-significant (g = 0.14, 95% CI = −0.10, 0.38) (10). Based on the available evidence, MT's single-dose effect on cortisol levels appears to be small, and possibly zero. Behavioral Dimension Distress behaviors. One study (No. 20 in Table 1) compared MT with standard care for children receiving burn treatment, and examined the children's distress behaviors before and during this painful procedure. The Children's Hospital of Eastern Ontario Pain Scale (34) was used to code distress behaviors before and during a dressing change (35). Six behavior categories were assessed including cry, facial, verbal, torso, touch and legs. The MT group showed only an increase in torso movements during the dressing change, whereas the control group showed an increase in five out of the six distress behaviors. Combined observational ratings made by nurses, who were blind to the group to which the children were assigned, favored MT, g = 0.37 (95% CI = −0.43, 1.17). Though this effect favors MT, the wide confidence interval, which may be partially or wholly attributable to the small sample (n = 24), makes this result difficult to interpret. In several other studies with behavioral outcomes where effect sizes could not be calculated, researchers reported improvements in fidgetiness, activity, vocalization and cooperation. The most commonly used instrument in these studies is a Behavior Observation Scale that was first used to assess behavior after relaxation therapy classes (36). Using this measure, behavior is observed three times during the 30 min prior to MT, during MT itself and during the 30 min after MT. Because the effect size data are inconclusive and researchers report positive behavioral improvements resulting from MT, further study on distress behaviors may be warranted. Multiple-Dose Effects Affective Dimension Depression. Depression, including motivational and cognitive deficits, vegetative signs, and disruptions in interpersonal relationships beyond those expected from ordinary unhappiness or poor mood (37), has been examined in pediatric MT studies, albeit with small samples. Two studies (No.'s 13 and 17 in Table 1), using a total of 54 participants, examined whether MT would reduce depression in comparison to either a wait-list control or to progressive muscle relaxation. The first study (No. 13) used the Children's Depression Inventory-Short Form, and was the only study reviewed to use that measure. It is an abbreviated version of a widely used self-report measure of depression for children and adolescents. The second study (No. 17) used the Center for Epidemiological Studies-Depression scale, which was also used in several studies that did not permit effect size calculation. Though the mean effect favored MT, g = 0.48, this effect is not statistically significant (95% CI = −0.06, 1.02). This result is inconclusive, but promising given that the confidence interval approaches significance, and also because the estimated effect is not greatly different from statistically significant reductions of depression found in adult samples (g = 0.62, 95% CI = 0.37, 0.88). It is also generally consistent with the conclusions reached by the authors of studies that did not permit effect size calculation. Still, the possibility that MT does not reduce depression for pediatric recipients to the extent it does in adults cannot be ruled out. At this point, further study on MT for pediatric depression is needed. Trait anxiety. One study (No. 13 in Table 1), using a sample of children and adolescents with attention-deficit hyperactivity disorder, examines MT's effect on trait anxiety, a dispositional, internalized proneness to be anxious (38), with enough detail to permit effect size calculation. This study, which also appears in the adult MT meta-analysis, uses the Revised Children's Manifest Anxiety Scale to compare MT with a wait-list condition and yields a large, statistically significant effect (g = 0.94, 95% CI = 0.20, 1.68) that is consistent MT's anxiolytic effects demonstrated in numerous studies with adult recipients. This result is consistent with author claims in the ‘Hurricane Andrew’ study, where improvements in children's trait anxiety at the end of the MT treatment are reported. Reduction of trait anxiety, resulting from a course of MT sessions, is certainly worthy of further study in pediatric populations. Pain. Four pediatric studies have been conducted that include pain as an outcome measure. Three of these studies used the Happy Faces Scale while only one study, a study of children with juvenile rheumatoid arthritis (and the only study with sufficient data for calculating effect sizes), used the Varni/Thompson Pediatric Pain Questionnaire. This study (No. 6 in Table 1) assessed pain in three ways—child self report, parent report and physician report. The children and a pediatric rheumatologist, who was blind to group assignment, generated reports of pain reduction that were consistent; parents' ratings were omitted from effect size calculation due to their non-blind status and proneness to bias. In comparison to a relaxation therapy group, MT yielded a large, statistically significant pain reduction (g = 1.33, 95% CI = 0.37, 2.29). This very large effect, in contrast with other studies that have shown mixed results for pain reduction resulting from MT, suggests that MT may be particularly well-suited to pain reduction for children with this condition. Three RCT's that do not permit effect size calculation looked at MT's impact on pain, and concluded that children experienced reductions of pain resulting from MT. Samples in these studies included children who experienced Hurricane Andrew (23), adolescents with attention-deficit hyperactivity disorder (39) and children with atopic dermatitis (26). These reports, combined with the findings in the juvenile rheumatoid arthritis study, indicate that MT for pain reduction in pediatric recipients is worthy of further examination. Physiological Dimension Muscle tone. Two recent studies that assess muscle tone (No.'s 23 and 24 in Table 1) yield g = 0.90 (95% CI = 0.23, 1.57). The first examines MT for children with cerebral palsy who received 30 min of MT 2 times per week for 12 weeks. This study also assesses spasticity, motor functioning, facial expressions and limb activity. The second study, of children with Down syndrome provided 30 min of MT 2 times per week for 8 weeks. Children's development, and fine and gross motor functioning, was evaluated. Both studies used the Arms, Legs and Trunk Muscle Tone Scale (ALT Muscle Tone Scale), which was designed during the pilot phase of the Down syndrome study. There is modest support that MT improves muscle tone, though it must be pointed out that the two studies diverge greatly in their individual results. Study number 23 in Table 1 had virtually no effect, while study number 24 had a huge effect. It must be noted that because these results come from a newly developed measure, the validity of the measure is not yet well-established. This may account for the divergent results across studies. Range of motion. One study (No. 23 in Table 1), with a sample of children suffering from cerebral palsy, yields non-significant improvements in range of motion, g = 0.31 (95% CI = −0.57, 1.19). The study notes that right and left hip extension, but not abduction, improved. Study authors note that increased muscle tone may have led to an increase in range of motion, but this result, based on a single study with 20 participants, is inconclusive. Immune measures. Two studies (No.'s 17 and 22 in Table 1) examined markers of immune system functioning in HIV+ adolescents or children. Though study authors are quick to point out that there were some within-group effects for those who received MT, a between-groups analysis that compares MT recipients with controls yields no effect (g = 0.06, 95% CI = −0.52, 0.63). In light of this result, specific MT effects on CD4 count were then examined separately, to account for the fact that the numerous immune system markers reported in those studies, and combined in the process of effect size calculation, might be obscuring this most important measure of disease progression in HIV+ persons. When an effect size for CD4 count only is generated from these two studies, the effect is only marginally larger and still non-significant (g = 0.24, 95% CI = −0.33, 0.82). Though these studies, and another study of MT for children with leukemia (29), make much of within-groups MT effects on immune system markers, the available between-groups data show that MT's effect on immune system function is probably small and possibly zero. Pulmonary function. One study (No. 14 in Table 1) examined whether MT might generate an improvement in peak air flow for a sample of 20 children with cystic fibrosis. In comparison to a reading control group, the children receiving MT had higher peak air flow, g = 0.47 (95% CI = −0.41, 1.35), indicating MT may be of specific value to children with this condition. Another study (27) that does not permit effect size calculation reports multiple measures of breathing improvement for asthmatic children ages 6–8 years, though older children (ages 9–14 years) showed fewer improvements. Overall, evidence for MT effects on pulmonary function is promising, and further study is warranted. Skin condition. Two RCT's examine MT's effect on skin condition. Children with atopic dermatitis were studied to determine the effect of MT on redness, scaling, lichenification, excoriation and pruritus (26). Though effect sizes cannot be calculated, study authors indicate some improvement following MT. Pediatric patients with hypertrophic scarring (HTS) were also studied and vascularity, pliability and height of the HTS revealed no appreciable effects. There is some evidence that pruritis decreased in certain patients with mature burn scars following MT. There is insufficient evidence to validate MT effects on skin condition in pediatric recipients. Glucose level. (No. 7 in Table 1) One study of children with diabetes examined the effect of MT on blood glucose levels (25). The authors conclude that MT lowers mean blood glucose levels and that compliance for insulin and food regulation improved, though data necessary for effect size calculation are not reported. Behavioral Dimension Developmentally appropriate functioning. Two studies (No.'s 23 and 24 in Table 1) used the Developmental Programming for Infants and Young Children (DPIYC) scale to assess developmentally appropriate functioning in recipients following MT. The average effect was small and non-significant (g = 0.24, 95% CI = −0.38, 0.86), though it must be pointed out that the studies had divergent results. Study number 24 yields a modest improvement in developmentally appropriate functioning in a sample of children with Down syndrome. Study number 23, with a sample of cerebral palsy sufferers, yields no effect. Possibly, these divergent results are the result of the different populations sampled within the studies. Taken together these results are inconclusive, but the encouraging results of the study with children who have Down syndrome indicate that further studies should be done with children having this or a related condition. Spasticity. One study (No. 23 in Table 1), with a sample of children suffering from cerebral palsy, examined whether MT had an effect on spasticity by using the Spasticity scale/modified Ashworth scale. The small, non-significant effect (g = 0.26, 95% CI = −0.62, 1.14) does not support an MT effect on spasticity. Because this study was very small (n = 20), further testing in this area may be beneficial. Hostility. One study (No. 21 in Table 1), with a small sample of aggressive adolescents, examined whether MT might reduce hostility compared with relaxation therapy. This study yields g = −0.85 (95% CI = −1.85, 0.15), a non-significant effect, but one that almost reaches significance in the wrong direction. The most tenable conclusion from this finding, a result based on a very small sample, is that MT has no effect on hostility beyond that provided by relaxation therapy to which it was compared. Though participants were randomly assigned, the MT group was approximately three points higher on the SCL-90R Hostility subscale prior to treatment. While scores decreased in the MT group during the treatment period, they also did in the relaxation therapy group, such that the MT group is still three points higher than the relaxation therapy group at the end of the treatment period. The authors' assertion that ‘by the end of the study [massaged adolescents] reported feeling less hostile’ obfuscates the fact that there was no between-groups effect. Related measures used in this study, such as the Overt Aggression Scale, were completed not by study participants, but by their legal guardians who were not blinded to group assignment, making their ratings prone to bias. Nevertheless, when all such measures included in the study are used to calculate an effect size, regardless of blinding, the result still converges on no effect. Classroom behavior. One study (No. 13 in Table 1) examined the classroom behavior of students with attention-deficit hyperactivity disorder using the Conners Teacher Rating Scale. Six factors of this scale were used including hyperactivity, conduct, emotional-indulgent, anxious-passive, asocial and daydream/attention problems. In comparison with a wait-list control, MT yielded g = 0.66 (95% CI = −0.07, 1.39), an effect that, while non-significant, is encouraging. Further studies of MT for this population are needed to confirm whether it might be of value in improving behavior, an outcome that seems possible given MT's relatively well-established potential to reduce anxiety. Studies that do not permit effect size calculation also suggest some improvements in classroom behavior following a course of MT. Improved classroom behavior and social relatedness are reported in studies of preschool children (40) and children with autism (41). Cognitive performance. (No. 11 in Table 1) A study of MT for children in preschool concludes that MT impacts cognitive performance (42). This study used three subtests of the Wechsler Preschool and Primary Scale of Intelligence—Revised, but does not report sufficient data to quantify the effect. Sleep and relaxation. (No.'s 1, 2 and 18 in Table 1) Though none permit effect size calculation, three studies have examined the effect of MT on sleep and relaxation. In one study, parents of children with autism recorded their children's sleep behavior in sleep diaries that included 5-point Likert scales (41). It is reported that the MT group showed decreases in fussing/restlessness, crying, self-stimulation behavior and getting out of bed. Nighttime sleep recordings were conducted for a previously mentioned study of MT for child and adolescent psychiatric patients (21). A video camera was set up on a tripod in the participant's room. The video tapes were subsequently coded for quiet sleep, active sleep, awake and lying quietly, and awake and active. Percentage of time asleep increased significantly from the first to last day of MT and the percentage of nighttime wakefulness decreased over the same time frame. Finally, children who experienced Hurricane Andrew were rated by an observer on a visual analogue scale (VAS), based on the child's apparent relaxation level (23). The VAS relaxation score increased significantly from the first to the last day of MT for the MT group. Possibly, MT promotes sleep and relaxation, but current studies do not permit this effect to be quantified. Discussion Available data reveals that MT provides benefit to pediatric recipients, though not as universally as has sometimes been reported. Benefits from both single-dose and multiple-dose sessions are evident. Most of the statistically significant effect sizes were observed for affective outcomes; findings for the behavioral and physiological dimensions were less consistent. These results parallel known MT effects in adult recipients, where multiple-dose reductions of depression and trait anxiety are the largest effects. In reviewing MT for pediatric recipients, we encountered several weaknesses endemic to the MT research literature that should be addressed in subsequent studies. These included (i) low statistical power, (ii) frequent failure to report basic descriptive statistics, (iii) descriptions of results that do not logically follow study designs, and (iv) lack of replication. We discuss these in turn. Low statistical power. Most pediatric MT studies were conducted with fewer than 30 participants. When studies are this small, only the largest effects have any likelihood of being uncovered. As researchers engaged in our own clinical MT studies, we are sensitive to the expense and difficulty involved in recruiting participants, but our empathy, unfortunately, does nothing to change the mathematics of the situation. The statistical power of MT studies must be increased (primarily by conducting studies with larger samples) if we wish to discover anything beyond the largest effects. Failure to report basic descriptive statistics. Many pediatric MT studies have appeared in journals with lax standards for the reporting of data; there is no other explanation for why studies would fail to report the most basic descriptive statistics. When a report does not include the standard deviations that describe the spread of the data, that report has almost no value as scientific evidence. This must not be allowed to continue in MT research, because it represents an enormous waste of resources. Simply put, researchers who take the time, effort and expense to perform a study should not allow the results of that study to be published without the statistics that permit an understanding of the outcome. If a journal does not require them, then the onus is on the researcher to ensure that the statistics are reported. Results that do not logically follow study designs. The importance of between-groups designs in MT has been noted (12). Why, then, do so many MT studies that employ a between-groups design emphasize within-group comparisons? The likely answer is that planned between-groups comparisons were non-significant (possibly due to low statistical power), so study authors may given in to the temptation to report the statistically significant, but misleading, within-group effect. This problem is rampant in MT research, and represents a real threat to the way this research will be perceived in the future. Many of the consumers of MT research are committed practitioners who have not had training in statistics and research design (though it must be noted that research literacy among MT practitioners is increasing). As such, they may be inclined to believe that coming across the sacrosanct ‘P < 0.05’ is proof of an MT effect, without realizing that this ‘effect’ may be nothing more than the effect of time, a placebo effect and/or regression to the mean. Knowledgeable consumers of MT research must learn to distinguish within-group effects from between-groups effects, and MT researchers must clearly present between-groups findings when their studies employ a between-groups design. Lack of replication. All but two of the pediatric MT studies we reviewed were conducted by the Touch Research Institute at the University of Miami. Their contribution to MT research has been considerable; however, scientific understanding is hampered when one laboratory is responsible for almost all of the results in an area of inquiry. Replication of results is a foundation of scientific progress, so it is necessary that other researchers contribute to this field. Given recent interest in CAM modalities, the small number of studies from other laboratories is surprising. We hope that this is about to change, and the success of the recent Highlighting Massage Therapy in CAM Research Conference (43) suggests that it will. Three pediatric MT RCT's from other research groups, currently in progress, support our optimism. Two of these examine MT for children and adolescents with cancer. While the results of a study that will examine pain, mood, stress and relaxation outcomes from MT and heat therapy for terminal children (44) cannot yet be reported, preliminary results of another study examining anxiety, fatigue, pain and nausea outcomes from MT in a sample of children with cancer (45) are promising; in particular, reductions of anxiety are expected to be consistent with what has been reported in this review. The third study, conducted by the first author of this review (S.B.) and researchers at the UCLA Pediatric Exercise Laboratory (46), evaluates pre-exercise MT for children with and without heart disease. Preliminary results show significant improvement in VO2 (oxygen consumption) for children who have an MT session prior to performing an exercise regimen on a stationary bicycle. If these results are validated in final analyses, they will be generally consistent with the finding that MT can improve pulmonary function. There are two other studies worth mentioning that have received IRB approval from Children's Memorial Hospital in Chicago and will be enrolling participants shortly. These studies are an evaluation of MT for reduction of pain, nausea and anxiety in pediatric cancer patients (47) and an evaluation of MT on immune function and heart rate variability in HIV-infected pediatric patients (48). The outcome measures for the first study are self-report ratings of pain, nausea and anxiety, whereas the second study evaluates T-Helper cells (CD4+), T-Suppressor cells (CD8+), Total B cells (CD3, CD19), Natural Killer cells (CD56+), viral load and heart rate variability. Conclusion Current research indicates that MT is not a panacea for conditions studied in the pediatric population. In contrast to what has sometimes been claimed, there is little to no evidence to date to support effects such as improved immune system functioning, reduction of spasticity, or amelioration of hostility. In addition, there is currently scant evidence that MT provides benefits by first reducing cortisol, as MT's effect on this stress hormone is seen to be small when analyzed correctly (i.e. in between-groups as opposed to within-group comparisons). There is, however, a set of MT effects that have been shown to have real value to the pediatric population. MT shows a considerable impact on the state and trait anxiety levels of children. Because these effects are strong, and also because they are consistent with the findings in adults, future research on the anxiolytic effects of MT on pediatric recipients does not need to simply replicate previous studies. The greatest progress can now be made by focusing on the mediators and moderators of MT effects on anxiety, and on testing explanatory theories of these outcomes. MT effects on arthritis pain and muscle tone also appear to be strong, but these results do need to be replicated, as they are based on single studies. Other pediatric outcomes that are promising, but in need of further study, include MT's effects on depression, negative mood, certain types of behavior (likely due to reductions of anxiety) and air flow in those suffering from pulmonary disorders such as cystic fibrosis. As increased statistical power in the form of additional studies is brought to bear on these potential benefits, it is likely that some will be quantitatively validated. Finally, it has been noted that prior MT research has not accounted for the communication that inevitably takes place between massage therapists and their recipients, nor has it examined the likelihood that therapists and recipients develop a therapeutic relationship during the course of MT (10). This is also true in pediatric MT studies. MT has important parallels (in both process and outcomes) to psychotherapy (10), a treatment that relies on communication and therapeutic relationship to provide effects. It seems likely that MT effects, especially those belonging to the affective category, are mediated or moderated by these previously unexamined factors. These should not be neglected in subsequent pediatric MT research. As adult consumers continue to explore and utilize all of their health care options, children will increasingly be recipients of MT. With this in mind, it is essential that we continue to study the benefits of MT for children, and the explanatory models that underlie them, so children's health and wellness can be maximized. The value of MT has been examined for many specific conditions that afflict children. It is our hope that this review has consolidated those findings, indicated areas that require further study, and led to an increased scientific understanding of pediatric MT.	[ "cam", "child", "health care", "pain management", "comfort", "touch", "integrative medicine", "kid", "quality of life" ]	[ "P", "P", "P", "P", "P", "P", "R", "U", "M" ]
Purinergic_Signal-4-2-2377320	Purinergic signaling in the lumen of a normal nephron and in remodeled PKD encapsulated cysts	The nephron is the functional unit of the kidney. Blood and plasma are continually filtered within the glomeruli that begin each nephron. Adenosine 5′ triphosphate (ATP) and its metabolites are freely filtered by each glomerulus and enter the lumen of each nephron beginning at the proximal convoluted tubule (PCT). Flow rate, osmolality, and other mechanical or chemical stimuli for ATP secretion are present in each nephron segment. These ATP-release stimuli are also different in each nephron segment due to water or salt permeability or impermeability along different luminal membranes of the cells that line each nephron segment. Each of the above stimuli can trigger additional ATP release into the lumen of a nephron segment. Each nephron-lining epithelial cell is a potential source of secreted ATP. Together with filtered ATP and its metabolites derived from the glomerulus, secreted ATP and adenosine derived from cells along the nephron are likely the principal two of several nucleotide and nucleoside candidates for renal autocrine and paracrine ligands within the tubular fluid of the nephron. This minireview discusses the first principles of purinergic signaling as they relate to the nephron and the urinary bladder. The review discusses how the lumen of a renal tubule presents an ideal purinergic signaling microenvironment. The review also illustrates how remodeled and encapsulated cysts in autosomal dominant polycystic kidney disease (ADPKD) and remodeled pseudocysts in autosomal recessive PKD (ARPKD) of the renal collecting duct likely create an even more ideal microenvironment for purinergic signaling. Once trapped in these closed microenvironments, purinergic signaling becomes chronic and likely plays a significant epigenetic and detrimental role in the secondary progression of PKD, once the remodeling of the renal tissue has begun. In PKD cystic microenvironments, we argue that normal purinergic signaling within the lumen of the nephron provides detrimental acceleration of ADPKD once remodeling is complete.	[ "signaling", "kidney", "adenosine", "atp", "nucleotides", "nucleosides", "polycystic kidney disease" ]	[ "P", "P", "P", "P", "P", "P", "P" ]
Ann_Biomed_Eng-3-1-2040176	Occupant Dynamics in Rollover Crashes: Influence of Roof Deformation and Seat Belt Performance on Probable Spinal Column Injury	Motor vehicle crashes are the leading cause of death in the United States for people ages 3–33, and rollover crashes have a higher fatality rate than any other crash mode. At the request and under the sponsorship of Ford Motor Company, Autoliv conducted a series of dynamic rollover tests on Ford Explorer sport utility vehicles (SUV) during 1998 and 1999. Data from those tests were made available to the public and were analyzed in this study to investigate the magnitude of and the temporal relationship between roof deformation, lap–shoulder seat belt loads, and restrained anthropometric test dummy (ATD) neck loads. Introduction Motor vehicle crashes are the leading cause of death in the United States for persons of every age from 3 through 33.18 Rollover crashes, in particular, have a higher fatality rate than any other crash mode. Of the 6,159,287 police reported crashes in 2005 in the United States, only 4.1% involved a rollover. Yet, rollovers accounted for 34.4% (10,816) of all passenger vehicle fatalities. The fatality rate (defined as deaths per 100,000 registered vehicles) in rollover crashes is more than two and one-half times higher in sport utility vehicles (SUVs) than in passenger cars (8.28 vs. 3.22, respectively). Another 149,406 individuals sustained serious injuries in rollover crashes in 2005, approximately 30% of whom were occupants of SUVs.18 Measurement tools are required to evaluate appropriate intervention strategies to reduce the mortality and morbidity associated with rollover crashes. The Abbreviated Injury Scale (AIS) and Harm value (HARM) are two such measures, which are used to compare injuries of different types and severities.9 The AIS reflects “threat to life” (or risk of death) and stratifies injuries into six categories (AIS 1–6) with AIS 6 defined as an “untreatable” injury. The maximum AIS (MAIS) score refers to the most severe injury sustained in a specific individual. The highest rate of seriously injured (MAIS 3+) children 4–12 years old occurs in rollovers, more than double any other crash mode.22 One study reported that despite the increased weight of SUVs, the risk of injury for children in SUVs was similar to that for passenger cars, which may be due to the increased tendency of SUVs to roll over.6 HARM is a mathematical measure that applies a weighting factor to each injured, yet surviving occupant, which includes both monetary costs of the injury (i.e., direct costs associated with the loss of wages) and comprehensive costs (i.e., monetary costs plus non-monetary costs).9,14 Importantly, the non-monetary costs include such measures as loss of functional capacity and quality of life. In rollover crashes, the highest HARM is reflected in the population of occupants who survive the crash, yet sustain catastrophic injury, including permanent brain and/or spinal cord injury.9 A major source of HARM in rollover crashes results from occupant contact with the vehicle roof, support pillars and side headers.9 According to Digges et al. “Countermeasures to reduce rollover casualties include increased belt use, and technological interventions to prevent ejection and reduce the severity of body contacts with the vehicle interior.”9 Hugh DeHaven outlined principles for designing vehicles for human transport in 1952, and stated that “the package (the passenger compartment) should not collapse under expected conditions of force, thereby exposing objects (and people) inside it to damage.”7 Closely related to this principle, DeHaven stated that “the packaging structures which shield the inner container must not be made of brittle or frail materials; they should resist force by yielding and absorbing energy applied to the outer container so as to cushion and distribute impact forces and thereby protect the inner container.” Franchini11 reported the importance of maintaining the integrity of the occupant compartment (referred to as “survival space”) for belted occupants during a crash event. Few studies, however, have experimentally evaluated the influence of dynamic roof deformation with concomitant loss of occupant survival space as a potentially preventable cause of head and cervical spine injuries in rollover crashes. Bahling et al.1 conducted rollover and drop tests of late 1980s model Chevrolet Malibu passenger cars using lap–shoulder belted anthropomorphic test dummies (ATDs) to evaluate the influence of roof strength on injury mechanics. In several studies, Moffatt and other investigators reported on a customized, dynamic test procedure with controlled roof impacts of passenger vans and sedans.3,5,15 Both groups concluded that “potentially injurious impacts” (defined by these authors as 2000 N recorded by the ATD upper neck sensor) occurred prior to significant roof crush and that enhanced roof strength made no difference in the likelihood of serious neck injury to restrained occupants. These conclusions have been contradicted in other literature, albeit without the benefit of scientific data from comparable full-scale, dynamic rollover crash tests.12,13 Recently, the raw data from a series of full-scale, dynamic rollover tests of 1998–1999 model year Ford Explorer SUVs, using fully instrumented ATDs, were made publicly available to the scientific community.2 These tests were conducted using a standardized test methodology specified by Federal Motor Vehicle Safety Standard (FMVSS) 208 of the U.S. Department of Transportation (FMVSS 208 49 CFR Ch. V (10-01-04 Edition)). These data provide the first opportunity to analyze the dynamics of belted occupants within a contemporary SUV in a standardized rollover test environment. The authors used the described data in the present study. The study’s objective was to investigate both the magnitude of and the temporal relationship between dynamic roof deformation, lap–shoulder seat belt loads, and restrained ATD neck loads in these FMVSS 208 dolly rollover tests of Ford Explorer SUVs. Materials and Methods Three full-scale dolly rollover tests of 1998–1999 Explorer SUVs were sponsored by Ford Motor Company and conducted at the North American Autoliv test facility1 according to the standardized FMVSS 208 test protocol. This standard, issued by the National Highway Transportation Safety Administration (NHTSA) in 1971, outlines vehicle crash performance requirements designed for improving occupant protection. Described within FMVSS 208 is a voluntary, dolly rollover test used to evaluate occupant protection during a lateral rollover. This standardized test procedure was used for all three SUV rollover tests. The FMVSS 208 test procedure involves accelerating a rollover dolly carrying a test vehicle to a set velocity and then rapidly decelerating the dolly, causing the vehicle to release, trip, and enter a lateral roll. The SUVs were first placed on a dolly fixed with a rigid, angled platform positioned 9 in. from the ground and 23° from the horizontal. To assure that the roll was driver-side leading (i.e., that the driver’s side contacted the ground first), each vehicle was situated on the platform so that its driver’s side tires rested along a 4 in. lip at the forward edge of the test platform, as shown in Figs. 1a and 1b. Figure 1(a) FMVSS 208 dynamic rollover pre-test setup of SUV test B190043 on a dolly rollover cart; (b) critical dimensions of FMVSS 208 dolly The vehicle and dolly were then accelerated along a straight path perpendicular to the longitudinal axis of the vehicle. The dolly was then decelerated rapidly (minimum deceleration rate of 20 g for 0.04 s), resulting in vehicle trip and subsequent rollover. The target speed for all Explorer tests was 30 mph. The vehicle identification number (VIN) and actual dolly launch velocity for each of the three tests are summarized in Table 1. Table 1Test conditionsAutoliv test numberTest descriptionData sampling rateSampling durationTest dateAccelerometerCoordinates (mm)XYZB190042FMVSS 208 Rollover (VIN1FMCU24E5VUC19292) (23°; 30.5 mph)20,000 Hz∼13,000 ms8/10/99Center of gravity2073.10−24.50975.00Driver’s rail at A-pillar2038.20−596.101770.0Driver’s rail at B-pillar2776.40−571.301863.30Driver’s rail at C-pillar3209.70−580.601871.40Passenger’s rail at A-pillar2011.80592.101758.60Passenger’s rail at B-pillar2786.60532.601857.10Passenger’s rail at C-pillar3234.40523.301868.20B190043FMVSS 208 Rollover (VIN1FMDU34E6VUB99290) (23°; 30.4 mph)20,000 Hz∼13,000 ms8/11/99Center of gravity2231.0−4.1761.3Driver’s rail at A-pillar2077.2−543.71626.8Driver’s rail at B-pillar2637.2−535.01644.0Driver’s rail at C-pillar3436.0572.41598.0Passenger’s rail at A-pillar2077.9573.51564.7Passenger’s rail at B-pillar2636.5569.71625.5Passenger’s rail at C-pillar3436.3−543.01617.6B180220FMVSS 208 Rollover (VIN1FMDU35P5VUC14510) (23°; 30.9 mph)12,500 Hz∼8000 ms12/10/98Center of gravity2231.0−4.1761.3Driver’s rail at A-pillar2077.2−543.71626.8Driver’s rail at B-pillar2637.2−535.01644.0Driver’s rail at C-pillar3436.0572.41598.0Passenger’s rail at A-pillar2077.9573.51564.7Passenger’s rail at B-pillar2636.5569.71625.5Passenger’s rail at C-pillar3436.3−543.01617.6 Vehicle and ATD Instrumentation Triaxial and biaxial accelerometers were installed in each SUV to measure vehicle structural dynamics during the rollover tests. Accelerations were recorded at each of the following vehicle structure locations in coordinate systems consistent with SAE J211 guidelines (SAE J211-1 REV DEC03), as shown in Figs. 2a and 2b, and Table 1.Center of Gravity (CG) (x, y, z)Left (driver) roof rail at A-pillar (y, z)B-pillar (y, z)C-pillar (y, z)Right (passenger) roof rail at A-pillar (y, z)B-pillar (y, z)C-pillar (y, z) Figure 2 (a) Exterior view of test vehicle indicating A, B, and C-pillars; (b) interior view of test vehicle—accelerometers were mounted at the roof rail-to-pillar junction at the A, B, and C-pillars for both driver and passenger sides of the Ford Explorer SUVs In each test, two Hybrid III 50th percentile male, instrumented ATDs were placed in the front seats of Ford Explorer SUVs and restrained with lap–shoulder belts that were affixed with load cells. Time synchronized with all other sensors and cameras in the experimental setup, these load cells continuously recorded tensile forces in the webbing of the lap and shoulder belts. Both ATDs were positioned in their respective seats according to reference FMVSS 208 Dummy Positioning Procedure with the seats located in the mid-track position. Given that the vehicle was positioned to roll driver-side leading, the driver ATD represented the “near-side” occupant and the passenger ATD, the “far-side” occupant in this test series. The following sensor outputs were recorded continuously throughout each rollover test for both driver and passenger ATDs:Head CG acceleration (x, y, z)Upper Neck Force (x, y, z)Moment (x, y, z)Lower Neck Force (x, y, z)Moment (x, y, z)Chest Acceleration (x, y, z)DeflectionPelvis Acceleration (x, y, z)Femur Force (Fz, right and left) Dynamic motion of the SUVs was visually recorded by up to 10 high speed external cameras, which were time synchronized with the cart release trigger and imaged at 500 frames/s. (Fig. 3) Clocks that were positioned along the test track were visible in the external camera views throughout the roll sequence. Targets were affixed to the rear aspect of the driver and passenger ATD heads and were visible by the same camera monitoring the onboard time clock. Up to three onboard, high speed cameras recorded ATD kinematics as a function of time using the onboard clock, which was also time synchronized with the cart release trigger, external cameras, and electronic sensors affixed to both ATDs, the seat belts, and the SUV structure. Figure 3External camera setup With the cart positioned according to FMVSS 208 specifications, the data acquisition system was set up and armed for trigger. Time zero for all data acquisition was provided by a contact switch at both shock absorbers to indicate the start of cart deceleration. The data-sampling rates for each test are provided in Table 1. Data Analysis The raw ASCII sensor data retrieved from the ATD and vehicle-mounted transducers were filtered according to the SAE J211 channel filtering class (CFC) requirements to remove the effects of crash related noise on the sensor readings. Each CFC specifies that the channel frequency response lies within limits detailed in SAE J211. Neck forces, neck moments, and vehicle data were filtered at CFC 1000, CFC 600, and CFC 60 accordingly. Subsequently, the filtered vehicle acceleration data were transformed into a vehicle-fixed, center-of-gravity, coordinate system using MATLAB® (registered trademark of The Mathworks, Inc.). Roof rail accelerations were transformed by subtracting the center of gravity accelerations from each measurement as a function of time. The result of the transformations was the acceleration of the roof rail relative to the vehicle’s center of gravity. The transformed roof rail acceleration values provided insight into when and in what direction the rail was moving. Transformed data was then compared to the ATD neck and seat belt loads as a function of time. The accelerations recorded by the rail sensors were influenced by both vehicle rotation (angular velocity) and localized pillar deformation. The contribution due to vehicle rotation was made at a low frequency given that the peak roll rate of the vehicle at any time during the test interval (0–1000 ms) was approximately 350–400°/s (i.e., less than 2 Hz). Thus, all higher frequency accelerations were necessarily due to localized structural deformation. The sensor outputs from the driver and passenger roof rail accelerometers positioned at the B-pillar were used as a quantitative definition of roof intrusion (i.e., “objective roof crush”) into the occupant survival space. The mathematically transformed driver and passenger roof rail acceleration tracings in the present study provided objective evidence of the time interval(s) when roof/pillar deformation occurred, and the time of deformation was corroborated through careful scrutiny of the high speed video images (“observable” roof crush). Objective roof crush in our study was defined under the following necessary, contemporaneous conditions:Vertical and/or lateral rail acceleration peak(s) “downward” and/or inboard toward the restrained ATDExternal camera video images consistent with SUV roof-to-ground contactOnboard camera video images consistent with a compromise in occupant survival space (i.e., reduced headroom) Roof rail accelerations, seat belt loads and peak ATD axial neck force (Fz, axial compression/tension) and moments (My, chin-to-chest flexion, and Mx, ear-to-shoulder lateral bending) were compared to the high-speed video data to validate the correlation in time between the presence of objective roof crush and the development of peak ATD neck and seat belt loads. Results Results are presented and discussed for the first full second of the roll sequence for each of the three SUV tests, which includes the initial driver-side-to-ground contact and full roof contact, followed by the first passenger-side-to-ground contact. In all three tests, the external high-speed cameras recorded continuous roof-to-ground contact from the first contact of the driver’s roof rail with the ground until the end of the 1000 ms period, which corresponded to approximately 5/8 roll. Onboard high speed video cameras recorded the front passenger compartment of each SUV, capturing the kinematics of the ATDs as well as the deformation of the roof header and side roof rails during the rollover event. These cameras recorded inboard displacement of roof rails (“observable” roof crush) of both driver and passenger-side roof rails during the respective roof rail-to-ground contact for all three tests. The onboard clock, which was recorded by the interior cameras, allowed for a time-synchronized comparison of this data to the sensor output of the ATD neck transducers and roof rail accelerometers. In all three tests, the head of each lap–shoulder belt restrained ATD (driver and passenger) contacted the roof two or more times during the first 1000 ms roll interval as indicated by increased neck compression loads (negative Fz values) recorded by the ATDs’ upper neck sensor (Fig. 4). The time of occurrence of each peak neck load for the driver and passenger ATD head-to-roof contacts (the “Local Maximum” as well as the “Absolute Maximum” or “Peak” neck loads) was consistent with the onboard high speed video data. For each of the three tests, the time recorded by the onboard camera during observable ATD head-to-roof contact was within 3 ms of the time recorded by the neck sensors for peak neck loads. Peak axial neck force data (Fz) were measured by the ATDs’ upper neck sensors and the moment loads (My and Mx) were measured by the ATDs’ lower neck sensors (Table 2). Figure 4Passenger upper neck Fz vs. time (Test B190042)Table 2Magnitudes of Local and Absolute maximum (Peak) neck loads (during time interval of continuous roof-to-ground contact)Test parameterDriverPassengerB190042B190043B180220B190042B190043B180220Max peak Fz (N)a−958−1962−1920−5933−3245NonebLocal peak (s), Fz−200−295−223−361−50200–260Max peak My (N m)5811094304178261Local peak (s), My2112–5412–2220–2410Max peak Mx (N m)−106−124−167689841Local peak (s), Mx−11 to −18n/a−20 to −4691219–21aFz was measured at the dummy upper neck load cell; My and Mx measured at the lower load cellbIn Test 180220, no absolute maximum Fz was identified as all peak neck compression loads were within the range of 200–260 N The near-side (i.e., the driver) ATD recorded Peak axial neck compression (Fz) loads of −2000 N or less for all three tests. Peak neck flexion (My) loads for the driver ATD resulted in chin-toward-chest head motion and were within a range of 58–110 N m for all three tests. In contrast, the far-side (i.e., the passenger) ATD recorded Peak neck Fz values in the range of −3244 to −5933 N for two out of three tests. In Test B180220, no Absolute Maximum Fz value was identified as all Local Maxima compression loads were within the range of −200 to −260 N. The Peak neck flexion (My) loads, recorded by the passenger ATDs, resulted in chin-toward-chest head motion and ranged from 178 to 304 N m for all three tests. The Peak lateral neck bending ear-to-shoulder (Mx) loads ranged from 41 to 98 N m for the passenger ATD, and the driver experienced Mx load magnitudes in the range of −106 to −167 N m, for all three tests. The time of occurrence of objective vertical and lateral roof crush was identified by the roof rail acceleration peak, for both the passenger and driver side roof rails in all three tests (Table 3). Objective roof crush occurred prior to the Absolute Maximum dummy neck loads (Fz, My, Mx) in each of the three tests for both driver and passenger ATDs (Table 3, Fig. 5). The far-side, passenger ATD recorded Peak neck loads within 10 ms following the time of occurrence of objective roof crush for all three tests. Peak neck loads for the near-side, driver ATD occurred up to 65 ms following the time of occurrence of objective roof crush for all three tests. Table 3Time of occurrence (ms) of roof/pillar deformation and absolute maximum neck loadsTest parameterDriverPassengerB190042B190043B180220B190042B190043B180220Objective roof/pillar deformation (vertical acceleration peaks)497513510730∼600–800a742Objective roof/pillar deformation (lateral acceleration peaks)497513494496512495Peak Fzc533540516730764n/abPeak My533541517729764751Peak Mx537548540783774760aNo isolated spike in acceleration was noted; however, a sustained level of roof deformation is reflected in the roof rail tracings and the video footage for this testbNo isolated absolute maximum Fz value was noted, which differed significantly from the other “local” maximums. This was consistent with the very high peak My in this test, however, which was noted at 751 ms (see Table 2)cFz was measured at the upper neck load cell, whereas My and Mx were measured at the lower neck load cellsFigure 5Driver upper neck Fz vs. vertical acceleration of the driver rail at B-pillar (Test B190043) Lap and shoulder belt loads were recorded for both driver and passenger ATDs in all three tests, with the exception of the driver’s shoulder belt for test B190043 and passenger shoulder belt for test B180220, both of which experienced sensor failures (Table 4). The Peak lap belt loads for the near-side, driver ATD occurred within 1 ms of the time of Peak Fz and My neck loads for two out of three tests (Figs. 6, 7). In the remaining test (B190042), the Peak lap belt load occurred 22 ms prior to the time that peak neck Fz and My neck loads occurred in the driver ATD. In contrast, the far-side, passenger ATDs’ Peak lap belt loads occurred 65–131 ms prior to the time of Peak neck compression (Fz) for two out of three tests and 65–145 ms prior to the time of Peak neck flexion (My) for all three tests (Fig. 8). In Test B180220, the passenger’s Peak lap belt load occurred 6 ms after the time of Peak neck compression, yet 145 ms before the Peak neck flexion load (My). Table 4Peak belt loads compared to belt loads at time of absolute maximum (peak) neck loadsTest parameterDriverPassengerB190042B190043B180220B190042B190043B180220Peak lap belt (N)548 (511 ms)795 (540 ms)940 (515 ms)797 (599 ms)705 (699 ms)779 (606 ms)Peak shoulder belt (N)395 (379 ms)N/Aa410 (373 ms)899 (587 ms)953 (689 ms)N/AaPeak FzLap belt (N)398 (533 ms)789 (540 ms)937 (516 ms) 359 (730 ms)115 (764 ms)771 (600 ms)Shoulder belt (N) 231 (533 ms)N/Aa236 (516 ms)337 (730 ms)426 (764 ms)N/AaPeak My(N m)Lap belt (N)393 (533 ms) 785 (541 ms)937 (517 ms)361 (729 ms)115 (765 ms) 218 (751 ms) Shoulder belt (N) 227 (533 ms)N/Aa236 (517 ms)341 (729 ms)417 (764 ms)N/AaaData not available due to transducer failureFigure 6Driver lap belt and neck load vs. time (Test B190043)Figure 7Inverted 50th percentile Hybrid III Driver ATD “diving” into roof with lap and shoulder belt providing restraining forces, which reduces neck loadFigure 8(a) Roof crush—passenger lap belt and neck load vs. time (Test B190043); (b) Roof crush—passenger shoulder belt and neck load vs. time (Test B190043) The driver ATDs’ Peak lap belt loads ranged from 548 to 940 N for the three tests, and the shoulder belt loads were 395–410 N for two tests (the driver’s shoulder belt sensor failed on Test B190043) (Table 4). The passenger ATDs’ Peak lap belt loads were similar to those of the driver ATDs’, ranging from 705 to 797 N for all three tests; the shoulder belt loads were 899–953 N for two of three tests (the passenger’s shoulder belt sensor failed on test B180220). The near-side, driver ATD lap belt loads at the time of occurrence of the Peak Fz and My neck loads were 72–99% of the ATDs’ respective Peak lap belt load for all three tests (Table 4). In contrast, the far-side, passenger ATD lap belt loads at the time of occurrence of the Peak neck compression (Fz) were less than 45% of the passenger ATDs’ respective Peak lap belt load recorded for two of the three tests. At the time of Peak neck flexion (My), the lap belt loads were less than 45% of the Peak lap belt load for all three tests. Shoulder belt loads for both driver and passenger ATDs at the time of occurrence of the Peak neck Fz and My were 37–58% of the ATDs’ respective Peak shoulder belt loads for two of three tests (load cells malfunctioned in the driver’s shoulder belt in test B190043 and passenger’s shoulder belt in test B180220). Discussion The three rollover tests of 1998–1999 Ford Explorer SUVs analyzed in the present study represent a unique dataset evaluating occupant dynamics in rollover crashes as represented by Hybrid III ATDs. To the authors’ knowledge, this study represents the first published test series of full-scale rollover crashes of a contemporary SUV with time synchronized sensor output from ATD neck transducers, roof rail accelerometers, an onboard high speed clock, and high speed external and internal video cameras. ATD Biofidelity The biofidelity of the Hybrid III ATD in rollover crash conditions is a significant limitation of this study as well as other published rollover studies1,15; however, ATD sensor data may be carefully evaluated within such limitations to yield objective insights. The ATDs used in this study were originally designed as a measurement tool in frontal crash tests, not rollovers. Thus, the Hybrid III ATD neck was designed and tested primarily to simulate dynamic head motion in the sagittal plane (neck flexion and extension).8,10 The Hybrid III ATD neck was not specifically designed to produce a biofidelic response under axial compression loading, typical of the neck loads resulting from the head-to-roof contacts observed in this study. Researchers have shown that with an axially directed impact to the crown of the ATD head, the ATD lower neck sensor records essentially the same magnitude axial force as the upper neck load cell, whereas the human cadaver neck lower load cell records a much reduced load magnitude compared to the upper load cell.24,29 Also, the Hybrid III ATD necks are neither frangible nor viscoelastic; therefore, they cannot accurately simulate individual vertebral displacement and/or fracture. Notwithstanding the biomechanical differences between the necks of human cadavers and ATD mechanical necks, the Fz measurements of the upper load cell of a restrained Hybrid III dummy may provide a reasonable estimate of the initial input load to a restrained occupant’s head when it contacts the roof in a rollover crash under similar dynamic conditions (e.g., impact velocity, impact duration). Given that the mass of a Hybrid III dummy head reasonably mimics that of the comparable human (50th percentile male in the present study), the inertial effect of the dummy head should not confound Fz readings from its upper neck load cell. In all three tests within the present study, Fz values for ATD axial neck load were evaluated solely from the upper neck load cells, thereby taking into consideration, as much as possible, the significant differences in axial stiffness between the ATD and cadaver human neck (the lower neck Fz sensor was not evaluated). These upper neck Fz sensor readings were interpreted relative to the head impact loads in cadaver tests under similar dynamic conditions (described below). Given that the Hybrid III ATD neck was specifically designed to be biofidelic in dynamic, sagittal plane bending (flexion and extension) loads, neck flexion loads (My) were evaluated in this study from the lower neck load cell of driver and passenger ATDs. Lateral bending loads (Mx) were recorded and reported, but not interpreted in this study due to the lack of available biomechanical tolerance data for the human spine when subjected to lateral bending. Additional research in this important area, for both rollover and side impact crashes, is needed. Injury Tolerance of the Human Cervical Spine Semantic confusion exists in the published literature regarding cervical spine tolerance or “failure load.” Some investigators have reported cervical spine “failure loads” resulting from impacts to cadaver heads as the output force measured at the lower cervical spine, thereby neglecting the kinetic energy dissipation associated with vertebral fracture and/or subluxation above the position of the sensor as well as the viscoelastic response of the intervertebral discs.19,20 Other studies have reported both the input and output loads in cadaver tests of the cervical spine.23–26 These data provide a quantitative comparison of the significant differences between the input loads measured at the cadaver head versus what might be more appropriately referred to as the residual force magnitude that is recorded at the base of the cervical spine segment. No experimental cadaver work has measured and/or estimated the input loads at the superior aspect of the human neck, which is the site of the upper neck load cell in the Hybrid III dummy, when the head is subjected to an axial impact on the crown of the head. Additional research in this area is needed. The confusion of terms regarding cervical spine failure loads in the published literature has a significant bearing on the proper interpretation of Hybrid III ATD axial neck load sensor output in controlled rollover tests such as the present study. Bahling et al. utilized 2000 N as a neck sensor Fz threshold for “potentially injurious impacts (PII)” in rollovers, which is in the range of the published residual force magnitude recorded at the lower cervical spine following an injurious impact to the cadaver head.1,19 In contrast, other investigators reported the actuator (input) force at the head that was required for cervical spine injury in axial compression.24 In interpreting the ATD neck sensor data in this study, an Fz value of 7000 N was utilized as a threshold of “probable (spinal) column injury,” which is in the range of the published head impact loads in cadaver studies. A value of 150 N m was used as the threshold for probable column injury for interpretation of My readings from the lower neck sensor. In reality, the true axial compressive failure load of the human cervical spine is likely slightly lower than that measured by the head input loads in cadaver tests (due to inertial effects of the cadaver head), yet substantially greater than the lower cervical spine force data reported in the same cadaver tests. Thus, the real world relevance of threshold values used for both PII and “probable column injury” as a predictor of human injury risk must be tempered, consistent with the known biofidelity limitations of the ATD neck (described previously) and the wide variation in injury tolerance of the human spine, as described below. The biomechanical literature suggests that the tolerance of the human cervical spine to serious injury is influenced by several mechanical factors, including individual tolerance variations due to age and gender, load magnitude and loading rate, pre-alignment (initial head-to-neck position), and end conditions of the cervical spine. Pintar et al.24 reported that serious (AIS 3+) cervical spine injury under axial compression loading conditions occurred at 5856 N with a loading rate of 3 m/s (for a 67-year-old female cadaver) and 11,242 N with a loading rate of 8 m/s (for a 50-year-old male cadaver). In the present study, none of the near-side, driver ATD sensors recorded upper neck Fz values exceeding even 2000 N neck compression, and only one far-side passenger dummy (Test B190042) exceeded a neck sensor reading of 5856 N compression. Controlling for loading rate, age, and gender in a sample of 25 human cadaver head-neck compression tests, Pintar and Yoganandan27 reported that compressive tolerance varied from 7 kN in the young (third decade) to 2 kN in the very old (ninth decade). “Failure loads” (recorded at the lower cervical spine) ranged from 2 kN at quasi-static loading rates to 5 kN at dynamic loading rates (8 m/s). Failure tolerance for the male population was 25% higher than the female population (without regard to age and rate of loading). The present study recorded roof rail-to-dummy head contacts in the range of 5 m/s (11.2 mph) in the first ground strike of the SUV rollovers.4 Thus, the roof rail-to-head impact speeds in the present study were within the range of loading rates to cadaver and Hybrid III ATD heads in the studies by Pintar et al.24,27 Pre-alignment of the head and neck complex when struck by (or when it strikes) the roof is a key variable in determining whether catastrophic injury will occur, and if so, the specific type of injury that will be sustained. Published biomechanical studies suggest that cervical spine pre-flexion yields a greater incidence of lower cervical compression and burst fractures than neutrally positioned spines.24,25 Restrained occupants in rollover collisions are typically pre-flexed, due to the initial locking of their belt restraint system, and thus are at higher risk of cervical spine injury in the presence of an intruding roof structure when compared to unrestrained occupants. This higher risk of roof impact for restrained vs. unrestrained occupants in rollovers was confirmed with U.S. field accident data.9 The degree of pre-flexion of the ATD necks prior to head-to-roof contact in the present study was not determined because the onboard cameras were positioned behind the ATDs during the rollover tests. The seat belt load cells (i.e., excluding the two sensors described previously that malfunctioned) recorded continuous tensile loads in both lap and shoulder belts throughout the test interval for both driver and passenger ATDs in all three tests. This belt load data indicated that the lap–shoulder belts locked and remained locked during the first 1000 ms test interval for all ATDs in all three tests. Nightingale et al.20 reported on the influence of boundary or end conditions of the cervical spine relative to specific injury types. When the head was unconstrained and free to translate and/or rotate away from the applied load, no cervical injury was sustained, despite high input loads to the head. In “full constraint” conditions, with the head pre-flexed along its stiffest axis (i.e., the removal of the normal cervical lordosis), buckling and burst fractures were noted. With rotational constraint (e.g., when the neck flexes forward to the point where the chin is against the sternum), bilateral locked facets typically occurred. Using cadaver spine specimens, Pintar et al.26 determined that the average peak My magnitude resulting in “major” neck injury was 97 N m at the specific site of injury in the cervical spine (i.e., not positioned as inferior as the lower neck load cell of the Hybrid III ATDs in the present study). The peak axial head impact force ranged from 3000 to 9700 N, with the peak flexion bending moment at the injured level ranging from 19 to 169 N m. Thus, the degree of constraint imposed by the contacting surface, such as an intruding roof, can be a major determinant for cervical spine injury. In all three tests of the present study, the far-side, passenger ATDs recorded lower neck flexion (My) loads that exceeded 175 N m. None of the near-side, driver ATD lower neck sensors recorded peak My loads exceeding 110 N m. No dynamic test data for cervical spine tolerance in lateral bending was identified; therefore, no interpretation of the peak Mx sensor outputs for any of the ATDs was made. The biomechanical impact environment to which the restrained ATDs were subjected in the present study was shown to be comparable to that of published cadaver studies on the basis of the following: head impact load magnitude and direction, loading rate, pre-alignment of the ATD head and neck, and degree of rotational head constraint. Thus, a comparison of the ATD sensor outputs for upper load cell Fz and lower load cell My values to probable column injury values of 7000 N and 150 N m, respectively appears scientifically reasonable, subject to the limitations described above. Roof Crush as a Correlate or Cause of Injury Two opposing views exist in the published literature regarding whether roof crush causes serious injury or whether it is simply associated or correlated with serious injury. One group of investigators has concluded that restrained occupants receive catastrophic head and neck injury from diving into the roof and making head-to-roof contact while the top of the vehicle/roof is striking/hitting the ground during a rollover.1,3,5,15,16 Other investigators have concluded that serious neck injuries to restrained occupants are directly caused by the dynamic intrusion of the roof structure into the occupant’s survival space during a rollover crash.12,13,28 The present study reports quantitative evidence of the temporal relationship between dynamic roof deformation, lap–shoulder seat belt loads, and restrained ATD neck loads, which provides further clarification in this scientific debate. Objective roof/pillar deformation occurred prior to the occurrence of Peak neck loads (Fz, My, Mx) for both driver and passenger ATDs in all three rollover tests. Prior to the occurrence of Peak neck loads, the driver and passenger ATD heads contacted the roof one or more times in all three tests; the magnitude of neck sensor responses ranged from 3 to 57% of the respective dummies’ Peak neck loads. Thus, the restrained driver and passenger ATDs did, indeed, “dive” into the roof while the roof was in contact with the ground, resulting in Local Maximum neck load values, which in no instance exceeded 57% of the Absolute Maximum (Peak) neck load. In particular, the “diving” neck loads (Local Maxima) for the far-side, passenger ATDs were only 2–13% of the Peak neck loads (Fz and My) for all three tests (Table 2, Fig. 4). In each test, the Peak neck loads of the far-side, passenger ATD occurred within 10 ms following the time of maximum objective roof crush for all three tests, which is consistent with published inertial effects of the ATD head in transmitting an axial force to the lower neck sensors.24 In contrast, Peak neck loads for the near-side, driver ATD occurred up to 65 ms following the time of occurrence of objective roof crush for all three tests. The onboard cameras also revealed observable roof crush into the far-side versus the near-side occupant compartment during the time of the peak neck loads, which was completely consistent with the timing and higher neck load magnitudes of the passenger, compared to driver, ATD for all three tests. The lap and shoulder belt load profiles for both driver and passenger ATDs provide further insights in the scientific debate regarding roof crush as a correlate versus a cause of serious injury. During the time interval of the Peak neck loads in the passenger ATDs for all three tests, the shoulder belt load, which should increase in magnitude as it resists torso augmentation, instead consistently decreased by 65–85% to only 44 N in some cases. The lap belt showed the same reduction in load during the time period that Peak neck load was recorded; it was reduced by 58% to approximately 156 N (Table 4, Figs. 8a and 8b). The belt load reductions occurred at the same time that on-board cameras clearly recorded the passenger ATD hanging upside down in the lap–shoulder belt and the roof crushing into the occupant space. This time of “observable” roof crush also occurred contemporaneously with the “objective” roof crush predicted by the accelerometers mounted on the SUV roof rails. The belt load that was recorded during each Peak neck load event for the far-side, passenger ATD was less than 50% of the Peak belt load in all three tests (Table 4). The shoulder and lap belt load cells indicated that the belts were off-loaded (i.e., loading decreased) as the roof crushed down on the ATD head and pushed the ATD back toward the seat cushion, away from the shoulder and lap belts, at the time of Peak neck loads (Fig. 9). Additionally, the deformation of the passenger (far-side) B-pillar lowered the upper attachment point (i.e., D-ring) for the belt, causing slack in the shoulder belt with a concomitant reduction in shoulder belt load. Figure 9The shoulder and lap belts are off loaded (i.e., load decreases) as the roof crushes down on the passenger head and pushes the dummy back toward the seat cushion, away from the shoulder and lap belts, at the time of injurious neck loads Repeatability and Reliability The FMVSS 208 dolly rollover test methodology has been criticized for its alleged lack of reliability and/or repeatability.15 The results of this study, however, demonstrated that the 208 rollover tests are very reliable when viewed from a vehicle-based, occupant protection frame of reference. Test-to-test comparisons of the FMVSS 208 rollover results revealed remarkable similarity in predicting the time of occurrence of Peak neck loads for both driver and passenger ATDs. The Peak neck Fz occurred in all three tests at 530 ± 15 ms for the driver ATD (Fig. 10 and Table 3) and 730 ± 15 ms for the passenger ATD (Table 3). The Peak neck My occurred in all three tests at 530 ± 18 ms for the driver ATD (Fig. 11 and Table 3) and 750 ± 21 ms for the passenger ATD (Table 3). The Peak value for Mx occurred in all three tests at 530 ± 18 ms for the driver ATD and 770 ± 13 ms for the passenger ATD. These small variations in time of occurrence of roof/pillar deformation and Peak neck loads are particularly remarkable given that (1) the data came from six different dummies and three different vehicles tested on three different days, (2) the differences in the time of occurrence of the Peak neck loads was ≤20 ms in an overall time interval of 1000 ms, and (3) each 1000 ms time interval included either a sampling rate of 20,000 data points (B190043 and B190042) or 12,500 (B180220). Figure 10The absolute maximum value for upper neck Fz occurred in all three tests at 530 ± 15 ms for the driver dummyFigure 11The absolute maximum value for lower neck My occurred in all three tests at 530 ± 18 ms for the driver dummy Quantitative analyses were undertaken to determine the significance probability of each pairwise comparison. Restated, the probability that the similarity of the maximum and minimum values of any two tests were related to something other than chance was analyzed (e.g., B190042-Fz vs. B190043-Fz, B190042-Fz vs. B190220-Fz, and B190043-Fz vs. B190220-Fz). Appendix 1 examines this issue and computes the probability for each pair of test metrics. This analysis revealed that there is a 93.1–98.6% probability that the differences in time occurrence of Peak neck loads between these results are NOT due to random chance alone. Stated more simply, there is a less than 7% probability that these differences are coincidental. Validity The constellation of driver vs. passenger neck loads measured in this test series of SUV rollovers was consistent with certain incidence trends of catastrophic head and neck injuries observed in real-world rollovers with restrained front seat occupants.9,21,28 None of the peak neck loads recorded in the near-side, driver ATD neck sensors exceeded the threshold values for probable column injury, which were used in this study. In contrast, the far-side, passenger ATDs recorded Peak neck loads that consistently exceeded the threshold for probable column injury due to neck flexion, My. These findings are consistent with 1992–1998 data obtained from NHTSAs National Accident Sampling System (NASS) database, which showed serious spinal injuries were more frequent for the far-side occupants (compared to the near-side), where the source was most often coded as roof, windshield, and interior.21 A probability of two common catastrophic spinal column injuries was predicted by the passenger ATD neck sensors in all three rollover tests. A combination load of compression and of flexion is associated with burst fractures and wedge compression fractures. This load profile was recorded by the upper and lower neck sensors in the far-side passenger ATD in B190042. Bilateral facet dislocation injuries have been associated with sagittal plane flexion.17 The flexion moments recorded by the passenger ATD lower neck sensors in tests B180220 and B190043 exceeded the threshold values for probable spinal column injury by 40–240%. Thus, the results of this rollover test series suggest that the FMVSS 208 dolly rollover test may be a valid predictor of serious spinal injuries to restrained occupants in real-world rollovers. The results of this study provide a unique data set that furthers understanding of probable spinal column injury mechanisms within a rollover crash environment. Such information may assist the scientific community and automotive engineers in recommending and designing appropriate intervention strategies to mitigate morbidity and mortality in rollover crashes. Moreover, these data may inform government agencies in formulating appropriate public safety policy to improve rollover crash protection for restrained occupants. Conclusions During each of the three FMVSS 208 dolly rollover tests of Ford Explorer SUVs, the far-side, passenger ATDs exhibited Peak neck compression and flexion loads, which indicated a probable spinal column injury in all three tests. In those same tests, the near-side, driver ATD neck loads never predicted a serious injury. In all three tests, objective roof/pillar deformation occurred prior to the occurrence of Peak neck loads (Fz, My) for far-side, passenger ATDs, and Peak neck loads were predictive of probable spinal column injury. The production lap and shoulder seat belts in the SUVs, which restrained both driver and passenger ATDs, consistently allowed ATD head contact with the roof while the roof was contacting the ground during this 1000 ms test series. Local peak neck forces and moments were noted each time the far-side, passenger ATD head contacted (“dived into”) the roof while the roof was in contact with the ground; however, the magnitude of these local maxima was only 2–13% of Peak neck loads in all three tests. “Diving-type” neck loads were not predictive of spinal column injury for either driver or passenger ATD in any of the three tests. When viewed from a vehicle-based, occupant protection frame of reference, the existing FMVSS 208 dynamic rollover test is a repeatable, reliable, and valid test method to evaluate the risk of probable spinal column injury in rollover crash environments.	[ "suv", "roof crush", "crash test", "restraint system", "spinal injury" ]	[ "P", "P", "P", "P", "P" ]
Breast_Cancer_Res-5-6-314403	Transforming growth factor beta-regulated gene expression in a mouse mammary gland epithelial cell line	Background Transforming growth factor beta (TGF-β) plays an essential role in a wide array of cellular processes. The most well studied TGF-β response in normal epithelial cells is growth inhibition. In some cell types, TGF-β induces an epithelial to mesenchymal transition (EMT). NMuMG is a nontransformed mouse mammary gland epithelial cell line that exhibits both a growth inhibitory response and an EMT response to TGF-β, rendering NMuMG cells a good model system for studying these TGF-β effects. Introduction Transforming growth factor beta (TGF-β) is the prototype of a large family of signaling molecules with more than 40 members. TGF-β signals through the type I receptor (TβRI) and the type II receptor (TβRII), and the Smad transcriptional regulators, as well as many other signaling pathways [1]. TGF-β plays an essential role in a wide array of cellular processes, including early embryonic development, cell growth, differentiation, motility, and apoptosis. In nontransformed epithelial cells, the most studied TGF-β response is growth inhibition. In some cell types, TGF-β induces an epithelial to mesenchymal transition (EMT). TGF-β plays an important, albeit complex, role in tumorigenesis. In many tumors of epithelial origin, cells become resistant to TGF-β-mediated growth inhibition. In parallel, TGF-β induces EMT in tumor cells, and increases tumor invasion and metastasis [2]. NMuMG is a nontransformed mouse mammary epithelial cell line that exhibits both a growth inhibitory and EMT response to TGF-β [3]. These properties render NMuMG cells a suitable model system for studying these TGF-β actions. Previous gene expression profiling studies of TGF-β action on cell lines utilized cDNA microarrays to characterize extracellular matrix-related genes in human dermal fibroblasts. The upregulation of these genes by TGF-β was Smad mediated [4]. Another study examining the genetic programs of epithelial cell plasticity directed by TGF-β utilized the human keratinocyte HaCaT cell line [5]. In this study, HaCaT cells were treated for up to 4 hours, and microarray gene expression profiling was performed using an array of 16,580 human cDNAs. The expression of 728 known genes (10% of those analyzed) was regulated by TGF-β within 4 hours of treatment. In the present study, we examined TGF-β regulation of gene expression in the NMuMG epithelial cell line. NMuMG cells were chosen as a model system because the results are more relevant to our studies employing transgenic mouse mammary tumor models, and because NMuMG cells are a well-established model system for studying both growth inhibition and EMT induced by TGF-β [3]. EMT represents a complex physiological process that includes dissolution of adherens junctions, a change to spindle-like cell morphology, cytoskeletal reorganization, increased cell motility, loss of epithelial markers, and induction of mesenchymal markers. EMT thus includes several cellular responses that may occur through different mechanisms and at different time points after TGF-β treatment. We consequently examined changes in gene expression at 1, 6, and 24 hours after TGF-β treatment. In agreement with previous studies with HaCaT cells [5], approximately 10% of the genes examined were regulated by TGF-β in our microarray studies that queried more than 15,000 mouse cDNAs. Our study demonstrates that by 1 hour, TGF-β suppresses the expression of multiple genes that play key roles in regulating cell cycle progression. At 6 and 24 hours, TGF-β enhances the expression of multiple genes involved in regulating cell shape and cell adhesion, and therefore may mediate TGF-β-induced EMT. Additionally, several novel TGF-β-regulated genes were identified that may play important roles in TGF-β responses. Materials and methods Cell culture and treatment NMuMG cells were cultured in Dulbecco's modified Eagle's medium containing 10% serum and 10 μg/ml insulin, until reaching 70–80% confluency. The cells were treated with 4 ng/ml TGF-β1 for 1, 6 or 24 hours. Untreated cells were used as reference samples at each time point, and were referred to as controls in the analyses. Four independent replicates of each experiment were performed. Cell morphology was examined under phase contrast using an Olympus CK40 microscope (Melville, NY, USA). Growth inhibition assay 3H-thymidine incorporation assays were performed using cells that had been treated for 24 hours with TGF-β. The cells were pulsed with 3H-thymidine for the final 2 hours of the treatment period. 3H-thymidine incorporation was quantitated using a LS 6500 Multi-Purpose Scintillation Counter (Beckman). Four replicates of each experiment were performed, and each 3H-thymidine incorporation experiment was performed in triplicate. RNA sample preparation NMuMG cells were lysed and homogenized in Trizol reagent (Invitrogen, Carlsbad, CA, USA), and total RNA was prepared according to the manufacturer's instructions. Two rounds of extraction were performed to isolate pure, high-quality total RNA. Thirty micrograms of total RNA from each sample were used for reverse transcription and cyanine (Cy)-dye incorporation. cDNA microarray hybridization The National Institutes of Aging (NIA) 15,000 cDNA microarray was printed on three contiguous slides by the Vanderbilt Microarray Shared Resource. The mouse 15,000 NIA microarray contains 15,247 genes of known and unknown function, developed at the NIA. Detailed descriptions of the gene array list, microarray hardware and procedures are available online . The reference RNA samples from nontreated NMuMG cells were labeled with Cy5, and the test RNA samples from NMuMG cells treated with 4 ng/ml TGF-β for 1, 6, and 24 hours were labeled with Cy3. The Cy3-labeled and Cy5-labeled samples were hybridized simultaneously to the same array. Four independent replicates of TGF-β treatment, RNA isolation and labeling, and microarray hybridization were performed. To estimate 'system noise' that may be due to differences in Cy-dye labeling between samples, we performed a 'self-to-self' hybridization, where the same control sample was labeled separately with Cy3 and Cy5, and then hybridized. Oligo dT for reverse transcription was synthesized by Invitrogen. SuperScript II reverse transcriptase (Invitrogen) was used. A Qiagen PCR purification kit was used for purification of probes (Qiagen, Valencia, CA, USA). Slides were prehybridized to eliminate nonspecific interactions in prehybridization solution: 1% BSA, 5 × sodium chloride/sodium citrate buffer (SSC), 0.1% SDS at 65°C for 45 min. Hybridizations were performed at 65°C for 14–16 hours in a humidified hybridization chamber (Corning, Acton, MA, USA). After hybridization, the slides were washed once each in solutions 1–3 (wash solution 1, 2 × SSC, 0.1% SDS; wash solution 2, 1 × SSC; wash solution 3, 0.1 × SSC) for 5 min at 55°C with gentle stirring/agitation. Washed slides were centrifuged in a conical tube for 5 min at 1600 rpm to dry them. Microarray slides were scanned using a Genepix 4000 B scanner (Axon Instruments, Union City, CA, USA) at a resolution of 10 μm, and the original data files were generated by GenePixPro software (version 4.0; Axon Instruments). Data analysis A merged 15,000 data file was constructed by combining the three original 5000 data files generated from the three contiguous slides on which the total 15,000 gene clones were printed. The GenBank accession number and the full annotation gene name for each gene were incorporated. Data analyses were performed using the GeneSpring 4.1.5 software package (Silicon Genetics, Redwood City, CA, USA). The normalization for the two-color cDNA microarray data was designed as follows. The net signal intensity in each channel (Cy3 or Cy5) was determined by subtracting the local background from signal intensity values. Each gene's measured intensity was divided by its control channel value (the reference RNA sample channel of Cy5). A Cy3/Cy5 ratio represents the relative abundance of a target transcript in TGF-β-treated and nontreated samples respectively. When the control channel value was below 1000.0, the data point was considered too weak and was discarded. Each sample was normalized to the 50th percentile of all measurements. The bottom 10th percentile was considered background, and was subtracted from all the other values. Each experimental result was interpreted as the average of the four replicates of each time point. Time-course gene expression profiles were required to be present in two of the four datasets, and to be present in at least two of the three time points. A signal to control channel ratio > 2.5 and a probability t test P < 0.05 was defined as upregulation by TGF-β. A signal to control channel ratio < 0.4 and a probability t test P < 0.05 was defined as downregulation. The annotated gene lists were constructed from all genes represented in the normalized dataset. The differentially expressed genes were classified into functional subgroups using Simplified Ontology in GeneSpring. For each pathway, pathway images were saved as GIF files and loaded into GeneSpring. Each pathway element was identified in the normalized data. Northern hybridization All cDNAs of the NIA mouse 15,000 microarray were cloned into the NotI/SalI sites of the ampicillin-resistant pSPORT1 vector (Invitrogen). The insert sizes ranged from 0.5 to 3 kb, with an average insert size of 1.5 kb. The Escherichia coli host strain used was DH10B. Plasmid DNA was isolated using the High Pure Plasmid Isolation Kit (Roche, Indianapolis, IN, USA). The inserts were verified for their correct annotation by DNA sequencing. The reverse pSPORT1 primer was used as the sequencing primer. To prepare cDNA probes, cDNA inserts were amplified using pSPORT1 primers: forward, GTTTTCCCAGTCACGACGTTG; reverse, TGA-GCGGATAACAATTTCACACAG. Plasmid DNA was used as the template for PCR amplification under the following conditions: initial denaturation at 95°C for 1 min; then 40 cycles of denaturation at 94°C for 30 s, annealing at 54°C for 30 s, and extension at 72°C for 3.5 min. A final extension was performed at 72°C for 7 min. To examine the quality and quantity of PCR products, they were separated on 1.2% agarose gels. Specific cDNA bands were recovered from the gels using QIAEX II Gel Extraction Kits (Qiagen). Seventy-five nanograms of cDNA was labeled with 32P using a Random Primed DNA Labeling Kit (Roche). For northern blot analysis, 10 μg total RNA was separated by electrophoresis on a 1% agarose, 0.66 M formaldehyde gel and transferred to Hybrid Nylon transfer membranes (Amersham Biosciences, Piscataway, NJ, USA). After transfer, the membrane was UV cross-linked (Stratalinker; Stratagene, La Jolla, CA, USA). All cDNA probes were hybridized overnight at 42°C in ULTRAhyb solution (Ambion, Austin, TX, USA). Each filter was washed twice for 5 min in 2 × SSC and 0.1% SDS at 42°C, followed by two 15 min washes each in 0.5 × SSC and 0.1% SDS at 42°C, followed by two 15 min washes in 0.1 × SSC, 0.1% SDS at 42°C. Equal loading of gel lanes was confirmed by hybridization with the house keeping gene 1B15. Results were visualized and quantitated using a FUJIFILM-FLA-5000 Phosphoimager (Fuji Photo Film Co. Ltd, Stamford, CT, USA). Results Biological and gene expression responses of NMuMG cells to TGF-β NMuMG cells react to TGF-β treatment with both an EMT response and a growth inhibitory response (Fig. 1). These characteristics render NMuMG cells an appropriate model system to study the signaling pathways mediating these TGF-β effects. To examine TGF-β regulation of genes involved in EMT and cell cycle control, we treated NMuMG cells with TGF-β for 1, 6, or 24 hours, isolated the total RNA, and performed gene expression profiling by cDNA microarray. The relative transcript abundance was expressed as Cy3/Cy5 ratios of signal intensities after background subtraction in each channel. Data analysis and quality control procedures are described in detail in Materials and methods. TGF-β treatment resulted in the upregulation or downregulation of 939 genes out of the 15,247 genes examined. Subtracting redundancy or signals too weak to score, this represents approximately 10% of the genes analyzed. The time course of TGF-β1-dependent changes in gene expression demonstrates a progressively greater alteration of gene expression over time (Fig. 2). In general, TGF-β1 regulation of cell cycle-related genes occurred with an onset of less than 1 hour, while TGF-β1 regulation of EMT-related genes occurred with a later onset, with changes first becoming apparent at 6 or 24 hours (see later). Figure 3 and Table 1 present the distribution of downregulated genes and of upregulated genes, and the sum of upregulated and downregulated genes at each time point. The total number of differentially expressed genes at each time point does not differ significantly. The 6 and 24 hour time points shared more TGF-β1-regulated genes than were shared between the 1 and 6 hour time points, or the 1 and 24 hour time points. Full lists of TGF-β-regulated genes at each time point can be accessed on our website . Functional class distribution of the novel and known TGF-β1-regulated genes Of the 939 genes regulated by TGF-β1, approximately one-third (350 genes) can be classified into functional groups. The classified genes were further separated into two tables of genes previously reported or not previously reported to be regulated by TGF-β. This was determined by performing a Medline search using the gene name and TGF-β as keywords. Tables of the complete dataset with each gene in its functional category are presented on our website . Tables 1 and 2 present some selected examples of known or unknown TGF-β-regulated genes, respectively. The genes involved in cell cycle regulation and cell adhesion (see later) are included. Northern analysis verification of microarray data To validate the microarray results, we subjected nine of the novel TGF-β-regulated genes to northern analysis. Of the nine genes examined, seven exhibited TGF-β1-induced changes in gene expression that matched those observed in the microarray experiments. The seven novel TGF-β1-regulated genes verified by northern analysis include Akt, mCalpain, RhoB, PR53, actinin 3, Ikki (Fig. 4) and the IQ motif-containing GTPase-activating protein 1 (IQGAP1). IQGAP1 showed upregulation in NMuMG cells after 1, 6 and 24 hours of treatment with TGF-β, supporting the microarray data. As reported by other investigators [6], we detected multiple IQGAP1 transcripts. The IQGAP1 northern blot is thus not shown in Fig. 4. Akt and RhoB were reported previously to be regulated by TGF-β at the translational level [7,8], but not at the transcriptional level as we show here. Some of the TGF-β1-induced changes in gene expression were also verified by real-time RT-PCR (data not shown). In further support of the reliability of the microarray experiments, genes that were represented multiple times in the array as different clones exhibited identical expression profiles (examples can be found in the supplementary tables at our website: . In addition, our expression profiling results are consistent with previously reported expression patterns of TGF-β1-responsive genes [1,5]. Examination of cell cycle control and cell adhesion-related genes by microarray There are multiple ways to explore the biological significance of results obtained in microarray experiments. One may organize the genes into functional groups or signaling pathways to examine the regulation of clusters of genes. We were interested in identifying TGF-β-regulated genes involved in the induction of EMT and cell cycle arrest. In our study in NMuMG cells, the expression of many cell cycle regulators was altered by TGF-β. One hour after TGF-β treatment, cyclin D2, cdk4, E2F5, cyclin A, and c-myc were downregulated. Six hours after TGF-β treatment, cyclin D2, p107, E2F5, c-myc, and Id2 were downregulated. Twenty-four hours after TGF-β treatment, cyclin D2, p107, E2F5, cyclin A, cyclin B, cyclin H, and Id2 were downregulated (Fig. 5). These data are consistent with these genes serving as mediators of TGF-β-induced growth arrest. The microarray results also revealed upregulation of a number of genes that play a role in cell–cell and cell–matrix adhesion and EMT. At 1 hour, c-Jun, β-integrin, and IQGAP were upregulated. At 6 hours, actin, calpain, β-integrin, fibronectin, collagen, myosin light chain, α-actinin, β-catenin and IQGAP were upregulated. At 24 hours, c-Jun, actin, calpain, α-integrin, β-integrin, fibronectin, collagen, α-actinin, β-catenin and IQGAP were upregulated. The only gene in this category observed to be downregulated was E-cadherin at 6 hours (Fig. 6). The results are consistent with these genes playing a role in TGF-β1-mediated changes in cell–cell and cell–substratum adhesion. Our results are consistent with previous studies, but we identified several novel TGF-β-regulated genes that could play important roles in both EMT and cell cycle regulation, including RhoB, mCalpain, actinin 3, IQGAP1, Ikki and protein phosphatase 2A (PP2A)-PR53. Discussion The mechanisms by which TGF-β signals are not fully understood. In this regard, gene expression profiling has been used successfully by a variety of investigators to explore genetic events involved in TGF-β signaling. Zavadil and colleagues detected genetic programs of epithelial cell plasticity directed by TGF-β [5]. Verrecchia and colleagues identified novel TGF-β/Smad gene targets in dermal fibroblasts [4]. Chen and colleagues found defective repression of c-myc in breast cancer cells, suggesting that this event is at the core of the TGF-β growth arrest program [9]. Shen and colleagues discovered that the activity of the guanine exchange factor NET1 is essential for TGF-β-mediated stress fiber formation [10]. In the present study, we have selected the NMuMG nontransformed mouse mammary gland epithelial cell line as a model system to identify TGF-β-regulated target genes involved in the control of growth inhibition and EMT. To identify TGF-β-regulated genes, we employed a cDNA microarray that represents 15,247 genes of known and unknown function. We found that the expression of 939 of the genes was altered by TGF-β regulation. This represents about 10% of the genes examined. Many of the genes identified were previously determined to mediate key biological responses to TGF-β. In addition, several novel TGF-β-regulated genes were identified. The present report provides a comprehensive view of the coordinated regulation of genetic programs induced by TGF-β during the processes of cell cycle arrest and EMT. A remarkably large number of genes undergo rapid changes in expression after TGF-β treatment. This report adds to the database of hundreds of genes that are regulated by TGF-β [5]. In addition, we have identified several novel TGF-β-regulated genes, and have verified these results in northern blotting experiments. Table 1 presents previously identified TGF-β-regulated genes, many of which are known to play important roles in TGF-β induction of EMT and cell cycle arrest. Briefly, in the cellular communication category, platelet-derived growth factor was upregulated by TGF-β. Multiple genes involved in cell adhesion were upregulated by TGF-β, including integrins (integrin α3A, integrin β, integrin β5), integrin-linked kinase, and the integrin ligand fibronectin. These results are consistent with TGF-β activating integrin-dependent cell adhesion. TGF-β upregulation of α-actinin, β-tropomyosin 2, and p120 catenin may play a role in the induction of EMT. RhoB was upregulated at 6 and 24 hours. RhoB is an endosomal small GTPase, and it functions as a negative regulator of integrin and growth factor signals [11]. Our previous results [8] and those of other workers [12] suggest a novel mechanism of tumor suppressive regulation by TGF-β, and implicate RhoB as a negative regulator of TGF-β signal transduction. TGF-β downregulation of the helix–loop–helix DNA binding protein Id2 and of the inhibitor of DNA binding Idb3 were previously identified as important events in TGF-β-regulated signaling. These proteins belong to a family of helix–loop–helix proteins that, in general, function as positive regulators of cell growth and as negative regulators of cell differentiation in many cell types [13]. TGF-β also regulates many genes involved in the progression of cancer. The proto-oncogene Akt is upregulated at 1 hour at the mRNA level. The Gro1 proto-oncogene is downregulated by TGF-β at all the time points examined. The roles that these gene products play downstream of TGF-β during tumorigenesis have not been extensively studied [14,15]. Well-established TGF-β responses such as downregulation of c-myc and upregulation of the Jun oncogene were also observed. TGF-β is a potent growth inhibitor of epithelial cells. TGF-β-mediated growth inhibition has been associated with effects on G1 phase cyclins, cyclin-dependent kinases (Cdks), Cdk-activating kinase, Cdc25A, and Cdk inhibitors. TGF-β potently inhibits the activities of both the cyclin D–Cdk4/Cdk6 and cyclin E–Cdk2 complexes, resulting in the hypophosphorylation of pRbs and in the decreased transcriptional activity of E2Fs. TGF-β does not affect the expression levels of Cdk4, Cdk6, or Cdk2 in exponentially proliferating cells. TGF-β decreases the expression levels of cyclin A, which is probably the result of cell cycle arrest in the late G1 phase. TGF-β has been associated with upregulation of Cdk inhibitors p15Ink4B, p21Cip1, and p27Kip1. The c-myc proto-oncogene has been implicated in TGF-β-mediated growth inhibition. TGF-β rapidly inhibits c-myc expression in a wide variety of cell types [1]. c-Myc downregulation is a key event in the TGF-β program of growth inhibition [9]. Additional mechanisms mediating TGF-β-induced cell cycle arrest continue to be uncovered. In our study, examination of cell cycle genes demonstrated TGF-β1-induced downregulation of cyclin D2, c-myc, Id2, p107, E2F5, cyclin A, cyclin B, and cyclin H. These data are consistent with these genes serving as mediators of TGF-β-induced growth arrest. Of the genes that could be placed into functional categories, the expression of the majority of the genes was not previously reported to be regulated by TGF-β. Table 2 presents some examples of genes that we have identified in this study to be regulated by TGF-β that are involved in cell signaling. Gadd45g was upregulated. Gadd45g specifically interacts with the Cdk1-cyclin complex, B1 resulting in inhibition of kinase activity [16]. The IQGAP1 gene is a target molecule of Cdc42 and Rac1. IQGAP negatively regulates the E-cadherin–catenin complex-based cell–cell adhesion by dissociating α-catenin and/or β-catenin, a key molecule that links the E-cadherin-catenin complex to the actin cytoskeleton. These data suggest that TGF-β upregulation of IQGAP may be one of the mechanisms by which TGF-β regulates cell–cell adhesion [17]. Calpains are a family of Ca2+-dependent intracellular cysteine proteases, including the ubiquitously expressed micro-calpains and m-calpains. Calpain activity may be regulated by targeting to specific adhesion-related substrates [18]. Our results also demonstrate that TGF-β regulates multiple genes involved in regulation of the actin cytoskeleton. Some of these components include actinin α3, the Arp2–Arp3 complex, and myosin light chain 2. Four different isoforms of tropomyosin were upregulated. Actinin α3 is a calcium-dependent cytoskeletal protein with an actin-binding domain. Actinin α3 is associated with adherens junctions and desmosomes, together with E-cadherin, α-catenin, β-catenin and γ-catenin, vinculin, α-actinin, and polymerized actin. The genes α-actinin, myosin light chain 1/myosin light chain 2, and tropomyosin are components of myofilaments involved in cell contractility and motility. During our studies we identified several genes not previously reported to be regulated by TGF-β that are regulated by TGF-β in NMuMG cells. These novel TGF-β-regulated genes include IQGAP1, mCalpain and actinin α3, which were discussed earlier, and also Ikki and PP2A-PR53. For the first time, we show in the present study that Ikki is downregulated by TGF-β within 1 hour of treatment, and that downregulation persisted for at least 24 hours. NF-κB activation depends on the phosphorylation and degradation of its inhibitor protein, IkB. The phosphorylation of IkBα is initiated by an IkB kinase complex that includes a catalytic heterodimer composed of IkB kinase 1 (IKK-1) and IkB kinase 2 (IKK-2) [19]. A novel inducible IkB kinase (IKKi) has recently been described. IKKi is functionally and structurally distinct from the constitutively expressed IKK-1 and IKK-2. In contrast to TGF-β, tumor necrosis factor alpha stimulated IKKi expression, an effect that persisted for at least 24 hours [20]. The biological significance of TGF-β downregulation of IKKi is currently under investigation. PP2A is a multifunctional serine/threonine phosphatase that is critical to many cellular processes, including cell cycle regulation and signal transduction. PR53 belongs to one of the families of PP2A regulatory subunits. The significance of TGF-β regulation of PR53 expression is unknown [21]. Altogether, our results are consistent with the important role of TGF-β in regulating focal adhesions, integrin-based adhesion, actin cytoskeletal architecture, and cell motility. TGF-β plays a profound role in the dedifferentiation of epithelial cells, causing depolarization, disruption of epithelial interactions, altered expression of extracellular matrix proteins, rearrangement of the cytoskeleton, and formation of actin stress fibers. These TGF-β responses provide cells with increased metastatic potential and with increased motility. Over the past few years, tremendous progress has been made in identifying signal transduction pathways activated by TGF-β family members. In the present study, we have combined microarray analysis with the examination of specific signaling pathways. Because of the important role that TGF-β plays in controlling cell proliferation and cell adhesion, we used the signaling elements involved in these processes as a template for the analysis of our microarray results. Many of the genes that we identified as being regulated by TGF-β were previously reported to be downstream targets of TGF-β. We also discovered several novel TGF-β target genes known to play roles in regulating cell adhesion, and thus EMT, and also cell cycle regulation. Examining the biological functions of these novel TGF-β target genes will increase our knowledge of the mechanisms by which TGF-β mediates its cellular effects. Furthermore, our results indicate several possible points of convergence between TGF-β signaling and other major intracellular and intercellular signaling systems. Conclusion A 15,247 cDNA microarray was used to examine TGF-β-regulated gene expression in mouse mammary NMuMG cells. Expression of 10% of the genes examined (939) was altered by TGF-β treatment. We have reported a comprehensive analysis of the coordinated regulation of genetic programs induced by TGF-β in mammary epithelial cells during the processes of cell cycle arrest and EMT. In addition, several genes previously not known to be regulated by TGF-β at the transcriptional level were identified, including Akt, RhoB, IQGAP1, mCalpain, actinin α3, Ikki and PP2A-PR53. Competing interests None declared. Abbreviations BSA = bovine serum albumin; Cdk = cyclin-dependent kinase; Cy = cyanine; EMT = epithelial to mesenchymal transition; IKKi = inducible IkB kinase; IQGAP1 = IQ motif-containing the GTPase-activating protein 1; NF = nuclear factor; NIA = National Institutes of Aging; PCR = polymerase chain reaction; PP2A = protein phosphatase 2A; RT = reverse transcriptase; SSC = sodium chloride/sodium citrate (buffer); TGF-β = transforming growth factor beta.	[ "mammary gland", "epithelial", "cdna microarray", "adhesion", "cell cycle control" ]	[ "P", "P", "P", "P", "P" ]
Histochem_Cell_Biol-4-1-2248609	Clathrin-independent endocytosis: from nonexisting to an extreme degree of complexity	Today it is generally accepted that there are several endocytic mechanisms, both the clathrin-dependent one and mechanisms which operate without clathrin and with different requirements when it comes to dynamin, small GTP-binding proteins of the Rho family and specific lipids. It should be noted that clathrin-independent endocytosis can occur even when the cholesterol level in the membrane has been reduced to so low levels that caveolae are gone and clathrin-coated membrane areas are flat. Although new investigators in the field take it for granted that there is a multitude of entry mechanisms, it has taken a long time for this to become accepted. However, more work needs to be done, because one can still ask the question: How many endocytic mechanisms does a cell have, what are their function, and how are they regulated? This article describes some of the history of endocytosis research and attempts to give an overview of the complexity of the mechanisms and their regulation. Introduction The existence of different endocytic mechanisms is becoming of increasing interest, for instance in relation to growth and differentiation, cell adhesion, regulation of the activity of signalling receptors, drug delivery, and entry of pathogens. The most well-studied endocytic mechanism involves clathrin-coated pits and vesicles, and impressive amounts of details are now known about clathrin-dependent endocytosis (Schmid and McMahon 2007; Benmerah and Lamaze 2007; Ungewickell and Hinrichsen 2007). Initially, the idea was that the formation of the clathrin basket was the driving force during membrane internalization, and quantitative studies suggested that the pathway could account for everything endocytosed (Doxsey et al. 1987). With basic ideas such as “biology is simple” and questions like, “how can a vesicle form without clathrin?”, it was difficult and it took time to get acceptance for the existence of alternative endocytic pathways (Montesano et al. 1982; Sandvig et al. 1985, 1987; Moya et al. 1985; Payne et al. 1988; Sandvig and van Deurs 1991, 1994; Damke et al. 1995). Clathrin-dependent endocytosis can be inhibited both by removing clathrin from the plasma membrane by a treatment involving hypotonic shock and depletion of cytosolic potassium (Larkin et al. 1983), and by “freezing” the clathrin coat at the surface by acidification of the cytosol (Sandvig et al. 1987, 1988). The finding that endocytosis of for instance the protein toxin ricin continued under both conditions, suggesting that clathrin-independent endocytosis was responsible for this uptake, was apparently difficult to accept for a number of investigators during a 10 year period (1985–1995). New techniques in molecular biology have been crucial for the elucidation of endocytic processes, and when in 1995 it was published that the dynamin-mutant K44A inhibited clathrin-mediated endocytosis and that the cells still endocytosed (Damke et al. 1995), the complexity of endocytosis started to become apparent to the community in general. Not only did clathrin-independent endocytosis exist, it could also be regulated (Damke et al. 1995). We now know that a number of signalling pathways can regulate clathrin-independent endocytosis (for review, see Sandvig and van Deurs 2005; Marsh and Helenius 2006; Mayor and Pagano 2007). Despite accumulating evidence for several endocytic mechanisms, partly obtained by studies of toxins and viruses (Marsh and Helenius 2006; Sandvig and van Deurs 2002), many text-books in cell biology still operate with only two endocytic mechanisms, uptake from clathrin-coated pits and caveolae. The aim of this brief review is to outline the extreme complexity of clathrin-independent endocytosis as it appears today. Clathrin-independent endocytosis: expansion of the field The number of clathrin-independent endocytic pathways revealed has been increasing with time. Mayor and Pagano (Mayor and Pagano 2007) recently subdivided the different mechanisms into four types: RhoA-regulated, cdc42-regulated, and Arf6-dependent mechanisms, and caveolar uptake. Of these, the RhoA-regulated mechanism and caveolar endocytosis also require the GTP-binding protein dynamin (Fig. 1a). However, there are already reports of mechanisms that may not necessarily fit into any of these four categories. For instance, clathrin-independent internalization of proteoglycans seems to depend not only on dynamin, but also on flotillin (Payne et al. 2007), and internalization of Herpes Simplex Virus Protein VP22 is independent of clathrin, but dependent on dynamin and Arf6 (Nishi and Saigo 2007). Also, studies of virus and toxin uptake indicate that there could be even more mechanisms (Marsh and Helenius 2006; Sandvig 2001; Sandvig and van Deurs 2005). Since cells may compensate for loss of one endocytic pathway by upregulating another (Damke et al. 1995), one can of course also question to which extent interfering with the endocytic machinery of physiologically relevant pathways in different ways, might create new pathways or cause upregulation of those normally playing only a minor role. Today, more specific methods such as RNA interference (RNAi) are used to interfere with a given process. However, also RNAi-mediated removal of a given molecule can induce upregulation of related molecules (Utskarpen et al. 2006). Moreover, upon removal of a membrane-associated molecule one may ask: To which extent do we actually know that the endocytic process is blocked? Could the rest of the machinery allow uptake to continue with different kinetics and be interpreted as a “new mechanism”? Similarly, an old objection to the notion that removal of clathrin by potassium-depletion (Larkin et al. 1983; Moya et al. 1985; Madshus et al. 1987; Sandvig et al. 1985) revealed endocytosis normally operating in parallel to the clathrin-dependent pathway, was that the rest of the endocytic machinery originally belonging to the clathrin-dependent machinery was still operating. A similar critical attitude should be applied on today’s interpretations. Importantly, both toxins and viruses can modify endocytic events, in some instances by inducing signalling, even though they may bind to glycolipid receptors (Marsh and Helenius 2006; Lauvrak et al. 2006; Römer et al. 2007). Fig. 1An overview of a factors involved in different types of endocytosis, and b schematic drawing of polarized MDCK cells where some proteins/enzymes involved in regulation of apical clathrin-independent endocytosis have been listed. It should be noted that apical clathrin-independent endocytosis may consist of more than one mechanism and that polarized MDCK cells have caveolae only at the basolateral surface. For references, see the text Internalization of membrane components and ligands by endocytosis thus have different requirements when it comes to dependency of dynamin, caveolin, flotillin, and small GTPases, and in addition, cholesterol and other specific lipids play a role (see below). To which extent the different endocytic mechanisms vary depending on cell type, growth conditions and the extent of cell polarization is to a large extent not known, and has to be addressed in future studies. It should be stressed that most of our knowledge about clathrin-independent endocytic mechanisms derives from cells in culture. The work of combining different results into clearly defined models is far from finished. In 2005 it was published that CtBP3/BARS drives membrane fission in processes where dynamin is not involved (Bonazzi et al. 2005). Interfering with the action of this protein reduces fluid phase uptake, as previously reported for cdc42-dependent uptake (Mayor and Pagano 2007). However, whether cdc42 and CtBP3/BARS affect the same process does not seem to be known. The possibility exists that CtBP3/BARS might indirectly affect endocytosis (Mayor and Pagano 2007). Furthermore, when inhibiting endocytic uptake of a ligand, it can be difficult to measure whether it is movement of the ligand into the membrane invagination giving rise to the endocytic vesicle that is reduced or whether it is the formation of the vesicle that is the step affected. To classify the different mechanisms properly, more information is required (see Fig. 1a, b). It is important to note that polarized cells have a differential regulation of clathrin-independent endocytosis at the two poles, and regulators of apical clathrin-independent endocytosis have been indicated in Fig. 1b (separate chapter below). Interestingly, caveolae are only present on the basolateral side in the polarized epithelial cell line MDCK (Verkade et al. 2000; Vogel et al. 1998). In the following sections we will describe various aspects of the clathrin-independent mechanisms in more detail. Membrane lipids required for clathrin-independent endocytosis In addition to a large number of proteins, also membrane lipids are important for clathrin-independent endocytosis. Interfering with sphingolipid synthesis may affect different types of clathrin-independent endocytosis by mechanisms poorly understood (Cheng et al. 2006). Importantly, sphingomyelin seems to be required both for recruitment of RhoA and cdc42 to the membrane and for endocytosis involving these GTP-binding proteins (Cheng et al. 2006). Cdc42-dependent uptake is independent not only of RhoA, but also of Rac, clearly separating it from macropinocytosis which requires Rac activity (for review, see (Hanzal-Bayer and Hancock 2007; Chadda et al. 2007). Interestingly, sphingolipid synthesis, or rather the production of glycosphingolipids was found to be required for transport of caveolin to the plasma membrane and for caveolae formation (Cheng et al. 2006). A change in the fatty acid composition of membrane lipids can also be expected to change the properties of the membrane and its interaction with cytosolic proteins of importance for endocytosis. It was recently found that incubation of cells with unsaturated fatty acids affect the uptake of the glycolipid-binding Shiga toxin which is internalized both by clathrin-dependent and clathrin-independent endocytosis (Spilsberg et al. 2007). A possible explanation for this finding is that there is a change in the amount and/or size of rafts or their ability to become associated with the endocytic machinery. Lipid rafts play a central role in membrane structure and function (Rajendran and Simons 2005), and it is therefore not surprising that the fatty acid composition may affect endocytosis. Raft-dependent uptake have recently been reviewed (Marsh and Helenius 2006; Lajoie and Nabi 2007; Hanzal-Bayer and Hancock 2007). One can often see the generalized statement: clathrin-independent endocytosis is cholesterol-dependent (Hanzal-Bayer and Hancock 2007). This is true for processes such as macropinocytosis (Grimmer et al. 2002), uptake from caveolae, Rho-dependent uptake of IL2 (a recent review has been written by Lajoie and Nabi 2007), and cdc42-dependent endocytosis where it was reported that there is a cholesterol-dependent activation of cdc42 which in turn mediates actin polymerization required for endocytosis via this pathway (Chadda et al. 2007). However, a normal concentration of cholesterol does not seem to be required for uptake of all ligands taken in by clathrin-independent endocytosis (Rodal et al. 1999). When enough cholesterol has been extracted from the plasma membrane for the caveolae to disappear and the clathrin-coated pits to flatten, endocytosis of the plant toxin ricin still occurs (Rodal et al. 1999). Furthermore, when Cholera toxin uptake into HeLa cells is reduced by expression of the dynamin-mutant K44A or by induction of antisense to clathrin heavy chain in BHK cells, endocytosis of this toxin can be increased by extraction of cholesterol with methyl-β-cyclodextrin (Sandvig et al. 2004). Similarly, uptake of Shiga toxin is increased by treatment with methyl-β-cyclodextrin after block of clathrin-dependent uptake by induced expression of antisense to clathrin heavy chain in BHK cells (Fig. 2). One possible explanation is that these toxins are retained at the cell surface by being associated with rafts, and that release from these structures by treatment with methyl-β-cyclodextrin actually permits endocytosis of the toxins. Also, uptake of proteoglycan-binding ligands, reported to be clathrin- and caveolin-independent, but dynamin- and flotillin-dependent, were unaffected by filipin and nystatin, drugs known to disrupt lipid rafts (Payne et al. 2007). It should be noted that cholesterol-depletion can inhibit the formation of invaginated clathrin-coated pits by a so far unknown process (Rodal et al. 1999; Subtil et al. 1999), and cholesterol-dependency is therefore not necessarily a tool to easily distinguish between endocytic processes. Fig. 2Shiga toxin endocytosis is strongly increased upon methyl-β-cyclodextrin-treatment of BHK cells induced for expression of antisense to clathrin heavy chain. BHK cells with inducible expression of antisense to clathrin heavy chain was preincubated with methyl-β-cyclodextrin (10 mM) for 30 min. Then biotin- and TAG-labeled Shiga toxin (25 ng/ml) was added and the cells were incubated for 20 min. Surface-bound and endocytosed Shiga toxin was quantified as previously described for Cholera toxin (Torgersen et al. 2001). The data are presented as internalized toxin as percent of total cell-associated toxin (mean ± SD, n = 4). Endocytosis of transferrin was performed in parallel to verify the inhibition of clathrin-dependent endocytosis upon induction of antisense to clathrin heavy chain. Upon induction, transferrin uptake was reduced by 95% Role of actin in endocytosis Recruitment of actin to clathrin-coated pits and subsequent actin-driven formation and internalization of endocytic vesicles and the appearance of actin “comet tails” on such vesicles is well-established and appears to be a highly complex process involving several protein systems (McPherson 2002; Kaksonen et al. 2005, 2006; Smythe and Ayscough 2006). Interestingly, a crucial role of actin in clathrin-independent endocytosis is now becoming evident as well (Kaksonen et al. 2006; Yarar et al. 2007; Lanzetti 2007), and it seems likely that actin will turn out to be involved in all clathrin-independent endocytic mechanisms. However, in some cases actin might function to facilitate the process and not be strictly required. It has for instance been reported that there is a difference between cells growing in monolayers and cells in suspension when it comes to actin requirement for clathrin-dependent endocytosis (Fujimoto et al. 2000). One of the most studied examples of a clathrin-independent endocytic mechanism that is based on actin polymerization, is macropinocytosis (see below), a major fluid-phase uptake pathway. Macropinocytosis An increasing complexity of the formation of macropinosomes is becoming evident. Thus, it has long been known that GTP-bound Rac stimulates accumulation of actin filaments at the plasma membrane and that it is involved in plasma membrane ruffling induced by growth factors (Ridley et al. 1992) (Fig. 1a). Rac in turn activates the Arp2/3 protein complex and WAVE, a process involved in ruffling (Gao et al. 2007). In addition to Rac, also Ras, Src and phosphoinositide (PI) 3-kinase promote macropinocytosis (Gao et al. 2007), and a role of N-WASP and the sorting nexin 9 (SNX9) in assembly and organization of actin during ruffle formation and macropinocytosis has recently been shown (Yarar et al. 2007). Also, histone deacetylase 6 (HDAC6) is associated with plasma membrane ruffles and formation of macropinosomes upon stimulation with growth factors, processes that were inhibited in HDAC6-deficient cells. Interestingly, a substrate of HDAC6, the heat shock protein Hsp90, seemed to be involved in ruffling and macropinocytosis as well (Gao et al. 2007). Moreover, a clathrin-independent process reminiscent of macropinocytosis has recently been reported (Orth et al. 2006). Here, actin-dependent circular dorsal ruffles or dorsal membrane waves are formed upon growth factor stimulation, leading to significant receptor internalization. Like in ordinary macropinocytosis, PI 3-kinase is involved, but interestingly, also dynamin, which is a part of the actin-polymerizing Arp2/3-N-WASP-cortactin system and a large number of other proteins are required for this internalization mechanism (Orth et al. 2006; Orth and McNiven 2006; McNiven 2006). RhoA-, cdc42- and CTBP3/BARS-dependent uptake The RhoA-dependent uptake of IL-2 receptors, γc-cytokine receptor and IgE receptor is a dynamin-dependent process (Mayor and Pagano 2007) (Fig. 1). Whether RhoA is required for the correct sorting of ligand or whether it is involved in the endocytic process as such is however being discussed (Mayor and Pagano 2007). Data from polarized cells showing that ricin uptake and fluid phase endocytosis at the apical side of polarized MDCK cells are also dependent on RhoA (Garred et al. 2001), suggest that it is the endocytic mechanism itself that is RhoA-dependent. A cdc42-regulated process, which operates independently of dynamin, has been reported to be involved in uptake of glycosylphosphatidylinositol (GPI)-anchored proteins as well as being responsible for a major fraction of fluid-phase uptake in the cell (Kalia et al. 2006). GPI-anchored proteins were found to be internalized into CHO cells by a cdc42-dependent, but Arf6-independent mechanism. The internalized membrane ended up in vesicles without EEA1, Rab4, or Rab5, but these vesicles were found to fuse with normal endosomal compartments via a Rab5/PI3-kinase dependent machinery. That a GPI-anchored protein and its ligand can enter independently of caveolae and dynamin and reach an acidic compartment, was first shown in studies of a GPI-anchored diphtheria toxin receptor (Skretting et al. 1999). The finding that diphtheria toxin bound to this GPI-linked receptor intoxicates cells, demonstrates that the toxin has been transferred to an acidic compartment facilitating low pH-induced conformational change of diphtheria toxin and translocation of the toxin to the cytosol. The finding that the cdc42-dependent mechanism is responsible for most fluid uptake (Kalia et al. 2006) is surprising, considering all the other uptake mechanisms now reported, not least macropinocytosis. It was recently shown that cdc42 is associated with the recycling endosome and that also Par proteins play a role in the endocytic process (Balklava et al. 2007). To which extent there is an indirect effect on the uptake from the plasma membrane when mutants of cdc42 are studied, is not quite clear. Also, as mentioned earlier, since CtBP3/BARS is involved in dynamin-independent endocytosis (Bonazzi et al. 2005), and also endocytosis regulated by this protein can account for a considerable amount of fluid phase uptake, it should be elucidated to which extent cdc42 and CtBP3/BARS regulated processes are related. Arf6-dependent uptake ADP-ribosylation factor 6 (Arf6) has been identified as a key regulator of a clathrin- and dynamin-independent endocytic pathway mainly studied in HeLa and COS-7 cells, and responsible for uptake and recycling of a collection of integral plasma membrane proteins devoid of conventional endocytic signals (see below). The number of cargo proteins reported to be internalized by Arf6-regulated endocytosis in these cells is increasing and includes the plasma membrane proteins Major Histocompatibility Complex Class I (MHCI) and interleukin-2 receptor α-subunit/Tac (Radhakrishna and Donaldson 1997), β1-integrin (Brown et al. 2001), M2 muscarinic acetylcholine receptor (Delaney et al. 2002) and certain GPI-anchored proteins, such as CD59 (Naslavsky et al. 2004). Internalized cargo can recycle to the plasma membrane via the tubular Arf6-positive recycling compartment (Radhakrishna and Donaldson 1997) or, alternatively, be delivered to Rab5 and EEA1 positive early endosomes for further transport to late endosomes and lysosomes for degradation (Naslavsky et al. 2003). Recently, rafts were reported to be recycled by an Arf6-dependent process (Balasubramanian et al. 2007). Interestingly, the Herpes Simplex Virus protein VP22 was in CHO-K1 and HeLa cells shown to be endocytosed by a mechanism dependent on Arf6, dynamin and lipid rafts, but independent of Rho-GTPases (Nishi and Saigo 2007). A new class of trafficking motifs in non-clathrin-dependent cargo proteins have been identified and suggested to prevent internalization and/or facilitate recycling of proteins carrying these, via the Arf6-regulated pathway by enhancing their association with the Arf6-specific guanine nucleotide exchange factor EFA6 (Gong et al. 2007). In certain cell types, rather than defining a separate endocytic pathway, Arf6 regulates clathrin-dependent uptake. In Hek293 cells, for example, Arf6 regulates uptake of several G-protein coupled receptors by clathrin-dependent endocytosis (Houndolo et al. 2005), and there is also evidence for complex formation between Arf6, the β-subunit of AP2 and the heavy chain of clathrin (Poupart et al. 2007). The data furthermore suggest that Arf6 regulates the angiotensin AA type 1 receptor by recruiting AP2 and clathrin. Arf6 has also in HeLa cells been proposed to have a direct function in clathrin coat recruitment to the plasma membrane. In fact, Arf6-GTP has recently been demonstrated to interact with AP-2 both in vitro and in vivo, and a clathrin-interacting Arf6-specific GTPase activating protein (called SMAP), affecting clathrin-dependent endocytosis only, has also been identified (Paleotti et al. 2005; Tanabe et al. 2005). Interestingly, in polarized MDCK cells, Arf6 associated with the guanine nucleotide exchange factor ARNO regulates endocytosis exclusively at the apical surface by recruiting actin to clathrin-coated pits (Hyman et al. 2006). The present picture is complicated as many of the roles ascribed to Arf6 within the endocytic system seem to be not only cell type dependent, but also determined by the exchange factor associated with Arf6 (Kalia et al. 2006). The Arf6 GTP exchange factor ARNO has previously been shown to be dependent on inositol-lipids for its membrane recruitment. However, the process seems to be even more complex as small G proteins of the Arl 4 family can function in recruitment of the Arf6 GEF cytohesin to the plasma membrane (Hofmann et al. 2007). Perhaps this might explain some of the apparent differences between cell types. Caveolae and endocytosis Caveolae are characteristic, small (50–100 nm), regular invaginations of the plasma membrane as seen in the EM (Fig. 3). Since caveolae have this distinct, easily recognizable ultrastructural morphology, they have been of particular interest in relation to clathrin-independent endocytosis. The distinct shape of caveolae early led to the assumption that caveolae were endocytic structures comparable to clathrin-coated pits and vesicles. Also, in thin sections for EM which are not cut perpendicular to the plasma membrane, one might get the impression of several free, i.e. not surface-connected, caveolar profiles, supporting the idea of caveolar endocytosis and a front of internalized caveolae moving into the cell (Fig. 3). Interestingly, more careful EM examination—for instance using Ruthenium Red as a surface marker—reveal that most if not all such vesicular profiles are indeed surface-connected caveolae (Fig. 4). Also, the observation that e.g. Cholera toxin localized to caveolae supported the idea that caveolae were responsible for the endocytic uptake of Cholera toxin, a notion that is certainly an oversimplification (Torgersen et al. 2001; Shogomori and Futerman 2001; Nichols et al. 2001; see discussion in Hommelgaard et al. 2005; see also Kirkham et al. 2005). Fig. 3Appearance of caveolae in myoepithelial cells. In (a) is seen a single caveola (Cav) and for comparison a clathrin-coated pit (Cp). b shows a group of caveolae at the plasma membrane; as the section is largely perpendicular to the membrane, most of the caveolae are clearly seen to be surface-connected. In (c) is shown an example of a section which is not perpendicular to the plasma membrane, and here many of the caveolae appears as free (not surface-connected) vesicles. If such “free” vesicles should be unequivocally established as caveolae, they can be immunogold-labeled using an antibody against caveolin, as shown on the ultracryosection in (d) (Cav). An unlabeled, Cp is also seen. Bars 500 nmFig. 4Caveolae are surface-connected structures. a–d show filter-grown MDCK cells postfixed with the electron-dense cell surface marker Ruthenium Red from the basolateral side. In (a) is seen a number of caveolae connected to the basolateral membrane (arrowheads). Also note that three caveolae (arrows) are connected to the plasma membrane (open arrow) via a larger membrane invagination (asterisk). b–d show clusters of caveolae (arrows) associated with elongated “vacuolar” structures (asterisks) apparently freely localized in the cytoplasm. However, the labelling with Ruthenium Red clearly demonstrates that these structure are indeed connected to the basolateral membrane, as revealed in (a). En, early and later endosomes; Nu, nucleus. Bar 500 nm However, fluorescence recovery after photobleaching (FRAP) analysis of cells expressing GFP-caveolin revealed that caveolae under normal conditions are not involved in endocytosis to any significant degree (Thomsen et al. 2002). Similarly, more recently live cell imaging of cells expressing GFP-caveolin showed that even high concentrations of EGF, which had been reported to lead to internalization of the EGFR via caveolae (Sigismund et al. 2005), do not mobilize caveolae above a low background level (Kazazic et al. 2006). Also, in a careful EM study where the setup allowed to discriminate between caveolin-associated and other vesicles as well as between surface-connected and free vesicles, Kirkham et al. (2005) showed that only about 2% of caveolae bud from the plasma membrane per minute. For comparison, it takes 50–150 s for a clathrin-coated pit to assemble and pinch off to form a free, clathrin-coated vesicle (Ungewickell and Hinrichsen 2007). This means that the entire population of coated pits present at a given time point will have been internalized within approximately 1–2.5 min. Thus, today it appears well-established and generally accepted that under normal, non-stimulating conditions, caveolae are quite stable structures at the plasma membrane that are not involved in endocytosis to any significant degree (van Deurs et al. 2003; Pelkmans et al. 2004; Hommelgaard et al. 2005; Parton et al. 2006; Kirkham et al. 2005; Lajoie and Nabi 2007). On the other hand, it is also evident from the above mentioned studies that a few caveolae may pinch off from the plasma membrane (Thomsen et al. 2002; Kirkham et al. 2005). This could simply be due to a basal turnover of caveolae at the plasma membrane. Normally, caveolae seem to be stabilized at, or anchored to, the plasma membrane by actin filaments (Stahlhut and van Deurs 2000; Hommelgaard et al. 2005; Thomsen et al. 2002; Pelkmans et al. 2002). The phosphatase inhibitor okadaic acid stimulates caveolar clustering and internalization in an actin-dependent manner (Parton et al. 1994; Thomsen et al. 2002). One could therefore also imagine that caveolae might be involved in a very slow, regulated endocytic uptake which under certain circumstances can be speeded up, for instance by controlled reorganization of the actin cytoskeleton. However, this remains speculative. Interestingly, it was recently reported that coassembly of flotillin-1 and -2 induced membrane curvature and formation of membrane invaginations morphologically similar to caveolae but different from these by not containing caveolin, the major caveolar protein. These flotillin-positive but caveolin-negative structures became internalized in contrast to “classical” caveolae (Frick et al. 2007). Caveolae perform short range “kiss-and-run” motility very close to the plasma membrane. The process most likely involves continuous internalization and recycling of caveolae immediately beneath the plasma membrane (probably in the narrow 100–200 nm zone between the plasma membrane and the stabilizing actin cytoskeleton) (Thomsen et al. 2002, Fig. 5, online movie; Pelkmans and Zerial 2005). The significance of this process is unclear but it may explain the widely accepted role of caveolae in endothelial transcytosis. Thus, since endothelial cells are often very flat (100–200 nm), caveolae that have pinched off from e.g. the luminal endothelial plasma membrane, could fuse immediately after with the ablumenal plasma membrane or, in cases where the endothelial cell is very flat, the fusion may take place even before the pinching has been completed, and temporary transendothelial channels may arise (van Deurs et al. 2003). When SV40 virus is added to cells, the virus particles are trapped in caveolae which subsequently become internalized by a process which requires tyrosine phosphorylation, reorganization of the cortical actin cytoskeleton, and recruitment of dynamin (Pelkmans et al. 2001, 2002). Characteristically, the caveolar internalization is a single wave phenomenon and it takes hours before caveolin/caveolae reappear at the plasma membrane. More recently, it was shown that SV40 virus is actually internalized much more efficiently by a caveolin-independent mechanism (Damm et al. 2005). In any case it is questionable whether the downregulation of plasma membrane caveolae induced by SV40 virus represents any normally occurring, inducible clathrin-independent endocytic mechanism. Rather, it might reflect clearance of the plasma membrane for caveolae occupied by cross-linked or aggregated receptors or ligands. Such downregulation of caveolae stimulated by receptor cross-linking is discussed in more detail elsewhere (see Hommelgaard et al. 2005). In this context, it is important to consider the many non-endocytic functions that have been associated with plasma membrane caveolae (see van Deurs et al. 2003). Thus, endothelial nitric oxide synthase (eNOS) is associated with caveolin-1, and caveolae play an important role in NO signaling, and are also involved in e.g. calcium signalling (Murata et al. 2007). Moreover, caveolae is important for cholesterol transport and homeostasis in adipocytes, and it was recently found that cholesterol-induced transport of caveolin to lipid droplets requires dynamin and PKC and very likely involves caveolar endocytosis (Le et al. 2006). Therefore, rather than considering caveolae in general as structures specialized for clathrin-independent endocytosis, they should be considered multifunctional membrane domains with physiological roles depending on cell type and situation (van Deurs et al. 2003). In addition, a recent report suggests that caveolins may even regulate cell function independently of caveolae (Head and Insel 2007). Regulation of endocytosis in polarized epithelial cells Studies of endocytic mechanisms in cell lines do not necessarily reflect the uptake mechanisms operating in vivo. Both cell density, cell adhesion and polarization can be regulators of endocytosis (del Pozo and Schwartz 2007; Sandvig 1978; Sandvig and van Deurs 2005; Eker et al. 1994; Holm et al. 1995; Llorente et al. 2000; Sandvig et al. 2000; Llorente et al. 1996). An early argument against clathrin-independent endocytosis was, not unexpectedly, that this could be a process occurring only in unpolarized cell lines grown on plastic support. However, investigations of polarized cells, such as MDCK I cells grown on filters, revealed that clathrin-independent endocytosis contributed to the endocytic uptake both on the apical and basolateral side (Eker et al. 1994). When using cytosolic acidification to block clathrin-dependent endocytosis (Sandvig et al. 1987), there was still uptake of the plant toxin ricin at both poles (Eker et al. 1994). The fact that there was a larger reduction of uptake at the apical side than at the basolateral side, does, however, not necessarily provide information about the fraction of clathrin-independent endocytosis at the two poles, since the low pH might affect the remaining uptake at the two poles differently. Importantly, apical clathrin-independent endocytosis turns out to be under complex regulation, and can be regulated independently of uptake at the basolateral side. Using either cells permeabilized at the basolaterial side or membrane permeable compounds (for review, see Sandvig and van Deurs 2005), we have shown that apical clathrin-independent endocytosis is regulated by protein kinase A (Eker et al. 1994), protein kinase C (Holm et al. 1995), phospholipase D (Sandvig et al. 2000), cyclooxygenase (Llorente et al. 2000) and calmodulin (Llorente et al. 1996). Interestingly, a Rho family member, apparently RhoA, seems to regulate apical endocytosis in a manner which can be supported by GTPγS (Garred et al. 2001). The RhoA-mediated mechanism is able to mediate uptake of both gold particles, ricin-HRP and HRP.	[ "endocytosis", "caveolae", "ricin", "flotillin", "shiga toxin", "cholera toxin", "rho proteins" ]	[ "P", "P", "P", "P", "P", "P", "R" ]
Biochem_Pharmacol-2-1-2279149	Mouse N-acetyltransferase type 2, the homologue of human N-acetyltransferase type 1	There is increasing evidence that human arylamine N-acetyltransferase type 1 (NAT1, EC 2.3.1.5), although first identified as a homologue of a drug-metabolising enzyme, appears to be a marker in human oestrogen receptor positive breast cancer. Mouse Nat2 is the mouse equivalent of human NAT1. The development of mouse models of breast cancer is important, and it is essential to explore the biological role of mouse Nat2. We have therefore produced mouse Nat2 as a recombinant protein and have investigated its substrate specificity profile in comparison with human NAT1. In addition, we have tested the effects of inhibitors on mouse Nat2, including compounds which are endogenous and exogenous steroids. We show that tamoxifen, genistein and diethylstilbestrol inhibit mouse Nat2. The steroid analogue, bisphenol A, also inhibits mouse Nat2 enzymic activity and is shown by NMR spectroscopy, through shifts in proton peaks, to bind close to the active site. A three-dimensional structure for human NAT1 has recently been released, and we have used this crystal structure to generate a model of the mouse Nat2 structure. We propose that a conformational change in the structure is required in order for ligands to bind to the active site of the protein. 1 Introduction Arylamine N-acetyltransferases (NATs, EC 2.3.1.5) have traditionally been identified as drug-metabolising enzymes responsible for the metabolism of arylamines, arylhydroxylamines and arylhydrazines. NATs were first identified in man and a range of eukaryotes and have had an important part to play in the early years of identification of pharmacogenetic variation in response to drug treatment [1]. The hydrazine drug isoniazid is polymorphically N-acetylated in humans to its therapeutically inactive form by the human isoenzyme now known as human NAT2. There are two human NAT isoenzymes: human NAT2, which metabolises isoniazid, and human NAT1, which does not metabolise isoniazid or other arylhydrazines [2], but does catalyse the acetylation of a distinct but overlapping series of arylamines [2]. The pattern of expression of the two human NAT genes also differs. As a result of genome-wide microarray [3–12] and proteomic studies [13], it is clear that human NAT1 is highly expressed in oestrogen receptor positive breast cancer [14]. There is also evidence that the level of expression of human NAT1 affects the growth of cultured breast cancer cells [13]. The role of the human NAT1 enzyme in breast cancer has not been extensively explored, although it has previously been demonstrated that the anti-oestrogen compound tamoxifen is an inhibitor of human NAT1 [15–17]. It has also been demonstrated that both human NAT1 and the oestrogen receptor are down-regulated in tissues in which p53 is mutated [5]. It appears that human NAT1 is a marker for oestrogen receptor positive breast cancer, although information is still accumulating on the relationship of this marker to others, including the oestrogen receptor itself [14]. However, if the human NAT1 enzyme is to be considered as a target for breast cancer therapy, it is essential that there is a suitable in vivo animal model for testing [18,19]. Mice represent the most convenient animal models of disease. At present, the status of mouse Nat2, the equivalent of human NAT1, in a model of breast cancer has not been explored. We have previously demonstrated that mouse Nat2 is expressed in the epithelial cells lining the mammary ducts [20]. This is the same location as the human NAT1 enzyme in normal breast tissue [13,21]. In order to provide a firm foundation for establishing such a model, we have generated pure recombinant mouse Nat2 and investigated its activity with a wide range of substrates and the effects of a range of potential inhibitors, including endogenous and exogenous steroids. 2 Materials and methods 2.1 Chemicals All chemicals were purchased from Sigma–Aldrich and all molecular biology reagents were purchased from Promega, unless otherwise stated. 2.2 Cloning and expression studies of mouse Nat2 The mouse Nat2 open-reading frame [22] was sub-cloned into pET28b(+) (Novagen) using the compatible restriction enzyme sites, NdeI and EcoRI. This allowed the production of recombinant mouse NAT protein with an N-terminal His-tag, for ease of downstream purification. The pET28b(+) plasmid containing mouse Nat2 was transformed into Escherichia coli strain JM109 and further transformed into Rosetta(DE3)pLysS (Novagen) after confirming the correct insert sequence (DNA Sequencing Facility, Biochemistry, University of Oxford). The positive transformant was grown in LB to mid-log phase at 37 °C with shaking (180 rpm) and stored at −80 °C with 10% (v/v) glycerol. Thawed glycerol stock (100 μL) was used to inoculate fresh LB media (100 mL) supplemented with kanamycin (30 μg/mL) and chloramphenicol (34 μg/mL) and the culture was incubated at 37 °C for 16 h with shaking (180 rpm). The starter culture was then diluted 50-fold into fresh LB media (2 L) supplemented with kanamycin (30 μg/mL), which was then incubated at 27 °C with shaking (180 rpm). When the absorbance at 600 nm reached 0.7, the expression of mouse Nat2 was induced by addition of isopropyl-β-d-thiogalactopyranoside (IPTG) to a final concentration of 0.1 mM. The culture was grown for a further 16 h and the cells were harvested by centrifugation (6000 × g, 4 °C, 20 min). The cell pellet was resuspended in 20 mL lysis buffer (300 mM NaCl, 20 mM Tris–HCl (pH 8.0) containing 1× EDTA-free Complete Protease Inhibitor (Roche)) and then stored at −80 °C. For expression of 15N-labelled mouse Nat2, the starter culture was diluted 50-fold in 15N-labelled Standard-E. coli-OD2 N medium (Silantes) supplemented with kanamycin (30 μg/mL) and grown at 37 °C until the absorbance at 600 nm reached approximately 0.6. The cells were further grown at 27 °C until the absorbance reached 0.7, when the expression was induced with 0.1 mM IPTG. After induction for 17 h, the cells were harvested and stored as described for unlabelled mouse Nat2. The resuspended frozen cells were thawed at 37 °C and sonicated on ice (25 cycles of 30 s on, 45 s off at 10 μm). The soluble lysate was isolated from the cell debris by centrifugation (12,000 × g, 4 °C, 20 min) and incubated (4 °C, 5 min) with 5 mL of Ni-NTA resin (Qiagen). The resin was sequentially washed with lysis buffer containing increasing imidazole concentrations in a step gradient at 4 °C (2 washes each of 10 mL of 0 mM, 10 mM, 20 mM, 50 mM and 100 mM imidazole). Fractions containing hexa-histidine tagged Nat2 (His-Nat2), were pooled and thrombin (5 U/mg NAT) was added to remove the His-tag. After 16 h incubation at 4 °C, the protein sample was dialysed against 20 mM Tris–HCl (pH 8.0), 1 mM dithiothreitol (DTT), 1 mM EDTA buffer, and concentrated to 1 mg/mL by centrifugation in a 15 mL concentrator (Amicon) after addition of 5% (v/v) glycerol. Protein samples in 100 μL aliquots were frozen in liquid nitrogen and stored at −80 °C. Recombinant hamster Nat2 was expressed and purified as previously described [23]. 15N-labelled hamster Nat2 was grown and expressed as described for 15N-labelled mouse Nat2 and purified by the same method as for the unlabelled hamster Nat2. The expression levels and activities of soluble Nat2 were monitored by visualizing the levels of recombinant protein expression on 12% acrylamide Tris–Glycine SDS-PAGE as well as by measuring the rate of acetylation of 4-aminobenzoic acid (PABA) as a substrate (see method below) [24]. 2.3 Enzymic assays 2.3.1 Acetylation of arylamines The rates of arylamine acetylation by NATs were determined colorimetrically as previously described [25], but with minor modifications. Each assay contained the enzyme, arylamine substrate (see each figure legends for concentrations) and AcCoA (400 μM) in a total volume of 100 μL in assay buffer (20 mM Tris–HCl (pH 8.0), 1 mM DTT). Initially, the enzyme and substrate mixtures were pre-incubated at 25 °C for 5 min and AcCoA was added to start the reaction. The reaction was quenched using 20% (w/v) trichloroacetic acid (TCA) at different time intervals. The stopped reaction was centrifuged (16,000 × g rpm, 10 min) to pellet the precipitated proteins. The stopped reaction mixture (200 μL) was added to 800 μL of 5% (w/v) 4-(N,N-dimethylamino)benzaldehyde (DMAB) in 9:1 acetonitrile:water to develop the colour. The absorption at 450 nm was measured with a Hitachi U-2001 UV–vis spectrophotometer, with the amount of residual substrate in the reaction was determined by comparison with a standard curve. 2.3.2 Hydrolysis of AcCoA (free CoA production) The rate of production of the free thiol Coenzyme A by NAT, in the presence of a range of known NAT acetyl-acceptor substrates, was determined by using Ellman's reagent, 5,5′-dithio-bis(2-nitrobenzoic acid) (DTNB) as previously described [2,26]. The substrate (500 μM) and purified recombinant NAT were pre-incubated (25 °C, 5 min) in 96-well flat-bottomed polystyrene plates (Costar®, Corning Inc.) in 20 mM Tris–HCl (pH 8.0). AcCoA (400 μM) was added to start the reaction in a final volume of 100 μL. The reaction was quenched with the addition of 25 μL of guanidine–HCl solution (6.4 M guanidine–HCl, 0.1 M Tris–HCl (pH 7.3)) containing 5 mM DTNB. The absorbance at 405 nm was measured on a plate-reader (Sunrise, Tecan). Assay buffer was used to replace substrate, AcCoA or NAT for control reactions. The amount of CoA produced in the assay was determined in comparison with a CoA standard curve. Substrate abbreviations used for the substrate selectivity assays are as follows: aniline (ANL), 4-aminobenzoic acid (PABA), 4-aminosalicylic acid (4AS), 5-aminosalicylic acid (5AS), 4-chloroaniline (CLA), 4-bromoaniline (BRA), 4-iodoaniline (IOA), 4-methoxyanline (ANS), 4-ethoxyaniline (EOA), 4-butoxyanline (BOA), 4-hexyloxyaniline (HOA), 4-phenoxyaniline (POA), 4-aminoveratrole (4AV), 2-aminofluorene (2AF), 4-aminobenzoyl-l-glutamate (pABGlu), sulfamethazine (SMZ), procainamide (PRO), 4-aminopyridine (APY), isoniazid (INH), hydralazine (HDZ) and phenylhydrazine (PHZ). The stock concentration of substrate was 100 mM dissolved in DMSO, such that the final concentration of DMSO was less than 5% in the assays. DMSO only controls were carried out. 2.4 NMR spectroscopy The 2D 1H-15N HSQC spectrum of mouse Nat2 was collected using a 750 MHz NMR spectrometer (Department of Biochemistry, University of Oxford). The spectrum was collected at 20 °C for a sample containing 30 mg/mL of uniformly 15N-labelled mouse Nat2 in 95% H2O/5% D2O 10 mM Tris buffer at pH 7.0. Spectral widths of 12,500 and 2500 Hz were used in F2 and F1, respectively. The dataset contained 1024 and 128 complex points in t2 and t1, respectively. 1D 1H NMR spectra for hamster and mouse Nat2 were collected at 20 °C using a jump-return sequence. This sequence enabled the observation of HN peaks which exchange to a significant extent with H2O. Spectra were collected using a 600 MHz spectrometer using a sweep width of 12,500 Hz. Unlabelled hamster and mouse Nat2 were used at a concentration of 10 mg/mL in the buffer described above. The titration of mouse Nat2 with bisphenol A was carried out by stepwise addition of small volumes (5–20 μL) of a concentrated solution of bisphenol A (100 mM in deuterated DMSO) to the mouse Nat2 solution in the NMR tube. The highest concentration of DMSO in the sample used was 2.8%. DMSO alone had no effect on the NMR spectrum of mouse Nat2 using concentrations up to 5%. 3 Results 3.1 Cloning and expression of mouse Nat2 The mouse Nat2 was cloned into E. coli expression vector pET28b(+) and expressed in E. coli Rosetta (DE3)pLysS with an N-terminal hexa-histidine tag. The recombinant mouse His-Nat2 was purified using immobilised metal affinity chromatography, using a Ni-NTA column (Novagen), and eluted (Fig. 1a). Pure Nat2 predominantly eluted in the 50 mM and 100 mM imidazole washes. The hexa-histidine tag was readily removed by thrombin digestion (Fig. 1b). The purified recombinant Nat2 was active, readily catalysing the acetylation of PABA, a well-studied arylamine substrate of mouse Nat2 [27] (Table 1). The enzyme was relatively stable, and the enzymic activity was maintained (>85%) after incubation at temperatures ranging between 4 °C and 25 °C over a 72 h period. 3.2 Substrate specificity of mouse Nat2 To further characterise the purified recombinant mouse Nat2, a panel of known NAT substrates (20 compounds) was tested to determine the substrate selectivity of recombinant mouse Nat2. In order to allow direct comparison with substrate selectivity profile for human NAT1 and NAT2, the assay for mouse Nat2 was performed under the same conditions as previously described [2]. Mouse Nat2 shows high specific activities with a broad range of arylamines, including PABA, 5AS, 4AS and 2AF (Fig. 2 and Supplementary Table 1). In contrast, mouse Nat2 shows low specific activities against arylhydrazine substrates such as HDZ, INH, and PHZ, and against arylamine drugs procainamide and sulfamethazine (Fig. 2). This overall trend is in agreement with the substrate selectivity observed in previous studies with impure recombinant mouse Nat2 using a subset of these compounds [22,27,28], although it is much less than the activity observed towards these substrates with human NAT2 (Fig. 2). There was low, yet measurable activity using arylhydrazine substrates with purified mouse Nat2, which has not been observed in previous studies using recombinant mouse Nat2 expressed in a transient expression system [28]. Pure recombinant mouse Nat2 also readily acetylates 2-aminofluorene, the substrate which has been used to identify mouse strains carrying an active form of mouse Nat2—the fast acetylating C57Bl/6 strain, rather than the inactive form as is found in the slow acetylating A/J strain which is unstable and likely to be subject to rapid degradation [29]. The specific activity pattern of human NAT1 [2] overlaps with the mouse Nat2 profile (Fig. 2), reconfirming that mouse Nat2 is very similar to human NAT1 in substrate specificity. Mouse Nat2 shows differences from the activity profile of human NAT1, however. It appears that mouse Nat2 generally has higher inherent specific activity than human NAT1 and also acetylates arylhydrazines, when the acetylation of these compounds by human NAT1 is not detected under the same conditions [2]. 3.3 Mouse Nat2 inhibition by steroidogenic compounds It had been demonstrated, with extracts of tissues and cells now known to express the human NAT1 gene, that tamoxifen is an inhibitor of human NAT1 [15–17]. In view of the relationship between the expression of human NAT1 and oestrogen receptor positivity in breast cancer [14], we wished to investigate the effects of steroidogenic compounds and xenobiotic oestrogenic compounds, including tamoxifen, on the activity of mouse Nat2. From the inhibition study using a broad range of steroidogenic compounds, it has been demonstrated that mouse Nat2 is inhibited selectively by oestrogenic compounds, with 17-hydroxy-β-estradiol identified as the most potent inhibitor (Table 2 and Supplementary Fig. 1). This was substantiated by the inhibition of mouse Nat2 activity by xenobiotic oestrogens, such as tamoxifen, and the IC50 values for these inhibitors are shown in Table 3. The phytoestrogens genistein, alpha-zearalenol and the synthetic oestrogen diethylstilbestrol show similar potency of mouse Nat2 inhibition to 17-hydroxy-β-estradiol, tamoxifen and 4-hydroxy tamoxifen. Bisphenol A was also shown to be an inhibitor of mouse Nat2, albeit with lower potency than some of the other compounds tested. One key practical advantage of using bisphenol A as an inhibitor is the improved solubility of this compound relative to the other phytoestrogens and steroids. 3.4 Structure of mouse Nat2 Until recently, there has been no structural information on eukaryotic NAT enzymes. However, recently the three-dimensional structure of a human NAT1 mutant (Phe/Ser substitution at position 125) was determined by X-ray crystallography (PDB accession code: 2IJA [30]). A non-mutated human NAT1 crystal structure has also been deposited in which the active-site cysteine residue was modified to S-(2-anilino-2-oxoethyl)-cysteine (PDB accession code: 2PQT). It is considered likely that both the mutation of residue 125 and the modification of the active-site cysteine residue stabilise the structure of the human enzyme, because thousands of attempts to generate a crystal from the native human NAT1 protein have been unsuccessful (A. Kawamura, unpublished results). The structure of human NAT1 is illustrated in comparison with the NAT from M. smegmatis (Fig. 3). The eukaryotic proteins contain a loop between the second and third domains. This inter-domain loop is not present in their prokaryotic counterparts [31], as highlighted in Fig. 3. It is interesting to note that in the structure of the human NAT1 enzyme, this inter-domain loop is folded back over the active site. Likewise, the C-terminus is shown to be folded over the active-site cleft. These observations indicate that a conformational change in the protein may be required to allow the substrates to access the active-site cysteine residue. However, a structure of human NAT2 with coenzyme A bound has also been deposited (PDB accession code: 2PFR), and the position of the coenzyme A is such that the C-terminus and inter-domain loops regions are not significantly changed. This is in contrast to the recently reported structure of Mycobacterium marinum NAT in complex with coenzyme A [32]. In this bacterial NAT, the C-terminus is shorter than in the eukaryotic structures, and there is no inter-domain loop, and the ligand is found in the space where these regions are located in the eukaryotic structure. We have generated a structural model of the mouse Nat2 protein, whose amino acid sequence is over 80% identical to human NAT1, by using the program Modeller 8v2 [33]. Fig. 4 shows the structural model of mouse Nat2 and a comparison of the mouse Nat2 and human NAT1 protein structures. The two structures share 1191 equivalent atoms, over which the root mean squared deviation is 0.75 Å. In mouse Nat2, as in human NAT1, the inter-domain loop and the C-terminus occlude the active site. Despite the evidence for coenzyme A binding to human NAT2 without major conformational rearrangement of the C-terminus and inter-domain loop regions, it has not been possible to perform ligand docking into the mouse Nat2 or human NAT1 structures without disturbing these regions of the protein structure. Therefore, it remains possible that the inter-domain loop regions and C-termini of the eukaryotic NAT proteins are conformationally flexible, in order to accommodate substrate or inhibitor binding. 3.5 NMR studies NMR studies are ideal for investigating protein interactions with substrates and inhibitors in solution. Therefore, we have generated mouse Nat2 uniformly labelled with 15N to allow 2D NMR investigations of the protein. The 2D 1H-15N HSQC spectrum of mouse Nat2 is shown in Fig. 5; this spectrum contains a peak for each backbone amide (1HN-15N) and additional peaks from the side chains of Asn, Gln and Trp. The HSQC spectrum is well resolved, as expected for a compact globular structure, and some cross peaks are relatively sharp for a 30 kDa protein, this may indicate some flexibility between the domains of mouse Nat2 or the presence of mobile loops. The HSQC spectrum of mouse Nat2 has been compared with the corresponding spectrum from hamster Nat2 for which a full assignment is available [34]. The pattern of peaks in the HSQC spectra is extremely similar for the two homologous proteins; this allows some peaks observed in the mouse Nat2 spectrum to be assigned on the basis of the hamster Nat2 spectrum. The downfield region of the 1H 1D spectra, collected with a jump-return sequence, of mouse and hamster Nat2 also show a similar pattern of peaks (Fig. 6A and B). The four peaks are compared in Table 4, together with their assignments. Two of the peaks (2 and 3) have been assigned previously for hamster Nat2 and can be assigned in the mouse spectrum by homology; these correspond to strongly hydrogen bonded tryptophan indole groups: the tryptophan associated with the P-loop (Trp132, see Fig. 7) and Trp67, which is located adjacent to the active-site Cys68, and appears to have a structural role. The strong hydrogen bonds are likely to be responsible for the large downfield shifts of these tryptophan peaks compared to random coil values (∼9.5 ppm). The two remaining peaks (1 and 4) are assigned to the imidazole 1HN of histidine on the basis of 15N decoupling experiments; these peaks have not been assigned previously for hamster Nat2. The side chain 1HN groups of histidine are usually not observed by NMR, particularly at pH 7, due to their rapid exchange with solvent protons. The observation of these peaks indicates that the histidine residues are probably buried within the protein and that the side chain 1HN are involved in hydrogen bonds. Analysis of the homology model of mouse Nat2 suggests that at least one of these peaks might arise from the active site histidine, His107. The homology model of mouse Nat 2 indicates that the side chain of His107 is involved in two hydrogen bonds in the active site and is the least accessible to solvent. In addition, recent enzymological studies with hamster Nat2 have confirmed that the active-site cysteine and histidine residues exist as a thiolate-imidazolium ion pair [35], in which the histidine has two tightly bound and H-bonded protons. 3.6 Interaction of mouse Nat2 with exogenous steroid inhibitors In order to probe the molecular interactions of mouse Nat2 and the exogenous steroid inhibitors, we have investigated the effect of bisphenol A on the 1D 1H NMR spectrum of mouse Nat2 (Figs. 8 and 9). The observation of changes in chemical shift for particular amino acid residues which result from the addition of a ligand are usually interpreted as an indication that these residues are located in close proximity to the ligand-binding site. When bisphenol A was titrated into a solution of mouse Nat2, the two downfield histidine peaks (1 and 4, Table 4) were observed to shift by more than 0.05 ppm (Fig. 9); peak 1 shifts downfield by ∼0.12 ppm and peak 4 shifts upfield by ∼0.08 ppm in the presence of 8 equivalents of bisphenol A. Additional small shifts were also observed for the side chain HN of Trp67 and Trp132 (Fig. 9). The observation of progressive changes in chemical shift upon addition of bisphenol A indicates fast exchange on the NMR timescale; this would be expected given the relatively weak affinity of mouse Nat2 for this ligand (IC50 = 290 ± 10 μM). We have postulated above that at least one of the histidine peaks arises from the active-site histidine, His107; therefore, these results show that binding of bisphenol A is likely to take place in close proximity to the active site of mouse Nat2. 4 Discussion The possibility that human NAT1 may serve as a marker for sub-dividing populations of different breast cancers is intriguing. It may also be that human NAT1 will find a role as a target for breast cancer therapies. In order to exploit these findings fully, it will be essential to have an animal model. There have been, for a long time, suggestions that human NAT1 has an endogenous role. The mouse equivalent of human NAT1 is mouse Nat2. Genetic deletion of mouse Nat2 has not created an overt phenotype; however, some of these studies in mice indicate that Nat2 deletion gives rise to skewed sex ratios both on mixed [36] and pure [37] genetic backgrounds, indicating a gender-dependent phenotype, discernable at the population level, that is consistent with a role in endocrine function. There is now also clear evidence that mouse Nat2 has a role in vivo in folate catabolism [38]. These studies, together with the data on human NAT1 and breast cancer, are intriguing and are likely to provide clues as to the endogenous role of NAT1. The data presented increases the knowledge of mouse Nat2 in comparison with human NAT1 and reinforce the notion that mouse Nat2 is a good homologue of human NAT1. Mouse Nat2 has a similar substrate specific activity profile to human NAT1. The inhibition studies show that mouse Nat2 is inhibited by endogenous steroids and, like human NAT1, is inhibited by tamoxifen. Bisphenol A, which is used for making polycarbonates [39], has been reported to activate the oestrogen receptor [40]. Like tamoxifen, it is also an inhibitor of mouse Nat2. We have used bisphenol A to look at the binding of an oestrogenic stimulatory compound to mouse Nat2 and show by NMR that bisphenol A interacts with mouse Nat2 close to the active site region. These studies could not be done directly with tamoxifen because of the high concentrations of protein (and hence ligand) required for NMR experiments under which conditions tamoxifen is insoluble. Both endogenous and xenobiotic oestrogens have been shown to modify breast cancer risk, and tamoxifen is widely used for breast cancer therapy (reviewed in [41]). The effect of different synthetic or natural oestrogens on cell proliferation has been analysed directly, in breast cancer cell lines, and in gene expression studies [41–44]. Although the dose–response curves for different oestrogenic compounds vary with the assay system, bisphenol A is generally less active than tamoxifen, genistein or diethylstilbestrol, which reflects the order of potency of mouse Nat2 inhibition. Circulating plasma concentrations of the phytoestogen genistein have been found to be as high as 18 μM after a soy-based meal [45], and intratumoral levels of tamoxifen can reach micromolar concentrations [46]. Whether the effects on inhibition of Nat2 activity in vitro may be physiologically relevant for breast cancer prevention or therapy will require further analysis. These studies pave the way for investigation of the role of mouse Nat2, the human NAT1 homologue, in studies of oestrogen receptor positive breast cancer. This is particularly relevant in view of the observation that mouse Nat2 is located histologically in the mammary gland epithelial cells [20]. The structural studies also provide a rational basis for identification of mouse Nat2 inhibitors. We have recently screened a 5000-strong compound library using human NAT1, human NAT2, mouse Nat1 and mouse Nat2; from these studies, several compounds been have identified which are potent and specific inhibitors of human NAT1 and its mouse homologue, Nat2. An accumulation of information on the eukaryotic NAT proteins at a structural level will allow further development of chemical tools for investigation of the role of this marker in breast cancer in humans, and its potential in animal models.	[ "breast cancer", "steroid", "xenobiotic", "arylamine n-acetyltransferase/nat", "enzyme inhibition", "selective estrogen receptor modulator/serm" ]	[ "P", "P", "P", "M", "R", "M" ]
Acta_Neuropathol-4-1-2386160	Cellular distribution of vascular endothelial growth factor A (VEGFA) and B (VEGFB) and VEGF receptors 1 and 2 in focal cortical dysplasia type IIB	Members of the vascular endothelial growth factor (VEGF) family are key signaling proteins in the induction and regulation of angiogenesis, both during development and in pathological conditions. However, signaling mediated through VEGF family proteins and their receptors has recently been shown to have direct effects on neurons and glial cells. In the present study, we immunocytochemically investigated the expression and cellular distribution of VEGFA, VEGFB, and their associated receptors (VEGFR-1 and VEGFR-2) in focal cortical dysplasia (FCD) type IIB from patients with medically intractable epilepsy. Histologically normal temporal cortex and perilesional regions displayed neuronal immunoreactivity (IR) for VEGFA, VEGFB, and VEGF receptors (VEGFR-1 and VEGFR-2), mainly in pyramidal neurons. Weak IR was observed in blood vessels and there was no notable glial IR within the grey and white matter. In all FCD specimens, VEGFA, VEGFB, and both VEGF receptors were highly expressed in dysplastic neurons. IR in astroglial and balloon cells was observed for VEGFA and its receptors. VEGFR-1 displayed strong endothelial staining in FCD. Double-labeling also showed expression of VEGFA, VEGFB and VEGFR-1 in cells of the microglia/macrophage lineage. The neuronal expression of both VEGFA and VEGFB, together with their specific receptors in FCD, suggests autocrine/paracrine effects on dysplastic neurons. These autocrine/paracrine effects could play a role in the development of FCD, preventing the death of abnormal neuronal cells. In addition, the expression of VEGFA and its receptors in glial cells within the dysplastic cortex indicates that VEGF-mediated signaling could contribute to astroglial activation and associated inflammatory reactions. Introduction The vascular endothelial growth factor (VEGF) family includes seven members which are structurally homologous, but display molecular and functional diversity [58, 81]. VEGFA, the most well known member of the VEGF family, is a crucial regulator of angiogenesis and vascular permeability in both physiological and pathological conditions such as tumor growth, chronic inflammation, and ischemia [10, 16, 19]. In addition to the unquestioned role in angiogenesis, it has recently been shown that VEGFA has direct trophic effects on neuronal and glial cells in the central nervous system [10, 22, 25, 53]. VEGFB is most closely related to VEGFA [46, 81]; however, the biological function of VEGFB is less well characterized than the function of VEGFA. VEGFB is expressed early during development and appears to have prominent expression in the central nervous system [1, 36]. Additionally, VEGFB has been shown to function as an angiogenic and neuroprotective protein [29, 63, 68] and recent evidence suggests a role for VEGFB in neurogenesis [69, 70]. The diverse functions of VEGF proteins can be explained by their differential binding to signaling VEGF receptors [VEGFRs; VEGFR-1 (Flt-1), VEGFR-2 (Flk1/KDR), and VEGFR-3 (Flt-4) [50, 81]]. VEGFA binds to VEGFR-1 and VEGFR-2, whereas VEGFB binds specifically to VEGFR-1 and not to VEGFR-2 [81]. The VEGF-signaling pathway, involving both neuronal and glial cells, has been implicated in several neurological disorders, including neurodegeneration, stroke, and cerebral and spinal trauma [10]. In addition, expression of VEGFA is upregulated in neuronal and glial cells after epileptic seizures in rats [13], suggesting a role for VEGFA in seizure disorders. A recent study points to a neuroprotective role for VEGFA following status epilepticus [45]. The relevance of these findings in animal models to human epileptic disorders is uncertain. Using serial analysis of gene expression (SAGE), we recently identified the VEGFB gene to be upregulated in human tissue from a patient with focal cortical dysplasia (FCD) and intractable epilepsy compared to control cortex (Boer et al., unpublished observations). Upregulation of VEGFA and its receptor has also been recently shown in the hippocampus of cases of human temporal lobe epilepsy (TLE) [56]. However, the distribution of VEGFA, VEGFB, and VEGFRs in epilepsy-associated human malformations of cortical development has not yet been defined. In the present study, we investigated the expression of both VEGFA and VEGFB and their receptors (VEGFR-1 and VEGFR-2) in patients with FCD, which is a developmental disorder known to be a major cause of intractable epilepsy [73]. We report the specific cellular distribution, including both the neuronal and the glial components of the dysplastic cortex, and we discuss the potential role of VEGFA, VEGFB, and their receptors in the histogenesis and epileptogenesis of this developmental lesion. Materials and methods Subjects The cases included in this study were obtained from the databases of the Departments of Neuropathology of the Academic Medical Center (University of Amsterdam; UVA) in Amsterdam and the University Medical Center in Utrecht (UMCU). We examined surgically resected tissue from nine patients undergoing epilepsy surgery for focal cortical dysplasia. Informed consent was obtained for the use of brain tissue and for access to medical records for research purposes. The tissue was obtained and used in a manner compliant with the Declaration of Helsinki. The classification system proposed by Palmini et al. [48] was used for grading the degree of FCD and only patients with FCD type IIB located in the temporal lobe were included. The clinical characteristics derived from the patient’s medical records are summarized in Table 1. The predominant type of seizure pattern was that of complex partial seizures, which were resistant to maximal doses of antiepileptic drugs (AEDs; carbamazepine, valproic acid, phenytoin, levetiracetam, oxcarbazepine, and clonazepam). Information concerning the exact time of last seizure occurrence prior to surgical resection was not available. However, all the patients included in our series did not have seizure activity in the last 24 h before surgery. The patients underwent presurgical evaluation [74]. Intraoperative ECoG was performed routinely in all operations for tailoring of surgery and we classified the post-operative seizure outcome according to Engel [17]. Follow-up period ranged from 1 to 9 years. Table 1Summary of clinical findings of patients with focal cortical dysplasiaPatient/sex/age (years)DiagnosisDuration of epilepsy (years)Seizure typeEngel class1/M/11FCD IIB11CPSI2/M/31FCD IIB20CPSI3/F/25FCD IIB9CPSI4/F/22FCD IIB21CPS/SGSI5/M/18FCD IIB14CPSI6/M/17FCD IIB10CPSI7/F/16FCD IIB11CPSI8/M/29FCD IIB21CPSI9/M/28FCD IIB21CPSIFCD focal cortical dysplasia, CPS complex partial seizures, SGS secondary generalized seizures Normal-appearing control cortex/white matter from temporal region was obtained at autopsy from five adult control patients (male/female: 2/3; mean age 42, range 17–55) without history of neurological diseases. All autopsies were performed within 12 h after death (post mortem delay: 11, 11.5, 9, 8.5, 6). The cause of death was represented by acute myocardial infarction. In addition, four of the nine FCD cases contained sufficient amount of perilesional zone (normal-appearing cortex/white matter adjacent to the lesion), for comparison with the autopsy specimens. This material represents good disease control tissue, since it is exposed to the same seizure activity, drugs, fixation time, and the age and gender are also the same. Tissue preparation Tissue was fixed in 10% buffered formalin and embedded in paraffin. Two representative paraffin blocks per case (containing the complete lesion or the largest part of the lesion resected at surgery) were sectioned, stained, and assessed. Paraffin-embedded tissue was sectioned at 6 μm, mounted on organosilane-coated slides (Sigma, St Louis, MO) and used for histological and immunocytochemical reactions as described below. Frozen tissue from control cortex and FCD tissue, stored at −80°C, was used for western blot analysis. Antibody characterization To document the presence of a heterogeneous population of cells, we used the following antibodies: glial fibrillary acidic protein (GFAP; polyclonal rabbit, DAKO, Glostrup, Denmark; 1:4,000; monoclonal mouse, DAKO; 1:50), vimentin (mouse clone V9, DAKO; 1:1,000), MAP2 (polyclonal rabbit; Chemicon; 1:500), neuronal nuclear protein (NeuN; mouse clone MAB377, Chemicon, Temecula, CA, USA; 1:2,000), non-phosphorylated neurofilament (SMI311; Sternberger monoclonals, Lutherville, MD; 1:1,000), human leukocyte antigen (HLA)-DP, -DQ, -DR (CR3/43; monoclonal mouse, DAKO; 1:400), CD68 (mouse clone PG-M1, DAKO; 1:200) and CD31 (mouse clone JC70A, DAKO; 1:100). For the detection of VEGFA, VEGFB, and their receptors, the following antibodies (Abs) were used: VEGFA (G153-694, monoclonal mouse; recognizing VEGF 165 and 189 [51], Pharmingen, CA, USA; 1:100), VEGFA (A-20, SC-152, polyclonal rabbit; raised against the N-terminus of VEGFA, recognizing VEGF 121, 165 and 189, Santa Cruz Bio., CA, USA; 1:100), VEGFB (H-70, SC-13083, polyclonal rabbit; raised against amino acids 1–70 of human VEGFB, Santa Cruz Bio.; 1:20), Flt-1 (VEGFR-1; C-17, SC-316, polyclonal rabbit, Santa Cruz Bio.; 1:100), Flk-1 (VEGFR-2; A-3, SC-6251, monoclonal mouse; Santa Cruz Bio.; 1:100). To allow comparative analysis, we used on frozen specimens of normal (n = 3) and FCD tissue (n = 2), two additional antibodies (which are not suitable for staining paraffin-embedded, formalin-fixed tissue): VEGFR-1 (clone Flt-19, 1:400; developed against the recombinant human extracellular domain of VEGFR-1) and VEGFR-2 (clone KDR-1, 1:400; developed against the recombinant human extracellular domain of VEGFR-2), kindly provided by Dr. H. A. Weich (National Research Center for Biotechnology, Braunschweig, Germany), and previously characterized on human tissues [49, 64, 80]. Similar immunoreactivity patterns were observed on paraffin-embedded and frozen tissue. The specificity of the antibodies used for immunocytochemistry on paraffin-embedded, formalin-fixed tissue (VEGFA, VEGFB, VEGFR-1, VEGFR-2; Santa Cruz Bio.), was further tested by performing western blot analysis of total homogenates of human control cortex. We also include one FCD case of which sufficient frozen material for blot analysis was available (Fig. 1). VEGFR1 and VEGFR-2 receptor proteins were detectable as a band of approximately 180 and 200 kDa, respectively; VEGFB was detectable as a band of approximately 40 kDa; VEGFA labeled a prominent band at approximately 48 kDa and a light band at 21 kDa (reducing conditions), as recently reported in human brain tissue (control hippocampus and FCD; [56]). All immunoreactive bands disappeared after preadsorption with the corresponding peptide. Fig. 1Representative immunoblot of VEGFA, VEGFB, VEGFR-1, and VEGFR-2 in total homogenates from control cortex and FCD tissue. Expression of β-actin (as reference protein) is shown in the same protein extracts For immunoblot analysis, human normal cortex (n = 3) and FCD (n = 1) samples were homogenized in lysis buffer containing 10 mM Tris (pH 8.0), 150 mM NaCl, 10% glycerol, 1% NP-40, Na-orthevanadate (10.4 mg/ml), 5 mM EDTA (pH 8.0), 5 mM NaF, and protease inhibitor cocktail (Boehringer Mannheim, Germany). Protein content was determined using the bicinchoninic acid method [65]. Non-reducing conditions were used to improve the detection of the VEGFA antibody, as previously reported [8]. For electrophoresis, equal amounts of proteins (30 μg/lane) were separated by sodium dodecylsulfate-polyacrylamide gel electrophoretic (SDS-PAGE) analysis in 7.5–12.5% gels. Separated proteins were transferred to nitrocellulose paper for 1 h and 30 min, using a semi-dry electroblotting system (BioRad, Transblot SD, Hercules, CA, USA). Blots were incubated overnight in TTBS (20 mM Tris , 150 mM NaCl, 0.1% Tween, pH 7.5)/5% non-fat dry milk, containing the primary antibody (VEGFA, VEGFR-1, and VEGFR-2, 1:1,000; VEGFB, 1:200). After several washes in TTBS, the membranes were incubated in TTBS/5% non-fat dry milk/1% BSA, containing the goat anti-rabbit coupled to horseradish peroxidase (1:2,500; Dako, Denmark) for 1 h. After washes in TTBS, immunoreactivity was visualized using an enhanced chemiluminescence kit (Amersham, Buckinghamshire, UK). Expression of β-actin (monoclonal mouse, Sigma, St Louis, MO; 1:50,000) was used as reference protein. Because of the limited availability of frozen material from FCD cases, a complete analysis with statistical comparison between control and FCD by immunoblot could not be performed. Immunocytochemistry Sections were deparaffinized, re-hydrated, and incubated for 20 min in 0.3% H2O2 diluted in methanol to quench the endogenous peroxidase activity. Antigen retrieval was performed by incubation for 10 min at 121°C in citrate buffer (0.01 M, pH 6.0); sections were washed with phosphate-buffered saline (PBS), and incubated for 30 min in 10% normal goat serum (Harlan Sera-Lab, Loughborough, Leicestershire, UK). We incubated the sections with the primary antibodies overnight at 4°C. Hereafter, sections were washed in PBS and we used the ready-for-use Powervision peroxidase system (Immunologic, Duiven, The Netherlands) and 3,3′-diaminobenzidine as chromogen to develop the staining. Sections were counterstained with hematoxylin, dehydrated, and coverslipped. Sections incubated without the primary antibody and excess of the antigenic peptide were essentially blank. For double-labeling studies, after incubation overnight at 4°C with the primary antibodies, sections were incubated for 2 h at RT with Alexa Fluor® 568-conjugated anti-rabbit IgG and Alexa Fluor® 488 anti-mouse IgG (1:1,000; Molecular Probes, The Netherlands). The VEGFR-2 antibody (monoclonal mouse) could only be combined with MAP2 and GFAP (polyclonal rabbit). Sections were analyzed by means of a laser scanning confocal microscope (Bio-Rad, Hercules, CA, USA; MRC1024) equipped with an argon-ion laser. Evaluation of immunostaining Semi-quantitative evaluation of immunoreactivity As previously reported [3, 54], a semi-quantitative analysis was done using an Olympus microscope and examining in each section, high-power non overlapping fields (of 0.0655 × 0.0655 mm width, each corresponding to 4.290 μm2), defined in the center of the lesion using a square grid inserted into the eyepiece. A total microscopical area of 858.050 μm2 was assessed per case. Neuronal cell bodies were differentiated from glia and glia-neuronal balloon cells on the basis of morphology. Balloon cells have eccentric nuclei and ballooned opalescent eosinophilic cytoplasm. The staining intensity was evaluated using a semi-quantitative three-point scale where immunoreactivity was defined as: − absent (0); + moderate (1); ++ strong staining (2); intensity score (Table 2). This score represents the predominant cell staining intensity found in each section for the different cell types (neurons, astrocytes, microglial cells, and balloon cells) as averaged from the selected fields (as previously described [3, 54]). Table 2VEGFA, VEGFB, VEGFR-1, and VEGFR-2 distribution in different cellular types in cases of FCD (% of cases with immunoreactive cells)Focal cortical dysplasia (n = 9)NeuronsAstrocytesBalloon cells−+++−+++−+++VEGFA022%78%011%89%011%89%VEGFB033%67%78%22%078%22%0VEGFR-1055%45%022%78%033%67%VEGFR-2011%89%55%45% 022% 45%33%FCD focal cortical dysplasia; immunoreactivity: − not present, + moderate, ++ strong Frequency of cell staining In each slice, we assessed the number of neurons and astrocytes labeled by a specific Ab on the total number of each cell type within the lesion using an ocular grid [4]. This frequency score was assigned using three distinct categories: (1) <10%, rare; (2) 11–30%, sparse; (3) >30%, high. The product of the intensity and the frequency scores was taken to give the total immunoreactivity score, as previously reported [23, 54]. For statistical analysis of data, SPSS for Windows was used. Data were compared using a non-parametric Kruskal–Wallis test followed by a Mann–Whitney test to assess the difference between groups. P < 0.05 was taken as the level of significance. Results Human material and histological features All nine patients had chronic pharmacoresistant epilepsy and were all seizure-free postoperatively (Engel’s class I; Table 1). The FCD cases included in this study have all the histopathological features of severe (type IIB) FCD, according to the classification of Palmini et al. [48]. The resected specimens consisted of disorganized neocortex containing immature neurons, giant neurons, dysmorphic neurons, and balloon cells. Neurons and balloon cells were also observed in the subcortical white matter and there was a prominent population of reactive astrocytes. Cells of the microglia/macrophage lineage were also observed within the dysplastic cortex, suggesting activation of inflammatory processes in FCD [6]. Expression of VEGF and VEGFR in normal temporal cortex and FCD Cellular distribution of VEGFA VEGFA staining was observed within the histologically normal cortex (Fig. 2a, b). The staining was strongest in pyramidal neurons, which displayed somatic staining and staining of the apical dendrites (Fig. 2b). Neuropil staining was weak and resting glial cells did not show VEGFA immunoreactivity (IR). Weak staining was observed in endothelial cells. Autopsy material and the perilesional cortex showed similar IR. Fig. 2VEGFA immunoreactivity in focal cortical dysplasia type IIB. a VEGFA immunoreactivity (IR) within the histologically normal adult cortex. b High magnification showing somatic staining in pyramidal neurons (inset: vascular staining). c VEGFA in focal cortical dysplasia (FCD; type IIB) showing strong IR within the dysplastic cortex. d High magnification showing VEGFA IR in dysplastic neurons (arrows). e A binucleated VEGFA positive dysplastic neuron. f VEGFA IR within the subcortical dysplastic region. g VEGFA expression in a balloon cell. h, i VEGFA expression in reactive astrocytes (arrows in i indicate perivascular astrocytic end-feet) j–l double-labeling of GFAP (green, j) with VEGFA (red, k) shows co-localization (yellow, l) in astrocytes. m–o Double-labeling of non-phosphorylated neurofilament (SMI311; NF, green, m) with VEGFA (red, n) shows co-localization (yellow, o) in dysplastic neurons. p Merged image showing co-localization of CD68 (green) with VEGFA (red) in macrophages. Scale bar in a a, c, f 400 μm; b, d 120 μm; e, g–i 35 μm; j–o 40 μm; p 18 μm In the majority of FCD cases, strong VEGFA immunoreactivity (IR) was observed in dysplastic neurons located throughout the dysplastic cortex (Fig. 2c–e; Table 2; Fig. 6). Strong staining was also detected in balloon cells, reactive astrocytes, and in perivascular astrocytic end-feet (Fig. 2f–i; Table 2; Fig. 6). Endothelial IR was weak. Double-labeling experiments confirmed expression in reactive astrocytes, neurons, and in CD68+ macrophages (Fig. 2p). Immunocytochemistry using two different antibodies to VEGF (Pharmingen and Santa Cruz Bio.) demonstrated similar patterns. Cellular distribution of VEGFB Histologically normal cortex displayed only weak VEGFB IR (Fig. 3a). Both autopsy and surgical specimens showed light staining in pyramidal neurons and in endothelial cells. Glial cells did not show VEGFB IR. Fig. 3VEGFB immunoreactivity in focal cortical dysplasia type IIB. a Histologically normal adult cortex, showing neuronal distribution of VEGFB with weak immunoreactivity (IR) in pyramidal cell neurons (high magnification of a pyramidal neuron is shown in the inset aI); weak staining was also observed in blood vessels (inset aII). b VEGFB in focal cortical dysplasia (FCD; type IIB) showing strong IR within the dysplastic cortex. c Strong VEGFB IR in dysplastic neurons of different size and shape (arrows). d Undetectable VEGFB IR in balloon cells (arrows). e–g Absence of co-localization between GFAP (green, e) with VEGFB (red, f) in astrocytes (g, merged image). h–j Double-labeling of non-phosphorylated neurofilament (SMI311; NF, green, h) with VEGFB (red, i) shows co-localization (yellow, j) in dysplastic neurons. k Merged image showing co-localization of CD68 (CD68; green) with VEGFB (red) in macrophages. Scale bar in a a, b 200 μm; c–j 40 μm; k 18 μm In FCD specimens, moderate to strong VEGFB IR was observed within the dysplastic cortex (Fig. 3b; Table 2; Fig. 6) with strong VEGFB IR in dysplastic neurons (Fig. 3c). In the majority of cases (seven out of nine), balloon cells and reactive astrocytes did not express VEGFB (Fig. 3d; Table 2; Fig. 6). Double-labeling experiments confirmed the absence of VEGFB IR in GFAP-positive cells (astrocytes), whereas co-localization was found with neurofilament in dysplastic neurons (Fig. 3h–j). VEGFB IR was also observed in CD68+ macrophages (Fig. 3k). Cellular distribution of VEGFR-1 Histologically normal cortex (autopsy and surgical specimens) displayed only weak VEGFR-1 IR, which was restricted to pyramidal neurons (Fig. 4a). IR in blood vessels was weak (Fig. 4b). Glial cells did not show VEGFR-1 IR. Fig. 4VEGFR-1 immunoreactivity in focal cortical dysplasia type IIB. a, b Histologically normal adult cortex (a) and white matter (b) showing weak immunoreactivity (IR) in neurons (a) and blood vessels (b; arrowheads). c, d VEGFR-1 in focal cortical dysplasia (FCD; type IIB) showing strong IR in dysplastic neurons (c; arrows) and in blood vessels (d; arrows). Inset in c Co-localization between non-phosphorylated neurofilament (SMI311; NF, green) with VEGFR-1 (red) in dysplastic neurons. Inset in d Co-localization between CD31 (endothelial marker; green) with VEGFR-1 (red) in blood vessels. e–g Strong IR in balloon cells of different size (arrows) and glial cells (arrowheads in e). h–j Co-localization between GFAP (green, h) with VEGFR-1 (red, i) in astrocytes (j, merged image). k–m Co-localization of CD68 (CD68; green, k) with VEGFR-1 (red, l) in macrophages (m, merged image). Scale bar in a a–d and h–m 40 μm; e–g 35 μm In FCD specimens, moderate to strong VEGFR-1 staining was observed within the dysplastic cortex (Fig. 4c–g; Table 2; Fig. 6). VEGFR-1 IR was observed in different cell types, including dysplastic neurons, astrocytes, and endothelial cells. Double-labeling experiments confirmed the co-localization of VEGFR-1 IR with neuronal (inset in Fig. 4c), endothelial (inset in Fig. 4d), and glial (Fig. 4h–j) markers. Additionally, expression of VEGFR-1 was observed in CD68+ macrophages (Fig. 4k–m). Cellular distribution of VEGFR-2 VEGFR-2 staining was observed within the histologically normal cortex (autopsy and perilesional zone) in pyramidal neurons (Fig. 5a). IR in blood vessels was weak (Fig. 5b). Glial cells did not show VEGFB IR. Fig. 5VEGFR-2 immunoreactivity in focal cortical dysplasia type IIB. a, b Histologically normal adult cortex (a) and white matter (b) showing moderate immunoreactivity (IR) in pyramidal cells (a); in the white matter, endothelial IR was weak and glial IR was under detection level (b). c, d VEGFR-2 in focal cortical dysplasia (FCD; type IIB) showing strong IR in dysplastic neurons (arrows in c), but weak IR in blood vessels (arrowheads in d). e, f Strong IR in a dysplastic neuron (arrow in e) and in a balloon cell (arrow in f), but weak IR in glial cells (arrowheads in e, f). Inset in e Co-localization between non-phosphorylated neurofilament (SMI311; NF, green) with VEGFR-2 (red) in a dysplastic neuron. g–i Co-localization between MAP-2 (green, g) with VEGFR-2 (red, h) in balloon cells (i, merged image). Scale bar in a a–d and g–i 40 μm; e, f 35 μm In the large majority of FCD cases (eight out of nine), VEGFR-2 was strongly expressed in dysplastic neurons (Fig. 5c, e; Table 2; Fig. 6). VEGFR-2 IR was also detected in balloon cells, but only three out of nine cases displayed strong staining for VEGFR-2 (Fig. 5f; Table 2). Endothelial expression was weak. In many FCD cases (five out of nine), expression of VEGFR-2 was undetectable in reactive astrocytes (Table 2). Double-labeling experiments confirmed the co-localization of VEGFR-2 IR with neuronal markers (neurofilament or MAP2, Fig. 5e, g–i) within the dysplastic cortex. Fig. 6Distribution of immunoreactivity scores (total score; see details in “Materials and methods” section) in neurons and astrocytes of control, perilesional, and FCD specimens. a, e VEGFA; b, f VEGFB; c, g VEGFR-1; d, h VEGFR-2. a–d Neurons; e–h astrocytes. IR scores of VEGFs and VEGFRs in neurons of FCD were greater than IR scores of VEGFs and VEGFRs of control and perilesional cortex; scores of VEGFA and VEGFRs in astrocytes of FCD were greater than IR scores of VEGFA and VEGFRs of control and perilesional cortex (P < 0.05). There were no significant differences in IR scores of VEGFs and VEGFRs between control and perilesional cortex Discussion In addition to their role in angiogenesis, VEGF proteins and their receptors have been implicated in several neurological disorders, including epilepsy [13, 44, 56]. In the present study, we demonstrate a prominent expression of VEGFA, VEGFB, and their signaling receptors in FCD type IIB, a malformation of cortical development associated with intractable epilepsy. This is particularly interesting in view of the recently proposed role of VEGFs and their signaling pathways during development and in epilepsy-associated pathologies [13, 26, 33, 69, 76]. Expression of VEGFA and VEGFB in normal temporal cortex In histologically normal temporal cortex (autopsy and perilesional zone), we have shown weak expression of both VEGFA and VEGFB in cortical neurons. Expression of VEGFA and VEGFB, including both mRNA and protein, has been demonstrated in neurons in adult rodent brain [39, 41, 42, 72]. In human adult brain, only few studies have described neuronal expression of VEGFA in control tissue [8, 78], which was similar to our observed staining pattern in the control temporal specimens. To our knowledge, previous studies of VEGFB protein expression in human control cortex have not been described. However, VEGFB mRNA was detected in human hippocampal cortex [24], and VEGFB mRNA and protein expression has been described in adult rodent brain [36, 42]. In control rat brain, VEGFB was constitutively expressed in endothelial cells [42]. In our study, we observed only weak endothelial VEGFB IR in blood vessels. In agreement with previous studies [8, 36, 78], immunoreactivity (IR) for both VEGFs was not observed in glial cells within control specimens. Differential cellular distribution of VEGFA and VEGFB in FCD In the present study, we provide evidence for a consistent expression of both VEGFA and VEGFB within the dysplastic cortex of patients with FCD. Both VEGFs are highly expressed in dysplastic neurons; however, only the VEGFA protein is prominently expressed in reactive astrocytes. Expression of VEGFA in astrocytes has been shown in several other pathologies associated with reactive gliosis, such as ischemic stroke, traumatic brain injury, neurodegenerative disorders, and the hippocampus following entorhinal deafferentation [8, 27, 59, 78, 79]. In addition, we previously reported upregulation of both neuronal and glial VEGFA expression in patients with hemimegalenchephaly, an epilepsy-associated malformation of cortical development [5, 7]. In the present study, we also observed expression of VEGFA in balloon cells, which are characteristic cell types of severe FCD [48]. Whether these cells are glial or neuronal in nature is still controversial [12]. Induction of both neuronal and astroglial VEGFA expression has been shown in different experimental models of seizures and human temporal lobe epilepsy (TLE) [13, 44, 56]. Rigau et al. [56] showed increased levels of VEGFA in the hippocampus of several cases of TLE, including two cases of focal dysplasia. All TLE cases showed VEGFA expression in pyramidal neurons and granule cells of the hippocampus [56]. However, immunocytochemical analysis of the temporal cortex and the FCD cases was not performed. The molecular mechanism underlying the induction of VEGFA expression after seizures remains unclear. One possible mechanism, which has been proposed to explain the association between seizure activity and the induction of VEGFA expression, is represented by the stabilization of the hypoxia inducible factor-1α (HIF-1α). HIF-1α is a transcription factor which upregulates VEGFA transcription under hypoxic conditions [21, 38, 61]. Hypoxia may occur during seizures, representing an important trigger in the induction of VEGFA expression, particularly in case of long lasting seizures, such as in status epilepticus models. However, VEGFA expression is already induced after acute seizures [44] and the mechanisms that regulate VEGFA expression are complex. Several transcription factors, including AP-1, HIF-1α, and NF-κB, have been identified to regulate VEGFA expression [31], and recently it has also been shown that inflammatory cytokines, such as interleukin-1β (IL-1β), activate HIF1α and VEGFA gene expression in primary human astrocytes [2]. Interestingly, increased expression of proinflammatory cytokines and related molecules has been reported in both animal models and human epilepsy-associated pathologies, including FCD [3, 15, 54, 77]. In addition, VEGFA has been demonstrated to be a key mediator of the inflammatory process [14, 35]. Thus, we might speculate that the prominent expression of VEGFA within the dysplastic cortex could be a critical component of the complex cascade of events leading to a chronic inflammatory state and the sustained seizure activity [76, 77]. With respect to inflammation, inflammatory cells, such as macrophages, can also release various angiogenic cytokines including VEGFs [71]. Accordingly, we observed expression of both VEGFA and VEGFB in macrophages (CD68-positive cells), as previously shown in animal models of brain ischemia [11, 42, 52]. VEGFA effects can also compromise the integrity of the blood–brain barrier (BBB; [60]). Interestingly, alterations of the BBB permeability have been recently observed in both human and experimental temporal lobe epilepsy with positive correlation between the increased vascular permeability and the occurrence of spontaneous seizures in chronic epileptic rats [55, 56, 75]. In contrast, several studies highlight a dichotomous function of VEGFA, also demonstrating a neuroprotective role [13, 47]. Administration of VEGFA and neuronal expression of VEGFA have been shown to stimulate neurogenesis in vitro and in vivo [9, 28]. In addition, it has been suggested that the neuroprotective effects of VEGFA are mediated by the neuronal VEGFR-2 and the subsequent activation of the PI3K/Akt survival pathway [32, 67]. VEGFB expression is not induced by hypoxia or several transcription factors known to regulate VEGFA expression [18], as the promoter region of VEGFB lacks HIF-1 and AP-1 sites [43, 62]. The regulation of the expression of VEGFB remains unknown. Since all cases examined were associated with epilepsy, we cannot exclude that chronic seizure activity could also contribute to the VEGFB expression in FCD specimens. Alternatively, since VEGFB expression has been shown to be prominent during early brain development [36], the strong neuronal expression of VEGFB could represent an intrinsic and immature feature of the dysplastic neuronal cells that could contribute to their survival. Recent studies using VEGFB knock-out mice demonstrate a neurotrophic and neuroprotective activity of VEGFB, exerting a direct action on neurons, and promoting neurogenesis [68, 69]. This is an observational study and we were, therefore, not able to investigate the spatio-temporal regulation of the VEGF system. Further research in animal models of cortical dysplasia is clearly needed to elucidate the role of VEGFs and their signaling pathways in the histogenesis or epileptogenesis of developmental disorders. Expression of VEGF receptors in normal temporal cortex In histologically normal temporal cortex (autopsy and perilesional zone), VEGFR-1 and VEGFR-2 showed a similar pattern of expression, with weak to moderate immunostaining in pyramidal neurons. Neuronal expression of VEGFRs’ mRNA and protein has been reported in adult human and rodent brain, with strong expression in the hippocampus [8, 11, 78, 79]. In agreement with these studies, we did not detect glial VEGFR expression in histologically normal cortex and only weak VEGFR expression was observed in endothelial cells. Differential cellular distribution of VEGFR-1 and VEGFR-2 in FCD Consistent expression of both VEGFR-1 and VEGFR-2 was detected within the dysplastic cortex of patients with FCD. Both receptors were upregulated in dysplastic neurons. Increased expression of the VEGFRs and, in particular VEGFR-1, was observed in reactive astrocytes. Upregulation of VEGFRs in neurons and reactive astrocytes has been shown in several other pathological conditions including ischemia, neurodegenerative diseases, and trauma [8, 11, 37, 39, 66, 79]. Recently, increased expression of VEGFR-2 has been shown in several cases of TLE, including two cases of cortical dysplasia [56]. Immunocytochemical analysis demonstrated expression only in endothelial cells, whereas neuronal VEGFR-2 IR was not detected in either control hippocampus or TLE specimens [56]. Differences in the phenotypes of cells expressing VEGFRs have been observed in several other studies [8, 30, 52]. These discrepancies may be caused by differences in experimental methods, tissue processing, or the use of different antibodies that recognize different epitopes. Our results support the neuronal expression of VEGFR-2 recently reported in human brain [8]. The similar expression pattern of VEGFA and VEGFR-2, with prominent neuronal IR, suggests autocrine/paracrine effects on dysplastic neurons, supporting the hypothesis of a mechanism to protect abnormal neurons from cell death associated with seizures. Autocrine and/or paracrine effects of VEGFA are supported by the observation that administration of VEGFA has been shown to induce mRNA and protein expression of both receptors in adult rat brain [34, 57]. A protective mechanism of VEGFA has been suggested in epileptic rats, showing that VEGFA may reduce spontaneous discharges in epileptic rats [40]. Therefore, upregulation of VEGFA could represent an endogenous compensatory mechanism to reduce excitability and to prevent cell loss after severe seizures. Accordingly, infusion of VEGFA into the hippocampus has been shown to protect against neuronal cell loss after pilocarpine-induced status epilepticus [45]. In the present study, we also provide evidence for the expression of VEGFR-1 in activated cells of the microglia/macrophage lineage, which have been shown to be present in FCD specimens [6]. This is in agreement with previous in vitro and in vivo studies [11, 20] showing VEGFR-1 expression in activated microglial cells. These observations suggest that the microglia/macrophage lineage is also a target for VEGF, which may affect chemotaxis and proliferation of these cells, contributing to the inflammatory state in the epileptic brain. There is substantial information about the function and the signaling through VEGFR-2; in contrast, signaling through VEGFR-1 remains poorly understood and has been a matter of discussion. A decoy role has been proposed for VEGFR-1, but more recently, functional signaling via VEGFR-1 has been reported (for reviews see [43, 67]). These observations may give rise to new therapeutic strategies focusing on VEGFR-1 specific ligands, such as VEGFB [43]. Although a rapid induction of VEGF and its receptors has been shown in different experimental models of seizures [13, 44, 56], seizures alone cannot account for changes in neuronal and glial expression in FCD since perilesional tissue was exposed to seizures but did not show significant upregulation of VEGFs and/or VEGFRs. Therefore, the lesion per se or the concomitant presence of the lesion and the epileptic activity, are likely to play a role in modulating the VEGF system in these developmental disorders. Conclusions Our observed cellular distribution of VEGFA, VEGFB, and their signaling receptors indicate that different cellular components of FCD are involved in VEGF-signaling. In this context, future studies, using both in vivo and in vitro models, will be important to achieve a better understanding of the role of the VEGF-mediated pathways in the histogenesis and epileptogenesis of developmental lesions associated with intractable chronic epilepsy. Presently, signaling via VEGF receptors is not targeted by existing therapies in epileptic patients, but it can be potentially useful in view of its involvement in the regulation of neurogenesis, inflammation, and BBB integrity. However, an effective therapeutic intervention based on modulation of the VEGF system has to take into consideration the specific role of VEGFA and VEGFB and the multiple effects (protective and/or detrimental) reported for VEGFA.	[ "vascular endothelial growth factor", "receptors", "epilepsy", "dysplastic neurons", "immunocytochemistry", "astrocytes" ]	[ "P", "P", "P", "P", "P", "P" ]
Arch_Dermatol_Res-4-1-2254658	Lymphocyte subsets in peripheral blood of patients with moderate-to-severe versus mild plaque psoriasis	In several studies peripheral blood T-cells have been quantified, yet few data are available on lymphocyte subsets in moderate-to-severe psoriasis (in terms of extent and activity of lesions) versus mild psoriasis. The objective is to compare lymphocyte subsets in peripheral blood of patients with moderate-to-severe disease (PASI-score ≥12) to patients with mild disease (PASI-score <12) and to healthy subjects. By means of flow cytometry method, lymphocytes in peripheral blood of 27 patients with psoriasis and 10 healthy controls were characterized. The absolute number of total lymphocytes was markedly decreased in patients with moderate-to-severe psoriasis as compared to patients with mild disease and normal subjects. Cellcounts of all analysed subsets were found to be increased in more severe psoriasis, except for CD8+CD45RO+ cells. The under-representation of CD8+CD45RO+ cells is compatible with the dynamics of acquired immunity, which requires a time log after the relapse of the lesions to differentiate from CD45RA+ naive cells. Introduction Psoriasis is a common inflammatory skin disease characterized by hyperproliferation of keratinocytes. It is a well established fact that T-cells play an important role in the pathogenesis of psoriasis [6, 9]. Indeed, treatments such as anti-CD4 [5, 15, 19], anti-CD11 [5, 13] and anti-CD25 [13], targeting at specific T-cells, have shown to be effective in psoriasis. Several T-cell subsets seem to play a primary role. Major classes are CD4+, CD8+, CD45RO+ and CD45RA+ T-cells. Recently, NK-T cells have been suggested to play an additional role in the regulation of immunity, through the release of cytokines [3, 14, 16]. NK-T cells are cells bearing both T cell receptors as well as natural killer-cell specific receptors such as CD161. It has been shown that circulating NK-T cells are significantly reduced in some autoimmune diseases [8, 17]. The large body of research on the role of T-cells in psoriasis has been focused on lesional T-cells. This has resulted in the hypothesis that there is a final common pathway responsible for the development of chronic plaque psoriasis, which may involve specific antigen recognition by T-cells that results in stimulation of keratinocyte proliferation. In psoriasis, CD4+ cells seem to be important mainly in the early phase of plaque development. These cells are found predominantly in the upper dermis, whereas CD8+ cells have proven to be relevant during chronic phases and are found predominantly in the epidermis [21], although others have found a more early involvement of CD8+ T-cells in the development of psoriatic plaques [22]. A large number of activated T-cells have been shown in clinically involved skin of psoriatic patients, but also uninvolved skin contains a significant number of T-cells. In healthy skin, however, T-cells can hardly be found. It has been suggested that circulating T-cells are activated and subsequently recruited from the circulation during the development of psoriatic plaques [2, 10]. Several studies have investigated peripheral blood T-cells of psoriatic patients [1–3, 7, 10, 12, 14, 20, 21]. These studies indicated that the total amount of T-cells in patients is comparable or slightly increased as compared to that found in normal subjects. No relevant differences have been shown with respect to total T-cell counts and T-cell subsets including CD4+, CD8+, CD45RO+ and CD45RA+ cell counts. Only few studies comprise a comparison between mild and more severe forms of psoriasis, using clinical severity indicators such as the Psoriasis Area and Severity Index (PASI-score) [4, 18]. The present study compares specific circulating lymphocyte subsets in patients with mild psoriasis, patients with moderate-to-severe psoriasis and normal subjects. Materials and methods Patients and controls Fifteen patients with moderate-to-severe psoriasis vulgaris (12 male and three female aged 19–66, mean age 46.2 years) and 12 patients with mild psoriasis (seven male and five female, aged 34–72, mean age 52.8 years) from our outpatient department participated in this study. Mean PASI-scores in both groups were 20.97 ± 2.55 (mean ± SEM) and 6.11 ± 1.27, respectively. Patients were classified into one of both groups based on their PASI-score. We considered all patients with a PASI-score <12 as having a “mild”, and all patients with a PASI >12 as having a “moderate-to-severe” psoriasis [4, 18]. Patients were free of any systemic therapy for at least 4 weeks and did not use any topical therapy in the last 2 weeks. Peripheral blood was obtained from all subjects with their written informed consent. Control samples were collected from 10 healthy volunteers without any history or signs of skin disease (four male and six female aged 24–49, mean age 33.8 years). Preparation of PBMC’s For each patient, the exact amount of blood withdrawn was determined by measuring the height of the column of blood in the tube, which was subsequently converted to the corresponding volume in microliter. PBMC’s were isolated from heparinized blood by density centrifugation on polyester gel (Becton Dickinson Vacutainer™ CPT™, Franklin Lakes NJ, USA). After filtering through a 70 μm cellstrainer, cells were washed twice. For flow cytometry, single-cell suspensions (concentration 5 × 105 cells ml−1) were stained in 1% fetal calf serum in phosphate-buffered saline (PBS) at concentrations recommended by the manufacturer. Additionally, 450 μl propidium iodide (PI) in PBS was added to each sample in order to exclude non-viable cells from analysis. The following monoclonal antibodies were used: fluorescein isothiocyanate (FITC)-conjugated anti-CD4, FITC-conjugated anti-CD8, phycoerythrin (PE)-conjugated anti-CD4, PE-conjugated anti-CD8, PE-conjugated anti-CD94 and PE-conjugated anti-CD161 (all from Immunotech, Marseille, France), as well as FITC-conjugated anti-CD45RA (Becton Dickinson, San Jose, CA, USA) and FITC-conjugated anti-CD45RO (DAKO, Glostrup, Danmark). Flowcytometric analysis Cells were analyzed with an EPICS Elite flow cytometer (Coulter, Luton, UK), using the forward scatter as a discriminator. Lymphocytes were identified by gating on CD45 and side and forward scatter properties. All samples were processed within 18 h of phlebotomy. The following (combinations of) PE and FITC-conjugated reagents were used to determine the expression of each antigen or antigen combination on lymphocytes derived from peripheral blood: CD4+ (marker for T-helper cells), CD8+ (marker for cytotoxic T cells), CD45RA+ (marker for naïve T-cells), CD45RO+ (marker for memory T-cells), CD94+ (marker for NK-cells, T cells), CD161+ (marker for NK-cells, T-cells with memory phenotype), CD4+ CD45RA+, CD4+ CD45RO+, CD8+ CD45RA+, CD8+ CD45RO+, CD8+ CD94+ and CD8+CD161+. Determination of absolute numbers of lymphocytes Enumeration of positive cells was performed by adding Flow Count Beads (Beckman Coulter, Fullerton, CA, USA) to the cell suspension of PBMCs. An automatic stop after a defined amount of beads was programmed on the flow cytometer. Absolute counts of lymphocyte subsets per blood sample were calculated by determining the ratio of the beads to the cell population and then multiplying this ratio by the number of beads in the tube. By dividing this count by the amount of microliters in the tube, a relative absolute lymphocyte count in cells μl−1 blood withdrawn was obtained. Analysis was performed with Verity software. Ratios for CD4+/CD8+ cells, CD45RO+/CD45RA+ cells, CD4+CD45RO+/CD4+CD45RA+ cells and CD8+CD45RO+/CD8+CD45RA+ cells were calculated afterwards, with data derived from this analysis. Statistical analysis Data-entry and analysis was performed using Statistica 6.0 software. Means and standard deviations were calculated for each parameter and were tested with one-way ANOVA. Differences were considered statistically significant at p < 0.05. Results In total, 27 patients with psoriasis and 10 normal subjects without signs or symptoms of skin disease were included in the present study. Table 1 summarizes the demographic details and psoriasis related characteristics of the patients with moderate-to-severe psoriasis, patients with mild psoriasis and normal subjects. Four out of 12 patients with mild psoriasis had shown a minimal increase in the extent of lesions; in the others the skin abnormalities were stable. With respect to the 15 patients with moderate-to-severe-psoriasis, 12 had an increase of the extent of lesions during the previous 4 weeks, whereas the other three had been stable. Table 1Demographic and psoriasis related characteristics of patient and subjects participating in the study (mean ± SEM)Moderate-to-severe psoriasisMild psoriasisNormal subjectsDemographicsAge (years)46.2 ± 3.4552.8 ± 3.4933.8 ± 3.1Male:female ratio 12:37:54:6Psoriasis related characteristicsPASI-score 20.97 ± 2.556.11 ± 1.27NADuration (years)16.7 ± 2.722.2 ± 2.5NA Total cell counts The total amount of gated cells (per microliter blood withdrawn) in normal subjects was 2,220 ± 301 cells μl−1 (mean ± SEM). The cell-count in mild psoriasis was 1,923 ± 230 cells μl−1. In contrast, patients with moderate-to-severe psoriasis had cell counts of 498 ± 95 cells μl−1. No statistically significant difference was observed between normal subjects and patients with mild psoriasis. However, a highly significant difference (p < 0.0001) was observed between mild and severe psoriasis. The CD4+/CD8+ ratios in all three groups were in the same range, meaning no shift occurred in the distribution of these subgroups of peripheral blood lymphocytes. The total number of lymphocytes, and the CD4+ and CD8+ counts of the patients with moderate-to-severe disease were also routinely measured at the GLP certified Central Haematology Laboratory in our hospital, in which the whole blood lysing method is the standard. Comparison of both methods reveals an acceptable resemblance in results, with lower counts of maximum 10%. Lymphocyte subsets Lymphocyte subsets as a percentage of total gated cells are summarized in Fig. 1. CD4+, CD8+, CD45RA+ and CD45RO+ cells are increased in patients with severe psoriasis as compared to patients with mild psoriasis and normal subjects (p < 0.0001 in all cases). Fig. 1CD4, CD8, CD45RO and CD45RA subpopulations, expressed as a percentage of total lymphocyte counts. Comparison between normal subjects, patients with mild- and patients with moderate-to-severe psoriasis. (: p < 0.0001) No significant differences were observed between patients with mild psoriasis and normal subjects. Double labeling of lymphocytes revealed that the percentages of CD4+CD45RO+, CD4+CD45RA+ and CD8+CD45RA+ cells were significantly increased in moderate-to-severe versus mild psoriasis (p < 0.001 in all cases), whereas the percentage of CD8+CD45RO+ cells was not increased significantly (Fig. 2). Cell ratios for CD4+/CD8+ cells, CD45RO+/CD45RA+ cells, CD4CD45RO+/CD4CD45RA+ cells and CD8CD45RO+/CD8CD45RA+ cells are summarized in Fig. 3. Fig. 2Expression of CD4+CD45RO+, CD4CD45RA+, CD8+CD45RO+ and CD8CD45RA+ cells, as percentage of the total amount of lymphocytes, in three subgroups: moderate-to-severe psoriasis, mild psoriasis and normal subjects (: p < 0.001)Fig. 3Ratios of CD4/CD8, CD45RO+/CD45RA+, CD4CD45RO+/CD4CD45RA+ and CD8CD45RO+/CD8CD45RA+ in normal subjects, mild and moderate-to-severe psoriasis (: p < 0.05) Lymphocytes expressing NK cell receptors CD94+ and CD161+ cells, expressed as a percentage of total cells were increased in patients with moderate-to-severe psoriasis, as compared to patients with mild psoriasis and normal subjects (p < 0.05 and p < 0.001, respectively). Figure 4 summarizes these observations. Again no statistically significant differences were observed between normal subjects and patients with a mild form of psoriasis. Fig. 4CD94 and CD161 subpopulations, expressed as a percentage of total lymphocyte counts. Comparison between normal subjects, patients with mild- and patients with moderate-to-severe psoriasis (: p < 0.05 and p < 0.001 respectively) With respect to double stained cells, we observed an increase in the expression of both CD8+CD94+ cells as well as CD8+CD161+ cells (markers for NK T subpopulations) in patients with a more extensive form of psoriasis, expressed as a percentage of the total amount of cells, as compared to patients with mild psoriasis and normal subjects (p < 0.008 and p < 0.009, respectively, results shown in Fig. 5). Fig. 5Cellcounts of CD8+CD94+ and CD8+CD161+ cells as percentage of total lymphocytes (*: p < 0.009) Correlation between PASI-score and lymphocyte subsets and cells expressing NK receptors Linear regression analysis revealed no significant correlation between the individual PASI-scores and the quantifications of lymphocyte subsets or cells with NK-cell receptors. Correlation analysis between disease activity and peripheral lymphocytes did not reveal a difference between patients with spreading lesions and those with stable plaque psoriasis. Discussion The results of this study show marked differences in absolute cellcounts between the three patient groups. Our main goal was to investigate whether there is a difference in the occurrence of biomarkers on circulating lymphocytes in mild psoriasis versus more severe forms of the disease. One of the criteria to qualify for a treatment with major systemic therapies such as a biological is often a minimum PASI-score of 12 [4, 18]. We were interested whether this cut-off point also implies a differentiation between mild versus moderate/severe psoriasis in terms of T-cell and NKT-cell subsets. The results of the present study have shown us that there are indeed marked differences between mild psoriasis and the more severe forms of disease defined by cut-off point PASI 12. A decrease of total lymphocyte counts, as percentage of PBMC, and an increase in the percentage of CD4+CD45RO+ cells in patients with moderate-to-severe psoriasis has been reported before [10]. These patients (32 subjects) had an average PASI of 25.7 (10–48), and all had had an exacerbation following infection. Patients had a mixed morphology, from guttata psoriasis to chronic plaque psoriasis. In the same study, however, a decrease in the amount of CD4+ cells was described, which is in contrast with the findings in our study. Other authors have reported no significant difference in total T-cells as percentage of PBMC counts between patients with psoriasis and normal subjects [3]. These patients (14 subjects) had an average PASI of 23.5 ± 10.6. However, no information was given whether these patients had relapsing psoriasis. In the present study, the patients with severe psoriasis had an extension of lesions during the wash out period (average increase in PASI: 4.9). Therefore, the severity in our patients not only comprises the extent of lesions but also activity of lesions with progression and therefore, approaches the characteristics of the patients reported by Lecewicz-Torún et al. [10], who also reported a decrease of T-cells in peripheral blood in psoriasis (expressed as percentage of total PBMC). The decrease of lymphocyte counts, expressed as a percentage, in peripheral blood in patients with moderate-to-severe psoriasis is the reverse of the substantial accumulation of lymphocytes in psoriatic skin. The most likely explanation for this paradox is an increased influx of cells into the skin of patients with severe disease. Further studies, in which simultaneous measurements of lesional and circulating T cells will be performed, need to be carried out to support this hypothesis. A striking observation is the increased percentage of both CD4+ cells and CD8+ cells in patients with moderate-to-severe psoriasis as compared to mild psoriasis and normal subjects, without a change in CD4/CD8 ratio. So far, reports on the CD4+/CD8+ ratio in psoriasis have been at variance [10, 11]. With respect to the double labeled lymphocytes, both the percentage of CD4+CD45RO+ cells as well as the percentage of CD4+CD45RA+ cells is increased in patients with moderate-to-severe psoriasis, in contrast to the absolute decrease of total lymphocytes. The relative count of both subsets is increased, whereas their ratio shows a tendency to decrease (p = 0.0509). Thus, with increasing severity of disease, the increase in CD4+CD45RA+ cells seems to exceed the increase in CD4+CD45RO+ cells. An explanation for this phenomenon is the preferential migration of CD4+CD45RO+ cells into lesional skin in extensive disease, resulting in a relative decrease of this subset of CD4+ lymphocytes in peripheral blood. This hypothesis supports earlier studies where CD4+CD45RO+ cells have been shown to be the major players in psoriatic plaques. Indeed, CD4+ cells appear in a very early phase of the psoriatic process. In a previous study, CD4+ cells have been detected in the distant uninvolved skin of patients with active psoriasis [22]. The percentage of CD8+CD45RO+ cells in patients with moderate-to-severe psoriasis approached the percentages in patients with mild psoriasis and normal subjects, in contrast with the CD8+CD45RA+ cells, which did show a significant increase in peripheral blood. The impressive under-representation of CD8+CD45RO+ lymphocytes as compared to CD8+CD45RA+ lymphocytes in moderate to severe psoriasis suggests a preferential influx of CD8+CD45RO+ cells into the skin. Both CD8+ cells, as well as CD45RO+ cells are abundantly present in psoriatic lesional skin. Previously, we have shown that in the outer margin of the spreading psoriatic plaque the number of CD8+ cells and CD45RO+ cells were increased [22]. Therefore, the relative under-representation in peripheral blood of CD45RO+ cells, in particular the CD8+ CD45RO+ cells, most likely reflect the influx of these cells into the skin in patients with moderate-to-severe psoriasis in an active phase of disease. However, further studies are needed to support this hypothesis. Lymphocyte counts, preferably T-cells, in skin as well as in the peripheral blood should be monitored during a longer period of time, e.g., during exacerbation and during disease-free intervals. The CD8+CD94+ cells and CD8+CD161+ cells show a marked increase in patients with moderate-to-severe psoriasis as compared to mild psoriasis and normal subjects. The increases of these lymphocytes expressing NK receptors in peripheral blood are accompanied with increases of these cells in the psoriatic lesion. The results of the present study suggest that in peripheral blood of patients with moderate-to-severe psoriasis, in contrast to the vast presence of T-cells in skin lesions, the absolute lymphocyte counts are profoundly decreased, without any change of CD4/CD8 ratio. The relative under-representation of CD45RO+ cells, as compared to CD45RA+ cells, in peripheral blood most likely results from a massive influx of memory effector T-cells into the psoriatic lesions in an active phase of disease. Both in blood and skin of patients with moderate-to-severe psoriasis the percentages of CD94+ and CD161+ cells are increased. These results are compatible with the functions of these cells. CD45RO+ cells differentiate from CD45RA+ cells and represent acquired immunity, whereas NK-T cells are able to act without prior sensitization, representing innate immunity [3]. In further research, more selective lymphocyte markers are required in order to investigate the role of peripheral blood T-cells in the pathogenesis of psoriasis, and correlation studies between peripheral blood T-cells and T-cells in skin are required in order to understand the dynamics of compartmentalization of T-cell subsets belonging to either the acquired or the innate immunity system in psoriasis.	[ "lymphocytes", "peripheral blood", "psoriasis", "t cells", "severity of disease" ]	[ "P", "P", "P", "P", "P" ]
Pituitary-3-1-1915635	Progressive improvement of impaired visual acuity during the first year after transsphenoidal surgery for non-functioning pituitary macroadenoma	Objective Improvement of visual field defects continues even years after the initial surgical treatment. Because this process of continuing improvement has not been documented for visual acuity, we audited our data to explore the pattern of recovery of visual acuity until 1 year after transsphenoidal surgery for non-functioning pituitary macroadenoma. Introduction Non-functioning pituitary macroadenomas (NFMA) are the most prevalent macroadenomas [1, 2]. Patients with NFMA mainly present with decreased visual acuity, visual field defects and hypopituitarism, caused by mass effects of the tumor [3–5]. Transsphenoidal surgery is the treatment of choice, resulting in improvement of visual field defects in 75–100% of all patients [3–5]. Accordingly, visual acuity improves in the majority of patients [6–10], although visual field defects and visual acuity may worsen in a limited number of patients after surgery [6, 7, 9, 10]. The process of visual fields recovery starts directly after surgery and can already be documented on the second postoperative day [11]. This process of recovery is probably due to restoration of the velocity of conduction. However, improvement of visual field defects continues even years after the initial surgical treatment [12, 13]. Although a reduced visual acuity is correlated with the extent of visual field defects [13, 14], postoperative changes in visual acuity do not parallel changes in visual fields in every case [6]. To our knowledge, the process of gradual visual improvement, has only been observed for visual field defects, but not for visual acuity [12]. This is, however, clinically relevant, since improvement of minor visual field defects can occur unnoticed, whereas improvement in visual acuity is almost invariably noticed, and can potentially overcome thresholds to previous impairments in daily life. We therefore audited our data to explore the pattern of recovery of visual acuity until 1 year after transsphenoidal surgery for non-functioning pituitary macroadenoma. Patients and methods Patient selection Forty-three patients, mean age 56 ± 14 years, were included in this retrospective analysis. In order to obtain a homogenous cohort for assessment, the inclusion criteria were the following:Compression of the optic chiasm on MRI by the pituitary tumor prior to surgeryTranssphenoidal surgery for non-functioning pituitary macroadenoma (diameter > 1 cm) for decompression of the optic chiasmAssessment of visual acuity at least once before surgery, and at least twice in the first year after surgeryThe availability of two postsurgical MR scans, within a time frame of maximal 3 months of the ophtalmological assessments We did not restrict the inclusion criteria to patients with visual field defects, but also included patients with clear compression of the optic chiasm without decreased visual function, for two reasons. First, we were interested in the pattern of change in visual function prior to surgery. Second, it is, at least theoretically, possible that in patients with compression of the optic chiasm, surgical treatment may decrease visual function. To overcome the potential effect of postoperative tumor-regrowth on visual outcome, patients were excluded from analysis if the 1 year post-operative scans revealed growth of residual tumor. For complete assessment, we reviewed the patient records of all departments involved in the treatment of NFMA (Endocrinology, Neurosurgery, Ophthalmology). Endocrine (pituitary function) and ophthalmologic data (visual acuity and visual fields) were assessed before surgery, 3 and 12 months after surgery. An MRI was performed before surgery, 3–6 months and 12–15 months after surgery. Transsphenoidal surgery was performed by one of two neurosurgeons Corrected visual acuity was determined by the Snellen chart [15, 16] and was scored for both eyes. Visual fields were assessed by Humphrey perimetry in all patients. Goldman perimetry was used as an additional tool to assess peripheral visual field defects. To increase the reliability of the visual field testing, assessment of the visual fields, and thereby also visual acuity, was preferably performed twice prior to surgical treatment. However, surgical treatment was not delayed due to double ophthalmologic testing in patients with severe impaired visual acuity. Definitions The diagnosis of non-functioning pituitary macroadenoma was based on two criteria: the presence of a pituitary macroadenoma (>1 cm) on MRI, and the absence of overproduction of any of the pituitary hormones. In all cases the diagnosis was histologically confirmed. Tumor size was classified according to Hardy [17]. Tumor extension was classified as suprasellar (Hardy II0, A, B, C) or parasellar/infrasellar (Hardy IID,E, Hardy III0-E and Hardy IV0-E). In all patients, visual acuity was scored on a scale between 0 and 1.25. Visual field defects were scored semi-quantitative and classified as mild, moderate or severe. Visual field defects were classified as mild, if there were peripheral defects in only one quadrant. Defects were classified as moderate if the upper quadrants were affected, whereas in combined upper and lower quadrant field defects, these were classified as severe. Growth hormone (GH) deficiency was defined as an insufficient rise in GH levels (absolute value < 3 μg/l) after stimulation during an insulin tolerance test (ITT). When secondary amenorrhoea was present for more than 1 year premenopausal women were defined as LH/FSH deficient. Postmenopausal women were defined as LH/FSH deficient, when gonadotropin levels were below the normal post-menopausal range (LH < 10 U/l, FSH < 30 U/l). In men, LH/FSH deficiency was defined, as a testosterone level below the reference range (8.0 nmol/l). TSH deficiency was defined as a total or free T4 level below the reference range. ACTH deficiency was defined as a basal cortisol level at 8.00 AM of <0.12 μmol/l and/or an insufficient increase in cortisol levels (absolute value < 0.55 μmol/l) after an ITT. Statistical analysis The paired t-test was used for paired samples. SPSS software version 12.0 (SPSS Inc., Chicago, IL, USA) was used. A P-value of <0.05 was considered statistically significant. Results Preoperative patient characteristics (Table 1) Forty-three patients, 51% male, were included in this study. Radiological imaging by MRI revealed a macroadenoma in all patients, with suprasellar extension in 100% and parasellar/infrasellar extension in 42% of cases. Hypopituitarism was present in 79% of all patients, and panhypopituitarism in 23%. Table 1Patient characteristics before transsphenoidal surgeryMale/female22/21Age (years ± SD)56 ± 14Visual acuity Right eye (mean ± SD)0.65 ± 0.37Left eye (mean ± SD)0.60 ± 0.32Visual field defectsSevere60%Moderate17%Mild14%None9%Pituitary functionGH deficiency76%LH/FSH deficiency69%ACTH deficiency43%TSH deficiency36%Panhypopituitarism23%MRI characteristicsSuprasellar extension100%Infrasellar/parasellar extension42% Pre-operative assessment of visual function (Table 1) Visual fields were normal in 4 patients. Transsphenoidal surgery in these patients was performed because compression of the optic chiasm was evident on the MR scan (n = 4). In these 4 patients visual acuity was at least 1.0 for both eyes. Of all patients with visual field defects 60% were classified as severe, 17% as moderate, and 14% as mild. Mean visual acuity was 0.65 ± 0.37 for the right eye, 0.60 ± 0.32 for the left eye. Visual acuity was assessed twice before surgery in 34 patients (79% of total). The median time between the first and the second pre-surgical assessment was 4 weeks (range 1–45). No significant decrease in visual acuity was observed within this time period. The mean visual acuity for the two pre-surgical assessments was: 0.65 ± 0.37 vs. 0.66 ± 0.38 (right eye), and 0.60 ± 0.32 vs. 0.62 ± 0.33 (left eye). Moreover, no decrease in visual acuity was observed in patients (n = 13) with a time interval of >8 weeks between the first and the second pre-surgical assessment. In 4 of these patients there were no visual field defects. Surgical treatment All patients were treated by transsphenoidal surgery. Repeat surgery within 6 months after initial treatment was performed twice (large residual tumor mass (n = 1) and persisting liquor leakage (n = 1)). Repeat surgery was performed by transcranial approach. Transsphenoidal surgery was followed by radiotherapy in 10 patients in order to prevent recurrence. Patients received 40 Gy (n = 5) or 46 Gy (n = 5). Three months postoperative assessment of visual function After transsphenoidal operation in patients with preoperative visual field defects improvement was observed in 60% and normalization of the visual fields in 30%. In 1 patient there was a slight increase in visual field defects. Visual acuity improved significantly 3 months after transsphenoidal surgery (Fig. 1). The mean visual acuity increased from 0.65 ± 0.37 to 0.75 ± 0.36 (P < 0.01) (right eye), and from 0.60 ± 0.32 to 0.82 ± 0.30 (P < 0.01) (left eye). Fig. 1The pattern of improvement of visual acuity after transsphenoidal surgery for non-functioning pituitary macroadenomas (n = 43) One-year postoperative assessment of visual function One year after initial surgical therapy visual field defects showed continuous improvement in 36% of patients, compared to early postoperative results. In 80% of them, this improvement was accompanied by continuous improvement of visual acuity. In 56% of all patients (n = 24) visual acuity shows continuing improvement until 1 year after surgery. In 8 of them, this improvement was not accompanied by further improvement of visual field defects. Visual acuity was improved 1 year after transsphenoidal surgery compared to the 3 months postoperative values (Fig. 1). The mean visual acuity increased from 0.75 ± 0.36 to 0.82 ± 0.34 (P < 0.05) (right eye), and from 0.82 ± 0.30 to 0.88 ± 0.27 (P < 0.05) (left eye). Discussion The main aims of surgery in non-functioning pituitary macroadenoma are restoration of visual acuity and visual field defects by decompression of the optic chiasm. Nonetheless, our data indicate that there is no necessity for immediate decompression, since postponement of surgery for several weeks did not result in deterioration of visual acuity. Moreover, in this series of 43 patients, we demonstrated a continuing improvement of visual acuity until 1 year after transsphenoidal surgery. There is a significant correlation between the severity of visual loss prior to surgery and persisting visual field defects [9, 13, 14]. In our patients in whom visual acuity was assessed twice before surgery, no decrease of visual acuity was observed with a median time interval between two measurements of 4 weeks. This indicates that postponing surgery for 1 month does not negatively influence visual outcome, which is in line with the slow growth pattern of non-functioning adenomas [4]. Nonetheless, the delay of surgery should not unnecessarily be prolonged because a significant, inverse, correlation between visual outcome and the prolonged duration of symptoms has been reported [9]. Moreover, especially in patients with rapid loss of visual function or decreased visual acuity due to apoplexy, urgent surgical intervention is indicated. The initial event in the pathogenesis of decreased visual function in pituitary macroadenomas is compression of the optic chiasm. Nerve compression leads to decreased conduction and to demyelination. In an experimental setting, the process of demyelination after nerve compression has been observed even after 2 days [18, 19]. In case of continuous nerve compression remyelination can be observed after several weeks, although remyelinated fibers do not seem to reach normal thickness and organization structure, and complete demyelinated fibers co-exist [18, 19]. Re-myelinated fibers restore conduction, at least partially, even if the causative nerve lesion is still existing [20, 21]. The improvement of visual dysfunction after surgical treatment is supposed to consist of two, or probably even three, phases [12]. There is an early phase, comprising the first hours and days after surgery. In this early fast phase, the improvement is caused by decompression of the visual pathways, leading to a restoration of signal conduction. Visual recovery has been demonstrated in the first days after surgical treatment [11, 12]. The second phase, i.e., delayed recovery, is pathophysiologically caused by restoration of axonal transport and remyelination, and based on remyelination of the optic nerve vessels. This phase of delayed recovery may last for several years [12, 13]. A precise boundary between the end of the fast phase of recovery and the start of the delayed recovery seems to be artificial, because these two phases reflect different pathophysiological mechanisms, which may co-exist for a certain time-period. The contribution of the first phase of recovery might be larger, given the fact that more than 50% of eventual recovery takes place within the first 3 months after surgery [13]. It is already known that visual acuity improves in the first months after surgical treatment [6–10] and that the improvement of visual field defects is a continuing process for at least 1 year [12, 13]. Kerrison et al. [12] showed progressive improvement of visual fields even more than 2 years after surgical decompression of the optic chiasm. However, they did not demonstrate this same pattern of recovery for visual acuity. This might be due to the relative small number of patients during prolonged follow-up. In the present study we demonstrate that also improvement of visual acuity continued 1 year after surgical treatment. The clinical consequences of the delayed phase of recovery for both visual field defects and visual acuity are obvious. Follow-up of patients after surgical treatment for pituitary macroadenomas should include ophthalmologic assessment within several weeks after surgery, as well as subsequent assessments after one and 2 years, in order to estimate the final effect of surgery on visual function. Moreover, patients should be told that visual function can continue to improve, at least until 1 year after surgery. The relevance of these findings are obvious and is of importance to all patients, given the impact of a decreased visual acuity as an independent predictor for a decreased quality of life [22, 23]. Moreover, these data are essential in order to evaluate potential effects of recurrent pituitary adenomas on visual function. Ten patients in our series received postoperative radiotherapy. However, it is unlikely that this treatment affected the results of our study. In a series reported by Gnanalingham et al. [13], in which 34% of all patients received postoperative radiotherapy, persistent improvement of visual field defects was documented even years after surgical therapy. In another series of 21 patients, 2 years after pituitary irradiation, there were no cases of radiation-induced visual field or visual acuity deterioration [24]. In conclusion, this study demonstrates that the improvement of visual acuity, after transsphenoidal treatment for non-functioning pituitary macroadenomas, consists of both an early and a delayed phase of recovery. After initial post-surgical recovery, a progressive delayed improvement of visual acuity at least until 1 year after transsphenoidal surgery, is likely to occur.	[ "visual acuity", "surgery", "non-functioning macroadenoma" ]	[ "P", "P", "R" ]
Evid_Based_Complement_Alternat_Med-4-2-1876616	Effectiveness of Massage Therapy for Chronic, Non-malignant Pain: A Review	Previous reviews of massage therapy for chronic, non-malignant pain have focused on discrete pain conditions. This article aims to provide a broad overview of the literature on the effectiveness of massage for a variety of chronic, non-malignant pain complaints to identify gaps in the research and to inform future clinical trials. Computerized databases were searched for relevant studies including prior reviews and primary trials of massage therapy for chronic, non-malignant pain. Existing research provides fairly robust support for the analgesic effects of massage for non-specific low back pain, but only moderate support for such effects on shoulder pain and headache pain. There is only modest, preliminary support for massage in the treatment of fibromyalgia, mixed chronic pain conditions, neck pain and carpal tunnel syndrome. Thus, research to date provides varying levels of evidence for the benefits of massage therapy for different chronic pain conditions. Future studies should employ rigorous study designs and include follow-up assessments for additional quantification of the longer-term effects of massage on chronic pain. Introduction Massage therapy has been defined as soft-tissue manipulation by trained therapists for therapeutic purposes (1). Massage therapy has a long history, being first described in China during the second century B. C. and soon thereafter in India and Egypt (1). More recently, massage therapy has been administered using mechanical devices in addition to hands-on treatment by therapists. Massage can be applied to single or multiple body parts or to the entire body. There are many different types of massage therapy including Swedish massage, Shiatsu, Rolfing, reflexology and craniosacral therapy. Most of the published trials on massage therapy have utilized Swedish or Swedish-type massage. Despite the growing popularity of massage, there is inconsistent empirical support for its effectiveness in chronic pain. Although the effects of massage therapy on chronic pain has been the subject of prior reviews (described subsequently), most of these reviews have focused on a single chronic pain condition and the level of supportive evidence appears to vary greatly, depending on the particular pain condition investigated. This review aims to bring together the existing data on the effectiveness of massage therapy for a wide range of chronic, non-malignant pain conditions. (Evidence regarding the effects of massage on acute pain is reviewed under ‘Meta-Analysis—Massage Therapy Effects on Pain. Chronic pain has been defined by the International Association for the Study of Pain (IASP) as continuous or recurrent pain that persists for longer than the normal time of healing, generally about 3 months (2). It is hoped that by providing an overview of the field, gaps in extant research may be identified in order to inform future clinical trials. Typically, massage is viewed as adjunctive therapy to help prepare the patient for exercise or other interventions, and is rarely administered as the main treatment (3). Following the methodology of the Cochrane Collaboration, the focus of this review will be on those studies in which massage for pain relief is delivered alone rather than as part of a treatment package, since it is difficult to draw conclusions regarding the effectiveness of massage when multiple treatments are involved. In addition, the emphasis of this review will be on randomized, controlled trials (RCTs) or quasi-RCTs of massage therapy. As noted above, the main purpose of this article is to provide a broad overview of extant literature on the application of massage to a wide range of chronic pain conditions. Prior reviews have typically focused on a single pain complaint, even though many chronic pain patients present with multiple pain conditions. Thus, the current review summarizes the findings of existing reviews and meta-analyses as well as key individual studies that have appeared since the publication of these comprehensive reviews. Although this approach is limited as it depends heavily on the methodology used in extant reviews and the quality of the methodology likely varied across reviews, it was considered the most feasible approach in order to synthesize the large number of studies examining massage for a broad array of chronic pain conditions. This review is organized as follows. First, empirical findings of effectiveness are presented according to the type of chronic pain condition examined; each of these sections concludes with a summary statement of the level of evidence for the specific pain condition studied. (A summary table of the overall findings is also presented in Table 6.) It should be noted that a review of the findings from the handful of studies on mixed chronic pain problems is also included. Second, the results of a meta-analysis of massage therapy for pain complaints are discussed. This study was unusual in that it examined massage therapy effects across a number of pain complaints. The review concludes with a summary of the findings across the various chronic pain conditions, together with a discussion of putative mechanisms, clinical implications and recommendations for future trials. Methods The PubMed, PsychInfo, CINAHL, and Cochrane Library databases searched up to July 2006 using the keywords ‘massage’, ‘pain’, ‘analgesia’ and ‘analgesics’. As noted above, the focus of this article is to provide a comprehensive overview of the evidence regarding massage therapy for chronic, non-malignant pain. Thus, due to the large number of trials uncovered using the present search strategy, reviews by the Cochrane Collaboration and other authors were used where applicable to identify relevant trials. Primary studies that were excluded by the Cochrane group or others due to methodological or other limitations were therefore not included in the present review. However, findings from relevant primary studies that had been published since these reviews appeared were included. Only those chronic, non-malignant pain conditions that had been examined by at least one controlled trial were included in this review. The studies included in this review focused on adult participants; the application of massage therapy for chronic pain in children has been discussed in a prior review (4,5). The study findings are summarized in Tables 1–6. Mean reductions in the main pain outcome measure are shown in the Tables. Since the most commonly used pain outcome measure was a visual analog scale or VAS rating of pain intensity, means for this outcome are reported whenever possible. Table 1.Summary of results for studies included in the Cochrane Review on massage therapy (MT) for low back painStudynPain durationControl conditionsOutcomesFindings (Mean reduction in pain)Hsieh et al. (1992) (6)63Not statedSpinal manipulation (SM); Corset (CT);TranscutaenousOswestry Low Back Pain Q’aireSM (20) > MT (9.2) muscular stimulation (TMS)Roland-Morris Activity Scale(SM = CT) > MT; SM > TMSCherkin et al. (2001) (11)26261%=Acupuncture (AC); Self-care education (SC)Symptom scale (0–10)MT (2.6) > SC (1.5)>1 year1 year F/UMT (3.0) = SC (2.3); MT > AC (1.7)Roland Disability ScaleMT > SC; MT > AC1 year F/UMT = SC; MT > ACUse of Medications (1 year F/U)MT > SC; MT > ACSF-12 (Physical Health)MT > SC1 year F/UMT = SC = ACSF-12 (Mental Health)MT > SC; MT > AC1 year F/UMT = SC = ACHernandez-Reif et al.24Not stated but >6 mProgressive muscle relaxation (PMR)VAS pain intensity (0–10)MT (3.9) > PMR (1.6) (2001) (10)McGill Pain QuestionnaireMT = PMRRange of Motion (ROM)MT > PMRHoehler et al. (1981) (13)9548–52% < 1 mSMPatient-rated pain (unspecified scale)MT = SM17–29% > 6 mStraight-leg raiseMT = SMPope et al. (1994) (7)*16429% < 6 mSM; TMS; CTVAS pain intensity (0–10)MT (17.2) = SM (24.1) = TMS (9.6) = CT (15.9)35% = 6–12 mROM (flexion/extension)MT = SM = TMS = CT36% > 2 yearsMaximum voluntary extension effortMT = SM = TMS = CTSorensen fatigue testMT = SM = TMS = CTGodfrey et al. (1984) (12)81Not stated but <14 daysSM; Low-level electrical stimulation (LES)Pain, stiffness, tenderness (0–4)MT = SM = LESDaily activities (0–4)MT = SM = LESSelf-reported limitations due to painMT = SM = LESFingertip flexion testMT = SM = LESMelzack et al. (1983) (14)41Mean = 36.2 wksTranscutaneous electrical nerve stimulation (TENS)McGill Pain Questionnaire (PRI)TENS (85%) > MT (38%) (>50% reduction on PRI)Straight-leg raiseTENS > MTBack FlexionTENS = MTPreyde et al. (2000) (9)98Mean range = 12.0–14.8 wksSoft-tissue manipulation (STM); Exercise (EX);Roland Disability QuestionnaireCMT (5.9) > EX (0.3), SL (0.4); STM (5.2) > EX, SLSham laser therapy (SL)1 month F/UCMT (6.8) > EX (1.5), SL (0.7); (STM (5.7) = EX) > SLNote: Above were compared with Comprehensive Massage Therapy (CMT) which included STM and EXLumbar ROMCMT = STM = EX = SLNote: *denotes high quality study per Cochrane review (3); denotes low quality study per Cochrane review (3). The Hsieh and Pope studies reported findings from the same trial. Table 2.Summary of findings for studies on massage therapy (MT) for headache painStudy and type of headachenPain durationComparison conditionsOutcomesFindings (Mean reduction in pain)Nilsson et al. (1997) (17)Cervicogenic Headaches54Not statedSpinal manipulation (SM) but ≥3 mNOTE: The above was compared with MT plus placeboVAS pain intensity (0–10)SMT (17.0) > MT (4.2)LaserHeadache durationSMT > MTWylie et al. (1997) (18)Migraine and67Mean = 10.2AcupuncturePain total index (hours × severity) Tension-type yearsNOTE: The above was comparedMigraineMT (171.7) > AC (128.0) with MT plus relaxationTension-typeMT (217.0) = AC (119.5)Both groups combinedMT = ACHeadache index (number × severity)MigraineMT > ACTension-typeMT = ACBoth groups combinedMT = ACMigraine DaysMT = ACHanten et al. (1999) (20)65Not statedResting position (RP); No treatment (CON)Tension-typeNOTE: The above was compared with CV-4 techniqueVAS pain intensity (0–10)MT (19.3) > [RP (11.2) = CON (7.8)]VAS pain affect (0–10)MT > (RP = CON)Note: this study was included in the Cochrane review of non-invasive physical treatments for chronic/recurrent headache (15). Table 3.Summary of findings for studies on massage therapy (MT) for neck and shoulder pain and carpal tunnel syndromeStudy and type of painnPain durationComparison conditionsOutcomesFindings (Mean reduction in pain)Irnich et al. (2001) (28)VAS pain intensity (0–100)AC (17.3) > MT (3.1); AC = SLA (11.4)Neck pain17721.5–22.6% > 5 yearsAcupuncture (AC);3 month F/UAC (15.0) = MT (8.1); AC = SLA (11.2)Sham laserRange of Motion (ROM)AC > MT; AC = SLA acupuncture (SLA)3 month F/UAC = MT; AC = SLAPressure pain threshold (algometer)AC = MT; AC = SLA3 month F/UAC = MT; AC = SLASpontaneous pain (7 point scale)AC = MT; AC = SLA3 month F/UAC > MT; AC = SLAMotion-related pain (7-point scale)AC > MT; AC > SLA3 month F/UAC > MT; AC = SLAGlobal complaints (7-point scale)AC > MT; AC > SLA3 month F/UAC > MT; AC = SLAHealth-related quality of life (SF-36)AC = MT; AC = SLA3 month F/UAC = MT; AC = SLADyson-Hudson et al. (2001) (24)Shoulder pain18Mean range = 13.4–16.2 yearsACWheelchair user's shoulder pain indexAC (23.3) = MT (21.7)ROMAC = MTMok and Woo (2004) (25)Shoulder pain102Not statedNo treatment control (CON)VAS pain intensity (unspecified scale)MT (14.6) > CON (0.2)State anxiety (STAI)MT > CONHeart rateMT > CONSystolic blood pressureMT > CONDisastolic blood pressureMT > CONvan den Dolder (2003) (27)Shoulder pain29Mean range = 26–30 weeksCONMcGill Pain QuestionnaireVAS pain intensity (0–10)MT (26.6) > CON (0.1)Present Pain Intensity (PPI) scaleMT = CONVerbal descriptorsMT > CONFunctional DisabilityMT > CONROMMT > CONField (2004) (32)Carpal Tunnel16Mean = 6.7 yearsCONVAS pain intensity (0–10)MT (3.2) > CON (.08) SyndromeGrip StrengthMT > CONState anxiety (STAI)MT > CONProfile of Mood States — DepressionMT > CONPhysician assessed CTS symptomsMT > CON Table 4.Summary of findings for studies on massage therapy (MT) for fibromyalgiaStudynPain durationComparison conditionsOutcomesFindings (Mean reduction in pain)Sunshine et al.30Not statedTranscutaneous electricalImmediate post-session (1996) (33) stimulation (TENS); Sham TENS (STENS)State anxiety (STAI)MT = TENS; TENS, MT > STENSProfile of Mood States (POMS) — DepressionMT = TENS; TENS, MT > STENSSalivary cortisolMT = TENS; TENS, MT > STENSPre- Post TreatmentSelf-Report Interview — Pain (unspecified scale)MT (3.3) > TENS (0.2); STENS (1.6)StiffnessMT > TENS; STENSFatigueMT > TENS; STENSSleepMT > TENS; STENSDolorimeter testMT > TENS; STENSPhysician assessment of conditionMT = TENS; TENS, MT > STENSDepression (CES - D)MT = TENS = STENSField et al.20Mean years inProgressive muscleImmediate post-session(2002) (34) treatment = 9.2 relaxation (PMR)STAIMT = PMRPOMS – DepressionMT = PMRPre- Post TreatmentVAS pain intensity (0–10)MT (2.3) > PMR (1.4)VAS fatigue (0–10)MT > PMRVAS stiffness (0–10)MT > PMRPhysician assessment of conditionMT > PMRNumber of tender pointsMT > PMRCES - DMT > PMRBratttberg (1999) (35)4886% > 5 years;No treatment control (CON)VAS pain intensity (0–100)MT (26.3) > CON (1.4)50% > 10 yearsDisability rating indexMT = CONSleep disturbance (0–5)MT = CONHospital anxiety and depression scaleAnxietyMT = CONDepressionMT > CONAlnigenis et al. (2001) (36)37Mean range = 5–10.4 yearsUsual care (UC);Arthritis impact measurement scaleUC plus phone calls (UCP)Pain (0–10)MT (0.9) = UC (1.0) = UCP (0.3)Rheumatology attitudes indexMT = UC = UCPCES-DMT = UC = UCPQuality of well being scaleMT = UC = UCP Table 5.Summary of findings for studies on massage therapy (MT) for mixed chronic painStudynPain durationComparison conditionsOutcomesFindings (Mean reduction in pain)Walach et al. (2003) (37)29Not stated but > 6 monthsUsual care (UC)Pain intensity (1–9 point scale)MT (1.0) > UC (0.1)Profile of Mood States (tiredness)MT > UCCES-DMT > UCState anxiety (STAI)MT > UCFrankfurt body concept scalesMT = UCHasson et al.129Not statedProgressive MusclePain (unspecified scale)(2004) (38)but > 3 monthsRelaxation (PMR)Post-treatmentMT (14.5) > PMR (2.1)3 month F/UMT (.03) = PMR (1.3)Mental energy (unspecified scale)Post-treatmentMT > PMR3 month F/UMT = PMRSelf-perceived health status (5-point scale)Post-treatmentMT > PMR3 month F/UMT = PMRPlews-Ogan et al. (2005) (39)30Not stated but > 3 monthsMindfulness-based stress reductionVAS pain unpleasantness ratings (0–10)MT (2.9) > UC (.13); MBSR (.7) = UC(MBSR); UC1 month F/UMT = MBSR = UCSF-12 physical healthMT = MBSR = UC1 month F/UMT = MBSR = UCSF-12 mental healthMT = MBSR = UC1 month F/UMBSR > UC; MBSR = MT Table 6.Summary of overall findingsType of painNumber of studiesResults (number of studies)Quality ranking (global assessment of quality)Low back pain8MT > relaxation (1), education (1), acupuncture (1)1 (good)MT = corsets (2), exercise (1)MT < spinal manipulation (4), TENS (1)Shoulder pain3MT > acupuncture (1), no treatment (2)2 (moderate)Headache pain3MT > rest for migraine (1) and tension-type (1)3 (moderate)MT < spinal manipulation for cervicogenic (1)Fibromyalgia4MT > TENS (1), relaxation (1)4 (preliminary)MT = no treatment (1), usual care (1)Mixed chronic pain3MT > usual care (1)5 (preliminary)MT = relaxation (1), meditation and usual care (1)Neck pain1MT = acupuncture (1)6 (preliminary)Carpal tunnel1MT > no treatment (1)7 (preliminary)Note: Quality Ranking refers to the overall ranking of the quality of the literature on the effectiveness of massage therapy for each chronic pain condition relative to the other chronic pain conditions discussed in this review. Results Empirical Findings Massage Therapy for Low Back Pain The Cochrane Collaboration recently published a report on the use of massage therapy for non-specific low back pain (LBP) (3). Their comprehensive review included studies published until May 2001 and was substantively amended at the end of January 2002. The review included randomized or quasi-randomized trials testing the use of any type of massage (using hands or mechanical device) as an intervention for LBP. The Cochrane review identified nine publications which reported the results of eight randomized trials (Table 1 for detailed information). Note that the studies by Hsieh et al. (6) and Pope et al. (7) listed in Table 1 reported results from the same trial. One study was in German (8) (results not shown in Table 1) and the remainder were in English. Numerous studies were excluded from the review; many studies were excluded because massage was tested within a treatment package combined with various other therapies. Standardized criteria were applied to the included studies to assess methodological quality. For the eight trials, five were judged to be of high methodological quality (denoted as ** in Table 1) and three were deemed to be of low quality (denoted * in Table 1). In one study (9), massage was compared with a placebo (sham laser). Massage was found to be superior to the placebo treatment. In the other seven trials, massage was compared with various active treatments. These studies showed that massage was superior to relaxation (10), acupuncture (11) and education (11); massage was equal to corsets (6,7) and exercises (9); massage was inferior to spinal manipulation (6,7,12,13) and transcutaneous electrical nerve stimulation (TENS) (14). The single German study showed that acupressure/pressure point massage techniques provided more pain relief than classical (Swedish) massage (8). Beneficial Effects for Subacute and Chronic Non-specific Low Back Pain The Cochrane review concluded that massage therapy may be beneficial for patients with subacute and chronic non-specific LBP, particularly when combined with exercises and education. They also noted that the results of one high quality study showed that the benefits of massage last as long 1 year following the end of active treatment (11). The benefit obtained from massage exceeded that achieved from relaxation, education or acupuncture. However, the beneficial effects may be less than that provided by spinal manipulation or TENS. The Cochrane Review noted that there is insufficient evidence regarding the effects of massage on acute back pain and on specific forms of massage for chronic LBP. Although the Cochrane review represents a synthesis of the most rigorous trials to date examining massage for LBP, it should be noted that the review based their conclusions on a relatively small number of studies. For example, their conclusions regarding the superiority of massage to relaxation and acupuncture were based on only a single study each, and therefore await further confirmation in future trials. On the other hand, the studies included in the review demonstrated therapeutic effects for massage that exceeded or equaled those obtained from various active treatment conditions. This level of evidence is more encouraging than that obtained from trials showing that massage is superior to no treatment or waitlist control. Pain Management for Headaches from Massage Therapy The Cochrane Collaboration also published a recent review of non-invasive physical treatments for chronic/recurrent headache (15). This review included studies published until November 2002 and was substantively updated in May 2004. The review studied five types of headaches (i.e. migraine, tension-type, cervicogenic, mixed tension-type and migraine and post-traumatic headache), and a broad range of treatments were examined. Of the 22 studies that met inclusion criteria, only one trial, reported in two studies (16,17) examined the impact of massage on headaches. In this trial, massage plus placebo laser was compared with spinal manipulation for cervicogenic headache. The findings of this study indicated that spinal manipulation was superior to massage for headache pain intensity, headache duration and medication use (Table 2). The Cochrane review concluded that there is moderate evidence that massage plus placebo laser is inferior to spinal manipulation for pain intensity and duration in relation to cervicogenic headache. The Cochrane review did not include the results of Wylie and colleagues (18) compared massage and relaxation to acupuncture for headache pain in patients with migraine or combined headache and patients with tension-type headaches. Patients received six sessions lasting 45 min. For the massage and relaxation condition, it is unclear how much of each session was devoted to relaxation which included muscle and breathing exercises as well as visualization techniques. All patients exhibited significant decreases in pain total index (PTI; monthly number of headache hours multiplied by severity) and headache index (HI; monthly number of attacks multiplied by severity). For patients with tension-headaches (n = 40), there was no difference in pain outcomes based on intervention type. However, for patients with migraines (n = 27), those who received massage had significantly lower PTI and HI scores than those who received acupuncture (Table 2). The number of migraine days did not vary across treatments. These findings suggest that massage may be superior to acupuncture for migraine headaches. But, because massage was combined with relaxation and other self-help techniques, it is not possible to draw definitive conclusions regarding the specific effects of massage based solely on these findings. Craniosacral therapy is based on the notion that movement restrictions in the cranial structures of the skull adversely impact rhythmic impulses conveyed through the cerebral spinal fluid from the cranium to the sacrum (19). Thus, craniosacral therapy is a form of massage that uses gentle pressure on the plates of the patient's skull. Few controlled studies have been conducted on craniosacral therapy and a recent review concluded that there is insufficient evidence to support the effectiveness of this approach (19). One study, not included in the review, examined the CV-4 craniosacral technique on tension-type headaches (20). The CV-4 technique moves with the narrowing and widening of the skull, or the cranial rhythm; the basis of the technique is the compression of the fourth ventricle. Sixty patients were randomly allocated to one of three conditions: (i) a 10 min session during which multiple still points were induced by the CV-4 technique; (ii) a 10 min session during which the head and neck were positioned according to the resting position technique; (iii) a 10 min no treatment control. Immediately following treatment, the CV-4 group reported less pain intensity and pain affect than the control group; there were no differences between the resting position group and controls. Moderate Evidence for Cranioscaral Massage in Managing Tension-Type Headache Pain There have been surprisingly few published RCT's on massage therapy for headache pain. The single study included in the Cochrane review of non-invasive treatments for recurrent/chronic headache found that spinal manipulation resulted in greater pain reduction than massage plus sham laser for cervicogenic headache. One other study reported that massage may be more beneficial than acupuncture for migraine headaches but that both approaches were equally effective for tension-type headaches. A single study found that craniosacral therapy led to superior pain reduction compared with rest positioning and no intervention for tension-type headaches. This latter study using CV-4 technique was included in a recent review of manual therapies for tension-type headaches (21). The review authors gave this study a score of 6 out of 10 possible points for methodological quality, suggesting that there is moderate evidence from this trial that the CV-4 technique exerts a beneficial effect on pain related to tension-type headaches. As noted in the review, given that headaches are among the most common problems seen in medical practice (22), there is an urgent need to establish the effectiveness of manual therapies, including massage in the treatment of headache pain. The review also noted that not only is tenderness of the pericranial myofascial tissues one of the prominent features of tension-type headache, myofascial tissues may play an important role in the genesis of such headaches (23). Because the aim of soft tissue manipulation is to alter mechanical stress caused by myofascial tissue disorders, the review concluded that massage techniques may therefore be an effective therapy for tension-type headaches. However, further large-scale studies are needed before conclusions regarding the effectiveness of massage for tension-type or other types of headaches can be drawn. Shoulder Pain — Moderate Support for the Use of Massage Therapy Three analyses examined the effects of massage for shoulder pain. The first study compared acupuncture to Trager Psychophysiological Integration (a form of massage) in 18 patients with chronic shoulder pain who used manual wheelchairs as their primary means of mobility (24). Each patient received 10 treatments over 5 weeks; the acupuncture sessions lasted approximately 20–30 min and the Trager sessions lasted approximately 45 min. By 5-week follow-up, both groups exhibited improvements in pain and range of motion (see Table 3); there were no significant differences between groups. The study authors concluded that both acupuncture and Trager were effective for shoulder pain in wheelchair users. In their meta-analysis of massage therapy effects (discussed subsequently) this study was included as evidence supporting the benefits of massage on the delayed assessment of pain (i.e. pain that is assessed following a period after which no treatment is delivered). Mok and Woo (25) analyzed hospitalized stroke patients with shoulder pain who were randomly assigned to receive slow-stroke back massage (SSBM) (26) or no intervention control. SSBM was administered for 10 min before bedtime for seven consecutive days. Patients who received SSBM experienced decreases in pain, anxiety, heart rate and blood pressure, compared with no such changes in controls (see Table 3). Another study (27) compared patients with shoulder pain who received six 15–20 min sessions of massage over two weeks (n = 15) to a waitlist control (n = 14). Patients who received massage improved significantly in functional disability, pain, and range of motion, whereas the control group evidenced no changes on these outcomes (Table 3). These results provide moderate support for the use of massage for shoulder pain. However, the patient characteristics varied greatly across studies (e.g. stroke patients; wheelchair users) suggesting that generalizability of the findings may be limited. Moreover, with the exception of Mok and Woo (25) these studies had small sample sizes and only the study by Dyson-Hudson et al. (24) included a follow-up assessment. Finally, two of these studies compared massage with no treatment control. Additional studies are needed to determine whether massage therapy leads to improvements in shoulder pain when controlling for non-specific effects of treatment (e.g. increased clinician time/attention). Replication of these findings of these three studies in larger samples and with longer follow-up evaluation periods is warranted. Comparison with Acupuncture for Neck Pain Another group (28) examined patients with chronic neck pain randomly assigned to one of the following conditions: acupuncture (n = 56), massage (n = 60), ‘sham’ laser acupuncture (n = 61). Patients received five 30-min treatments over 3 weeks. For the main outcome, pain VAS (0–100) ratings, acupuncture was superior to massage, but no different than sham laser immediately after treatment. However, by 3-month follow-up, there were no differences between acupuncture and massage or between acupuncture and sham laser (Table 3). Among the secondary measures, acupuncture showed more improvement than massage across most measures assessed immediately post-treatment. Nevertheless by 3-month follow-up, acupuncture was superior to massage only on spontaneous pain, global complaints and motion-related pain. Effectiveness of Massage Therapy for Neck Pain Remains Unclear The authors of the previous work concluded that acupuncture appears to be effective in the short-term for chronic neck pain, it should be noted that by the 3-month follow-up, acupuncture was no more effective than massage or sham laser acupuncture on most outcome measures. Unfortunately, the study did not include direct comparisons between the massage and sham laser conditions. Therefore, it is not possible to determine whether massage was superior to a placebo condition. An unpublished master's thesis (29) has been cited in a recent meta-analysis as providing support for the longer-term effects of massage on pain (30). These studies were included in a very recent Cochrane Review on massage for mechanical neck disorders (31), which also included a broad array of interventions such as traditional Chinese massage, ischemic compression, self-administered ischemic pressure using a J-knob cane, and occipital release, among others. The review authors noted that many of these approaches were of questionable value. The review concluded that no practice recommendations could be made since the effectiveness of massage for neck pain remains unclear. Preliminary Support for Pain Relief in Carpal Tunnel Syndrome Patients with carpal tunnel syndrome (CTS) were randomly assigned to either a 4-week course of massage therapy or usual care (32). The massage group received a 15-min massage once a week from a massage therapist and were also taught self-massage to be done daily at home prior to bedtime. The massage group evidenced improvements in pain, grip strength, anxiety and depression compared with no such improvements in the control group (Table 3). Physician assessments of carpal tunnel symptoms also indicated significant improvements in the massage group versus no change in the control group. The findings of this single study provide preliminary support for the application of massage to CTS. However, the sample size was very small, and it is unclear whether the physicians assessing the patients were aware of group assignment. Moreover, it is unclear to what extent patients practiced self-massage at home and whether the amount of massage administered was related to treatment response. Further work with larger samples and more rigorous study methodology are needed to determine the effectiveness of massage therapy for CTS. Equivocal Support for Analgesic Effects in Fibromyalgia The effects of massage therapy on fibromyalgia have been examined in four investigations. Two of these studies were conducted by Tiffany Field and colleagues. In the first study (33), women with fibromyalgia were randomly assigned to receive massage, TENS or sham TENS for 30 min twice a week for 5 weeks. Immediately following treatment on the first and last days, the massage therapy group evidenced reductions in anxiety, depression and salivary cortisol. The TENS group experienced the same improvements but only on the last day. The sham TENS groups demonstrated no such improvements. By the end of treatment, the massage therapy group reported less pain, stiffness, fatigue and difficulty sleeping (as assessed via interviews), as well as improvements in dolorimeter test value and physician's assessment of clinical condition (Table 4). The TENS group improved on this latter measure only. Limitations of this study include inadequate information regarding the interview items. Moreover, it was unclear whether interviewers and physicians were blind to patient group assignment. The Field group (34) also randomly assigned fibromyalgia patients to receive either massage therapy or progressive muscle relaxation (PMR) for 30 min twice a week for 5 weeks. Both groups reported reductions in anxiety and depression immediately following treatment on the first and last days. By the end of treatment, the massage group evidenced significant reductions in self- and physician-assessed pain and symptoms, as well as reductions in the number of tender points and substance P levels (Table 4). No significant improvements were seen in the PMR group. Brattberg (35) randomly allocated patients with fibromyalgia to massage therapy or no treatment control. The massage group received 15 massages of indeterminate length over 10 weeks. Comparisons immediately post-treatment revealed greater improvements in pain, depression and quality of life in the massage group relative to controls, but no differences in disability, sleep disturbance or anxiety (Table 4). Brattberg maintained that there was a 37% reduction in pain following the massage treatment but that 30% of the improvement in pain had disappeared by 3-month follow-up and 90% of the reduction in pain was gone by the 6 month follow-up. Thus, it appears that the benefits of massage for fibromyalgia do not persist over the longer-term after the termination of active treatment. Brattberg recommended that following an initial treatment of 15 sessions, maintenance therapy may be instituted (e.g. once or twice per month). Another study randomly assigned 37 patients with fibromyalgia to one of three conditions: massage, usual care, usual care with follow-up phone calls from a nurse (36). The massage group received 10 treatments of indeterminate length over 24 weeks. Unfortunately, only 16 patients completed the full study protocol (six patients in the each of the two usual care groups and four patients in the massage group). Although the massage group showed a trend towards greater improvement in pain and self-efficacy for managing their condition, there were no between-group differences by the end of treatment (Table 4), likely due to the small cell sizes. Summary — Massage Therapy for Fibromyalgia The evidence supporting the use of massage for fibromyalgia is mixed. Whereas the Field group has found that massage leads to improvements in pain and symptoms compared with relaxation (33) or TENS (34), their work suffers from methodological limitation such as small sample size, inadequate blinding of assessors and an absence of follow-up assessments. The two other studies found either no benefits for massage (36) or only short-term benefits that eroded over time (35). Replication of the positive results reported by the Field group in an independent research group would increase confidence in their findings. Application to Mixed Chronic Pain Conditions Whereas the majority of existing trials of massage therapy have examined patients with discrete chronic pain syndromes, three studies have investigated the impact of massage on patients with a variety of chronic pain complaints. Walach and colleagues (37) conducted a randomized controlled trial comparing massage therapy (10–20 min sessions administered twice weekly for 5 weeks) to usual care for patients with various chronic pain symptoms (i.e. lower back, neck, shoulders, headaches). By 3-month follow-up, patients who received massage reported less pain, depression, anxiety and tiredness relative to controls (Table 5). However, the study authors noted that their study was limited due to lack of equivalence across groups on demographic characteristics. Moreover, it is unclear whether the groups were similar in terms of the type of pain complaints represented or important clinical characteristics such as the duration and/or severity of pain. Another study randomly assigned patients with chronic pain to receive either massage or relaxation (listening to a PMR tape recording) (38). The massage group received 6–10 sessions lasting 30 min each; patients were treated 1–3 times per week. The relaxation group listened to the audiotape twice a week for 5 weeks. Although the massage group evidenced improvements in pain, mental energy and self-perceived health status compared with the relaxation group immediately following treatment, by 3-month follow-up, there were no differences between groups. In a third analysis, patients with chronic musculoskeletal pain were randomly assigned to mindfulness-based stress reduction (MBSR), massage or usual care (UC) (39). MBSR involved 8 weekly 2.5-h sessions in a group format, with audiotaped meditation exercises assigned as daily home practice. The massage group received 1-h massage once per week for 8 weeks. At post-treatment, the massage group reported less pain unpleasantness and improved mental health compared with the usual care group. However, by 1-month follow-up, there were no differences among the groups in pain intensity or pain unpleasantness (Table 5). Moderate Support for Short-term Benefits of Massage Therapy for Mixed Chronic Pain Conditions These reports provide modest support for the immediate benefits of massage for a variety of chronic pain complaints. However, it appears that these treatment gains were not maintained following the end of active treatment. A potential difficulty with such studies including heterogeneous pain complaints is ensuring that treatment and control conditions are equivalent on key clinical and demographic characteristics. On the other hand, evidence of therapeutic effects across a variety of pain conditions supports the generalizability of the findings to a potentially broader group of patients. Future studies incorporating samples with mixed chronic pain conditions may also examine which types of pain conditions may benefit most from massage therapy in order to promote a more targeted approach to treatment. Meta-Analysis — Massage Therapy Effects on Pain As noted above, the majority of primary analyses and review articles have focused on the application of massage to discrete pain conditions. Moyer and colleagues (30) have published a recent meta-analysis of massage therapy research that examined the effects of massage for a variety of both chronic and acute pain symptoms. Data across trials were aggregated to investigate these effects across pain conditions on the immediate and delayed assessment of pain. According to Moyer et al. (30), immediate assessment of pain pertains to ‘single dose effects’ or short-term effects observed on the same day following massage therapy. The delayed assessment of pain reflects ‘multiple dose effects,’ which refers to outcomes that were assessed at various time points after treatment has been discontinued and following multiple sessions of massage. It should be noted that several of the studies included in this meta-analysis were also included in the Cochrane reviews discussed earlier in this article. Moyer and colleagues (30) concluded that massage therapy did not exhibit a significant effect on the immediate assessment of pain. The studies included in this category examined a wide range of pain problems ranging from back pain, neck pain, cancer pain, headaches and fibromyalgia, as well as acute pain in relation to surgery and other procedures (e.g. amniocentesis; cardiac catheterization). However, Moyer et al. (30) did conclude that massage evidenced a significant effect for the delayed assessment of pain. They maintained that patients who received a course of massage and were evaluated several days or weeks after the end of active treatment exhibited levels of pain that were on average 62% lower than controls. This conclusion was based on the results of 5 studies—two of these were conducted with patients experiencing LBP (9,11), one was conducted with patients with tendonitis (40), one was conducted with patients experiencing shoulder pain (24) (discussed earlier in this article), and one was an unpublished master's thesis (29) examining patients with neck pain. It should be noted that the Moyer et al. (30) analysis of the delayed assessment of pain did not include three of the trials discussed earlier in this article that did not find long-term benefits for fibromyalgia (35) and mixed chronic pain (38,39). The latter two studies were published after the Moyer review was completed; it is unclear why the study on fibromyalgia was not included in the meta-analysis. Nevertheless, the rigorous approach employed by Moyer and colleagues, which expressly included the calculation of between-group effect sizes and the aggregation of data across numerous trials lends confidence to their overall conclusion that multiple applications of massage therapy appears to confer lasting benefits on pain. Their work did not however, speak to the time period over which such analgesic effects are maintained, or the rate at which such effects decay. These considerations warrant further examination in additional studies. Discussion The existing literature provides varying levels of support for the effectiveness of massage therapy for chronic pain. The most abundant and rigorous evidence was found for the effects of massage on non-specific LBP. The Cochrane Collaboration (3), concluded that massage therapy may be beneficial for patients with subacute and chronic non-specific LBP, especially when combined with exercises and education. Whereas the evidence supporting the application of massage for LBP is fairly robust, there is less support for the use of massage for the other chronic pain conditions reviewed. This review suggests that the level of evidence for massage therapy effects by pain condition is (in order from most to least): LBP, shoulder pain, headache pain, fibromyalgia, mixed chronic pain, neck pain and CTS. Although shoulder pain has been the subject of only three studies, all of the studies yielded positive outcomes for pain and were methodologically rigorous, with one study including over 100 patients (25). Moreover, one of these studies was cited by Moyer et al. (30) in their meta-analysis indicating that massage therapy demonstrates significant effects on the long-term assessment of pain. Like shoulder pain, massage for headache pain has only been the subject of three studies but the data are somewhat weaker. The Cochrane Collaboration (15) concluded that there is moderate evidence that spinal manipulation is superior to massage plus placebo laser for pain related to cervicogenic headache, although these conclusions were based on the results of a single trial reported in two studies (16,17). Two additional studies provided preliminary evidence for the benefits of massage and craniosacral therapy in the treatment of pain related to migraine headaches (18) and tension-type headaches (20), respectively. However, in the migraine study (18), massage was combined with relaxation and other self-help techniques, making it difficult to draw conclusions regarding the specific effects of massage. There is considerably less support for the effectiveness of massage therapy in treating the remaining chronic pain conditions. Of the four studies examining massage therapy for fibromyalgia, only two studies, both by the same research group, revealed therapeutic effects (33,34), whereas the other two studies found no benefits (36) or improvements that attenuated over time (35). Thus, there is only modest evidence for the effectiveness of massage for pain related to fibromyalgia. For mixed chronic pain, the three studies to date provide somewhat conflicting findings. Whereas one study found that massage was superior to usual care (37), two other studies found that by follow-up, massage was no better than relaxation (38), mindfulness meditation or usual care (39). Taken together, these studies provide fairly weak support for the application of massage to mixed chronic pain. For neck pain, one trial using conventional massage techniques found that massage was similar to acupuncture by 3-month follow-up. A recent Cochrane review which included a broad array of massage techniques, many of which were considered questionable, reported that no firm conclusions could be drawn regarding the effectiveness of massage for neck pain (31). Only one published trial has investigated massage therapy effects on CTS; this study found that massage was superior to no treatment. Based on these findings, there is only preliminary evidence to support the effectiveness of massage for both neck pain and CTS. Putative Mechanisms of Massage Therapy for Chronic Pain The precise mechanism of action in massage therapy is not known. It has been proposed that increased parasympathetic activity (41) and a slowed-down physiological state may underpin the behavioral and physiological processes associated with massage. As discussed by Wright and Sluka (42), massage is thought to induce a variety of positive physiological effects that may contribute to tissue repair, pain modulation, relaxation, and improved mood. For example, these authors point to research showing that massage has beneficial effects on arterial and venous blood flow and edema (43). In addition, they note that vigorous massage has been shown to increase local blood flow and cardiac stroke volume (44), as well as improve lymph drainage (45); massage also appears to have an anticoagulant effect (46). Finally, Wright and Sluka maintain that massage may activate segmental inhibitory mechanisms to suppress pain and that some techniques may activate descending pain inhibitory systems (43), as suggested by gate theory (discussed subsequently). The main theories regarding the analgesic effects of massage include gate theory, the serotonin hypothesis, and the restorative sleep hypothesis (47). According to gate theory (48), pressure receptors are longer and more myelinated than pain fibers, and thus pressure signals from massage are transmitted faster, closing the gate to pain signals. The serotonin hypothesis maintains that massage increases levels of serotonin, a neurotransmitter that modulates the pain control system (49). The restorative sleep hypothesis holds that because substance P, a neurotransmitter associated with pain is released in the absence of deep sleep, the ability of massage to increase restorative sleep reduces substance P and consequent pain (50). There is little definitive data to support these major theories concerning the mechanisms underlying the analgesic benefits of massage. Clinical Implications: The Application of Massage Therapy for Chronic Pain The existing literature suggests that massage therapy may be a useful approach for pain relief in a number of chronic, non-malignant pain conditions, particularly musculoskeletal pain complaints (e.g., shoulder pain, low back pain). Massage is typically administered as adjunct therapy to help prepare the patient for exercise or other interventions and is rarely administered as the main treatment (3). Thus, massage is not usually considered a first line treatment, but rather as a complement to other conventional first line approaches (e.g., physical therapy; medications). It should be noted that the studies reviewed above did not specifically report on findings regarding possible interactions of massage therapy with other CAM or conventional medicine approaches. Nevertheless, the increasing popularity of massage and the fact that it is typically used as an adjunctive approach with other established treatments suggests that massage may be successfully integrated into the treatment of a variety of chronic or recurrent non-malignant pain conditions. The paucity of data on negative side effects pertaining to massage does not necessarily mean that such effects do not exist. Future work should focus on systematically characterizing those patients for whom massage is not indicated. Future Directions: Critical Issues for Studies on Massage Therapy for Chronic Pain This review highlights the need for continued rigorous research on the effectiveness of massage therapy for chronic, non-malignant pain conditions. Somewhat surprisingly, this review indicated that very few studies to date have focused on massage for pain related to chronic/recurrent headaches and chronic neck pain. Given that massage promotes relaxation, it would appear to be a particularly appropriate therapy for tension-type headaches as well as migraine related to increased stress. Moreover, at pointed out above, massage therapy may alter the mechanical stress caused by myofascial tissue disorders (21) which have been implicated in tension-type headaches (23). In light of the contradictory findings noted above, future work may also continue to examine massage therapy effects on pain related to fibromyalgia which involves wide-spread, diffuse pain that is often not responsive to traditional approaches. Numerous methodological problems were noted in the studies reviewed including small sample sizes, lack of equivalence across treatment and control groups, and inadequate blinding of assessors. However, one of the most notable limitations of the literature as a whole is that very few studies included follow-up assessments. As indicated by Moyer et al. (30) in their meta-analysis, the beneficial effects of massage therapy on pain are predominately evidenced after the end of active treatment. They concluded that such delayed effects on pain were substantial, with patients who were evaluated several days/weeks after treatment cessation exhibiting on average 62% less pain than controls and one study on LBP revealed significant benefits from massage persisting 1 year after the cessation of active treatment (11). It should be noted however, that the conclusions by Moyer et al. (30) were based on only 5 studies, suggesting that future trials of massage therapy should include follow-up assessments in order to further quantify such delayed effects. Moreover, additional studies may focus on examining the optimal time periods for the scheduling of ‘booster’ sessions to maintain treatment gains. Previous work has suggested that psychological treatment delivered according to a schedule with increasing time intervals between sessions (e.g. 1, 4, 10 intervening days) is more effective over the long-term compared to a uniform schedule (e.g. 5, 5, 5 intervening days) of treatment delivery (51). Thus, future research may also examine the optimal treatment schedule for delivery of massage therapy with a view to enhancing longer-term analgesic effects. Limitations of the Current Review and Concluding Statements The main limitation of the current study is its reliance on existing reviews and meta-analyses. Thus, many of the conclusions drawn in this article are based on the findings of other authors. Relatedly, the soundness of the methodological approach of these existing reviews may have been limited (e.g. due to improper exclusion of specific studies) as well as highly variable across reviews. Nevertheless, the reliance on extant reviews was considered necessary in order to synthesize a vast and diverse literature examining a broad array of chronic pain conditions. Another limitation of the present study is that only those pain conditions that were the subject of at least one controlled trial of massage therapy were included. Thus, not all chronic pain problems were examined in this review. It is possible that positive effects for massage therapy on other chronic pain conditions may have been reported in uncontrolled trials and/or case studies. In sum, this review identified important areas for future research on the effectiveness of massage therapy for chronic, non-malignant pain. Whereas there is fairly robust support for the analgesic effects of massage for non-specific LBP, there is only moderate support for such effects on shoulder pain and headache pain. Extant literature provides only modest, preliminary support for massage in treating fibromyalgia, mixed chronic pain, neck pain and CTS. One of the most important methodological considerations that should be addressed in future trials is the inclusion of follow-up assessments in order to allow further quantification of the longer-term effects of massage therapy on pain. Another key methodological consideration is the inclusion of comparison conditions that control for non-specific effects including physical contact and therapist time and attention. Moyer et al. (30) in their meta-analysis maintained that their positive findings for delayed assessment of pain are consistent with the notion that massage may promote pain reduction by enhancing restorative sleep. However, they note that data on sleep patterns was not included in the studies reviewed and therefore, this possibility remains to be tested. Thus, the careful consideration of potential mechanisms may inform future research, particularly with respect to the inclusion of key outcome variables as well as the examination of possible moderators and mediators of treatment response. Additional rigorous research is needed to establish massage therapy as a safe and effective intervention for the treatment of chronic, non-malignant pain.	[ "low back pain", "headache pain", "recurrent pain", "musculoskeletal pain", "fibroymalgia" ]	[ "P", "P", "P", "P", "U" ]
Pediatr_Nephrol-3-1-1766474	Hemodialysis in children: general practical guidelines	Over the past 20 years children have benefited from major improvements in both technology and clinical management of dialysis. Morbidity during dialysis sessions has decreased with seizures being exceptional and hypotensive episodes rare. Pain and discomfort have been reduced with the use of chronic internal jugular venous catheters and anesthetic creams for fistula puncture. Non-invasive technologies to assess patient target dry weight and access flow can significantly reduce patient morbidity and health care costs. The development of urea kinetic modeling enables calculation of the dialysis dose delivery, Kt/V, and an indirect assessment of the intake. Nutritional assessment and support are of major importance for the growing child. Even if the validity of these “urea only” data is questioned, their analysis provides information useful for follow-up. Newer machines provide more precise control of ultrafiltration by volumetric assessment and continuous blood volume monitoring during dialysis sessions. Buffered bicarbonate solutions are now standard and more biocompatible synthetic membranes and specific small size material dialyzers and tubing have been developed for young infants. More recently, the concept of “ultrapure” dialysate, i.e. free from microbiological contamination and endotoxins, has developed. This will enable the use of hemodiafiltration, especially with the on-line option, which has many theoretical advantages and should be considered in the case of maximum/optimum dialysis need. Although the optimum dialysis dose requirement for children remains uncertain, reports of longer duration and/or daily dialysis show they are more effective for phosphate control than conventional hemodialysis and should be considered at least for some high-risk patients with cardiovascular impairment. In children hemodialysis has to be individualized and viewed as an “integrated therapy” considering their long-term exposure to chronic renal failure treatment. Dialysis is seen only as a temporary measure for children compared with renal transplantation because this enables the best chance of rehabilitation in terms of educational and psychosocial functioning. In long term chronic dialysis, however, the highest standards should be applied to these children to preserve their future “cardiovascular life” which might include more dialysis time and on-line hemodiafiltration with synthetic high flux membranes if we are able to improve on the rather restricted concept of small-solute urea dialysis clearance. Introduction The European Paediatric Dialysis Working Group was established in 1999 by pediatric nephrologists from different European countries with a major interest in dialysis. The group has already published guidelines, mainly on peritoneal dialysis [1, 2]. Hemodialysis practices for children have improved over the ensuing 20 years, especially because of technological developments and the evolution from “minimum/adequate” to “optimum/maximum” dialysis prescription [3]. Therefore, new general recommendations seem necessary. These guidelines were initiated and discussed at meetings of the group and refined by e-mail discussion to develop a consensus of opinion, on the basis of cumulative clinical experience and reported studies. This paper will discuss the main factors affecting hemodialysis prescription and management in children. In some European countries hemodialysis (HD) is often preferred for children over the age of five years [3]. In contrast, peritoneal dialysis (PD) is offered to the younger children especially under the age of two years or weighing less than 10 kg. A multicenter European study has, however, found that 73% of 189 children were older than five years when peritoneal dialysis was started [4]. Factors ranked as first priority for choice of therapy [4] included age of the child (30%), parent choice (27%), distance from unit (14%), patient choice (11%), social condition (7%), and unable to do one mode (6%). Nevertheless important differences appear in the individual countries. Usually, however, HD is not offered to children less than 5 years old unless there are important contra-indications for PD [1]. For older children HD is applied for drop-outs from the PD program or if there are medical (rare) or psychosocial (more often) reasons for not performing PD. Choosing a mode of dialysis, either HD or PD, for a child requires consideration, among other factors, of the probable impact of either mode of dialysis on the maintenance of residual renal function (RRF), because of its specific impact on patient outcome. Although there is no general consensus, peritoneal dialysis has been associated with less risk of RRF loss [5, 6]. Overall the choice of the mode of dialysis is just a part of the integrated care model, each child should be considered for a combined dialysis-transplantation program. Provision of adequate vascular access remains the single greatest obstacle to successful HD, especially in infants. Unlike in the USA, where patients frequently use a central catheter for vascular access [7], in Europe an arteriovenous fistula is the most common vascular access for chronic/long term dialysis [8]. According to the K-DOQI guidelines, the percentage of catheters in a dialysis unit for adults should be less than 10%, although many pediatric centers do not meet this standard, because of the difficulty of creating fistulas in smaller children, especially in children less than 2 years of age. During the past two decades there have been many improvements in the technology [3]: bicarbonate used as buffer in the dialysis solution, volumetrically controlled ultrafiltration, smaller dialysis lines and synthetic membranes useful even for babies, modeling of ultrafiltration rate and dialysate composition, on line hemodiafiltration and the concept of ultrapure dialysate, i.e. sterile and pyrogen free. Non invasive technologies to assess patient target dry weight and access flow offer a potential decrease in dialysis morbidity and costs [9]. Recently marketed medications to improve anemia, for example erythropoïetin even darbepoietin, and iron infusion, contribute to the clinical improvement of the hemodialysis session [10]. Dialysis adequacy quantification by urea kinetic modeling enables a more specific approach to dialysis dosing and indirect assessment of protein intake, despite the limited value of small-solute clearance [11]. Nevertheless, it has been widely accepted that clinical results depend at least in part on the dialysis dose delivered [12, 13, 14, 15]. In fact, a single center experience shows the beneficial impact of longer dialysis duration on clinical outcome in children [13]. However, it is now becoming more and more evident that increasing the dialysis dose when delivered only three times weekly is an unphysiological strategy, self limited by the potential increased risk of hemodynamic and electrolytic disturbances [16, 17, 18]. In adult care, there is a growing interest in the use of daily dialysis, because long term experience has shown good results [16, 17, 18]. In children only a pilot study in one center supported the positive impact of daily dialysis in very non-compliant adolescents [19]. In children the hemodialysis prescription should be individualized. Choice of the mode of hemodialysis should take into account the presumed waiting time before kidney transplantation as a “ justification” for the use of “ the best available” mode having the highest cost and, conversely, being supported by very limited/preliminary studies only [13, 19]. The importance of the choice of material used for dialysis and its application should not obviate the need for management of the entire child with ESRF, especially regarding optimum nutrition [20]. Because dialysis per se is not able to correct completely the numerous functions of the kidney lost during ESRF, medications and dietary recommendations are needed in children on hemodialysis [20]. Recombinant growth hormone is often needed considering the growth velocity rate of children on chronic dialysis [3]. Guideline 1: the dialysis unit hemodialysis should be delivered in a “pediatric” dialysis center with a multidisciplinary support team which supports individualized and integrated therapy nutrition, growth, and educational support are of major importance Because of the specific needs of children, hemodialysis should be delivered at the best, and probably only, in a pediatric dialysis unit [3, 4, 7] This includes the treatment of adolescents up to the age of 18 years and beyond depending upon their physical and psychological development and transition arrangements to adult units [21]. Taking care of a child with ESRF necessitates an engaged team consisting of doctors, nurses, dietician, psychologist, school teacher, play therapist, and social worker [22]. This “second family or support team” should be multidisciplinary and immediately available to the chronically ill child, both close and distant enough to stimulate normal family life, supporting a proper (school) education, leaving all possibilities open for “full” integration into society in the future. Hemodialysis, in contrast with peritoneal dialysis, is usually performed in an hospital setting, with a frequency of three times per week for most patients. This frequency may be increased to address the specific needs of babies and/or adolescents requiring “more dialysis” [3, 13, 19]. Guideline 2: water quality adequate in terms of biochemical composition free from microbiological contamination The dialysis machine needs water for dialysate production adequate in terms of biochemical composition and free from microbiological contamination, i.e. germs and endotoxins (Table 1). Water purification depends on the disposable water quality. Usually filtration with charcoal and the small sieving coefficient associated with reverse osmosis produces water for dialysis in accordance with the recommendations [23] (Table 2). Currently, all new dialysis machines have the ability to filter the dialysate through a high flux membrane, which increases microbiological purity.Table 1 Water contaminants and associated complicationsaDissolved organic materialComplicationsContaminants:- Pesticides, herbicidesNo documentation during dialysis- Chloramines, chlorine compoundsSevere hemolytic anemiaBacteria and pyrogens:- BacteriaBacteremia or septicemiaFever, chills, shakingHypotension and death- PyrogensPyrogenic reaction-feverChills, uncontrollableShaking, vomiting, hypotensionaNot exhaustiveTable 2 Definitions of water and dialysate quality (levels given as an upper limit for water-quality definition [70])Bacterial growth (cfu mL−1)Endotoxin (EU mL−1)Cytokine-inductionAAMI, water2005+European pharmacopoeiaRegular water1000.25+Ultra-pure0.010.03–Sterile10−60.03− In hemodiafiltration using an on-line technique [24, 25] with direct production from the dialysate of the hemofiltration substitution fluid, the dialysate benefits from double ultrafiltration, producing an ultrapure dialysate which is sterile and endotoxin free, at least at detectable levels. This ultrapure dialysate should limit the risks related to microbiological contamination, i.e. inflammatory process induction with both acute and chronic consequences [23]. This level of ultrapure dialysate is also required for synthetic high-flux membrane use even or especially when used in a conventional hemodialysis mode. Decontamination or sterilization by chemical agents or by heating should be performed in line with water, before final dialysate production by the dialysis machine, by filtration and osmosis installation and water distribution, without any break between sterilization and final dialysate. Quality control of the water for the dialysate should be performed regularly with regard to chemical composition (at least once per year), and final dialysate purity should be assessed with regard to bacteria and endotoxins (more regularly, depending in part on the mode of dialysis, weekly for high-flux membrane use) (Table 2) [23]. Guideline 3: the dialysis machine volumetric ultrafiltration control option for both single and double-needle dialysis In the last decade numerous innovations in equipment have been developed by different manufacturers [3]. But the relevance to child outcome remains unknown, because of the absence of sufficient controlled study results. Nevertheless the following innovations seem essential: dialysate production by double dilution pumps using volumetric ultrafiltration control and blood pumps with double pumps available for single-needle dialysis. Other “high-tech” innovations only deserve mention because of their limited application in “expert” centers: individual modeling of the dialysis session with monitoring of ultrafiltration and dialysate solute concentration (i.e. sodium, bicarbonate); polyvalency machine which enables not only conventional dialysis but also hemofiltration and hemodiafiltration providing the highest standard in terms of tolerance and efficiency [24, 25]. Newer dialysis machines provide monitoring of hematocrit variation as a major promising innovation [9, 26] and direct urea kinetic monitoring [27]. There is a restricted offer for blood thermal monitoring to avoid loss of calories to the dialysate or to prescribe cooled dialysate [28]. All these innovations enable individualized hemodialysis for the children, but their regular application should take into consideration the balance between the expected benefits and the costs. Guideline 4: blood lines available in infants/babies size biocompatible material A range of blood lines are available for dialysis of babies to dialysis of the largest adolescent. They should be considered for their biocompatibility, type of sterilization (ethylene oxide-free), and the blood volume required [3]. Guideline 5: principles of blood purification small solute clearance and more, from diffusion process (urea) to convection (other uremic toxins “middle molecules”) mass transport hemodiafiltration is an option to consider to obtain “maximum” dialysis efficiency Uremic toxin extraction in dialysis [3, 24] is related to a combination of the diffusion process and convection mass transport (Table 3). In hemodialysis (HD), blood purification depends mostly on a diffusion process secondary to a concentration gradient, which ensures the best elimination of small molecules (urea). HD clearance (KHD) correlates directly with blood flow rate. In hemofiltration (HF), uremic toxin extraction is mostly dependent on convection mass transport secondary to a pressure gradient, which optimizes the elimination of both low and middle-molecular-weight compounds. HF clearance (KHF) directly correlates with ultrafiltration flow rate which is limited by the blood flow rate. In the post dilution mode, i.e. replacement fluid in the venous line chamber located after the dialyzer membrane, maximum filtrate flow rate is less than half the blood flow rate; it is usually one third, to limit the risks of excessive hemoconcentration. In the predilution mode, i.e. replacement fluid perfusion in the arterial line chamber, which is situated before the dialyzer membrane, maximum filtrate flow rate should be two thirds of or equal to the blood flow rate. Hemodiafiltration (HDF) combines HD and HF simultaneously, which enables blood purification by both a diffusive process and convective mass transport. HDF clearance (KHDF) in post-dilution mode is measured by use of the Granger formula [24]:Table 3 Dialyzer membrane permeability: diffusion and convectionDiffusion processConvection mass transportMembrane areaUltrafiltrate flow (QUF)Mass-transport coefficientHydraulic permeabilityConcentration gradientTransmembrane pressure (TMP; mmHg)Blood flow×extraction coefficientMembrane areaSieving coefficient (S)ci and co are inlet and outlet solute concentrationsMolecular permeabilityCUF is the ultrafiltrate solute concentration(postdilution)(predilution)KHD is hemodialysis clearance and KHF is hemofiltration clearance On replacement of QUF×S by KHF and Qb by Kmax (maximum achieved clearance) the formula for KHDF becomes: Thus it is clear that in terms of blood purification KHDF enhances the clearance of a uremic toxin if HF or HD clearances are lower than the Kmax (equal to the blood flow rate). HDF with a highly permeable membrane is as efficient as HD for low-molecular-weight compounds, but is more efficient than HF for low-molecular-weight compounds [29]. Moreover, HDF, besides its blood purification efficiency, is associated with a lower intradialytic morbidity rate [30, 31], as is HF [3]. On-line HDF [24, 25], in which filtered dialysate free of toxins and pyrogens is used as replacement fluid, enables use of an elevated convection fluid rate, especially in predilution mode, and facilitates a dialysis dose increase without a cost increase. The use of ultrapure dialysate, i.e. sterile and pyrogen free, as used for on-line HDF dialysate should reduce the diseases possibly associated with chronic inflammation related to contaminated dialysate [23], e.g. β2 microglobulin amyloidosis, hypercatabolism with loss of lean body mass, decreased growth rate, fibrosis and cardiovascular diseases. A high flux membrane [32], with an elevated ultrafiltration coefficient of permeability enabling backfiltration from the dialysate to the blood compartment, which is called retrofiltration, increases these risks, especially with contaminated dialysate [23]. Guideline 6: extracorporeal blood access and circulation fistula vascular access is preferred for long-term chronic hemodialysis in young children, less than 15 kg, the time needed to develop a fistula before it can be used could be some months the double-needle technique is the standard, but single needle with double pump system is an alternative a single lumen catheter with clamps offers for small children an acceptable compromise between a very low extracorporeal blood volume and valuable dialytic efficacy total extracorporeal blood volume (needles, tubing and dialyzer) should, approximately, be less than 10% of patient total blood volume anticoagulation in the extracorporeal circuit is achieved either with conventional heparin or with low-molecular-weight heparin an extracorporeal blood flow rate (QB) of 150–200 mL min−1 m−2 or 5–7 mL min−1 kg−1 is often sufficient The success of chronic hemodialysis depends on good vascular access: internal arteriovenous fistulae (AVF), shunt (AVS), graft (AVG) or central venous catheter. The type of access used is variable depending on factors in different units and countries, for example surgical experience, patient age and size, the time available before dialysis must be started, and the presumed waiting time before transplantation. Patient choice plays a major part, especially with adolescents. A catheter is more commonly used in the USA than in Europe [7, 8]. A catheter can be a primary access particularly in acute renal failure or chronic renal failure with acute presentation, in small children and in the case of a presumed short period on chronic hemodialysis. Internal jugular vein catheter access is superior to subclavian vein; it admittedly preserves the future arteriovenous fistula implantation on the arm. Femoral catheter access should be used only for “rescue and transient” access if intensive care is needed: it is easy to perform but with a higher risk of infection and thrombosis. A double lumen cuffed catheter, at least 8 French, is mostly preferred for children and has been reported to have a survival rate as high as 60 to 85% in one year [33], or as low as 30% [34]. Nevertheless in small infants a single lumen catheter used with the alternative clamps technique offers an acceptable compromise between recirculation and both the amount of extracorporeal blood volume and the achieved blood flow [35]. Thrombosis, a major cause of catheter failure, is reported to be between 9 and 46% [34]. Thrombosis causing poor flow can be corrected to salvage the catheter by different methods: catheter replacement over guidewire, systemic oral anticoagulation and local urokinase or tissue plasminogen activator instillation [36]. Loss of catheter access related to infection has decreased during the last decade; the aggressive use of antibiotics and perhaps antibiotic lock therapy, although not universally accepted, account for this lower rate of infection related catheter loss [34, 36, 37]. Microsurgery enables creation of a functional AVF at the wrist in most children, even small ones [8] but only a few surgeons are trained for vascular microsurgery, which therefore is rarely used. Creation of a fistula at the elbow is a second-choice vascular access. With a non functional cephalic vein, a basilic vein transposition, i.e. superficialization, is possible [38]. Synthetic grafts should be reserved for children who have exhausted autologous veins and should be used in children only very rarely. For all these reasons preoperative evaluation of the vessels to determine the correct choice of vein before the operation is mandatory. The non-dominant arm should be regarded as first choice of fistula implantation. The survival rate for a AVF is higher than the survival rate for a catheter, with more than two thirds of the children having a functioning AVF at four years [8]. With a basilic vein superficialization the fistula should not be used before full healing (2 to 6 weeks) to avoid a dissecting hematoma. Otherwise the time needed for venous development before use depends on the age of the patient and the place of the AVF (distal or proximal). In small children this period of time is often a delay of months. Before surgery it is essential to avoid venopuncture of the selected arm in the weeks before AVF creation. It is of interest to protect the dominant arm from the beginning of taking care of a child with “chronic dialysis risk” to enable, if necessary, implantation of a fistula. Such venoprotection should not be forgotten for peritoneal dialysis children, even babies/infants. For a period of time before surgery, especially for small children, [8] dilatation of the veins by immersion of the forearm in hot water is advantageous, a maneuver enhanced by placement of a tourniquet. A proximal AVF with a high blood flow, usually close to 1000 mL min−1 m−2 , is a risk factor for cardiac failure. Nevertheless, the major complication is thrombosis, consequent to local stenosis. Therefore, follow up of the access flow is essential, on the one hand clinically: auscultation (the sound of the AVF is maximum at the surgical site and decreases with distance from the fistulae), observation (elevation of the forearm should induce emptying of the previous dilated veins, and on the other hand by Doppler ultrasound or vascular access flow monitoring [9]. Application of regular access flow monitoring can be used to detect vascular stenosis before complete AVF thrombosis [9]. But it should be remembered that “Transonic” access flow monitoring can only be performed with double-pump dialysis and is not available for pediatric-sized blood lines. The extracorporeal blood flow rate is achieved through venous puncture, most often via two needles, one for blood aspiration called the arterial needle, one for venous reinjection called the venous needle. The distance between the needles should be sufficient to limit recirculation, which is best prevented by opposite orientation of the needles: the arterial one toward the fistula, the venous one in the opposite direction. Usually the needle size is 17-gauge at initiation of dialysis; thereafter considering patient need and fistula development 16 or 14-gauge needles, particularly in adolescents, can be used to achieve a sufficiently high blood flow rate. Pain related to the puncture should be prevented by anesthetic cream (Emla or Amelop); this advance is important for both the children and nurses [39]. An extracorporeal blood flow rate (QB) of 150–200 mL min−1 m−2 , 5–7 mL min−1 kg−1, is often sufficient to achieve the targeted goals with double needle dialysis; in small children QB is determined using body weight (BW, kg): (BW+10)×2.5=QB (mL min−1). The arterial blood aspiration pressure should be monitored if possible and kept between 150–200 mmHg to limit endothelial trauma. For single-needle dialysis in children the highest blood flow rate is obtained with a double pump system (venous flow higher than arterial flow) monitored by the pressure, system called time pressure regulation. The risk of recirculation is important with the latter; some machines limit this risk more than others, especially with the addition of clamps. Conversely for small infants a single lumen catheter used with the alternative clamps technique is an acceptable compromise between recirculation and both the extracorporeal blood volume and the achieved blood flow [35]. The total extracorporeal blood volume (needles, tubing, and dialyzer) should preferably be less than 10 % of patient total blood volume. This is essential for small children; however, the relative normal hemoglobin level obtained with erythropoïetin therapy enables this volume to be exceeded slightly without significant hypotension at the end of dialysis session when the patient reaches dry body weight. Nevertheless, it should be kept in mind that the higher the extracorporeal blood volume, the higher the volume of returned fluid, which will load the patient with fluid at the end of the dialysis session. (In very small children the substitution by air may be necessary to limit blood loss on one side and high substitution volume on the other side, but is very dangerous and should be strictly monitored.) System priming with saline, albumin, and sometimes blood should be applied in the first dialysis sessions with babies or small infants. Anticoagulation of the extracorporeal blood volume is performed either by use of conventional, heparin with continuous infusion of 20 to 30 IU kg−1 h−1, or with low-molecular-weight heparin at 1 mg kg−1 as a bolus at the beginning of the dialysis session. If the hematocrit is over 35%, the risk of clotting is increased. Regional citrate anticoagulation is sometimes used especially when acute dialysis is needed [2]. Predilution treatment, feasible in either hemofiltration or hemodiafiltration, reduces the risk of clotting and even enables dialysis without anticoagulation in some circumstances. In the presence of thrombopenia heparin-toxicity is to be suspected. The venous blood line has a pediatric size air-trap chamber to limit extracorporeal blood volume. The dialysis membrane is protected by an arterial chamber of expansion which in small children is often not incorporated in the line to reduce the extracorporeal blood volume. Prevention or treatment of ethylene oxide allergy is possible by using steam sterilization of needles, lines, and membranes; this is becoming the preferred option throughout Europe. Guideline 7: which dialyzer membrane to “choose” synthetic membrane, low flux, capillary configuration high-flux membrane use requires use of ultrapure dialysate removal of urea and other uremic toxins dialytic should be considered, especially in chronic, long-term dialysis Three general types of membrane are available at present [32]: unmodified cellulose (low flux and so-called bioincompatible membranes), modified/regenerated cellulose (low flux or high flux; so-called relatively biocompatible), synthetic (low flux or high flux; so called relatively biocompatible). The choice of a dialyzer membrane should take into account the following (Table 4):Table 4 Dialyzer membranes: practical parameters of choice- Type of membrane: biocompatibility toward complement system- Initial blood volume needed, i.e. area-related, quality of restitution- Molecular permeability: maximum clearance for urea and the other uremic toxins, e.g. phosphate, related to potential patient osmotic risk- Hydraulic permeability: possibility of use for HF or HDF procedure, but related to back filtration risk, high flux membranes need ultrapure dialysate- Adsorption capacity on to the membrane (a characteristic of synthetic membranes)- Costthe biocompatibility of the material towards leucocytes and complement activationthe blood volume priming requirement, which is membrane area-relatedthe permeability, determined in the most simple way by two characteristics:hydraulic permeability (CUF) measured in mL per mmHg of transmembrane pressure achieved per hour, i.e. either low permeability, CUF under 5 mL mmHg−1 h−1 (low-flux membrane), and high permeability, CUF over 15 to 20 mL mmHg−1 h−1 (high-flux membrane)molecular permeability determined at least by the molecular weight of the molecule considered, usually between 0.8 and 0.9 for urea and lower for the other uremic toxins with a cut off of zero for albumin. In practice this cut off is often under a molecular weight of 20,000 Daltons. The profile of this molecular permeability [40, 41] is a specific characteristic of each manufactured dialysis membrane. Highly permeable membranes give the theoretical potential for middle-molecular-weight (Babb theory; 500 to 2,000 Daltons) uremic toxins being removed during dialysis. In adult dialysis patients the clinical benefits of improved removal of middle molecules by high flux, large pore, biocompatible membranes, more or less established, are [41]: reduction of uremia related amyloidosis, maintenance or residual renal function, and reduction of inflammation, malnutrition, anemia, dyslipidemia, and mortality.The absorption capacity on to the membrane, (IL1, TNF, IL6, β2 microglobulin):a characteristic of synthetic membranes For conventional dialysis low-flux membranes are suitable, but to achieve hemofiltration or hemodiafiltration high-flux membranes are necessary. The higher the hydraulic permeability, the higher is the backfiltration risk; this process could be limited both by permanent convective flow from the blood compartment to the dialysate compartment, as ultrafiltration (HF, HDF, or at least weight loss) and by use of ultrapure dialysate. Synthetic membranes seem the best theoretical choice but clinical justification of the relatively higher cost is uncertain [32]. Justification for use of high-flux synthetic membranes, as used in on-line HDF, remains a matter of debate for children on dialysis for short periods only while waiting for their kidney transplant.Reuse of the membrane is not applied in practice for children. Guideline 8: the dialysate bicarbonate buffered, low calcium level (1.25 mmol L−1) becomes the standard, glucose concentration at physiological level, dialysate quality control (germs and endotoxins) is required The dialysate is prepared as a dilution of concentrate with water, ideally with ultrapure water. The composition of the dialysate has changed over the last two decades [42]. Acetate as buffer has been replaced by bicarbonate, with the development of machines with two separate dilution pumps, one for bicarbonate concentrate free from calcium, often as a powder, and one for the acid concentrate containing residual levels of acetate and the electrolytes (Na, K, Cl, Ca). The current use of oral calcium carbonate as a phosphate binder has mandated the need to decrease the calcium concentration of the dialysate, usually at a low rate, 1.25 mmol L−1 Ca2+, less often at a normal rate, 1.5 mmol L−1, avoiding the “historically” high level of 1.75 mmol L−1 Ca2+. In fact, the use of calcium carbonate combined with a high concentration of calcium in the dialysate, often led to an elevated Ca×P serum product, compared with the current recommendation of a product less than 5 mmol2 m−2 (60 mg2 dL−2) [43, 44]. This Ca×P serum product seems to be an important factor implicated in the vascular calcifications seen in the dialyzed patients [43], affecting even the dialyzed children [44, 45]. The need for glucose in the dialysate is of importance [46] and should be near the physiological concentration. Higher glucose concentrations or the introduction of parenteral feeding during dialysis will drive the potassium into the cells, leading to ineffective potassium-extraction [41]. Potassium-free dialysate is rarely used because of the theoretical risk of hypokalemia [42]. Therefore “low” (1–1.5 mmol L−1), “normal” (2–2.5 mmol L−1), and “high” (3–3.5 mmol L−1) potassium dialysate are available enabling individual adaptation and prevention of the arythmogenic potential of dialysis [42]. Nevertheless special attention should be devoted to avoiding any confusion among the “potassium charged” dialysates. Sodium concentrations have increased from the previous classical level of 132 mmol L−1 to a more physiological level of 138 to 144 mmol L−1. Newer machine capabilities enable dialysate profiles to change during a dialysis with respect to sodium and ultrafiltrate profiles [47, 48] to increase tolerated weight loss; and bicarbonate profiles [49], to enhance phosphate removal. Intermittent ultrafiltration rates, enabling better plasma refilling is the most common profile used. Similarly, the dialysate flow rate can be adapted to need, usually in the range 300 to 800 mL min−1. In general practice, 500 mL min−1 is used. The dialysate flow is usually in the opposite direction of the blood flow, separated by the membrane of the dialyzer. Dialytic thermal exchanges seem of importance especially for babies and/or high-flow dialysate use, leading to a risk of patient hypothermia. Control of thermal exchanges during a dialysis session is therefore available on a new machine [3, 28]. Guideline 9: post-dialytic dry weight assessment and adjustment particularly difficult to define in growing children no “unique” optimum method, importance of a clinical “pediatric” experience need for regular assessment in a growing child close collaboration with pediatric renal dietician Patient dry weight is defined as the weight at the termination of a regular dialysis session, below which the patient will become symptomatically hypotensive. Incorrect estimation of dry weight will lead either to chronic fluid overload or chronic dehydration. Estimation of dry weight is particularly difficult in children for many reasons. First, the hypotensive tendency during a dialysis session is multifactorial and not only related to the ultrafiltration rate but also to the plasma refilling rate capacity [47, 48]. Second, body composition, i.e. total body water ratio to total body mass, is variable with age, especially during infancy and puberty. In infants and in adolescents dry weight must be assessed almost monthly to follow rapid body composition changes during a rapid growth period. This is also important under anabolic conditions such as with growth hormone treatment, and conversely under catabolic conditions such as the ill child with intercurrent infections or reduced food intake. Clinical criteria used to assess hydration status are important but not always reliable. Therefore, different approaches have been proposed: assessment of total body water by bioelectrical impedance analysis [50], continuous measurement of hematocrit variations by non-invasive methods during dialysis [9, 26], plasma atrial natriuretic peptide or cyclic guanosine monophosphate determination [51], and, last, by echography of the inferior vena cava (IVC) [52, 53, 54, 55]. Measurement of the diameter of the IVC (IVCD) by ultrasound, expressed as an index to body surface-area in mm m−2, and the decrease on deep inspiration, called the collapse index, expressed as a percentage (%) seems to be an accurate non-invasive method easily performed serially. An IVCD between 8.0 and 11.5 mm m−2 and a collapse index between 40 and 75 % is considered as representing normovolemia [52, 53, 54, 55]. However, unlike body impedance, interstitial volume and sodium balance are not reflected by IVCD [55]. In fact all these approaches have to be balanced by clinical judgment and experience and combined with nutritional support. Achievement of dry weight during ultrafiltration is associated with a drop of the hematocrite level. Ultrafiltration is well tolerated until a certain level of decrease of initial hematocrite, called “crash hematocrite” a patient individual characteristic, usually over 10% blood volume reduction over a 3-h session [56]. If the hematocrite curve is flat over time during a dialysis session, the patient could be considered as being over his optimum dry weight [9, 56]. In practice, monitoring of hematocrit (or blood volume) and guided ultrafiltration should avoid both fluid overload and hypotensive “crash hematocrit” and consequently approach more precisely the patient dry weight [9, 56]. Guideline 10: urea dialytic kinetic, dialysis dose, and protein intake assessment (nutrition) Urea kinetic modeling (UKM) has been widely accepted as a method for dialysis dose assessment despite its limited value as a unique measure of dialysis adequacy. Does small solute clearance really matter? [11]. In adult patients the HEMO study suggested that increasing urea clearance above currently accepted target ranges does not lead to improved patient outcome [11]. Although urea per se is not toxic in concentrations normally encountered in dialysis patients, it may serve as a marker of unknown toxins of uremia, some of which are called “middle-molecular-weight” uremic toxins [11, 41]. UKM facilitates identification of underdialyzed patients and recognition of dietary compliance. The measures most widely used to gauge dialyzer treatment are Kt/V, that is dialyzer urea clearance (K) multiplied by duration (t) of the dialysis session and divided by urea volume (V) of distribution, and the normalized protein catabolic rate (nPCR) [57, 58, 59]. Urea dialytic reduction rate (URR) is derived from the pre and post-dialysis serum urea values and quantitates urea removal by dialysis. URR expressed as the ratio post/pre should be at least equal to or lower than 0.35 and when expressed as the difference between pre and post-urea, divided by the predialysis value, should at least equal to or higher than 0.60 [60]. URR is proportional to dialysis efficiency, and thus to urea dialytic clearance. URR is inversely proportional to the urea refilling rate of the blood compartment and the extracellular space (EC) from the intracellular space (IC), called the transcellular urea mass transfer coefficient (Kie). URR is also correlated with the amount of urea dialytic removal (Kt) compared to the amount of urea body content (V) and thus to Kt/V. Usually urea dialytic clearance in children is low in comparison with the high Kie which is between 200 to 1000 mL min−1 (6 to 12 mL min−1 kg−1 BW) [58, 61]. Nevertheless, after dialysis the concentration of urea in plasma increases rapidly in an initial period, usually until 60 min postdialysis [62]. This postdialytic urea rebound (PDUR) is a multifactorial event [63, 64]. Vascular access and cardiopulmonary recirculation occurs within the first 2 to 3 min of discontinuing hemodialysis and account for 60 to 70% of total PDUR. Subsequently, tissue rebound occurs, because of intercompartmental, i.e. IC versus EC, urea dysequilibrium at the end of the dialysis session and tissue re-equilibration which is usually complete within one hour postdialysis, reaching the equilibrated postdialytic plasma urea concentration. For highly diffusible substances such as urea, distribution in total body water (TBW) seems to be limited by cardiovascular flow rather than diffusion [64]. The apparent IC–EC two-pool model should perhaps be the result of a regional blood flow distribution system in which approximately 80% of TBW (and thereby urea) is located in muscle, bone, and skin, with organs receiving only 20 to 30% of the cardiac output, i.e. low-flow system. The remaining organs contain only 20% of TBW (hence urea) but receive 70–80% of the cardiac output, high-flow system. One would expect the urea concentration in these organs to fall quite rapidly during dialysis. This flow system, and the vascular resistance model could explain the great variability of the PDUR among patients for which the IC–EC two-pool model is not accurate. Do some patients have thicker cell walls than others? By contrast the possible causes of vascular resistance variability, i.e. hypovolemia, hypertension, heart failure, hematocrit, alkalosis or acidosis, low-temperature dialysate, can explain PDUR variability. The URR variability could also be explained by vascular resistance changes over the dialysis session, at least for urea [64]. Kt/V calculation based on a single pool urea model neglects compartmental urea distribution in the body, hence PDUR, resulting in overestimation of actual Kt/V. Therefore, a two pool model should be applied, using instead of the urea plasma concentration at the end of the dialysis, the equilibrated urea, i.e. 60 min postdialysis [65, 66]. Other improvements from the initial formula are proposed to provide a more accurate Kt/V calculation: weight loss (UF/BW) and urea generation during the dialysis session (0.008td), leading to the Daugirdas and Schwartz formula proposed in 1994 [60]:where td is the dialysis time (h), Cpre and Ceq are, respectively, the predialysis and equilibrated postdialysis urea concentrations, and UF/BW is the ultrafiltrate-to-body weight ratio (L kg−1) The predialysis blood sample should be taken from the arterial line, before any rinsing. Because of the practical difficulty in obtaining the postdialysis equilibrated urea sample 60 min after the end of the dialysis, different indices have been proposed to estimate Ceq, for example using a 6 min [67] or a 15 min [68, 69] post treatment sample. The most important rule of the urea end dialysis sample should be the use of the “stop dialysate flow method” [67], measuring urea 6 min after angio access was removed and cardiopulmonary recirculation completed. The other major cause of error for the Kt/V calculation is determination of V. The V, hence the TBW, could be calculated from a formula (Table 5) [71] or determined by bioimpedance measurements [50].Table 5 Formulas enabling calculation of the volume of distribution of urea in liters (total body water) using height, weight, sex and age (from Ref. [67])Boys:Ht<132.7 cmV=1.927+0.465/BW (kg)+0.0045/ht (cm)Ht>132.7 cmV=−21.1933+0.406/BW (kg)+0.209/ht (cm)Girls:Ht<110.8 cmV=0.076+0.507/BW (kg)+0.013/ht (cm)Ht>110.8 cmV=-10.313+0.252/BW (kg)+0.154/ht (cm) Guideline 11: dialysis dose and outcome only “small solute urea clearance” prescription? a minimum Kt/V urea level of 1.2–1.4 is thought to be desirable; adequacy tests should be performed monthly dialysis and residual renal small-solute clearance are not equivalent dialysis prescription should be adequate before being optimum, not only a “urea dialysis dose” Although the optimum level of Kt/V required is matter of debate, a minimum Kt/V level of 1.2–1.4 is now thought to be desirable [11]. Overall, this Kt/V as an index of dialysis dose should only be analyzed in comparison with the nPCR, hence the diet, protein and caloric intake (Fig. 1). Because of the mathematical relationship between Kt/V and nPCR [72, 73] the real impact of these variables for a given patient would determine the therapy necessary for a patient to achieve an “urea dialysis dose”. Nevertheless increasing dialysis dose seems to have a direct impact on nutrition [74] and the combination of increased dialysis dose and adequate nutrition can promote normal growth in children treated with long-term hemodialysis [75]. Therefore malnutrition should be avoided [76] by using a diet survey, anthropometric measurements, and perhaps IGF1 determination [77].Fig. 1 Dialysis prescription balance Does small solute clearance, i.e. urea, really matter? [11]. Because of the limited number of children on chronic dialysis the relationship between optimum urea dialysis dose and patient outcome will be “difficult” to establish. It is, however, known that blood purification, dialysis and residual renal small solute clearance, are not equivalent [11, 78] giving more importance to other uremic toxins, whose removal is enhanced using high-flux dialyzer membranes with on-line hemodiafiltration [3, 24, 32, 41]. Even if the pediatric data only seem to be unique center experiences, the case for a greater urea dialysis dose [12] could be correlated with both growth rate enhancement [12, 13, 19, 78] and improved cardiac function [14, 19]. The duration of each hemodialysis session is also matter of debate, long duration being able to induce regression of left ventricular hypertrophy in adult patients [16] and being able in children [13] to promote growth and well being. In the same way daily dialysis seems related to better clinical results both in adults [17, 18, 77, 79] and in adolescents [19]. Dialysis prescription should be adequate before being optimum (Table 6) [80]. In long term chronic dialyzed children the individualized prescription should consider all the available new strategies to fully preserve at the best “the life chances” [13, 14, 19].Table 6 Hemodialysis prescription for children: adequate, before optimum- Dialysis modality should enable achievement of blood pressure control (without antihypertensive medications for most children), normal myocardial morphology and function- Dialysis dose prescription should not only be an urea dialysis dose. Removal of the other uremic toxins should be considered, not only middle molecules but overall phosphate- Dialysis frequency and duration must be adjusted to the tolerance of ultrafiltration to reach the dry weight. Ultrafiltration rate should not exceed 1.5±0.5% of body weight per hour (in theory no more than 5% BW loss per whole session ). Blood volume (hematocrite) guided ultrafiltration secure- A regular diet survey is essential to maintain adequate protein and calorie intakes. Urea kinetic assessment enables not only urea dialysis dose calculation, i.e. Kt/V, but also estimation of protein intake by use of the PCRn calculation (protein catabolic rate). Fasting to enable a short duration three times a week dialysis schedule is inadequate care management- Too fast ultrafiltration can induce hypotension and cramps during dialysis, usually during the second half time session, and fatigue and/or hang over after dialysis- A small solute, e.g. urea, clearance which is too high is a factor of disequilibrium syndrome occurring during dialysis, usually after the first half/or one hour session time with headache, even seizures, nausea, vomiting, sleepiness or a hypertensive tendency with a narrow range between systolic and diastolic pressure values. Symptoms usually disappear a few hours after the end of the dialysis Guideline 12: the dialysis session, prescription, and monitoring individual prescription is required: babies/infants/children specificities assessment and adjustment is needed regularly in small/growing children psychological preparation of the child and his family is needed, pain prevention is essential The first dialysis session is of importance to induce child and parent confidence, therefore appropriate preparation is needed. The site of the puncture of the fistula, most often with a double needle, size gauge 17, is carefully chosen and determined so that the needles are sufficiently separated to limit recirculation. Pain prevention is essential by application of a xylocaine ointment (Emla) one hour before needle insertion [39]. Psychological preparation of the child and family is also needed to limit “anxious stress” [22]. An aseptic procedure is essential. The extracorporeal circulation is adapted to the level of arterial aspiration pressure if measurable by the machine to prevent endothelial vascular trauma (not less than −150 mmHg). The venous return pressure should not be more than +200 mmHg to prevent endothelial vascular trauma. During the first dialysis session, the blood flow rate is maintained at a low level to prevent the dysequilibrium syndrome secondary to too efficient solute removal during this first session. Therefore, the blood flow rate should be approximately 3 mL kg−1 BW (or 90 mL m−2), or even less, so that urea clearance will be less than 3 mL min−1 kg−1 BW, which is usually well tolerated even in small children and limits the development of the dysequilibrium syndrome. The duration of the first dialysis session should be short, no more than 3 h, or adapted to the ultrafiltration need. The dysequilibrium syndrome is most often only symptomatic after one to two hours of dialysis, with variable symptoms such as headache or seizure, vomiting, fatigue, sleepiness, or a hypertensive tendency with a narrow range between systolic and diastolic pressure values. If needed, mannitol infusion (1 g kg−1 BW over 1 to 2 h during dialysis) is effective in preventing the syndrome. Symptoms usually disappear a few hours after the end of the dialysis. The extracorporeal blood flow rate, the duration of the session, and the number of sessions a week is progressively increased to individual patient need. Usually a blood flow rate of 150 to 200 mL min−1 m−2 and three sessions per week for 3 to 4 h per session achieve the minimum target prescription of 1.2 to 1.4 Kt/V [11]. The duration of a dialysis session is often prescribed to reach the anticipated dry weight at the end of the session. The total amount and the rate of ultrafiltration needed must be tolerable. A weight loss per hour of 1.5 to 2% of the BW is standard [3, 80] and most often well tolerated. Intermittent ultrafiltration with bicarbonate buffered dialysate which is not too warm (so called “cooled dialysate”), a normal “high” level of sodium (140 to 144 mmol L−1), which is not more than the normal concentration of sodium per liter plasma water, a normal hematocrit over 30% and optimally near 35% but not higher [78], and a mode of dialysis based on hemofiltration, i.e. (optimally HDF) are some of the major “tricks” used to enhance ultrafiltration tolerance [3, 48]. Intolerance of weight loss usually becomes symptomatic at the end of the dialysis session, when the patient is near his dry weight. Continuous blood volume monitoring during the session should become a real clinical support to enable optimum ultrafiltration tolerance (notion of crash hematocrit) [26, 56]. This information is limited to blood compartment changes. The interstitial space, which is mostly sodium-dependent, is better estimated by clinical assessment of edema or body weight [54, 55]. Rarely bed scales are used to assess more precisely the weight changes over a dialysis session. For most infants and children weighing less than 10 kg the need for more than three sessions a week may become evident to enable adapted nutrition, i.e. milk that is ”water”, hence 4 to 5 sessions a week are frequently prescribed [3, 11]. The adequate number and duration of each session should avoid partial fasting to achieve the weight needed to facilitate a short dialysis duration [7, 21, 80]. The volume of fluid used for extracorporeal blood replacement at the end of the session should be limited, and preferably a glucose solution instead of saline solution be used, especially in infants without residual renal function [3]. At the beginning of the dialysis session clinical manifestations of bioincompatibility may occur. This first use reaction is related to the biocompatibility of the material in the extracorporeal circuit, i.e. membrane, lines or even the needle either during the first session, first contact with the “extracorporeal” material or thereafter for example, in a new dialysis center for holidays. The major positive diagnostic criteria is the onset within 20 min of starting dialysis, of the major symptoms of dyspnea, burning heat throughout the body or access site, angioedema, flushing or vascular collapse, or with minor symptoms such as itching, rhinorrhea, lacrymation, urticaria, or abdominal cramping. Even if its occurrence is rare, or underestimated in the event of intermittent minor symptoms only during the first hour of session, the risk could be substantial. Biocompatible membranes, steam-sterilized material, adequate flushing of the circuit before blood connection, are some of the most important prevention factors [32]. The dialysis per se should be regarded as part of an overall strategy for care including dietary adequacy and interdialytic therapy [1, 21]. A weight gain over 10% dry BW during the interval of two sessions is often correlated with global non-compliance [3, 80]. In these cases, major outcomes could even occur: first acute, i.e. hyperkalemia or pulmonary edema, second chronic, i.e. hyperparathyroidism, and third long term, i.e. cardiovascular and coronary involvement [44, 45]. Conclusions Hemodialysis in children has benefited from major progress over the last 20 years. The morbidity of the sessions has decreased, even disappeared, seizures being exceptional, hypotensive episodes or headaches rare, and pain related to the fistula puncture effectively prevented by xylocaine ointment. The development of urea kinetic modeling enables calculation of the dialysis dose and indirect assessment of protein intake, nPCR. Even if the validity of these values is questioned their combined analysis provides an assessment and therefore is a “good thing”. The patient also benefits from the technological revolution. The newer machines enable precise control of ultrafiltration volumetric assessment and continuous blood volume monitoring during the session, buffered bicarbonate has become a standard technique, synthetic more biocompatible membranes and specific material available for babies/infants have been developed. Non invasive intervention, for example blood volume guided ultrafiltration have provided more adequate dialysis sessions and better dry weight assessment [81]. Last, the availability of erythropoietin [82] and of growth hormone and the promising results from enhanced dialysis dose on both growth and cardiac function [13, 20], all give the dialyzed child a real increased quality of life. In theory, reduction of dialysis prescription to only a urea dialysis dose achieved by three short (3-h) dialysis sessions, should be abandoned for long term dialyzed children and replaced by optimum dialysis obtained with longer (4 and more hours) and/or more frequent (daily: 5 to 6) sessions [13, 20, 79, 80]. But for such a daily dialysis strategy all the costs must be considered. On the one hand the financial cost cannot be neglected. For the patient bearing the burden, on the other hand, such an intensive dialysis prescription is acceptable only as an integrated therapy life project, a dialysis–transplantation program (HD, PD) with special regard for prevention of the vascular calcification [83]. Daily hemodialysis is one approach, perhaps the only one, to achieve phosphate purification [16, 17, 18, 19] and thereby maintain the calcium×phosphorus product in the optimum range of 3.3 to 4.4 mmol2 mL−2 [43].	[ "hemodialysis", "children", "guidelines" ]	[ "P", "P", "P" ]
Planta-4-1-2413075	Pepper pectin methylesterase inhibitor protein CaPMEI1 is required for antifungal activity, basal disease resistance and abiotic stress tolerance	Pectin is one of the main components of the plant cell wall that functions as the primary barrier against pathogens. Among the extracellular pectinolytic enzymes, pectin methylesterase (PME) demethylesterifies pectin, which is secreted into the cell wall in a highly methylesterified form. Here, we isolated and functionally characterized the pepper (Capsicum annuum L.) gene CaPMEI1, which encodes a pectin methylesterase inhibitor protein (PMEI), in pepper leaves infected by Xanthomonascampestris pv. vesicatoria (Xcv). CaPMEI1 transcripts are localized in the xylem of vascular bundles in leaf tissues, and pathogens and abiotic stresses can induce differential expression of this gene. Purified recombinant CaPMEI1 protein not only inhibits PME, but also exhibits antifungal activity against some plant pathogenic fungi. Virus-induced gene silencing of CaPMEI1 in pepper confers enhanced susceptibility to Xcv, accompanied by suppressed expression of some defense-related genes. Transgenic ArabidopsisCaPMEI1-overexpression lines exhibit enhanced resistance to Pseudomonas syringae pv. tomato, mannitol and methyl viologen, but not to the biotrophic pathogen Hyaloperonospora parasitica. Together, these results suggest that CaPMEI1, an antifungal protein, may be involved in basal disease resistance, as well as in drought and oxidative stress tolerance in plants. Introduction Plant cell wall, the first barrier of defense against invading pathogens, is composed of cellulose microfibrils cross-linked by hemicellulose, pectin and extensin. Pectin comprises a highly heterogeneous group of polymers that includes homogalacturonans and rhamnogalacturonans I and II (Willats et al.2001a, b). In pectin polymers, the galacturonic acid carboxyl groups can be methylesterified by a group of pectinases. These galacturonic acid methylesters are hydrolyzed by pectin methylesterase (PME) (Hagerman and Austin 1986; Pelloux et al. 2007). Many physiological processes, such as fruit maturation, microsporangenesis, seed germination and pollen growth, are affected by the degree to which PME changes the methlyesterification of galacturonic acid (Tieman and Handa 1994; Kagan-Zur et al. 1995; Ren and Kermode 2000; Bosch et al. 2005). The number and distribution of free and unesterified galacturonate carboxyl groups along the homogalacturonan chain has a great influence on the pectin properties and cell wall firmness (Willats et al. 2001a, b). Plant pathogens hydrolyze the cell wall components of plants using extracellular pectinolytic enzymes (Collmer and Keen 1986), and PME is found in many plant pathogenic bacteria and fungi (Asoufi et al. 2007). The black filamentous fungus Aspergillus niger secretes a set of pectin-degrading enzymes that include PME, polygalacturonase and pectin lyase, and these decompose the plant cell wall to establish infection and absorb nutrients from the host (de Vries and Visser 2001). In soft rot disease caused by Erwinia chrysanthemi, bacterial PME activity is induced during infection of Saintpaulia plants; however, PME-deficient mutants are noninvasive to the host cells (Boccara and Chatain 1989). PMEs have also been found in higher plants; they play significant roles in physiological processes and interactions with pathogens (Micheli 2001). In tobacco plants, host cell pectin methylesterases are required for the tobacco mosaic virus movement protein to transfer the viruses between host cells (Chen et al. 2000). Furthermore, PME-degraded polygalacturonans are associated with recognition of fungal pathogens (Wietholter et al. 2003). PME is also involved in symbiosis-specific functions (Lievens et al. 2002). For example, plant PME isoenzymes may undergo organism-specific post-translational processing for structural and functional integrity during interactions with various microorganisms (Micheli 2001). The PME enzyme activity is modulated specifically by inhibitor proteins such as the pectin methylesterase inhibitor (PMEI; Micheli 2001). Moreover, the PMEIs that inhibit demethylesterification of highly heterogeneous polymers (pectins) are the plant invertase inhibitor-related proteins, which are inhibitors of important metabolic enzymes (Koch 1996). Plant invertase inhibitor-related proteins play key roles in wounding, the plant defense reaction and developmental transitions (Raush and Greiner 2004), as well as during osmotic stress, senescence and seed development (Greiner et al. 1998, 1999). Investigation of gain- and loss-of-function mutants of tobacco cell wall invertase inhibitor (NtCIF) protein demonstrated that these inhibitor proteins play a role in seed development (Raush et al. 1998). Overexpression of the tobacco vacuolar invertase inhibitor protein (NtVIF) gene in transgenic potatoes is of potential use in the field of food technology (Greiner et al. 1998). However, there is little known about the in vivo functions of PMEI protein. Proteinaceous inhibitors have been purified from kiwi (Actinidia deliciosa) (Giovane et al. 2004), Arabidopsis (Wolf et al. 2003; Raiola et al. 2004), rice (Han et al. 2005) and the jelly fig (Ficus awkeotsang cv. Makino) (Jiang et al. 2001, 2002). The kiwi PMEI is specific for PME (Balestrieri et al. 1990) and is active against PMEs from several plants, including kiwi, orange, apple, tomato, apricot, carrot, potato and banana (Ly-Nguyen et al. 2004). Four Cys residues conserved in several isoforms of PMEI are involved in the formation of disulfide bridges (Camardella et al. 2000). PME and PMEI form a stoichiometric 1:1 complex, in which the interaction between the PME and the inhibitor occurs in close proximity to the putative active site (Di Matteo et al. 2005). Since PME activity can be modulated by pH, the stability of the PME–PMEI complex is also affected by pH (Denès et al. 2000). Crystallographic work has revealed that an α-helical hairpin motif plays a structurally important role in PMEI activation (Hothorn et al. 2004). Many cDNAs encoding PMEIs have been isolated and functionally characterized from plants (Rausch and Greiner 2004); however, their role in plant defense remains relatively unknown. To date, the functional analyses of genes associated with defense responses in plants have utilized reverse-genetics approaches based on loss-of-function via double-stranded RNA interference (Robertson 2004) or gain-of-function via transgenic gene expression (Clough and Bent 1998). Virus-induced gene silencing (VIGS) has been proven to be a useful method for assessing the function of target genes in Solanum species (Brigneti et al. 2004). In particular, VIGS studies have been used to investigate disease resistance signaling and defense-related genes such as SGT1 (Liu et al. 2002c; Peart et al. 2002b), EDS1 (Liu et al. 2002b; Peart et al. 2002a) and NPR1/NIM1 (Liu et al. 2002b) in Nicotiana benthamiana. AtPGIP1 is among the genes encoding pectic enzyme-related proteins; it encodes polygalacturonase (PG)-inhibiting protein (PGIP), and an antisense AtPGIP1 gene was used to silence its expression in transgenic Arabidopsis plants (Ferrari et al. 2006). Previous studies have revealed that overexpression of two closely related genes, AtPGIP1 and AtPGIP2, conferred resistance against Botrytis cinerea infection (Ferrari et al. 2003). Silencing of AtPGIP1 resulted in enhanced susceptibility to infection, as well as reduced activity of PGIP (Ferrari 2006). However, gene-silencing techniques such as VIGS and antisense RNA have not yet been used to investigate the role played in plant defense by other PMEI-encoding genes. Here, we used a macroarray technique to isolate and functionally characterize a pectin methylesterase inhibitor gene, CaPMEI1, from a cDNA library of pepper (Capsicum annuum L.) leaves infected with Xanthomonas campestris pv. vesicatoria (Xcv; Jung and Hwang 2000). Local and systemic induction of CaPMEI1 was investigated in pepper plants following inoculation with pathogenic and non-pathogenic bacteria. We also examined the involvement of CaPMEI1 in defense-related signal transduction cascades via exogenous application of abiotic elicitors to pepper plants. Recombinant CaPMEI1 proteins were expressed in E. coli and exhibited antifungal activity against plant pathogenic fungi. Since it is difficult to transform pepper plants, we performed gene silencing and CaPMEI1 overexpression in pepper and Arabidopsis, respectively, to identify the cellular functions of the CaPMEI1 gene. The functional data obtained by VIGS and transgenic ectopic expression of CaPMEI1 suggest that this pepper pectin methylesterase inhibitor is involved in plant defense and abiotic stress responses. Materials and methods Plant materials and growth conditions Pepper (Capsicum annuum L. cv. Nockwang) plants were grown at 28°C under a 16 h day at 70 μmol photons m−2 s−1. Plants were seeded into a plastic tray (55 × 35 × 15 cm3) containing steam-sterilized soil mix (peat moss, perlite and vermiculite; 5:3:2, v/v/v) and loam soil (1:1, v/v). At the two-leaf stage, seedlings were transplanted into plastic pots (5 × 15 × 10 cm3) containing the soil mix previously described. Arabidopsisthaliana (ecotype Columbia) plants were grown in pots containing vermiculite, peat moss and perlite (1:1:0.5, v/v/v) in a growth chamber under a 12 h light/12 h dark photoperiod (130 μmol photons m−2 s−1) at 24°C and 60% relative humidity. Prior to sowing, the seeds were surface-sterilized using 1% sodium hypochlorite and vernalized at 4°C for 3 days to break dormancy. Pathogens, inoculation procedures, disease rating and tissue staining Xanthomonas campestris pv. vesicatoria (Xcv) strains Ds1 and Bv5-4a were used in this study. Bacteria were cultured overnight in yeast-nutrient (YN) broth (5 g L−1 yeast extract, 8 g L−1 nutrient broth) at 28°C. Prior to inoculation, bacteria were harvested by centrifugation and resuspended in sterile tap water (108 cfu mL−1). Pepper plants were inoculated at the six-leaf stage by infiltrating the bacterial suspension into the abaxial side of fully expanded leaves using an atomizer. Infected plants were then incubated for 16 h at 28°C in a moist chamber with 100% relative humidity. Bacteria-infected leaves were sampled at various time intervals after inoculation. To evaluate systemic induction in the upper leaves, bacterial suspensions (108 cfu ml−1) were infiltrated into the lower leaves of pepper plants at the two-leaf stage using a needless syringe. The bacterial strains used for this study included: virulent and avirulent strains of X. campestris pv. vesicatoria (Ds1 and Bv5-4a); a non-pathogenic strain (Pseudomonasfluorescens ATCC13525); and Escherichiacoli JM109. Leaves of 6-week-old Arabidopsis plants were infiltrated with a suspension (OD600 = 0.001) of virulent Pseudomonas syringae pv. tomato strain DC3000 (Pst DC3000). The bacteria were cultured overnight at 28°C and suspended in 10 mM MgCl2. To determine bacterial growth, leaf discs were cut from infected leaves at different time intervals after inoculation. Bacterial growth was monitored by performing serial dilutions onto KB agar containing 100 μg mL−1 rifampicin. Each experiment was replicated three times. Hyaloperonospora parasitica isolate Noco2 was propagated by weekly subculturing on 7- to 10-day-old Arabidopsis seedlings. The 7-day-old seedlings were inoculated with an H. parasitica asexual inoculum (5 × 104 conidiosporangia mL−1). The seedlings inoculated with H. parasitica were covered with a plastic dome to maintain a high relative humidity (80–100%) and grown in a growth chamber at 17°C. Seven days after inoculation, disease rating was scored for more than 50 plants per treatment. A visual disease rating was expressed as the number of sporangiophores on each cotyledon and was divided into five classes: 0–5, 6–10, 11–15, 16–20 and >20 sporangiophores per cotyledon. The cotyledons from inoculated plants were stained with lactophenol-trypan blue (10 mL lactic acid, 10 mL glycerol, 10 g phenol and 10 mg trypan blue, dissolved in 10 mL of distilled water) to assess H. parasitica growth. At 2–5 days after inoculation, the infected cotyledons were boiled for 5 min in the staining solution and de-stained overnight in chloral hydrate (2.5 g chloral hydrate dissolved in 1 mL distilled water). The destained cotyledons were subsequently mounted in 70% glycerol for microscopic observation. Isolation and sequence analysis of pathogen-induced cDNAs To construct a pathogen-induced cDNA library, pepper leaves were inoculated with the avirulent strain X. campestris pv. vesicatoria Bv5-4a. The pathogen-induced cDNA library was constructed using 5 μg poly(A)+ mRNA extracted from inoculated pepper leaves (Kim and Hwang 2000). To isolate pathogen-inducible cDNAs from the pepper cDNA library, we performed differential hybridization, as described previously by Jung and Hwang (2000). Digoxigenin (DIG)-labeled, single-stranded cDNA probes were generated from total RNA of healthy and Bv5-4a-infected leaves using RT-PCR. Nylon membranes were pre-hybridized at 65°C for 3 h in 5× SSC, 0.1% sodium lauroylsarcosine, 0.02% SDS and 1% blocking reagent (Boehringer Mannheim, Mannheim, Germany). Hybridization was then performed overnight at 65°C in the same buffer with single-stranded cDNA probes. Hybridized membranes were rinsed twice for 5 min with 2× SSC and 0.1% SDS at room temperature, and twice for 10 min with 0.1× SSC and 0.1% SDS at 65°C. The hybridization signals were detected according to the manufacturer’s protocol (Boehringer Mannheim). We selected cDNA clones that were expressed strongly in pathogen-infected leaves, compared with those of healthy leaves. Clones were sequenced with an ABI 310 DNA sequencer (Applied Biosystems, Foster City, CA, USA) using the PRISM Big Dye™ Terminator Cycle Sequencing Ready Reaction Kit (PE Biosystems, Foster City, CA, USA). Sequencing results were analyzed using BLAST (National Center for Biotechnology Information; Altschul et al. 1997). Treatment with abiotic elicitors and environmental stresses The leaves of pepper plants at the six-leaf stage were sprayed with 5 mM salicylic acid (SA), 100 μM methyl jasmonate (MeJA) or 100 μM absicisic acid (ABA). Pepper plants treated with methyl jasmonate were sealed tightly in plastic bags. For ethylene treatment, whole plants were removed from soil, and then placed in a water-containing glass chamber, followed by injection of ethylene gas (5 μL L−1). For cold stress treatment, plants were placed at 4°C in a cold room. For wounding stress, the leaves were pricked with needles. To impose drought stress, the plants were removed from the soil and then incubated at room temperature without water. H2O2 treatment was performed by spraying leaves with 100 mM H2O2 solution. Leaves treated with various elicitors and abiotic stresses were removed from the plants, frozen in liquid nitrogen and stored at −70°C until used for RNA isolation. RNA isolation and RNA gel blot analysis Total RNA was extracted from pepper leaves, stems, roots, flowers and fruits using the guanidine isothiocyanate method (Chomczynski and Sacchi 1987). Frozen tissues (1 g) were ground to a powder and homogenized in 10 mL extraction buffer (4 M guanidine isothiocyanate, 25 mM sodium citrate [pH 7.0], 0.55% [w/v] N-laurylsarcosine and 0.1 M 2-mercaptoethanol). A mixture of 2 M sodium acetate (pH 4.0), water-saturated phenol and chloroform–isoamylalcohol (24:1) was added to the homogenate, followed by precipitation. Total RNA (20 μg) was separated by 1.2% formaldehyde-agarose gel electrophoresis and then blotted onto Hybond-N+ membranes (Amersham, Buckinghamshire, UK). Transferred RNA was fixed to the membrane using UV cross-linking. The 3′ UTR region of CaPMEI1 was amplified for use in the generation of a gene-specific probe. The primers used for amplification of the CaPMEI1 gene-specific region were 5′-CATGGGTAAGTGCTGCCTTGACGGAC-3′ and 5′-GTTAACAAATGCATA TGGAACATTT-3′. The CaBPR1 coding region was amplified with the primers 5′-ATGGGACACTCTAATATTGCC-3′ and 5′-GACATCAGTTGGAAGTTCCAA-3′. The CaSAR82A coding region was amplified using the primers 5′-ATGGTTTCCAAAAGT AGTATTTTTATTT-3′ and 5′-TATGCTTAACAATTATTACTGAATA TAATC-3′. PCR-amplified products were 32P-labeled using a random priming kit (Boehringer, Mannheim). Hybridization was performed overnight at 65°C in 5% dextran sulfate, 0.25 M disodium phosphate (pH 7.2), 7% (w/v) sodium dodecyl sulfate (SDS) and 1 mM EDTA. Following hybridization, the membranes were washed twice for 10 min with 2× SSC and 0.1% SDS at room temperature, and then twice for 5 min with 0.1× SSC and 0.1% SDS at 65°C. Equal loading of RNA was confirmed by ethidium bromide-staining of ribosomal RNA. In situ RNA localization In situ RNA localization was performed as described previously (Lee et al. 2000). Leaf tissue was fixed for 2 h in a solution of 1× phosphate-buffered saline (PBS), 4% paraformaldehyde and 1 μL mL−1 Triton X-100 by vacuum infiltration for 10 min and shaking for 2 h at room temperature. The fixed samples were washed with 1× PBS, dehydrated through a graded ethanol and xylol series, and then embedded in liquid paraplast at 57°C (Sherwood Medical, St. Louis, MO, USA). Paraplast-embedded sections (10 μm in thickness) were placed on glass slides coated with poly-l-lysine (Sigma, St. Louis, MO, USA) and incubated at 42°C. Sections were deparafinated using xylene and 1-propanol, followed by rehydration with serial dilutions of ethanol. Tissue samples were treated with 0.01 M Tris–HCl (pH 8.0) and 1% bovine serum albumin (BSA) for 10 min, followed by incubation in 100 mM Tris–HCl solution (pH 8.0) containing proteinase K (5 mg mL−1) and 50 mM EDTA, for 30 min at 37°C. Sections were treated with 0.25% acetic anhydride in 100 mM triethanolamine (pH 8.0) for 10 min at room temperature to inhibit non-specific signals. Digoxigenin (DIG)-labeled probes were prepared using the Dig High Prime DNA Labeling and Detection kit, according to the manufacturer’s instructions (Boehringer Mannheim). Sections were prehybridized and hybridized at 42°C in 50% formamide, 4× SSC, 150 μg mL−1 tRNA and 0.5% blocking reagent (Boehringer Mannheim). After hybridization, the sections were washed twice with 50% formamide and 4× SSC at 42°C, twice with 4× SSC and then once with diethyl pyrocarbonate (DEPC)-treated water. The DIG signal was detected, according to the manufacturer’s instructions (Boehringer Mannheim). Color reactions were developed overnight at 37°C with nitro-blue tetrazolium chloride (NBT) and 5-bromo-4-chloro-3-indolyl phosphate (BCIP), and reactions were stopped with TE buffer (10 mM Tris–HCl, 1 mM EDTA, pH 8.0). Sections were photographed with Kodak ISO 100 film under a bright field Olympus BH-2 microscope (Olympus, Tokyo, Japan). To demonstrate the specificity of in situ hybridization, control hybridizations were performed without DIG-labeled probes. Purification of recombinant CaPMEI1 protein The CaPMEI1 coding region, including the stop codon, was amplified by PCR using the forward and reverse primers 5′-GAATTCATGGAAGGTGGCAATTTTCT-3′ and 5′-CTCGAGTAGCCGTGAAGGGCAGCCAGACG-3′, respectively. Amplification products were cloned into pCR2.1-TOPO (Invitrogen, Carlsbad, CA, USA), which was digested with EcoRI and XhoI and ligated into the similarly digested pET32a (Novagen, Madison, WI, USA). Escherichia coli BL21 (DE3) pLysS (Novagen), which is defective for thioredoxin reductase, was used as a host for recombinant protein expression. Cultures were started from single colonies, grown in LB broth at 37°C, and then diluted 1:100 at OD600 = 0.6. After dilution, bacteria were grown to a density of OD600 = 0.6, induced with 10 mM IPTG and grown for a further 5 h at 37°C. Cells were harvested by centrifugation for 15 min at 5,000g, extracted with denaturation buffer (8 M urea, 20 mM sodium phosphate buffer [pH 7.8], 500 mM NaCl) and disrupted by sonication. Following centrifugation at 5,000g for 15 min at 4°C, the supernatant was loaded onto a 1.5 mL column of Ni-NTA resin (Qiagen, Hilden, Germany), which was washed with an initial denaturing wash buffer (8 M urea, 20 mM sodium phosphate buffer [pH 6.0], 500 mM NaCl), followed by a second denaturing wash buffer (8 M urea, 20 mM sodium phosphate buffer [pH 5.3], 500 mM NaCl). Bound fusion protein was eluted with a final denaturing elution buffer (8 M urea, 20 mM sodium phosphate buffer [pH 4.0], 500 mM NaCl). Recombinant proteins were dialyzed against a buffer containing 10 mM Tris (pH 8.0) and 0.1% Triton X-100, according to the manufacturer’s protocol. As a control, thioredoxin was purified using a native buffer that did not contain urea. The purified CaPMEI1 recombinant protein was digested for 16 h at room temperature with recombinant enterokinase (1 U/5 μg recombinant protein; Novagen). SDS polyacrylamide gel electrophoresis CaPMEI1 was dissolved in 1× SDS sample buffer (0.9 g glycerol, 5% SDS, 1% bromophenol blue, 0.1 mL mercaptoethanol and 1 L H2O), separated by 12% SDS-PAGE as described previously (Laemmli 1970), and stained with Coomassie Brilliant Blue R-250. Molecular weights were estimated using 6.5 to 200.5 kDa marker proteins (Bio-Rad, Hercules, CA, USA). Pectin methylesterase enzyme inhibitor assay The inhibitory effect of CaPMEI1 on the enzymatic activity of pectin methylesterase (PME) was assayed under standard conditions (Grsic-Rausch and Rausch 2004). The reaction mixture comprised 894 μL 0.4 mM NAD in 50 mM phosphate buffer (pH 7.5), 80 μL 5% (w/v) pectin (Sigma) in H2O, 8 μL formaldehyde dehydrogenase (0.35 U; Sigma) and 8 μL alcohol oxidase (1.0 U; Sigma). After mixing, the reaction was started with the addition of 10 μL (7.8 mU) PME from orange peel (Sigma). To analyze CaPMEI1 inhibition of PME, 1 μL inhibitor solution (0.5 mg CaPMEI1 fusion protein, 10 mM Tris buffer [pH 7.5], 0.1 M NaCl) was mixed with 10 μL PME (7.8 mU) and 0.5 μL 3 M K-acetate buffer (pH 5.3), followed by incubation for 15 min at room temperature. All reaction temperatures were maintained at 25°C. Reaction rates were recorded continuously at 340 nm using a DU650 UV-visible spectrophotometer (Beckman, Fullerton, CA, USA). PME–PMEI interactions were determined by measuring the rate of NADH formation per minute at pH 7.5 and 25°C. In vitro antifungal activity To determine the antifungal activity of CaPMEI1, we examined its effect on plant pathogenic fungi. Fusarium oxysporum f.sp. matthiole and Alternaria brassicicola were incubated on potato dextrose agar (PDA) at 28°C for 1–2 weeks; B. cinerea was incubated at 20°C for 1–2 weeks. Fungi were grown in 48-well plates (Cell Wells™, Corning Glass Works, Corning, NY, USA) containing sterile 4× potato dextrose broth (PDB, 100 μL) and a range of concentrations of purified CaPMEI1 protein (0–500 μg mL−1). Spore suspensions (104 spores mL−1) were prepared and 100 μL of inoculum was added to each microwell. Plates containing B. cinerea were incubated at 20°C; the other fungi were incubated at 28°C for 4–7 days. To determine the inhibitory effect of CaPMEI1 on spore germination and hyphal growth, CaPMEI1 (0.1–500 μg mL−1) or the thioredoxin control were added to F. oxysporum f.sp matthiolae spore suspensions (105 spores mL−1), placed on glass slides and incubated for 12 h at 28°C. After 6 h incubation, 100 germinated spores were examined using a haemocytometer and the lengths of 50 individual hyphae were determined. The experiment was repeated three times. Plasmid construction and plant transformation The CaPMEI1 coding region was PCR-amplified without stop codon using the following primers to generate XbaI and BamHI sites: 5′-TCTAGAATGGAAGGTGGCA ATTTTCTCACA-3′ and 5′-GGATCCGCCGTGAAGGGCAGCCAGACGGT -3′. The fragment was inserted into pCR2.1-TOPO (Invitrogen) and digested with XbaI and BamHI. The construct was confirmed by sequencing. To generate the reporter construct p35S-CaPMEI1-GFP, the fragment was inserted into the XbaI–BamHI sites of the binary vector p35S-smGFP. p35S-smGFP was generated by fusing the gene encoding smGFP from the vector 326-GFP into pBIN35S digested with HindIII and EcoRI. The CaMV35S promoter-CaPMEI1-smGFP construct was introduced into the Agrobacterium tumefaciens strain EHA105 using electroporation. Transgenic Arabidopsis plants were generated by floral-dipping wild-type (Col-0) plants into an A. tumefaciens culture containing the appropriate construct (Clough and Bent 1998). Agrobacterium was grown at 28°C and 250 rpm in YEP medium (10 g Bacto peptone, 10 g yeast extract, 5 g NaCl) supplemented with kanamycin (25 μg ml−1). Cells were harvested by centrifugation for 20 min at 5,500g and resuspended in inoculation media containing 5.0% sucrose and 0.05% Silwet L-77 (OSi Specialties, Inc., Danbury, CT, USA) to OD600 = 0.8. Arabidopsis plants with 2–10 cm long bolts were inoculated by dipping into an Agrobacterium suspension (OD600 = 0.8) and then left in the dark overnight, prior to return to a growth chamber. Seeds from the transformed Arabidopsis plants were collected and screened for kanamycin resistance. Virus-induced gene silencing (VIGS) The TRV-based VIGS system was used for silencing of CaPMEI1 in pepper plants, as described previously (Liu et al. 2002a; Chung et al. 2004). The pepper CaPMEI1 coding region was cloned into pTRV2 to generate the construct pTRV2:CaPMEI1. The fully extended cotyledons of 2-week-old pepper seedlings were co-infiltrated with A. tumefaciens GV3101 carrying pTRV1 or pTRV2:CaPMEI1 (OD600 = 0.2 for each construct). Plants were placed in a growth room at 25°C under a 16 h light/8 h dark photoperiod, to allow for growth and viral spread. Experiments were performed 5–6 weeks after the induction of silencing. RT-PCR analysis The RT reactions (20 μL) were performed at 42°C with total RNA (2 μg), oligo p(dT)15 primer (Roche, Mannheim, Germany) and AMV reverse transcriptase (Roche). Aliquots (1 μL) of the RT reaction products were used for RT-PCR analysis with the following gene-specific primers: 5′-CAGGATGCAACACTCTGGTGG-3′ and 5′-ATCAAAGGC CGGTTGGTC-3′ for CaBPR1 (accession no. AF053343); 5′-TGTCGAAGGTGGTCC AATAAA-3′ and 5′-TAGACAGAAGGATTGGCGAGG-3′ for CaPR10 (accession no. AF244121); 5′-ATCTGTACCAGCTTGCACGTGT-3′ and 5′-CCCTCACTGTGGCCT TGG-3′ for CaPOA1 (accession no. AF442387); and 5′-CAGGGAGATGAATTCTGA GGC-3′ and 5′-CATATGAACCTCTATGGATTTCTG-3′ for CaSAR82A (accession no. AF313766). RT-PCR conditions were 95°C for 10 min and 30 cycles of 95°C for 30 s, 52°C for 30 s, and 72°C for 1 min and 30 s. Single bands for PCR products were confirmed on an agarose gel. Drought stress treatment and evaluation For germination tests, seeds from Arabidopsis wild-type and CaPMEI1-OX transgenic lines were surface-sterilized and placed on MS media (Murashige and Skoog 1962) containing 200 or 600 mM mannitol. The seeds were maintained at 4°C for 48 h under dark conditions to synchronize germination, and then transferred to a growth chamber. Experiments were repeated at least three times, using approximately 100 seeds. Surface-sterilized seeds were also used for examination of relative root length. Seedlings were grown on plates placed vertically in growth chambers and root length was estimated for 14 days. To assess mannitol tolerance, seedlings were grown for 7 days in 1× MS agar media supplemented with 1% sucrose, and then transferred to wells containing 1× MS liquid media supplemented with 100, 200, 300 or 400 mM mannitol. For drought treatment, 4-week-old soil-grown plants were deprived of water for 15 days and then re-watered on day 16. To minimize experimental variation, both the transgenic and control plants were grown in the same tray. Experiments were repeated at least three times. For transpiration rate measurements, leaves of 4-week-old plants were detached and maintained at room temperature. Leaf weight was determined every 20 min for 2 h, and then every 1 h thereafter. Each measurement was performed using eight leaves. Experiments were repeated at least three times with similar results. Oxidative stress treatment and evaluation For analysis of oxidative stress tolerance, surface-sterilized seeds from Arabidopsis wild-type, vector control and CaPMEI1-OX transgenic lines were placed on MS medium containing different concentrations of methyl viologen (MV; Sigma). After synchronized germination, seeds were maintained in a growth chamber. To investigate whether or not MV treatment caused retardation of seedling development, 7-day-old seedlings were randomly transferred into well-plates containing 1× MS liquid medium supplemented with 0.05, 0.1 or 0.5 μM MV, and then grown for a further 2 weeks. To determine leaf senescence during MV treatment, fully expanded leaves were detached from 4-week-old plants and rinsed briefly with 70% ethanol. The detached leaves were floated on MS medium containing MV for 24 h following extraction with liquid nitrogen. Their chlorophyll content was measured spectrophotometrically according to the formula (Ca + b = 5.24 × A664 + 22.24 × A648), where C is the chlorophyll concentration (Ca + b) in micrograms per mL and A is absorption (Lichtenthaler 1987). Results Isolation and sequence analysis of CaPMEI1 cDNA in pepper To identify the molecular mechanisms involved in plant defense against the invasion of microbial pathogens, we performed a differential hybridization screening of a cDNA library constructed from pepper leaves undergoing an incompatible interaction with Xcv (Jung and Hwang 2000). PR-1, PR-10, SAR8.2 and Myb transcription factor were among the pathogen-induced genes isolated by this screen. In addition, we isolated a full-length cDNA (834 bp) designated CaPMEI1 (Capsicum annuumpectin methylesterase inhibitor protein, accession no. AF477956), which contains a 54 bp 5′ untranslated sequence and a 174 bp 3′ untranslated region. The predicted open reading frame encodes a full-length protein of 200 amino acid residues, with an estimated molecular mass of 21.39 kDa and an isoelectric point (pI) of 6.51. BLAST searches were performed between the predicted amino acid sequence of CaPMEI1S (accession no. AF477956) from C. annuum and the EMBL/GenBank database. These sequence analyses revealed that CaPMEI1 contains a conserved pectin methylesterase inhibitor (PMEI) domain of 156 amino acids (Fig. 1). Four cysteine residues were conserved in all the aligned proteins (Fig. 1), and these are predicted to be involved in the formation of disulfide bridges (Camardella et al. 2000). The predicted amino acid sequence of CaPMEI1 is 80% identical to that of Nicotiana tabacum DC1.2 protein (accession no. BAA95794) and has 50% identity with Arabidopsis thaliana ripening-related protein (accession no. BAA97200), 36% with Pinus radiata pectinesterase homolog (accession no. T08112) and 24% with A. thaliana invertase homolog (accession no. NP_201267). Fig. 1Amino acid sequence alignments of pepper CaPMEI1 with Nicotiana tabacum DC1.2 protein (accession no. BAA95794), A. thaliana ripening-related protein (accession no. BAA97200), A. thaliana invertase homolog (accession no. NP201267) and Pinus radiata pectinesterase homolog (accession no. T08112). The boxed amino acid sequences represent the pectin methylesterase inhibitor (PMEI) domain. The conserved cysteine residues are marked by asterisks and the shaded regions represent conserved amino acid residues Organ-specific expression of CaPMEI1 in pepper tissues RNA gel blot analysis was performed to examine the organ-specific expression of CaPMEI1 in pepper plants. Although high levels of CaPMEI1 transcription were observed in stem tissues, only relatively low levels of expression were found in leaf, root, flower and green fruit (Fig. 2a). CaPMEI1 transcripts were not detected in red fruit. Fig. 2RNA gel blot analysis of expression of CaPMEI1, CaSAR82A and CaBPR1 in pepper plants. Membranes were hybridized with probes from the 3′ UTR region of pepper CaPMEI1 cDNA or full-length CaBPR1 cDNA. Equal loading (20 μg) was verified by visualizing RNA on a gel stained with ethidium bromide. H healthy, M mock-inoculated or mock-treated. aCaPMEI1 expression in various organs of pepper plants. b Expression of CaPMEI1 and CaBPR1 in pepper leaves at various time intervals after inoculation with C. coccodes or virulent strain Ds1 (compatible interactions with pepper: susceptible response) and avirulent strain Bv5-4a (incompatible interactions with pepper: resistant response) of Xcv.cCaPMEI1 expression in lower (local) infected and upper (systemic) uninfected leaves at various time intervals after inoculation with the virulent and avirulent strains Ds1 and Bv5-4a of Xcv, P.fluorescence ATCC13525 and E.coli JM109. The lower leaves of pepper plants were inoculated at the 6-leaf stage. For the mock-inoculation, the lower leaves were infiltrated with 10 mM MgSO4. d Expression of CaPMEI1 and CABPR1 in pepper leaves at various time intervals after treatment with salicylic acid (SA, 5 mM), methyl jasmonate (MeJA, 100 μM), ethylene (5 μl L−1) and abscisic acid (100 μM). e Expression of CaPMEI1 and CaSAR82A in pepper leaves at various time intervals after treatment with drought, wounding, cold and H2O2 CaPMEI1 induction by pathogen infection To assess the accumulation of CaPMEI1 transcripts in pepper leaves during the compatible and incompatible interactions with pathogens, leaves were inoculated with C. coccodes and Xcv Ds1 (virulent, compatible), as well as Xcv Bv5-4a (avirulent, incompatible). CaPMEI1 transcripts were not detected in leaves mock-inoculated with sterilized water (Fig. 2b, c). During fungal infection at the four- and eight-leaf stages, CaPMEI1 expression was induced at 48 and 12–48 h after inoculation, respectively, suggesting that the adult plant stage is more resistant to C. coccodes (Fig. 2b, upper panel). CaPMEI1 transcripts increased rapidly within 30 min after infection by both the virulent (Ds1) and avirulent (Bv5-4a) strains of Xcv (Fig. 2b, lower panel). High transcript levels accumulated by 18 h after infection with the virulent strain (Ds1) that causes a susceptible response; however, they started to decline rapidly by this time following infection with the avirulent strain (Bv5-4a) that causes a resistant response. CaBPR1 (Capsicum annuum basic PR1) expression was observed to confirm the success of individual inoculations. CaBPR1 transcripts were detected at 6 h after inoculation with the virulent and avirulent strains of Xcv and increased gradually during the 24 h following inoculation (Fig. 2b). Local and systemic expression of CaPMEI1 in infected pepper plants The local and systemic CaPMEI1 expression was analyzed in pepper leaves inoculated with Xcv and the non-pathogenic bacteria Pseudomonas fluorescens ATCC13525 and E. coli JM109 (Fig. 2c). In leaves infected with the virulent (Ds1) and avirulent (Bv5-4a) strains of Xcv, CaPMEI1 transcripts were detected 3 h after inoculation and levels decreased thereafter. Infection of lower (local) leaves by either Xcv strain induced systemic accumulation of CaPMEI1 transcripts in the upper leaves. The systemic induction by the avirulent strain (Bv5-4a) infection was stronger than that by the virulent strain (Ds1) infection. Local or systemic induction of CaPMEI1 transcripts was not observed 30 h after infection. In response to the non-pathogenic bacteria P. fluorescens and E. coli, CaPMEI1 transcripts were detected 3 h after inoculation and decreased subsequently in both local and systemic leaves. CaPMEI1 induction following treatment with plant hormones Plant hormones are involved in defense-related signal transduction pathways (Hammond-Kosack and Jones 1996; Chung et al. 2007). We examined whether or not CaPMEI1 expression is induced by plant hormones using pepper leaves treated with salicylic acid (SA), ethylene, methyl jasmonate (MeJA) or abscisic acid (ABA) (Fig. 2d). Salicylic acid strongly induced expression of CaPMEI1 at 1 and 2 h, as well as between 18 and 24 h after treatment. Following treatment with ethylene and MeJA, CaPMEI1 expression was strongly induced for 24 h. In addition, strong CaPMEI1 induction was observed 1 h after treatment with ABA. In comparison, SA only induced CaBPR1 transcription 18–24 h after treatment (Fig. 2d), although transcripts were detected 6 h after treatment with ethylene and MeJA. In response to ethylene treatment, CaBPR1 transcripts increased rapidly between 18 and 24 h after treatment, but increased gradually between 6 and 24 h after MeJA treatment. CaBPR1 expression was induced between 2 and 12 h after ABA treatment. CaPMEI1 induction by abiotic stresses Abiotic stresses resulted in differential induction of CaPMEI1 in pepper leaves (Fig. 2e). The CaPMEI1 gene was responsive to mechanical injury, and expression was observed between 15 min and 6 h after wounding, although no transcription was detected thereafter. CaPMEI1 transcription also increased between 2 and 18 h after cold treatment. CaPMEI1 transcription was induced at 1 h of drought stress, but was undetectable thereafter. In response to H2O2 treatment, CaPMEI1 transcript levels increased for the first 6 h and declined thereafter. The marker genes CaBPR1 and CaSAR82A were used for comparison, as they also exhibit differential induction by these abiotic stresses. In situ localization of CaPMEI1 transcripts We performed an in situ hybridization using a CaPMEI1 probe to examine spatial expression of CaPMEI1 in pepper leaf tissues (Fig. 3). No hybridization signals were observed in non-treated leaves (Fig. 3a, b). Intense localization of CaPMEI1 transcripts was observed in the xylem area of the vascular bundle in non-treated stems (Fig. 3c). CaPMEI1 transcripts were detected in the vascular bundle of leaves inoculated with C. coccodes (Fig. 3e) and in the vascular bundles and upper epidermis of leaves treated with ethylene (Fig. 3g). No transcript signals were observed in any pepper tissues hybridized with the CaPMEI1 sense DIG-labeled RNA probe (Fig. 3b, d–f). These in situ hybridization results are supported by the temporal expression patterns observed in the RNA gel blot analysis of CaPMEI1 expression (Fig. 2). Fig. 3In situ localization of CaPMEI1 transcripts in pepper leaf and stem tissues. Cross sections of leaf tissues were hybridized with CaPMEI1 antisense (a, c, e, g) and sense (b, d, f, h) DIG-labeled RNA probes, and then photographed under bright-field conditions. The transcript signal is purple. a, b Untreated leaf tissues. c, d Untreated stems. e, f Leaf tissues at 24 h after inoculation with C. coccodes. g, h Leaf tissues treated with 5 μl L−1 ethylene. P phloem, X xylem, UE upper epidermis, LE lower epidermis, Vs vascular bundle, C cortical cell Inhibition of pectin methylesterase (PME) activity by CaPMEI1 protein To prepare recombinant CaPMEI1 protein, the CaPMEI1 coding region was PCR-amplified and cloned into pET32a. E. coli strain BL21 (DE3) pLysS was used as a host for the recombinant expression of the CaPMEI1 construct and an empty vector control, yielding recombinant CaPMEI1 and thioredoxin, respectively. The purified CaPMEI1 (Fig. 4a, lane 5) and thioredoxin (Fig. 4a, lane 3) were examined by SDS-PAGE analysis. The purified thioredoxin–CaPMEI1 fusion protein formed a single band with an apparent molecular mass of 42 kDa (Fig. 4a, lane 5). Following cleavage of thioredoxin by enterokinase digestion, the purified CaPMEI1 protein (Fig. 4a, lane 6) was used for biological function determination. The cleaved recombinant CaPMEI1 protein was found exclusively in the insoluble fraction as inclusion bodies (data not shown). After attempting unsuccessfully to renature the cleaved protein using Triton X-100 as a detergent (Kim and Hwang 1994), we performed the PME inhibition assay using crude and purified CaPMEI1 fusion proteins. Fig. 4Inhibition of pectin methylesterase (PME) activity by the pepper pectin methylesterase inhibitor (CaPMEI). a Recombinant CaPMEI1 expression in E. coli BL21 (DE3) pLysS. Cells were grown in LB media, and recombinant protein expression was induced with 1 mM IPTG. M, molecular marker (kDa); Lane 1, Uninduced E. coli BL21 cell extracts; Lane 2, Soluble fraction of E. coli BL21 cell extracts encoding thioredoxin protein after IPTG induction; Lane 3, Purified thioredoxin; Lane 4, Crude protein extracts of E. coli cells producing the thioredoxin–CaPMEI1 fusion protein after IPTG induction; Lane 5, Purified thioredoxin–CaPMEI1 fusion protein; Lane 6, Cleaved CaPMEI1 and thioredoxin proteins following enterokinase digestion. Protein staining was performed using Coomassie brilliant blue. b Inhibition of PME activity following treatment with CaPMEI1. For the inhibition assay, crude CaPMEI1 (0.5 μg) was mixed with 7.8 mU of orange peel PME in a total volume of 11 μL, and then preincubated at 25°C for 15 min, followed by addition to the reaction solution CaPMEI1 contains a pectin methylesterase inhibitor (PMEI) domain (Fig. 1) and shares similarity with a plant pectin methylesterase inhibitor related to regulation of pectin degradation (Raiola et al. 2004). In addition, PMEIs have been found to play crucial roles in plant and microbial enzyme regulation of pectin degradation by de-esterification (Raiola et al. 2004). To determine whether or not orange peel PME activity is inhibited in the presence of crude CaPMEI1 fusion protein, we used the PME enzyme assay to measure the effect of PME inhibitor on the formation of NADH (Fig. 4b). Both crude and purified CaPMEI1 (ca. 0.5 μg) strongly inhibited PME enzyme activity, and a higher concentration of purified CaPMEI1 (ca. 1.0 μg) exhibited the strongest inhibition. However, thioredoxin alone did not exhibit any inhibitory effect. These results suggest that CaPMEI1 functions as an inhibitor of PME activity. Antimicrobial activity of CaPMEI1 CaPMEI1 exhibited antifungal activity against the three plant pathogenic fungi examined: F. oxysporum f.sp. matthiole, A. brassicicola and B. cinerea (Fig. 5a). The thioredoxin–CaPMEI1 fusion protein (50 μg mL−1) suppressed mycelial growth of the three plant pathogenic fungi, whereas thioredoxin alone did not. Furthermore, the purified thioredoxin–CaPMEI1 (500 μg mL−1) fusion protein inhibited spore germination and hyphal growth of F.oxysporum f.sp. matthiole, whereas thioredoxin alone did not (Fig. 5b). Increasing concentrations of recombinant thioredoxin–CaPMEI1 fusion protein increasingly inhibited both spore germination and hyphal growth (Fig. 5c). Fig. 5Assay of thioredoxin–CaPMEI1 fusion protein antimicrobial activity. a Inhibitory effects of the CaPMEI1–thioredoxin fusion protein on mycelial growth of the plant pathogenic fungi F. oxysporum f.sp. matthiolae, A. brassicicola and B. cinerea. In each plate, the upper wells (a) were treated with purified thioredoxin and the lower wells (b) were treated with the purified thioredoxin-CaPMEI1 fusion protein. b Inhibition of germination and hyphal growth of F.oxysporum f.sp. matthiolae. Fungal spores were allowed to germinate and grow in 100 μL potato dextrose broth medium alone (top), or with 500 μg mL−1 thioredoxin (middle) or 500 μg mL−1 thioredoxin–CaPMEI1 fusion protein (bottom). Photographs were taken after incubation for 10 h at 28°C. Bars 20 μm. c Inhibition of spore germination and hyphal growth of F.oxysporum f.sp. matthiolae by CaPMEI1. The percentage of germinated spores and the length of fungal hyphae were determined by light microscopy. Data represent means ± SD from three independent experiments Enhanced susceptibility of CaPMEI1-silenced pepper plants to Xcv infection Since inoculation with Xcv strongly induced CaPMEI1 expression in pepper plants, we performed the virus-induced gene silencing (VIGS) technique (Liu et al. 2002a; Chung et al. 2004) to examine its cellular function during pathogen infection. The full-length ORF of CaPMEI1 was used to construct pTRV2:CaPMEI1. Five to six weeks after induction of silencing, empty vector control (TRV:00) and CaPMEI1-silenced (TRV:CaPMEI1) pepper plants were inoculated with virulent and avirulent strains Ds1 and Bv5-4a of Xcv, respectively. To assess the efficiency of VIGS, CaPMEI1 transcript levels were examined by RT-PCR (Fig. 6). CaPMEI1 transcripts were nearly undetectable in both non-inoculated empty vector control and CaPMEI1-silenced plants. However, at 12 h after inoculation with virulent and avirulent Xcv, we observed strong induction of CaPMEI1 in the empty vector control plants, whereas only weak or undetectable transcript levels were found in CaPMEI1-silenced plants. These results confirm the effective silencing of the target gene in pepper plants. Fig. 6RT-PCR analysis of expression of CaPMEI1 and several defense-related genes in empty vector control (TRV:00) and CaPMEI1 gene-silenced (TRV:CaPMEI1) pepper plants 12 h after inoculation with the virulent (Ds1; C, compatible) and avirulent (Bv5-4a; I, incompatible) strains of Xcv (5 × 106 cfu mL−1). 18S rRNA levels were visualized as a loading control. This experiment was repeated three times with similar results. H healthy leaves, CaBPR1 basic pathogenesis-related protein 1, CaPR10 putative ribonuclease-like protein, CaPOA1 ascorbate peroxidase 1, and CaSAR82A SAR8.2 To determine whether or not the expression of defense-related genes is affected by bacterial infection in the silenced plants, we used RT-PCR to analyze transcript levels of some defense-related genes (Fig. 6). In the empty vector control plants, both CaBPR1 (basic pathogenesis-related protein 1) and CaPR10 (putative ribonuclease-like protein) were strongly induced by virulent and avirulent Xcv infection. However, induction of these genes was slightly reduced in CaPMEI1-silenced plants infected with virulent Xcv, but was unaffected in plants infected with avirulent Xcv. Thus, CaPMEI1 gene silencing did not alter the induction of CaPOA1 (ascorbate peroxidase) or CaSAR82A (SAR 8.2) in response to Xcv infection. To examine the function of the CaPMEI1 gene in basal defense or gene-for-gene resistance, empty vector control and CaPMEI1-silenced pepper plants were infected with virulent or avirulent Xcv (Fig. 7). CaPMEI1 gene silencing significantly increased susceptibility to virulent Xcv infection, but not to avirulent Xcv infection. The CaPMEI1-silenced pepper leaves exhibited more severe disease symptoms 5 days after virulent Xcv inoculation than did empty vector control plants, and these symptoms were accompanied by severe chlorosis and enlarged water-soaked lesions (Fig. 7a). However, we did not observe any phenotypical changes in the cell death response of CaPMEI1-silenced plants following avirulent Xcv inoculation (Fig. 7a). At 3 days after inoculation with virulent Xcv, bacterial growth was tenfold greater in CaPMEI1-silenced plants than in the empty vector controls (Fig. 7b). However, silencing of CaPMEI1 conferred a slightly enhanced susceptibility to avirulent Xcv. These findings suggest that CaPMEI1 may function in basal resistance of pepper plants against Xcv infection, rather than gene-for-gene resistance. Fig. 7Enhanced disease susceptibility of CaPMEI1-silenced pepper plants to infection by the virulent Xcv strain Ds1, but not the avirulent Xcv strain Bv5-4a. a Disease symptoms developed on the leaves at different time points after inoculation with the virulent Xcv strain Ds1 (5 × 106 cfu mL−1) and the avirulent Xcv strain Bv5-4a (various bacterial concentrations). b Bacterial growth in leaves of empty vector control (TRV:00) or CaPMEI1-silenced (TRV:CaPMEI1) pepper plants at different time points after inoculation with the virulent Xcv strain Ds1 or the avirulent Xcv strain Bv5-4a (104 cfu mL−1). Data represent the mean ± SD from three independent experiments Enhanced resistance of CaPMEI1-OX plants to Pst DC3000 The p35S-CaPMEI1-GFP and p35S-GFP overexpression (OX) plants were generated by transformation of A. thaliana using the floral dipping method (Clough and Bent 1998). Plants transformed with the empty vector p35S-GFP were used as a control. Seeds (T1) were collected from each transformed plant and screened for resistance to kanamycin. Northern blot analysis was performed on T1 plants exhibiting kanamycin resistance to determine the integrity of the inserted transgene. Three T2 plants showing strong CaPMEI1 expression were selected for analyses in planta (Fig. 8a). Fig. 8Responses of wild-type (Col-0) Arabidopsis and CaPMEI1-OX transgenic plants to infection with P. syringae pv. tomato DC3000. a RNA gel blot analysis confirming CaPMEI1 overexpression (OX) in the transgenic Arabidopsis lines. Total RNA (10 μg) was loaded into each lane. The 3′ UTR region of pepper CaPMEI1 cDNA was used as a probe. b Growth of Pst DC3000 in the leaves of wild-type and transgenic plants. The mature leaves of the 6-week-old plants were infiltrated with a Pst Dc3000 suspension (105 cfu mL−1), and the degree of bacterial growth was rated at 0, 2 and 4 days after inoculation. c Disease symptoms on leaves of 6-week-old plants infiltrated with virulent Pst DC3000 (105 cfu mL−1). d Expression of pathogen-related (PR) genes in transgenic plants. Northern blot analyses were performed with 10 μg total RNA prepared from 5-week-old leaves of the wild-type (WT), vector control (smGFP) and transgenic (CaPMEI1::smGFP) plants. The samples were collected at 5, 15 and 25 h following pathogen infiltration with a suspension of the virulent strain Pst DC3000 (105 cfu mL−1) To determine the contribution of CaPMEI1 to Arabidopsis resistance, wild-type and CaPMEI1-OX plants were inoculated with virulent P. syringae pv. tomato DC3000 (Pst; 105 cfu mL−1). Bacterial titers were determined 4 days after inoculation. CaPMEI1-OX plants exhibited much less growth of Pst bacteria than wild-type or empty vector control plants (Fig. 8b). In the transgenic plants, this reduced bacterial multiplication was most pronounced at 4 days after inoculation. At 6 days after inoculation, wild-type (Col-0) and empty vector control plants developed typical chlorotic symptoms in the infected leaves, whereas the transgenic plants displayed few disease symptoms (Fig. 8c). Expression of defense-related genes in CaPMEI1-OX plants To gain insight into the role played by CaPMEI1 in PR gene induction, we examined the expression patterns of the well-established marker genes AtPR1a, AtPR2 and AtPR5 in the defense responses of wild-type, vector control and the CaPMEI1-OX Arabidopsis plants infected with virulent Pst DC3000 (Fig. 8d). The expression of AtPR1a and AtPR2 was very similar in the CaPMEI1-OX, wild-type and vector control plants. AtPR5 expression was not induced in either un-inoculated or inoculated, wild-type or vector control plants. However, there was significant expression of AtPR5 in the transgenic CaPMEI1-OX plants at 15 and 25 h after Pst DC3000 infection. AtPR1a, AtPR2 and AtPR5 expression is known to be regulated via the SA-dependent pathway in Arabidopsis (Uknes et al. 1992), and in nahG plants, lack of SA reduces expression of these three PR genes during pathogenesis (Delany et al. 1994; Nawrath and Metraux 1999). In our experiments, AtPDF1.2 (defensin) transcripts were not detected in wild-type, vector control or CaPMEI1-OX plants (data not shown). Responses of wild-type and CaPMEI1-OX plants to Hyaloperonospora parasitica We examined whether or not ectopic CaPMEI1 expression in Arabidopsis plants affected resistance to the virulent biotrophic oomycte pathogen H. parasitica isolate Noco2 (Fig. 9). Over 100 seedlings of both wild-type (Col-0) and CaPMEI1-OX transgenic lines were inoculated with spores of H. parasitica isolate Noco2 (5 × 104 conidiosporangia mL−1). At 7 days after inoculation, over 50 plants in each line were sampled to estimate the disease and assess the degree of asexual sporulation, which was quantified and expressed as the mean number of sporangiophores per cotyledon. Fig. 9Responses of wild-type (Col-0) Arabidopsis and CaPMEI1-OX transgenic plants to infection with H. parasitica isolate Noco2. a Disease symptoms and trypan blue-stained pathogen structures on 7-day-old cotyledons of wild-type and transgenic plants 7 days after inoculation; dpi days post-inoculation. Bars 0.5 mm. b Quantification of asexual sporangiophores per cotyledon for at least 50 cotyledons of wild-type and transgenic plants 7 days after inoculation. The average number of sporangiophores produced on the cotyledons of wild-type and transgenic lines are shown below each of the lines tested As shown in Fig. 9a, the cotyledons of wild-type and transgenic seedlings responded to H. parasitica infection by stimulating similar levels of asexual parasite sporulation. Their trypan blue-stained cotyledons also exhibited abundant hyphae, conidiospores and oospores (Fig. 9a). Quantitative disease ratings are shown in Fig. 9b. Both the transgenic and the wild-type plants showed a high level of susceptibility to H. parasitica isolate Noco2, and similar levels of heavy asexual sporulation (>20 sporangiophores per cotyledon) were observed on both lines. Drought tolerance of CaPMEI1-OX plants To investigate a possible role for CaPMEI1 in the dehydration response, we tested seed germination and seedling growth under osmotic stress. The seeds of wild-type, vector control and CaPMEI1-OX plants were placed on MS media supplemented with various concentrations of mannitol. We observed no significant differences in seed germination between these lines (Fig. 10a). However, treatment with 200 and 600 mM mannitol strongly inhibited germination in the wild-type and vector control plants compared with the CaPMEI1-OX plants. Fig. 10Transgenic ArabidopsisCaPMEI1-OX lines exhibit enhanced tolerance to drought stress. a Seed germination in wild-type, smGFP and transgenic plants on the MS media containing 0, 200 and 600 mM mannitol. The data represent the mean ± SD of 100 seeds for each line tested. b Relative root length of wild-type, smGFP and transgenic lines in MS agar medium containing different concentrations of mannitol. Three independent experiments were performed with 40 seedlings of both wild-type and transgenic lines. c Drought tolerance test of transgenic seedlings. Wild-type, smGFP and transgenic lines were germinated and grown in 1× MS agar medium. Each seedling was transferred to liquid medium containing 100 mM mannitol. d Wild-type, smGFP and the CaPMEI1 transgenic Arabidopsis plants after 15 days without water. e Water loss from the excised leaves of wild-type, smGFP and transgenic plants. Data represent the mean ± SD from three independent experiments We also tested the sensitivity of root growth to osmotic stress (Fig. 10b). Drought tolerance was observed during post-germination growth. The root growth of wild-type and vector control seedlings was inhibited in the presence of 150 and 200 mM mannitol. However, root elongation of CaPMEI1-OX plants was less sensitive to mannitol-induced osmotic stress and in comparison with wild-type and vector control plants, mutant seedlings grew well in liquid medium supplemented with 100 mM mannitol (Fig. 10c). In addition, we noticed that adult CaPMEI1-OX plants exhibited enhanced drought tolerance. Following 14 days without water, wild-type and vector plants had withered severely, whereas the CaPMEI1-OX plants remained healthy. To determine the effect of CaPMEI1 on survival, these plants were rewatered on day 16; wild-type and vector control plants died, whereas the transgenic lines survived (Fig. 10d). Next, we examined transpiration rates by measuring fresh weight loss in detached leaves. The leaves of CaPMEI1-OX lines exhibited slightly slower water loss than those of wild-type or vector control plants (Fig. 10e). Together, these results indicate that CaPMEI1 overexpression enhanced water stress resistance. Oxidative tolerance of CaPMEI1-OX plants To investigate the response of CaPMEI1-OX plants to oxidative stresses, seeds of wild-type, vector control and CaPMEI1-OX plants were exposed to MS medium containing methyl viologen (MV) (Fig. 11). Treatment with 5 or 10 μM MV significantly inhibited germination of wild-type and vector control seeds compared with CaPMEI1-OX seeds (Fig. 11a). Seven-day-old seedlings were transferred to a medium containing different MV concentrations (0–0.5 μM) and grown for 2 weeks. Wild-type and vector control seedlings turned white and started to die after 15 days of MV treatment, but CaPMEI1-OX lines were much less affected by the treatment (Fig. 11b). Furthermore, these differences in plant phenotype were also observed with respect to the higher fresh weight of transgenic seedlings compared to those of wild-type and vector control plants (Fig. 11c). Detached leaves of 4-week-old plants were treated with 10 μM MV for 24 h, after which chlorophyll content was measured (Fig. 11d). The leaves of transgenic lines treated with MV retained more chlorophyll than those of wild-type or vector control plants. Fig. 11Transgenic ArabidopsisCaPMEI1-OX lines exhibit tolerance to oxidative stress. a Effects of methyl viologen on the seed germination of transgenic lines. Seeds from wild-type, vector control and transgenic lines were plated on media with or without methyl viologen (MV, 5 and 10 μM) and incubated for 3 days. The data represent mean ± SD of 100 seeds for each line tested. b Phenotypes of wild-type, vector control and transgenic line seedlings treated with different concentrations of MV. c Fresh weights of seedlings grown in the indicated concentrations of MV for 2 weeks. The results are presented as the average fresh weight per seedling. Data represent mean ± SD from three independent experiments. d Chlorophyll content of MV-treated leaves of wild-type, vector control and transgenic plants, which were floated on 0, 0.05, 0.1 and 0.5 μM MV in MS medium and then incubated for 24 h in a growth chamber Discussion Plants possess a diverse range of cell wall-modified enzymes, which are post-transcriptionally regulated by numerous inhibitor-related proteins (Tymowska-Lalanne and Kreis 1998; Raush and Greiner 2004). Several invertase inhibitor-related proteins have been isolated from higher plants such as kiwi, tobacco and Arabidopsis (Greiner et al. 1998; Wolf et al. 2003; Giovane et al. 2004; Raiola et al. 2004). Attempts to characterize the activity of invertase inhibitor proteins from plant species other than kiwi or Arabidopsis have been either unsuccessful or resulted in the isolation of invertase inhibitors that share structural similarities with pectin methylesterase inhibitor proteins (PMEI), but which represent completely different target enzymes (Greiner et al. 1998; Scognamiglio et al. 2003). In this study, we identified and functionally characterized a novel pepper CaPMEI1 gene encoding a PMEI. This CaPMEI1 protein contains the four cysteine residues that are conserved among other PMEI proteins (Camardella et al. 2000; Sato et al. 2000). These residues are expected to be engaged in two disulfide bridges, which constitute a common structural motif within the PMEI domain (Camardella et al. 2000). At the molecular level, CaPMEI1 expression was induced in pepper leaves by infection with bacterial pathogens and treatment with plant hormones such as SA, ethylene, MeJA and ABA. In particular, these hormone treatments strongly induced CaPMEI1 transcription, suggesting that this gene may be involved in the early stages of the active defense responses to bacterial pathogen infection and exogenous treatment with plant hormones. In enzymatic assays, purified CaPMEI1 proteins significantly inhibited activity of plant pectin methylesterase (PME). In addition, CaPMEI1 exhibited antifungal activity against a broad range of plant pathogenic fungi, including F. oxysporum f.sp. matthiolae, A. brassicicola and B. cinerea. To penetrate the cuticular layer, plant fungal pathogens produce plant cell wall-degrading enzymes such as polygalacturonase, pectin lyase and cellulase (Collmer and Keen 1986). An aggressive Phaeosphaeria nodorum isolate was shown to produce high amounts of xylanase, cellulase, polygalacturonase and butyrate esterase in vitro (Lalaoui et al. 2000). Recently, it was found that plant PMEIs do not inhibit the PMEs produced by plant pathogens (Giovane et al. 2004; Di Matteo et al. 2005). However, Arabidopsis plants expressing either AtPMEI-1 or AtPMEI-2 showed reduced infection by B. cinerea (Lionetti et al. 2007). This finding suggests that the increased level of pectin methylesterification caused by overexpression of AtPMEI-1 results in the inhibition of fungal endopolygalacturonase activity. Thus, the increase in PMEI activity resulted in reduced accessibility for fungal pectin degrading-enzymes and hence provided increased resistance to pathogens (Boudart et al. 1998; Lionetti et al. 2007). Consistent with these findings, our results suggest that CaPMEI1 may function as part of a new group of plant pectin methylesterase inhibitors, which restrict fungal pathogen infection in plants. We used virus-induced gene silencing (VIGS) to investigate the effect of CaPMEI1 loss-of-function in pepper plants during Xcv infection. The CaPMEI1-silenced plants were susceptible to Xcv infection, and in particular to infection with the virulent strain, which resulted in enhanced bacterial growth and reduced PR1 and PR10 gene expression. Basal resistance is activated during the compatible bacterial interaction which restricts the spread of pathogens in the host plants to a certain extent (Glazebrook 2001). In addition, basal resistance is also effective in retarding proliferation of a wide range of microbial pathogens (Chisholm et al. 2006), but it is dependent upon SA accumulation (Cao et al. 1994, 1997; Kinkema et al. 2000). Therefore, we conclude that CaPMEI1 expression may be involved in basal resistance by triggering downstream PR gene induction in pepper plants. To determine the effect of CaPMEI1 gain-of-function inplanta, we generated the CaPMEI1-OX Arabidopsis transgenic lines and investigated their response to P. syringae and H. parasitica infection, because these well-known model pathogens have been used extensively for the study of disease resistance mechanisms in Arabidopsis (Quirino and Bent 2003; Slusarenko and Schlaich 2003). The CaPMEI1-OX lines were resistant to Pst DC3000 infection, but not to infection by the biotrophic oomycete H. parasitica, which uses living cells as a nutrient source during the infection cycle (Alfano and Collmer 1996; Heath 2002). Moreover, as CaPMEI1 transcripts localize intensively in the xylem of vascular bundles in leaf tissues, they may not affect the accessibility of host plant cells to H.parasitica. In contrast, we hypothesize that intercellular growth of P. syringae may be restricted by the extracellular secretion of CaPMEI1 into host cells, resulting in triggering of the basal resistance response. We found that fungal growth was inhibited by treatment with the recombinant CaPMEI1 protein in vitro, suggesting that CaPMEI1 may interfere directly with pathogen infection of host plants. Plant pathogenic microorganisms have been shown to produce a variety of pectinolytic enzymes that macerate and kill plant tissues (Collmer and Keen 1986). Cell wall fragments released by these pectinolytic enzymes may elicit the plant defense response (D’Ovidio et al. 2004). The polygalacturonase-inhibiting protein (PGIP) plays an important role in the recognition and inhibition of fungal polygalacturonase (PG). Overexpression of PGIPs in Arabidopsis not only significantly reduces disease symptoms, but also enhances defense gene activation during pathogen infection (Ferrari et al. 2003). Thus, we suggest that CaPMEI1 can disrupt invading pathogenic microorganisms by inhibiting pectin methylesterases produced by these pathogens. The reduced bacterial growth observed in CaPMEI1-OX Arabidopsis lines may result from the expression of SA-inducible genes and CaPMEI1 overexpression. The SA-inducible genes such as PR1, PR2 and PR5 are activated in the SA defense pathway (Uknes et al. 1992). PR5 proteins are similar to thaumatin, which is a sweet-tasting protein from Thaumatococcus daniellii (Hu and Reddy 1997), and several pathogens can induce these proteins in a wide range of plant species (Ward et al. 1991; Hu and Reddy 1997; Reuber et al. 1998). Members of the PR5 group have been shown to exhibit antifungal activity against a broad spectrum of fungal pathogens (Coca et al. 2000) and to participate in the coordinated induction of systemic acquired resistance (SAR) against TMV (Ward et al. 1991). Therefore, the concomitant induction of PR genes may contribute to the enhanced resistance of CaPMEI1-OX Arabidopsis plants to bacterial pathogens. The CaPMEI1-OX Arabidopsis lines showed a strong tolerance to drought stress, and CaPMEI1 overexpression resulted in reduced transpiration and enhanced root elongation. In contrast, transgene overexpression did not cause any obvious phenotypic differences under optimal growing conditions. Dehydration factors such as a mannitol and polyethyleneglycol (PEG) have been used to evaluate the effects of decreased water availability and simulate drought conditions in wild-type and CaPMEI1-OX Arabidopsis plants (Gupta and Kaur 2005; Verslues et al. 2006). Other plant invertase inhibitor-related protein genes such as NtCIF and NtVIF are also strongly induced by treatment with PEG or ABA (Rausch and Greiner 2004). However, the molecular and genetic roles played by PMEIs in drought and osmotic stress tolerance remain poorly understood. Since plant responses to different abiotic stresses may be related to the accumulation of ROS and the mechanisms for their detoxification, the role played by ROS in stress signaling has been studied extensively (Apel and Hirt 2004). Methyl viologen binds to the thylakoid membranes of chloroplasts and in the presence of light, transfers electrons to O2 in a chain reaction causing continuous formation of superoxide radicals and oxidative stress (Asada 1996). The ArabidopsisCaPMEI1-OX lines exhibited tolerance to oxidative stress, both during seed germination and seedling growth. This tolerance to oxidative stress may reduce the damage caused by other stresses via the antioxidizing system, which suggests that CaPMEI1 overexpression results in detoxification of endogenous superoxide. Here, we have determined that CaPMEI1 from pepper plays a role as an antifungal protein and has an inhibitory effect on PME. Furthermore, we have shown that Arabidopsis CaPMEI1-OX lines are resistant to bacterial pathogens. In addition, they exhibit tolerance to drought and oxidative stress. In conclusion, these multivariate functions of CaPMEI1 provide valuable insights into understanding the physiological significance of PMEIs in plant disease resistance and abiotic stress tolerance. Electronic supplementary material Below is the link to the electronic supplementary material. S1 (DOC 19 kb)	[ "pectin methylesterase inhibitor protein", "antifungal activity", "disease resistance", "capsicum annuum", "oxidative stress tolerance", "drought tolerance" ]	[ "P", "P", "P", "P", "P", "P" ]
J_Gastrointest_Surg-3-1-1852386	Bromodeoxyuridine Labeling Index as an Indicator of Early Tumor Response to Preoperative Radiotherapy in Patients with Rectal Cancer	Purpose Assessment of tumor proliferation rate using Bromodeoxyuridine labeling index (BrdUrdLI) as a possible predictor of rectal cancer response to preoperative radiotherapy (RT). Introduction In specialized centers, a refined surgical technique has resulted in high local control figures in rectal cancer. However, local recurrence rates after “standard” surgery are generally high, with figures ranging between 20 and 40%1,2, although after adopting the total mesorectal excision (TME) concept they fell down to 10–12%3,4. Radiotherapy in addition to surgery significantly diminishes the risk of local failure by more than half, from 8 to 2% after 2 years3. Therefore, combined treatment: radiotherapy (RT) and surgery in the treatment of patients with resectable rectal cancer has been proposed in many trials using either preoperative5,6 or postoperative irradiation7,8. Better results of preoperative RT for 5 days (25 Gy in five fractions) in comparison with postoperative 60 Gy in 30 fractions were achieved by a Swedish group4,9,10, with respect to the local recurrence rate11 and overall survival11,12. A corresponding improvement in overall survival has not been demonstrated after postoperative radiotherapy alone13. Graf’s12 study provided a clinically significant biologic effect of a short preoperative course of radiotherapy on the tumor size and on the incidence of nodal metastases; however, this effect was minimized if surgery was performed immediately after radiotherapy. The effect is most likely caused by death of tumor cells in the primary tumor and in the involved nodes. A short treatment course of radiotherapy, i.e., 5 × 5 Gy is desirable, and this regimen is currently considered as the gold standard in many centers. However, using this schedule it is difficult to observe a down-staging and/or downsizing of the tumor, which is of importance for the selection of patients for sphincter-preserving surgery (anterior resection). In clinical practice there are no certain methods able to predict tumor response to preoperative radiotherapy (RT). The optimal timing of surgery after preoperative radiotherapy in rectal cancer is unknown. However, it was shown that a long interval (6–8 weeks) between preoperative radiotherapy (39 Gy in 13 fractions) and surgery was associated with a significantly greater clinical tumor volume reduction than a short interval (2 weeks)14. On the other hand, it was shown that subclinical pelvic deposits of rectal cancer could grow rapidly during preoperative radiation therapy and during the radiotherapy–surgery interval, with an adverse influence on the rate of pelvic tumor control from protracting the overall treatment time15. Graf et al12 showed that low doses in short RT only offer clinically relevant reduction in the risk of pelvic relapses if the overall radiation treatment time is short. Thus, the rate of cancer cell proliferation seems to be a very important prognostic factor. The aim of this study is to evaluate BrdUrd LI and S-phase fraction (SPF) as the possible indicators of tumor proliferation rate and predictors of the tumor response to neoadjuvant RT in patients with rectal cancer, and to suggest an optimal interval between short RT course and surgery. Methods and Materials Patients Between November 2003 and January 2006 we recruited 92 patients with resectable rectal carcinoma for whom curative surgery was planned. Patients were eligible for the trial if they were less than 75 years old, had a histopathologically proved adenocarcinoma (T2/T3)16 situated less than 12 cm from the verge of the anus, and gave informed consent for their participation. The protocol was approved by the Ethical Committee of the Center of Oncology, and each patient gave written consent. The criteria for exclusion were: locally nonresectable tumor; plan to perform only local tumor excision; known metastatic disease; previous radiotherapy of pelvis region; other malignant disease; and patient’s refusal. Preoperative Radiotherapy The patients assigned to preoperative radiotherapy received a total tumor dose of 25 Gy. The treatment was given in five fractions over 5 days, one posterior and two lateral wedged fields were irradiated with photons of maximum 6 MV energy. According to the random selection surgery was performed the following week (schedule I) or after longer interval of 4–5 weeks (schedule II). Surgery Anterior resection of rectum or abdominoperineal excision was performed within a week or a month after the completion of RT. Type of surgery was resection of the rectum and lower sigmoid with involved adjacent tissue and regional lymph nodes up to or above the origin of inferior mesenteric artery. A minimal touch technique was used with high tight ligation of the inferior mesenteric artery. The decision whether the patient should have an abdominoperineal resection or a sphincter-preserving surgery was made by the surgeon during the operation. An abdominoperineal resection of rectum was performed in 41 (44.6%) of the patients, and sphincter preserving surgery was performed in 51 (55.4%). Biological Assessment of Tumor Response Tumor samples were taken twice: before radiotherapy (through a rectoscope) and during surgery from the same place, i.e., at the lowest edge of the tumor mass. Each biopsy was divided into two parts: one was used for BrdUrd LI assessment, and the second was used for immunohistochemical analysis (these results will be the subject of a separate study). Bromodeoxyuridine Labeling Index Incorporation of BrdUrd in tumor samples from a biopsy (0.3–0.5 cm3) was carried out in vitro according to the high-pressure oxygen method. The BrdUrd staining procedure and flow cytometry have been described in detail elsewhere17. The stained preparations were analyzed with a FACS Calibur flow cytometer (Becton Dickinson Immunocytometry Systems, Sunnyvale, CA, USA) by one coauthor (AG) and 20 × 103 events were collected in each histogram. The BrdUrdLI was calculated as a percentage of BrdUrd-labeled cells in a sample, which incorporated BrdUrd during 1 h of incubation at 37°C (with discrimination of diploid subpopulation in aneuploid tumors). The tumor ploidy and SPF were calculated from the DNA profile with ModFit software running on a MacIntosh computer. Apoptotic cells were identified as objects with a fractional DNA content not less than 20% of the 2n DNA content. Cell death was calculated as the sum of apoptosis and debris. The tumor ploidy was estimated by evaluating the DNA index, i.e., the ratio of the modal DNA fluorescence of abnormal to normal G1/0 cells. Aneuploidy was assessed in cases in which the normal and neoplastic cell populations gave two separate peaks. Human lymphocytes were used for the reference peak. Tumors with BrdUrdLI >8.5% (median value) were considered as fast, and those with BrdUrdLI ≤8.5% were considered as slowly proliferating tumors. Clinical Assessment of Tumor Response Tumor size before RT was assessed basing on measures taken during rectoscopy, and endorectal sonography. Tumor regression after RT was assessed at the time of operation by surgeons according to the following Response Evaluation Criteria in Solid Tumors (RECIST)18: Complete response (CR): 100% disappearance; partial response (PR): 30–99% decrease; stable disease (SD): neither CR, PR or PD criteria met; progression of disease (PD): 20% increase in sums of tumor longest diameters. Pathological Assessment of Tumor Response Tumor regression after RT was evaluated by a pathologist on the excised tumor mass. The following criteria of tumor regression assessed by Dworak et al.19 were applied: D0—no regression; D1—dominant tumor mass with obvious fibrosis and/or vasculopathy; D2—dominantly fibrotic changes with few tumor cells or groups; D3—very few (difficult to find microscopically) tumor cells in fibrotic tissue with or without mucous substance; D4—no tumor cells, only fibrotic mass (total regression or response). Statistical Methods Statistical analysis was performed with STATISTICA vs.5. Intergroup differences in the ordinal data were tested with ANOVA test or Student’s t test. P values of less than 0.05 were considered to indicate statistical significance. Linear regression was applied for assessing differences between fast and slowly proliferating tumors in relation with OTT, and its significance was determined by testing the difference between two correlation coefficients. Stratification by BrdUrd LI level was introduced and tested by the inclusion of dummy variable in the regression model. Results Patients A total of 92 patients were included in the study. Twenty-eight (23.3%) out of 120 patients initially qualified for this study were excluded from the analysis because of discontinuation of treatment, metastatic tumor noticed at operation, or no tumor samples taken for biological assessment during surgery. Mean age for the entire group of patients was 61.6 years (range 30–75). There were 68 men and 24 women. There were no statistical differences between the two groups at the time of recruitment for prognostic factors such as: sex, age, histologic grade, or tumor stage (Table 1). Table 1Selected Characteristics of Patients and Treatment ParametersCharacteristicsSchedule ISchedule IITotalAge mean (±SD) years(38)a 61.2 ± 12.0(54) 61.9± 9.5(92) 61.6 ± 10.6Sex Male303868 Female81624Histological grade G162026 G2293463 G3303Tumor stage T181927 T2253055 T35510PTNM 1162541 28614 3131730 4123Interval between RT and surgery Mean (range) days(38)a 8.8 (2–14)(54) 32.9 (17–45)(92) 22.9 (2–45) OTT mean (range) days(38) 13.8 (7–19)(54) 37.9 (22–50)(92) 27.9 (7–50)Surgery Sphincter-preserving20 (52.6 %)31(57.4 %)51 Abdominoperineal resection182341aNumber of patients In our series of patients, there were 27 stage 1 (29.3%), 55 were T2 (59.8%), and 10 were T3 (10.9%). In 26 patients, tumor cells well differentiated (G1), 63 moderately differentiated (G2), and three poorly differentiated (G3) (Table 1). Thirty-eight patients were treated according to schedule I, in which time interval between end of irradiation and surgery averaged 8.8 days (range 2–14; Table 1). In 54 patients, schedule II was applied, in which mean break was 32.9 days (range 17–45). Because the interval between RT and surgery appeared to be longer than planned, overall treatment time (OTT), e.g., time from the beginning of RT to surgery, was calculated and it appeared to be 7–50 days (Table 1). Biologic, Pathologic, and Clinical Assessment of Tumor Response Mean BrdUrd LI before RT was 8.5% (range 1.0–24.2%) and SPF was 22.0% (range 3.8–49.9%) and the mean values did not differ between the two schedules (Table 2). Poorly differentiated tumors showed statistically significant higher BrdUrd LI than grades 1 and 2 tumors (P = 0.015; Table 3). After RT, tumors treated according to both schedules showed statistically significant growth inhibition (reduction of BrdUrd LI and percentage of SPF cells) in comparison with the values obtained before RT (Table 2). Radiation induced inhibition of tumor proliferation was expressed as a percentage of the after RT to before RT BrdUrd LI, and SPF as after/before RT percentage. This ratio ranged from 2.5 to 514% for BrdUrd LI (Fig. 1) and from 5.8 to 522.2% for SPF. When we stratified patients into two groups according to their biological RT response, those radioresponsive with reduction of pretreatment values after radiotherapy above 50% and those less responsive with reduction below 50%, it appeared that the mean values (of the after/before RT ratios of BrdUrd LI and SPF) for the more radioresponsive tumors were significantly higher than for the less responsive ones. Therefore, these ratios were presented separately for fast (BrdUrd LI >8.5%, SPF >22.0%) and slowly (BrdUrd LI ≤8.5%, SPF ≤22.0%) proliferating tumors. Mean BrdUrd LI value after RT for fast proliferating tumors (41 cases) showed statistically significant (P = 0.027) reduced pretreatment percentage (46.8%) in comparison with slowly proliferating tumors (85.3%, 51 cases). The same was true for SPF of fast (56.4%, 55 cases) and slowly (113.8%, 37) proliferating tumors (P = 0.006). Table 2Status of Biological Parameters Before and After RTGroupBrdUrd LI (%) Mean (range)S-phase fraction (%) Mean (range)Apoptosis (%) Mean (range)All patientsBefore RT8.5 (1.0–24.2)22.0 (3.8–49.9)5.9 (0–52.8)After RT4.1* (0.4–18.3)16.8 (1.5–101.0)9.8* (0–45.9)RT schedule IBefore RT8.4 (1.1–24.2)21.5 (6.1–49.2)6.6 (0–32.4)After RT3.8(0.8–12.6)14.1*** (1.5–47.9)10.5 (0–43.3)RT schedule IIBefore RT8.6 (1.0–20.0)22.3 (3.8–49.9)5.4 (0–52.8)After RT4.5* (0.4–18.3)17.2 (2.6–101.0)9.5***** (0–45.9)P = 0.000P = 0.015P = 0.010P = 0.002**P = 0.042Table 3The Relationship Between Tumor Biological Parameters and Histological GradeHistological gradeNBrdUrd LI (%) Mean (range)S-phase fraction (%) Mean (range)Apoptosis (%) Mean (range)G1268.5 (1.1–17.1)23.7 (5.8–49.9)3.4 (0–32.4)G2618.2 (1.0–20.0)21.0 (3.8–45.6)7.2 (0–52.8)G3316.2, ** (9.5–24.2)27.7 (18.9–34.7)2.1 (0.4–4.8)P = 0.015, difference between G1 and G3*P = 0.013, difference between G2 and G3Figure 1The association between biological tumor response for slowly (BrdUrd LI ≤ 8.5%; closed symbol) and faster proliferating tumors (BrdUrd LI > 8.5%; open symbol) and overall treatment time. Insert shows linear regression performed separately for each of the tumor subgroups for OTT > 30 days. P value shows difference between two correlation coefficients. Next, the after/before RT ratios for BrdUrd LI and SPF were correlated with OTT. For SPF, statistical difference between linear regression coefficients for fast and slowly proliferating tumors was not obtained (P = 0.446), therefore the data for BrdUrd LI only are shown (Fig. 1). Insert on Fig. 1 shows a significant (P = 0.033) difference in proliferation rate between fast and slowly proliferating tumors treated within OTT >30 days. At that time slowly proliferating tumors, contrary to fast proliferating ones, show no inhibition but accelerated proliferation of tumor cells. This phenomenon was also confirmed by increased fraction of S-phase cells in tumors treated with longer RT schedule (Table 2). The influence of BrdUrdLI level has been also tested by the extended regression model between OTT and the percentage of after/before RT BrdUrd LI. BrdUrd LI level higher than 8.5% has been coded as dummy variable. It appeared to be significant (P = 0.025) in the relation between OTT and the percentage of after/before RT BrdUrd LI. The partial regression coefficient indicates that the average decrease of the percentage after/before RT BrdUrdLI for fast proliferating tumors (BrdUrd LI >8.5%) equals 39%. All 92 irradiated rectal tumors were reviewed by the same pathologist (KN). The tumors were classified according to the World Health Organization classification of intestinal carcinoma16 and staged according to the TNM classification20. Of the 92 rectal tumors four had no pTNM classification, 41 were pT1 (46.6%), 14 were pT2 (15.9%), 30 were pT3 (34.1%), and three were pT4 (3.4%). Regional lymph node metastases were found in 27 (30.7%) patients, and 27 (30.7%) patients had their tumor down-staged. Pathologic assessment of tumor regression after RT according to classification described by Dworak19 was performed in 90 out of 92 patients (for two patients the assessment was impossible). The analysis showed no tumor regression (D0) in 18 (20.0%) tumors, dominant tumor mass (D1) in 46 (51.1%) tumors, a few tumor cells in fibrotic mass (D2) in 18 (20.0%) tumors, single tumor cells (D3) in four (4.4%), and no tumor cells were observed in four (4.4%) of the examined tumors (Fig. 2a). In 25 (27.8%) out of 90 patients marked pathologic down-staging (no residual tumor confined to the rectal wall) was visible. Pretreatment BrdUrdLI and SPF were not correlated with early clinical and pathologic tumor response. However, patients having tumors with LI >8.5% were more radioresponsive (showed significant reduction in proliferative rate after radiotherapy) than patients with BrdUrdLI ≤8.5% tumors, although statistically significant difference between the two tumor subgroups was seen only for D0–D1 grade (Fig. 2b). Figure 2Association between biological and pathological assessment (Dworak classification) of early tumor regression for (A) total group of patients and (B) for slowly (BrdUrdLI ≤ 8.5%) and fast proliferating (BrdUrdLI > 8.5%) tumors. Mean values ± SE are shown. For stages D0–D1, statistically significant lower inhibition of tumor cell proliferation after RT was observed for slowly than fast proliferating tumors. In the clinical assessment of tumor mass resected during surgery, 34 (36.9%) tumors showed stable disease, 12 (13.0%) showed progressive disease, 41 (44.6%) showed partial response, and four (4.3%) showed complete response (Fig. 3a). And again, in fast proliferating tumors, greater inhibition in tumor proliferation rate (reduction of pretreatment BrdUrd LI value >50%) was observed in fast than in slowly proliferating tumors; however, this difference was not statistically significant (Fig. 3b). As the observed correlation between clinical assessment and SPF was weaker than for BrdUrd LI, the data were not shown. Partial and total tumor regression was observed in 45 (48.9%) tumors. However, tumor proliferation status was not in agreement with the kind of surgery. Sphincter-preserving surgery was performed in 51 out of 92 patients: in 22 (23.9%) fast proliferating and in 29 (31.5%) slowly proliferating tumors. Figure 3Association between biological and clinical assessment of early tumor regression after RT for all tumors (A) and separately for slowly and fast proliferating tumors (B). Mean value ± SE are shown. Discussion This study provides evidence of a clinically significant biological effect of a short preoperative course of RT on tumor proliferation rate. The impact of irradiation on biological tumor response was assessed by BrdUrd LI, SPF, and the degree of subsequent pathologic and clinical down-staging of the tumors after surgery. The study showed differences in the pretreatment proliferation rate of the tumor. Mean BrdUrd LI before RT was equal to 8.5% and ranged from 1 to 24.2%. Mean SPF was 22.0% and ranged from 3.8 to 49.9%. The proportion of cells in S-phase as estimated by the DNA content overestimates the labeling index determined by the uptake of BrdUrd. This may be so because the exposure time is quite short and there may be subpopulations in the tumors that are synthesizing DNA at a very slow rate, or there may indeed be cells with an S-phase DNA content that are not synthesizing DNA (as a result of nutrient or oxygen supply, lack of growth factors, inadequate vascularity). Mean value of the BrdUrd LI obtained in this study was lower than the one estimated by Bergstrom et al.21, Palmqvist et al.22, and Terry et al.23, and can be explained by a different method used by these authors: in vivo incorporation of iodouridine/bromodeoxyuridine, which can cause longer exposure of the tracers to S-phase cells. The differences in the LI value might be caused also by heterogeneity in proliferation within the tumor. It was shown by Bergstrom et al.21 that rectal tumors are polarized, having the superficial surface toward the lumen of the gut and the other toward deep structures facing totally different environments. Apart from Bergstrom et al., none of the above-mentioned authors gave account of site from where the tumor samples were taken. In each tumor analyzed by us, all the samples were taken from the same region, i.e., the bottom part of the mass. In our study, pretreatment BrdUrd LI or SPF was not predictive for early clinical and pathologicl tumor response, probably because of different tumor microenvironment. However, BrdUrd LI after/before RT ratio gave information on the different significant biological processes that take place after irradiation, and have impact on cell death like redistribution, repopulation, and reoxygenation. BrdUrd LI after RT decreased to mean 4.1% independently of the time interval between RT and surgery. Magnitude of LI reduction after RT was correlated with tumor proliferation rate. Greater reduction of BrdUrd LI value was observed in fast proliferating (LI >8.5%) tumors (to mean 46% of the pretreatment value) than in slowly (LI ≤8.5%) proliferating tumors (to mean 85.3% of pretreatment value). What then is the justification for better RT response of fast proliferating tumor cells? According to current knowledge on tumor proliferation, radiation therapy should preferentially inactivate rapidly dividing cells, leaving behind a population biased toward slow proliferation. However, recruitment is a known effect of cytotoxic treatment, and new cells from quiescent cell populations are recruited into active proliferation after irradiation. Probably, slowly proliferating tumors might have greater propensity to recruit cells into rapid cycle in response to treatment than fast proliferating tumors, which might have little reserve capacity for further accelerating their cell cycle24. That might be why we observed acceleration of proliferation rate in slowly proliferating tumors from 5 weeks after RT (basing on after/before RT BrdUrd LI ratio), which followed temporary reduction of the number of DNA-synthesizing cells, 4–5 weeks after the start of RT. Accelerated proliferation was confirmed by increased S-phase fraction. However, better biological tumor responsiveness of fast proliferating tumors on cellular level did not find confirmation on tissue level that is in surgery because a fewer number of sphincter saving resections were performed in patients with fast (22) than those with slowly proliferating tumors (29). Regression of rectal carcinoma after preoperative irradiation varies, likely reflecting differences in the physical and biologic properties of these tumors. Apart from biological characteristics discussed here, tumor down-staging depends on the total irradiation dose, the fractionation, and the interval between irradiation and surgery25. We showed association of tumor proliferation rate after RT with tumor response basing on BrdUrd LI. SPF, considered as a less sensitive method of tumor proliferation, did not show such a correlation. The after/before radiotherapy BrdUrd LI ratios correlated, however nonsignificantly, with the degree of pathologic and clinical down-staging, which indicates that more radiation-induced cell death occurred in tumors that expressed high levels of BrUrd LI, or that an increased rate of tumor clearance occurred in more rapidly proliferating tumors. This effect was reflected by significantly higher incidence of apoptosis observed after RT only in fast proliferating tumors (4.1% vs 11.1%; P = 0.000). However, patients having tumors with LI >8.5% did not show higher rate (11.2%) of tumor pathological down-staging (D2–D4) than patients with BrdUrd LI ≤8.5% (16.8%) tumors, which may be suggestive of significant impact on tumor response also by biological processes other than proliferation. In the Spanish study26, high proliferative activity of rectal cancer, as determined by PCNA immunostaining, was predictive of response to preoperative chemoradiotherapy. Willett et al.27, in the same tumor type treated with higher RT dose (47–52 Gy) and 5 FU, showed that patients having tumors with extensive Ki-67 staining had also a higher rate of tumor down-staging (36%) 4–6 weeks after treatment than patients with minimal to moderate Ki-67 staining tumors (22–23%). These authors show that elevated postirradiation tumor proliferative activity correlated strongly with improved survival28. These authors, in contrast to our study, did not consider the proliferation profile of pre- and postirradiation for individual patients. The correlation of down-staging and higher survival rates was also found by other authors29,30. In our study, even in totally regressed tumors (D4), the percentage of the after/before radiotherapy BrdUrd LI was about 50%, which may not indicate tumor but normal cell proliferation, mainly a fraction of activated fibroblasts or cycling endothelial cells in capillaries high in colorectal carcinoma31. Our study showed complete pathologic response (D4) similar to that in a Norwegian study (4.5%)32, where histological tumor slides were analyzed after treatment with a dose of 31.5 Gy in 18 fractions and 2–3 weeks interval between RT and surgery. However, it should be stressed that in this study, a high incidence (31.3%) of recurrences was observed at late follow-up. Our analysis showed that patients having fast proliferating tumors, as assessed by BrdUrdLI, experienced higher rates of regression than patients with slowly proliferating tumors, which could suggest a more frequent possibility of performing sphincter-preserving procedures in these tumors. However, this was not confirmed in surgical procedures. Therefore, we do not know yet if pretreatment BrdUrd LI assessment will be a good predictor for a locoregional failure. Berger et al.25, analyzing tumor sterilization after preoperative RT for rectal cancer, did not find a predictive factor for complete pathological response among such factors as age, sex, tumor stage, and pathologic grade. However, they found favorable influence of higher doses (>44 Gy) on pathologic stage. There is no known optimal time for the interval between RT and surgery. The Swedish group keeps the interval at about a week; however, in other institutions, using longer RT treatments and higher total dose, longer intervals—4 to 6 weeks were adopted14,25. The main reason for a longer interval is tumor regression, which makes sphincter preservation possible. Similar to Francois et al14, we observed higher clinical and pathologic response rate after longer interval between RT and surgery. However, these authors14 showed nonsignificantly better overall survival for patients treated with shorter interval. Withers and Haustermans33 estimated the interval between long course of fractionated RT (40–54 Gy) and surgery and stated that the interval is not critical to either local recurrence or distant metastases. The authors offered the following arguments: the tumor cells do not disseminate until the primary tumor is large enough to be clinically detectable (probably 80% of patients whose rectal tumors have not metastasized to lymph nodes will be free of metastases). Irradiation with a dose of 40 Gy in 2 Gy fractions (equivalent to 25 Gy in five fractions) reduces tumor cell survival by about six decades, e.g., from 1010 to 104 cells. However, we have to remember that although the short overall treatment duration in the 25 Gy in five-fraction regimen provides a radiobiological advantage, this is a relatively low dose34, which causes about a 66% reduction in the rate of local recurrence11. A retrospective analysis of published results of preoperative radiation therapy for rectal cancer showed that local control probability curves were displaced toward higher doses as the overall duration of preoperative radiation therapy was increased15. Therefore, longer intervals between short RT schedule (25 Gy) and surgery may be inappropriate in case of patients with incomplete resection (cut-through) of primary tumor, in whom the average subclinical cancer cell burden increases during long interval. Also, subclinical disease beyond the future surgical margins, may be a potential target for future recurrences. Longer intervals after short RT can be dangerous because of potential subclinical tumor, which may grow more quickly than primary tumor15,33, the and risk of developing distant metastases. If we imply that moderately differentiated adenocarcinoma cells have different metastatic and proliferative activities from poorly differentiated cancer cells, which was shown by Taniyama et al35, then we could have an indication to adjuvant chemotherapy for patients with differentiated tumors. The authors35 indicated that moderately differentiated cancer cells are associated with hematogenous metastases to the liver, and the loss of tubular formation of cancer cells in poorly differentiated tumors may be fundamentally related to lymph node metastases and infiltrative growth. Therefore, particularly in patients with moderately differentiated and slowly proliferating tumors, adjuvant chemotherapy could be suggested after OTT shorter than 4 weeks, to prevent developing metastases to the liver. In conclusion, our study shows that pretreatment BrdUrd LI or SPF were not predictive for early clinical and pathologic tumor response. After/before BrdUrd LI ratios showed inhibition of proliferation in responsive tumors, but this was not reflected in the number of sphincter preserving procedures performed. As 1 month after RT, accelerated proliferation of tumor cells is observed only in slowly proliferating tumors, we think that longer interval between RT and surgery is inadvisable. If late tumor response confirms that patients having tumors with increased proliferative activity have statistically significantly less recurrences and improved survival rates compared with patients with less proliferative tumors, then we will be able to suggest a prognostic factor for individual rectal cancer patient, and a basis for selection to postoperative adjuvant chemotherapy.	[ "early tumor response", "radiotherapy", "rectal cancer", "proliferation rate", "brdurdli" ]	[ "P", "P", "P", "P", "P" ]
Eur_J_Pediatr-4-1-2254655	Pneumomediastinum in the neonatal and paediatric intensive care unit	The incidence, aetiology and pathophysiology of pneumomediastinum (PM), an uncommon and potentially serious disease in neonates and children, were evaluated. A retrospective chart review of all patients diagnosed with PM who were hospitalised in the intensive care unit of the University Children’s Hospital Zürich, Switzerland, from 2000 to 2006, was preformed. We analysed the incidence, severity and causes of PM and investigated the possible differences between neonatal and non-neonatal cases. Seven children and nine neonates were identified with PM. All patients had a good outcome. Six cases of PM in the group of children older than 4 weeks were deemed to be caused by trauma, infection and sports, whereas one case was idiopathic. All nine neonatal cases presented with symptoms of respiratory distress. We were able to attribute four cases of neonatal PM to pulmonary infection, immature lungs and ventilatory support. Five neonatal cases remained unexplained after careful review of the hospital records. In conclusion, PM in children and neonates has a good prognosis. Mostly, it is associated with extrapulmonary air at other sites. It is diagnosed by chest X-ray alone. We identified mechanical events leading to the airway rupture in most children >4 weeks of life, whereas we were unable to identify a cause in half of the neonates studied (idiopathic PM). Introduction Pneumomediastinum (PM) is defined as a mediastinal air leak. The experimental works of Macklin and Macklin provided insights into its pathophysiology [4, 5]; alveolar rupture occurs because of a pressure gradient between the alveolus and the surrounding tissues. This gradient develops either through overinflation of the alveolus or a reduction of interstitial pressure. The air that subsequently leaks into the interstitial tissue diffuses toward the peribronchial and perivascular tissue, and then towards the mediastinum, the neck and into the subcutaneous tissue. However, due to pressure equalisation between the affected and adjacent alveoli in the lungs, the interalveolar walls remain intact and the lungs inflated. The diagnosis of PM is confirmed by frontal chest roentgenogram, including the cervical region. Typical radiological signs of PM include the continuous diaphragm sign (interposition of air between the pericardium and the diaphragm, which becomes visible in the central mediastinal part) and linear bands of mediastinal air parallelling the left side of the heart and the descending aorta (pleura is shown as a fine opaque line) with extension superiorly along the great vessels into the neck. In infants, the “spinnaker sign” (an upwards and outwards deviation of thymic lobes) can be seen when the thymus is raised above the heart by pneumomediastinal air that elevates the thymus and separates it from the cardiac silhouette beneath [2]. Various causes of PM are found in the literature, such as airway obstruction (e.g. foreign body aspiration), iatrogenic (e.g. mechanical ventilation), infections (e.g. pneumonia), obstructive lung disease (e.g. asthma), toxic effects (e.g. smoking), trauma (e.g. chest trauma), Valsalva manoeuvres (e.g. vomiting) and the weakness of tissue (e.g. anorexia nervosa). In spontaneous PM, the underlying lung is healthy and the air leak is thought to be atraumatic [3]. In neonates, known predisposing factors are mixed lung diseases, such as pneumonia or meconium aspiration syndrome, with coexisting atelectasis and airway obstruction [1]. However, only scarce literature is found about neonates with PM. In this study, we retrospectively analysed the incidence, severity and causalities of PM in neonates and children >4 weeks of life admitted to our intensive care unit, and we investigated the possible differences between the groups. Material and methods We retrospectively reviewed all records of children diagnosed with PM who were hospitalised in the interdisciplinary neonatal and paediatric intensive care unit of the University Children’s Hospital in Zürich, Switzerland, between January 2000 and September 2006. The patients were divided into two groups according to their age: neonates (under 4 weeks of age) and children (over 4 weeks of age). We were interested in the causes of PM as documented by the treating physicians, the types and results of radiologic investigations performed, any invasive interventions used to treat PM, the severity of the PM and the length of stay in the intensive care unit. Results About 1,200 children are admitted to our intensive care unit per year. The incidence of PM in our intensive care unit was 0.08% for children >4 weeks of age and 0.1% for neonates. In all patients, PM was diagnosed by chest X-ray and all had a positive outcome related to the PM. All five patients with pneumopericardium (PP) did not suffer from any complications (e.g. pericardial tamponade). Seven children were >4 weeks of age (Table 1). Their mean age was 7.1 years (range 1.3–15.8 years). In addition to PM, two children of this group had subcutaneous emphysema (SE), two a pneumothorax (PT) and two a PP. Different causes were found for the air trapping. There were two traumatic aetiologies (rib fracture after a severe car accident, lesion in the hypopharynx after a fall). Two children were diagnosed with obstructive bronchitis and in one child, barotrauma occurred intraoperatively due to a clamped expiratory tube during mechanical ventilation (Fig. 1). One child had exercised vigorously three days before hospitalisation, which may have caused the PM. In one adolescent, PM occurred spontaneously. All children were hospitalised in the intensive care unit for one to seven days, depending on the severity of their underlying disease. Only two children required pleural drainage and intubation. All other children were treated for their underlying conditions and received oxygen therapy. Diagnostics for the PM other than chest X-rays were performed in four patients. All of these had received a thoracic CT scan. In one child, who also had a huge subcutaneous emphysema and dysphagia, the reason for the air trapping could only be found by means of a laryngotracheoscopy, which showed a traumatic lesion in the hypopharynx. The patient’s history revealed that she had fallen onto a piece of wood by her neck. Table 1Results for the group of children >4 weeks of lifeAir leakAetiologyAge (years)Intubation (after diagnosis)Pleural drainageDiagnostic testsDays in ICUPM, PP, SESpontaneous/3 days earlier intensive sport15.8NoNoChest X-ray 3×2PM, PTTraumatic7.53 daysYesCT/chest X-ray 3×5PM, SE Traumatic (lesion in hypopharynx)1.3NoNoCT/chest X-ray 3×, oesophagogram with contrast medium, laryngotracheoscopy4PM, PPIatrogenic: equipment failure with barotrauma during mechanical ventilation2.3NoNoChest X-ray 2×2PMSpontaneous15.4NoNoChest X-ray/CT1PM Obstructive bronchitis5.6NoNoChest X-ray 2×/CT2PM, PTObstructive bronchitis1.97 daysYesChest X-ray 7×7Fig. 1Pneumomediastinum (PM), subcutaneous emphysema (SE) and pneumopericardium (PP) in a 2-year-old intubated patient The group of children older than 4 weeks stayed in the intensive care unit for a mean of 3.2 days (range 1–7 days), depending on the severity of the PM and the underlying disease. Compared to the neonatal group, the length of stay in the intensive care unit was shorter. However, most neonates stayed in the intensive care unit longer, primarily because of comorbid conditions and not because of the PM. We found nine neonates who were diagnosed with PM (Table 2); two premature and seven term infants, all of whom presented with signs of respiratory distress. Three neonates were also diagnosed with a PP, one with SE and five with a PT. Birth weight ranged from 2,150 g to 4,140 g (mean 3,340 g). Five children were born vaginally and four by caesarean section. All children were vigorous at birth and none required resuscitation with bag mask ventilation or surfactant. Before arriving in the intensive care unit, where the diagnosis of PM was confirmed by chest X-ray, two infants had received ventilatory support by CPAP and one of the premature infants had to be intubated for respiratory failure. During hospitalisation in the intensive care unit, two children deteriorated and required mechanical ventilation for three and four days, respectively, and two other children needed CPAP for a few hours. Only one child received pleural drainage. All children received oxygen therapy and specific therapy for their underlying disease. The age at admission to the intensive care unit ranged from a few hours to four days. One neonate was admitted to the intensive care unit due to convulsions and developed a PM on day six of life. The treating physicians felt that the PM may have been associated with a Valsalva manoeuvre, which occurred during the seizure. Other causes of PM were a pulmonary infection due to maternal infection and a possible barotrauma due to peak inspiratory pressure of 25 cm H2O in a mechanically ventilated premature neonate. Two newborns had PM related to CPAP and four neonates were diagnosed with spontaneous PM. Neonates stayed in the intensive care unit for 3–13 days (mean 5.6 days), depending on the severity of the underlying diseases. Table 2Results for the group of neonatesAir leakAetiologyBirth weightGestational age (weeks)Mode of deliveryMechanical ventilation before diagnosisDuration of ventilatory support after diagnosisPleural drainageDays in ICUPMSpontaneous4,140 g40 0/7VaginalNoNoNo4PM, PP, SEPremature lungs, barotrauma2,150 g34 4/7Caesarian section Pip max. 25 cm H203 days (intubation)No4PM, PT, PPPremature lungs, spontaneous or CPAP2,480 g35 6/7VaginalCPAP4 days (intubation)Yes6PM, PTSpontaneous3,485 g38 1/7Caesarian sectionNoNoNo3PM, PTSpontaneous3,440 g37 5/7VaginalNo6 hours (CPAP)No3PM, PPSpontaneous2,830 g39 1/7Caesarian sectionNoNoNo13PM, PTSpontaneous3,970 g38 5/7Caesarian sectionNoNoNo9PM, PTPulmonary infection due to maternal infection3,440 g38 5/7VaginalCPAP1 day (CPAP)No5PMConvulsions or spontaneous4,130 g40 5/7VaginalNoNoNo4 Discussion All children with PM had a good outcome without any complications due to air trapping. In the group of children older than 4 weeks, only two children developed a respiratory insufficiency, leading to mechanical ventilation. In both of them, respiratory failure was related to their underlying condition (polytrauma with haematothorax and severe obstructive bronchitis, respectively). These were also the children who stayed longest in the intensive care unit. All other children were treated with oxygen only and stayed in the intensive care unit until they improved clinically and radiographically. Regarding radiologic diagnostics, four patients of the group of children >4 weeks of life had CT scans (three of them had been done in outside clinics from where the patients had been admitted to our intensive care unit). Retrospectively, the utility of the CT scans was put into question, as these scans did not change the patient management. The only patient in whom the CT scan changed management was the child with polytrauma. In this patient, other intrathoracic injuries needed to be ruled out. In the group of neonates, it was much more difficult to find the aetiology of the PM, since all neonates had presented with respiratory disease, and radiologic investigations were partially performed only after the use of CPAP or tracheal intubation. Five of the nine neonates had a spontaneous PM without risk factors, such as mechanical respiratory support (bag mask ventilation after birth, CPAP, mechanical ventilation) or restrictive lung disease. Three of these five babies were delivered by caesarean section. In the remaining four newborns, possible mechanical incidents leading to the air leak could be revealed: mechanical ventilation with high inspiratory pressure, CPAP, pulmonary infection and convulsion. Further investigations are needed to find the aetiology of spontaneous PM in healthy, term neonates. In conclusion, PM in children and neonates has a good prognosis. Mostly, it is associated with extrapulmonary air at other sites. It is diagnosed by chest X-ray alone. Whereas in older children mechanical events leading to the airway rupture can be revealed in most cases, about half of the neonates in our series suffered from PM without obvious reason.	[ "pneumomediastinum", "air leak", "pneumopericardium", "subcutaneous emphysema", "respiratory distress syndrome" ]	[ "P", "P", "P", "P", "R" ]
Law_Hum_Behav-4-1-2175020	Policy Forum: Studying Eyewitness Investigations in the Field	This article considers methodological issues arising from recent efforts to provide field tests of eyewitness identification procedures. We focus in particular on a field study (Mecklenburg 2006) that examined the “double blind, sequential” technique, and consider the implications of an acknowledged methodological confound in the study. We explain why the confound has severe consequences for assessing the real-world implications of this study. One of the most interesting products of the first wave of wrongful convictions exposed by DNA has been a vigorous debate over potential changes in the design and execution of the lineups and photographic arrays, familiar to every television viewer, that police rely on to probe memory in eyewitness cases, the category that dominates the exoneration lists. All of the current proposals for change in investigative practice derive from extensive laboratory inquiry, and they have at their cores the novel “double-blind, sequential” technique for conducting eyewitness identification procedures. In this technique the law enforcement personnel conducting an identification procedure are “blind” concerning which person in the lineup or photo array is the police suspect, and they present the “fillers” and the suspect to the witness individually (“sequentially”) rather than in a group (“simultaneously”), as in the traditional practice. The changes from current procedure are designed to ensure that witnesses discern no inadvertent cues as to which individual they should or should not identify, to encourage witnesses to compare each individual they see to the remembered image of the criminal (rather than to make a relative, “looks-most-like,” judgment comparing the individuals displayed to each other), and to eliminate unnecessary “feedback” to witnesses who have made a selection and might look to the lineup administrator for confirmation or contradiction. Everyone agrees that proposed changes in investigative practice should be tested in the field, but moving from the laboratory to the field has always been problematic. The proper design of field studies used to evaluate new procedures in the field has become an important issue. Two recent efforts at field-testing the “double-blind, sequential” option have taken place. The first, conducted by several departments in Hennepin County, Minnesota, produced results consistent with those predicted by the laboratory scientists, but made no explicit comparison to traditional practices, and it has not been controversial. (Klobuchar et al. 2006). The second field study, conducted in three Illinois jurisdictions under the direction of the general counsel for the Chicago Police, Sheri Mecklenburg, and documented at length in a report (usually referred to as “The Mecklenburg Report,” after its author) appeared to contradict both the laboratory scientists’ predictions and the sparse existing field data on eyewitness performance (Mecklenburg 2006). The Mecklenburg report stated that in two of the three jurisdictions reporting, the traditional method of an aware, “not-blind” detective displaying the suspect and “fillers” in a group to the witness produced a lower rate of identifications of innocent fillers and a higher rate of identifications of suspects than did the lab-generated “double-blind, sequential” technique. The recommendation of the Mecklenburg Report, in other words, was that the system should not institute changes on the basis of the laboratory science. The Mecklenburg Report was vigorously publicized, and it immediately drew both determined support and sharp criticism from psychologists who had long been interested in the issue of eyewitness investigative procedures. Unfortunately for criminal justice practitioners who must decide whether procedures should be changed, the early scientific commentaries on the Mecklenburg Report generally aligned with the views on the potential of these particular procedural innovations that the commentators had announced throughout their long careers of involvement with the issue of eyewitness memory. Seizing on this, partisans on both sides of the debate over procedures have unfairly dismissed some criticism and praise of the Mecklenburg Report as reflecting nothing more than the scientific commentators’ stubborn loyalty to their own pre-existing beliefs. A standoff has arisen. Although everyone agrees that further field studies are required, practitioners considering future field studies have been left to wonder whether they should simply repeat the Illinois Study described in the Mecklenburg Report, or attempt to find a new design. We have read the materials related to the Mecklenburg study, including the Mecklenburg Report, its Addendum and Appendices, the supportive comments of Dr. Roy Malpass (2006) and Dr. Ebbe Ebbesen (2006), and the critical comments of Dr. Gary Wells (2006) and Dr. Nancy Steblay (2006). The Report indicates, and all commentators seem to agree, that the study does contain a confound: a non-blind simultaneous procedure is compared with a blind sequential procedure. The bottom line issue here, or at least the one that drew our group’s attention, concerns the importance of the confound. It is easy to understand the sentiment expressed by Mecklenburg in her Addendum that not all variables can be controlled in a field study such as the one she designed and describes in the Report. Confounds can occur in laboratory studies as well as field studies. The issue that always arises in such cases concerns the implications of the confound: Is it critically related to interpreting the major outcome of the study? Or is the confound incidental to the main conclusion, such that even though the confound is acknowledged, the major results of the study are still interpretable? The Mecklenburg Report asserts that “The Illinois Pilot Study was properly designed to answer the question: how do the current procedures compare with the proposed procedures, both in terms of identification rates and implementation?” From this perspective, the confound between blind/non-blind and sequential/simultaneous would not be critical, because non-blind simultaneous reflects the current procedure to which the blind/sequential procedure is compared. Unfortunately, this perspective seems seriously problematic. Our reading of the materials forces us to conclude that the confound has devastating consequences for assessing the real-world implications of this particular study. If it is the case that the better outcome from the non-blind/simultaneous procedure is partly or entirely attributable to subtle, unintentional cues provided by the administrator, then the Illinois results may simply underscore that the present procedure produces a biased outcome that may ultimately result in the increased conviction of innocent individuals. Stated slightly differently, it is critical to determine whether the seemingly better result from the simultaneous procedure is attributable to properties of the simultaneous procedure itself, or to the influence of the non-blind administrator. We should note that under these testing conditions, if the results had shown the sequential lineup to be superior, one would not know whether it was really the use of the sequential lineup or the use of a blind investigator conducting the lineup that produced the result. Of course, any difference between conditions could be due to some combination of the factors. Even if no difference in outcome occurred between the procedures, one could not safely conclude there is no difference between them if the detectives were informed in one condition and not in the other. Thus, although the conditions used in the study made some sense from a practical standpoint, the design guaranteed that most outcomes would be difficult or impossible to interpret. The only way to sort this out is by conducting further studies including, at a minimum, a blind/simultaneous condition (it would also be desirable to include a non-blind/sequential condition to fill out the design, but it is not an absolutely necessary condition for the present purposes). In the materials we have reviewed, the Mecklenburg Report’s detractors (including Wells) and its advocates (including Mecklenburg herself) disagree on whether the misidentification rates in two of the three participating jurisdictions (a zero rate of “filler” identifications) are suspiciously low. However, Wells cites enough evidence that they may be low to justify the concern that administrator bias is operating, either consciously or unconsciously; either by failing to count tentative “filler” choices, or in steering witnesses away from fillers, or toward suspects. The problem is that we cannot know on the basis of the Mecklenburg study whether such bias is operating, even though the entire interpretation of the significance of the study for real-world practices hinges on this issue. Mecklenburg states in her Addendum that the question of how blind administrators affect simultaneous lineups is one of several questions to be addressed in future studies. We certainly hope so. But the statement that follows is problematic: “However, the Illinois Pilot Program was not intended to answer those questions and any attempt to discredit the Illinois study on that basis is misguided.” If the Illinois study was not designed to address the question of what happens in a blind/simultaneous line-up, given its centrality to the issue, then our assessment is that the Illinois study addressed a question (comparing blind/sequential and non-blind/simultaneous) that is not worth addressing, because the results do not inform everyday practice in a useful manner. No single field study can produce a final blueprint for procedural reform; we will need many. The design of these studies, however, will be crucial. A well-designed field study that avoids the flaw built into the Illinois effort, can be an important first step toward learning what we need to know about the best practices in identification procedures.	[ "eyewitness identification", "field studies", "double blind sequential procedure" ]	[ "P", "P", "R" ]
Matern_Child_Health_J-2-2-1592248	The National Summit on Preconception Care: A Summary of Concepts and Recommendations	The Centers for Disease Control and Prevention (CDC) and 35 partner organizations have engaged in developing an agenda for Preconception Health. A summit was held in June 2005 to discuss the current state of knowledge regarding preconception care and convene a select panel to develop recommendations and action steps for improving the health of women, children, and families through advances in clinical care, public health, and community action. A Select Panel on Preconception Care, convened by CDC, deliberated critical related issues and created refined definition of preconception care. The panel also developed a strategic plan with goals, recommendations, and action steps for improving preconception health. The recommendations and action steps are specific to the implementation of health behavior, access, consumer demand, research, and surveillance activities for monitoring and improving the health of women, children and families. The outcome of the deliberations is the CDC publication of detailed recommendations and action steps in the Morbidity and Mortality Weekly Report series, Recommendations and Reports. Introduction In June 2005 the Centers for Disease Control and Prevention (CDC) and the March of Dimes, in collaboration with 35 professional and governmental organizations, convened a 3-day summit to discuss an agenda for preconception care programs, research, and policy. The summit was the result of internal workgroup discussions at CDC and an initial meeting with external stakeholders in November 2004. Prior to the November meeting, subject matter experts in 19 programs from 8 centers throughout CDC conducted a detailed review of the relevant literature. The 3-day summit was divided evenly into two components: a presentation series from preconception care practitioners, followed by a select panel meeting. More than 60 presentations on preconception care research and programs were made and subsequently discussed by the select panel to generate recommendations for improving maternal and child health outcomes through better preconception care. The goals of the select panel discussions were to define current scientific knowledge, to identify best practices, and to highlight key issues needing further attention that could be used as the basis for formulating recommendations and action steps. Deliberations of the select panel The select panel consisted of nationally recognized experts from a variety of disciplines representing different perspectives on preconception care services designed to promote women's health and reduce adverse pregnancy and perinatal outcomes. The panel included experts in the fields of obstetrics, family practice, pediatrics, public health, nursing, reproductive health, toxic exposures, and chronic and infectious disease. The initial conceptual framework for deliberations and discussions was based on the acronym “ACT” (Accessible health care, Comprehensive care, and Timely provision of care) proposed by Dr. Jennifer Howse, president of the March of Dimes. It was then expanded to include six “A’s”: access, availability, affordability, acceptability, accuracy, and appropriateness. Five organizing themes (clinical practice, social marketing and health promotion, public health and community, public policy and finance, data and research) served as the basis for discussion. One key topic of debate was the potential conflict in characterizing preconception care as preparation for pregnancy, as opposed to the broader promotion of women's health. Some panelists believed strongly that the focus must be on both women and infant health outcomes. In moving forward, the panel was conscientious not to carve preconception care out separately from good routine primary care; however, panel members also asserted that even though the focus should be on preconception care, providers should also recognize that the scope of care should include comprehensive women's health services. As one panel member said, “Preconception care should be happening at every interaction with a woman or man of reproductive age…. It's part of what we’re already doing in primary care. We’re just trying to get providers to reframe their thinking so we’re achieving preconception health.” Some, while agreeing, argued that the ultimate goal was to improve perinatal outcomes, making this a part of, but distinct from, all well-woman's health care. The status of current research emerged as a central issue. Discussions often centered around the concern that insufficient scientific evidence currently exists for many preconception care interventions, for the best methods of integrating them into primary care, and for effectively delivering interventions as a package—or even if they work. Many panelists cited the need for intervention trials. In contrast, other members believed that existing research findings were sufficient, and that moving in the direction of translation and action should be the next step. This theme was captured in one member's comment, “Research is important; there's no doubt about it. But instead of just creating more new knowledge, let's ask ourselves: Why is it that the knowledge we created before isn't getting translated into action?” Ultimately, the panel's recommendations reflected both areas where sufficient evidence exists and those where more research is needed. Many panel members supported recommendations that would advocate provider assistance for a woman and her partner in developing a reproductive life plan and in communicating her/their intentions through the implementation of that plan. As one panel member described it, “[Primary care providers can] develop the ability to help a woman write her reproductive health plan and facilitate her ability to carry this out, including the upgrading of this plan throughout her reproductive years.” Some participants suggested that the recommendations should strongly encourage all providers not only to ask a woman about her reproductive intentions, but to ask in a manner that conveyed to the woman that her decisions should be based on her personal preferences. This would allow the provider to help her choose the best contraceptive for meeting her life goals and to provide appropriate preconception awareness information. Panelists generally agreed that the recommendations should focus on feasible actions that could have the greatest positive impact on health outcomes. This resulted in much discussion of fundamental issues, including 1) who gets preconception care—women at high risk for adverse pregnancy outcomes or all women; 2) what would yield the greatest return on investment (e.g., focus on the preconception period or the interconception period; focus on all women or high-risk women); 3) whether interventions could be packaged and targeted to reach specific populations or address specific needs of women; and 4) how finance case and services, as well as the potential cost effectiveness and cost benefit of the proposed interventions? Defining preconception care The select panel worked to refine the definition of preconception care. The participants agreed that it is important to recognize what preconception care is and what it is not. Preconception care is not a single visit but a continuum of care designed to meet the needs of a woman throughout the various stages of her reproductive life. The goal of the preconception care process is to make sure that the woman is healthy as she attempts to become pregnant [1–6], and to promote her health and the health of her children throughout her reproductive lifespan. Based on the Select Panel deliberations and its subsequently published recommendations, preconception care is defined here as a set of interventions that aim to identify and modify biomedical, behavioral, and social risks to a woman's health or pregnancy outcome through prevention and management, emphasizing those factors that must be acted on before conception or early in pregnancy to have maximal impact. Thus, it is more than a single visit and less than all well-woman care. It includes care before a first pregnancy or between pregnancies (commonly known as interconception care) [7]. Preconception care and interventions are designed to reduce perinatal risk factors and, for optimal effectiveness, must be successfully implemented before the start of pregnancy [4]. The components of preconception care, including an array of interventions to address medical, psychosocial, and environmental risks associated with childbearing, augment routine well-woman care. Current Guidelines for Perinatal Care (jointly issued by the American Academy of Pediatrics and American College of Obstetricians and Gynecologists) recommend that “All health encounters during a woman's reproductive years, particularly those that are a part of preconception care, should include counseling on appropriate medical care and behavior to optimize pregnancy outcomes” [8, 9]. There is scientific evidence that certain preconception interventions are effective, yet at present they are not routinely and systematically applied. Fig. 1Ecological model The framework for preconception care recommendations The framework for the recommendations and action steps to improve preconception health is based on two well-established models for community health and behavior change, the ecological and the lifespan models. The ecological model for medical care is constructed with the woman being the central player, while recognizing the significant roles of the larger community, the providers, and the institutions in ensuring the best outcomes for her (Fig. 1). The four levels of this model interact and are not distinct. Each level has primary responsibility for different aspects of health and well-being; however, all levels are integrated and each plays a supportive role in the success of the other levels and the system as a whole. If one level is not fully engaged, the system will weaken, and the result may be poor health for the individual woman. In many regards the community and institution levels overlap; institutions operate within the community, just as the community determines what services the institutions provide. Using this framework, the proposed recommendations outline actions and responsibilities at each level and demonstrate the role that all sectors have in improving preconception care and ultimately the health of women and their families. The recommendations and action steps described here aim to address needs and gaps at each level. The second model influencing the framework for preconception care is the lifespan perspective [10]. This framework recognizes that health is determined by factors across the lifespan and is intergenerational. Improving preconception health will require, first, that the woman herself is achieving optimal health throughout her life. With women living in a healthy state, healthy children are more likely to result. As one panel member said, “We should build on the pediatric model of anticipatory guidance… [and] have consistent messages and care for women through the lifespan that will address women's and preconception health before, between and beyond.” Furthermore, good health in infancy is likely to carry through to childhood, adolescence, young adulthood and older adulthood so that healthy infants become healthy adults and achieve their reproductive goals. This lifespan framework extends beyond the individual to promote health in the family and community. Thus, the recommendations and action steps both address health needs across the lifespan and recognize that the individual, family, community, and institutions each have roles in ensuring that all women have lifelong optimal health. Challenges for implementing the proposed frameworks Maximizing the potential impact of preconception interventions will require integration across the various levels of care and areas of expertise [5, 11–13]. Changes in individual behavior, clinical practice, community programs, and public policies are recommended. In the context of clinical care, much can be done by primary care providers, but evidence suggests that many women have two primary care providers—one with expertise in obstetrics and gynecology and one with general medical training. In addition, the U.S. health-care system is fragmented into highly specialized components. Nevertheless, bundling the various preconception interventions has the potential to improve perinatal outcomes and reduce costs associated with adverse outcomes. These recommendations call for more integrated delivery of health promotion and care services for women rather than for greater fragmentation of the health-care system. Integration is critical to moving forward, In consideration of competing priorities and limited resources, integration may achieve health-care economies of scale that produce more efficient delivery of services and more effective results in improving preconception health. For example, primary care providers potentially could routinely and efficiently use the tools available to screen for smoking, alcohol abuse, genetic risks, and occupational hazards; endocrinologists, geneticists, and nutritionists could become an integral part of a preconception care strategy. Moreover, settings such as family planning or sexually transmitted disease clinics must become part of an integrated approach in order to reach many women at high risk for adverse pregnancy outcomes. New models of care and well-designed quality-improvement efforts could foster the integration of preconception interventions. At the same time, preconception care must be tailored to meet the needs of the individual woman. Given that preconception care is a process, not a visit, some recommendations will be more relevant to women at specific stages in their lives or those with varying levels of health risks. Health promotion efforts, risk screening, and interventions would be different for a young woman who has never experienced pregnancy than for a 35-year-old woman who has had three children. Women with chronic diseases, prior pregnancy complications, or behavioral risk factors may need more intensive interventions [5, 6, 11, 12]. At the clinical level, it is especially important that individual interventions be tailored to the specific needs of the woman. Such variations in interventions also place constraints on how interventions can and should be bundled or prepackaged. This special issue of the Maternal and Child Health Journal, as well as the MMWR Recommendations and Reports, emphasizes components of preconception care supported by research and an evidence base that documents the efficacy of specific interventions [7]. For example, that increasing pre-pregnancy intake of folic acid reduces neural tube defects, and controlling diabetes prior to pregnancy yields better outcomes for mothers and babies [14, 15]. As with prenatal care or well-child care, however, evidence may not be forthcoming that one regimented process of preconception care significantly improves outcomes. Rather, the process of providing individualized screening, health education, and necessary interventions that can yield better outcomes. Finally, implementing these recommendations to improve preconception health will take time. Diffusion of innovation theory demonstrates how concepts and best practices tend to be disseminated and adopted [16]. The innovators who lead current efforts to advance preconception care and the early adopters who run model programs are making only the first steps toward larger societal change. Whereas the select panel's action steps were designed to be short-term efforts, better implementing professional standards of care, modifying physician behaviors, developing effective health promotion messages, and adjusting payment mechanisms all will take time. A strategic plan for improving women's health and pregnancy outcomes Based on the June summit presentations and subsequent expert panel deliberations, 10 recommendations were developed (Table 1). These recommendations were based on four overarching goals that were identified as critical to help women reach optimal health and realize their reproductive goals. These four broad goals are 1) to improve both men's and women's knowledge, attitudes, and behaviors related to preconception health; 2) to ensure that all U.S. women of childbearing age receive preconception care services—screening, health promotion, and interventions—that will enable them to begin a pregnancy in optimal health; 3) to reduce risks indicated by a prior adverse pregnancy outcome through interventions during the interconception (inter-pregnancy) period that can prevent or minimize health problems for a mother and her future children; and 4) to reduce the health disparities in adverse pregnancy outcomes. Table 1.Panel Recommendations and ActionsRecommendation 1. Individual responsibility across the lifespan. Each woman, man and couple should be encouraged to have a reproductive life plan.Action Steps• Develop, evaluate, and disseminate reproductive life planning tools for women and men in their childbearing years, respecting variations in age; literacy, including health literacy; and cultural/linguistic contexts• Conduct research leading to development, dissemination, and evaluation of individual health education materials for women and men regarding preconception risk factors, including materials related to biomedical, behavioral, and social risks known to affect pregnancy outcomesRecommendation 2. Consumer awareness. Increase public awareness of the importance of preconception health behaviors and preconception care services by using information and tools appropriate across various ages; literacy, including health literacy; and cultural/linguistic contextsAction Steps• Develop, evaluate, and disseminate age-appropriate educational curricula and modules for use in school health education programs• Integrate reproductive health messages into existing health promotion campaigns (e.g., campaigns to reduce obesity and smoking)• Conduct consumer-focused research necessary to develop messages and terms for promoting preconception health and reproductive awareness• Design and conduct social marketing campaigns necessary to develop messages for promoting preconception health knowledge and attitudes, and behaviors among men and women of childbearing age• Engage media partners to assist in depicting positive role models for lifestyles that promote reproductive health (e.g., delaying initiation of sexual activity, abstaining from unprotected sexual intercourse, and avoiding use of alcohol and drugs)Recommendation 3. Preventive visits. As a part of primary care visits, provide risk assessment and educational and health promotion counseling to all women of childbearing age to reduce reproductive risks and improve pregnancy outcomesAction Steps• Increase health provider (including primary and specialty care providers) awareness regarding the importance of addressing preconception health among all women of childbearing age• Develop and implement curricula on preconception care for use in clinical education at graduate, postgraduate, and continuing education levels• Consolidate and disseminate existing professional guidelines to develop a recommended screening and health promotion package• Develop, evaluate, and disseminate practical screening tools for primary care settings, with emphasis on the 10 areas for preconception risk assessment (e.g., reproductive history, genetic, environmental risk factors)• Develop, evaluate, and disseminate evidence-based models for integrating components of preconception care to facilitate delivery of and demand for prevention and intervention services• Apply quality improvement techniques (e.g., conduct rapid improvement cycles, establish benchmarks and brief provider training, use practice self-audits, and participate in quality improvement collaborative groups) to improve provider knowledge and attitudes, and practices and to reduce missed opportunities for screening and health promotion• Use the federally funded collaboratives for community health centers and other FQHC to improve the quality of preconception risk assessment, health promotion, and interventions provided through primary care• Develop fiscal incentives for screening and health promotionRecommendation 4. Interventions for identified risks. Increase the proportion of women who receive interventions as follow-up to preconception risk screening, focusing on high priority interventions (i.e., those with evidence of effectiveness and greatest potential impact)Action Steps• Increase health provider (including primary and specialty care providers) awareness concerning the importance of ongoing care for chronic conditions and intervention for identified risk factors• Develop and implement modules on preconception care for specific clinical conditions for use in clinical education at graduate, postgraduate, and continuing education levels• Consolidate and disseminate existing professional guidelines related to evidence-based interventions for conditions and risk factors• Disseminate existing evidence-based interventions that can be used in primary care settings (e.g., brief interventions for alcohol misuse and smoking)• Develop fiscal incentives (e.g., pay for performance) for risk management, particularly in managed care settings• Apply quality improvement techniques and tools (e.g., conduct rapid improvement cycles, establish benchmarks, use practice self-audits, and participate in quality improvement collaborative groups)Recommendation 5. Interconception care. Use the interconception period to provide additional intensive interventions to women who have had a previous pregnancy which ended in adverse outcome (e.g., infant death, fetal loss, birth defects, low birthweight or preterm birth)Action Steps• Monitor the percentage of women who complete postpartum visits (e.g. using HEDIS measures for managed care plans and Title V Maternal Child Health Block Grant state measures), and use these data to identify communities of women at risk and opportunities to improve provider follow-up• Develop, evaluate, and replicate intensive evidence-based interconception care and care coordination models for women at high social and medical risk• Enhance the content of postpartum visits to promote interconception health• Use existing public health programs serving women in the postpartum period to provide or link to interventions (e.g., family planning, home visiting, and WIC)• Encourage additional states to develop preconception health improvement projects with funds from the Title V Maternal Child Health Block Grant, Prevention Block Grant, and similar public health programsRecommendation 6. Prepregnancy check up. Offer, as a component of maternity care, one prepregnancy visit for couples and individuals planning pregnancyAction Steps• Modify third party payer rules to permit payment for one prepregnancy visit per pregnancy, including development of billing and payment mechanisms• Consolidate existing professional guidelines to develop the recommended content and approach for such a visit• Educate women and couples regarding the value and availability of prepregnancy planning visitsRecommendation 7. Health insurance coverage for women with low incomes. Increase public and private health insurance coverage among women with low incomes to improve access to preventive women's health, preconception, and interconception careAction Steps• Improve the design of family planning waivers by permitting states (by federal waiver or by creating a new state option) to offer interconception risk assessment, counseling, and interventions along with family planning services. Such policy developments would create new opportunities to finance interconception care• Increase health coverage among women who have low incomes and are of childbearing age by using federal options and waivers under public and private health insurance systems and the State Children's Health Insurance Program• Increase access to health-care services through policies and reimbursement levels for public and private health insurance systems to include a full range of clinicians who care for womenRecommendation 8. Public health programs and strategies. Integrate components of preconception health into existing local public health and related programs, including emphasis on interconception interventions for women with previous adverse outcomesAction Steps• Use federal and state agency support to encourage more integrated preconception health practices in clinics and programs• Provide support for CDC programs to develop, evaluate, and disseminate integrated approaches to promote preconception health• Analyze and evaluate the preconception care activities used under the federal Healthy Start program and support replication projects• Convene or use local task forces, coalitions, or committees to discuss opportunities for promotion and prevention in preconception health at the community level• Develop and support public health practice collaborative groups to promote shared learning and dissemination of approaches for increasing preconception healthRecommendation 9. Research. Increase the evidence base and promote the use of the evidence to improve to preconception healthAction Steps• Prepare an updated evidence-based systematic review of all published reports on science, programs, and policy (e.g., through the Agency for Healthcare Research and Quality)• Encourage and support evaluation of model programs and projects, including integrated service delivery and community health promotion projects• Conduct quantitative and qualitative studies to advance knowledge of preconception risks and clinical and public health interventions, including knowledge of more integrated practice strategies and interconception approaches• Design and conduct analyses of cost-benefit and cost-effectiveness as part of the study of preconception interventions• Conduct health services research to explore barriers to evidence-based and guidelines-based practice• Conduct studies to examine the factors that results in variations in individual use of preconception care (i.e., barriers and motivators that affect health-care use)• Support activities to translate research into clinical practice and public health actionRecommendation 10. Monitoring improvements. Maximize public health surveillance and related research mechanisms to monitor preconception healthAction Steps• Apply public health surveillance strategies to monitor selected preconception health indicators (e.g., folic acid supplementation, smoking cessation, alcohol misuse, diabetes, and obesity)• Expand data systems and surveys (e.g., PRAMS and NSFG) to monitor individual experiences related to preconception care• Use geographic information system techniques to target preconception health programs and interventions to areas where high rates of poor health outcomes exist women of reproductive age and their infants• Use analytic tools (e.g., PPOR) to measure and monitor the proportion of risk attributable to the health of women before pregnancy• Include preconception, including interconception, health measures and population-based performance monitoring systems (e.g., in national and state Title V programs)• Include a preconception measure in the Healthy People 2020 objectives• Develop and implement indicator quality improvement measures for all aspects of preconception care. For example, use HEDIS measures to monitor the percentage of women who complete postpartum visits For each of the 10 recommendations, the panelists identified specific actions. For each action step, those persons primarily responsible for implementation as well as those who had supportive roles were listed. The action steps were designed as feasible and practical activities that could be undertaken in the near future and could result in change in the next 2 to 5 years. These goals, recommendations, and action steps form a strategic plan for improving preconception health and pregnancy outcomes. This can be illustrated as a pyramid (Fig. 2), building from the smaller attainable actions steps to fulfilling a recommendation, to attainment of the four goals, and finally to the combination of the goals to reach the pinnacle of the pyramid—a vision for better women's health and improved pregnancy outcomes. Fig. 2Preconception care pyramid The recommendations presented here focus on individual health knowledge and behavior, clinical care, public health programs, and health-care policies, improving preconception health will require the involvement of and changes in other sectors, including education, housing, urban planning, and environmental health. Moving forward, these sectors should be included as part of the comprehensive solution to improving women's health and, by extension, the health of families. In all of the action steps, women and their families have a critical role to play in engaging the system and developing the consumer demand and culture to normalize preconception care as part of women's health promotion and routine care. The focus of the recommendations The 10 recommendations outlined in Table 1 address individual, provider, system, and research needs that mirror the different levels described in the ecological model. Under each of the recommendations there are multiple actions steps that describe specific activities geared towards achieving the recommendations. The first two recommendations focus on individual responsibility for preconception health. Under these recommendations, action steps are focused on the development, evaluation, and dissemination of tools to help women and their partners make decisions regarding their reproductive health across the lifespan. The action steps clearly highlight the need to ensure that these tools are age-appropriate and culturally relevant and cover both general health topics and specific risk behaviors. The action steps recognize the importance of being able to integrate preconception health-care messages with existing health promotion activities whenever possible. This is especially important in an environment with limited resources for health promotion activities. Both recommendations 1 and 2 call for action to be taken at both the individual and community level so that societal norms shift toward supporting optimal preconception health behaviors. Recommendations 3, through 6 call for actions to improve health-care services and are particularly focused on changing provider knowledge, attitudes, and behaviors. Because the existing knowledge base has not been widely disseminated, the action steps call for new and continuing education activities to help all providers improve their skills and coverage of preconception care services. The use of quality improvement tools and techniques are suggested as a means of changing professional practices. In the current market-driven environment, it is clear that incentives such as pay for performance and risk management activities also must be developed to help encourage the provision of these services. Recommendations 5 and 6 particularly focus more on improving preconception care and health for specific groups of women. Recommendation 5 focuses on interconception care and the opportunity for the prevention of adverse pregnancy outcomes among those who are at risk for or have previously had pregnancy complications. In addition to action steps aimed at modifying provider practices, others encourage public health programs to do more identification and follow-up of women at risk. Recommendation 6 suggests pre-pregnancy checkups or visits for couples to focus on their health and risks when they are trying to conceive. The action steps include consumer education and consolidated professional guidelines, as well as better third-party health insurance coverage of such visits as a part of prenatal and maternity coverage. Action steps for recommendation 7 call for increased health-care coverage among uninsured, low-income women, specifically through Medicaid. Currently, most low-income women do not qualify for Medicaid unless they are pregnant; unfortunately, this is often too late for some health promotion and therapeutic interventions to have maximum impact on the health of the woman and her child. Recommendation 8 describes a series of action steps that public health and community programs can take to improve preconception health by increasing the access to and use of preconception care services. These steps include using publicly funded women's and children's health programs to promote preconception health, to screen for health risks, and to refer women at risk to appropriate clinicians. For example, preconception health promotion messages are complementary to the purposes of family planning and HIV/STD clinics. However, the use of publicly funded women's and children's health programs for these purposes will require that federal and state public health agencies minimize the categorical restrictions that often prevent true integration at the community level. The second two action steps encourage the adaptation, implementation, and evaluation of community-based programs that provide preconception services. There are model programs currently being implemented that could be modified and used to meet the unique needs of specific populations. A key component of these actions is to engage the community to develop the best methods for development, implementation, and evaluation of local preconception care programs. Without community support and designs that truly represent the needs of the communities, programs will be ineffective. Recommendations 9 and 10 include action steps focused on the continuous quality improvement and planning feedback loop that supports excellence in health-care systems. Research and evaluation are critical to examining progress toward achieving goals. Currently, a need exists for updating the systematic reviews of existing literature on preconception care interventions as well as increasing the knowledge base through qualitative and quantitative research projects. Economic analyses of preconception care interventions are not often conducted, and a great deal of work must be done to understand the impact of interventions. Limitations exist in many of the interventions currently available; and effective interventions have not been developed in some areas. Meeting the need for research and evaluation is a key factor in the ability to continue moving the field forward. In addition to research activities, ongoing monitoring systems—typically managed by public health agencies—are critical to program planning, development, and evaluation. This recommendation includes the use of surveillance systems such as the Pregnancy Risk Assessment and Monitoring System (PRAMS) and the Health Employee Data and Information Set (HEDIS) measures. Conclusion Improving preconception health requires changes in the knowledge, attitudes, and behaviors of individuals, families, communities, and institutions (e.g., government, health-care settings). The purpose is to improve the health of each woman before any pregnancy and to thereby affect the future health of the woman, her child, and her family. Through the CDC Preconception Health Initiative and the deliberations of the Select Panel on Preconception Care, a set of goals, recommendations, and specific actions steps have been developed to form a strategic plan. The framework incorporates both an ecological model and a lifespan perspective on health. This work also recognizes the unique contributions and challenges faced by individual women and their families, communities, and institutions. As one Select Panel member put it, “If you raise the level of health of women in a society, you’ve raised the level of health of families and the community at the same time.” The successful implementation of the recommendations will help achieve the vision for preconception health and pregnancy outcomes where All women and men of childbearing age have high reproductive awareness (e.g., understand risk factors related to childbearing).All women have a reproductive life plan (e.g., whether or when they wish to have children, how they will maintain their reproductive health).All pregnancies are intended and planned.All women of childbearing age have health-care coverage.All women of childbearing age are screened prior to pregnancy for risks related to the outcomes of pregnancy.Women with a prior pregnancy loss (e.g., infant death due to very low birthweight or preterm birth) have access to intensive interconception care aimed at reducing their risks.	[ "preconception care", "health behavior", "maternal health", "research and surviellance", "access to care" ]	[ "P", "P", "R", "M", "R" ]
Matern_Child_Health_J-2-2-1592148	Preconception Care Between Pregnancies: The Content of Internatal Care	For more than two decades, prenatal care has been a cornerstone of our nation’s strategy for improving pregnancy outcomes. In recent years, however, a growing recognition of the limits of prenatal care and the importance of maternal health before pregnancy has drawn increasing attention to preconception and internatal care. Internatal care refers to a package of healthcare and ancillary services provided to a woman and her family from the birth of one child to the birth of her next child. For healthy mothers, internatal care offers an opportunity for wellness promotion between pregnancies. For high-risk mothers, internatal care provides strategies for risk reduction before their next pregnancy. In this paper we begin to define the contents of internatal care. The core components of internatal care consist of risk assessment, health promotion, clinical and psychosocial interventions. We identified several priority areas, such as FINDS (family violence, infections, nutrition, depression, and stress) for risk assessment or BBEEFF (breastfeeding, back-to-sleep, exercise, exposures, family planning and folate) for health promotion. Women with chronic health conditions such as hypertension, diabetes, or weight problems should receive on-going care per clinical guidelines for their evaluation, treatment, and follow-up during the internatal period. For women with prior adverse outcomes such as preterm delivery, we propose an internatal care model based on known etiologic pathways, with the goal of preventing recurrence by addressing these biobehavioral pathways prior to the next pregnancy. We suggest enhancing service integration for women and families, including possibly care coordination and home visitation for selected high-risk women. The primary aim of this paper is to start a dialogue on the content of internatal care. For more than two decades, prenatal care has been a cornerstone of our nation’s strategy for improving pregnancy outcomes. In recent years, however, the effectiveness of this strategy has been called into question [1]. Increasing use of early and adequate prenatal care has not led to a significant decline in low birthweight (LBW) or prematurity, which are leading causes of infant mortality in the U.S.. Nor has closing the access gap in prenatal care led to a significant reduction in racial-ethnic or socioeconomic disparities in these adverse birth outcomes. There is now a growing consensus that further reduction in this nation’s LBW and prematurity rates cannot be achieved solely by improving access to prenatal care [2]. A limitation on the effectiveness of prenatal care may be its timing. Many of the pathophysiological pathways leading up to adverse birth outcomes may have their onset early in pregnancy, possibly at or even before implantation. By the time prenatal care is initiated, it may be too little too late to significantly alter the course or outcome of the pregnancy. Indeed, there is now a growing recognition that birth outcomes are the end product of not only the nine months of pregnancy, but the entire life course of the mother leading up to the pregnancy [3]. Improving birth outcomes, therefore, takes promoting maternal health not only during pregnancy, but before and between pregnancies and, indeed, across the woman’s lifespan. The recognition of the importance of maternal health before pregnancy has led to a growing interest in preconception care [4]. A special subtype of preconception care is internatal care. Internatal care refers to a package of healthcare and ancillary services provided to a woman and her family from the birth of one child to the birth of her next child. It is, in essence, preconception and prenatal care for a subsequent pregnancy. We prefer the term “internatal care” over the more popular term “interconception care” (from the conception of one pregnancy to the conception of the next pregnancy) because 1) internatal care represents an extension of prenatal care with which the public is familiar; 2) the point of entry into care is the birth of one child; and 3) the endpoint of care is the birth of the next child. And while in this paper we will focus primarily on the interpregnancy interval (from the termination of one pregnancy to the conception of the next pregnancy), we prefer the term “internatal care” over “interpregnancy care” because internatal care suggests a continuity of care into a subsequent pregnancy. Despite the growing interest, for most people it remains unclear what is meant by internatal care. For example, many Healthy Start programs now offer internatal care, and yet the content of care varies greatly across sites. There have also been a few intervention studies on internatal care, most notably in Denver [5], Atlanta [6], and Philadelphia [7], with vastly different approaches. The lack of some standards of care makes it difficult to move forward research, practice, and policy, or even discussion, on internatal care. The purpose of this paper is to begin to define the content of internatal care. For healthy mothers, internatal care offers an opportunity for wellness promotion between pregnancies. Presently many women lack such opportunity, particularly among low-income and minority women. In most states, Medicaid programs terminate pregnancy-related healthcare coverage for most low-income mothers at 60 days postpartum. Other than the one recommended postpartum visit, many of these women will have little or no access to healthcare between pregnancies. Internatal care can help close the gap in healthcare for these women. Even for women with healthcare coverage, it is unclear how much and what internatal care they are currently getting. In this paper, we begin to define the core contents of universal internatal care that all women of childbearing age should receive between pregnancies. While we believe that internatal care should be universally available to all women between pregnancies, it may be particularly beneficial for high-risk mothers, such as women with chronic diseases or prior adverse pregnancy outcomes. Many such outcomes (e.g. prematurity, fetal death) carry a high recurrence risk in a subsequent pregnancy, and the biobehavioral risk factors are often carried from one pregnancy to the next [8]. For these mothers, internatal care offers an opportunity for risk reduction before their next pregnancy. In this paper, we also begin to describe the content of enhanced internatal care for high-risk mothers. We will use women with chronic health conditions (hypertension, diabetes, or weight problems) and women with a prior preterm delivery as examples of what more can be done for high-risk mothers during internatal care than under current prevailing practice which, for many women, consists of the one recommended six-week postpartum visit, or none at all. Two decades ago, the U.S. Public Health Service assembled an expert panel to define the contents of prenatal care. In its report Caring for Our Future: The Content of Prenatal Care, the Expert Panel identified four basic components of prenatal care consisting of 1) early and continuing risk assessment, 2) health promotion, 3) clinical interventions, and 4) psychosocial interventions [9]. We now use the same framework to outline the contents of internatal care, based on our search of the literature for clinical standards, best practices, and proven or promising strategies. The focus of this paper is on the content of internatal care, particularly during clinical visits. Other important topics, such as financing, organization, delivery, motivation, and community interventions, are beyond the scope of this paper and will be addressed in a series of papers to follow. It is not our intention in this paper to propose a complete and final model of internatal care. Instead, we present here the start of a framework for internatal care, to be filled in through more dialogues and consensus-building which we hope this paper will help move forward. Recommended core contents of universal internatal care for all women following a pregnancy Most pregnant women in the United States are healthy and experience healthy pregnancy and birth outcomes. For these women, internatal care is still important because it offers an opportunity for wellness promotion. In this section, we describe the core contents of internatal care, including risk assessment, health promotion, clinical interventions, and psychosocial interventions (Table 1). These components should be offered to every woman following a pregnancy, regardless of her risk status. We recommend expanding the current six-week postpartum visit to three or more internatal visits, at 2 weeks, 6 weeks, and 6 months postpartum, with annual follow-up beginning at one year postpartum. We will expand on the rationale for the recommended schedule of visits, and address issues related to systems integration later in this section. Table 1.Recommended content and schedule of internatal care, and organizations recommending these practice standardsRisk assessmentOrganizationsaSchedulebCore content of internatal care for all womenFamily violenceJACHO, AMA1, 2, 3 Infections ChlamydiaUSPSTF, ACPM2 PeriodontalADA2 Immunizations Diptheria-tetanus boosterUSPTSF, ACPM, HHS2 Hepatitis BUSPTSF, ACPM, HHS2 Measles and mumpsUSPTSF, ACPM, HHS, ACIP2 RubellaUSPTSF, ACPM, HHS, ACIP2 VaricellaUSPTSF, ACPM, HHS, ACIP2 NutritionUSPSTF1,2, 3 DepressionUSPSTF, ACOG1, 2, 3 Stress1, 2, 3Health promotion BreastfeedingAAP, HP20101, 2, 3 Back-to-SleepAAP, NICHD, HP20102 ExerciseNHLBI, HHS, HP20102 Exposures Smoking, alcohol, drug useHHS, ACOG1, 2, 3 MercuryFDA/EPA2, 3 LeadCDC, AAP2, 3 DioxinsIOM2, 3 Indoor/outdoor pollutants & allergensNHLBI2, 3 Family planningACOG/AAP1, 2, 3 Folate supplementationCDC, AAP3Clinical interventions Height and weight measurementsICSI, ACOG1, 2, 3 Blood pressureICSI, ACOG2 Total skin examinationACS, ACPM2 Clinical breast examinationACS, ACOG3 Papanicolau smear and pelvic examinationACS, ACOG2 Screening mammographyACS, ACOG3Psychosocial interventions Social services1, 2, 3 Clinical support1, 2, 3 Parenting support1, 2, 3Enhanced content of internatal care for high-risk womenChronic hypertension Risk assessment, health promotion, clinical & psychosocial interventionsNHLBIGestational hypertension Risk assessment, health promotion, clinical & psychosocial interventionsNHLBI, ACOGPregestational diabetes Risk assessment, health promotion, clinical & psychosocial interventionsADAGestatational diabetes Risk assessment, health promotion, clinical & psychosocial interventionsADA, ACOGOverweight/obesity Risk assessment, health promotion, clinical & psychosocial interventionsNHLBI, HHS USPSTFPreterm birth Risk assessment Reproductive historyACOG Family historyACOG Medical assessmentACOG Nutritional assessmentACOG Social assessmentACOG Health Promotion Smoking cessationHHS Substance abuse treatmentACOG Optimal prepregnancy BMIIOM Ensure adequate omega-3 fatty acids intakeISSFAL, IOM Family planningACOG/AAP Clinical interventions Progesterone use in subsequent pregnancyACOG Psychosocial Interventions Services for children with special health care needsAAP Home visitationAAPaReferences to these recommendations available through the authors. Abbreviations: JACHO, Joint Commission on Accreditation of Healthcare Organizations; AMA, American Medical Association; USPSTF, United States Preventive Services Task Force; ACPM, American College of Preventive Medicine; ADA, American Dental Association; HHS, Department of Health and Human Services; ACIP, Advisory Committee on Immunization Practices (CDC). ACOG, American College of Obstetricians and Gynecologists; AAP, American Academy of Pediatrics; HP2010, Healthy People 2010; FDA, Food and Drug Administration; EPA, Environmental Protection Agency; CDC, Centers for Disease Control and Prevention; IOM, Institute of Medicine; NHLBI, National Heart, Lungs, and Blood Institute; ICSI, Institute for Clinical Systems Improvement; ACS, American Cancer Society; ADA, American Diabetes Association.bRecommended schedule for each component of core content of internatal care; 1= 2-week visit, 2= 6-week visit; 3= 6-month visit . Schedule for enhanced content varies according to need Risk assessment The primary objective of risk assessment is to identify on-going problems that need to be addressed. Most pregnant women in the U.S. are healthy and thus should have no major problems postpartum. However, some medical, behavioral or psychosocial problems may have been overlooked by their prenatal care providers. Other problems may develop during the internatal period. Providers of internatal care should pay particular attention to five problems that are now commonly missed during prenatal or postpartum care: family violence, infection/immunization, nutrition, depression and stress (FINDS). Family violence The experience of family violence may be quite common among pregnant and postpartum women in the U.S. Most studies of the prevalence of physical abuse during pregnancy report estimates in the range of 4 to 8 percent, though higher estimates (around 20%) have been reported in some populations [10]. Less is known about the prevalence of violence after pregnancy, with estimates at 3 months postpartum ranging from 3.2% in a population-based survey to 21% in a prospective study of adolescent mothers [11, 12]. In the latter study [12], three of four women who reported intimate partner violence (IPV) had not reported IPV prior to delivery. Family violence puts maternal health and child development at great risk. The literature contains solid links between intimate partner violence and child abuse [13]. Because violence is frequently missed by healthcare providers including prenatal care providers, we recommend routine screening for family violence during every internatal visit. Several brief screening instruments, such as the 4-item Hurt Insulted Threatened or Screamed at (HITS), the 3-item Partner Violence Screen (PVS), and a 3-item Abuse Assessment Screen (AAS), can be effectively used to screen for IPV in a clinical setting [13]. Infection/immunization Some infections may go unrecognized or untreated during pregnancy. Chronic, untreated infections, such as periodontal disease, could pose a threat not only to the next pregnancy [14], but also to the mother’s long-term health as well [15]. Most women are not screened for periodontal disease during prenatal care, and many women lack access to oral health services following pregnancy. While presently there is insufficient evidence to justify universal screening for asymptomatic infections among “low-risk” women in the internatal period (other than the U.S. Preventive Services Task Force (USPSTF) recommendation for Chlamydia screening in women under the age of 25) [16], in some populations routine screening for periodontal, sexually-transmitted and some urogenital tract infections may be warranted. Internatal care also offers an important opportunity to update women’s immunization status. The USPSTF and other national advisory committees recommend that 1) all adults receive periodic diptheria-tetanus toxoids booster every 10 years, 2) all young adults receive a series of Hepatitis B vaccines, 3) all adults born after 1956 who lack evidence of immunity receive vaccination against measles and mumps, and 4) all women of childbearing age be screened for rubella and varicella susceptibility; susceptible nonpregnant women of childbaring age should be offered vaccination and avoid conception for at least 28 days after vaccination; susceptible pregnant women should be vaccinated immediately after delivery (Table 1). Immunization status should be assessed and updated at the six-week internatal visit. Nutrition Nutrition can play an important role in promoting maternal health and child development, and yet women’s nutritional needs are often overlooked by prenatal or postpartum care providers. One recent population-based survey in California found that nearly one in three pregnant women who were income-eligible for Women, Infants, and Children (WIC) program reported being food-insecure; of these, one in four reported not being enrolled in WIC [17]. Nutritional screening should address the ABCD’s of nutritional risks: anthropometric (e.g. low or high body mass index), biochemical (e.g. for anemia or folate deficiency in some populations), clinical (e.g. eating disorder) and dietary (e.g. content, pattern, food insecurity). An abbreviated 6-item version of the Household Food Security Scale can be used to screen for food insecurity. Inquiry should be made about folate or multivitamin supplementation (see under Health Promotion). We recommend routine screening of nutritional risks at every internatal visit. Depression Maternal depression affects a large number of women and their children. A recent systematic review reported combined estimates of the point prevalence of major and minor depression ranging from 8.5 to 11 percent during pregnancy, and 6.5 to nearly 13 percent during the first year postpartum [18]. Maternal depression can have negative long-term impact on both maternal health and child development, but is often missed by prenatal and postpartum care providers. We recommend routine screening for maternal depression at every internatal visit. Fairly accurate and feasible screening measures are available, including the Postpartum Depression Screening Scale (PDSS), Edinburgh Postnatal Depression Scale (EPDS), Beck Depression Inventory (BDI), and the Center for Epidemiologic Studies Depression Scale (CES-D) [18]. The evidence available, albeit limited, suggests that providing some form of psychosocial support to pregnant and postpartum women at risk of having a depressive illness may decrease depressive symptoms [18]. Stress Maternal stress is perhaps the hardest to screen; clinicians are often at a loss as to how to screen or what to do with a positive screen. Yet chronic stress can pose significant risks to maternal health and child development. For the mother, chronic stress can cause wear and tear to her body’s adaptive systems, what McEwen refers to as allostatic load, which could lead to more rapid deterioration in her health and function over time [19]. For the infant (or the fetus in a subsequent pregnancy), maternal stress could result in the re-programming of the child’s basic autonomic rheostat and alteration of immune regulation, leading to increased vulnerability to disease and dysfunction later in life [20]. There is an important need for the development of a multi-dimensional screening tool for stress that can be feasibly implemented in a clinical setting. For now, clinicians need to ask about major stressors in women’s lives (e.g. homelessness, unemployment, family violence, social isolation) at every internatal visit, and find resources and supports to help the family deal with these stressors. Thus even for healthy mothers with healthy infants, much can be done to reduce health risks and promote well-being during internatal care. In addition to the five areas we highlighted (FINDS–family violence, infections/immunization, nutrition, depression, and stress), a battery of reproductive, medical, family, genetic and psychosocial risks should be routinely assessed during internatal care. Other areas of risk assessment can be added to the contents of internatal care as the model becomes more fully developed. A word of caution about risk screening Presently a major limitation on the effectiveness of risk screening is the lack of available or accessible services for those with a positive screen. The benefit of depression screening is limited when there are no mental health services available or accessible to refer mothers who screen positive for depression. And a lack of support services (and coordination thereof) for abuse victims is identified by many providers as a major deterrent to screening for family violence. For risk screening to be effective, it needs to be followed up by effective interventions. This means having all the necessary resources and service capacity (e.g. oral health services for those with periodontal disease, nutritional support for those who are food-insecure, housing assistance for homeless mothers and families) to assist those with a positive screen. Health promotion The primary objective of health promotion is to promote the health and wellbeing of the mother, infant, and family. We will highlight six components health promotion during internatal care: breastfeeding, back-to-sleep, exercise, exposures, family planning and folic acid (BBEEFF). Breastfeeding Despite well-known benefits of breastfeeding for both mother and infant [21], at least one-third of mothers in the U.S. never initiate breastfeeding [22]. Among women who initiated, nearly one in five will stop nursing by one month, and nearly two-thirds will stop by six months [22], despite current recommendation to breastfeed for at least one full year [21]. Internatal care offers an important opportunity to promote breastfeeding, particularly the initial two-week visit to address nascent problems, the six-week visit to anticipate problems related to transition back to work, and the six-month visit to encourage continued breastfeeding with introduction of solid foods. Back-to-sleep With the introduction of the Back-to-Sleep campaign by the American Academy of Pediatrics in 1992, the overall incidence of sudden infant death syndrome (SIDS) has decreased by almost 50% [23]. Despite this success, SIDS continues to be a leading cause of infant death in the U.S., as well as racial disparity in infant mortality [23]. SIDS rate among black infants remains more than twice that of white infants, even after the implementation of the Back-to-Sleep campaign [23]. Black infants are more likely than white infants to be placed in the prone position and soft sleep surfaces for sleep [23]. Providers of internatal care need to discuss the infant’s sleep environment with the parents. Because the risk of SIDS peaks at 2 to 4 months of age, the six-week visit offers an important opportunity to talk about back-to-sleep. Exercise Health education about physical activities should be part of an on-going effort for wellness promotion during internatal care. The 1988 National Maternal and Infant Health Survey revealed that 25 percent of white women and 45 percent of black women were heavier by 4.1 kg (nearly 9 pounds) or more at 10 to 18 months post partum than they were before pregnancy [24]. Clinicians can help women develop a weight loss program that consists of healthy dieting, physical activities, and behavioral modifications if needed. The National Heart, Lung, and Blood Institute recommends that all adults should set a long-term goal to accumulate at least 30 min or more of moderate-intensity physical activity on most, and preferably all, days of the week [25]. For women in the postpartum period, prepregnancy exercise routines may be resumed gradually as soon as it is physically and medically safe. Clinicians need to routinely discuss physical activity with the woman, particularly at the six-week and six-month visits. Exposures Cigarette smoking poses a significant threat to the health of the mother, her infant, and her subsequent pregnancy. Approximately one in eight (11.2%) U.S. women who gave birth in 2002 reported smoking cigarettes during pregnancy [26]. Furthermore, up to 70% of women who quit smoking during pregnancy will relapse within 6 months of delivery [27]. Effective models of clinical intervention for smoking cessation (e.g. 5A’s) and relapse prevention have been developed and should be incorporated into internatal care [28]. Prenatal exposure to alcohol and other substances like cocaine and amphetamines have been associated with birth defects as well as adverse pregnancy and neurodevelopmental outcomes. According to the Behavioral Risk Factor Surveillance System, in 1999 the prevalence of any alcohol use and binge drinking among pregnant women in the U.S. was 12.8 and 2.7%, respectively [29]. Postpartum use and recidivism rates are less well described. Several screening questionnaires have been developed to detect problem drinking which may also prove helpful in detecting substance abuse, include T-ACE (tolerance, annoyed, cut down, eye opener) questions, the CAGE (cut-down, annoyed, guilty, eye-opener) questionnaire, and the brief MAST questionnaire [30]. All women should be screened at the time of their first internatal visit, and referral for evaluation and treatment should be offered to those who screen positive. Additionally, mothers whose infants are at risk for neurobehavioral problems from prenatal exposures should be provided resources for early developmental evaluation and interventions for their infants [30]. A large number of environmental exposures during the internatal period can affect maternal health and child development. For example, chronic exposures to indoor and outdoor pollutants and allergens, including household molds and dust mites, may increase the child’s future risk for atopy, allergies, and asthma. Less is known about the long-term effects of such exposures to maternal health and subsequent pregnancy outcomes. Lead is a potent developmental neurotoxicant and immunotoxicant, and household exposures can come from old paint, leaching from plumbing, and ceramic dinnerware and glazes. Certain measures can be taken to reduce household exposures to pollutants, allergens and lead. Exposure to mercury can be reduced by adherence to the EPA/FDA joint advisory related to consumption of fish and shellfish, and exposure to dioxins can be reduced by decreasing dietary fat consumption (Table 1). Clinicians need to assess and discuss with their patients avoidance of exposures that could adversely impact on maternal health and child development at every internatal visit, but particularly at the 6-week and 6-month visits. Family planning Family planning is vital to the health and wellbeing of women and their families [31]. Unintended pregnancy and short interpregnancy intervals are associated with increased risk for adverse birth outcomes [31, 32]; family planning can reduce the risk by promoting effective contraceptive use and optimal birth spacing. An unintended pregnancy also puts future maternal health and social wellbeing at risk; family planning can protect maternal health and choice [31]. Discussion of family planning should begin prenatally and prior to discharge from the hospital; it should be addressed again at the two-week and six-week internatal visits, including risks, benefits and side effects of available contraceptive options and plans for future childbearing. Adherence to method and side effects should be reassessed at the six-month visit. Folate supplementation Folate is important to both maternal and child health. Maternal folate deficiency has been linked to future risk for cardiovascular and other chronic adult diseases, as well as neural tube and other birth defects in the offspring. Periconceptional (preconceptional and in early pregnancy) use of folic acid has been demonstrated to reduce the occurrence and recurrence of neural tube defects [33]. Thus the U.S. Public Health Services recommends that all women of childbearing age consume 400 micrograms of folic acid daily [34]. Even though there are several ways of getting this amount of folic acid everyday, two-thirds of U.S. women of childbearing age still do not consume enough folic acid, and surveys continue to find substantial knowledge gaps regarding the benefits of folic acid. Folic acid should be promoted at every internatal visit, or at the minimum during the six-month visit. We feel that the recommendation for folic acid can be broadened to include daily multivitamin supplementation given the potential benefits of a number of other vitamins (e.g. B12) and elements to maternal health and subsequent pregnancy outcomes, particularly for women with nutritional deficiencies. Table 1 summarizes the core contents of health promotion during internatal care, with an emphasis on BBEEFF (breastfeeding, back-to-sleep, exercise, exposures, family planning and folic acid). Other topics can be added to health promotion as the model of internatal care becomes more fully developed. Clinical interventions Most women in this category are, by definition, healthy and require little in the way of clinical interventions during internatal care, other than the standard preventive health services recommended for all women of childbearing age (Table 1). These include periodic height, weight and blood pressure measurements, total skin examination, clinical breast examination, Pap smear and bimanual pelvic examination every 1 to 3 years. Mammography should also be performed in women ages 40 and above. We will address clinical interventions for women with chronic conditions in the next section. Psychosocial interventions Many healthy women who gave birth to a healthy infant will nonetheless need some psychosocial support. There are three types of social support that need to be made available as part of the core package of internatal care services: 1) social services, 2) clinical support, and 3) parenting support (Table 1). Social services may include assistance with obtaining certain public insurance (e.g. SCHIP) or benefits (e.g. TANF, EITC). Some women may not be aware of their eligibility for child care subsidies (e.g. CCDF) or availability of early childhood education programs. Women experiencing intimate partner violence may need multiple medical, legal, and social services and could use some help with service coordination. The internatal visits provide a platform for accessing these social services. Women who experience postpartum depression or other affective disorders may benefit from some forms of psychological support and therapy during the internatal period. Couples who experience problems with marital or sexual relationship in the internatal period may benefit from counseling or interventions. Women with alcohol or drug problems could also use clinical support and treatment. These and other professional clinical support services should be made available and accessible as part of standard internatal care services. Many parents could benefit from some parenting support, such as mother support groups or parenting classes. A current innovative idea in prenatal care is “centering” in pregnancy, which provides social support to pregnant women through group prenatal care [35]. While this idea remains to be tested, we believe that group internatal care–a “centering” in internatal care model–may provide additional support for some women, particularly around issues related to parenting practices and the stress of parenting. A note on psychosocial interventions Throughout this paper, the readers will notice an imbalance in the levels of details we used to describe clinical and psychosocial interventions. That is, clinical recommendations are, for the most part, described with much greater specificity than psychosocial interventions. It is certainly not our intention to propose a predominantly biomedical model of internatal care. Rather, the lack of specificity in our description of psychosocial interventions reflect both the complexity of psychosocial issues that families–particularly low-income families–face for which there are often no easy solutions, as well as the dearth of well-designed and evaluated intervention programs addressing these psychosocial issues. For example, mental health services alone are unlikely to be effective in treating the postpartum depression of a homeless mother, and housing assistance alone is unlikely to find her stable housing without sustainable income and employment, but few well-designed intervention studies have critically evaluated the impact of a comprehensive package of mental health services, housing assistance, work- and life-skills training and childcare on maternal and family health. Our model of internatal care is unlikely to be effective without effective psychosocial interventions, but we will not know what these are as long as we keep looking for quick biomedical fixes. Future intervention studies on internatal care need to take a more integrative, comprehensive approach to improving family and women’s health. Service coordination and integration As aforementioned, presently some referral services (e.g. mental health services, oral health services) are unavailable or inaccessible to many women between pregnancies. But even if the services are available, they are often poorly coordinated. Fragmentation in service delivery can deter access to care, particularly for low-income women with many other competing needs. Providers of internatal care need to consider how to better coordinate and integrate services in order to improve access. This can be accomplished by establishing a well-developed referral network. In some populations, care coordination or case management provided by a nurse or social worker may be needed. Ideally these services should be provided at one location to increase service coordination and integration. For example, the two-week internatal visit can be provided at the same time and location as the two-week well-baby check-up, with on-site WIC, health education, and social services that allow for “one-stop shopping.” We will take up issues related to the organization and delivery of internatal care that will maximize access and utilization of services in a series of papers to follow. Schedule of visits In developing a schedule of visits for internatal care, we had to first wrestle with two questions. First, shouldn’t the schedule of visits be individualized according to individual needs? While we support this view, we nonetheless feel that it may be useful to recommend some minimum standards to start a dialogue about this topic. For healthy women with healthy infants, we propose expanding the current six-week postpartum visit to three or more internatal visits, at 2 weeks, 6 weeks, and 6 months postpartum, with annual exams beginning at one year postpartum. For some activities, the six-week postpartum visit may be too little too late. For example, one in five mothers who initiated breastfeeding will quit nursing within the first month [22]; a 2-week postpartum visit may offer more timely encouragement and interventions for maintenance of breastfeeding than a 6-week visit could. The 2-week visit also offers an opportunity to follow-up on wound care for cesarean delivery. Whenever possible, this visit should be coordinated with the two-week well-baby check-up, preferably at the same location, to increase motivation for attendance. We feel that it is important to keep the current six-week postpartum visit as the second internatal visit. This is a well-established visit in obstetrics, with well-defined contents. This visit, however, can be made more useful by expanding its contents. We feel that a third internatal visit at 6 months is important for follow-up on a number of issues, such as breastfeeding and family planning. The standard annual visits should begin at one year postpartum (Table 1). Second, shouldn’t any routine visit by a woman who may, at some time, become pregnant again be viewed as an opportunity to emphasize the importance of internatal health and habits [36]? While we support this view, we nonetheless feel that it would be preferable to recommend a schedule of internatal visits rather than relying on “any routine visit.” Several components of internatal care are time sensitive (e.g. counseling about breastfeeding and family planning at 2 weeks or about SIDS at 6 weeks) which should not wait for any routine visit. Scheduled internatal visits may be particularly important for women who are not making routine visits, or who are seeking care through the emergency department or urgent care clinics which lack the resources and capacity to provide comprehensive internatal care. We would also like to see internatal care become reimbursable visits, with standardized contents and schedule (rather than as an add-on to a visit for mastitis or cystitis), which may increase provider incentives to provide internatal care. Issues related to patient motivation and provider reimbursement will be addressed in a series of papers to follow. Recommended contents of enhanced internatal care for high-risk women In this section, we begin to define the contents of enhanced internatal care for high-risk women (Table 1). We focus on four groups of high-risk women: 1) women with chronic hypertension or history of hypertensive disorders during pregnancy, 2) women with pregestaional or gestational diabetes mellitus, 3) women who are underweight, overweight or obese, and 4) women with prior preterm birth. We call attention to the first 3 groups because they are often underserved in the internatal period, especially if they gave birth to a healthy infant. Their continued health needs may not be met because they had a good birth outcome. Some women will lose healthcare coverage (e.g. pregnancy-related Medicaid), and will have little on-going care despite their chronic conditions. We call attention to the 4th group because of their high recurrence risk [8] and the significance of recurrence. The goals of internatal care for these women are to provide ongoing care for their health problems and optimize their health prior to next pregnancy. In addition to the core contents of universal internatal care recommended above for all women of childbearing ages, we will examine what enhanced contents may benefit their long-term health, as well as the outcome of their subsequent pregnancy. Women with chronic hypertension or hypertensive disorders during pregnancy Hypertensive disorders are the most frequently reported medical risk factor during pregnancy [11]. Among the 4 million U.S. women who gave a live birth in 2002, more than 150,000 had “pregnancy-associated” (gestational) hypertension, more than 3,000 had chronic hypertension, and nearly 13,000 had eclampsia [11]. Among women with chronic hypertension (i.e. hypertension diagnosed prior to twentieth week of gestation and probably antedates the pregnancy) that persists after pregnancy, guidelines for evaluation and treatment are well established and can be incorporated into the contents of internatal care [37]. Risk assessment has three major objectives: 1) to assess lifestyle and identify other cardiovascular risk factors or concomitant disorders that may affect prognosis and guide treatment 2) to reveal identifiable causes of hypertension, and 3) to assess the presence or absence of target organ damage and cardiovascular disease. Initial evaluation should include medical history, physical examination (including fundoscopic and cardiovascular examinations), laboratory testing and other diagnostic procedures (including an electrocardiogram and testing for renal function and lipoprotein profile). Health promotion should promote weight reduction, DASH (Dietary Approaches to Stop Hypertension) eating plan, dietary sodium reduction, physical activities, moderation of alcohol, and smoking cessation. Clinical interventions should include drug therapy for all women with hypertension, as well as prehypertensive women with compelling indications (e.g. chronic kidney disease or diabetes). Follow-up should be monthly until the blood pressure goal is achieved, and then at 3–6 months interval thereafter during the internatal period. Psychosocial interventions should address chronic stressors that may exacerbate the chronic hypertension, as well as barriers (e.g. financial, motivational) to adherence with treatment protocol. The contents of internatal care for women with gestational hypertension, pre-eclampsia, and eclampsia are less well defined, especially if the hypertension does not persist following pregnancy. In most women with gestational hypertension, the blood pressure will become normal during the first week postpartum; for women with preeclampsia the hypertension will take longer to resolve. If the hypertension has not remitted by the 6-week visit, the women should be reexamined 6 weeks later, when persisting pathology will probably be chronic [38]. During internatal care, providers need to discuss with their patients recurrence risk in the subsequent pregnancy. They also need to be aware that preeclampsia has been associated with increased risk of maternal cardiovascular diseases later in life [39], and that attention to cardiovascular risk factors is required in all subsequent internatal care and wellness visits. Presently there is no proven internatal clinical intervention for the prevention of recurrent gestational hypertension or preeclampsia. During a subsequent pregnancy, antihypertensive therapy should be used only for pregnant women with severe hypertension, as pharmocologic treatment of women with uncomplicated mild chronic hypertension has not been found to be beneficial [38]. Several studies failed to demonstrate a benefit for low-dose aspirin in preventing preeclampsia among women with chronic hypertension [38]. When chronic hypertension is complicated by intrauterine growth restriction or preeclampsia, fetal surveillance is warranted. For women with preeclampsia during a previous pregnancy, low-dose aspirin started in the second trimester of pregnancy did not reduce the recurrence risk of preeclampsia [38]. Further studies are needed to identify effective interventions during the internatal period and subsequent pregnancy to prevent recurrent gestational hypertension or preeclampsia. Women with pre-gestational or gestational diabetes mellitus Diabetes complicating pregnancy is the second most frequently reported medical risk factor during pregnancy [11]. In 2002, more than 130,000 women reported having diabetes during pregnancy [11]. For women with pre-gestational diabetes, they are at increased risk for microvascular, neuropathic and cardiovascular diseases, especially if the diabetes is long-standing and poorly controlled. Their risks are reduced with tight glycemic control. Additionally, their fetuses are at increased risk for stillbirth and congenital anomalies. Preconception glycemic control among women with pre-gestational diabetes has been shown to reduce the risk of congenital anomalies [40]. Women who had gestational diabetes have 30–70% chance of recurrence during their next pregnancy [41] and an increased risk of developing Type II diabetes later in life [42]. Gestational diabetes increases the risk of fetal macrosomia, birth trauma, newborn hypoglycemia and hyperbilirubinemia. Evidence from animal models suggests that gestational diabetes, especially if poorly controlled, may also program insulin and leptin resistance in the fetus, leading the greater susceptibility for diabetes and obesity in the offspring [43]. For women with pre-gestational diabetes, internatal care should follow the guidelines issued by the American Diabetes Association on preconception care of women with chronic diabetes [44]. At the initial office visit, risk assessment should include a complete history, physical examination (including cardiovascular, neurological, and dilated retinal exams), and laboratory evaluation (including glycosylated hemoglobin (HbA1c), serum creatinine and 24-h urinary excretion of total protein and/or albumin, and measurement of thyroid stimulating hormone and/or free thyroxine level in women with type 1 diabetes because of the 5 to 10% coincidence of hyper- or hypothyroidism) and electrocardiogram if the diabetes has been long-standing. Health promotion should emphasize practical self-management skills essential for glycemic control and preparation for pregnancy including the use of an appropriate meal plan, self-monitoring of blood glucose (SMBG), self-administration of insulin and self-adjustment of insulin doses, treatment of hypoglycemia (patient and family members), incorporation of physical activity, and development of stress reduction techniques. Patients should be counseled about fetal and neonatal complications (and possibly long-term health consequences) of maternal diabetes, including the risk of congenital malformations, and the need for effective contraception and optimal glycemic control prior to the next pregnancy. Clinical interventions include dietary control and pharmacological treatment using oral hypoglycemics or insulin. Glycemic control can be self-monitored at home, and glycosylated hemoglobin levels should be evaluated at 1- to 2-month intervals until stable. Follow-up visits should be at 1- to 2-month intervals, and frequent phone contact for adjustment of insulin doses and other aspects of the treatment regimen is advised. Psychosocial interventions should address barriers to adherence with treatment protocol. Referral to and coordination with a dietician and programs such as W.I.C. or Sweet Success in California may enhance treatment efficacy. Stress management should be emphasized because stress can affect dietary control as well as glucose homeostasis. For women with gestational diabetes in a previous pregnancy, guidelines for internatal care are less well established. The American Diabetes Association recommends that a 75-g oral glucose tolerance test (GTT) be performed 6 to 8 weeks after delivery [41]. The extent to which this recommendation is adhered is presently unknown. We believe that much more can be done in the internatal period to promote maternal health (as well as fetal health in a subsequent pregnancy) among women with gestational diabetes. Because these women are at increased risk for developing type II diabetes, risk assessment should include a fasting glucose once a year. If clinical assessment suggests polycystic ovarian syndrome and/or metabolic syndrome, periodic testing for lipoprotein profiles, fasting glucose, and possibly glucose tolerance test may be warranted [45]. Patients should be evaluated for glucose intolerance before their next pregnancy. For health promotion, all women who have had gestational diabetes should be encouraged to exercise and lose weight if they are obese. Dietary counseling is recommended given their increased risk for future type II diabetes. Presently there are no specific clinical interventions recommended in the internatal period for women with a history of gestational diabetes. During a subsequent pregnancy, women with pregestational diabetes using oral hypoglycemic agents should be switched to insulin because the safety and efficacy of currently available oral hypoglycemics have not been evaluated for use in pregnancy among women with pregestational diabetes. For women with gestational diabetes in a prior pregnancy, the use of a second-generation oral hypoglycemic agent (i.e. glibenclamide or glyburide) may be considered if gestational diabetes recurs in a subsequent pregnancy. The agent does not cross the placenta and has been shown to be comparable to insulin in improving glucose control for women with gestational diabetes [46]. Weight gain during pregnancy should be closely monitored. Women who are underweight, overweight, or obese A growing number of women of reproductive age in the U.S. are overweight or obese. According to the 1997 National Health Interview Survey, one-third of women ages 18 to 24 and 43% of women ages 25 to 44 are overweight [47]. The survey also found one in eight women ages 18 to 24 and one in six women ages 25 to 44 are obese [47]. Based on the CDC 2003 Pregnancy Nutrition Surveillance System (PNSS) of nearly 728,000 low-income pregnant women, 43% had a prepregnancy body mass index (BMI) that is considered overweight or obese, up from less than 30% in 1989 [48]. Maternal obesity poses a threat to not only subsequent pregnancy outcomes, but also long-term maternal health. Pregnancy itself is a risk factor for maternal obesity. According to the 2003 PNSS, more than 44% had gained more than the ideal weight gain [48] and, as previously cited [24], weight retention following pregnancy is common. Weight retention needs to be addressed during internatal care, particularly for women who are overweight or obese. According to the 2003 PNSS, one in eight pregnant women were underweight before pregnancy, and one in four had gained less than the ideal weight gain during pregnancy [48]. Both low pre-pregnancy BMI and poor pregnancy weight gain have been shown to be risk factors for preterm delivery and low birth weight [49]. For women who are underweight, overweight, or obese, an important goal of internatal care is to optimize their weight prior to the next pregnancy. For risk assessment, the BMI can be used because it provides an acceptable approximation of body fat and fair correlation with disease outcomes. Potential causes of underweight (e.g. eating disorders, food insecurity), overweight or obesity (e.g. polycystic ovarian syndrome) need to be evaluated. For health promotion, women who are underweight should be encouraged to gain weight through healthy nutrition; for women who are overweight or obese, clinical guidelines have been established for the identification, evaluation, and treatment of overweight and obesity [50]. These guidelines should be followed for internatal care. The initial goal of weight loss therapy should be to reduce body weight by approximately 10 percent from baseline, and weight loss should be about 1 to 2 pounds per week for a period of about 6 months. This amount of postpartum weight loss in overweight women does not appear to affect the growth of their infants, even if they are exclusively breastfed [51]. Weight loss and weight maintenance therapy should employ the combination of low-calorie diets, increased physical activity, and behavior therapy. Clinical interventions may include weight loss drugs as an adjunct to diet and physical activity for patients with a BMI of ≥30, as well as for patients with a BMI of ≥27 with concomitant risk factors or diseases. After successful weight loss, the likelihood of weight loss maintenance is enhanced by a program consisting of dietary therapy, physical activity, and behavior therapy which should be continued until the next pregnancy. Psychosocial interventions should address barriers to healthy nutrition and coordination with a dietician, WIC and other community programs. The schedule of internatal visits should be individualized; the literature suggests that frequent contacts between professional counselors and patients help promote healthy weight maintenance. In a subsequent pregnancy, the Institute of Medicine guidelines for weight gain during pregnancy should be promoted based on prepregnancy BMI [52]. Women who had a prior preterm birth Presently this group of women is most often targeted for internatal care. Preterm birth is a leading cause of infant mortality and long-term disabilities in children in the U.S.. Prior preterm birth is the strongest and most consistent predictor of a subsequent preterm birth. This may be due to the fact that many of the biobehavioral risk factors for preterm birth are carried from one pregnancy to the next. The goal of internatal care for these women is to prevent recurrence of preterm birth by addressing these continuing biobehavioral risks prior to their next pregnancy. Risk assessment Women who had a preterm birth should undergo a comprehensive risk assessment, starting with a complete review of their obstetrical history at the two-week visit. Distinction should be made between spontaneous and induced preterm deliveries. Details about the preterm delivery should be obtained. Work-up should be guided by known etiologic pathways (e.g. infectious-inflammatory, vascular, neuroendocrine and pathologic uterine over-distension) [53]. To identify infectious-inflammatory causes, medical records should be reviewed for untreated infections (e.g. asymptomatic bacteriuria), physical examination should be performed to uncover on-going, chronic infections (e.g. periodontal or reproductive tract infections), and placental pathology should be reviewed for evidence of chorioamnionitis [54]. To identify vascular causes, placental pathology should be reviewed for evidence of uteroplacental ischemia and thromboses [54]. The presence of these findings, in the absence of known causes, may indicate work-up for hereditary or acquired thrombophilia [55]. To identify neuroendocrine causes, inquiries should be made about psychosocial stress and support during pregnancy. Special attention should be paid to FINDS (family violence, infections, nutrition, depression, and stress), since these risk factors may be involved in one or more etiologic pathways leading to the occurrence or recurrence of preterm birth. Review of past medical history should focus on conditions that could increase the risk of preterm deliveries (e.g. systemic lupus erythematosus or renal disease). Review of family history should identify other preterm deliveries in the family (particularly among first-degree relatives), as well as other conditions associated with preterm deliveries. A thorough review of reproductive history, including other preterm births or pregnancy losses, sexual history including sexually transmitted infections, past and current contraceptive use, and plans for future childbearing and child spacing should be conducted. Health promotion The goal of health promotion is to promote protective factors and reduce risk factors for recurrent preterm births. We will highlight three areas of health promotion: smoking cessation, healthy nutrition, and family planning. First, exposure to cigarette smoking significantly increases the risk of preterm delivery. Smoking cessation (by women and their partners) should be encouraged during internatal care. As previously discussed, effective models of clinical interventions (e.g. 5A’s) have been developed [28]. Second, healthy nutrition should be promoted. Low pre-pregnancy BMI has been associated with increased risk for spontaneous preterm birth, while high pre-pregnancy BMI has been associated with increased risk for indicated preterm birth [49]. For women with a history of preterm birth, achieving a healthy weight prior to their next pregnancy may reduce the risk of recurrent preterm birth. A number of programs have demonstrated effectiveness in helping women achieve and maintain healthy weight [56]. Additionally, nutrition may also play an important role in modulating several of the etiologic pathways leading to preterm birth. For example, increased dietary intake of omega-3 polyunsaturated fatty acids may help modulate inflammation [57]. Women should be encouraged to increase consumption of foods rich in omega-3 (the International Society for the Study of Fatty Acids and Lipids recommends that during pregnancy and lactation women must ensure a minimum daily intake of 300 mg of Docosahexaenoic Acid or DHA, an important source of omega-3 fatty acids [58]), and decrease consumption of foods high in saturated fatty acids, trans-fats, and partially-hydrogenated oils. Increasing dietary intake of fruits and vegetables rich in anti-oxidants and certain phytochemicals should also be encouraged. These recommendations could arguably be adopted on the basis of promoting women’s health alone, even in the absence of data supporting their effectiveness in preventing recurrent preterm birth. Third, the importance of family planning cannot be over-emphasized, particularly for women with a history of preterm birth. Both unintended pregnancies and short interpregnancy intervals are associated with increased risk for preterm delivery. Family planning has been shown to reduce unintended pregnancies and promote optimal child spacing. Other BBEEFF topics, such as breastfeeding and back-to-sleep, are also important to address with mothers of preterm babies. Clinical interventions To our knowledge, there has been only one randomized controlled trial for the prevention of recurrent preterm birth in which a specific clinical intervention is initiated during the interpregnancy period [59]. In that study, women with a prior spontaneous preterm birth at less than 34 weeks of gestation were randomized at 3 months postpartum to a course of oral azithromycin and sustained-release metronidazole or placebo. The regimen was repeated every 4 months until conception of another pregnancy. The investigators found no significant difference in the rates of preterm birth between the study and control groups. In fact, women in the treatment group, on average, had lower birth weight babies and delivered 2.4 weeks earlier [59]. In the absence of evidence of effectiveness for preventing recurrent preterm birth, presently no specific clinical intervention can be recommended for women with prior preterm birth in the interpregnancy period. Several studies have examined the effectiveness of clinical interventions applied to women with prior preterm birth during a subsequent pregnancy for the prevention of recurrent preterm birth. Perhaps the most promising is the use of progesterone. A large randomized controlled trial in the U.S. [60] and two recent meta-analyses [61, 62] found evidence of effectiveness of progestational agents (e.g. 17-hydroxyprogesterone caproate) for the prevention of preterm or recurrent preterm birth. However, “because unresolved issues remain, such as optimal route of drug delivery and long-term safety of the drug,” the American College of Obstetricians and Gynecologists (ACOG) has issued a committee opinion that “when progesterone is used, it is important to restrict its use only to women with a documented history of a previous spontaneous birth at less than 37 weeks of gestation” [63]. Another promising clinical intervention is the use of fish oil supplementation for the prevention of recurrent preterm birth. In the Fish Oil Trials in Pregnancy (FOTIP) study, fish oil supplementation in a subsequent pregnancy significantly reduced recurrence risk of preterm delivery [64]. There is also emerging evidence that scaling and root planing in pregnant women with periodontitis may reduce preterm birth [65], though no published studies have evaluated the effectiveness of periodontal screening in women with prior preterm birth. Further studies are needed to demonstrate the effectiveness of these interventions. Perhaps the most controversial clinical intervention is antibiotic treatment of asymptomatic bacterial vaginosis among women with a history of preterm birth [66, 67]. While some studies have shown a benefit, others have not. Efficacy of treatment may depend on the agent used, the route of administration, the timing of treatment, test of cure, and a number of host factors. In at least four antibiotic trials in pregnancy involving women with prior preterm birth [59, 65, 68, 69], antibiotic treatment (of bacterial vaginosis, Trichomonas vaginalis, or periodontal infection) was associated with higher, not lower, incidence of recurrent preterm birth. Clearly there is a need for further intervention studies, guided by better understanding of pharmacokinetics and pharmacogenetics, disease mechanisms and host response, as this remains a potentially important area of intervention for the prevention of recurrent preterm birth. Psychosocial interventions A growing body of evidence suggests that maternal psychosocial stress is an important risk factor for preterm delivery, and that risk can be mitigated by reducing stress and/or increasing support. However, most extant studies of social support during pregnancy suffer from methodological flaws [70], and there is a paucity of research on the effectiveness of preconception or internatal support. The provision of social, clinical and parenting support during the internatal period, as previously described, may be quite valuable to mothers of a preterm infant, who often has special healthcare and other developmental needs that require additional support services. Such services may also be particularly beneficial to low-SES mothers, who may face greater burdens with less personal and social resources to deal with them. More research is needed to identify effective models of internatal psychosocial interventions for women with a history of preterm delivery. Service coordination and integration Women who have had a preterm birth may require multiple services during the internatal period, particularly if their children have significant disabilities and special healthcare needs. Providing integrated services, including linkages to early intervention or other early child development programs for preterm or LBW infants, should be a primary objective of care for these women. Several models of care coordination and service integration have been developed [71, 72]. One of the most successful models is the Olds nurse home visitation program [73]. The program provided nurse home visitation during pregnancy and the first two years postpartum. Evaluation of the Memphis/Shelby County program [73] found women who received home visits by nurses had fewer closely spaced subsequent pregnancies, longer internatal intervals, and fewer months of using public assistance programs, compared to a control group of women. There were, however, no statistically significant program effects on LBW or stillbirth in the subsequent pregnancy. The program is technically not an “internatal care” program and not targeted specifically to women who had a preterm birth, but this model could potentially be adapted by Healthy Start or other public health or community-based programs to provide service coordination and integration for internatal care using community nurses. We recommend that all women with prior preterm birth be offered care coordination and/or home visitation during internatal care. Schedule of visits The schedule of internatal visits for women with prior preterm birth needs to be individualized (with a minimum of three visits plus annual visits, as previously described), depending on the clinical and psychosocial needs of the woman and family. During their subsequent pregnancy, these women should be supervised by a maternal fetal medicine specialist, either directly or indirectly through consultations. In sum, given the large impact of preterm birth on infant mortality and childhood disabilities and the high rates of recurrence in a subsequent pregnancy, women with prior preterm birth could be one of the most critical target groups for enhanced internatal care. Other high-risk women (e.g. women with prior fetal death), could also benefit from such program; we are unable to describe the content of their care in this paper due to space limitation and will do so in a series of papers to follow. The goal of such program is to prevent recurrence of preterm birth by addressing known biobehavioral pathways (e.g. infections/inflammation, neuroendocrine, vascular, etc) prior to the next pregnancy. There is a great need for more comprehensive and systematic risk assessment and health promotion during internatal care, as well as more basic, clinical and intervention research to identify effective clinical and psychosocial interventions for the prevention of recurrent preterm birth. Conclusion In this paper, we began to define the contents of internatal care. We recommended expanding the current six-week postpartum visit to three or more internatal visits (at two weeks, six weeks, and six months postpartum, and annual visits beginning at one-year postpartum). We suggested some core contents that all women should receive during internatal care, including risk assessment (e.g. FINDS), health promotion (e.g. BBEEFF), clinical and psychosocial interventions. For women with chronic health conditions such as hypertension, diabetes, or weight problems, we identified clinical guidelines for their evaluation, treatment, and follow-up during the internatal period. For women who had a preterm birth, we proposed an internatal care model based on known etiologic pathways, with the goal of preventing recurrence by addressing these biobehavioral pathways prior to the next pregnancy. We suggested enhancing service integration for women and families, including possibly care coordination or home visitation for women with prior preterm birth. We were limited in our effort to define the contents of internatal care by the paucity of research on the internatal period. Many pre-disease pathways leading to recurrence of adverse birth outcomes have not yet been elucidated. More importantly, even less is known about the trajectories to long-term women’s health or child health, or how to alter these pathways and trajectories. There is a need for more intervention studies evaluating the effectiveness of the various components of internatal care. More importantly, we need more research on mechanisms of disease to guide the design of internatal interventions. Our proposal will need to be updated and revised continually as new research data emerge. We were also limited by the paucity of research on community-level or public health interventions during the internatal period. While our paper focused on the contents of the clinical visits, we believe that provision of internatal care should not be limited to the clinician’s office. For example, clinical interventions to control hypertension may be more effective if they are supported by community-level or public health interventions, such as implementation of the American Public Health Association resolution that the food manufacturers and restaurants reduce sodium in the food supply by 50 percent over the next decade [37]. The 5A’s of smoking cessation may be more effective if they are supported by multilevel interventions at the interpersonal (e.g. targetting partner smoking), community and institutional (e.g. billboards and marketing of tobacco products, access to smoking cessation programs), and policy levels (e.g. ban on smoking in public space, promotion of smoke-free environments, and increased taxation on tobacco products) [3]. We believe that the ultimate success of internatal care may hinge on the development of a much broader and more integrated conception of internatal care than currently prevails with prenatal care [2]. Presently there are numerous public health programs offering some components of internatal care, including many Healthy Start programs. They could provide the ideal social laboratory settings for implementing and evaluating various community-level interventions. While there are significant challenges to this type of research, it needs to be done. Identifying what internatal interventions work in the community may be just as important as identifying what works in the clinical setting. Such research needs to be guided by the principles and methods of community-based participatory research (CBPR) [74] and be held to scientifically rigorous standards. Although more research is needed to identify effective internatal interventions, we believe we cannot wait for the perfect studies to be done. We must begin to act. Most of our recommendations can be justified on the basis of providing good preventive and primary care for women and children, even in the absence of conclusive evidence supporting their effectiveness in improving subsequent pregnancy outcomes. For the most part they reflect current practice standards recommended by major national organizations (Table 1); most of these are considered standards of care or best practices in women’s and children’s health care. The only recommendations without the endorsement of a national organization–screening for psychosocial stress and provision of psychosocial support–are based on our collective opinions; we strongly believe these are important to reduce allostatic load on women’s health and improve subsequent pregnancy outcome. In this paper we did not attempt to address all the institutional components and challenges in developing a successful model of internatal care, only the content of such a program. This paper reflects our effort to start the process of defining the contents of internatal care. We hope that this paper can be used to stimulate further dialogues about internatal care. Such dialogues, we believe, are needed to move forward research, practice, and policy on internatal care, and to move this nation one step closer toward the provision of a more universal, longitudinally-integrated continuum of women’s healthcare.	[ "preconception care", "content", "internatal care", "interpregnancy", "preterm birth" ]	[ "P", "P", "P", "P", "P" ]
Mycorrhiza-4-1-2245993	Tricholoma matsutake 1-Ocen-3-ol and methyl cinnamate repel mycophagous Proisotoma minuta (Collembola: Insecta)	Two major volatiles produced by the mycelia and fruiting bodies of Tricholoma matsutake (1-octen-3-ol and methyl cinnamate) repel a mycophagous collembolan, Proisotoma minuta. Aggregation of the collembolans on their diet was significantly inhibited by exposure to 1 ppm methyl cinnamate or 10 to 100 ppm 1-octen-3-ol. The aggregation activity decreased dose-dependently upon exposure to 1-octen-3-ol at concentrations higher than 0.01 ppm. Aggregation in the presence of methyl cinnamate exhibited three phases: no significant effect at concentrations ranging from 0.001 to 0.1 ppm, significant inhibition from 1 to 100 ppm, and strong inhibition at 1,000 ppm. These results may explain why certain collembolan species do not prefer T. matsutake fruiting bodies. Introduction Fungal fruiting bodies are more nutritious than their mycelia (Stark 1972; Cromack et al. 1975; Vogt and Edmonds 1980) and are exploited by various arthropods, including dipteran larvae, collembolans, and acarine and oribatid mites (Hammond and Lawrence 1989). Collembolans are the most abundant group of insects found on agaric fruiting bodies (Yamashita and Hijii 2003), and numbers often reach hundreds to thousands on a suitable fruiting body (Sawahata et al. 2000, 2001, 2002; Nakamori and Suzuki 2005a). Collembolans attack the hymenial area, often consuming 50% or more of it in suitable fruiting bodies (Sawahata 2006). Additionally, collembolans are able to break the hyaline and thin-walled basidiospores of some fungi (Ponge and Charpentie 1981; Sawahata et al. 2001; Nakamori and Suzuki 2005b). Therefore, collembolans are likely to adversely affect spore production by fungal species with these types of spores. The fruiting body of the ectomycorrhizal fungus Tricholoma matsutake (S. Ito and Imai) Sing., which has the common name matsutake, is one of the favorite and most valuable food items produced by forests in Japan because of its characteristic odor and texture, but artificial cultivation of T. matsutake has not yet succeeded. Sawahata and Narimatsu (2006) reported that collembolan populations on the fruiting bodies of T. matsutake are smaller than those on other mushrooms growing in the same forest. The gut contents of collembolans collected from the fruiting bodies suggested that they fed on the fungal tissue on the gill surface (Sawahata and Narimatsu 2006). In our preliminary studies, Folsomia candida Willem and Proisotoma minuta Tullberg fed on sliced samples of the fruiting body of T. matsutake in a laboratory experiment (data not shown). These findings suggested that the mushroom is not poisonous to collembolans and that it is simply avoided as a source of food when alternatives are available. Collembolans use fungal volatiles to recognize potential sources of food (Bengtsson et al. 1988; Hedlund et al. 1995). Bengtsson et al. (1991) suggested that the volatile compounds 1-octen-3-ol and 1-hepten can aid the collembolans in their search for a palatable fungus. The major volatiles produced by T. matsutake are 1-octen-3-ol and methyl cinnamate (Ohta 1983; Terashita et al. 1991). The former is a primary volatile produced by many species of fungi (Kaminski et al. 1974; Pyysalo 1976), and the latter determines the characteristic odor of this fungus (Terashita et al. 1991). Hence, methyl cinnamate may prevent aggregation and feeding on the fruiting body by some collembolans. Further, high concentrations of 1-octen-3-ol can repel fungivores (Pfeil and Mumma 1993; Wood et al. 2001). The fruiting body of T. matsutake contains 1-octen-3-ol at concentrations ranging from 2 to 186 ppm on a fresh-weight basis (Ohta 1983; Terashita et al. 1991), but it is unknown whether 1-octen-3-ol repels collembolans at high concentrations. The present study aimed at investigating the effects of the major volatile components of T. matsutake (1-octen-3-ol and methyl cinnamate) at different concentrations on the aggregation of mycophagous collembolans by biological assay (feeding test). Our goal was to provide an explanation for why collembolans aggregate in relatively small numbers on T. matsutake fruiting bodies. Materials and methods Proisotoma minuta was extracted from forest soils near the Matsunoyama Museum of Natural Science in Niigata Prefecture (307 m above sea level, 37°05′N, 138°36′E), in central Japan. This species was used because it fed on various ectomycorrhizal fungi in a choice experiment (Schultz 1991; Hiol Hiol et al. 1994), and it is amenable to laboratory culture. The collembolans were reared on dry yeast at 20°C in a cylindrical chamber (11 cm in diameter, 7 cm in height) whose bottom (1.0 to 1.5 cm in thickness) was covered by a mixture of plaster of Paris and charcoal (10:1 v/v). We followed the methods used in previous food choice experiments (e. g., Schultz 1991; Kaneda and Kaneko 2004). Each assay (with six replications) used 70 individuals (1 month after eclosion from eggs, 0.8 to 1.1 mm in body size) at a time in the same type of chamber. The major volatile compounds of T. matsutake have already been determined to be 1-octen-3-ol and methyl cinnamate by several researchers (e.g., Ohta 1983; Terashita et al. 1991). We substituted laboratory-grade versions of these chemicals for the natural volatile compounds. 1-Octen-3-ol (Wako Pure Chemical Industries, Osaka, Japan) was diluted with distilled water to prepare solutions of 0.001, 0.01, 0.1, 1, 10, 100, and 1,000 ppm. Methyl cinnamate (Wako) was diluted similarly but with dichloromethane. Controls were created using distilled water and dichloromethane, respectively. Into each diluted solution, three pieces of filter paper (3 cm in diameter) were dipped for long enough that the paper became completely saturated. The wet 1-octen-3-ol filter papers were placed 1 cm apart on the bottom of the assay chamber to form a triangle (Fig. 1). The wet methyl cinnamate filter papers were kept in air for 15 min to evaporate the solvent and then moistened with distilled water before bioassay by the same procedure. Fig. 1Bioassay method used to assess the ability of 1-octen-3-ol and methyl cinnamate to repel the collembolan Proisotoma minuta (n = 70). Plugs of potato-dextrose agar (PDA) were placed at the center of each sample-treated filter paper on a mixture of plaster of Paris and charcoal (10:1 v/v) at the bottom of the cylindrical chamber On each sample of filter paper, including the control, a plug (1 cm in diameter) of 3% potato dextrose agar medium (a PDA plug) was provided as food (Fig. 1). Seventy insects were introduced into the chamber and held for 90 min in the dark at 20°C; then, the individuals on each PDA plug were counted without disturbance. The results represent means of 18 samples (six replicates of three PDA plugs per chamber) among the dilution series (eight levels, including the control) for the two test compounds. The mean numbers of collembolans in each treatment were compared by F test, one-way analysis of variance, and a multiple-range test (Fisher’s PLSD). Statistical analyses were performed with Stat View 5.0J software (SAS Institute, Cary, NC, USA). Results Both compounds at 1,000 ppm almost completely inhibited the aggregation of P. minuta on the PDA plugs (Figs. 2 and 3), leaving means of fewer than 1.0 and 0.5 individuals per PDA plug on methyl cinnamate and 1-octen-3-ol, respectively. Statistically significant reductions in aggregation were also observed at several lower concentrations. Fig. 2Influence of the concentration of methyl cinnamate on the number of Proisotoma minuta aggregated to feed on PDA plugs. Gray bars indicate the means of 18 samples (six replicates of three PDA plugs per dish), and the error bars indicate the standard error (SE) of the mean. Means labeled with different letters differ significantly (Fisher’s PLSD test, P < 0.05)Fig. 3Influence of the concentration of 1-Octen-3-ol on the number of Proisotoma minuta aggregated to feed on PDA plugs. Gray bars indicate the means of 18 samples (six replicates of three PDA plugs per dish), and the error bars indicate the standard error (SE) of the mean. Means labeled with different letters differ significantly (Fisher’s PLSD test, P < 0.05) Discussion Methyl cinnamate exhibited three ranges of concentrations with different abilities to inhibit the aggregation of P. minuta (Fig. 2): no significant effect (a mean of 13.3 to 16.1 individuals per plug) from 0.001 to 0.1 ppm, incomplete but significant inhibition (a mean of 1.2 to 8.2 individuals per plug) from 100 to 1,000 ppm, and strong inhibition at 1,000 ppm. Thus, methyl cinnamate repels collembolans at high concentrations. Methyl cinnamate has previously been shown to be a principal volatile component of the fruiting body of T. matsutake (Terashita et al. 1991), and its content increases during fungal development to reach 22 to 154 ppm on a fresh weight basis by the time the veil breaks (Ohta 1983). High levels of the compound (150 to 360 ppm, with a mean of 236 ppm) are localized in the lamellae and spores (Ohta 1983), where collembolans are frequently observed feeding on the spores and hyphae of the hymenia (Sawahata 2006). The high content of methyl cinnamate in the lamellae may thus prevent aggregation and grazing on gill surfaces by collembolans. 1-Octen-3-ol has previously been shown to be another principal volatile component of the fruiting body of T. matsutake, at concentrations in the fruiting bodies ranging from 2 to 186 ppm on a fresh weight basis (Ohta 1983; Terashita et al. 1991). This volatile component exhibited a clear dose–response relationship (Fig. 3): As the concentration of 1-octen-3-ol increased from 0.001 to 1000 ppm, aggregation of the insects decreased continuously, and the difference from the control became significant at concentrations of 10 ppm and higher. At concentrations of 100 and 1,000 ppm, 1-octen-3-ol significantly reduced aggregation of P. minuta (to an average of 1.2 individuals per PDA plug). These results suggest that Collembola show a different response pattern at different concentrations, with a slight and nonsignificant increase in aggregation compared with the control at concentrations ranging up to 0.01 ppm; at higher concentrations, aggregation decreased significantly (to less than 50% of the control), and at 1000 ppm, aggregation was almost completely eliminated. Bengtsson et al. (1991) also found a weak but nonsignificant attractive effect of 1-octen-3-ol on collembolans (individuals of Onychiurus armatus Tullb.) in soil fungi. Further studies with new methods are needed to explain the attractive effect of low levels of 1-octen-3-ol among a range of collembolans. The present study provides an explanation of why certain collembolans avoid aggregation on T. matsutake fruiting bodies in the field. Both major components of the volatiles present in the fruiting bodies of this species (methyl cinnamate and 1-octen-3-ol) defend against collembolan attack at high concentrations. Wood et al. (2001) reported that 1-octen-3-ol prevents feeding by banana slugs at a concentration of 28 ppm. Pfeil and Mumma (1993) reported that 1-octen-3-ol and 3-octanone seemed to deter gravid females of Megaselia halterata Wood (Phoridae) at high concentrations. Therefore, some fungal species may produce high concentrations of 1-octen-3-ol to defend against mycophagous insects.	[ "matsutake", "methyl cinnamate", "collembola", "1-octen-3-ol", "repellent activity of tricholoma." ]	[ "P", "P", "P", "P", "R" ]
Eur_J_Epidemiol-4-1-2190784	Rise in seroprevalence of herpes simplex virus type 1 among highly sexual active homosexual men and an increasing association between herpes simplex virus type 2 and HIV over time (1984–2003)	Objectives Herpes simplex virus type 1 and type 2 (HSV-1 and HSV-2) are both highly prevalent. The rate of genital HSV-1 transmission is reportedly increasing over time. HSV-2 is considered to be an important risk factor for HIV transmission. We therefore studied changes in the HSV-1 and HSV-2 prevalence in a large cohort of men who have sex with men (MSM) over a 20-year time period. Methods Among 1847 HIV-infected and HIV-uninfected MSM participating in the Amsterdam Cohort Studies, seroprevalence of HSV-1 and HSV-2 was determined and prevalence rate ratios (PRR) and 95% confidence intervals were calculated. Results Between 1984 and 2003 the HSV-1 and HSV-2 prevalence decreased among HIV-uninfected MSM (P < 0.001), but remained stable among HIV-infected MSM. HSV-1 prevalence increased among men with at least 200 sexual partners over lifetime (PRR: 1.49, P < 0.001). The association between HIV infection and HSV-2 became stronger over time (PRR: 3.45, P < 0.001). Conclusions Seroprevalence of HSV-1 and HSV-2 remained high among HIV infected MSM from 1984 to 2003. The association of HIV and HSV-2 increased during the HIV epidemic. Since the proportion of sexual transmission of HSV-1 is rising, it is important to study the potential role of HSV-1 as risk factor for HIV acquisition. Introduction Herpes simplex virus type 1 (HSV-1) is widespread in the general population, while herpes simplex virus type 2 (HSV-2) is more restricted to risk groups such as men who have sex with men (MSM). HSV-1 prevalence is around 70% in the general population [1, 2]. Transmission usually occurs during childhood through oral contact and normally causes oropharyngeal infection. Childhood HSV-1 transmission has declined in industrialised countries, resulting in a lower prevalence of HSV-1 and leaving a larger population of adolescents at risk for sexual transmission of HSV-1. Earlier studies have reported sexually related risk factors for HSV-1 infection in women [3]. Among those persons attending the STD clinic and blood donors, HSV-1 infection is associated with younger age of first intercourse [4]. However, less is known about sexually related risk factors for HSV-1 infection among MSM [5]. HSV-2 infection is usually transmitted sexually and is considered as a marker for sexual risk behaviour in populations [6]. In HIV-infected MSM, the prevalence of HSV-2 is as high as 61% [1], while being 15–25% in the general population [7, 8]. HSV-2 is a risk factor for HIV acquisition, especially in the African setting [9] and in MSM [10]. HSV-2 infected persons are more susceptible for HIV [11, 12]. Moreover, HIV-infected persons are more likely to have subclinical reactivation of HSV-2 and are therefore more likely to transmit the virus [13]. We previously demonstrated a decline in the prevalence of HSV-2 among MSM in Amsterdam between 1984–1997, which could be explained by a decrease in sexual risk behaviour [5]. However, in the second half of the 1990s their sexual risk behaviour increased after effective HIV therapy became generally available. This may have caused an increase in the prevalence of HSV-2 and possibly also in HSV-1 since 1996. We here studied the trend in HSV-1 and HSV-2 prevalence among homosexual men over a 20-year time period (1984–2003) and whether risk factors for infection changed during this period. Methods Study population In 1984 an open and prospective cohort study on HIV seroconversion and AIDS among sexually active HIV-negative and positive homosexual men was started. The Amsterdam Cohort Study (ACS) is still ongoing, although entry criteria with respect to HIV status and age have changed over time. From 1984 until May 1985, both HIV-positive and HIV-negative men were included. From May 1985 until February 1988, only HIV-negative men were allowed in the study. From February 1988 through 1994, HIV-positive and HIV-negative men could enter the study, but since 1995, they must be ≤30 year of age. At an ACS visit, a standardised questionnaire is administered regarding demographics, sexual behaviour, and medical history for sexually transmitted infections (STI). Blood samples are collected for immunologic and virologic testing and for storage. For this study, stored sera (collected at the first cohort visit) taken from ACS participants with at least two cohort visits (1847/2100 (88%)) were tested for HSV-1 and HSV-2. Laboratory methods Sensitive and specific FDA approved serological assay for HSV-1&2 was used (HerpeSelect by FOCUS technologies, USA). Its manufacturer recommends an index value > 1.1 as positive. However, there is evidence that using this cut-off value in HSV-2 studies yields a high rate of false positive results in populations with multiple infections, such as those in Africa [14]. Raising the positive cut-off will increase the specificity [14]. Since the optimal cut-off for our target population has not been established, 100 samples with results in the range of 0.9 and 3.5 were re-tested with a highly specific Western blot. HSV-2 ELISA and Western blot results were concordant for 80/100 samples. The proportion of samples that were positive with both ELISA and Western blot increased with increasing index value (Fig. 1). Based on these results, we consider a cut-off value of ≥2.1 as being positive and an index value <2.1 is classified as negative. The HSV-2 index value <2.1 had 36% concordance with Western blot results, while the HSV-2 index value of ≥2.1 showed 93% concordance. There were no differences in Western blot outcomes between HIV-infected and HIV uninfected MSM. Fig. 1Comparison of HSV-2 serology by ELISA and Western blot. 100 samples were tested with HSV-2 ELISA and re-tested with the Western blot to identify discrepancies between the two assays Blood samples were tested for HSV-1 and HSV-2 at the Public Health Laboratory of the Health Service of Amsterdam. The Western blot was conducted at the Institute for Pathology and Medical Research (ICPMR) in Sydney, Australia. Blood samples are also tested for HIV antibodies by enzyme linked immunosorbent assay (ELISA) (Abbot Laboratories, North Chicago, Illinois, USA; Vironostika, Organon, Teknika, Boxtel, the Netherlands), and when positive, are confirmed by Western blot. Variables and statistical analyses The statistical analyses were based on the data collected at entry of the cohort. Variables used in this study were calendar year of ACS entry, HIV-status, age, nationality, education, age of first homosexual contact, lifetime sexual partners, and self-reported history of syphilis and gonorrhoea in the past 5 years. Variables concerning sexual practices included orogenital, anogenital and oroanal contact in the prior 6 months. Some changes in the questions were made in the questionnaires between 1984 and 2003 regarding the sexual practices. Oroanal contact was not asked for from 1889 to 1994, resulting in approximately 25% missing values on this variable. From 1995 onward, the most important difference was that orogenital contact with ejaculation was asked, while in the previous years orogenital contact in general was asked. The percentage of participants not having orogenital contact was somewhat higher for the years 1995 and 1996. The prevalence of HSV-1 and 2 at the ACS entry was determined and risk factors for HSV-1 and 2 were assessed by calculating prevalence rate ratios (PRR), with their 95% confidence interval (CI). Odds ratios could not be interpret as relative risks, since the rare event assumption was not reached. Therefore PRR’s were directly estimated, using a modified Poisson regression approach [15]. This approach provided a correctly estimated standard error for the estimated relative risk. Since inclusion criteria with respect to age and HIV-status changed over time, all risk factor analyses were adjusted for age and HIV-status. Variables that were statistically significant were included in the multivariate model, using a stepwise forward approach forcing age and HIV-status in the model. We tested whether risk factors changed over time by testing for interaction between variables under investigation and calendar time in the multivariate model. Calendar time therefore was categorised as 1984–1986, 1987–1991, 1992–1996, 1997–2003. Confounding was defined to be present when the included variable caused a change of the prevalence ratio by more than 10%. Interaction was defined to be present when the addition of an interaction term improved the original model and the P-value was <10%. Statistical significance was defined as a P-value <0.05. To reduce residual confounding when measuring the association between HSV and sexual practices, three variables measuring sexual practices over the prior 6 months were included in the model at the same time, together with the lifetime sexual partners. We modelled time trends in the HSV-1 and 2 prevalence with calendar time as a continuous variable using restricted cubic splines with four knots, resulting in a smoothly varying curve. For 77 MSM the HSV-1 index value was missing and 91 MSM had a missing HSV-2 index value. Participants with a missing index value for both HSV-infections were not included in the analyses. MSM with a missing index value for one HSV-infection, but with a known HSV status of the other HSV-infection were included in the analyses for the known HSV serostatus. Finally, sensitivity analyses for HSV-2 were conducted by using the cut-off value of 1.1, as recommended by the manufacturer and by using the cut-off value of 3.5, excluding those with an index value in the grey area (between 0.9–1.1 and 0.9–3.5). However, time trends in prevalence and risk factors found were comparable to when 2.1 was the cut-off value (data not shown). Results General characteristics Between 1984 and 2003, a total of 1847 MSM had at least two visits. General characteristics of the total study group are presented in Table 1. Of the 1847, 1207(65%) MSM were HSV-1 antibody positive, while 759/1847(41%) of the men were HSV-2 antibody positive. Of the total group, 558(30%) were positive for both. Participants were predominantly of Dutch nationality (86%) and had a median age of 29 years (interquartile range: 25–36). Table 1Demographic and sexual characteristics of 1847 homosexual men, for the total study group, and for HSV-1-infected participants and HSV-2-infected participants separately, between 1984 and 2003, with the prevalence ratios for HSV-1 and HSV-2 with their 95% confidence intervalsaCharacteristicsTotalHSV-1-infectionPRR (95% CI)Overall P-valueHSV-2-infectionPRR (95% CI)Overall P-valueTotal18471207 (65)759 (41)Year of study entry<0.0001<0.00011984–1986943675 (72)1461 (49)11987–1991165113 (68)0.87 (0.78–0.97)92 (56)0.84 (0.72–0.98)1992–1996222138 (43)0.88 (0.79–0.96)80 (36)0.66 (0.56–0.77)>1997517281 (54)0.76 (0.70–0.84)126 (24)0.47 (0.40–0.55)Index value<0.9–5358530.9–1.1–29321.1–2.1–100112≥2.1–1107759Missing7791Age<0.0001<0.0001<30 years100257012521≥30 years8456371.20 (1.13–1.29)5071.98 (1.75–2.23)Nationality:0.0060.62Dutch1586 947 (60)1615 (39)1Northern/central Europe116 72 (62)1.04 (0.92–1.18)54 (47)1.08 (0.90–1.38)Non-European145 117 (81)1.17(1.07–1.28)75 (52)1.04 (0.91–1.23)Education0.0050.007Low9677 (80)157 (59)1Middle669408 (61)0.85 (0.76–0.95)239 (36)0.72 (0.59–0.87)High971574 (59)0.81 (0.73–0.81)387 (40)0.79 (0.66–0.94)Missing11577 (67)54 (47)Sexual partners in lifetime1–20860497 (58)1<0.0001277 (32)10.5121–200531360 (68)1.12 (1.04–1.22)212 (40)2.30 (0.93–1.21)>200443339 (77)1.24 (1.15–1.34)262 (59)1.07 (0.95–1.20)Age of first homosexual contact (median, IQR)18 (15–20)17 (15–20)1.01(1.01–1.02)^<0.000117 (15–20)0.99 (0.99–1.01)c0.32HSV co-infection5685681.29 (1.14–1.47)<0.00015681.167 (1.08–1.24)<0.0001HIV infection (%)513 367 (72)1.11 (1.03–1.18)0.007312 (61)1.12 (1.00–1.24)0.05History of gonorrhoea in the past 5 years1053 666 (63)0.88 (0.83–0.93)0.0005424 (40)0.88 (0.79–0.97)0.02History of syphilis in the past 5 years278 219 (79)1.15 (1.07–1.24)0.001197 (71)1.51 (1.36–1.69)<0.0001Orogenital contact in the past 6 monthsb1363892 (65)0.99 (0.82–1.22)0.46565 (41)0.69 (0.55–0.87)0.01Anogenital contact in the past 6 monthsb1222819 (67)1.11 (1.14–1.60)0.002551 (45)1.36 (1.16–1.60)0.0002Oroanal in the past 6 monthsb1081700 (65)1.00 (0.91–1.09)0.99438 (41)0.98 (0.85–1.12)0.80aSince ACS inclusion criteria have changed over times all analyses were adjusted for age (per 10-year increase) and HIV status at cohort entrybAnalyses are also adjusted for the sexual techniques and number of partners to exclude residual confoundingcPer 10 year of increase Prevalence of HSV-1 and HSV-2 over time There was an overall decline in the prevalence of both HSV-1 and 2 between 1984 and 2003 (Tables 1, 2). To investigate time trends in HSV seroprevalence we included an interaction term between time and HIV-status (Fig. 2a, b). Among HIV-negative MSM, the HSV-1 prevalence decreased significantly over time, P < 0.001 (Fig. 2a). Among HIV-positive MSM, the HSV-1 prevalence remained stable over time, P = 0.35(Fig. 2a). Table 2Multivariate model of risk factors associated with HSV-1 infectionHSV-1Overall P-value(a) HSV-1 infectionYear of study entry<0.00011984–198611987–19910.91 (0.70–1.18)1992–19960.83 (0.66–1.02)>19970.75 (0.63–0.90)Age1.13(1.07–1.18)<0.0001HIV serostatusNegative10.01Positive1.10 (1.02–1.18) NationalityDutch10.0006Northern or Central European1.05 (0.92–1.20)Non European1.62 (1.12–1.36) Education 0.25Low1Middle 0.90 (0.79–1.00)High0.84 (0.76–1.06)Sexual partners in lifetime1–2010.00321–2001.13 (1.05–1.25)>2001.13 (1.04–1.23)History of Gonorrhoea in the past 5 years0.97 (0.90–1.03) 0.11History of Syphilis in the past 5 yearsOrogenital contact in the past 6 months1.12 (0.88–1.43)0.42Anogenital contact in the past 6 months1.08 (0.97–1.20) 0.20Oroanal contact in the past 6 months1.02 (0.70–1.13)0.81HSV-2Overall P-value(b) HSV-2 infectionYear of study entry<0.00011984–198611987–19910.86 (0.70–1.06)1992–19960.58 (0.48–0.71)>19970.47 (0.39–0.56)HIV serostatusNegative1<0.0001Positive1.50 (1.37–1.68)HSV coinfection1.15 (1.02–1.30)0.02History of Syphilis in the past 5 years1.21 (1.08–1.36)0.001Orogenital contact in the past 6 months0.69 (0.56–0.84)<0.0001Anogenital contact in the past 6 months1.20(1.08–1.42)0.02Oroanal contact in the past 6 months1.00(0.87–1.15)0.72Fig. 2(a) HSV-1 prevalence in the Amsterdam Cohort Study among MSM, according to the HIV status (1 = HIV positive, 0 = HIV negative) and the 95% confidence interval. (b) HSV-2 prevalence in the Amsterdam Cohort Study among MSM, according to the HIV status (1 = HIV positive, 0 = HIV negative) and the 95% confidence interval The HSV-2 prevalence significantly decreased among HIV-negative men, P < 0.001(Fig. 2b). Results from the regression models showed that, after adjustment for changes in age, nationality, education and changes in sexual risk behaviour, the decline in HSV-2 prevalence among HIV negative MSM remained significant (PRR adjusted 0.92, P < 0.001). In contrast, the HSV-2 prevalence remained stable over time for men infected with HIV (P = 0.12). Again, this result was observed after controlling for age, demographic characteristics and sexual behaviour. Risk factors Risk factors for HSV-1 and HSV-2 infection, adjusted for age and HIV-status, are presented in Table 1. In the final model calendar year, HIV-status, nationality, and number of lifetime sexual partners remained independent predictors for HSV-1 infection (Table 2). For HSV-2 infection, earlier year of study entry, positive HIV-status, HSV-1 co-infection, a history of syphilis and sexual behaviour remained independent predictors (Table 2). Changing risk factors over time Different interaction terms were included in the model. It appeared that the effect of calendar year differed between HIV-infected and HIV-uninfected MSM for both HSV-1 and HSV-2. As shown in Fig. 2a, the association between HIV and HSV-1 became stronger over time. This was due to the decline in HSV-1 prevalence over time among HIV-negative MSM but not among HIV-positive MSM (Fig. 2a). For HSV-2 the association with HIV infection increased with calendar year and was highest after 1996 (Fig. 2b). For HSV-1, also the effect of calendar time differed with respect to nationality, number of lifetime sexual partners, and with HSV-2 co-infection. As shown by the regression model, a decrease in HSV-1 infection over time was observed only among MSM with Dutch or Northern/Central-European nationality. The association between HSV-1 and having non-European origin became stronger over time. The adjusted PRR was 1.08 (P = 0.3) before 1986 and the adjusted PRR became 1.58 (P < 0.0001) for the time period after 1996. Also the association between HSV-1 and a higher number of lifetime sexual partners became stronger after 1996. Figure 3a shows the prevalence of HSV-1 infection over time according to the number of sexual lifetime partners. A decrease in the HSV-1 prevalence was seen among MSM with fewer than 21 partners (P < 0.0001), while among MSM with more than 200 partners, the HSV-1 prevalence increased between 1988 and 2003 (P = 0.01). Fig. 3(a) Prevalence of HSV-1 and the 95% confidence interval over time among HIV negative MSM, according to number of lifetime sexual partners; 1 = 1–20, 2 = 21–200, 3>200 partners. (b) Prevalence of HSV-2 and the 95% confidence interval over time among HIV negative MSM, according to number of lifetime sexual partners; 1 = 1–20, 2 = 21–200, 3 > 200 partners For HSV-2, the effect of calendar time differed with respect to the number of lifetime sexual partners and with HSV-1 co-infection. A large number of lifetime partners also was strongly associated with HSV-2. However, a decrease in the HSV-2 prevalence was seen among all categories of lifetime sexual partners (Fig. 3b), but this decrease was stronger for MSM with fewer than 21 partners and for those MSM with 21–200 partners. Discussion In the present study, we demonstrated an overall decrease in HSV-1 and HSV-2 prevalence among HIV-negative MSM, but not among HIV-positive MSM. In the 1984–2003 period, the association between HSV-2 and HIV among MSM became stronger over time, and HSV-1 prevalence increased in highly sexually active HIV-negative MSM. To our knowledge, this is the first study based on almost 20 years of HSV-1 and HSV-2 prevalence data among MSM. The decrease seen in the seroprevalence of HSV-1 and HSV-2 over time, could not be explained by changes in demographic characteristics or sexual behaviour. The decrease in HSV-1 is likely to reflect a decrease in childhood transmission by the oropharyngeal contact of HSV-1. Since fewer individuals are infected in childhood with HSV-1, there is a growing population of persons at risk at the time they become sexually active, resulting in a larger proportion of sexual transmission of HSV-1. Two risk factors for HSV-1 infection that could be important in its sexual transmission were identified by this study. First, HIV infection is associated with HSV-1 seropositivity. Second, the prevalence of HSV-1 was higher among highly sexual active MSM (at least 200 lifetime sexual partners). An association also shown earlier by others [3]. HIV infection in this respect may reflect an epidemiological marker for sexual risk behaviour for HSV-1 transmission. HSV-1 prevalence did not decrease in those infected with HIV, and we consider that genital HSV-1 infection has a growing role in the acquisition of HIV. Likewise, HSV-2 prevalence did not decline over time among those infected with HIV, whereas a decline was noted among HIV uninfected MSM. Although HSV-2 is sexually transmitted, we did not find an association between a higher number of lifetime partners and HSV-2 infection. HSV-2 was also highly prevalent among MSM with 1–20 lifetime partners. Mainly between 1984 and 1995, there were no major differences in the proportion of MSM infected with HSV-2. This suggests that HSV-2 is highly sexual transmissible and when having a low number of life time partner the risk of receiving a HSV-2 infection is still very high. The results of this study show a protective effect of orogenital contact for HSV-2 infection, which might be explained by the fact that anogenital contact is a stronger predictor for HSV-2 infection. All variables measuring sexual practices are included in the analyses at the same time. Since most MSM practised all the techniques during the same time period these practices could not be analysed as independent risk factors. The stronger effect of anogenital contact might have overruled the effect of orogenital contact, resulting in a protective effect of orogenital contact. The overall prevalence of HSV-2 in this study is similar to that among MSM in San Francisco in 1989, but higher than the prevalence found in more recent studies in the US [16–18]. The lower prevalence in those more recent studies probably reflects the decline in HSV-2 prevalence over time, as found in our study. Russell et al., found high prevalence rates of HSV-2 among HIV-infected MSM in Australia [1]. The HSV-2 seroprevalence was more than twice as high as among HIV-uninfected MSM; there was no significant difference in HSV-1 prevalence between HIV-infected and HIV-uninfected MSM. Several epidemiological studies have described an association between HIV and HSV-2 [10]. HSV-2 is recognised as a risk factor for HIV acquisition in MSM. In addition, HSV-2 may up-regulate HIV and increase local HIV replication on mucosal surfaces, leading to an increased risk of HIV transmission. Our study is the first to show an increase in the association between HSV-2 and HIV since 1996, suggesting that HSV-2 may play a growing role in driving the HIV epidemic in MSM. If this is the case, prevention of HSV-2 may well contribute to the prevention of HIV among highly sexually active MSM. Although serological screening for HSV-2 among MSM is still under debate, the increasing association between HSV-2 seropositivity and HIV is an argument in its favour. Several reasons against screening have been raised, such as the lack of a reliable serological test. We are aware of the low specificity of the various HSV-2 serological assays. However, these serological assays might be useful as a screening tool, when used with an increased cut-off value less individuals will be classified as false positive. A second argument against serological screening is that HSV-2 infection is largely asymptomatic and condom use appears only partially protective against HSV transmission. These factors complicate the prevention of HSV-2 and genital HSV-1 infection. However, it has been shown that half of the patients, initially unaware of their HSV-2 infection are able to recognise symptoms after being educated to do so [19]. Also, knowledge of the HSV-2 status of a sexual partner has been associated with a reduced risk of HSV-2 transmission [20]. Antiviral drugs used as suppressive therapy will lower the frequency of recurrences by 70–80% [21, 22]. A combined approach of offering serological screening to highly sexual active MSM together with encouraging condom use to reduce the risk of HSV transmission and using suppressive therapy among those with recurrent lesions might eventually play an effective part in controlling the HIV epidemic among MSM. One limitation of our study is its cross-sectional design. As a consequence, we cannot reveal the relation between HIV and HSV infection, being unable to determine which occurred first. As both HIV and HSV are sexually transmitted diseases, their association may well reflect shared sexual behavioural practices leading to transmission as well as a biological relation. Longitudinal studies, in which incident HIV and HSV cases are captured are therefore needed to give more insight into the relationship between HIV and HSV as affected by changes in sexual risk behaviour. The results of this study have two implications for HIV and HSV research among highly sexually active MSM. First, it appears that HSV-2 and HIV are now more strongly related than in the early days of the HIV epidemic. As a vaccine against HSV-2 for MSM is not yet available, a determination of the extent to which the prevention of HSV-2, specially aimed for MSM at high risk for HIV, can contribute to controlling the HIV epidemic is needed. Second, since the extent of sexual transmission of HSV-1 is rising, we need to clarify its potential role as a risk factor for HIV acquisition in longitudinal studies.	[ "prevalence", "hiv", "hsv-1", "hsv-2", "msm" ]	[ "P", "P", "P", "P", "P" ]
Bioprocess_Biosyst_Eng-2-2-1705492	Dynamics of amino acid metabolism of primary human liver cells in 3D bioreactors	The kinetics of 18 amino acids, ammonia (NH3) and urea (UREA) in 18 liver cell bioreactor runs were analyzed and simulated by a two-compartment model consisting of a system of 42 differential equations. The model parameters, most of them representing enzymatic activities, were identified and their values discussed with respect to the different liver cell bioreactor performance levels. The nitrogen balance based model was used as a tool to quantify the variability of runs and to describe different kinetic patterns of the amino acid metabolism, in particular with respect to glutamate (GLU) and aspartate (ASP). Introduction Bioreactor technology for extracorporeal liver support using primary human liver cells has been developed within the last decade [1]. The design of a bioreactor for maintaining the hepatocyte’s full functionality is of great importance. The used multi-compartment bioreactor consists of three interwoven, independent capillary membrane systems. Two of them (compartments 1 and 2, in the following aggregated to the ‘perfusion compartment’) provide decentralized plasma flow and the third one (compartment 3) provides oxygen supply to the cells, which are localized in the extracapillary space (compartment 4, ‘liver cell compartment’). This three-dimensional spatial structure represents an artificial equivalent of the hepatic vasculature at the lobular level. The bioreactor is integrated into a perfusion system that enables monitoring and control of system conditions (see Fig. 1). Previous studies have shown that primary liver cells reconstitute to tissue-like structures after inoculation into the bioreactor and that they maintain metabolic activities over several weeks [2–4]. This bioreactor therefore provides a valuable tool to analyze the dynamics and network structures of the physiological and molecular interactions of liver cells under standardized conditions that closely reflect the situation in the natural organ. Fig. 1Scheme of the liver cell bioreactor with the perfusion circuit, the two inflow streams and the outflow stream. Measured data were acquired from the waste Recently, data from this bioreactor system has been analyzed statistically as well as by fuzzy cluster and rule based data mining and pattern recognition methods in order to identify early performance predictors for the bioreactor’s long-term performance [5–7]. With respect to the prominent role of hepatocytes in the amino acid metabolism, the analysis and modeling of metabolic pathways of amino acids could provide important information on the functional state of hepatocytes cultured in vitro. In initial studies the role of amino acid regulation in extracorporeal liver support systems was analyzed [8]. More recently, three different modeling paradigms, i.e. correlation networks, Bayesian networks and systems of differential equations were applied to characterize the concentration profiles of six amino acids and related nitrogen-containing compounds in bioreactor cultures of primary human liver cells [6]. The present work reports on the model based analysis of the measured kinetics of 18 amino acids as well as ammonia (NH3) and urea (UREA) of 18 liver cell bioreactor runs. Materials and methods Cells and bioreactors Primary human liver cells were isolated according to a method described elsewhere [9] from human donor livers (n=18) that were not suitable for transplantation due to organ injury. After isolation, cells were cultured within bioreactors over 7–34 days. (This broad variability in duration is caused by their use for liver support from the seventh day onwards and by the performance. Low performance runs were finished earlier. High performance runs were continued for a longer period of time. However, only the first 6 days of the runs were analyzed in this study.) Concentrations of free amino acids in the culture perfusates were determined by an automated reversed phase high performance liquid chromatography system (RP-HPLC) with precolumn derivatization using the ophthaldialdehyde method [10]. NH3 and UREA concentrations were determined using routine clinical analyzers (Roche Diagnostics, Heidelberg, Germany). With respect to the model based analysis, two compartments of the multi-compartment capillary membrane bioreactor system were considered: The ‘liver cell compartment’ with the volume V2=600 mL contains the liver cells in the inter-capillary space. The ‘perfusion compartment’ with the volume V1=900 mL supplies a stream through the inside of the capillaries. This perfusion stream carries the concentrations of the compounds that are supplied to or removed from the bioreactor. Decentralized mass exchange at low gradients is achieved by independent perfusion of two medium capillary systems, enabling different perfusion modes [10]. In this study, counter-directional medium flow was performed to facilitate rapid matter distribution. Due to the high flow rate of the perfusion stream (250 mL min−1), an almost ideal mixing within the perfusion compartment can be assumed (Fig. 1). To this perfusion compartment two time-variant inflow streams are added with the flow rates FA(t) and FB(t) as defined by Eq. (1). FA(t) follows a step function from FA1=150 mL h−1 down to FA2=50 mL h−1 switching at tA(=1 day). FB(t) switches from FB1=0 up to FB2=1 mL h−1 at the time tB (see Table 4). The outflow rate F0(t) to the waste equals the sum of both inflow rates. (This is realized by overflow keeping a constant pressure in the perfusion compartment). The inflow rate FA(t) carries the 18 amino acids and NH3 at the concentrations cAi (see Table 1). The inflow rate FB(t) carries only the amino acid aspartate (ASP) at the concentration cB15=1,500 μmol L−1, i.e. cBi=0 for all i≠15. The time courses of the concentrations c0i(t) in the outflow stream may be considered to describe the response of the medium to the inoculation of the bioreactor with cells. These concentrations are in steady-state at c0i(0)=cAi prior to inoculation. Due to the almost ideal mixing in the perfusion compartment, the concentrations c0i(t) in the perfusion stream are the same as in the outflow stream. Table 1Model variables ci and c0,i for the liver cell compartment and the perfusion compartment, respectively, their initial values ci(0) and c0,i(0) before inoculation of the bioreactor with liver cells, the concentration cAi in the inflow stream fed with the flow rate FA(t) according to Eq. (1) and the stoichiometric coefficients si for nitrogenNameModel variables ci and c0,Ici(0)=c0,i(0)=cAi [μmol L−1]Stoichiometric coefficient siLEUc1, c0,122581HISc2, c0,29403ARGc3, c0,326044VALc4, c0,49871TRPc5, c0,54642PHEc6, c0,613001ILEc7, c0,75001ALAc8, c0,818251TYRc9, c0,922151LYSc10, c0,103272METc11, c0,112591SERc12, c0,128611GLYc13, c0,1319571THRc14, c0,148591ASPc15, c0,152841ASNc16, c0,161682GLUc17, c0,172651GLNc18, c0,186872NH3c19, c0,19411UREAc20, c0,2002PROTc21, c0,210– Data A data set with the elements ci,j,k (i=1,...,20; j=1,...,18; k=1,...,K) for 20 kinetic variables and 18 bioreactor runs at K time points tk was analyzed. The kinetics over the first 6 days of the bioreactor runs were analyzed here (tk=6 days). Concentrations of NH3 and UREA were measured daily (tk=0–6 days). Amino acid concentrations were determined up to the third day daily and every third day afterwards (tk=0, 1, 2, 3, 6 days). The samples for the measurement of the ci,j,k were taken from the waste of the liver cell bioreactor system (see Fig. 1) which was emptied daily after accumulation of the outflow stream over the period of 24 h. The model based analysis described focuses on the amino acid metabolism as quantified by the measured time series of the concentrations of 18 amino acids (i=1,..., 18; see Table 1) as well as of NH3 (i=19) and UREA (i=20). Each run was labeled by Lj ∈ {‘low’, ‘medium’, ‘high’} describing the performance with respect to the long-term maintenance of the functionality of the liver cells within the bioreactor. Seven runs were labeled ‘high’ (j=1,..., 7), seven runs ‘medium’ (j=8,..., 14) and four runs ‘low’ (j=15,..., 18). This performance labeling was provided by an expert based on his assessment of altogether 99 variables that were measured to quantitatively characterize the system. Differential equation system The differential equation system (2.1–2.10) was developed to describe the measured kinetics of the 18 amino acids as well as of NH3 and UREA (Table 1). This model takes into account two compartments as described in Cells and bioreactors, the ‘perfusion compartment’ and the ‘liver cell compartment’. The Eq. (2.1) describe the dynamics of the components i in the perfusion compartment by four terms: The first and second term represent the fresh medium inflow with the volumetric rates FA(t) and FB(t), respectively, from the reservoir into the perfusion compartment with the volume V1. The reservoir concentrations cAi of the components i are specified in Table 1. The volumetric rates FA(t) and FB(t) are specified by Eq. (1). The third term in Eq. (2.1) denotes the outflow from the perfusion compartment into the waste with the volumetric rate F0(t). The last term describes the diffusion between the perfusion and the liver cell compartment with its rate being proportional to the difference of the concentration c0,i in the perfusion compartment and the concentration ci in the liver cell compartment. The parameter p0 is proportional to the diffusion coefficient and assumed to be the same for all compounds to simplify parameter identification. The Eqs. (2.2–2.9) formulate the changes of the concentrations ci in the liver cell compartment with the volume V2 due to the exchange with the perfusion compartment (the first term on the right side of each differential equation) and the metabolic reactions (the following term(s) on the right side of each differential equation) as drawn schematically in Fig. 2. The model drawn in Fig. 2 and its Eqs. (2.1–2.10) were constructed based on the mean kinetics of 20 measured variables averaged over the seven high-performance runs H1–H7 as shown in Fig. 3. The kinetics of these seven individual runs are highly correlated with their correlation coefficients being greater than 0.8 for at least six of the seven high-performance runs and the 11 amino acids MET, SER, THR, ARG, HIS, GLY, TRP, LEU, VAL, PHE, and ILE [6]. The dynamics of these amino acids are characterized by monotonously, almost exponentially decreasing kinetics [6]. Their dynamics were therefore described by first order reactions as formulated by the second term in Eq. (2.2). The observed decreasing kinetics of amino acid concentrations is the net result of anabolic and catabolic reactions, where the catabolism of amino acids (which was considered here only) surpasses the amino acid synthesis and release by proteolytic activities (which were neglected here). The second term with the parameters p1,..., p14 in Eq. (2.2) can be interpreted mainly as amino acid uptake by transamination and oxidative deamination (see Table 2). In addition to the 11 amino acids with highly correlated and monotonously decreasing kinetics, the time courses of the concentrations of the amino acids ALA, LYS, and TYR were, for simplification, also modeled using Eq. (2.2). The kinetics of the amino acids ASP, ASN, GLU, and GLN were modeled in greater detail by Eqs. (2.3–2.6). Glutamate (GLU) is known to be the central compound of amino acid metabolism. Its kinetics were found in [6] to be highly correlated with ASP. ASP was fed by the inflow rate FB(t) (see Cells and bioreactors) to avoid or compensate the exhaustion of ASP which is essential for NH3 elimination via the urea cycle. The degradation of amino acids is interpreted in a simplified form by transamination to GLU for the amino acids LEU, HIS, ARG, VAL, TRP, PHE, ILE, ALA, TYR, and LYS (denoted by ‘AAg’ in Fig. 2). The amino acids MET, SER, GLY, and THR (denoted by ‘AAn’ in Fig. 2), however, are not transaminated forming GLU, but deaminated via different pathways forming NH3. The parameters with the indices 15–18, 21–23 as shown in Fig. 2 and Table 2 describe the metabolic fluxes between ASP, ASN, GLU, GLN, and NH3. They can be interpreted in a simplified way by reactions catalyzed, e.g., by the enzymes asparaginase, aspartate aminotransferase (AST), glutamate oxalacetate transaminase (GOT), glutamine synthetase (GS) and glutamate dehydrogenase (GLDH). In general, the biochemical reactions were assumed to be either of linear or bilinear nature (first or second order). In two cases, however, additional assumptions were required to obtain sufficient model fits: (1) For the aggregated modeling of the NH3 elimination via the urea cycle in the presence of ASP a non-linear Michaelis-Menten-type kinetics was used (see Eq. 2.8). This reflects the observation that the UREA formation rate does not significantly increase any more at strongly elevated ASP concentrations (see ASP and UREA in Fig. 3). (2) Protein synthesis, i.e. anabolism, seems to be switched on or is strongly increased after the third day. This was modeled using Eq. (2.10). The total protein (PROT) was not measured, but a corresponding variable c21 was hypothetically introduced by Eq. (2.9) to improve the model fit of the other 20 variables to the measured data for t>3 days. Fig. 2Structure of the model Eqs. (2.2–2.10). The numbers within the circles denote the indices i of the corresponding concentrations ci and the numbers at the arrows denote the indices m of the corresponding model parameters pm. (The diffusion process modeled by the parameter p0 is not shown here.) ‘AAg’ denotes the amino acids LEU, HIS, ARG, VAL, TRP, PHE, ILE, ALA, TYR, and LYS; ‘AAn’ denotes the amino acids MET, SER, GLY, and THRFig. 3Mean kinetics (± SD) of 20 measured variables (18 amino acids, NH3, UREA) averaged over the kinetics of the n=7 high performance runs H1–H7 (n=7)Table 2Parameters pm of the model (2) with interpretation, estimated values and confidence intervals [pmlow, pmhigh]; enzyme activities: AST aspartate aminotransferase, GOT glutamate oxalacetate transaminase, GS glutamine synthetase, GLDH glutamate dehydrogenase; values and units of the parameters pm of the model (2) as identified by the model fit to the mean kinetics averaged over the seven high performance runs (Fig. 3); PC_1, PC_2: first and second principal componentmInterpretation/enzymespm [pmlow, pmhigh]UnitPC_1PC_20Diffusion49628 [34681, 132980]mL h−10.18820.35301Aminotransferases (Transamination)LEU1.816 [0.000, 4.629]day−10.2679−0.01872HIS16.29 [3.42, 620.16]day−10.0251−0.29483ARG100.00 [88.75, 2655]day−1−0.14300.22534VAL1.147 [0.428, 23.112]day−10.2506−0.06175TRP15.99 [7.42, 833.88]day−10.2022−0.10636PHE4.148 [1.740, 9.520]day−10.2595−0.03647ILE3.723 [0.000, 6.213]day−10.2491−0.02688ALA1.573 [0.000, 6.213]day−10.25140.01359TYR1.135 [0.559, 3.450]day−10.1866−0.273010LYS0.192 [0.000, 2.164]day−10.23580.005311Other specific reactionsMET34.53 [7.27, 284.76]day−10.25870.117112SER10.80 [3.34, 102.71]day−10.26980.039913GLY10.76 [2.69, 251.25]day−10.26740.084914THR9.326 [3.16, 420.70]day−10.2543−0.054615Asparaginase0.0009 [0.0000, 0.002]day−1 μmol−1 L0.18250.4175163.593 [1.215, 8.505]day−10.22540.297117AST50.08 [39.25, 59.37]day−1−0.06500.034118GOT56.73 [50.35, 71.85]day−10.18190.145619Urea Cycle Km (ASP)225.76 [187.7, 378.9]day−1 μmol−1 L0.2000−0.3185205.00 [0.9375, 6.17]μmol L−1−0.10230.181321GS0.0033 [0.001, 0.007]day−1 μmol−1 L0.1198−0.1839220.0205 [0.001, 0.039]day−10.1017−0.391723GLDH, Protein Synthesis40.68 [24.12, 50.38]day−10.11210.0330240.5475 [0.465, 0.813]–−0.0057−0.1114 The model parameters were identified by fitting the model to the measured data ci,j,k minimizing the scaled mean square error (mse) as defined by Eq. (3) for run j. The model fitting error mse was scaled by the square of the maximum of the respective measured variable. According to Eq. (3), the kinetics c0,i(t) obtained from the simulation of Eqs. (2.1–2.10) with the initial values listed in Table 1 were averaged over 24 h (i.e. over the time interval of the accumulation of the bioreactor outflow in the waste where the samples for the measurements were taken from daily) and then compared with the measured data ci,j,k. The differential equations were solved using a Runge-Kutta fourth order algorithm. The parameter identification by model fitting minimizing the mse was performed using a simplex search method. MATLAB tools (The MathWorks, Inc., Natick, MA, USA) were used for all calculations. Results and discussions The mean time profiles of the 20 measured variables each averaged over the high performance runs H1–H7 are shown in Fig. 3. The results of the model fitting to these averaged data are shown in Fig. 4 and the identified model parameters p0,..., p24 are listed in Table 2 (third column). The confidence intervals for the parameter values (fourth and fifth column in Table 2) were identified by repeated model fitting to the randomly disturbed time series data with means and standard deviations as shown in Fig. 3. The fit of the modeled kinetics to the measured data is acceptable for 18 of the 20 variables with the exception of lysine (LYS) and alanine (ALA). The initial increase of the LYS kinetics could be hypothetically explained by proteolytic activities. The final increase of ALA can be explained by the ALA aspartate transferase activity. The decreasing kinetics of 11 out of the 18 amino acids is explained by transamination and oxidative deamination forming GLU and NH3. The feeding of ASP after the third day results not only in an increase of the concentration of ASP but also of asparagine (ASN), GLU, and glutamine (GLN) due to the activities of the enzymes AST, GOT, GS, and asparaginase (alternatively or additionally the activity of asparagine synthetase could be included in the model). Fig. 4Measured and simulated kinetics of the concentrations of 18 amino acids, NH3 and UREA. The measured kinetics (dots) are those of the mean profiles shown in Fig. 3 (n=7). The simulated kinetics c0i(t): (thin lines) are those obtained from the model (2). The simulated kinetics c0i(t) were averaged according to Eq. (3) over the past 24 h (thick lines) in order to use them for model fitting to the measured data that were acquired from the waste (outflow accumulated over 24 h) Table 3 lists the model parameter values pm with the respective model fitting error mse as identified by individual model fitting to the seven high performance runs. Figs. 5–7 show the results of the model fitting to the data for the individual high performance runs H1, H5, and H7, respectively. The kinetics of run H2 (not shown here) is very similar to the mean kinetics of the seven high performance runs shown in Fig. 4. The quasi-stationary concentrations of several amino acids, such as MET and LEU, are lower in run H5 (see Fig. 6) than in run H2 (compare Fig. 4). This results in a p0 value, which is about 6.5× higher for run H5 than for run H2 (see Table 3). While the variability of the parameter values p1,..., p24 can be explained by differences in the liver cell material obtained from different individual donors (having different age, weight, liver damage, etc.), it is not entirely clear at this stage how to explain such different values of the parameter p0 that represents diffusion and potentially other phenomena taking place across and beyond the membrane between the perfusion and the liver cell compartment. These issues are addressed in greater detail in [11] as well as in studies described in [12] to elucidate the dynamics of electrolyte distribution in the bioreactor without liver cells using tracer experiments and input/output systems analysis.Table 3Values of the parameters pm of the model (2) with the values of the scaled mean square error (mse) as identified by individual model fits to the seven high performance runs H1–H7mH1H2H3H4H5H6H70937624859915419060636318770949398101012.43642.11022.75503.75785.85762.930202224.49445.133536.905850.803938.03622.913921.94173342.2608285.530379.992797.785355.361589.3504278941.50421.94421.33832.24653.23482.2254058.391812.479319.586688.639644.309819.81934.866662.86253.79438.639110.571211.08434.07560.834574.83146.49384.00106.018610.17706.62230.010283.54400.91921.77712.03255.82352.7174092.13891.56292.44631.57461.29080.854801000.02.21511.52131.69980.470801110.800524.378383.028053.8904116.586825.381312.99081210.07679.470921.121816.373531.182010.25363.38151310.05735.139727.421322.631545.111610.22233.25321411.66507.21708.506922.044225.209312.49543.0131150.00080.00070.00180.00070.01290.00080.0007165.15803.47423.85547.125130.91683.79471.20831740.468559.405837.907663.556960.347343.376069.34681869.917457.329772.578270.4225109.786752.520163.280919291.5727192.9329341.5933334.0218238.6410263.7827111.2798205.56698.42293.25044.82613.82513.78088.6594210.00140.00.01140.00260.00060.00520.0034220.05130.00.00310.02800.00040.01570.00002352.597035.600151.898249.646767.233839.360326.0489240.57070.68330.19630.59330.61320.68230.6408mse0.01620.01430.02020.01150.01350.01450.0321Fig. 5Measured data of the 20 variables of bioreactor run H1 with the simulated kinetics of the model fitted to the data of H1Fig. 6Measured and simulated kinetics as in Fig. 5, but for run H5 The model, which was developed using the kinetic patterns of high performance runs was also fitted to the data of the 11 runs with medium and low performance. The aim of these model studies is to elucidate causes of low liver cell bioreactor performance. The fit of the model with 25 parameters to 18 bioreactor runs results in an array of 25×18 parameters. Three selected parameters with their values as identified by fitting the model to the individual 18 runs are listed in Table 4. Considering a tolerable model fitting error mse of about 2% or less, the model fit proved satisfactory for six of the seven high performance runs (with the exception of run H7) and for two of the seven medium performance runs (M8 and M10). However, the model fit was not acceptable for the high performance run H7, the medium performance runs M9, and M11–M14 as well as all four low performance runs L15–L18. The main reason for the insufficient model fits is that the measured kinetics of leucine (LEU), isoleucine (ILE), valine (VAL), ALA, and other amino acids are not decreasing in these runs. This behavior is shown in Fig. 7 for run H7. In these cases some parameter values, e.g., for p1, were found to be zero (negative parameter values were set to zero; see Table 4). These zero parameter values cause that the medium and low performance runs displayed in Fig. 8 are artificially close to each other. Run H7 labeled as a high performance one appears to be a medium performance run according to the model simulation results (see Figs. 7 and 8).Table 4Parameters for the 18 liver cell bioreactor runs: tB as used in Eq. (1); the star () for run 13 denotes that the feed FB(t) was switched on during the period from the second to the fourth day; the scaled mean square error (mse) and the parameters p1, p2, and p22 as used in Eq. (2) were identified by model fitting to the data of the high performance runs H1–H7 (averaged) and of the individual high performance runs H1,..., H7, medium performance runs M8,..., M14 and low performance runs L15,..., L18Run No.tB[d]mse [-]p1 [d−1]p2[d−1]p221,000 [d−1]H1-H7 (averaged)30.00991.81616.2920.5H130.01622.4364224.494451.2761H230.01432.11025.13350H330.02022.755036.90583.1020H430.01153.757850.803928.0101H530.01355.857638.03620.4057H630.01452.93022.913915.7285H730.0321021.94170.0186M830.01471.03678.77420.0497M950.033002.86110.0381M1050.01990.000811.29470.0578M1130.03130.001032.36580.2347M1230.02590.9387168.22990.0722M132–4*0.15590.000311.64390.1391M1430.04920.000310.97870.1615L1500.2747039.92250.0572L1660.080302.59350.0268L1730.08640137.27210.0131L1830.132101.26130.0560Fig. 7Measured and simulated kinetics as in Fig. 5, but for run H7Fig. 8Biplot of the first and second principal components of the model parameter matrix {p0,...,p24} as identified for the 18 liver cell bioreactor runs (high performance runs H1,...,H7, medium performance runs M8,...,M14, low performance runs L15,...,L18) Figure 8 displays the identified model parameters of the 18 liver cell bioreactor runs after principal component analysis (PCA) of the parameter array for the first and second principal components in a biplot [13]. The first principal component represents 53% of the total parameter variance and the first and second principal components together represent almost 64% of this variance. Most of the parameters p1–p14 (with the exception of p2, p3, p9) representing transamination and oxidative deamination activities mainly constitute the first principal component of the parameter array (see Table 2: PC_1>0.2 for these parameters). These 11 parameters are correlated: The 55 correlation coefficients calculated between the 11 parameters estimated over the 18 runs have values between 0.54 and 0.987 with the median of 0.87 and the mean of 0.84. The values of these 11 parameters are high for six of the seven high performance runs and low for run H7 as well as for the medium and low performance runs as shown in Table 4 for parameter p1. The value of the parameter p1 is strongly correlated with the bioreactor performance (p-value 0.00017 by the two-sided t-test of pm values of high versus medium or low performance runs). The mean value of the parameter p1 averaged over the high performance runs equals 2.84 (±1.77). The parameter is zero for all low performance runs. In Fig. 8 the medium and low performance runs are displayed in the left part describing low transamination and oxidative deamination activities whereas six of the seven high performance runs (with the exception of run H7) are displayed in the right part representing high transamination and oxidative deamination activities. Some parameters are highly correlated over the 18 runs with a correlation coefficient r of more than 0.98, e.g., the parameters p11 and p13 (representing the catabolism of methionine and glycine) as well as p4 and p7. Also, the parameter p1 is highly correlated (r>0.96) with the parameters p4 and p7. The high correlation of the parameters p1, p4, and p7 representing the transamination of LEU, ILE and VAL can be hypothetically explained by the fact that LEU, ILE and VAL are branched-chain amino acids which cannot be transaminated in hepatocytes. To transaminate these amino acids, the activity of non-parenchymal liver cells is required. Among the high performance runs there are further individual differences that are displayed in Fig. 8 by a low second principal component PC_2 (e.g., high values of the parameters p2 and p22 for run H1) and a high PC_2 (e.g., low values of the parameters p2 and p22 for runs H2, H5, and H7). The parameters p2 and p22 represent the uptake of histidine and glutamine (see Table 2). As shown in Fig. 5 for run H1, the high parameter value p2 results in a low stationary concentration of HIS and the high parameter value p22 results in high NH3 and low GLU levels. These effects could be related to the amino acid transport system N (SN1) that mediates specifically the uptake of histidine and glutamine [14]. The parameter p21 that represents the activity of GS was found to be very low for the high performance runs H2 and H5 (see Table 3). GS plays an important role in the spatial organization (zonation) of the liver and is exclusively expressed in pericentrally located hepatocytes [15]. The low GS activity in the high performance runs H2 and H5 could therefore be caused by a high proportion of periportal versus pericentral hepatocytes in the cell preparations for these two runs. Conclusion The kinetics of 18 amino acids and the related nitrogen-containing compounds NH3 and UREA in a primary human liver cell bioreactor were analyzed and modeled using a differential equation system. The model focuses on the kinetics of GLU and ASP as well as on the formation and elimination of NH3 and the synthesis of UREA. It describes the degradation of amino acids by transamination, oxidative deamination and other specific reactions. In addition, the activities of selected enzymes such as AST, GOT, GS, and GLDH as well as, in a more aggregated form, the activities of urea cycle enzymes are included. The differential equation system does not represent a fully mechanistic but rather a phenomenological model since essential metabolic activities had to be neglected because they cannot be identified based on the measured data. The differential equation system allows the analysis of a number of representative liver cell functions in terms of their kinetic behavior. The identification of the model parameters by fitting the model responses to the measured data was used to generate hypotheses about the causes of specific differences between the bioreactor runs. The model fits were found to be very satisfactory for eight high and medium performance runs. The model is however inadequate for low performance runs and with respect to the LYS kinetics also for high performance runs. This is probably caused by the neglection of proteases activities in the model that appear to be relevant for low and medium performance runs. Both, protein synthesis and degradation could not be modeled in detail due to the lack of representative protein measurements. The applied model based analysis of data obtained from the bioreactor system can be used to quantitatively evaluate the functional state of liver cell cultures under high performance conditions intended for clinical application in extracorporeal liver support systems. The approach can also be used to study the effect of several exogenous factors, e.g., of hormones or drugs, on hepatocyte metabolism in vitro. The model based analysis methods applied here therefore provide suitable tools for in silico studies supplementing in vitro studies of hepatocyte functions in a systems biological way.	[ "bioreactor", "liver support", "systems biology", "metabolic network" ]	[ "P", "P", "P", "R" ]
Osteoporos_Int-4-1-2267486	Development and application of a Japanese model of the WHO fracture risk assessment tool (FRAX™)	Summary The present study estimated the 10-year probability using the Japanese version of WHO fracture risk assessment tool (FRAX™) in order to determine fracture probabilities that correspond to intervention thresholds currently used in Japan and to resolve some issues for its use in Japan. Introduction Fractures related to osteoporosis have become a major health and economic burden in Asian countries just as they have in North America and Europe. An estimated 117,900 cases of hip fracture occurred in 2002 [1], and the incidence in Japan has increased in the past 10 years [1, 2]. Asia will be expected to have the highest absolute increase in fracture number because it has the largest population. Early detection of individuals with high fracture risk using clinical risk factors would have a substantial impact on reducing the burden of fractures in Asia. A series of meta-analyses on prospective population-based cohorts has identified a number of clinical risk factors that contribute to fracture risk independently of BMD at the femoral neck [3]. The integration of these risks would, therefore, enhance the predictive value of BMD [4]. The risk factors comprise age, sex, bone mineral density, body mass index (BMI), long-term use of glucocorticoids, parental history of hip fracture, history of fragility fracture, smoking, alcohol consumption (3 or more units/day), and secondary osteoporosis such as rheumatoid arthritis. A WHO scientific group has proposed that the 10-year probability for fracture is used to express fracture risk for clinical assessment [5] and to determine intervention thresholds [3] The aim of this study was to create a fracture probability model based on the methodology of the WHO risk assessment tool (FRAX™) [6] calibrated to the epidemiology of Japan. In addition, several problems need to be resolved before the FRAX™ model is applied to Japan. First, the FRAX™ tool inputs femoral neck BMD and the Z-score or T-score is based on the NHANES III reference data base. In Japan, BMD at the lumbar spine is widely used clinically because the physical size of Japanese people is smaller than that of Western people, giving rise to a view based on little evidence that the reproducibility of measurements at the femoral neck BMD would be poorer than that at the lumbar spine. Furthermore, data on the young adult mean (YAM) and the mean at each age are installed in the DXA systems in Japan, and programmed to calculate T- and Z-scores from Japanese reference data. In addition, the Japanese Society for Bone and Mineral Research [7, 8] provide recommendations for the diagnosis of osteoporosis and intervention based on YAM, and these are widely used in clinical practice. Against this background, additional aims of the present study were to provide fracture probabilities based on the FRAX™ tool that were equivalent to currently accepted intervention thresholds, explore the impact of using Japanese-specific normative data for femoral neck BMD, and reassess the respective performance characteristics of BMD at the femoral neck and lumbar spine. Methods Models were constructed to compute the 10-year probability of hip fracture and a major osteoporosis-related fracture in Japan. A major osteoporosis-related fracture was defined as a clinical spine, hip, proximal humeral and forearm fracture. Poisson modelling was used to calculate the hazard functions. The relationship between probability and hazard functions were used to calculate the 10-year probability or fracture for a combination of the risk factors. The mortality estimates for Japan were those published by the World Health Organization for 1999, which accord with estimates from Japan [9]. The incidence of hip fractures was taken from previously published sources [1] as was the incidence of fractures at the proximal humerus and distal forearm [10]. Since the incidence of a clinical vertebral fracture was not known in Japan, we assumed that the ratio of clinical vertebral fracture incidence to that of a vertebral fracture diagnosed by radiographic surveys [11] would be the same in the Japan as it was for Sweden [12]. The relationship of clinical risk factors to fracture outcomes was assumed to be the same as that determined in a large meta-analysis of risk factors of 190,000 patient years from nine prospectively studied population-based cohorts from Europe, Australia, North America and Asia [3]. The relationship has been validated in a further 11 cohorts of population-based samples with 1.2 million patient years of observation from the same regions [4]. The independent contribution of each risk factor was used to compute probabilities of fracture in the absence of clinical risk factors or in the presence of any combination [13, 14]. In Japan, the criteria for the diagnosis of osteoporosis prepared by the Japanese Society for Bone and Mineral Research [7] are based on BMD measurements expressed as a percentages of the young adult mean (YAM) for women. In patients with no prior fragility fracture a diagnosis of osteoporosis is made where the BMD is less than 70% of YAM. In patients with a previous fracture, osteoporosis is diagnosed where the BMD is less than 80% of YAM. These diagnostic thresholds, derived by maximising sensitivity and specificity for fracture detection, are also used as intervention thresholds. In order to compare intervention thresholds using YAM with probabilities derived from the FRAX™ algorithm, T-score equivalents were used. The T-score equivalent to 70% and 80% of YAM for Japanese people is −2.7 SD and −1.8 SD, respectively, using the NHANES III reference for BMD at the femoral neck in Caucasian women aged 20–29 years [15]. The relative performance characteristics of BMD at the lumbar spine and femoral neck were examined in a population-based prospective study in Hiroshima. The Hiroshima cohort comprised 2,596 men and women (69% female, 9,803 person years, mean age 65.1 years). Details of the cohort have been previously published [11]. In brief, the participants received measurement of lumbar spine and femoral neck BMD using dual X-ray absorptiometry (DXA, QDR-2000, Hologic) during the period from 1994 to 1995 and were followed for a mean period of 4 years. Information about hip fracture, fracture of the distal radius, proximal humeral fracture and clinical spinal fracture was collected at interview by trained nurses and physicians during the biennial health examinations. One hundred eighty-six fractures were detected during the follow-up period, of which 89 were categorized as osteoporotic fractures and 31 were hip fractures. The gradient of fracture risk (increase in fracture risk per SD change in Z-score for BMD and increase in fracture risk per 0.1 g/cm2 change in BMD) at the two sites was determined by the use of Poisson models [16]. The fracture outcomes were calculated for hip fracture, a major osteoporosis-related fracture (femoral neck, distal radius, proximal humerus and clinical spine fracture) and all fractures. The mean height and body weight for Japanese women in the Hiroshima cohort was 150 cm and 52.6 kg, respectively, giving a BMI of 23.4 kg/m2. The calculation of fracture probability was made at this BMI, but differences in BMI have little effect on predictive value for fracture risk assessment in the presence of BMD [17]. Japanese-specific T- and Z-scores and those derived from NHANES III were compared using the database of the Japanese Society for Bone and Mineral Research [7]. Data using both methods of calculation were entered into the FRAX™ tool. Results Ten-year probability of fracture The 10-year probability of a major osteoporotic fracture for individuals without clinical risk factors is shown in Fig. 1 grouped by sex, age and T-score. The 10-year probability increased with age and with decreasing T-score. At younger ages, the fracture probability was similar in men and women. With advancing age, probabilities rose as expected, but the increase was greater in women than in men. In men aged 80 years, the 10-year probability for osteoporosis-related fracture exceeded 10% at a T-score of −3 SD, whereas in women of the same age, fracture probabilities exceeded 10% with T-score of −1 SD. Fig. 1Ten-year probability (%) of osteoporotic fracture (hip, clinical spine, humerus, forearm) in Japanese men and women without clinical risk factors according to age and T-score for BMD at the femoral neck The contribution of clinical risk factors to fracture probability is shown in Fig. 2 for women aged 65 years with a BMI of 23.4 kg/m2. In women without clinical risk factors, the 10-year probability for an osteoporosis-related fracture was 7.5%. The 10-year probability was higher in the presence of clinical risk factors. Smoking and alcohol were relatively weak risk factors, the use of long-term glucocorticoids of intermediate weight, and a parental history of hip fracture or a prior fragility fracture were associated with the highest risks. For example, the 10-year probability was 8.1% for smokers and 14.5% for individuals with a prior fracture. The 10-year probability for hip fracture was 1.1% in women without a clinical risk factor, 1.6% in smokers and 2.7% in women with a previous fracture (see Fig. 2). Fig. 2Ten-year probability for osteoporotic (hip, clinical spine, humerus, forearm) and hip fracture (%) according to the presence of a clinical risk factor, in women at the age of 65 years and with a BMI of 23.4 kg/m2 Fracture probabilities were computed in women at the diagnostic threshold recommended in Japan. Thus, the cut-off level of BMD was set at 70% of YAM in women without a previous fracture and at 80% of YAM in those with a previous fragility fracture. In women aged 50, 60, 70 and 80 years without clinical risk factors and with BMD equivalent to 70% of YAM, the 10-year probability was 5.4%, 8.7%, 13.8% and 23%, respectively. In women having BMD equivalent to 80% of YAM and existing fracture but no other clinical risk factors, the 10-year probability was 7.1%, 10.5%, 14.7% and 23.4% at the same ages, respectively. Thus, at each age, the fracture probability was similar using the two diagnostic criteria. In contrast, the fracture probability equivalent to the diagnostic threshold in Japan rose with age, and at the age of 80 years was about four times higher than that at age 50 years (Fig. 3). Similar findings were apparent for hip fracture probability in that probabilities equivalent to the diagnostic threshold in Japan rose with age. The increase with age was more marked than for all major fractures and at the age of 80 years was about 6–40 times higher than that at age 50 years depending on the threshold used (see Fig. 3). Fig. 3Ten-year probability of osteoporotic (hip, clinical spine, humerus, forearm) and hip fracture based on women at the threshold for the diagnosis of osteoporosis using the criteria of the Japanese Bone Mineral Metabolism Association Comparison of lumbar spine and femoral neck BMD The gradient of fracture risk for spine BMD and femoral neck BMD in the Hiroshima cohort indicated that lumbar spine measurements predicted all fractures, osteoporosis-related fracture and hip fracture with approximately equal gradients of risk that ranged from 1.25/SD for all fractures to 1.17/SD for hip fractures. There was no difference in the gradient of risk between men and women. In the case of hip fracture risk, the gradient of risk in men and women combined was not statistically significant with BMD measured at the lumbar spine. BMD at femoral neck had a similar or slightly higher gradient of risk for fractures compared with spine BMD, particularly in the case of hip fracture (Table 1). There was no significant difference in gradient of risk between lumbar spine BMD and femoral neck BMD with the exception of that for hip fracture where the gradient of risk was significantly higher for measurements made at the femoral neck. When gradient of risk was standardized to a constant denominator (i.e., RR/0.1 g/cm2) the findings remained unchanged (see Table 1). Table 1Predictive ability of spine and femoral neck BMD for any, osteoporotic and hip fracture in men and women from Japan RR/SDRR/0.1 g/cm2RR95% CIRR95% CIa. Any fractureFemoral neck1.451.23–1.701.431.22–1.68Lumbar spine1.251.13–1.391.381.19–1.61b. Osteoporosis-related fracturesFemoral neck1.401.09–1.781.381.09–1.74Lumbar spine1.201.04–1.401.301.05–1.61c. Hip fractureFemoral neck2.08b1.34–3.222.111.38–3.23Lumbar spine1.170.91–1.501.250.87–1.80a Hip, clinical spine, forearm and proximal humerusb Significantly higher than lumbar spine ( = 0.049) Japanese reference values The reference mean in women aged 20–29 years at the femoral neck was 0.858 g/cm2 (SD = 0.120 g/cm2) using the NHANES III data. When young normal values were computed from the Japanese population the mean BMD was 0.786 g/cm2 (SD = 0.107 g/cm2). Thus the threshold for osteoporosis using the NHANES III data was 0.558 g/cm2 and that derived from the Japanese data was 0.519 g/cm2. The thresholds for osteopenia (WHO definition) were 0.738 g/cm2 and 0.679 g/cm2, respectively. Thus there were systematic differences in the T-score derived from the two data sets. A comparison of fracture probabilities computed from the z-scores using the two approaches is shown in Fig. 4 for different combinations of risk factors. The differences in probabilities were relatively modest, but as expected, the use of Japanese reference values overestimated fracture probabilities. Fig. 4Correlation between 10-year fracture probabilities (%) in women without clinical risk factors computed from normative data using NHANES III reference values and Japanese-derived reference values for femoral neck BMD. BMI is set at 23.4 kg/m2 Discussion This paper describes the development of the WHO fracture risk assessment tool calibrated to the epidemiology of Japan. The incidence of hip fracture, distal radius fracture and proximal humeral fracture in Japan is lower than that in North America or Northern Europe [1, 10]. However, the prevalence and incidence of spinal fracture are higher in Japan [11, 18]. A minority of all cases of morphological spinal fracture are assumed to be clinical spine fractures in the Japanese FRAX™ models. The multiplier is age and sex specific. For men the multiplier goes from approximately 33% at age 50 to 48% at age 85. For women the corresponding figures are 19% and 24%. These estimates, derived from the epidemiology of fracture in Sweden [12], have been shown to hold true for Japan [19]. The FRAX™ algorithm is suitable for assessment in men and women from the age of 40 years and calculates the 10-year probability for both hip fracture and a major osteoporosis-related fracture. One of its strengths is that it can capture the independent contribution of several clinical risk factors to fracture risk and can be used with or without information on femoral neck BMD. A more detailed account of the properties of the FRAX™ models is provided elsewhere [13]. In this paper, we focused on its application to decision-making in clinical practice with the estimation of intervention threshold i.e., the fracture probability at which intervention is currently considered to be worthwhile. The approach should be distinguished from intervention thresholds that are based on health economic analysis. The WHO makes no specific recommendation concerning intervention thresholds, since these depend on many local factors [6]. Rather, it is suggested they should be determined by each country, based on the local healthcare situation and cost-effectiveness of the treatment of osteoporosis. Intervention thresholds, based on cost-effectiveness have been formulated in the UK, the USA and in Sweden [20–22]. In Japan, diagnostic thresholds are used as intervention thresholds. When the probabilities of osteoporosis-related fracture were determined at these thresholds, they varied with age (see Fig. 3), ranging from approximately 5% at the age of 50 years to more than 20% at the age of 80 years. Against this background, a 10-year probability of 10% for osteoporosis-related fracture may be an acceptable intervention threshold for Japan, though an optimization should take account of health economic consequences for individuals and for the health care budget. The FRAX™ tools are designed to be extensively used in the world as a means of identifying individuals with elevated risk for fracture and aid in the determination of the threshold for therapeutic intervention, but there will be hurdles to be faced in the ease of its acceptance. Such hurdles are likely to differ from country to country. In Japan, the choice of clinical risk factors is not at issue since the risk factors adopted in the FRAX™ algorithms included the data from the Hiroshima cohort, Japan [11, 23, 24], and the validation included the Japanese Miyama cohort [13]. More problematic is the inclusion of femoral neck BMD, since the lumbar spine measurement is the most widely used in Japan. Asian physiques are smaller than those of Caucasians, and the geometric characteristics of the femoral neck in Japanese differ from those in American women, in that the femoral neck length is shorter than in the Japanese [25, 26]. Because of uncertainty regarding measurement of femoral neck BMD, measurement of lumbar BMD is widely used in Japan. The present study indicates that concerns over the use of femoral neck BMD are unfounded. Femoral neck BMD was superior in its ability to predict hip fractures compared with spine BMD, and BMD at the spine and femoral neck had similar predictive value for fractures other than hip fracture. Fujiwara et al. [11] have also shown that the ability to predict the risk for morphological spinal fracture was similar between femoral neck BMD and lumbar spine BMD. These data on gradients of fracture risk, derived in Japan, did not differ from those reported in Western countries. Indeed, the meta-analysis used to inform the FRAX™ tool showed no evidence for heterogeneity in gradients of risk between cohorts [16]. Thus the evidence suggests that fracture risk assessment is not disadvantaged by the use of femoral neck BMD. Indeed, the converse may be true. A further hurdle, unique to Japan, is that diagnostic thresholds differ from the WHO description of osteoporosis which defines osteoporosis on the basis of a fixed T-score threshold (≤ −2.5 SD) using an international reference standard for young (aged 20–29 years) Caucasian women [15]. In Japan, diagnostic thresholds are also derived by reference to a young population, but differ from the WHO in that a local (i.e., Japanese) standard is used and that the criteria differ in patients with or without previous fracture. In view of the widespread use of data derived from the Japanese population, the question arises whether T-scores or Z-scores derived from Japanese databases could be used in the FRAX™ tool, rather than those derived from the international reference base. In the present study, mean BMD at the femoral neck was lower in Japanese women than in the NHANES III sample from the USA, as previously shown [27]. The difference was not, however, large (approximately a half a SD). There were also differences in the SD which was smaller in Japanese women than in the NHANES III sample (0.107 and 0.120 g/cm2, respectively). Not surprisingly, the DXA-based T-score obtained from Japanese and USA populations differed as did the computed probabilities. Although the differences were small at low T-scores, when applying the FRAX™ model to Japan, it is preferable to program the system so that the Japanese T- and Z-scores are converted into the appropriate T- and Z-scores based on NHANES III. In conclusion, a FRAX™ tool has been developed to compute fracture probabilities calibrated to the epidemiology of Japan. The tool has been used to determine possible thresholds for therapeutic intervention, based on equivalence of risk with current guidelines. The approach will need to be supported by appropriate health economic analyses. The present study indicates that the femoral neck BMD is suitable for prediction of the risk for fracture among Japanese people. However, when applying the FRAX™ model to Japan, T-scores and Z-scores should be converted into those derived from the international reference.	[ "japan", "fracture", "fracture risk assessment tool", "fracture probability", "intervention thresholds", "bone mineral density" ]	[ "P", "P", "P", "P", "P", "P" ]
Matern_Child_Health_J-2-2-1592243	Do We Practice What We Preach? A Review of Actual Clinical Practice with Regards to Preconception Care Guidelines	Objectives: To review what past studies have found with regard to existing clinical practices and approaches to providing preconception care. Methods: A literature review between 1966 and September 2005 was performed using Medline. Key words included preconception care, preconception counseling, preconception surveys, practice patterns, pregnancy outcomes, prepregnancy planning, and prepregnancy surveys. Results: There are no current national recommendations that fully address preconception care; as a result, there is wide variability in what is provided clinically under the rubric of preconception care. Conclusions: In 2005, the Centers for Disease Control and Prevention sponsored a national summit regarding preconception care and efforts are underway to develop a uniform set of national recommendations and guidelines for preconception care. Understanding how preconception care is presently incorporated and manifested in current medical practices should help in the development of these national guidelines. Knowing where, how, and why some specific preconception recommendations have been successfully adopted and translated into clinical practice, as well as barriers to implementation of other recommendations or guidelines, is vitally important in developing an overarching set of national guidelines. Ultimately, the success of these recommendations rests on their ability to influence and shape women's health policy. There is a strong body of evidence to demonstrate that preconception care can modify behavioral, medical, and other health risk factors known to impact pregnancy outcomes. For example, preconception care efforts and interventions have been shown to improve folate status among women planning to conceive and to reduce the risk of fetal alcohol syndrome [1, 2]. Controlling known medical problems prior to pregnancy not only helps to optimize the mother's health, it may also improve neonatal and pediatric outcomes [3]. Preconception care efforts and interventions have been shown to lower the risk for both major and minor birth defects among the children of women with pre-existing diabetes [4]. Women with epilepsy and their offspring benefit from preconception care counseling that manages medications, optimizes seizure control and prescribes folic acid for neural tube defect (NTD) prevention [5, 6]. While there is a burgeoning knowledge base on how to improve pregnancy outcomes, there are few studies examining the effects of this knowledge base on actual practice patterns. It is not well known to what extent health care providers are translating preconception care knowledge into practice for all women capable of becoming pregnant (whether planning pregnancy or not). Material and methods A literature review between 1966 and September 2005 was performed using Medline to examine the impact of preconception care trials and recommendations on the clinical practice patterns of health care providers. Key words included preconception care, preconception counseling, preconception surveys, practice patterns, pregnancy outcomes, prepregnancy planning, and prepregnancy surveys. Relevant review articles regarding preconception care were also included. All studies had to be done in the United States. Studies that examined existing clinical practices and approaches directly to providing preconception care were eligible for inclusion. Studies that used indirect methods of measurement, e.g., interviews with women as a means of assessing preconception care practices of providers, were not included. A total of 11 studies were located and included analyses of practice patterns of various health care providers and specialties (see Appendix). Results Studies researching health care provider awareness, knowledge and practices regarding preconception care in the United States are infrequently performed, so it is difficult to fully assess health care provider approaches and practices related to preconception care. A small pilot study involving family practice nurse practitioners was done in 1987 to determine what preconception care issues were discussed with women of childbearing age during a well woman visit. Audiotapes of the visits were compared to an investigator-developed model for preconception counseling. Based on the results, the authors concluded that current obstetrical care had not expanded to include preconception counseling [7]. A study among family physicians conducted in 1991 using the Comprehensive Prevention Knowledge and Applications Survey Instrument found that only 37% of providers reported counseling women of childbearing age about preconception behaviors >75% of the time [8]. When asked how often they felt ready to counsel women on preconceptional factors, about 70% of providers said they were prepared at least three-fourths of the time. However, perceived preparedness did not equate to counseling success. Among providers who did counsel, only about one-third reported being successful at helping their patients change preconceptional behaviors at least 75% of the time. A 1991 study conducted among family practice and internal medicine residents in an inner-city public hospital found that both groups of residents reported low levels of knowledge and management decision skills regarding preconception care when compared with standard recommendations, despite their favorable attitudes towards preconception care [9]. In this survey, close to 50% of the residents did not mention family planning during counseling sessions with women of reproductive age and 74% would not raise the issue of congenital defects in their diabetic patients seeking to become pregnant. While both family practice and internal medicine residents had favorable attitudes towards preconception care, the family practice residents scored higher in positive attitude about preconception care. There was no difference between the groups in terms of management skills, and there was no sustained difference between groups in the area of knowledge. Previous rotation through a high-risk perinatal clinic increased scores in areas of management and knowledge compared to family practice or internal medicine residents who had not rotated through such a clinic, but these differences were not significant [9]. In an effort to assess the practices of genetic counseling and screening for consanguineous couples, their pregnancies, and their children, 1,582 surveys were mailed to board certified genetic counselors and medical geneticists in the United States in 1999 [10]. While the response rate was very poor (only 20%) there was wide variability in suggested screening practices for consanguineous couples before conception, during pregnancy, after birth, and for children placed for adoption. Respondents seemed to generally agree to do screening based on ethnicity but there was no consistency regarding which genetic disorders to include in these screening efforts. A mailed survey of obstetrician-gynecologists in 2000 focused on issues around folate for the prevention of NTDs [11]. Almost 97% of participants knew that when taken early enough, folic acid reduced the incidence of NTDs. While two-thirds of respondents said they routinely screened their pregnant patients for folate intake, only 53% screened nonpregnant women of childbearing age. The authors of the study concluded that most obstetrician/gynecologists are aware of the link between folate intake and NTDs although the data showed that physicians who routinely screened for folate intake correctly answered more survey questions than those who did not. As part of a larger effort to improve the documentation and delivery of preconception care, Bernstein and colleagues evaluated the knowledge base and awareness of preconception care for women in their childbearing years seeking care in an inner city outpatient gynecology clinic in 2000 [12]. Providers in the clinic included physicians and nurse practitioners. A pre-intervention chart review found that many of the providers were not addressing family planning services, domestic violence, nutrition and medical risk factors, medication use, appropriate counseling and use of referral services during gynecologic visits. The authors concluded that the providers did not take the opportunity to discuss preconception care during routine gynecologic visits. A series of studies looking at provider knowledge and practice regarding preconception and prenatal screening for cystic fibrosis (CF) was conducted during 2001–2004 using the Collaborative Ambulatory Research Network (CARN) [13]. Results showed that almost one-half of the responders did not ask nonpregnant women of childbearing age about their family history of CF, nor did they provide information on CF screening. Close to 90% of respondents did offer CF screening or inquire about a family history of CF among pregnant women. Of those Ob/Gyns who did selectively screen for CF among their pregnant patients, about 25% utilized all of the selection criteria noted in the American College of Obstetricians and Gynecologists (ACOG) recommendations. In 2002 and 2003 the March of Dimes examined folic acid knowledge and practice patterns of health care providers in obstetric/gynecology and family/general practice settings [14]. Survey results found that, while knowledge about the need for supplementation and timing of folic acid was high, increased knowledge about unintended pregnancy rates and correct doses of folic acid for prevention of occurrence and recurrence of NTDs is needed. A little more than half of all providers knew the correct dose of folic acid (400 mcg daily) for the prevention of NTDs. This percentage dropped dramatically when providers were asked the correct dose of folic acid for a woman with a history of a previous NTD-affected pregnancy. Almost two-thirds of providers did not know the folic acid dose for recurrence prevention (4 mg daily). In the surveys, the vast majority of providers reported always recommending folic acid supplementation for women who expressed interest in becoming pregnant, approximately 60% of respondents who provide prenatal care reported seeing less than one fifth of their prenatal patients for a preconception care visit [14]. At annual gynecologic or well-woman examinations as well as other types of patient visits, less than 60% of respondents always or usually addressed supplementation. Lack of knowledge (39%) and lack of time (30%) during a busy exam schedule were the most cited reasons providers gave for not always recommending folic acid. In 2004 a survey was sent to Florida health care providers to establish their baseline knowledge and practice behavior regarding folic acid and NTDs [15]. At baseline, 97% of providers reported awareness of the protective nature of folic acid during preconception and early in the first trimester; however, less than half of providers reported that they discuss folic acid consumption with all women of child-bearing age. After a statewide educational program, the same providers were surveyed again to see if there had been any change in knowledge base or practice patterns. The survey done after the educational outreach efforts did demonstrate an increase in awareness and an increase in the percentage of providers who recommended peri-conceptional use of folic acid to prevent NTDs. The authors note, however, that despite these modest gains, the need for continued education and modification of provider practice patterns remains. A recent survey by ACOG was done to assess the opinions and practices of obstetrician-gynecologists regarding preconception care, and their perception of patients’ receptiveness to preconception services [16]. The vast majority of physicians (97.3%) stated that they provide preconception care for their patients, although obstetrician-gynecologists were more likely to do so than providers who only practiced gynecology. The majority of physicians (87%) agreed with the definition of preconception care as “Specialized pre-pregnancy care that focuses on issues not typically addressed during a routine exam which are specific to ensuring an optimal pregnancy outcome.” Most of the respondents indicated that women are more likely to present for preconception care to ensure a healthy pregnancy than because of an elevated risk for a birth defect or developmental disorder. The majority of physicians agreed that preconception care is an important issue and that it has a positive effect on pregnancy outcomes, though only 21% agreed that it was a high priority in their workload. Half of the physicians said there was not enough time to provide preconception care to all women with reproductive potential, nor were there reimbursement incentives to do so. Discussion Preconception care- care that is initiated before pregnancy- is advocated to help women reduce their risk for adverse pregnancy outcomes and make informed decisions regarding their readiness for and timing of pregnancy [17]. Appropriate preconception care improves pregnancy outcomes and has several components. These include, but are not limited to: the systematic identification of preconceptional risks through an assessment of the woman's reproductive, family, and medical history; the family and medical history of the father; the woman's nutritional status, social concerns, and any drug or substance exposures she or the father may have (had); discussions regarding possible effects of any existing medical problems and potential interventions; screening for infectious diseases with treatment and immunization where indicated; discussions about environmental exposures, both occupational and household; a review of the circumstances of the woman's life and behavioral patterns; counseling and discussion about birth spacing including real and perceived barriers to achieving her desires; and inquiry and education regarding contraceptive use. Traditionally, preconception care endeavors have focused on women who have a chronic medical condition, a history of a poor outcome in a previous pregnancy, or who are planning to become pregnant in the near future. For women with chronic medical conditions, the effect(s) of their illness(es) range from minimal to limited activity to hospitalization. Drawing on information for the 1996 National Health Interview Survey (NHIS), 9.9% of women ages 18–44 years had some limitations placed on their activity levels due to chronic disease [18]. There is clear evidence that the initiation of preconception counseling and care for women with some chronic health conditions, e.g., diabetes, will positively impact pregnancy outcomes [19]. For women with a history of a poor pregnancy outcome in the past, pre- or interconceptional intervention strategies have been demonstrated to be effective [20]. Limiting preconception care endeavors to women actively planning pregnancy, however, reaches a limited audience as almost 50% of pregnancies in the U.S. are unintended [21]. Comprehensive preconception care, however, encompasses much more than just these risk categories and is inclusive of all women with the potential to become pregnant. Women of childbearing age visit their physician an average of about three times per year; these visits represent opportunities to deliver preconception care and messages related to preconception care [22]. In reviewing the literature related to the clinical practice of preconception care, it is clear that most of these opportunities are either missed or foregone. In an article focusing on preconceptional health promotion, Moos proposes 5 categorical reasons for why providers are inconsistent in applying proven prevention strategies for poor pregnancy outcomes: lack of knowledge regarding the incidence of unintended pregnancy, inadequate provider education, lack of confidence that preconceptional health counseling is valuable, a belief that women will ‘know’ to seek the care appropriate to their needs, and concerns over lack of reimbursement coverage for preconceptional visits [23]. Changing behavior patterns is no small feat, but understanding the ‘hows’ and ‘whys’ of change(s) in provider practice patterns is becoming better understood. For example, in diffusion theory, diffusion is the process through which an innovation is adopted for use or application by a community [24]. One proposed method for facilitating diffusion of innovation utilizes an expert opinion leader process whereby a designated (or identified) local physician leader communicates information about new innovation(s) or knowledge to colleagues and then follows up with them about their personal experiences with the information [25]. This approach has demonstrated effectiveness for eliciting changes in practice patterns in certain situations and settings, but it is not uniformly effective in all instances [26–29]. In his assessment of how research findings get translated into “best practices”, Green postulates that there are three areas representing barriers to the adoption of best practice: 1) accessibility gap, i.e., practitioners do not have the same resources available as the researchers; 2) credibility gap, e.g., a comparison of differences in practice settings or populations that might ‘explain’ why the research is not applicable to the general practitioner; 3) expectations gap, e.g., the practitioner views the research findings as unnecessary goals to set for their own practices [30]. These barriers are not insurmountable, and there is evidence that different approaches to changing provider behavior patterns are effective [26, 27]. In an article by Cullum, several model programs are reviewed that have demonstrated changes in the practices of health care providers related to preconception care [31]. Studies evaluating effective preconception care interventions and strategies are numerous, and continue to grow. There is also an increasing amount of information about how patients assimilate and act upon recommendations regarding preconception care. There is little data, however, that analyzes the impact and integration of preconception care innovations on daily clinical practice(s). One limitation of this literature review is that articles with secondary goals of evaluating actual clinical practices may have been overlooked. Efforts were made to include all relevant search terms but it is possible that articles written before 1985 may have been inadvertently excluded if different keywords were used at that time. Nonetheless, there is a large need for further research into the contrast between the science of preconception care and the reality of actual clinical practice. Conclusion Preconception care seeks to promote the health of women of childbearing age prior to conception and to improve pregnancy-related outcome. Various authors propose the integration of preconception care into routine wellness care for all women of reproductive potential [23, 32–34]. The question is how do we raise the level of importance and prioritization of preconception care during individual encounters as well as on a national health policy level? Proving the efficacy of a practice does not guarantee an actual change in practice and simply prompting providers to incorporate preconception care will not suffice. Creating the expectation that preconception care is an ongoing process during a woman's reproductive life span will require a myriad of efforts concurrently directed at providers, patients, payors, and policy makers. In 2005, the Centers for Disease Control and Prevention sponsored a national summit regarding preconception care with a broad cross-section of stakeholders. Efforts are currently underway to develop a uniform set of national recommendations and guidelines for preconception care. Understanding how preconception care is presently incorporated and manifested in current medical practices should help in the development of these national guidelines. Knowing where, how, and why some specific preconception recommendations have been successfully adopted and translated into clinical practice, as well as barriers to implementation of other recommendations or guidelines, is vitally important in developing an overarching set of national guidelines. Ultimately, the success of these recommendations rests on their ability to influence and shape women's health policy.	[ "preconception care", "preconception counseling", "preconception surveys", "practice patterns", "pregnancy outcomes", "prepregnancy planning", "prepregnancy surveys" ]	[ "P", "P", "P", "P", "P", "P", "P" ]
Eur_Radiol-3-1-2039780	MRI of the kidney—state of the art	Ultrasound and computed tomography (CT) are modalities of first choice in renal imaging. Until now, magnetic resonance imaging (MRI) has mainly been used as a problem-solving technique. MRI has the advantage of superior soft-tissue contrast, which provides a powerful tool in the detection and characterization of renal lesions. The MRI features of common and less common renal lesions are discussed as well as the evaluation of the spread of malignant lesions and preoperative assessment. MR urography technique and applications are discussed as well as the role of MRI in the evaluation of potential kidney donors. Furthermore the advances in functional MRI of the kidney are highlighted. Introduction The role of computed tomography (CT) and magnetic resonance imaging (MRI) in the evaluation of renal abnormalities is ever increasing. Although multidetector helical CT has taken the largest leap, MRI can be used in case of compromised renal function, severe contrast allergy, or in case radiation exposure is a problem, such as in children and pregnant women. Furthermore, MRI can be used as a problem-solving modality when the CT findings are nondiagnostic. Attempts are being made to use MRI for imaging of renal function, including perfusion [1, 2], glomerular filtration rate [1, 3, 4] and intrarenal oxygen measurement [1, 5]. MRI technique In MRI of the kidneys, fast imaging techniques are essential because of respiratory motion of the kidneys [6]. When possible the scan should be performed within one breath-hold. The patient should get clear instructions on breath-hold technique. If the patient has difficulty with breath-holding, a short period of hyperventilation before breath-holding may be helpful. The scan should be performed during expiration because the kidney position is more constant in expiration than in inspiration. If the sequence is too long to perform in one breath-hold, respiratory triggering can be used [6]. Another technique of respiratory motion control is respiratory gating by use of a navigator pulse. In this technique the movement of the diaphragm is monitored by a very fast 1D MRI sequence. If breath-holding is not possible, signal averaging can be used, but the quality of the images will be limited. The use of a phased array body coil is preferable because of the improved signal-to-noise ratio. To prevent aliasing in coronal imaging, the patient’s arms should be raised above the head, or the arms may be supported by cushions, anterior to the coronal plane through the kidneys. The imaging protocol for evaluation of the kidney at our institution on a 1.5-Tesla (T) MRI system consists of the following sequences: Coronal T2-weighted half Fourier single-shot turbo spin echo sequence (HASTE) (TR infinite, TE 120 ms, flip angle 90°, breath-hold), serving as a localizer, but also supplying valuable T2-weighted information. The limitation of this sequence is a relatively low signal-to-noise ratio.Axial T2-weighted turbo spin echo sequence with fat suppression (TR 2,000 ms, TE 100 ms, flip angle 90°, respiratory triggering). This sequence provides for more detailed T2-weighted information. The T2-weighted sequence is especially helpful in characterizing cysts and intraparenchymal abscesses and in evaluating hydronephrosis. Furthermore, the T2-weighted sequence is helpful in detecting solid lesions.Axial T1-weighted gradient echo sequence, in-phase and opposed-phase (TR 180 ms, TE 2.3 ms/4.6 ms, flip angle 90°, breath-hold), preferably as a dual-echo sequence. Many solid renal lesions are hypointense compared to the renal parenchyma on T1-weighted images, but lesions with hemorrhage, lesions with macroscopic fat, melanin-containing lesions and cysts with high protein content may show hyperintense signal [6]. Opposed-phase T1-weighted gradient echo sequences can be used to prove the presence of small amounts of fat.Axial T1-weighted gradient echo sequence for dynamic imaging (TR 130 ms, TE 1.0 ms, flip angle 90°), using 30 ml intravenous gadolinium contrast, immediately followed by three breath-hold periods with four scan series per breath-hold. In this way pre-contrast and post-contrast images in arterial and nephrographic phase are obtained. Gadolinium-enhanced images are used for lesion detection and characterization.Coronal 3D fast gradient echo with fat suppression, obtained immediately after the dynamic series for delayed contrast-enhanced images (TR 3 ms, TE 2 ms, flip angle 15°). This sequence can be used for renal venous anatomy, for the analysis of (tumor) thrombus and for evaluation of extent of the tumor in the perinephric fat. Currently 1- to 1.5-T systems are generally used for abdominal imaging, but the advent of 3-T MRI systems brings a twofold increase in the signal-to-noise ratio (SNR). The increase in SNR can be spent on higher resolution or on even faster imaging. When combined with parallel imaging techniques such as sensitivity encoding (SENSE), the speed of any sequence can be increased by up to a factor of four or higher. However, although 3-T MRI is promising, only a limited amount of research has been published on 3-T MR imaging for renal lesions, and its value has still to be established [7, 8]. Renal lesions The main goal in the evaluation of renal lesions is to differentiate surgical lesions from nonsurgical lesions. Most simple cysts are easily recognized and don’t need further analysis. Complicated or multiloculated cysts need more attention in order to differentiate them from cystic carcinomas. In most solid renal lesions, neither CT nor MRI is able to reliably distinguish benign from malignant. Some solid lesions, however, may be identified as benign with high confidence, like angiomyolipomas. In general, if a lesion cannot be characterized as benign or malignant, it should be considered malignant [9]. Malignant renal lesions Renal cell carcinomas Renal cell carcinomas account for 3% of all malignancies in adults. Almost 50% are detected incidentally. As many as 85% of suspicious renal lesions are malignant [10]. Features indicating potential malignancy of a renal lesion are size of the lesion, the presence of calcifications, the distribution of the calcifications within the lesion, wall thickness and the presence of septa in case of a cystic lesion (Fig. 1), inhomogeneity of the lesion, extension of the tumor beyond Gerota’s fascia, and last but not least, enhancement after contrast administration. Concerning calcifications in cystic lesions, recent research suggests that the importance of these calcifications as a determinant of malignancy is relatively low [11]. Fig. 1Gradient echo images with intravenous gadolinium at baseline (a), 8 months later (b) and 15 months later (c). The complicated cortical cyst (arrow) in the left kidney on image (a) in a patient with Von Hippel Lindau disease progresses into a frank renal cell carcinoma with multiple enhancing internal septations (c). Several simple cysts are visible. Images courtesy of Roy S. Dwarkasing Duchene found in 186 renal tumors that all tumors greater than 7 cm (n = 48) were malignant. About 80% of tumors smaller than 3 cm were malignant [10]. The differential diagnosis of solid renal lesions smaller than 7 cm consists of oncocytoma, angiomyolipoma, hemangioma, leiomyoma, and focal xanthogranulomatous pyelonephritis. Of these lesions, only cysts and angiomyolipomas can often be positively identified as benign lesions. Generally, MRI is performed only after a renal lesion has been detected by ultrasound or CT. CT may be followed by MRI if the enhancement at CT imaging is indeterminate (10–20 Hounsfield Units) [12] or in case of suspected pseudo-enhancement. At CT examination, simple cysts may show pseudo-enhancement after intravenous contrast administration, which is an increase in attenuation of more than 10 Hounsfield Units, not caused by administered contrast or by partial volume effect but by technical factors [13]. Assessment of enhancement The main MRI feature indicating potential malignancy of a renal tumor is enhancement after intravenous gadolinium administration, differentiating the lesion from a cyst. However, enhancement at MRI cannot be measured as easily as enhancement at CT. The MR signal is not calibrated, in contrast to density at CT imaging. The MRI signal depends not only on tissue characteristics, but also on the size of the patient, the gain setting of the MR system, the pulse sequence and the coils. The presence of enhancement can be assessed subjectively, by subtraction imaging, and by quantitative assessment. Subjective assessment of enhancement has been shown to be accurate in detecting renal cell carcinomas [14]. However, in a cystic lesion with only a small solid component, subjective assessment may be difficult and subtraction images may be used to better assess the presence of enhancement. It is important to realize that renal cell carcinomas may be hypovascular and therefore may show less enhancement than the surrounding renal parenchyma [15]. Also in these hypovascular lesions and in lesions that are hyperintense on T1-weighted imaging, subjective assessment may be difficult, and subtraction images may be of particular use [16]. Some investigators have studied quantitative assessment of enhancement by calculating the relative enhancement defined as the signal intensity increase after contrast administration compared to the signal intensity before contrast administration. Ho used relative signal intensity enhancement to differentiate cysts from malignant lesions. Using a threshold of 15% relative signal intensity enhancement after administration of intravenous gadolinium, Ho found in 74 patients with renal lesions a sensitivity of 100% and a specificity of 94% in the detection of renal cell carcinomas [17]. The relative enhancement peak was maximum between 2 and 4 min after injection of gadolinium contrast. It is worthwhile to note that cysts also showed a mean enhancement change of up to 5%. This pseudo-enhancement may be attributed to motion artifacts and volume averaging [14]. Subtypes of renal cell carcinoma Several histological subtypes of renal cell carcinoma are recognized. The most frequent subtype is the conventional or clear cell carcinoma, which comprises 88% of all renal cell carcinomas, followed by papillary carcinoma (10%) and chromophobe carcinoma (2%). Collecting-duct carcinoma is very rare. The first three subtypes do not differ significantly in prognosis and show a 5-year survival rate of 73–88% [18]. Several attempts have been made to distinguish the subtypes of renal cell carcinoma by imaging features using CT characteristics [19, 20]. There are only a few studies using MRI to differentiate the subtypes of renal cell carcinoma. Outwater et al. described that clear cell carcinomas may show loss of signal intensity on opposed-phase images compared to in-phase images, due to intracellular lipid [21]. This effect is caused by the presence of fat and water protons in the same voxel, resulting in cancellation of the signal on the opposed-phased sequence. The intracellular lipid contributes to the histological appearance of the clear cell. Because oncocytoma and transitional cell carcinoma do not contain diffuse lipid, the loss of signal on opposed-phase images may allow differentiation between clear cell carcinoma on the one hand and oncocytoma and transitional cell carcinoma on the other hand. However, this distinction is of little clinical use. If the tumor does not show signal loss, it can still be a clear cell carcinoma [21]. Fat-containing renal cell carcinomas Although rare, it has been claimed that renal cell carcinomas may contain macroscopic fat. Since intralesional fat has long been considered diagnostic for angiomyolipoma, these rare fat-containing renal cell carcinomas may easily be confused with angiomyolipoma. A few cases of fat-containing renal lesions, suggestive of angiomyolipoma, that appeared to be renal cell carcinoma have been reported [22–27]. With a few exceptions [25, 27], most of the reported cases also contained intratumoral calcifications [22–26], which is very rare in angiomyolipoma. Therefore, the presence of calcification in a lesion with macroscopic fat should be a warning that the lesion may very well be a carcinoma [28]. Because the accuracy of MRI in detecting calcifications is relatively low, it is important always to assess earlier ultrasound or CT examinations of the patient, which will likely have been made in most cases. Less than 5% of renal cell carcinomas are cystic [9] (Fig. 2). On CT imaging, cystic renal lesions are classified using the Bosniak classification system [29]. This system was developed for CT, but recently its use has been evaluated in MRI by Israel et al. [12]. The CT and MRI findings in this study were similar in 81% of 69 renal masses. In the remaining 19% of cases, MRI showed more septa, increased wall thickness or increased enhancement compared to CT. This resulted in an upgrade of the Bosniak classification in seven cases (10%), of which two cases were upgraded to category 3 and two cases to category 4. The latter two cases appeared to be malignant at surgery. One of the cases upgraded to category 3 appeared to be benign, and the second one showed progression (patient refused surgery). The authors state that the Bosniak grading system is appropriate for use in MRI, but acknowledge that further research is necessary [12]. Fig. 2T2-weighted turbo spin echo sequence (a) showing a cystic multiloculated lesion in the left kidney (arrow) with high signal intensity content. The pre- and post-gadolinium images (b and c) show enhancement of the septations. Pathologic examination showed a cystic renal cell carcinoma Staging There are two staging systems for renal cell carcinoma, both based on the degree of tumor spread beyond the kidney. The Robson staging system [30] (Table 1) is still in use, but it is being replaced by the TNM staging system, developed by the American Joint Committee on Cancer (AJCC) [31] (Table 2). The TNM staging system is similar to the Robson system, but provides a more detailed description. Staging is usually performed using CT. Hallscheidt compared the performance of CT and MRI in the TNM system and found a similar accuracy in the staging of renal cell carcinoma [32]. Table 1Robson staging system for renal cell carcinoma [30]StageTumor extentITumor confined to the kidneyIITumor extension through the capsule of the kidney in the perirenal fat including the adrenal gland, no involvement of Gerota’s fasciaIIIaTumor extension into the renal vein or inferior vena cavaIIIbInvolvement of regional lymph node(s)IIIcInvolvement of regional lymph node(s) and extension into the renal vein or inferior vena cavaIVaTumor extension beyond Gerota’s fascia into adjacent organsIVbDistant metastasisTable 2AJCC TNM staging system for renal cell carcinoma (2002, sixth edition) [31]StageDescriptionTxNo information on primary tumor availableT0No evidence of primary tumorT1aTumor size 4 cm or less, limited to the kidneyT1bTumor size more than 4 cm but no more than 7 cm, limited to the kidneyT2Tumor size more than 7 cm, limited to the kidneyT3aTumor extension into the perinephric fat and/or renal sinus fat or the adrenal gland, but not beyond Gerota’s fasciaT3bTumor grossly extends into the renal vein or its segmental (muscle-containing) branches, or inferior vena cava below the diaphragmT3cTumor grossly extends into the inferior vena cava above the diaphragm or invasion of the IVC wallT4Tumor extension beyond Gerota’s fasciaNxNo information on regional lymph nodes availableN0No regional lymph node metastasisN1Metastasis in a single regional lymph nodeN2Metastasis in more than one regional lymph nodeMxNo information on distant metastases availableM0No distant metastasesM1Distant metastases Determination of the extent of the tumor into the renal vein and the inferior vena cava (IVC) (TNM stage T3b and T3c) is important for the surgical approach. Involvement of the IVC is reported to occur in 4–10% of renal cell carcinoma patients [33]. If the tumor extends into the large veins, the upper level of the tumor thrombus dictates the surgical approach. If the extension is limited to the renal vein, it is necessary to know the distance to the IVC, in order not to dissect the renal vein through the intraluminal tumor. Extension into the IVC requires cavotomy with clamping of the IVC and the contralateral renal vein (Fig. 3). If a tumor thrombus extends above the hepatic veins, the liver should be mobilized to control the IVC above the hepatic veins. Extension into the right atrium generally requires cardiopulmonary bypass during surgery with hypothermic cardiac arrest [33]. Fig. 3T1-weighted gradient echo sequence after intravenous contrast of a large renal cell carcinoma in the upper pole of the left kidney with tumor thrombus extending into the IVC up to the level of the liver For differentiation between bland thrombus and tumor thrombus, the use of gadolinium contrast is indicated, since the enhancement of the thrombus indicates a tumor, whereas the lack of enhancement indicates a clot [34]. Before the introduction of multidetector CT scanners, MRI was considered to be superior to CT in the assessment of intravenous tumor extension, especially at the level of the intrahepatic IVC. However, multislice helical CT scanners are nowadays able to generate multiplanar reconstructions in any direction at high resolution. Recent studies comparing multidetector CT (MDCT) and MRI showed no difference in accuracy of MDCT and MRI in assessing the extension of thrombus in the IVC [35, 36]. Invasion of the IVC wall (TNM stage T3c) is important to detect because in that case the surgeon has to partially resect and reconstruct the IVC. The accuracy of MRI in IVC wall invasion has not yet been well documented. One study reported a sensitivity of 100% and a specificity of 89% in the detection of IVC wall involvement, however, this study consisted of only 12 patients and no inclusion criteria were reported [33]. In this study it was noted that altered signal in the vessel wall and wall enhancement were nonspecific. The most reliable sign of IVC wall invasion was tumor signal both inside and outside the vessel wall [33]. The detection of lymph node metastasis by CT mainly relies on the size of the lymph nodes. In recent years, several studies have been published on the use of ultra-small superparamagnetic iron oxide particles (USPIO) as a negative contrast agent for the detection of small lymph node metastasis. The USPIO particles consist of an iron-oxide core covered with a low-molecular-weight dextran coating. After intravenous injection, the USPIO particles accumulate in healthy lymph nodes. The USPIOs are ingested by macrophages through phagocytosis and cause a decrease in signal intensity on T2- and T2-weighted images. Gradient echo sequences are the most sensitive for these susceptibility effects. Lymph node metastases displace the macrophages in the lymph node and therefore do not show the loss in signal intensity seen in normal lymph nodes. To our knowledge no studies have been published on the use of USPIOs in renal cell carcinoma. However, a recent meta-analysis by Will et al. [37] describes the pooled results of several studies on USPIOs for several types of metastasis. In these studies, MRI with and without ferumoxtran, a first-generation USPIO, was compared with histology. MRI with ferumoxtran significantly improved the diagnostic precision compared to MRI without ferumoxtran. It was shown to be both sensitive and specific, especially in the detection of abdominal and pelvic metastasis. MRI in nephron-sparing surgery In recent years, interest in nephron-sparing surgery has been growing. At first, partial nephrectomy was mainly performed in case of a solitary kidney or diminished renal function, in order to preserve as much function as possible. Due to improving techniques and to increasing application of modern imaging modalities, the number of small, incidentally detected renal tumors is increasing [38]. This development has encouraged surgeons to use nephron-sparing surgery also for patients with normal renal function. The long-term follow-up data suggest that survival after partial nephrectomy of small renal cell carcinomas is comparable to total nephrectomy [39]. The ideal tumor for partial nephrectomy is smaller than 3 cm, is confined to the parenchyma of the kidney and has a peripheral location [40]. The presence of a pseudocapsule around a renal tumor is a sign of lack of perinephric fat invasion and therefore a favorable sign for partial nephrectomy [38, 41]. A pseudocapsule consists of compressed renal tissue and fibrous tissue. The sensitivity of CT in depicting a pseudocapsule is low (10–26%), whereas MRI shows a moderate to high sensitivity in depicting the pseudocapsule (54–93%) [41, 42]. On MRI, a pseudocapsule presents as a hypointense rim around the tumor on both T1-weighted and T2-weighted images, but can be best seen on the T2-weighted images and sometimes on gadolinium-enhanced GRE images [40]. Transitional cell carcinoma Transitional cell carcinoma of the kidney is usually evaluated by intravenous urography, CT and endoscopy. However, MRI may play a role if CT and endoscopy are not feasible. CT may not be possible in case of poor renal function, and in case of ureteral obstruction, contrast excretion may be too limited to allow tumor detection. The ureter may not be accessible for endoscopy because of fibrosis and stricture of the ureter and ureter ostium. This may especially be the case in patients who have been treated for bladder cancer. These patients are at particularly increased risk for upper urinary tract transitional cell carcinoma [43]. Transitional cell carcinoma of the pyelum or ureter will generally show as an irregular mass projecting in the lumen. In the collecting system, transitional cell carcinoma is usually confined to the lumen (Fig. 4), but infiltrative growth into the renal parenchyma occurs and typically does not distort the renal contour. Chahal et al. applied MR urography (MRU) in 23 patients with high clinical suspicion of upper tract transitional carcinoma and hydronephrosis that could not be explained with other imaging modalities. MRU showed five renal pelvic transitional cell carcinomas and eight ureteral transitional cell carcinomas, confirmed by histology. In the remaining patients, no sign of transitional cell carcinoma was observed during 1-year follow-up [43]. Although these results are promising, the number of publications on MRI for transitional cell carcinoma is limited and more research is required to determine the value of MRI for the detection of transitional cell carcinoma. Fig. 4T1-weighted gradient echo images before (a) and after (b) gadolinium administration. A mass in the renal pelvis (arrow) shows moderate enhancement after gadolinium administration. A transitional cell carcinoma was suspected, which was confirmed after nephrectomy Benign renal lesions Oncocytoma Oncocytomas are benign, most often asymptomatic renal tumors. They represent 2–12% of renal masses. On MRI, oncocytomas show variable low signal intensity on T1-weighted images (Fig. 5) and heterogeneous high signal intensity on T2-weighted images. In 33–54% a central scar with low signal intensity on both T1- and T2-weighted images is visible. After contrast enhancement a spoke-wheel-like pattern may be observed [44]. Both a central scar and spoke-wheel pattern, however, may also be seen in renal cell carcinomas [38, 45] and are therefore not specific for oncocytomas. Oncocytomas may show a pseudocapsule, consisting of compressed renal parenchyma and fibrous tissue. One should be aware that a pseudocapsule is not specific for oncocytoma, since renal cell carcinomas can also be surrounded by a pseudocapsule. Because the characteristics of oncocytomas show considerable overlap with the characteristics of renal cell carcinomas, the therapy for a suspected oncocytoma is usually surgical [38]. Fig. 5Post-contrast fat-suppressed T1-weighted gradient echo images, arterial phase (a) and nephrographic phase (b) of a central oncocytoma (arrow). The tumor shows a hypointense central scar in the arterial phase. In the nephrographic phase, the central scar is slightly hyperintense Angiomyolipoma Angiomyolipomas are benign hamartomatous tumors, consisting of fat, smooth muscle and blood vessels. Angiomyolipomas are the only solid renal tumors that can be positively characterized using MRI [46]. Angiomyolipomas are identified by demonstrating macroscopic fat in the lesion (Fig. 6). The ability to differentiate angiomyolipomas is especially urgent in patients with tuberous sclerosis, since angiomyolipomas develop in about 80% of these patients, and at the same time these patients are at an increased risk of developing renal cell carcinomas. Macroscopic fat in a renal lesion can be detected by CT using density measurement and by MRI using fat-suppression techniques. On the opposed-phase gradient echo images, macroscopic fat is demonstrated by a hypointense rim surrounding the fat (India ink artifact) [46]. It must be noted that the India ink artifact also occurs at the interface between tumors that don’t contain fat and the surrounding perinephric fat, if the tumor extends into the perinephric fat. Fig. 6T1-weighted gradient echo image (a), post-contrast fat-suppressed T1-weighted gradient echo image (b), and T2-weighted HASTE (half Fourier single-shot turbo spin echo) image of the left kidney (c). The hyperintense parts of the tumor in the lower pole (black arrow) on the T1-weighted image show a drop in signal intensity on the post-contrast fat-suppressed T1-weighted image, proving the presence of fat, while the hypointense parts of the tumor enhance after gadolinium. The fatty portions are hyperintense on the T2-weighted sequence, but not as high as the cyst (white arrow) in the midportion of the kidney. The MRI characteristics of the tumor in the lower pole are consistent with angiomyolipoma If the amount of intralesional fat is small, the differentiation between angiomyolipoma and renal cell carcinoma may be difficult (Fig. 7). Kim et al. showed for CT that homogeneous enhancement and a prolonged enhancement pattern were significantly more prevalent in angiomyolipoma with minimal fat than in renal cell carcinoma [47]. Using both CT findings as a criterion for the differentiation of angiomyolipoma with minimal fat from renal cell carcinoma, they found a positive predictive value of 91% and a negative predictive value of 87%. It is likely that this is also the case in contrast-enhanced MRI, however this has not yet been proved. Fig. 7Lesion in the upper pole of the right kidney in a patient with tuberous sclerosis (arrow). The lesion shows low signal intensity on T2-weighted turbo spin echo images (a) and intermediate signal intensity on in-phase T1-weighted gradient echo images (b). No signal loss is observed on the out-of-phase T1-weighted gradient sequence (c). The lesion shows moderate enhancement after intravenous gadolinium administration (d). Pathologic examination after resection of the lesion showed an angiomyolipoma. In this unusual case, no macroscopic fat was detected on MRI It has been shown that clear cell carcinomas may show signal loss on opposed-phase images compared to in-phase images, due to intracellular lipid [21]. This loss of signal intensity should be distinguished from the signal loss at the interface of macroscopic fat and surrounding tissue in angomyolipomas on opposed-phase images, especially if the amount of macroscopic fat is small. If in doubt, the in-phase gradient echo images are often helpful because the fatty portion of angiomyolipomas will be hyperintense, whereas renal clear cell carcinomas are generally hypo- or isointense [6, 46]. Unfortunately, clear cell carcinomas are incidentally hyperintense on T1-weighted images [48]. In these cases spectral fat-suppression images should be used to prove the presence of macroscopic fat in the angiomyolipoma. Attention should be paid to the possibility that carcinomas sometimes contain hemorrhage, causing high signal intensity on in-phase T1-weighted images. In these cases, opposed-phase images and spectral fat suppression will not show a drop in signal intensity. Xanthogranulomatous pyelonephritis Xanthogranulomatous pyelonephritis is a rare chronic pyelonephritis, which may result in severe renal impairment. It is most common in middle-aged women, but it may also occur in children. Often Proteus or E. coli species are involved and the pyelonephritis is often accompanied by calculi, most typically staghorn calculi. It may be accompanied by calyx obstruction and parenchymal abscesses. In xanthogranulomatous pyelonephritis, the affected renal parenchyma is replaced by lipid-laden macrophages, resulting in the typical yellow-gray appearance of the lesion at macroscopy. It may involve the whole kidney, or it may be focal. Especially when it is focal, xanthogranulomatous pyelonephritis may be mistaken for a renal carcinoma. On T1-weighted images, the solid component of the lesion may be isointense or hyperintense, which can be attributed to the fatty component. On T2-weighted images, the signal intensity of the solid component is isointense to slightly hypointense. The parenchymal cavities filled with fluid and pus show high signal intensity on T2-weighted images and low signal intensity in T1-weighted images, varying according to the protein concentration in the cavity [49, 50]. In xanthogranulomatous pyelonephritis, the perirenal fascia may be thickened and show enhancement after gadolinium administration [49]. The absence of hyperintense signal on T2-weighted images of the solid components may be helpful in the differentiation between xanthogranulomatous pyelonephritis and renal tumor [50]. MR urography In MR urography, the pyelocalyceal system and the ureters are visualized using heavily T2-weighted images or T1-weighted images with gadolinium contrast. On the heavily T2-weighted images, the urine in the pyelocalyceal system and ureters is hyperintense because of its long T2 relaxation time, whereas the surrounding tissue is hypointense. The HASTE or single-shot fast spin echo (SSFSE) is very suitable for this purpose; they are very fast with sufficient in-plane resolution [51]. The thin sections can be used for detailed evaluation; a maximum intensity projection is useful for overview of the urinary tract. For T2-weighted MR urography, it is essential that the urinary tract is sufficiently filled with urine. Therefore it is often necessary to use a diuretic if the urinary tract is not dilated [51]. Hagspiel et al. evaluated whether MRI of the urinary system of potential renal donors was feasible without diuretic stimulation or compression and found sufficient visualization of the urinary collecting system in only 14% of the rapid acquisition with relaxation enhancement (RARE) urograms and in 26% of the gadolinium-enhanced 3D fast low-angle shot (FLASH) urograms [52]. In contrast-enhanced MR urography, intravenous gadolinium can be combined with a T1-weighted 3D gradient echo sequence. In case the patient is unable to hold his breath, a fast 3D GRE EPI sequence can be used that offers the additional advantage of reduced ghost artifacts caused by ureteral peristalsis [53]. However, the conventional GRE images provide better high resolution images compared to the EPI images [53]. In contrast-enhanced MR urography, the images are acquired in the excretory phase, typically 5–8 min after intravenous gadolinium injection [51]. Additional use of a diuretic is advisable to increase excretion and to dilute the excreted contrast: if the excreted gadolinium is too concentrated, the T2 effect may cause signal loss [51]. The accuracy of MR urography in assessing renal obstruction is similar to CT urography [54, 55]. MRI has the advantage that it is better able to detect perirenal edema as a secondary sign of obstruction [54]. In a study by Sudah et al., gadolinium-enhanced MR urography showed renal calculi with considerably higher sensitivity than T2-weighted MR urography [55]. On MR urography, a calculus appears as a signal void, which is nonspecific: blood clots, gas, sloughed papilla and tumors may also appear as a low signal within the bright signal of urine [51]. If a signal void is not clearly detached from the wall of the pelvis or ureter, additional T1-weighted and contrast-enhanced images are necessary to further characterize the lesion. MRI of potential donor kidneys Due to the increasing demand for donor kidneys and the relative shortage of cadaver kidneys, the importance of living donors is increasing. Nephrectomy is increasingly performed by laparoscopic surgery to keep the burden for the donor as low as possible. A thorough pre-operative evaluation of the donor kidney is essential to keep the risks as low as possible. Especially for endoscopic nephrectomy, it is important for the surgeon to be informed about the arterial and venous vasculature of the kidney, about the presence of accessory vessels and about abnormal vessel location, like extrahilar branching and retrocaval position of vessels. Moreover, the surgeon needs to be informed about the presence of an abnormal collecting system and the presence of cysts or tumors [56]. In the imaging protocol, special attention should be focused on arterial and venous imaging, in addition to the standard parenchymal imaging. For the arterial MR angiography, a 3D fast GRE with intravenous gadolinium after timing bolus can be used (TR 4.7 ms, TE 1.4 ms, flip angle 30°), with coronal thin-section reconstructions. A relatively large flip angle (up to 40°) can be used to minimize background signal around the high signal of the renal arteries [56]. The venous angiography sequence should follow the arterial angiography immediately. Due to the excretion of gadolinium by the kidneys, the concentration of gadolinium in the renal veins is lower than in the renal arteries, causing lower contrast of the veins compared to the background. A lower flip angle (15°) can be used to compensate for the lower gadolinium concentration, at the expense of more background signal [56]. The 3D dataset can be used for the reconstruction of thin 2D sections for detailed evaluation, as well as for maximum intensity projection (MIP) reconstructions of the vessels (Fig. 8). Fig. 8Pre-operative imaging of potential donor kidneys. a Gradient echo image after intravenous contrast, arterial phase, TR 3 ms, TE 1 ms, slice thickness 2.2 mm, flip angle 27°, coronal. b Gradient echo image after intravenous contrast, nephrographic phase, TR 4 ms, TE 1 ms, slice thickness 2.2 mm, flip angle 15°, coronal. The lower flip angle in the depiction of the renal veins was chosen to compensate for the lower gadolinium concentration in the renal veins Although the time–of-flight technique can be used to evaluate renal arteries, it is not recommended for the detection of accessory renal arteries of small caliber [57]. It may be used to clarify intraluminal filling defects potentially caused by flow artifacts on gadolinium-enhanced MR angiography [6]. Phase contrast imaging can be used for detection and grading of renal artery stenosis [58, 59] but has limitations in evaluating potential kidney donors. The problem of motion artifacts in phase contrast imaging caused by the long acquisition time can be overcome by using interleaved gradient echo-planar technique, shortening the acquisition time from minutes to about 30 s [59]. However, it is difficult to select the proper velocity encoding gradient [6], and tortuous venous anatomy and low flow limit the use of phase contrast MRI in potential kidney donors [56]. Most studies on the accuracy of MR angiography (MRA) in the evaluation of renal vessels compared to digital subtraction angiography and CT angiography show similar or even better results [60–63], although not all studies confirm these findings [64]. In a recent study on 111 MRA examinations for donor nephrectomy, nine accessory arteries were missed, requiring anastomosis in four arteries. Of 14 kidneys with more than one vein, only 4 were identified by MRA, requiring anastomosis in 1 case [65]. MRI for imaging of renal function In recent years, progress has been made in the use of MRI for the evaluation of renal function. Renal disease often causes impairment of renal function. Measurement of renal function can be used as an indicator of severity of disease and can direct therapy. The most simple tests of renal function are serum creatinine and creatinine clearance. However, these tests do not provide information about the function of each individual kidney. This information can be important in case of a living renal kidney donor, prior to nephrectomy or in case of renal artery stenosis. Renal scintigraphy provides information about the function of each kidney, but this test lacks anatomic detail. CT has the disadvantage of ionizing radiation and the use of potentially nephrotoxic contrast, which generally isn’t a problem in healthy kidneys, but may cause deterioration of renal function in diseased kidneys. MRI has the potential to combine the functional and anatomic information about each kidney individually [66]. Perfusion Measurement of renal perfusion may be a tool to assess the significance of renal artery stenosis and to assess ischemic nephropathy, and it may be used in renal transplant assessment. Several techniques have been studied to measure renal perfusion [1]. The maximum slope Gd-DTPA technique uses the maximum slope of the Gd-DTPA enhancement curve in relation to the maximum Gd-DTPA concentration in the aorta to calculate renal blood flow. The Gd-DTPA concentration in the aorta can be calculated using pre-contrast and post-contrast relaxation times measured in the aorta, assuming a linear relationship between gadolinium concentration and the inverse of the pre- and post-contrast T1 difference. Calibration of the signal intensity is performed using phantoms with increasing gadolinium concentration [67]. Furthermore, attempts have been made to assess renal perfusion by arterial spin labeling. To prevent leakage of the contrast medium into the extravascular space, albumin-bound contrast agent is used [68]. However, the different MRI renal perfusion techniques still need to be validated so more research is needed to assess the clinical usefulness. Glomerular filtration rate The glomerular filtration rate (GFR) is a parameter that is used to assess renal function. The serum creatinine level is a rough indicator of the glomerular filtration rate, it is easy and cheap to obtain, but it provides no information about each individual kidney. For the measurement of the GFR by MRI, several techniques have been investigated. The first technique used MR spectroscopy to measure the T1 relaxation times of urine and serum samples taken at intervals after intravenous gadolinium administration. Calculation of the glomerular filtration rate was based on the linear relationship of 1/T1 to the serial dilution measurements of the serum and urine samples [69]. This technique still had the disadvantage of measuring the GFR of both kidneys together. Another MR technique assessing each individual kidney was developed by calculating the extraction fraction (EF) of Gd-DTPA, which is the difference between renal arterial and renal venous Gd-DTPA concentration, normalized to the arterial Gd-DTPA concentration: The gadolinium concentrations can be calculated after measurement of the T1 relaxation time of arterial and venous blood before and after intravenous gadolinium administration, using the formula where R is the relaxivity of gadolinium, and T1pre and T1post are the T1 relaxation times before and after intravenous gadolinium. The glomerular filtration rate can then be calculated according to the formula where RBF is renal blood flow, which can be measured by phase contrast flow quantification [3], and Hct is hematocrit [66]. An alternative MRI technique to measure glomerular filtration rate also uses contrast-enhanced dynamic MRI and is based on a the time-dependent concentrations of Gd-DTPA in the cortex and the medulla—a two compartment model [1]. The different techniques to measure glomerular filtration rate still need to be validated, and their role in clinical practice needs to be established. Diffusion-weighted MRI In diffusion-weighted MR imaging, the image contrast is influenced by the Brownian motion of water molecules. The signal intensity is high if water molecules are restricted in their motion, which can be caused by cell membranes or, in the case of free fluid, by high viscosity. The MR signal intensity is low if water molecules can diffuse freely. Diffusion-weighted imaging has found its place in neuroradiology, especially for the early detection of ischemic brain lesions. Diffusion MRI can also be used for fiber tract mapping in cerebral white matter, by measuring the directional components of the diffusion. Because diffusion-weighted images are inherently T2-weighted, the images are influenced by the T2-shine-through effect. This is the presence of high signal intensity in restricted water, caused by the T2 effect. To cope with this effect, the apparent diffusion coefficient (ADC) may be calculated from two images acquired with different gradient duration and amplitude (b-values) and used for ADC mapping. In renal MR imaging, the role of diffusion MRI is not yet as clear as in neuro-imaging. Fukuda et al. showed that diffusion in the kidney is anisotropic, due to the radial orientation of the tubules in the pyramids and the blood vessels in the renal cortex [70] (Fig. 9). Diffusion-weighted MRI has been applied in patients with solid renal masses as well as in pyelonephritis and renal failure [71, 72]. ADC mapping showed differences between the lesions and normal tissue, but more research is needed to know whether diffusion-weighted MRI can actually help in characterizing different abnormalities. Diffusion-weighted imaging has been applied to differentiate between hydronephrosis and pyonephrosis. Chan et al. found in a limited group of patients that diffusion-weighted imaging showed a hypointense pyelocalyceal system in hydronephrosis and a hyperintense pyelocalyceal system in pyonephrosis [73]. The hyperintensity in pyonephrosis is thought to be due to the high viscosity of the pus, whereas the free-moving molecules in hydronephrosis cause low signal intensity. The results of diffusion-weighted MRI in the kidney are still preliminary, and more research should reveal the value of its clinical application. Fig. 9Diffusion-weighted tensor image of the right kidney on a 3T system. The renal pyramids show lower signal intensity than the surrounding parenchyma because of the radial orientation of the tubules in the pyramids, restricting the Brownian motion of the water molecules to one direction Conclusion The role of MRI in renal imaging is still mainly in differentiating benign lesions versus malignant lesions in patients who cannot undergo CT scanning with intravenous iodinated contrast media, or in cases with nondiagnostic CT results. MRI and CT show comparable accuracy in detection and characterization of most renal lesions. MRI can have additional diagnostic value in the evaluation of lesions with minimal amounts of fat or with intracellular fat. Data suggest that MRI has a higher sensitivity in evaluating complicated cysts, however, the clinical implications still have to be studied. There is evidence to suggest that MRI has a higher accuracy than CT in the evaluation of early lymph node spread. MRI is a suitable tool in the preoperative work-up of potential kidney donors. Functional MRI of the kidney has not yet found broad clinical application, but it has great potential. Through the ongoing development of functional MRI techniques, we may expect an increasing role for functional MRI in the management of patients with renal disease.	[ "magnetic resonance imaging", "humans", "kidney diseases", "kidney neoplasms" ]	[ "P", "U", "R", "M" ]
Matern_Child_Health_J-2-2-1592159	Oral Health in Women During Preconception and Pregnancy: Implications for Birth Outcomes and Infant Oral Health	The mouth is an obvious portal of entry to the body, and oral health reflects and influences general health and well being. Maternal oral health has significant implications for birth outcomes and infant oral health. Maternal periodontal disease, that is, a chronic infection of the gingiva and supporting tooth structures, has been associated with preterm birth, development of preeclampsia, and delivery of a small-for-gestational age infant. Maternal oral flora is transmitted to the newborn infant, and increased cariogenic flora in the mother predisposes the infant to the development of caries. It is intriguing to consider preconception, pregnancy, or intrapartum treatment of oral health conditions as a mechanism to improve women's oral and general health, pregnancy outcomes, and their children's dental health. However, given the relationship between oral health and general health, oral health care should be a goal in its own right for all individuals. Regardless of the potential for improved oral health to improve pregnancy outcomes, public policies that support comprehensive dental services for vulnerable women of childbearing age should be expanded so that their own oral and general health is safeguarded and their children's risk of caries is reduced. Oral health promotion should include education of women and their health care providers ways to prevent oral disease from occurring, and referral for dental services when disease is present. Introduction A woman's preconception and pregnancy experience with the two most prevalent diseases of the mouth—periodontal disease and dental caries—not only influences her own oral health status but also may increase her risk of other diseases such as atherosclerosis [1–4], rheumatoid arthritis [5], and diabetes [6], impact pregnancy outcome [7–9], and her offspring's risk of developing early and severe dental caries [10–13]. Although largely preventable through evidence-based interventions, both periodontal disease and caries in women of childbearing age are highly prevalent, particularly among low-income women and members of racial and ethnic minority groups. In addition, both periodontal disease and caries are typically asymptomatic for long periods of time with only intermittent painful exacerbations. The combination of high prevalence, insufficient treatment rates, missed preventive opportunities, and intermittent symptoms led the US Surgeon General to publish a report in 2001 on oral health in America characterizing dental and oral disease as a “silent epidemic” [14]. Socioeconomic factors, lack of resources to pay for care, barriers to access to care, and lack of public understanding of the importance of oral health and effective self-care practices all represent underlying reasons cited for observed inadequacies in oral health. An important oral condition affecting many pregnant women is periodontal disease. Periodontal disease is a destructive inflammatory condition of the gingiva and bone that supports teeth. It is most commonly associated with a gram-negative anaerobic infection of these structures. Fluid that bathes the tooth at the gingival margin, known as gingival crevicular fluid, often contains inflammatory mediators and oral pathogens associated with periodontal disease. The mechanisms underlying this destructive process involve both direct tissue damage resulting from plaque bacterial products, and indirect damage through bacterial induction of the host inflammatory and immune responses [15]. Destructive periodontal disease affects up to 15% of the population of childbearing age, with a relatively high proportion of pregnant women demonstrating some degree of periodontal disease [7, 8, 16]. Advancing age, smoking, and diabetes are risk factors for the development of periodontal disease [17]. Whereas periodontal disease is a chronic, local oral infection, systemic inflammation may also occur [15]. The second oral disease important to women of childbearing age because of its maternal-child health associations is dental caries. Dental caries is the pathologic process by which teeth “decay” and develop “cavities.” It occurs when acid is produced at the tooth surface by cariogenic bacteria in the dental plaque that metabolize dietary carbohydrates. Acquisition of these cariogenic bacteria, dietary practices that govern the caries process, use of fluorides that dampen the caries process, and utilization of dental care all link mothers’ and children's experience with tooth decay through biological, behavioral, and social pathways. Is maternal oral health linked to pregnancy outcome? Preterm birth Preterm birth, delivery at less than 37 weeks’ gestation, occurs in approximately 12% of all births [18, 19]. Prematurity is the leading cause of neonatal morbidity and mortality in non-anomalous infants [20]. There are numerous and heterogeneous factors associated with preterm birth, such as low maternal body mass index, maternal smoking, and maternal infections [21]. In 1996, Offenbacher and colleagues [7] first reported a potential association between maternal periodontal disease and delivery of a preterm/low birthweight infant. In a case-control study of 124 pregnant women, they observed that women who delivered at less than 37 weeks’ gestation or an infant <2500 g had significantly worse periodontal disease than control women. The adjusted odds ratio for delivery of a preterm, low birth weight infant was ∼7; these data led the authors to conclude that periodontal disease may represent a previously unrecognized and clinically significant risk factor for delivery of a preterm low birth weight infant [7]. Extrapolation from these data suggested that 18% of the preterm, low birth weight infants born annually might be attributable to periodontal disease, and thus account for a significant proportion of the $5.5 billion annual hospital costs associated with the care of preterm/low birthweight infants. In a subsequent case-control study, Dasanayake et al. studied 55 pairs of women. Logistic regression indicated that mothers with ‘healthy gingiva’ were at lower risk for low birth weight infants [22]. Women in both of these case-control studies were examined at the end of pregnancy or after delivery, which does not convincingly prove an antecedent exposure and thus causality. Despite this limitation, these early studies led to the hypothesis that periodontopathic bacteria, primarily Gram-negative anaerobes, may serve as a source for endotoxin and lipopolysaccharides, which then increases local inflammatory mediators including PGE2, and cytokines, and that this increases systemic inflammatory mediators that can then lead to preterm birth [23]. Additionally, Jeffcoat et al. [9] examined the relationship between maternal periodontal disease and spontaneous preterm birth among 1313 pregnant women, and found that moderate/severe maternal periodontal disease identified early in pregnancy was associated with an increased risk for spontaneous preterm birth, independent of other traditional risk factors [9]. Despite these compelling data, it is important to recognize that other studies have failed to demonstrate any association between maternal periodontal disease and preterm birth. In a case control study conducted in London, Davenport et al. [24] examined 236 infants born at <37 weeks’ gestation or <2500 g and compared them to a random sample of 507 control infants born at ≥38 weeks’ gestation and weighing ≥2500 g. The authors found no evidence for an association between delivery of a preterm, low birth weight infant and periodontal disease and somewhat surprisingly, found that deeper mean tooth pocket depths at delivery was associated with a reduction in the risk of delivery of a preterm, low birth weight infant [24]. The authors surmised that these discrepant findings might be due at least in part to racial differences in study populations. In a follow-up longitudinal study of 3738 women, Moore et al. [25] found no association between maternal periodontal disease and preterm birth. However, there was an increase in second trimester fetal loss rates among women with periodontal disease [25]. In an effort to better understand the possible mechanism behind the association between periodontal disease and preterm delivery, Offenbacher and colleagues [23] measured gingival crevicular levels of PGE2 and IL-1ß in 48 mothers who delivered preterm, low birth weight infants compared to control women and discovered that gingival crevicular fluid levels of PGE2 were significantly higher in case compared to control women. Furthermore, among the primiparous women delivering preterm, low birth weight infants, a significant inverse association was demonstrated between birthweight and gestational age and gingival crevicular PGE2 levels [23]. It is not yet clear whether the relationship between periodontal disease and adverse pregnancy outcomes is causal or is a surrogate for another maternal factor. As further evidence to support the concept that maternal oral health is important for normal pregnancy outcome, other investigators have examined the effect of antepartum treatment of periodontal disease on preterm birth risk. Three published studies of antepartum versus delayed (postpartum) treatment of maternal periodontal disease demonstrate promise for this intervention for preterm birth prevention. The effect of periodontal interventions on pregnancy outcome was assessed in a prospective study designed to examine the relationship between periodontal disease and preterm low birthweight infants in a cohort of young, minority, pregnant and postpartum women. Of 164 women for whom birth outcome data were available, 74 were subjected to oral prophylaxis during pregnancy, and 90 received no periodontal treatment. The preterm/low birthweight rate was lower among women who received periodontal treatment compared to those who did not (13.5% vs. 18.9%) [26]. Lopez et al. conducted a randomized clinical trial to assess the impact of periodontal treatment initiated during pregnancy versus delayed until postpartum on preterm low birthweight infant rates. The incidence of preterm/low birthweight infants in the antepartum treatment group was 1.8% (3/163) and in the delayed/postpartum group was 10.1% (19/188), (odds ratio [OR] 5.5, 95% confidence interval [CI] 1.7–18.2, P=0.001). Multivariable logistic regression analysis showed that periodontal disease was the strongest factor related to delivery of a preterm/low birthweight infant (OR 4.7, 95% CI 1.3–17.1) [27]. The data from these two studies suggest that treatment of periodontal disease during pregnancy could reduce preterm/low birthweight infant rates [26, 27]. In a pilot intervention trial designed to assess the feasibility of conducting a trial to determine whether treatment of periodontal disease reduces the risk of spontaneous preterm birth, Jeffcoat et al. found that among women at high risk for preterm birth and presence of periodontal disease, scaling and root planning therapy initiated during pregnancy is tolerated by pregnant women and may reduce spontaneous preterm birth [28]. Other adverse pregnancy outcomes Preeclampsia is a hypertensive disorder of pregnancy responsible for significant, maternal and fetal morbidity and mortality. Preeclampsia affects up to 5% of pregnant women [29]. The etiology of preeclampsia remains elusive. The underlying pathology may be related to a generalized intravascular hyperinflammatory state [29]. Some investigators have hypothesized a potential role for maternal periodontal disease as a risk factor for preeclampsia. In a retrospective analysis of data collected as part of the Oral Conditions and Pregnancy Study, Boggess et al. reported that women were at higher risk for preeclampsia if they had severe periodontal disease at delivery (adjusted odds ratio 2.4, 95% confidence interval 1.1, 5.3), or if they had periodontal disease progression during pregnancy (adjusted odds ratio 2.1, 95% confidence interval 1.0, 4.4) [30]. In a case-control study, Canakci et al. found that pre-eclamptic patients were 3.5 (95% CI=1.1–11.9) times more likely to have periodontal disease than normotensive patients [P < 0.01) [31]. In a study of 30 pregnant women, significantly higher periodontal probing depth and clinical attachment level scores were found among preeclamptic women compared with non-preeclamptic women. Gingival crevicular fluid levels of PGE2, TNF-α, and IL-1β levels were all significantly higher in the preeclamptic group [32]. Further study on the maternal and fetal inflammatory responses to chronic oral infection and on placental pathology in women with periodontal disease is needed to determine whether the relationship between periodontal disease and preeclampsia is causal or simply associative. If the relationship between maternal periodontal disease and preeclampsia risk proves causal in nature, then prevention of periodontal disease before pregnancy or treatment of periodontal disease during pregnancy may represent a novel approachs to the prevention of preeclampsia. Is maternal oral health linked to children's experience with tooth decay? Cariogenic bacteria are typically acquired by young children through direct salivary transmission from their mothers [10]. Factors influencing transmission are the levels of these bacteria in maternal salivary reservoirs, frequency and efficiency of transmission, and the child's receptivity to implantation, which is largely diet dependent. Additional factors include timing of transmission, which is affected by the window of infectivity and the age of the child, and the composition and flow of the child's saliva. The earlier the transmission and the more caries-supportive the diet, the earlier and more substantial the transfer will be. For this reason, mothers who have themselves experienced extensive tooth decay and therefore most likely harbor high titers of mutans streptococci in their saliva will more effectively transmit this infection vertically, thereby putting their young children at elevated risk for early childhood caries. Although maternal cariogenic bacteria can be isolated in the pre-dentate infant's mouth [11], these organisms become established in the dental plaque on the tooth surface only after teeth first appear at around six months of age. Because oral flora tends to remain stable over time, a woman's cariogenic flora before and during pregnancy anticipates her flora during the child's first years of life as well as the likelihood of transmitting infection early to her offspring. The lag time between infection and expression of a discernable cavity in a tooth depends upon additional factors, including the frequency of simple carbohydrate exposure in a child's diet, oral hygiene, and exposure to fluorides. The evidence that caries is frequently established as a pathologic process in the mouths of very young children is strong, as 28% of US children, over 4 million toddlers and preschoolers, experience one or more frank cavities by ages 2–5 years [33]. Given the biological and behavioral pathways that govern intergenerational transmission of caries activity, disease management, and use of dental care, it is not surprising that disparities in dental caries among adults are mimicked among their children. As with adults, children of color and children of low-income families experience substantially more extensive and severe disease and less treatment than their peers without these risk factors [33]. Fortunately, despite the high prevalence of caries in women and children, this disease is readily preventable or manageable though early and regular dental care, exposure to fluoridated water, use of appropriate topical fluorides including those in toothpastes, application of sealants to primary teeth, and adoption of a health-promoting diet like that suggested in the Dietary Guidelines for Americans [34]. Is preconception preventive oral health care the answer? It is intriguing to consider preconception, pregnancy, or intrapartum treatment of oral health conditions as a mechanism to improve women's oral and general health, pregnancy outcomes, and their children's dental health. Evidence is currently weakest for interventions that seek to reduce the incidence of preterm low birth weight through oral care. The mechanism of periodontal disease-associated adverse pregnancy outcomes is as yet unclear, and althoughit is hypothesized that if the ‘insult’ occurs early (either at conception or implantation) the risk is greater, no direct evidence to confirms that this is the case. However, given the strong relationship between oral health conditions and periodontal disease and general health and well-being, oral health care should be a goal in its own right for all individuals. If treatment of periodontal disease is going to impact pregnancy outcomes, then it is likely that the therapy will be of greatest benefit before or in very early pregnancy. The science supporting interventions before, during, and after pregnancy to reduce caries transmission is much stronger. Educational and behavioral interventions that reduce caries activity through appropriate use of fluorides, dietary guidelines, chlorhexidine gels and varnishes, and xylitol [35], can reduce a woman's caries activity and salivary cariogenic flora, thereby improving her own oral health and, at the same time, also reducing the risk of transmission to her offspring. In two landmark Swedish studies [12, 13], children of mothers who had their cariogenic oral flora suppressed were less likely to experience cavities, more likely to develop cavities later if they were affected, and had fewer cavities than children of control mothers. Pregnancy is itself often regarded as an opportune time for anticipatory guidance and oral health education, and is a suitable time, particularly during the second trimester, for dental repair. Access to oral health care during pregnancy The CDC's Pregnancy Risk Assessment Monitoring System (PRAMS) reported that only 23–43% of pregnant women received dental care during their pregnancies [36, 37]—a rate only half to two-thirds of US women's overall use of dental services (67%) (14). The PRAMS data revealed that overall, pregnant women covered by Medicaid were 24%–53% less likely to obtain a dental visit during pregnancy than women who are privately insured. Similarly, women who initiated prenatal care later than the first trimester, who did not intend the pregnancy, and who were poor were also less likely to obtain care [36]. In contrast, the Behavioral Risk Factor Surveillance System (BRFSS) revealed that 70% of pregnant women in the years 1999 and 2002 had received a dental visit in the prior 12 months [38]. One possible explanation for the higher level reported by BRFSS is that it includes three or more months of pre-pregnancy time, during which dental care utilization would be expected to reflect the national norm for women. However, in contrast to recognized disparities in dental care utilization [33], race and ethnicity were not significantly associated with dental care during pregnancy in the BRFSS study. The authors suggest that, “the prevalence of dental visits … probably reflects factors such as prevailing attitudes toward dental care, provider availability and practice norms, and salient features of medical and dental care delivery within the state.” Overall, women covered by Medicaid were 24%–53% less likely to obtain a dental visit during pregnancy than women who are privately insured. Similarly, women who initiated prenatal care later than the first trimester, who did not intend the pregnancy, and who were poor were also less likely to obtain care. An important additional consideration is that dentists are reportedly reluctant to provide care to pregnant women because of concern about possible risks [39]. Current practice typically limits non-urgent dental treatment of pregnant women to the second trimester, as there is concern about possible teratogenic consequences during the first trimester and about the woman's comfort in the dental chair during the third trimester. A single study relating antepartum dental radiography with full term low birth weight raised concern about the safety of dental care during pregnancy [40], but was criticized for its methodology. Neither professional associations nor government agencies have promulgated any authoritative guidance regarding dental care of pregnant women, although multi-center NIH clinical trials are underway that will determine the impact of dental care for periodontal disease during pregnancy on preterm low birth weight outcomes. Currently, the New York State Department of Health is developing “Guidelines for Prenatal Care, Oral Health, and Child Health Professionals” that promotes routine use of dental care during pregnancy. Independent of pregnancy, the presence and source of dental insurance coverage is an important predictor of dental care utilization with publicly insured adults experiencing higher levels of oral diseases but less access to dental care. Medicaid is particularly significant to dental care of pregnant women as this program covers approximately 1/3rd of births in the US. However, states vary widely in adult Medicaid dental coverage, and at present only 7 jurisdictions providing comprehensive care to eligible adults. In contrast, low income pregnant women seeking dental services find themselves with no coverage in 8 states, coverage for only relief of pain or infection in 18 states or eligible for a limited range of services in 18 states. Three states (UT, LA, CA) have recently expanded dental benefits specifically to pregnant women in anticipation of reduced rates of unfavorable pregnancy outcomes. Pregnancy may be the only time that some low-income woman can readily obtain dental care as some state Medicaid programs provide adult dental coverage only to pregnant women or enhanced coverage during pregnancy. Conclusions and future directions Data are emerging to support a role for maternal periodontal disease as an infectious risk factor for preterm birth and other adverse outcomes of pregnancy. The prevalence of periodontal disease and the possibility of preterm birth prevention by treatment of oral infection make this a novel approach to improve the health and well being of our mothers and their soon-to-be born children. Further studies to better understand the mechanism of periodontal disease-associated preterm birth will enable us to tailor treatment to those women who might benefit the most. Data on the relationship between maternal and child experience with dental caries is well established. Therefore, regardless of the potential for improved oral health to improve pregnancy outcomes, public policies that support comprehensive dental services for vulnerable women of childbearing age should be expanded, so not only their own oral and general health is safeguarded but also so that their children's risk of caries is reduced. Particularly if NIH trials confirm that treating pregnant women for periodontal disease reduces the incidence of unfavorable birth outcomes, the Centers for Medicare and Medicaid Services should build on its September 2004 coverage expansions for pregnant women by stimulating the states to similarly expand oral health services for pregnant women. The power of prevention needs to be brought to bear, as both periodontal disease and caries are overwhelmingly preventable through well recognized strategies including regular and effective home care for periodontal disease and use of fluorides and sealants for caries. To the degree that pregnancy provides a “teachable moment” in self-care and future child-care, prenatal education should universally adopt an oral health component. This educational intervention should prioritize those mothers who have suffered significantly from dental caries so that they can learn to effectively prevent transfer of this disease to their children. To be effective, oral health promotion must first seek to educate women and their health care providers about the importance of oral health and must promote an understanding of their ability to prevent and manage both periodontal disease and caries and to thereby limit the personal and intergenerational consequences of both conditions.	[ "oral health", "pregnancy", "infant caries" ]	[ "P", "P", "R" ]
Behav_Processes-2-1-2396234	The exploratory behaviour of rats in the hole-board apparatus: Is head-dipping a valid measure of neophilia?	The exploratory behaviour of laboratory rodents is of interest within a number of areas of behavioural pharmacology. However, how best to measure exploratory behaviour in rodents remains a contentious issue. Many unconditioned tests, such as the open field, potentially confound general locomotor activity with exploration. The hole-board apparatus appears to avoid this confound, as head-dipping into holes in the floor is assumed to be a valid measure of the subject's attraction towards novelty (neophilia). This study aimed to investigate whether head-dipping should be considered a valid measure of neophilia by comparing performance of adult male and female Lister hooded rats on the hole-board task (a) over repeated sessions and (b) when novel objects were absent or present underneath the holes. The results show that head-dipping initially decreased across repeated exposures, while time spent in the aversive central area increased. No change in head-dipping was seen in response to objects being placed underneath the holes. Rather than being a measure of neophilia, these results support the hypothesis that head-dipping represents an escape response, which declines as the subject becomes less fearful. These results are compared with previous studies of repeated exposure to other novel environments. 1 Introduction When faced with an unfamiliar environment or object, animals often exhibit behaviour patterns that broadly can be termed exploration, such as locomoting around the environment, orientating towards novelty, and touching or sniffing novel objects (Berlyne, 1950, 1960; Glickman and Sroges, 1966; Welker, 1957). Exploration potentially provides an animal with new information about food sources, shelters or mating opportunities. However, by entering a new environment or attending to a novel stimulus, an animal might also increase it's risk of predation, aggression from conspecifics or other hazards. Whether an animal investigates or avoids novelty has been described as the outcome of an approach–avoidance conflict (Montgomery, 1954, 1955; Montgomery and Monkman, 1955) or as a balance between neophilic and neophobic tendencies (Greenberg, 2003). In motivational terms, neophilia can be defined as the attraction that an animal displays towards an object or place simply because it is novel, while neophobia is the aversion that an animal shows towards approaching a novel object or place (Greenberg, 2003). In behavioural terms, neophilia and neophobia can be considered respectively as curiosity-based approach to, and fear-based avoidance of, a novel stimulus (Hughes, 2007). The exploratory behaviour of rodents has gained recent interest within a number of areas of behavioural pharmacology. For instance, researchers studying drug addiction are interested in the neural mechanisms underlying neophilia due to the apparent overlap with the neural mechanisms involved in the rewarding effects of drug-taking (Bardo et al., 1996). However, considerable controversy still surrounds the question of how best to measure neophilic and neophobic responses in laboratory animals. One of the most commonly used behavioural tests for laboratory rodents is known as the open field. Originally, the open field apparatus consisted of a flat, raised platform (Hall, 1934, 1936), although the term open field is now commonly used to refer to any enclosed arena that can range in size from a small box to a large playing field (Crawley, 1985; Whishaw et al., 2006). In such an arena, the overall level of locomotion and time spent in the centre of the arena (which is assumed to be aversive to rodents) are often interpreted as measures of exploratory behaviour. However, some researchers have argued that forcing an animal to be in an enclosed area, or on an open platform, does not allow the animal to exhibit its ‘motivation’ to explore an unknown environment, as the task evokes a strong fear response (Birke and Sadler, 1986; Denenberg, 1969; Renner, 1990; Walsh and Cummins, 1976). Corticosterone levels have been found to rise in rodents on exposure to a novel open field environment (e.g. Marin et al., 2007; Matzel et al., 2006; Rees et al., 2006), and open field behaviour is influenced by some anxiolytic (anxiety-reducing) substances (Prut and Belzung, 2003). In particular, benzodiazepines and serotonin receptor agonists, which have anxiolytic effects in human beings, generally increase the proportion of entries into the centre of the open field in rodents (Prut and Belzung, 2003). Together, these data suggest that the open field may provide valid behavioural measures of fearfulness, but may have limited value for researchers interested in measuring neophilia. Another common concern with interpreting open field tests is that differences in the performance of animals in the open field may result simply from differences in overall locomotor activity, which could be unrelated to differences in exploratory behaviour (Berlyne, 1960; Birke and Archer, 1983). In the 1970s, researchers began to use the hole-board apparatus, which consists of an enclosed arena with holes in the floor into which an animal can poke it's head, referred to as head-dipping (e.g. File and Wardill, 1975a,b; Nolan and Parkes, 1973). The frequency and duration of head-dipping are assumed to provide measures of neophilia (or directed exploration) that are independent from the general locomotor activity of the animal (File and Wardill, 1975a; Ljungberg and Ungerstedt, 1976). This apparatus has been argued, therefore, to avoid the difficulties of interpreting general locomotion that prove problematic in the open field, and a number of studies have shown that head-dipping and locomotion can vary independently of each other (e.g. Abel, 1995; Durcan and Lister, 1989; File, 1977; Lister, 1987; Rogers et al., 1999). In general, high levels of head-dipping are interpreted as indicative of neophilia, while low levels are assumed to result from a lack of neophilia or are assumed to reflect a high anxiety-like state in the animal (Crawley, 1985; Takeda et al., 1998). The hole-board task is currently being used as a test of neophilia in many areas of behavioural pharmacology (Kliethermes and Crabbe, 2006a). Researchers have attempted to validate head-dipping as a measure of neophilia by administering different classes of drugs and by comparing different genetic strains of rodents in their performance on the hole-board task. For example, if head-dipping is a neophilic response that is suppressed by an anxiety-like response, treatment with anxiolytic agents is predicted to increase head-dipping. Such studies have produced conflicting evidence; for instances, treatment of rodents with anxiolytic benzodiazepines has been reported to increase (rats: File, 1977; mice: Nolan and Parkes, 1973; Takeda et al., 1998), decrease (rats: Pellow et al., 1985) or have no effect (rats: Sayin et al., 1992) on the frequency of head-dipping. A recent review has suggested that the effects of anxiolytic compounds on head-dipping behaviour are generally confounded by changes in overall locomotion, despite the claims that head-dipping is unrelated to locomotor activity (Kliethermes and Crabbe, 2006a). Similarly, a study of several inbred mouse strains reported that head-dipping and locomotion are highly correlated (Kliethermes and Crabbe, 2006b). Whether head-dipping can be interpreted as a valid measure of neophilia remains unresolved (Bilkei-Gorzó and Gyertyán, 1996; Renner, 1990). The aim of this study was to investigate whether head-dipping behaviour should be considered a valid measure of neophilia by comparing performance of rats on the hole-board task (a) over repeated sessions and (b) when objects are placed underneath the holes. Repeated exposure to a novel apparatus is expected to produce a reduction in exploration as the animal becomes familiar with the environment, a process commonly referred to as habituation (Leussis and Bolivar, 2006). If head-dipping behaviour is a measure of neophilia, the frequency of head-dipping is therefore predicted to decrease over repeated sessions. In early study by Nolan and Parkes (1973), head-dipping by young mice (21–25 postnatal days) was reported to be lower on a second exposure to the hole-board apparatus compared to the first exposure. Two recent studies have also provided evidence that head-dipping by mice and rats decreases on repeated exposure to the hole-board apparatus (Gagliano et al., 2008; Mayeux-Portas et al., 2000). If this behaviour is a valid measure of neophilia, head-dipping is also predicted to be higher in the presence, than in the absence, of objects. Although an early study reported that rodents head-dip more frequently when objects are present (File and Wardill, 1975a), this finding has not received recent replication. Given that male and female rodents are reported to exhibit behavioural differences on the hole-board task (e.g. Aguilar et al., 2003; Ray and Hansen, 2004), subjects of both sexes were included. 2 Materials and methods 2.1 Subjects and housing The subjects of this experiment were eight male and eight female adult Lister Hooded rats (supplied by Harlan, U.K.). The animals were housed in a single room, which was controlled for temperature and humidity and was maintained on a 12-h light:dark cycle (lights on at 07:00). The animals were housed in same-sex pairs in plastic and wire mesh home-cages (measuring 25 cm × 45 cm × 15 cm) with ad libitum access to rodent pellets and water. All guidelines and requirements set out in the Principles of Laboratory Animal Care (National Institutes of Health, U.S.A., Publication No. 86-23, revised 1985) and the U.K. Animals (Scientific Procedures) Act 1986 were followed. 2.2 Apparatus and experimental design The hole-board apparatus consisted of a wooden, grey box, measuring 68 cm × 68 cm. The walls were 40 cm high, and the box was raised 28 cm above the ground on a metal stand. Four holes (4 cm in diameter) were cut into the floor of the apparatus; each hole was 28 cm from a corner of the box along the diagonal from the corner to the centre. The floor of the box was marked out into four outer areas and one central area using black masking tape. The central area was delineated by four lines of tape each 20 cm from one of the walls, while the four outer areas were marked out by diagonal lines of tape running from the corners of the floor to the corners of the central square. The four holes were thus located at the corners of the central square. The apparatus was located in a small testing room with dimmed white lighting. The stand of the apparatus was open on all sides, allowing the floor or objects to be dimly lit. Each subject was tested ten times in the hole-board apparatus, once per day during two sets of five consecutive days (Monday–Friday and the following Monday–Friday). During the first set of five trials, no objects were present underneath the holes of the apparatus; during the second set of five trials, an object was placed on the floor under each of the four holes prior to the start of the trial, approximately 20 cm below each hole. The objects were all distinct from each other but were similar in size (approximately 10 cm in length or diameter: a black-and-white rubber ball, a purple plastic star, a red-and-white rubber pet toy, and a yellow, rubber dumb-bell shaped pet toy). All trials were carried out between 09:00 and 17:00 h, and trials on males and females were alternated throughout the day. At the beginning of each trial, a subject was placed in one corner of the apparatus (always the corner closest to the door of the room), facing the centre of the arena. Each trial lasted 10 min. At the end of the trial, the subject was immediately placed into a carrying box and returned to the home cage. Between each trial, the floor and walls of the apparatus and the novel objects, if present, were cleaned with 70% alcohol solution. 2.3 Behavioural measurements During each 10-min trial, behavioural data were recorded by the observer (C.N.) onto a spreadsheet that was divided into 60 10-s time blocks. Inter-observer reliability between two independent observers (C.N. and G.R.B.) was over 80%. The following behaviour patterns were recorded:(i)enter a new area: the animal moves from one area of the open field to another (all four paws had to be placed on the floor of a new area);(ii)head-dip: the animal places it's head into one of the holes, to a minimum depth such that the ears were level with the floor of the apparatus (a new bout of head-dipping was recorded if the animal raised it's head fully out of the hole before resuming);(iii)rear: the animal is stationary on it's backpaws and raises it's forepaws off the ground, extending it's body vertically. The data on entries into a new area were used to calculate the total amount of locomotion (number of entries into all areas summed together) and the percentage of entries that were in made into the central area. The location of the animal during each of the 10-s time intervals was used to estimate the percentage of time spent in the central area. 2.4 Statistical analyses The data were analysed using repeated-measure ANOVAs, with ‘objects’ (with or without) as a within-subject variable, ‘trial’ (trials 1–5) as a within-subject repeated measure and ‘sex’ (male or female) as a between-subject variable. Significant interactions were analysed further using simple effects post hoc tests (Howell, 2007). 3 Results 3.1 Total locomotion The total amount of locomotion did not vary across trials (F4,56 = 0.096, n.s.) or vary with the presence or absence of objects (F4,56 = 0.071, n.s.). Females locomoted more on average than males (F1,14 = 4.871, p = 0.045; Table 1). There were no interactions between trial and sex (F4,56 = 1.433, n.s.), trial and object (F4,56 = 2.615, n.s.) or object and sex (F4,56 = 0.559, n.s.), and the three-way interaction between these variables was not significant (F4,56 = 1.174, n.s.). 3.2 Locomotion into the central area The percentage of entries that were into the central area differed significantly across trials (F4,56 = 14.842, p < 0.001) and varied with the presence or absence of objects (F1,56 = 84.240, p < 0.001). There was also a significant interaction between trials and the presence or absence of objects (F4,56 = 7.662, p < 0.001). Post hoc analyses indicated that the proportion of entries into the centre increased over the first five trials and remained steady thereafter (Fig. 1a). The percentage of entries into the central area varied between the sexes (F1,14 = 5.209, p = 0.039), with females exhibiting an overall greater proportion of entries into the centre than males (Table 1). There were no interactions between trial and sex (F4,56 = 0.761, n.s.) or object and sex (F4,56 = 1.956, n.s.), and the three-way interaction between these variables was not significant (F4,56 = 2.567, n.s.). 3.3 Time spent in the central area The percentage of time spent in the central area differed significantly across trials (F4,56 = 8.085, p < 0.001) and varied with the presence or absence of objects (F1,56 = 35.393, p < 0.001). There was also a significant interactions between trials and the presence or absence of objects (F4,56 = 3.493, p = 0.013). Further analyses indicated that the time spent in the centre increased over the first five trials and remained steady thereafter (Fig. 1b). The percentage of time spent in the central area varied between the sexes (F1,14 = 5.032, p = 0.042), with females exhibiting an overall greater percentage of time in the centre than males (Table 1). There were no interactions between trial and sex (F4,56 = 0.083, n.s.) or object and sex (F4,56 = 0.076, n.s.), and the three-way interaction between these variables was not significant (F4,56 = 1.001, n.s.). 3.4 Frequency of head-dipping The frequency of head-dipping differed significantly across trials (F4,56 = 2.626, p = 0.043), and there was a significant interaction between trials and the presence or absence of objects (F4,56 = 4.482, p = 0.013). Further analyses indicate that head-dipping decreased over the first five trials and slightly increased towards the end of the experiment (Fig. 2). There was no main effect of the presence or absence of objects (F1,14 = 0.142, n.s.). The frequency of head-dipping varied significantly between the sexes (F1,14 = 15.401, p = 0.002), with females exhibiting an overall greater frequency of head-dipping than males (Table 1). There were no interactions between trial and sex (F4,56 = 0.324, n.s.) or object and sex (F4,56 = 0.553, n.s.), or a three-way interaction between these variables (F4,56 = 0.779, n.s.). 3.5 Frequency of rearing The frequency of rearing differed significantly across trials (F4,56 = 3.204, p = 0.019), with frequency increasing slightly across sessions (Fig. 3). There was no main effect of the presence or absence of objects (F1,14 = 3.898, n.s.) or sex (F1,14 = 1.332, n.s.; Table 1). There were no interactions between trials and the presence or absence of objects (F4,56 = 0.457, n.s.), between trial and sex (F4,56 = 0.111, n.s.) or object and sex (F4,56 = 0.006, n.s.), or a three-way interaction between these variables (F4,56 = 1.708, n.s.). 4 Discussion The aim of this study was to investigate whether head-dipping behaviour should be considered a valid measure of neophilia by comparing performance on the hole-board task (a) over repeated sessions (trials 1–10) and (b) when no objects were present (trials 1–5) and when objects were placed underneath the holes (trials 6–10). The results show that head-dipping was high during the first test, decreased over the following two trials and remained relatively stable during the rest of the experiment. The initial drop in head-dipping following the first trial could be interpreted in two ways. First, head-dipping could be indicative of a neophilic response that decreases as the animal becomes familiar with the apparatus, i.e. head-dipping represents directed exploratory behaviour that drops as the apparatus loses its novelty. If this interpretation is correct, we would also predict that head-dipping would be greater in the presence of objects; however, there was no evidence of an increase in head-dipping behaviour when objects were present underneath the holes. These results do not support the hypothesis that head-dipping is a valid measure of neophilia. The second interpretation of the initial drop in head-dipping frequency is that head-dipping could represent a fearful, neophobic response, such that, on first exposure to the apparatus, the animal actively attempts to find an escape route (Renner, 1990). Adult male rats have been shown to exhibit an increase in circulating corticosteroid levels following a single exposure to the hole-board apparatus (Márquez et al., 2005, 2006), suggesting that testing in this apparatus is a stressful event. If this interpretation of head-dipping is correct, we would also predict that, as head-dipping behaviour declines, fearfulness would also decline. In favour of this interpretation, while head-dipping frequency declined over the first few tests, the amount of locomotion into the central area of the hole-board, and the time spent in the central area, greatly increased over these trials. Therefore, as fearfulness apparently decreased, head-dipping also decreased. If we assume that the fear experienced on exposure to a novel apparatus can be equated to normal or ‘state’ anxiety (Belzung and Griebel, 2001), these results contradict the assumption that head-dipping behaviour is suppressed by an anxiety-like response, in which case we might have expected head-dipping to vary in the opposite direction to anxiety-like behaviour. A small number of previous studies on rodents have also presented evidence that the frequency of head-dipping decreases on second exposure to the hole-board apparatus (Gagliano et al., 2008; Mayeux-Portas et al., 2000; Nolan and Parkes, 1973). The early study by Nolan and Parkes (1973) investigated whether previous experience with the hole-board apparatus with a solid floor inserted would influence frequency of head-dipping on second exposure to the apparatus. The results indicated that prior exposure to the plain board reduced the frequency of head-dipping on subsequent exposure as much as prior exposure to the board with holes (Nolan and Parkes, 1973). Again, these results do not support the idea that head-dipping provides a valid measure of neophilia, rather than a neophobic response induced by an unfamiliar test situation. In our study, head-dipping did increase in frequency towards the end of the experiment, after eight or more exposures to the apparatus. Therefore, we cannot reject the possibility that, as the subjects became very familiar with the apparatus, they engaged in a greater level of visual exploration through the holes. Rearing behaviour is also commonly interpreted as an activity by which an animal obtains information about distal environmental cues (Lever et al., 2006). Given that rearing behaviour gradually increased in frequency across the sessions, as the subjects presumably became less fearful, the amount of visual assessment of the testing room, both above and below the apparatus, may have increased. This interpretation is supported by an experiment carried out by Bilkei-Gorzó and Gyertyán (1996). This study reported that the effects of the anxiolytic benzodiazepine chlordiazepoxide on head-dipping behaviour in rats varied with the light intensity during testing, such that treatment decreased head-dipping in very bright light and increased head-dipping under normal light. The authors argued that head-dipping during the aversive testing condition (bright light) represents an attempt by the subject to find an escape route from the apparatus, while head-dipping in less aversive conditions represents visual exploration of the apparatus. Future studies could investigate whether head-dipping increases with extended familiarisation to the apparatus even in the absence of novel objects, and whether head-dipping increases when novel objects are brought closer to the holes, when new novel objects are placed under the holes for each test, or when objects are re-located beneath different holes. Our results show that repeated exposure of subjects to a novel hole-board apparatus greatly affects the behavioural response, and that the neophobic response experienced by subjects during the first exposure to an apparatus apparently declines over further exposures. Repeated exposure to another commonly used behavioural test, the elevated maze, has been reported to result in a decrease in total locomotion and a decrease in the amount of time spent on aversive open sections of the maze (e.g. Cook et al., 2002; Dawson et al., 1994; Rodgers et al., 1996). In contrast, in our study, no change of overall locomotion was found on repeated exposure to the hole-board, and the amount of time spent in the aversive central area of the hole-board apparatus increased over sessions. The reason for the differences in the behavioural effects of repeated exposure to these two types of novel environments may be related to differences in the design of the two types of apparatus. Repeated exposure to an elevated maze, which consists of open areas and closed areas, may have resulted in the subjects retreating into the darker, closed arms once an initial investigation had found no potential escape routes from the apparatus; such a response was not an option in the hole-board test. If a shelter were added to the hole-board apparatus, locomotion might be predicted to decrease over repeated exposure, with subjects choosing to remain under a shelter once potential escape routes had been investigated. In support of previous reports (e.g. Aguilar et al., 2003; Ray and Hansen, 2004), our results indicate that female rats head-dip more frequently, locomote more and spend more time in the centre of the hole-board apparatus than males. These sex differences in behaviour were not affected by repeated exposure to the apparatus, or by the presence of novel objects. Given that the locomotor patterns of female rodents are influenced by the stage of the estrus cycle and by experimental manipulation of hormones such as estradiol (e.g. Garey et al., 2001; Morgan and Pfaff, 2002), researchers have questioned whether gonadal hormones might influence general levels of arousal that might impact on how females respond to novelty compared to males (Morgan et al., 2004). However, while reports of sex differences in hole-board performance have been interpreted as showing that females are more exploratory than males, females have been reported to locomote more than males in familiar, as well as novel, environments (e.g. Cortright et al., 1997; Eckel and Moore, 2004), and, as discussed above, head-dipping may not represent neophilic behaviour. Therefore, the conclusion that females are more exploratory or neophilic than males should not be drawn from such data. The results of this study indicate that the assumption that head-dipping on the hole-board task represents a neophilic response is not necessarily accurate. At least on first exposure to the apparatus, head-dipping is likely to result from an attempt by the subject to find a potential escape routes. This issue highlights a more general problem with unconditioned behavioural tests of exploratory behaviour, in that high levels of locomotion around an environment should not immediately be interpreted as neophilia or a positively rewarded aspect of novelty-seeking (Welker, 1957). Any behaviour in a novel environment will be influenced by both neophilia and neophobia, such that a single behavioural measure is unlikely to be purely indicative of either neophilic or neophobic tendencies. Rather than being at polar ends of a continuum, neophilia and neophobia should be thought of as two orthogonal factors that can vary independently (Greenberg, 2003). For researchers interested in rodent exploration, simple measures, such as head-dipping, are unlikely to expose by themselves the complex interaction between these two factors.	[ "exploration", "rat", "hole-board", "head-dipping", "neophilia", "neophobia" ]	[ "P", "P", "P", "P", "P", "P" ]
J_Urban_Health-3-1-2134844	Quality of Water the Slum Dwellers Use: The Case of a Kenyan Slum	As a result of rapid urbanization in a context of economic constraints, the majority of urban residents in sub-Saharan Africa live in slums often characterized by a lack of basic services such as water and sewerage. Consequently, the urban poor often use inexpensive pit latrines and at the same time may draw domestic water from nearby wells. Overcrowding in slums limits the adequate distance between wells and pit latrines so that micro-organisms migrate from latrines to water sources. Sanitary practices in these overcrowded slums are also poor, leading to contamination of these wells. This study sought to assess sanitary practices of residents of a Kenyan urban slum and fecal contamination of their domestic water sources. This cross-sectional study involved 192 respondents from Langas slum, Kenya. Forty water samples were collected from the water sources used by the respondents for laboratory analysis of coliforms. Of these 40 samples, 31 were from shallow wells, four from deep wells, and five from taps. Multiple-tube fermentation technique was used to enumerate coliform bacteria in water. The study found that most people (91%) in the Langas slum used wells as the main source of domestic water, whereas the rest used tap water. Whereas most people used pit latrines for excreta disposal, a substantial percentage (30%) of children excreted in the open field. The estimated distance between the pit latrines and the wells was generally short with about 40% of the pit latrines being less than 15 m from the wells. The main domestic water sources were found to be highly contaminated with fecal matter. Total coliforms were found in 100% of water samples from shallow wells, while 97% of these samples from shallow wells were positive for thermotolerant coliforms. Three out of the four samples from deep wells were positive for total coliforms, while two of the four samples were positive for thermotolerant coliforms. None of the samples from taps were positive for either total or thermotolerant coliforms. Because the presence of thermotolerant coliforms in water indicates fecal contamination, facilitated by the proximity between the wells and pit latrines, the study suggests that the pit latrines were a major source of contamination of the wells with fecal matter. However, contamination through surface runoff during rains is also plausible as indiscriminate excreta disposal particularly by children was also common. Owing to the fecal contamination, there is a high possibility of the presence of disease pathogens in the water; thus, the water from the wells in Langas may not be suitable for human consumption. To address this problem, treatment of the water at community or household level and intensive behavioral change in sanitary practices are recommended. Efforts should be made to provide regulated tap water to this community and to other slums in sub-Saharan Africa where tap water is not accessible. However, more sampling of different water sources is recommended. INTRODUCTION Rapid urban growth in a climate of economic constraints has resulted in the majority of residents in Africa’s large cities, and an increasing proportion of Africans overall, living in overcrowded slums and shantytowns. In these slums and shantytowns, health conditions and livelihood opportunities are poor.1–3 Available evidence indicates that the poor urban residents of Africa exhibit higher morbidity, have poor access to health services, and consequently exhibit higher mortality rates than residents of other areas including rural residents.4–8 The situation in Kenya is similar to other situations in Africa. The proportion of urban population in Kenya nearly doubled between 1980 and 1998, increasing from 16 to 31%.9 Rapid urbanization amid economic degradation in Kenya has resulted in an increased proportion of people living in absolute poverty in the urban areas.10 Therefore, poverty has increasingly become a crucial urban problem in Kenya leading to mushrooming of informal settlements in the urban parts of Kenya where the urban poor find shelter. This has overwhelmed the environmental health resources in urban areas. Because of their illegal status, residents of informal settlements in Kenya do not receive government services such as water, drainage, sewerage, and rubbish collection. Consequently, informal settlements are characterized by poor environmental conditions that predispose their inhabitants to poor health outcomes.4 Evidence shows that children of poor families in urban areas of Kenya exhibit poorer health conditions than their rural counterparts. According to a report by African Population and Health Research Center (APHRC) in 2002,4 infant and child mortality risks are particularly higher in the slums of Nairobi than those observed in other urban areas and in rural Kenya. For instance, the under five mortality was 35% higher among slum residents in Nairobi than among the rural population in Kenya. The report attributes these patterns to poor water and sanitation in these slum settlements.4 An adequate supply of safe drinking water is universally recognized as a basic human need. Yet millions of people in the developing world do not have ready access to an adequate and safe water supply. By 1996, the number of people without access to safe water in urban areas was rising sharply in developing countries as a result of rapid urbanization, much of which was occurring in peri-urban and slum areas.11 Because the United Nations projects a rapid population growth in urban areas between 2000 and 2030,12 access to safe drinking water and adequate sanitation in urban areas is likely to worsen unless there is a drastic policy change to cater to the needs of the urban poor. Human excreta and the lack of adequate personal and domestic hygiene have been implicated in the spread of many infectious diseases including cholera, typhoid, hepatitis, polio, cryptosporidiosis, ascariasis, and schistosomiasis. It is estimated that one-third of deaths in developing countries are caused by the consumption of contaminated water and on average as much as one-tenth of each person’s productive time is sacrificed to water-related diseases.13 The World Health Organization estimates that 2.2 million people die annually from diarrhea diseases and that 10% of the population of the developing world are severely infected with intestinal worms related to improper waste and excreta management.14,15 In Kenya, diarrheal diseases are among the major illnesses affecting children of the slum residents. According to the report by APHRC in 2002, prevalence of diarrhea was 32% among children below 5 years of age in the slums, which is double the rate for Nairobi and the national average.4 Where ground water is used as a source of domestic water, use of pit latrines is not recommended because the two are incompatible unless the water table is extremely low and soil characteristics are not likely to contribute to contamination of ground water. Where they coexist, although it is difficult to give a general rule for all soil conditions, the commonly used guideline is that the well should be located in an area higher than and at least 15 m from the pit latrines and should be at least 2 m above the water table. Available evidence shows that increased lateral separation between the source of pollution and groundwater supply reduces the risk of fecal pollution.16 Coexistence of on-site sanitation and use of underground water has in the past been mainly confined to the rural areas where there is adequate land to allow for adequate distance between pit latrines and shallow wells. With the rapid urbanization and rapid expansion of slum settlements in sub-Saharan Africa, on-site sanitation and underground water are used in some urban areas because they are affordable options in the absence of government-supplied services. However, the congestion in the urban slums does not allow for adequate distance between the wells and the pit latrines, which allows micro-organisms to migrate from fecal contents into the underground water sources. Furthermore, poor sanitary practices (for example, disposal of human excreta) in these slum areas lead to contamination of water and consequently water-borne diseases. It is in this context and in the context of high levels of diarrheal diseases in the urban slums in Kenya that this study sought to assess the sanitary practices and the fecal contamination of domestic water sources in an urban slum in Eldoret, Kenya. STUDY CONTEXT The study was conducted between January and June 1999 in Langas, an urban slum in Eldoret municipality, Kenya, less than 10 km from Eldoret town. Eldoret town is located in the Rift Valley Province, about 330 km north west of Nairobi. Eldoret, the headquarter of Uasin Gishu District, is one of the fastest growing urban areas in Kenya. Langas falls under high density, low-income areas of the Eldoret municipality. It is divided into four administrative blocks that are further subdivided into about 2,500 plots. Each plot (1/8 of an acre) has between one and 30 households each with an average of six occupants. Settlement in Langas began in 1965 and at the time there were no basic services. Early settlers dug shallow wells for their water needs.17 The water table is high, and this raises the possibility of ground water contamination where on-site sanitation systems are in use. METHODOLOGY A cross-sectional study design was used and a sample of 192 households1 was selected through multistage sampling technique as follows: two out of the four administrative blocks were randomly picked and from the two blocks, 192 plots were picked. From each of the selected plots, one household was selected to participate. The 192 households were selected as follows: starting from one corner of each of the blocks and walking across the two blocks in a more or less a diagonal line, about every plot on the diagonal line was selected until the sample of 192 was reached. From each of these plots, one household was selected to represent the plot (this depended mainly on availability of respondents and their willingness to participate), and the first household to be contacted in a selected plot was considered for recruitment into the sample. From the household, a credible respondent was interviewed. A credible respondent was a resident of the selected household who was aged 18 years or above. For water sampling, 20 samples were taken from water points in each of the two blocks to make a total of 40 samples. This number for water samples was determined by financial and logistical feasibility. In most circumstances, residents of a plot shared one water source and there was roughly one water point per plot. In a few cases, several plots shared a water point. To get the 40 samples, every fourth household2 of the 192 households included in the interview sample was asked their source of water and a water sample was taken from this source so long as it had not already been taken. In the event that a previous fourth household shared the source with the current household, the source for the immediate next household in the study sample was considered. Of the 40 water samples, 31 were from shallow wells (defined as a hand-dug well), four from deep wells (defined as a drilled well) and five from taps (referring to tap water from the municipal council distribution system). Using the above described criteria, only one deep well was selected and purposive sampling was then used to get three other such wells, consequently including all the deep wells that were used by the study sample. Questionnaires were administered to the 192 selected households to obtain information on the type of toilet facility used, major source of domestic water, method of human waste disposal, whether drinking water was boiled, and the perceptions of possible sources of water contamination in the area. For the method of excreta disposal and water source, the main method and source were considered in instances where there was more than one method or source, respectively. The distance between the pit latrines used by the 192 households and the wells (in cases where they used wells) was estimated. We also observed sanitation practices. Water samples were collected aseptically with sterile sampling bottles. The samples were transported within 2 hours of collection in a cool box containing ice packs to the Faculty of Health Sciences, Moi University microbiology laboratory for analysis. Fecal contamination of the water was determined through isolation of indicator organisms, total coliforms, and then thermotolerant (fecal) coliforms, through multiple-tube fermentation (MTF) technique. Probability tables (McCrady tables) were used to determine the Most Probable Number (MPN) estimates of the coliform organisms per 100 ml of water. Analysis of data was generally descriptive, involving determination of frequencies. Stata statistical package was used to analyze the data. RESULTS Method of Excreta Disposal The majority of respondents (98%) said that adults used pit latrines, whereas the rest said adults defecated indiscriminately. Similarly, a majority of respondents (70%) said that children used pit latrines, whereas 30% said children used open field/defecated indiscriminately. Most of the pit latrines (95%) in the community were traditional, whereas the rest were ventilated improved pit latrines (VIP latrines). A walk through the community confirmed the report that some people excreted indiscriminately as human excreta was observed strewn all over the compounds. Source of Domestic Water Most people (89%) said they used shallow wells as the major source of domestic water, whereas 2% said they used water from deep wells and the rest said they used tap water from the municipal council (Table 1). The shallow wells often had no concrete slab and often the aperture was not covered at all or was poorly covered with a loose lid that was not lockable, whereas the deep wells had a piped system. Those who used deep wells were mainly the more affluent people in the community who often owned the plot in which the well was situated. Tap water was mainly from water kiosks where water was being sold to the slum residents. Respondents who did not use the tap water said that the water from water kiosks was expensive and unaffordable to be used for domestic purposes.3 Problems of unreliability were mentioned as hindering use of tap water from the kiosks as some respondents said that sometimes the kiosk near their house could remain closed for a whole day or more. Some of the respondents reported that the nearest water kiosk was too far from their homes. TABLE 1Source of domestic water among 192 sampled households in Langas Slum, KenyaSource of domestic waterNumberPercentTap178.9Shallow well17189.1Deep well42.1Total192100 Distance between Pit Latrine and Well The wells were very close to the pit latrines. In many circumstances (38%), the distance between the wells and the pit latrines was estimated to be less than 15 m (the commonly used guideline is that the distance should be at least 15 m). Most wells (about 59%) were estimated to be at a distance between 15 and 30 m from the pit latrines (Table 2). The distance between pit latrines and wells for the wells from which water samples were taken was similar to that of other wells (Table 3). TABLE 2Distance between pit latrine and wells for all wells in the study in Langas Slum, KenyaDistanceNumberPercent1–15 m6738.315–30 m10358.930 m and above52.9Total175100This total reflects only wells, and thus excludes taps.TABLE 3Distance between pit latrine and well for wells whose water was sampled in Langas Slum, KenyaDistanceNumberPercent1–15 m144015–30 m195430 m and above26Total35100This total reflects only wells whose water was sampled, and thus excludes taps. Boiling of Drinking Water Despite the short distance estimated between the pit latrines and the wells and the poor sanitary practices like indiscriminate excreta disposal, only 42% of those who reported using wells said they boiled their drinking water, when asked if they did. Residents’ Perception of Contamination Respondents pointed out various possible sources of contamination of the water sources in the area. These included children dipping dirty objects into water source (34%) as the main source of contamination, drawing water from the source with dirty containers (27%), domestic animals defecating around water sources (19%), and people washing clothes (5%) at the water source, among others. Interestingly, no one mentioned closeness of the well to the pit latrines as a possible source of contamination. Water Analysis Results Forty water samples were analyzed: 31 from shallow wells, four from deep wells, and five from taps. Wells Shallow Wells All the samples (31) taken from shallow wells were positive for total coliforms. The most probable number of total coliforms for most of the samples (71%) was 1,100+ per 100 ml. The minimum number of total coliforms was 63, whereas the maximum was 1,100+ coliforms per 100 ml water. Only one of the samples had no thermotolerant coliforms at all, whereas most had 1,100+ thermotolerant coliforms per 100 ml. Deep (Drilled) Wells Of the four samples taken from the deep wells, three were positive for total coliforms, whereas one was negative. The highest number of total coliforms was 240. For thermotolerant coliforms, two samples were negative. The maximum number of thermotolerant coliforms was 23. Taps For the five samples taken from taps (municipal tap water), all of them were negative for total coliforms and consequently thermotolerant coliforms. DISCUSSION This paper provides evidence on the extent of contamination of main domestic water sources in the Langas slum and suggests the most probable sources of this contamination. The evidence reveals that the most probable sources of contamination are hardly mentioned among the many sources perceived to contaminate the water sources by the residents of the slum, which impacts how interventions ought to be developed. The results indicate that the majority of the community members used pit latrines and at the same time used wells as the major source of domestic water. The conditions found in Langas do not fulfill the recommendations given for coexistence of onsite sanitation and use of ground water for domestic purposes, which indicate that there should be an adequate lateral separation between the pit latrine and the well to reduce chances of fecal contamination of the ground water.16 The distance between the wells and the pit latrines was estimated to be generally short with nearly 40% of the pit latrines estimated to be at a distance of less than 15 m from the wells. This raises the risk of contamination of the water sources as coliforms migrate from the pit latrines to the wells. The presence of indicator organisms (Escherichia coli or thermotolerant coliform bacteria) in water indicates recent contamination of the water source with fecal matter and hence possible presence of intestinal pathogens. According to World Health Organization (WHO) guidelines, E. coli or thermotolerant coliform bacteria should not be detectable in any water intended for drinking.18 The laboratory analysis results of water samples in this study show that fecal matter heavily contaminated the water sources and especially the shallow wells. None of the shallow wells met the WHO requirements for water intended for drinking. The presence of indicator organisms in the water samples collected from the wells indicates the coliforms migrated from fecal matter in the pit latrines through the soil to the water sources, facilitated by the very short distance between the pit latrines and the wells. Poor sanitary practices are also likely sources of pollution of the water sources. Sanitary practices were found to be generally poor from observation and from responses from the respondents. Thirty percent of children excreted indiscriminately and consequently there were a lot of indiscriminately disposed excreta observed. Rains are likely to wash off indiscriminately disposed excreta into shallow wells particularly if the wells are not protected. Therefore, this may have also contributed to the contamination of the generally open shallow wells with fecal matter. Other studies have also attributed contamination of water sources to wet seasons.19 Despite the contamination of water, it was evident that it was not a common practice for the slum dwellers to boil the water, as only 42% of those who used water from wells said they boiled drinking water. The illegal status of the slum areas in Kenya has hindered the expansion of municipal services to serve them. This has resulted to the poor being denied access to safer drinking water and proper sanitation. The results of this study suggest that water from the tap (which was from the municipal council) was safe for human consumption according to WHO guidelines.18 However, it is important to note that although the presence of thermotolerant coliforms indicates the presence of fecal contamination and potential presence of pathogens, absence of the same does not necessarily mean absence of pathogens, and further investigations would be worthwhile. Langas is not the only area with the problem of safe drinking water; other urban centers of the developing world have experienced similar problems.19–22 Similar to Langas, in the slums of Nairobi, Kenya, although wells are not a common source of water, slum residents are forced to buy tap water at exorbitant prices from vendors who operate without regulatory mechanisms.23,24 If the Millennium Development Goals of reducing by half the proportion of people without sustainable access to safe drinking water by 2015 and achieving a significant improvement in lives of at least 100 million slum dwellers by 202025 are to be met, there is a dire need for reconsideration of the slum areas in the developing world as far as water supply is concerned, as these slums are home to about 70% of all urban residents in sub-Saharan Africa.26 However, improving the water quality at source alone may not be the ultimate solution because improving water quality at source may not always ensure a reduction in the transmission of water-related diseases. Studies27–30 have shown significant deterioration in water quality between the source and the point of use. Esrey31 concluded that improving water had no health impact if the sanitation was not improved and that improving both water and sanitation together were synergistic in producing larger impacts than either alone. Although the results of this study suggest the need for provision of safer water sources, in this community and in many other slum communities with evident poor sanitary practices, intensive behavior change communication on sanitary practices is also paramount as this has been found effective in the reduction of water-borne diseases elsewhere.32–35 Whereas the findings from this study are worthwhile and act as an eye opener for the situation of quality of water in the rapidly growing informal settlements in the urban areas in Kenya and in the rest of Africa, more sampling of different water sources is highly recommended. CONCLUSION It is evident that most of the sources of domestic water in Langas slum are contaminated with fecal matter and do not meet the WHO guidelines for drinking water quality. This poses a health hazard to the residents of the slum as they are at risk of water-borne diseases. The results of this study also suggest that tap water may be safer, but additional sampling is needed. The ideal intervention in the long-run may therefore be provision of adequate piped water to all slum dwellers. However, this may take sometime, and simpler interventions could be put in place in the mean time. Basic sanitary improvement may be worthwhile at the moment. Covering the shallow wells and possibly installing hand pumps or mechanical pumps at the wells could improve the situation. Basic treatment of the water at the community or household level by chemical disinfection using chlorine, filtration using simple household filters, and boiling should also be promoted. These interventions may have a great impact on the health of the slum dwellers as access to safe drinking water and basic sanitation services for populations at risk would result in 200 million/year fewer diarrheal episodes and 2.1 million/year fewer deaths caused by diarrhea.35	[ "water", "slums", "kenya", "urbanization", "africa", "urban poor", "sanitation", "sanitary practices", "contamination", "coliforms", "poverty", "water quality" ]	[ "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P" ]
Environ_Health_Perspect-113-12-1314901	The Tobacco Industry and Pesticide Regulations: Case Studies from Tobacco Industry Archives	Tobacco is a heavily pesticide-dependent crop. Because pesticides involve human safety and health issues, they are regulated nationally and internationally; however, little is known about how tobacco companies respond to regulatory pressures regarding pesticides. In this study we analyzed internal tobacco industry documents to describe industry activities aimed at influencing pesticide regulations. We used a case study approach based on examination of approximately 2,000 internal company documents and 3,885 pages of U.S. Environmental Protection Agency documents obtained through Freedom of Information Act requests. The cases involved methoprene, the ethylene bisdithiocarbamates, and phosphine. We show how the tobacco industry successfully altered the outcome in two cases by hiring ex-agency scientists to write reports favorable to industry positions regarding pesticide regulations for national (U.S. Environmental Protection Agency) and international (World Health Organization) regulatory bodies. We also show how the industry worked to forestall tobacco pesticide regulation by attempting to self-regulate in Europe, and how Philip Morris encouraged a pesticide manufacturer to apply for higher tolerance levels in Malaysia and Europe while keeping tobacco industry interest a secret from government regulators. This study suggests that the tobacco industry is able to exert considerable influence over the pesticide regulatory process and that increased scrutiny of this process and protection of the public interest in pesticide regulation may be warranted. Tobacco is a pesticide-intensive crop. With nearly 27 million pounds of pesticides (including insecticides, herbicides, fungicides, and suckercides) applied to the U.S.-grown crop from 1994 to 1998, it ranks sixth in terms of the amount of pesticides applied per acre [U.S. Government Accounting Office (GAO) 2003]. The tobacco industry regards pesticides as essential to tobacco production, stating that “the crop could not be produced economically without them” (Davis 1989; Philip Morris 1990b). According to industry documents, government-imposed limitations on pesticide use “may present a serious impediment” to the international tobacco trade (Hill 1989). Internal tobacco industry documents provide a window into the tobacco industry’s activities regarding pesticide regulations. These case studies drawn from industry documents describe the tobacco industry’s responses to pesticide regulatory action. The documents also provide insight into the relationships between the tobacco industry and pesticide regulatory agencies and tensions between business and public health interests. The Tobacco Industry Documents Litigation against the tobacco industry has resulted in the release of nearly 7 million previously secret tobacco industry documents (Bero 2003; Malone and Balbach 2000). Scanned PDF versions of original handwritten, typed, or printed documents have been archived at the University of California, San Francisco, library in electronic repositories, searchable using basic keywords (http://legacy.library.ucsf.edu). Between July 2003 and February 2004, we searched the archives using a “snowball” sampling strategy, beginning with broad search terms (“pesticide” and “crop protection agent”) and using retrieved documents to identify more specific search terms (such as names of specific pesticides, people, and regulatory agencies). Table 1 provides examples of keyword searches and the number of documents yielded. This process produced nearly 300,000 documents relating to many different pesticides. The first author reviewed these documents’ index entries and excluded duplicates and documents unrelated to pesticide regulatory issues. The final sample size was approximately 2,000 documents, spanning 1974–2001. We also filed Freedom of Information Act (FOIA) requests with the U.S. Environmental Protection Agency (EPA) on pesticide issues raised by information in the industry documents, resulting in 3,885 pages of government documents. Finally, we reviewed public health agency reports based on industry documents (Zeltner et al. 2000). We analyzed the industry, government, and public health agency documents by assembling chronologically constructed case studies, a method common to sociology, political science, and anthropology (e.g., analyses of a corporation’s organizational structure, a social movement, or a tribe) (Hill 1993; Yin 1994) (Table 2). The pesticides chosen for inclusion [methoprene, the ethylene bisdithiocarbamates (EBDCs), and phosphine] were those for which sufficient information related to regulatory activities was available in the archives to construct a case study. Pesticides and Tobacco Pesticides used on tobacco are also used regularly on food crops. As with food crops, trace amounts of pesticides remain on tobacco leaves after treatment; typically, residue levels decline during the drying and manufacturing process, although additional pesticides may be applied to the finished product (U.S. GAO 2003). Although pesticides increase production of tobacco and food crops, pesticide exposure may harm humans; thus, regulatory agencies such as the U.S. EPA may set limits on the amount of pesticide residue permitted in or on food and tobacco and establish standards for workers handling pesticides. Because tobacco is burned and the smoke inhaled, active and passive smokers are exposed to pyrolyzed pesticide residues (U.S. GAO 2003). The U.S. EPA has concluded that this exposure poses no short-term risk, but little is known about the long-term health effects (U.S. GAO 2003). Methoprene In 1974, Philip Morris formed a partnership with the chemical company Zoecon to market a new insecticide (Manzelli 1975). The insecticide’s active ingredient, methoprene, acted as an endocrine disruptor in cigarette beetles and tobacco moths, preventing their larvae from maturing into adult insects (Manzelli 1975). Philip Morris anticipated that methoprene would replace phosphine, a common warehouse fumigant (Philip Morris 1988) and pledged to assist Zoecon in introducing methoprene “in as many countries as we can” (Seligman 1982). Some countries have regulations that require the establishment of maximum residue limits (MRLs) for pesticides on crops; however, Philip Morris determined that MRLs were not required in all countries, especially for pesticides on nonfood crops such as tobacco (Ryan 1991). Philip Morris asked Zoecon “to not force this issue and submit for MRLs when not required” (Lindahl 1992b). In April 1991 Zoecon alerted Philip Morris’s director of research that the Malaysian pesticide board had recently set an MRL of 1.0 ppm for methoprene on tobacco (Hutney 1991). Zoecon considered 1.0 ppm too low to enable the effective use of methoprene; the level supported by the labeled application rate was 10 ppm (Ryan 1992). Philip Morris requested that Zoecon ask for an even higher MRL of 15 ppm to allow for application errors (Greenberg and Transon 1992; McCuen 1992). Zoecon representatives met with government authorities and requested a change to 15 ppm (Hutney 1991). A Zoecon representative informed Philip Morris that “in order to avoid surprises of this nature in the future,” he had directed Zoecon’s pharmaceutical group to obtain information from health authorities in other countries regarding the commodities for which methoprene tolerances were assigned (which could include foods such as rice and mushrooms as well as tobacco) (Hutney 1991). Assigning this task to the pharmaceutical group instead of the pesticide group, the Zoecon representative wrote, “will not arouse the curiosity of the health directorates and will allow us to keep our promise to the tobacco industry, namely, that we won’t initiate queries that may cause the health authorities to direct attention to tobacco” (Hutney 1991). In April 1992, George Lindahl of Zoecon faxed a letter to Bob McCuen, head of Philip Morris’s biochemical research, outlining some of his concerns about Philip Morris’s approach to establishing MRLs for methoprene on tobacco (Lindahl 1992b). In regard to Zoecon’s effort to establish an MRL of 15 ppm in Malaysia, Lindahl explained that I know we simply argued this case without any data to support our request. In more advanced countries, this tactic will not succeed. … All our data demonstrate the need for a 10 ppm MRL. If a higher value is desired then we will require data from real field operations showing that a worse [sic] case scenario for faulty application will result in a 15 ppm residue, and hence the need for this value. (Lindahl 1992b) In a fax following this one, Lindahl asked Philip Morris to provide such data; a handwritten comment from a Philip Morris employee who reviewed the fax noted that “data doesn’t [sic] exist” (Lindahl 1992a). Initially, the Malaysian authorities agreed to increase methoprene’s MRL to 10 ppm (Lindahl 1992c); subsequently, it was raised to 15 ppm (Mueller and Ward 1998). Philip Morris continued to advocate (through Zoecon) for MRLs of 15 ppm in Italy and Germany (Greenberg and Transon 1992). In the meantime, anticipating the creation of a single European market with uniform pesticide regulations, Philip Morris had asked the longtime tobacco industry law firm, Shook, Hardy, and Bacon, to prepare a document with MRL recommendations for possible submission to the European Community (Kemna 1991). Philip Morris first provided a draft of recommended MRLs to the Scientific Working Group of the Confederation of European Community Cigarette Manufacturers (CECCM) (Philip Morris 1991c). At their June 1991 meeting, members of this group (including representatives of Philip Morris, British American Tobacco, R.J. Reynolds, Gallaher, and Rothmans) recommended that the document be rewritten as a voluntary code of practice “to be used pre-emptively … in advance of any EC [European Community] initiative” to impose formal regulations on pesticide residue limits on tobacco (Philip Morris 1991a). A meeting participant reported, “It is hoped that, by implementing this Code, the EC Commission would not any longer see the need to develop a formal EC regulation on pesticide residues in tobacco (products)” (Mueller 1991). Manuel Bourlas, Philip Morris’s director of research and development, was appointed chair of a subgroup of tobacco company representatives who were to assist in preparing the code (Philip Morris 1991a). This voluntary code underwent numerous revisions throughout 1991 and 1992 (CECCM 1991, 1992a, 1992b, 1992c, 1992d, 1992e; Philip Morris 1991b). Although 236 regulated and unregulated tobacco pesticides were in use at the time (Mitchell 1991b), the voluntary code proposed MRLs for only 25–27 pesticides [including chlordane, dichlorodiphenyltrichloroethane (DDT), lindane, dithiocarbamates, methoprene, and maleic hydrazide]. According to British American Tobacco’s Terry Mitchell, “many of the substances in the list are no longer recommended for tobacco production” (e.g., DDT) (Mitchell 1991a). Moreover, this list did not impose “any constraint automatically on non-specified substances” (Mitchell 1991a). Mitchell noted that this lack of limits was “highly desirable” (Mitchell 1991a). In December 1992, Walter Russell, a legal assistant, reported that the code “has undergone two more revisions (by SHB) [Shook, Hardy, and Bacon] and it [is] currently watered down, but still causing much agitation” (Philip Morris 1992). Russell pointed out that the code set MRLs that Philip Morris “might have trouble complying with” if they were to become international standards (Philip Morris 1992). In addition, “failure to comply with tolerances written by the tobacco industry which might come up during litigation would put the tobacco industry at great disadvantage” (Philip Morris 1992). He indicated that Philip Morris had decided to withdraw its support from the voluntary code (Philip Morris 1992). In 1993, the tobacco companies suspended work on the document due to “principle disagreements both within and between participating companies” (R.J. Reynolds 1993). Throughout the 1990s, the tobacco industry continued to anticipate European Union harmonization of tobacco pesticide MRLs (Philip Morris 1995); as of April 2004, the European Union had established community-level MRLS for 150 pesticides, but none specifically applied to tobacco (European Union 2004). EBDC Fungicides In 1987, the U.S. EPA initiated a review of EBDC fungicides, prompted by the agency’s determination that a breakdown product of EBDCs, ethylene thiourea (ETU), was a probable human carcinogen (U.S. EPA 1987). Anticipating the U.S. EPA’s cancellation of many EBDC uses, U.S. manufacturers voluntarily withdrew EBDC registrations for all but 13 food crops in 1989, including wheat and corn (U.S. EPA 1989). At least one company continued to hold registrations for EBDCs on tobacco, but only for seed bed use, not plants (Arce 1989). In internal documents, the tobacco industry expressed concern that the U.S. EPA’s action could result in the “imposition of potentially crippling product residue tolerances” in Europe [Centre de Coopération pour les Recherches Scientifiques Relatives au Tabac (CORESTA) 1989b; Mitchell 1990]. EBDCs were regarded as vital to control blue mold outbreaks in Europe (Philip Morris 1990a). In October 1989, members of CORESTA, an international tobacco research organization with members drawn largely from the tobacco industry, established a subcommittee to “provide regulatory agencies with a sound basis for the development of tobacco agro-chemical regulations” (CORESTA 1989a, 1989b). As discussed in a larger World Health Organization (WHO) report on tobacco industry influence at that agency, the subcommittee hired a consultant, Gaston Vettorazzi, to provide advice on influencing regulation (CORESTA 1990b; Zeltner et al. 2000). Vettorazzi was a former WHO toxicologist and former technical secretary of the Joint Food and Agriculture Organization/WHO Meeting on Pesticide Residues (JMPR), an international meeting of scientists whose decisions often formed the basis of international law (Zeltner et al. 2000). Selected partly for his “old boys’ contacts” (Reif 1991b), Vettorazzi’s initial duties were to provide a review and analysis of toxicologic data on EBDCs and ETU (CORESTA 1990a). Some CORESTA members were concerned that Vettorazzi’s review might conclude that EBDCs were unsafe (Beuchat 1990). However, according to one member’s notes, at his first meeting with the subcommittee in April 1990, Vettorazzi stated that “someone has to lay the red carpet for [me], otherwise [I] can spoil more than help” (Reif 1990). Vettorazzi’s initial review concluded that ETU was neither carcinogenic nor genotoxic (Vettorazzi 1991a). Some of the tobacco industry scientists commented that this statement was “too strong in light of the NTP feeding studies”—a reference to the U.S. National Toxicology Program’s conclusion that animal studies showed clear evidence of ETU’s carcinogenicity (Reif 1991a). Vettorazzi subsequently revised his conclusions, stating that ETU’s “toxicity, including carcinogenicity, can be explained by the known mechanisms of action characteristic of thyroid-function inhibiting agents” (Vettorazzi 1991b). Thus, he stated, a threshold could be set below which ETU did not cause thyroid tumors (Vettorazzi 1991b). CORESTA authorized the distribution of Vettorazzi’s revised report to his former colleagues at WHO, once all references to tobacco and CORESTA were removed (CORESTA 1992). WHO’s JMPR was scheduled to review EBDCs/ETU in 1993; if this review were favorable, the tobacco industry would be assured continued access to EBDCs in Europe (Zeltner et al. 2000). With CORESTA funding ($100,000 a year) and approval, Vettorazzi offered to assist J. Herrman, of the JMPR WHO Secretariat, with JMPR toxicologic reviews, without disclosing his tobacco industry ties (Herrman 1991; Vettorazzi 1991c, 1992a). Vettorazzi wrote and reviewed several working papers on compounds to be discussed at the 1992 JMPR, including the EBDC thiram (Herrman 1992; Vettorazzi 1992b). One outcome of that meeting was the reestablishment, at a higher level, of the previously cancelled Acceptable Daily Intake (ADI) for thiram (Vettorazzi 1992b). Vettorazzi continued his work with WHO in 1993, supplying his CORESTA-funded reviews to the adviser responsible for drafting the working paper that would form the basis of the September JMPR on EBDCs/ETU without revealing their sponsor (Zeltner et al. 2000). Vettorazzi also attended the September meeting as an invited “temporary adviser” (Zeltner et al. 2000). The meeting’s outcome reflected Vettorazzi’s conclusions. In contrast to the U.S. EPA, JMPR determined that ETU was not genotoxic, and thus raised the ADI level from 0.002 to 0.004 mg/kg body weight (Black 1993). CORESTA considered this “a very positive result for the industry,” since it “clearly indicates that the ‘carcinogenicity’ of [ETU] is not really a burning issue any longer” (CORESTA 1994; Mueller 1993). JMPR’s safety standard became part of international trade law, preserving tobacco industry access to EBDCs (Zeltner et al. 2000). Soon after the JMPR meeting, CORESTA extended Vettorazzi’s contract for 18 months, listing one of his duties as providing “information about the activities of pesticide action groups” (CORESTA 1993). He was to be paid another $100,000 (CORESTA 1993). Vettorazzi continued working for CORESTA until at least 2001, when the organization paid him $30,000 to monitor international activities related to tobacco pesticide residues and registrations (CORESTA 2001). Phosphine Phosphine is a fumigant used on stored commodities, including nuts, seeds, grains, coffee, tobacco, and finished cigarettes to kill insects. Because of the risks it poses, applicators are advised to wear respirators and protective clothing, and warehouses must be sealed to prevent leaks that contribute to air pollution and endanger nearby residents (U.S. EPA 1998b). By the early 1990s, several case reports had been published noting sometimes fatal phosphine poisoning among workers and community members (Garry et al. 1989, 1993; Heyndrickx et al. 1976; Schoonbroodt et al. 1992; Wilson et al. 1980). In December 1998, the U.S. EPA proposed a series of 15 risk mitigation measures (RMMs) for phosphine. The U.S. EPA’s primary concern was the risk that phosphine posed to applicators and community residents (U.S. EPA 1998b). Thus, the RMMs included a threshold limit value of 0.03 ppm of phosphine during fumigation (reduced from the existing 0.3-ppm standard), the establishment of a 500-foot buffer zone around all fumigated structures, and prior notification of all residents living within 750 feet of a fumigated structure (U.S. EPA 1998a). The Tobacco Association of the United States, in a letter to the U.S. EPA, stated that the economic burdens imposed by the RMMs would “make it virtually impossible for our industry to continue to fumigate stored tobacco” (Ward 1999). The Tobacco Association, R.J. Reynolds, Philip Morris, and > 150 other organizations with a stake in the continued use of phosphine formed a lobbying group, the Commodity Industry Coalition for Phosphine Fumigation (Harrell 1999). R.J. Reynolds, represented primarily by toxicologist Joel Seckar, took an active role in the Commodity Industry Coalition (Seckar 1999c). The company calculated that complying with the U.S. EPA’s buffer zone requirement would cost approximately $50 million in new land and warehouse purchases (R.J. Reynolds 1999a). Increasing the time required to aerate warehouses before employee reentry to comply with the worker exposure limit of 0.03 ppm would increase costs, as would the possibility of liability suits brought by nearby residents notified of phosphine use (Degesch America 1998; R.J. Reynolds 1999d). Coalition members lobbied Congress, released media statements, worked closely with the U.S. Department of Agriculture, and attended U.S. EPA-sponsored stakeholder meetings (Goldman 1998; Lyon 1999; R.J. Reynolds 1999b, 1999c). Their message was that the proposed RMMs were overly conservative, based on “anecdotal information and hypothetical risk” rather than on “sound science” (Lyon 1999; Ong and Glantz 2001). To challenge the scientific basis of the U.S. EPA’s proposals, the coalition decided to hire an expert whose research would support existing standards (Seckar 1999h). They chose Sciences International, a consulting firm specializing in health and environmental risk assessment. It was headed by Elizabeth Anderson, a former director of the Carcinogen Assessment Group and the Office of Health and Environmental Assessment at the U.S. EPA (Sciences International 2005). She was also an experienced expert defense witness, having served in that capacity in a number of environmental lawsuits brought against corporations (Anderson 1999c). To support the Commodity Industry Coalition’s assertion that the proposed exposure level of 0.03 ppm was too conservative, Sciences International focused on the inter-species uncertainty factor. The U.S. EPA had first determined from a published subchronic toxicity study of rats that there were no observed effects attributable to inhaled phosphine at 3 ppm (Seckar 1999a). To extrapolate to humans, the U.S. EPA had then used two 10-fold uncertainty factors, one for intraspecies variability and one for interspecies variability, to arrive at a maximum exposure level of 0.03 ppm (Sciences International 1999c). Documents indicate that Sciences International’s strategy was to convince the U.S. EPA that the interspecies uncertainty factor was unnecessary, showing that because a number of animal species reacted in the same manner to phosphine, humans were similar enough that the interspecies uncertainty factor could be removed (Seckar 1999a, 1999b). This would leave only the intraspecies factor of 10, which would result in a maximum exposure level for humans of 0.3 ppm, the existing standard. In April 1999, the U.S. EPA representatives met with a small group of Commodity Industry Coalition members, including R.J. Reynolds’s Seckar and Sciences International’s Anderson (Seckar 1999a). Anderson questioned the U.S. EPA’s interspecies uncertainty factor, citing several animal studies and an epidemiologic study to suggest that the U.S. EPA’s calculations were too conservative (Seckar 1999a). In an e-mail, Seckar noted that Anderson’s presentation was very effective, as evidenced by the fact that U.S. EPA representatives were now informing coalition members that the 0.03 ppm standard “was not ‘set in stone,’” a direct contradiction of earlier statements to the U.S. Department of Agriculture (Bair 1999; Seckar 1999d). (Despite Freedom of Information Act requests, we were unable to obtain U.S. EPA documents related to its meetings with the coalition.) Soon after, Sciences International asked the Commodity Industry Coalition for additional funding to turn its phosphine report into a peer-reviewed journal article (Turim 1999). In a memo to Seckar, Anderson (1999b) explained that My experience is that consultant reports funded by those being regulated, and written expressly for the EPA, are easily and frequently ignored or dismissed by the Agency, no matter how scholarly. However, a paper or article that is peer-reviewed and published, or in the peer review process for publication, in an accepted scientific journal can neither be ignored nor dismissed. Anderson suggested that since she was editor-in-chief of Risk Analysis, “perhaps the peer review process could be expedited if we decide that it is the journal of choice” (Anderson 1999b). R.J. Reynolds, Brown and Williamson, and several other tobacco companies agreed to fund most of the cost of this work (Seckar 1999e). The paper was published in Risk Analysis in 2004, with the acknowledgment that “This work was supported by the Phosphine/Metal Phosphide Coalition, consisting of the producers and users of phosphine and metal phosphides for the control of insects in stored commodities” (Pepelko et al. 2004). Coalition members also pursued other strategies. At a meeting with U.S. EPA representatives in March 1999, the Commodity Industry Coalition proposed that the U.S. EPA participate in a series of small, coalition-sponsored focus groups to “educate [EPA] on the issues involved with … fumigations” (Seckar 1999g). One such group met in May 1999, when tobacco companies demonstrated a tobacco warehouse fumigation (Ward and Cowan 1999). The following month, several companies conducted additional emissions tests to show that the proposed 500-foot buffer was unnecessary (Bridges 1995). However, an e-mail message from a Philip Morris employee indicated that Philip Morris’s test coordinator had “some reservations regarding the quality of the test design/data generation” and that he himself believed that “the test plan and methods will provide, literally, no information, so it won’t hurt us to do it” (Bridges 1995). In June 1999, Sciences International submitted a first draft of its phosphine toxicity review to some coalition members (Sciences International 1999a). A reviewer from the coalition’s lobbying firm pointed out that the animal studies cited did little to support the idea that the interspecies uncertainty factor should be eliminated “since most [of the animals] appear to be rat or mouse strains with similar breathing characteristics” (Wilkinson 1999). Instead, the studies cited by Sciences International seemed to support the idea that phosphine called for a conservative standard, as they indicated that “phosphine is a very toxic material to most species tested” (Wilkinson 1999). Another reviewer noted that the uncertain and tentative tone of the report “will trigger concerns by EPA and they will say ‘if [an] expert in the field states that there remains great uncertainty, maybe we are on solid ground by being very conservative’” (Barolo 1999a). Sciences International staff revised the report, removing tentative statements and asserting that their work to date supported reducing the interspecies uncertainty factor to 1 (effectively eliminating it), thus preserving the existing exposure standard of 0.3 ppm (Sciences International 1999b). They submitted this revised interim report to the U.S. EPA in July 1999 (Sciences International 1999b). At a Commodity Industry Coalition meeting that same month, coalition consultant Dan Barolo, former director of the U.S. EPA’s Office of Pesticide Programs (OPP), reportedly urged members to speed their efforts because phosphine is quite hazardous when used improperly. The more the Coalition slows the process, the greater the chance for an accident with possible fatalities, which would send EPA back into conservative mode and make it far more difficult for them to publish reasonable RMMs. (Seckar 1999f) In August, John Whalan, a toxicologist at the U.S. EPA’s Health Effects Division, summarized in a memo his analysis of Sciences International’s interim report (Whalan 1999). He noted that there is no precedent for using an [interspecies uncertainty factor] of 1 when establishing … an inhalation regulatory value in the Health Effects Division. The only time an interspecies [uncertainty factor] is not applicable is when human data are used. The available data do not support deviating from Agency policy, and the Coalition did not provide any new data. (Whalan 1999) He also pointed out that Sciences International’s review of animal studies, intended to show that phosphine toxicity was relatively constant across species, was largely “irrelevant” because it did not include a comparison of toxicity for a small versus large mammal. In September 1999, phosphine registrants and several coalition members again met with U.S. EPA officials to discuss alternative RMMs proposed by the coalition (Seckar 1999i). Instead of a 500-foot buffer and a 750-foot neighbor notification requirement, the coalition recommended a “site management plan” that required companies to develop emergency preparedness measures. The U.S. EPA asked the Commodity Industry Coalition to reword its proposals to specify how and when workers and bystanders would be informed of danger (Seckar 1999i). On the exposure limit for workers, the U.S. EPA now proposed a 0.1-ppm standard (reflecting a reduction from 10 to 3 in the interspecies uncertainty factor) based upon Sciences International’s interim report (despite the weaknesses noted by Whalan) (Seckar 1999i). (The U.S. EPA failed to provide memos or notes regarding this decision.) In several fall 1999 memos to Seckar, Sciences International staff explained that they thought it would be difficult to convince the U.S. EPA to drop the interspecies uncertainty factor without human exposure studies (Anderson 1999a; Gray 1999). Commodity Industry Coalition members expressed reluctance to commit to human studies without confirmation that this would convince the U.S. EPA to “give up” the uncertainty factor (Barolo 1999b). Barolo commented to Seckar, “I do not believe it will be easy for OPP to abandon both safety factors. There are too many unknowns from children to endocrine to reliability of studies to absence of dog/monkey study. … Some day they are going to figure out there is a 0.1 ppm standard in other countries and the door will close” (Barolo 1999c). Although Sciences International had not yet submitted to the U.S. EPA its full report on phosphine, in December 1999, the U.S. EPA made its final decision (Sharp 1999). (This decision was published in the Federal Register in February 2001 [U.S. EPA 2001]). The U.S. EPA now mandated a “fumigation management plan” like that proposed by the Commodity Industry Coalition (U.S. EPA 2000). The agency also eliminated the inter-species safety factor and left the old 0.3-ppm standard in place, on condition that phosphine registrants conduct additional research if Sciences International’s review was found to be inadequate (U.S. EPA 2000). A coalition member noted that “it is important to point out that this additional work will take years and that the current 0.3 ppm threshold will stay in place during that time” (Sharp 1999). R.J. Reynolds credited its leadership on the scientific issues with saving the company “many millions of dollars” (R.J. Reynolds 2000). Conclusion Although others have charged that agencies responsible for protecting human health and the environment are unduly influenced by the industries they regulate (Abraham 2002; Huff 2002), it is rare to be able to study this process from the perspective of the regulated industry. This study provides documentation of the behind-the-scenes activities of an industry as it attempts to influence the regulatory process on matters that have a direct bearing on public health. Our analysis has limitations. Given the sheer volume and limited indexing of the documents, it is impossible to ensure that we located all potentially relevant documents. Some may have been destroyed or concealed by the tobacco companies (Liberman 2002); others may have never been obtained in the legal discovery process. In addition, we had no access to pesticide company documents, except those in the tobacco documents archives. Finally, despite properly filed Freedom of Information Act requests, we were unable to obtain from the U.S. EPA documentation of its meetings with the industry’s Commodity Industry Coalition. All minutes of meetings with stakeholders should be part of the public record. Despite these limitations, the case studies discussed here provide insight into tactics that the tobacco industry applies to a regulatory agency when trying to influence the outcome of a decision. These tactics go significantly beyond the usual approaches—such as participation in public comment periods and public meetings—to influence scientific and regulatory decision making. Tobacco industry tactics described in these cases include: Encouraging a chemical company (Zoecon) to advocate for high MRLs without any supporting data and directing that same company to gather information about international regulatory efforts on methoprene in a manner designed to hide the interest of the tobacco industry in this chemical; Attempting to forestall regulatory efforts on tobacco pesticides in the European Community by creating voluntary industry MRLs for a subset of chemicals; Hiring an ex-WHO scientist to participate (without disclosing his funding source) in the WHO regulatory effort on EBDCs; Hiring several ex-U.S. EPA scientists to influence the U.S. EPA’s regulatory decision making on phosphine; Hiring scientific consultants with instructions to marshal data to support the tobacco industry’s a priori arguments and funding consultants to publish a report supporting these arguments in a journal over which the consultants had influence; Staging fumigations for the U.S. EPA with the knowledge that the methodology was flawed and the results would show no emissions problem. Yet, as the case of European MRLs showed, the tobacco industry does not always work together effectively to influence regulations. Tobacco companies may disagree about regulatory strategies or conclude that inaction is preferable to action that might have unintended consequences. Moreover, the fact that even voluntary, industry-friendly pesticide guidelines posed significant problems for Philip Morris underscores tobacco industry motivation for resisting or influencing more stringent, government-imposed regulations. This study also raises questions about industry influence over regulatory agencies. In the case of WHO deliberations on EBDCs, the tobacco industry coordinated covert actions, hiding the financial ties and involvement of CORESTA. Rigorous disclosure requirements and oversight might have allowed the WHO’s agencies to judge more accurately the potential for bias related to conflicts of interest. In the case of the U.S. EPA’s review of phosphine, a regulatory agency appears to have been quite willing to cooperate with the industry and its consultants. This is a reminder of why regulatory processes were designed to be transparent and open to the public, and why “closed-door” meetings between regulators and industry have been ruled illegal (Federal Advisory Committee Act 1972; Registration Standards 2004; Special Review Procedures 2002). Protection of the public interest hinges on an open process and regulatory agencies’ willingness to stand up to pressure from regulated industries. When these are in doubt, public confidence in the fairness and efficacy of regulations may be unwarranted. The resource disparities between powerful industries and public health organizations may also make it difficult to ensure that the public interest is fairly represented, particularly when discussions occur behind closed doors, as apparently occurred at the U.S. EPA. Increased public and media scrutiny of these processes could help ensure that public health considerations are weighed at least as heavily as commercial ones. Finally, given the deadly epidemic of tobacco-caused disease, which kills an estimated 5 million people annually worldwide (WHO 2004), is it in the public interest for regulatory agencies today to continue facilitating standards that make it easier and less costly to grow, transport, store, and manufacture tobacco products?	[ "tobacco industry", "pesticide regulation", "environmental protection agency", "methoprene", "ethylene bisdithiocarbamates", "phosphine", "world health organization" ]	[ "P", "P", "P", "P", "P", "P", "P" ]
Purinergic_Signal-4-2-2377325	Purinergic signalling in the subretinal space: a role in the communication between the retina and the RPE	The retinal pigment epithelium (RPE) is separated from the photoreceptor outer segments by the subretinal space. While the actual volume of this space is minimal, the communication that occurs across this microenvironment is important to the visual process, and accumulating evidence suggests the purines ATP and adenosine contribute to this communication. P1 and P2 receptors are localized to membranes on both the photoreceptor outer segments and on the apical membrane of the RPE which border subretinal space. ATP is released across the apical membrane of the RPE into this space in response to various triggers including glutamate and chemical ischemia. This ATP is dephosphorylated into adenosine by a series of ectoenzymes on the RPE apical membrane. Regulation of release and ectoenzyme activity in response to light-sensitive signals can alter the balance of purines in subretinal space, and thus coordinate communication across subretinal space with the visual process. Introduction The retinal pigment epithelium (RPE) lies between the outer segments of the photoreceptors and the choroidal blood supply (Fig. 1). The RPE combines the functions of epithelial and glial cells, providing a barrier while also supporting the neural photoreceptors and modulating their function. Tight communication between photoreceptors and the RPE is critical to coordinate the multiple levels of interaction, and the purinergic contribution to this communication is becoming apparent. The relevance of this purinergic input is emphasized by the many functional effects of P1 and P2 receptor stimulation and by the multiple mechanisms in place to regulate subretinal levels of purine agonists. As the dynamics of ATP release and extracellular conversion into adenosine will modify agonist availability, the modulation of these processes can exert a temporal control on purinergic signaling. The following review will outline the main interactions between the RPE and photoreceptors, describe the effects of stimulating purinergic receptors on both sides of subretinal space, and summarize how levels of ATP, ADP, and adenosine are manipulated in this microenvironment. Fig. 1Schematic illustration of the key components of purinergic signaling in the subretinal microenvironment. Stimulation of P2 receptors on the RPE can enhance transepithelial fluid absorption while P1 receptors can modulate phagocytosis. ATP released through CFTR and other Cl− channels can stimulate P2 receptors or be converted to ADP, AMP, and adenosine (Ado) by a series of ectonucleotidases present on the apical membrane of the RPE. By controlling the balance of extracellular purines available to stimulate these receptors these mechanisms can control levels of endogenous purines available to activate the receptors. While theoretically possible, it remains to be determined whether these subretinal purines can actually stimulate photoreceptors Purines and subretinal space RPE-photoreceptor interactions across the subretinal space The outer segments of the rods and cones are responsible for the initial stages of vision, converting photon energy into a series of enzymatic reactions that close the light-sensitive channels on the photoreceptor plasma membrane, hyperpolarize the cells, and reduce the release of glutamate from the synaptic terminals [1, 2]. Efficient photoreceptor function depends upon both short-term and long-term support from the RPE. The critical nature of these interactions is evident from the rapid degeneration of photoreceptors in the absence of a healthy RPE layer and by the RPE localization of defective gene product in some forms of hereditary photoreceptor degeneration [3]. The apical membrane of the RPE is separated from the plasma membrane of the outer segments by an extracellular space of only 10–20 nm [4]. Although small, this subretinal space contains a highly structured matrix which ensheathes the outer segments and extends to the RPE [5, 6]. The presence of enzymes within this interphotoreceptor matrix emphasizes that this extracellular space itself is functionally active [7, 8]. This intimate anatomical relationship between photoreceptors and the RPE reflects multiple functional interactions. For example, the RPE accepts, recycles, and exports central components of the phototransduction process [9]. The outer segments are continuously resynthesized, and the phagocytosis, degradation, and processing of shed outer segment tips by the RPE cells is central to this renewal [10]. The ion channels and transporters on the apical membrane of the RPE help regulate the ionic composition of the subretinal space [11]. As extracellular levels of ions can modify the ionic driving forces across the photoreceptor plasma membrane, these RPE transporters can influence the state of neural activity. The transport of fluid and ions from the apical membrane to basolateral membrane of the RPE is also one of the main forces keeping the retina attached [12]. The control of photoreceptor activity by light gives a rapid temporal dependence to some interactions between the photoreceptors and the RPE. The volume of subretinal space increases in response to light [13], linking small changes in the ionic composition of the subretinal space with activity of the RPE transport mechanisms which maintain this volume [14, 15]. Other processes are controlled on a diurnal cycle. The shed tips of the outer segments are phagocytosed by the RPE soon after the onset of light [16, 17]. These processes can both be modulated by purine levels in subretinal space, indicating purinergic regulation is important over multiple time scales. Purinergic receptors on photoreceptors A2 adenosine receptors were localized to both the inner and outer segments of photoreceptor outer segments over a decade ago by Blazynski and colleagues [18], with more recent reports emphasizing their functional role. A2 agonists inhibit the L-type Ca2+ channel on rod outer segments [19] and can inhibit the synaptic release of glutamate from rods, suggesting changes in adenosine levels in subretinal space could modulate light sensitivity [20]. The A2 agonist DPMA and the adenosine deaminase inhibitor EHNA reduce mRNA for opsin in rods, suggesting that endogenous levels of adenosine can downregulate opsin message at night [21]. EHNA and the A2A receptor agonist CGS21680 also increase the survival of chick embryonic photoreceptors in culture [22], indicating a long-term neuroprotective role for adenosine. P2 receptors are also present in the photoreceptors. mRNA for the P2X2 receptor is expressed in the photoreceptor cell bodies, with immunohistochemical localization of the protein to outer segments [23]. In situ hybridization indicates the photoreceptor layer has the highest level of P2Y2 receptor of any region in the rabbit retina, although staining was not pronounced in monkey [24]. P2X7 receptors have recently been localized to photoreceptor synaptic terminals, with evidence for ecto-ATPase activity in the synapse, and functional evidence suggesting ATP augments transmission of the light response by rods [25]. It was suggested that ATP might be co-released from photoreceptors with glutamate, although this remains to be tested directly. Purinergic receptors on the RPE Stimulation of P1 receptors can have a considerable impact on RPE cells. A2 receptors have been recognized on cultured and fresh RPE cells for some time [26, 27], with in situ hybridization confirming the presence of A2A receptors in rat RPE [28]. Stimulation of A2 receptors reduces the rate of rod outer segment phagocytosis by RPE cells [29], while application of adenosine to the apical membrane of chick RPE cells increases the basolateral Cl- conductance, the transepithelial potential, and the c-wave, and decreases the hyperpolarization of the basal membrane in response to light [30]. Although adenosine alone does not increase intracellular Ca2+ levels [31], adenosine acts synergistically with ATP to elevate Ca2+ levels in human RPE cells by stimulating both A1 and A2A receptors [32, 33]. Stimulation of A1 receptors with high doses of NECA increases the active transport of fluorescein across the RPE, while activation of A2A receptors decreases this transport, and by extension, transport of the ions that underlie fluid movement [34]. Stimulation of A1 and A2A receptors produces analogous increases and decreases, respectively, in the absorption of subretinal fluid blebs. This is consistent with the negative coupling of the A1 receptor and the positive coupling of the A2 receptors to adenylate cyclase, as increasing cAMP inhibits the transport of fluid across the RPE towards the choroid [35–37]. The agonist 2-Cl adenosine reverses the deficit in phosphoinositide metabolism found in diabetic RPE cells [38], suggesting effects on metabolism in addition to transport and phagocytosis. Multiple P2 receptors have been localized to the RPE. The P2Y2 receptor was initially characterized in cultured human RPE [31], with subsequent reports localizing transcript for P2Y1, P2Y2, P2Y4, and P2Y6 in the rat RPE/choroid [39] and for P2Y1 and P2Y12 receptors in ARPE-19 cells [40], and functionally identifying a P2X receptor in rat RPE cells [41]. ATP, ADP, and UTP induce numerous effects on RPE physiology [32, 33, 42, 43]. While these effects likely involve multiple different receptor types, the contributions of the P2Y2 receptor have been explored in most detail to date. The P2Y2 receptor has been specifically localized to the apical membrane of fresh bovine RPE cells, and addition of ATP to this membrane transiently elevates Ca2+, activates a basolateral Cl- conductance, inhibits an apical K+ conductance, and increases the apical to basolateral flow of fluid [43]. This increased absorption of fluid from the subretinal space suggests P2Y2 receptor stimulation could reduce the excess fluid that accumulates in retinal edemas, and several reports have supported this theory. ATP, UTP, and the P2Y2 receptor agonist INS37217 decrease the size of subretinal fluid blebs when injected into subretinal space of rats [44]. In both normal and rds +/- mice with experimentally induced detachment, INS31217 improves the ERG recovery and decreased cell death [45]. INS37217 also reduces subretinal blebs in rabbits [46]. Injection of another P2Y2 agonist, INS542, increases the active transport of fluorescein across the RPE, consistent with this upregulation of ion and fluid transport across the tissue [47]. Together these experiments emphasize the clinical potential of treating retinal edema with P2Y2 agonists. Regulation of purine levels in subretinal space While synthesized purinergic agonists may prove useful in treating some ocular disorders, the endogenous activation of P1 and P2 receptors in the subretinal microenvironment will ultimately be determined by availability of agonists. These levels are largely controlled by the release of ATP into the subretinal space, its conversion into other purines including adenosine, and the manipulation of adenosine by enzymes or transporters. Recent work has increased our understanding of both the stimuli that initiate changes in subretinal purine levels and the mechanisms that mediate these changes. Release of ATP by the RPE At least some of the ATP capable of stimulating the purinergic receptors on RPE cells is released from the RPE itself. The resulting autocrine stimulation ensures local delivery, and control, of purines to initiate the physiologic changes in the RPE. The release of ATP by RPE cells is triggered by multiple stimuli including osmotic stress, bFGF, UTP, NMDA, glutamate, and ATP [39, 40, 48–51]. The ATP release following activation of NMDA receptors by glutamate may have the most interesting implications for communication across subretinal space, given that glutamate confers the light signal from photoreceptors to the rest of the visual system. Glutamate and the specific receptor agonist NMDA triggers ATP release from ARPE-19 cells, with the release inhibited by NMDA antagonist MK-801, and by DCKA, which inhibits the glycine B binding site on NMDA receptors [51, 52]. Although NMDA raises intracellular Ca2+ levels, this increase is prevented by eliminating ATP with apyrase, indicating autostimulation through released ATP is responsible for this Ca2+ signal. NMDA also triggers a release of ATP when applied to the intact bovine RPE eyecup [51]. The NMDA receptors and the ATP release sites have been functionally identified to the apical membrane of the bovine RPE, suggesting the neurotransmitter interactions could amplify the signal from any glutamate reaching subretinal space. The ability of both UTP and ATP to stimulate release of ATP from the RPE supports the theory that the system acts to amplify signals. When applied at greater than 1 μM, ATP triggers a secondary release of ATP peaking 10 min after the initial stimuli [40]. UTP also initiates a release in extracellular ATP with a similar delay [48]. The rise in ATP triggered by UTP is inhibited by the Cl- channel blocker NPPB, and UDP is much less effective at triggering release than UTP; both observations suggest the P2Y2 receptor contributes to the increase in ATP more than diphosphokinase, although influence from the enzyme cannot be ruled out [53]. Recent evidence suggests that ischemia may lead to the release of ATP from RPE cells. Chemical ischemia triggers a substantial ATP release from cardiac myocytes [54], while changes in oxygen levels trigger ATP release in central chemoreceptors [55]. We found that exposure to sodium cyanide led to a rapid release of ATP from ARPE-19 cells (Fig. 2). As hypoxic and/or ischemic challenge may lead to changes in the expression of growth factors in RPE cells during certain ocular disorders such as macular degeneration [56], and as purines can induce expression of VEGF in other cells [57], this ATP release may contribute to growth factor signaling by the diseased RPE. Fig. 2Chemical ischemia triggers ATP release from ARPE-19 cells. ATP release was measured in the bath directly from cells plated in 96-well plates to which the luciferin- luciferase reaction mixture was added [51]. Left Levels of ATP in the bath after addition of 5 mM NaCN to the cells. Measurement began 1 min after addition of NaCN or control solution to wells. Right Levels of ATP measured at the peak, 3 min after addition of NaCN (n = 12). Levels were normalized to concurrent levels in control (n = 14). Symbols and bars represent mean ± SE, *p < 0.001 The particular conduit for ATP release varies with the stimuli. The release in response to hypotonic challenge is largely dependent upon CFTR, as it was prevented by the specific CFTR inhibitor CFTR172 in addition to the more general blocker glybenclamide [50]. While the precise mechanisms by which CFTR contributes to this release are not yet known, a role for CFTR in ATP release into subretinal space is consistent with the reduction of certain ERG components in cftr -/- mice [58] and with the ability of apical ATP to activate conductances associated with these ERG components [43]. The release of ATP is also largely blocked by the vesicular transport inhibitor brefeldin A, suggesting the two processes occurred in series whereby ATP efflux follows the insertion of vesicles containing CFTR into the plasma membrane. Although the Ca2+ chelator BAPTA blocks this ATP release [50], raising Ca2+ alone with ionophore ionomycin does not itself initiate release [48]. This necessary but not sufficient contribution of Ca2+ also supports a role for vesicular insertion. In contrast to the release following hypotonic challenge, the ATP release in response to NMDA does not involve CFTR [51]. Release is blocked by NPPB, however, suggesting another type of anion channel could serve as a conduit for ATP release. The presence of parallel mechanisms coexisting on the same cell for ATP release triggered by either agonists or by cell swelling has also been reported in astrocytes [59] and may reflect the multiple roles of purinergic signaling within a given tissue. As both stimuli lead to release across the apical membrane into subretinal space, both are expected to influence signaling in the microenvironment. Interconversion of purines in subretinal space The interconversion of nucleotides and nucleosides each capable of stimulating distinct receptors makes the purinergic signaling system of particular interest in a confined region such as the subretinal space. The main enzymes responsible for dephosphorylating extracellular ATP on the RPE cells have been analyzed and a basic understanding of their regulation has begun. This section first describes the enzymes that act on ATP and ADP, followed by enzymes which convert AMP into adenosine. The dephosphorylation of extracellular ATP by RPE cells involves enzymes from multiple families [40], as found in airway epithelial cells [60]. Degradation of ATP by the apical membrane of the fresh bovine eyecup and by ARPE-19 cells is inhibited by ARL67156 or βγmATP. Message for eNPP1, eNPP2, and eNPP3 is present in ARPE-19 cells, consistent with the preference of βγmATP for members of the eNPP family [61]. The cells also express NTPDase2, and NTPDase3, although the intermittent presence of NTPDase1 likely reflects a regulated process [40]. Ecto-alkaline phosphatase has no effect on ATP degradation in RPE cells, in contrast to its considerable contribution in airway epithelium [62]. The putative contribution from diphosphokinases to interconversion of subretinal purines is presently unknown. Extracellular AMP is rapidly dephosphorylated into adenosine in subretinal space. The production of adenosine from ATP at the apical membrane of the bovine RPE eyecup is inhibited by the ecto-5′-nucleotidase inhibitor αβmADP, confirming a role for this enzyme [63]. The enzyme is localized to rat RPE and ARPE-19 cells immunohistochemically. Degradation of 5′AMP is highest near the subretinal space of rat retina [63], although localization in mouse indicated larger amounts of ecto-5′-nucleotidase at the tips of adjacent Müller cells [64]. Levamisole does not inhibit the dephosphorylation of 5′AMP by the RPE, consistent with the absence of substantial ecto-alkaline phosphatase in subretinal space. The presence of light may alter the levels of adenosine in subretinal space. Epinephrine is released at the onset of light [65] and stimulation of the RPE with epinephrine can decrease activity of ecto-5′-nucleotidase [63]. While norepinephrine and phenylephrine lead to similar decreases in enzyme activity, prazosin and corynanthine block the effects of norepinephrine, implicating the α1 epinephrine receptor in the inhibition of ecto-5′-nucleotidase [63]. The kinetics of inhibition are consistent with cleavage of the nucleotidase from its GPI anchor. The phagocytosis of rod outer segments is maximal shortly after light onset [16], and this phagocytosis is inhibited by adenosine [29]. The ability of epinephrine released by the illuminated retina to reduce ecto-5′-nucleotidase activity and consequently adenosine levels may relieve this inhibition and enhance the rate of phagocytosis at light onset. Physiologic effects of subretinal purines on the RPE and photoreceptors The number of purinergic receptors on both photoreceptor and RPE membranes suggests purines make multiple contributions to the physiology of the outer retina. Our increased understanding of how agonist levels in subretinal space are controlled has begun to indicate how and when this contribution may occur. Future research will involve applying these findings from isolated systems to intact RPE-photoreceptor models, and pursuing the role of defective purinergic regulation in ocular disease. While it is unlikely that ATP released across the apical membrane of the RPE can diffuse to these P2 receptors in the outer plexiform layer given the ecto-ATPase activity in the synaptic clef [25], stimulation of receptors elsewhere on the photoreceptor membrane is possible. It would be interesting to determine whether ATP released from the RPE and converted to adenosine by ecto-nucleotidases can actually modulate the response to light by stimulating the A2A receptors on photoreceptor outer segments. The impact of purinergic signaling on chronic ocular diseases is also of interest, such as the role of ischemia-driven ATP release in VEGF production. While the small size of subretinal space can complicate pharmacologic manipulation within the intact RPE-photoreceptor complex, molecular approaches may provide new insight into how endogenous purines in subretinal space affect the physiology, and pathophysiology, of both RPE and photoreceptors.	[ "rpe", "photoreceptors", "microenvironment", "glutamate", "atp release", "cftr", "ecto-5′-nucleotidase", "ntpdase" ]	[ "P", "P", "P", "P", "P", "P", "P", "U" ]
Pediatr_Radiol-4-1-2292498	Voiding urosonography with ultrasound contrast agents for the diagnosis of vesicoureteric reflux in children	Voiding urosonography (VUS) encompasses examination of the urinary tract with intravesical administration of US contrast agent (UCA) for diagnosis of vesicoureteric reflux (VUR). The real breakthrough for US examination of VUR came with the availability of stabilized UCAs in the mid-1990s. This article presents a comprehensive review of various procedural aspects of VUS. Different US modalities are available for detecting the echogenic microbubbles: fundamental mode, colour Doppler US, harmonic imaging and dedicated contrast imaging with multiple display options. The reflux is graded (1 to 5) in a similar manner to the system used in voiding cystourethrography (VCUG). The most commonly used UCA for VUS, Levovist, is galactose-based and contains air-filled microbubbles. The recommended concentration is 300 mg/ml at a dose of 5–10%, or less than 5%, of the bladder filling volume when using fundamental or harmonic imaging modes, respectively. There are preliminary reports of VUS using a second-generation UCA, SonoVue. Here the UCA volume is less than 1% of the bladder filling volume. There is no specific contraindication to intravesical administration of UCA. The safety profile of intravesical Levovist is very high with no reports of side effects over a decade of use in VUS. Introduction Diagnostic imaging for vesicoureteric reflux (VUR) is a common procedure in children. Currently, three modalities are available for reflux diagnosis, namely voiding cystourethrography (VCUG), radionuclide cystography (RNC) and voiding urosonography (VUS). The last of these is carried out using US in combination with intravesical administration of US contrast agent (UCA). This allows the use of ionizing radiation to be avoided, which is not the case in RNC and VCUG. However, in VCUG a marked reduction in radiation dose has been achieved with the introduction of digital pulsed fluoroscopy. The first attempts at the implementation of US for the diagnosis of VUR began in the mid-1970s. A comprehensive account of the evolution of this undertaking over the subsequent two decades has been presented by Darge [1]. The indirect methods for reflux diagnosis were based on US of the urinary tract, without administration of any kind of substance into the bladder. These included depicting various sonomorphological changes of the urinary tract as a result of VUR, detecting newly appearing or an increase in existing ureteral or pelvicalyceal dilatation during voiding and assessing ureteric jet changes with duplex and colour Doppler US. The direct means used to diagnose VUR required instilling different substances intravesically. The most frequently administered fluid was physiological saline solution. Ballooning of the renal pelvis during the filling of the bladder was the criterion for diagnosis of VUR. Application of air bubbles, by shaking the normal saline before administration or adding carbon dioxide, were also tried. US studies were also carried out, in which the empty bladder was solely filled with air. In addition to low diagnostic accuracy, all the above methods had major procedural drawbacks making them impractical for widespread integration into routine imaging. The intravesical use of a UCA consisting of sonicated albumin (Albunex; Molecular Biosystems, San Diego, Calif.) for VUS in a child was first reported in 1994 [2]. Another UCA used in the past was Echovist (Schering, Berlin, Germany), which is composed of galactose with incorporated microbubbles [3]. Its very short imaging window of approximately 5 min, however, prevented its routine application. The breakthrough in US diagnosis of VUR in children came about the mid-1990s with the availability of UCAs containing stabilized microbubbles. Levovist (Levograf, Schering Spain, Madrid, Spain; SHU-508-A, Schering, Berlin, Germany) was the first such UCA to become available for clinical use in Europe. This opened the door for rapid development of VUS and its introduction as part of the routine diagnostic imaging option of VUR. A number of different names and acronyms have been put forward to denote US examination for the diagnosis of VUR using intravesical UCAs. These include simply “sonography/ultrasound” [4–6], “reflux sonography” [7], “cystography” [8–10], “cystosonography” [11–15], “cystourethrosonography” [16] and “urosonography” [17–23]. They are more often used in combination with one or more of the following terms: “echo-enhanced”, “contrast-enhanced” and “voiding”. Depending on the US technique employed further descriptions such as “colour Doppler US” or “harmonic imaging” are added. The most widely applied name “voiding urosonography” with the abbreviation “VUS” was proposed for the first time in 2000 [24]. The selection of this particular designation was based on careful consideration of various factors. The prefix “uro-” is used to denote the bladder, ureters and kidneys. “Sonography” was chosen rather than “ultrasonography” as it is a shorter form. Thus “urosonography” correctly describes the fact that in this examination US of the bladder, ureters and kidneys takes place independent of the absence or presence of VUR. “Voiding” was selected instead of “micturating” as it is used more commonly in the medical literature. Despite the fact that for VCUG x-ray contrast agent is necessary, this is not directly added to the name VCUG in the form of, for example, contrast-VCUG. Likewise, the terms “contrast” or “echo-enhanced” were not added by default to the basic term “voiding urosonography”. The abbreviation “VUS” is not only similar to the most common acronym used for the radiological reflux examination, namely VCUG, but had also not been used in the medical literature to denote something else [25]. For the sake of minimizing the confusion of names and facilitating communication, literature search and procedural standardization, the use of one name for the same procedure, “voiding urosonography” (“VUS”) is recommended. This article is part I of a comprehensive review of all currently available literature on VUS and prepares the ground for objective evaluation and decision-making. In this part (part I) a detailed procedural description is presented, including discussion of the pros and cons of the various examination steps and imaging modalities for VUS. The measures undertaken to optimize VUS are elaborated. Furthermore, studies dealing with adverse events of intravesical administration of UCAs are reviewed. In part II a review of comparative studies between VUS and RNC/VCUG is presented [26]. US contrast agent The most widely used UCA for VUS is Levovist. This first-generation UCA was introduced for intravenous (IV) use in the mid-1990s and from 1999 started to obtain approval for intravesical application in children successively in 13 European countries and Australia. Levovist consists of dry granules made of galactose and palmitic acid [27]. The granules easily disintegrate into microparticles upon preparation. They form porous structures, which are necessary for the formation of bubbles. Gas bubbles are formed within the pores during suspension of the granules, while dissolution of the galactose particles begins simultaneously. The Levovist bubble is a microbubble of air—65% nitrogen and 35% oxygen—stabilized by palmitic acid. One gram of Levovist granules contains 999 mg d-galactose and 1 mg palmitic acid. The concentrations of microbubbles in freshly prepared batches (300 mg/ml) are within the range of approximately 1–2×108 microbubbles per millilitre of suspension. The Levovist suspension should always be freshly prepared prior to intravesical administration. The steps of preparation, as outlined by the manufacturer, must be strictly adhered to in order to avoid any reduction of contrast enhancement due to improper handling [28]. The administration of freshly prepared UCA has to be carried out within 30 min [29]. Concentration The recommended concentration of Levovist for VUS is 300 mg/ml [28]. This concentration has been used for VUS in most studies. A concentration of 200 mg/ml seems to result in faster dissipation of microbubbles, as found in in-vitro experiments (unpublished data). In only three studies has a concentration of 200 mg/ml been used [13, 15, 30]. Bosio initially used this concentration, but later changed to the higher one as the conspicuity of refluxing microbubbles in fundamental mode was inadequate [13]. In another study in adults with transplanted kidneys this concentration in combination with an increased volume, i.e. 10–15% of bladder filling, was used [30]. The combination of colour Doppler US with the burst technique administering 200 mg/ml selectively in infants has been reported, but the results using just this concentration were not analyzed separately [15]. The utilization of the higher concentration of 400 mg/ml has only been reported once without any additional advantages being mentioned [31]. Probably, the disadvantage would be an increase in UCA volume leading to an unwarranted increase in cost. Dosage The volume of UCA administered has been rather variable. Some have applied a fixed dose for all, independent of the patient’s age, weight and bladder volume [4, 5]. In one study the volume was calculated in terms of body weight [13]. In most studies the volume of bladder filling is taken as the decisive factor [1, 12, 28]. This is very logical when considering intravesical administration of UCA and contrast enhancement of bladder content. It is possible to directly measure the bladder volume using the formula for an ellipse (length×width×height×0.5], note the total volume of normal saline administered in the bladder and also calculate in a simple way the maximum bladder capacity: volume in millilitres = (age in years +2)×30 [32]. Furthermore, using the bladder volume for calculation of UCA dose makes comparison between studies much easier. Nakamura et al. [21] carried out simultaneously VCUG and VUS in 56 children. The fundamental (i.e. conventional) imaging modality was used for VUS. They initially administered Levovist into an almost empty bladder followed by continuous infusion of the radiographic contrast agent. They simultaneously monitored the appearance of reflux during VUS, relating it to the concentration of the UCA in the bladder. The concentration of UCA in the bladder at one point in time ranged form 1.8% to 23%. All false-negative results in the VUS were associated with a Levovist volume of <5% of the total bladder filling. Thus a volume of UCA between 5% and 10% of the bladder filling is required when performing VUS using fundamental imaging. As there is a clear correlation between the experience of the sonographer and the sensitivity of VUS, it is advisable for beginners to start with the highest recommended volume of 10% and with increasing experience to reduce the amount, but without going below 5% of the bladder filling when using fundamental imaging. It seems possible to reduce the UCA volume when adding colour Doppler US [15]. There is clear evidence that when using harmonic imaging, 5% or less of UCA with respect to the bladder filling will suffice [18, 33]. Preliminary experience with dedicated contrast imaging modalities such as cadence agent detection imaging (ADI) (Acuson; Siemens, Mountain View, Calif.) point to the potential of further dose reduction [34]. Physicochemical properties UCAs were primarily developed for IV use and, consequently, preclinical studies were aimed at elucidating problems that could potentially have been encountered during administration via this route. Intravesical use is different and has its own peculiarities one must be aware of. An in-vitro study was carried out with Levovist to measure how variations in US machine setting, transducer choice, mode of application and bladder content may affect the microbubbles [35]. The effects of power output, transducer frequency and injection rate were found to be comparable to those during IV administration. The physicochemical properties of particular relevance for intravesical use are described below. Interaction with normal saline solution Unlike in VCUG and direct RNC, normal saline plays an important role in VUS. It is the actual bladder filling medium in which the microbubbles are injected. In an in-vitro study the median contrast duration when the UCA was mixed with normal saline solution from glass containers was 30 s, whereas when the same UCA was added to normal saline solution from a plastic container the median contrast duration increased to 11 min [36]. None of the plastic containers were sealed under vacuum. In contrast, most of the glass containers were sealed under vacuum. The mean pO2 of the normal saline solutions from vacuum-sealed containers was found to be 50% less than the pO2 of the normal saline solutions from plastic containers. The air-filled microbubbles of Levovist are permeable to air, and gas is exchanged with the atmosphere until equilibrium is reached [27]. Thus, in normal saline solution with low air saturation the microbubbles tend to collapse promptly [37]. Collapse of the microbubbles means that echo enhancers are not available. In practice, the use of normal saline solution from plastic containers is preferred. Interaction with urine Contrast duration of the UCA has been found to be more than four times longer in urine than in normal saline. It has been shown that simply adding urea to buffer the solution significantly increases contrast duration [38]. It is postulated that urea affects the bonding between water molecules and facilitates the formation of more microbubbles. Thus, having some urine in the bladder when performing VUS actually has a positive effect on imaging. Ascent of microbubbles in the ureter Once reflux of microbubbles takes place from the bladder into the distal ureter, the question as to whether the microbubbles can ascend passively to the proximal ureter and renal pelvis while the refluxed fluid stays in the distal ureter has been raised [39]. Given constancy of gravity, the density difference between gas and fluid and the viscosity of fluid, the velocity of ascent is proportional to the square of the microbubble radius. The typical microbubble radius of Levovist is 1–2.5 μm. Calculation of velocity of ascent indicates that for all practical purposes passive ascent can be excluded. Furthermore, in an in-vitro simulation of VUS with a set-up imitating the worst-case scenario (90°, glass tube (no adsorption), constant stirring of fluid suspension, and absence of counter-flow) the possibility of passive ascent was evaluated using an ultraviolet spectrometer for the detection of microbubbles. The result of this experiment was also negative. These findings suggest that the microbubbles in the ureter do not ascend passively and that reflux pressure is necessary for propagation, particularly in vivo with a constant counter-flow of urine from the renal pelvis to the bladder. Procedural details VUS entails four major basic steps: (1) scan of the urinary tract before contrast agent administration, (2) intravesical administration of prewarmed physiological saline solution and UCA, (3) scan of the urinary tract after administration of UCA, and (4) scan of the urinary tract after administration of UCA during and after voiding [28, 40]. Optionally, a transperineal US of the urethra and/or cyclical filling of the bladder may be added. Various US modalities are available for performing VUS. The main differences between the modalities are in the degree of conspicuity of the microbubbles and sensitivity of reflux detection. US scan modalities Fundamental imaging This conventional modality in B-mode is the earliest and most widespread modality employed for VUS. In the case of fundamental imaging it is important to use the same scan planes and magnification of the ureters and renal pelves before and after UCA administration in order to facilitate comparison of the images (Fig. 1). This is particularly important when the reflux is not obvious. Fig. 1Scans in fundamental mode before (a, c) and after (b, d) contrast agent administration of a dilated left distal ureter (a, barrow) and pelvicalyceal system (c, d). In the postcontrast scans echogenic microbubbles fill the distal ureter (b) and are also detected in the pelvicalyceal system (d) Colour Doppler US The microbubbles increase the backscattered signal from the urine giving it a characteristic chromatic effect, e.g. a set of blue and red punctiform colour collection [32]. This makes recognition of the urine flow direction easier and enhances the visualization of the microbubbles in the pelvicalyceal system. It is recommended that the Doppler US settings be optimized to perform VUS [31, 41]. Currently, there are five reports in which colour Doppler US alone or in combination with fundamental imaging have been used for VUS [11, 14, 31, 41, 42]. The diagnostic accuracy of colour Doppler US was compared with fundamental imaging in only one study [41]. In the same group of patients VUS without and with colour Doppler US followed by VCUG was carried out. With VCUG as the reference method, the diagnostic accuracy was found to be 96% with colour Doppler US and 90% without. In another study, no significant increase was found in the sensitivity and specificity of VUS when adding colour Doppler US [42]. In one study a modified colour Doppler US examination was used employing a high mechanical index (MI) burst contrast technique based on stimulated acoustic emission (SAE), in which the microbubbles are made to burst creating strong acoustic signals [15]. It was possible to detect more reflux with the burst colour Doppler US compared to regular colour Doppler US or fundamental imaging and even VCUG. No systematic evaluation is available regarding the application of power Doppler US for VUS. Overall, adding optimized colour Doppler US to fundamental imaging seems to have some advantage and should be tried when other contrast-specific modalities are not available. Harmonic imaging The nonlinear propagation property of US waves is the basis of harmonic imaging [43]. This may be optimized to receive harmonics from either tissue or microbubbles, thus creating “tissue-specific harmonic imaging” or “contrast-specific harmonic imaging”, respectively. Some US machines only use the second harmonic for imaging (narrow band), whereas others are capable of implementing a wider range of harmonics (wide band). There may be some degree of difference in image quality between these different technical approaches, in general the latter being of higher quality. Harmonic imaging increases contrast and spatial resolution and also results in a reduction of artefacts [43]. The resulting images are clearer and crisper. The advantage of this modality compared to fundamental imaging for scanning the urinary tract in children has already been shown [44]. Due to attenuation, at a certain depth the positive effect of harmonic imaging will disappear. Regarding contrast-enhanced imaging, there is distinctly increased conspicuity of the microbubbles with harmonic imaging (Fig. 2). Air-filled bowel can be a disturbing factor, particularly when imaging the retrovesical space, requiring adjustment of the gain. Darge et al. [17] compared fundamental and harmonic imaging options in 54 children undergoing VUS. In all cases the conspicuity of the microbubbles was much higher with harmonic imaging than with the fundamental modality. Moreover, there was an increase in the reflux detection rate of 30% (from 19 to 27 pelvic-ureter units, PUUs). In this study the volume of Levovist used was 10% with respect to bladder filling. Even reducing the volume of UCA, more reflux episodes were detected in VUS with harmonic imaging than in VCUG [18, 33]. If available, harmonic imaging should be chosen in preference to the fundamental modality for VUS [17, 18, 33]. Fig. 2Scans after contrast agent administration in fundamental mode (a, c) and with harmonic imaging (b, d) of the bladder and right dilated ureter (a, bdotted circle) and a duplex kidney (c, d) with a multicystic dysplastic upper moiety. Reflux in the right ureter and in the lower moiety of the duplex kidney (grade II, arrow) are much more conspicuous with harmonic imaging (b, d). Note also the crisper depiction of the cysts in the upper moiety with harmonic imaging Dedicated high-MI contrast imaging with multiple display options In recent years, there have been major developments in US technologies for contrast-enhanced US. One such US modality uses high-MI imaging resulting in destruction of microbubbles and depicts the bubble destruction signature as a colour overlay with the possibility of visualizing the grey-scale image alone, the grey-scale image together with the microbubbles with colour overlay or just the microbubbles with colour overlay alone. Additional features may be real-time dual imaging in which two of the above modalities can be used in parallel and the possibility to switch between the three different presentations after having acquired an image in just one option (Figs. 3 and 4). This modality is named differently by different manufacturers: for example, agent detection imaging (ADI; Sequoia, Acuson Siemens), contrast tissue enhancement imaging (CTEI; Technos MPX, Esaote), rate subtraction imaging (RSI; Aplio, Toshiba), true agent detection (TAD; Logiq 9, GE), etc. [45]. The increase in microbubble conspicuity with this modality is so striking that even a beginner should find the examination easy to perform. The potential for further reduction of the dose of UCA and duration of examination are enormous and require future evaluation. If available, this US technique is preferred for VUS using Levovist. This feature is also available with low-MI imaging, which is important when using newer generation UCAs. Fig. 3VUS with the application of dedicated contrast imaging modality using high-MI (agent detection imaging, ADI). Reflux into the right ureter (arrow) and right renal pelvis (grade II). a Using grey-scale display alone the refluxing microbubbles are not easily demonstrated but the renal parenchyma is seen well. b Using “grey scale + contrast” visualization of both the reflux with a colour overlay and the bladder and kidney are noticeably improved. c Using “contrast only” the tissue part has been subtracted and only the microbubbles are seenFig. 4VUS with the use of a dedicated contrast imaging modality using high-MI (agent detection imaging, ADI). Note the time and number of the images (arrow). Once just one image has been documented it is possible as a postprocessing option to switch between the different modalities for display: a “grey-scale + contrast” and b “contrast only” options. In this case of grade III reflux marked intrarenal reflux is present in the upper pole (arrowhead) Procedural steps Precontrast scans of the bladder, ureters and kidneys The necessity to perform a detailed scan of the urinary tract arises when the VUS is combined with a follow-up US of the urinary tract, for example after pyelonephritis or when there is a need for comparison with the postcontrast images. The scan before contrast agent administration should be carried out in accordance with standard US of the urinary tract in the supine position (and/or the prone position), in both the transverse and longitudinal planes [28]. Particular attention should to be paid to documentation of subtle changes in the retrovesical region, the vesicoureteric junction and any dilated ureter. The renal pelves (and/or the calyces) are imaged with maximum magnification. A scan in the supine position may suffice if adequate visualization of the renal pelves on both sides is possible, even though scanning the kidneys with the child in the prone position definitely allows better demonstration of the pelvicalyceal system. It should always be kept in mind that not performing a precontrast scan reduces the duration of the whole examination. This should be the case whenever harmonic or other dedicated contrast imaging is used. Intravesical administration of normal saline and US contrast agent The UCA and normal saline are administered via a transurethral catheter, but suprapubic puncture is also possible. The administration of UCA is carried out under US monitoring. The aim should be to have homogeneous strong contrast agent density of the bladder contents. Early studies advocated filling the bladder to the maximum prior to injecting the UCA [13, 28]. This makes the calculation of UCA volume to be administered easier, but has the disadvantage that low-pressure reflux may be obscured. Moreover, in neonates and infants due to repeated voiding at small bladder filling volume, there is insufficient time to administer UCA and scan the urinary tract. Increasingly, a more fractional approach to administering UCA is being practised [21, 40]. It is important to note that emptying the bladder for the sake of UCA administration is not mandatory as explained above. It is helpful to inject UCA into a bladder that is not completely empty in order to have better visualization of the microbubbles and avoid a strong dorsal acoustic shadow. Using a three-way valve it is easy to alternately administer normal saline and UCA adjusting the volume of UCA to the bladder filling volume and in the end reaching the recommended dose. The Levovist suspension should be administered slowly for two reasons: first, to minimize the destruction of the microbubbles and second, to reduce settling of the suspension at the bottom of the bladder that would create an acoustic shadow that temporarily obscures the retrovesical region [28]. In such a case, turning the child several times from side to side hastens the homogeneous distribution of the UCA in the bladder. The administration of normal saline is continued until the child has the urge to micturate or there is the first slight sign of back pressure to the infusion or injection. Postcontrast scan of the bladder, ureters and kidneys The US scan after contrast agent administration is basically carried out in the same manner as that before contrast agent administration. There are minor procedural variations depending on the type of US modality utilized. The diagnosis of reflux is made when echogenic microbubbles appear in a ureter or renal pelvis. When a ureter is distinctly visible behind the bladder, the microbubbles render the echo-free lumen echogenic. When the ureter is not clearly visible as a round echo-free structure, VUR can be demonstrated if microbubbles are seen entering the vesicoureteric junction or if just behind this junction, one can depict a round echogenic spot in transverse section, which can be differentiated from its surroundings by its strong echogenicity and possibly constant movement of microbubbles [28]. When the reflux reaches the kidney the echogenic microbubbles can be detected in the pelvicalyceal system. In a recent study by Kopac et al. [23] the need for UCA to demonstrate reflux was again emphasized. They compared indirect non-contrast VUS with contrast-enhanced VUS in 47 children with 93 PUUs. In the former there was no catheterization and bladder filling and any increase in renal pelvic and proximal ureteric width during voiding was considered as a sign of VUR. Compared to VUS with UCA this was found to have low diagnostic accuracy of only 59%, which is not sufficiently reliable for routine use. Postcontrast scan of the bladder, ureters and kidneys during and after voiding The US examination is continued during and after voiding in a similar manner as above, always scanning the right and left kidneys alternately, and also the bladder if the patient’s position allows. If cyclic filling is not planned, the catheter can be removed prior to micturition. Most children can void around a thin catheter on the examination table while lying supine, prone or on the side [28]. Other alternatives to consider are to have the child sit on a potty and to scan the kidneys from the back. A “music potty” that signals voiding with music is helpful. In older boys the option to micturate into a urine bottle while standing and being scanned from the back may be offered. At the end the bladder is checked for residual urine. Postcontrast scan of the urethra during voiding (urethrosonography) The main focus in VUS has been on detection of reflux. In the past, transperineal imaging of the urethra was not given as much consideration during VUS. The lack of urethral imaging in VUS was regarded as a drawback compared to VCUG [46, 47]. In recent years there has been an increased interest in the inclusion of urethral imaging. To date, six studies are available that deal exclusively or partly with contrast-enhanced voiding US of the urethra [16, 19, 48–51]. In these studies the urethra was examined in a total of 847 children comprising 647 (76%) boys and 200 (24%) girls (age range 1 day to 15 years). Levovist in combination with fundamental imaging was used in all children. In half of the studies, both boys and girls were included and in the remainder only boys [19, 48, 49]. An unbiased comparison of all cases with VCUG was carried out in only two studies [48, 49]. In three studies just those children with pathological findings of the urethra by voiding urethrosonography underwent VCUG [19, 50, 51]. In the remaining one study the selection for comparison was not clearly presented [16]. The diagnosis of posterior urethral valve was made in 22 boys (3.3%) using transperineal contrast-enhanced voiding urethrosonography (Fig. 5). This was also found to be suitable for assessment of the urethra after valve resection [48, 49]. Other diagnoses in boys were anterior urethral valve and urethral stenosis [48]. All diagnoses were confirmed by VCUG. However, the urethral pathologies detected were not only too few, but also the range of pathologies encountered was small [47]. Other urethral pathologies such as paraurethral cysts, diverticula, double urethra, urethral fistula, and complex anorectal and cloacal malformations are still missing in these series. Urethrovaginal reflux was evaluated in two studies and was found to be present in 67 of 165 girls (40.6%) [19, 49]. Fig. 5Transperineal voiding urethrosonography (a) as part of VUS in comparison with (b) VCUG. To facilitate the comparison the US image (a) is presented upside down. Note in the transperineal US (a) the microbubbles in the bladder (B) and in the massively dilated posterior urethra (pU). The anterior urethra (aU) is depicted as very thin in the presence of a posterior urethral valve (arrow). The finding was confirmed on VCUG (b) (courtesy of Dr. M. Bosio, Milan, Italy) The most important study comparing contrast-enhanced voiding urethrosonography and VCUG is that by Berrocal et al. [49]. In this prospective study a total of 146 children (87 boys, 59 girls) with a mean age of 3.3 years (8 days to 14 years) were recruited. In the voiding part of VUS, primarily transperineal US of the urethra was performed. During voiding the catheter was removed. The diameters of the anterior and posterior parts of the urethra were measured at maximum dilatation during voiding. In boys the normal value was found to be 6.1 ± 0.8 mm (2.8–7.1 mm) and 6.3 ± 0.67 mm (3.7–7.2 mm), respectively. In girls the urethral diameter was 4.2 ± 1.0 mm (2.5–7.8 mm). This was the basis for evaluation of pathological changes in the urethra. In all children the US study was followed by VCUG. There were three boys with posterior urethral valves detected on US with dilated posterior urethra, reduced diameter of the anterior urethra and delay in UCA flow at the valve. Furthermore, urethral stenosis was diagnosed in two boys. Seven children were evaluated after resection of the posterior urethra valves and were found to have a normal posterior urethral diameter in the presence of dilatation of the anterior part without any flow delay. In all children these findings were confirmed on VCUG. All girls and 75 boys showed a normal urethra at both transperineal US and VCUG. Thus sensitivity and specificity were each 100%. The approximate mean duration for VUS including transperineal voiding US was 30 min. In conclusion, transperineal contrast-enhanced VUS of the urethra has not only been shown to be possible but also to be a potential adjunct to VUS in routine examinations. Cyclical filling of the bladder VUR is an intermittent phenomenon. Differences in presence and degree of reflux on repeated examinations have been reported in both VCUG and RNC [52, 53]. In general, cyclical filling results in an increase in reflux detection rate. Repeated filling of the bladder may be necessary when there is marked discrepancy in kidney size and/or intermittent dilatation of the ureters on US and a negative result during the initial examination. VUS has a specific advantage in this regard as it is not associated with repeated exposure to radiation. Three studies have evaluated the value of cyclical filling in VUS [54–56]. In two of the studies, two cycles of bladder filling including Levovist administration were carried out with scanning in fundamental mode [54, 55]. One study was performed in 27 patients with transplanted kidneys [54]. This group comprised eight children (age range 12.2–18.3 years) and 19 adults (age range 19.2–54.8 years). The reflux detection rate in the first and second cycles was 17 and 16 of 27, respectively, and thus not very different, but there were significantly higher reflux grades in the second cycle. So the conclusion was that cyclical filling in VUS with fundamental imaging did not result in increased reflux detection rate. In another study with a greater number and wider spectrum of patients, exclusively children, the opposite was found. Novljan et al. [55] evaluated cyclical filling in 49 children (age range 1.4–15.8 years, mean 4.1 years) with 98 PUUs. A total of 35 PUUs with reflux were detected when taking both cycles into account. In the first cycle 28 of 35 and in the second 33 of 35 of the PUUs were positive. In the first cycle only just 2 of 35 refluxes (both grade II) were detected. The VUR diagnosed in the second cycle alone comprised six PUUs with grade II and one PUU with grade III reflux. Not only were 25% more refluxing units detected in the second cycle but also 50% more grade III refluxes. When the bladder is completely emptied one can often still find residual echogenic UCA lining the bladder mucosa which together with the hypoechoic wall resembles the mouth part of a “smile sign” (Fig. 6). This so-called “smile sign” indicates that there is still enough UCA in the bladder to facilitate repeat filling with normal saline alone and allow another cycle of VUS. Papadopoulou et al. [56] found that in 112 of 117 children (96%) undergoing cyclical harmonic VUS the residual UCA in the bladder was sufficient to allow a second filling with normal saline only. In the 112 children (224 PUUs) they compared the results of the first cycle with Levovist and normal saline administration with those of a second cycle with administration of only normal saline. VUR was detected in 57 PUUs in the first cycle, and of the remaining 167 nonrefluxing PUUs, 12 showed VUR in the second cycle (one grade I, nine grade II, two grade III). Only one PUU reflux (grade II) was diagnosed in the first cycle but missed in the second cycle. Considering all refluxing units as true positives, the sensitivities of the first and second cycles were 87% and 98% and the negative predictive values 84% and 98%, respectively. Thus a second cycle of harmonic VUS with no addition of UCA reveals significantly more PUUs with VUR at almost no additional cost for the examination. From the current standpoint of available data, cyclical filling in VUS, particularly when using harmonic imaging, seems to be a very promising adjunct. Fig. 6At the end of VUS and bladder emptying there can still be a “layer” of UCA lining the mucosa of the empty bladder. The configuration of the bladder in transverse section in combination with the remaining echogenic UCA resembles the mouth part of a “smile sign”. This sign indicates that there is still sufficient UCA in the bladder and if considered necessary a second filling with normal saline only would suffice to carry out cyclical VUS Grading of reflux VUR grading is necessary because the severity of reflux correlates with prognosis and consequently serves as a basis for therapeutic decision making. It is also essential to have a uniform grading system for comparison between different research studies. The reflux gradings are based on the degree of pelvicalyceal and ureteral dilatation. Intrarenal reflux is not taken into consideration for grading purposes. It is to be noted that it is possible to depict intrarenal reflux in VUS, particularly when using harmonic imaging and other high-MI modalities [13, 57]. No systematic evaluation of intrarenal reflux in VUS is available. The first grading system for VUS was proposed in 1985 by Beyer et al. [58]. Here grading was for VUS with administration of normal saline only into the bladder. This sonographic grading has five levels based on the extent of dilatation of the renal pelvis, calyces and ureters. Taking this grading system and in conformity with the international reflux grading for VCUG, a five-level grading system was also adapted for contrast-enhanced VUS [59, 60]. The diagnosis of reflux is based on the presence of microbubbles and the severity of the reflux determined by taking primarily the pelvicalyceal and secondarily the ureteric dilatation into account (Table 1). This grading system for VUS has gained widespread acceptance. With the severity of reflux being a continuum, no sonographic measurements have been proposed for differentiating between the grades. Unavoidably, as in VCUG, there is a certain element of subjectivity in the grading system and also some reflux episodes do not fall precisely within one of the five grades. The attempt to differentiate between reflux in dilated and nondilated systems has not become widely used, probably due to lack of immediate consequence of such an addition [60]. Table 1Reflux grading in contrast-enhanced VUSGradeDefinitionIMicrobubbles only in the ureterIIMicrobubbles in the renal pelvis; no significant renal pelvic dilatationIIIMicrobubbles in the renal pelvis + significant renal pelvic dilatation + moderate calyceal dilatationIVMicrobubbles in the renal pelvis + significant renal pelvic dilatation + significant calyceal dilatationVMicrobubbles in the renal pelvis + significant renal pelvic dilatation and calyceal dilatation + loss of renal pelvis contour + dilated tortuous ureters Duration of VUS examination In four comparative studies, in which VUS and VCUG were carried out successively, the duration of each procedural step was recorded [12, 28, 31, 42]. In two of the studies VUS was carried out using fundamental mode [12, 28]. The other two studies incorporated colour Doppler US [31, 42]. In these studies the durations of VUS, including catheterization, were in the range 26–34 min with the precontrast US taking up almost one-third of the time. The durations of VCUG, adding the time for catheterization, were in the range 13–20 min. In one other study VUS and VCUG were carried out simultaneously and here the duration of one such examination including catheterization was 9 min [61]. Novljan et al. [55] performed VUS with cyclical filling (two fillings) and the mean duration of the VUS was 24 min (range 15–30 min). Overall, these studies show that the duration of VUS shows a wide range and is longer than that of VCUG. As mentioned above the precontrast US may not be necessary when using harmonic or other dedicated contrast imaging modalities, resulting in a reduction of the duration of VUS by almost one-third. This would put the duration of VUS equivalent to that of VCUG. VUS with a second-generation US contrast agent (SonoVue) In 2001 a second-generation UCA—SonoVue (Bracco, Milan, Italy)—was approved in the European Union for IV use in adults. SonoVue is now widely used, the main application being for the evaluation parenchymal abdominal lesions [62]. Even though it has not been approved yet for use in children one in-vitro and four clinical studies point out potential advantages compared to the use of the first-generation UCA, Levovist [63–67]. SonoVue is composed of a stabilized aqueous suspension of sulphur hexafluoride (SF6) microbubbles with a phospholipid shell and is available as a package comprising one vial with granules and a 5-ml syringe prefilled with normal saline [62]. In an in-vitro comparative study with Levovist (5% volume) the contrast duration with SonoVue was seven times longer at a dose that was 80% lower [63]. It was found that UCA at 0.25–1.0% of the filling volume would provide adequate contrast. Moreover, the contrast duration of a freshly prepared suspension of SonoVue was stable over 6 h, whereas Levovist showed a significant reduction after 30 min. The first clinical studies [64–67], in total comprising 210 children, have shown that low-MI imaging is the most favourable US modality for VUS when using SonoVue [67] (Fig. 7). Although it is possible to use all other modalities with this UCA, the highest contrast difference between tissue and microbubbles seems to be achieved with low-MI imaging, in which the tissue is suppressed and the microbubbles become more conspicuous. During intravesical administration of SonoVue there is a minor difference compared with Levovist. With the patient in the supine position at the beginning of injection the UCA accumulates at the roof of the bladder and thus does not reach the vesicoureteric junctions [65]. Only after starting normal saline infusion can a homogeneous distribution of the microbubbles be observed. Not much acoustic shadowing of the retrovesical space is encountered. The intravesical dosages that have been successfully used in clinical studies are 1% of bladder filling [65] and 1 ml per bladder filling [67]. This implies that with a 5-ml suspension from one vial that is stable over 6 h there is the potential for performing several studies from the one vial. If one vial can be used for more than one patient, the cost of the UCA, a major obstacle to the widespread use of VUS, could be reduced. It is important to note that in some countries there may be restrictions regarding the use of one vial for more than one patient. Fig. 7VUS using the second-generation UCA, SonoVue. The scan before contrast agent administration (a) uses the tissue harmonic imaging (THI) modality and the scan after contrast agent administration (b) uses echo contrast imaging (ECI). In the former the MI (arrow) is high (1.1) while in the latter a low-MI (0.2) technique is applied. In THI the kidney and particularly the renal pelvis are very well depicted. The refluxing microbubbles of the second-generation UCA are displayed most conspicuously using the low-MI imaging modality Contraindications and adverse events from intravesical administration of US contrast agents There are no specific contraindications as such to the intravesical administration of Levovist. There are five clinical studies including a total of 626 patients with an age range of 2 days to 20 years, in which systematic monitoring for possible adverse events related to intravesical administration of Levovist were carried out [12, 20, 28, 41, 42]. The evaluation incorporated all or some of the following: various levels of vital sign monitoring [12, 28], assessment for possible signs and symptoms during and after the procedure, observation for up to 12 h as inpatient [42], and request to the patient and parents to report any symptoms in the next 24 h and an active 24-h follow-up by phone. This evaluation was carried out within the context of comparative studies, in which VCUG was performed following VUS. All patients were catheterized for the examinations. There were 15 patients with transient visible haematuria during or at completion of voiding at either VUS or VCUG and one patient each with transient mild abdominal and urethral pain. For these observed adverse events the catheterization was primarily to blame rather than the UCA [68]. No substance-specific adverse events definitely related to the intravesical administration of Levovist were observed. The osmolality of Levovist in solution is about five times higher than that of blood [37]. This osmolality corresponds to that of previously used ionic radiographic contrast agents, which were likewise instilled into the bladder. No adverse events attributable to the osmolality were observed with these agents. Apart from this, the amount of UCA administered is small and in the bladder there is usually urine and/or normal saline, resulting in dilution of the hyperosmolar solution. Hyperosmolality of the UCA can, therefore, be ignored in VUS, particularly as it lies approximately in the physiological range of osmolality of urine. No animal study has been published on the potential effect of insonation of the urothelium in the presence of microbubbles, particularly at high MI settings. Potential alterations of the urothelium cannot be excluded with absolute certainty. Over the last 10 years, predominantly in Europe, it is roughly estimated that some 20,000 VUS examinations have been conducted in children (publications, reports and personal communications). There have been no reports of clinical adverse effects related to the intravesical administration of Levovist. Experience with the new UCA, SonoVue, for VUS is still very limited. Contraindications for the IV administration are known hypersensitivity to sulphur hexafluoride and certain cardiopulmonary disorders [29]. Recently, a large-scale retrospective analysis showed that IV SonoVue has a good safety profile in abdominal applications [69]. In the small number of VUS studies with a total of 210 children no adverse events directly related to the UCA have been reported [64–67]. There are, as yet, no detailed studies regarding any possible interaction with urine and the urothelium. Conclusion It is important to realize that a breakthrough to implement US for reflux diagnosis, which had began in the late 1970s, only came with the availability of a stable UCA. Advances in US technology resulting in marked improvement in the depiction of microbubbles have facilitated many procedural aspects in addition to the diagnostic accuracy of the modality. Furthermore, the first results of VUS with the use of a more stable second-generation UCA point to further procedural and diagnostic progress. Like VCUG and direct RNC, it is still necessary to catheterize the bladder or perform suprapubic puncture for VUS. This is a major drawback of all direct reflux examinations and one that makes them invasive. Attempts to generate bubbles exogenously using high-intensity focused US would solve this problem, but have not yet moved beyond the stage of animal studies [70].	[ "voiding urosonography", "ultrasound contrast agent", "vesicoureteric reflux" ]	[ "P", "P", "P" ]
Acta_Neuropathol-3-1-2080348	Gray matter injury associated with periventricular leukomalacia in the premature infant	Neuroimaging studies indicate reduced volumes of certain gray matter regions in survivors of prematurity with periventricular leukomalacia (PVL). We hypothesized that subacute and/or chronic gray matter lesions are increased in incidence and severity in PVL cases compared to non-PVL cases at autopsy. Forty-one cases of premature infants were divided based on cerebral white matter histology: PVL (n = 17) with cerebral white matter gliosis and focal periventricular necrosis; diffuse white matter gliosis (DWMG) (n = 17) without necrosis; and “ Introduction Periventricular leukomalacia (PVL), a major disorder of the immature cerebral white matter, has long been considered the underlying neuropathologic substrate of cerebral palsy in premature infants who survive into childhood [41]. The substrate of the cognitive impairments in these children, however, is less certain, given that cognition is typically attributed to gray matter (neuronal cell body), as opposed to white matter (oligodendrocyte), function, and the incidence of gray matter injury in PVL has historically been considered minimal [1, 4, 8]. Understanding the neuroanatomic basis of impaired cognition is of increasing importance in the care of premature infants: currently, in extremely premature infants (born at <1,500 g), the most common neurological disabilities involve cognition and learning, affecting 20–50% of such infants, compared to cerebral palsy which affects 10% [39, 42]. Attempts have been made to correlate neuropathologic findings in autopsied brains with neurologic sequelae in survivors of premature birth, however, much of this work has focused on the role of white matter injury and hemorrhages and not gray matter injury [13]. In the seminal paper of the neuropathology of PVL in 1969, Banker and Larroche reported only “mild neuronal injury” in the cerebral cortex, hippocampus, subiculum, basis pontis, and cerebellar dentate nucleus and Purkinje cells in the majority of PVL brains [4]. Subsequent neuropathologic studies reported only rare “anoxic neuronal injury” in association with PVL [1, 8], or downplayed the importance of observed gray matter lesions and did not emphasize their potential importance to neurologic sequelae in survivors. Thus, the conventional teaching has been that cerebral white matter is especially vulnerable to hypoxia-ischemia in the premature brain, with relative sparing of the gray matter, and that gray matter injury dominates only in older infants, children, and adults [23, 24], even though recent evidence shows that neuronal/axonal injury is common in the perinatal brain [5, 25]. This teaching is increasingly challenged by modern quantitative volumetric MRI studies of premature infants who exhibit reduced volumes of the cerebral cortex [35], thalamus [17, 18], basal ganglia [17, 18], and hippocampus [19, 34]. The anatomic substrate for the gray matter volumetric deficits associated with the apparent white matter disease remains unknown. In the following study of the neuropathology of premature infants autopsied at this institution in the modern era of neonatal intensive care, we aimed to determine: (1) whether gray matter abnormalities in premature infants are more common in the presence of PVL than in the absence of this lesion; and (2) whether the constellation of any gray matter abnormalities provides insight into the basis of the cognitive impairments in living premature infants. In this study, we stratified the cases according to three patterns of cerebral white matter histology: (1) PVL (n = 17), defined as diffuse cerebral white matter gliosis combined with focal (macro- and/or microscopic) periventricular necrosis and; (2) diffuse white matter gliosis (DWMG) (n = 17), defined as diffuse gliosis without focal necrosis; and (3) “Negative” (n = 7), without diffuse gliosis and focal necrosis. While the pathogenesis of the white matter gliosis in the DWMG group is unknown, we considered it in a separate category from PVL (gliosis combined with focal necrosis) because the cerebral white matter pathology is distinct, without macro- and/or microscopic periventricular cysts. In the following study, we analyzed the three groups independently of each other. We hypothesized that PVL cases have a significantly greater incidence and degree of gray matter injury than non-PVL, i.e., DWMG and Negative cases; that this injury involves structures critical for cognition and learning, i.e., deep gray nuclei, cerebral cortex, and hippocampus; and that the pattern of gray matter injury mimics the pattern of volume reduction in the deep gray nuclei and cerebral cortex detected by neuroimaging studies in long-term survivors. Materials and methods Case selection criteria The neuropathology of all premature infants (<37 gestational weeks at birth) autopsied between 1997–1999 at Children’s Hospital Boston was retrospectively reviewed. This time-frame was selected because it represents a modern era of intensive care management of premature infants in a Level 3 neonatal intensive care nursery. It also represents a period at our hospital when the brain and spinal cord were extensively sampled in a relatively standardized fashion, whether or not macroscopic lesions were apparent, thereby permitting a systematic neuropathologic survey in a large dataset. Parental authorization of the use of autopsy human tissue for research was given in each case. Microscopic slide review A median of 15 (range 9–20) hematoxylin–eosin (H&E) or H&E/Luxol-fast-blue stained sections was examined from each case. These sections included cerebral cortex from all lobes, thalamus (at the level of lateral geniculate nucleus and including the dorsomedial and lateral posterior nuclei), hypothalamus, caudate, putamen, globus pallidus, hippocampus (level of lateral geniculate nucleus), amygdala, cerebellar dentate nucleus, cerebellar cortex (including Purkinje cells and granule cell layer), midbrain, pons, and medulla. We scored the density of neuronal necrosis, neuronal loss, and gliosis in the most severely affected high power fields of each of the gray matter sites. Two of three neuropathologists (CRP, RDF, HCK) reviewed the slides together at any one time, with a three-way consensus achieved on difficult cases. The brainstem tegmentum was graded separately from the basis pontis and inferior olive in recognition of their susceptibility to injury in the perinatal period [23]. We defined neuronal necrosis as hypereosinophilic neurons with pyknotic nuclei, or, in cases with neuronal immaturity and scant cytoplasm, karyorrhexis [2]. We used the following scale: 0, no necrosis; 1, a few scattered necrotic neurons in a high-powered field (hpf; 400×); 2, isolated clusters of necrotic neurons/hpf; and 3, larger, confluent areas of necrotic neurons/hpf. We scored neuronal loss as: 0, no neuronal loss/hpf; 1, mild, scattered neuronal dropout/hpf; 2, moderate, focal areas of neuronal dropout/hpf; and 3, severe, confluent areas of neuronal dropout/hpf. We interpreted neuronal necrosis as an acute or agonal change, occurring within 24–48 h of death [2], while neuronal loss and gliosis were interpreted as markers of subacute or chronic injury, indicative of insult occurring 3–5 days or more prior to death [2]. We combined cases for analysis that had grades 2/3 and 3/3 of neuronal loss because these grades are unequivocally recognized by standard microscopic examination, and therefore represent, in our opinion, a substantial degree of injury. We scored gliosis as: 0, no reactive astrocytes/hpf; 1, 1–10 reactive astrocytes/hpf; 2, 11–20 reactive astrocytes/hpf; and 3, >20 reactive astrocytes/hpf. Reactive astrocytes were defined as stellate configured cells with abundant (“hypertrophic”) eosinophilic cytoplasm and an enlarged, often eccentrically placed nucleus with delicate chromatin. The cytoplasm of these cells was immunopositive for glial fibrillary acidic protein (GFAP), the well-established astrocytic marker. In scoring neuronal necrosis, neuronal loss and gliosis, we evaluated the entire available region of each gray matter structure in the section, and assessed the most severely injured region, which was virtually always representative of all fields. We assessed white matter in the cerebral lobes, corpus callosum, posterior limb of the internal capsule, and cerebellum for PVL and diffuse white matter gliosis (DWMG). PVL is defined by the combined presence of: (1) focal necrosis in the periventricular region; and (2) diffuse reactive gliosis in the surrounding white matter [24]. DWMG is defined by the presence of gliosis in the cerebral white matter unaccompanied by periventricular foci of necrosis [24]. Both PVL and DWMG are characterized by reactive astrocytes throughout the white matter, and are thus associated with a pattern of “diffuse” white matter injury. The density of white matter gliosis was scored according to the same scale used for gray matter gliosis (grades 0–3). Glial fibrillary acidic protein (GFAP) immunohistochemistry Four-micron thick formalin-fixed paraffin-embedded sections of frontal cortex could be cut from 27 of the 41 cases and were immunostained with mouse anti-GFAP antibody (1:500, #SM1-22R, Covance, Berkeley, CA). Negative controls were performed without primary antibody. Scoring of GFAP stained sections was performed by counting positive cells/hpf, in the most intensely immunopositive region of frontal cortex after a survey of all fields. Reactive astrocytes were defined as those cells with substantial cytoplasmic GFAP staining around a nucleus. Other non-reactive cortical astrocytes with limited cytoplasmic staining had GFAP positive processes that were generally perpendicular to the glial limitans were counted separately. The grading system was: 0, no staining; 1, 1–10 cells/hpf; 2, 11–20 cells/hpf; and 3 > 20 cells/hpf. Two observers (CRP, HCK) scored each case without knowledge of the white matter group. O4 and GFAP double-labeling immunofluorescence Myelination gliosis is commonly encountered in the newborn brain and must be discerned from reactive gliosis. So-called myelination glia are oligodendrocyte precursors that form during myelination; these cells have large nuclei with chromatin intermediate in density between oligodendrocytes and astrocytes and large slightly basophilic cell bodies so they can potentially be mistaken for reactive astrocytes [36]. To help discern reactive gliosis from myelination gliosis fresh tissue that was immediately fixed in 4% paraformaldehyde and sectioned at 40–50 μm was available from 4 PVL, 1 DWMG and 4 of the Negative cases. Double labeling was performed sequentially beginning with the mouse anti-O4 monoclonal antibody (1:750; gift from Dr. Steven Pfeiffer) to detect developing oligodendrocytes and followed with rabbit anti-GFAP antibodies (1:200, Z0334, Dako) to detect astrocytic differentiation. Relevant secondary antibodies conjugated with FITC or Texas Red were used and sections were visualized with Nikon Eclipse E800 microscope (Nikon, Melville, NY) outfitted with Spot image capture software (Diagnostics Instruments Incorporated, Sterling Heights, MI). Statistical analysis The 41 cases were divided into three groups according to cerebral white matter histology: (1) a PVL group; (2) a DWMG group; and (3) a “Negative” white matter group with no diffuse gliosis or focal periventricular necrosis in the cerebral white matter [23, 24]. Demographic characteristics were compared between the PVL, DWMG and Negative groups using Wilcoxon rank sum tests for continuous variables and χ2 tests for categorical variables (Table 1). The number of gray matter sites involved with an injury, i.e., acute neuronal necrosis, neuronal loss, and gliosis, was counted for each case in each group, and plotted relative to postconceptional age (i.e., gestational age plus postnatal age, PCA; in weeks) to graphically depict the total number of gray matter sites that were injured in each case. To test the hypothesis that the incidence and severity (grade 2–3/3) of neuronal necrosis, neuronal loss, and gliosis in all gray matter sites analyzed varies significantly among PVL, DWMG, and Negative cases, Fisher exact tests were used. To control for the potential impact of age on these analyses, analysis of covariance of diagnosis on lesion severity were performed, controlling for PCA. Nonsignificant interaction effects between diagnosis and age were subsequently eliminated from the models. In all analyses, P < 0.05 was considered significant. Table 1Clinicopathologic variables of the three white matter study groups for comparison of gray matter injuriesMean ± SD; median, (range) or percentP valuePVL group n = 17DWMG group n = 17Negative group n = 7Three-wayPVL versus DWMGDemographics Gestational age (weeks)32.8 ± 3.1; 34, (26–36)31.6 ± 3.8; 33, (24–36)26.5 ± 2.3; 27, (23–30)0.0030.282 Postnatal age (weeks)3.7 ± 4.1; 2.3, (0.1–15)3.4 ± 4.0; 1.2, (0.1–12.0)0.8 ± 1.2; 0.1, (0.1–3.0)0.141 Postconceptional age (weeks)36.5 ± 5.4; 35.5, (26–52)34.8 ± 4.5; 35.0, (28.4–48.0)27.3 ± 2.8; 28.3, (23.1–30.3)0.0010.293 Length of ICU stay (days)13.8 ± 15.9; 9.5, (0–56) 23.8 ± 34.6; 5.0, (1–96) 5.0 ± 8.9; 1.5, (0.04–23.0) 0.419 Race: Caucasian8/17; 47%9/17; 53%5/7; 71%0.801 African-American2/17; 12%2/17; 12%1/7; 14% Unknown7/17; 41%6/17; 35%1/7; 14% Percent male9/17, 53%7/17, 41%4/7, 57%0.702Postmortem interval (h)16.0 ± 7.7; 18, (2–30)23.4 ± 19.6; 17, (2–72) 36.3 ± 42.4; 22, (16–132)0.364Birth weight (g)1576 ± 717; 1380, (780–2700)1832 ± 1048; 2100, (610–3800)925 ± 408; 930, (440–1400)0.232Brain weight (g)245.0 ± 81.1; 257.5, (110–340)255.0 ± 72.1; 245, (124–380) 127.6 ± 40.7; 140.0, (63.8–178.0)0.0030.905Body weight (g)2423.1 ± 1165.0; 2225, (760–4875)3276.2 ± 4706.0; 2290, (1150–21300)947.9 ± 359.2; 915, (440–1400)0.0020.815Body length (cm)42.0 ± 7.8; 41.5, (26–56)43.2 ± 4.4; 44.0, (36–51)34.7 ± 4.0; 32.5, (30–41)0.0070.589Twin or other multiple gestation3/17, 18%4/17, 24%2/7, 29%0.824Multiple congenital anomaliesa4/17, 24%2/17, 12%3/7, 43%0.242Cesarean section10/17, 59%12/17, 71%3/7, 43%0.436Cardiorespiratory factorsApgar score at 1 min4.1 ± 2.4; 4, (1–8) 4.7 ± 2.9; 5, (0–9) 3.5 ± 2.2; 3, (1–7) 0.691Apgar score at 5 min5.2 ± 2.9; 6, (1–9) 6.6 ± 2.9; 7, (0–10) 5.0 ± 2.8; 5.5, (1–8) 0.405Acute respiratory distress syndrome7/17, 41%10/17, 59%5/7, 71%0.344Mechanical ventilation13/17, 77%13/17, 77%7/7, 100%0.359Length of mechanical ventilation (days)8.4 ± 11.6; 3.0, (0–38)3.5 ± 6.3; 1.0, (1–19) 4.4 ± 8.2; 1.0, (0.04–23.0) 0.383Extracorpeal membrane oxygenation1/17%, 6%0/17%, 0%0/7, 0%0.485Continuous positive airway pressure0/17, 0%3/17, 18%1/7, 14%0.202Cardiopulmonary resuscitation any time during hospitalization5/17, 29%7/17, 41%5/7, 71%0.165Congenital heart diseasea2/17, 12%2/17, 12%1/7, 14%0.983Infectious/inflammatory factorsPneumoniaa4/17, 24%2/17, 12%0/7, 0%0.303Maternal fever at delivery0/15, 0%2/9, 22%0/4, 0%0.103Maternal history of urinary tract infection 0/15, 0%0/10, 0%1/4, 25%0.039NAChorioamnionitis1/15, 7%2/10, 20%1/5, 20%0.562Necrotizing enterocolitisa1/17, 6%4/17, 24%2/7, 29%0.415Clinical diagnosis of sepsis8/17, 47%8/17, 47%3/7, 43%0.980If a three-way P value was significant (P < 0.05), a two-way P value between the PVL and DWMG groups was performed. NA, not applicable since the incidence was 0 in the PVL and DWMG groups NS, not significanta diagnosed at autopsy Results Clinical and autopsy data Seventeen cases (41%) fulfilled the criteria for PVL, while 17 cases (41%) had DWMG, and there were 7 so-called Negative cases (17%). Pregnancy, labor and delivery The Negative group was significantly younger in terms of gestational age (GA) than the PVL and DWMG groups, but there was no significant difference in GA between the PVL and DWMG groups (Table 1). In terms of PCA, the PVL and DWMG groups were significantly older than the Negative group (Table 1). The Negative group had significantly lower brain weight and body weight, and shorter body length, but not birth weight, than the PVL and DWMG groups. The low somatic and brain measurements in the Negative group compared to the PVL and DWMG groups reflect the early gestational age at birth and younger postnatal age at death (Table 1). The incidence of various clinical variables, e.g., chorioamnionitis, maternal fever at delivery, history of Cesarean section, and congenital anomalies were not significantly different among the three groups. The mean Apgar scores were less than seven at 1 and 5 min in all three groups (Table 1). Syndromes Twenty-four percent of cases had a constellation of findings classified as a genetic/developmental syndrome, e.g., Treacher-Collins syndrome, Potter’s sequence, and osteogenesis imperfecta in three PVL cases; Down’s syndrome in a DWMG case; and Fryns syndrome, Beckwith-Wiedemann syndrome, and VACTERL association in three Negative cases. Excluding these cases from the analysis had no significant effect on the results for the different acquired lesions analyzed semi-quantitatively, i.e neuronal loss and gliosis (data not shown), and thus, their data were combined with that of the non-syndromic cases in the complete analysis reported below. Postnatal period Although there were no statistically significant differences in postnatal age (PNA) among the three groups, the median PNA was only 0.1 week in the Negative group in contrast to 2.3 weeks in the PVL group and 1.2 weeks in the DWMG group (Table 1). The younger PNA and significantly different PCA and GA lead us to limit our gray matter comparisons to those between the PVL and DWMG groups, although the data from the Negative group is reported in all of the tables for completeness. Cardiorespiratory disorders were common in all three groups, with no significant differences (Table 1). Acute respiratory distress syndrome was noted in 41% of PVL cases, 59% of DWMG cases, and 71% of Negative cases. Seventy-seven percent of PVL and DWMG cases, and 100% of the Negative cases required ventilation (Table 1). The duration of ventilation and incidence of cardiopulmonary resuscitation was not significantly different among the three groups. Infectious and inflammatory disorders e.g., pneumonia, necrotizing enterocolitis, and sepsis, occurred in all groups, and did not differ significantly among the groups (Table 1). White matter findings Eighty-five percent of all cases studied had diffuse gliosis in the cerebellum, and 82% in the cerebral hemispheres. Forty-one percent of the cases had PVL, and 41% had DWMG. Macroscopically evident periventricular cysts (<5 mm in diameter) were noted in one PVL case, while chalky-white necrotic foci (2–3 mm) were visible in two cases. In all of the other PVL cases (82%), necrotic foci were only detected microscopically, and were <1 mm in diameter. Necrotic foci were typically found within a few millimeters from the ventricles. All cerebral lobes demonstrated a similar incidence of PVL (23–28%), except for the temporal lobe (12%). DWMG particularly involved the internal capsule (67%) and corpus callosum (64%) (Table 2). Necrotic foci were also identified in nonperiventricular regions, notably the internal capsule (19%) and corpus callosum (9%) (Table 2). The incidence of PVL increased with age, but only significantly so in the frontal lobe (Table 2). In contrast, the incidence of DWMG significantly increased in almost all sites with increasing gestational and postnatal ages (Table 2). By 37+ weeks, DWMG was present in 100% of the cases in the fronto-parieto-temporal lobes, 90% of the cases in the cerebellum, and 80–86% of the cases in the internal capsule and corpus callosum (Table 2). The degree of severity of DWMG significantly increased with both GA and PNA in all cases in the frontal, parietal, temporal, occipital, and cerebellar white matter (P < 0.04; data not shown). Moreover, the coefficients in the regression models for GA and PNA were similar, indicating that the degree of gliosis increased the same amount for each extra week of gestation and for each extra week of postnatal life (data not shown). In effect, the degree of gliosis increased constantly with PCA (gestational age plus postnatal age). Table 2Distribution of white matter lesions in 41 autopsied premature infants by anatomic site and postconceptional age (weeks)White matter siteOverall incidenceIncidence by postconceptional age (weeks) P value Logistic Regression of incidence and postconceptional weeks23–2930–3637+Frontal lobePVL11/40 (28%)0/8 (0%)6/22 (27%)5/10 (50%)0.025DWMG31/40 (78%)2/8 (25%)19/22 (86%)10/10 (100%)0.004Temporal lobePVL4/33 (12%)0/6 (0%)3/22 (14%)1/5 (20%)NSDWMG22/33 (67%)1/6 (17%)16/22 (73%)5/5 (100%)0.006Parietal lobePVL7/28 (25%)0/6 (0%)6/16 (38%)1/6 (17%)NSDWMG20/28 (71%)1/6 (17%)13/16 (82%)6/6 (100%)0.010Occipital lobePVL8/35 (23%)1/5 (20%)5/23 (22%)2/7 (29%)NSDWMG28/35 (80%)3/5 (60%)19/23 (83%)6/7 (86%)NSCorpus callosumPVL2/22 (9%)0/3 (0%)1/14 (7%)1/5 (20%)NSDWMG14/22 (64%)0/3 (0%)10/14 (71%)4/5 (80%)0.049Internal capsulePVL5/27 (19%)0/2 (0%)3/18 (17%)2/7 (29%)NSDWMG18/27 (67%)0/2 (0%)12/18 (67%)6/7 (86%)NSCerebellumPVL3/40 (8%)0/8 (0%)2/22 (9%)1/10 (10%)NSDWMG34/40 (85%)4/8 (50%)21/22 (95%)9/10 (90%)0.012P values denote significant differences in the incidence of DWMG with postconceptional age.NS, not significant; PVL, periventricular leukomalacia; DWMG, diffuse white matter gliosis in the cerebral and cerebellar hemisphere To exclude the possibility that myelination glia were scored along with reactive astrocytes in these infants immunofluoresence staining was performed to co-localize O4 a marker of the developing oligodendrocyte, and GFAP, an astrocytic marker. No co-localization of these proteins was identified in of the three groups of cases studied (Fig. 1). Fig. 1Immunoflourescence images of parieto-occipital white matter from a PVL case at 39 postconceptional weeks. O4 labeling (a) is shown in red and GFAP labeling (b) is shown in green. Merged images (c) show no co-localization of O4 and GFAP suggesting two distinct cell populations, i.e. GFAP-positive astrocytes and O4- positive oligodendrocyte precursors, are present. The scale bar represents 50 μm Gray matter lesions associated with PVL in the premature infant Acute neuronal necrosis Acute neuronal necrosis, which is considered a marker of terminal/agonal injury, was common, and occurred diffusely across gray matter regions in all three groups. Sixty-six percent of PVL, 59% of DWMG, and 43% of Negative cases had two or more gray matter sites with acute neuronal necrosis (data not shown). Significant differences in the incidence of acute neuronal necrosis between the three groups were noted only in the cerebellar cortex (53%, PVL; 13%, DWMG; 0%, Negative; P = 0.008) and frontal cortex (56%, PVL; 41%, DWMG; 0%, Negative; P = 0.039). The incidence of neuronal necrosis was not significantly different at any gray matter site when adjusted for GA and PNA (data not shown). Neuronal loss and gliosis Neuronal loss and gliosis, considered markers of subacute and chronic injury, were more prevalent and of greater severity in PVL cases compared to non-PVL cases (DWMG and Negative groups) (Figs. 2 and 3, Tables 3 and 4). PVL cases showed more damage to the deep nuclear structures than was encountered in non-PVL cases. In PVL cases, the thalamus and globus pallidus had significantly higher incidences of neuronal loss (38 and 33%, respectively) and more severe neuronal loss (38 and 33%, respectively) than did the DWMG and Negative groups (both, 0%) (Fig. 2, Table 3). The incidence of gliosis was also significantly higher in the thalamus (56%), caudate (60%), putamen (50%) and globus pallidus (60%) in PVL than in DWMG (12–47%) and Negative cases (0–14%). The cerebellar dentate nucleus showed a significantly higher incidence of neuronal loss in PVL (29%) compared to the DWMG (6%) and Negative (14%) groups. PVL cases (29%) had significantly more severe neuronal loss in the cerebellar dentate compared to the DWMG and Negative groups (both, 0%) (Table 3). Gliosis of the basis pontis was seen in 100% of PVL cases and only 79% of DWMG and 29% of Negative cases (P = 0.001; Table 4). The hippocampus also had substantial neuronal loss and gliosis (Tables 2 and 3). PVL cases showed relatively mild cerebral cortical neuronal loss, compared to other neuroanatomic sites, while the incidence of gliosis ranged from 20% (temporal cortex) to 31% (frontal cortex) (Tables 3 and 4). By contrast, the cerebral cortex in all lobes from DWMG and Negative cases was totally free of neuronal loss and was infrequently gliotic (all <10%) (Tables 3 and 4). Fig. 2Photomicrographs from thalami illustrating neuronal loss scores of 0 (a), 1 (b), 2 (c) and 3 (d). The asterisk in panel b denotes a focal area of neuronal loss. The scale bar represents 20 μmFig. 3Photomicrographs from the inferior olivary nuclei depicting gliosis scores of 0 (a), 1 (b), 2 (c) and 3 (d). Arrows in panels b and c indicate some of the reactive astrocytes that are present. The scale bar represents 20 μmTable 3Incidence and severity of neuronal loss in PVL, DWMG and Negative casesOverall incidenceIncidence of severity 2–3PVLDWMGNegativeP valuePVLDWMGNegativeP valueNeuronal lossCerebral cortexFrontal cortex13% (2/16)0% (0/17)0% (0/7)0.4776% (1/16)0% (0/17)0% (0/7)0.575Temporal cortex0% (0/15)0% (0/13)0% (0/6)1.0000% (0/15)0% (0/13)0% (0/6)1.000Parietal cortex8% (1/13)0% (0/11)0% (0/6)1.0008% (1/13)0% (0/11)0% (0/6)1.000Occipital Cortex0% (0/15)0% (0/16)0% (0/4)1.0000% (0/15)0% (0/16)0% (0/4)1.000Deep gray nucleiThalamus38% (6/16)0% (0/17)0% (0/7)0.00538% (6/16)0% (0/17)0% (0/7)0.005Hypothalamus20% (2/10)0% (0/10)0% (0/2)0.56710% (1/10)0% (0/10)0% (0/2)1.000Caudate13% (2/15)0% (0/16)0% (0/7)0.32913% (2/15)0% (0/16)0% (0/7)0.329Putamen13% (2/16)0% (0/17)0% (0/7)0.47713% (2/16)0% (0/17)0% (0/7)0.477Globus pallidus33% (5/15)0% (0/15)0% (0/6)0.02833% (5/15)0% (0/15)0% (0/6)0.028Cerebellum and relay nucleiBasis pontis21% (3/14)0% (0/14)0% (0/7)0.20614% (2/14)0% (0/14)0% (0/7)0.341Inferior olive15% (2/13)8% (1/13)20% (1/5)0.8078% (1/13)8% (1/13)20% (1/5)0.549Cerebellar cortex24% (4/17)6% (1/16)14% (1/7)0.44924% (4/17)6% (1/16)14% (1/7)0.449Dentate29% (4/14)0% (0/15)0% (0/6)0.03129% (4/14)0% (0/15)0% (0/6)0.031Limbic structuresHippocampus33% (5/13)0% (0/14)14% (1/7)0.05533% (5/15)0% (0/14)14% (1/7)0.055Amygdala0% (0/6)0% (0/3)0% (0/2)1.0000% (0/6)0% (0/3)0% (0/2)1.000Substantia inominata29% (2/7)0% (0/3)0% (0/1)1.00029% (2/7)0% (0/3)0% (0/1)1.000Brainstem14% (2/14)0% (0/14)0% (0/5)0.62914% (2/14)0% (0/14)0% (0/5)0.629PVL, periventricular leukomalacia; DWMG, diffuse white matter gliosis in the cerebral and cerebellar hemisphere. P values denote differences in the incidence or severity of neuronal loss at these neuroanatomic sites between PVL and DWMG groups with postconceptional ageTable 4Incidence and severity of gliosis in PVL, DWMG and Negative casesOverall incidenceIncidence of severity 2–3PVLDWMGNegativeP valuePVLDWMGNegativeP valueGliosisCerebral cortexFrontal cortex31% (5/16)6% (1/17)0% (0/7)0.10213% (2/16)0% (0/17)0% (0/7)0.477Temporal cortex20% (3/15)8% (1/13)0% (0/6)0.4950% (0/15)0% (0/13)0% (0/6)1.000Parietal cortex23% (3/13)9% (1/11)0% (0/6)0.4998% (1/13)9% (1/11)0% (0/6)1.000Occipital cortex27% (4/15)0% (0/16)0% (0/4)0.05413% (2/15)0% (0/16)0% (0/4)0.395Deep gray nucleiThalamus56% (9/16)18% (3/17)14% (1/7)0.03119% (3/16)0% (0/17)14% (1/7)0.161Hypothalamus40% (4/10)10% (1/10)50% (1/2)0.26420% (2/10)0% (0/10)50% (1/2)0.130Caudate60% (9/15)19% (3/16)14% (1/7)0.02813% (2/15)6% (1/16)14% (1/7)0.659Putamen50% (8/16)12% (2/17)14% (1/7)0.04419% (3/16)0% (0/17)0% (0/7)0.130Globus Pallidus60% (9/15)47% (7/15)0% (0/6)0.04020% (3/15)7% (1/15)0% (0/6)0.492Cerebellum and relay nucleiBasis pontis100% (14/14)79% (11/14)29% (2/7)0.00136% (5/14)21% (3/14)14% (1/7)0.684Inferior olive92% (12/13)92% (12/13)80% (4/5)0.54962% (8/13)54% (7/13)20% (1/5)0.400Cerebellar cortex29% (5/17)6% (1/16)14% (1/7)0.25912% (2/17)6% (1/16)0% (0/7)1.000Dentate43% (6/14)13% (2/15)17% (1/6)0.17721% (3/14)0% (0/15)0% (0/6)0.125Limbic structuresHippocampus47% (7/15)7% (1/14)29% (2/7)0.05620% (3/15)0% (0/14)29% (2/7)0.143Amygdala50% (3/6)0% (0/3)0% (0/2)0.32730% (0/6)0% (0/3)0% (0/2)1.000Substantia inominata29% (2/7)0% (0/3)0% (0/1)1.0000% (0/7)0% (0/3)0% (0/1)1.000Brainstem43% (6/14)20% (3/14)20% (1/20)0.5187% (1/14)0% (0/14)0% (0/5)1.000PVL, periventricular leukomalacia; DWMG, diffuse white matter gliosis in the cerebral and cerebellar hemisphere. P values denote differences in the incidence or severity of gliosis at these neuroanatomic sites between PVL and DWMG groups with postconceptional age Glial fibrillary acidic protein immunohistochemistry performed on sections of frontal cortex from PVL (n = 10), DWMG (n = 13) and Negative cases (n = 4) showed astrocytes of two different general morphologies (Fig. 4). Greater numbers of reactive astrocytes with abundant GFAP positive cytoplasm tended to occur in the frontal cortex of PVL cases (mean score 0.75 ± 0.22), compared to DWMG cases (0.48 ± 0.19) or (0.59 ± 0.38); however, when controlled for PCA this trend was not statistically significant. Non-reactive astrocytes were found in similar frequency among PVL, DWMG and Negative cases, and no such trend, as that noted for reactive astrocytes appeared (data not shown). Fig. 4GFAP immunohistochemical staining of frontal cortex illustrating non-reactive astrocytes (a) with a linear GFAP-positive process that is perpendicular to glial limitans, which is at the right of this image (not depicted) and a reactive astrocyte (b), with abundant GFAP-positive cytoplasm, and an eccentrically placed, enlarged nucleus. The scale bar represents 20 μm Discussion This neuropathologic analysis shows clearly that gray matter abnormalities are more common in the presence of PVL than in its absence (summarized in Fig. 5). Moreover, and remarkably, neuronal loss and gliosis in the cerebral cortex and deep nuclear structures are essentially confined to those infants with PVL. Thus, not a single infant with DWMG exhibited neuronal loss in the cerebral cortex, hippocampus, and deep gray nuclei. Similarly, gliosis was very unusual in these areas in the infants with DWMG. Although the incidence and severity of lesions in this autopsy series may not be completely representative of the brain pathology in premature infants who survive beyond the perinatal period, the findings are nevertheless important to understand the neuroanatomic substrate and pathogenesis of the neurological sequelae in long-term survivors. Fig. 5Summary diagram comparing gray matter sites with a significantly higher incidence (percentages) of neuronal loss (a) and gliosis (b) in PVL (right of panel) and DWMG (left of panel) cases. Gliosis of the cerebral and cerebellar white matter, basis pontis, brainstem tegmentum and inferior olives is depicted by small red dots, and focal, periventricular necrosis in the cerebral white matter (PVL) is denoted by a large red periventricular circle The pathogenesis of the gray matter lesions in PVL is likely due to the same phenomena implicated in the white matter lesion. The pathogenesis of PVL likely involves cerebral ischemia-reperfusion in the respiratory compromised preterm infant, in combination with one or more infectious/inflammatory and other, yet to be defined, derangements [24]. Thus, the topographic patterns of gray and white matter damage in the premature brain likely reflect a complex interplay of the differential vulnerabilities of the regions to glutamate, free radical, and cytokine toxicity. These differential vulnerabilities appear to be based upon the maturational stage of neurons and oligodendrocytes [3, 23, 40], i.e., the targeted cell types in gray and white matter injury, respectively, and upon the developmental profiles of glutamate and cytokine receptors [9, 23] and antioxidant systems [9], as well as multiple related factors [23, 40]. The term “perinatal panencephalopathy” best describes, in our opinion, the combined gray and white matter injury delineated in this study that is typical of perinatal neuropathology of prematurity. We regard PVL as a major part of this disorder that should now be considered, we believe, in the context of total brain injury. Since the majority of patients in the PVL, DWMG, and Negative groups required mechanical ventilation of comparable durations, and showed substantial involvement by inflammatory/infectious processes, it is difficult to decipher the factors responsible for the substantial brain injury in the PVL group. It is very likely that there are specific clinical factors at work that we do not yet know, or were not analyzed in this study, e.g., lowest oxygen levels, alterations in acid-base status, and dysfunction in cerebral autoregulation, which are difficult to analyze in a meaningful way from complicated neonatal records. Thus, this study is hypothesis-generating for a prospective analysis of the key clinical factors involved in the pathogenesis of perinatal panencephalopathy. The presence of isolated hypertrophic astrocytes in the cerebral white matter of premature infants, as reported in this series in the DWMG group, has been recognized for decades, but its significance remains unknown. Focal necrosis and diffuse hypertrophic astrocytes that are associated with “globules” and “acutely damaged glia” have been considered histological manifestations of the same disorder of immature cerebral white matter for which the term acquired perinatal telencephalopathy (PTL) has been coined [27]. Yet, hypoxic–ischemic white matter injury may follow a continuum of damage, from mild (gliosis [hypertrophic astrocytes] alone) to severe (periventricular necrosis combined with gliosis) [27]. Astrocytes, however, may also normally undergo hypertrophy in the late fetal and perinatal white matter as an obligatory developmental change, potentially due to the “physiological oxidative stress” of active myelin sheath synthesis, and thus may not be a marker of pathology at all [15]. These so-called myelination glia are immature oligodendrocytes that express markers such as O4, and are morphologically similar to GFAP positive reactive astrocytes. Coimmunofluoresence studies show no overlap in the expression of O4 and GFAP. The significant differences in age and survival encountered among the PVL, DWMG and Negative groups in this study precluded using the Negative group as a control representing “no white matter injury”. The Negative group consisted of infants who were born after significantly shorter gestational periods and who survived for significantly shorter time-periods postnatally than those in the PVL or DWMG groups. Thus, it is possible that the infants in the Negative group showed no white matter gliosis because the white matter is not vulnerable to injury at this early age, immature astrocytes are not capable of mounting a hypertrophic reaction to injury at this early time-point, and/or the patients did not survive long enough for astrocytic hypertrophy to develop. Further studies are needed to examine the significance of astrocytic hypertrophy in developmental pathology. The challenge is heightened by the unavoidable fact that live-born infants dying during the last half of gestation are not “normal”, but rather, typically die in intensive care units with multiple complications of prematurity that are known to adversely affect the brain. This study suggests that neuronal loss and/or gliosis in the perinatal period in gray matter sites critical to cognition, memory, and learning, i.e., thalamus [7, 38], basal ganglia [33], hippocampus [7, 26], and cerebellum [21, 37], play a role in cognitive defects in long-term survivors of prematurity. The neuroanatomic structures involved with neuronal loss and/or gliosis correlates well with the neuroimaging data, which has shown volumetric deficits in the thalamus and basal ganglia [17, 18], and to a lesser degree, the hippocampus [19, 34], and cerebral cortex [35] in survivors of prematurity. This thalamic damage could be important in the pathogenesis of subsequent cognitive impairments. Of note, afferent thalamocortical axons fail to reach the cortex when the subplate neurons are ablated and abnormal cortical lamination results [10–12, 14, 22, 30]. Selective subplate neuronal loss occurs in hypoxic-ischemic injury in neonatal rats [31], underscoring the possibility of homologous injury in human premature infants [16] and the need for in depth studies of the subplate-thalamic-cortical unit in humans. Premature infants are also at high risk for cerebellar injury [6, 20, 28, 29, 32], given the mounting evidence that the cerebellum plays a role in cognition [21, 37], our finding of substantial damage in this structure and its brainstem relay nuclei suggests that it could contribute to cognitive defects in survivors. In addition, injury to the cerebellum, as well as the thalamus and basal ganglia (globus pallidus), may contribute to the motor deficits of prematurity. Traditionally, the spastic motor deficits, i.e., cerebral palsy, in preterm infants has been attributed to damage to axons in the necrotic foci in PVL that are coursing through the periventricular white matter from the motor cortex to the spinal cord [41]. Our data suggest that at least some of the common, less severe motor deficits are due to gray, as well as white, matter damage. In conclusion, this study draws attention to the combination of white and gray matter injury in the brains of preterm infants dying in the perinatal period by the term “perinatal panencephalopathy”. Our findings suggest that future treatment strategies should target both white and gray matter damage to prevent the neurologic deficits in survivors of prematurity.	[ "white matter gliosis", "thalamus", "basal ganglia", "brainstem", "perinatal panencephalopathy", "neurodevelopmental disability", "perinatal hypoxia–ischemia" ]	[ "P", "P", "P", "P", "P", "M", "M" ]
Ann_Biomed_Eng-2-2-1705490	Automatic Regulation of Hemodynamic Variables in Acute Heart Failure by a Multiple Adaptive Predictive Controller Based on Neural Networks	Automated drug-delivery systems that can tolerate various responses to therapeutic agents have been required to control hemodynamic variables with heart failure. This study is intended to evaluate the control performance of a multiple adaptive predictive control based on neural networks (MAPCNN) to regulate the unexpected responses to therapeutic agents of cardiac output (CO) and mean arterial pressure (MAP) in cases of heart failure. The NN components in the MAPCNN learned nonlinear responses of CO and MAP determined by hemodynamics of dogs with heart failure. The MAPCNN performed ideal control against unexpected (1) drug interactions, (2) acute disturbances, and (3) time-variant responses of hemodynamics [average errors between setpoints (+35 ml kg−1 min−1 in CO and ±0 mmHg in MAP) and observed responses; 6.4, 3.7, and 4.2 ml kg−1 min−1 in CO and 1.6, 1.4, and 2.7 mmHg (10.5, 20.8, and 15.3 mmHg without a vasodilator) in MAP] during 120-min closed-loop control. The MAPCNN could also regulate the hemodynamics in actual heart failure of a dog. Robust regulation of hemodynamics by the MAPCNN was attributable to the ability of on-line adaptation to adopt various responses and predictive control using the NN. Results demonstrate the feasibility of applying the MAPCNN using a simple NN to clinical situations. INTRODUCTION Hemodynamic variables in a critical care patient with heart failure must be monitored during and after cardiac surgery. The hemodynamic conditions must then be regulated using infusion of several drugs. In particular, cardiac output (CO) and mean arterial pressure (MAP) are primary target variables because increased CO is required with suppressing myocardial oxygen consumption and MAP must be kept at the lowest level that can adequately maintain coronary circulation with decreasing systemic vascular resistance (SVR).14 Combined infusion of an inotropic agent such as dobutamine (DBT) or dopamine and a vasodilator such as sodium nitroprusside (SNP) or nitroglycerin has proven effective for patients with heart failure to regulate hemodynamics.12,13,25 Inotropic agents increase the force and velocity of cardiac muscle contraction and result in enhancing CO. Vasodilators reduce SVR and result in the decrease of the afterload of the heart, which thereby decreases MAP and increases stroke volume secondarily.3,14 Multivariable automated drug-delivery systems have been developed to help busy physicians or anesthesiologists use several drugs in many critical tasks to regulate the various hemodynamics that occur during heart failure.11,21,22,24,28,32,33 In simulation and animal studies undertaken during early system development, adaptive controllers demonstrated the feasibility of application of a multivariable drug-delivery system to simultaneously control CO and MAP using a combination of an inotropic agent and a vasodilator.11,24,28 In subsequent stages, multiple models and adaptive predictive controllers have been developed to adequately adjust the hemodynamic parameters in the presence of nonlinear physiological responses5,32 and drug interactions.21,22,33 Fuzzy controls have also shown optimal performance in regulation of multivariable hemodynamics with heart failure.6,7,31 However, patients with heart failure may have nonlinear and time-variant responses under unexpected variability to drugs and drug interactions with disturbances. Under such unknown conditions, the model-based controllers might require the preparation of numerous linear models to describe the various heart-failure patient responses to drugs,21,22,33 and the fuzzy controllers might require many rules based on the expert knowledge and heuristics of physicians or anesthesiologists for drug therapy of heart failure.6 The controllers that can adjust adaptively to unexpected responses with heart failure are required in actual clinical situations. Because neural networks (NN) might be one of the simple tools used for nonlinear and time-variant system responses in the presence of the unknown response variability and interactions with exogenous perturbation,15,27 the application of the NN to drug-delivery systems has been desired.8,17 To our knowledge, the multivariable controller using the NN for hemodynamic variables with heart failure has not been tested whereas the NN controllers have engendered the MAP controls such as post-operative hypertension1 and during acute hypotension9. Therefore, this study is intended to develop a multiple adaptive predictive control based on NN (MAPCNN) for hemodynamic variables and to evaluate its control performance under unexpected responses to drugs during CO and MAP regulation using SNP and DBT in heart failure. To be assured of a control algorithm before animal experiments, the MAPCNN was tested under largely unexpected (1) drug interactions, (2) acute disturbances, and (3) time-variant hemodynamic responses which can be freely operated. Finally, the performance of the MAPCNN was tested by actual hemodynamics of a canine left heart failure. MATERIALS AND METHODS This section fundamentally includes three parts: (i) ‘Modeling of Pharmacological Response’, which explains the method for the development of a computer response model that allows testing of the MAPCNN, (ii) ‘Development of Controller’, and (iii) ‘Evaluation of Controller’, which is the actual test of the MAPCNN. In addition, the animal study was divided into two parts. First, to model the pharmacological response in (i), animal experiments were performed using dogs with heart failure. Five of eleven dogs were used for every drug infusion. Then, for additional validation of the developed controller after computer simulations, an animal experiment was performed using one dog with heart failure. Modeling of Pharmacological Response To produce the response models to therapeutic agents in acute heart failure, the following animal study, which conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health, was performed. Microsphere embolization of the left main coronary artery induced acute ischemic heart failure in dogs (n = 5, 26–32 kg) that were anesthetized with pentobarbital sodium and ventilated artificially. A double-lumen catheter was introduced into the right femoral vein for administration of pharmaceutical agents using a computer-controllable infusion pump (CFV-3200; Nihon Kohden, Tokyo, Japan). An in-line electromagnetic flow probe (MFV-2100; Nihon Kohden) was used to measure CO; MAP was measured through a fluid-filled catheter and a pressure transducer (DX-200; Nihon Kohden). The CO and MAP were digitized at a 10-Hz sampling rate through a 12-bit digital-to-analog converter connected to a laboratory computer. The total number of all animals used for modeling was 11. The orders of drug infusions were the following four: (i) 3, 6, and 9 μg kg−1 min−1 in DBT and 1, 2, and 4 μg kg−1 min−1 in SNP (n = 2 of 11 animals); (ii) 3 and 6 μg kg−1 min−1 in DBT (n = 3); (iii) 9 μg kg−1 min−1 in DBT (n = 3); and (iv) 1, 2, and 4 μg kg−1 min−1 in SNP (n = 3). Specifically, five of all animals were used for each drug infusion rate. The hemodynamic variables [n = 9 except for the two animals in the above case (iii), the 9 μg kg−1 min−1 in DBT (n = 3), because of no measurements] were changed to −60.2 ml kg−1 min−1 in CO (p < 0.01, paired t-test) and +4.9 mmHg in MAP (not significant) before (CO, mean ± S.E.M = 132.0 ± 11.2 ml kg−1 min−1; and MAP, 92.8 ± 3.2 mmHg) and immediately after (CO, mean ± S.E.M = 71.8 ± 6.2 ml kg−1 min−1; and MAP, 97.7 ± 4.3 mmHg) the heart failure; for hemodynamics immediately before each drug infusion, see Fig. 1A.). The period between two trials of drug infusions was 10 min in each animal. To prevent the washout process of a catheter in the above case (i), both the drug infusions of DBT and SNP (n = 2), a double-lumen catheter was used. FIGURE 1.Single-drug dose responses in canine left heart failure (n = 5). A. Step responses of the change of cardiac output (ΔCO, left) and mean arterial blood pressure (ΔMAP, right) from the baseline after induced acute heart failure during 10-min infusion of (a) dobutamine (DBT) at 3, 6, and 9 μg kg−1 min−1 and (b) sodium nitroprusside (SNP) at 1, 2, and 4 μg kg−1 min−1. Data digitized at 10 Hz were averaged every 30 s. Data are mean ± S.E.M. B. Unit impulse responses of ΔCO (left) and ΔMAP (right) calculated from data of (a) DBT infusion at 3, 6, and 9 μg kg−1 min−1 and (b) SNP infusion at 1, 2, and 4 μg kg−1 min−1 in dogs. CO response of SNP at 1 μg kg−1 min−1 was eliminated because of impossible fit to step response. In this study, component models comprising a first order dynamic system cascaded with a nonlinear sigmoidal function were used to model the responses of CO and MAP with heart failure. After pharmacodynamics for the evaluation of controllers was represented by a linear first-order transfer function,18 a sigmoidal function was applied to the linear model to express the nonlinear characteristic with a component model approximating the positive step response.4,30 The procedure to produce the model responses can be described as follows. First, simple models for responses to therapeutic agents were produced from experimental data in canine left-heart failure. Figure 1A shows the step responses of CO and MAP changed (ΔCO and ΔMAP) from baseline values immediately before each drug infusion after inducing the acute heart failure during 10-min (a) DBT and (b) SNP infusions. The step responses of ΔCO and ΔMAP during infusion of DBT at 3, 6, and 9 μg kg−1 min−1 or SNP at 1, 2, and 4 μg kg−1 min−1 were averaged every 30 s. Then, each single-input single-output response [Δŷ(t)] of the four step responses (DBT-CO, DBT-MAP, SNP-CO, and SNP-MAP loops as input–output relationships) was approximated to the linear first-order delay system with a pure time delay in the continuous-time domain, as where K is a proportional gain, L is a pure time delay, and Tc is a time constant. If t < L then Δŷ(t) = 0. The fitted parameters to the averaged step responses (n = 5) in the single infusion of DBT or SNP were acquired by least squares method and used for calculation of the following unit impulse response (i.e., the response to the infusion of 1 μg kg−1 min−1 of a given drug). Second, the linear-model response [Δy(t)] was calculated by the convolution integral in the discrete-time domain: where . In those equations, u(t) is the drug infusion rate, ΔT is the sampling interval, and Nm is the finite number of terms in the model for the unit impulse response. The unit impulse response is g(t), as calculated from the derived values of the step response of Eq. (1). The proportional gain of a unit impulse response is Ku; Tc and L are the same values as Eq. (1). For the simulation study, ΔT and Nm were set, respectively, to 30 s and 20. According to this method, the unit impulse responses of ΔCO and ΔMAP during infusions of DBT at 3, 6, and 9 μg kg−1 min−1 and SNP at 1, 2, and 4 μg kg−1 min−1 were calculated from the fitting parameters to the average step responses (Fig. 1B, the CO response to SNP at 1 μg kg−1 min−1 was excepted because of the impossible fit to the step response). Although some differences existed among unit impulse responses at those infusion rates, as shown in Fig. 1B, in the present study, median values (i.e. 6 μg kg−1 min−1 in DBT and 2 μg kg−1 min−1 in SNP) were used for the following simulations because they characterized the effects of drugs on hemodynamics well. Numerical data in Ku, Tc, L, and g(t) used for the simulation study are shown in Table 1 (left). TABLE 1.Model parameters in linear-fitting and nonlinear-fitting functions.Drug-responseLinearNonlinefsarProportional gain (Ku)Time constant (Tc)Pure time delay (L)R2Response range (p1)Coefficient of gain (p2)R2DBT-CO15.8164.3300.98105.30.0280.99DBT-MAP4.465.2300.7522.80.1450.98SNP-CO3.040.6600.2837.70.0510.80SNP-MAP−12.5209.4600.96-26.2−0.0850.99Fitting parameters (Ku, Tc, L, p1, and p2) in a single-input single-output relationship. The Ku [ml μg−1 in DBT-CO and SNP-CO loops or mmHg (μg kg−1 min−1)−1 in DBT-MAP and SNP-MAP loops], Tc (s), and L (s) are the parameters of Eq. (2). The p1 and p2 are the parameters of Eq. (3). R2 shows a multiple coefficient of determination. Third, to express the nonlinear response to a single drug infusion, the Δy(t) in Eq. (2) as the linear model response was modified through a sigmoidal function:30 in which p1 is the parameter of the response range, which shows the difference between the maximum and minimum values of Δy′(t) as the nonlinear model response, and p2 is the parameter of the coefficient of gain. Parameters p1 and p2 were determined by nonlinear least-squares method for the simulation study [Table 1 (right) and Fig. 2A]. Figure 2A contains the average values in ΔCO and ΔMAP responses of the final 30 s during 10-min infusions of DBT at 3, 6, and 9 μg kg−1 min−1 and SNP at 1, 2, and 4 μg kg−1 min−1 (Fig. 1A). The ΔCO and ΔMAP responses tested in simulations were over the ranges of the averaged responses in actual hemodynamics of dogs with heart failure. FIGURE 2.A. Nonlinear responses of ΔCO and ΔMAP in infusion of DBT or SNP. Circles (•) show experimental data for dogs with heart failure; solid lines are curves fitted to the averaged data. Dashed lines show the limits of the tested range [one-third (a1 = a2 = b1 = b2 = 1/3) to three-times (a1 = a2 = b1 = b2 = 3) responses compared with the averaged data (a1 = a2 = b1 = b2 = 1) in Eq. (4)] in simulations. B. Multi-input multi-output model responses for ΔCO and ΔMAP. System inputs are the infusion rates of DBT and SNP; outputs are the ΔCO and ΔMAP. The outputs were determined by the first-order dynamic systems cascaded with nonlinear sigmoidal functions. The proportional gains of patient sensitivities to drugs and drug interactions are shown as a1, a2, b1, and b2. Finally, the model responses containing patient sensitivities to therapeutic agents and drug interactions are expressed as , where ΔCOmod(t) is the model response containing sensitivities to drugs and drug interactions of DBT and SNP; ΔCO1′(t) and ΔCO2′(t), respectively, indicate the nonlinear responses to single infusions of DBT and SNP; and a1 and a2, respectively, represent the proportional gain of patient sensitivity to DBT and SNP. The model response containing sensitivities to drugs and drug interactions of SNP and DBT is ΔMAPmod(t). The respective nonlinear responses to single infusions of SNP and DBT are ΔMAP1′(t) and ΔMAP2′(t). The respective proportional gains of patient sensitivity to SNP and DBT are denoted as b1 and b2 (see Fig. 2B). In particular, a2 and b2 can be defined as the strength of the drug interaction when the two treatments for DBT-CO and SNP-MAP loops are performed. Development of Controller Control Design Figure 3 portrays a block diagram of a MAPCNN system for adaptation to various patient responses with heart failure. Treating the multiple loops in the therapy for heart failure separately allows setting of a clear goal of NN learning for the various patient responses during closed-loop control. On the other hand, in the completely separated controllers, the total control performance will be late because one controller performs the drug therapy after detection of the drug interaction disturbance induced by performing an action taken by the other controller. Therefore, the MAPCNN in this study includes two module controllers for the DBT-CO loop considering the effects of SNP on CO and the SNP-MAP loop considering the effects of DBT on MAP. FIGURE 3.A block diagram showing multiple adaptive predictive control using neural networks (MAPCNN) to regulate CO and MAP. The r is a target value, and e(t) is the error between the target value and observed value. The value e(t + i) represents the error between the target value and output predicted by the NN. Thick lines show the learning loop in the NN; dotted lines show the prediction loop using the NN. Figure 4A depicts one of the two NN structures in MAPCNN tested for the simulation study. The MAPCNN is a control system in which the NN recursively learns patient characteristics using their observed responses to drug infusions only once every 30 s during closed-loop control (Learning Loop). It subsequently determines the future outputs using the learned NN (Prediction Loop). A multilayer feed-forward NN with two hidden layers (ΔyNN) emulated the nonlinear responses in ΔCOmod and ΔMAPmod. The ΔyNN is predicted through NN as where Δymod(t−1) and Δymod(t−2) are model responses of past ΔCO or ΔMAP; u1(t−1),..., and u1(t−6) or u2(t−1),... and u2(t−6) represent the past 3-min infusion rates of DBT or SNP. Here, to determine the length of the history of model response and infusion rates as inputs to the NN, the accuracy of NN learning was tested under various lengths of components (for a detailed protocol, refer to the following paragraph “Learning of Initial Weights in NN”). Figure 4B shows the average values (final 500 points) in the absolute error between the ΔCOmod and ΔCONN responses (left) or ΔMAPmod and ΔMAPNN responses (right) from the trained NN [100,000 times, a1 = a2 = b1 = b2 = 1 in Eq. (4)] under various input–output components to the NN input. The numbers of the input units (m) to a single NN component and of the units in the first (n1) and second (n2) hidden layers of the NN were set to the same as that of components to the NN input. For example, if the inputs to the NN were 14 (the past 3 min in both the two drug inputs and the past 1 min in the model response), then m = n1 = n2 = 14. Learning rates of the two NN in both the CO and MAP controls were set to Kn1 = Kn2 = 0.2. The starting weights of NN were given at random every trial. In the past history of infusion rates, the error between the NN and model responses in the past 2-min infusion rates (eight components in Fig. 4B: four in DBT and four in SNP) showed adequate accuracy, although that in the past 1-min infusion rates (four components) was not demonstrably accurate in both the CO and MAP responses. Here, it was predicted that the pure time delays to drug inputs were different among patients and that the NN was required to express the characteristics of the transient response to a drug input adequately considering the effect of the other drug input. Therefore, to adjust the NN response to the changes of hemodynamics with noise and disturbances during the real-time control, the infusion rates of the past 3 min in both drugs (total 12 components: 6 components in each drug) were, on the safe side, selected for both the CO and MAP controls in the present study. In the past history of model response, the accuracy of NN learning was the almost equal among the cases of the past 1, 1.5, and 2 min (2, 3, and 4 components in Fig. 4B) in both the CO and MAP responses, whereas that in the past 30 s (1 component) showed inadequate accuracy. Accordingly, the length of the past history of model response was determined as 1 min (2 components) in both the CO and MAP controls. FIGURE 4.A. A single component of a four-layer feed-forward NN with two hidden layers in MAPCNN to emulate the characteristics of a patient. The number of units in each hidden layer of the NN was set to 14 (numerically equal to the input units). A hyperbolic tangent function was used as the output of each unit. B. Average values (final 500 points) in the absolute error between the ΔCOmod and ΔCONN responses (left) or ΔMAPmod and ΔMAPNN responses (right) from the trained NN [100,000 times, a1 = a2 = b1 = b2 = 1 in Eq. (4)] under various lengths of input–output components to NN input. C. The absolute error between the model response and the predicted response by NN in (a) ΔCO (left) or (b) ΔMAP (right). In the simulation study, the number of input units to a single NN component was m = 14; the numbers of units in the first and second hidden layers of the NN were n1 = n2 = 14, being equal to the input units. The weights in the single NN were 435: 196 in the input to first hidden layer, 196 in the first to second hidden layer, 14 in the second hidden layer to output layer, and 29 for biases (Fig. 4A). The two NN for the controls of DBT-CO and SNP-MAP loops had identical structures. In the learning loop, a single NN was trained by the output of ΔCOmod or ΔMAPmod to the random inputs of DBT and SNP using the backpropagation algorithm in an on-line mode, showing that the error function is calculated after presentation of an input. The prediction loop in the MAPCNN determines the optimal DBT or SNP infusion rate that minimizes the cost function using the updated NN through the learning loop every 30 s. Therein, the cost function [J1(t), J2(t)] comprises the weight of input change (q1, q2), the prediction range (Np1, Np2), and the setpoint (r1, r2) in the controllers for the DBT-CO loop or the SNP-MAP loop. The physiological responses predicted by the NN are indicated as ΔCONN and ΔMAPNN. The controller based on the NN for each loop predicts future outputs using past inputs of infusion rates of DBT and SNP. The optimization of the infusion rates [u1(t), u2(t)] was performed using a Nelder–Mead Simplex algorithm.16,27 Determination of Control Parameters Learning of Initial Weights in NN The two NN, respectively, learned ΔCOmod response in the DBT-CO loop and ΔMAPmod response in the SNP-MAP loop to determine the initial weights in the NN for the MAPCNN. The starting weights in the NN before learning the model response were given at random between −1 and 1. Subsequently, the infusion rates of DBT and SNP were given at random between −4 and 6 μg kg−1 min−1. Here, both the plus and minus signs as drug inputs (artificial infusions) to the NN were used because the learning of NN was inferred to be more effective than that under a plus sign alone as the drug input to the NN in the trial and error and previous studies.9,27 Learning of the NN for ΔCOmod or ΔMAPmod responses was repeated 100,000 times. The ΔCOmod was divided by 200, and ΔMAPmod was divided by 100 for normalization during NN learning. Figure 4C shows the absolute error between the ΔCOmod and ΔCONN responses (left) or ΔMAPmod and ΔMAPNN responses (right) from the trained NN in the average patient sensitivity [a1 = a2 = b1 = b2 = 1 in Eq. (4)]. Learning rates of the two NN for the following simulation study were set to Kn1 = Kn2 = 0.2 showing a suitable number by trial and error in both the ΔCOmod and ΔMAPmod. Learning results of the NN, respectively, showed errors of 2.5 ml kg−1 min−1 in the DBT-CO loop and 1.5 mmHg in the SNP-MAP loop. Controller Tuning Optimal values of the range of prediction (Np1, Np2) and the weight of input (q1, q2) in the cost function (6) of the controller were explored using the model patient responses, ΔCOmod and ΔMAPmod, to determine the initial controller parameters. The prediction range was set to Np1 = Np2 = 4, 8, or 12, and the weight of input was set to q1 = q2 = 0.01, 0.1, or 1. The learning rates of ΔCOmod and ΔMAPmod were fixed at Kn1 = Kn2 = 0.2. To regulate the control speed and stability simultaneously, the setpoint of ΔCO was guided by the linear function r1 = 35t/600 ml kg−1 min−1 during the 10 min following the start of the closed-loop control. Thereafter, it was maintained at r1 = 35 ml kg−1 min−1. The setpoint of ΔMAP was set to r2 = ±0 mmHg. The duration of the closed-loop control was set to 40 min. Figure 5 shows simulation results of the MAPCNN using average responses of ΔCOmod and ΔMAPmod [a1, a2, b1, and b2 in Eq. (4) were set to unity]. The controller suppressed a control speed instead of facilitating stable control with the increase of the weight of input (q1 = q2 = 0.01→1) at each range of prediction (Np1 = Np2 = 4, 8, or 12). On the other hand, when the weight of the input is small (q1 = q2 = 0.01), the controller performed a slightly aggressive control with the decreased range of prediction (Np1 = Np2 = 12→4). The relationship at the large weight of input (q1 = q2 = 1) was opposite because of the strong effects of the input weight at the small range of prediction (Np1 = Np2 = 4). For subsequent simulations, the parameters (Np1, Np2, q1, q2) in the MAPCNN were set to Np1 = Np2 = 12 and q1 = q2 = 0.01 considering the settling time which reflects control speed and stability. In this case, the settling time within ±3 ml kg−1 min−1 in ΔCOmod was 660 s; its time within ±2 mmHg in ΔMAPmod was 990 s. The average absolute value of error between a setpoint and model response in ΔCOmod or ΔMAPmod over the entire control period (average error) for 40 min was 2.8 ml kg−1 min−1 in ΔCO or 0.5 mmHg in ΔMAP. FIGURE 5.Simulation results of the MAPCNN under average responses of ΔCOmod and ΔMAPmod [a1 = a2 = b1 = b2 = 1]. The input weight was changed to q1 = q2 = 0.01, 0.1 or 1 fixing the range of prediction at Np1 = Np2 = (A) 4, (B) 8, or (C) 12. The NN learning rate was set to Kn1 = Kn2 = 0.2 under all conditions. Evaluation of Controller Simulations To evaluate the control performance, a simulation study in MAPCNN, which expressed the repeatability and freely operated physiological parameters such as nonlinearity and interaction,22,30 was performed under unexpected changes of patient responses to therapeutic agents with acute disturbances using the model response based on experimental data of canine heart failure. Increased CO to more than 95 ml kg−1 min−1, while keeping MAP within the normal range (80–100 mmHg) is desirable to treat acute heart failure.6,7 Therefore, the control objectives in this study were to increase the low CO (mean ± S.E.M. = 67.5 ± 3.4 ml kg−1 min−1, Fig. 1A) at the setpoint of +35 ml kg−1 min−1 (ΔCO = +35t/600 within 10 min after the start of closed-loop control) using DBT infusion and to simultaneously maintain the normal MAP (mean ± S.E.M. = 96.7 ± 2.0 mmHg, Fig. 1A) at the setpoint (ΔMAP = ±0 mmHg) using SNP infusion when hypertension is induced by treatment for DBT-CO loop in acute heart failure. The infusion rates were basically bounded as 0 ≤ u1(t) ≤ 10 in DBT and 0 ≤ u2(t) ≤ 6 in SNP to avoid an overdose or drug toxicity.5,22 Hemodynamic control was simulated under the following cases: drug interactions between DBT and SNP, acute disturbances, and time-variant changes of physiological parameters. Drug Interactions Grasping and estimating the acts of drug interactions are difficult for hemodynamic control using multiple drugs.6,30 To examine the controller’s robustness for wide ranges of patients’ sensitivity to drugs and drug interactions, parameters (a1, a2, b1, b2) in Eq. (4) were changed to 2, 1/3, or 3 every 40 min for 120-min control (Fig. 6A). Hemodynamic responses were unknown to the NN because the NN learned only the average model responses (a1 = a2 = b1 = b2 = 1). The controller was also tested under various combinations for sensitivities to drugs and drug interactions. The model parameters, a1, a2, b1, and b2 in Eq. (4), were set to one of 1/3 (Low), 1 (Mid.), and 3 (High), and all combinations were tested with or without the limitations of drug infusion rates. The control duration was 40 min. FIGURE 6.Simulation results of MAPCNN with unexpected patient sensitivities and drug interactions. A in the top graph displays changes of parameters (a1, a2, b1, b2) in Eq. (4). B(a) and C(a) in the graph display setpoints, ΔCOmod and ΔMAPmod (solid lines), and predicted outputs by NN (dashed lines). B(b) and C(b) are the time courses of weights changed from the baseline at the starting time in the NN for controllers in the DBT-CO and SNP-MAP loops: weights between input and first hidden layers (top), first and second hidden layers (middle), and second hidden and output layers (bottom). D. Infusion rates of DBT (solid line) and SNP (dashed line). To know the limitation of the control performance, the MAPCNN was tested under very low (a1 = a2 = b1 = b2 = 1/5 or 1/10) and high (a1 = a2 = b1 = b2 = 5 or 10) sensitivities to drugs and drug interactions. The limitations of infusion rates of drugs were eliminated to emphasize the control performance. The control duration was 40 min. Acute Disturbances Bolus infusion of drugs during and after cardiac surgery often introduces severe disturbances to a controller. Hemorrhage, patient-position changes, and changes in anesthesia levels will modulate patient response characteristics.10,20 To examine the controller’s performance during acute disturbances, the MAPCNN against acute hypertension was simulated using the ΔCOmod and ΔMAPmod responses for 120 min. Exogenous perturbations were added to the patient responses ranging within ±10 ml kg−1 min−1 in ΔCOmod and +20 mmHg in ΔMAPmod. To mimic physiological variation, random noises within ±3 ml kg−1 min−1 in ΔCOmod and ±2 mmHg in ΔMAPmod were also added [Fig. 9A(a) and B(a)]. To implicate the limitation of the control performance, the MAPCNN was tested under very severe situations with random noise and exogenous perturbations over physiological responses: (i) very huge amplitudes of random noise, (ii) very large and (iii) acute disturbances. First, (i) three levels of huge amplitudes of random noise (level 1: ±5 ml kg−1 min−1 in ΔCOmod and ±5 mmHg in ΔMAPmod, level 2: ±25 ml kg−1 min−1 and ±25 mmHg, and level 3: ±50 ml kg−1 min−1 and ±50 mmHg) were added to model responses (a1 = a2 = b1 = b2 = 1) during drug treatment for 40 min [Figs. 10A(a) and B(a)]. The random noises among the three levels had the same pattern except for the amplitude. Next, (ii) the MAPCNN was evaluated under very huge disturbances (level 1: −20 ml kg−1 min−1 in ΔCOmod and +20 mmHg in ΔMAPmod, level 2: −50 ml kg−1 min−1 and +50 mmHg, and level 3: −100 ml kg−1 min−1 and +100 mmHg; a1 = a2 = b1 = b2 = 1; Fig. 11A). Finally, (iii) the MAPCNN was tested under very acute disturbances with model responses in high sensitivities to drugs and drug interactions (a1 = a2 = b1 = b2 = 3). At 20 min after the start of control for 80 min, the acute disturbances of three levels [10 (level 1), 5 (level 2), and 0 (level 3) min to reaching the offset values of −50 ml kg−1 min−1 in ΔCOmod and +50 mmHg in ΔMAPmod after the start of the disturbances] were added (Fig. 11B). In those simulations, the limitations of infusion rates of drugs were eliminated to elucidate the limitations of the control performance. Time-variant Changes Patient responses to therapeutic agents have a nonlinear and time-variant nature. If the delay of a plant, which reflects the infusion rate of drugs through the catheter, internal patient circulation and perfusion delay, and the drug-recirculation characteristics of a patient,2,26,30 is not known accurately or changes during drug infusions, infusion delays will engender an unstable condition. Therefore, the controller robustness was tested under the change of the infusion delay in this study. The pure time delays [L in the unit impulse response of Eq. (2)] of patient responses to therapeutic agents were varied from 30 to 90 s in CO responses and 60 to 120 s in MAP responses to DBT and SNP infusions during the closed loop control [Fig. 12A(a)]. The parameters (a1, a2, b1, b2) in Eq. (4) were varied from 1/3 to 3 to examine the controller performance under time-variant and wide ranges of patient sensitivities to drugs and drug interactions [Fig. 12A(b)]. The exogenous perturbations were added to the time-variant patient responses ranging within ±10 ml kg−1 min−1 in ΔCOmod and +20 mmHg in ΔMAPmod. Random noises within ±3 ml kg−1 min−1 and ±2 mmHg were added to ΔCOmod and ΔMAPmod responses [Fig. 12B(a) and C(a)]. The control duration was 120 min. Animal Study To evaluate the control performance in MAPCNN under the unknown physiological responses such as nonlinearity and drug interaction, the simultaneous control of CO and MAP was performed using a dog with acute heart failure. Acute ischemic heart failure in an anesthetized and ventilated dog (23 kg) was induced by microsphere embolization of the left main coronary artery. A double-lumen catheter was introduced into the right femoral vein for administration of drugs using an infusion pump (CFV-3200; Nihon Kohden, Tokyo, Japan). CO was measured by an electromagnetic flow probe (MFV-2100; Nihon Kohden), and MAP was measured through a pressure transducer (DX-200; Nihon Kohden) at a 10-Hz sampling rate through a 12-bit digital-to-analog converter. The control objective was to increase the low CO (62.4 ml kg−1 min−1) at the setpoint (+40%) using DBT infusion and to maintain the MAP (73.9 mmHg) at the setpoint (ΔMAP = ±0 mmHg) using SNP infusion in acute heart failure. The closed-loop control duration was 60 min. The infusion rates were bounded as 0 ≤ u1(t) ≤ 10 in DBT and 0 ≤ u2(t) ≤ 6 in SNP. RESULTS Simulations Drug Interactions Figure 6 shows simulation results of closed-loop control by MAPCNN (Np1 = Np2 = 12, q1 = q2 = 0.01, and Kn1 = Kn2 = 0.2) under unknown patient sensitivities to drugs and drug interactions. The ΔCO in the DBT-CO loop converged on the setpoint (+35 ml kg−1 min−1) within approximately 15 min, according to the guided setpoint during the first 40-min control period, regardless of unanticipated patient sensitivities and drug interactions (a1 = a2 = b1 = b2 = 2). The control for the ΔMAP in the SNP-MAP loop minimally suppressed the hypertension (+4.2 vs. +15.0 mmHg with or without SNP infusion) induced by the DBT infusion. At 40 and 80 min of the closed-loop control, patient sensitivities and drug interactions were widely changed (a1 = a2 = b1 = b2 = 2 to 1/3 and 1/3 to 3). The ΔCO converged on the setpoint robustly, whereas the ΔCO showed transient and large changes (+35.0 showing the setpoint to +14.9 and to +99.3 ml kg−1 min−1 at approximately 50 and 85 min). Although ΔMAP was decreased acutely by the change of patient sensitivity and drug interaction (±0 to −20.7 mmHg at approximately 85 min), it returned robustly to a normal level. The average errors between setpoints and observed responses were 6.4 ml kg−1 min−1 in CO and 1.6 mmHg (10.5 mmHg without SNP infusion) in MAP during 120-min closed-loop control. Weights and biases of two NN were adjusted to optimal values [Fig. 6B(b) and C(b)] when the unexpected changes occurred. The infusion rates of DBT and SNP were adjusted smoothly to optimal levels corresponding to the unknown patient responses to drugs. Table 2 shows average errors between setpoints and model responses in ΔCO and ΔMAP during closed-loop control by MAPCNN (Np1 = Np2 = 12, q1 = q2 = 0.01, and Kn1 = Kn2 = 0.2) under the various sensitivities to drugs and drug interactions. The control performance was overall accurate and a tendency excited for dependence on the sensitivity of CO to DBT (i.e., parameter a1); the control performance was increased as the sensitivity of CO to DBT increased (a1 = 1/3→3). However, in some cases, great hypertension was observed during control, as displayed in Fig. 7A; although the CO control was smoothly completed under such cases, the great hypertension was induced because of the high interaction of DBT infusion used for CO treatment (i.e. b2 = 3). The maximum values of the hypertension in cases 1 (a1 = ‘Low’, a2 = ‘Low’, b1 = ‘Low’, and b2 = ‘High’), 2 (a1 = ‘Low’, a2 = ‘Low’, b1 = ‘Mid.’, and b2 = ‘High’), and 3 (a1 = ‘Low’, a2 = ‘Mid.’, b1 = ‘Low’, and b2 = ‘High’) were 60.3, 39.4, and 33.8 mmHg, respectively. Here, ‘Low’ = 1/3, ‘Mid.’ = 1, and ‘High’ = 3. In case 2, the infusion rates of DBT were slightly disturbed and the estimation error between the NN and ΔCO responses oscillated. However, the ΔCO in case 2 was unaffected by the oscillation of DBT infusion because of the low sensitivity of CO to DBT (a1 = ‘Low’). Infusion rates of SNP in all cases were saturated at around 8 min because of the limitation of drug infusion. After the saturation, the infusion rate of SNP in case 1 was decreased at around 12 min. TABLE 2.Average errors between setpoints and model responses in CO and MAP under various sensitvities to drugs and drug interactions.CO (a1):LowMid.HighCO (a2):MAP (b1)MAP (b2)LowMid.HighLowMid.HighLowMid.HighLowLow5.5 (0.5)4.8 (0.4)3.4 (0.3)2.9 (0.3)2.7 (0.3)2.2 (0.2)1.3 (0.1)1.2 (0.1)1.2 (0.1)Mid.5.3 (10.3)4.1 (4.8)3.0 (1.0)3.3 (1.9)2.6 (1.0)2.0 (0.7)1.3 (0.5)1.2 (0.5)1.2 (0.4)High5.7 (45.0)3.8 (25.1)4.2 (3.0)3.5 (17.3)2.9 (6.1)2.2 (1.6)4.2 (3.4)3.9 (3.5)1.5 (1.2)Mid.Low6.0 (1.2)5.7 (1.2)5.8 (1.1)3.2 (0.6)2.9 (0.6)2.4 (0.5)1.5 (0.4)1.4 (0.4)1.3 (0.3)Mid.5.5 (0.6)4.7 (0.5)3.5 (0.4)3.2 (0.5)2.8 (0.5)2.1 (0.4)1.6 (0.4)1.4 (0.4)1.3 (0.3)High5.3 (29.6)3.8 (12.2)3.4 (1.9)3.3 (3.9)2.7 (1.4)2.0 (1.0)1.9 (1.1)1.7 (0.9)1.3 (0.7)HighLow6.0 (2.5)5.9 (2.5)6.2 (2.3)2.8 (0.7)2.8 (0.6)2.7 (0.7)1.3 (0.6)1.3 (0.5)1.3 (0.4)Mid.6.0 (4.1)6.2 (3.9)7.0 (3.4)2.9 (0.6)2.8 (0.6)2.5 (0.6)1.4 (0.6)1.3 (0.5)1.3 (0.4)High5.6 (1.0)4.8 (0.7)6.4 (4.3)3.2 (0.8)2.9 (0.8)2.3 (1.3)2.0 (1.6)1.8 (1.6)1.5 (1.5)The numbers show the average errors in CO (left) and MAP (right). ‘Low’ = 1/3, ‘Mid.’ = 1, and ‘High’ = 3 in the parameters in Eq. (4). The sign of shows the cases of large hypertension (for detailed results, see Fig. 7).FIGURE 7.Simulation results of MAPCNN under cases of great hypertension (A) with and (B) without the limitation of drug infusion rates. Parameters (a1, a2, b1, b2) in Eq. (4) are a1 = ‘Low’, a2 = ‘Low’, b1 = ‘Low’, and b2 = ‘High’ in case 1, a1 = ‘Low’, a2 = ‘Low’, b1 = ‘Mid.’, and b2 = ‘High’ in case 2, and a1 = ‘Low’, a2 = ‘Mid.’, b1 = ‘Low’, and b2 = ‘High’ in case 3. ‘Low’, ‘Mid.’, and ‘High’ are 1/3, 1, and 3, respectively. The hemodynamic responses (top), the error between the model response and the predicted response by NN during control (middle), and infusion rates of DBT and SNP (bottom) in ΔCO (left) or ΔMAP (right). To elucidate the control performance without the limitation of infusion rates of drug inputs, the MAPCNN was also tested under the same cases 1, 2, and 3 (Fig. 7B). However, great hypertension was not improved by SNP infusion without the drug input limitation (63.8, 39.8, and 38.4 mmHg in cases 1, 2, and 3) compared with those with the limitation of the drug input. In cases 2 and 3, the drug infusion rates in DBT were disturbed and the estimation errors between the NN and ΔCO responses had some oscillations. In addition, in all cases, the infusion rates of SNP were decreased after increasing to around 10 μg kg−1 min−1, irrespective of the remaining the hypertension. Figure 8 shows the closed-loop control by MAPCNN (Np1 = Np2 = 12, q1 = q2 = 0.01, and Kn1 = Kn2 = 0.2) under (A) very low (a1 = a2 = b1 = b2 = 1/5 or 1/10) and (B) high (a1 = a2 = b1 = b2 = 5 or 10) sensitivities to drugs and drug interactions. Under very low sensitivities to drugs and drug interactions, the ΔCO response was not able to reach the setpoint (±35 ml kg−1 min−1) because of the slight sensitivity of CO to DBT (a1 = 1/5 or 1/10), but it finally converged on a stable value (Fig. 8A, left). Estimation errors between the NN and model responses in ΔCO had oscillations under such conditions. On the other hand, under very high sensitivities to drugs and drug interactions, both the ΔCO and ΔMAP responses oscillated depending on the degree of sensitivities to drugs [maximum amplitudes between observed values and setpoints: (+)9.2 ml kg−1 min−1 and (−)18.8 mmHg in a1 = a2 = b1 = b2 = 5 or (+)35.9 ml kg−1 min−1 and (−)54.1 mmHg in a1 = a2 = b1 = b2 = 10, Fig. 8B]. However, the MAPCNN had a tendency to suppress oscillations in the ΔCO and ΔMAP responses gradually. The estimation error between the NN and model responses and the infusion rates of drugs were also disturbed. FIGURE 8.Simulation results of MAPCNN with (A) very low (a1 = a2 = b1 = b2 = 1/5 or 1/10) or (B) very high (a1 = a2 = b1 = b2 = 5 or 10) patients’ sensitivities to drugs and drug interactions. The hemodynamic responses (top), the error between the model response and the predicted response by NN during control (middle), and infusion rates of DBT and SNP (bottom) in ΔCO (left) or ΔMAP (right). The limitations of drug infusion rates were eliminated. Acute Disturbances Figure 9 shows closed-loop control by MAPCNN (Np1 = Np2 = 12, q1 = q2 = 0.01, and Kn1 = Kn2 = 0.2) under unexpected acute disturbances with background noise. The ΔCO converged on the setpoint (+35 ml kg−1 min−1) within approximately 12 min according to the guided setpoint during the first 40-min control period, regardless of the added random noise (±3 ml kg−1 min−1) with expected patient responses (a1 = a2 = b1 = b2 = 1). The appropriate infusion of SNP for control of ΔMAPmod with random noise (±2 mmHg) suppressed hypertension (+3.5 vs. +15.4 mmHg with or without SNP treatment) induced by DBT infusion. At 40 and 80 min of the closed-loop control, acute disturbances were added to the ΔCOmod (+10 and −10 ml kg−1 min−1) and ΔMAPmod (+20 and +10 mmHg) responses. The MAPCNN suppressed the transient change of CO minimally and the CO converged on the setpoint as quickly as possible. The induced transient hypertension was suppressed robustly by optimal infusion of SNP (+6.7 vs. +31.0 mmHg with or without SNP treatment in the added disturbance at 40 min). Average errors between setpoints and observed responses were 3.7 ml kg−1 min−1 in CO and 1.4 mmHg (20.8 mmHg without SNP) in MAP during 120-min closed-loop control. Weights and biases of two NN were adjusted to the optimal values when the unexpected changes occurred [Fig. 9A(c) and B(c)]. Changes of biases were linked to those of the acute disturbances and physiological variation. The infusion rates of DBT and SNP were adjusted smoothly to optimal levels corresponding to the unknown disturbances. FIGURE 9.Simulation results of MAPCNN with unknown disturbances. In the graph, A(a) and B(a) show changes of acute disturbances and random noise added to ΔCOmod and ΔMAPmod responses. A(b) and B(b) represent setpoints, ΔCOmod and ΔMAPmod (solid lines), and predicted outputs by NNs (dashed lines). A(c) and B(c) indicate weight changes in NN for the controllers in DBT-CO and SNP-MAP loops: weights between the input and first hidden layers (top), first and second hidden layers (middle), and second hidden and output layers (bottom). C. Infusion rates of DBT (solid line) and SNP (dashed line). Figure 10 shows the closed-loop control by MAPCNN (Np1 = Np2 = 12, q1 = q2 = 0.01, and Kn1 = Kn2 = 0.2) under very large amplitudes of random noise [levels 1 (±5 in both ΔCOmod and ΔMAPmod), 2 (±25), and 3 (±50), Fig. 10A(a) and B(a); a1 = a2 = b1 = b2 = 1]. Although the ΔCO and ΔMAP responses [Figs. 10A(b) and B(b)] approached the setpoints, they depended directly on the appearance patterns and the amplitudes in large random noise. The infusion rates in DBT and SNP and the estimation errors between the NN and model responses were also reflected by changes of random noise.FIGURE 10.Simulation results of MAPCNN with large random noise. A(a) and B(a) show changes of large random noise added to ΔCOmod and ΔMAPmod responses. Noise level 1 (±5 in ΔCOmod and ±5 mmHg in ΔMAPmod), level 2 (±25 ml kg−1 min−1 and ±25 mmHg), and level 3 (±50 ml kg−1 min−1 and ±50 mmHg). A(b) and B(b) represent setpoints, ΔCOmod and ΔMAPmod responses (a1 = a2 = b1 = b2 = 1). A(c) and B(c) indicate the error between the model response and the predicted response by NN during control. A(d) and B(d) are Infusion rates of DBT and SNP. The limitations of drug infusion rates were eliminated. Figure 11A shows the closed-loop control by MAPCNN (Np1 = Np2 = 12, q1 = q2 = 0.01, and Kn1 = Kn2 = 0.2) under huge disturbances [levels 1 (−20 ml kg−1 min−1 in ΔCOmod and +20 mmHg in ΔMAPmod), 2 (−50 ml kg−1 min−1 and +50 mmHg), and 3 (−100 ml kg−1 min−1 and +100 mmHg); a1 = a2 = b1 = b2 = 1]. In the disturbance of level 1, the ΔCO and ΔMAP responses converged on the setpoints, whereas the estimation error between the NN and model responses in ΔCO exhibited some oscillations around 60–65 min. At level 2, the CO control was completed well. However, the MAP did not reach the setpoint because of the large disturbance and the interaction from DBT used for the CO control. In addition, the infusion rate of SNP was decreased at around 30 min irrespective of the hypertension of approximately 40 mmHg. The MAP response did not oscillate and converged on a steady state being far from the setpoint. At level 3, neither the ΔCO nor ΔMAP responses reached the setpoints. The infusion rate of SNP was increased instantaneously by receiving a great disturbance at 20 min, and it was decreased at around 25 min regardless of remaining the large hypertension of approximately 120 mmHg. The estimation error between the NN and model responses in ΔCO showed oscillation at approximately 40 min.FIGURE 11.Simulation results of MAPCNN with (A) very huge and (B) acute disturbances. A: huge disturbance level 1 (−20 ml kg−1 min−1 in ΔCOmod and +20 mmHg in ΔMAPmod), level 2 (−50 ml kg−1 min−1 and +50 mmHg), and level 3 (−100 ml kg−1 min−1 and +100 mmHg). B: acute disturbance level 1 (10 min to the added disturbance of −50 ml kg−1 min−1 in ΔCO and +50 mmHg in ΔMAP), level 2 (5 min), and level 3 (0 min); a1 = a2 = b1 = b2 = 3 in Eq. (4). The hemodynamic responses (top), the error between the model response and the predicted response by NN during control (middle), and infusion rates of DBT and SNP (bottom) in ΔCO (left) or ΔMAP (right). The limitations of drug infusion rates were eliminated. Figure 11B shows the closed-loop control by MAPCNN (Np1 = Np2 = 12, q1 = q2 = 0.01, and Kn1 = Kn2 = 0.2) under very acute disturbances with high sensitivities to drugs and drug interactions [levels 1 (10 min to the added disturbances of −50 ml kg−1 min−1 in ΔCO and +50 mmHg in ΔMAP), 2 (5 min), and 3 (0 min); a1 = a2 = b1 = b2 = 3]. At levels 1 and 2, although the ΔCO and ΔMAP responses showed small oscillations because of the acute disturbances, they converged on the setpoints by 80 min. In level 3 (step input of disturbance), the ΔCO and ΔMAP received the effects of the step input of the very acute disturbances at 20 min directly. Although ΔCO and ΔMAP responses were disturbed until 70 min, those values tended to converge on the setpoints eventually. In particular, the infusion rate of SNP at level 3 was increased acutely to 20 μg kg−1 min−1 against the large hypertension of 50–60 mmHg induced by the very acute disturbance with high amplitude of 50 mmHg at 20 min; in turn, the excessive hypotension after the acute hypertension was induced by the high sensitivity of MAP to the SNP used for the MAP control (b1 = 3). However, the infusion rate of SNP was adjusted to the optimal value at approximately 70 min. Time-variant Changes Figure 12 shows the closed-loop control by MAPCNN (Np1 = Np2 = 12, q1 = q2 = 0.01, and Kn1 = Kn2 = 0.2) under the unknown time-variant responses containing time delays to therapeutic agents with acute disturbances. The ΔCO converged on the setpoint (+35 ml kg−1 min−1) within approximately 10 min according to the guided setpoint during the first 40-min control, irrespective of time-variant patient sensitivities and drug interactions with added noise; it showed only slight oscillation (±5 ml kg−1 min−1). The SNP infusion for the control of ΔMAP suppressed the hypertension (+6.1 and +22.8 mmHg with and without SNP treatment) induced by DBT infusion beforehand. At 40 min of the closed-loop control, acute disturbances (−10 ml kg−1 min−1 in CO and +20 mmHg in MAP) were added to the patient responses. The ΔCO and ΔMAP quickly converged on the setpoint within approximately 10 min, whereas the transient hypertension was induced (+9.4 vs. +38.9 mmHg with or without SNP). The average errors between setpoints and observed responses were 4.2 ml kg−1 min−1 in CO and 2.7 mmHg (15.3 mmHg without a vasodilator) in MAP during 120-min closed-loop control. The weights and biases of two NN were adjusted to optimal values when the unexpected changes occurred [Fig. 12B(c) and C(c)]. Changes of biases were linked to those of the time-variant responses and acute disturbances. The infusion rates of DBT and SNP were adjusted to optimal levels corresponding to the unknown time-variant responses. FIGURE 12.Simulation results of MAPCNN under unknown time-variant responses with disturbances. A(a) in the graph displays changes of the parameters of the time delays (L) in the unit impulse response of Eq. (2). A(b) indicates changes of the parameters (a1, a2, b1, b2) in Eq. (4). B(a) and C(a) show changes of acute disturbances and random noise added to ΔCOmod and ΔMAPmod responses. B(b) and C(b) denote setpoints, ΔCOmod and ΔMAPmod (solid lines), and predicted outputs by NN (dashed lines). B(c) and C(c) are the weight changes in NN for controllers in DBT-CO and SNP-MAP loops: weights between input and first hidden layers (top), first and second hidden layers (middle), and second hidden and output layers (bottom). D. Infusion rates of DBT (solid line) and SNP (dashed line). Animal Study Figure 13 shows results of closed-loop control by MAPCNN (Np1 = Np2 = 8, q1 = q2 = 0.3, and Kn1 = Kn2 = 0.2) under the actual response of canine left heart failure. The CO in the DBT-CO loop converged on the setpoint within approximately 10 min regardless of the unknown response containing nonlinear, drug interaction, and the effects of arterial baroreflex with physiological variation in actual heart failure. The control for the ΔMAP in the SNP-MAP loop suppressed the acute hypertension (+20 mmHg) induced by the DBT infusion. Whereas the large disturbance like arrhythmia was induced at around 38 min during the control, the CO and MAP were appropriately controlled by the MAPCNN adjusting the infusion rates of DBT and SNP to the optimal levels. The average errors between setpoints and observed responses were 7.3 ml kg−1 min−1 in CO and 6.4 mmHg in MAP during 60-min closed-loop control. FIGURE 13.Results of hemodynamic regulation by means of MAPCNN under canine left heart failure. Raw data of CO (top) and MAP (middle) responses during DBT and SNP infusions. Bottom in the graph is the Infusion rates of DBT (solid line) and SNP (dashed line). DISCUSSION The development of automatic drug-delivery systems requires a controller that can adapt to the various patient responses in clinical situations. The MAPCNN was confirmed to be robust with respect to uncertainty in drug interactions, acute disturbances, time-variant responses containing time delays to therapeutic agents (Figs. 6, 9, and 12), and the actual response of a dog with heart failure (Fig. 13) because of its ability to learn nonlinear and time-variant changes of the system during the on-line control. The infusion of DBT increased MAP as well as CO in acute left heart failure of dogs [Fig. 1A(a)]. The DBT infusion does not generally act on the SVR whereas both the CO and SVR affect MAP.14 Therefore, the increase of MAP during DBT infusion in this study would have resulted mainly from increasing the CO induced by the actions of the beta receptors (β1, β2) of cardiac smooth muscle30 rather than the SVR. On the other hand, the SNP infusion resulted in the decrease of MAP and the increase of CO between middle and high doses [Fig. 1A(b)] because of the decrease of SVR (afterload of a heart) by SNP and secondarily increased CO.6 In addition, the SNP treatment might have suppressed the increase in preload through the decrease of SVR because of increasing venous compliance for retaining the blood in the veins and lowering the venous return to the heart in case of congestive heart failure. In the simultaneous regulation of CO and MAP, the control for CO induced hypertension (Figs. 6, 9, 12, and 13) because the primary control target in this study was the increase of low CO in acute heart failure. The MAPCNN was able to suppress the hypertension using optimal infusion of SNP as well as increasing CO using DBT to an optimal target value because the combined infusion of an inotropic agent and a vasodilator would have acted effectively.12,13,25 Application of MAPCNN to a multiple hemodynamic control accomplished the regulation of CO and MAP under various changes of the patient’s responses to drugs and disturbances (Figs. 6, 9, 12, and 13). In particular, regardless of the large change of patient sensitivities as shown in Fig. 6, the MAPCNN robustly adjusted the acute and large changes to generate a stable condition. Similarly to the previous controllers,7,10 the MAPCNN suppressed those disturbances performed stabilized control under unexpected acute disturbances (Figs. 9 and 12). Under time-variant patient responses with pure time delays, which are a crucial obstacle to stable control,2,26 the MAPCNN provided sufficient control performance (Fig. 12). Regardless of actual nonlinear response, drug interaction, and partial disturbances with arrhythmia, the MAPCNN could regulate the CO and MAP simultaneously (Fig. 13). The superior control performance resulted from the function to adjust the weights and biases of the NN to optimal points during the on-line control (Figs. 6, 9, and 12). Only the two-NN models of average responses with heart failure were considered in the calculation of the appropriate multiple drug infusion rates of DBT and SNP (Fig. 3) to mitigate the enormous number of trials associated with the control design. Model predictive controllers or fuzzy controllers might require an extremely lengthy set-up stage to prepare the model banks as linear models of patients’ responses to drugs or to provide the experienced rules describing various cases in clinical settings whereas the controllers are an effective means of adjusting to various patients’ sensitivities to drugs21 and describing nonlinear responses to drugs5. A controller based on NN solves those problems because it decreases the number of models required for the control design of the various changes of hemodynamics clearly. Irrespective of the wide range over physiological responses (Fig. 2) in simulations and actual response of a dog with heart failure containing the effects of baroreflex and the full renin–angiotensin system induced by long-term control,14 CO and MAP in the MAPCNN promptly approached the setpoints because of the optimization of both the stability and speed for the MAPCNN (Figs. 6, 9, 12, and 13). Therefore, the designed MAPCNN will be feasible for application to automatic drug therapy in heart failures. However, when rapid treatment using drugs against more acute and large disturbances is required during hemodynamic controls, another supplemental system might be required.9 Diagnoses of characteristics of patients’ responses to drugs or tuning weights of NNs during closed-loop controls1 may also be effective for hemodynamic controls to accelerate the NN learning speed. In addition, because the fluid infusion, blood transfusion, anesthesia, and muscular blockade as well as the therapeutic agents controlled in this study are common in clinical practice,22,29 the controllers that can adjust physiological responses to further multiple drugs will be desired. The MAPCNN tested herein can be extended simply to multivariate control systems under such clinical conditions for drug therapy with heart failure. The MAPCNN was tested under various conditions over physiological responses to elucidate the limitations of the control performance (Figs. 7, 8, 10, and 11). Regardless of such severe conditions, the hemodynamics during MAPCNN learning very large changes of the sensitivities to drugs and drug interactions and the disturbances using NN tended to converge on the setpoints with some oscillations observed, insofar as those responses were within the possible range of the control using DBT and SNP (e.g., Figures 8B and 11B). On the other hand, there existed cases for which it was obviously impossible for MAPCNN to control the hemodynamics (Figs. 7, 8A, and 11A). For example, when the interaction from DBT for the CO control to MAP was very large (b2 = 3) and the sensitivity of MAP to SNP was small (b1 = 1/3 or 1, Fig. 7), the MAPCNN was actually incapable of attenuating the induced hypertension. Those cases resulted from nonlinear model responses to drugs in the present study; the hemodynamic responses, therefore, would have saturated because of the range of nonlinear model responses, regardless of the increase of drug infusion rates. The NN in the MAPCNN under the severe conditions seems to have tried to learn and adapt to the situations during real-time control. In cases of large error between actual responses and predicted responses where the NN in the MAPCNN learned the average responses before the control (e.g. the CO responses of case 2 in Fig. 7A, cases 2 and 3 in Fig. 7B, and cases 1 and 3 in Fig. 11A), the hemodynamics during the control exhibited oscillation and the NN in the MAPCNN would have tried to learn the severely changed situation again. In addition, when the hemodynamic response showed incorrect or opposite response to the drug input compared with the learned average response by the NN before the control [e.g. cases of not reaching the setpoint and converging on a stable state regardless of the DBT infusion for CO control because of very low sensitivity to the drug (a1 = 1/5 or 1/10, Fig. 8A) and of the opposite effect of SNP on MAP, speciously, because of the very large disturbances (levels 2 and 3 in Fig. 11A) compared with the learning response], the drug infusion rate of the MAPCNN was decreased despite the remainder of the low CO or hypertension during the closed-loop control with real-time NN learning. Therefore, these simulation results suggest the following two points. First, in those cases such as the high interaction over the drug effect on the target physiological parameter and the incorrect or opposite responses to drugs compared with those of previously learned NN, improvement of the strategy in drug treatment would be required; alternatively, the MAPCNN would fall into a situation of control impossibility. Second, the physiological variations related with responses to anesthesia, antiarrhythmic drug, and muscle relaxant, the external disturbances, and the artificial background noise must be set to the smallest possible values to bring out the best performance of the controller. Several limitations are apparent in present study. First, the modeling for the CO and MAP responses to drugs might depend on the protocols of animal experiments such as the order and washout period of drug infusions. The protocol in the present animal study was, therefore, described in detail (reference ‘Modeling of Pharmacological Response’ in the ‘MATERIALS AND METHODS’ section). Second, one (low CO and normal MAP) of the heart-failure conditions was tested using the present animal study. The kinds of heart failures are various in actual patients. Therefore, further animal studies will be required. Finally, an electromagnetic flow probe was used for CO measurement in the present study. However, in an actual clinical setting, the common technique for CO measurement (e.g., the thermodilution technique using a pulmonary artery catheter19) has a slower response than that of an electromagnetic probe. The high accuracy of a flow probe, such as the time resolution (at least 30 s as used for this experiment), and the pure time delay from an actual response (as few response delays as possible) would be required to acquire the results that were obtained in the present study. CONCLUSIONS The MAPCNN was designed and evaluated in simulation and animal studies to regulate the nonlinear responses of CO and MAP in acute heart failure using DBT and SNP under unexpected changes of patient sensitivities to drugs, drug interactions, acute disturbances, and time-variant responses to therapeutic agents. The MAPCNN showed robust control performance irrespective of various unexpected responses to drugs over actual physiological responses (Fig. 2) and actual response of a dog in heart failure. Flexibility of a NN coupled with an adaptive control mechanism will enable the regulation of various physiological responses to drugs with heart failures.	[ "therapeutic agents", "cardiac output", "mean arterial pressure", "an automated drug infusion system" ]	[ "P", "P", "P", "R" ]
Ann_Surg_Oncol-4-1-2277442	The Microanatomic Location of Metastatic Breast Cancer in Sentinel Lymph Nodes Predicts Nonsentinel Lymph Node Involvement	Background The majority of sentinel node (SN) positive breast cancer patients do not have additional non-SN involvement and may not benefit from axillary lymph node dissection (ALND). Previous studies in melanoma have suggested that microanatomic localization of SN metastases may predict non-SN involvement. The present study was designed to assess whether these criteria might also be used to be more restrictive in selecting breast cancer patients who would benefit from an ALND. Axillary nodal status is among the most important prognostic factors in breast cancer patients. SN biopsy with an intensive pathological assessment of selectively removed lymph nodes is currently a highly accurate, minimally invasive technique to assess nodal status.1,2 It reduces the morbidity of breast cancer surgery by avoiding unnecessary axillary lymph node dissection (ALND) in patients with negative sentinel nodes (SNs).3 The optimal treatment of patients with a positive SN is however less clear. The few SNs can be cost-effectively analyzed by multiple-level evaluation and immunohistochemistry (IHC) which increases the likelihood of detecting small metastases.4 The decision to proceed with an ALND in patients with macrometastatic SN involvement does not pose a major clinical dilemma. The need for routine ALND in patients with minimal SN involvement however continues to be debated since only a minority of these patients (10–15%) show non-SN involvement.5 Predicting the chance of involvement of the non-SN would facilitate the selection of patients with a potential therapeutic benefit of ALND. Several features of the primary tumor and the involved SNs have been investigated as potential predictors for non-SN involvement. Primary tumor size, palpability, presence of peritumoral lymphovascular invasion, number of tumor-involved SNs, size of the SN metastases, and extracapsular extension (ECE) correlate with non-SN status.6–10 However, none of these factors are sensitive and reproducible enough to reliably identify a subgroup of patients who might be spared ALND. Several melanoma studies have reported that the microanatomic pattern of SN involvement and the penetrative depth (defined as the maximum distance of cancer cells from the inner margin of the SN capsule) predict non-SN involvement.11–14 In breast cancer, this has been studied less extensively.15We therefore set out to study the predictive value of microanatomic location and penetrative depth of SN metastatic deposits for non-SN metastases, accurately assessed by morphometry, in a large series of SN-positive breast cancer patients. Patients And Methods Patients A retrospective database was analyzed, including patients with invasive breast cancer and a tumor positive axillary SN followed by ALND, treated at the University Medical Center Utrecht or the St Antonius Hospital in Nieuwegein from January 2000 to May 2007 (n = 357), including patients from our previous study.16 Exclusion criteria were multicentric tumors, neoadjuvant chemotherapy, and a total of fewer than six lymph nodes examined. SN Biopsy Technique The technical aspects used for the SN procedure are described in detail elsewhere.16 Briefly, before surgery, SN identification was performed by peritumoral injection of 120 MBq 99mTc-Nanocolloid (Amersham Cygne, Eindhoven, The Netherlands) in a maximal volume of 0.5 mL. Dynamic and static scintigraphic images were subsequently obtained. On the same day, immediately preoperatively 0.5 mL Patent blue dye (Guerbet, Aulnay-sous-Bois, France) was injected intradermally and intra/peritumorally. The SN was identified after careful dissection of blue lymphatic channels and detection of radioactivity with a handheld gamma ray detection probe. Palpation of the open axilla was performed to detect enlarged non-SN metastases. Histopathological Evaluation The SNs were processed according to the protocols described previously.16–18 SNs were lamellated according to their size, fixed in neutral buffered formaldehyde, and completely embedded. Step sections 5 μm thick were cut at five levels with 250 μm intervals for staining with haematoxylin and eosin (H&E). In the absence of apparent metastases by H&E examination, immunohistochemistry was performed with CAM 5.2 (Beckton Dickinson, Franklin Lakes, New Jersey, USA) or CK AE1/3 (Dako, Glostrup, Denmark) at each level. All non-SNs were identified visually or by palpation, dissected, processed routinely, and examined at one level with H&E staining. All SNs and non-SNs were examined initially by multiple pathologists at the two institutions, reviewed histologically, and reclassified according to the current 6th edition of the American Joint Committee on Cancer (AJCC) staging system by one observer (CHMvD). All cases were evaluated without knowledge of non-SN involvement. Clinicopathological Features Clinicopathological features recorded included age, pT (TNM system of the AJCC), histological subtype (according to the WHO), histological grade (defined according to the Nottingham modified Bloom–Richardson score based on the percentage of tubule formation, nuclear pleomorphism and mitotic activity), mitotic activity index (MAI),19 steroid receptor and HER-2/neu status (not routinely determined before 2005). SN and Non-SN Characteristics SN characteristics included metastatic size according to the 6th edition of the AJCC staging system [isolated tumor cells (ITC) (≤0.2 mm), micrometastases (>0.2 mm and ≤2 mm), macrometastases (>2 mm)], the number of SNs, ECE, maximal diameter of the largest metastases, microanatomic location of the metastatic deposit, and the penetrative depth. If multiple but distinct deposits were identified in the same SN, the largest metastasis was recorded. If single tumor cells, cluster of nests were continuous, or separated by a few cells distance, they were measured as one focus.20 If more than one SN was involved in an individual patient, the most extensive and/or deepest metastatic deposit was recorded. In case both axillary and internal mammary SNs were involved, the features were measured in the axillary SN. The microanatomic location of metastatic deposits within each SN was classified as subcapsular, combined subcapsular and parenchymal, parenchymal or extensive. Extensive SN involvement, as defined in the study of Ruiter et al.,21 was a deposit >5 mm in diameter (Fig. 1). The centripetal depth was, according to Starz et al.,11 defined as the maximal depth at which tumor cells have infiltrated the SN, as measured from the inner margin of the capsule (Fig. 2), further denoted tumor penetrative depth according to the proposal of Scolyer et al.12 Multifocality was defined as two or more separated metastatic deposits at some distance from each other. All measurements were calculated microscopically in the plane of the tissue sections using interactive video morphometry systems (Q-PRODIT, Leica, Cambridge, UK or Research Video Assistant, Baarn, The Netherlands). FIG. 1.Pattern of distribution of metastatic breast cancer deposits in SNs.FIG. 2.The tumor penetrative depth of metastases was defined as the maximal distance of breast cancer cells from the inner margin of the SN capsule (arrow) (H&E, original magnification × 10). Non-SN characteristics included the total number of non-SNs, maximal tumor diameter, AJCC classification, and ECE. If more than one non-SN was involved, the largest diameter was recorded. Statistical Analysis Statistical analysis was performed using SPSS 13.0 for Windows. Patients were divided into groups with and without non-SN involvement. The Pearson chi-square test was used to determine the relationship between categorical variables (histological type and grade, steroid receptor and HER-2/neu status, number of SNs, AJCC classification, ECE) on the one hand and non-SN status on the other. Continuous data (age, diameter primary tumor, MAI, SN tumor diameter) were analyzed using the Mann–Whitney U-test. P-values <0.05 were considered significant. All relevant variables that were associated with the presence of positive non-SNs were included in a multivariate logistic regression model. SN metastatic characteristics (diameter and penetrative depth) were further compared by receiver operating characteristic (ROC) analysis, calculating the area under curve (AUC) as a measure of discriminative value. Results The median age was 53 years (range 22–86 years) and the mean histological invasive tumor size was 2.4 cm (range 0.3–9.0 cm). Overall, 571 SNs were obtained (mean 1.6 SN per patient) of the 357 patients with a positive SN as well as 4939 non-SNs (mean 14 per patient). The metastatic deposits were subcapsular in 167 patients (47%), combined subcapsular and parenchymal in 40 patients (11%), parenchymal in 11 patients (3%), and extensive in 139 patients (39%) (Fig. 1). Of the 357 patients, 24 (7%) patients had ITC only in the SN, whereas 112 (31%) had micrometastases and 221 (62%) had macrometastases. Localisation of ITCs was subcapsular (96%) or parenchymal (4%). SN micrometastases were located subcapsularly (81%), parenchymally (5%) or had combined localisation (13%). The overall prevalence of non-SN involvement was 38% (136/357 patients). Other descriptive characteristics of the study population are listed in Tables 1 and 2. TABLE 1.Clinicopathological characteristics of 357 invasive breast cancer patients with a positive SN and subsequent axillary lymph node dissectionFeatureNo.%Mean age (range)54 y (22–86)Primary tumorMean tumor size (range)2.4 cm (0.3–9.0) pT118752 pT215042 pT3195 pTx10Histological subtype Invasive ductal cancer29884 Invasive lobular cancer3610 Others236Histological grade (B&R) 17421 216145 312234MAI, mean/2 mm2 (range)13 (0–102)Steroid receptor status ER – positive ‡32290 ER – negative3410 ER – unknown 1 PR – positive ‡28480 PR – negative7020 PR – unknown3HER-2/neu status Positive2615 Negative15085 Unknown181‡ ≥10% immunoreactive neoplastic cells.TABLE 2.SN and non-SN characteristics of 357 invasive breast cancer patients with a positive SN and subsequent axillary lymph node dissectionFeature No. %SNTotal number of SNs571Mean number of SNs 1.6Total number of positive SNs419Mean diameter SN metastases 4.7 mmAJCC classification of SN metastatic size ITC247 Micrometastasis11231 Macrometastasis22162Extracapsular extension No25772 Yes10028Microanatomic location Subcapsular16747 Parenchymal 113 Combined4011 Extensive13939Mean penetrative depth2.9 mmNon-SNsTotal number of non-SNs4939Mean number of non-SNs (range)14 (5–38)Total number of positive non-SNs474Total number of negative non-SNs4465Size of non-SN metastases ITC32 Micrometastases3324 Macrometastases10074 The following factors were significant predictors of non-SN metastases by univariate analysis: primary tumor size, number of involved SNs, ECE, AJCC classification, diameter, and penetrative depth of the SN metastatic deposit (Tables 3 and 4). None of the other classic variables (age, histological subtype and grade, steroid receptor and HER-2/neu status, MAI) of the primary tumor correlated significantly with non-SN involvement. TABLE 3.Comparison of categorical clinicopathological and SN characteristics in invasive breast cancer patients without and with non-SN metastases by Pearson chi-square testWithout non-SN metastasesWith non-SN metastasesP valueFeatureNo (%)No (%)Total 221 (62)136 (38)Histological subtype0.914 Ductal183 (51)115 (32) Lobular22 (6)14 (4) Others16 (5)7 (2)Histological grade (B&R)0.798ER positive198 (56) 124 (35)0.715 Negative22 (6)12 (3)PR positive179 (50)106 (30)0.589 Negative41 (12)29 (8)HER-2/neu positive16 (9)10 (6)0.670 Negative83 (47)67 (38)No. of SN0.246 1131 (37)89 (25) >190 (25)47 (13)No. of involved SN 0.050 1198 (55)112 (31) >123 (6)24 (7)AJCC SN metastases<0.001 ITC21 (6)3 (1) Micrometastasis86 (24)26 (7) Macrometastasis114 (32)107 (30)Extracapsular extension<0.001 No175 (49)82 (23) Yes46 (13)54 (15)Microanatomic location SN metastases<0.001 Subcapsular126 (35)41 (11) Combined23 (6)17 (5) Parenchymal8 (2)3 (1) Extensive64 (18)75 (21)Multifocality0.087 No176 (49)118 (33) Yes45 (13)18 (5)TABLE 4.Comparison of continuous clinicopathological and SN characteristics in invasive breast cancer patients without and with non-SN metastases by Mann–Whitney U-testWithout non-SN metastasesWith non-SN metastasesP valueFeatureTotal no. (%)221 (62)136 (38)Age (years, mean)54530.249Primary tumor diameter, cm (mean)2.32.50.019MAI, 2mm2 (mean)13120.693SN tumor diameter, mm (mean) 3.66.5<0.001SN tumor penetrative depth, mm (mean) 2.23.9<0.001 The primary tumor features histological grade and diameter were associated with multiple SN tumor deposits (P = 0.03 and 0.05, respectively). Frequency of non-SN metastases in patients with SN ITC (N = 24), micro- (N = 112) and macrometastases (N = 221) was 12.5%, 23%, and 48%, respectively (P < 0.001, Fig. 3). Of those three patients with SN ITC and involved non-SNs, two had a non-SN micro- and one had a non-SN macrometastasis. One of these ITC was located in the parenchyma, the other two patients had a subcapsular location. Two of these three patients with SN ITC and non-SN involvement showed multiple small cell clusters and single cells in the SN. FIG. 3.Flow chart showing distribution of SN AJCC classification according to non-SN involvement. The microanatomic localization of SN metastatic deposits correlated with non-SN involvement. Patients with subcapsular (N = 167), combined subcapsular and parenchymal (N = 40), parenchymal (N = 11) and extensive (N = 139) tumor deposits showed non-SN involvement in 25%, 42%, 27%, and 54% of cases, respectively (Table 3). Morphometrically assessed penetrative depth of SN metastases was also associated with non-SN involvement. Frequency of non-SN metastases in patients with a SN tumor penetrative depth <0.155 mm (N = 29), 0.155–2.7 mm (N = 181) and >2.7 mm (N = 147) was 10%, 28%, and 56%, respectively (P < 0.001, Table 5). In ROC analysis, diameter (AUC = 0.686) and the penetrative depth (AUC = 0.680) of the SN tumor deposit had comparable discriminative value (Fig. 4). TABLE 5.Predictive value of penetrative depth of SN metastases for non-SN involvement in patients with invasive breast cancerSN penetrative depth (mm)NNo. of patients with non-SN involvement (%)<0.155293 (10)0.155–2.718151 (28)>2.714782 (56)FIG. 4.ROC curves showing the sensitivity and specificity of SN tumor diameter and penetrative depth as predictors of non-SN involvement. The larger the area below the curve, the more accurate the prediction of non-SN involvement (P < 0.001). In multivariate analysis, SN tumor diameter (P = 0.032) and SN tumor penetrative depth (<0.155 mm versus 0.155–2.7 mm versus >2.7 mm) (P = 0.015) were significant in predicting non-SN involvement. Cutoff values were interactively statistically as those best discriminating between low and high risk of non-SN metastases. Risk stratification by combining these features identified a low-risk group, an intermediate-risk group, and a high risk group for non-SN metastases with frequency of non-SN metastases in 11%, 29%, and 56% of patients, respectively (Table 6). Frequencies of non-SN involvement in patients with SN micrometastases and a subcapsular (N = 91), combined subcapsular and parenchymal (N = 15), and parenchymal location (N = 6) were 21%, 40%, and 17%, respectively. Frequencies of non-SN involvement in patients with SN macrometastases and a subcapsular (N = 54), combined subcapsular and parenchymal (N = 25), parenchymal (N = 4) and extensive location (N = 138) were 39%, 44%, 25%, and 54%, respectively. TABLE 6.Predictive value of a combination of AJCC classification of SN metastases and penetrative depth of SN metastases for non-SN involvement in patients with invasive breast cancerNNo. of patients with non-SN involvement (%)ITC/SN penetrative depth < 2.7 mm or Micrometastases/penetrative depth < 0.155 mm384 (11)Micro- or macrometastases/penetrative depth 0.155–2.7 mm17250 (29)Micro- or macrometastases/penetrative depth >2.7 mm14782 (56) Discussion The lymphatic spread of breast cancer cells has been shown to follow an orderly progression via the SN to non-SNs, which implies that the risk of spread of tumor from the SN to the non-SN may depend on the extent of SN involvement. While many studies have focused on the maximal diameter of the metastatic tumor deposit22, 23 we investigated the microanatomic location and the penetrative depth of the metastatic deposits as a putative predictor of non-SN involvement. This was based on the concept that within the SN, tumor cells also follow an orderly route, arriving in the subcapsular sinuses through an afferent lymph vessel. Later, there is subcapsular outgrowth of malignant cells in the marginal sinuses and into the cortical parenchyma. Finally, these cells extend to the deeper zones of the lymph node parenchyma, frequently following the medullary sinuses to efferent lymph vessels.24 Consistent with this concept we found that the microanatomic location, the size, and the penetrative depth of SN tumor deposits were correlated significantly with non-SN involvement. This finding is consistent with recent melanoma studies. Startz et al.11 proposed a micromorphometric classification, based on the depth of the metastasis from the capsule and the number of 1-mm slices containing melanoma. This classification was a highly significant predictor for distant metastases and overall survival. Dewar et al.13 also recorded that the microanatomic location of melanoma SN metastases predicts non-SN involvement, and proposed that it would be possible to safely avoid a lymph node dissection in patients with subcapsular deposits only. Indeed, subcapsular location and small tumor penetrative depth correlated with less non-SN involvement in our breast cancer study, although no subgroup of patients could be selected without non-SN involvement. The measurement of the SN tumor penetrative depth was difficult in many cases, especially in case of extensive tumor deposits. In these cases the tumor deposits frequently extended beyond the center of the lymph node, making it difficult to determine which edge of the SN capsule should be used to measure the tumor penetrative depth. Similar difficulties were encountered when the SN had a lobulated outline. Further studies which more rigorously define the tumor penetrative depth may strengthen its predictive power and reproducibility. In conclusion, patients with breast carcinoma and SN involvement can be stratified into subgroups at significantly different risk for non-SN involvement, according to microanatomic localization of the SN metastatic deposits and penetrative depth into the SN. However, based on these features no subgroup of patients could be selected without non-SN involvement.	[ "breast cancer", "sentinel node", "morphometry", "axillary lymph node metastases" ]	[ "P", "P", "P", "R" ]
Int_J_Colorectal_Dis-4-1-2225996	NF-κB-dependent synergistic regulation of CXCL10 gene expression by IL-1β and IFN-γ in human intestinal epithelial cell lines	Background and aims Little is known about the intestinal epithelial expression and secretion of CXCL10 (IP-10), a chemokine involved in recruiting T cells and monocytes. We aimed to study CXCL10 gene expression and regulation by the pro-inflammatory cytokines interleukin (IL)-1β, interferon (IFN)-γ and tumour necrosis factor (TNF)-α in intestinal epithelial cell lines. Introduction Intestinal epithelial cells (IECs) form the first immunologic barrier of the organism while the villus surface is continuously in contact with various agents. IEC gene expression is tightly regulated to induce appropriate responses to pathogens and to avoid deleterious reactions to microbial flora. When stimulated with pro-inflammatory cytokines or infected with microbial pathogens, IECs may secrete chemokines, a set of low-molecular weight cytokines, directing migration and activation of leucocytes, which play a major role in the perpetuation of inflammatory processes [1–5]. The final composition of leucocytes present in the inflamed intestine is most likely due to both secreted chemokines and the relative expression of chemokine cell surface receptors on different cell types. Chemokine production by colonic epithelial cells is thought to contribute to the characteristic increased infiltration of selected populations of leucocyte cells in inflammatory bowel disease (IBD) [6, 7]. CXCL10 is a CXC chemokine, known to be secreted by endothelial cells [8], keratinocytes [9], eosinophils [10] and neutrophils [11]. It is also constitutively expressed by normal human colon epithelium [12] and chemoattracts activated T cells as well as monocytes by binding to the CXCR3 receptors present on the respective cells [13, 14]. While CXCL10 up-regulation may occur in acute inflammation in response to pathogens, it is also known to be permanently over-expressed in IBD patients [15, 16]. Blocking CXCL10 could prevent mice from acute colitis, suggesting this chemokine to play an important role in IBD pathogenesis [17]. Moreover, CXCL10 is believed to possess anti-tumoural properties by promoting anti-tumour T cell immunity [18, 19]. Knowledge of its regulation in colonic cell lines may therefore give also insight into the pathobiology of colorectal cancer. Most previous studies on chemokine expression in IEC concentrated on CXCL8 (interleukin [IL]-8), a neutrophil chemoattractant [20, 21], but there were only few studies regarding the expression and secretion of CXCL10 in IECs [12]. Although it is well known that interferon (IFN)-γ is a major inducer of CXCL10 and that CXCL10 expression may be enhanced by pathogens and other inflammatory cytokines, the relative contribution of cytokines on CXCL10 induction, their synergistic action and the kinetics of cytokine-induced CXCL10 expression in intestinal epithelial cells has not been fully elucidated. Especially, the role of IL-1β in CXCL10 gene regulation in IECs and also in other cell types has not been fully evaluated so far. Intestinal mucosa is composed of cell populations in continuous change from a proliferative and undifferentiated stage in the basal parts of crypts to mature surface villus epithelial cells [22, 23]. Little is known about the alteration of immunological functions as IECs mature. It has been shown that cellular differentiation affects IEC responsiveness to IL-1β [24]. In the present study, we aimed to elucidate the role of the pro-inflammatory cytokines IL-1β, tumour necrosis factor (TNF)-α and IFN-γ, which have established role in IBD pathogenesis, in inducing CXCL10 gene expression in IEC lines. Especially, the role of IL-1β, one of the most abundantly expressed cytokines, both in normal and inflamed mucosa [25], in inducing CXCL10 expression by IECs and its possible signalling mechanisms was given importance. To consider possible influences of cellular differentiation on cytokine responsiveness, we performed studies on three different colonic epithelial cell lines. Our findings will get an insight into the role of intestinal epithelium in immune responses and in evaluating pathways, which might be targets of present and future pharmacotherapy of IBD and possibly colorectal cancer. Materials and methods Biological reagents The recombinant cytokines IL-1β, TNF-α and IFN-γ were purchased from Roche applied science (Mannheim, Germany). An enzyme-linked immunosorbent assay (ELISA) kit for CXCL10 was purchased from RnD systems (Weisbaden, Germany). The nuclear factor (NF)-κB inhibitor BAY11-7082 was purchased from Calbiochem (San Diego, CA). Cell culture and stimulation protocols The human colon adenocarcinoma cell lines Caco-2, HT-29 and DLD1 were obtained from DSMZ (Braunschweig, Germany). Caco-2 cells were grown in Eagle’s minimal essential medium (BioWhittacker, Germany) containing 20% foetal calf serum (FCS) supplemented with 100 U/ml each of penicillin and streptomycin and 1% non-essential amino acids at 37°C and 5% CO2. HT29 and DLD1 were grown in Roswell Park Memorial Institute medium containing 10% FCS and 100 U/ml penicillin and streptomycin at 37°C and 5% CO2. Caco-2, HT29 and DLD-1 cells were plated into six-well plates at a density of 5 × 105 cells per well for real-time polymerase chain reactions (PCRs), ELISA, Northern blotting and electrophoresis mobility shift assay (EMSA) experiments unless until stated and grown till they reached 70–80% confluence. These cells were then stimulated with IL-1β (0.1, 0.5, 1, 2, 10 and 50 ng/ml), TNF-α (2, 10 and 50 ng/ml) and IFN-γ (2, 10 and 50 ng/ml) based on the type of experiments. RNA isolation, real-time PCR Ribonucleic acid (RNA) was isolated by using Qiagen RNeasy mini kit according to the manufacturer’s protocol. The RNA concentrations were determined photometrically using a Gene Quant RNA/deoxyribonucleic acid (DNA) calculator (Pharmacia, Freiburg, Germany). RNA was subsequently used for real-time PCR and Northern blot analysis. Reverse transcription of messenger RNA (mRNA) was performed using 1 μg of total cellular RNA. To determine the mRNA expression of CXCL10, real-time PCR was carried out using gene-specific primers for human CXCL10 (Invitrogen GmbH, Karlsruhe) forward 5′-CCA GAA TCG AAG GCC ATC AA-3′, reverse 5′-CAT TTC CTT GCT AAC TGC TTT CAG-3′ and β-actin (Invitrogen GmbH) forward 5′-CTG GCA CCC AGC ACA ATG-3′, reverse 5′-CCG ATC CAC ACG GAG TAC TTG-3′ in an ABI Prism 7000 sequence detection system. PCR reaction was set up with Sybr® Green PCR Master mix (Invitrogen) containing 0.3 μmol/l primers each and 1 μl of RT-product in 25 μl volume. A two-step amplification protocol was chosen consisting of initial denaturation at 95°C for 10 min followed by 45 cycles with 15 s denaturation at 95°C and 30 s annealing/extension at 60°C. Finally, a dissociation protocol was performed to control specificity of amplification products. Relative expression of CXCL10 was then calculated using the comparative threshold-cycle (CT) method. The amount of target mRNA in each sample was normalized to the amount of β-actin mRNA designated as calibrator, to give ΔCT (CT β-actin − CT CXCL10). The relative expression of CXCL10 was calculated as the . MRNA expression is presented as fold increase calculated in relation to unstimulated cells after normalization to β-actin. Transient transfections Caco-2, HT29 and DLD1 cells were plated in 24-well plates, 24 h before transfection (Costar, Corning, NY), at a density of 50,000/well in 1 ml medium. After 24 h, the cells were transfected with 250 ng of respective plasmids along with 20 ng of renilla plasmid using the non-liposomal formulation FuGENE (Roche Molecular Biochemicals, Mannheim, Germany). On the day of transfection, fresh medium was added. FuGENE was added to the plasmid DNA at a ratio of 3 μl/μg DNA. FuGENE was prediluted in 100 μl serum-free medium and added drop-wise to the concentrated plasmid DNA. After 15–25 min at room temperature, the mixture was added to the cells. The plasmids TGL-IP10, TGL-IP10 κB2 Mut were kind gifts from Dr. Richard M. Ranshoff (Cleveland, OH) and were described elsewhere [26]. Luciferase assays Twenty-four hours after transfection, cells were stimulated with IL-1β (0.1,1, 2, 10 and 50 ng/ml), TNF-α (2, 10 and 50 ng/ml), IFN-γ (2, 10 and 50 ng/ml), IL-1β (1 ng/ml) + TNFα (50 ng/ml), IL-1β (1 ng/ml) + IFN-γ (50 ng/ml) and TNFα (50 ng/ml) + IFN-γ (50 ng/ml) for 24 h, harvested and lysed, and luciferase was assayed by using a luciferase kit (Promega). Values were normalised to renilla luciferase. Enzyme-linked immunosorbent assay To study the synergistic effects of cytokines on CXCL10 secretion, supernatants were collected at 0, 3, 6, 12 and 24 h in Caco-2 and HT29 cells stimulated with IL-1β (1 ng/ml), TNF-α (50 ng/ml), IFN-γ (50 ng/ml), IL-1β (0.5 ng/ml) + TNFα (50 ng/ml), IL-1β (0.5 ng/ml) + IFN-γ (50 ng/ml) and TNFα (50 ng/ml) + IFN-γ (50 ng/ml) and stored at −70°C until measured. CXCL10 protein secretion was measured by ELISA according to the manufacturer’s protocol. Preparation of nuclear extract Caco-2 cells were grown for 24 h, pre-treated for 1 h with BAY 11-7082 (10 μM) and then stimulated with respective cytokines for 30 min. Cells were washed with 10 ml cold phosphate-buffered saline (PBS) twice, harvested, centrifuged at 1,000 × g for 5 min and resuspended in 1 ml of cold PBS. Centrifuged briefly, excess PBS was removed, and the pellet was resuspended in ice-cold extraction buffer (20 mM hydroxyethyl piperazineethanesulfonic acid [HEPES] pH 7.8, 10 mM KCL, 0.1 mM ethylenediamine tetraacetic acid [EDTA], 1 mM dithiothreitol and 0.5 mM phenylmethanesulphonylfluoride [PMSF]) by gentle pippeting. Cells were then kept on ice for 15 min, and 30 μl of 10% Nonidet NP-40 was added and vortexed briefly for 10 s. The lysed cell suspension was centrifuged for 30 s, and the supernatant was discarded. The pellet was resuspended in 50 μl extraction buffer containing 400 mM NaCl, agitated vigorously for 15 min at 4°C. Then, the cell suspension was centrifuged for 5 min at 4°C, and the supernatant containing the nuclear extract was collected and stored at −80°C until further use. Electrophoretic mobility shift assay and super-shift assay For binding reactions, 5 μg of nuclear extracts were incubated in 20 μl reaction mixture containing 40 mM HEPES pH 7.5, 50 mM NaCl, 1 mM EDTA, 1mM DithitreitolHSCl, 1 mM PMSF, 1.28 μg/ml pBluescript vector DNA as half of the reaction and the rest being nuclear extract plus other components. Oligonucleotide duplex probes (Promega) were end-labelled with T4 polynucleotide kinase and [γ32P] adenosine triphosphate; 1 × 104 cpm were then added to the reaction mixtures and incubated overnight at 4°C. Reaction products were analysed by non-denaturing electrophoresis in a 5% polycralamide gel with 0.5× Tris–borate–EDTA buffer at room temperature. Gels were then exposed directly to X-ray film at −70°C for autoradiography. For competition, unlabelled oligonucleotides were added in molar excess at room temperature for 15 min before adding radiolabelled probe. In super shift experiments, nuclear extracts were incubated with anti-p65 anti-body (Santa Cruz Biotechnology, Santa Cruz, CA) overnight at 4°C. Statistical analysis Statistical analyses were carried out using the Prism software packet (Version 3.0, Graphpad Software, San Diego, CA). Comparisons between two or more different treatment groups were made with the non-paired t test or analysis of variance (ANOVA) where appropriate. For time kinetics experiments, a two-way ANOVA test followed by the Bonferroni post-test was performed. In case of RNA expression, a log transformation was performed before. Statistical differences were regarded significant at a p level below 0.05. Data were expressed as means ± standard error of the mean (SEM). Results Differential expression of CXCL10 mRNA in intestinal epithelial cell lines under basal conditions In the present study, the basal mRNA expression of CXCL10 in three different IECs, Caco-2, HT29 and DLD1 cells, were studied first. Real-time PCR experiments using gene-specific primers for CXCL10 revealed that the basal mRNA expression of CXCL10 was highest in Caco-2 followed by HT29 and DLD1 (Fig. 1). Fig. 1Basal level expression of CXCL10 in IECs. Caco-2, HT29 and DLD1 cells were plated at a density of 5 × 105 cells in six well plates, grown for 24 h and total RNA from the cells was isolated. First-strand complementary DNA was prepared from 1 μg of total RNA. Real-time PCR was performed for CXCL10 and β-actin was used as an internal control. The bars indicate the mean of six individual experiments. Asterisk corresponds to p < 0.01 IL-1β, TNF-α and IFN-γ dose-dependent influence on CXCL10 gene expression in IECs Caco-2, HT29 and DLD1 cells were tested for their ability to respond to pro-inflammatory cytokines by the induction of CXCL10 mRNA expression. Cells were stimulated for 4 h with various concentrations of cytokines, and CXCL10 expression was analysed by real-time PCR. In response to IFN-γ and TNF-α, Caco-2, HT29 and DLD1 cells showed a dose-dependent induction of CXCL10 mRNA (Fig. 2). Increased expression of CXCL10 mRNA was observed when 2 ng/ml IL-1β was used, while a higher concentration did not further enhance expression. Further, we checked the response of IECs to lower doses of IL-1β showing a dose response in the range of 0.1 to 1 ng/ml in the induction of CXCL10 mRNA. IL-1β at a concentration of 0.5 ng/ml was able to induce maximum expression of CXCL10 mRNA at 4 h (data not shown). Fig. 2Dose-dependent expression of CXCL10 mRNA in IECs stimulated with cytokines. 5 × 105 cells were plated into six-well plates and grown for 24 h and then stimulated with IL-1β, TNF-α and IFN-γ at an increasing doses of 2, 10 and 50 ng/ml. The cells were harvested after 4 h, total RNA from the cells was isolated, and first-strand cDNA was prepared from 1 μg of total RNA. Real-time PCR was performed for CXCL10, and β-actin was used as an internal control. Graph represents the mean ± SEM of four individual experiments. All cytokine-stimulated expressions were significantly different from unstimulated controls (p < 0.05). Asterisk, no significant difference between different IL-1β concentrations (p < 0.05) IL-1β, TNF-α and IFN-γ time-dependent influence on CXCL10 gene expression in IECs Once we confirmed the responsiveness of IECs to pro-inflammatory cytokines, we next examined the mRNA kinetics of CXCL10 induced by the cytokines alone or in combination with each other. Each cell line had a different time course of response when stimulated with respective cytokines and their combinations (Fig. 3a–c). In Caco-2, stimulated with IL-1β or TNF-α, CXCL10 mRNA was induced within the first 40 min, and the induction reached to a peak after 4 h, while IFN-γ had a later initial effect inducing CXCL10 mRNA after 2 h and reached to a maximum at 8 h. Combination of IL-1β + IFN-γ had a stronger synergistic effect in the induction of CXCL10 mRNA within 20 min and reaching to maximum after 4 h, whereas TNF-α + IFN-γ synergistically led to a later induction of CXCL10 mRNA starting at 40 min but had a strong synergistic effect after 8 h when compared to IL-1β + IFN-γ (Fig. 3a). In HT29 (Fig. 3b) and DLD1 (Fig. 3c), CXCL10 mRNA time kinetics deviated from that in Caco-2 (Fig. 3a). CXCL10 mRNA was induced after 1 h of stimulation with the respective single cytokines in DLD1. In case of HT29, either TNF-α or IFN-γ alone led to delayed induction of CXCL10 mRNA after 1 h, whereas IL-1β alone led to an induction starting at 2 h. HT29 and DLD1 cells treated with combinations of IL-1β + IFN-γ showed a strong synergistic induction of CXCL10 mRNA within 40 min, whereas the TNF-α + IFN-γ combination led to an even stronger induction of CXCL10 mRNA within 20 min in HT29 and within 40 min in DLD1. IL-1β + TNFα in combination had no synergistic effect on CXCL10 mRNA induction in all the cell lines used for this study. The real-time PCR results could be confirmed by Northern blot analysis in Caco-2 and HT29 cells. However, the detection threshold was not low enough to display basal and single cytokine-induced CXCL10 mRNA (data not shown). Fig. 3Time kinetics of CXCL10 mRNA expression in IECs stimulated with cytokines. a Caco-2, b HT29 and c DLD1 cells were plated at a density of 5 × 105 cells in six-well plates and grown for 24 h and then stimulated with respective cytokines and their combinations. Cells were harvested at different time points as indicated, and total RNA from the cells was isolated and first-strand cDNA was prepared from 1 μg of total RNA. Real-time PCR was performed for CXCL10, and β-actin was used as an internal control. Graphs represent the mean ± SEM of three to five experiments. IFN-γ led to a significant increase in CXCL10 expression from 2 h onwards. A significant increase (p < 0.05) from control was observed in all expect in those marked as N.S. (non-significant). Number sign indicates time points with significant differences (p < 0.05) of cytokine combinations or IL-1β compared to IFN-γ alone IL-1β, TNF-α and IFN-γ time-dependent influence on CXCL10 protein secretion in IECs To test whether cytokine stimulation of IECs would lead not only to the induction of CXCL10 mRNA but also to the synthesis and secretion of the protein, Caco-2, HT29 and DLD1 cells were incubated with respective cytokines (from 3 to 24 h), and secretion of CXCL10 was detected by ELISA. Under basal conditions, Caco-2 cells secreted 0.012 ± 0.003 ng/ml CXCL10. Treatment with IL-1β, TNF-α and IFN-γ led to a time-dependent increase in CXCL10 protein, with IL-1β being the most potent stimulus during the first 6 h (Fig. 4a). Of importance, after 12 h, CXCL10 secretion was stimulated strongly by IFN-γ. Co-incubation of Caco-2 with respective cytokines revealed a synergistic effect of IL-1β + IFN-γ and TNF-α + IFN-γ on CXCL10 secretion, reaching levels of 3.354 ± 0.690 and 7.808.1 ± 0.173 ng/ml after 24 h. The combination of IL-1β + TNF-α had no synergistic effect on CXCL10 secretion when compared to the respective cytokines alone. These time kinetic studies revealed that in Caco-2, IL-1β alone or in combination with IFN-γ induced the secretion of CXCL10 very early compared to the combination of TNF-α + IFN-γ, which induced a delayed but stronger synergistic secretion of CXCL10 protein, suggesting that in case of IL-1β + IFN-γ, a synergism could take place very early, e.g., at the level of intracellular signal cascades. In HT29 cultures, CXCL10 secretion was not detectable under basal condition. TNF-α or IFN-γ alone led to an induction of CXCL10 protein secretion within 3 h (0.0172 ± 0.005 and 0.011 ± 0.001 ng/ml), whereas IL-1β could not induce the secretion of CXCL10 protein. IL-1β, TNF-α and IFN-γ led to a time-dependent increase in CXCL10 concentrations in HT29 cell supernatants reaching maximum levels of 0.039 ± 0.0057, 0.632 ± 0.0222 and 19.929 ± 0.838 ng/ml showing that IFN-γ is the strongest inducer of CXCL10 protein secretion when compared to other two cytokines in this cell line (Fig. 4b). The combinations of TNFα + IFN-γ and IL-1β + IFN-γ induced CXCL10 protein secretion to 1.822 ± 0.109 and 0.410 ± 0.027 ng/ml within 3 h and led to a maximum of 492.728 ± 83.776 and 87.158 ± 7.913 ng/ml after 24 h (Fig. 4b). Fig. 4Time kinetics of CXCL10 protein secretion in Caco-2 and HT29 cells stimulated with cytokines. 5 × 105 cells were plated into six-well plates and grown for 24 h and then stimulated with IL-1β, TNF-α and IFN-γ as indicated. Supernatants were collected at respective time points from a Caco-2 and b HT29 cells and ELISA for CXCL10 protein was performed. Data represent means of three to five experiments ± SEM done in duplicates for each sample. Asterisk indicates time points with significant differences (p < 0.05) of cytokine combinations compared to IFN-γ alone CXCL10 promoter activation by cytokines TGL-IP10, a promoter-reporter containing 435 bp of human CXCL10 sequence upstream of the transcriptional start site, was used to evaluate the role of IL-1β, TNF-α and IFN-γ alone or in combination with each other in the transcription of CXCL10. When TGL-IP10-transfected Caco-2 cells were stimulated with single cytokines, IL-1β (1 ng/ml) led to a stronger induction (6.53 ± 1.65-fold increase) than TNF-α (50 ng/ml) and IFN-γ (50 ng/ml) alone (2.76 ± 0.53 and 3.14 ± 1.2-fold increase). The combination of IL-1β + IFN-γ showed a stronger synergistic induction (26.18 ± 11.68) than the combination of TNF-α + IFN-γ (13.97 ± 3.25-fold increase) (Fig. 5a). In HT29 cells, when stimulated with single cytokines, IFN-γ (50 ng/ml) led to a maximum induction (6.52 ± 0.58-fold increase), followed by TNF-α (4.44 ± 0.16-fold increase) and IL-1β (1.43 ± 0.05-fold increase). The combination of IL-1β + IFN-γ synergistically led to 12.76 ± 3.21-fold induction, and TNF-α + IFN-γ led to 26.70 ± 7.5-fold induction of promoter activity (Fig. 5b). In DLD1 cells (Fig. 5c), IFN-γ (50 ng/ml) led to a maximum induction (8.06 ± 0.51-fold increase) of CXCL10 promoter activity followed by TNF-α (50 ng/ml) induction (3.59 ± 0.08-fold increase) and IL-1β (1 ng/ml) induction (2.32 ± 0.12-fold increase). The combination of IL-1β + IFN-γ led to a 40.72 ± 12.19-fold induction, and TNF-α + IFN-γ led to a 20.56 ± 6.27 fold induction. In all three cell lines, the combination of IL-1β + TNF-α had no synergistic effect on CXCL10 promoter activity. Fig. 5Transient transfection analysis of TGL-IP10 in IECs. The CXCL10 promoter containing reporter construct, pTGL-IP10 and pRL-TK (renilla) plasmids, was transiently transfected into a Caco-2, b HT29 and c DLD1 cells. 24 h after transfection, cells were stimulated with cytokines as indicated for 24 h, harvested, and promoter activity was monitored by luciferase assay. Data were normalised with renilla luciferase values and included as fold increase vs control. Data represent the means ± SEM of three to five independent experiments done in duplicates for each sample. Asterisk indicates an induction significantly different (p < 0.05) from control (C). Significant differences between combinations of cytokines and the respective single cytokines are indicated with bars with a number sign (p < 0.05). N.S. is not significant Evaluating the role of NF-κB in CXCL10 gene expression To evaluate the possible role of NF-κB in cytokine-induced CXCL10 gene expression, experiments with a specific NF-κB inhibitor were performed. The effect of BAY 11-7082 on NF-κB inhibition was evaluated by reporter assays. BAY11-7082 was able to dose-dependently inhibit IL-1β-induced pNF-κB-SEAP gene reporter activity in Caco-2 cells (data not shown). Based on this experiment, we could determine that 10 μM of BAY11-7082 was sufficient to inhibit the NF-κB activation. The induction of CXCL10 gene expression by cytokines was inhibited by addition of the NF-κB-specific inhibitor BAY11-7082 in Caco-2, HT29 and DLD1 cells, with a varying degree of inhibition in the individual cell lines. Inhibition of NF-κB led to almost complete inhibition (approximately 90–95% inhibition) of CXCL10 mRNA and protein secretion induced by cytokines in Caco-2 and DLD1 cells (Fig. 6a and c). In DLD1 cells, IL-1β (1 ng/ml) did not have any inducing effect on CXCL10 secretion. In HT29, inhibition of NF-κB had approximately 60–75% inhibitory effect on cytokine-induced CXCL10 mRNA expression and secretion (Fig. 6b). Fig. 6Role of NF-κB in cytokine-induced CXCL10 mRNA expression by IECs. 5 × 105 cells were plated and grown for 24 h. Then, the cells were pre-treated with 10 μM of BAY 11-7082 for 1 h and then stimulated with respective cytokines and their combinations as stated in the figure for 2 and 24 h. After 2 h, cells were harvested for RNA isolation to perform real-time PCRs, and at 24 h, cell supernatants were collected and stored at −70°C until ELISA was performed for CXCL10. Data presented are the means ± SEM of three independent experiments done in duplicates for each sample in a Caco-2, b HT29 and c DLD1 cells. Significant CXCL10 inhibition by BAY 11-7082 is indicated by an asterisk (p < 0.05). N.S. is not significant To further evaluate the role of NF-κB, transient transfections were done with a CXCL10 promoter construct bearing a mutated NFkB-binding site (TGL-IP10 κB2 Mut). In Caco-2 cells, TGL-IP10 κB2 Mut was not responsive to any of the cytokines and their combinations, suggesting that CXCL10 gene induction is strongly dependent on NF-κB in this cell line (Fig. 7a). In HT29 and DLD1 cells, the combination of TNF-α + IFN-γ led to 6.49 ± 2.13 and 4.87 ± 1.43-fold increase, and the combination of IL-1β + IFN-γ led to 2.26 ± 0.63 and 4.75 ± 0.53-fold increase of the TGL-IP10 κB2 Mut promoter activity (Fig. 7b and c), revealing that a smaller portion of CXCL10 gene expression might be regulated independently from NF-κB. Fig. 7Transient transfection analysis of TGL-IP10 and TGL-IP10 κB2 mutant in IECs. TGL-IP-10 and TGL-IP10-κB2 Mut plasmid was transiently transfected into a Caco-2, b HT29 and c DLD1. 24 h after transfection, cells were stimulated with cytokines for 24 h, and promoter activity was monitored by luciferase assay. Data were normalised with renilla luciferase values. Data presented are the means ± SEM of three to five independent experiments done in duplicates for each sample. Significant differences vs controls (C) are indicated by an asterisk (p < 0.05). Significant differences of wild type vs mutant are indicated by a number sign (p < 0.05). N.S. is not significant To confirm further our results from promoter and inhibitor studies, we did gel-shift and super shift assays to show that NF-κB is activated by cytokines and that the activated NF-κB complex contains the NF-κB functional sub-unit p65. In the nuclear extracts of Caco-2 cells, we could clearly see a shift in the samples treated with IL-1β, TNF-α alone or the combinations of IL-1β + IFN-γ and TNF-α + IFN-γ. The protein-DNA complex was efficiently competed by unlabelled κB oligonucleotides. There was no shift in the samples treated with the NF-κB inhibitor BAY11-7082, stating that the shift was indeed because of NF-κB binding. Furthermore, antibody super shift assay using anti-p65 antibody had clearly shown that the complex contains NF-κB functional sub-unit, p65 (Fig. 8). Fig. 8Gel shift and super shift analysis for NF-κB. Caco-2 cells were seeded and grown for 24 h before pre-treating with BAY 11-7082 (10 μM) for 1 h and then stimulated with respective cytokines for 30 min; cells were then harvested, and nuclear extracts were prepared. 5 μg of nuclear extracts were used for gel shifts with a wild-type κB oligonucleotide. Super shifts were performed with anti-p65 antibody. Similar results were obtained in two separate experiments Discussion In addition to its many homeostatic functions such as barrier protection, mucus production and resorption, the intestinal epithelium plays an important role in regulating local inflammation and immune responses. The intestinal epithelium participates in inflammatory responses in part through the generation of numerous cytokines and chemokines that mediate recruitment and activation of inflammatory cells. The epithelium is involved in local cytokine networks allowing first response to noxious agents as well as cross-talk with immune cells to help provide an effective inflammatory response. These bi-directional inflammatory communications are necessary for normal host defence, but they also play an important role in the pathogenesis of IBD and might be also of importance in modulating anti-tumoural immunity in colonic malignancies. In this report, we demonstrate that pro-inflammatory cytokines, which may be produced by sub-mucosal immune cells, can differentially and co-ordinately regulate CXCL10 gene expression in IECs. CXCL10 is one of several chemokines produced by IEC. Previous studies could show that IFN-γ in combination with TNF-α caused induction of RANTES (regulated upon activation, normal t-cell expressed, and secreted protein) and monocyte chemotactic protein-1 in colorectal tumour cell lines, while IFN-γ had no effect on CXCL8 (IL-8) secretion [27], the latter observation being confirmed in all three cell lines in the present study (data not shown). Dwinell et al. [12] had previously shown that stimulation of IECs with TNF-α or IL-1α strongly potentiated IFN-γ-induced CXCL10. It is well known that among IL-1α and IL-1β, the latter is secreted and has an active role in inflammation. Importantly, IL-1β was shown to be over-expressed in IBD patients [25, 28]. This makes IL-1β an interesting cytokine to study its effect on CXCL10 gene regulation. Although the synergistic effects of TNF-α + IFN-γ on CXCL10 gene expression and the signalling mechanisms involved were studied in astrocytes [26] and their synergistic effect on inducible nitric oxide synthase expression in IECs and macrophages [29] had been studied in detail, the synergistic action of these cytokines on CXCL10 gene expression and the signalling mechanisms involved have not been evaluated in IECs. Especially, the role of IL-1β alone and its synergism with both TNF-α and IFN-γ on CXCL10 gene expression in IECs has been given no attention. Previously, it has been shown that IECs may express elevated levels of CXCL10 [12], and its expression is increased in the mucosa of patients with IBD [16]. A critical role for CXCL10 in IBD has also been shown in animal models [30, 31]. Apart from this, CXCL10 has been shown to be expressed constitutively in colorectal carcinomas and primary metastatic sites exerting anti-tumour activity by recruiting mononuclear cells via the CXCR3 receptor. However, at the same time, it has also been shown that CXCL10 promoted metastatic adhesion and tumour cell migration [32] so that care should be taken considering CXCL10 as an anti-tumoural therapy option. In light of these findings, we assessed whether stimulation of IECs in culture by inflammatory mediators could induce CXCL10 gene expression. All three IEC lines we used expressed CXCL10 mRNA in a concentration-dependent manner when stimulated with IFN-γ and TNF-α alone. These data support a vast body of evidence that demonstrates that IFN-γ can induce CXCL10 expression in a variety of cells types, including granulocytes, T lymphocytes, bronchial epithelial cells and keratinocytes [33, 34], while TNF-α is a potent inducer for CXCL10 in hepatocytes [17]. Importantly, in addition to these data, we could also show that Caco-2 cells stimulated with IL-1β up-regulate CXCL10 mRNA to a significant level, whereas in HT29 and DLD1 cells, induction of CXCL10 mRNA was less pronounced. A number of authors have reported functional synergy between TNF-α and IFN-γ in promoting gene expression for proteins such as RANTES, IFN regulatory factor 1 (IRF-1), MIG and intercellular adhesion molecule-1[35–38] in a variety of cell types and have demonstrated that this synergy depends on the co-existence of TNF-α-responsive NF-κB-binding sites and IFN-γ-responsive signal transducers and activators of transcription (STAT) protein-binding elements within the promoters of the genes of interest [35, 36]. In our study, when IFN-γ and TNF-α were combined, the accumulation of CXCL10 mRNA and protein was elevated compared to unstimulated cells or cells stimulated with either cytokine alone, implying a synergistic effect on transcription of the CXCL10 gene. A synergistic effect by these two cytokines on CXCL10 expression was also demonstrated in astrocytes, and this synergy has been shown to be at least in part dependent on NF-κB activation and binding to a NF-κB-binding site within the CXCL10 promoter [26]. The synergistic action of IL-1β with IFN-γ in IECs has not been characterised so far. Recently, Takami et al. [39] showed the synergistic induction of hepatocyte growth factor in human skin fibroblasts by IL-1β + IFN-γ+ IFN-γ [39]. We here report for the first time that in intestinal epithelial cell lines, IL-1β alone or in combination with IFN-γ induced CXCL10 gene expression. In Caco-2 cells, the combination led to an early synergistic induction of both CXCL10 mRNA and protein compared to the combination of IFN-γ+TNF-α. This early induction of CXCL10 mRNA and protein by IL-1β+IFN-γ in Caco-2 might be due to the early convergence of signalling pathways that activates transcription factors like NF-κB, IRF-1, STAT1 or other factors that mediate transcription factor interactions. Transient transfections with human CXCL10 promoter [26], in Caco-2, HT29 and DLD1 cells, showed that the synergistic effect of IL-1β + IFN-γ and IFN-γ + TNF-α on CXCL10 gene expression in these cell lines was indeed because of the transcriptional synergy employed at the level of the CXCL10 promoter. A number of studies had revealed that cytokines like IL-1β, TNF-α or enteroinvasive bacteria and the bacterial cell wall product lipopolysaccharide induce the expression of several inflammatory genes through the transcription factor NF-κB in IECs [40–43]. To find out the signalling mechanism involved in CXCL10 gene expression in IECs, we explored the role of transcription factor NF-κB. Inhibition of NF-κB with BAY11-7082 clearly inhibited both CXCL10 mRNA and protein induced by single cytokine alone or in combinations in Caco-2, HT29 and DLD1 cells. Gel shift in Caco-2 could confirm that NF-κB is activated and binds to consensus κB oligonucleotides in cells treated with either IL-1β or TNF-α alone and in combination with IFN-γ. It was shown previously that the binding complex of NF-κB to the κB2-binding site of the CXCL10 promoter contains the homodimer of p65 [26]. In our study, super shift assays in Caco-2 cells treated with IL-1β alone or in combination with IFN-γ confirmed the presence of p65 in the complexes bound to the consensus κB oligo nucleotides. In addition to these experiments, transient transfection experiments with a κB2 mutant of the CXCL10 promoter [26] clearly showed that in Caco-2 cells, none of the cytokines and their combinations were able to induce the CXCL10 promoter activity. In HT29 cells, the combination of IFN-γ + TNF-α led to a weaker induction of TGL-IP10κB-2 mutant activity, whereas in DLD1cells, the combination of IL-1β + IFN-γ and IFN-γ + TNF-α led to a weak induction of TGL-IP10 κB-2 mutant activity. Based on these data, we can conclude that NF-κB plays a critical role in CXCL10 gene expression induced by cytokines in IECs, and this is mainly dependent on the κB-2 site present in its promoter. Moreover, IFN-γ-induced CXCL10 mRNA and protein secretion was inhibited upon NF-kB inhibition. Although the present results may reflect only a small aspect in the overall concert of chemokines, it seems apparent that the basic regulatory principles elucidated here might be shared also by a number of other chemokine genes like that of CCL19, which has recently been shown to be regulated by multiple NF-κB and IRF family transcription factors in human monocyte-derived dendritic cells [44]. Recently, Hiroi M et.al. [45] showed that co-operation between IFN-γ induced STAT1, and the constitutively active or inducible NF-κB is necessary for the transcriptional activity of IFN-γ-inducible genes like MIG and CXCL10. Therefore, we suggest a similar mechanism for IFN-γ-induced CXCL10 in IECs. The responsiveness to cytokines might depend on the state of differentiation. Schlottmann et. al. [46] recently showed that in contrast to HT29 where IFNγ up-regulated CXCL8 secretion, the same was unaffected in Caco-2 cells or even down-regulated in primary colonic epithelial cells. Future studies on chemokine regulation should therefore include primary cells, keeping well in mind that the presence of contaminating immune cells, culture conditions and cell stress during isolation may influence results. Taken together, our data suggest that while IFN-γ, TNF-α and IL-1β may work individually through disparate signalling pathways in CXCL10 gene regulation, a combination of these cytokines enhances transcription of the CXCL10 gene in an NF-κB-dependent manner. The fact that IL-1β via activation of NF-κB is able to mimic the effect of TNF-α in augmenting CXCL10 expression might be of importance in patients with IBD, who do not respond to anti TNF-α antibody therapy. In those cases, an additional inhibition of IL-1β might be desirable as a future therapy option in treating IBD.	[ "intestinal epithelial cells", "inflammatory bowel disease", "cytokines and chemokines", "nuclear factor kappa b", "colon cancer" ]	[ "P", "P", "P", "M", "R" ]
Graefes_Arch_Clin_Exp_Ophthalmol-4-1-2292476	The prescribing of prisms in clinical practice	The use of prisms in cases of decompensated heterophoria is an established treatment modality. The clinical literature lacks consensus upon the appropriate use of prisms, and fails to provide the necessary evidence base. While the experimental literature can guide the practitioner, the lack of double-blind, placebo-controlled clinical studies needs to be addressed. The use of prisms in cases of binocular dysfunction is an important treatment modality for dealing with such patients [1]. The decision to prescribe a prism, and what value of prism to give, is subject to varying clinical opinions and practices [2]. The conservative view in clinical practice is that prisms should not be prescribed in the absence of symptoms of binocular dysfunction [1]. Furthermore, the prescribing of prisms is contraindicated in certain cases, due to the risk of exacerbating an existing condition through the process of vergence adaptation [3–6]. The control of the vergence response has been described previously using linear systems modelling [7], and can aid our understanding of the aetiology of binocular visual dysfunction, and the effect of clinical intervention. The normal vergence response operates in a closed-loop feedback system, where the total response is the sum of the outputs of two control elements with differing temporal properties; a fast (phasic) control element and a slow (tonic) adaptive control element [7]. Previous reports have suggested that the strength of the adaptive vergence control element is often related to the presence of symptoms in patients with binocular dysfunction, and in severe cases the output of the adaptive vergence controller may be reduced to almost zero [8, 9]. In less severe cases, vergence adaptation may be present at a sub-normal level, but sufficient to produce an adaptive response to the introduction of a prism [8]. In such patients, strengthening the adaptive vergence control element is the treatment of first choice [8, 9]. Other studies have shown that in elderly patients vergence adaptation is very limited, and that while this may often be the cause of binocular dysfunction in these patients, remedial action with prisms is clinically viable because adaptation is very unlikely [10]. Indeed, we observe in our own binocular vision clinic that treatments designed to improve the strength of the adaptive vergence controller are generally unsuccessful in elderly patients. Typical clinical measures of binocular function, three of which are investigated by Otto et al. [11] in the current issue, are: Heterophoria (or dissociated phoria).This represents the fusion-free position of the eyes, and therefore the magnitude of the deviation which has to be overcome by the vergence system.Fixation disparity.First investigated by Ogle [12], this represents a small misalignment of the visual axes during binocular viewing, normally measured in seconds of arc. Many clinicians take the view that fixation disparity is indicative of stress within the binocular system [1]. Other authors regard fixation disparity to be a purposeful error, necessary for the vergence control system [5].Associated heterophoriaMallett [13, 14] is the author associated with describing the clinical characteristics of this measurement, which measures the amount of prism required to reduce any fixation disparity (in 2 above) present to zero. In Mallett’s opinion, the presence of an associated phoria in a patient with symptoms of binocular vision dysfunction is indicative of stress within the binocular system, and requires treatment [15].Fusional vergence reserves.This is a clinical measure of the overall ability of the vergence system to control heterophoria. It is generally used in the calculation of Sheard’s Criterion [16], which requires the vergence reserves to be at least twice the size of any heterophoria present. There is evidence that the vergence reserves are related to the strength of the adaptive component of the vergence system [8, 9]. Each of these measures assess one specific aspect of binocular function, and while some practitioners prefer to rely on a particular measure to assess binocular function, others may assess a combination of the measures above to decide the most appropriate clinical intervention for the patient [2]. There is a lack of consensus in the literature about which of these measures correlates most closely with symptoms of binocular dysfunction, and which provides the most accurate basis for prescribing a prism when this is appropriate. Previous work has shown that no single measure correlates well with patient symptoms in all types of binocular anomaly, suggesting that neither fixation disparity, nor associated phoria, provide a universal indicator of binocular dysfunction [17]. This work also found that measures of vergence reserves often showed a better correlation with symptoms [17]. The imperative question for the clinician remains; is there a reliable method to determine the magnitude of prism required to compensate the heterophoria in patients suitable for this method of treatment? Mallett [15] suggested that the associated heterophoria identified the uncompensated portion of the binocular anomaly, and therefore this value represented the prism required to give the patient binocular comfort and stability. Other studies have questioned this [5, 17], and indeed the presence of an associated phoria in patients without binocular visual problems casts doubt on the general applicability of this measure [18]. Sheard’s criterion provides a rationale for prescribing prisms [1, 18]; however, there is still a lack of randomised controlled clinical trials to prove the efficacy of any treatment of binocular vision [2], although a recent study has begun to address this [19]. The use of a measure such as the “comfortable prism”, while clinically appealing, is highly subjective in nature, and would also require a clinical evidence base to prove efficacy. There is considerable work still to be done in the field of clinical management of binocular vision dysfunction in order to provide the required evidence base. In all cases of binocular dysfunction, the choice to prescribe prisms lies with the clinician, who must judge what is most appropriate for each individual patient. A relatively straightforward procedure has been described to assess the probability of vergence adaptation occurring and rendering any prescribed prism ineffective [20]. Where prism treatment is being considered, obtain a prior measurement of the associated phoria, insert the prism to be prescribed into the trial frame and measure the associated phoria with the prism in situ. Allow the patient to wear the prism for 10 minutes; if the associated phoria returns to the value measured prior to insertion of the prism, then the patient’s vergence adaptation mechanism has been strong enough to overcome the prism, and it will have little clinical benefit. Conversely, if no vergence adaptation is observed, the patient is likely to derive significant clinical benefit from this treatment.	[ "prisms", "binocular", "vergence adaption" ]	[ "P", "P", "P" ]
Histochem_Cell_Biol-4-1-2386527	Actin: its cumbersome pilgrimage through cellular compartments	In this article, we follow the history of one of the most abundant, most intensely studied proteins of the eukaryotic cells: actin. We report on hallmarks of its discovery, its structural and functional characterization and localization over time, and point to present days’ knowledge on its position as a member of a large family. We focus on the rather puzzling number of diverse functions as proposed for actin as a dual compartment protein. Finally, we venture on some speculations as to its origin. Introduction A common incident, ubiquitous and frequently observed by the experienced teacher: a student wants to purify protein X from an eukaryotic organism Y. Sooner or later the undergraduate stares at SDS polyacrylamide gels and is surprised, disappointed, depressed, furious about those horrible actin contaminations that seem to be everywhere. The student does not care about the advisor’s soothing explanations: “actin is the most abundant protein”, “actin exists as globular monomer (G-actin) in the soluble supernatant”, “but it is also sedimentable as filamentous actin (F-actin)”. It will be difficult to calm the young scientist down. Looking at these copious amounts of the protein, the student cannot believethat a long time ago it was a scientific breakthrough to show that actin exists at all,that it took years and years to convince truculent groups of actin pioneers that a non-muscle cell contains actin as well,that actin is not a singular and once for all invention in evolution, but is the founder of a large family of isoforms and related proteins, andthat quite a number of these proteins can be found in different subcellular compartments. For example, it has been established as a regular component of both, the cytoplasm and the nucleus; but this perception is the result of a fight that went on for decades. The dawn: actin as a muscle component The first traceable report that described contractile substances in muscle goes back to 1859 and Wilhelm Friedrich Kühne (Kuehne 1859). He belonged to the best known physiologists of his time and coined modern designations like “enzyme” or “myosin”. The latter one he described as a substance that could be isolated from frog muscle and forms a contractile clot under certain conditions. Today we would call this clot a contractile pellet, consisting mainly of actomyosin. Halliburton followed shortly afterwards and was intrigued by this clot formation (Halliburton 1887). He compared it to blood clotting and the transformation of fibrinogen to fibrin. In tedious experiments he extracted mammalian muscle tissue at different time scales, temperatures, salt concentrations and found that myosin apparently needed an additional component to perform fast aggregation and contraction. Without really knowing it, Halliburton worked with actin and produced with this so-called “myosin ferment” an actomyosin precipitate. In our days one reads Halliburton’s publication with some amazement. The report covers 70 pages and looks with its descriptions and tables like a lab-book, e.g.: “Extract #1—10 a.m.—Diluted with an equal amount of water—10 a.m. next day—No change”. Surely, at that time editors were not as busy as they are today. Muscle actin comes of age Then, it took almost 60 years until Brúnó F. Straub in Albert Szent-György’s laboratory at the Hungarian University of Szeged was able to purify actin in good qualities and quantities. At that time ATP was known and available, i.e. Straub could study the influence of ATP on the viscosity of myosin preparations in the presence and absence of other extractable muscle components. The decrease in viscosity upon addition of ATP was considered “activation” of myosin, the protein responsible for this activation was named “actin”, and the complex of these two proteins was “actomyosin”. Straub and Szent-Györgyi discovered these interrelationships 1941/1942 in the middle of World War II and, therefore, published the data only in local periodicals (“Studies from the Institute of Medical Chemistry”, University of Szeged). The research on actin and myosin, the political situation and the excitement in Szent-Györgi’s group in the early 1940s are nicely summarized by W.·F. Mommaerts who was a witness of the original experiments (Mommaerts 1992), and by A.G. Szent-Györgyi himself (Szent-Gyorgyi 2004). It is far beyond the scope of this review to cover the avalanche of actin and myosin research that started after those first molecular characterizations. Of course, muscle was and still is the major source to purify actin from, and, consequently, the research during the following decades used exclusively muscle actin. Actin conquers the cytoplasm But as it is in science, one opens one door and discovers ten closed ones. Keeping that in mind a scientist should always be suspicious regarding dogmas of all sorts. For some time it was almost a law that actin exists only in skeletal muscle and that the well-known conventional myosin is the only motor protein which turns the world go round. Today it is a common place that both assumptions were wrong, although it was quite a task to convince the scientific community. Many minor publications pointed at the presence of actin in non-muscle cells. For example, one found actomyosin-like components and Mg2+ATP-dependent viscosity changes in extracts from sea urchin eggs (Ohnishi 1962), and an actin-like protein from calf thymus nuclei (Ohnishi et al. 1963), and it was assumed that the filaments seen in fibroblasts are analogous to filaments in muscle, responsible for cell motility (Buckley and Porter 1967). Most of these publications never had a really strong impact, especially since they argued against a rock-solid preconception. But to be fair: these tiny reports were the pebbles that paved the road for better equipped groups and more detailed molecular studies. Notably, it was a slime mould, Physarum polycephalum that was used as a model organism to prove the existence of actin and myosin in non-muscle cells. The key experiments have been performed in Japan, in Sadashi Hatano’s group (Hatano and Oosawa 1966a, b; Hatano and Tazawa 1968) and the data were confirmed shortly afterwards by other laboratories. Especially the decoration of actin filaments in non-muscle cells with myosin was a breakthrough technique that unequivocally proved the existence of the actomyosin system in many non-muscle cells (for reviews see: Huxley 1973; Pollard and Weihing 1974; Pollard 1981; Tilney 1975). Actin as the prototype of a large family Today conventional actin is thoroughly studied, with respect to its polymerization and depolymerization equilibrium, its function in the cytoskeleton as a morphological stabilizer, its role during motile activities of a cell, its three-dimensional structure, its binding partners, etc. But the traitorous word in this sentence is “conventional”. What hides behind “conventional” actin? And what is an actin-related protein? A “conventional” actin consists of 375 highly conserved amino acids, give or take a few. The polypeptide sequence is folded into a roughly U-shaped molecule, exposing well-characterized binding sites for many different ligands. There are several isoforms as products of different genes, coined alpha, beta and gamma actin, all of which are polymorphic proteins capable to form polymers. “Conventional actins” display a particular fold which enables them to interact with a variety of partners, to form specific suprastructures involved in intracellular motility, adhesion and locomotion (see below). Figure 1 shows the dynamics of the actin cortex in a Dictyostelium discoideum amoeba during random migration, pinocytosis, and phagocytosis of a yeast cell. An actin-related protein harbours an actin domain, but it differs from “conventional” actin in size and overall amino acid sequence, usually cannot form polymers and may differ from the conventional form in subcellular distribution and its physiological activity. The increasing number of completely sequenced genomes is an enormous help towards our understanding of actin and actin-related proteins. One can take any sequenced genome to analyze actin as a founder of a large protein family. We chose the recently unravelled D. discoideum genome since it represents the genome of a model organism that can locomote as a single cell and in cell assemblies, and can differentiate into simple tissues in a developmental cycle. Hence, it is a social amoeba at the evolutionary border between unicellular and multicellular organisms (Eichinger et al. 2005). Notably, the number of actin genes does not tell us very much about the complexity of an organism. Whereas the genome of budding yeast contains only one single and, not surprisingly, essential gene that codes for a conventional actin, mouse harbours 35, and the plant Arabidopsis thaliana 10 actin genes. The D. discoideum genome contains 33 genes that code for a bona fide conventional actin. Most puzzling, 17 of these actins share identical amino acid sequences, but are encoded by 17 distinct genes, and many of these are developmentally regulated. Why would evolution allow this seemingly luxurious feature? Elimination of redundant genes can only be avoided if they represent a selective advantage. If the identical actins are translated in a developmental pattern, then posttranslational modifications might play a much larger role than we are aware of today. Indeed, actin’s posttranslational modifications are numerous. Actins are acetylated, acylated, arginylated, Ser/Thr/Tyr-phosphorylated, ubiquitinylated etc. This suggests that a developmentally regulated expression of actin genes requires a similarly regulated expression of enzymes that catalyze posttranslational modifications. Such a hypothesis opens a completely new search for our understanding of actin and its isoforms.Fig. 1Dynamics of the actin cortex during amoeboid movement, pino- and phagocytosis. D. discoideum was transformed with a truncated LimE construct, tagged to green fluorescent protein (GFP). LimE binds specifically to filamentous actin (Bretschneider et al. 2004). The transformed cells are depicted in phase contrast (left panels) and fluorescence (right), which reveals the concentration of actin filaments by GFP-LimE. Upper panels: optical section through the actin cortex during random amoeboid movement. Bottom panels: during uptake of a rhodamine-labelled yeast cell the actin cortex forms a phagocytic cup around the particle. The engulfment of liquids by macropinocytosis requires the actin system as well (arrow). Size bar 5 μm The picture became even more complicated after the discovery of the actin-related proteins (Arps). Calculation of the putative structures shows very nicely that actin and Arps share a common three-dimensional fold (Fig. 2, taken from Muller et al. 2005). Most characteristic is the nucleotide-binding cleft which divides the protein in roughly two halves. The depicted structures all show the same orientation. For actin itself, this means that, when incorporated into a “conventional” actin filament, the bottom part of the folded monomer points towards the fast growing (“barbed”) end, the upper part towards the slow growing (“pointed”) end (see also Fig. 3). Figure 2 shows strikingly (1) that all Arps can be moulded into a structure quite similar to the overall structure of conventional actin and (2) that they display a clearly distinct pattern of patches with conserved amino acids. As seen in the comparative scale (Fig. 2, right) Arps 1, 2 and 3 are most closely related to actin, whereas all the other Arps show a much lower percentage of identity. Arps 2 and 3 have been shown to form a tight complex that, after activation through several actin binding proteins and ATP, binds to actin filaments and induces their branching (Kiselar et al. 2007). However, it should be noted that the role of the Arp2/3 complex for branching of actin filaments in vivo is still heavily challenged (Koestler et al. 2008). It is quite amusing to read Vic Small’s statement in the Research Roundup of J Cell Biol: (Feb 25, 2008): “We haven’t disproven branching yet. We’ll need 3D imaging to put the nail in that coffin. But what we’ve seen makes the branching model unlikely”. Undoubtedly, we are still on thin ice with our knowledge on Arp2/3 function in vivo.Fig. 2Structural conservation in the ARP subfamily (adapted from Muller et al. 2005, with permission). The conservation scale weights identity from 0 (blue) to 100% (red). All Arps can adopt a global actin conformation (upper left) but only Arps 1, 2 and 3 show a pronounced sequence homologyFig. 3Generation of different actin polymers as a result of different subunit interactions (adapted from Jockusch et al. 2006, with permission). The subdomains of the actin monomer, as deduced from the crystal structure, are numbered 1–4. In the process that leads to the “conventional” actin filament (F-actin), three such monomers (a–c) form intermolecular contact sites involving the subdomains as indicated left of the trimeric seed. Under physiological conditions, these trimers convert subsequently into the polar F-actin filaments that are depicted in the electron micrograph (top right). In a more rapid process, the actin monomers may form an antiparallel dimer which, at least in vitro, can form bipolar filaments (LD filaments) that show a strikingly different appearance in the electron microscope (bottom right) Let us risk some sort of a heretical view. Trivial as it may be, one has to repeat it again and again: not the three-dimensional fold, but the surface of a protein determines its function. There are curious examples for this. Hisactophilin, a histidine-rich actin-binding protein from D. discoideum, has a structure nearly identical to that of interleukin-1β and fibroblast growth factor, despite its unrelated amino acid sequence. Luckily for the authors, these data were published in a high ranking journal (Habazettl et al. 1992), although there was no functional similarity between these three proteins whatsoever. Today, their structural similarity is not considered important at all and, like many others, they are just members of the large trefoil protein superfamily (Liu et al. 2002). Sippl and coworkers systematically analyzed novel proteins based on known structures (Koppensteiner et al. 2000). They came to the conclusion that almost 30% of the studied proteins have similar structures but different functions. Coming back to actins and Arps. Is it possible that the name “actin” for all of these proteins is utterly misleading? Are we wasting time in our search for cytoskeleton-related activities just because the fold triggered a wrong name? Is the so-called actin fold another example for a successful superfamily which contains only a few proteins that play a role in the cytoskeleton, while many other members are e.g. subunits of larger protein complexes in chromatin or hexokinases or proteins with yet unknown functions? Actin-like proteins in the nucleus: the early period The first reports on actin as a nuclear protein appeared already 5 years after its acceptance as a ubiquitous cytoplasmic component. Nuclear filament bundles resembling microfilaments were observed in oocytes when transcription was inhibited by actinomycin D (Lane 1969), and biochemical analyses identified a prominent protein present in isolated nuclei of the multinucleated plasmodia of Physarum polycephalum (Jockusch et al. 1971; Jockusch et al. 1974; Lestourgeon et al. 1975). The concentration of actin in the nucleus fluctuated with the cell cycle phase of the plasmodia, leading to suggestions that nuclear actin might either have a role in the constriction of the nuclear membrane in Physarum plasmodia that exhibit an intranuclear mitosis (Jockusch et al. 1971), or with changes in transcriptional activity during the cell cycle (Lestourgeon et al. 1975), or during transition from active growth to the formation of sklerotia (spherules) from plasmodia after starvation (McAlister et al. 1977). All this remained speculation at that time, and the majority of the cell biological community dismissed these findings as an artifact, caused by a contamination of nuclei with cytoplasmic actin. However, between 1979 and 1984 three reports appeared which supported the concept of actin being involved in transcription: (1) actin was identified in a complex of RNA polymerase II from Physarum (Smith et al. 1979), (2) it stimulated markedly the transcriptional activity of RNA polymerase II purified from HeLa cell extracts at the preinitiation phase (Egly et al. 1984), and (3) antibodies specific for actin, when injected into the large amphibian oocyte nuclei, caused a dramatic collapse of lampbrush chromosome loops, concommitant with a complete stop of mRNA transcription (Scheer et al. 1984). However, these observations still did not overcome the scepticism at that time—it took another two decades to recognize actin as a dual compartment protein that can execute different functions by selectively associating with ligands specific either for the cytoplasm or the nucleus. Actin and Arps in the nucleus In the last 20 years, studies on nuclear actin became fashionable, and now a new area began which firmly established actin, its relatives and binding partners as nuclear components. Several important steps served as signposts in this process:Cell biologists began to understand that proteins are not necessarily confined to a single intracellular compartment. Numerous examples proved that proteins may shuttle between intracellular regions and even adopt different, compartment-specific activities. Shuttling of actin and actin-related proteins between the nucleus and the cytoplasm cannot be considered any longer as an exceptional, exotic behaviour.Genetic studies, for example performed with yeast or Drosophila, showed that activities specific for the nucleus, like chromatin remodelling or transcription required the presence of the beta actin isoform, in a configuration that could bind to nuclear protein complexes, but were inhibited with actin mutants that failed in this respect.The development of cell biological and immunological methods over the past 20 years, like cellular expression of flurorochrome-tagged proteins in conjunction with video microscopy, and the generation of epitope-characterized monoclonal antibodies allowed to follow actin trafficking throughout the cell and its location in the nucleus.And finally, some old observations made in structural biology came back into people’s mind: actin is a highly polymorphic protein, which can give rise to several different and distinct polymers (Aebi et al. 1981; Millonig et al. 1988). Thus, while there are functionally quite unrelated proteins that may share the same structure (see above), it is also conceivable that a particular protein may adopt slightly different “unconventional forms”, possibly induced by specific binding partners. Such “distortions” may then trigger the formation of specific oligomers or polymers and complex formation with still other partners. Thus, intranuclear actins need not necessarily adopt the same structures found so abundantly in the cytoplasm, like G- or F-actin. This topic is discussed in detail in (Pederson and Aebi 2002), and this concept is supported by monoclonal anti-actins that specifically decorate nuclear actin (Gonsior et al. 1999; Schoenenberger et al. 2005). In Fig. 3 (taken from Jockusch et al. 2006), we show actin in two different forms that both can form polymers. The conventional actin filament requires the interaction of three actin monomers in a specific orientation. Such a trimeric seed will then grow into the polarized F-actin filaments (Fig. 3, top). However, there is also evidence for another pathway: two actin monomers can bind to each other in another orientation, and this “lower dimer” may then lead to nonpolar, quite different filaments (Fig. 3, bottom). It is of course tempting to speculate that such “unconventional” actin polymers may play a role in vivo, for example in the nucleus, but so far, there is no solid evidence for this assumption. Not surprisingly, the wealth of data on nuclear actin and its relatives stimulated numerous hypotheses on their function. The relevant findings and the conclusions on putative functions are the topic of a number of interesting reviews, and in this article, we can only refer the reader to them and to the list of original references covered there (Bettinger et al. 2004; Blessing et al. 2004; Franke 2004; Jockusch et al. 2006; Pederson and Aebi 2002; Rando et al. 2000). The brave reader who works himself through all this information will arrive at two conclusions: the actin-like proteins have definitely conquered a firm position among the group of respectable, important nuclear proteins, and they seem to be involved in a plethora of different activities. While the tasks for actin in the cytoplasm all are connected with the conventional G- and F-actin structures and the dynamic equilibrium between these two states, nuclear functions of these proteins seem associated with several different forms or configurations which are ill or even not at all defined. In the following paragraphs, we will briefly mention the most important tasks proposed for nuclear actins. Actin filaments and a caryoskeleton It is generally accepted that nuclear activities require a precise topographical arrangement of chromatin, to spatially separate hetero- from euchromatin and allow for the complex mechanism of chromatin remodelling during gene activation and transcription. Hence, when electron microscopy provided elegant images of networks of intranuclear filaments, preferably in large objects like the amphibian oocyte nuclei (germinal vesicles), it seemed quite plausible to conclude that a “nuclear matrix” or scaffold is composed of F-actin filaments (Clark and Merriam 1977; Clark and Rosenbaum 1979; Gard 1999) that could provide mechanical stability to the nucleus and serve as a platform to anchor, at least temporarily, chromatin constituents. Indeed, there is evidence from both earlier and more recent investigations that actin associates with filamentous structures: actin antibodies and myosin subfragment 1 decorate such filaments in amphibian oocyte nuclei (Clark and Rosenbaum 1979; Gard 1999; Scheer et al. 1984), and fluorescently labelled actin is found in a polymeric form in Xenopus and HeLa cell nuclei (Kiseleva et al. 2004; McDonald et al. 2006). Yet, in general, these filaments do not stain with fluorescent phalloidin, a small drug binding with high affinity to actin filaments of the conventional “F-actin” type. It is only after subjecting cells to stress or fixation protocols that nuclear filaments bind phalloidin, thus, the suspicion arises that such treatments rearrange actin into conventional filaments (Gall 2006; Jockusch et al. 2006). This, of course, does not preclude the concept that actin participates in a nuclear matrix, especially in the very large, actin-rich nuclei of amphibian oocytes. There is good evidence that it engages in providing mechanical support to this cellular compartment (Bohnsack et al. 2006; Stuven et al. 2003). The nuclear Arps from yeast, Arp 7 and Arp 9, have been shown to build heterodimers and have the potential to form filaments from tetramers (Szerlong et al. 2003). However, it is not clear whether they do in vivo, and whether they might coassemble with nuclear actin in polymers (Blessing et al. 2004). Actin/Arps and their connection with the nuclear envelope Several structural proteins that are essential for the assembly and maintenance of an intact, functional nuclear envelope and its underlying lamina, display well defined binding motifs for actin. These include spectrin, protein 4.1, the nesprins, lamin A and emerin, (Krauss et al. 2003; Shumaker et al. 2003, more references in Blessing et al. 2004; Rando et al. 2000). Hence, it has been speculated, but not proven, that at least part of the nuclear actin is involved in the structural organization of the nuclear envelope. Actin/Arps and the nucleo-cytoplasmic traffic Actin-containing, intranuclear filaments were seen to connect to nuclear pore complexes in the amphibian oocyte nucleus (Hofmann et al. 2001; Kiseleva et al. 2004). In the Xenopus oocyte, but also in dipteres and mammalian cells, actin is engaged in the transport and nucleocytoplasmic export of mRNA (Hofmann et al. 2001) and RNA-protein complexes, in particular in those containing a subset of hnRNPs (Percipalle et al. 2001, 2002, reviewed in Bettinger et al. 2004; Pederson and Aebi 2005). Actin/Arps and chromatin remodelling Actin and nuclear Arps were both reported as being critically involved in chromatin remodelling (reviewed in (Bettinger et al. 2004; Blessing et al. 2004; Olave et al. 2002; Percipalle and Visa 2006; Rando et al. 2000). Chromatin remodelling complexes operate as large, multiunit machines in mammals, insects, yeast and plants to reorganize the genetic material by unravelling nucleosomes and converting the genetic material into a form suitable for transcription. Many of these steps require energy, and thus many of these protein complexes contain ATPases, comprised of several subunits, that can be grouped into subfamilies. Actin was identified in complex with specific subunits of most ATPases, together with four nuclear Arps (4, 5, 6 and 8) in all organisms, with the exception of yeast. Here, the two yeast-specific nuclear Arps (7 and 9) are found in the corresponding ATPases. One of the histone acetyl transferases that also contribute to chromatin remodelling and subsequent transcription (NuA4, Doyon et al. 2004) is highly conserved from yeast to man and is also found in a complex with actin and Arp 4 (Harata et al. 2002). Structural organization and function of the actin and Arp molecules in these ATPase and acetyl transferase complexes are largely unknown, and speculations for their respective roles range from stabilizing the enzymatic and a chaperoning activity to connecting these gigantic protein complexes to nuclear filamentous scaffolds. Actin/Arps and transcription Actin has been identified as a regular component of all the three nuclear RNA polymerases (Pol I, II and III), apparently interacting with two subunits that are shared among all three enzyme complexes. In vitro transcription by all three enzymes is actin-dependent (reviewed in Grummt 2006; Percipalle and Visa 2006). In the nucleus, pre-mRNA is complexed with ribonucleoproteins to form hnRNPs, and, presumably, there is co-translational recruitment of beta actin to these particles during their formation, as deduced from findings that several hnRNP components bind directly actin. Actin’s association with the RNA polymerases may precede, be simultaneous with or follow the initiation of actin-dependent chromatin remodelling by ATP-dependent complexes and/or histone acetylation, as described above (Grummt 2006; Percipalle and Visa 2006). Most of the data available today are derived from studies with Pol II, in insects and vertebrates, supporting and confirming the conclusions drawn more than two decades earlier from work on Xenopus oocyte and HeLa cell nuclei (Egly et al. 1984; Scheer et al. 1984). There is solid evidence that actin’s role in Pol II-dependent transcription is executed by the beta actin isoform (Hofmann et al. 2004; McDonald et al. 2006), and no other isoform seems involved. While actin’s contribution to the formation of pre-initiation complexes and subsequent transport of pre-mRNPs through the nucleus towards their export through nuclear pores is well accepted today (Bettinger et al. 2004; Franke 2004), it is completely unknown whether actin might mediate such intranuclear movements in filamentous or even contractile structures (Pederson and Aebi 2005). Actin ligands in the nucleus The nuclear constituents harbouring actin, like those of the nuclear matrix, the nuclear envelope, the nuclear pores, chromatin remodelling complexes, RNA polymerases and pre-mRNPs are all highly insoluble, gigantic structures. However, there are also actin ligands that are found in the cytoplasmic as well in the nuclear compartment, possibly in rather dynamic and temporary complexes with actin. Since the discovery of actin as a prominent component in the cytoplasm of practically all eukaryotic cells, a plethora of such actin-binding proteins has been described and characterized. They arrange actin into the various suprastructures required for cytoplasmic functions, such as intracellular motility and locomotion, by either regulating the balance between monomeric and filamentous actin, or generating networks and bundles of filaments (Winder and Ayscough 2005). Remarkably, many of these actin ligands are evolutionary well conserved, like actin itself (Korn 1982; Schleicher et al. 1988). Members of the various subfamilies, with affinity to either monomeric or filamentous actin, are dual compartment proteins that can shuttle between the nucleus and the cytoplasm, their main location frequently being dependent on the differentiation state of cells. Comprehensive lists of these proteins that have previously been described as cytoplasmic actin ligands but were also identified in the nucleus are given in (Rando et al. 2000; Bettinger et al. 2004; Pederson and Aebi 2005). Among them are several which form complexes with actin and interfere with the formation of conventional actin filaments, like profilin, binding to monomeric actin, and cofilin, a ligand for both, monomeric and polymeric actin. Nuclear profilin is apparently involved in the regulation of the level of nuclear actin, as profilin–actin complexes are recognized and exported from mammalian nuclei by a specific exportin (Stuven et al. 2003), while actin free of profilin can apparently be exported by a different exportin, due to its nuclear export sequences (Wada et al. 1998). Profilin and cofilin both can sligthly change the fold of actin. Profilin induces a form which faciliates the exchange of bound ATP in G-actin (reviewed in Jockusch et al. 2007). Cofilin, a phosphoprotein, contains a nuclear location sequence, and in its dephosphorylated state can transport actin “piggy-back” into the nucleus (Pendleton et al. 2003). When bound to actin polymers, it distorts their conformation such that these filaments do not bind phalloidin anymore. Thus, these proteins might be critical in creating forms specific for nuclear actin, as detected by specific antibodies (Gonsior et al. 1999; Jockusch et al. 2006; Schoenenberger et al. 2005). Among the actin ligands solely found in the nuclear compartment are a nuclear actin binding protein identified in Acanthamoeba (Rimm and Pollard 1989), an actin binding protein composed of two different subunits in mammalian cells (Ankenbauer et al. 1989), and a nucleus-specific form of myosin I (Pestic-Dragovich et al. 2000). The latter is a small, monomeric myosin that does not form filaments as would be needed for contraction of putative actin filaments in the nucleus. It is apparently involved in transcription of ribosomal genes by Pol I, where it binds to a transcription initiation factor but also to components of a chromatin remodelling complex. It has been speculated that in this location it may recruit actin associated with Pol I to the site of rRNA transcription, but so far, there is no evidence that these two proteins interact directly in vivo (reviewed in (Grummt 2006; Percipalle and Farrants 2006). Conclusions So, have we now lost the student being confronted with actin for the first time in his scientific education, or have we been able to attract his curiosity for this abundant, highly conserved, mysterious protein? Actin has come a long way from the times of its discovery in muscle and we have learned to know it as a member of a large family of structurally related proteins, some of which are dual compartment proteins. Concomitant with this, there is an overwhelming expansion of their putative functions, and both, the fine structural conformation as well as the different functions are probably determined by the numerous actin binding proteins. As a cartoon, Fig. 4 summarizes the most important players in the cytoplasmic and the nuclear compartment. It is obvious that the putative activities in the nucleus outnumber those in the cytoplasm. Taken together, these findings allow for the view that actin and its relatives were originally engaged in nuclear activities, and that their appearance in the cytoplasm is related to a specialization in functions connected solely with intracellular motility and locomotion (F-actin filament formation, actin dynamics). Further functional narrowing would then result in the special case of skeletal muscle, where a specific isoform, alpha actin, is overexpressed and knows nothing else than to allow myosin-driven contraction.Fig. 4The manifold functions of nuclear actin. The G-/F-actin equilibrium is only a highly specialized function of actin in the cytoplasm (top); in the nucleus (bottom) the actin system plays a quite different regulatory and structural role. Actin can execute its tasks as monomer or as nucleus-specific polymer. It interacts with heterogenous ribonucleoproteins (hnRNPs) and is thought to function as a track for RNA transport through NPCs. It remains to be shown how nuclear actin cooperates with nuclear myosin-1 to enhance movement of RNA polymerases or to re-organize chromatin (Nunez et al. 2008). Positioning of the nucleus in the cell is achieved by nuclear membrane (NM) proteins that, like the large nesprins, hook the nucleus to the actin cytoskeleton. Anchorage of nesprins is guaranteed by SUN proteins in the inner nuclear membrane (Worman and Gundersen 2006). The stabilization of the nucleoskeleton involves the interaction of inner nuclear membrane proteins like emerin or nesprins with lamins as well as polymeric actin (Libotte et al. 2005). It should, however, be borne in mind that the level of nuclear actin is apparently tightly regulated in mammalian cells by exportin 6 which selectively transports profilin–actin complexes from the nucleus into the cytoplasm (Bohnsack et al. 2006). The nuclear functions of many dual compartment actin binding proteins are not well understood, but one has to assume that their interaction can discriminate between nuclear and cytoplasmic actin (Pederson and Aebi 2005). For example, the recently described cyclase associated protein CAP2 is nuclear in myoblasts but sarcomeric in myotubes (Peche et al. 2007). Actin-related proteins are only summarized in one symbol and are either nucleus specific subunits in chromatin remodelling complexes (Arp4, Arp8 etc.), cytoplasmic or belong also to the group of dual compartment proteins If one was to speculate further, one might imagine that prokaryotic precursors of present days’ actin, related to the proteins that spatially and temporally control macromolecular trafficking, chromosome segregation and cell polarity in recent bacteria (Carballido-Lopez 2006), appeared first as “nuclear actin” in an ancestor eukaryotic cell as part of the endosymbiontic engulfment. But, it will require several decades and lots of industrious students to decide whether this is science fiction or facts about actin.	[ "transcription", "chromatin remodelling", "actin history", "cytoplasmic motility", "nucleocytoplasmic traffick" ]	[ "P", "P", "R", "R", "R" ]
J_Urban_Health-2-2-1705505	Respondent-Driven Sampling in a Study of Drug Users in New York City: Notes from the Field	Beth Israel Medical Center (BIMC), in collaboration with the Centers for Disease Control (CDC) and the New York State Department of Health (NYSDOH), used respondent-driven sampling (RDS) in a study of HIV seroprevalence among drug users in New York City in 2004. We report here on operational issues with RDS including recruitment, coupon distribution, storefront operations, police and community relations, and the overall lessons we learned. Project staff recruited eight seeds from a syringe exchange in Lower Manhattan to serve as the initial study participants. Upon completion of the interview that lasted approximately 1 h and a blood draw, each seed was given three coupons to recruit three drug users into the study. Each of the subsequent eligible participants was also given three coupons to recruit three of their drug-using acquaintances. Eligible participants had to have: injected, smoked or snorted an illicit drug in the last 6 months (other than marijuana), aged 18 or older, adequate English language knowledge to permit informed consent and complete questionnaire. From April to July 2004, 618 drug users were interviewed, including 263 (43%) current injectors, 119 (19%) former injectors, and 236 (38%) never injectors. Four hundred sixty nine (76%) participants were men, 147 (24%) were women, and two (<1%) were transgender. By race/ethnicity, 285 (46%) were black, 218 (35%) Hispanic, 88 (14%) white, 23 (4%) mixed/not specified, and four (<1%) native American. Interviews were initially done on a drop-in basis but this system changed to appointments 1 month into the study due to the large volume of subjects coming in for interviews. Data collection was originally proposed to last for 1 year with a target recruitment of 500 drug users. Utilizing RDS, we were able to recruit and interview 118 more drug users than originally proposed in one quarter of the time. RDS was efficient with respect to time and economics (we did not have to hire an outreach worker) and effective in recruiting a diverse sample of drug users. Background A variety of sampling methods have been used to recruit hard-to-reach populations, such as drug users, into research studies. These methods include snowball sampling, targeted sampling, and time-space sampling.1–3 Despite widespread use of these methods, they are not without important limitations. Snowball sampling begins with a set of initial study participants, called seeds, who refer other eligible respondents. Once these persons are interviewed, they are also asked to bring in or provide references for other potential respondents. This process continues until the sample size is achieved. Snowball sampling can provide easy access to hidden populations, but it is most often biased because it is usually done out of convenience rather than randomly.1–3 Targeted sampling involves thoroughly collecting preliminary data to determine various characteristics such as locations where a sample congregates and demographic characteristics such as race, gender, and age.1,2,4 Similar to snowball sampling, using targeted sampling to study drug users may result in a selection bias because less-visible, isolated drug users may not be found at the sites where other drug users are sampled and, thus, may be overlooked. Time-space sampling involves preliminary research by ethnographers to determine when and where hidden populations congregate.3,5 Once a list of times and sites is determined, researchers then randomly choose times to visit those sites to collect data. Time-space sampling has limited use with drug users because making an exhaustive list of where drug users congregate could be a very protracted and expensive process, as drug-use sites change frequently in response to environmental factors such as police presence. Respondent-driven sampling (RDS),1 a relatively new method being used to draw probability samples of hidden populations, incorporates some methods of snowball sampling, such as chain referral sampling, but includes additional provisions to minimize bias. It produces population estimates that are asymptotically unbiased, which means that bias is only on the order of 1/[sample size], so bias is minimal in samples of substantial size.6 This allows for probability-based inferences based on the social network of the sample. This paper reports on the use of RDS to recruit participants for a study of risk behavior and HIV seroprevalence among drug users in New York City. The study was conducted by Beth Israel Medical Center in collaboration with the Centers for Disease Control and Prevention and the New York State Department of Health. We discuss the operational issues faced with RDS such as recruitment, coupon distribution, storefront operations, police and community relations, and the overall lessons learned. Methods In April 2004, Beth Israel Medical Center staff recruited eight seeds from a syringe exchange in Lower Manhattan. Each seed was screened for eligibility before being given a coupon to come back for an interview. The seeds were asked about their drug use, including mode of use and drug preparation techniques. Those who claimed to be an injection drug user (IDU) were also asked to show track marks. Although not a requirement of RDS, the seeds were recruited to resemble the race and gender profile of drug users in Lower Manhattan.7 They were asked to come to a research storefront in Lower Manhattan the following day to complete a computer-assisted interviewer-administered personal interview (CAPI) and to have their blood drawn for an HIV test. Each seed, and subsequent study subject, received $20 compensation for their time. When subjects arrived at the storefront, they were questioned by the study screener to ensure eligibility. Once they were deemed eligible by the screener, each subject was assigned a unique code to serve as their study identification. This code included the following information: the first two letters of the last name, first letter of mother’s first name, the last two digits of the birth year, one letter for the person’s race, and one letter for gender. This code was used to identify blood work and questionnaires and could be regenerated if the subject forgot it. No names or other identifying information were asked. After being assigned a study code, subjects met with an interviewer to be consented and interviewed. The interview had three parts, an hour-long questionnaire, administered by the interviewer, HIV pre-test counseling and a blood draw for an HIV test. The interview consisted of a structured questionnaire which took approximately 1 h and asked about drug-use frequency, drug and sexual risk behavior, syringe acquisition, and knowledge of HIV and hepatitis B and C. After the interview, HIV counseling, and blood collection, each subject was given three coupons to recruit three other drug users into the study. The subjects were briefly trained on how to recruit others, with specific emphasis on the recruitment of friends and acquaintances who use drugs. Eligible respondents had to have done the following: injected, smoked, or snorted an illicit drug in the past 6 months (those who smoked only marijuana were not eligible for the study); turned age 18 years or older by the time of the interview; been able to speak English adequately to consent to the study and complete the questionnaire; and lived, bought, and/or used drugs on the Lower East Side of Manhattan. Subjects could have, however, lived in the larger New York metropolitan area, as long as they bought or used their drugs on the Lower East Side of Manhattan. Coupons contained the time that the storefront opened (9:00 a.m.) and a unique number (to make each one distinct). To make them difficult to duplicate, coupons were printed on thick cardstock paper with color images. To track the coupons and payment for each respondent, we used custom-developed software for RDS called IRIS Plus. Information such as respondent’s unique code, physical traits, coupon number, and the numbers of the coupons each respondent distributed were all recorded in IRIS Plus. This information enabled us to link coupons together, determine when respondents should be paid and who gave coupons to whom. This software also helped to prevent the redemption of duplicated coupons because the database would not accept duplicate coupon numbers. When respondents came in to make an appointment, the coupon was checked in the IRIS Plus database to verify that it had not been used previously. Additionally, each person’s code was checked in IRIS Plus after being screened to determine whether a person with that code had previously been enrolled in the study. If another study subject had that code, the screener looked at the physical traits listed for that person for verification. Additionally, if the screener felt that a person looked familiar, she would search for the person's physical traits in IRIS Plus to see if another person with similar characteristics was previously enrolled.1 On July 2, 2004, we ceased coupon distribution to give potential subjects 2 weeks to redeem remaining coupons. On July 16, 2004, we concluded data collection, and the study officially concluded on July 30, 2004. The last 2 weeks of July were set aside to allow the remainder of subjects to come in for their HIV test results. Results From April to July 2004, 618 respondents were interviewed, including 263 (43%) current injectors (had injected drugs within the past 6 months), 119 (19%) former injectors (had injected drugs more than 6 months prior to the interview—who used drugs in the past 6 months by other modes of administration), and 236 (38%) never injectors (those that had never injected any drugs but used non-injection drugs in the past 6 months). The mean age of respondents was 44 years, with a mean age of first drug use at 19 years and mean age of first injection drug use at 22 years (among current and former drug injectors). By gender, 469 (76%) were men, 147 (24%) were women, and two (<1%) were transgender. By race/ethnicity, 285 (46%) were black; 218 (35%), Hispanic; 88 (14%), white; 23 (4%), mixed race or unspecified race; and four (<1%), native American. Seed Recruitment To control the flow of recruitment, we recruited only eight seeds: two white men, one white woman, two Hispanic men, one Hispanic woman, one black man, and one black woman, all of whom identified themselves as active IDUs. A relatively small number of seeds was recruited so that we could assess how productive they were and then recruit more if necessary, without being too overwhelmed with respondents. Respondent Recruitment Our initial goal was to interview 500 drug users within 1 year. After 4 months, 618 drug users had been interviewed. Respondent recruitment (those people recruited after the seeds) happened much faster than we had expected. In the first couple of weeks, recruitment was slow but quickly picked up as more coupons got out into the community. Interviews were initially done on a drop-in basis, but this system changed to appointments after 1 month into the study due to the large volume of respondents coming in for interviews. As the flow of potential subjects continued to be rapid and at times overwhelming, we were also forced to post-date all coupons by 2 days. Therefore, when a respondent was given a coupon, that person could not be interviewed or make an appointment until the date written on the coupon. This system was helpful in pacing the number of people coming into the storefront for interview appointments but perhaps choosing a smaller number of seeds may have alleviated the need to implement these changes altogether. Police and Community Relations As previously noted, as the study progressed and more coupons were distributed in the community, the number of people coming into the storefront for interviews increased dramatically. For the first 2 weeks, the interviewers would arrive in the morning to find 10 to 20 people waiting for them. People reported arriving as early as 7:00 a.m. so that they could be interviewed that day. Once crowds began to form outside the storefront, neighboring businesses complained to our staff or called the police to file complaints. The police came to speak with staff on three separate occasions regarding complaints of noise and loitering. At that point we decided to move to an all-appointment system. Instead of telling people that their recruits could show up and be interviewed, recruits were told to first come into the storefront to be screened, and that if they passed the screener, they could make an appointment. Lessons Learned Were we to do another RDS study in New York City, we would do the following: We would have a phone number so that screening and appointments for potential participants could be done by telephone. This would lessen the amount of traffic in the storefront and reduce some confusion. Once participants arrived for their interviews, we would then screen them about their drug use. If we were looking to recruit a large number of drug users in a short amount of time, we would again recruit eight seeds. The eight seeds recruited for this study seemed to be an appropriate number because we were trying to maximize our time in the field for data collection. Although the first 2 weeks of the study were relatively slow-paced, it gave the field staff time to adjust to their new environment and new roles and make any necessary changes before they became too overwhelmed. Drug users in New York City are generally a very overstudied population, and thus we expected a large portion of the seeds to show up for their interviews. In cities where the population is not as familiar with research, researchers may want to recruit more seeds, considering that many may not return for the initial interview. Another alternative, however, if time permits, is to recruit a smaller number of seeds, give them a few weeks to return for their interview and then recruit more seeds, if necessary. It was our intention to recruit more seeds if our initial eight were not productive, but this was not necessary. We would start with an appointment-only system. We began with a drop-in system for interviews because our time to collect data was very limited and this was the most efficient use of the interviewers’ time. For instance, if a person dropped in and wanted to be interviewed, we could do it on the spot if the person with a scheduled interview had not shown up. Even though this system worked for us initially, it soon became unwieldy, and we were forced to change a month into the study. People adjusted very quickly, and many people reported being relieved because they didn’t have to wait in front of the storefront for 2 h in the morning before staff arrived. We would be very rigid about our appointments. In this study we were very strict about appointment times and conveyed this to each participant. If persons with appointments were more than 5 min late and someone else arrived to schedule an appointment but was free to complete it then, we took the person who dropped in rather than waiting for the person with the scheduled appointment to show up. We rarely had interviewers sitting around with nothing to do. There was almost always someone who dropped in when someone else was late for an appointment. We would meet with local business owners to inform them of who we are and what we were doing prior to the start of the study. We would invite them to tell us if they had any concerns. We would not disclose the type of research being done or with whom, but it would be helpful to let the surrounding business owners know that they may see an increase in the number of people in or around the storefront. We would remain at the storefront 6 weeks after ending coupon distribution. Coupon distribution ended a full 4 weeks before we vacated the storefront, and data collection ended 2 weeks before. The timing of both of these allowed for a fair amount of those with coupons to redeem them, for those with outstanding incentives to collect their money, and for those with outstanding test results to come in for them. Two additional weeks would have been helpful to allow for remaining coupons to be redeemed. Because the storefront we used stayed open for other research studies, we were able to leave contact information for our interviewers in case any respondents showed up for their HIV test results. Within the month after we left the storefront, we received two additional calls for HIV test results. Our interviewers met the respondents at a park and a coffee shop to give them their results. The other storefront staff reported a handful of people coming in with coupons after the study closed, but they did not report any incidents with those people that were unable to redeem them. Finally, we did not pay out any uncollected incentives after July 30, 2004. Some of the lessons that we learned about conducting a study using RDS may be specific to New York City. Because there was an abundance of willing and able participants, we had to take several measures to make data collection more manageable. These measures may not be necessary or practical for researchers in cities where the drug-using population is smaller, less willing to participate in research, or both. Conclusions Using RDS, we were able to recruit and interview 118 more drug users than originally proposed in one third of the time. In our experience, RDS was efficient with respect to time and economics (we did not have to hire an outreach worker) and effective in garnering a diverse sample of drug users. We were able to interview many more drug users than expected in a short period of time with very minimal recruitment effort. In summary, RDS can be an efficient and effective form of recruitment, particularly for research involving drug users or other hidden populations.	[ "drug users", "respondent-driven sampling (rds)", "time-space sampling", "capi", "iris plus" ]	[ "P", "P", "P", "P", "P" ]
Pflugers_Arch-3-1-1839769	The contribution of refractoriness to arrhythmic substrate in hypokalemic Langendorff-perfused murine hearts	The clinical effects of hypokalemia including action potential prolongation and arrhythmogenicity suppressible by lidocaine were reproduced in hypokalemic (3.0 mM K+) Langendorff-perfused murine hearts before and after exposure to lidocaine (10 μM). Novel limiting criteria for local and transmural, epicardial, and endocardial re-excitation involving action potential duration (at 90% repolarization, APD90), ventricular effective refractory period (VERP), and transmural conduction time (Δlatency), where appropriate, were applied to normokalemic (5.2 mM K+) and hypokalemic hearts. Hypokalemia increased epicardial APD90 from 46.6 ± 1.2 to 53.1 ± 0.7 ms yet decreased epicardial VERP from 41 ± 4 to 29 ± 1 ms, left endocardial APD90 unchanged (58.2 ± 3.7 to 56.9 ± 4.0 ms) yet decreased endocardial VERP from 48 ± 4 to 29 ± 2 ms, and left Δlatency unchanged (1.6 ± 1.4 to 1.1 ± 1.1 ms; eight normokalemic and five hypokalemic hearts). These findings precisely matched computational predictions based on previous reports of altered ion channel gating and membrane hyperpolarization. Hypokalemia thus shifted all re-excitation criteria in the positive direction. In contrast, hypokalemia spared epicardial APD90 (54.8 ± 2.7 to 60.6 ± 2.7 ms), epicardial VERP (84 ± 5 to 81 ± 7 ms), endocardial APD90 (56.6 ± 4.2 to 63.7 ± 6.4 ms), endocardial VERP (80 ± 2 to 84 ± 4 ms), and Δlatency (12.5 ± 6.2 to 7.6 ± 3.4 ms; five hearts in each case) in lidocaine-treated hearts. Exposure to lidocaine thus consistently shifted all re-excitation criteria in the negative direction, again precisely agreeing with the arrhythmogenic findings. In contrast, established analyses invoking transmural dispersion of repolarization failed to account for any of these findings. We thus establish novel, more general, criteria predictive of arrhythmogenicity that may be particularly useful where APD90 might diverge sharply from VERP. Introduction Hypokalemia exerts important clinical effects on cardiac function that in some respects resemble those seen in the congenital long-QT syndromes (LQTS). Thus, both conditions result in electrocardiographic QT prolongation [12, 23] and premature ventricular depolarizations (PVDs), which may result in the initiation of an arrhythmic activity [41, 52]. In contrast to the cardiac effects of hypokalemia, arrhythmic activity in LQTS has been extensively studied and has often been attributed to after-depolarizations occurring against a background of re-entrant substrate [2, 36, 44]. Re-entry may take place as a result of inhomogeneities producing regions of conduction block, which lead to wave-break and circus movement [21, 37] or altered repolarization gradients, which lead to wave reflection [1]. In this situation, depolarization propagates from active cells into previously active adjacent regions, establishing re-entrant circuits. These may become established either locally or over larger regions of the myocardium, such as across the thickness of the myocardial wall. Tendencies to transmural re-entrant excitation in models of LQTS have been previously analyzed in terms of transmural dispersions of repolarization (TDR) obtained from the positive part of the difference between respective endocardial and epicardial stimulation to repolarization times [36, 44, 45]. In human LQTS, increases in the interval between the peak and full recovery of electrocardiographic precordial T waves (Tpeak to Tend), previously shown to reflect TDR [54], are indeed associated with arrhythmic activity [33]. Certainly, recent reports correlate Tpeak to Tend to arrhythmic risk more closely than more widely accepted indicators such as corrected QT interval and QT dispersion [53]. However, such re-excitation may also be limited by recovery from refractoriness; re-entrant excitation would require this to precede the return of the membrane potential to threshold [40]. Certainly, class 1 antiarrhythmic drugs such as lidocaine are known to increase ventricular effective refractory period (VERP) [28]. Yet, such use of spatial differences in action potential repolarization times to quantify arrhythmic substrate neither explicitly considers changes in VERP nor applies such criteria to potential local as opposed to transmural re-excitation. This paper associates for the first time the proarrhythmic effect of hypokalemia with a significant decrease in VERP, despite contrasting prolongation of action potentials, in agreement with computer-modeling studies of action potential waveforms using established data on the various effects of hypokalemia on ionic conductivity properties of ventricular myocytes. Furthermore, it associates the antiarrhythmic effects of lidocaine with a significant increase in VERP, despite having little effect on action potential duration, in agreement with clinical observations. Analyses using TDR were insufficiently sensitive to account for any of these arrhythmogenic findings. This study accordingly established more general novel criteria that would provide necessary conditions for local and transmural and epicardial and endocardial re-excitation incorporating not only action potential duration but also VERP and conduction times that may be particularly useful when action potential duration differs sharply from VERP. These criteria successfully accounted for all the arrhythmogenic findings. Materials and methods Experimental animals Mice were housed in an animal facility at 21 ± 1°C with 12 h light/dark cycles. Animals were fed sterile chow (RM3 Maintenance Diet, SDS, Witham, Essex, UK) and had free access to water. Wild-type 129 Sv mice aged 3–6 months were used in the experiments. All procedures complied with UK Home Office regulations (Animals [Scientific Procedures] Act 1986). Solutions All solutions were based on bicarbonate-buffered Krebs-Henseleit solution (mM: NaCl 119, NaHCO3 25, KCl 4, KH2PO4 1.2, MgCl2 1, CaCl2 1.8, glucose 10 and Na-pyruvate 2; pH adjusted to 7.4) bubbled with 95% O2/5% CO2 (British Oxygen Company, Manchester, UK). Hypokalemic (3.0 mM K+) solutions were prepared by reducing the quantity of KCl added. Lidocaine-containing normokalemic and hypokalemic solutions were prepared by adding lidocaine (Sigma–Aldrich, Poole, UK) to a final concentration of 10 μM. Preparation A Langendorff-perfusion protocol previously adapted for murine hearts [4] was used. In brief, mice were killed by cervical dislocation (Schedule 1: UK Animals [Scientific Procedures] Act 1986), and hearts were then quickly excised and placed in ice-cold bicarbonate-buffered Krebs-Henseleit solution. A short section of aorta was cannulated under the surface of the solution and attached to a custom-made 21-gauge cannula filled with the same solution using an aneurysm clip (Harvard Apparatus, Edenbridge, Kent, UK). Fresh Krebs-Henseleit solution was then passed through 200 and 5 μm filters (Millipore, Watford, UK) and warmed to 37°C using a water jacket and circulator (Techne model C-85A, Cambridge, UK) before being used for constant-flow retrograde perfusion at 2–2.5 ml/min using a peristaltic pump (Watson-Marlow Bredel model 505S, Falmouth, Cornwall, UK). Hearts were regarded as suitable for experimentation if, on rewarming, they regained a healthy pink colour and began to contract spontaneously. Electrophysiological measurements An epicardial monophasic action potential (MAP) electrode (Hugo Sachs, Harvard Apparatus) was placed against the basal region of the left ventricular epicardium. In addition, a small access window was created in the interventricular septum to allow access to the left ventricular endocardium [9]. A custom-made endocardial MAP electrode composing two twisted strands of high-purity Teflon-coated 0.25 mm diameter silver wire (Advent Research Materials, UK) was constructed. The Teflon coat was removed from the distal 1 mm of the electrode, which was then galvanically chlorided to eliminate DC offset, inserted and placed against the septal endocardial surface. MAPs were amplified, band-pass filtered (0.5 Hz to 1 kHz: Gould 2400S, Gould-Nicolet Technologies, Ilford, Essex, UK), and digitized at a sampling frequency of 5 kHz (micro1401, Cambridge Electronic Design, Cambridge, UK). Analysis of MAPs was performed using Spike II software (Cambridge Electronic Design). Experimental protocol A bipolar platinum stimulating electrode (1 mm interpole spacing) was placed on the basal surface of the right ventricular epicardium. Square-wave stimuli (Grass S48 stimulator, Grass-Telefactor, Slough, UK) of 2 ms duration and with amplitudes of twice the excitation threshold were initially applied to hearts at a constant cycle length of 125 ms for at least 10 min and until MAPs showed stable baselines, rapid upstroke phases that reached consistent amplitudes and smooth repolarization phases [30]. Hearts were then exposed to test solutions for 20 min, during which time stimulation was continued, before subsequent recordings were made. Intrinsically evoked MAPs were recorded in the absence of stimulation while action potential duration (at 90% repolarization, APD90) and stimulation to depolarization latency were determined during regular stimulation at a constant interstimulus interval of 125 ms. Hearts were then subjected to an adapted form of an extrasystolic electrical stimulation procedure previously used to assess arrhythmogenicity and refractoriness in both human [43] and murine [22] studies of congenital LQTS, described in detail later. The possibility that events evoked by extrasystolic stimuli rather represented motion artifacts was excluded by their being reproducible between hearts and appearing identical in both electrodes. All data are presented as means±standard errors of the means and include both the number of repetitions and the number of hearts. Comparisons were made using analysis of variance (significance threshold set at P ≤ 0.05). Modeling The charge-difference model of Fraser and Huang [16, 18] was adapted to permit computational modeling of the murine ventricular cardiac myocyte using ion channel equations and parameters from the model of Bondarenko et al. [6] with the Na+/K+-ATPase model of Hernandez et al. [24]. The use of charge-difference modeling allowed the model to reach a true beat-to-beat steady state that was independent of initial intracellular ion concentrations [17], thus permitting simulation of the influence of changes in extracellular ion concentrations that are well recognized to influence Na+/K+-ATPase activity, and hence steady-state intracellular ion concentrations. Model cells were studied under normokalemic (5.2 mM K+) conditions with normal ion permeabilities, hypokalemic (3 mM K+) conditions with these same ion permeabilities and hypokalemic (3 mM K+) conditions with the K+ permeabilities of channels carrying the repolarizing currents IK1 and Ito reduced by 20%, replicating the effect of such hypokalemia on transmembrane K+ permeabilities observed experimentally by Killeen et al. [29]. Stimulation was applied at a regular 125 ms interstimulus interval at an amplitude of twice the diastolic threshold, as for the experimental preparations. After beat-to-beat stability was achieved, APD90 was measured under each condition. Refractory periods were then determined using a similar protocol to that used in the experiments, every eighth (S1) stimulus being followed by an extrasystolic (S2) stimulus. S1S2 interval was initially 70 ms and was subsequently decremented by 1 ms with each successive cycle until an S2 stimulus failed to initiate an action potential. Results After-depolarizations initiate arrhythmic activity in bradycardic hypokalemic hearts In initial experiments, isolated perfused hearts were stimulated at a constant interstimulus interval of 125 ms for 20 min after 20 min exposure to test solutions. This demonstrated stable trains of MAPs, under all normokalemic (5.2 mM K+, n = 7; five hearts), hypokalemic (3.0 mM K+, n = 8; five hearts), or lidocaine-treated (10 μM) normokalemic (n = 6; five hearts) or hypokalemic (n = 7; five hearts) conditions: After-depolarizations and arrhythmic activity were consistently absent throughout. The subsequent experiments then examined arrhythmic properties at the longer cycle lengths (between 224 and 271 ms) that occurred in the absence of extrinsic stimulation (Fig. 1) and that have previously been reported to be proarrhythmic both under hypokalemic conditions and in the congenital LQTS [11]. Intrinsic cycle length did not differ significantly (P > 0.05) between normokalemic (250 ± 21 ms, six hearts), hypokalemic (253 ± 16 ms, seven hearts), lidocaine-treated normokalemic (248 ± 22 ms, six hearts), and lidocaine-treated hypokalemic (260 ± 10 ms, eight hearts) hearts. Fig. 1After-depolarizations and arrhythmic activity in spontaneously contracting hypokalemic hearts. Epicardial monophasic action potential recordings in the absence of extrinsic stimulation in hearts exposed to normokalemic (5.2 mM K+, a) and hypokalemic (3.0 mM K+, b) test solutions and normokalemic (c) and hypokalemic (d) test solutions containing lidocaine (10 μM) for 20 min Epicardial MAPs then retained morphologically consistent waveforms and were entirely free of after-depolarization and arrhythmic phenomena through 116 min of recordings over six normokalemic hearts (Fig. 1a). In contrast, 46 ± 7% of MAPs showed after-depolarizations early in their repolarization phases during 140 min of recordings that led to episodes of arrhythmic activity in 52 ± 3% of cases in five out of seven hypokalemic hearts (P < 0.01 as compared to normokalemic hearts, Fig. 1b). However, MAPs showed consistent waveforms without such after-depolarizations or arrhythmic activity during 118 min of recordings over six lidocaine-treated normokalemic hearts (Fig. 1c). Finally, 40 ± 9% of MAPs showed after-depolarizations occurring late in the repolarization phase in 40 ± 9% of cases more than 98 min of recordings in six out of eight lidocaine-treated hypokalemic hearts (P > 0.05 as compared to hypokalemic hearts). After-depolarizations occurred more frequently in those instances where intrinsic cycle length was long. However, these events were never followed by arrhythmic activity (P < 0.01 as compared to hypokalemic hearts, Fig. 1d). Extrasystolic stimulation immediately after recovery from refractoriness initiates arrhythmic activity in hypokalemic hearts A programmed electrical stimulation protocol recently shown to predict arrhythmogenicity in clinical LQTS [43] and previously adapted for use in murine models of LQTS [22] confirmed the above arrhythmogenic tendencies in hypokalemic hearts (Fig. 2). This comprised regular (S1) stimulation at a constant interstimulus interval of 125 ms interrupted by an extrasystolic (S2) stimulus after every eighth S1 stimulus. The S1S2 interval was decremented in 1 ms steps with each successive stimulus cycle from an initial value of 120 ms until the S2 stimulus either appeared to initiate arrhythmic activity, confirmed during an imposed 250 ms pause, or failed to initiate a MAP suggesting that the VERP has been reached. Accordingly, VERP values are reported to the nearest millisecond. Fig. 2Arrhythmic activity in hypokalemic hearts after extrasystolic stimulation applied close to the refractory period. Epicardial monophasic action potential recordings resulting from application of extrasystolic (S2) stimuli at S1S2 intervals greater than the action potential duration at 90% repolarization (A), just greater than the ventricular effective refractory period (VERP; B), and just less than the VERP (C) in hearts exposed to normokalemic (5.2 mM K+, a) and hypokalemic (3.0 mM K+, b) test solutions and normokalemic (c) and hypokalemic (d) test solutions containing lidocaine (10 μM) for 20 min. Single vertical lines indicate the timing of S1 stimuli, and double lines indicate the timing of S2 stimuli Normokalemic hearts (Fig. 2a, A–C) again were consistently free from arrhythmic activity after S2 stimulation after any S1S2 interval (n = 11; eight hearts). This also applied to hypokalemic hearts when S2 stimuli were delivered when MAPs had reached 90% repolarization (Fig. 2b, A, n = 7; five hearts). However, S2 stimuli delivered within the period just after recovery from refractoriness consistently initiated arrhythmic activity under these conditions (Fig. 2b, B). In contrast, S2 stimuli delivered before recovery from refractoriness failed to elicit MAPs, and this was followed by the resumption of stable rhythms (Fig. 2b, C). Finally, S2 stimulation did not result in arrhythmic activity in lidocaine-treated hearts whatever the S1S2 interval, whether under normokalemic (Fig. 2c, A–C, n = 8; six hearts) or hypokalemic (Fig. 2d, A–C, n = 8; five hearts) conditions. When taken together, the presence or absence of arrhythmogenicity in these experiments parallels clinical findings. Arrhythmic tendency in hypokalemia correlates with increased local critical intervals One hypothesis for the tendency towards either local or transmural re-excitation during action potential repolarization might consider the relationship between the time course of the recovery of membrane voltage and the corresponding time course of recovery of excitability from total refractoriness to a finite threshold for excitation in the myocardial regions concerned. These parameters were approximated by action potential duration at 90% repolarization (APD90) and VERP, respectively. Both these were measured during the procedures of the kind illustrated in Fig. 2, allowing for the delay between endocardial and epicardial excitation where appropriate. Firstly, the risk of local reexcitation of either the epicardium or the endocardium would be reflected in a critical interval given by the relevant APD90–VERP. Secondly, the risk of transmural re-excitation of either the epicardium by the endocardium (or the reverse) would require incorporation of the delay between endocardial and epicardial excitation given by the difference between endocardial and epicardial stimulation to depolarization latencies, Δlatency. This would give critical intervals of (endocardial APD90 + Δlatency−epicardial VERP) and (epicardial APD90 + Δlatency−endocardial VERP), respectively. Figures 3a and 4a show typical epicardial and endocardial action potential waveforms during regular stimulation under each of the four above conditions (A–D). Figures 3b and 4b show the corresponding APD90s (vertical solid lines and dense hashing), VERPs (vertical broken lines and sparse hashing), and local critical intervals (shading). Asterisks indicate values that are significantly (P < 0.05) larger and daggers those that are smaller than those recorded in normokalemic hearts. Neither epicardial (46.6 ± 1.2 ms) nor endocardial (58.2 ± 3.7 ms) APD90s were significantly different (P > 0.05) from the corresponding VERPs (41 ± 4 and 48 ± 4 ms) under normokalemic conditions (Figs. 3A and 4A, n = 10; eight hearts). This resulted in local critical intervals taking small positive values of 5.4 ± 4.3 ms in the epicardium and 9.8 ± 5.3 ms in the endocardium. In contrast, epicardial (53.1 ± 0.7 ms) but not endocardial (56.9 ± 4.0 ms) APD90 increased significantly (P < 0.05), whereas both epicardial (29 ± 1 ms) and endocardial (29 ± 2 ms) VERPs decreased significantly under hypokalemic conditions (Figs. 3B and 4B, n = 6; five hearts). This resulted in significant positive shifts (P < 0.01) in both epicardial (23.7 ± 1.2 ms) and endocardial (28.5 ± 4.6 ms) local critical intervals, in fitting with the occurrence of arrhythmic activity under hypokalemic conditions. Fig. 3Changes in epicardial action potential duration, ventricular effective refractory period, and local critical interval after exposure to hypokalemia and to lidocaine. Epicardial MAP waveforms during regular stimulation in hearts exposed to normokalemic (5.2 mM K+, A) and hypokalemic (3.0 mM K+, B) test solutions and normokalemic (C) and hypokalemic (D) test solutions containing lidocaine (10 μM) for 20 min comparing action potential duration at 90% repolarization; APD90 (vertical solid lines), VERP (vertical broken lines), and local critical interval (shading; a). Action potential duration at 90% repolarization, APD90 (dense hashing), VERP (sparse hashing) and critical interval (shading) under these conditions (b). Asterisks indicate values that are significantly (P < 0.05) larger and daggers those that are smaller than those recorded in normokalemic heartsFig. 4Changes in endocardial action potential duration, ventricular effective refractory period, and local critical interval after exposure to hypokalemia and to lidocaine. Endocardial MAP morphologies during regular stimulation in hearts exposed to normokalemic (5.2 mM K+, A) and hypokalemic (3.0 mM K+, B) test solutions and normokalemic (C) and hypokalemic (D) test solutions containing lidocaine (10 μM) for 20 min comparing action potential duration at 90% repolarizationl; APD90 (vertical solid lines), VERP (vertical broken lines), and local critical interval (shading; a). Action potential duration at 90% repolarization, APD90 (dense hashing), VERP (sparse hashing), and critical interval (shading) under these conditions (b). Asterisks indicate values that are significantly (P < 0.05) larger and daggers those that are smaller than those recorded in normokalemic hearts The opposing effects of hypokalemia on APD90 and VERP can be explained in terms of alterations in conductances of repolarizing K+-channels The above MAP findings concerning APD90 and VERP were in close agreement with the predictions of established ion channel equations and parameters from the model of Bondarenko et al. [6] with the Na+/K+-ATPase model of Hernandez et al. [24] using charge-difference modeling in simulated single cells. The model simulated the effects of regular stimulation at a 125 ms interstimulus interval until a steady state was reached, as reflected in beat-to-beat stability. Action potential characteristics were then simulated under normokalemic conditions, hypokalemic conditions with normal K+ permeabilities, and hypokalemic condition with the 20% reduction in the permeabilities of channels carrying the repolarizing K+ currents IK1 and Ito (C) as reported in recent experimental results [29] (Fig. 5). Figure 5a demonstrates the predicted steady-state action potential waveforms under each condition. Figure 5a and b also show APD90 (vertical solid lines and dense hashing), VERP (vertical broken lines and sparse hashing), and critical intervals (shading). Under normokalemic conditions (A), the resting membrane potential was −83 mV. APD90 (22 ms) was shorter than recorded in the whole hearts in keeping with previous results from microelectrode studies [7, 20]. Nevertheless, APD90 was shorter than VERP (27 ms) resulting in a critical interval of −5 ms. Figures 5B (a and b) demonstrate the consequences of an altered Nernst potential for K+ alone both upon resting membrane potential and the time course of a subsequent action potential. Hypokalemia hyperpolarized the membrane potential (−93 mV) and shortened both APD90 (19 ms) and VERP (25 ms), thus having little effect on the limiting criterion for re-excitation (−6 ms). In contrast, Fig. 5C (a and b) additionally demonstrate the combined effects of hypokalemia on both the Nernst potential and K+ permeabilities of the respective channels carrying IK1 and Ito. Although there was no additional effect on the resting membrane potential (−93 mV), hypokalemia increased both APD90 (29 ms) and VERP (27 ms) causing a positive shift in critical interval from −5 ms to +2 ms (Fig. 5C). Fig. 5Computational modeling of murine ventricular action potentials showing changes in action potential duration, ventricular effective refractory period, and critical interval after exposure to hypokalemia. Action potential morphologies during regular stimulation in cells under normokalemic (5.2 mM K+, A) and hypokalemic (3.0 mM K+) conditions comparing action potential duration at 90% repolarization; APD90 (vertical solid lines), VERP (vertical broken lines), and critical interval (shading; a). In B, permeabilities of ion channel are under normokalemic conditions. In C, permeabilities of ion channels carrying the repolarizing K+ currents IK1 and Ito are reduced by 20%. APD90 (dense hashing), VERP (sparse hashing), and critical interval (shading) under these conditions (b) The abolition of arrhythmic tendency by lidocaine correlates with negative shifts in local critical intervals Exposure of normokalemic hearts to lidocaine (Figs. 3, C and 4C) significantly increased (P < 0.05) epicardial (54.8 ± 2.7 ms), although not endocardial (56.6 ± 4.2 ms), APD90 and significantly increased (P < 0.01) both epicardial (84 ± 5 ms) and endocardial (80 ± 2 ms) VERPs (n = 5; five hearts). This resulted in significant (P < 0.01) negative shifts in local critical intervals in both the epicardium (−31.7 ± 5.3 ms) and endocardium (−23.4 ± 4.7 ms). Lidocaine exerted concordant effects on hypokalemic hearts (Figs. 3D and 4D): Epicardial APD90 increased to 60.6 ± 2.7 ms, and endocardial APD90 remained unchanged (63.7 ± 6.4 ms), whereas both epicardial (81 ± 7 ms) and endocardial (84 ± 4 ms) VERPs were significantly increased (P < 0.05, n = 6; five hearts). The resulting significant negative shifts (P < 0.01) in local critical intervals in both epicardium (−20.4 ± 7.5 ms) and endocardium (−20.7 ± 7.6 ms) paralleled the antiarrhythmic effect of lidocaine. Arrhythmic tendency in hypokalemia also correlates with increased transmural critical intervals Figures 6a and 7a show epicardial and endocardial action potential waveforms during regular stimulation under each condition. Figures 6b and 7b show APD90 (vertical solid lines and dense hashing) and VERPs (vertical dotted lines and sparse hashing), as detailed above, together with Δlatencies (horizontal arrows and horizontal shading) and transmural critical intervals (shading) under the four conditions studied. Asterisks indicate values that are significantly (P < 0.05) larger and daggers those that are smaller than recorded in normokalemic hearts. Figure 6 thus compares endocardial APD90s with epicardial VERPs, allowing for Δlatency to describe critical intervals for epicardial re-excitation. In contrast, Fig. 7 compares epicardial APD90s with endocardial VERPs, allowing for Δlatency to describe critical intervals for endocardial re-excitation. Fig. 6Changes in endocardial action potential duration, epicardial ventricular effective refractory period, transmural conduction time, and epicardial transmural critical interval after exposure to hypokalemia and to lidocaine. Epicardial and endocardial MAP morphologies during regular stimulation cross-comparing epicardial and endocardial waveforms and the relationship between the action potential duration at 90% repolarization; APD90 (vertical solid lines) and VERP (vertical broken lines) of one waveform and the decay of the other, indicating critical intervals (shading). Hearts were exposed to normokalemic (5.2 mM K+, A) and hypokalemic (3.0 mM K+, B) test solutions and normokalemic (C) and hypokalemic (D) test solutions containing lidocaine (10 μM) for 20 min. Horizontal arrows indicate the time taken for depolarization to spread from epicardium to endocardium (a). APD90 (dense hashing), VERP (sparse hashing), transmural conduction time (horizontal hashing), and critical interval (shading) under these conditions (b). Asterisks indicate values that are significantly (P < 0.05) larger, and daggers those that are smaller than those recorded in normokalemic heartsFig. 7Changes in epicardial action potential duration, endocardial ventricular effective refractory period, transmural conduction time, and endocardial transmural critical interval after exposure to hypokalemia and to lidocaine. Epicardial and endocardial MAP morphologies during regular stimulation cross-comparing epicardial and endocardial waveforms and the relationship between the action potential duration at 90% repolarization; APD90 (vertical solid lines) and VERP (vertical broken lines) of one waveform and the decay of the other, indicating critical intervals (shading). Hearts were exposed to normokalemic (5.2 mM K+, A) and hypokalemic (3.0 mM K+, B) test solutions and normokalemic (C) and hypokalaemic (D) test solutions containing lidocaine (10 μM) for 20 min. Horizontal arrows indicate the time taken for depolarization to spread from epicardium to endocardium (a). APD90 (dense hashing), VERP (sparse hashing), transmural conduction time (horizontal hashing), and critical interval (shading) under these conditions (b). Asterisks indicate values that are significantly (P < 0.05) larger and daggers those that are smaller than those recorded in normokalemic hearts Epicardial (21.0 ± 0.82 ms) and endocardial (58.2 ± 3.7 ms) latencies were statistically indistinguishable in normokalemic hearts giving a Δlatency of 1.6 ± 1.4 ms. This contributed to significant (P < 0.05) negative shifts in transmural critical intervals in both the epicardium and endocardium (−14.3 ± 6.1 and −3.3 ± 4.2 ms, respectively, Figs. 6A and 7A). Hypokalemia had no significant effect (P > 0.05) on latencies or on Δlatency (1.1 ± 1.1 ms) but resulted in significant (P > 0.01) positive shifts in transmural critical intervals in both the epicardium (23.6 ± 2.6 ms) and endocardium (29.2 ± 6.0 ms; Figs. 6B and 7B). The abolition of arrhythmic tendency by lidocaine also correlates with significantly decreased transmural critical intervals In contrast, exposure of normokalemic hearts to lidocaine significantly increased both epicardial and endocardial latencies (33.8 ± 2.6 and 46.3 ± 5.6 ms, respectively) but still did not significantly alter (P > 0.05) Δlatency (12.5 ± 6.2 ms). However, transmural critical intervals became significantly negative in both the epicardium (−15.2 ± 6.6 ms) and endocardium (−16.6 ± 8.8 ms; Figs. 6C and 7C). This was also true when hypokalemic hearts were exposed to lidocaine: Both epicardial and endocardial latencies were significantly increased (26.3 ± 3.2 and 33.8 ± 1.3 ms, respectively) but Δlatency was not significantly altered (P > 0.05, 7.6 ± 3.4 ms). Again, transmural critical intervals were significantly decreased (P < 0.05) in both the epicardium (−12.8 ± 5.9 ms) and endocardium (−10.4 ± 1.7 ms; Figs. 6D and 7D). Arrhythmogenesis occurs in the absence of significant alterations in the TDR The analysis above thus established four critical intervals, which, when taken together, provided a clear prediction of arrhythmogenicity under these circumstances when changes in APD90 did not correspond to changes in VERP. In contrast, an analysis in terms of TDR, previously shown to predict arrhythmogenicity in both congenital and acquired forms of the LQTS [2], gave insufficiently sensitive predictions. TDR was calculated as the time between stimulation and 90% repolarization in the endocardium minus the time between stimulation and 90% repolarization in the epicardium. TDR values did not significantly alter with the presence or otherwise of arrhythmogenicity (Fig. 8). Thus, in normokalemic hearts (Fig. 8, A) epicardial repolarization time (67.6 ± 1.5 ms) was significantly shorter (P < 0.05) than endocardial repolarization time (77.6 ± 3.9 ms), giving a TDR of 10.0 ± 4.2 ms (n = 10; eight hearts). Epicardial and endocardial repolarization times, as well as the resulting TDR, remained unchanged under hypokalemic conditions (Fig. 8, B). Finally, although treatment with lidocaine significantly increased (P < 0.05) epicardial and endocardial repolarization times under both normokalemic (Fig. 8, C, to 86.1 ± 2.8 and 103.0 ± 7.0 ms, respectively, n = 6; five hearts) and hypokalemic (Fig. 8, D, to 86.9 ± 4.1 and 97.6 ± 6.6 ms, respectively, n = 5; five hearts) conditions, in neither case did it affect TDR. Fig. 8Changes in transmural dispersion of repolarization after exposure to hypokalemia and to lidocaine. Epicardial (up-sloping hashing) and endocardial (down-sloping hashing) stimulation to repolarization times, and the difference between these values giving transmural dispersion of repolarization (open bars) in hearts exposed to normokalemic (5.2 mM K+, A) and hypokalemic (3.0 mM K+, B) test solutions and normokalemic (C) and hypokalemic (D) test solutions containing lidocaine (10 μM) for 20 min. Asterisks indicate values that are significantly (P < 0.05) larger than those recorded in normokalemic hearts These findings indicate that despite similarities, criteria that have been established to predict arrhythmogenicity in LQTS do not necessarily apply in hypokalemia. However, explicit inclusion of refractory behavior yields novel criteria, which may constitute more sensitive general predictors of arrhythmogenicity and may prove particularly useful in situations where APD90 diverges sharply from VERP. Further, these novel criteria provide a physiological basis for the participation of refractoriness in arrhythmogenicity. Discussion In clinical situations, hypokalemia is associated with arrhythmogenesis initiated by PVDs [52] and accompanied by prolongation of the electrocadiographic QT interval, reflecting increased action potential duration [23]. Furthermore, class 1 antiarrhythmic agents are effective in suppressing arrhythmic activity in hypokalemic patients [42]. These features thus resemble corresponding characteristics of the congenital LQTS [12, 41], where arrhythmic activity is thought to result from re-entrant excitation [2]. This has been attributed to the propagation of depolarization from active cells to previously active adjacent regions subsequently triggering spread of excitation and thus establishing re-entrant circuits [1, 3, 37]. We sought to study the physiological basis for the arrhythmogenicity observed under hypokalemic conditions, particularly the extent to which this resembles or differs from the corresponding features of LQTS, in intact isolated perfused murine hearts. Epicardial and endocardial MAPs were first recorded from hypokalemic hearts, confirming that stimulation at a regular interstimulus interval of 125 ms (S1 stimulation) resulted in stable rhythms. In contrast, at the long intrinsic cycle lengths occurring in the absence of extrinsic stimulation, frequent after-depolarizations were observed and were often followed by the initiation of an arrhythmic activity. This is consistent with the known proarrhythmic effect of bradycardia [10, 13, 46, 51]. Previous studies correlating results from single-cell and whole-heart preparations have attributed such after-depolarizations to inward currents flowing through reactivated voltage-operated Ca2+ channels [27, 35], Ca2+-coupled inward Na+ currents via the Na+–Ca2+ exchanger [47], or Ca2+-induced Ca2+-release from intracellular stores [26]. In the presence of the class 1b antiarrhythmic agent lidocaine, after-depolarizations persisted. These events were especially common in hearts where intrinsic cycle length was particularly long, again in agreement with the known proarrhythmic effect of bradycardia [10, 13, 46, 51]. However, in the presence of lidocaine, these events were never followed by the initiation of arrhythmic activity. Thus, the experiments demonstrated that hypokalemic murine hearts showed arrhythmogenic properties in agreement with clinical findings and established conditions in which arrhythmogenicity was and was not observed. A quantitative assessment of MAP waveforms and refractory characteristics associated with this arrhythmogenicity was then performed using an extrasystolic stimulation (S2) procedure previously established in the assessment of arrhythmogenicity in both clinical [43] and murine [22] studies of LQTS. S2 stimulation reproducing the effect of after-depolarizations immediately after recovery from refractoriness failed to initiate arrhythmic activity in normokalemic hearts. However, such stimulation consistently resulted in arrhythmic activity in hypokalemic hearts. This is in agreement with previous reports that after-depolarizations and S2 stimulation early during action potential repolarization are particularly arrhythmogenic [14, 38] and also parallels clinical observations that PVDs coincident with T-waves frequently initiate arrhythmic activity [48]. In contrast, S2 stimuli did not elicit arrhythmic activity in lidocaine-treated hearts, whether studied under normokalemic or hypokalemic conditions. These are consistent with after-depolarizations having occurred late in action potential repolarization, and thence, failing to initiate arrhythmic activity. Previous studies in murine [30], canine [15], and human [31] ventricles have consistently reported that maneuvers, which alter action potential duration, also produce corresponding changes in refractory period, with the notable exception of exposure to class 1 antiarrhythmic drugs [39]. However, the effect of isolated reduction in [K+]o on these parameters has not been studied. Exploration of the effect of varying S1S2 interval demonstrated for the first time that although action potential duration (quantified at 90% repolarization, APD90) was increased in hypokalemia, VERP was decreased. Furthermore, exposure to lidocaine had no effect on APD90 in the epicardia or endocardia of normokalemic hearts and significantly increased APD90 only in the endocardia of hypokalemic hearts, despite significantly increasing VERP in all cases. Thus, the proarrhythmic effect of hypokalemia was associated with recovery from refractoriness occurring earlier in action potential repolarization, whereas the antiarrhythmic effect of lidocaine was associated with recovery from refractoriness occurring later in action potential repolarization. Reduction of [K+]o might be expected to increase outward K+ currents and thereby decrease APD90. Indeed, computer modeling of single ventricular myocytes confirmed that reduction of [K+]o per se resulted in decreased APD90. Furthermore, the Nernst equation would predict that reduction of [K+]o should hyperpolarize the resting membrane potential, thereby increasing the proportion of sodium channels available for activation [19] and decreasing the VERP: our model replicated this effect. However, incorporation of recent data from our group demonstrating that reduction of [K+]o decreases the repolarizing K+ currents IK1 and Ito [29] resulted in VERP returning to its normokalemic value and APD90 being increased beyond its normokalemic value. Thus, although reduction in K+ permeability compensates for the change in VERP, it overcompensates for the change in APD90, rendering the APD90 longer than the VERP, in fitting with experimental results and with the proarrhythmic effect of hypokalemia. Previous studies have reported that although exposure to lidocaine increases VERP through an effect on the gating of fast Na+-channels [32, 34], it has a proportionately smaller effect on action potential duration [5]. Exposure to lidocaine is thus established to result in postrepolarization refractoriness [39], in fitting with experimental results and with the antiarrhythmic effect of lidocaine. We then applied an analytical scheme to provide a simple physiological explanation for these findings. Subject to electrotonic coupling between cells [25], the simplest condition for local re-excitation between adjacent cells within either epicardium or endocardium would require membrane potential to exceed threshold at some point during action potential repolarization, the recovery of membrane potential lagging behind the recovery of excitability. Recovery of excitability was approximated by the VERP, measured using a standard stimulus of consistent amplitude and duration. Recovery of membrane potential was approximated by the action potential duration at 90% repolarization (APD90). Accordingly, APD90−VERP gives a critical interval that reflects tendency towards local re-excitation and arrhythmogenicity. Positive shifts in this interval would reflect a relatively proarrhythmic state, whereas negative shifts in this interval would reflect an antiarrhythmic state. The corresponding analytical condition for transmural re-excitation across the thickness of the myocardial wall with a transmural conduction time given by Δlatency [40] yields transmural critical intervals of (endocardial APD90 + Δlatency−epicardial VERP) for the epicardium and (epicardial APD90 + Δlatency−endocardial VERP) for the endocardium, again subject to electrotonic spread of current between cells [25]. Although exposure to lidocaine increased both epicardial and endocardial stimulation to depolarization latencies, attributable to its established effect on conduction velocity [8], it had no significant effect on Δlatency. Hypokalemia resulted in significant positive shifts in the magnitude of all four critical intervals, whereas exposure to lidocaine resulted in significant negative shifts, in precise agreement with the presence or absence of arrhythmogenicity. Modelling of single ventricular myocytes predicted shorter APD90s than were recorded from whole-heart preparations, in agreement with previous experimental observations in such single-cell preparations [7, 20, 30]. Nevertheless, hypokalemia when similarly modeled for such single-cell preparations resulted in a positive shift in critical interval, in common with our experimental observations from whole hearts. Although changes in all four critical intervals correlated with arrhythmogenicity, TDR, previously shown to predict arrhythmogenicity in LQTS [36, 44, 45], proved an insufficiently sensitive predictor. This finding may be attributable to sharp differences between APD90 and VERP: it is possible that APD90 and VERP were in close agreement in previous studies on LQTS [49, 50]. In the latter event, the present analysis would have yielded identical results to one adopting TDR. Thus, we establish for the first time that hypokalemia decreases VERP despite increasing APD90 and attribute this surprising finding to effects of reduced [K+]o on ion channel gating. Secondly, we establish novel indices incorporating VERP as general criteria for re-entrant arrhythmogenicity that additionally provide a physiological basis for the association between changes in epicardial and endocardial VERP and APD90 and susceptibility to arrhythmogenesis. Thirdly, we establish that such analyses provide more sensitive indications of arrhythmogenicity than previous analyses invoking TDR.	[ "action potential duration", "refractory period", "conduction time", "transmural dispersion of repolarization", "critical intervals", "arrhythmia" ]	[ "P", "P", "P", "P", "P", "U" ]
Histochem_Cell_Biol-4-1-2228382	Morphogenesis of post-Golgi transport carriers	The trans-Golgi network (TGN) is one of the main, if not the main, sorting stations in the process of intracellular protein trafficking. It is therefore of central importance to understand how the key players in the TGN-based sorting and delivery process, the post-Golgi carriers (PGCs), form and function. Over the last few years, modern morphological approaches have generated new insights into the questions of PGC biogenesis, structure and dynamics. Here, we present a view by which the “lifecycle” of a PGC consists of several distinct stages: the formation of TGN tubular export domains (where different cargoes are segregated from each other and from the Golgi enzymes); the docking of these tubular domains onto molecular motors and their extrusion towards the cell periphery along microtubules; the fission of the forming PGC from the donor membrane; and the delivery of the newly formed PGC to its specific acceptor organelle. It is now important to add the many molecular machineries that have been described as operating at the TGN to this “morphofunctional map” of the TGN export process. Introduction The Golgi complex serves as the central station in the biosynthetic pathway, where proteins are sorted for their different destinations, such as various domains of the cell surface and the endosomal–lysosomal system. This delivery of cargo proteins from the Golgi complex to their target compartments is carried out by dynamic, membrane-bound organelles that are frequently called either “transport carriers” or “post-Golgi carriers” (PGCs) (Luini et al. 2005). As these PGCs have an important role in the process of intracellular transport, their morphology, living dynamics and molecular compositions have become the focus of significant interest, particularly over the last few years. These PGC transport organelles were originally defined through the development of green fluorescent protein (GFP) technology and living-cell imaging (Lippincott-Schwartz et al. 2000), whereby the first fluorescent cargo protein that was followed in living cells revealed a new world of images of these relatively large and highly dynamic structures that travel from the Golgi complex to the plasma membrane (Wacker et al. 1997; Hirschberg et al. 1998; Nakata et al. 1998). Then, with time, the list of cargo molecules that could be visualized in vivo expanded, further exposing the unexpected complexity of the post-Golgi transport pathways. Finally, further technical advances resulted in the combination of video and electron microscopy (correlative light-electron microscopy, CLEM), which provided the means to define the morphogenesis of these PGCs at the ultrastructural level. As a result of these studies we now know that PGCs arise from specific membrane domains of the Golgi complex that lack resident Golgi enzymes, forming what are known as the “PGC precursors” (Hirschberg et al. 1998; Keller et al. 2001; Polishchuk et al. 2003; Puertollano et al. 2003). The shapes and sizes of PGCs that even carry the same cargo can vary across a wide range, and under the light microscope, most of them are seen to be clearly larger that plasma-membrane associated clathrin vesicles and 100-nm fluorescent beads (Hirschberg et al. 1998). The smaller PGCs usually have a size of 300–400 nm, although some of the larger ones can reach several microns in length. Video microscopy has also revealed that many of these carriers appear globular, although they frequently stretch out into tubular shapes during their translocation through the cytosol; thus PGCs have frequently been termed as “pleiomorphic” structures. In transporting their cargoes to the correct acceptor compartments, the movement of PGCs is mediated via microtubules. However, carriers can also form and support post-Golgi transport without microtubules, although the correct targeting of their cargo proteins is usually compromised under these conditions (Rindler et al. 1987; Kreitzer et al. 2003). Finally, the life cycle of a PGC can be schematically imagined to consist of three stages: (1) its formation (which can in turn be further divided into several substages; see below); (2) its transition through the cytosol; and (3) its docking and fusion with the target membrane (Polishchuk et al. 2000). The first of these steps, the formation of the PGC, appears to be the most complex, and it also probably remains the least understood. Since the question of the morphogenesis of a PGC is closely intertwined with that of the structure of the organelle from which it originates, the trans-Golgi network (TGN), both of these issues will now be discussed in an integrated fashion. The trans-Golgi network and the origin of PGCs The process of PGC morphogenesis has been characterized in detail using video and electron microscopy, as well as with the combination of these techniques (CLEM). As indicated above, this process comprises the formation, the extrusion and the fission of the export domain from the TGN, thus generating the free carrier (Polishchuk et al. 2003; see also Fig. 1). The first step (the formation of the tubular export domains) also includes the segregation of the cargo proteins from the Golgi resident enzymes. These events appear to be common to the proteins directed towards different post-Golgi compartments, such as basolateral plasma membrane (Hirschberg et al. 1998; White et al. 2001; Polishchuk et al. 2003), the apical plasma membrane (Keller et al. 2001) and the endosomal–lysosomal system (Puertollano et al. 2003). These export domains usually contain TGN markers; i.e. they are part of the TGN (Polishchuk et al. 2003; Puertollano et al. 2003). Of note, these three post-Golgi compartments are the ones that have been best characterized for the PGCs leaving the TGN, with the total number and types of pathways for cargo exit from the TGN at present not known, although there are at least six, and probably more (Rodriguez-Boulan et al. 2005). Fig. 1Formation of post-Golgi transport carriers. Subsequent frames extracted from a time-lapse sequence illustrating the different stages of PGC biogenesis (arrows): a formation of the tubular domain containing the cargo VSVG-YFP and devoid of the Golgi-resident protein galatosyltransferase-CFP. b Extrusion of this tubular domain from the Golgi complex. c, d Fission of the domain from the parental Golgi membranes. e Thin section of a cell expressing the TGN38-HRP construct. Arrows indicate tubular PGC precursor that has been pulled out of the TGN area of the Golgi complex Although the existence of the TGN has been known for many years (Griffiths and Simons 1986), its precise structure and identity still remain to be ascertained. In one view, which predominated in the field until recently, the TGN was considered to consist of essentially an anastomosing tubular network that emanates from (or that results from the breakdown of) the trans-most Golgi cisterna, and that projects mainly in the trans direction (Griffiths et al. 1989, 1985; Clermont et al. 1995). Rambourg and colleagues have thoroughly described the TGN in several different cell types. In these descriptions, this trans-most cisterna also tends to “peel off” from the rest of the Golgi stack, and together with the other morphological characteristics of the Golgi stack, this is a feature that is suggestive (Clermont et al. 1995) of the cisternal progression-maturation trafficking model (Bannykh and Balch 1997; Mironov et al. 1997; Bonfanti et al. 1998; Losev et al. 2006; Matsuura-Tokita et al. 2006). Thus, in a simple version of this scheme, the TGN would result from the final stages of maturation of the Golgi cisternae. This would include the partial transformation of this trans-most cisternae into a tubular network, and the various carriers that leave the TGN for their different destinations would all originate form different domains of this tubular network. A more complex and more recent view of the organization at the trans face of the Golgi complex comes from studies using electron microscopy tomography of cryofixed, freeze-substituted cells (Ladinsky et al. 1999, 2002). These studies have indicated that the TGN derives not only from the last trans cisterna, but instead from the three trans-most cisternae (Roth et al. 1985), from where tubules emanate into the trans space of the Golgi stacks. Remarkably, only the trans-most cisterna exhibits clathrin-coated buds, with the other buds appearing to have different, yet-to-be identified, types of coats (Ladinsky et al. 1999). Thus, only the trans-most cisterna would be responsible for the clathrin- and AP-1-dependent trafficking towards the endo-lysosomes, and although all of these three trans-most cisternae of the Golgi stacks would serve as the classical TGN, each of them could be specialized in the packaging and export of specific cargo proteins. This might have important functional and mechanistic implications for the proteins sorting and export process. Another remarkable TGN feature that has been revealed by these tomographic studies is that these three trans-most cisternae can intercalate with the cisternae of the endoplasmic reticulum (ER) (Ladinsky et al. 1999, 2002). These have been proposed to have important roles in the transfer of lipids (through specialized contact sites) between the ER itself and the Golgi complex. It will be important to establish whether the morphological discrepancies described above reflect fundamental differences in the organization of the export of cargo in different cell systems. Indeed, it is possible that the first model (the TGN as a network resulting from the tubular disassembly of the last trans-most cisterna) might actually only represent a special case of the more complete three-trans-cisterna model, and that both of these two different models of TGN may apply, depending on the cell type and the functional conditions. In support of this possibility, it is clear that the TGN can vary significantly in both size and composition across different cell types. For example, cells with a well-developed endo-lysosomal system and without secretory granules exhibit an extensive tubular TGN, while the tubular component of the TGN is reduced in cells that are specialized in regulated secretion (Clermont et al. 1995). This appears to occur in the latter because most of the TGN membranes are used for the packaging of secretory granules, and the tubules only form the thin bridges between the granule precursors. Such an organization simplifies the release of the granules, which occurs through the rupture of these tubular elements (Clermont et al. 1995). Similarly, human fibroblasts have large collagen-containing distensions that are connected to tubular elements in the trans-Golgi area (Polishchuk et al. 2003). Thus, the morphology and size of the TGN apparently depend on the predominant type and amount of cargo protein departing from the Golgi complex. The morphology of the TGN depends also on its secretory status. When exit from the TGN is blocked by a lowering of the temperature to 20°C in the presence of abundant cargo (Matlin and Simons 1983), the volume and surface areas of the TGN increase greatly (Griffiths et al. 1989); while in the absence of cargo, the 20°C block results in all three of the trans-most cisternae producing bulging exit domains (Ladinsky et al. 2002). At physiological temperatures, three-dimensional (3D) analyses of the Golgi stacks have revealed well developed, tubular–reticular, TGN-like membranes at the trans side of the Golgi complex in actively secreting cells (Trucco et al. 2004), while the TGN is nearly absent in quiescent cells (Trucco et al. 2004). Thus, the TGN is a very dynamic structure, the shape and size of which is potently regulated by the extent of traffic flowing through it. To discover the regulatory mechanisms is an important challenge for the future. Another challenge will be to define whether the three-cisterna organization really underlies a functional specialization of these cisternae in the sorting and packaging of different cargo classes, as suggested by the presence of clathrin buds only on the last cisterna of the Golgi stack. So far, the lack of immunolabeling studies does not allow determining whether different trans cisternae really do contain different transport proteins. In addition to being an organelle of the biosynthetic pathways, the TGN is involved in endocytic transport routes (Griffiths and Simons 1986; Mellman and Simons 1992; Pavelka et al. 1998). Several studies with different cell types showed that the TGN and Golgi stacks can contain plasma-membrane constituents and internalized materials (Stoorvogel et al. 1988; van Weert et al. 1997). These exchanges with the endocytic systems could well introduce a further level of structural complications, since the recycling endosomes, in particular, reside in the Golgi area and are tubular in nature. In conclusion, both the dynamics and the structure of the TGN are variable and incompletely defined, thereby presenting an additional layer of difficulty to our understanding of the biogenesis of the carriers that depart from this organelle. Formation of PGC precursors at the TGN There are two key aspects to the formation of the PGC precursors: their morphogenesis at the TGN; and the segregation between different types of cargo proteins that are targeted for different destinations, as well as between the Golgi-resident proteins, such as the Golgi enzymes. Regarding PGCs morphogenesis, in principle, the tubular carrier precursors might form through the mechanical pulling force exerted by microtubule-based motors on a flatter, parent, membrane domain (e.g., a Golgi cisterna). This has been shown to be mechanistically possible for both artificial and natural membranes (Roux et al. 2002). If this is the case, the PGC precursors would be expected to be essentially simple linear tubular structures (Roux et al. 2002). The second possibility is that these precursors are actually formed from the tubular subdomains of the TGN, which are generated prior to docking and extrusion by microtubules. In this case, the GPCs should reflect the morphologically complex structure of their parent membranes. Recent studies have provided compelling support in favor of the latter model (Polishchuk et al. 2003). First, both free PGCs and their precursors at the TGN comprise complex tubulo-reticular structures, which have often been described as the main component of the TGN in many cell types (Clermont et al. 1995). For example, PGCs carrying VSVG show a mostly tubular morphology, and can have a complex structure that even contains clearly visible fenestrae (Polishchuk et al. 2003); this would be expected of membranes that derive from protrusions of the TGN. Indeed, PGC precursors that have been visualized using CLEM, appear to be comprised of tubular segments that are interconnected with complex branching and fenestrated membranes; they are also seen to be continuous with the parent membranes of the Golgi stack (Polishchuk et al. 2003). Similarly, carriers containing the apical cargo protein hemagglutinin (HA) frequently have a tubular morphology as well as HA-positive domains at the TGN (Puertollano et al. 2001). Thus structural similarities between PGC precursors and the TGN appear to be a common feature of different types of PGCs. This strongly suggests that PGCs form via the fission of an entire precursor domain (or a large part thereof) from the rest of the TGN membranes. Thus the question regarding the mechanism by which the originally flat Golgi membranes are converted into highly bent, tubular–reticular TGN structures arises. Essentially, this can be achieved by one of two mechanisms: either by the action of proteins that can bend these membranes into tubules, in a manner that is possibly similar to that of amphiphysin, endophilin, sorting nexins and others (Antonny 2006), or via the alteration of the lipid composition of these trans-Golgi membranes, which can be mediated, in turn, by several processes. For instance, changes in lipid content, and hence membrane curvature, can be modulated via the lipid-metabolizing enzymes that reside at the Golgi complex (for review, see De Matteis and Godi 2004; Luini et al. 2005). The transmembrane or inter-organelle transfer of lipids can contribute to the generation of particular lipid environments in the membranes of the TGN. In this respect, it is important to note that numerous contact sites between the ER and the trans-cisternae of the Golgi complex have been detected by electron microscopy tomography (Ladinsky et al. 1999, 2002). Such contact sites can be explored for lipid transfer between the ER and trans-Golgi that is mediated by specific lipid-transfer proteins. Also, since the cisterna-like morphology of Golgi compartments can be stabilized by large polymers of Golgi enzymes (Nilsson et al. 1996), the loss of oligomerized Golgi enzymes in the trans-Golgi should, in principle, favor cisterna-to-tubule transformation of cargo-containing membranes. Moreover, this process can be accompanied by the loss of the stacking mechanisms at the trans side of the Golgi complex. Here, GRASP65 and GRASP55 are two proteins that have been suggested to be involved in the maintenance of cisterna juxtapositioning, and they are located mainly at the cis and medial Golgi, rather than the trans-Golgi (Barr et al. 1997; Shorter et al. 1999). So this intercisternal “glue” may be gradually lost as a cisterna progresses towards the trans pole of the Golgi complex. This has been confirmed both in mammals and yeast by the observation that the trans-most cisterna frequently peels off from the main Golgi stack (Clermont et al. 1995; Mogelsvang et al. 2003). It is possible that more than two or more mechanisms act in synergy to provide this transition from flat Golgi cisternae to the tubulo-reticular TGN morphology. In addition to the formation of tubular domains at the exit face of the Golgi complex, this is also the level at which cargo proteins that are directed to different post-Golgi destinations should be sorted. The classical view in the membrane transport field implies that sorting at the TGN (as well as throughout the whole secretory pathway) is driven mainly by the coat-adaptor-protein machinery, which interacts specifically with amino-acid signals of certain transmembrane cargo proteins; this then provides the mechanical force for budding and fission of transport vesicle (Mellman and Warren 2000). This holds true for the endo-lysosome-directed carriers that have been thoroughly characterized. These carriers consist of clusters of clathrin-coated buds that are connected by tubular regions, and thus exhibit a grape-like structure. In contrast, PGCs carrying a cargo like the G-protein of vesicular stomatitis virus (VSVG) form in a coat and an AP-independent manner. Both PGCs and their precursors do not show β-COP or γ-, δ- and ε-adaptins at their membranes (Polishchuk et al. 2003). Other adaptors, such as the GGAs, are excluded from VSVG carriers as well (Puertollano et al. 2003; Polishchuk et al. 2003). Similarly, coats and adaptors have also never been detected on PGCs that are carrying proteins to the apical surface in polarized cells (Kreitzer et al. 2003). Thus, these carriers should form either by virtue of some still-unknown adaptors that cannot yet be visualized by electron microscopy, or by their association with specific lipid microdomains that are involved in sorting (Schuck and Simons 2004). This might be the case for proteins directed to the apical surface in polarized epithelial cells, the concentration of which at the TGN appear to be through their partition into cholesterol- and sphingolipid-rich membrane domains that are known as “rafts” (Schuck and Simons 2004). The scission of PGCs from the TGN The morpho-dynamics of the fission process are also fairly complex. Observations in living cells have revealed that the fission of PGCs frequently coincides with mechanical pulling of carrier precursor from the TGN along microtubules. Apparently, the pulling force which the molecular motors such as kinesin (see below) can apply to the TGN membranes is important to facilitate the extension of PGC precursors from the Golgi body and for the later fission of the PGC (Kreitzer et al. 2000; Polishchuk et al. 2003). In cell-free systems, the addition of kinesin to Golgi membranes (and even to liposomes) together with microtubules induces the formation of tubule-like membranes that are similar to PGC precursors (Roux et al. 2002), while a block of kinesin function by microinjection of an inhibitory antibody (Kreitzer et al. 2000) or expression of the headless kinesin mutant (Nakata and Hirokawa 2003) prevent PGC formation from the Golgi complex. Kinesin has been seen to be associated with the tip of PGC precursors, although it can also attach to other points along the PGC precursor membrane (Polishchuk et al. 2003). The movement of kinesins along microtubules can then create tension within the PGC precursor that will facilitate the fission process. Indeed, based on in vitro data, membranes under tension have recently been proposed to have an important role in fission (Roux et al. 2006). However, PGCs can also form when microtubules have been destroyed by nocodozole treatment; in this case, the pulling force to create membrane tension in fission-prone regions might be provided by actin motors (Warner et al. 2003; Sahlender et al. 2005). Live cell imaging and CLEM have also shown that fission does not take place randomly along the membranes of PGC precursors. In this case, which regions of a PGC precursor can be defined as prone to fission? First, this depends on the geometry of the precursor membranes. Our data suggest that fission usually takes place at the thinnest parts of the PGC precursor (Fig. 2a), which at the electron microscopy level corresponding to thin tubular segments of membranes (Polishchuk et al. 2003). In contrast, fission does not take place at the TGN regions with a complex morphology (i.e., in those containing tubular networks and branching tubules, or in thick vacuolar regions). Obviously, the precise points of fission will define not only the compositions of the PGC carriers, but also their morphology. If fission occurs close to the tip of the TGN tubule, a carrier will be smaller in size. In contrast, larger PGCs can form by cleavage at the base of the PGC precursors (Fig. 2a). Similarly, endosome-directed PGCs can apparently detach from the TGN as simple cargo-containing vesicles if the fission occurs at the neck of the clathrin-coated buds (Fig. 2b). However, many clathrin-positive PGCs have a grape-like morphology (tubule with several buds), suggesting that entire chunks of TGN membranes containing 2–3 clathrin-coated buds can be cleaved from the Golgi complex (Polishchuk et al. 2006). Fig. 2Fission of post-Golgi transport carriers. a TGN precursors of post-Golgi carriers are pulled along microtubules by kinesin. The fission (red line) of the carriers occurs at the thinnest parts of the PGC precursor, which correspond to thin tubular segments of the TGN membrane at the electron microscopy level. In contrast, fission does not take place at the TGN regions with a complex morphology (i.e., containing tubular networks and fenestrae, or in thick vacuolar regions). If fission occurs close to the tip of a PGC precursor, the carrier will be smaller in size (1). In contrast, larger PGCs can be formed by cleavage at the bottom of a PGC precursor ( 2). b PGCs directed to endosomes detach from the TGN as simple clathrin-coated vesicles if fission (red line) occur at the neck of the clathrin-coated bud (1). Alternatively, entire chunks of the TGN membrane containing 2-3 clathrin-coated buds can be cleaved from the Golgi complex Another factor that might be important for PGC shaping is the type of cargo that is embedded in the PGC. For example, procollagen-I usually forms quite large aggregates that are visible within 300–400-nm-diameter membrane distensions of the Golgi membranes. As a consequence, similar distensions have been detected in most collagen-containing PGCs (Polishchuk et al. 2003; Canty et al. 2004). Movement of PGCs from the Golgi complex to the target membrane After fission from the TGN, PGCs move to the acceptor membrane. Different members of the kinesin superfamily (Kamal et al. 2000; Nakata and Hirokawa 2003; Teng et al. 2005), and also the other microtubule motor dynein (Tai et al. 1999), have been shown to drive post-Golgi transport of specific cargo to various destinations. This high fidelity of cargo selection by molecular motors at the TGN and their further delivery to the correct surface or intracellular domain may be regulated by interactions of motor protein directly with the cargo (Kamal et al. 2000; Teng et al. 2005) or with components of the sorting machinery at the TGN (Nakagawa et al. 2000). As an example, transport of HA and annexin 13b to the apical surface in epithelial cells relies on raft-associated motor KIFC3 (Noda et al. 2001). KIF13A operates in the other post-Golgi route that is used for the transport of the mannose-6-phosphate receptor (Nakagawa et al. 2000). A number of neuronal proteins, including bAPP, GAP43 and vamp-2, require KIF5 for their correct targeting (Nakata and Hirokawa 2003), while the microtubule minus-end-directed motor dynein has been shown to support rhodopsin transport in rod photoreceptors (Tai et al. 1999; Yeh et al. 2006). Selection of specific cargoes by motors could be driven by various mechanisms. The simplest would use a direct interaction between the motor and a specific domain of a cargo protein, as is seen for the dynein light chain and the cytoplasmic tail of rhodopsin (Tai et al. 1999). Alternatively, adaptor proteins could serve as a bridge between a motor and its cargo. For example, KIF13A transports the mannose-6-phosphate receptor through its interaction with the AP-1 complex (Nakagawa et al. 2000). Finally, both motor and cargo could associate with the same specific lipid microdomain, as for instance, for KIFC3 and the apically targeted annexin XIIIb. Another issue that needs to be addressed is whether any of these sorting processes take place in the PGCs while they are moving toward their acceptor compartment. This happens, for example, with the maturation of secretory granules when the mannose-6-phosphate receptor is concentrated and sorted from the secretory granules by clathrin-coated vesicles (Klumperman et al. 1998). So several strategies have been used to determine whether similar sorting events happen with PGCs. Mature exocytic carriers can be arrested before their fusion with the plasma membrane either by microinjection of an anti-NSF antibody or by treatment with tannic acid (which fixes the plasma membrane but does not penetrate inside the cell). In contrast to secretory granules, the comparison of mature and newly formed PGCs did not reveal significant changes in either their ultrastructure or their composition (Polishchuk et al. 2003, 2004). Similarly, mature Golgi-to-endosome carriers were accumulated in cells upon endosome inactivation. However, they did not show any significant transformation, except for a very moderate reduction in the area covered by clathrin (Polishchuk et al. 2006). Live-cell imaging of subconfluent MDCK cells has shown that PCGs that initially contained both a basolateral marker (VSVG-CFP) and an apical marker (GPI-YFP) did not sort out either of these cargoes into any separate structures, and instead delivered both of the proteins to the plasma membrane (Polishchuk et al. 2004). GPI-GFP was then sorted from the basolateral surface to the apical membrane through transcytosis (Polishchuk et al. 2004). On the other hand, the partitioning of two proteins from a common PGC into separate carriers has also been reported (Jacob and Naim 2001). This suggests that sorting from the PGC may exist, but that it should depend on the nature of the cargo proteins being transported. Intermediate stations in post-Golgi transport The complexity of sorting events in the post-Golgi space appears to be even worse since the discovery that certain cargoes may pass through the endosomal compartments before their arrival at the plasma membrane. Sporadic reports in the past have suggested that some secretory proteins do not move from the Golgi complex directly to the plasma membrane, but instead pass through an endocytic intermediate on their way to the cell surface (Leitinger et al. 1995). This indirect “through-endosome” delivery of cargo to the plasma membrane might be also facilitated by the intimate association of TGN membranes with number of endocytic compartments in the perinuclear area of the cell (Marsh et al. 2001). The list of the proteins using this pathway has been recently updated, and it has now been shown that in MDCK epithelial cells, VSVG, the LDL receptor and E-cadherin can be detected in the endosomes before their exit to the plasma membrane (Ang et al. 2004; Lock and Stow 2005). These findings, however, promote a number of further questions. The first is whether this transport route is ubiquitous (i.e., does it exist in different cells?). The second question is whether different cargoes move through the same endosomal compartment on their way to the cell surface in epithelial cells? A number of proteins (such as VSVG and the LDL receptor) have been reported to use a Rab8-positive sub-population of endosomes as an intermediate station on their way to the basolateral membrane in epithelial cells (Ang et al. 2003). These proteins require the AP-1B adaptor complex to be properly sorted from these Rab8 endosomes towards the basolateral surface domain (Ang et al. 2003). Other cargoes (such as E-cadherin, for example) move to the plasma membrane through a Rab11 endocytic compartment (Lock and Stow 2005). It remains to be determined, however, whether there is any cross-talk between these Rab8- and Rab11-dependent routes. This possibility apparently exists, since Rab8 and Rab11 endosomes are both accessible to transferrin (Ang et al. 2004; Lock and Stow 2005). Finally it is important to clarify as to what extent this through-endosome transport route is used by different cargo proteins in epithelial cells. This issue has been partially addressed by the silencing of the μ1B subunit of the AP-1B adaptor complex, which resides at the endosomes and is required for the correct delivery of many basolateral proteins. The use of RNA interference has revealed, however, that a number of cargoes (such as transferrin and Fc receptors, for example) can be targeted correctly even in the absence of AP-1B, which suggests that these proteins can move directly from the Golgi complex to the basolateral surface without crossing the endocytic routes (Gravotta et al. 2007). The existence of more than one basolateral pathway has also been supported by the observation that transport of different basolateral proteins can be selectively regulated by different PKD isoforms (Yeaman et al. 2004). Thus, further efforts need to be made to understand to what extent an endocytic post-Golgi compartment is involved in the sorting and transport events of cell-surface proteins. Conclusions The extensive characterization of PGC morphology by video and electron microscopy has provided a framework for the positioning of the molecular machineries in the morpho-functional maps of various trafficking segments. At the same time, many molecular players in TGN-to-plasma membrane transport have been identified. The attribution of each of the molecular components to each of the pathways is probably now the main challenge. We believe that the development of specific assays will significantly assist in the achieving of this objective. Video microscopy of GFP-tagged cargo proteins has allowed us to evaluate the process of PGC formation, as well as the speed and directionality of PGC movement through the cytosol. Other assays use tannic acid treatment (Polishchuk et al. 2004; Jakob et al. 2006) or horse-radish-peroxidase-based endosome immobilization (Ang et al. 2004; Polishchuk et al. 2006) to prevent the fusion of PGCs with their target membranes. These conditions allow us to trap mature PGCs and to compare their compositions to newly formed carriers and the TGN membranes. Moreover, preferential sites of PGC docking and fusion can also be easily identified in this way. The combination of these experimental approaches with specific molecular inhibitors now allows us to attribute a protein of interest to one of the steps in post-Golgi transport.	[ "post-golgi transport", "trans-golgi network", "post-golgi carriers", "golgi complex" ]	[ "P", "P", "P", "P" ]
Neuroimage-2-1-2330063	Tractography of the parahippocampal gyrus and material specific memory impairment in unilateral temporal lobe epilepsy	Introduction Temporal lobe epilepsy (TLE) is associated with disrupted memory function. The structural changes underlying this memory impairment have not been demonstrated previously with tractography. Introduction Temporal lobe epilepsy (TLE) is the most common form of refractory focal epilepsy (Crawford, 2000). Parahippocampal structures, which are critically implicated in the generation and propagation of seizures in TLE (Avoli et al., 2002; Bertram, 2006; Du et al., 1993; Plate et al., 1993; Rutecki et al., 1989; Spencer and Spencer, 1994), are essential for declarative memory (Eichenbaum, 2000). Longitudinal neuropsychological studies have shown that persisting epilepsy is associated with progressive memory impairment (Dodrill, 2002; Helmstaedter et al., 2003). Those who undergo anterior temporal lobe resection (ATLR) are at further risk of memory impairment, the nature of which depends on whether surgery is on the dominant or non-dominant side (Ivnik et al., 1987; Spiers et al., 2001). Neuropsychological assessment, quantitative MRI, and latterly functional MRI (fMRI) indicate the role of the medial temporal lobe structures (MTL) in sustaining material specific memory functions (Powell et al., 2005), and the reorganisation of memory that occurs with TLE (Powell et al., 2007a). These and lesion deficit studies (Frisk and Milner, 1990; Smith and Milner, 1981) have shown that memory deficit after ATLR is related to the functional integrity of the parahippocampal structures. The strength of connections, or structural connectivity, of the parahippocampal gyrus in TLE has not been evaluated, or related to function. Tractography is a technique which uses diffusion tensor imaging (DTI) data to delineate white matter tracts, and to quantify their volume, and infer structural characteristics (Johansen-Berg and Behrens, 2006). Using one such tractography technique, Probabilistic Index of Connectivity or PICo (Parker et al., 2003; Parker and Alexander 2003), we evaluated the structural connectivity of the parahippocampal gyrus in TLE, with the hypothesis that this would be impaired ipsilaterally to the seizure focus, and that the degree of any such impairment would correlate with material specific memory function. Methods Subjects We studied 18 patients (median age 33.5 years; range 22–47 years; 11 males) with medically refractory TLE undergoing pre-surgical evaluation at the National Hospital for Neurology and Neurosurgery, London, UK. All patients had undergone structural MRI at 1.5T (Duncan, 1997). Of the eight left TLE patients, seven had hippocampal sclerosis (HS) (one also had a ganglioglioma in the left fusiform gyrus) and one had a MTL dysembryoblastic neuroepithelial tumour (DNET). Of the ten right TLE patients, seven had HS, one had a right MTL glioma, another had a right MTL DNET, and the other had right superior temporal focal cortical dysplasia (FCD). Video-EEG had confirmed seizure onset in the MTL ipsilateral to the clinically defined seizure site, and all patients had a normal, contralateral hippocampus based on qualitative and quantitative MRI criteria (Woermann et al., 1998). All patients were on anti-epileptic medication, and were fluent English language speakers. Handedness was determined using the Edinburgh handedness inventory (Oldfield, 1971), and language dominance was determined using a range of fMRI tasks which have been described previously and include the use of verbal fluency, and reading tasks (Powell et al., 2006). All patients underwent a standardised pre-surgical neuropsychological assessment (Baxendale et al., 1998). Patient demographics, neurological test results, and surgical outcome data are listed in Table 1. The ILAE classification of post-operative seizure outcome following epilepsy surgery was used (Wieser et al., 2001). We also studied 10 right-handed native English speaking, healthy volunteers (median age 29.5 years; range 23 to 50 years; 7 females). The study was approved by the National Hospital for Neurology and Neurosurgery and the Institute of Neurology Joint Ethics Committee, and informed written consent was obtained from all subjects. MR data acquisition MRI studies were performed on a 1.5-T GE Signa Horizon scanner (General Electric, Wakashua, Milwaukee, Wisconsin, USA). Standard imaging gradients with a maximum strength of 22 m Tm- 1 and slew rate 120 Tm- 1s- 1 were used. All data were acquired using a standard quadrature birdcage head coil for both RF transmission and reception. The scanning protocol also included a coronal T1-weighted volumetric acquisition sequence with 1.5-mm-thick slices, and hippocampal volumes were determined according to a previously described method (Moran et al., 1999). Diffusion tensor imaging The DTI acquisition sequence was a single-shot spin-echo planar imaging (EPI) sequence, cardiac gated (triggering occurring every QRS complex) (Wheeler-Kingshott et al., 2002), with TE = 95 ms. Sets of 60 contiguous 2.3-mm thickness axial slices were obtained, covering the whole brain, with diffusion sensitizing gradients applied in each of 54 non-colinear directions (maximum b value of 1148 mm2 s- 1 (δ = 34 ms, Δ = 40 ms, using full gradient strength of 22 mTm- 1)) along with 6 non-diffusion weighted (b = 0) scans. The field of view was 24 cm, and the acquisition matrix size was 96 × 96, zero filled to 128 × 128 during reconstruction so that the reconstructed voxel size was 1.8 × 1.8 × 2.3 mm3. The DTI acquisition time for a total of 3600 image slices was approximately 25 min (depending on the heart rate). We used the method of Parker and Alexander (Parker et al., 2003; Parker and Alexander, 2003) to reduce fibre orientation ambiguities in voxels containing fibre crossings. Voxels in which the single tensor fitted the data poorly were identified using the spherical-harmonic voxel-classification algorithm of Alexander et al. (Alexander et al., 2002). In these voxels a mixture of two Gaussian probability densities was fitted and the principal diffusion directions of the two diffusion tensors provided estimates of the orientations of the crossing fibres (Tuch et al., 2002). In all other voxels a single tensor model was fitted. For all voxels, fractional anisotropy (FA) maps were generated from the single tensor fit (Pierpaoli et al., 1996; Pierpaoli and Basser, 1996). Tractography All scans were transferred to a Unix workstation for processing. We used the PICo algorithm extended to cope with crossing fibres (Parker et al., 2003; Parker and Alexander, 2003) to track from anatomically defined regions of interest (ROIs) within the parahippocampal gyrus. This algorithm adapts the commonly used streamline approach to exploit the uncertainty due to noise in one or more fibre orientations defined for each voxel. This uncertainty is defined using probability density functions (PDFs) constructed using simulations of the effect of realistic data noise on fibre directions obtained from the mixture model (Parker and Alexander, 2003). The streamline process is repeated using Monte Carlo methods to generate maps of connection probability or confidence of connection from the chosen start region(s). The anatomical definition of the ROI was based on a previously published tractography analysis of the parahippocampal gyrus in healthy subjects (Powell et al., 2004). Viewing the FA images in three orthogonal planes using MRIcro (http://www.psychology. nottingham.ac.uk), the centre of the white matter tract just anterior to the brainstem, and posterior to the cerebral peduncles was selected, such that the parahippocampal gyrus was defined at its longest in the corresponding sagittal view (Fig. 1). The corresponding coronal slice was then used to select two adjacent voxels in a left-right direction, such that they both lay within the white matter tract on axial and sagittal views. This process was repeated in one anterior and one posterior coronal slice. This method was chosen as the parahippocampal gyrus runs anterior to posterior, inferior to superior and medial to lateral within the medial temporal lobe. The principal eigenvector of each voxel, when viewed using PICo and projected on the axial plane was orientated anterior-posterior. A threshold of FA ≥ 0.1, and curvature threshold of 180° were set for tractography. Neuropsychological tests The list learning and design learning tests were used to assess material specific memory function (Coughlan and Hollows, 1985). In the verbal learning task the subject is read a list of 15 words five times, and on each presentation attempts to recall as many of the words as possible. The overall percentage of correct responses was used as the measure of verbal memory efficiency. For non-verbal memory we employed a design learning task; the subject is presented with a visual design on five occasions with recall being tested after each presentation. The percentage of correct responses over the five trials was used as a second measure of non-verbal memory efficiency. These tests form part of our presurgical memory assessment in TLE cases, and have proven least affected by performance anxiety, have a good test-retest reliability and are sensitive indicators of medial temporal lobe function (Baxendale et al., 1998). Neuropsychological test results are listed in Table 2. Data analysis All data were analysed using SPSS (11.0.0). It was first verified whether all parameters were normally distributed using the Kolmogorov–Smirnov test for normal distribution. Group differences for age were determined by a one-way analysis of variance (ANOVA), and gender distribution was assessed using the Pearson's χ2 test. The age of onset of epilepsy, duration of epilepsy, and frequency of complex partial (CPS) and secondarily generalized seizures (SGS) in right and left TLE patients were compared using the Mann–Whitney U test. Each subject's output tractography connection probability map was spatially normalised by mapping into a standard space using the MNI template provided by SPM2 (Wellcome Department of Imaging Neuroscience, London; http://www.fil.ion.ucl.ac.uk/spm). Binary masks at a threshold connection probability value of 0.05 were then constructed. Our group has previously demonstrated in this same group of patients and control subjects, that the threshold 0.05 strikes a balance between losing non-specific low probability connections, while retaining the main body of the pathways (Powell et al., 2006). Binary masks at this threshold were therefore averaged across each group, to produce variability (or commonality) maps indicating the degree of spatial variability and overlap of the identified connections (Parker et al., 2005). A voxel commonality value C of 1.0 indicates that every individual had a connection identified in this voxel, while a C value of 0 indicates that none of them did (Parker et al., 2005). Tracts were then assessed visually in all three planes for visual symmetry and size/extent using MRIcro. Normalised tract volumes were calculated for the connecting tracts from the left and right PHG of each control and patient at a threshold of 0.05 (Toosy et al., 2004). An asymmetry index for volume (AIvol) defined as AIvol = [100 × (Right Volume − Left Volume)] / [(Right Volume + Left Volume)/2] was calculated (Jutila et al., 2001), and the mean values between groups compared using a one way ANOVA analysis. Comparisons between the control AI values, and left and right TLE AI values were carried out using post-hoc Dunnett t-tests. Two way mixed ANOVA with one between subjects factor (group –— controls or TLE (both left and right)) and one within subjects factor (hemisphere — left or right) was used to test for the effect of interaction between group and hemisphere on volume, and unpaired t-tests were used to compare the patient and control group tract volumes. The mean FA of the connected volume was calculated in native space for the left and right tracts in controls and patients. This was carried out by multiplying the native, thresholded, binarised images with that subject's whole brain FA image, in order to calculate the mean intensity value of the voxels. An asymmetry index for FA (AIFA) defined as AIFA = [100 × (Right FA - Left FA)] / [(Right FA + Left FA)/2] (Jutila et al., 2001) was calculated, and the mean values between groups compared using one way ANOVA analysis. Comparisons between the control AI values and left and right TLE AI values were carried out using post-hoc Dunnett t-tests. Two way mixed ANOVA with one between subjects factor (group — controls or TLE (both left and right)) and one within subjects factor (hemisphere — left or right) were used to test for the effect of an interaction between group and hemisphere on FA, and unpaired t-tests were used to compare the patient and control group tract FA values. Pearson's correlation test was used to evaluate the evidence for a correlation of hippocampal volume with tract volume and FA, and tract volume with tract FA, ipsilateral and contralateral to the seizure focus. Performance on material specific memory measures was investigated for evidence of a correlation with tract volume and FA in right and left TLE groups, omitting patient 14 because of his atypical language dominance. Results Demographic analysis There was no significant difference in the mean age or gender distribution of participants in the three groups (controls, left TLE, right TLE). There was no significant difference in the age of onset, duration of epilepsy or frequency of CPS and SGS between left and right TLE patients. Qualitative tract analysis In controls, PHG connections were visually symmetric. Connections between the para-hippocampal gyrus and anterior temporal lobe, orbitofrontal areas and posterior temporal and extrastriate occipital areas were observed as documented previously (Powell et al., 2004). There was a clear decrease in ipsilateral compared with contralateral connections in left TLE (Fig. 2), though an ipsilateral reduction was not evident in the commonality map of the right TLE group (Fig. 3). Quantitative tract analysis There was a significant difference in AIvol between groups [F(2,25) = 3.31 p = 0.05]. The AIvol was greater in left TLE patients than in controls (p = 0.05) with a mean 22% reduction in volume on the left (Table 3). There was no significant difference between the AIvol in right TLE patients and controls. There was no significant interaction between group and hemisphere for tract volume, and no significant differences between left and right volumes in left/right TLE patients, and controls. There was a significant difference in AIFA between groups [F(2, 25) = 4.92 p = 0.02], attributable to reduced FA on the left in left TLE (p = 0.02 against controls). There was no significant difference between the AIFA in right TLE patients and controls. There was a significant interaction between group and hemisphere on tract FA [F(2, 25) = 4.35 p = 0.02], with FA being lower on the left in left TLE patients, compared with controls (p = 0.03). Although a similar trend was present in right TLE patients, with FA being lower on the right side compared with controls, this was not significant (p = 0.06). Correlation of tract volume and FA with pre-surgical material specific memory All but one patient were left hemisphere dominant for language. Patient 14 was left handed and right hemisphere dominant on both fMRI and intracarotid amytal testing. He was therefore omitted from the correlation analysis. In the left TLE patients, FA of the left PHG connections were correlated significantly with pre-surgical verbal learning (r = 0.88, p = 0.002), and right FA correlated significantly against pre-surgical design learning (r = 0.63, p = 0.05) (Fig. 4). Left and right tract volumes were not significantly correlated with verbal learning and design learning, respectively. In the right TLE patients, there were no significant correlations between left or right tract FA or volume, with pre-operative verbal learning or design learning respectively. Correlations of hippocampal volumes with tract volumes and tract FA There was no significant correlation between ipsilateral hippocampal volume and ipsilateral tract volumes or FA nor between contralateral hippocampal volumes and contralateral tract volume or FA. Correlation analysis of tract volume with tract FA There was a significant correlation between tract FA and tract volume (r = 0.61, p = 0.008) ipsilateral to seizure focus (Fig. 5), but not contralateral to seizure focus. Discussion Our principal finding was that in TLE the white matter connections of the parahippocampal gyrus ipsilateral to the seizure focus had smaller volumes and decreased FA. This was statistically significant in left but not right TLE patients. Furthermore, in left TLE, decreased FA was associated with poorer performance on both material specific memory measures. These results are consistent with the hypothesis that TLE involves dysfunction and structural changes in a network that includes the parahippocampal gyrus. The connections of the human parahippocampal gyrus, visualized with tractography, have been described in a control population (Powell et al., 2004). Using a different tractography algorithm (Fast Marching Tractography or FMT) connectivity was found between the parahippocampal gyrus and the anterior temporal lobe, orbitofrontal areas, posterior temporal lobe and extra-striate occipital lobe via the lingual and fusiform gyri. These findings are similar to those in this control group. This is the first study to use tractography to quantitatively assess the structural changes in the parahippocampal gyrus in TLE. Several volumetric MRI (Bernasconi et al., 2003; Bonilha et al., 2003) and voxel-based morphometry (VBM) studies (Bernasconi et al., 2004; Bonilha et al., 2004; Keller et al., 2002b,a, 2004) have investigated the extra-hippocampal structural changes in unilateral TLE. The para-hippocampal gyrus or its sub-regions including the entorhinal cortex, parahippocampal cortex and perirhinal cortex have been shown to be affected in TLE. Several studies suggested that the degree of atrophy and regional distribution was more extensive in left TLE, though none were specifically designed to evaluate this (Bernasconi et al., 2004; Bonilha et al., 2004; Keller et al., 2002b). It has however been recently demonstrated that the distribution and severity of extra-hippocampal grey matter loss is more extensive in left TLE patients than right TLE patients (Bonilha et al., 2007). Our study is the first to show that the white matter connections of the parahippocampal gyrus are affected by unilateral TLE. This process appears to be more severe in left TLE compared with right TLE. A similar pattern was observed in another tractography study that assessed language pathways in TLE patients (Powell et al., 2007b), although in this case there was also asymmetry of connections in controls, with more extensive connections in the speech dominant hemisphere. Ultimately, longitudinal quantitative MRI studies will be necessary to determine the effects of underlying cause, seizures, medications, and co-morbidity on the white matter structures and connections of the brain. Parahippocampal connections and memory in temporal lobe epilepsy The memory deficits associated with TLE, particularly left TLE, can be disabling. Consequently, much attention has been focused on the effects of both chronic TLE and ATLR on memory. While the hippocampus plays a critical role in the initial formation of memories (Eichenbaum, 2000), the parahippocampal region is thought to be involved in the intersection between perception and memory, and the translation of material into a more permanent storage in the cortical association areas (Eichenbaum, 2000; Murray and Bussey, 1999). Animal models have shown that the parahippocampal region is important for recognition memory (Brown and Aggleton, 2001; Eichenbaum, 2000), and that selective lesions to the parahippocampal area can severely impair memory (Suzuki et al., 1993). Furthermore, the functions of the medial temporal lobe are highly lateralized and the classic model of material specific memory predicts that lesions in the left hippocampal system impair verbal memory retrieval (Hermann et al., 1997), while those in the right hippocampal system affect non-verbal memory, though these findings are less consistent (Alessio et al., 2004). In left TLE patients there was a significant correlation between parahippocampal tract FA and material specific memory measures on both the left and right sides. No significant correlations were seen with respect to tract volume in left TLE, and neither volume nor FA in right TLE. The finding of fewer occipital connections in the left TLE group compared with right TLE patients and controls may be significant in this respect (Figs. 2 and 3). Animal models suggest that potential roles for these connections may involve the priming of mesial temporal lobe structures to facilitate the consolidation of visual memory, or enhancing the visual processing of emotionally significant stimuli (Suzuki et al., 1993; Suzuki and Amaral, 1994). Visual cues may be important not only for non-verbal memory, but may also be useful in verbal memory. This could explain the correlation of FA found with verbal and non-verbal memory in left TLE patients. In the right TLE patients the occipital connections did not seem to be as affected, and hence these patients may have been able to use visual cues to aid both verbal and non-verbal memory. Several other studies have also observed correlations between MRI findings and memory dysfunction in TLE. Some have shown a correlation between left and right hippocampal MRI volumes, and verbal and non-verbal memory respectively (Baxendale et al., 1998; Kalviainen et al., 1997; Lencz et al., 1992). Others have shown correlations between hippocampal T2-signal, or abnormalities of MR spectroscopy and material specific measures (Gadian et al., 1996; Wendel et al., 2001). Lui et al. (Lui et al., 2005) examined 18 TLE patients and found correlations between apparent diffusion coefficients (ADC) in the left and right hippocampi and parahippocampal gyri and, verbal and non-verbal memory respectively. Others have demonstrated correlations between localized diffusion measures and domains of cognitive functioning in diseases as varied as schizophrenia (Nestor et al., 2004), HIV related dementia (Ragin et al., 2005), age related cognitive decline (Charlton et al., 2006), mild cognitive impairment (Rose et al., 2006), and Alzheimer's disease (Yoshiura et al., 2002). There are also studies that have observed that the relationship between the side of pathology and memory dysfunction appears to be more evident in those with left TLE than right TLE. Alessio et al evaluated the relationship between several medial temporal lobe structures and memory in 39 patients, and found a correlation between the degree of left sided hippocampal atrophy and verbal memory deficits, but not between right sided hippocampal atrophy and visual memory deficits (Alessio et al., 2004). In this study, it also appears that tractography derived parahippocampal FA is a more sensitive marker than volume of the functional integrity of tracts that form the hardwiring of circuits needed for memory (Charlton et al., 2006). Other tractography studies have shown that FA is a more sensitive and robust measure than volume of pathology in white matter tracts (Heiervang et al., 2006). In biological terms the interpretation of reduced anisotropy is complex (Beaulieu, 2002), and depends on the context or disease in which it is found (Alexander et al., 2007). In epilepsy, it may represent neuronal loss, gliosis, and structural disorganisation (Wieshmann et al., 1999). The association between hippocampal volume and parahippocampal Volume/FA No association between hippocampal volume, and tract volume or FA was found in this study. Other studies have reported mixed findings. While Bonhila et al found no correlations between the severity of hippocampal atrophy and grey matter volume in the parahippocampal region (Bonilha et al., 2003), Jutila et al reported positive correlations but only in a subgroup analysis of those patients with the most severe hippocampal atrophy (Jutila et al., 2001). This suggests that despite the fact that the hippocampus and parahippocampal gyrus are significantly interconnected, the structural size relationship between the two structures is not linear. There was also a significant correlation between ipsilateral tract volume and FA in the current study. Limitations of Study Only modest numbers of patients and controls were available for this study, as the scanner used was decommissioned and replaced with one of higher field strength, making comparison with more recent data impossible. These small numbers, particularly in the left TLE group, may have contributed to the pattern of observed changes in left but not right TLE patients. Furthermore, for this reason several morphological and functional factors that can influence cognition and memory in chronic epilepsy were not included as co-variates in the statistical analysis. The pathological basis of TLE was not homogenous throughout the group, and the degree of hippocampal atrophy was varied. Secondly the spatial resolution of the tractography was limited and we did not separately evaluate the entorhinal cortex, parahippocampal cortex, and perirhinal cortex (Duvernoy, 1998). The entorhinal cortex is considered to be the route by which data reaches the hippocampus. The perirhinal, and parahippocampal cortex on the other hand, provide the incoming connections to the entorhinal cortex, conveying information from the polymodal and unimodal cortices (Squire, 1991). A sub-regional analysis is an area that should be explored in future tractography studies. Conclusion This tractography study has shown disruption of the architecture and atrophy of the connections of the parahippocampal gyrus ipsilateral to the seizure focus in patients with refractory TLE, and these structural changes were associated with memory deficits evident on psychometric testing. This information has both diagnostic and prognostic implications. Larger, longitudinal studies at a higher resolution will enable both sub-regional analysis, and the investigation of other factors that may contribute to neuronal loss and structural changes, and subsequent memory impairment in patients with TLE.	[ "tractography", "parahippocampal gyrus", "memory", "epilepsy" ]	[ "P", "P", "P", "P" ]
Pediatr_Nephrol-3-1-1915616	Interactions of Shiga-like toxin with human peripheral blood monocytes	The cytotoxic effect of Shiga-like toxin (Stx; produced by certain Escherichia coli strains) plays a central role in typical hemolytic uremic syndrome (HUS). It damages the renal endothelium by inhibiting the cellular protein synthesis. Also, the monocyte has a specific receptor for Stx but is not sensitive for the cytotoxic effect. In this work, monocytes were studied as a potential transporter for Stx to the renal endothelium. Coincubation of isolated human monocytes loaded with Stx and target cells (vero cells and human umbilical vascular endothelial cells) were performed. Transfer was determined by measuring the protein synthesis of target cells and by flow cytometry. Furthermore, the effect of a temperature shift on loaded monocytes was investigated. Stx-loaded monocytes reduced the protein synthesis of target cells. After adding an antibody against Stx, incomplete recovery occurred. Also, adding only the supernatant of coincubation was followed by protein synthesis inhibition. Stx detached from its receptor on the monocyte after a change in temperature, and no release was detected without this temperature shift. Although the monocyte plays an important role in the pathogenesis of HUS, it has no role in the transfer of Stx. Introduction Hemolytic uremic syndrome (HUS) is a clinical syndrome consisting of three characteristic features: hemolytic anemia, thrombocytopenia, and acute renal failure [1]. In the new classification of HUS, infections due to Shiga-like toxin (Stx)-producing bacteria belong to the category “etiologically advanced” [2]. This work focuses on the form in which Stx-producing Escherichia coli is the most common pathogen [1]. It can produce several types of Stx, of which Stx1, Stx2, and Stx2c are most frequently associated with HUS [3, 4]. Stx plays a crucial role in the pathogenesis because of its cytotoxic effect on the renal endothelium. Both renal tubular epithelial cells and glomerular visceral epithelial cells (podocytes) are also sensitive to the toxic effect of Stx [5, 6]. It can inhibit the protein synthesis of these cells after specifically damaging the ribosomal RNA [7]. However, the question of how this toxin is targeted mainly to the kidney remains unsolved. Stx was never detected in the serum of patients, but it was detected in renal biopsy material of patients with HUS [8]. As a specific treatment for HUS is still lacking, more insight into the transport of this toxin might lead to new intervention strategies. After oral ingestion of the bacteria through contaminated food or water, the noninvasive bacteria adhere to the intestinal epithelial cells of the distal small bowel and colon. This leads to a rearrangement of the morphology of the cells and initiates inflammation [9, 10]. Bacterial flagellin plays an important role in this process [11]. Stx can probably reach the circulation because of active transport in these cells and also passively after damage to the intestinal cells [12]. Subsequently, it has to be transported in the circulation to reach its primary target, the renal endothelium. It is very tempting to look at the blood cells as a carrier for the toxin. Stx can bind to a specific receptor, which is a globotriaosylceramide (Gb3, Pk Antigen, CD77) [13]. This receptor is present on renal endothelial cells but also on blood cells. Stx binding has been described on red blood cells [14], B lymphocytes [15], and platelets, which also have an additional binding possibility (glycolipid, band 0.03) [16]. Several groups showed the existence of a specific binding of Stx on monocytes [17, 18, 19]. After binding to its receptor, Stx can be internalized. Whereas in epithelial cells the toxin follows the retrograde transport route and becomes cytotoxic, in monocytes it is targeted to the lysosomes and will get degraded [19]. During this transport, the monocyte becomes activated. This will lead to an increase of transcription factors, such as nuclear factor kappa B (NF-κB) and activator protein 1 (AP-1), and an upregulated production of cytokines such as interleukin (IL)-1β, tumor necrosis factor (TNF)-α, IL-6, and IL-8 [17, 20]. These events will have a pro-inflammatory effect. We postulated that, as the monocyte has a specific receptor, it might also function as a carrier to transport Stx to the renal endothelium. To investigate this hypothesis, Stx was loaded to isolated monocytes from healthy donors and coincubated with target cells [vero cells and human umbilical cord venous endothelial cells (HUVEC)]. The level of transfer was determined by measuring the protein synthesis of these target cells and the transfer of fluorescein isothiocyanate (FITC)-labeled B subunit of Stx1 with flow cytometry. Materials and methods Materials Stx2 was kindly provided by Dr. M. Karmali (Public Health Agency of Canada, Ontario, Canada). FITC-labeled Stx1B subunit and 125I-Stx1B subunit were a gift from Dr. L. Johannes (Institut Curie, Paris, France). Stx1B subunit is a useful tool for studying binding in monocytes [19]. It is the binding part of the toxin, whereas the enzymatic A subunit will only stimulate the uptake of the toxin and does not affect binding [21]. Vero cell medium consists of M199 (Gibco; Paisley/UK), fetal calf serum (FCS, Greiner Bio-One; Kremsmunster/Austria), penicillin/streptomycin (Gibco, Paisley/UK), and glutamine (MP Biomedicals; Eschwege/Germany). HUVEC medium is made of M199, human serum (HS; Cambrex; Walkersville/USA), newborn calf serum (NBCS; Gibco, Paisley/UK), penicillin/streptomycin, glutamine, heparine (Leo Pharma BV, Breda/The Netherlands), and endothelial-cell growth factor [22]. Ethylenediamine tetraacetic acid (EDTA) tubes were purchased from BD Vacutainer (Alphen aan de Rijn/The Netherlands). The MACS kit for negative selection of monocytes was provided by Miltenyi Biotec (Bergisch Gladbach/Germany). Hank’s balanced salt solution (HBSS) was ordered from MP Biomedicals (Eschwege/Germany). Human serum albumin (HSA) from Sanquin (Amsterdam/The Netherlands) and porcine gelatin from Fluka (Neu-Ulm, Germany) was used. Trichloroacetic acid and TNF-α was obtained from Sigma-Aldrich Chemie B.V. (Zwijndrecht/The Netherlands). The antibody against Stx2 (TMA-15) is well characterised [23]. It was a kind gift from Dr. Yamagami from the Department of Biomedical Research from the Teijin Institute, Tokyo, Japan. 3H-leucine and Ficoll-paque PLUS was purchased from Amersham Biosciences (Uppsala, Sweden). Culture plates were ordered from Corning Inc. (Corning, USA). Culture of vero cells and HUVEC Vero cells (renal epithelial cells of the African green monkey) were grown to confluency on 24-well plates (ordered from ATCC; Middlesex, UK). These cells have a high basal expression of Stx-receptor CD77. HUVEC were isolated, and these cells were grown to confluence on gelatin-coated 24-well plates [24]. Every 2 days, fresh medium was added to the cells. In contrast, HUVEC were preincubated for 24 h with TNF-α (10 ng/ml) to upregulate the expression of CD77. Isolation of monocytes and loading with Stx2 Fresh venous blood (20 ml) from 40 healthy donors was collected into EDTA tubes. Monocytes were isolated by negative selection using antibody-labeled beads (CD3, CD7, CD16, CD56, and CD123). After centrifugation of blood over Ficoll (20 min 400 g without break at room temperature), the interphase (containing monocytes, lymphocytes, and platelets) was collected. Platelets were removed by centrifugation (200 g 10 min at room temperature) before adding the beads. The purity of monocytes after the magnetic isolation was (as determined by flow cytometry) 80–85%. After the isolation, the monocytes were resuspended in HBSS with 1% HSA and placed on ice. In every experiment, monocytes from one donor were used. To load the monocytes with Stx2, the toxin was added to a concentration of 10 nM during a time period of 3 h [17]. The cells remained on ice. After proper washing to remove all unbound Stx2, the cells were resuspended in vero cell or HUVEC medium (in which HS and NBCS were substituted by FCS). Monocytes stayed viable after this loading, as determined by trypan blue exclusion. Coincubation Stx2-loaded monocytes and target cells These experiments were performed in two different experimental settings: with transfer of loaded monocytes from 4° to 37°, and without change of temperature. To start with the first setting, the Stx2-loaded monocytes were added to a monolayer of target cells [HUVEC (n = 10) or vero- cells (n = 9)] in a concentration of 1 × 106 per well. This was performed during 24 h at 37°C. For comparison, also monocytes without Stx were used. To determine the specificity of the effect of Stx2, the loaded monocytes were preincubated with a well-characterized antibody against Stx2 (TMA-15, 1 μg/ml, approximately 150× excess). All experiments were performed in duplicate. To measure the transfer of Stx2 from the monocyte, the protein synthesis of the target cells was determined by adding 3H-Leucine (0.67 μCi/ml). Subsequently, intracellular proteins were precipitated by treatment with trichloroacetic acid (TCA), and the radioactivity was measured in a liquid scintillation counter. In the other experimental setting, monocytes were loaded in a similar way with FITC-labeled Stx1B subunit; 1 × 106-loaded monocytes were coincubated with 1.25 × 105 vero cells in suspension for 3 h at 4°C (n = 5) while being continuously rotated. At this temperature, bias due to internalization of the toxin could be avoided [19, 25]. After 3 h, the presence of Stx1B subunit FITC on the vero cells was determined by flow cytometry. For every experiment, at least 1,000 cells were measured. Study of supernatant Stx-loaded monocytes To study the effect of the supernatant of Stx2-loaded monocytes, it was collected after 16 h of coincubation on vero cells (n = 11). The supernatants were centrifuged to remove possible monocytes and added again to fresh vero cells. Protein synthesis was measured by adding 3H-leucine and measuring the incorporation after 24 h of incubation at 37°C. Also, the antibody against Stx2 was used to determine a possible effect of free Stx2. Furthermore, monocytes were loaded with 125I-Stx1B subunit to investigate whether it could be released after a change in temperature from 4° to 37°C (n = 3). For each experiment, between 3 and 4 × 106 monocytes are loaded with 200 nM Stx1B subunit (3,600 cpm/ng protein). After loading, unbound Stx1B subunit was removed by centrifugation. Subsequently, the monocytes were placed at 37°C for 2 h. They were then centrifuged, and cells and supernatant were measured separately for the presence of 125I-Stx1B with a gamma counter. This was compared with the amount of binding to the monocytes before the temperature change. The experiment was performed in duplicate. Statistics All data presented are expressed as a range with the median. Significance of increase or decrease of protein synthesis compared with controls was analyzed using the Wilcoxon signed ranks test. The statistical level of significance was defined as P < 0.05. Results Coincubation Stx2-loaded monocytes and target cells To investigate whether monocytes can transfer Stx2 to target cells (vero cells and HUVECs), Stx2-loaded monocytes were added to a monolayer of these cells. After coincubation, protein synthesis of vero cells and HUVEC was measured. If the Stx2 was transferred to the target cells, there was protein synthesis inhibition, as this is the biological effect of Stx2 in both types of target cells. After 24 h of coincubation at 37°C, in both cell types, there was protein synthesis inhibition (Fig. 1). In vero cells, greater protein synthesis reduction was measured than in HUVEC. Directly adding 10 nM Stx2 to vero cells reduced protein synthesis to 5.5% (data not shown). When an antibody against Stx2 was added to the monocytes before coincubation with vero cells, protein synthesis inhibition was partly restored (Fig. 2). Only in one out of five experiments was there complete recovery. Fig. 1Coincubation of Shiga-like toxin (Stx)2-loaded monocytes with target cells [vero cells and human umbilical cord venous endothelial cells (HUVEC)]. To investigate whether monocytes could transfer Stx2, toxin-loaded monocytes were coincubated with vero cells (n = 9) and HUVEC (n = 10). This led to protein synthesis inhibition, as could be determined with the incorporation of 3H-leucine. The coincubation with unloaded monocytes was used as a control and set at 100%. P < 0.01Fig. 2Effect of the addition of an antibody against Shiga-like toxin (Stx)2. The monocytes of five healthy donors were coincubated with vero cells (Stx2-loaded monocytes, white bar; Stx2-loaded monocytes with Stx2 antibody, black bar). Addition of the antibody against Stx2 led to partial protein synthesis recovery in four donors and complete recovery in one donor. Coincubation with unloaded monocytes was used as a control and set at 100% In equal experiments performed with HUVEC (n = 3), TMA-15 also partly prevented protein synthesis inhibition (data not shown). Study of supernatant Stx-loaded monocytes Because the addition of an antibody against Stx2 was not sufficient for complete protein synthesis recovery in four out of five tested donors, we investigated the possible presence of an additional inhibitor in the supernatant after 16 h of coincubation. This was performed by re-adding the supernatant of coincubated Stx2-loaded monocytes and a monolayer of vero cells to new vero cells. Figure 3 shows that this incubation again led to an inhibitory effect on vero-cell protein synthesis in contrast to the supernatant of unloaded monocytes. But this reduction was less than with direct coincubation. The effect could be partly blocked by adding the antibody against Stx2. This means that there is unbound Stx2 present in the supernatant. Next, we investigated whether the temperature change of the Stx-loaded monocytes from 4° to 37°C could lead to toxin release. For this reason, we loaded the monocytes with 125I-Stx1B subunit. The amount of radioactivity on or inside the monocytes was measured before and after incubation of the cells for 2 h at 37°C (Table 1). Vero cells were used as a control. In vero cells as in monocytes, the toxin was released from its receptor after the incubation. Fig. 3Addition of supernatants coincubation to vero cells. After coincubation of Shiga-like toxin (Stx)-loaded and unloaded monocytes, supernatants were collected and re-added to fresh vero cells (n = 11). The supernatant from the toxin-loaded monocytes induced, again, protein synthesis inhibition. This inhibition could be partially restored with a Stx2 antibody (AB). The incubation of supernatant with unloaded monocytes was used as a control and set at 100%. P < 0.01Table 1Amount of 125I- Shiga-like toxin (Stx)1B subunit on cells after change of temperatureStx-loaded cellsBefore 4→37°C (cpm)After 4→37°C (cpm)Vero cells1,034,383671,060Monocytes donor 11,437427Monocytes donor 21,070374Monocytes donor 31,121677cpm counts per minute Transfer of Stx to target cells without change in temperature A change in temperature from 4° to 37°C released Stx from its receptor on the monocyte. To study again the possibility of Stx transfer from monocyte to target cells, we loaded isolated monocytes with Stx1B subunit labeled with FITC. Coincubation of these monocytes with vero cells in suspension without a change in temperature did not result in a transfer of the B subunit to the vero cells (n = 5, Fig. 4). Fig. 4Transfer of Shiga-like toxin (Stx)-loaded monocytes to vero cells without temperature shift (white: normal monocytes; black: Stx-loaded monocytes; grey: vero cells). The x-axis represents the intensity of the fluorescein isothiocyanate (FITC) signal, the y-axis represents the number of cells. a FITC-labeled Stx1B subunit was bound to monocytes from a healthy donor, showing an increase in intensity. b Coincubation of vero cells and unloaded monocytes; note the high basal signal of vero cells. c After 2 h of coincubation with vero cells in suspension at 4°C, there was no transfer of the FITC signal. The cells remained in position Discussion After coincubation of Stx2-loaded monocytes with target-cells (vero cells and HUVEC), protein synthesis inhibition could be detected. As the biological effect of Stx is protein synthesis inhibition, a transfer was expected. This effect could not be due to the presence of monocytes, because it is well described that Stx has no inhibitory effect on peripheral blood monocytes [17]. However, the addition of an antibody against Stx only partly restored the inhibition. We hypothesized that in parallel to transfer of Stx2, a possible additional inhibiting factor was present in the supernatant during the experiment. To further investigate this possibility, the supernatant of this coincubation was re-added to fresh vero cells. Again, there was protein synthesis inhibition, which could partly be decreased by blocking with the Stx2 antibody. The conclusion was made that there another inhibitory factor needed to be present (possibly released by activated monocytes), but unbound Stx2 was also present [26]. Apparently, Stx1B subunit is released from its receptor when there is shift of temperature from 4° to 37°C, as shown in this study and also by Ramegowda and Tesh [18]. This finding stresses the caution that must be taken when drawing conclusions from in vitro experiments performed at 4°C. In Fig. 5, all findings are schematically summarized. Fig. 5Schematic summary of performed experiments. a Monocytes loaded with Shiga-like toxins (Stx) at 4°C. b After shifting Stx-loaded monoytes to target cells at 37°C, Stx was released from its receptor. This led to protein synthesis inhibition of target cells. However, transport could not be excluded (but this was performed in experiments without a shift in temperature). Probably, some toxin was internalized and monocytes became activated. This could lead to cytokine and proteases production (possible cytotoxic factors). c When the supernatant is re-added to new target cells, there are released Stx and secreted products present. This induced, again, protein synthesis inhibition As these in vitro experiments were not suitable to investigate a possible transfer, experiments were performed without a change of temperature. No transfer of the binding part of the toxin could be detected in this setting. Experiments could not be performed at 37°C because the toxin is internalized by the monocytes during 3 h of incubation. Because of these in vitro experiments, we believe that the monocyte cannot function as a transporter for Stx in the circulation. However, monocytes still seem to play an important role in pathogenesis. As a component of the innate immune system, they play a central role in immunity and inflammation. Fernandez et al. showed that patients in the acute period of HUS have monocytes with phenotypic (reduced expression of CD14, CD64, and CD11b) and functional [decreases lipopolysaccharide (LPS)-induced TNF-α production and Fcγ-dependent cytotoxicity] differences compared with healthy children and acute uremic children [27]. Some unanswered questions remain. What is the inhibitory factor released by monocytes after loading with Stx? Why is there still unbound Stx2 present in the supernatant after 16 h of coincubation? It is surprising that Stx2 is not completely bound and internalized by the numerous receptors on monocytes and vero cells. Also, how Stx is transported in the circulation remains unsolved. Kimura et al. described that serum amyloid P (SAP) can bind Stx2 and function as a neutralizing factor [28]. However, in humans, there was no correlation between circulating SAP and the development of HUS [29]. As renal endothelial damage is already present after the occurrence of clinical symptoms of HUS, it is very important to develop efficacious early prevention. Understanding the mechanism in which the toxin is specifically targeted to the kidney can lead to novel intervention strategies.	[ "shiga-like toxin", "monocyte", "hemolytic uremic syndrome", "vero cells", "acute renal failure", "huvec" ]	[ "P", "P", "P", "P", "P", "P" ]
Eur_Arch_Otorhinolaryngol-4-1-2254469	Fiberoptic endoscopic evaluation of swallowing in intensive care unit patients	Aspiration in critically ill patients frequently causes severe co-morbidity. We evaluated a diagnostic protocol using routine FEES in critically ill patients at risk to develop aspiration following extubation. We instructed intensive care unit physicians on specific risk factors for and clinical signs of aspiration following extubation in critically ill patients and offered bedside FEES for such patients. Over a 45-month period, we were called to perform 913 endoscopic examinations in 553 patients. Silent aspiration or aspiration with acute symptoms (cough or gag reflex as the bolus passed into the trachea) was detected in 69.3% of all patients. Prolonged non-oral feeding via a naso-gastric tube was initiated in 49.7% of all patients. In 13.2% of patients, a percutaneous endoscopic gastrostomy was initiated as a result of FEES findings, and in 6.3% an additional tracheotomy to prevent aspiration had to be initiated. In 59 out of 258 patients (22.9%), tracheotomies were closed, and 30.7% of all 553 patients could be managed with the immediate onset of an oral diet and compensatory treatment procedures. Additional radiological examinations were not required. FEES in critically ill patients allows for a rapid evaluation of deglutition and for the immediate initiation of symptom-related rehabilitation or for an early resumption of oral feeding. Introduction Pulmonary aspiration is a serious cause of morbidity and mortality in patients with a depressed sensorium, patients with neuromuscular discoordination, or patients having structural disorders of the upper aerodigestive tract. It is a leading cause of nosocomial infection in the intensive care unit (ICU). The most common manifestations of pulmonary aspiration are pneumonia, pleuropulmonary infection, and acute airway obstruction [1]. Hospital-acquired pneumonia is the second most common nosocomial infection in the critically ill patient and is associated with the greatest mortality and increased morbidity and cost of care. Incidence of HAP varies in different populations of critically ill patients and generally ranges from 9 to 20%. The etiologic agents differ according to the population studied, duration of hospital stay, time after intubation, and prior antimicrobial therapy [2]. Risk factors include nonmodifiable factors like age, chronic obstructive pulmonary disease, severe head trauma, and multiple trauma, and modifiable factors like large-volume gastric aspiration, duration of mechanical ventilation, elevated gastric pH, histamine type two blocker therapy, ventilator circuit change frequency, self-extubation, and silent aspiration following scheduled extubation [2]. When patients are admitted to the ICU, they are at increased risk of more frequent aspiration events because of a variety of factors, including depressed level of consciousness (often caused by excess analgesia and sedation), forced supine position, and the presence of nasogastric or endotracheal tubes [3]. There is a growing body of evidence suggesting that intubation for longer than 48 h may cause at least transient injury to the larynx with a subsequent reduction in the protective mechanism and increased incidence of oropharyngeal secretions once the patient is extubated. The presence of an orotracheal tube has been shown to alter the mechanoreceptors and chemoreceptors of the pharyngeal and laryngeal mucosa, causing dysfunction of the swallowing reflex. The prevalence of swallowing dysfunction post-extubation has been reported to occur in between 20 and 83% of those patients intubated longer than 48 h. This wide range of estimate has partly been attributed to the variation in the diagnostic tools [4]. The impact that diagnosis using invasive diagnostic techniques may have on the epidemiological characteristics of HAP in critically ill patients is unknown, but may potentially improve aspiration-related specific therapy and ultimately clinical outcome. Based on these considerations, we prospectively evaluated a diagnostic protocol using routine fiberoptic endoscopic evaluation of swallowing (FEES) in critically ill patients at risk to develop aspiration following temporary transnasal intubation or tracheotomy and tried to define the impact of post-extubation FEES on the initiation of aspiration-related treatment. Material and methods We conducted a prospective, interventional, clinical study at Klagenfurt General Hospital. The hospital is a 1,400-bed tertiary referral centre that serves a population of approximately 1,000,000 and runs seven ICUs (paediatrics and neonatology, 1; neurology, 1; neurosurgery, 1; cardiology, 1; general internal medicine, 1; and anaesthesiology, 2) with a total of 84 beds. We offered a standardized endoscopy protocol for critically ill patients considered to be at risk for silent aspiration following transnasal intubation or tracheotomy for ventilation. Patient selection for endoscopic assessment of swallowing was done by ICU physicians using a protocol that we designed to help in the early recognition of risk factors for, or overt signs of, aspiration. This protocol included a taxonomic summary of symptoms and clinical signs potentially related to aspiration. Rating was applied after the stop of mechanical ventilation. Symptoms were defined as:Weight loss,unexplained fever >38.0° centigrade,coughing,bronchitis/pneumonitis,impaired voice,witnessed regurgitation/aspiration event at the bedside accompanied by coughing, choking, and/or expectoration of material,prolonged oral feeding,aversion for oral intake of liquids and solids,disturbance of bolus transport,frequent postural changes during oral intake, andregurgitation.If such conditions were observed in previously intubated patients following extubation, or in patients with tracheotomies, physicians at ICUs were encouraged to request endoscopic assessment of swallowing. The FEES procedure was consequently provided by staff members of the department of Oto-Rhino-Laryngology, Head and Neck Surgery. The team consisted of a laryngologist and a nurse who had both previously gone through a specialized training in performing FEES procedures in patients with deglutition disorders. Over the 45-month period from 1 January 2002 to 30 September 2005, we were called to perform 913 endoscopic examinations in 553 patients treated in one of the ICUs. Four hundred and 46 patients underwent one single endoscopy, and 107 had repeated examinations. Two hundred and 95 patients had shortly before been extubated after transnasal tracheal intubation, and 258 patients had indwelling tracheotomies. For further data interpretation, patients were sub-grouped according to their route of ventilation (transoral/transnasal intubation vs. tracheotomy). Findings and recommendations from the initial FEES procedure in all 553 patients were accumulated for this study. The results of 360 control procedures were not further analyzed. Details of the FEES procedure have previously been described in detail. Therefore, only a short summary of the procedure shall be given here: The FEES procedure is a portable examination, easily taken to bedside in ICUs. Since our patients were usually bedridden, we performed the examination in a in bed with the head of bed elevated to approximately 70° with the bend of the bed is at the patient’s lower back (904/913 patients, i.e., 99%). In four patients (0.4%), the examination was carried out with the patient in reverse Trendelenberg at 30°–45°, and five patients (0.6%) were able to sit on a chair during the procedure. A fiberoptic laryngoscope was passed transnasally to the oropharynx, where the larynx and surrounding structures could then be visualized. Patients were led through various tasks to evaluate the sensory and motor status of the pharyngeal and laryngeal mechanism. Stained liquid and semi-liquid boluses were then given to determine the integrity of pharyngeal deglutition. The interior larynx and airway were examined for evidence of food penetration within the laryngeal vestibule and aspiration of food below the true vocal folds before and after each swallow. In each case, the nasogastric tube was removed prior to the procedure. We assessed structural changes of the larynx and pharynx, timing and direction of movement of the bolus through the pharynx, the ability to protect the airway and to uphold airway protection for a some seconds, the capability to clear the bolus during deglutition, presence of pooling and residue of material in the hypopharynx, and timing of bolus flow and laryngeal closure. Aspiration was defined as the entry of material into the airway below the levels of the true vocal cords. Silent aspiration was defined as aspiration occurring in the absence of acute symptoms (i.e., lack of cough or gag reflex as the food or liquid bolus passed into the trachea). Findings were documented in a standardized form to allow for monitoring therapeutic interventions over time, and for later data analysis and evaluation. Results Over a 45-month period, we were called to perform 913 endoscopic examinations in 553 patients. Of these, 295 patients had previously been extubated, and 258 patients had indwelling tracheotomies. Silent aspiration or aspiration with acute symptoms (cough or gag reflex as the bolus passed into the trachea) was detected in 383 (69.3%) of all patients during initial FEES. Prolonged non-oral feeding via a naso-gastric tube was initiated in 49.7% of all patients. In 13.2% of patients, a percutaneous endoscopic gastrostomy was initiated as a result of FEES findings, and in 6.3% an additional tracheotomy to prevent aspiration had to be set off. In 59 out of 258 patients (22.9%), tracheotomies were closed, and 30.7% of 553 patients could be managed with the immediate onset of an oral diet ± compensatory treatment procedures without further diagnostic evaluation. Logopedic instruction of the patient and/or rehabilitation using compensatory strategies to reduce the risk of aspiration were an integral part of our rehabilitation program and included one or more of the following:Modifications in volume and tempo of food presentationoral feeding with consistency modificationshead rotationholding the chin down during deglutition to narrow the airway entrancesupraglottic swallow: This technique uses simultaneous swallowing and breath-holding, closing the vocal cords and protecting the airway.One hundred and seven patients were scheduled for repeated FEES, usually to assess the progress of the rehabilitation program. Signs and symptoms encountered during initial FEES are condensed in Table 1. Table 2 stratifies the route of feeding in the two subgroups of our study population. Table 3 classifies the type of aspiration detected during FEES. Table 4 summarizes the recommendations derived from our endoscopic findings for the further management of patients with aspiration. Table 1Symptoms of aspiration in 553 patientsPatients after transnasal tracheal intubation (n = 295)Patients with indwelling tracheotomies (n = 258)Unexplained fever17 (5.8%)39 (15.1%)Coughing68 (23.1%)31 (12.0%)Bronchitis/pneumonitis57 (19.32%)39 (15.1%)Impaired voice57 (19.32%)0 (0%)Witnessed regurgitation/aspiration event 74 (25.1%)149 (57.8%)Others16 (5.4%)23 (8.9%)No data69 (23.3%)19 (7.4%)Since more than one symptom or sign could apply per patient, results sum up to more than 100%Table 2Route of feeding for 553 patients on initial FEESPatients after transnasal tracheal intubation (n = 295)Patients with indwelling tracheotomies (n = 258)Nasogastric tube224 (75.9%)186 (72.1%)PEG15 (5.1%)59 (22.9%)Parenteral12 (4.1%)5 (1.9%)Oral diet44 (14.9%)8 (3.1%)PEG percutaneous endoscopic gastrostomyTable 3Classification of aspiration for 553 patients on initial FEESPatients after transoral or transnasal tracheal intubation (n = 295)Patients with indwelling tracheotomies (n = 258)Silent aspiration51 (17.3%)95 (36.8%)Aspiration115 (39.0%)126 (48.8%)No aspiration129 (43.7%)37 (14.4%)Aspiration = inhalation of material into the airway below the level of the true vocal cords, with acute symptoms (cough or gag reflex as the bolus passed into the trachea), silent aspiration = aspiration occurring in the absence of acute symptomsTable 4Recommendation for further treatment in 553 patients after initial FEESPatients after transoral or transnasal tracheal intubation (n = 295)Patients with indwelling tracheotomies (n = 258)Non-oral feeding (naso-gastric tube) ± logopedic (functional) therapy144 (48.8%)131 (50.8%)Indication for PEG27 (9.1%)46 (17.8%)Indication for non-oral feeding plus tracheotomy35 (11.9%)NAOral feeding ± logopedic (functional) therapy89 (30.2)22 (8.5%)Decannulation and oral feeding ± logopedic (functional) therapyNA59 (22.9%)PEG percutaneous endoscopic gastrostomy, NA not applicable Discussion Aspiration is a common event. The North American Summit on Aspiration in the Critically Ill Patient, in a consensus statement published in 2002 [5], estimates the frequency of aspiration in various patient populations is as follows:45% in normal individuals during sleep70% in patients with impaired consciousness0–40% in patients on ETF50–75% in patients with endotracheal tubesWhat determines the clinical consequences of an aspiration event are host factors (age, immune status, underlying disease process, and comorbidities) and the nature of the aspirate. The morbidity and mortality associated with aspiration (with or without subsequent pneumonia) is difficult to ascertain and may vary greatly depending on the patient population that is under study [5]. Despite these general considerations, aspiration is generally accepted as a leading cause of nosocomial pulmonary infection in the ICU [6]. Ajemian reported a 56% (27/48) incidence of swallowing dysfunction following prolonged intubation; 12 (25%) of 48 patients were silent aspirators [7]. In a study by Leder et al. [8], FEES demonstrated swallowing dysfunction in 33% of critically ill trauma patients after prolonged intubation, with 20% having silent aspiration. In a recent study by El Solh and co-workers, aspiration was documented in 52% of ICU patients aged over 65 years and in 36% of those up to age 65. No significant difference in the co-morbidity index and the length of mechanical ventilation was found between aspirators and non-aspirators [4]. Non-specific methods to monitor for and detect aspiration are usually subjective, not standardized or validated, and inaccurate, and there are few outcome valid data. Blue dye has been commonly used to color the feeding solutions, but this method is insensitive and non-specific for assessing aspiration and pneumonia. Testing of pharyngeal secretions with glucose oxidase strips also is not standardized and is too insensitive and non-specific to test for aspiration of feeding solution. The best single test for aspiration remains undetermined to date. Therefore, clinical studies are needed to determine the potential role of particular tests suited for specific indications [9]. Video-fluoroscopy has traditionally been accepted as the “gold standard” for evaluation of a swallowing disorder for the comprehensive information it provides. However, it is not very efficient and accessible in certain clinical and practical situations. This is particularly true for critically ill patients treated at ICUs, who are almost invariably bedridden. FEES has been shown to be safe and effective for assisting in swallowing evaluation, and in therapy as a visual display to help patients learn various swallowing manoeuvres. [4, 10–13]. It can be used in a wide variety of settings, ranging from office evaluation of swallowing, inpatient and outpatient services to critically ill patient treated at the ICU [14–20]. FEES is not just a screening procedure, it can achieve a complete assessment of the pharyngeal stage of swallowing. It includes five components: assessment of structural changes in the larynx and pharynx, assessment of movement and sensation, assessment of secretions management, direct visualization of swallowing function for food and liquid, and response to therapeutic interventions [12–15, 17]. Aviv and co-workers suggested combining FEES with sensory testing of the laryngeal and pharyngeal mucosa using discrete pulses of air delivered to the epithelium innervated by the internal branch of the superior laryngeal nerve (FEESST: Fiberoptic endoscopic evaluation of swallowing with sensory testing). They performed 1,340 consecutive FEESST procedures on 1,076 adult patients with dysphagia, among them 172 examinations in critically ill patients at the ICU (mostly following cardiac surgery). The authors concluded that FEESST for the purpose of swallowing function can be done safely and yields a 96.5% success rate in terms of completed examinations [19, 21]. Our experience suggests that FEES can be performed easily at the bedside of ICU patients. It is associated with minimal complications for those patients recently liberated from mechanical ventilation, and for patients with indwelling tracheotomy treated at the ICU. However, we cannot determine from our data whether adopting FEES prevented aspiration pneumonia. Barquist et al. [22] conducted a randomized, prospective clinical trial of FEES versus routine clinical management in 70 patients after prolonged intubation. The addition of a FEES examination did not change the incidence of aspiration or postextubation pneumonia. Larger prospective randomized trials of FEES versus conventional patient management will be required to determine whether a difference in the rate of post-extubation pneumonia can be demonstrated. However, in our study, the observations recorded during FEES had immediate impact on the further treatment of patients: prolonged non-oral feeding via a naso-gastric tube was initiated in 49.7% of patients, percutaneous endoscopic gastrostomy in 13.2%, and an additional tracheotomy to prevent aspiration in 6.3%. On the other hand, 81 out of 258 pre-existing tracheotomies were immediately closed, and 30.7% of all 553 patients with endoscopy-proven aspiration could be treated with the immediate onset of an oral diet and compensatory treatment procedures. After assessing the critically ill patient for risk of aspiration, the clinician still must decide if the patient is ready to be fed. The goal is to identify critically ill patients who are likely to tolerate enteral nutrition and attempt to minimize complications [23]. FEES is certainly a versatile tool in achieving timely and appropriate clinical decisions in patients at risk for aspiration-related morbidity. In contrast to most of the recently published material on FEES in critically ill patients, we report our observation in a large proportion of patients with indwelling tracheotomies (258 out of 553, i.e., 46.7%) at the time of the examination. Tracheotomy is a commonly performed procedure in ventilator dependent patients. Many critical care practitioners believe that performing a tracheostomy early in the postinjury period decreases the length of ventilator dependence as well as having other benefits such as better patient tolerance and lower respiratory dead space [24]. Table 4 demonstrates that silent aspiration as well as clinically overt aspiration was more frequently encountered in patients with tracheotomies as compared to those that had previously been intubated via an oral or transnasal route. However, the clinical impressions that a tracheotomy or tracheotomy tube increases aspiration risk or that decannulation results in improved swallowing function have not been supported by a recent study. Rather, the need for a tracheotomy indicated comorbidities (e.g. respiratory failure, trauma, stroke, advanced age, reduced functional reserve, and medications used to treat the critically ill) that by themselves predispose patients for dysphagia and aspiration [25]. These data confirmed similar findings previously published by same authors: Neither the presence of a tracheotomy tube nor decannulation affected aspiration status in early, postsurgical head and neck cancer patients. In the acute care setting, no causal relationship between tracheotomy and aspiration status was exhibited [26]. Our results allow no conclusion regarding the impact of tracheotomies on deglutition, since the initial decision to perform a tracheotomy had not been randomly assigned to patients in our two subgroups. The decision to perform tracheotomies was not consistent over different ICUs and largely depended on the anticipated length of assisted ventilation. In our study, FEES was used to determine the need for maintaining tracheotomy, and was accepted by ICU physicians as an important criterion in deciding to close tracheotomies following assisted ventilation. Conclusion After the introduction of a standardized endoscopy protocol for critically ill patients considered being at risk for silent aspiration, ICU physicians soon requested FEES routinely for their patients. FEES in critically ill patients allows for a rapid evaluation of deglutition, for targeted further diagnostic procedures if needed, and for the immediate initiation of symptom-related rehabilitation or for an early resumption of oral feeding. FEES is now accepted at our institution as an important tool in achieving timely and appropriate clinical decisions for ICU-patients at risk for aspiration-related morbidity. Laryngologists should be encouraged to offer FEES procedures to responsible coordinators of ICUs.	[ "swallowing", "aspiration", "tracheotomy", "deglutition", "rehabilitation", "endoscopy", "dysphagia", "laryngology" ]	[ "P", "P", "P", "P", "P", "P", "P", "U" ]
Eur_J_Appl_Physiol-3-1-1914240	Conventionally assessed voluntary activation does not represent relative voluntary torque production	The ability to voluntarily activate a muscle is commonly assessed by some variant of the twitch interpolation technique (ITT), which assumes that the stimulated force increment decreases linearly as voluntary force increases. In the present study, subjects (n = 7) with exceptional ability for maximal voluntary activation (VA) of the knee extensors were used to study the relationship between superimposed and voluntary torque. This includes very high contraction intensities (90–100%VA), which are difficult to consistently obtain in regular healthy subjects (VA of ∼90%). Subjects were tested at 30, 60, and 90° knee angles on two experimental days. At each angle, isometric knee extensions were performed with supramaximal superimposed nerve stimulation (triplet: three pulses at 300 Hz). Surface EMG signals were obtained from rectus femoris, vastus lateralis, and medialis muscles. Maximal VA was similar and very high across knee angles: 97 ± 2.3% (mean ± SD). At high contraction intensities, the increase in voluntary torque was far greater than would be expected based on the decrement of superimposed torque. When voluntary torque increased from 79.6 ± 6.1 to 100%MVC, superimposed torque decreased from 8.5 ± 2.6 to 2.8 ± 2.3% of resting triplet. Therefore, an increase in VA of 5.7% (from 91.5 ± 2.6 to 97 ± 2.3%) coincided with a much larger increase in voluntary torque (20.4 ± 6.1%MVC) and EMG (33.9 ± 6.6%max). Moreover, a conventionally assessed VA of 91.5 ± 2.6% represented a voluntary torque of only 79.6 ± 6.1%MVC. In conclusion, when maximal VA is calculated to be ∼90% (as in regular healthy subjects), this probably represents a considerable overestimation of the subjects’ ability to maximally drive their quadriceps muscles. Introduction When a supramaximal electrical stimulus is applied to the nerve of a muscle during a voluntary isometric contraction, the contractile response of any motor units not yet (fully) recruited will lead to a force increment. Merton (1954) first reported a linear decrease of this superimposed force increment as voluntary contraction intensity increased in the adductor pollicis muscle. Hence, as voluntary activation (VA) increases, the force increment as a result of electrical stimulation decreases. For the most accurate assessment of maximal VA, Behm et al. (1996) recommended using the maximal voluntary contraction (MVC) with a superimposed electrical stimulus, which is now common practice (Allen et al. 1995; Babault et al. 2003; Becker and Awiszus 2001; Behm et al. 2001; de Ruiter et al. 2004; Maffiuletti and Lepers 2003; Newman et al. 2003; Oskouei et al. 2003; Suter and Herzog 1997). Maximal VA is generally quantified by expressing the ‘twitch’ force increment on an MVC as a percentage of resting twitch force. Subtraction from 100% results in a value for maximal VA. Generally, high levels of VA (∼90%) for the knee extensors are achieved by regular healthy subjects (e.g. Babault et al. 2003; Behm et al. 2001; Bulow et al. 1993; de Ruiter et al. 2004; Newman et al. 2003; Oskouei et al. 2003; Suter and Herzog 1997). An important implicit meaning of a VA level of 90% is that 90% of the maximal torque capacity (MTC) is reached, and that at most, a 10% further increase in torque is possible. In our experience, however, a greater increase in torque can be observed than the obtained VA level suggested when subjects succeed in improving their MVC in a later attempt during the same session. This suggests that VA can be overestimated using the conventional method of superimposed stimulation. This would have important consequences for conclusions drawn regarding calculated VA. The inability of patient groups to access their muscles’ potential could be strongly underestimated. This has implications for the estimations made of the effects of paralytic poliomyelitis on strength and endurance for instance (Grabljevec et al. 2005). Furthermore, many studies have been performed using target force levels expressed as %MVC (de Ruiter et al. 2006; Grabljevec et al. 2005; Hisaeda et al. 2001; Kalmar and Cafarelli 1999; Ng et al. 1994; Place et al. 2005). For example, the %MVC at which full occlusion of blood flow of the knee extensors occurs during isometric contraction (de Ruiter et al. 2006). Clearly, knowledge of the potential of a far greater torque increase than, for example, the 10% implied by a VA of 90% is very important. In our experience, in subjects capable of consistently high levels of maximal VA (∼97%), the superimposed torque increment is already very small at high contraction intensities (∼70%MVC). Further increases in voluntary torque show only minor reductions of the superimposed torque increment. This is in accordance with the modelling study of Herbert and Gandevia (1999); for the adductor pollicis muscle they show that large increases of motoneuronal excitation near MVC will be accompanied by only minor reductions in the size of the superimposed force. At high contraction intensities, therefore, the %VA does not necessarily represent the same percentage of the muscles’ maximal torque capacity. The first aim of the present study was therefore to assess the relationship between calculated VA and voluntary torque of the knee extensors at high contraction intensities. To achieve the best possible assessment of the relationship between conventionally assessed VA and voluntary torque (as %MVC), in the present study, only those individuals capable of consistently high levels of maximal VA were included. An additional concern of using the ITT is the dependence of the calculated VA on the size of the resting twitch. Following a single electrical stimulus, tendon slack must first be taken up before resting peak twitch torque is reached. Therefore, during an isometric contraction, where most of the tendon slack has already been taken up, the twitch torque increment will be relatively large compared to the twitch torque at rest (Belanger and McComas 1981). By scaling the superimposed twitch to resting twitch torque, calculated VA will thus be lower than it otherwise would have been. The effect of tendon slack on the resting twitch has been reported to be of minor influence on the relationship between the superimposed response and voluntary torque, at flexed knee angles (Behm et al. 2001; Bulow et al. 1993; Rutherford et al. 1986). However, tendon slack is greater at short compared to long muscle length (Muraoka et al. 2004). Therefore, particularly at short muscle lengths, scaling the superimposed torque response to the response obtained at rest may lead to a relatively lower calculated VA at short (extended knee angles) compared to long muscle lengths (flexed knee angles). In subjects with a high ability for maximal VA, a neural drive may be reached that is usually unattainable for regular healthy subjects. Moreover, during voluntary fast isometric torque development, rectified surface EMG levels of the knee extensors may be substantially higher than at the torque plateau of an MVC (de Ruiter et al. 2004). These findings indicate a potential to increase neural drive at the torque plateau of a MVC (see also Herbert and Gandevia 1999). However, based on the shape of the stimulation frequency–force (de Ruiter et al. 1999; Deutekom et al. 2000) and pCa–force curve (Stephenson and Williams 1982) at high force levels large increases in neural drive and calcium are required for a comparatively small increase in force. As EMG is often measured to quantify the neural drive at different contraction intensities, this has important consequences for the EMG–torque relationship. EMG is generally normalized to the EMG obtained during MVC, a disproportionate increase in EMG at very high force levels will consequently lead to a more curvilinear EMG–torque relationship. However, this can only be demonstrated in subjects who are able to consistently reach high levels of neural drive. The shape of a subjects’ EMG–torque relationship would therefore be determined by the subjects’ ability for VA. This would explain why both linear (Woods and Bigland-Ritchie 1983) and non-linear (Alkner et al. 2000) EMG–torque relationships have been reported. Therefore, this study further aims to investigate the EMG–torque relationship of the knee extensors in subjects with high ability for VA. Methods Subjects Seven healthy male subjects (23–32 year, 69–83 kg, and 1.73–1.93 m) volunteered to be subjects for this investigation. Before participation, each subject was thoroughly informed about the procedures and provided written informed consent prior to testing. Only subjects capable of >95%VA across knee angles (see below) were included in the present study. The study was performed according to the Declaration of Helsinki and approved by the local ethics committee. Subjects did not perform any fatiguing exercise 48 h prior to measurements. Experimental procedures The subjects visited our laboratory on two occasions with at least 1 day in between. All subjects were familiar with the set-up, electrical stimulation, and isometric knee extension from previous studies (de Ruiter et al. 2004; Kooistra et al. 2005). During the first experimental day, several contractions with superimposed stimulation (see below) were performed at 30° (extended) and 60 and 90° (flexed) knee angles. At the same time, surface EMG of the rectus femoris (RF), vastus lateralis (VL), and vastus medialis (VM) was recorded. To gain information concerning test-retest variability, the experimental procedure performed on experimental day 1 was repeated during experimental day 2. Torque measurements Isometric knee extension torque of the right leg was measured using a custom-built dynamometer (Kooistra et al. 2005). Subjects were seated with their hips at 70° (0° = full extension). Shoulders, hips, and lower thigh were strapped to the dynamometer. By very tight strapping of the hips and lower thigh, a contribution of the hip extensors to knee extension was absolutely impossible. Moreover, during contractions, subjects had to lean forward about 10 cm with their shoulders (while grasping the strap around their shoulders) pushing their lower back into the back rest. This further guaranteed that no hip extension occurred. The distal part of the shank was strapped to a force transducer (KAP, E/200 Hz, Bienfait B.V. Haarlem, The Netherlands) that was attached to the lever arm of the dynamometer. A shin guard ensured subjects could exert maximal forces without discomfort at the shin. The compliance of the dynamometer at the position of the force transducer was 1.4 × 10−4 deg/Nm. The backrest, force transducer height, and its medio-lateral position were adjusted for precise alignment of the knee axis with the axis of rotation of the dynamometer arm. A crank enabled changing of the knee angle of the subject. Knee joint angles were determined with a handheld goniometer (model G300, Whitehall Manufacturing) using the greater trochanter and lateral epicondyle of the femur, and the lateral malleolus of the fibula as references. The dynamometer arm was positioned so that the indicated knee angles were angles in an active state, with subjects delivering ∼50% of MVC. During pilot experiments using an electrogoniometer attached to the lateral side of the knee, the changes (active–passive) in knee angle were found to be 3–7°, independent of knee angle, with minimal (<1°) changes above ∼50% of MVC. Real-time force applied to the force transducer was displayed on line on a computer monitor and digitally stored (1 kHz) on computer disc. The force signals were automatically corrected for gravity at each angle: the average force applied by the weight of the limb to the transducer during the first 50 ms after the start of a recording, with the subject seated in a relaxed manner, was set to zero force by the computer program. Extension torque was calculated by multiplication of force with the individual lever arm and data will be presented as torque. Electrical stimulation A cathode (self-adhesive stimulation electrode, 5 × 5 cm, Schwa-Medico, The Netherlands) was placed over the femoral nerve. The anode (13 × 8 cm) was placed over the gluteal fold. The quadriceps femoris muscle was stimulated transcutaneously with rectangular pulses of 100 μs using a computer-controlled constant current stimulator (Digitimer DS7H, Digitimer Ltd., Welwyn Garden City, UK). Stimulation current was increased until torque measured in response to a triplet (three 100 μs pulses applied at 300 Hz) levelled off at each knee angle tested. The current (in mA) was then increased by a further 50 mA to ensure supramaximal stimulation (range: 250–600 mA). At this point, we considered all muscle fibres of the knee extensors to be activated. Pilot studies had shown that triplet torque did not increase after an MVC at any knee angle, indicating that any potential potential post-activation potentiation did not result in a triplet torque increase (see also Discussion). Usually, torque generated by the triplet is about ∼35% of the MTC (Kooistra et al. 2005). Furthermore, in pilot studies, the stimulation current was increased far beyond supramaximal levels (up to 200%) without a subsequent decrease in triplet torque. Hence, the hamstring muscles were assumed not to be activated during supramaximal stimulation of the knee extensors (see also Discussion). Voluntary activation During experimental day 1, the level of maximal VA was determined in random order at 30°, 60°, and 90° knee angles. At each knee angle, subjects were asked to maximally generate isometric torques for about 3–4 s to determine MVC extension torque. Real-time torque was visible on a computer monitor and subjects were vigorously encouraged to exceed their maximal value, which was also displayed. MVC torque was defined as highest torque within a stable 1,000 ms plateau of the torque signal. MVC torque was taken as the highest value, which did not exceed preceding attempts by >5%, allowing a maximum of four attempts to prevent fatigue. In the incidental case a trial had no stable plateau, the trial was repeated. Subjects then performed a series of superimposed measurements consisting of a triplet applied to a fully relaxed muscle and another triplet superimposed on the stable part of the torque signal (Fig. 1). If the torque level was unstable (>1% torque fluctuation) prior (100 ms) to stimulation, the contraction was repeated. As all our subjects were very experienced at maintaining stable torque levels at all contraction intensities, it was rarely necessary to repeat a measurement. As an indication, for all subjects, less than 1 in 10 contractions (across all knee angles and intensities) were repeated. Several different contraction intensities with an emphasis toward higher intensities were chosen (25, 50, 70, 80, 90, and 100%MVC). At each knee angle, the order of the contraction intensities with superimposed triplet was randomized, and the order of the knee angles was randomized per subject. A 3-min rest period separated contractions. For each superimposed contraction, the accepted formula for the interpolated twitch torque technique (Allen et al. 1995; Babault et al. 2001, 2006; Becker and Awiszus 2001; de Ruiter et al. 2005; Kooistra et al. 2005; Maffiuletti and Lepers 2003; McKenzie et al. 1992; Newman et al. 2003; Oskouei et al. 2003) was used to determine the level of voluntary activation (VA): Fig. 1Typical example of a superimposed contraction. The resting triplet torque as a result of the electrical stimulation as well as the triplet torque increment on a 100%MVC contraction at the 90° knee angle are shown. The timing of the triplet stimulation is shown by the vertical arrows and vertical dashed lines. Voluntary activation (VA) is generally calculated using the following equation: VA (%) = 100–[(triplet torque increment/resting triplet torque) × 100]. In this particular example, voluntary activation was calculated to be: 100–[(4.2/94.6) × 100] = 95.6% With this method, the triplet torque increment is expressed as a percentage of the resting triplet torque, and subtraction from 100% results in a value for VA (Fig. 1). Maximal VA was obtained from the 100%MVC superimposed contraction. To illustrate the relationship between superimposed and voluntary torque from 90 to 100%VA, only subjects capable of an average maximal VA of >95% across knee angles were included in the present study. MVC torques as close to the muscles’ MTC as possible were thus obtained. Due to our strict inclusion criteria (average VA of >95%) at each knee angle, we were able to select a contraction at a lower than maximal torque level where the superimposed triplet was ∼10% of resting triplet. At this torque level, VA is conventionally calculated to be ∼90%, and we expressed this torque level as a percentage of MVC. We could thus compare the theoretical 10% torque increase implied by a conventionally calculated VA level of 90% with the torque increase actually achieved by our subjects during MVC. This provides us with a good indication of the potential torque increase that may be possible in regular healthy subjects with a maximal VA level of ∼90%. The ∼90% is a VA level that is frequently reported in the literature for regular healthy subjects (Babault et al. 2001, 2003; Bulow et al. 1993; de Ruiter et al. 2004; Newman et al. 2003; Oskouei et al. 2003; Suter and Herzog 1997). Furthermore, for each subject and for each superimposed contraction, the increment in torque as a result of the superimposed triplet was plotted as a function of the torque reached just prior to the superimposed stimulation (e.g. Figs. 2, 3a). Curvilinear relationships between voluntary and superimposed torque are often reported in the literature for group data (e.g. Behm et al. 2001; Dowling et al. 1994; Oskouei et al. 2003; Suter and Herzog 1997). In the present study, however, neither a linear nor a curvilinear fit closely followed the data points for any of the individual subjects at any knee angle. However, linear regression was performed on the highest four contraction intensities (70, 80, 90, and 100%), and R2 values were calculated at each knee angle for each subject. This was done to show that although superimposed torque increments were small at high contraction intensities, the torque increment continued to decrease with contraction intensity (e.g. Fig. 3a, b). Fig. 2Torque increment as result of the triplet versus voluntary torque delivered at the 30° knee angle by a normal healthy subject in a pilot study. The empty circles denote the first day of testing, and the black triangles the second day. The dashed and solid lines represent linear regression fits for the data points on experimental days 1 and 2, respectively, excluding the resting triplet torque at 0 Nm. The resting triplet torque is clearly underestimated as it is very similar to the triplet torque increment at ∼30 Nm. The dotted lines represent backward extrapolation of the regression fits for both experimental days and illustrate the underestimation of the resting triplet torque as a result of tendon slack. For each experimental day, the respective equations and R2 values are shown. Using the interpolated twitch torque technique, voluntary activation (VA) is calculated to be quite high for experimental days 1 (84.4%) and 2 (94.7%)Fig. 3a The torque increment as a result of the triplet on a voluntary contraction (y-axis) is shown as a function of the voluntarily delivered torque just before the triplet (x-axis). The data points shown are those of subject no. 1, a subject with high ability for maximal voluntary activation, at the 30° knee angle. The open circles denote results from experimental day 1 and the filled triangles from experimental day 2. It is immediately apparent that using a linear or a curvilinear fit over all data points is incorrect for this subject. The dashed and solid lines illustrate linear regression fits that have been performed for a voluntary torque level of ∼50–150 Nm for experimental days 1 and 2, respectively. The dotted lines represent backward extrapolation of the regression fits for both experimental days and illustrate the potential effect of tendon slack on the resting triplet torque, especially when compared to the triplet torque increment at ∼60 Nm. For clarity, and to illustrate the continued decrease of the triplet torque increment with increasing voluntary torque, the data points of torque levels above 140 Nm have been replotted in Fig. 3b. b Data points from Fig. 3A have been replotted for torque levels above 140 Nm. The torque increment as a result of the triplet is shown as a function of the voluntarily delivered torque just before the triplet (x-axis) for four high intensity contractions (70, 80, 90, and 100%MVC). The open circles represent data points obtained on experimental day 1 and the filled triangles experimental day 2. A linear regression line shows the continued decrease in triplet torque increment with increase in contraction intensity for each set of four data points with corresponding R2 values Surface electromyography Electromyographic activity of the RF, VL, and VM muscles was recorded using surface EMG electrodes (Blue Sensor, Ambu, Ølstykke, Denmark, lead-off area: 1.0 cm2). After shaving, roughening, and cleansing the skin with 70% ethanol, electrodes were placed on the muscle belly in a bi-polar configuration in line with the muscle fibre direction with a centre to centre inter-electrode distance of 25 mm. Reference electrodes were placed on bone structures, on each patella, and on the lateral epicondyle of the femur of the right leg. Surface EMG signals were amplified (×100), digitized (1 kHz), and stored with the force signal on computer disc. All EMG signals were band-pass filtered (10–400 Hz). Rectified surface EMG amplitude (rsEMG) was calculated for the RF, VL, and VM for 1,000 ms segments just before the superimposed stimulation. RsEMG values obtained during the highest torque level on each experimental day were set to 100%. Statistics All results are presented as mean ( SD. Knee angle effects were tested for significance with repeated measures ANOVA. If significant main effects were observed, Bonferroni tests were performed for post hoc analysis. Test-retest reliability was tested for using the intraclass correlation coefficient. The level of significance of all statistical analyses was set at P < 0.05. Results Torque levels during the superimposed contractions at the 25, 50, 70, 80, 90, and 100%MVC levels were not different from the intended values or different between experimental days. Linear extrapolation of the torque increment as a result of the triplet on the high intensity contractions (70, 80, 90, and 100%MVC) provided high R2 values at each knee angle (30°: 0.89 ± 0.09; 60°: 0.90 ± 0.10; and 90°: 0.87 ± 0.10 averaged over days), which were not different between knee angles or the two experimental days (e.g. Fig. 3b). The continuing consistent decrement of the triplet torque increment with increase in contraction intensity was denoted by the overall negative slope (−0.13 ± 0.05 across days and angles), which was not different between days and angles. The relationship between the triplet torque increment and voluntary torque is shown in Fig. 4 for the 30, 60, and 90° knee angles. Fig. 4The torque increment as result of the triplet is shown as a function of voluntary torque for subject no. 5 on experimental day 1. Note that at the higher contraction intensities (>70%), the shape of the curve is similar at the 30° (black triangles, solid line), 60° (white squares, dashed line), and 90° (grey circles, dotted line) knee angle at the higher (>70%MVC) contraction intensities There was no difference in resting triplet torque or MVC between experimental days and average values across days are presented. At the 60° knee angle, MVC was significantly (P < 0.05) greater compared to both 30 and 90° knee angles (Table 1). Resting triplet torque, however, was significantly (P < 0.05) lower at 30° compared to 60 and 90° (Table 1). Reproducibility for both MVC and resting triplet torque was very high, which was illustrated by high ICC values (Table 1). Table 1TorqueKnee angle30°60°90°ICCMVC (Nm)186.1 ± 36.6278.4 ± 44.5189.6 ± 19.70.99Triplet torque (Nm)65.9 ± 10.091.6 ± 15.880.1 ± 10.80.96**All values are means ± SD, averages across days shown, n = 7. Maximal voluntary isometric knee extension (MVC) and resting triplet torque for the 30°, 60°, and 90° knee angle Significantly different from 30° to 90° knee angle (P < 0.05) Significantly different from 60° to 90° knee angle (P < 0.05)* Significant ICC, the intraclass correlation coefficient (ICC) was calculated for the measurements made on experimental days 1 and 2 and is shown in the last column Voluntary activation Overall, there was no difference in maximal VA, determined by applying the ITT to the highest MVC, between experimental days and average values across days are presented. The significantly lower triplet torque at 30° compared to the 90° knee angle did not result in a lower maximal VA at 30° compared to 90° At each knee angle, maximal VA was very high and similar between knee angles (Fig. 5). Furthermore, at each knee angle, at torques where the triplet increment was ∼10% (8.5 ± 2.6%), VA was calculated to be ∼90% (91.5 ± 2.6%, across angles). This VA level turned out to be significantly higher compared with the relative voluntary torque level at which the triplet was superimposed. The torques at which VA was calculated to be 91.5 ± 2.6% represented only 79.6 ( 6.1% of MVC (across knee angles), a finding that was similar at each knee angle (P > 0.05). Note that at this torque level, the triplet torque increments were already small (3.2 ± 1.1% MVC) compared to the torque generated when the triplet was applied on the resting muscle (∼37% MVC). The key point we want to address with our study is illustrated in Fig. 6. For both subject nos. 1 and 7, a VA level of ∼90% was calculated at the 60° knee angle at a torque level of ∼250 Nm. This suggests a potential further increase of maximal torque by 10%. However, subject no. 1 was able to significantly increase his torque production by more than 30%. Fig. 5Average (2 days) maximal voluntary activation (VA) per subject for the 30° (black triangles), 60° (white squares), and 90° (grey circles) knee angles. Note that only 2 out of a total of 21 data points are below 95%VA. The horizontal dotted line indicates the 90%VA level that is generally obtained for the knee extensors by regular healthy subjectsFig. 6The torque increment as result of the triplet is shown as a function of voluntary torque for subject no. 1 (grey squares) and 7 (open circles) at the 60° knee angle on day 1. At the 60° knee angle, the maximal VA level of subject no. 7 calculated at 250 Nm was 90.7%, which resembled that of a regular subject, and is exceptionally low in our study (see lowest point in Fig. 5). A VA of 90.0% is also calculated at 240 Nm for subject no. 1. In both subjects, this implies a further potential increase in torque of 10%, yet a >30% torque increase (to 330 Nm) could be demonstrated in subject 1. This is most likely due to the exceptional neural drive of subject 1 (∼98%VA) during his best attempts at this knee angle EMG The rsEMG values obtained during MVC were similar between days at each knee angle (P = 0.96). For all submaximal contraction intensities, normalized rsEMG values for each muscle (RF, VL, and VM) at each contraction intensity was similar between days; hence, averaged values across days are presented. Furthermore, no difference between knee angles or muscles was found, and with each increase in contraction intensity, there was a significant increase in normalized rsEMG (Fig. 7a, VL shown). An increase of normalized rsEMG of ∼21% (across angles and muscles) was seen for a 25% increase in torque from 25 to 50%MVC (Fig. 7a). When contraction intensities approached MVC, there was a much larger increase in normalized rsEMG (∼34%) relative to the torque increase of only ∼18% at all knee angles (Fig. 7a). In regular healthy subjects, the highest MVC would have occurred at torque levels where VA would have been calculated to be ∼90%, and consequently the EMG–torque relationship would most likely have been considerably less curvilinear. This is illustrated by the two EMG–torque relationships shown in Fig. 7b for the VL at the 90° knee angle. The first EMG–torque relationship includes all contraction intensities measured, whereas for the second EMG–torque relationship, the 90 and 100%MVC contraction intensities have been removed. In the latter case, all rsEMG values are normalized to a torque level that corresponds to the maximal VA level of ∼90% (generally obtained by regular healthy subjects), the result being a near linear EMG–torque relationship (Fig. 7b). Fig. 7a Normalized rsEMG levels averaged over the two experimental days versus normalized torque for the vastus lateralis (VL) muscle at 30° (black triangles), 60° (white squares), and 90° (grey circles) knee angles. *Significantly different (P < 0.05) from preceding intensity level. At the lower contraction intensities, ΔEMG1 denotes the increase in EMG (21%, across muscles) that is accompanied by a comparatively larger increase torque (∼25%), which is denoted by ΔTorque1. Conversely, ΔEMG2 denotes the much larger increase in EMG (∼34%, across muscles) that is accompanied by a comparatively smaller increase in normalized torque (∼18%, denoted by ΔTorque2) as the contraction intensity approaches MVC. Note that on average for ΔTorque2, calculated voluntary activation (VA) increased from 91.5 ± 2.6% to 97.2 ± 2.3%. Thus, a 34% increase in normalized rsEMG was accompanied by an 18% increase in torque, for which only a ∼5.7% increase in VA (denoted by ΔVA) was calculated. As a consequence of the relatively large increase in EMG as MVC is approached, the rsEMG of contraction intensities below MVC are normalized to a relatively large value and are located well beneath the line of identity. b Normalized rsEMG levels averaged over days versus normalized torque for the vastus lateralis (VL) muscle at the 90° knee angle. The white circles represent rsEMG values that have been normalized to the MVC of subjects with a very high ability for voluntary activation (VA). With the black circles, the EMG–torque relationship for regular healthy subjects has been predicted. The black circles denote rsEMG values that have been renormalized to the rsEMG value reached at 90%VA. This is similar to the usual maximal VA in regular individuals. Note that the EMG–torque relationship predicted for regular healthy subjects is linear and closer to the line of identity compared to subjects with very high ability for voluntary activation As a consequence of the disproportionate increase in EMG as MVC is approached, normalized rsEMG values obtained at intensities below MVC are normalized to a relatively large value. Hence, these normalized rsEMG values are relatively smaller compared to the relative torques at which they have been obtained. For example, at 25 and 50%MVC, normalized rsEMG for the RF, VL, and VM muscle was significantly less than would be expected based on the %MVC at which they were obtained at all knee angles (Fig. 7a, VL shown). By normalizing the rsEMG levels to the rsEMG values obtained at a torque level that corresponds to the VA level of 90% as found in regular healthy subjects, normalized rsEMG values are closer to the line of identity, and a more linear EMG–torque relationship is seen (Fig. 7b). Discussion In the present study, the relationship between calculated VA and voluntary torque was assessed in subjects with high ability for maximal VA at different knee angles. At the same time, the EMG–torque relationship in these selected subjects was examined. The first and main finding of the present study was that when VA is calculated to be ∼90% (as in regular healthy subjects), this probably represents a considerable overestimation of the subjects’ ability to maximally drive their quadriceps muscles. An additional finding was that, although resting triplet torque was lower at 30° versus 90°, this was of minor influence on the calculated maximal VA in our subjects. Furthermore, a relatively large increase in normalized rsEMG was observed as MVC was approached, making the shape of the normalized EMG–torque relationship of the knee extensors curvilinear. Maximal voluntary activation and knee angle Subjects in the present study exhibited a consistent high level of maximal VA of the knee extensors (Fig. 5), which is higher than the maximal VA levels of 90% that are generally reported for the knee extensors in regular healthy subjects when applying the ITT (Babault et al. 2001, 2003; Bulow et al. 1993; de Ruiter et al. 2004; Newman et al. 2003; Oskouei et al. 2003; Suter and Herzog 1997). Although full activation of the knee extensors has been reported in the past (Bigland-Ritchie et al. 1986; Rutherford et al. 1986), in those studies a single superimposed twitch was used. Due to the declining signal-to-noise ratio with increase in contraction intensity, the detection of a single superimposed twitch is difficult, and 100% VA may mistakenly have been assumed (Dowling et al. 1994). To improve the signal-to-noise ratio, multiple stimuli have been suggested (Gandevia and McKenzie 1988; Suter and Herzog 2001). In the present study, triplet stimulation provided considerable resting triplet torque levels at each knee angle (Table 1), which ensured a good signal-to-noise ratio. Furthermore, pilot studies had shown no increase in triplet torque before and after an MVC at any knee angle. By using supramaximal and high frequency (300 Hz) triplet stimulation, we avoided the potential influence of post-activation potentiation on our data. A limitation of the ITT may be the scaling of the twitch increment to the resting twitch. When muscle is stimulated by a twitch at rest, all tendon slack has to be taken up; conversely, for a twitch superimposed on an ongoing isometric contraction, no further slack has to be taken up (Suter and Herzog 1997). Since calculated VA is determined by the scaling of the triplet torque increment to the resting triplet, calculated VA will be reduced by the influence of tendon slack. As the effect of tendon slack is greater at short versus long muscle length (Muraoka et al. 2004), VA will be influenced to a greater degree at short versus long muscle length. Besides tendon slack, length-dependent Ca2+ sensitivity may also influence the size of the resting triplet by shifting the resting triplet torque–length relationship to shorter muscle length compared to the MVC torque–length relationship. The length-dependent effect of tendon slack and Ca2+ sensitivity were reduced by the use of a triplet (Kawakami et al. 2001), but not abolished as triplet torque was significantly smaller at the 30° versus 90° knee angle (Table 1), whereas MVC levels were similar. This implies that despite using triplet stimulation, compared to the MVC, resting triplet torque remains relatively lower at the 30° versus 90° knee angle. The latter is supported by Figs. 2 and 3A, where the triplet response is similar at 0 and 25%MVC at the 30° knee angle. It is evident that calculated VA will be influenced by tendon slack, as the resting triplet is influenced by tendon slack, especially at short muscle length, whereas the triplet torque increment is not (Suter and Herzog 1997). Despite a lower resting triplet torque at the 30° compared to the 90° knee angle, maximal VA was not different between the 30° and 90° knee angle in our subjects. From the literature, it is unclear whether maximal VA is knee angle-dependent. Kubo et al. (2004) report a lower level of maximal VA at extended compared to flexed knee angles, whereas Suter and Herzog (1997) find maximal VA to be highest at the most extended knee angle (15°). Newman et al. (2003) and Babault et al. (2003), however, find no effect of knee angle on maximal VA, which is in line with our current findings. Note that due to our selection criteria (average maximal VA of 95%, and at least 90% at each knee angle), maximal VA was very high (∼97%) and close to 100%. Potential knee angle-dependent differences in maximal VA could consequently only occur within a very small range. Furthermore, the small size of the superimposed triplet torque makes the calculation of maximal VA insensitive to large differences in the size of the resting triplet torque. By hypothetically increasing the resting triplet torque at 30° by 20%, making it similar to the 90° knee angle, an increase in maximal VA of less than 0.5% is calculated at the 30° knee angle. In subject populations possessing lower maximal VA levels (e.g. patient groups), a relatively greater increment will be obtained on an MVC. This greater increment is more susceptible to the muscle-length dependent influence of, for example, tendon slack on the size of the resting triplet. The influence of tendon slack would have been greater still had a twitch instead of a triplet been used (Kawakami et al. 2001), leading to an even greater difference in resting twitch size at 30° versus 90°. Combined with the use of a resting twitch, therefore, a muscle length-dependent effect of tendon slack is expected to significantly influence the calculation of VA in this subject population. Calculated voluntary activation in relation to relative voluntary torque For the most accurate assessment of maximal VA, Behm et al. (1996) recommended using the MVC with a superimposed electrical stimulus. As mentioned previously, in regular healthy subjects a maximal VA level of ∼90% is often reported for the knee extensors using this method (Babault et al. 2001, 2003; Bulow et al. 1993; de Ruiter et al. 2004; Newman et al. 2003; Oskouei et al. 2003; Suter and Herzog 1997) implying that the MTC has almost been reached. In the present study, with our selected subject group, we were able to demonstrate that 90%VA is already calculated at torque levels that represent only ∼79% of MVC (across knee angles). In some subjects, this phenomenon is very pronounced. For the subject shown in Fig. 3a and b, already at a torque of ∼148 Nm, VA using the ITT was calculated to be 95.3 and 97.1% for experimental day 1 and 2, respectively. These are very high VA levels that have been calculated at 148 Nm, especially when compared with the MVC of >200 Nm that was obtained on both experimental days. Note that the higher part of the curve (in this example over 140 Nm), as shown in Fig. 3b can usually not be obtained in regular healthy subjects. In our experience, we can only demonstrate the existence of this part of the curve in subjects with a very high ability for maximal VA (>99% in this extreme example, subject no. 1 in Fig. 5). This finding strongly suggests that the 90%VA reported for regular healthy subjects is a large underestimation of the MTC of the muscle. A far greater torque increase therefore seems possible than the 10% implied by a calculated VA of 90% (Fig. 6). The main new finding, therefore, is that the present study for the first time quantifies the difference between calculated VA and relative voluntary torque. Experimental and methodological factors When relating the triplet torque increment to the triplet torque obtained at rest, it is important to maintain supramaximal stimulation during both conditions. In a previous study examining superimposed stimulation of the knee extensors (Behm et al. 1996), the authors reported significantly lower M-wave amplitudes during a superimposed stimulation on an MVC compared to during rest. They presumed that the contraction of the knee extensors resulted in a displacement of the stimulating electrode from its optimal position over the femoral nerve. In pilot studies of the present investigation, M-wave amplitude at rest and superimposed on an MVC was similar. Moreover, a substantial (50%) increase in stimulation current did not lead to an increase in the size of the triplet increment on high intensity contractions (70 and 80%MVC). This strongly suggests that even during high intensity contractions, there was supramaximal stimulation of the femoral nerve during superimposed triplet stimulation. By using multiple stimuli during superimposed stimulation, spinal reflexes may have more time to diminish the superimposed response (Herbert and Gandevia 1999; Herbert et al. 1997). Modeling suggests that the use of twin stimuli at 100 Hz has a minimal effect on the estimates of VA (Herbert and Gandevia 1999). By applying a stimulation frequency of 300 Hz in the present study, triplet stimulation occurred within 10 ms, which is even less when compared to the use of twin stimuli at 100 Hz. The influence of spinal reflexes on the superimposed stimulation is therefore considered negligible. Furthermore, although the current passes through the sciatic nerve that innervates the hamstring muscles, its stimulation was regarded negligible. This is illustrated by the very small EMG recording of the m. biceps femoris compared to the VL muscle during stimulation of the knee extensors [Fig. 8, data from a previous study of ours (Kooistra et al. 2005)]. We accordingly assumed a marginal activation, if at all, of the hamstring muscles during superimposed stimulation and therefore no influence on our calculation of VA. In a previous study by our group (de Ruiter et al. 2004), we measured the contribution of co-activation during isometric knee extensions during brief MVCs and found it to be very small (5–10%). As this finding is in accordance with previous findings (Newman et al. 2003; Psek and Cafarelli 1993), we did not measure coactivation in the present study and considered it minimal. Fig. 8Representative EMG recordings are shown for the vastus lateralis (VL, upper panel) and biceps femoris (BF, lower panel) muscle. On the left, the M-wave for the VL as a result of supramaximal twitch stimulation applied to the n. femoralis. The simultaneous EMG recording for the BF muscle is shown in the lower left panel. On the right, the EMG recording during maximal voluntary extension (top, right) and flexion (lower, right) torque is shown to illustrate that the lower M-wave of the BF is not due a lower sensitivity of the BF recordings compared to the VL recordings The EMG–torque relationship As expected, when MVC was approached we observed a relatively large increase in normalized rsEMG (Fig. 7a). This large increase makes the shape of the EMG–torque relationship of the knee extensors curvilinear. This is in contrast to the findings of Woods and Bigland-Ritchie (Woods and Bigland-Ritchie 1983) who report a linear EMG–torque relationship, yet in accordance with Alkner et al. (2000) and Pincivero and Coelho (2000) who show a non-linear relationship for the knee extensors. Based on a modelling study of motor-unit pools, Fuglevand et al. (1993) indicate that the difference between muscles that exhibit the linear as opposed to the nonlinear form of the EMG–torque relationship may be related to differences in firing rate behaviour rather than due to differences in recruitment organization. It is reasonable to expect that only subjects with a high ability for maximal VA (high neural drive) will achieve very high motor unit firing rates. For this large increase in firing rate, however, little gain in torque may be predicted (Herbert and Gandevia 1999). The large increase in excitation, as represented by the disproportionate increase in normalized rsEMG when MVC was approached, confirms the model study by Herbert and Gandevia (1999). However, in contrast to their model where essentially 100%MVC was achieved at 60% of maximal voluntary excitation, in the present study, we experimentally determined that the disproportionate increase in excitation lead to a more substantial increase in torque (∼18%). However, many factors may affect the EMG–torque relationship that have not been taken into account. Those include, signal cancellation from overlapping positive and negative phases of action potentials (Day and Hulliger 2001); the sigmoid relationship between motor unit force and firing rate (Herbert and Gandevia 1999); and the nonlinear distribution of recruitment thresholds (Fuglevand et al. 1993). Interestingly, in a review, de Luca (1997) states that the amplitude of the EMG signal should be normalized to values less than 80%MVC. Above this level, the EMG signal is said to be exceptionally unstable and, hence, is unable to provide a suitable reference point. As shown in the present study, very large increases in EMG coincide with very small increases in calculated VA (Fig. 7a). This is consistent with unstable EMG measurements for similarly high torque levels. EMG, superimposed torque and voluntary activation When combining the non-linearity of the EMG–torque relationship and non-linearity of the superimposed triplet-torque relationship, the EMG–torque relationship indicates that a large amount of extra EMG produces relatively little extra torque at high intensities. The triplet given to evoke the superimposed triplet could be considered as extra EMG in that three extra action potentials are added. As voluntary EMG is reduced in its effectiveness at producing torque in strong voluntary contractions, the extra activation from the triplet might be reduced similarly. This is in accordance with a model study of the adductor pollicis motoneuron pool (Herbert and Gandevia 1999). The authors predict large increases in motoneuronal excitation for minor increases in voluntary force and VA (as shown by small decreases in the superimposed twitch, their Fig. 9c). As mentioned above, this is consistent with the EMG–torque data in the present study (Fig. 7a). The small decreases in modelled interpolated twitch amplitude are also consistent with the current findings, where the triplet torque increment was only 3.2 ± 1.1%MVC at a torque level that corresponded to 79.6 ± 6.1%MVC and showed only a minor further reduction in size (to 1.0 ± 0.9%MVC). Note that the minor reduction in triplet increment with increasing voluntary torque is, and can only be, observed in those subjects consistently able to achieve high levels of VA, as would be expected from the model (Herbert and Gandevia 1999). This would indicate that these subjects are capable of excitation levels that (almost) evoke their MTC. Having noted this, it is important to point out that a 100% excitation may not be necessary to evoke the MTC as “only” 60% excitation was modelled to elicit 100% MTC (Herbert and Gandevia 1999). This is in agreement for example with the much greater excitation (EMG) levels attained during voluntary fast isometric knee extensions compared to during the plateau of an MVC (de Ruiter et al. 2004). How can we be certain that our selected subjects did indeed approach their true MTC? The only way to truly determine the MTC of a muscle is by the use of tetanic nerve stimulation. In two subjects with exceptional ability for maximal VA (>99%), supramaximal tetanic nerve stimulation (2 s, at 150 Hz) of the knee extensors was performed and torque levels showed no further increase above MVC (data not presented). Applying supramaximal tetanic nerve stimulation is not recommended, however, as it is highly unpleasant and could be harmful to certain vulnerable subject groups (e.g. subjects with an ACL deficiency). Currently, the ITT is the only feasible available technique used to determine VA of the knee extensors and, as mentioned previously, it is used extensively (Allen et al. 1995; Babault et al. 2003; Becker and Awiszus 2001; Behm et al. 1996, 2001; de Ruiter et al. 2004; McKenzie et al. 1992; Newman et al. 2003; Oskouei et al. 2003; Suter and Herzog 1997). The findings of the present study indicate that for the knee extensors, despite a commonly reported maximal VA of ∼90% for regular healthy subjects, a far greater relative torque increase seems possible than the 10% implied by 90%VA. This has important implications for conclusions that are drawn regarding calculated VA. Especially because many studies have used target force levels expressed as %MVC to investigate: effects of caffeine neuromuscular function (Kalmar and Cafarelli 1999), the %MVC at which full occlusion of the blood supply occurs (de Ruiter et al. 2006), and fatigue of the quadriceps in patients with multiple sclerosis for instance (de Haan et al. 2000). Clearly, knowledge of the potential of a far greater torque increase than, for example, 10% implied by a VA of 90% is very important. In conclusion, when maximal VA is calculated to be ∼90% (as in regular healthy subjects), this probably represents a considerable overestimation of the subjects’ ability to maximally drive their quadriceps muscles. Moreover, the effect that the length dependent size of the resting triplet has on the calculation of VA with the conventional method is only minimal when VA is greater than 95%. Furthermore, a curvilinear shape of the EMG–force relationship may be caused by a disproportionately large increase in normalized rsEMG when MVC is approached in subjects with very high capacity to drive their muscles maximally.	[ "voluntary activation", "twitch interpolation", "emg", "muscle length", "neural drive" ]	[ "P", "P", "P", "P", "P" ]
Pflugers_Arch-3-1-2040175	Dysregulation of the expression and secretion of inflammation-related adipokines by hypoxia in human adipocytes	The effect of hypoxia, induced by incubation under low (1%) oxygen tension or by exposure to CoCl2, on the expression and secretion of inflammation-related adipokines was examined in human adipocytes. Hypoxia led to a rapid and substantial increase (greater than sevenfold by 4 h of exposure to 1% O2) in the hypoxia-sensitive transcription factor, HIF-1α, in human adipocytes. This was accompanied by a major increase (up to 14-fold) in GLUT1 transporter mRNA level. Hypoxia (1% O2 or CoCl2) led to a reduction (up to threefold over 24 h) in adiponectin and haptoglobin mRNA levels; adiponectin secretion also decreased. No changes were observed in TNFα expression. In contrast, hypoxia resulted in substantial increases in FIAF/angiopoietin-like protein 4, IL-6, leptin, MIF, PAI-1 and vascular endothelial growth factor (VEGF) mRNA levels. The largest increases were with FIAF (maximum 210-fold), leptin (maximum 29-fold) and VEGF (maximum 23-fold); these were reversed on return to normoxia. The secretion of IL-6, leptin, MIF and VEGF from the adipocytes was also stimulated by exposure to 1% O2. These results demonstrate that hypoxia induces extensive changes in human adipocytes in the expression and release of inflammation-related adipokines. Hypoxia may underlie the development of the inflammatory response in adipocytes, leading to obesity-associated diseases. Introduction White adipose tissue has traditionally been considered as primarily a vehicle for the storage of fuel in the form of triacylglycerols, but several other functions are now recognised for the tissue. In particular, adipose tissue is a major endocrine organ secreting several key hormones, notably leptin and adiponectin [11, 26, 29, 37, 38]. These hormones are part of the large group of protein signals and factors secreted by adipocytes, termed adipokines, many of which are linked to immunity and the inflammatory response [14, 26, 29, 38]. These encompass major cytokines and chemokines, including TNFα, IL-1β, IL-6, IL-10, MCP-1 and MIF [11, 26, 38]. Obesity, which is characterised by a major expansion in adipose tissue mass, is associated with a state of chronic mild inflammation, there being increased circulating levels of inflammatory markers such as C-reactive protein, IL-6, IL-18, haptoglobin, MIF and PAI-1 [11, 14, 26, 38]. The production of a number adipokines increases during the expansion of adipose tissue mass in the obese (for example, leptin, MIF, TNFα and PAI-1), with the result that there is an inflammatory state within the tissue [11, 14, 26, 38]. An exception is adiponectin, which has an anti-inflammatory action [25, 49]. Inflammation in WAT is considered to be causal in the development of type 2 diabetes and the metabolic syndrome linked to obesity [11, 14, 26, 29]. There has, however, been little focus on why the rise in adipose tissue mass in the obese should lead to the increased production of inflammatory adipokines. One possibility, which we have recently suggested, is that it is a response to relative hypoxia in clusters of adipocytes distant from the vasculature, inflammation serving to increase blood flow and stimulate angiogenesis [38, 39]. Hypoxia, which occurs, for example, in solid tumours and during wound healing, induces a series of adaptive responses by cells, and a key role in the transmission of the hypoxic response is played by the hypoxia-inducible transcription factors, particularly hypoxia-inducible factor-1 (HIF-1) [6, 13, 32, 33, 44]. HIF-1 is composed of two subunits, HIF-1β, which is constitutively expressed, and HIF-1α, which is recruited in response to low O2 tension to yield the functional transcription factor [6, 32, 44]. HIF-1α is present in murine clonal adipocytes, and hypoxia has been shown to induce an increase in the expression of leptin and vascular endothelial growth factor (VEGF) in these cells [21]. Recently, hypoxia has also been reported to induce the production of PAI-1 and to inhibit the synthesis of adiponectin by 3T3-L1 adipocytes [4]; it is also reported to induce the expression of visfatin in these cells [31]. HIF-1α has been identified in human adipose tissue and is reported to be increased in obesity [3]. However, the extent to which human adipocytes respond to hypoxia and which genes are hypoxia-sensitive are unknown. In the present study, we have examined the effects of hypoxia on the expression and secretion of key adipokines linked to inflammation in human adipocytes differentiated from fibroblastic preadipocytes in culture. The results demonstrate that hypoxia, whether through low O2 tension or induced chemically by CoCl2, leads to a marked recruitment of HIF-1α in human adipocytes and that expression of the GLUT1 facilitative glucose transporter in adipocytes is hypoxia-sensitive. Importantly, the expression and secretion of several adipokines, including FIAF/angiopoietin-like protein 4, IL-6, leptin, MIF and VEGF, is stimulated by hypoxia in human adipocytes, while adiponectin and haptoglobin are inhibited. Materials and methods Human adipocyte culture Human subcutaneous preadipocytes, derived from adipose tissue pooled from seven female subjects, were obtained (together with culture media) from Zen-Bio (USA). The patients had a mean body mass index of 25 (range 22.5–28.2) and average age of 41 years (range 27–51 years). Cells were trypsinized from a 75-cm2 flask and plated at a density of 40,000/cm2 in a 24-well plate and maintained in preadipocyte medium containing Dulbecco’s modified Eagle’s medium (DMEM)/Ham’s F12 (1:1, v/v), 10% fetal calf serum (FCS), 15 mM HEPES, 100 U/ml penicillin, 100 μg/ml streptomycin and 0.25 μg/ml amphotericin B at 37°C in a humidified atmosphere of 95% air/5% CO2. Cells were induced at confluence by incubation in differentiation medium composed of adipose medium (AM) supplemented with 0.25 mM isobutyl methylxanthine and 10 μM of a PPARγ agonist for 4 days. The cells were then cultured with AM containing DMEM/Ham’s F-12 (1:1, v/v), 3% FCS, 1 μM dexamethasone, 100 U/ml penicillin, 100 μg/ml streptomycin and 0.25 μg/ml amphotericin B. The medium was changed every 3 days. Fully differentiated cells at day 15 post-induction were treated with CoCl2 or exposed to 1% O2 for up to 24 h. For CoCl2 treatment, wells incubated without CoCl2 were used as controls; 75–200 μM CoCl2 was used in initial experiments, and 100 μM CoCl2 was employed in a time course study. For exposure to hypoxia, the cells were transferred to a MIC-101 modular incubator chamber (Billups-Rosenberg, USA), which was flushed with 1% O2, 94% N2 and 5% CO2, and sealed and placed at 37°C for up to 24 h as indicated. The control cells were cultured in a standard incubator (21% O2 and 5% CO2). The cells were harvested in 700 μl of Trizol (Invitrogen, UK) or 150 μl of lysis buffer per well at the stated time points. Media were collected and centrifuged at 1,000 rpm for 10 min to remove cell debris and the supernatant stored at −20°C until required for analysis. All incubations at each time point were performed in replicates of up to six wells. RNA extraction and cDNA synthesis Total RNA was isolated from cells using Trizol, and 1 μg of RNA was treated with DNase I (Invitrogen) according to the manufacturer’s instructions. RNA concentration was quantified from the absorbance at 260 nm; all samples had a 260/280 nm absorbance ratio of 1.7–1.9. One microgram of DNase I-treated RNA was reverse-transcribed using a Reverse-iT™ 1ST strand synthesis kit (Abgene, UK) in the presence of anchored oligo dT in a total volume of 20 μl. Real-time PCR Quantitative real-time polymerase chain reactions (PCRs) were carried out in a final volume of 12.5 μl consisting of 12.5–50 ng of reverse-transcribed cDNA mixed with optimal concentrations of primers and probe and qPCR™ core kit (Eurogentec, UK) in 96-well plates on a Mx3005P detector (Stratagene, USA). The primer and probe sets were designed using Primer Express software (Applied Biosystems) and synthesized commercially; the primers and TaqMan probes were from Eurogentec. TaqMan probes were labelled with a reporter fluorescent dye (FAM: 6-carboxyfluorescein) at the 5′-end and a fluorescent dye quencher (TAMRA: 6-carboxy-tetramethyl-rhodamine) at the 3′-end. The sequence and optimal concentrations of primers and probes, together with the size of products, are as detailed previously [41, 42], with the exception of GLUT1, FIAF, HIF-1α, MIF and POLR2A which were as follows. GLUT1 (93 bp): 5′-ATACTCATGACCATCGCGCTAG-3′ (forward), 5′-AAAGAAGGCCACAAAGCCAAAG-3′ (reverse) and 5′-FAM-TGGAGCAGCTACCCTGGATGTCCTATCTGA-TAMRA-3′ (probe); FIAF (117 bp): 5′-GATGGCTCAGTGGACTTCAACC-3′ (forward), 5′-CCCGTGATGCTATGCACCTTC-3′ (reverse) and 5′-FAM-CCAGACCCAGCCAGAACTCGCCGT-TAMRA-3′ (probe); HIF-1α (75 bp): 5′-TCCAGTTACGTTCCTTCGATCA-3′ (forward), 5′-TTTGAGGACTTGCGCTTTCA-3′ (reverse) and 5′-FAM-CACCATTAGAAAGCAGTTCCGCAAGCC-TAMRA-3′ (probe); MIF (74 bp): 5′-AGCCCGGACAGGGTCTACA-3′ (forward), 5′-GCGAAGGTGGAGTTGTTCCA-3′ (reverse) and 5′-FAM-CTATTACGACATGAACGCGGCCAATGT-TAMRA-3′ (probe); POLR2A (81 bp): 5′-ATGGAGATCCCCACCAATATCC-3′ (forward), 5′-CATGGGACTGGGTGCTGAAC-3′ (reverse) and 5′-FAM-TGCTGGACCCACCGGCATGTTC TAMRA-3′ (probe). Typically, the amplification started with 2 min at 50°C, 10 min at 95°C and then 40 cycles of the following: 15 s at 95°C and 1 min at 60°C. Human POLR2A was used as an endogenous reference. This housekeeping gene was selected based on comparison with β-actin, where POLR2A demonstrated no significant changes in expression when exposed to CoCl2 and 1% O2 (data not shown). Relative quantitation values were expressed using the method (see user bulletin no. 2, ABI Prism 7700, pp 11–15, Applied Biosystems), as fold changes in the target gene normalised to the reference gene (POLR2A) and related to the expression of the untreated controls. The PCR efficiency in all runs was close to 100%, and all samples were analysed in at least duplicate. Measurement of HIF-1α and adipokines by ELISA Total HIF-1α in cell lysates was measured with an enzyme-linked immunosorbent assay (ELISA) kit (R&D Systems, UK) according to the manufacturer’s protocol. Cells were solubilised in lysis buffer consisting of 50 mM Tris (pH 7.4), 300 mM NaCl, 10% (w/v) glycerol, 3 mM ethylenediaminetetraacetic acid, 1 mM MgCl2, 20 mM β-glycerophosphosphate, 25 mM NaF, 1% Triton X-100, 25 μg/ml leupeptin, 25 μg/ml pepstatin and 3 μg/ml aprotinin. Before measuring HIF-1α by ELISA, the total protein content of the lysates was determined by the BCA protein assay reagent (Sigma, UK). Adiponectin, IL-6, leptin, MIF and VEGF were measured in cell culture media using commercial ELISA kits (R&D Systems). The assays were conducted in 96-well microplates according to the manufacturer’s instructions. Measurement of HIF-1α by Western blotting Samples were prepared as described above for the HIF-1α ELISA assay. Fifteen micrograms of protein were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis and then transferred to a nitrocellulose membrane. The membranes were blocked and probed with polyclonal goat anti-human HIF-1α (R&D Systems) or mouse monoclonal anti-α-tubulin (Sigma, UK) as the primary antibody, then subjected to HRP-conjugated anti-goat IgG (R&D systems) or anti-mouse IgG (Santa Cruz Biotechnology) as the secondary antibody. Specific proteins were visualised with the enhanced chemiluminescence reagent (Amersham, UK). Statistical analysis The results are expressed as mean values±SE. Differences between groups were analysed by unpaired two-tailed Student’s t tests. Results HIF-1α expression during differentiation of human adipocytes We first investigated whether human white adipocytes express the HIF-1α gene and synthesize the encoded protein when differentiated in culture. Quantitative changes in HIF-1α gene expression after the induction of adipocyte differentiation were analysed by real-time PCR. Although the HIF-1α gene was expressed both before and after differentiation, there was a marked decrease in HIF-1α mRNA level after differentiation was induced (Fig. 1a): By day 2, the mRNA level fell to <25% of that seen at day 0 and remained low. The expression of a classical adipokine–leptin— was determined as a reference gene. As expected, leptin mRNA was differentiation-dependent and detectable only at day 2 after the induction of differentiation, increasing progressively thereafter (Fig. 1b). Fig. 1Time course of quantitative changes in HIF-1α mRNA level (a), leptin mRNA level (b) and HIF-1α protein (c) during the differentiation and development of human adipocytes. Confluent human preadipocytes (day 0) were differentiated and cultured for up to 14 days. Relative mRNA levels were normalised to human POLR2A at day 0 (or day 2 when no signal was evident at day 0). Total protein concentration of HIF-1α in cell lysates was measured by ELISA. Results are means±SE (n = 3) HIF-1α protein expression was examined at day 0 and day 14 using a specific ELISA and was detected at both time points. In parallel with the mRNA level, HIF-1α protein was substantially higher in preadipocytes than in adipocytes (Fig. 1c). Response to CoCl2 treatment Hypoxic effects can be mimicked by the divalent transition-metal ion cobalt. In the next experiments, differentiated human adipocytes (day 15 post-induction) were treated with CoCl2 at concentrations of 75 and 200 μM for 24 h. CoCl2 treatment resulted in a marked increase in total HIF-1α protein, the level being fivefold higher at both doses compared to untreated cells (Fig. 2a). In contrast, HIF-1α mRNA level was reduced three to fivefold by CoCl2 treatment (Fig. 2b). Fig. 2Effect of CoCl2 at two dose levels on HIF-1α protein concentration (a) and on the mRNA levels of HIF-1α, GLUT1 and adipokines (b) in human adipocytes. Differentiated human adipocytes at day 15 were incubated in medium containing 75 or 200 μM CoCl2 for 24 h. Total HIF-1α protein and mRNA level of adipokines were measured as described in Fig. 1. Results are means±SE (n = 6). In a,P < 0.01, P < 0.001 compared with controls; in b †P < 0.001 compared with controls. Open bars, controls; shaded bars, 75 and 200μm CoCl2 (light and dark bars, respectively) To assess whether human adipocytes respond to the CoCl2-induced increase in HIF-1α protein by increasing the expression of ‘classical’ hypoxia-sensitive genes, the mRNA level of the GLUT1 facilitative glucose transporter was examined as a reference; this gene is recognised in other cells to be upregulated under hypoxic conditions. Treatment with CoCl2 induced a sixfold increase in GLUT1 mRNA level in the adipocytes (Fig. 2b). The expression of a series of inflammation-related adipokine genes was then examined in the CoCl2-treated adipocytes, and three distinct responses were observed. No effect was found on TNFα or adipsin mRNA level at either of the two doses of CoCl2. There was, however, a significant reduction in mRNA level for adiponectin, MCP-1 and haptoglobin (Fig. 2b). Adiponectin mRNA level was decreased three to fourfold, MCP-1 by up to fivefold and haptoglobin mRNA by less than half with both concentrations. In the case of IL-6, the high dose of CoCl2 reduced mRNA level threefold, but there was no response with the lower dose (Fig. 2b). In contrast, CoCl2 treatment resulted in a marked increase in mRNA level for FIAF, leptin, MIF, PAI-1 and VEGF. The increase was six to sevenfold for VEGF and FIAF, and two to fourfold for PAI-1, MIF and leptin (Fig. 2b). Time course of response to CoCl2 The previous experiment indicated that in most cases, there was little difference in the effect of 75 or 200 μM CoCl2 on HIF-1α protein level and adipokine gene expression. However, with leptin, MIF and PAI-1 in particular, effects were noted with the lower, but not the higher, dose. This might suggest the possibility of additional, or toxic, responses to CoCl2 at higher doses. Consequently, in the subsequent time course study, a concentration of 100 μM was employed at day 15 post-induction. There was a rapid and significant accumulation of HIF-1α protein in human adipocytes after the addition of CoCl2. The protein level was increased 4.7-fold by 2 h of CoCl2 treatment and peaked at 8 h, at which point it was increased 7.7-fold before gradually falling (Fig. 3a). HIF-1α protein level was also assessed by Western blotting, and the pattern of results was similar to that obtained by ELISA (Fig. 3b). Fig. 3Time course of effects of CoCl2 (100 μM) on total HIF-1α protein level (a and b) and GLUT1 mRNA levels (c) in human adipocytes. Cells were cultured and treated as in Fig. 2. Results are means±SE (n = 4). P < 0.01, *P < 0.001 compared with 0-h controls GLUT1 was again used as a hypoxia-sensitive reference gene, and the pattern of the time course of the changes in GLUT1 mRNA level was similar to that of HIF-1α protein. The mRNA level was significantly increased (2.3-fold) by 2 h, peaking at 8 h, at which point it was 14-fold higher, gradually declining thereafter (Fig. 3c). The time course of the effect of CoCl2 on the expression of genes encoding adipokines was next examined. A group of adipokines including FIAF, IL-6, leptin, MIF, PAI-1 and VEGF showed marked increases in mRNA level in response to CoCl2, some of which were both substantial and rapid. FIAF mRNA level was increased 30-fold by 2 h after the addition of CoCl2, peaking at 8 h, at which point it, had risen 210-fold (Fig. 4a). There was a fall thereafter, but even at 24 h, FIAF mRNA was still elevated 30-fold. A similar pattern was also evident with IL-6, PAI-1 and VEGF, but the changes were not as large (Fig. 4a). In comparison to the above genes, there was a more gradual response to CoCl2 with leptin and MIF, the mRNA for each reaching a peak (fourfold) at 16 h, with little change for the rest of the 24-h incubation period (Fig. 4a). Intriguingly, MCP-1 mRNA level exhibited a substantial acute increase, being elevated >20-fold at 2 h, but there was a rapid decline thereafter such that at 16 and 24 h the levels were similar to the controls (Fig. 4a). Fig. 4Time course of effects of CoCl2 (100 μM) on adipokine mRNA levels of FIAF, IL-6, leptin, MCP-1, MIF, PAI-1, TNF-α and VEGF (a) and adiponectin, haptoglobin and adipsin (b) in human adipocytes. Relative mRNA levels. Cells were cultured and treated as in Fig. 2. Results are means±SE (n = 4). P < 0.05, P < 0.01, P < 0.001 compared with 0-h controls In contrast, to the genes described so far, no effect was found on mRNA level throughout the 24-h incubation period in the case of adipsin (Fig. 4b), while TNFα mRNA was also unchanged except for a transitory increase at 2 h only (Fig. 4a). There was, however, a significant reduction in mRNA level for adiponectin and haptoglobin after 8–16 h of the treatment; by 24 h, adiponectin and haptoglobin mRNA was decreased by 15.7-fold and threefold, respectively (Fig. 4). Induction of HIF-1α protein and GLUT1 gene expression by low O2 tension In the next set of experiments, the direct effects of low O2 tension were examined. Human adipocytes at day 15 post-differentiation were exposed to normoxic conditions (21% O2) or to 1% O2 for up to 24 h. Some cells were returned to normoxia for 16 h after 8 h of exposure to 1% O2 to examine the reversibility of responses. Total HIF-1α protein level was measured by both ELISA and Western blotting. HIF-1α concentration in cells was rapidly and substantially upregulated on exposure to 1% O2, being 7.8-fold higher by 4 h than in control cells cultured under normoxia. The HIF-1α level fell markedly after 8 h, but at 24 h, it was still 3.8-fold higher than in control cells (Fig. 5a). The transfer of adipocytes back to normoxic conditions after 8 h at 1% O2 led to a rapid fall in HIF-1α, the level of the protein at 1 h being threefold lower than in cells maintained in 21% O2 throughout (the cells were returned to normoxia after 8 h, rather than 24 h, because the high level of HIF-1α at the earlier time point facilitated assessment of changes with normoxia). By 16 h after the return to normoxia, the level of HIF-1α was similar to that in the control adipocytes. The pattern of HIF-1α level as determined by ELISA was mirrored by Western blot analysis (Fig. 5b). Fig. 5Time course of effects of hypoxia (1% O2) on total HIF-1α protein level (a, b) and GLUT1 mRNA (c) levels in human adipocytes. Differentiated adipocytes at day 15 post-induction were exposed to 1% O2 for up to 24 h; some cells were transferred back to normoxic conditions (21% O2) for 1 or 16 h after 8 h of hypoxia (8- to 1- and 8- to 16-h groups). Results are means±SE (n = 4). P < 0.05, P < 0.01, P < 0.001 compared with controls GLUT1 mRNA level was markedly increased (fivefold) after 4 h of exposure to 1% O2 (Fig. 5c) and reached a peak at 24 h, being 14-fold higher than in cells maintained under normoxia. After exposure to hypoxia for 8 h, the level of GLUT1 mRNA returned to normal after 16 h in 21% O2 (Fig. 5c). Regulation of adipokine gene expression by low O2 tension The effect of hypoxia on the expression of genes encoding key inflammation-related adipokines was next examined. The first group of adipokines, which encompassed FIAF, IL-6, leptin, MIF, PAI-1 and VEGF, exhibited rapid and marked increases in mRNA level in response to hypoxia, as with CoCl2 treatment. FIAF mRNA level was increased 3.3-fold by 4 h after exposure to 1% O2, rising to 11-fold by 24 h (Fig. 6a). FIAF mRNA level was completely normalised by 16 h of reexposure to normoxia after 8 h in 1% O2. A similar pattern was also evident with five other adipokines. The changes were even more substantial in the case of leptin and VEGF. After 4 h of hypoxia, leptin and VEGF mRNA levels were increased four to fivefold and rose such that at 24 h they were 29-fold and 23-fold greater than in controls, respectively (Fig. 6a). Again, after 16 h of reexposure to normoxia after 8 h in 1% O2, the mRNA level of these two adipokines had returned to normal. Fig. 6Time course of effects of hypoxia (1% O2) on adipokine gene expression in human adipocytes.Relative mRNA levels of FIAF, IL-6, leptin, MIF, PAI-1, VEGF (a) and adiponectin, haptoglobin, MCP-1, adipsin, TNF-α (b). Cells were cultured as in Fig. 5; some cells were transferred back to normoxic conditions (21% O2) for 1 or 16 h after 8 h of hypoxia (8- to 1- and 8- to 16-h groups). Results are means±SE (n = 4). P < 0.05, P < 0.01, P < 0.001 compared with controls IL-6, MIF and PAI-1 mRNA levels also showed significant increases in response to hypoxia, but these increases (2.5-fold to fivefold at 24 h), were less than those exhibited by FIAF, leptin and VEGF. Returning the adipocytes to 21% O2 led to a reversal by 16 h of the hypoxia-induced increases in IL-6, MIF and PAI-1 mRNA levels; in the case of IL-6, the mRNA was lower than in control cells at 16 h after the transfer from hypoxic to normoxic conditions. In contrast to the genes described so far, there was a significant reduction in adiponectin and haptoglobin mRNA levels by the end of the 24-h exposure period, the decrease being two to threefold (Fig. 6b). This reduction was reversed after 1 h of reexposure to 21% O2, but surprisingly was evident again after 16 h of reexposure. Adipsin mRNA level showed a significant, but very small, increase at 8 and 16 h (1.5 and 1.7-fold, respectively). There were no significant changes in the level of TNFα mRNA during exposure to 1% O2, while MCP-1 mRNA was also unchanged after 4 and 24 h of hypoxia. For reasons that are not apparent, there was a significant reduction in MCP-1 mRNA at 8 h, and this was sustained 1 h after return to normoxia (Fig. 6b). Adipokine secretion in hypoxia The secretion into the medium of a selected group of adipokines was examined during the exposure to hypoxia using specific ELISAs. A small but statistically significant reduction in adiponectin secretion was observed after 8 and 24 h incubation under 1% O2 (Fig. 7a). This reduction was not evident after reexposure to 21% O2 for 16 h (Fig. 7a). Fig. 7Time course of effects of hypoxia (1% O2) on the release of adipokines from human adipocytes. Adiponectin secretion (a) and IL-6, leptin, MIF and VEGF release (b–e) after exposure to hypoxia. Adipokine concentration was measured by ELISA in the medium of cells cultured as in Fig. 6; some cells were transferred back to normoxic conditions (21% O2) after 8 h of hypoxia and the adipokine concentration then measured 16 h later (8- to 16-h group). Values are means±SE (n = 6). P < 0.05, P < 0.01, *P < 0.001 compared with controls In marked contrast to adiponectin, there was a significant increase in leptin, MIF and VEGF release after both 8 and 24 h exposure to hypoxia (Fig. 7c–e). The increases in secretion were modest for MIF, but substantial in the case of both leptin and VEGF. Thus, after 24 h of hypoxia, leptin release was tenfold higher, while that of VEGF was fourfold greater. Elevated rates of secretion were still evident after transfer of the adipocytes from hypoxic to normoxic conditions, indicating a long-term carryover effect of low O2 tension. The secretion of IL-6 was not altered after 8 h of exposure to hypoxia, but it was significantly increased at 24 h (Fig. 7). Discussion An important recent development in our understanding of obesity is the emergence of the concept that it is characterised by a state of chronic low-grade inflammation [7, 9, 38]. White adipose tissue produces a number of adipokines linked to inflammation, including adiponectin, IL-1β, IL-6, TNFα, MCP-1 and MIF [11, 14, 26, 38], and their synthesis is generally substantially increased in obesity—with the exception of adiponectin whose production and circulating levels fall [15, 45]. We have recently proposed that hypoxia may occur in adipocytes distant from the vasculature as adipose tissue mass expands and that this underlies the inflammatory response exhibited by the tissue [38]. The central aim of the present study was to examine the effects of hypoxia on the expression and secretion of key inflammation-related adipokines in human adipocytes. The transcription factor HIF-1 is a key signal in the cellular response to hypoxia, the α-subunit of which is highly induced by hypoxia [13, 32, 44]. HIF-1α expression was evident in human adipocytes, as noted previously in murine fat cells, and the level of both the mRNA and the protein fell after the induction of differentiation. It has been reported previously that HIF-1α mRNA level in 3T3-L1 cells peaked at 3 h after the addition of induction medium and then rapidly decreased [16]. Immunoreactive HIF-1α has been reported in murine adipocytes, and hypoxia results in an increase in the amount of the protein in cultured fat cells [4, 21]. We have observed that the HIF-1α gene is expressed in human, mouse and rat WAT depots, and expression occurs in both the mature adipocytes and in the cells of the stromal vascular fraction. A previous study has suggested that HIF-1α is predominantly expressed in the stromal vascular fraction of obese subjects [3]. Hypoxia stabilises HIF-1α protein, which is otherwise (under normoxic conditions) degraded by an ubiquitin-dependent proteasome [32]. Total HIF-1α protein level in human adipocytes increased rapidly and substantially under hypoxic condition, both with 1% O2 and by treatment with CoCl2. The elevated HIF-1α protein rapidly fell on return to 21% O2, and the level was even decreased compared to controls after 1 h of reoxygenation. Thus, the accumulated HIF-1α in human adipocytes is rapidly degraded on returning to normoxia. Indeed, it was observed in another human adipocyte system (SGBS cells) that HIF-1α protein declined to control levels by just 10 min after return to 21% O2 (Wang, unpublished results). CoCl2, a known inducer of HIF-1α, has been widely employed as a hypoxia mimic and was used here to examine the effects of chemically induced hypoxia on adipokine gene expression. Previous studies have shown that at a global gene expression level, both CoCl2 and ambient hypoxia regulate a similar group of genes, and the observed similarity in gene expression appears to be dependent on functional HIF-1α [40]. CoCl2 stabilises HIF-1α under normoxia via inhibition of the iron-containing HIF prolyl hydroxylase enzyme, which plays a critical role in mediating normal hypoxic signalling by modifying HIF-1α and targeting it for degradation. Our results demonstrate that CoCl2 rapidly induces an accumulation of HIF-1α and influences the expression of key inflammation-related adipokines in human adipocytes. Other studies have also found that CoCl2 (100 μM) causes HIF-1α accumulation in various cell types, including bone marrow stromal cells, cancer cells and brown adipocytes [20, 24, 27]. In contrast to protein levels, HIF-1α mRNA level was inhibited by CoCl2 treatment. This indicates that CoCl2 modulates HIF-1α at the posttranslational rather than the transcriptional level. There is, however, evidence for hypoxic induction of HIF-1α mRNA in some cells [17, 43]. Under hypoxia, GLUT1 gene expression has been shown to increase in several cell types, including rat liver cell lines and fibroblasts [1, 5], Chinese hamster ovary cells [48] and in human cancer cells [12]. Our data demonstrate that GLUT1 expression is also substantially upregulated by hypoxic conditions in human adipocytes. Measurement by Western blotting (using a specific antibody) of GLUT1 in total tissue lysates of adipocytes exposed to 1% O2 demonstrated that the level of the protein was also increased by hypoxia; at 24 h, GLUT1 protein was eightfold higher than in the normoxic controls (Wood, unpublished results). This phenomenon presumably reflects an adaptation of glucose metabolism to hypoxic conditions, with a marked stimulation of glycolysis in the face of a fall in aerobic metabolism. During the adaptive response to hypoxia, the expression of several genes encoding glycolytic enzymes is recognised to increase [34]. To sustain higher levels of glycolysis, there is a need for an increase in glucose uptake, and this would be aided by an increase in the level of GLUT1 expression in hypoxia. GLUT1 is a key facilitative glucose transporter in white adipocytes, but other transporter isoforms such as GLUT4, GLUT8, GLUT10 and GLUT12 are also expressed in the cells [46, 47]. The main focus of the present study was the effects of hypoxia on the expression and secretion of adipokines, and qualitatively similar results were obtained over the same time scale with both 1% O2 and CoCl2 treatment. Given the mechanism of action of CoCl2, this is consistent with the effects of hypoxia on adipokine production being mediated by HIF-1. Of the adipokines studied, the expression of TNFα and adipsin were not altered, and these genes would seem not to be hypoxia-sensitive. However, the expression of both adiponectin and haptoglobin (an acute phase protein) in human adipocytes was reduced by hypoxic conditions, and in the case of adiponectin, this was also evident at the level of the secretion of the protein into the medium. The reduction, albeit modest, in adiponectin production in human adipocytes by hypoxia is consistent with a recent report in murine 3T3-L1 adipocytes [4]. There was a discrepancy in the present study between the changes in the scale of alterations in adiponectin mRNA level and the amount of adiponectin secreted into the medium. This may, however, reflect the lag between changes in mRNA and alterations in secreted protein. In marked contrast to adiponectin, hypoxia stimulated the expression of the genes encoding FIAF, IL-6, leptin, MIF, PAI-1 and VEGF. VEGF is involved in the development of the vascular system, promoting angiogenesis, and is well known to be activated by hypoxia [50]. In the present study, an increase in VEGF protein secretion, as well as mRNA level, was observed in human adipocytes with hypoxia. This is in agreement with studies on rat omental adipose tissue and on murine 3T3-F442A adipocytes [21, 51]. Leptin is the most extensively studied adipokine, and adipose tissue is quantitatively the main leptin-producing organ. Activation of leptin gene expression by hypoxia resulted in a substantial increase in leptin release by human adipocytes. This would be expected to reduce food intake and stimulate energy expenditure, and as such, hypoxia could provide a mechanistic basis for the increase in leptin as adipose tissue mass expands in obesity. Leptin has been reported to be an angiogenic factor [35], and the physiological significance of a hypoxia-mediated induction of leptin secretion by adipocytes may relate to a local effect to stimulate the development of new blood vessels during adipose tissue expansion. Therefore, increased leptin production, together with that of VEGF, under hypoxic conditions could also reflect an adaptive process to promote angiogenesis, allowing appropriate oxygenation of adipose tissue. It has been previously established that leptin is a hypoxia-inducible hormone, and its production is increased by hypoxia in tissues such as the placenta, cardiac myocytes and cancer cells [8, 10, 23] as well as in murine 3T3-F442A adipocytes [21]. FIAF (also known as angiopoietin-like protein 4) is expressed particularly in adipose tissue and the liver, and has been suggested to play a role in adipose differentiation, systemic lipid metabolism and energy homeostasis [18]. In parallel with the activation of leptin and VEGF, FIAF expression was highly activated by hypoxia in human adipocytes; FIAF expression has also been shown to be hypoxia-sensitive in cardiomyocytes through a HIF-1-dependent pathway [2]. The major upregulation of FIAF by hypoxia in adipocytes may again relate to angiogenesis, this being one of the functions suggested for this protein [22]. However, a major effect of FIAF is in the clearance of plasma triglycerides through an inhibition of lipoprotein lipase, and it also appears to stimulate lipolysis [18]. Thus, the hypoxia-induced stimulation of FIAF production may alter lipid utilisation. The present study shows that the expression and release of MIF, a potent macrophage migration inhibitory factor which has recently been shown to be secreted from human adipocytes [36], was also stimulated by hypoxia. MIF has been described as a pro-inflammatory agent, and elevated expression has been seen in many inflammatory pathologies. Although MIF may play a role in the normal adaptive response to hypoxia, it is likely that exaggerated expression of MIF may contribute to the inflammatory state. It has also been proposed that MIF may play an important role in oncogenic transformation and tumour survival, and recent studies have implicated MIF in the regulation of VEGF expression [28]. In this context, during hypoxia, elevated MIF expression may play an important role in enhancing angiogenesis as well as macrophage recruitment. In contrast to MIF, no consistent pattern was observed with MCP-1, an important factor in macrophage recruitment; the acute, but transitory, increase in MCP-1 mRNA observed with CoCl2 may reflect an effect of the metal unrelated to hypoxia. The expression and release of the classic inflammatory factor IL-6 was significantly upregulated by hypoxia. Thus, hypoxia can directly affect key components of the inflammatory cascade within adipose tissue. The expression of PAI-1, an acute phase protein involved in fibrinolysis, was modestly increased in human adipocytes by hypoxia, consistent with the effects observed in 3T3-L1 adipocytes [4]. Induction of PAI-1 expression by hypoxia has also been reported in HepG2 cells [19] and in vascular smooth muscle cells [30]. Overall, the present results indicate that a number of genes are hypoxia-sensitive in human adipocytes and suggest that HIF-1 is involved in the transmission of the response to low O2 tension in these, as in other, cell types. The hypoxia-sensitive genes in adipocytes include those involved in basic metabolic processes, such as glucose transport, as well as those linked to energy balance, inflammation and angiogenesis. The stimulation of IL-6, leptin, MIF, PAI-1 and VEGF expression, together with the downregulation of adiponectin synthesis, is consistent with our recent proposal that hypoxia may underlie the inflammatory response in adipose tissue in obesity and play a causal role in the development of obesity-associated diseases [38]. There was, however, no effect of hypoxia on the expression of TNFα, a key pro-inflammatory cytokine. The hypoxia-induced changes in IL-6 and adiponectin, with its anti-inflammatory action, in particular, might be expected to impact on insulin sensitivity.	[ "inflammation", "adipokines", "hypoxia", "obesity", "adipose tissue", "metabolic syndrome" ]	[ "P", "P", "P", "P", "P", "P" ]
Knee_Surg_Sports_Traumatol_Arthrosc-4-1-2190783	Effects of a contoured articular prosthetic device on tibiofemoral peak contact pressure: a biomechanical study	Many middle-aged patients are affected by localized cartilage defects that are neither appropriate for primary, nor repeat biological repair methods, nor for conventional arthroplasty. This in vitro study aims to determine the peak contact pressure in the tibiofemoral joint with a partial femoral resurfacing device (HemiCAP®, Arthrosurface Inc., Franklin, MA, USA). Peak contact pressure was determined in eight fresh-frozen cadaveric specimens using a Tekscan sensor placed in the medial compartment above the menisci. A closed loop robotic knee simulator was used to test each knee in static stance positions (5°/15°/30°/45°) with body weight ground reaction force (GRF), 30° flexion with twice the body weight (2tBW) GRF and dynamic knee-bending cycles with body weight GRF. The ground reaction force was adjusted to the living body weight of the cadaver donor and maintained throughout all cycles. Each specimen was tested under four different conditions: Untreated, flush HemiCAP® implantation, 1-mm proud implantation and 20-mm defect. A paired sampled t test to compare means (significance, P ≤ 0.05) was used for statistical analysis. On average, no statistically significant differences were found in any testing condition comparing the normal knee with flush device implantation. With the 1-mm proud implant, statistically significant increase of peak contact pressures of 217% (5° stance), 99% (dynamic knee bending) and 90% (30° stance with 2tBW) compared to the untreated condition was seen. No significant increase of peak contact pressure was evaluated with the 20-mm defect. The data suggests that resurfacing with the HemiCAP® does not lead to increased peak contact pressure with flush implantation. However, elevated implantation results in increased peak contact pressure and might be biomechanically disadvantageous in an in vivo application. Introduction In 50% of the patients undergoing routine arthroscopy, articular cartilage defects are detected [6]. The majority of cartilage lesions are observed in patients over the age of 40 years [6, 11]. Approximately 20% of these lesions were categorized as full thickness cartilage defects (Grade IV) [6]. The average defect size in 1,000 knee arthroscopies was 2.1 cm2 with 42% of these defects being between 2 to 4 cm2 in area [11]. Location of the lesions are predominantly on the medial femoral condyle [6, 11]. Clinical and experimental studies have shown that a relationship exists between larger defect size and degeneration of the articular surface. Linden et al. [19] demonstrated that the presence of a significant chondral defect is associated with a much higher incidence of osteoarthritis than occurs in the general population. Hughston et al. [13] showed that larger osteochondral defects are associated with a poorer clinical outcome than smaller defects. The treatment for articular cartilage defects offers a great variety of options depending on the severity of the damage. The main factors to be considered are the patient age, the size of the lesion, the location of the defect, and further damage to the joint. Biological repair techniques, such as autologeous chondrocyte transplantation, osteochondral transplantation (OATS, Mosaicplasty) and marrow stimulation techniques are proposed for the treatment of localized cartilage defects in younger patients with a normal weight bearing axis and no further damage to the joint [3, 8]. However, many middle aged patients are affected by localized cartilage damage and are neither fit for biological repair methods, nor for traditional resurfacing techniques such as unicompartimental or total joint replacement. A contoured articular prosthetic unicondylar femoral resurfacing prosthesis (HemiCAP®, Arthrosurface Inc., Franklin, MA, USA) was developed to offer a treatment option among the currently used modalities. It is intended for use as a partial resurfacing device of the femoral condylar surface of the knee when only one compartment is affected by posttraumatic, degenerative disease or necrosis associated with large unstable articular defects with significant subchondral bone exposure. However, it is unknown if the femoral resurfacing device leads to increased peak contact pressure to the opposing biological structures such as meniscus and articular cartilage of the tibial plateau. The purpose of our study was to determine peak contact pressures in the medial compartment of human cadaveric knees under different loading conditions comparing the normal articular surface to flush and proud HemiCAP® implantation and defect. Based on the experimental study, biomechanical device safety on opposing structures can be assessed and a guideline might be given to the clinician for prosthetic implantation while avoiding potentially damaging effects. Materials and methods Knee joints A total of eight fresh-frozen knee cadaver specimens (3 pairs, 2 single) were used for data collection in this study. The specimens were obtained from donors, who consented in writing during their lifetime to the use of their body for research and education. The average age of the seven male and one female specimen was 71 years (range 61–81) with an average weight of 71 kg (range 62–85). Specimens were selected after inspection of the medial compartment according to the following criteria: Intact femorotibial cartilage, intact meniscus, and intact collateral and cruciate ligaments. Thirteen specimens were excluded following these criteria. Another six specimens were excluded due to specimen failure during the testing procedure: The reason for exclusion was fracture of the femur or tibial plateau (2 specimens), rupture of the anterior cruciate ligament (2 specimens) or rupture of the patella tendon (2 specimens). All specimen failures were female human cadaver knees. Technology A specially designed knee simulator was used for this study (Fig. 1). Similar to in vivo conditions, the main system composed of artificial muscle, force transducer sensor, the joint angle detection and the ground reacting force form a closed loop. The ground reaction force is adjustable according to the donor’s weight. The knee simulator consists of a loading frame (MTS 858 Bionix, MTS Systems, Eden Prairie, MN, USA) with a long stroke main actuator driven by a hydraulic pump unit (MTS 505.11 silent flow) to simulate body weight and the vertical hip displacement in the mechanical axis of the lower limb. Ankle joint simulation is performed with a hinge joint with one free motion axis. The possible rotation during the movement occurs in the artificial hip joint as if standing with fixed shoe contact. A load transducer is fixed between the mounting plate and ankle joint to detect the vertical ground reacting force (U3 load cell, Hottinger-Baldwin, Darmstadt, Germany). Two smaller actuators apply loads, which simulate the quadriceps force. The tendons of the quadriceps muscle are attached to customized curved cryo-clamps, which avoid patella tilting. These cryo-clamps are connected to a waterproof force transducer (SSM-AJ 500, Interface, Scottsdale, AZ, USA) and connected to an artificial muscle (Fluidic muscle MAS, Festo, Esslingen, Germany). The mathematical model shows that the properties of this fluidic muscle is comparable with a skeletal muscle [28]. Fig. 1Schematic drawing of the knee simulator that was used for this study A 0.1 mm thin electronic pressure measuring sensor (K-scan 4000, Tekscan, Boston, MA, USA) was placed in the medial compartment above the menisci and fixed with sutures in a manner that no displacement was possible. The sensor consists of load-sensing regions oriented in a grid with 1.27 mm spacing between rows and columns. Each region is referred to a sensel with piezo-resistive pigments to determine the total compressive load within that region. The size of the sensor is 28 × 33 mm with 62 sensels/cm2. The K-scan sensor was successfully used in comparable applications by several authors [18, 21, 25, 31]. The HemiCAP® implant (Arthrosurface Inc., Franklin, MA, USA) is a contoured articular prosthetic (CAP) unicondylar femoral resurfacing prosthesis consisting of two components, a fixation component and an articular component, that mate together via a taper interlock to provide stable and immobile fixation of the implant and stress-bearing contact at the bone/prosthetic interface. The fixation component is a modified titanium cancellous screw with a tapering distal tip, a full-length cannulation, and a proximal female taper bore. The articular component is a dome-shaped component manufactured of a Cobalt–Chromium–Molybdenum alloy with titanium plasma spray coverage on the underside for bony in-growth. Each diameter comes in a variety of incremental offset sizes which correspond to the superior/inferior and medial/lateral radius of curvatures at the implant site. The size of the implant used in this study was 20 mm in diameter with different offset sizes, matching the individual specimen joint curvature. Testing protocol The specimens were aligned using a fixed laser beam to achieve correct alignment in the mechanical axis of the lower limb. The mechanical axis was defined by a line through the center of the head of the artificial hip joint, the center of the knee joint and the center of the hinge joint representing the artificial ankle. For calibration of the sensor the ankle hinge joint was secured with two aluminum plates perpendicular to the ground. Thereby the knee was fixed in full extension. Each sensor was individually preconditioned and calibrated intra-articulary with a two-point calibration method at 700 and 1,500 N according to manufacturer’s guidelines. Definitions of the correct angles of the actual knee position were adjusted with a custom-made goniometer and by the displacement-controlled main rod. During the test cycles the cadaver were sprayed with saline solution to prevent dehydration. The specimens were tested in four different conditions: (1) Untreated knee, (2) Flush HemiCAP® (20 mm) implanted in the central weight bearing area of the medial femoral condyle, (3) 1 mm proud implantation to adjacent cartilage, (4) 20 mm defect (Fig. 2a–d). Each knee was tested in static knee stance positions (5°, 15°, 30°, 45°) with body weight ground reaction force, 30° flexion with two times body weight ground reaction force and during a knee-bending dynamic cycle (10 cycles) with body weight ground reaction force. The setting of ground reaction force to be achieved during the trial was adjusted to the living body weight of the donator of the cadaver (e.g. 70 kg∼700 N). The loading rate for exploring the dynamic contact pressures was 0.1 Hz/s. Fig. 2Different testing conditions (a Untreated; b Flush; c 1 mm; d Defect) Operative technique All specimens were released at their femoral fixation during preparation of a new testing condition. The exact position was marked and recorded to maintain specimen position in the knee simulator across all testing conditions. The position of the sensor did not change during the preparation. All procedures were performed by the same investigator (CB). The medial femoral condyle was exposed by a medial parapatellar incision. The knee was flexed to 90° to expose the central weight-bearing portion. A drill guide was used to place a pin perpendicular to the joint surface representing the center of the defect. The center of the defect was determined by measuring the condylar width and bisecting the distance. The cannulated instrumentation set ensured that the vertical axis was maintained throughout the procedure. After drilling a pilot hole, the fixation component was inserted. A contact probe determined the radius of curvature in two planes. With a matching reamer, the site for the implantation was prepared and a sizing trial with corresponding offsets inserted. The selected device was oriented in the correct planes and connected to the anchoring screw with a tapered lock. In order to allow for careful removal of the articular component, the guide wire was initially drilled through the condyle and remained inside the cannulated screw during the trial. After testing with flush implantation (Fig. 2b), the device was removed with the guide wire. The anchoring screw was elevated with a counter-clockwise quarter turn representing 1 mm and the device re-inserted for positioning 1 mm above the adjacent cartilage (Fig. 2c). Defect condition was tested after removing the guide wire and device leaving a 20-mm osteochondral defect of 3–4 mm depth (Fig. 2d). Data analysis Data were obtained using I-Scan software 4.23 (Tekscan, Boston, MA, USA). Peak contact pressure was assessed and recorded as the highest value at each stance position and during the dynamic knee-bending cycle. Mean, median and standard deviation values were evaluated using MS-Excel 2003 (Microsoft Inc., Redmond, WA, USA) and SPSS 11.0 (SPSS Inc., Chicago, IL, USA). For statistical analysis a paired sampled t test to compare means (significance, P ≤ 0.05) was used. Results Continuous data were obtained at every trial. No difference in the quality of data collection was seen comparing the stance positions or dynamic testing. The highest mean peak contact pressures were noted with the implant being 1 mm proud (Fig. 3; Table 1). Statistically significant increase of peak contact pressures of 217% (5° stance; P ≤ 0.004), 99% (dynamic knee bending; P ≤ 0.02) and 90% (30° stance with 2tBW; P ≤ 0.03) compared to the untreated condition was seen. Average results of 45° static testing demonstrated only slightly higher values compared to the other testing positions, indicating that the implant was non-weight bearing and lost contact to the sensor at this position (Fig. 3; Table 1). Fig. 3Mean peak contact pressures with one standard deviationTable 1Peak contact pressures at different testing conditionsTesting positionTesting conditionPeak contact pressure (Mpa)Mean difference to untreated (%)P value5°Untreated3.09 ± 0.86 (3.05, 1.92–4.86)N/AN/AFlush3.98 ± 2.19 (3.59, 2.15–9.02)29NS1 mm proud9.80 ± 4.37 (9.36, 4.70–18.37)217≤0.004Defect3.35 ± 1.39 (3.39, 1.90–5.83)8NS15°Untreated3.01 ± 0.81 (2.95, 2.01–4.11)N/AN/AFlush3.69 ± 1.79 (3.89, 1.38–7.06)23NS1 mm proud 9.19 ± 9.27 (5.87, 2.10–30.49)205NSDefect2.94 ± 1.07 (2.84, 1.69–5.09)−2NS30°Untreated3.18 ± 0.94 (3.52, 1.46–4.26)N/AN/AFlush2.89 ± 0.85 (3.02, 1.03–3.83)−9NS1 mm proud 7.07 ± 5.97 (5.24, 1.03–19.81)122NSDefect3.31 ± 1.06 (3.11, 2.05–5.34)4NS45°Untreated4.77 ± 1.85 (4.23, 2.66–7.90)N/AN/AFlush4.96 ± 2.02 (4.18, 2.79–7.92)4NS1 mm proud 5.79 ± 2.76 (5.54, 2.01–11.82)21NSDefect4.94 ± 1.81 (4.65, 2.05–7.54)4NSDynamicUntreated5.84 ± 2.12 (5.14, 3.85–9.82)N/AN/AFlush6.02 ± 2.05 (5.21, 3.46–8.97)3NS1 mm proud 11.61 ± 6.39 (10.50, 6.22–25.46)99≤0.02Defect5.68 ± 1.76 (5.11, 3.76–8.24)−3NS30° 2tBWUntreated6.57 ± 2.31 (6.01, 4.09–11.15)N/AN/AFlush6.05 ± 1.40 (5.82, 4.68–8.97)−8NS1 mm proud 12.49 ± 8.02 (9.67, 5.83–30.49)90≤0.03Defect7.38 ± 4.68 (5.31, 4.75–17.70)12NSValues given as mean ± SD (median, range)NS not significant On average, comparison of the untreated normal knee with flush device implantation demonstrated neither statistically significant differences in peak contact pressure during the dynamic knee-bending cycle nor static testing (5°, 15°, 30°, 45°) or two times body weight GRF at 30° static testing (Table 1). Six out of eight specimens demonstrated similar peak contact pressures in both testing conditions, two specimens showed a noticeable increase in peak contact pressures (flush device implantation) averaging 29% during 5° static testing and 23% in 15° static testing. Data evaluation of the two respective trials with outlying measurements demonstrated edge loading at the border of the implant to the adjacent cartilage (Figs. 4b, 5b). Fig. 4Peak contact pressure at 15° static knee stance position with single body weight of specimen No. 2. The color represents the spectrum of pressures (high pressure-red; low pressure-blue). From left to right the different testing conditions are displayed (a Untreated, b Flush, c 1 mm proud, d Defect). Marked increase of peak contact pressure at the edge of the implant to the adjacent cartilage is demonstrated (b)Fig. 5Picture of specimen No. 2 before testing with flush implantation of the device in front view (a) and top view (b). The arrow indicates the area of peak contact pressure displayed in Fig. 4b No significant increase of peak contact pressures was evaluated with the 20-mm osteochondral defect. However, average increase by 8% (5° stance) and 12% (30° stance with 2tBW) were found compared to untreated-condition (Table 1). Furthermore, maximum values were higher compared to untreated condition (except 45° stance position and dynamic testing cycle). Discussion The patient aged over 40 years with a full thickness chondral or osteochondral defect reflects a serious problem for the orthopedic surgeon. Considered as being too old for biological repair of the defect, primarily the patients are mostly managed with conservative, non-surgical treatment including weight reduction, physical therapy to increase and support musculature, unloading braces and medications such as NSAID’s, intraarticular injections (Corticosteroids, Hyaluronic acid, etc.) and dietary supplements. However, conservative treatment at best ameliorates the symptoms. Biomechanical studies have shown that untreated osteochondral defects may result in increased contact pressures [4, 7, 17]. Animal models proved that untreated osteochondral defects undergo progressive degenerative changes [5, 14]. Whereas smaller defects might have the capacity for healing [5], it was shown that larger defects resulted in resorption of the osseous walls of the defect, the formation of a large cavitary lesion, and the collapse of the surrounding articular cartilage and subchondral bone as well as to degeneration of the opposing tibial articular surface [5, 14]. A“threshold” effect for localized full thickness defects was described with rim concentration around the lesion becoming a factor for defects greater than 10 mm in diameter (0.79 cm2) [7]. Defect repair is therefore regarded crucial to prevent or delay progressive degenerative joint destruction. However, unicompartimental or total knee arthroplasty represent procedures of final resort for some of the affected patients. The HemiCAP® resurfacing prosthesis (Arthrosurface Inc., Franklin, MA, USA) offers an interim or alternative treatment strategy for the middle-aged patient with a full thickness cartilage defect. However, effects of a metallic implant articulating with intact opposing tibial articular cartilage and meniscus remain largely unanswered to date. An experimental study assessing the functional and biological response to its use in a goat model resulted in good clinical outcomes. One year after implantation, gross necropsy and histologic data implied the biocompatibility and functionality of the implant. No ongoing degenerative joint disease was apparent [15]. Macroscopic and histological analysis showed that the cartilage around the implant was largely intact although some focal fraying and erosion and limited meniscal damage was observed. The opposing tibial plateau cartilage surface was generally intact but exhibited some focal erosions of variable depth [15]. The outcome was substantially better by comparison to other reported experimental animal studies in goats with untreated full-thickness defects [14, 29]. A patellar resurfacing prosthesis resulted at an average of 8.1 years in 71% good or excellent cases with no progressive degenerative changes in the non-resurfaced apposing femoral articular cartilage as long as 16 years [9]. A successful clinical outcome with a metallic distal femoral prosthesis articulating directly against the menisci and proximal tibial plateau over 30 years was documented in a case report [12]. Our results revealed significant increase of peak contact pressures with the implant being proud to the surrounding cartilage compared to the untreated condition with an average maximum increase of 217% and 205% at 5° and 15° stance compared to untreated knees. Results in a biomechanical model using osteochondral plugs for the treatment of osteochondral defects demonstrated increased contact pressures up to 57% compared to intact condition in elevated or angled plugs with an edge placed higher than the adjacent cartilage [16, 17]. Thus, increased peak contact pressures may suggest biomechanical disadvantages and may cause damage to opposing structures. Several in vitro studies have shown increased chondrocyte apoptosis and matrix deformation after peak stress loading of bovine and human articular cartilage [2, 20, 22, 23]. However, quantitative thresholds above which elevated pressure is detrimental are not defined yet. Flush implantation showed no statistical increase of peak contact pressures compared to untreated. However, in two specimens we found marked increase of peak contact pressures at the edge of the implant. Evaluation of the pictures of the specimen after the trial showed that the implant did not appear to be proud to the adjacent cartilage in the front view (Fig. 5a). However, in the top view it seemed that the implant was in level with the adjacent cartilage or even slightly elevated at the point of increased peak contact pressure (Fig. 5b). Reduced quality of the adjacent cartilage directly next to the measured peak contact pressure might have lead to less resistive cartilage capacity allowing for edge loading. Thus, it appears to be crucial to have healthy surrounding cartilage around the implant and to spend special care during the implantation process so that the implant is not elevated above the adjacent cartilage at any point. Koh et al. [16] showed in their biomechanical study, that slightly recessed osteochondral grafts with the highest edge placed flush to neighboring cartilage demonstrated nearly normal contact pressures, whereas elevated angled grafts produced increased contact pressures. Our results confirm the observation concluding that slightly recessed implantation should be considered if flush implantation cannot be accomplished. Defect situation in our study did not result in significant increase of peak contact pressures. Rim stress concentration has been described by several authors in their biomechanical models evaluating the contact pressures in full thickness cartilage defects [4, 5, 7, 16, 17, 26]. However, results were uncertain whether the peak contact pressure increases at the rim of the defect. Whereas Koh et al. [17] reported average increase of peak contact pressure by 23% compared to the untreated condition, Raimondi et al. [26] and Nelson et al. [24] found no elevated contact stresses at the rim of the defect. These different observations might be caused by the different deformation of defect rims under a given load, especially in larger defects than in smaller defects. The relationship of defect size to the condylar surface may be also a factor [7]. The methodology concerning the model (animal or human being cadaver), the experimental setup (loading mechanics, dynamic or static, etc.) and the pressure-sensitive sensor and performance of the trial have also to be considered. Limitations of this study have to be considered and include the following. (1) This is a human cadaver study and biomechanical model. Only approximation of the living system can be achieved. However, the pre-selection of the applied load by the ground reaction force gave the opportunity to load the specimens with the known body weight of the donor, and thus better approximate the forces that occur in the living system [30]. Furthermore, the knee bending dynamic measurement may have provided a more accurate reproduction of physiologic weight-bearing activity. However, the complexity of weight-bearing motions including all the muscle groups for knee motion could not be reproduced. Applied forces were too high for female human cadaver knees resulting in a high specimen’s failure rate. (2) Peak contact pressures were determined by an electronic pressure sensitive sensor. The reliability of the K-scan sensor was verified in several studies [10, 18, 21, 31]. In comparison to other measuring devices (Fuji Photo Film), the Tekscan K-sensor proved to be superior [10, 27]. Limitations of the sensor include the thickness (0.1 mm), its sensitivity to temperature changes, its disposition for crinkling and the establishment of the position [1]. Although the position of the sensor was accurately secured, a small amount of displacement and crinkling could not be excluded. A certain amount of loss of sensitivity during the different testing conditions and imprecise calibration due to the different surface materials (metal on cartilage, cartilage on cartilage) have to be considered. Some drop of data point was observed at some specimen. However, peak contact values seemed not to be affected. (3) Stable, reproducible ground reaction force in full extension could not be established by the simulator. The quadriceps tendon could not be adequately tensioned in this position. (4) Implantation of the device was performed in the central weight-bearing area. Results might be different for the implantation in different areas of the medial femoral condyle or lateral condyle. In conclusion, the data suggest that resurfacing with the HemiCAP® with flush implantation does not lead to significantly increased peak contact pressure. However, elevated implantation results in significantly increased peak contact pressure and might be biomechanically disadvantageous in an in vivo application. Further research is necessary to evaluate the effects of the prosthetic device on contact pressures after loss of meniscus function and in longer continuous dynamic testing.	[ "articular prosthetic device", "peak contact pressure", "tibiofemoral joint", "osteochondral defect", "knee biomechanics" ]	[ "P", "P", "P", "P", "R" ]
Matern_Child_Health_J-2-2-1592246	Preconception Care for Improving Perinatal Outcomes: The Time to Act	Introduction A healthy baby and a healthy mother are valued hopes and dreams of families of all cultural heritages. National health goals in many countries around the world prioritize infant and maternal mortality and morbidity. In the United States, as in many other industrialized countries, pregnancy outcomes and maternal and infant health indicators have improved dramatically. At the turn of the 20th century, 7.28 women and 96 infants died for every 1,000 babies born alive [1]. By the end of the year 2000, the maternal mortality rate had decreased to less than 0.08 and infant mortality had decreased to less than 7 deaths per 1,000 live births [2]. These changes were characterized as one of the ten “great public health achievements” of the 20th century [3]. Much of this progress can be attributed to changes in social and living conditions as well as the development and implementation of effective medical interventions. Despite this progress, during the last two decades of the 20th century, even in the presence of significant breakthroughs in medical technology and its application, improvements in maternal and infant pregnancy ouctomes slowed down significantly, and in some cases, outcomes deteriorated. For example, from 1960 to 1980, the maternal mortality rate decreased from 32.1 to 9.4 deaths per 100,000 live births, a decrease of 70.7%. From 1980 to 2000, the maternal mortality rate decreased only 12.8%, from 9.4 to 8.2 (Figure 1) [2]. At the same time, from 1980 to 2000, the proportion of babies born preterm, very preterm, low birth weight, and very low birth weight increased by 26%, 8.2%, 14.7% and 25.9%, respectively (Figure 2) [4]. However, unlike other maternal and infant health indicators, from 1980 to 2000 the infant mortality rate continued to decrease almost at a similar pace as that from 1960 to 1980, dropping by 45.2% from 12.6 to 6.9 infant deaths per 1,000 live births, compared with a drop of 51.5% from 26.0 to 12.6 per 1,000 live births for 1960 to 1980 (Figure 1) [2]. Finally, it is estimated that the proportion of babies born in the United States who have a serious or major structural defect that can have adverse effects on their health or development continues to be about 3% [5]. Associated with this slowing rate of improvement (and, in some cases, deterioration) in pregnancy outcomes and maternal and infant health indicators is a shift in the leading causes of infant mortality. In 1960, maternal complications of pregnancy (including complications of placenta, cord, and the membranes) did not appear on the list of the 10 leading causes of infant mortality [1]. By 1980, maternal complications of pregnancy became the fifth leading cause of infant mortality, and the third leading cause of infant death in 2002, after congenital anomalies and low birth weight/preterm delivery [6, 7]. In 2002, congenital anomalies, low birth weight, preterm delivery, and maternal complications of pregnancy accounted for 46.4% of all infant deaths in the United States (12,996 infant deaths) (Figure 3) [7]. Although some of these infant deaths might have been prevented through interventions targeted at improving the health of mothers and modifying behaviors contributing to adverse pregnancy outcomes, poor maternal health, behaviors contributing to adverse pregnancy outcomes, and maternal complications of pregnancy continue to be prevalent. For example, 28.6% of women who gave birth between 1993 and 1997 were reported to have an obstetric complication, 4.1% had a preexisting medical condition, and 43.0% experienced some form of maternal morbidity (i.e., an obstetric complication, a preexisting medical condition, a cesarean section, or any combination of the three types of morbidity) [8]. In 2002, 26.1% of all deliveries were performed using cesarean section, presumably because of maternal or infant complications [8]. Fig. 1Maternal and infant mortality rates, United States, 1960–2002Fig. 2Percent of all live births that are preterm delivery, very preterm delivery, low birth weight, and very low birth weight, United States, 1980–2002 Early prenatal care is too late One of the reasons that progress in improving pregnancy outcomes has slowed down, and in some cases reversed direction, is that we have failed to intervene before pregnancy to detect, manage, modify, and control maternal behaviors, health conditions, and risk factors that contribute to adverse maternal and infant outcomes. Although we know many interventions that, if delivered before pregnancy, could improve pregnancy outcomes, we have failed to make those services and interventions available to couples and women in need. Women of childbearing age suffer from a variety of chronic conditions that could potentially contribute to adverse pregnancy outcomes. For example, in 2002, the U.S. Department of Health and Human Services reported that 6.1% of women of reproductive age have asthma, 5% are obese, 3.4% have cardiac disease, 3.0% are hypertensive, 9.3% are diabetic, and 1.4% have thyroid disorder [9]. Moreover, a substantial proportion of women continue to enter pregnancy with risks proven to contribute to adverse pregnancy outcomes: in 2002, 11.4% of pregnant women smoked during pregnancy, a risk factor for low birth weight [4]; at the same time, 10.1% of pregnant women and 54.9% of women at risk of getting pregnant consumed alcohol, a risk factor for fetal alcohol syndrome [10]. Finally, using the Perinatal Periods of Risk approach, researchers in three cities (New York City; Tulsa, Oklahoma; and Kansas City, Missouri) concluded that racial and ethnic disparities in feto-infant mortality were largely related to maternal health, and, interventions to reduce feto-infant mortality should include preconception care and improvements in women's health [11–13].Fig. 3Leading causes of infant mortality, United States, 1960, 1980, and 2002. *IMR = Infant Mortality Rate The prenatal care “revolution” of the 1980s resulted in an increase in the proportion of women receiving early prenatal care (defined as prenatal care begun in the first trimester) [2]. After decreasing from 76.3% in 1980 to 75.8% in 1990, the proportion of women receiving early prenatal care increased to 83.2% in 2000 [2]. However, for many women, “early prenatal care is too late” [14]. By the time a pregnant woman makes it to her first early prenatal visit, most fetal organs are already been formed, and many interventions to prevent birth defects or adverse maternal and infant outcomes come too late to have any effect. As a result, many national organizations now recommend routine preconception care. For example, the March of Dimes recommends that “as the key physician/primary care provider, the obstetrician/gynecologists must take advantage of every health encounter to provide preconception care and risk reduction before and between conceptions—the time when it really can make a difference” [14]. The American Academy of Pediatrics (AAP) and the American College of Obstetricians and Gynecologists (ACOG) recommend that “all health encounters during a woman's reproductive years, particularly those that are a part of preconception care, should include counseling on appropriate medical care and behavior to optimize pregnancy outcomes” [15]. Preconception interventions work Interventions designed to lower preconception risks are known collectively as preconception care. These interventions are characterized by the need to start—and sometimes complete—a designated intervention before conception occurs. To simplify matters, ACOG and AAP have grouped the main components of preconception care under four categories of interventions: maternal assessment (e.g., family history, behaviors, obstetric history, general physical exam); vaccinations (e.g., rubella, varicella and hepatitis B); screening (e.g., HIV, STD, genetic disorders); and counseling (e.g., folic acid consumption, smoking and alcohol cessation, weight management) [15]. Many of these interventions are currently available to women and, when indicated, to couples; however, they are not systematically delivered. A comprehensive review of the literature identified a long list of risk factors suggested to be included in comprehensive preconception care (Table 1). The strength of evidence supporting the efficacy of the various components of preconception care varies greatly [16]. Clinical practice guidelines (CPGs) have been developed for reducing the risk related to some of these factors. We identified 14 conditions for which CPGs exist and for which there is scientific evidence demonstrating effectiveness in improving pregnancy outcomes (Table 2) [17--74]. Table 1Risk factors (by Category) identified in the literature as needing attention during the preconception period1Chronic diseases: Diabetes; heart disease; high blood pressure; thyroid disease; asthma; anemia; kidney disease; metabolic and hematological disorders; depression and other mental disorders; autoimmune disease; and physical disability (access)2Infectious diseases: Vaccine-preventable diseases (rubella, hepatitis B, varicella, influenza, and tetanus); HIV/AIDS; syphilis, chlamydia, and other sexually transmitted diseases; periodontal disease; toxoplasmosis, and cytomegalic inclusion virus3Reproductive concerns: Unplanned pregnancies; contraception; infertility; adverse past pregnancy outcomes (preterm delivery, birth defects, fetal/infant death, maternal complications)4Genetic/inherited conditions: Sickle cell anemia; thalassemia; Tay-Sachs disease; fragile X syndrome; Down syndrome; cystic fibrosis; muscular dystrophy; hearing and vision loss associated with genetic predisposition5Medications and medical treatment: Prescription medications contraindicated in pregnant women (FDA's Category X Drugs, 117 products in 2001 PDR, antiepileptic drugs, oral anticoagulants for maternal clotting disorders, and Accutane); diagnostic radiation exposures6Personal behaviors and exposures: Smoking; alcohol consumption; illicit drug use; overweight/underweight; folic acid supplement use; domestic violence; eating disorders; exposure to infections; exposures to chemicals and other environmental toxins; consumption of over-the-counter medications; hyperthermia (e.g., from sauna use)Note. This list is not exhaustiveTable 2Selected preconception risk factors for adverse pregnancy outcomes for which clinical practice guidelines have been developed•Folic acid. Daily use of vitamin supplements containing folic acid has been demonstrated to reduce the occurrence of neural tube defects by two thirds [17–24]•Rubella seronegativity. Rubella vaccination provides protective seropositivity and prevents the occurrence of congenital rubella syndrome [25–27]•Diabetes (preconception). The three-fold increase in the prevalence of birth defects among infants of women with type 1 and type 2 diabetes is substantially reduced through proper management of diabetes [28–31]•Hypothyroidism. The dosages of Levothyroxine™ required for treatment of hypothyroidism increases in early pregnancy. Levothyroxine™ dosage needs to be adjusted for proper neurologic development [32–34]•HIV/AIDS. If HIV infection is identified before conception, timely treatment can be administered and women (or couples) can be given additional information that can influence the timing of the onset of pregnancy [35–41]•Maternal phenylketonurea (PKU). Women diagnosed with PKU as infants have infants with mental retardation. However, this adverse outcome can be prevented when mothers adhere to a low phenylalanine diet before conception and continue it throughout their pregnancy [42, 43]•Oral anticoagulant. Warfarin, which is used for the control of blood clotting, has been demonstrated to be a teratogen. To avoid exposure to warfarin during early pregnancy, medications can be changed to a nonteratogenic anticoagulant before the onset of pregnancy [15, 44–46]•Anti-epileptic drugs. Certain anti-epileptic drugs are known as teratogens. Before conception, women who are on a regimen of these drugs and who are contemplating pregnancy should be prescribed a lower dosage of these drugs [15, 47–51]•Isotretinoins (Accutane®): Use of isotretinoins in pregnancy to treat acne results in miscarriage and birth defects. Effective pregnancy prevention should be implemented to avoid unintended pregnancies among women with childbearing potential who use this medication [15, 52–54]•Smoking. Preterm birth, low birthweight, and other adverse perinatal outcomes associated with maternal smoking in pregnancy can be prevented if women stop smoking during early pregnancy. Because only 20% of women successfully control tobacco dependence during pregnancy, cessation of smoking is recommended before pregnancy [55–58]•Alcohol misuse. No time during pregnancy is safe to drink alcohol, and harm can occur early, before a woman has realized that she is or might be pregnant. Fetal alcohol syndrome and other alcohol-related birth defects can be prevented if women cease intake of alcohol before conception [59–64]•Obesity. Adverse perinatal outcomes associated with maternal obesity include neural tube defects, preterm delivery, diabetes, cesarean section, and hypertensive and thromboembolic disease. Weight loss before pregnancy reduces these risks [15, 65–69]•STD. Chlamydia trachomatis and Neisseria gonorrhea have been strongly associated with ectopic pregnancy, infertility, and chronic pelvic pain. STDs during pregnancy might result in fetal death or substantial physical and developmental disabilities, including mental retardation and blindness [70, 71]•Hepatitis B. Vaccination is recommended for men and women who are at risk for acquiring hepatitis B virus (HBV) infection. Preventing HBV infection in women of childbearing age prevents vertical transmission of infection to infants and eliminates risk for infection and sequelae, including hepatic failure, liver carcinoma, cirrhosis, and death [72–74] However, the best evidence for the effectiveness of specific components of preconception care is seen when the focus was on a single intervention and often not in the context of improving pregnancy outcomes. For example, effectiveness documented by the U.S. Preventive Services Task Force for interventions related to smoking, alcohol misuse, and obesity are based on studies of interventions delivered in primary care settings that was not complicated by the additional delivery of other components of preconception care [56, 61, 66]. One study reported the effectiveness of comprehensive preconception care but inferences for the United States are limited because the setting of intervention delivery was in Hungary [75]. Priorities for preconception care and preconception health The immediate priority is to ensure that evidence-based preconception interventions are implemented to further improve infant and maternal pregnancy outcomes. Many national professional organizations have developed and published CPGs for specific components of preconception care. For example, the American Diabetes Association developed CPGs to be started before pregnancy for women with preconception diabetes [28]. The American Association of Clinical Endocrinologists has developed preconception CPGs for women with hypothyroidism [34]. CPGs have also been developed for women being treated with teratogenic medications to guide the transition to safer medications. CPGs for women using antiepileptic drugs or oral anticoagulants have been developed by the American Academy of Neurology [50], the American Heart Association, and the American College of Cardiologists, respectively [44, 46]. Guidelines, recommendations, and strategies for reducing alcohol misuse and prevention of fetal alcohol syndrome have been developed and published by the U.S. Public Health Service, U.S. Preventive Service Taskforce, ACOG, and the National Taskforce on Fetal Alcohol Syndrome and Fetal Alcohol Effects [55–59]. However, moving forward towards more universal preconception care is not without its challenges. There is no national policy, nor do standard tools exist, for the delivery of these services. Some have suggested that challenges could be overcome by promoting preconception health to all women of reproductive age at each and every encounter with the health care system. The generally recommended mechanisms would include risk assessment (screening), health promotion (education and counseling), and intervention or referral. However, the existing recommendations to provide these three essential components have not been translated into practical tools for action and decision-making in the clinical setting. Stated another way, primary care providers do not have the tools they need to implement what is known to work for improving preconception health. Challenges and opportunities Whereas clinical practice and the promotion of preconception care services could potentially be shaped by guidelines, widespread adoption of the guidelines requires a more active approach [76, 77]. The literature on diffusion of innovation and translation of research into practice clearly indicates that changing primary care providers’ knowledge, attitudes, and practices requires multiple steps including consolidation and distillation of guidelines, active organizational support, clinically relevant decision tools, use of quality improvement techniques, and performance monitoring [78, 79]. A growing body of research indicates that providers are more likely to engage in evidence-based practices following participation in quality improvement projects (e.g., rapid improvement cycles or collaborative groups). The federally funded Health Disparities Collaborative, the Perinatal Care and Patient Safety Collaboratives, and similar quality improvement efforts for community health centers demonstrate how such approaches are being used to better serve low-income and uninsured patients. The need to develop innovative approaches to deliver and finance bundles of preconception interventions must be viewed as a high priority. However, although strong evidence exists for many preconception care interventions and shows they are effective, discussion continues regarding the benefits of delivering comprehensive preconception care [16, 80, 81], and the effectiveness or added value of “packaging” or “bundling” these interventions remains to be determined. For example, preconception care interventions could potentially be bundled into behavioral modification services (e.g., smoking and alcohol cessation, weight control, exercise), screening and assessment (e.g., vaccines, general physical examination), and specialized care (e.g., management of diabetes and hypertension). Alternatively, interventions could be bundled into visits (e.g., do x at each well-woman visit, do y every 2 years, do z only for women at risk). Some evidence already suggests that bundling could be of added value. For example, an evaluation of the National Centers of Excellence in Women's Health found that, compared with other settings, women received better-quality primary health care at those centers having better integrated and coordinated services [82]. However, more such research, focused specifically on integration of preconception care components into primary care, is clearly warranted. Efforts to promote healthy behavior and improve service delivery should be accompanied by improvements in health care coverage and financing for women of childbearing age. Affordability of care is a major concern for many women [83], and more can be done to improve access to preconception care. Many women under age 65 are uninsured, including 40% of poor women, one third of near-poor women (with income between 100% and 200% of the federal poverty level), 50% of women with disabilities, and 29% of young women ages 19–24 years [84]. Expanding health care coverage for low-income women, through public programs such as Medicaid, Medicaid waivers, the State Children's Health Insurance Program (SCHIP), is particularly important. Medicaid offers particularly important opportunities. In 2003, 12% of all women of childbearing age (15–44 years) and 37% of poor women in that age group relied on Medicaid for health care coverage [83]. Many low-income women do not, however, qualify for Medicaid because they do not have children under age 18, are not over age 65, or are undocumented. As states seek to expand Medicaid coverage to low-income, uninsured adults, women of childbearing age should receive high priority. Although two thirds of women in Medicaid are of childbearing age [19–44, 85], many additional low-income women do not qualify for Medicaid benefits under current state eligibility rules unless they are pregnant. Without coverage before and between pregnancies, low-income women typically miss preventive visits [83]. Recognizing these gaps and missed opportunities for prevention, 21 states have developed special programs (under federal Medicaid waivers) that cover some women who do not otherwise qualify for Medicaid, particularly to offer family planning and interconception care. Specifically, these states extend eligibility to women who lose coverage after the birth of a baby or starting a job, while other states offer family planning coverage based on income status to men and/or women [86]. An evaluation of “family planning waiver” projects prepared for the federal Center for Medicare and Medicaid Services (CMS) found that they resulted in significant savings to both the federal and state governments [87]. Greater potential savings and prevention, however, could undoubtedly result if states offered more preconception risk screening, health promotion, and interventions; to do so, however, states need permission from the federal government to include such services in their waivers or need for Congress to approve interconception care as an optional benefit. Finally, special attention should be given to the continuing, and sometimes increasing, racial gap in poor outcomes, especially in terms of access to services, and quality of care. These priorities can potentially complicate decision-making for public health policy. The best evidence of intervention effectiveness is often based on studies of health care systems. However, in developing programs and policies for the delivery of preconception services, one must consider the fact that poor access to health care can be a major determinant of health disparities. Preconception care: the next maternal and child health frontier The significant improvements in infant and maternal pregnancy outcomes during the past 40 years have resulted largely from the intensified focus on family planning in the 1960s and 1970s, on child health in the 1970s and 1980s, and on prenatal care in the 1980s and 1990s. Over the past 40 years, many public health and clinical care providers and their organizations have also recognized that, although it is important to offer family planning services to prevent unplanned pregnancies, it is not enough to worry about a pregnant woman's health only during her pregnancy or a child's health only after its birth, because many of the factors that contribute to the mother's and child's health can be identified and often successfully managed before pregnancy. Today, the greatest opportunities for further improvement in pregnancy outcomes—in improving the health of women and their children—lie in prevention strategies that must be implemented prior to conception to be effective. In 1989, the U.S. Public Health Service Expert Panel identified preconception care as an essential part of prenatal care [88]. National organizations of health professionals, such as the AAP, ACOG, and the American Academy of Family Physicians (AAFP), all recommend specific components of preconception care that should be delivered within the context of health care systems [15, 89]. The time has come to move forward to the next maternal and child health frontier of “prevention” by acting on the recommendations of professional organizations and implementing scientifically proven interventions to further improve pregnancy outcomes. Action is urgently required in an array of clinical settings, in health promotion campaigns, and in academic research settings. Because of the wide range of interventions included under the umbrella of preconception care, such interventions may be delivered in both primary care and specialty care practices. However, opportunities also exist to improve preconception health through wellness care, through maintenance care for women with chronic health conditions associated with increased preconception risk (e.g., maternal diabetes), and in settings where women seek medical support for one specific risk such as smoking or obesity. In addition, more research is needed to understand how a women's motivation for healthy behavior can be used to initiate the reduction of a number of preconception risks. For example, comparison of concurrent versus sequential interventions for smoking and weight control requires additional clarification [90]. Above all, health care providers and health agencies must understand that early prenatal care is too late for many women and babies. In particular, some evidence-based interventions recommended for implementation during pregnancy could be more effective and beneficial if implemented before conception. For example, although smoking cessation is recommended for all pregnant women, more than one attempt to stop may be necessary to achieve success. Only about 20% of pregnant women who smoke are able to stop smoking during pregnancy [90]. Starting smoking cessation treatment — and successfully completing it—should be recommended before pregnancy is begun. Similarly, the only definitive way of preventing fetal alcohol syndrome is to stop drinking even before pregnancy. In the same manner, HIV testing and genetic counseling are recommended during pregnancy, but additional options are available to women who use these interventions prior to conception, including the decision not to become pregnant. Conclusion There is ample evidence that individual elements of preconception care work. Equally important, however, is the evidence that some interventions work best and others only work if provided before pregnancy. Given these realities, to further improve perinatal outcomes, the United States must commit to improving preconception health and to providing preconception care to all women of reproductive age. Even though sufficient knowledge and evidence is present to take action now, further efforts are needed to identify best practices and the most effective means of delivering integrated preconception services.	[ "preconception care", "perinatal outcomes", "women's health" ]	[ "P", "P", "P" ]
Pflugers_Arch-3-1-2082651	Criteria for arrhythmogenicity in genetically-modified Langendorff-perfused murine hearts modelling the congenital long QT syndrome type 3 and the Brugada syndrome	The experiments investigated the applicability of two established criteria for arrhythmogenicity in Scn5a+/Δ and Scn5a+/− murine hearts modelling the congenital long QT syndrome type 3 (LQT3) and the Brugada syndrome (BrS). Monophasic action potentials (APs) recorded during extrasystolic stimulation procedures from Langendorff-perfused control hearts and hearts treated with flecainide (1 μM) or quinidine (1 or 10 μM) demonstrated that both agents were pro-arrhythmic in wild-type (WT) hearts, quinidine was pro-arrhythmic in Scn5a+/Δ hearts, and that flecainide was pro-arrhythmic whereas quinidine was anti-arrhythmic in Scn5a+/− hearts, confirming clinical findings. Statistical analysis confirmed a quadratic relationship between epicardial and endocardial AP durations (APDs) in WT control hearts. However, comparisons between plots of epicardial against endocardial APDs and this reference curve failed to correlate with arrhythmogenicity. Restitution curves, relating APD to diastolic interval (DI), were then constructed for the first time in a murine system and mono-exponential growth functions fitted to these curves. Significant (P < 0.05) alterations in the DI at which slopes equalled unity, an established indicator of arrhythmogenicity, now successfully predicted the presence or absence of arrhythmogenicity in all cases. We thus associate changes in the slopes of restitution curves with arrhythmogenicity in models of LQT3 and BrS. Introduction Mutations in the SCN5a gene encoding the α-subunit of the cardiac voltage-gated Na+ channel are well established to underlie hereditary arrhythmic syndromes which may result in sudden cardiac death. On the one hand, gain-of-function mutations which increase inward depolarising currents and consequently prolong the ventricular action potential are associated with arrhythmogenicity in the congenital long QT syndrome type 3 (LQT3) [23]. On the other hand, loss-of-function mutations which decrease such currents are associated with arrhythmogenicity in the Brugada syndrome (BrS) [1]. In this case, spatial heterogeneities in the degree of shortening of the ventricular action potential result in characteristic electrocardiographic changes, namely a partial right bundle branch block pattern with ST segment elevation in the right precordial leads [2]. Despite the similar clinical outcomes of LQT3 and BrS, class 1A and 1C, anti-arrhythmic agents, exemplified by quinidine and flecainide respectively, exert contrasting effects on arrhythmogenicity in these two syndromes [31]. Thus in LQT3, quinidine is pro-arrhythmic [33] and flecainide is anti-arrhythmic [22]. In contrast, in BrS quinidine is anti-arrhythmic [24], while flecainide is pro-arrhythmic [7] and used to unmask the arrhythmic phenotype in asymptomatic patients. These contrasts are precisely reflected in results from our murine models of LQT3 [39] and BrS [38], emphasizing the translatability of results obtained from such systems. Furthermore, both quinidine and flecainide are known to have the potential for pro-arrhythmic effects [12, 30]. These similarities and differences have parallels in the physiological changes thought to underlie arrhythmogenicity in LQT3 and BrS. Thus circus-type re-entry [2], requiring both an initiating trigger and a re-entrant substrate, has been implicated in arrhythmogenicity in both these conditions. Of these requirements, both conditions are associated with triggering extrasystoles [3]. In LQT3, these are initiated by early after-depolarizations attributable to action potential prolongation [42], while in BrS they are initiated by phase 2 re-entry [20, 46]. Furthermore, in both LQT3 and BrS, arrhythmia occurs against a re-entrant substrate [3]. In LQT3 this is provided by increased transmural dispersion of repolarization [13, 27, 42]. While increased transmural dispersion of repolarization has also been demonstrated in some studies on BrS [3], increased epicardial dispersion of repolarization [2] and slowed conduction [29] are also implicated. Thus studies in murine hearts have associated alterations in transmural dispersion of repolarization, reflected in relationships between epicardial and endocardial action potential durations, with arrhythmogenicity in models of LQT3 but failed to do so in models of BrS [21, 33, 38, 39, 42]. Arrhythmogenicity has often been associated with increases in the slopes of restitution curves, plotting action potential duration against the preceding diastolic interval, to greater than unity [14, 17–19, 28, 40]. Possible mechanisms that might underlie this association have been discussed in detail on a number of occasions [11, 25, 43, 44]. Furthermore, decreases in the slopes of such curves are associated with the anti-arrhythmic effects of several clinically important agents including amiodarone, bretylium and procainamide [11, 26, 32]. Such associations have previously been demonstrated in a pharmacological model of LQT2 [45] but has not been assessed in LQT3 or in BrS. The present study compares an analysis examining transmural dispersion of repolarization, assessed by comparing epicardial and endocardial action potential durations obtained over a range of heart rates and previously established to predict arrhythmogenicity in hypokalaemic Langendorff-perfused murine hearts [34], with another examining effects on restitution curve parameters. This represents the first occasion on which restitution curves have been constructed in a murine system, in relation to BrS or LQT3, or indeed in any genetic model of an hereditary arrhythmic syndrome. To this end, the monophasic action potential (MAP) technique and an established Langendorff-perfused system [15, 34–36] are used to study genetically modified mice modelling LQT3 (Scn5a+/Δ) [13] and BrS (Scn5a+/−) [29] before and after exposure to quinidine and flecainide. We proceed to demonstrate that while the former criterion fails, the latter produces predictions in full agreement with the arrhythmogenic findings in every case and thus suggest that alterations in the slopes of restitution curves are associated with arrhythmogenicity in both LQT3 and BrS. Materials and methods Experimental animals Wild-type (WT), Scn5a+/Δ and Scn5a+/− 129 Sv mice aged 3–6 months were housed at 21 ± 1°C with 12 h light/dark cycles. They were fed sterile chow (RM3 Maintenance Diet, SDS, Witham, Essex, UK) and had free access to water. All procedures complied with the UK Animals (Scientific Procedures) Act 1986. Solutions Solutions were based on bicarbonate-buffered Krebs-Henseleit solution (119 mM NaCl, 25 mM NaHCO3, 4 mM KCl, 1.2 mM KH2PO4, 1 mM MgCl2, 1.8 mM CaCl2, 10 mM glucose and 2 mM Na-pyruvate; pH adjusted to 7.4) and were bubbled with 95% O2/5% CO2 (British Oxygen Company, Manchester, U.K). Drug-containing solutions were prepared by adding flecainide (Sigma-Aldrich, Poole, UK) to final concentration of 1 μM and quinidine (Sigma-Aldrich, Poole, UK) to final concentrations of 1 and 10 μM: these concentrations were chosen to reflect clinically therapeutic values (effective free and total plasma concentrations of 0.79 and 1.48 μM, respectively, for flecainide and 0.74 and 7.9 μM, respectively, for quinidine) [9]. Preparation A Langendorff-perfusion protocol previously adapted for murine hearts [4] was used. Briefly, mice were killed by cervical dislocation [Schedule 1: UK Animals (Scientific Procedures) Act 1986] before hearts were quickly excised and placed in ice-cold bicarbonate-buffered Krebs-Henseleit solution. Short sections of aorta were cannulated under the surface of the solution and attached to a custom-made 21-gauge cannula filled using an aneurysm clip (Harvard Apparatus, Edenbridge, Kent, UK), within 90 s of excision. Fresh Krebs-Henseleit solution was passed through 200- and 5-μm filters (Millipore, Watford, UK) and warmed to 37°C using a water jacket and circulator (Techne model C-85A, Cambridge, UK). This solution was then used for constant-flow retrograde perfusion, thereby circumventing any changes in coronary vascular resistance and therefore flow that might otherwise result from the effects of the drugs studied [10], at 2–2.5 ml/min, driven by a peristaltic pump (Watson-Marlow Bredel model 505S, Falmouth, Cornwall, UK). Hearts were regarded as suitable for experimentation if they regained a healthy pink colour and began to contract spontaneously on re-warming. Electrophysiological measurements An epicardial MAP electrode (Hugo Sachs, Harvard Apparatus, UK) was placed against the basal left ventricular epicardium. A custom-made endocardial MAP electrode consisting of two twisted strands of high-purity Teflon-coated silver wire of 0.25 mm diameter (Advent Research Materials, UK) was manufactured. A small access window was created in the interventricular septum to allow access to the left ventricular endocardium [8]. The Teflon coating was removed from the distal 1 mm of the electrode; this was then galvanically chlorided to eliminate DC offset, inserted and placed against the septal endocardial surface. MAPs were amplified, band-pass filtered (0.5 Hz to 1 kHz: Gould 2400S, Gould-Nicolet Technologies, Ilford, Essex, UK) and digitised at a sampling frequency of 5 kHz (micro1401, Cambridge Electronic Design, Cambridge, UK). Analysis of MAPs was performed using Spike II (Cambridge Electronic Design, Cambridge, UK). Experimental protocol A bipolar platinum stimulating electrode (1 mm inter-pole spacing) was placed on the basal surface of the right ventricular epicardium. Square-wave stimuli (Grass S48 stimulator, Grass-Telefactor, Slough, UK) of 2 ms duration and amplitudes of twice the excitation threshold were initially applied to hearts at a constant baseline cycle length (BCL) of 125 ms until MAPs showed stable baselines, rapid upstroke phases that reached consistent amplitudes and smooth repolarization phases [16] and for at least 10 min. Before subsequent recordings were made, hearts were then exposed to test solutions for 20 min while stimulation was continued. In initial experiments, MAPs were recorded during regular stimulation at a BCL of 125 ms. Arrhythmogenicity was then assessed both during regular stimulation and using an extrasystolic stimulation protocol [37] previously adapted for the study of arrhythmogenesis in murine hearts incorporating the Scn5a+/Δ and Scn5a+/− mutations [13, 38] and described below. Hearts were then subjected to an adapted dynamic pacing protocol [17] explained in detail below, and the data obtained used to construct restitution curves. Data analysis All data are presented as means ± standard errors of the means and include the number of hearts studied. Comparisons between data sets used analysis of variance (significance threshold set at P ≤ 0.05). Curve fitting of particular functions to data sets used a Levenberg-Marquardt algorithm (OriginPro 7.5, OriginLab, MA, USA). Results The experiments made direct measurements of MAPs from intact Langendorff-perfused murine hearts. They compared the extents to which relationships between epicardial and endocardial action potential durations and the slopes of restitution curves are associated with arrhythmogenicity in genetic models of the congenital LQT3 (Scn5a+/Δ) and of the BrS (Scn5a+/−), and in wild-type (WT) controls. They began by successfully demonstrating that these experimental models recapitulated arrhythmic phenotypes observed clinically as well as confirming the anticipated effects of treatment with clinically-relevant [6] concentrations of flecainide (1 μM) and quinidine (1 or 10 μM). Hearts were exposed to solutions for 20 min while stimulated at a constant BCL of 125 ms before recording was commenced. Arrhythmogenicity in WT, Scn5a+/Δ and Scn5a+/− hearts subject to both regular and extrasystolic stimulation recapitulates clinical observations Figures 1 and 2 show results obtained from WT (Figs. 1a, 2a), Scn5a+/Δ (Figs. 1b, 2b) and Scn5a+/− (Figs. 1c, 2c) hearts treated with control solution (A) and with solutions containing 1 μM flecainide (B), 1 μM quinidine (C) and 10 μM quinidine (D). Figure 1 illustrates MAP waveforms recorded from the epicardia of hearts during the extrasystolic stimulation procedure showing arrhythmic waveforms in situations where arrhythmic activity was observed in >50% of cases; Fig. 2 quantifies the incidence of such arrhythmic activity both during regular stimulation (i) and during the extrasystolic stimulation procedure (ii). Fig. 1Arrhythmic activity resulting from extrasystolic stimulation. Epicardial monophasic action potential (MAP) recordings illustrating the result of the application of extrasystolic stimuli to WT (a), Scn5a+/Δ (b) and Scn5a+/− (c) hearts treated with control solution (A) and with solutions containing 1 μM flecainide (B), 1 μM quinidine (C) and 10 μM quinidine (D). Traces show arrhythmic activity where this was observed in ≥50% of casesFig. 2Incidences of arrhythmic activity in WT, Scn5a+/Δ and Scn5a+/− hearts subjected to the extrasystolic stimulation procedure. Percentage of WT (a), Scn5a+/Δ (b) and Scn5a+/− (c) hearts demonstrating arrhythmic activity during regular stimulation (i) and during the extrasystolic stimulation procedure (ii). In both these cases, MAPs were obtained during treatment with control solution (A) and with solutions containing 1 μM flecainide (B), 1 μM quinidine (C) and 10 μM quinidine (D). Hashing indicates hearts demonstrating arrhythmic activity, and dots indicate hearts not demonstrating such activity The frequencies of arrhythmic phenomena in hearts subjected to regular stimulation at a BCL of 125 ms were relatively low irrespective of the genetic or pharmacological conditions but were nevertheless consistent with clinical expectations. First, zero out of seven WT hearts (~0%, Fig. 2a, i) exposed to control solution (A) demonstrated arrhythmic activity, while treatment with flecainide (B) led to arrhythmic activity in only one out of six (~17%) cases. Furthermore, such hearts never showed arrhythmic activity during treatment with 1 μM quinidine (C, zero out of six hearts, ~0%) and again did so in only one out of six cases (~17%) with 10 μM quinidine (D). Secondly, arrhythmic activity did occur in one out of six (~17%) Scn5a+/Δ hearts (Fig. 2b) treated with control solution (A) and was absent in all cases (zero out of six hearts, ~0%) during treatment with flecainide (B). Arrhythmic activity was not observed (zero out of four hearts, ~0%) during treatment with 1 μM quinidine (C) but did occur in one out of five hearts (~20%) exposed to 10 μM quinidine (D). Thirdly, two out of nine Scn5a+/− hearts (~22%, Fig. 2c) exhibited arrhythmic activity during exposure to control solution (A). Such arrhythmic activity was also observed in one out of five hearts (~20%) treated with flecainide (B). Arrhythmic activity was then absent after treatment with either 1 μM (C) or 10 μM (D) quinidine (zero out of five, 0%, and zero out of six, 0%, hearts, respectively). In contrast, extrasystolic stimulation (Figs. 1 and 2, ii) revealed clearer trends in arrhythmogenicity in full agreement with clinical expectations. First, while zero out of seven WT hearts (0%, Fig. 2a, ii) exposed to control solution (A) exhibited arrhythmic activity, treatment with flecainide (B) increased the incidence of such events to two out of six hearts (~33%). Furthermore, exposure to 1 μM quinidine (C) resulted in arrhythmic activity in four out of eight hearts (~50%), while exposure to 10 μM quinidine (D) resulted in arrhythmic activity in three out of six hearts (~50%). Secondly, one out of five Scn5a+/Δ hearts (~20%, Fig. 2b, ii) demonstrated arrhythmic activity during exposure to either control solution (A) or flecainide (B). However, exposure to quinidine resulted in arrhythmic activity in three out of five hearts (~60%) at 1 μM (C) and four out of four hearts (~100%) at 10 μM (D). Thirdly, Scn5a+/− hearts (Fig. 2c, ii) demonstrated such arrhythmic activity in three out of nine cases (~33%) during treatment with control solution (A) but in five out of six cases (~83%) during treatment with flecainide (B). Furthermore, arrhythmic activity was observed in only one out of six such hearts (~17%) during treatment with 1 μM quinidine (C) and in zero out of six hearts during exposure to 10 μM quinidine (D). Graphical analyses of arrhythmogenic tendency In an effort to explore the possible basis for these arrhythmic phenomena, subsequent experiments made systematic measurements of epicardial and endocardial action potential durations (at 90% repolarization, APD90s) during a dynamic pacing protocol [17]. This consisted of cycles each comprising 100 stimuli, initially delivered at a BCL of 175 ms. BCL was then decremented in 50-ms steps with each subsequent cycle until a reproducible sequence of consistently shaped MAP waveforms was no longer obtained. Steady states were consistently reached within the first 50 responses, and therefore mean values of APD90 were obtained from the final 50 action potentials in each cycle. This approach provided data that could be used to evaluate over the full range of BCLs the relationship between arrhythmogenicity and two possible electrophysiological properties. Thus it was possible first to evaluate transmural repolarization gradients and secondly to evaluate slopes of restitution curves relating APD90 to the preceding DI, given by BCL minus APD90 [17, 25, 28, 43]. Relationships between epicardial and endocardial action potential durations are not consistently associated with either the presence or absence of arrhythmogenicity Alterations in relationship between epicardial and endocardial APD90 determined at a single BCL has previously been associated with arrhythmogenicity in hearts obtained from our Scn5a+/Δ mice [42]. The present study now extends this analysis from a single BCL to a range of BCLs and applies it to Scn5a+/− hearts for the first time, plotting endocardial APD90 on the abscissa and epicardial APD90 on the ordinate over the full range of BCLs investigated (triangles, Fig. 3). Such plots have previously been used to assess arrhythmogenicity in hypokalaemic WT murine hearts [34] and provided possible criteria for arrhythmogenicity. The control plot obtained from WT hearts perfused with control solution (six hearts, Fig. 3a, A) gave a quadratic relationship between endocardial and epicardial APD90 (solid line), in agreement with previous work [34]. Thus the curve y = ax2 + bx + c, where a = 0.0097 ± 0.0028, b = 1.46 ± 0.27 and c = −5.06 ± 6.34, gave a reduced χ2 of 0.49. Repeating the regression analysis this time fitting the line y = mx + c gave a reduced χ2 of 0.86, and permitted the derivation of an F statistic for the appropriate, N + 1 = 17 data points, given by in which the subscripts on the right-hand side of the equation denote the number of parameters involved [5]. This yielded a probability of exceeding F of <0.05, suggesting that inclusion of the quadratic term was justified, again in full agreement with the previous study [34]. Fig. 3Relationships between epicardial and endocardial action potential durations (at 90% repolarization, APD90s) obtained over a range of baseline cycle lengths. WT (a), Scn5a+/Δ (b) and Scn5a+/− (c) hearts were exposed to control solution (A) and to solutions containing 1 μM flecainide (B), 1 μM quinidine (C) and 10 μM quinidine (D) and stimulated at baseline cycle lengths of between 80 ms and 175 ms. Broken lines indicate reference curves constructed by least-squares fitting of the equation y = ax2 + bx + c to data points obtained from wild-type hearts treated with control solution Data points obtained under the range of genetic and pharmacological conditions studied could then be compared with this reference curve (broken lines, Fig. 3) with the expectation that the departure of points to the left of this curve might be associated with arrhythmogenicity [34]. However, data sets obtained before and after the pharmacological manoeuvres satisfied this expectation in WT but clearly failed to do so in the remaining genetically modified hearts. First, treatment of WT hearts (Fig. 3a) with flecainide (five hearts, B), 1 μM quinidine (six hearts, C) and 10 μM quinidine (six hearts, D) all resulted in points falling to the left of the reference curve, in parallel with the occurrence of arrhythmic activity under these conditions (Fig. 2a, B, C and D). Secondly, data points obtained from Scn5a+/Δ hearts (Fig. 3b) treated with control solution (five hearts, A) again fell to the left of the reference curve in contrast with the low incidence of arrhythmic activity in this setting (Fig. 2b, A). This pattern was maintained when such hearts were exposed to flecainide (five hearts, Fig. 3b, B), once again in contrast with the low frequency which arrhythmic activity was observed under these conditions (Fig. 2b, B). Exposure to 1 μM quinidine (five hearts, Fig. 3b, C) and 10 μM quinidine (six hearts, D) again both resulted in points falling to the left of the control curve, this time in parallel with the high incidence of arrhythmic activity in these settings (Figs. 2b, C and 3b, D). Thirdly, data points obtained from Scn5a+/− hearts (Fig. 3c) treated with control solution (six hearts, A) consistently fell on the reference curve, in contrast with the occurrence of arrhythmic activity in this setting (Fig. 2c, A). Furthermore, treatment with flecainide (six hearts, Fig. 3c, B) failed to result in a left-shift of data points, in contrast with the large increase in the incidence of arrhythmic activity (Fig. 2c, B). In addition, points obtained during treatment with either 1 μM quinidine (five hearts, Fig. 3c, C) or 10 μM quinidine (seven hearts, D) again fell to the left of the reference curve, in contrast with the anti-arrhythmic effect of this agent demonstrated in Fig. 2c, C and D. Alterations in the slopes of restitution curves correlate with the presence and absence of arrhythmic activity Restitution curves (data points, left ordinate) were then plotted using the APD90 (left ordinate) and DI data obtained during the dynamic pacing protocol from the epicardia (circles, Fig. 4) and endocardia (squares, Fig. 5) of WT (Figs. 4a, 5a), Scn5a+/Δ (Figs. 4b, 5b) and Scn5a+/− (Figs. 4c, 5c) hearts for the first time. Hearts were studied during treatment with control solution (A) and with solutions containing flecainide (B), 1 μM quinidine (C) or 10 μM quinidine (D). Under all conditions, epicardial and endocardial APD90 both decreased as DI decreased. Increases or decreases in the slopes of such relationships were in full agreement with the presence or absence of arrhythmic activity under all conditions studied. Fig. 4Epicardial restitution curves plotting action potential duration (at 90% repolarization, APD90) against preceding diastolic interval (DI). Data were obtained from the epicardia (circles) of WT (a), Scn5a+/Δ (b) and Scn5a+/− (c) hearts exposed to control solution (A) and to solutions containing 1 μM flecainide (B), 1 μM quinidine (C) and 10 μM quinidine (D). Curves are fitted with mono-exponential growth functions obtained by least squares fitting to the experimental values of APD90 and DI (solid lines, left ordinates). Gradients were obtained by differentiation of the fitted functions (broken lines, right axes). Shaded boxes indicate ranges of DI values at which such gradients exceed unityFig. 5Endocardial restitution curves plotting action potential duration (at 90% repolarization, APD90) against preceding diastolic interval (DI). Data were obtained from the endocardia (squares) of WT (a), Scn5a+/Δ (b) and Scn5a+/− (c) hearts exposed to control solution (A) and to solutions containing 1 μM flecainide (B), 1 μM quinidine (C) and 10 μM quinidine (D). Curves are fitted with mono-exponential growth functions obtained by least squares fitting to the experimental values of APD90 and DI (solid lines, left ordinates). Gradients were obtained by differentiation of the fitted functions (broken lines, right axes). Shaded boxes indicate ranges of DI values at which such gradients exceed unity Restitution curves yield parameters predictive of arrhythmogenicity To permit quantitative analysis, simple mono-exponential growth functions, as used previously [28], were fitted to the experimental data. These took the form: where y represents APD90, x represents DI, and y0, A and τ are constants obtained by least squares fitting to the experimental values of APD90 and DI in each case. These fitted parameters, along with corresponding reduced χ2 values, are shown in Tables 1 (epicardium) and 2 (endocardium). Curve fits performed on data sets obtained by separating successive odd- and even-numbered action potentials yielded statistically indistinguishable (P > 0.05) parameters in all cases (Tables 3 and 4) thus eliminating the possibility of error arising as a result of averaging. The corresponding gradients were then given by: and take their largest values at the shortest BCLs studied. Table 1Fit parameters for epicardial restitution curvesConditiony0 (ms)A (ms)τ (ms)χ2WT (6 hearts)1.6 ± 4.456.2 ± 2.458.4 ± 9.81.4WT + flecainide (5 hearts)−16.3 ± 9.4104.7 ± 7.182.8 ± 28.05.9WT + 1 μM quinidine (6 hearts)−38.3 ± 17.492.4 ± 14.745.1 ± 9.42.5WT + 10 μM quinidine (6 hearts)−863.0 ± 1,387.1981.6 ± 1,369.127.1 ± 15.9113.3Scn5a+/Δ (5 hearts)−19.6 ± 7.290.0 ± 4.453.1 ± 10.714.0Scn5a+/Δ + flecainide (5 hearts)−16.1 ± 6.582.2 ± 12.395.3 ± 42.13.5Scn5a+/Δ + 1 μM quinidine (5 hearts)−677.1 ± 306.1748.0 ± 304.715.2 ± 2.210.0Scn5a+/− (6 hearts)−674.8 ± 329.9729.4 ± 329.114.2 ± 2.18.4Scn5a+/− + flecainide (6 hearts)−56,359.8 ± 2,559.752,677.3 ± 2.27.8 ± 2.514.8Scn5a+/− + 1 μM quinidine (5 hearts)−118.3 ± 57.4202.5 ± 49.436.4 ± 10.516.7Scn5a+/− + 10 μM quinidine (7 hearts)−1.7 ± 44.7131.9 ± 62.3158.4 ± 262.210.8Results of least-squares fits to the function under the conditions indicatedTable 2Fit parameters for endocardial restitution curvesConditiony0 (ms)A (ms)τ (ms)χ2WT (6 hearts)−11.8 ± 3.773.8 ± 1.765.2 ± 8.41.2WT + flecainide (5 hearts)−769.0 ± 244.1827.3 ± 243.714.5 ± 1.31.4WT + 1 μM quinidine (6 hearts)−65.8 ± 100.4113.5 ± 98.324.9 ± 12.011.7WT + 10 μM quinidine (6 hearts)−41.9 ± 71.578.2 ± 68.732.8 ± 19.313.6Scn5a+/Δ (5 hearts)−31.9 ± 9.0109.9 ± 5.149.6 ± 9.010.3Scn5a+/Δ + flecainide (5 hearts)−561.1 ± 1.2631.6 ± 15.118.7 ± 0.36.7Scn5a+/Δ + 1 μM quinidine (5 hearts)−499.7 ± 143.8563.0 ± 142.916.3 ± 1.656.0Scn5a+/− (6 hearts)−281.7 ± 9.843.5 ± 6.448.5 ± 19.04.3Scn5a+/− + flecainide (6 hearts)−1,050.6 ± 1,558.51,100.1 ± 1,557.811.7 ± 4.010.0Scn5a+/− + 1 μM quinidine (5 hearts)−261.2 ± 117.9350.5 ± 107.834.2 ± 9.321.7Scn5a+/− + 10 μM quinidine (7 hearts)−220.1 ± 322.0319.1 ± 316.826.2 ± 12.745.5Results of least-squares fits to the function under the conditions indicatedTable 3Fit parameters for endocardial restitution curves separating successive odd- and even-numbered action potentialsConditionOddEveny0 (ms)A (ms)τ (ms)χ2y0 (ms)A (ms)τ (ms)χ2WT (6 hearts)13.4 ± 7.640.5 ± 12.950.3 ± 21.46.38.1 ± 11.258.3 ± 25.448.2 ± 32.34.6WT + flecainide (5 hearts)−14.7 ± 8.9105.2 ± 9.088.0 ± 31.35.8−44.6 ± 18.5117.4 ± 14.249.0 ± 9.91.5WT + 1 μM quinidine (6 hearts)−20.6 ± 9.975.8 ± 6.256.4 ± 13.23.8−34.2 ± 11.789.2 ± 9.345.9 ± 7.83.3WT + 10 μM quinidine (6 hearts)−1,391.7 ± 3,195.91,496.0 ± 3,182.021.1 ± 14.3117.9−393.1 ± 5.3.0532.8 ± 481.741.1 ± 27.6115.7Scn5a+/Δ (5 hearts)−23.9 ± 8.392.0 ± 5.347.8 ± 9.012.4−24.4 ± 8.692.6 ± 5.647.6 ± 8.911.8Scn5a+/Δ + flecainide (5 hearts)14.6 ± 6.377.1 ± 8.484.1 ± 32.23.4−33.2 ± 46.4110.1 ± 41.936.4 ± 13.12.0Scn5a+/Δ + 1 μM quinidine (5 hearts)−250.2 ± 12.8483.0 ± 24.518.8 ± 7.812.4−254.7 ± 321.2328.1 ± 318.320.0 ± 7.75.3Scn5a+/− (6 hearts)−850.4 ± 282.7904.2 ± 282.113.5 ± 1.34.3−913.3 ± 283.3968.6 ± 282.713.7 ± 1.23.7Scn5a+/− + flecainide (6 hearts)−1,473.8 ± 3,144.51,534.3 ± 2,141.214.7 ± 7.634.3−2,108.8 ± 4,489.92,171.3 ± 4,486.814.0 ± 6.735.0Scn5a+/− + 1 μM quinidine (5 hearts)−46.8 ± 51.2146.8 ± 21.359.5 ± 43.59.4−46.7 ± 50.6149.2 ± 20.261.0 ± 44.29.3Scn5a+/− + 10 μM quinidine (7 hearts)7.3 ± 34.2167.7 ± 264.8272.8 ± 78.310.82.8 ± 44.6143.1 ± 121.3192.2 ± 405.812.0Results of least-squares fits to the function under the conditions indicatedTable 4Fit parameters for endocardial restitution curves separating successive odd- and even-numbered action potentialsConditionOddEveny0 (ms)A (ms)τ (ms)χ2y0 (ms)A (ms)τ (ms)χ2WT (6 hearts)−17.4 ± 5.373.8 ± 2.867.3 ± 13.54.7−38.9 ± 15.491.6 ± 13.046.5 ± 8.51.8WT + flecainide (5 hearts)−49.2 ± 26.2113.8 ± 23.035.7 ± 8.27.6−367.3 ± 171.3426.9 ± 170.617.7 ± 2.41.3WT + 1 μM quinidine (6 hearts)−393.2 ± 368.6437.7 ± 368.614.9 ± 4.318.1−1,337.3 ± 1,206.41,384.5 ± 1,205.711.5 ± 2.514.7WT + 10 μM quinidine (6 hearts)−17.7 + −39.053.2 ± 35.042.6 ± 28.911.2−28.0 ± 62.766.6 ± 58.239.6 ± 29.717.3Scn5a+/Δ (5 hearts)−149.6 ± 131.4221.0 ± 128.626.4 ± 7.60.9−42.3 ± 10.5115.3 ± 8.240.8 ± 5.02.6Scn5a+/Δ + flecainide (5 hearts)−1,133.8 ± 784.21,201.4 ± 783.115.2 ± 2.62.0−2,164.0 ± 1,195.72,233.3 ± 1,195.013.6 ± 1.61.4Scn5a+/Δ + 1 μM quinidine (5 hearts)−2,720.4 ± 3,917.82,779.7 ± 3,917.210.4 ± 2.83.4−3,527.2 ± 4,674.783,603.0 ± 4,674.111.1 ± 2.73.9Scn5a+/− (6 hearts)−134.6 ± 28.9195.3 ± 28.022.8 ± 2.13.3−149.1 ± 33.5212.0 ± 32.622.9 ± 2.23.6Scn5a+/− + flecainide (6 hearts)−13.8 ± 29.266.2 ± 27.132.5 ± 11.11.8−55.4 ± 55.5108.6 ± 54.623.6 ± 5.91.2Scn5a+/− + 1 μM quinidine (5 hearts)−1,979.9 ± 1,709.02,051.5 ± 1,704.615.9 ± 2.813.4−108.4 ± 15,101.810,911.4 ± 15,098.311.8 ± 3.014.8Scn5a+/− + 10 μM quinidine (7 hearts)−332.1 ± 447.2431.4 ± 439.825.2 ± 12.7116.0−160.0 ± 289.5262.2 ± 281.920.1 ± 18.148.1Results of least-squares fits to the function under the conditions indicated Figures 4 and 5 thus superimpose optimisations of Eq. 1 to the data shown in Tables 1 and 2 (solid lines, left-hand ordinates). Furthermore, they show the corresponding gradients calculated using Eq. 2 (broken lines, right-hand axes). Shaded regions where present indicate DIs less than the critical DI at which these gradients exceed unity: the existence of such a region is an established criterion for arrhythmogenesis [25]. Critical DI was calculated as: Figure 6 proceeds to show these values of critical DI derived from the fitted values of A and τ for the epicardia (filled bars) and endocardia (open bars) under each of the above conditions. In contrast to the findings from the analysis of the relationships between epicardial and endocardial APD90s described above, the present analysis of epicardial and endocardial restitution properties together provided datasets in complete agreement with the observed incidences of arrhythmic activity. Fig. 6Critical diastolic intervals obtained from restitution curves. Critical diastolic intervals at which gradients equalled unity obtained from epicardial (filled bars) and endocardial (open bars) restitution curves shown in Figs. 4 and 5 in WT (a), Scn5a+/Δ (b) and Scn5a+/− (c) hearts exposed to control solution (A) and to solutions containing 1 μM flecainide (B), 1 μM quinidine (C) and 10 μM quinidine (D). Asterisks indicate values that are significantly (P < 0.05) larger than those recorded in WT hearts exposed to control solution First, in WT hearts (Fig. 6a) exposed to control solution (A) the critical DI took a value of −63.63 ± 14.41 ms in the epicardium and −15.43 ± 4.97 ms in the endocardium. Treatment with flecainide (B) significantly (P < 0.05) increased these values to 19.45 ± 9.39 ms in the epicardium and 58.75 ± 18.71 ms in the endocardium. This critical DI was attained in the endocardium (shaded region in Fig. 5a, B), in parallel with the increased incidence of arrhythmogenesis observed during the extrasystolic stimulation procedure under these conditions (Fig. 2, ii). Treatment with 1 μM quinidine (Fig. 6a, C) also significantly (P < 0.05) increased the critical DI in both the epicardium and endocardium (32.30 ± 7.80 ms and 27.80 ± 14.53 ms respectively) in parallel with the observed arrhythmogenicity on extrasystolic stimulation under these conditions (Fig. 2, ii), although such values were not attained during the dynamic pacing protocol. Furthermore, treatment with 10 μM quinidine (Fig. 6a, D) again significantly (P < 0.05) increased these values to 97.20 ± 57.90 ms in the epicardium and to 28.50 ± 34.52 ms in the endocardium. This DI was attained in the epicardium in parallel with the observed arrhythmogenicity during the extrasystolic stimulation procedure (Fig. 2, ii). Thus, increases in the critical DI are precisely paralleled with arrhythmogenicity in WT hearts treated with flecainide and quinidine. Secondly, critical DIs obtained from Scn5a+/Δ hearts (Fig. 6b) were always significantly greater than those in WT controls (P < 0.05 in all cases) in parallel with the increased incidence of arrhythmic activity observed during the extrasystolic stimulation procedure (Fig. 2, ii). Thus, in hearts treated with control solution (Fig. 6b, A) the critical DI increased to 28.04 ± 8.08 ms in the epicardium and 39.38 ± 10.28 ms in the endocardium: These values were attained in both the epicardium and endocardium in association with the increased incidence of arrhythmic activity on extrasystolic stimulation observed (Fig. 2, ii) and in full agreement with clinical expectations [23]. Treatment with flecainide (Fig. 6b, B) decreased the epicardial value (−14.09 ± 9.05 ms, P < 0.05 as compared to hearts treated with control solution) but did not significantly affect the endocardial value (21.41 ± 8.58 ms, P > 0.05 as compared to hearts treated with control solution), having no effect on the observed incidence of arrhythmic activity on extrasystolic stimulation (Fig. 2, ii). However, treatment with 1 μM quinidine (Fig. 6b, C) significantly (P < 0.05) increased both epicardial and endocardial values (59.09 ± 26.91 ms and 57.64 ± 16.68 ms, respectively) as compared to those obtained from such hearts treated with control solution, in association with the observed increases in the incidence of arrhythmic activity resulting from extrasystolic stimulation (Fig. 2, ii) and again in agreement with expectations from clinical work [33]. It was not possible to obtain MAPs over a sufficiently wide range of BCLs to permit the construction of such curves in Scn5a+/Δ hearts treated with 10 μM quinidine. Thirdly, critical DI values were also increased in Scn5a+/− hearts (Fig. 6c) in association with the increased incidence of arrhythmic activity observed. Thus, in hearts treated with control solution (A), the critical DIs were significantly (P < 0.05) increased as compared to WT controls in both the epicardium and endocardium to 56.00 ± 27.79 ms and 52.59 ± 9.50 ms, respectively. These values were attained in both the epicardium and endocardium, fulfilling clinical expectations [1]. Treatment with flecainide (B) did not significantly alter these values (68.14 ± 24.09 ms in the epicardium and 58.56 ± 14.75 ms in the endocardium). These were again attained in both epicardium and endocardium in association with the arrhythmic activity observed on extrasystolic stimulation (Fig. 2, ii). Furthermore, treatment with quinidine at a concentration of 1 μM (Fig. 6c, C) had no significant effect on critical DIs in either the epicardium or endocardium (62.45 ± 29.74 ms and 79.63 ± 39.31 ms, respectively). These values were again attained in both epicardium and endocardium. However, treatment with quinidine at 10 μM significantly decreased both epicardial and endocardial critical DIs, such that they became statistically indistinguishable from WT controls (29.01 ± 9.27 ms and −80 ± 3.58 ms, respectively). This was associated with the complete absence of arrhythmogenesis during the extrasystolic stimulation procedure (Fig. 2, ii). Thus, while relationships between epicardial and endocardial action potential durations obtained over a range of cycle lengths are poorly correlated with arrhythmogenicity, restitution curves are, in general, predictive of arrhythmogenicity under the genetic and pharmacological conditions studied. Discussion The congenital LQT3 and the BrS constitute two important examples of genetic causes for ventricular arrhythmia and sudden cardiac death [3]. These conditions result from similarly targeted mutations in the SCN5a gene coding for the cardiac Na+ channel: gain-of-function mutations are characteristic of LQT3, while loss-of-function mutations are characteristic of BrS [1, 23]. The present study began by using the MAP technique to verify that Langendorff-perfused murine hearts modelling LQT3 (Scn5a+/Δ) [13] and BrS (Scn5a+/−) [29] recapitulated features of the incidence of arrhythmic activity reported clinically. These results were in full agreement with previous physiological and pharmacological findings from our LQT3 [39] and BrS [38] models. Epicardial and endocardial MAP recordings were then used to evaluate the applicability of two established criteria in predicting arrhythmogenicity in these models over the full physiological range of heart rates (quantified as baseline cycle lengths, BCLs). Transmural dispersion of repolarization was studied over a range of BCLs by comparing epicardial and endocardial action potential durations (at 90% repolarization, APD90s) in the two genetic models. Such an analysis had previously been demonstrated to reveal arrhythmogenicity in hypokalaemic Langendorff-perfused murine hearts [34]. Furthermore, changes in this relationship investigated at a single BCL had previously been associated with arrhythmogenicity in our Scn5a+/Δ hearts [42]. Plots of epicardial against endocardial APD90 in WT hearts treated with control solution gave reference relationships best fitted by a quadratic function, in full agreement with previous work [34]. Treatment of such WT hearts with either quinidine or flecainide resulted in the data points falling to the left of the reference curves, reflecting increased transmural repolarization gradients, in association with increased incidences of arrhythmic activity and again in full agreement with a previous study [34]. However, such plots failed to consistently predict the presence or absence of arrhythmogenicity in the Scn5a+/Δ and Scn5a+/− hearts. Such left-shifting of points was observed in Scn5a+/Δ hearts under all pharmacological conditions studied, despite the established anti-arrhythmic effect of flecainide in this setting. Furthermore, data points obtained from Scn5a+/− hearts treated with control solution fell on the reference curve despite the increased incidence of arrhythmic activity observed. Treatment with flecainide failed to result in a left-shift of points, despite its established pro-arrhythmic effect; treatment with quinidine did result in such a shift, despite its anti-arrhythmic effect. These discrepancies prompted an investigation for alternative criteria for arrhythmogenicity. We therefore explored the applicability of a second approach involving restitution curves plotting APD90 against the preceding DI. Arrhythmic activity occurring in a range of settings has previously been associated with increases in the slopes of such curves to greater than unity [14, 17–19, 28, 40]. Such an association has been demonstrated in a canine pharmacological model of the congenital long QT syndrome type 2 [45] but has not previously been studied in any models of LQT3 or BrS in any species. Furthermore, decreases in the slopes of such curves have been associated with the effects of clinically important anti-arrhythmic drugs, including the class 1A agent procainamide in a canine model [32]. However, such slopes have not been examined in relation to either quinidine or flecainide, nor have any such studies previously been attempted in murine models. This study accordingly proceeded to construct epicardial and endocardial restitution curves for the first time in a murine system and in relation to LQT3 or BrS in any species. In contrast to the previous criterion, epicardial and endocardial restitution curves taken together yielded parameters in precise agreement with the presence or absence of arrhythmic phenomena demonstrated. In WT hearts, treatment with either quinidine or flecainide significantly increased the critical DI at which the slopes of the restitution curves exceeded unity, in full agreement with the increased incidences of arrhythmic activity observed. Not only did the Scn5a+/Δ and Scn5a+/− mutations result in significant increases in critical DI in association with increased rates of arrhythmic activity, but quinidine and flecainide exerted contrasting effects on this critical DI in agreement with clinical expectations in both cases. Thus, in Scn5a+/Δ hearts, treatment with quinidine increased critical DIs in association with its pro-arrhythmic effect. Furthermore, in Scn5a+/− hearts, treatment with quinidine decreased critical DIs in association with the decreased incidence of arrhythmic activity observed, while treatment with flecainide had the opposite effect. Thus class 1A agents may result in either increases (quinidine in WT and Scn5a+/Δ murine hearts) or decreases (quinidine in Scn5a+/− murine hearts and procainamide in canine hearts) [32] in the critical DI: Such changes are in full agreement with effects on arrhythmogenicity in all cases. Taken together, these results demonstrate that criteria considering effects on the slopes of restitution curves successfully account for arrhythmogenicity in murine models of arrhythmic syndromes resulting from mutations in the Scn5a Na+ channel, as well as for the effects of quinidine and flecainide on such arrhythmogenicity. These observations might form the basis for future clinical explorations of the utility of restitution curves obtained using the MAP technique [28, 40] in the assessment of arrhythmic propensity in hereditary arrhythmic syndromes. Such a robust means of assessing arrhythmogenicity might be particularly useful given the known variable expressivity of the arrhythmic phenotype in patients harbouring such mutations [41].	[ "heart", "action potential", "arrhythmia", "heart rate", "heart excitation", "venticular muscle", "ventricle" ]	[ "P", "P", "P", "P", "R", "U", "U" ]
Exp_Brain_Res-3-1-1914263	Perception of limb orientation in the vertical plane depends on center of mass rather than inertial eigenvectors	We performed two experiments to test the hypothesis that the perception of limb orientation depends on inertial eigenvectors (ei) against the alternative hypothesis that it depends on the center of mass vector (CM). Whereas ei constrains the dynamic torques involved in angular rotation, CM constrains the static torque necessary to keep the limb aloft in the gravitational field. Hence, possible effects of ei and CM on kinesthetic judgments must be related to the dynamic and static torques, respectively, involved in moving and positioning a limb. In the first experiment, blindfolded participants matched, with upper arms supported, the orientation of their forearms while the forearms’ ei and CM were manipulated relative to the elbow. The manipulation of the vector CM alone induced a matching bias, as did the combined manipulation of ei and CM, whereas the manipulation of ei alone did not. In the second experiment, participants positioned their unseen and unsupported right arm at an indicated spatial configuration while ei and CM of the right forearm were manipulated as in Experiment 1. As in the first experiment, forearm positioning was affected by the independent manipulation of CM and the combined manipulation of ei and CM, but not by the independent variation of ei. Moreover, none of the manipulations affected upper arm positioning. These results refute the claim that the perception of limb orientation (in the vertical plane) is based on ei and demonstrate, for the first time, the implication of a limb segment’s CM in the perception of its orientation. Introduction How do we perceive where our limbs are in space without having to look at them continuously? Although this question is fundamental to the understanding of both perception and motor control, it is far from resolved. Investigations of the neurophysiological basis of kinesthesis have produced many relevant findings, but they have also highlighted the need for psychophysical concepts that pertain to the role of mechanical, in particular kinetic, information in kinesthetic experiences, as is illustrated by a brief overview of the pertinent literature. Given that the relative position of limb segments is specific to a particular set of joint angles, the neurophysiological basis of kinesthesis has been sought primarily in neural signals providing geometric information related to joint angles. In principle, such signals may come from mechanoreceptors in joint, skin, and muscle. The role of joint receptors proved to be modest at best as they were found to be mostly silent in the mid-range of motion and to fire only at the extremes of a joint’s movement range (e.g., Burgess and Clark 1969; Clark and Burgess 1975; Grigg and Greenspan 1977). Similarly, no prominent role of cutaneous receptors in signaling joint angles was established (Gandevia 1996), leaving muscle receptors as the primary candidates for signaling limb geometry. Although skeletal muscles contain both Golgi tendon organs and muscle spindles, only the latter are currently thought to reliably signal muscle length and, hence, inform about limb geometry (Kandel et al. 2000). Goodwin and colleagues firmly established this view by showing that tendon vibrations targeted at primary and secondary spindle endings induced marked illusory joint displacements (e.g., Goodwin et al. 1972a, b). More recently, Ribot-Ciscar et al. (2003) showed that the exclusive reliance on primary muscle spindle afferents may in theory lead to the accurate perception of joint angle. It is thus beyond doubt that geometrical information pertaining to joint angles, primarily signaled by muscle spindles, plays an important role in kinesthesis. Yet it has become apparent that our sense of limb position and movement is not based exclusively on geometric information. Signals of a kinetic nature (i.e., relating to force or effort), which presumably cannot be conveyed by muscle spindles, probably provide positional information as well. Rymer and D’Almeida (1980) demonstrated such kinetic influences by showing that errors in perceived finger orientation occur upon the generation of isometric contractions. They attributed this effect to a central mechanism receiving muscle force information from Golgi tendon organs. Results of Worringham and Stelmach (1985) suggested that limb kinesthesis depends on gravitational torque (Ng), which would also imply that it has a kinetic component. The significance of information related to muscular force or effort was amplified further by recent findings indicating a kinesthetic effect of muscular effort (Proske et al. 2004; Walsh et al. 2004; Winter et al. 2005), motor outflow signals (Gandevia et al. 2006), and muscle activity (Prud’homme and Kalaska 1994). Given the accumulating evidence that signals related to muscular force or effort are involved in kinesthesis, one may wonder how information relevant to kinesthetis is extracted from such kinetic signals. This issue is far from trivial because, unlike muscle spindle activity, tendon organ activity or motor outflow corrolaries bear no direct relation to muscle length or joint angle. This fact may have contributed to the now widespread view, put forward by McCloskey (1981), that motor outflow corollaries are only used to filter out spindle discharges related to changes in muscle activity so as to obtain accurate muscle length information (see Gandevia 1996). Yet it is possible that both afferent and efferent kinetic signals contribute to kinesthesis independent of muscle spindle activity. Information about muscular force directly reflects the dynamic and static torques involved in actively moving or positioning a limb. Through Newton’s laws of motion, those torques are linked to the limb’s mass distribution in space. Therefore, a promising psychophysical approach may be to postulate that an independent kinetic foundation of kinesthesis resides directly in specific characteristics of the limb’s mass distribution. Taking such an approach, Pagano and colleagues (e.g., Pagano and Turvey 1995; Pagano et al. 1996; Garrett et al. 1998) hypothesized that the perception of a limb’s spatial orientation depends on its inertial eigenvectors (ei), which represent a characteristic of a limb’s mass distribution related to the direction in which it resists rotation (see the appendix for a more detailed explanation of the physical meaning of inertial eigenvectors). Relative to a given point in space, any rigid object has three orthogonal eigenvectors: e1, e2, and e3. For a forearm rotating around the elbow, e1 and e2 describe a plane through the elbow, orthogonal to the forearm’s longitudinal axis. The axis of minimal resistance against rotation, e3, roughly coincides with the forearm’s longitudinal axis. The possible kinesthetic role of ei in general and e3 in particular resides in the fact that they reflect the forearm’s spatial orientation. In studying limb kinesthesis, it is essential to recognize that information about ei is only available when a perceiver actively rotates his or her forearm. After all, ei is related to the resistance against rotation and thus exclusively affects the dynamic torques involved in rotation. When a forearm is held stationary, the required muscular torque only has a static component and is therefore independent of ei. The so-called inertial eigenvector hypothesis was introduced by Pagano and Turvey (1995) and corroborated in numerous subsequent studies (Pagano et al. 1996; Garrett et al. 1998; Pagano and Turvey 1998; Turvey 1998; Pagano 2000; Riley and Turvey 2001; Riley and Pagano 2003; Bernardin et al. 2005; Riley et al. 2005). All pertinent experiments adopted a similar method in which a single load was attached to the limb at a distance from its longitudinal axis, thereby breaking its coincidence with e3. With the exception of a recent study by Craig and Bourdin (2002, but see Riley and Pagano 2003; Riley et al. 2005) it was found that the perception of limb orientation was biased towards e3, which was taken as evidence for the hypothesis in question. The studies by Pagano et al. (1996) and Garrett et al. (1998) are of particular interest for the present study because they specifically addressed the perception of limb orientation in a vertical (i.e., gravitational) plane and found that it was affected by the aforementioned manipulation of e3. However, close scrutiny of this manipulation reveals that an alternative explanation is possible. The placement of a single load off the forearm’s longitudinal axis not only introduces a rotation of the vector e3 relative to the elbow, but also displaces the forearm’s center of mass. Similar to e3, we view the center of mass as a vector originating at the elbow (CM), and its displacement as a rotation of this vector. Importantly, CM and e3 constitute principally different characteristics of the forearm’s mass distribution. Whereas ei exclusively affects the dynamic torques involved in limb rotation, the vector CM exclusively affects the static gravitational torque (Ng). A dependence of kinesthesis on CM would thus imply a fundamentally different kinetic basis than a dependence on ei. Yet the possibility that CM, rather than e3, governed the results of Pagano et al. (1996) and Garrett et al. (1998) has never been tested experimentally. In order to disentangle the effects of both variables, and thus to test the inertial eigenvector hypothesis against the center of mass hypothesis, ei and CM must be varied independently. We accomplished such a manipulation and experimental test in the experiments reported below. Experiment 1 In the first experiment, we tested the inertial eigenvector hypothesis against the center of mass hypothesis using a forearm-matching task similar to that employed by Pagano et al. (1996) and Garrett et al. (1998). The inertial eigenvector hypothesis predicts that variation in e3 is sufficient to induce a matching bias. Hence, according to this hypothesis a matching bias should occur in all conditions in which e3 is manipulated. According to the center of mass hypothesis, a pointing bias should occur upon rotation of the vector CM, irrespective of the orientation of e3. In the present experiment, the mass distribution of the forearms was manipulated such that e3 and CM either varied independently or covaried (as was the case in the experiments of Pagano and colleagues), thus allowing for a critical test of the two hypotheses of interest. Method Twenty healthy participants (11 female and 9 male; all right-handed; mean age 28.5 years, SD 6.0 years) participated voluntarily in the experiment. They were not familiar with the type of experiment or the rationale behind it. The experiment, which was conducted in accordance with the 1964 Declaration of Helsinki, was approved formally by the ethical committee of our faculty and carried out with the adequate understanding and written informed consent of all participants. Participants were blindfolded and sat on a stool with their upper arms resting on a wooden surface. They were to match the orientation of their unseen forearms, to which carbon fiber frames with two brass loads were fixed in order to achieve the desired orientations of e3 and CM (see Fig. 1). Fig. 1The experimental setup of Experiment 1. Participants were blindfolded and sat on a stool with their upper arms resting on a wooden surface oriented at a 60° angle with the horizontal. Participants’ armpits touched the upper edge of the wooden surface and their upper arms were positioned in parallel and flush with the wooden surface. The carbon fiber frames, used to attach loads to the forearms (see text for a detailed description), were to remain in parallel vertical planes while forearm orientation was matched by flexion and extension of the elbows. Two straws with a length of 5 cm extended from the upper end of the proximal crosspieces, preventing the loads from touching the upper arms and marking the maximal allowed elbow flexion. The straws are visible inside the dotted circle Each carbon fiber frame consisted of a stem, 30 cm in length and 1 cm in diameter, and two parallel crosspieces, both 40 cm in length and 0.6 cm in diameter, that were pierced through the stem at a 90° angle. These crosspieces were placed 22 cm apart with the most distal crosspiece at 4 cm from the distal end of the stem. The total mass of each frame was 32 g. The frames were fixed to the ventral side of the forearm, along its longitudinal axis, and their distal tips protruded from the closed hand between the ring finger and the middle finger. The crosspieces were positioned at a distance of 11 and 33 cm, respectively, from the medial epicondyl of the humerus. A thin straw with a length of 5 cm protruded from the upper part of the proximal crosspiece (see Fig. 1), preventing the upper arm from contacting the loads. Moreover, by touching the upper arm, it signaled the smallest elbow angle allowed in the experiment (approximately 80°). Forearm matching occurred in eight experimental conditions and a control condition. In the experimental conditions, two cylindrical brass loads were attached to the frames to achieve the desired orientations of e3 and CM. The experimental manipulations are illustrated in Fig. 2 and the exact masses and positions of the loads in the eight experimental conditions are reported in Table 1. The magnitude and direction of e3 rotation induced by the loads was calculated by adding the loads’ inertia tensors relative to the elbow to that of the unloaded forearm (using the parallel axis theorem) and subsequently calculating the new orientation of e1, e2, and e3 by diagonalizing the resulting tensor. In conditions 1 and 2, e3 was thus manipulated 5° toward flexion in one arm (the right arm in condition 1; the left arm in condition 2) and 5° toward extension in the other, without manipulating CM, resulting in a 10° difference between the eigenvectors of the forearms. In conditions 3 and 4, CM was manipulated 5° toward flexion in one arm (the right arm in condition 3; the left arm in condition 4) and 5° toward extension in the other, without manipulating e3, resulting in a 10° CM difference between the forearms. In conditions 5 and 6, both e3 and CM were manipulated such that there was a 10° e3 and CM difference between both forearms. In conditions 7 and 8, one arm remained unloaded (the left arm in condition 7; the right arm in condition 8) and loads were attached to the other arm in a symmetrical way, i.e., without inducing a rotation of either e3 or CM. A ninth condition, in which both forearms remained unloaded, served as control condition. The maximal value of Ng about the elbow was equal in both arms in all conditions, except for conditions 7 and 8, which were included to explicitly test for the effect of Ng suggested by Worringham and Stelmach (1985). In conditions 1 through 6, the placement of the loads merely induced an asymmetrical mass distribution relative to the forearm’s longitudinal axis. Fig. 2The experimental manipulations of the vectors e3 and CM. CM is the vector from the elbow joint to the average position of the (loaded) forearm’s mass (i.e., the effective point of origin of gravitational force). The eigenvector e3 is the axis through the elbow joint about which the (loaded) forearm’s rotational inertia is minimal. Panel a shows the loads as they were attached to one of the forearms in conditions 7 and 8 (in which the other forearm remained unloaded). The loads are equal in mass, are placed at equal but opposite distances from the forearm’s longitudinal axis, and have equal distances to elbow. They thus do not induce a rotation of either CM or e3 relative to an unloaded arm. In panel b (conditions 5 and 6), the asymmetrical load placement causes both vectors e3 and CM to be rotated towards the loads. In panel c (conditions 1 and 2), the loads have equal mass and are placed at equal but opposite distances from the forearm’s longitudinal axis, so that no rotation of CM is induced. Yet the loads have different distances to the elbow, so that they do induce a rotation of e3. Panel c thus represents the independent variation of e3. Finally, in panel d (conditions 3 and 4), the two loads are placed equidistant from the forearm’s longitudinal axis and have different distances to the elbow, as in panel c, but now their masses differ with the heaviest load being closest to the elbow. This leads to CM being rotated towards the greatest load, while e3 still coincides with the longitudinal axis. Panel d thus represents the independent variation of CMTable 1Masses and positions of the loads and corresponding rotations of e3 and CM in the experimental conditions of experiment 1Right forearmLeft forearmLoad 1Load 2Load 1Load 2ConditionMass (g)x, ya (cm)Mass (g)x, y (cm)Mass (g)x, y (cm)Mass (g)x, y (cm)Δe3b ΔCMc 115011, −18.815033, 18.815011, 18.815033, −18.810°0°215011, 18.815033, −18.815011, −18.815033, 18.8−10°0°330011, 13.810033, −13.830011, −13.810033, 13.80°10°430011, −13.810033, 13.830011, 13.810033, −13.80°−10°515011, 5.815033, 12.815011, −5.815033, −12.810°10°615011, −5.815033, −12.815011, 5.815033, 12.8−10°−10°710033, 18.810033, −18.8––––0°0°8––––10033, 18.810033, −18.80°0°ax refers to the distance of the load’s center of mass from the elbow along the arm’s longitudinal axis; y refers to the orthogonal distance of the load’s center of mass from the arm’s longitudinal axis, where a positive (negative) distance indicates a position on the radial (ulnar) side of the forearmbΔe3 indicates the angle between the smallest eigenvector (e3) of the left and the right forearm. A positive angle indicates that e3 of the right forearm was manipulated towards elbow flexion and e3 of the left forearm towards elbow extensioncΔCM indicates the angle between the center of mass vector (CM) of the left and the right forearm. A positive angle indicates that CM of the right forearm was oriented towards elbow flexion and CM of the left forearm towards elbow extension Participants performed the nine conditions in nine corresponding trial blocks. Before each trial block, they assumed a position in which one arm was flexed at an elbow angle of 80°, which was achieved by letting the tip of the straw just touch the upper arm, and the other arm was fully extended. Note that, in this starting position, a perceptual reference was ensured both in the flexed arm (by the straw) and in the fully extended arm (by the end of the elbow’s movement range), preventing any drift effects over trials. From this position, one of the forearms (the target arm) was moved towards the other forearm until the experimenter called out “stop”. The experimenter ensured that this stop signal was given at a different arm orientation in each trial. The other arm (the matching arm) was then moved towards the target arm until the participant perceived the orientation of the two forearms to be identical. At this moment, the participant stopped the movement and called out “ja” (“yes”). After registering the orientation of both forearms (see below), the experimenter instructed the participant to assume the starting position for the next trial: The arm that was flexed at the start of the previous trial was now extended and vice versa. The target arm was alternately the left and the right arm. Participants were instructed to keep both arms and the crosspieces they enclosed with their hands in a vertical (i.e., sagittal) plane at all times. Each trial block was started with either the left or the right arm extended, and with either the extended or the flexed arm as the target arm. The four resulting starting configurations were counterbalanced across participants, with each individual participant starting all trial blocks from the same assigned configuration. Each trial block consisted of two series of eight matching trials. In one series the extended arm was the target arm and in the other the flexed arm was the target arm. All participants performed 144 matching trials in total (9 trial blocks; 16 trials per block). The duration of an experimental session was approximately 45 min. Forearm orientation was measured using a 3D active movement registration system (Optotrak 3020, Northern Digital Inc., Waterloo, Canada), which was calibrated using a coordinate frame with one axis aligned with the gravitational vertical, and one parallel to the horizontal axis described by participants’ elbow and shoulder joints. During the experiment, the position of four infrared markers was registered: two markers on the distal tip of each frame stem, and two markers on the horizontal axes described by the two elbow joints and the two shoulder joints, respectively. The latter two markers were placed on an adjustable carbon fiber frame to the left of the participants after they assumed the correct starting position. In this starting position, both armpits contacted the upper edge of the wooden board, which ensured that the elbow joints as well as the shoulder joints described a horizontal axis. The experimenter registered the position of the four markers each time a participant indicated that the orientation of the forearms was matched. For each participant and for each trial, the angle of the two elbow joints, projected onto a sagittal plane, was calculated from the position data of the four Optotrak markers. The elbow angles of both arms were subsequently averaged to obtain the angle around which matching occurred in each trial. To obtain the direction and magnitude of matching errors, the elbow angle of the right arm was subtracted from that of the left arm. A positive matching error thus indicated that elbow angle was smallest in the right arm, that is, that the right arm had a flexion bias relative to the left arm. Finally, matching errors in the control condition were subtracted from those in each experimental condition to obtain matching biases due to the eight experimental manipulations. We first examined the range of elbow angles around which matching occurred in each condition, the pattern of matching errors within conditions, and the matching errors in the control condition. Subsequently, we analyzed the matching errors according to a repeated measures analysis of variance (ANOVA) with condition (9 levels) and repetition (16 levels) as within-subject factors. Finally, we tested the matching biases due to the eight experimental manipulations using one-sample two-tailed t tests. Results and discussion The average and the range of the elbow angles around which matching occurred in each of the nine conditions was 115.5° (SD over conditions 1.0°) and 26.9° (SD over conditions 1.3°), respectively. The low standard deviations indicate that both the average and the range of matching angles only differed marginally between conditions. Before turning to the effects of our experimental manipulations on the matching errors, we first examined matching errors within conditions. Participants showed a significant overshoot of the target arm’s orientation with the matching arm (t(19) = 2.41, P = 0.027, η2 = 0.23), which is in keeping with Worringham and Stelmach (1985), but not with Pagano et al. (1996) who did not find such an effect. Matching errors of a representative participant in the control condition are shown in Fig. 3. The alternating pattern of errors in this figure is a manifestation of the aforementioned overshoot effect. After all, the left and the right arm alternatingly assumed a flexed starting position, so that an overshoot of the target arm would alternatingly lead to a positive and a negative matching bias. In this way, matching biases due to overshoot canceled out after averaging over trials. Accordingly, in the control condition, matching errors did not deviate significantly from zero (P = 0.53) after averaging over trials (mean 0.5°; SD across participants 3.7°). However, partly due to a tendency to overshoot the target arm, error variability across trials was considerable (SD 5.7°; see Fig. 3). This trial-to-trial variability is comparable to that reported by Soechting (1982). Fig. 3Matching errors of a representative participant in the control condition of Experiment 1, in which no loads were attached to the forearms. The alternating pattern of errors over trials reflects a tendency to overshoot the target arm with the matching arm The ANOVA performed on the matching errors in all conditions revealed that they were affected significantly by condition (F(8, 152) = 10.5, P < 0.001, ηp2 = 0.36), but neither by repetition (P = 0.80) nor by the condition × repetition interaction (P = 0.46). Figure 4 shows the average matching biases (with 95% confidence intervals) due to the eight experimental manipulations. No significant effect of e3 was found (P = 0.72 and P = 0.69, in conditions 1 and 2, respectively). In contrast, the 10° difference in CM orientation resulted in significant matching biases of −1.5° in condition 3 (t(19) = −2.92, P = 0.009, η2 = 0.31) and of 2.2° in condition 4 (t(19) = 2.70, P = 0.014, η2 = 0.28). The combined manipulations of e3 and CM had similar effects: A 10° difference of e3 and CM orientation between the arms resulted in significant matching biases of −1.9° in condition 5 (t(19) = −3.63, P = 0.002, η2 = 0.41) and of 2.5° in condition 6 (t(19) = 4.13, P = 0.001, η2 = 0.47). A negative (positive) matching bias implied that the flexion of the right arm relative to the left arm was smaller (greater) than in the control condition. The direction of the significant effects of conditions 3 through 6 was consistent with the hypothesis that perceived orientation would be biased towards CM. Finally, loading one of the arms symmetrically (i.e., without rotating either e3 or CM) while leaving the other arm unloaded did not induce significant matching biases relative to the control condition (t(19) = −1.87, P = 0.077, η2 = 0.16 for condition 7; t(19) = 1.50, P = 0.15, η2 = 0.11 for condition 8), although there was a trend towards an extension bias in the loaded arm (see Fig. 4). Such an extension bias would be consistent with the direction of the effect of Ng (or muscular effort) reported in the literature (Worringham and Stelmach 1985; Proske et al. 2004; Walsh et al. 2004; Winter et al. 2005). More importantly, the significant effect of CM in conditions 3 and 4, together with the absence of a significant effect of e3 in conditions 1 and 2, indicates that the results of Pagano et al. (1996) and Garrett et al. (1998) should be interpreted in retrospect as a sensitivity to CM rather than ei. Fig. 4Average matching biases relative to the control condition in the different manipulation conditions of Experiment 1. A positive bias implies that the flexion of the right arm relative to the left arm was greater in that manipulation condition than in the control condition. Error bars show the 95% confidence interval of the matching biases. Only e3 was rotated in conditions 1 and 2, only CM was rotated in conditions 3, and 4, and e3 and CM were rotated together in conditions 5 and 6. In conditions 1, 3, and 5, rotations were toward flexion in the right arm and toward extension in the left arm (and vice versa in conditions 2, 4, and 6). In conditions 7 and 8, e3 and CM were not rotated; only Ng differed between the two forearms: in condition 7 (8), it was greater in the right (left) arm Given that people can perceive the orientation of their limbs in space, and not only relative to each other, one may wonder whether the present results, favoring the center of mass hypothesis, generalize to the perception of the orientation of an unsupported arm in extrinsic space. In this situation, loading the forearm not only affects the torques around the elbow joint, as was the case in the present experiment, but also the torques around the shoulder joint. In principle, this could restrict or alter the implication of CM in the perception of limb orientation. These possibilities were investigated in Experiment 2. Experiment 2 In the second experiment, we manipulated e3 and/or CM of the forearm in the same manner as in Experiment 1, but this time to test their effects on the perception of the orientation of an unsupported arm in a vertical plane. Based on the results of Experiment 1, we hypothesized that perceived forearm orientation would be affected by CM and not by e3. We further hypothesized that the manipulations would not affect the perceived orientation of the upper arm, even though we recognized that loading the forearm does affect the torque at the shoulder when the upper arm is not supported. The latter hypothesis was motivated from the insight that the shoulder torque can only be informative about the orientation of the upper arm when the elbow torque is taken into account as well. We reasoned that manipulation of the forearm’s CM would induce torques at the elbow indicating an altered forearm orientation. Given that the corresponding change in the torque pattern at the shoulder is consistent with this change in forearm orientation, we further reasoned that it too would indicate an altered forearm orientation, rather than an altered upper arm orientation. Method Twenty healthy participants (13 female and 7 male; all right-handed; mean age 26.0 years, SD 4.3 years) participated voluntarily in the experiment. They were not familiar with the type of experiment or the rationale behind it. The experiment, which was conducted in accordance with the 1964 Declaration of Helsinki, was approved formally by the ethical committee of our faculty and carried out with the adequate understanding and written informed consent of all participants. The experimental setup is shown in Fig. 5. Participants were seated on a stool besides a vertical wooden board. Their right arm and shoulder were placed through a circular hole in the board between two overlapping pieces of lycra cloth that prevented them from looking through the hole to the other side of the board. Participants were wearing a T-shirt, of which the right sleeve was rolled up to just above the acromion to allow the placement of a marker on the shoulder (see below). The stool was adjusted such that the projection of the forearm on the vertical plane could coincide with line segments 1 through 3, and that of the upper arm with line segment 4. The experimental task consisted of matching the orientation of the unseen forearm with line segment 1, 2, or 3, while maintaining the upper arm parallel to line segment 4. Line segments 1, 2, and 3 were oriented at angles of 15°, 0°, and −15°, respectively, relative to horizontal. Line segment 4 was oriented at an angle of −60° relative to horizontal, which corresponded to the orientation of the upper arm support in Experiment 1. Brass weights were attached to the forearm by means of a carbon fiber frame—as in Experiment 1—to achieve the desired orientations of the vectors e3 and CM (see Fig. 2). The straw, used in Experiment 1 to prevent arm orientations smaller than about 80°, was not used in the present experiment because the risk that loads would contact the upper arm was much smaller than in Experiment 1. Fig. 5The experimental setup of Experiment 2. Participants sat on a stool besides a vertical wooden board. Their right arm and shoulder were placed through a circular hole in the board between two overlapping pieces of lycra cloth. Participants were wearing a T-shirt, of which the right sleeve was rolled up to uncover the shoulder. The stool was adjusted such that the projection of the right forearm on the vertical plane could coincide with line segments 1, 2 and 3 (15°, 0° and −15° relative to horizontal, respectively) and the projection of the upper arm with line segment 4 (−60° relative to horizontal). The experimental task consisted of matching the orientation of the unseen forearm alternately with line segments 1, 2 and 3, while maintaining the unsupported upper arm parallel to line segment 4. The carbon fiber frame, used to attach loads to the forearm (see text for a detailed description), was to remain parallel to the board at all times Arm positioning was required in seven experimental conditions and a control condition. Load placements in the experimental conditions were equal to those in experiment 1. The exact masses and positions of the loads can thus be found in the five leftmost columns of Table 1 (i.e., those referring to the right arm). In conditions 1 and 2, e3 was independently manipulated by 5° towards extension and 5° towards flexion, respectively, resulting in a 10° difference in e3 orientation between the two conditions. A 10° difference in CM orientation was achieved between conditions 3 and 4, and a 10° difference in the orientation of both e3 and CM was achieved between conditions 5 and 6. In condition 7, loads were attached to the forearm in a symmetrical way, i.e., without inducing a rotation of either e3 or CM. An eighth condition, in which the forearm remained unloaded, served as control condition. The eight conditions were performed in eight corresponding trial blocks. Before each trial block, participants assumed a position in which the upper arm was oriented parallel to line segment 4, and the arm was either fully flexed or fully extended (counterbalanced across participants). The experimenter then called out one of the numbers 1 through 3 to indicate the target line segment for the forearm, upon which the participant rotated his or her forearm around the elbow towards the target line segment. Participants were instructed to stop moving the arm and push a button with the left hand when they perceived their forearm to be parallel to the target line segment and their upper arm to be parallel to line segment 4. If the arm was initially fully flexed (extended), participants subsequently fully extended (flexed) their arm. The experimenter then called out a new target line segment and the forearm was again rotated about the elbow (albeit in opposite direction) to match the orientation of the indicated line segment. Participants thus alternated between full elbow flexion and full elbow extension, and assumed the instructed arm configuration as a pause in each elbow flexion and extension movement. They were instructed to keep the crosspieces they enclosed with their right hand parallel to the wooden board at all times. In each trial block, participants matched the orientation of their forearm with each of the three target line segments four times, resulting in a total of 12 trials per block. Throughout each trial block, the upper arm was to remain parallel to line segment 4. All participants thus performed a total of 96 matching trials (8 trial blocks; 12 trials per block). The duration of an experimental session was approximately 30 min. The configuration of the arm was measured using Optotrak. The position of six infrared markers was continuously registered at 100 Hz during each trial block: a marker on the tip of the carbon fiber frame (marker 1), a marker on either side on the distal crosspiece (markers 2 and 3), a marker on the lateral epicondyl of the humerus (marker 4), a marker just below the acromion on the deltoid muscle (marker 5), and finally a marker (marker 6) connected to the button in participants’ left hand. The latter marker only emitted infrared light when the button was pushed. For each participant and for each trial, the orientation of the upper arm and forearm was calculated from the position data of markers 1, 4, and 5. The target orientations of forearm (15°, 0°, or −15° relative to horizontal) and upper arm (−60° relative to horizontal) were subtracted from the actual orientations that forearm and upper arm had at each moment the button was pushed. Errors in forearm and upper arm orientation were analyzed according to a repeated measures analysis of variance (ANOVA). For forearm orientation errors, we analyzed condition (8 levels), target (3 levels, corresponding to line segments 1, 2, and 3; see Fig. 5), and repetition (4 levels) as within-subject factors. For errors in upper arm orientation, we analyzed only the factors condition and repetition. The difference in positioning errors between conditions 1 and 2 (e3 manipulation), 3 and 4 (CM manipulation), 5 and 6 (combined e3 and CM manipulation), and 7 and 8 (Ng manipulation without e3 and CM rotation), collapsed over targets, were subsequently compared using paired-samples two-tailed t tests. Note that in the present experiment four t tests rather than eight (as in Experiment 1) sufficed, because only one arm was manipulated rather than two. Results and discussion The average error in aligning the forearm with line segments 1, 2, and 3 was 4.0° in upward direction. The ANOVA performed on these errors revealed that they were affected significantly by condition (F(7, 133) = 8.4, P < 0.001, ηp2 = 0.31) and target (F(2, 38) = 7.0, P = 0.003, ηp2 = 0.27), but not by repetition (P = 0.15). The target effect indicated that a lower line segment was associated with smaller upward forearm positioning errors than a higher line segment. None of the two-way interactions was significant (all P’s > 0.5), whereas the three-way interaction just reached significance (F(42, 798) = 1.4, P = 0.042, ηp2 = 0.07). As the latter effect had a marginal effect size and had no readily apparent origin, we abstained from seeking an account for it. The results of the paired t tests performed on forearm positioning errors are shown in Fig. 6. No significant effect of adding mass in a symmetrical configuration was found (P = 0.18). The effect of e3 was also non-significant (P = 0.08). In contrast, CM rotation significantly affected forearm orientation (t(19) = 4.1, P = 0.001, η2 = 0.47), as did rotation of e3 and CM together (t(19) = 3.9, P = 0.001, η2 = 0.44). A 10° difference of CM orientation alone was accompanied by a 2.4° difference in forearm orientation. A 10° difference of both e3 and CM orientation corresponded to a 2.6° difference in forearm orientation. As expected, a more downward (upward) orientation of the forearm’s CM always accompanied a more upward (downward) forearm orientation. Fig. 6Difference in forearm orientation between conditions 1 and 2 (e3 manipulation), 3 and 4 (CM manipulation), 5 and 6 (combined e3 and CM manipulation), and 7 and 8 (Ng manipulation) in Experiment 2. In conditions 1, 3 and 5 (2, 4 and 6), e3 and/or CM were manipulated downward (upward). In condition 7, in which loads were attached to the frame symmetrically, Ng was larger than in condition 8, in which no loads were attached to the frame. For e3 and/or CM manipulation (three leftmost bars), a positive difference indicates that downward manipulation was associated with a more upward forearm orientation. For Ng manipulation (rightmost bar), a negative difference indicates that a greater Ng was associated with a more downward forearm orientation. Error bars show the 95% confidence interval of the orientation differences The average error in aligning the upper arm with line segment 4 was 7.4° in downward direction. The ANOVA performed on these errors revealed that they were only significantly affected by repetition (F(11, 209) = 2.4, P = 0.008, ηp2 = 0.11). The main effect of condition and the condition × repetition interaction just failed to reach significance (P = 0.056 and P = 0.093, respectively). The paired t tests revealed that neither CM rotation nor e3 rotation significantly affected upper arm orientation errors (both P’s > 0.7). The effect of symmetrical mass addition was also non-significant (P > 0.3). Finally, no effect of the combined e3 and CM manipulation was revealed, albeit that this effect approached significance (P = 0.071). In sum, forearm orientation was affected in a very similar way as in the matching task of Experiment 1, whereas the unsupported upper arm was not affected, or only marginally at best, by the manipulation of e3 and/or CM. The results thus corroborate both hypotheses forwarded in the introduction to the present experiment. General discussion With the overarching aim to uncover the kinetic foundation of kinesthesis, we tested the inertial eigenvector hypothesis against the alternative hypothesis that the perception of limb orientation in the vertical plane depends on CM. To this end, two experiments were conducted, one in which the orientations of the forearms had to be matched (Experiment 1) and one in which the orientation of a single arm had to be matched to an external reference configuration (Experiment 2). The results of both experiments supported the center of mass hypothesis and were inconsistent with the inertial eigenvector hypothesis. Experiment 1 revealed that the results of Pagano et al. (1996) and Garrett et al. (1998) should be interpreted in retrospect as a sensitivity to CM rather than ei. Experiment 2 generalized the results of Experiment 1 for forearm matching with the upper arm supported to orienting a single, unsupported limb with respect to extrinsic space. In particular, the results of Experiment 2 indicated that forearm manipulation in a vertical plane shifted perceived forearm orientation in that plane without causing a significant shift in the perceived orientation of the upper arm. The combined results of Experiments 1 and 2 raise the crucial question how CM might mediate limb kinesthesis. We address this question at two levels, first extensively in terms of the torques involved and then briefly and more tentatively in terms of possible neurophysiological mechanisms. In both experiments, placing mass symmetrically (see Fig. 2a) had no significant effect on the perception of forearm orientation, but did show a trend towards an extension bias in the heavier arm. It may be that adding mass symmetrically only had a small effect, which could explain the inconsistency of previous findings in this regard. In particular, whereas the results of Worringham and Stelmach (1985) indicate an effect of symmetrical mass placement, Soechting (1982) as well as Darling and Hondzinski (1999) failed to find such an effect. More important for the present discussion is the suggestion in the latter studies that Ng does not play a significant role in limb kinesthesis. This suggestion stands in stark contrast with the marked effects of CM found in the present experiments, which necessarily imply that Ng plays a role in limb kinesthesis, albeit in a different manner than considered previously. After all, as we already noted in the introduction, Ng is the only detectable variable affected by CM rotation. So how does CM affect Ng, and how is this different from the Ng manipulation in previous studies and our conditions with symmetrical mass addition? In an unloaded arm, Ng is a sinusoid function of arm inclination that reaches its maximum at a horizontal arm orientation and becomes zero when the arm is vertical (see solid gray curve in Fig. 7). Adding loads symmetrically (as shown in Fig. 2a) may be viewed as a scaling of this relationship. That is, it increases Ng at each inclination angle by a constant factor so that the maximum (minimum) of Ng still occurs at a horizontal (vertical) arm orientation (see the dotted black curve in Fig. 7). Yet a rotation of CM by asymmetric mass addition (as shown in Fig. 2b, d) not only scales but also horizontally shifts this relationship, so that the maximum of Ng no longer occurs at a horizontal arm orientation (see the solid black curve in Fig. 7). Hence, this horizontal shift reflects the rotation of the vector CM. One may therefore conclude that Ng plays an indirect role in the perception of limb orientation, namely as a mediator of CM orientation, rather than Ng magnitude itself being informative about limb orientation. Fig. 7Schematic depiction of the relationship between Ng and arm inclination angle (i.e., the angle of the arm’s longitudinal axis with the horizontal) in an unloaded arm (gray solid curve), a symmetrically loaded arm (black dotted curve) and an asymetrically loaded arm in which CM is rotated (black solid curve). Whereas symmetrical mass addition only changes the scaling of the relationship, CM rotation also induces a horizontal shift. We propose that the latter shift governs the sensitivity to CM demonstrated in the present study In a vertical plane, the perception of limb orientation relative to extrinsic coordinates (e.g., the gravitational vertical) has often been found to be more accurate than perception with respect to intrinsic coordinates (e.g., joint angle, or the trunk- and head-longitudinal axes; Soechting 1982; Worringham and Stelmach 1985; Worringham et al. 1987; Darling 1991; Darling and Hondzinski 1999). It follows from the present results that the use of extrinsic coordinates in limb kinesthesis may substantially benefit from CM-related information, conveyed by a horizontal shift of Ng as a function of arm inclination, which could explain this higher accuracy in extrinsic coordinates. The issue of extrinsic versus intrinsic coordinate systems is also a relevant dimension when comparing the center of mass hypothesis with the inertial eigenvector hypothesis. Unlike CM, ei cannot aid the perception of limb orientation in extrinsic, earth-fixed coordinates. After all, as we noted in the introduction, the possible perceptual role of ei resides exclusively in its effect on the dynamic torques involved in limb rotation, which are independent of a limb’s orientation in extrinsic space. Hence, the perception of limb orientation through the detection of inertial eigenvectors necessarily implies an intrinsic coordinate system, similar to that involved in the detection of muscle length or joint angle. This important point was overlooked by Garrett et al. (1998), who found that their manipulation of ei (and implicitly CM) affected the perception of forearm orientation relative to the gravitational vertical and interpreted this as an effect of ei. It follows from the preceding argument that this interpretation was invalid because ei, being gravity independent, cannot convey such information. Hence, irrespective of the present empirical evidence, Garrett et al. (1998) could have attributed their findings to gravity-dependent information. Because the experimental tasks in the present study allow for the use of extrinsic as well as intrinsic axes, the apparent dominance of CM over ei cannot be explained in terms of their respective coordinate systems. So what did underlie the difference in perceptual effect between ei and CM in the present experiments? The answer may reside in their respective signal-to-noise ratios, or saliences (cf. van de Langenberg et al. 2007). Recall that information about ei is only reflected in the dynamic torques involved in active limb movement. In the present experiments, as well as in Pagano et al. (1996) and Garrett et al. (1998), the arms were moved at a relatively low angular velocity, implying that the contribution of dynamic torques to the muscular tension was small relative to the static torque Ng. It follows that the salience of ei must have been low relative to that of CM, which exclusively affects Ng. In general, one can say that during slow movements or stationary postures, the salience of CM will be superior to that of ei. During fast movements, the salience of ei will increase at the cost of the salience of CM, which leaves open the possibility that the perception of arm orientation in fast movements, at least in intrinsic coordinates, is affected by ei as well. The results of Experiment 2 generalized the role of CM from matching the orientation of contralateral forearms to orienting a single arm in extrinsic space. They further showed that the perception of upper arm orientation was unaffected by the manipulation of the forearm’s CM, even though the manipulation affected Ng at both the elbow and the shoulder due to the absence of an upper arm support. The latter finding may be explained in two ways: Participants may have exploited Ng at the shoulder to obtain information about the configuration of the whole arm in space, as we anticipated in the introduction to Experiment 2, or they may have discounted it, only taking Ng into account at the elbow. The latter alternative would suggest that the role of CM in the perception of limb orientation pertains only to the most distal rigid segment, whereas the former alternative would suggest that its role is more general, supporting the perception of the orientation of both proximal and distal limb segments. This issue may be resolved in experiments in which the center of mass vectors of both distal and proximal limb segments (e.g., both the forearm and the upper arm) are manipulated in tasks similar to that used in Experiment 2. In seeking an encompassing account of the perception of limb orientation, it is important to note that we only tested the effect of CM and e3 manipulations in the vertical plane. Manipulation of CM in the horizontal plane would only affect the direction of Ng, not its magnitude, and hence cannot be mediated by the aforementioned horizontal shift in the relationship between arm orientation and Ng magnitude (see Fig. 7). In a recent study (van de Langenberg et al. 2007), we tested the effect of horizontal CM and e3 manipulations in a pointing task similar to that adopted in Experiments 1 and 2 of Pagano and Turvey (1995). We found essentially the same pattern of results as in the present study: Perceived arm orientation was affected by the horizontal manipulation of CM but not by that of e3. It thus appears that humans are sensitive to changes in both torque magnitude and direction, and hence to CM rotations in 3D. Another important point of discussion is that the observed perceptual biases were consistently smaller than the magnitude of CM rotation, as was the case in previous studies on the effect of vertical (CM and) e3 rotations (Pagano et al. 1996; Garrett et al. 1998). However, the relative sizes of our effects (i.e., 15–26% of the actual CM rotation) were smaller than those observed in those previous studies (i.e., 40–45% of the actual CM rotation). We suspect that this difference is related to differences in the magnitude of CM manipulation, which was 10° in our Experiment 1 and only about 5° in Pagano et al. (1996) and Garrett et al. (1998). In a study on the effect of horizontal e3 (and CM) rotations, Bernardin et al. (2005) adopted three manipulation magnitudes in a single experiment and indeed found a strong negative relationship between effect size and manipulation magnitude (upon an increase of CM manipulation from about 1.3° to about 5.5° effect size decreased from 38 to 13%). In addition to manipulation magnitude, other factors, such as manipulation direction (i.e. horizontal or vertical) and exploration style, may also influence the magnitude of CM’s effect on the perception of limb orientation. These factors should be explored further in future research. Regardless of their precise contributions, however, one can already conclude that the perceptual biases introduced by manipulating CM and e3 are in general incomplete. It follows from this general observation that additional information, unrelated to either CM or e3, must have been employed. The use of information unrelated to CM is also apparent in conditions in which the arm’s CM cannot be detected, such as when the limb is moved passively to a certain position (see e.g., Lee et al. 2003; Ulkar et al. 2004) or when gravity is absent (see e.g., Lackner and DiZio 2000). Notwithstanding the preceding qualifications, the present findings clearly underscore that there is a kinetic component to limb kinesthesis, which has important implications for its possible neural basis. More specifically, the apparent dependence of limb kinesthesis on CM points to an important role of neural signals related to muscular torque or effort. As noted in the introduction, several other psychophysical studies provided support for this view (Rymer and D’Almeida 1980; Prud’homme and Kalaska 1994; Proske et al. 2004; Walsh et al. 2004; Winter et al. 2005; Gandevia et al. 2006). Although these studies clearly challenge the view that limb kinesthesis relies on muscle spindle activity alone (see Ribot-Ciscar et al. 2003), they do not discard the possibility that signals related to muscular torque or effort only serve to accurately interpret spindle discharges in terms of muscle length changes, as proposed by McCloskey (1981, see also Gandevia 1996). The effect of CM on limb kinesthesis indicates that, in addition to such an indirect role, kinetic signals may convey information about the distribution of a limb’s mass in space and hence directly affect kinesthesis, independent of geometric information about muscle length or joint angle. Golgi tendon organs seem particularly suited for fulfilling this function, given that their activity is closely associated with muscle force (Kandel et al. 2000). This implies that the possible implication of Golgi tendon organs in limb kinesthesis may need to be reconsidered and that a mechanism akin to that suggested by Rymer and d’Almeida (1980) may indeed be possible.	[ "center of mass", "inertial eigenvectors", "kinesthesis", "information", "gravitational torque", "limb position sense", "proprioception" ]	[ "P", "P", "P", "P", "P", "R", "U" ]
J_Urban_Health-2-2-1705507	Respondent-Driven Sampling of Injection Drug Users in Two U.S.–Mexico Border Cities: Recruitment Dynamics and Impact on Estimates of HIV and Syphilis Prevalence	Respondent-driven sampling (RDS), a chain referral sampling approach, is increasingly used to recruit participants from hard-to-reach populations, such as injection drug users (IDUs). Using RDS, we recruited IDUs in Tijuana and Ciudad (Cd.) Juárez, two Mexican cities bordering San Diego, CA and El Paso, TX, respectively, and compared recruitment dynamics, reported network size, and estimates of HIV and syphilis prevalence. Between February and April 2005, we used RDS to recruit IDUs in Tijuana (15 seeds, 207 recruits) and Cd. Juárez (9 seeds, 197 recruits), Mexico for a cross-sectional study of behavioral and contextual factors associated with HIV, HCV and syphilis infections. All subjects provided informed consent, an anonymous interview, and a venous blood sample for serologic testing of HIV, HCV, HBV (Cd. Juárez only) and syphilis antibody. Log-linear models were used to analyze the association between the state of the recruiter and that of the recruitee in the referral chains, and population estimates of the presence of syphilis antibody were obtained, correcting for biased sampling using RDS-based estimators. Sampling of the targeted 200 recruits per city was achieved rapidly (2 months in Tijuana, 2 weeks in Cd. Juárez). After excluding seeds and missing data, the sample prevalence of HCV, HIV and syphilis were 96.6, 1.9 and 13.5% respectively in Tijuana, and 95.3, 4.1, and 2.7% respectively in Cd. Juárez (where HBV prevalence was 84.7%). Syphilis cases were clustered in recruitment trees. RDS-corrected estimates of syphilis antibody prevalence ranged from 12.8 to 26.8% in Tijuana and from 2.9 to 15.6% in Ciudad Juárez, depending on how recruitment patterns were modeled, and assumptions about how network size affected an individual’s probability of being included in the sample. RDS was an effective method to rapidly recruit IDUs in these cities. Although the frequency of HIV was low, syphilis prevalence was high, particularly in Tijuana. RDS-corrected estimates of syphilis prevalence were sensitive to model assumptions, suggesting that further validation of RDS is necessary. Introduction Injection drug use is a growing problem in cities located along the U.S.–Mexico border. Approximately 70% of U.S. cocaine originating in South America passes through the Central America–Mexico corridor.1 Cities and towns positioned on drug trafficking routes often experience epidemics of injection drug use.2,3 Injection drug users (IDUs) are at high risk of blood-borne infections, such as hepatitis C virus (HCV) and human immunodeficiency virus type-1 (HIV-1) infection, and of acquiring HIV-1 and other sexually transmitted infections (STIs) through high rates of unprotected sex.4–10 Mexico is currently considered a country of low HIV/AIDS prevalence11 (180,000 adult cases in 2005, a seroprevalence in the general population of 0.3%12), and the HIV epidemic has been mainly confined to men who have sex with men.13 Although injection drug use appears to have played only a minor role in the epidemic on a country-wide level,14 injection drug use appears to be increasingly important as a risk factor for HIV infection in some Mexican cities bordering the U.S. Viani et al.15 noted that the prevalence of HIV among pregnant women giving birth at Tijuana General Hospital rose from 0.29% in 1998 to 1.02% in 2001 and, in a subsequent study, showed that pregnant HIV-infected women were more likely to either inject drugs or to have a spouse/partner who injected drugs.16 In 2002, Valdez et al.17 reported that 21% of female sex workers in Ciudad (Cd.) Juárez injected illicit drugs, whereas a study by Patterson et al.18 in 2005 showed that over half of female sex workers in Cd. Juárez injected drugs, suggesting increasing overlap between sexual and IDU networks. Overlap between injection drug use and the trade of sex for money or drugs may contribute to elevated risk of STIs other than HIV, such as syphilis. Syphilis has been associated with higher HIV seroprevalence in a number of populations and is considered a cofactor of HIV transmission.19–22 In contrast to HIV, syphilis has been present in Mexico since at least the time of the Spanish invasion; however, the number of reported cases has decreased from 40,607 in 1945 (190.5 per 100,000) to 2,608 in 1990 (3.2 per 100,0000).23 Aggregate figures for syphilis prevalence belie the sub-epidemics occurring within specific risk groups. Several studies have been conducted with female sex workers in Mexico and have found varying syphilis prevalence levels. In 1990, 23.7% of 1,386 sex workers in four Mexican states had a reactive syphilis test.24 In 1993, testing of 826 sex workers in Mexico City showed an overall prevalence of 6.4%, with different syphilis rates associated with different patterns of sex work: 1.3% for massage parlor workers, 4.4% for bar girls, and 9.6% for streetwalkers.25 The prevalence of syphilis among 3,100 female sex workers tested at an AIDS clinic during 1992 and 1993 was 8.2%.26 In contrast, syphilis prevalence was low (2.3 and 1.1%) among gynecological outpatients in two Mexican cities between the years 1994 and 1995.27 However, little is known about syphilis prevalence in IDU populations in Mexico. In order to estimate the prevalence of HIV, HCV, and syphilis among IDUs, we conducted a cross-sectional study of IDUs in the border cities of Tijuana and Cd. Juárez, Mexico. Both cities are located on major drug trafficking routes and have large IDU populations (c. 6,000), with a similar sex ratio among the IDUs (c. 80% male).3 As stigma surrounding injection drug use makes it difficult to obtain a representative sample of injection drug users, we recruited individuals using respondent-driven sampling (RDS).28,29 By collecting data on individuals’ personal network sizes, RDS attempts to correct for biases in the sampling process, in order to obtain unbiased estimates of parameters such as the prevalence of a disease. In this study, we report on patterns of recruitment and the prevalence of HIV, HCV, and syphilis (and hepatitis B infection, for Cd. Juárez) in the context of sexual risk. Materials and Methods Study Population From February through April 2005, IDUs were enrolled in a cross-sectional study in Tijuana and Cd. Juárez, Mexico. Eligibility criteria for the study included: having injected illicit drugs within the past month, confirmed by inspection of injection stigmata (‘track marks’); aged 18 years or older; willing and able to provide informed consent; and not having been previously interviewed for the study. Subjects gave their written informed consent to participate in the study. Study methods were approved by the Institutional Review Board of the University of California, San Diego and the Ethics Board of the Tijuana General Hospital, which has one of the few federal-wide assurances in Mexico. Recruitment RDS methods were used to recruit participants.28,29 A diverse group of “seeds” (heterogeneous in age, gender, and geographic location) were selected to initiate the process. After providing informed consent, seeds underwent an interview, were educated on how to refer other eligible IDUs, and were given three uniquely coded coupons to refer their peers. Coupons were given to participants until approximately 150 participants were recruited in order to obtain a target sample size of approximately 200 per site. On each coupon, the study name, locations where they could participate, and a brief explanation was printed. In Cd. Juárez, interviews were conducted at a clinic run by Programa Compañeros, A.C., which is a trusted and well-respected non-governmental organization (NGO) that has been providing services to and conducting studies with IDUs in the city for decades. In Tijuana, staff from both COMUSIDA, the municipal HIV/AIDS program, and the Centro de Integración y Recuperación para Enfermos de Alcoholismo y Drogadicción “Mario Camacho Espíritu”, A.C. (CIRAD), an NGO that began working with drug users in 1991, made weekly trips to three geographically diverse ‘colonias’ (i.e., neighborhoods) in the city: Zona Norte, Grupo México, and Sepanal, using a modified recreational vehicle that operated as a mobile clinic (the ‘Prevemovihl’). Monetary reimbursements were given to participants to cover transportation costs and to compensate them for their time. The study staff in each site proposed the incentive levels based on their experience with this population and the incentives for previous studies. Participants in Cd. Juárez received $20 U.S. dollars (USD) for participation in the baseline visit and $5 USD when receiving laboratory test results at a one month follow-up visit. In Tijuana, $10 USD was given at baseline and $5 for the follow-up visit. In addition, participants at both sites were given $5 for each eligible person they recruited. These levels were not regarded as high. Data Collection Upon enrollment, trained staff administered quantitative surveys eliciting information on topics such as socio-economic and demographic profiles, drug use practices, sterile syringe access, barriers to sterile syringe use, experience with drug abuse treatment and incarceration history, health status, and HIV knowledge and testing history. We also asked about sexual behaviors and condom use with regular, casual and client partners of the opposite and same sex. Questions pertained to lifetime risk behaviors and those occurring in the prior six months. For RDS purposes, we measured network size using the question “En los últimos 6 meses, ¿cuántas personas conoce de nombre o de apodo que se han inyectado drogas?” (“In the past 6 months, how many people do you know by name or street name who have injected drugs?”). To determine the relationship between recruiter and recruitee, we asked “¿Cuál es su relación con la persona que le entregó el cupón?” (“What is your relationship to the person who gave you the coupon?”). Participants were given the choice of: “parientes” (relative); “pareja sexual” (sex partner/spouse); “amigo(a)” (friend); “conocido” (acquaintance); “desconocido” (stranger); and “otro” (other). To determine the size of individuals’ networks with respect to injection drug use, we asked “En los últimos 6 meses, ¿con cuántas personas diferentes acostumbra inyectarse?” (“In the last 6 months, on average how many different people did you usually inject with?”). After the interview, blood was drawn for antibody testing of HIV, HCV, HBV (Cd. Juárez only), and syphilis. Pre- and post-test counseling, and referral to treatment where indicated, was provided to all participants. Laboratory Samples Blood samples were obtained by venipuncture and serum was stored at the municipal health clinic in Tijuana or Cd. Juárez before being shipped frozen to the San Diego County public health laboratory or New Mexico State Laboratory, respectively. All participants were screened on-site in Mexico for HIV with the Determine rapid test (Abbott Laboratories). For the Tijuana samples, in the event of an HIV-positive or indeterminate test, results were confirmed with a Western blot, HIV enzyme immunoassay (EIA), and HIV immunofluorescence assay. For samples from Cd. Juárez, the HIV EIA was conducted on all samples, and a confirmatory Western blot was performed on positive or indeterminate sample. Cd. Juárez samples were also tested for hepatitis B antigen (Genetic Systems HBsAg EIA 3.0, Bio-Rad Laboratories) and antibody (DiaSorin ETI-AB-COREK PLUS). All samples were tested for syphilis with the rapid plasma reagin (RPR) test (Macro-Vue, Becton Dickenson) and if reactive, confirmed by a Treponema pallidum particle agglutination assay (TPPA; Fujirebio Diagnostics). Statistical Methods Obtaining estimates of population proportions of groups using RDS involves combining three kinds of data: the sample proportion of each group, the crosstabulation of groups between pairs of recruiters and recruitees, and differences in network size between groups. To estimate equilibrium proportions of different groups, and to estimate the pattern of mixing between groups, we assumed that the recruitment process followed a first order Markov process.28,29 Under this model, the relationship between the state of the recruiter and recruitee can be modeled using log-linear models applied to a two-way table of counts.30,31 We classified individuals by sex and syphilis seropositivity and fitted a series of hierarchical log-linear models of increasing complexity to the data to determine patterns of nonrandom mixing between groups along each recruitment tree, choosing the best model as that which had the lowest value of Akaike’s Information Criterion.32 For the purposes of analysis, we considered all individuals with positive syphilis tests based upon RPR, and did not classify individuals further into those with TPPA titers greater than or equal to 1:8 (who may represent infectious cases) and those with titers of 1:1 to 1:4 (who may represent past infection). To derive RDS-corrected estimates of syphilis seropositivity in men and women in Tijuana and Ciudad Juárez, we estimated recruitment weights for each group (as the ratio of the equilibrium to sample proportions of each group). We estimated the equilibrium fraction as previously described.29 We used both raw counts and predicted counts based on the best fitting log-linear model. Degree weights were estimated using linear least squares.29 We used both unadjusted and adjusted estimates of personal network size.33 An overall sampling weight was derived for each group, from which population-level estimates were obtained. Pre-processing of the data was performed using Stata v. 8.2 (Stata Corporation, College Station, TX). Networks and trees were generated using scripts written in Python and visualized using GraphViz (AT&T Research, Florham Park, NJ). Statistical analyses and summary statistics of the recruitment network were generated in R,34 and RDS based corrections were calculated using Maxima (http://maxima.sourceforge.net). We chose to develop our own programs rather than use RDSAT (http://www.respondentdrivensampling.org) primarily to familiarize ourselves with the statistical theory underlying RDS-based corrections. All code is available from the first author on request. Results Study Population Table 1 summarizes some basic data relating to the Tijuana (15 seeds, 207 recruits) and Cd. Juárez (9 seeds, 197 recruits) study populations. Both populations were predominantly male, with participants in their early to mid-30s. Crude HCV seroprevalence was extremely high (>95%) in both cities. Hepatitis B seroprevalence was only determined for Cd. Juárez, where it was high (85% overall); only one individual was positive for HBV antigen. Crude HIV seroprevalence was low, but the crude prevalence of syphilis was high, especially among women. Table 1.Summary statistics of age, parameters pertaining to risk of STI, and seroprevalence of HIV, HCV and syphilis, by city, sex, and by whether individuals were seeds, or recruitsCityTijuanaCiudad JuárezType of recruitSeedsRecruitsSeedsRecruitsGender (n)Male (9)Female (6)Male (194)Female (13)Male (6)Female (3)Male (184)Female (13)Age (median, range)35 (29–54)28 (22–32)34 (18–56)32 (24–48)42 (33–49)33 (30–42)33 (18–62)36 (20–48)Age at sexual debut (median, range)15 (12–18)16.5 (9–27)15 (7–28) [n=191]14 (11–15)13.5 (8–18)16 (15–18)15 (5–23) [n=181]14 (12–18)No. of lifetime male sexual partners (median, range)2 (0–15)8 (0–10)0 (0–100) [n=186]20 (2–500) [n=11]0 (0–2)5 (3–6)0 (0–15) [n=178]5 (1–100) [n=11]No. of lifetime female sexual partners (median, range)10 (4–300)0 (0–1)10 (0–500) [n=191]0 (0–4)22 (8–200)5 (0–10) [n=2]10 (1–300) [n=175]0 (0–7) [n=12]Prostitution as main source of income over the last 6 months (fraction, %)0/9 (0)1/6 (16.7)1/191 (0.5)6/13 (46.2) 0/6 (0)1/3 (33)0/184 (0)1/13 (7.7)Had sex in last 6 months (fraction, %)5/9 (55.6)4/6 (66.7)106/194 (54.6)12/13 (92.3)4/6 (66.7)2/3 (66.7)108/184 (58.7)10/13 (76.9)Ever been given (bought) sex in the last 6 months (fraction, %)1/3 (33.3)2/3 (66.7)24/67 (35.8)8/8 (100)0/2 (0)0/0 (N/A)10/57 (17.5)4/5 (80.0)Ever given (sold) sex in the last 6 months (fraction, %)1/3 (33.3)2/3 (66.7)17/67 (25.4)2/8 (25)1/2 (50)0/0 (N/A)23/57 (40.4)3/5 (60.0)Injected drugs with sex partner in last 6 months1/9 (11.1)2/6 (33.3)10/194 (5.15)6/13 (46.2)0/6 (0)1/3 (33.3)15/184 (8.2)3/13 (2.3)Syphilis RPR + (fraction, %)1/7 (14.3)2/6 (33.3)20/194 (10.3)8/13 (61.5)0/4 (0)2/3 (66.7)3/175 (1.71)2/11 (18.2)HIV antibody (fraction, %)1/7 (14.3)1/6 (16.7)4/194 (2.1)0/13 (0)0/6 (0)0/3 (0)8/182 (4.4)0/12 (0)HCV antibody (fraction, %)6/7 (85.7)6/6 (100)188/194 (96.9)12/13 (92.3)6/6 (100)3/3 (100)174/181 (96.1)10/12 (83.3)HBV antibody (fraction, %)N/AN/AN/AN/A6/6 (100)3/3 (100)154/181 (85.1)8/12 (66.7) Recruitment Dynamics RDS was an effective means of recruiting IDUs in both cities. The number of individuals recruited increased rapidly following the first interview, especially in Cd. Juárez (Figure 1a), where many individuals interviewed the same day as their recruiter (Figure 1b). Apart from these differences in the tempo of recruitment, patterns of recruitment were very similar between the two cities; recruitment was highest in the fourth wave of recruitment, with some individuals being recruited after eight waves, suggesting that despite rapid recruitment, good sociometric depth was obtained (Figure 1c). After excluding individuals who were not given coupons, the number of recruits per recruiter showed a bimodal distribution, with many individuals either recruiting zero or three recruits (Figure 1d), suggesting the presence of a mixed population of ineffective and effective recruiters. In both cities, approximately one half of participants were recruited via referral trees originating from two seeds (Figure 1e). The relationship between recruiter and recruit was usually ‘friend,’ ‘acquaintance,’ or another close relationship such as a family member or a sex partner (Figure 1f), which is important as RDS-based estimates assume that these relationships are reciprocal.28,29Figure 1Summary of the dynamics of recruitment, by city. (a) The cumulative number of recruits over time. (b) The interval between the interview of the recruiter and that of their recruitee (omitting those individuals who did not recruit). (c) The number of recruits in each recruitment wave from the seed. (d) The number of recruits per recruiter (excluding individuals who were not given any coupons). (e) The number of recruits from each seed. (f) The relationship between recruiter and recruitee. Recruitment Trees Figure 2 shows a ‘forest’ of recruitment trees for each city, with the syphilis antibody status indicated by shading and the gender of each individual indicated by different symbols. This figure illustrates that the sample prevalence of syphilis was higher in Tijuana than in Cd. Juárez and that syphilis cases appeared to cluster in the recruitment trees. Although there was a low frequency of women in the sample, a disproportionate number of women also had syphilis. Figure 2Recruitment networks (strictly speaking, a forest of recruitment trees) for the RDS based samples of IDUs in (a) Tijuana and (b) Ciudad Juárez. Seeds are shown at the top of the figure, and arrows indicate the direction of recruitment. Syphilis serostatus is shown by shading: black- syphilis antibody positive, white- syphilis antibody negative, gray- missing data. The gender of participants is indicated by the shape of the symbol: square for female and circle for male. The size of the symbol is related to the reported network size: the larger the symbol, the larger the network size. Symbols marked with an ‘×’ denote individuals who were given coupons, but did not recruit. To test whether syphilis prevalence differed by sex and whether cases of syphilis were clustered, we analyzed the relationship between the sex and syphilis status of the recruiter and recruitee in Tijuana and Cd. Juárez (Table 2). A set of hierarchical log-linear models were used to test whether there was an association between syphilis seropositivity and sex and between the state of the recruiter and that of the recruitee, with the best fitting model chosen using Akaike’s Information Criterion, AIC,32 with lower values indicating better fit (Table 3). This approach is functionally equivalent to calculating the RDS homophily index, for which 1 denotes a perfect association, 0 a zero association, and −1 a perfect negative association. The reason for using log-linear models is that they offer a solid statistical framework with which we can compare different models of association between the characteristics of the recruiter and those of the recruitee. Table 2Relationship between recruiter and recruitee in terms of sex and syphilis antibody status TijuanaCiudad JuárezFromToMaleFemaleMaleFemaleRPR−RPR+RPR−RPR+RPR−RPR+RPR−RPR+MaleRPR−145 (145.2)15 (14.66)2 (2.33)5 (4.82)147 (146.7)3 (2.30)7 (7.43)1 (1.53)RPR+15 (15.23)5 (4.91)1 (0.24)1 (1.61)1 (1.59)0 (0.2)0 (0.08)1 (0.13)FemaleRPR−8 (7.19)0 (0.11)1 (1.29)0 (0.41)7 (6.50)0 (0.10)0 (0.33)0 (0.07)RPR+6 (6.38)0 (0.32)1 (1.14)2 (1.16)3 (3.17)0 (0.40)1 (0.16)0 (0.27)Smoothed estimates are shown in parentheses and are based upon the best fitting log linear model for each sample (Table 3).Table 3.Fit of 12 log-linear models to the data shown in Table 2ModelpTijuanaCiudad JuárezlnLAICLRT, G2lnLAICLRT, G2Symmetric(sex1,sex2) + Symmetric(rpr1,rpr2)5−42.0694.1240.4 (p<0.001)−28.8467.6829.3 (p=0.001)sex1 + rpr1 + sex2 + rpr25−46.27102.548.8 (p<0.001)−29.9169.8231.4 (p=0.001)sex1 + rpr1 + sex2 * rpr26−37.2786.5530.8 (p=0.001)−27.2566.4926.1 (p=0.004)sex1 * rpr1 + sex2 * rpr27−30.3274.6416.9 (p=0.05)−19.2152.4310.0 (p=0.35)sex1 * rpr1 + sex2 * rpr2 + sex1 * sex28−27.4770.9511.2 (p=0.189)−19.1254.249.81 (p=0.28)sex1 * rpr1 + sex2 * rpr2 + rpr1 * rpr28−28.3172.6112.9 (p=0.115)−18.1652.327.89 (p=0.44)sex1 * rpr1 + sex2 * rpr2 + sex1 * sex2 + rpr1 * rpr29−26.0170.028.31 (p=0.306)−18.1454.287.85 (p=0.35)sex1 * rpr1 + sex2 * rpr2 + sex1 * sex2 + rpr1 * rpr2 + sex1 * rpr210−23.8767.734.02 (p=0.674)−16.4852.964.53 (p=0.61)sex1 * rpr1 + sex2 * rpr2 + sex1 * sex2 + rpr1 * rpr2 + rpr1 * sex210−25.6371.277.56 (p=0.272)−16.1752.353.92 (p=0.69)sex1 * rpr1 + sex2 * rpr2 + sex1 * sex2 + rpr1 * rpr2 + sex1 * rpr2 + rpr1 * sex211−23.6769.343.62 (p=0.604)−16.6553.294.87 (p=0.56)Symmetric(sex1 * rpr1, sex2 * rpr2)10−25.9171.828.1 (p=0.23)−15.0352.061.63 (p=0.90)Saturated model (sex1 * rpr1 * sex2 * rpr2)16−21.8675.71NA−14.2160.43NAThe variables in the ‘Model’ column correspond to the sex of the recruiter (sex1) and recruitee (sex2) and the syphilis serostatus of the recruiter (rpr1) and recruitee (rpr2). The term ‘Symmetric(a, b)’ applies to a table which is symmetric about the diagonal of the table a by b. The number of parameters is denoted p, the log-likelihood of the model (higher is better) by lnL, Akaike’s Information Criterion (lower is better) by AIC, and a likelihood ratio test for goodness of fit against a saturated model by LRT. The best fitting model for each city, based on the lowest AIC, is indicated by underlining. For both samples, the best fitting model included a significant (positive) association between the syphilis serostatus of the recruiter and the syphilis serostatus of the recruitee and was not significantly different from a ‘saturated’ model, in which each cell in the table is modeled with a single parameter. For the Tijuana sample, there was a positive association between the sex of the recruiter and the sex of the recruit, and, independently of this association, syphilis antibody positive women were disproportionately less likely to recruit syphilis positive men than syphilis negative men. Network Size There was great variation in both measures of network size, with the average number of known IDUs an order of magnitude higher than the number of injecting partners. There was no correlation in network size between recruiter and recruitee, measured using either the number of IDUs known (Spearman’s rho = 0.0328 and 0.0204 for Tijuana and Cd. Juárez, respectively) or the number of injecting partners as a measure of network size (Spearman’s rho = −0.006 and −0.0121 for Tijuana and Cd. Juárez, respectively; Figure 3a, b). Network sizes were similar when the sample was grouped by syphilis serostatus and sex (Figure 3c, d). As chain referral samples are biased towards individuals with larger network sizes, we adjusted the distribution of network size by weighting the distribution of network sizes by the inverse of the network size.33 This reweighting led to a significant drop in estimated network size, from a median of 25–50 to 4–5 in Tijuana and from 15–20 to 8–10 in Cd. Juárez (Figure 3e). Figure 3Summary of the network size distribution. (a) Scatterplot of the number of IDUs known by name or street name between recruiter and recruitee. (b) Scatterplot of the number of injecting partners between recruiter and recruitee. (c) Boxplot of the number of IDUs known by sex (M male, F female) and by syphilis antibody status (+ positive, − negative). (d) Boxplot of the number of injecting partners by sex (M male, F female) and by syphilis antibody status (+ positive, − negative). (e) Cumulative distribution of the number of IDUs known before (solid line) and after (dashed line) adjusting for biased sampling of individuals. Correcting for Sampling Bias Information on network size, and on who recruited whom, collected as part of RDS, allows population estimates to be generated from the sample, despite biases in the sampling process. To do so, we need to estimate the pattern of mixing between different groups along the recruitment network and to determine how different the network size is in each group. The seroprevalence of HIV was too low, and the seroprevalence of HCV was too high in order to obtain meaningful correction factors; hence we concentrated on obtaining population estimates of the prevalence of syphilis antibody in men and women in the two cities. In order to obtain unbiased population estimates of the prevalence of syphilis antibody, we used a poststratification process, in which we calculated recruitment weights and degree weights, which can be combined to give an overall sampling weight.46 As RDS is a chain referral method, the prevalence of syphilis in the sample may have been different if recruitment had continued for further waves. Assuming that recruitment follows a first order Markov process, the ‘equilibrium prevalence’ of syphilis can be calculated from the cross-tabulations of the syphilis status of the recruiter and the recruitee. Using these estimates, we calculated recruitment weights, as the ratio of the equilibrium to sample frequencies of each group (Table 4); the closer these weights are to 1, the more representative the sampling of the group. For the Tijuana sample, recruitment weights were close to 1 for three groups, except for syphilis negative women, whose recruitment weight was 0.87 (i.e., this group was oversampled). For the Cd. Juárez sample, recruitment weights were close to 1, with the exception of syphilis positive women (1.25), who were undersampled. Table 4.RDS-corrected estimates of syphilis seroprevalence, using the raw transition data, and adjusted degreesCityTijuanaCiudad JuárezSexMaleFemaleMaleFemaleSyphilis antibodySyphilis RPR−Syphilis RPR+Syphilis RPR−Syphilis RPR+Syphilis RPR−Syphilis RPR+Syphilis RPR−Syphilis RPR+n1742058158382Sample (%)84.169.662.423.8692.41.754.681.17Equilibrium (%)84.279.82.113.8292.31.754.461.46Recruitment weight11.010.870.99110.951.25Degree54.9664.83059.1247.5419.3345.1115Adjusted degree14.638.4518.294.8914.6714.418.6213.33Adjusted degree weight0.861.670.372.750.9810.713.16Adjusted sampling weights0.861.690.322.720.9810.673.95Adjusted population (%)72.3716.350.7710.590.471.753.164.62 Samples obtained using RDS can also be biased due to differences in network size between the groups. To compensate for this effect, we calculated degree weights, based upon the reported network size of known IDUs using the adjusted mean estimates of network size33 in each group (Figure 3f). After calculating degree weights for each group, and multiplying them by recruitment weights to generate an overall sampling weight, the RDS corrected estimates of syphilis antibody prevalence were higher than those in the overall sample for both cities (Table 4). Sensitivity of RDS Estimates To determine the sensitivity of point estimates of the prevalence of syphilis antibody to modeling assumptions, we also obtained estimates using unadjusted rather than adjusted network sizes and using a ‘smoothed’ transition matrix based on the best fitting log-linear model, rather than the raw counts. We found that RDS-based estimates were highly sensitive to these assumptions (Table 5). Estimated syphilis seroprevalence ranged from 12.4 to 26.8% in Tijuana and from 2.9 to 15.6% in Cd. Juárez, depending on how the pattern of recruitment was modeled and how reported network size was assumed to affect an individual’s probability of being included in the sample. However, our results suggest that syphilis seroprevalence is higher among women than men and higher in Tijuana than in Cd. Juárez and that sample proportions of syphilis using RDS in these populations may be underestimates of the true population seroprevalence. Table 5.Sensitivity of the estimated population prevalence of syphilis antibody among men and women to model assumptionsModel assumptionsPopulation prevalenceTijuanaCiudad JuárezMaleFemaleMaleFemaleTransition matrixNetwork sizesSyphilis RPR−Syphilis RPR+Syphilis RPR−Syphilis RPR+Syphilis RPR−Syphilis RPR+Syphilis RPR−Syphilis RPR+UnadjustedUnadjusted84.799.342.083.880.613.763.7611.86UnadjustedAdjusted72.3716.350.7710.590.471.753.164.62SmoothedUnadjusted84.818.982.753.4587.943.974.693.41SmoothedAdjusted73.3515.941.049.6793.281.743.721.25Smoothed transition matrices are obtained by using the fitted values from the best fitting model in Table 3. Adjusted network sizes are obtained as described in Salganik and Heckathorn.33 Conclusions Respondent driven sampling offers the promise of a probability sample of individuals from hidden and hard-to-reach populations. RDS was originally developed in the context of recruiting IDUs35–39 and, in our context, was an efficient method to recruit IDUs in two Mexican cities bordering the U.S. Recruitment was extremely rapid in Cd. Juárez compared to Tijuana, which may be due to greater access to the study site, higher monetary incentives, and that Programa Compañeros is more established in Cd. Juárez than CIRAD in Tijuana and had carried out studies in the past with monetary renumeration. In contrast, Mueller et al.40 report much slower recruitment of IDUs in Las Cruces, NM using RDS, despite similar methodology and the same eligibility criteria. The sample seroprevalence of HIV was relatively low in both cities. HIV-1 seroprevalence in IDUs in Tijuana recruited through RDS was similar to that found in IDUs studied by Güereña-Burgueño et al.41 in the early 1990s and in a study by Magis-Rodriguez et al.42 in 2003 that used time-location sampling methods, although the absolute number of HIV-positive cases was too low to perform reliable RDS corrections. In contrast to HIV, syphilis prevalence was extremely high, especially in women and in Tijuana. Unlike many adaptive sampling schemes43 in which the sampling process is controlled by the investigator, RDS enables the study subjects to control the sampling process. While this facilitates the recruitment process, it makes statistical inference more difficult. We found that estimates of syphilis seroprevalence were extremely sensitive to modeling assumptions. First, as recruitments between low-frequency groups are relatively rare, estimates of the recruitment rates may be biased. Smoothing these estimates using a statistical model can lead to different estimates. Although simulations and analytical results show that RDS-based estimates are unbiased in large populations, errors in RDS based estimates may be so high for small populations and/or low frequencies of groups as to render the use of RDS impractical.33 Secondly, the estimated prevalence of syphilis was sensitive to the assumption of how inclusion probability depends on reported network size. Estimates of network size may well have been different had we asked “How many people do you currently know by name or street name that inject drugs?” Estimating group-level network sizes is compromised by high variances, the small size of some of the subpopulations, and the poor ability of individuals to estimate the size of their personal networks.44 Although RDS controls for differences in network sizes, a sampling bias long known to be inherent in chain-referral samples, it is important that this information is as accurate as possible. It might be argued that prior to the advent of RDS, there was little incentive to accurately measure relative network sizes in epidemiological studies; given that this information plays a crucial role in the post-stratification process of RDS, we encourage further research to determine how best to collect this information accurately. RDS also has some inherent limitations in terms of inferences that can be drawn from the data45: it does not generate estimates of the absolute size of the population, only proportions, and it exploits social ties between individuals, limiting what one can conclude about sexual or drug-injecting networks from RDS data. Furthermore, without comparison of RDS to other types of sampling, we cannot conclude that obtaining a sample through RDS gives us a more representative sample than other methods. Nevertheless, RDS, or a modified version thereof, has the potential to efficiently recruit hidden populations such as IDUs, and creates avenues through which interventions can reach members of these populations. In the context of this study, prevention and treatment of syphilis is clearly an important public health concern.	[ "injection drug users", "hiv and syphilis prevalence", "respondent driven sampling." ]	[ "P", "P", "R" ]
Appl_Microbiol_Biotechnol-4-1-2270352	Lactic acid production from lime-treated wheat straw by Bacillus coagulans: neutralization of acid by fed-batch addition of alkaline substrate	Conventional processes for lignocellulose-to-organic acid conversion requires pretreatment, enzymatic hydrolysis, and microbial fermentation. In this study, lime-treated wheat straw was hydrolyzed and fermented simultaneously to lactic acid by an enzyme preparation and Bacillus coagulans DSM 2314. Decrease in pH because of lactic acid formation was partially adjusted by automatic addition of the alkaline substrate. After 55 h of incubation, the polymeric glucan, xylan, and arabinan present in the lime-treated straw were hydrolyzed for 55%, 75%, and 80%, respectively. Lactic acid (40.7 g/l) indicated a fermentation efficiency of 81% and a chiral l(+)-lactic acid purity of 97.2%. In total, 711 g lactic acid was produced out of 2,706 g lime-treated straw, representing 43% of the overall theoretical maximum yield. Approximately half of the lactic acid produced was neutralized by fed-batch feeding of lime-treated straw, whereas the remaining half was neutralized during the batch phase with a Ca(OH)2 suspension. Of the lime added during the pretreatment of straw, 61% was used for the neutralization of lactic acid. This is the first demonstration of a process having a combined alkaline pretreatment of lignocellulosic biomass and pH control in fermentation resulting in a significant saving of lime consumption and avoiding the necessity to recycle lime. Introduction Lactic acid is used throughout the world in manufacturing of food, chemicals, and pharmaceutical products. Recently, there is a lot of interest in biodegradable poly-lactic acid, which is an alternative to petrochemically derived plastic (Drumright et al. 2000). Chiral pure lactic acid is produced commercially by microbial fermentation of the carbohydrates glucose, sucrose, lactose, and starch/maltose derived from feedstocks such as beet sugar, molasses, whey, and barley malt (Narayanan et al. 2004). The choice of feedstock depends on its price, availability, and on the respective costs of lactic acid recovery and purification (Datta et al. 1995; Vaidya et al. 2005). As an alternative to these traditional feedstocks, lignocellulosic biomass is an inexpensive and widely available renewable carbon source that has no competing food value. Lignocellulose consists primarily of cellulose and hemicellulose; polymers build up of mainly hexose sugars and pentose sugars, which are embedded in a matrix of the phenolic polymer lignin. The main pathway to derive fermentable sugars from lignocellulose is through enzymatic hydrolysis by cellulolytic and hemicellulolytic enzymes. A mechanical and chemical pretreatment of the lignocellulose is required to reduce particle size, to modify and/or to remove the lignin, and with that to enhance the accessibility of the polysaccharides for enzymatic hydrolysis (Claassen et al. 1999). Various chemical pretreatments of biomass have been studied in research and development of lignocellulose-to-ethanol production technology (Mosier et al. 2005). One is the use of lime (calcium hydroxide) at relatively mild temperature conditions (Chang et al. 1998). Lime as a pretreatment agent has promising potential because it is inexpensive, safe, and its use hardly results in sugar degradation products such as furfural and hydroxymethyl furfural. Nevertheless, this alkaline pretreatment features a relatively high pH value (>10) of the treated biomass, and at these pH levels, the activity of common cellulolytic and xylanolytic enzymes, necessary for the depolymerization of (hemi)-cellulose, is negligible low. Therefore, lowering the pH is essential to achieve an efficient enzymatic hydrolysis of the polysaccharides. One approach to remove calcium hydroxide is by washing the lime-treated biomass before enzymatic hydrolysis (Chang et al. 1998); however, this leads to the use of high amounts of water. Another way to lower the pH of the pretreated material is by neutralizing calcium hydroxide with sulfuric acid. Yet, this results in the formation of the low value byproduct gypsum. As an alternative improvement to these approaches, we propose to use the calcium hydroxide present in lime-treated biomass as neutralizing agent for organic acids produced in microbial fermentation processes. To examine this proposed concept, lime-treated wheat straw (LTWS) was added fed-batch-wise during a simultaneous saccharification and fermentation (SSF) process in a 20-l controlled stirred fermenter containing hydrolytic enzymes and Bacillus coagulans DSM 2314, a thermophilic bacterium capable to convert both hexoses and pentoses homofermentative to l(+)-lactic acid (Otto 2004; Patel et al. 2006). The objective of this research was to evaluate whether high alkaline-treated lignocellulosic biomass (without neutralization) can be used directly in a SSF process by (1) providing a carbon source for enzymatic hydrolysis and fermentation and (2) providing a source of alkali to control the pH in the fermentation process. Materials and methods Feedstock and pretreatment Wheat straw was selected as a lignocellulose model feedstock and was purchased from a farm in the Northeast of The Netherlands. The wheat straw was air dried (89.5% [w/w] dry matter [DM]) and ground through a 2-mm screen. The lime pretreatment was performed by filling two 15-l mixers (Terlet, The Netherlands), both with 1,650 g ground wheat straw, 13 kg tap water, and 165 g calcium hydroxide. This wheat straw suspension was heated and kept at 85°C for 16 h under continuously stirring at 30 rpm. The LTWS suspension was subsequently cooled to 30°C, dehydrated by placing the LTWS in a cotton bag, and pressing the suspension using a manual piston press at pressure up to 9.7 kg/m2. After dehydration, an amount of 11.45 kg LTWS with an average DM content of 27.0% (w/w) and pH 11.8 was obtained and served as substrate for further experiments. The chemical composition of LTWS was determined as described by van den Oever et al. (2003). Enzyme preparation The enzyme preparation GC 220 (Genencor-Danisco, Rochester, USA) containing cellulase, cellobiase, and xylanase activity of 116, 215, and 677 U/ml, respectively (Kabel et al. 2006), was used for this study. The preparation had a specific gravity of 1.2 g/ml and contained 4.5 mg/ml glucose, 2.9 mg/ml mannose, and 0.8 mg/ml galactose. Microorganism and preculture The bacterium B. coagulans strain DSM 2314 was used as the lactic acid-producing micro-organism. Bacterial cells were maintained in a 10% (w/w) glycerol stock solution and stored at −80°C. Chemicals, unless indicated otherwise, were purchased from Merck (Darmstadt, Germany). Gelrite plates were prepared with a medium containing (per liter): glucose, 10 g; Gelrite, 20 g (Duchefa, Haarlem, The Netherlands); yeast extract, 10 g (Duchefa); (NH4)2HPO4, 2 g; (NH4)2SO4, 3.5 g; Bis–Tris, 10 g (USB, Ohio, USA); MgCl2 6H2O, 0.02 g; and CaCl2.2H2O, 0.1 g. Glucose and Gelrite were dissolved in stock solution A (four times concentrated). The pH of this stock solution was adjusted to 6.4 with 2 M hydrochloric acid and autoclaved for 15 min at 125°C. The remaining nutrients were dissolved in stock solution B (1.33 times concentrated), which was also adjusted to pH 6.4 with 2 M hydrochloric acid but was filter sterilized (cellulose acetate filter with pore size of 0.2 μm, Minisart, Sartorius). After sterilization, the medium was prepared by combining stock solutions A and B and Gelrite plates were poured. The bacteria were cultivated on Gelrite plates for 48 h at 50°C. An isolated colony was used to inoculate a 100-ml broth with similar composition and preparation as described above but without the addition of Gelrite. The culture was incubated statically for 24 h at 50°C and functioned as the inoculum for a 1,400-ml broth. This culture was incubated also statically for 12 h at 50°C and served as a 10% (v/v) preculture for the SSF experiments. Simultaneous saccharification and fermentation The SSF of LTWS was carried out in a 20-l fermenter (Applikon, Schiedam, The Netherlands) with pH and temperature control (biocontroller ADI 1020). At the start of SSF, the fermenter was filled with 6.0 kg tap water and 1,400 g dehydrated LTWS (DM content of 27.0% [w/w]). The following nutrients were then added to the LTWS suspension: yeast extract, 150 g (Duchefa); (NH4)2HPO4, 30 g; (NH4)2SO4, 52.5 g; MgCl2 6H2O, 0.3 g; and CaCl2 2H2O, 1.5 g. The LTWS suspension was then heated to 50°C, and the pH was adjusted to 6.0 with 101 g 3 M sulfuric acid (~30 g H2SO4). The SSF process of LTWS to lactic acid consisted of three phases: (1) the prehydrolysis phase of preloaded LTWS, (2) the fed-batch phase with automatic feeding of LTWS from a screw feeder, and (3) the batch phase with pH control by a calcium hydroxide suspension and no LTWS feeding. A schematic representation of the experimental setup is shown in Fig. 1. The prehydrolysis was initiated by the addition of 40 ml enzyme preparation (88 mg enzyme/g DM substrate) to the LTWS suspension and was incubated for 2 h at 50°C under continuously stirring at 250 rpm. The fed-batch phase was initiated by the addition of 1,500-ml preculture of B. coagulans DSM 2314 to the fermenter. The lactic acid produced by the bacteria was neutralized by the automatic addition of 8,623 g dehydrated LTWS (DM of 27.0%) to the fermenter through a feeder (K-Tron Soder Feeders, Canada) and was regulated by the pH of the medium, which was set at 6.0. Throughout the fed-batch phase, an amount of 280 ml of enzyme preparation (total enzyme loading of 98 mg/g DM substrate) was added proportional to the LTWS addition rate into the fermenter. During the batch phase, the pH was controlled at 6.0 by the addition of 20.0% (w/v) calcium hydroxide suspension. Samples were withdrawn for DM, substrate, and (by)product analysis. Fig. 1Schematic representation of the simultaneous saccharification and fermentation of lime-treated wheat straw to lactic acid Analytical methods For the analysis of monomeric sugars, the fermentation broth samples were centrifuged (3 min at 17,400 × g), and the pH of the supernatant was adjusted to 5.0 with barium carbonate using a pH indicator (Bromophenolblue) followed by filtration of the liquid. The analysis was performed by high-performance anion-exchange chromatography using a Carbopack PA1 column (column temperature of 30°C) and a pulsed amperometric detector (ED50; Dionex, Sunnyvale, CA, USA). Before injection, the system was equilibrated with 25.5 mM NaOH for 10 min at a flow rate of 1.0 ml/min. For the separation of monomeric sugars, at injection, the mobile phase was shifted to deionized water for 30 min. Postcolumn addition of sodium hydroxide was used for detection of the neutral monomeric sugars. The determination of soluble oligomeric sugars was performed by centrifugation for 5 min at 3,000 rpm (Centaur 2, Beun de Ronde, The Netherlands) of preweighed samples and freeze drying the supernatant overnight. Pellets were weighed and hydrolyzed with sulfuric acid, and neutral monomeric sugars were determined according to the method as described by van den Oever et al. (2003). For the calculations, an average molecular weight of oligomers from glucan and xylan of 166 and 132 g/mol, respectively, were applied, resulting in a hydrolysis factor of 1.08 and 1.14, respectively. For the analysis of insoluble polymeric sugars, samples of 25 g were centrifuged for 5 min at 3,000 rpm (Centaur 2, Beun de Ronde); the supernatant was removed, and the pellet was washed by resuspension in 25 ml fresh demineralized water followed by a centrifugation step of 5 min at 3,000 rpm (Centaur 2, Beun de Ronde). The sequence of resuspension and centrifugation was repeated three times. After the last removal of the supernatant, the pellets were freeze dried overnight. The pellets were weighed (values used for DM calculation), polymeric material was hydrolyzed with sulfuric acid, and neutral sugars were analyzed according to the method as described by van den Oever et al. (2003). For the calculations, a molecular weight of glucan and xylan of 162 and 132 g/mol, respectively, were applied, resulting in a hydrolysis factor of polymer to monomer of 1.11 and 1.14, respectively. The analysis of organic acids was performed by high-pressure liquid chromatography according to the procedure described by Maas et al. (2006). The chiral purity (%) of lactic acid was determined by derivatization of all lactates using methanol, after which both enantiomers of methyl lactate were separated on a chiral gas chromatography column and detected using a flame ionization detector. The chiral purity was expressed as the area of the main enantiomer divided by the sum of areas of both enantiomers. Calculations The theoretical maximum lactic acid (LAtheor. max. [g]) production was calculated according the following equation (Eq. 1): where DMsubstrate = the total dry matter of substrate LTWS (g), Fpolysacch. = fraction polysaccharides per substrate (g/g), HFmonsacch./polysacch. = hydrolysis factor of polysaccharides, incorporation of water results in 1.11 g hexose from 1.00 g glucan and 1.14 g pentose from xylan and arabinan (g/g), and FF = fermentation factor of 1.00 g lactic acid per gram of monomeric sugar. The efficiency of the enzymatic hydrolysis (%, w/w) was based on the amount of hydrolyzed polysaccharides (g; calculated by the difference between initial amounts and analyzed insoluble amounts) divided by the amount of polysaccharides (g) initially present in the substrate. The fermentation efficiency (%, w/w) is expressed as the amount of lactic acid produced (g) divided by the amount of monomeric sugars consumed (g) by the bacteria. The overall efficiency of the SSF (%, w/w) was calculated by the amount of lactic acid produced (g) divided by the theoretical maximum amount of lactic acid (g) determined as described in Eq. 1. Results Simultaneous saccharification and fermentation of LTWS to lactic acid The polysaccharide composition of the LTWS consisted mainly of glucan, xylan, and arabinan of 33.0%, 19.0%, and 2.0% (w/w), respectively, whereas the remaining mass constituted of lignin, ash, extractives, and uronic acids. Some of the soluble components in wheat straw were partially removed by the solid/liquid separation (dehydration) of the LTWS. The focus of this study was on the conversion of glucan, xylan, and arabinan, which are the predominant polysaccharides present in LTWS and accounted for 99.8% (w/w) of the total polymeric sugars. Previous work showed that the cellulase preparation GC 220, used for the saccharification of polysaccharides, functioned optimally at 50°C and pH 5.0 (Maas et al., submitted for publication), whereas growth conditions for B. coagulans DSM 2314 were 54°C and pH 6.5 (Otto 2004). In this study, both the enzymatic hydrolysis and the fermentation occurred simultaneously in the same reactor at compromising conditions, which were set at 50°C and pH 6.0. The SSF of LTWS to lactic acid was studied in a 20-l controlled stirred fermenter. Previous results showed that when this process was performed without a prehydrolysis of an initial amount of LTWS, the concentration of monomeric sugars was low and resulted, therefore, in relatively low lactic acid productivity. As a consequence, the fed-batch addition rate of the alkaline substrate to neutralize the produced lactic acid was low (results not shown). To start the fermentation with a substantial initial amount of fermentable sugars (>2 g/l), a prehydrolysis of 378 g LTWS and enzyme preparation (88 mg per g DM LTWS) in approximately 6 l volume at pH 6.0 for 2 h was introduced. This resulted in glucose, xylose, and arabinose concentrations of 2.0, 0.4, and 0.3 g/l, respectively (Fig. 3a). The second phase (II) was initiated by introducing a 1,500-ml preculture of B. coagulans DSM 2314. A minor amount of lactic acid produced in the preculture caused a slight pH decrease and was automatically neutralized by the addition of LTWS (Fig. 2a,b). After a lag phase of 4 h, the dissolved oxygen concentration decreased rapidly within 1 h from 100% to oxygen-limiting conditions of below 1% (results not shown), and lactic acid production started. At that moment, concentrations of glucose, xylose, and arabinose of 3.3, 0.7, and 0.3 g/l, respectively, were present (Fig. 3a). These sugars were consumed simultaneously where glucose was utilized faster than xylose and arabinose. Simultaneous with the consumption of these monomeric sugars, lactic acid was produced, which was neutralized by the automatic addition of alkaline LTWS. By the addition of alkaline substrate throughout the fed-batch phase, the pH was maintained accurately at 6.0 ± 0.1 (Fig. 2a,b). At the end of phase II, a total amount of 10,023 g dehydrated LTWS (~2,706 g DM LTWS) and 320 ml of enzyme preparation was added to the fermenter. A lactic acid concentration of 20.5 g/l supernatant was detected (Fig. 3b), corresponding to a total of 342 g lactic acid. The chiral L(+) purity of lactic acid was determined at 99.4%, which is similar to that obtained with xylose as the sole carbon source (Otto 2004). Fig. 2Control of pH (a) during simultaneous saccharification and fermentation of lime-treated wheat straw by commercial enzyme preparation GC 220 and B. coagulans DSM 2314 (b). The areas between the dotted lines represent the prehydrolysis phase (I), the fed-batch phase (II) with pH control by addition of alkaline LTWS and enzymes, and the batch phase (III) with pH control by addition of Ca(OH)2 suspension. Extra enzyme preparation GC220 was added at the times indicated by the arrowsFig. 3Profiles of glucose (empty square), xylose (empty diamonds), arabinose (empty triangles) (a) and lactic acid (filled diamonds) (b) in simultaneous saccharification and fermentation of lime-treated wheat straw by commercial enzyme preparation GC 220 and B. coagulans DSM 2314. The areas between the dotted lines represent the prehydrolysis phase (I), the fed-batch phase (II), and the batch phase (III). Extra enzyme preparation GC220 was added at the times indicated by the arrows At the end of phase I, a low acetic acid concentration was detected in the medium, which increased to 1.5 g/l throughout phase II but remained constant during phase III (results not shown). This indicates that acetic acid was most likely not a fermentation product formed by B. coagulans. Acetic acid can be released upon solubilization and hydrolysis of hemicellulose during chemical pretreatment (Palmqvist et al. 1999). By the dehydration procedure of the LTWS, part of the acetic acid was easily separated from the substrate by removing the press water. Apparently, a remaining amount of acetic acid was fed together with the substrate to the fermenter. Furthermore, minor traces of other organic acids such as succinic acid and formic acid (<0.5 g/l) were detected in the fermentation broth. Phase III was initiated by changing the pH control from the addition of alkaline LTWS to a 20% (w/v) calcium hydroxide suspension. To maintain the pH at 6.0, the addition of calcium hydroxide suspension occurred relatively fast but shifted, however, after a few hours to a lower addition rate indicating a decline of the volumetric lactic acid productivity (Figs. 2b, 3b). To exclude limitation (e.g., by inactivation) of enzymes, an extra dosage of enzyme preparation (80 ml) was added to the fermenter after 23.5 h of incubation. This resulted immediately in a slight acceleration of the calcium hydroxide addition rate indicating an increased lactic acid productivity and limitation of enzymatic activity (Fig. 3b). Nevertheless, after 29.7 h of incubation, a decline of the calcium hydroxide addition rate was observed again. Therefore, a second extra dosage of the enzyme preparation (240 ml) was added and resulted this time in a slight accumulation of glucose and xylose of 1.5 and 1.0 g/l (Fig. 3a), respectively, indicating that microbial conversion instead of enzymatic hydrolysis was rate limiting. After 32 h of incubation, a lactic acid concentration of 37.1 g/l was obtained, with a chiral l(+)-lactic acid purity of 99.4%. Continuation of the SSF process to a total incubation period of 55 h resulted in a slightly increased lactic acid concentration of 40.7 g/l supernatant (~37.8 g lactic acid/kg fermentation broth) with an overall volumetric lactic acid productivity of 0.74 g l−1 h−1. At this stage, a chiral l(+)-lactic acid purity of 97.2% was analyzed. This slight decline in lactic acid purity is possibly a result of infection with other undesired lactic acid-producing microorganisms. Because the substrate used was not sterile and also the chemical pretreatment and fermentation occurred in an open system under nonsterile conditions, microbial contamination throughout the SSF process is possible. Conversion efficiency The efficiency of the enzymatic hydrolysis of the polymeric material present in LTWS is shown in Fig. 4. The insoluble polymeric fraction was determined at various time points throughout the SSF experiment. At the end of the prehydrolysis (2 h) of 378 g LTWS, 36% of the insoluble glucan (Fig. 4a), 55% of xylan (Fig. 4b), and 62% of arabinan (Fig. 4c) were converted to soluble saccharides including monomeric sugars and oligomeric sugars. After the fed-batch phase (13 h), 2,706 g LTWS was added and resulted in a conversion of 42% of glucan, 57% of xylan, and 63% of arabinan to products including soluble saccharides and lactic acid. Between 13 and 32 h of incubation, further hydrolysis of the polymeric sugars was observed. However, during the last 23 h of the SSF, minor hydrolysis of the polysaccharides occurred, and this corresponded with the decline in lactic acid productivity during this phase. After 55 h, 398 g of glucan, 130 g of xylan, and 11 g of arabinan was still present as insoluble polymeric material. With these values, the hydrolysis efficiency of the initial glucan, xylan, and arabinan present in LTWS were calculated as 55%, 75%, and 80%, respectively. Fig. 4Insoluble fraction (empty bars) and hydrolyzed soluble fraction (filled bars) (g; calculated by the difference between initial amounts and analyzed insoluble amounts) of the polysaccharide glucan (a), xylan (b), and arabinan (c) at various time points during the simultaneous saccharification and fermentation of lime-treated wheat straw. The figure represents also the percentage of polysaccharide hydrolyzed into soluble products (filled triangles). The error bars denote the deviation of duplicate analysis The monomeric sugars, derived from the LTWS, were partly converted to lactic acid (711 g) by B. coagulans and accounted for 81% (w/w) of the theoretical maximum, indicating the formation of other products such as microbial biomass and carbon dioxide. An overall conversion yield of 43% (w/w) of the theoretical maximum was calculated according to Eq. 1. The fate of polysaccharides initially present in LTWS after 55 h of incubation is shown in Table 1. A part of the polysaccharides present in LTWS remained as insoluble polysaccharides (37% w/w), whereas a minor part was converted into soluble oligomeric (5% w/w) and monomeric (3% w/w) sugars. Another part of the initial polysaccharides present in the LTWS was not recovered in the form of saccharides or lactic acid and was therefore ascribed as ‘unaccounted.’ Table 1Fate of polysaccharidesa initially present in lime-treated wheat straw after 55 h of simultaneous saccharification and fermentationFractionPercentage (% w/w)Polysaccharides (insoluble)b37Oligosaccharides (soluble)5Monosaccharides (soluble)3Lactic acid (soluble)43Unaccounted (insoluble/soluble)c13Presented values are averages based on duplicate analytical measurements.aTotal of glucan, xylan, and arabinanbPart of the initial polysaccharides remained present as insoluble polysaccharides.cPart of the initial polysaccharides was not recovered and therefore denoted as ‘unaccounted.’ Neutralization of acid by alkaline substrate The lactic acid produced (342 g) during the fed-batch phase (II) was neutralized with alkaline-pretreated wheat straw. During this phase, an amount of 2,328 g LTWS was added to the fermenter. Together with this substrate, an amount of 230 g calcium hydroxide was added to the fermenter and accounted for a ratio of 0.67 g calcium hydroxide per gram of lactic acid. The lactic acid (369 g) produced during the batch phase (III) was neutralized with 163 g calcium hydroxide resulting in a ratio of 0.44 g lactic acid per gram calcium hydroxide. Discussion Lignocellulosic feedstocks are considered as potential attractive substrates for the production of bulk chemicals. Pretreatment of biomass is required to break open the lignocellulosic matrix, and an enzymatic hydrolysis is necessary for the hydrolysis of polymeric carbohydrates. The lime pretreatment has proven to enhance enzymatic digestibility of the polysaccharides present in lignocelluloses (Chang et al. 1998; Kaar and Holtzapple 2000) and results, in comparison to other pretreatment routes, in minor inhibitor formation. However, before the enzymatic hydrolysis, it is essential to adjust the pH to a level optimal for enzymatic activity. In this study, the reduction in pH by washing or neutralization was omitted by using the alkaline character of LTWS to neutralize lactic acid produced by microbial fermentation during a SSF process. The results showed that the largest part of the polysaccharides in LTWS was converted enzymatically and the resulting sugars were fermented simultaneously to mainly lactic acid by B. coagulans DSM 2314. Between 10 and 30 h of incubation, the bacteria utilized the monomeric sugars, as soon as they appeared in the medium, resulting in relatively low monomeric sugar concentrations (<2 g/l). This indicates that throughout this period, the enzymatic hydrolysis was the rate-controlling step. The highest lactic acid productivity was observed during the fed-batch phase and the initial hours of the batch phase and declined rapidly after approximately 20 h of incubation, as shown in Fig. 3b. An extra addition of enzyme preparation showed a slight improvement of the volumetric lactic acid productivity but shifted within a few hours again to a relatively low production rate. A second extra enzyme addition did not affect the lactic acid productivity significantly (Fig. 3b). This addition of new enzymes resulted in a modest liberation of hemicellulose sugars (xylose, arabinose), but no further hydrolysis of glucan occurred. This shows that the remaining glucan was too recalcitrant or not accessible for further hydrolysis, resulting in decreasing lactic acid productivity. Another possible explanation of the decreased lactic acid productivity is the inhibition of enzymes and/or bacteria by the increasing lactic acid concentration. A lactic acid concentration of 40.7 g/l supernatant (~37.8 g lactic acid/kg fermentation broth) with a relatively high chiral purity was determined after 55 h of incubation, corresponding to an overall lactic acid yield of 43% of the theoretical maximum. Moreover, the efficiencies of the enzymatic saccharification and the fermentation were both determined. These calculations showed that, based on residue analysis, at the end of the SSF process (55 h), 55% of the glucan, 75% of the xylan, and 80% of the arabinan present in LTWS was enzymatically hydrolyzed, which agree well with previously obtained results from experiments aiming to convert LTWS to ethanol. To improve the yield, it is necessary to decrease the recalcitrance or improve the accessibility of polymeric sugars in the LTWS by optimization of the pretreatment procedure. The concentrations of soluble monosaccharides and oligosaccharides in the medium were relatively low, which can be expected in a SSF process. A fermentation yield of 81% was determined (related to the amount of monosaccharides released from the LTWS) and is slightly better than the results obtained by Otto (2004) who reported the production of 35 g/l lactic acid from 50 g/l xylose as the sole carbon source. Because no other soluble fermentation products were detected, the remaining 19% of the LTWS-derived monomeric sugars were most presumably converted to bacterial biomass and some carbon dioxide during the aerobic part of the fermentation. Several process parameters can be listed for enhancement of the overall lactic acid yield and productivity such as improving the accessibility of polysaccharides by a more severe lime pretreatment, enzyme dosage, type of enzymes, B. coagulans strain, size and growth phase of inoculum, pH gradient in SSF, and in situ product removal of lactic acid. These issues will be subject to further studies. During the fed-batch phase (II), it was possible to counterbalance the pH decrease caused by lactic acid production by the addition of the alkaline feedstock. This suggests that it is possible to combine lime treatment with the production of other organic acids from lignocellulosic biomass. Throughout this phase, the ratio of calcium hydroxide in LTWS added per produced lactic acid was determined at 0.67 g/g. The theoretical stoichiometric neutralization of 1.00 g lactic acid requires 0.41 g calcium hydroxide. Therefore, only 61% of the calcium hydroxide initially added to the wheat straw was used for lactic acid neutralization. On the other hand, throughout the batch phase (III), an alkaline/acid ratio of 0.44 g/g was calculated corresponding to 93% of the added calcium hydroxide suspension used for lactic acid neutralization. The low efficiency of the calcium hydroxide added with the LTWS for lactic acid neutralization during phase II has three possible explanations. First, part of the calcium hydroxide could have been used during the chemical pretreatment of the wheat straw such as the neutralization of acetic acid or other organic acids and/or irreversible binding to the lignin. Second, the calcium hydroxide might be released slowly from the insoluble wheat straw fibers and could therefore partly have been used for lactic acid neutralization in the fed-batch phase. Finally, besides lactic acid production, other acidification reactions could have contributed to the decrease in pH and therefore the demand of alkaline substrate, for instance, the decrease in pH caused by the consumption of ammonium as the nitrogen source by microorganisms (Guebel et al. 1992). The results in this paper show that it is possible to use lignocellulosic materials for the production of lactic acid. Lignocellulosic biomass is a relatively inexpensive substrate, and this affects feedstock costs for lactic acid production positively. Nevertheless, in comparison to the traditional relatively ‘clean’ feedstocks with a well-defined composition, using heterogenic lignocellulosic substrates will require a more intensified downstream processing (DSP) to recover and purify the lactic acid from the complex fermentation broth. The costs of feedstock materials and operational costs of the DSP contribute considerably to the overall production costs of lactic acid (Åkerberg and Zacchi 2000). Whether the cost decrease in using lignocellulosic feedstocks outweighs the potential increasing costs of DSP was not analyzed at the moment. In summary, LTWS was converted into l(+)-lactic acid by B. coagulans throughout a SSF process at a 20-l bench scale. The pentose and hexose sugars derived from the polymeric material were utilized simultaneously by B. coagulans resulting in a final lactic acid concentration of 40.7 g/l supernatant, which accounted for 43% (w/w) of the theoretical yield. To our knowledge, this is the first paper demonstrating a process having a combined alkaline pretreatment of lignocellulosic biomass and pH control in organic acid fermentation resulting in a significant saving of lime consumption and avoiding the necessity to recycle lime.	[ "lactic acid", "lime-treated wheat straw", "ph control", "calcium hydroxide", "simultaneous saccharification and fermentation" ]	[ "P", "P", "P", "P", "P" ]
Clin_Oral_Investig-3-1-2099161	Three-year survival of single- and two-surface ART restorations in a high-caries child population	The aim of this study was to evaluate the survival of single- and two-surface atraumatic restorative treatment (ART) restorations in the primary and permanent dentitions of children from a high-caries population, in a field setting. The study was conducted in the rainforest of Suriname, South America. ART restorations, made by four Dutch dentists, were evaluated after 6 months, 1, 2, and 3 years. Four hundred seventy-five ART restorations were placed in the primary dentition and 54 in first permanent molars of 194 children (mean age 6.09 ± 0.48 years). Three-year cumulative survivals of single- and two-surface ART restorations in the primary dentition were 43.4 and 12.2%, respectively. Main failure characteristics were gross marginal defects and total or partial losses. Three-year cumulative survival for single-surface ART restorations in the permanent dentition was 29.6%. Main failure characteristics were secondary caries and gross marginal defects. An operator effect was found only for two-surface restorations. The results show extremely low survival rates for single- and two-surface ART restorations in the primary and permanent dentitions. The variable success for ART may initiate further discussion about alternative treatment strategies, especially in those situations where choices have to be made with respect to a well-balanced, cost-effective package of basic oral health care. Introduction The concept of minimal invasive dentistry has evolved as a consequence of an increased understanding of caries and the development of adhesive restorative materials [27]. Within this concept, prevention and hard tissue preservation are the primary goals, and dentists are encouraged to prefer a more conservative and biological approach rather than a surgical approach, although the latter is sometimes unavoidable. The atraumatic restorative treatment (ART) technique is part of a minimal invasive approach and, as such, a technique that meets the specific goals mentioned above. In brief, with ART, soft demineralized carious tooth tissue is removed using hand instruments only, followed by restoration of the tooth with an adhesive restorative material, often glass ionomer cement [4, 7]. Because neither electricity nor running water is required for this treatment approach, ART can be applied in almost any setting. Although initially developed to provide restorative dental treatment in outreach or rural areas, ART or modified ART techniques are increasingly introduced into dental clinics in industrialized countries [1, 11, 14]. Since its introduction in the mid-1980s, ART has been evaluated in several community field trials. These studies served mainly to obtain information on technical aspects of the process, handling characteristics of the restorative material, and on the survival of the restorations. They led to improvement of the technique [20] and to the development of new, more appropriate glass ionomer restoration materials, especially for ART purposes. Studies focussing on the survival of ART restorations have shown that the ART approach is very successful in restoring single-surface dentine lesions in the permanent dentition: 3-year survival rates of 71–92% have been reported [5, 6, 9, 10, 12, 20]. Regarding the survival rates of ART restorations in the primary dentition, only a few field studies were performed. They showed acceptable survival rates (65–96.7%) for single-surface ART restorations, but generally low success rates (31–76.1%) for multi-surface ART restorations, even with the newer glass ionomer materials [2, 3, 11–13, 19, 23, 24, 28]. Although its performance under multisurface conditions is disappointing, ART is considered a valuable approach towards the treatment of dental caries. The use of ART has resulted in the retention of many teeth that would otherwise have been extracted in a later stage. Nevertheless, there still remain some controversies towards the technique, presumably based on the inconsistency in survival results. Moreover, a recent study, investigating the influence of dental treatment on the oral health of a Surinamese child population, concluded that performing ART restorations only, did not contribute significantly to an improvement of the oral health, suggesting that ART alone is not a sufficient solution in the battle against dental decay (van Gemert-Schriks et al., submitted, 2007). Frencken et al. [8] described comprehensively that ART should be part of a basic package of oral care in which prevention and urgent care are also represented. However, within this package, these three components should be geared to one another as much as possible and the individual effects of all three components must be sufficient and beneficial under different circumstances. When the success of either component, particularly ART, cannot be guaranteed, its contribution in the package should be reduced. Thus, the evaluation of ART in different countries or communities, among different kinds of caries-risk populations and under diverging conditions remains useful. Therefore, the aim of this study is to evaluate the survival of both single- and two-surface ART restorations in the primary and permanent dentitions of children from a high-caries population in a field setting on a longitudinal base. Materials and methods This cohort study was conducted in the rainforest of Suriname, South America. It was part of a large-scale project investigating the influence of dental treatment on the oral health of children (van Gemert-Schriks et al., submitted, 2007). Within the scope of that particular project, 380 6-year-old children were divided randomly among four different treatment groups. Material presented in the current article concerns only those children who received restorative treatment, according to the ART method, either in their primary or permanent dentitions. The restorative treatments were performed in accordance with the ART guidelines [4, 7] and took place in empty classrooms where four children were treated at the same time. Ketac-Molar (3M-ESPE®) was used as the restorative material of choice. The treatments were carried out by four Dutch dentists who were trained in ART during a 1-week ART course and by using ART in children from their own practices, for a period of 3 months, before the start of the treatment phase of the study. They were assisted by six Surinamese health care assistants from the Medical Mission who completed an ART course supplemented with some basic dental knowledge. The dentists were asked to note any contamination with blood and/or saliva during the restoration of the cavity. Furthermore, the presence or absence of adjacent teeth was noted. During the treatment, one of the authors (MGS), who was not involved in the treatment phase, observed and classified the overall behavior of the child, based on a modified Venham scale [23, 29]. Before the study, this observer was trained in using the Venham behavior scale by scoring 42 videotapes of children in a dental situation. These observations were compared to the consensus score of two calibrated observers. This comparison resulted in a Cohen’s Kappa of 0.87, implying an excellent agreement. Restorations were not assessed at the time of placement (T0). The children were revisited for evaluation of the ART restorations 6 months (T1), 1 year (T2), 2 years (T3), and 3 years (T4) after the initial treatment. The same author and dentist mentioned above (MGS), evaluated the restorations according to the ART criteria (Table 1) using a CPITN probe, a mouth mirror and a headlamp. Before the study, this person was calibrated against a “gold standard” (Kappa 0.94). This gold standard was achieved by the consensus of two experienced dentists during the assessment of 24 extracted molars with ART restorations. Restorations scored code 00 or 10 were considered successful, codes 11–40 were classified as failures, and codes 50–90 were assigned in case the tooth was unavailable for evaluation. If a tooth or restoration showed multiple defects, a marginal defect dominated an over- or underfilled restoration (10, 11 > 12, 13), secondary caries dominated a marginal defect (20, 21 > 10, 11), absence of a restoration dominated secondary caries (30 > 20, 21), and an overfilled cavity dominated an underfilled cavity (13 > 12). Table 1Evaluation criteria for the ART restorationsCodeEvaluation characteristics00Restoration present, correct10Restoration present, slight marginal defect/wear of surface (<0.5 mm). No repair needed.11Restoration present, gross marginal defect/wear of surface (>0.5 mm). Repair needed.12Restoration present, underfilled (>0.5 mm). Repair needed.13Restoration present, overfilled (>0.5 mm). Repair needed.20Secondary caries, discoloration in depth, surface hard and intact, caries within dentin. Repair needed.21Secondary caries, surface defect, caries within dentin. Repair needed.30Restoration not present, bulk fracture, moving or partial lost. Repair needed.40Inflammation of the pulp; signs of dentogenic infection (abscesses, fistulae, pain complaints). Restoration might still be in situ. Extraction needed.50Tooth not present because of extraction60Tooth not present because of shedding70Tooth not present because of extraction or shedding90Patient not present Statistical analysis Statistical analyses were performed using SPSS for Windows, version 12.0.1 (SPSS, Chicago, USA). All significant differences were detected at a 95% confidence level. Kaplan–Meier survival analyses were performed on the censored data of both single- and two-surface restorations. The significance of differences between survival curves was determined with log-rank tests. Possible confounding variables were taken into account using a Cox regression analysis. Results As stated in the “Materials and methods” section, the children in this study were derived from a larger study population of children participating in another project. The overall caries prevalence, expressed in terms of decayed, missing and filled surfaces (dmfs) among that group of children, was 11.51 (SD 10.5; range 0–53) in the primary dentition and 0.20 (SD 0.62; range 0–5) in the permanent dentition. According to the standards of the World Health Organisation [17], this denotes a high-caries child population based on the caries prevalence in the primary dentition. Within the larger group, 194 children (mean age 6.09 ± 0.48 years) received ART restorations in either their primary or permanent teeth, or both. Only these children were included in the current study. Their baseline caries prevalence was 12.75 (SD 9.88; range 0–53) in the primary dentition and 0.23 (SD 0.67; range 0–5) in the permanent dentition. At baseline (T0), 475 ART restorations were placed in the primary dentition (mainly first and second molars) and 54 in the first permanent molars (predominantly mandibular). Table 2 presents data for the ART restorations, performed at baseline. A Mann–Whitney U test showed that children who received two-surface restorations scored higher on the Venham behavior scale (p = 0.005) than children that received single-surface restorations, in the primary dentition. Furthermore, dentists reported more contamination (chi-square = 25.02, df = 1, p < 0.001) when placing two-surface restorations than single-surface restorations. Table 2Baseline data for the ART restorations Primary dentitionPermanent dentitionNumber of filled surfaces121Number of restorations13334254Number of children (N)6114734Mean number of restorations per child (SD; range)3.50 (1.61; 1–7)3.64 (1.73; 1–8)2.07 (0.97; 1–4)Dentist143 (32.3%)74 (21.6%)12 (22.2%)241 (30.8%)84 (24.6%)16 (29.6%)334 (25.6%)89 (26.0%)7 (13.0%)415 (11.3%)a95 (27.8%)19 (35.2%)Adjacent tooth presentYes117 (88.0%)303 (88.6%)45 (83.3%)No16 (12.0%)39 (11.4%)9 (16.7%)Contamination blood/salivaYes13 (9.8%)110 (32.2%)a5 (9.3%)No120 (90.2%)232 (67.8%)49 (90.7%)Venham behavior score050 (37.6%)78 (22.8%)8 (14.8%)144 (33.1%)137 (40.1%)27 (50.0%)226 (19.5%)82 (24.0%)13 (24.1%)313 (9.8%)33 (9.6%)6 (11.1%)4–12 (3.5%)a–5–––SD Standard deviationaStatistical significant difference at p = 0.005. The lost-to-follow-up percentage of the restorations originally placed was 4.63%. After 3 years, the cumulative survival of the single-surface ART restorations in the primary dentition was 43.4% (standard error (SE) 10.9%). For the two-surface restorations, a cumulative survival of 12.2% (SE 2.99%) was observed. The survival curves, with censored data, are presented in Fig. 1a and b. The cumulative survival of the single-surface ART restorations in the permanent dentition was 29.6% (SE 8.2%) after 3 years (Fig. 2). Fig. 1a Survival curve single surface ART restorations, primary dentition. b Survival curve multisurface ART restorations, primary dentitionFig. 2Survival curve single-surface ART restorations, permanent dentition Table 3 represents the failure characteristics for the restorations in both primary and permanent dentitions at 3 years. The main failure characteristics of both single- and two-surface ART restorations in the primary dentition were gross marginal defects (score 11) and total or partial losses (score 30). For restorations in the permanent dentition, the main failure characteristics were secondary caries (score 21) and gross marginal defects (score 11). Table 3Failure characteristics for the ART restorations at 3 years Primary dentitionPermanent dentition1 surface2 surface1 surfaceRestorations baseline (N)13334254Failures (N)4225132Failure scoreGross marginal defect (11)21 (15.8%)86 (25.1%)13 (24.1%)Restoration present, underfilled (12)1 (0.8%)9 (2.6%)–Restoration present, overfilled (13)2 (1.5%)11 (3.2%)2 (3.7%)Sec. caries, discoloration (20)–––Sec. caries, surface defect (21)5 (3.8%)1 (0.3%)14 (25.9%)Total or partial loss (30)13 (9.8%)120 (35.1%)3 (5.6%)Pulpal inflammation (40)–24 (7.0%)–Restoration missing, extracted (50)–––Scores 60–90 were not included (censored data) A log-rank test indicated that there were no statistically significant differences in survival times between the four dentists regarding single-surface restorations in both primary and permanent teeth. However, regarding the two-surface restorations in the primary dentition, statistically significant differences between the four dentists appeared (log-rank statistic 11.7, df = 3, p = 0.009). The separate survival curves are presented in Fig. 3. Fig. 3Survival curves per dentist, multi-surface ART restorations primary dentition To detect any confounding variables on the survival of the ART restorations, a Cox regression analysis was performed. No significant relation could be found, indicating that neither the presence or absence of an adjacent tooth, nor contamination with blood and/or saliva, nor the behavior of the child during the restorative phase of the treatment had an influence on the 3-year survival of the restorations in the primary dentition. No effect also could be found regarding the number of restorations per child. Neither of these variables had an effect on the survival rates in the permanent dentition, except for the presence of adjacent teeth. Restorations in teeth where no adjacent tooth was present were found to be more likely to fail (hazard ratio = 6.53, 95% CI 2.66–16.02, p < 0.001). Discussion In contrast with other studies, the results of this study show extremely low survival rates for both single- and two-surface ART restorations in the primary and permanent dentitions. An operator effect was observed for two-surface restorations only. Neither the behavior of the child during restoration, and the number of restorations per child, nor the contamination of preparations with blood or saliva had a significant influence on the survival of the restorations in this study. This field study was performed correctly and the statistical power was sufficiently high to detect at least medium effects. However, because it was part of a large-scale randomized controlled clinical trial, no comprehensive criteria were formulated beforehand regarding, for example, the number of restorations per patient, and the location and the size of the cavities. This aspect is inherent to many cohort studies and it does not imply an inferior study quality, but it limits a meaningful comparison with other survival studies. Although all possible efforts were exercised to trace the participating children over the evaluation period, 22 restorations (4.63%, eight children), all in primary molars, could not be evaluated at any of the recall visits. Either the children did not show up, or the teeth concerned had exfoliated before the first evaluation. These restorations were regarded as missing data and, therefore, excluded from further analysis. Twenty-six restorations (5.47%, 21 children) were “lost” for evaluation because the teeth either exfoliated or the child moved to another district during the course of the study, but after the first evaluation. These restorations (scores 60–90) were treated as censored data and not as true failures because they survived up to a certain moment. Many causative factors could be suggested that might explain the failure of the ART restorations, such as secondary caries, cervical margin gaps, material properties, and field conditions (outside temperature, atmospheric humidity). However, many other ART studies face these or comparable problems and, therefore, these factors cannot sufficiently explain the extremely low survival rates found in this particular study. The operator difference for the survival rates of the two-surface restorations was not unique, and not a sufficient explanation for the disappointing survival results. Operator effects are often found in ART studies [4, 9, 15, 21, 26] and, as in every profession, there will always be individual differences in technical skills. The finding that the absence of an adjacent tooth was related to a lower 3-year survival of single-surface ART restorations in permanent molars could not be explained. One can only speculate about possible reasons for this relationship, such as that these freestanding molars experience larger occlusal forces. A possible influence of the relatively high caries prevalence on the survival of the restorations could be hypothesized, but is very doubtful. A study in Indonesia, where the child population exhibited a much higher caries prevalence, also found disappointing survival rates for two-surface ART restorations [28], but these rates were not as extreme as those found in the current study. The survival rates for single-surface ART restorations, derived from other earlier cited studies, were all very promising regardless of the caries profile of the study populations. Furthermore, no effect on the survival rates of the restorations was found when the number of restorations per child was included in the analysis. The ART protocol prescribes not to eat or drink within at least 1 h after the completion of the restorative treatment [7]. The children in the current study were not supervised after they received restorative treatment and consequently, their food intake could not be controlled. Future studies should take this aspect into account. Other patient-related factors that may influence the survival of the restorations are the behavior and saliva flow of the child. The survival of the ART restorations in this study was analyzed at the restoration level. This method requires independency of the restoration data and, with respect to the mentioned patient-related possible bias, this assumption could not be guaranteed. To control for this lack of independency, the survival analyses also were performed at the patient level, including only one randomly selected restoration per child. These analyses did not render higher survival rates. The predominant failure characteristics for both single- and two-surface ART restorations in the primary dentition were gross marginal defects and total or partial losses. This agrees with previous studies concerning the survival of ART restorations in the primary dentition [6, 13, 25, 26]. Gross marginal defects could be induced by occlusal forces or insufficient wear resistance of the restorative material. Ketac-Molar was specifically developed for ART purposes [13], and it has shown excellent results for posterior restorations in the primary dentition [16, 22]. Glass ionomer restorations can be dislodged for a number of reasons, such as insufficient cleaning and conditioning of the cavity, and improper mixing of the material. None of these conditions was recorded at the time the tooth was restored. However, all dentists and chair-side assistants followed the ART guidelines and the manufacturer’s instructions as much as possible under the given circumstances. The main reasons for failure of the single-surface restorations in the permanent dentition were gross marginal defects and secondary caries. This latter finding is somewhat surprising and contrasts with earlier ART studies [5, 6, 10, 26]. Glass ionomer cement has been the restorative material of choice for the ART technique, based mainly on its fluoride releasing and, thus, caries-preventive properties [7]. Many studies underline these characteristics of glass ionomer [18, 26, 30, 31]. The extremely low survival of the ART restorations observed in this study remains unexplained. Circumstances that were not recognized as possible interfering factors at the start of the study might have played an important role, including cultural and seasonal dietary influences. People living in the rainforest of Suriname eat seasonal fruits such as mangos and fruits of the fiber palm (Awarra). In particular, the latter may influence the survival of the restorations, given the frequency and method in which they are consumed. The authors have seen unusual wear patterns, also in adult dentitions, which might have been caused by excessive consumption of Awarras. A possible causality between these dietary habits and the survival of the ART restorations can only be disclosed by future controlled studies. Although previous studies have suggested that ART should not be considered as a routine procedure to restore multisurface cavities [13, 24], based on the results of this study, even the ART restoration of single-surface cavities might be reconsidered. This study underlines the inconsistency and variation in the success of the treatment. Apparently, certain conditions must be fulfilled to make ART successful. These conditions can be approached, but not always achieved, under all circumstances. Conclusion The uncertain predictability for the success of ART may introduce further discussion about alternative treatment strategies, especially in those situations where choices have to be made with respect to a well-balanced, cost-effective package of basic oral health care. To gain insight into factors determining the cumulative success rate of ART restorations, future studies should focus in more detail on variables that could possibly contribute to the failure of restorations.	[ "art", "primary dentition", "cumulative survival", "oral health" ]	[ "P", "P", "P", "P" ]
Reprod_Biol_Endocrinol-1-_-153492	Luteal regression: a redefinition of the terms	The corpus luteum is a transient endocrine gland that is specialized for the production of progesterone and that plays a critical role in the establishment and maintenance of pregnancy. The life span of the corpus luteum varies between species and, within a species, can be dramatically altered by events such as mating or pregnancy. Regardless of the duration of its life span, the corpus luteum eventually enters a dynamic regression process during which it loses the capacity to produce progesterone and undergoes structural involution. The overall process of luteal regression has been referred to by a variety of terms over the last several decades. In Table 1, we show the results of a cursory MEDLINE/PubMed search on luteal regression performed in November of 2002, which reveals that the most frequently utilized term for this process is "luteolysis," followed by "luteal regression," and then, to a lesser extent, by "functional luteolysis" and "structural luteolysis." Other terms, such as "luteal involution," "functional luteal regression," or "structural luteal regression" lag far behind in usage. The necessity of using several keywords to search for information on luteal regression is illustrated by the fact that in our sample MEDLINE/PubMed search, only 47% of the manuscripts retrieved using the keyword "luteal regression" were also retrieved when using the keyword "luteolysis." This inconsistency in terminology can result in serious under-retrieval and under-citation of papers using less popular keywords. Table 1 Results of a cursory MEDLINE/PubMed search performed on November 2002 using the terms as listed. Rank Terms Total Number of Articles Retrieved Retrieval Period Number of Articles Retrieved for Period 2001–2002 1 "luteolysis" 1251 1967–2002 96 2 "luteal regression" 403 1966–2002 35 3 "functional luteolysis" 82 1973–2002 5 4 "structural luteolysis" 72 1969–2002 14 5 "luteal involution" 41 1978–2002 8 6 "functional luteal regression" 10 1984–2002 0 7 "structural luteal regression" 5 1988–2002 1 To quote Rossdale and Cox in their communication entitled Terminology: a mark of scientific progress [1]: "Scientific terminology must be as exact as is possible within the state of knowledge available." As our knowledge of the molecular mechanisms participating in the process of luteal regression evolves, some of the commonly used terms cited above have been rendered obsolete, inappropriate, and even incorrect. The foremost of these is "luteolysis," still one of the most common terms used to define the process of luteal regression, most probably because it is listed among Medical Subject Headings (MeSH), and, therefore, used most frequently. MeSH is the U.S. National Library of Medicine's controlled vocabulary used for indexing articles in PubMed. MeSH terminology provides a consistent way to retrieve information that may use different terminology for the same concepts. In MeSH, "luteolysis" is defined as "degradation of corpus luteum" and further described as "...characterized by the involution and cessation of its endocrine function." From a cell biology point of view, "lysis" means "rupture of cell plasma membrane, leading to the release of cytoplasm and the death of the cell" [2]. If interpreted with strict accordance to the Latin, lysis refers to "disintegration," as recently discussed by Davis and Rueda [3]. Over the last decade, it has become clear that the involution of the corpus luteum is associated with a phenomenon of programmed cell death or apoptosis [4]. Because apoptosis is an organized process that does not involve a major inflammatory response, and because the majority of cells are removed prior to rupture, the term "lysis" when applied to luteal regression is largely inaccurate. Therefore, while the term luteolysis remains very popular – 1251 manuscripts were retrieved using the word "luteolysis" and only 403 were retrieved using the terms "luteal regression" in the MEDLINE/PubMed search (see Table 1) – this term no longer accurately reflects the mechanisms involved in luteal regression. In 1999, McCraken et al. [5] stated that "the term luteolysis may be something of a misnomer;" but also referred to the deep entrenchment of this term in the literature. However, the exponential increase in the number of scientific manuscripts published in recent years ensures that a prompt move to replace the older term with a more accurate descriptor will rapidly result in an overturn of the current ranking displayed in Table 1 and that only a minority of papers will continue to be retrieved using the older term. Despite the fact that a strict definition of "luteolysis" appears to regard only the structural disintegration of the corpus luteum, broader interpretations are rife within the literature. Rothchild [6] defined luteolysis as "stopping the secretion of progesterone by the corpus luteum." Two other recent reviews take a middle road. Hoyer [7] refers to the fall in progesterone secretion as functional regression but considers structural regression, or cell death, the "true" regression. Niswender et al. [8], while referring to luteal regression as the structural demise of the corpus luteum, go on to describe the loss of progesterone synthesis and secretion – not a structural event – as part of the regression of the corpus luteum. As the previous paragraph shows, the adjectives "structural" and "functional" cannot be overlooked when addressing terminology in the field of luteal regression. These words have had a consistent presence in literature concerning the regression of the corpus luteum since Malven and Sawyer [9] described the process of luteal regression in the rat as having two components: "1) an initial termination of luteal secretory function and 2) a subsequent morphological regression of the nonfunctioning corpora lutea." A few years later Malven et al. [10] used the term "structural luteolysis" to describe the morphological regression of persistent, nonfunctional corpora lutea in the hypophysectomized rat. The definition of "nonfunctional" provided in this manuscript expanded upon the 1966 description of the first component of luteal regression in the rat by explaining that the corpora lutea "did not secrete enough progesterone to induce deciduoma formation following endometrial traumatization." Thus, the corpora lutea were "nonfunctional" in terms of progesterone production and the potential for establishment of pregnancy. Since the appearance of these early papers, the terms functional and structural regression of the corpus luteum (or the more popular phrasing "functional and structural luteolysis") have gained relative acceptance and usage. The cursory MEDLINE/PubMed search shown in Table 1 turned up 82 references containing the phrase "functional luteolysis," of which 5 were 2001 and 2002 publications, and 72 references containing the phrase "structural luteolysis," of which 14 were 2001 and 2002 publications. The prevalence of the term "structural luteolysis" over "functional luteolysis" in recently published articles may be due in part to a greater disparity in the perceived meaning of the latter term. This stems largely from studies in rats, in which it was revealed that corpora lutea which no longer secreted significant quantities of progesterone retained many functions, including active steroidogenesis [11-13]. It is our contention that these two frequently described components of the process of luteal regression (i.e. cessation of progesterone secretion and disappearance of the structure) merit equal attention. However, the timing and inter-relationship of these two types of regressive changes appears to vary by species. As new information on the molecular steps triggered during luteal regression is obtained, these mechanisms will need to be defined as they relate to these facets of the overall regressive process. Davis and Rueda [3] advised that the terms functional and structural used in the context of luteal regression should be adequately defined by the authors in the context of the particular species and reproductive stage being studied, with which we concur. It is not extreme to predict that with the advancement of knowledge in the process of luteal regression researchers will find increasing difficulties in establishing the limits as to what mechanisms belong to the "functional" facet of luteal regression, and which ones are part of the "structural" facet, as luteal regression is a process of synchronized events. For example, it now appears that progesterone itself is responsible for preventing the onset of apoptosis and structural disintegration [14-17]. Conclusions These authors strongly recommend that the term luteal regression should be used to refer to the process of regression of the corpus luteum. Luteal regression represents a broad definition of the process of demise of the corpus luteum that is capable of accommodating all new knowledge evolved on the molecular mechanisms activated or inhibited during the process of regression of the corpus luteum regardless of species or reproductive stage. Further, according to the etymology of the word, luteolysis has a very limited meaning that no longer describes properly the complex sequence of synchronized molecular events associated with the demise of the corpus luteum. The modifying terms "functional" and "structural" can be used in conjunction with the term luteal regression to further identify aspects of the overall process but must be adequately defined within the context of the publication.	[ "luteal regression", "corpus luteum", "luteolysis", "functional luteolysis", "structural luteolysis" ]	[ "P", "P", "P", "P", "P" ]
Eur_J_Clin_Pharmacol-4-1-2426925	Pharmacokinetics and tolerability of artesunate and amodiaquine alone and in combination in healthy volunteers	Objectives The WHO recommends artemisinin-based combination therapies for treatment of uncomplicated falciparum malaria. At least 15 African countries have adopted artesunate plus amodiaquine as treatment policy. As no pharmacokinetic data on this combination have been published to date, we investigated its pharmacokinetic interactions and tolerability in healthy volunteers in Africa. Introduction Since Plasmodium falciparum has developed resistance to almost all anti-malarial drugs, the use of combinations of effective antimalarials with different mechanisms of action is required to improve cure rates and delay drug resistance. The World Health Organisation (WHO) now recommends treatment of uncomplicated falciparum malaria with one of several artemisinin-based combination therapies (ACTs), including artesunate plus amodiaquine [1]. To date at least 15 African countries have adopted artesunate plus amodiaquine as first-line malaria treatment policy [2]. Although the limited data available suggest that the combination of artesunate and amodiaquine is well tolerated as short-course treatment, there are two safety concerns that have not yet been fully addressed. Neutropaenia and hepatitis curtailed the use of amodiaquine as prophylaxis in the 1980s [3–5]. A recent randomised, controlled trial comparing treatment with amodiaquine alone or in combination with artesunate found marked, asymptomatic neutropaenia on day 28 (NCI grade 3–4, absolute neutrophil counts 306–900 μL−1) in 6% of 153 children with uncomplicated falciparum malaria, all of whom had normal baseline absolute neutrophil counts [6]. We have previously reported a case of hepatitis in this group of volunteers that developed after the second dose of amodiaquine and was considered probably related to amodiaquine [7]. A similar report of delayed-onset asymptomatic hepatitis has been described in healthy American volunteers who had received artesunate plus amodiaquine alone followed by artesunate plus amodiaquine plus efavirenz, resulting in early study discontinuation [8]. In that study, efavirenz was thought to have caused an increase in amodiaquine concentrations by competitive inhibition of the P450 cytochrome enzymes that metabolise amodiaquine. With the widespread use of artesunate plus amodiaquine treatment in Africa, these studies support the need for laboratory monitoring of the treatment’s safety, particularly if antiretrovirals are used concomitantly [1]. Pharmacokinetic data are still sparse for amodiaquine. After oral administration, amodiaquine hydrochloride is rapidly absorbed and undergoes rapid and extensive metabolisation to desethylamodiaquine, the main active metabolite. Amodiaquine has a short elimination half-life (approximately 5 h), while that for desethylamodiaquine is much longer (6–18 days) [9, 10]. There are several pharmacokinetic studies on oral artesunate in healthy volunteers and malaria patients that give broadly consistent results [11, 12]. Artesunate is rapidly absorbed, rapidly converted to its main active metabolite dihydroartemisinin and rapidly cleared (elimination half-life generally <1 h). Dihydroartemisinin has similar pharmacokinetic characteristics but has a slightly longer elimination half-life. There is wide inter-individual variation in artesunate and dihydroartemisinin pharmacokinetic parameters. No pharmacokinetic interaction data have yet been published on amodiaquine when used in combination with artesunate. In view of the forthcoming widespread use of this combination in Africa, our study aimed to investigate potential pharmacokinetic interactions between these antimalarials and assess their tolerability in healthy volunteers in Africa. Methods Study subjects and study design Male or female healthy normal volunteers who provided written informed consent and met the following criteria were eligible to participate in the study: aged 18–45 years; no abnormalities on medical history, clinical examination, laboratory safety assessment (full blood count, differential white cell count, routine liver and renal function tests) or electrocardiogram; and a negative pregnancy test (for female volunteers). Volunteers were excluded if they were smokers (>5 cigarettes/day), had taken antimalarials or been in a malarial area in the preceding 8 weeks, had malaria parasites on a thick smear, used recreational drugs, or had ingested any alcohol or any medicines (including over-the-counter preparations) in the week preceding study commencement. This was a randomised three-phase crossover study. All volunteers took artesunate (4 mg/kg as Arsumax, 50 mg tablets, Sanofi-Aventis), the drug with the shorter elimination half-life, in the first phase, prior to the administration of amodiaquine. In the second phase, 7 days later, the volunteers were randomly allocated to one of two treatment groups: group 1 received a single oral dose of artesunate (4 mg/kg) plus amodiaquine (10 mg base/kg as Camoquin, 200 mg tablets, Parke Davis), and group 2 received a single oral dose of amodiaquine (10 mg/kg) alone. In phase three, 21 days later, the groups received the alternative regimen. All trial drugs were given under direct supervision with 200 ml tap water on an empty stomach after an overnight fast. No caffeinated drinks were allowed during the study period. Standardised meals were offered, commencing with breakfast at 2 h post-dose. Volunteers in phase one had blood samples collected before taking artesunate and at 0.25, 0.50, 0.75, 1.0, 1.5, 2, 3, 4, 6, 8 and 12 h after artesunate administration. In phases two and three (i.e. volunteers receiving artesunate plus amodiaquine or amodiaquine alone) venous blood samples were collected pre-dose and at 0.25, 0.5, 0.75, 1, 1.5, 2, 3, 4, 6, 8, 12, 24, 48 and 72 h, and days 4, 5, 7, 10, 14 and 20 post-dose. At each time two 5-ml venous blood samples were collected in tubes containing sodium heparin anticoagulant, while three 5-ml samples were collected at each time point for the first 12 h from those volunteers receiving both drugs. Each sample was centrifuged within 5 min of collection and the plasma transferred to separate plastic cryotubes and frozen at −70°C until analysed for artesunate, dihydroartemisinin, amodiaquine and desethylamodiaquine. Volunteers remained in the study ward for the first 12 h of each phase; thereafter specimens were collected at outpatient visits. A clinical examination and laboratory safety tests were performed during screening and 6 days after artesunate administration and 20 days after administration of both amodiaquine and amodiaquine plus artesunate. A 12-lead electrocardiogram (ECG) was recorded for each volunteer during screening and approximately 2 h after each drug dose. The corrected QT interval (QTc) was calculated by dividing the QT interval by the square root of the R-R interval. ECGs were reviewed by C.O. and by a consultant cardiologist. At follow-up visits, the volunteers were asked to report any new symptoms. A targeted physical examination was conducted if indicated. An adverse event together with its relationship to study drugs was defined according to the WHO guidelines [13] and the severity graded using the National Cancer Institute (NCI) Common Toxicity Criteria [14]. All adverse events and concomitant medications taken were recorded. Any significant clinical or laboratory abnormalities or the use of a concomitant medication with a previously described interaction resulted in the volunteer being withdrawn from the study. Ethical approval This clinical study was conducted in accordance with the principles laid down by the World Health Assembly of 1975 on Ethics in Human Experimentation and the Helsinki Declaration. The study adhered to the standards established for good clinical practice. The protocol was approved by the University of Cape Town Research Ethics Committee and the WHO Secretariat Committee for Research Involving Human Subjects (SCRIHS). Each volunteer was informed of the objectives, nature and possible risks of the trial. Written informed consent was obtained from every volunteer participating in the study. The volunteers were informed that they were free to withdraw consent at any time. Drug assays Artesunate and dihydroartemisinin concentrations in plasma were analysed using the method of Batty and colleagues [15] with minor modifications. Solid-phase extraction of samples was performed using Bond-Elut phenyl extraction cartridges (Varian, Harbour City, CA, USA). After washing the column with 2 ml of water, samples were eluted from the column with 1.5 ml of acetonitrile, which was then evaporated to dryness under nitrogen. Samples were reconstituted in 100 μl of mobile phase (45% acetonitrile, 0.05 M acetic acid, pH 6.0), and 50 μl was injected onto the column. HPLC was carried out on a Symmetry C 8, 5 μm, 15 cm × 4.6 mm HPLC column (Waters, Milford, MA, USA) using a Waters 600 pump and WISP 712 auto-sampler. The post-column reagent (1.2 M KOH in 90% methanol) was introduced using a Waters reagent delivery module and the mixture passed through a 1-ml reaction coil (Waters) kept at 69°C. Detection was at 290 nm using a Waters 484 Tunable Absorbance Detector, which produced a linear response over the range of the standard curve (20–1,000 ng/ml). Quality control samples of 15, 75 and 150 ng/ml for artesunate and 75, 450 and 750 ng/ml for DHA were included in each run. Within- and between-day coefficients of variation were below 14%. The lower limit of quantification was 24 ng/ml for both artesunate and dihydroartemisinin. Amodiaquine and desethylamodiaquine were analysed by LC mass spectrometry using an Agilent 1100 Series LC/MS system. Protein was precipitated from plasma samples (200 ul) using three volumes of acetonitrile. Supernatant (5 μl) was injected onto the HPLC. Chromatography was carried out using a 50 × 4.6 mm C18 Xterra column (Waters). The mobile phase comprised 75% acetonitrile, 0.02 M ammonia, pH 10.2. The extracted ion for amodiaquine was m/z 356 and for desethylamodiaquine m/z 328. The calibration curve for amodiaquine was linear in the range 5–100 ng/ml and for desethylamodiaquine in the range 5–400 ng/ml. Quality control samples of 25, 75 and 250 ng/ml were used for desethylamodiaquine and 2.5, 7.5 and 25 ng/ml for amodiaquine. Within- and between-day coefficients of variation were below 12%. The lower limit of quantification was 5 ng/ml for both amodiaquine and desethylamodiaquine. Statistical analysis Pharmacokinetic parameters were calculated from the artesunate, dihydroartemisinin, amodiaquine and desethylamodiaquine concentration profiles using WinNonLin, version 3.3 (Pharsight), with concentrations outside the specified limits of quantification regarded as missing values. Clearance (Cl) was calculated as 0.693 × Vd/T1/2 where Vd is the apparent volume of distribution and T1/2 is the elimination half-life. Because there was no intravenous comparator arm in this study, equivalent bioavailability was assumed for comparisons of apparent volume of distribution (Vd/f) and clearance (Cl/f), where f is the oral bioavailability or fraction of the drug absorbed (unknown). All four compounds were first analysed separately. Because both artesunate and dihydroartemisinin are highly parasiticidal in malaria patients, the pharmacokinetic parameters associated with therapeutic response [area under the plasma concentration time curve (AUC) and maximum clearance (Cmax)] are also reported together as dihydroartemisinin equivalents, defined as the sum of the measured dihydroartemisinin concentrations plus the artesunate concentration converted to dihydroartemisinin equivalents (using molecular weights of 384 and 284 for artesunate and dihydroartemisinin respectively) [16, 17]. As amodiaquine undergoes a rapid and extensive conversion to its active metabolite, desethylamodiaquine was considered the primary analyte [9, 10, 18]. The pharmacokinetic parameters following administration of artesunate and amodiaquine alone and in combination were compared using Stata, version 9.0 [Stata, College Station, TX, USA]. Data were log-transformed and then compared using the analysis of variance (ANOVA) for a cross-over design to take into account the repeated measures by study subject, treatment period and treatment groups and, for amodiaquine and desethylamodiaquine, a sequence effect. The treatment effects generated from the ANOVA were exponentiated in order to express comparisons between monotherapy and combination therapy as a ratio. Any apparent discrepancies between the difference in the group means and these ratios are due to the fact that the ratios are based on the within-patient differences in log-transformed values and not the group arithmetic means. Given the multiple testing, statistical significance of results should be interpreted with caution. For the safety analysis, means and 95% confidence intervals (95% CI) for each haematological parameter (haemoglobin, haematocrit, platelets, white cell count, absolute lymphocyte and neutrophil count) were calculated at baseline and at the end of each treatment phase. Changes in these values from baseline were determined using a mixed effect regression model that took into account repeated measures within-patient and were adjusted for the period and period-treatment interaction effects. Results Subject demographics Nineteen volunteers were screened; two were excluded because of neutropaenia and two withdrew consent. Fifteen healthy normal volunteers (10 male, 5 female) entered the study with a mean age, weight and height of 24.4 years, 67.3 kg and 171 cm, respectively. The mean dose of amodiaquine was 10.72 mg/kg and that of artesunate was 4.26 mg/kg. Safety analysis included all 15 volunteers; only data from the 13 volunteers who completed the study were included in the pharmacokinetic analyses. A female volunteer was withdrawn due to a new prescription of fluoxetine for depression, as this drug could potentially interact with artesunate and amodiaquine by inhibiting cytochrome P450 enzymes. One male volunteer was withdrawn due to a possible hypersensitivity reaction on his first exposure to amodiaquine. Pharmacokinetic parameters Effect of amodiaquine on artesunate pharmacokinetic parameters Data were analysed using a non-compartmental model. Figure 1a shows the mean plasma concentration time curves of artesunate and dihydroartemisinin following administration of artesunate alone and artesunate plus amodiaquine. The AUC, Cmax and time to Cmax (Tmax) of artesunate were similar following monotherapy and combination therapy (Table 1). Figure 2a illustrates the impact of adding amodiaquine on the dihydroartemisinin AUC. Following the combination therapy, dihydroartemisinin AUC was lower [ratio 67% (95% CI 51–88%); P = 0.008], Cmax was lower [ratio 51% (95% CI 33–78%); P = 0.005], T1/2 was longer [ratio 157% (95% CI 115–213%); P = 0.008] and Vd was larger [ratio 192% (95% CI 133–275%); P = 0.003], when compared with artesunate monotherapy (Table 1). These ratios have been adjusted for repeated measures by subject and period effects. As a measure of the total exposure to artemisinin derivatives, the Cmax and AUC of the combined DHA-equivalents were also analysed and found to be lower following combination therapy [Cmax ratio 51% (95% CI 32–84%), P = 0.011; AUC ratio 72% (95% CI 54–95%), P = 0.023] when compared with monotherapy. Cl and Tmax of dihydroartemisinin were similar when monotherapy and combination therapy were administered (Table 1). Fig. 1Profiles of mean drug concentrations over time following monotherapy (mono) and combination (combo) therapy for a artesunate (AS) and dihydroartemisinin (DHA) and b amodiaquine (AQ) and desethylamodiaquine (DEAQ)Table 1Pharmacokinetic parameters for artesunate and dihydroartemisinin when artesunate was administered alone and with amodiaquine (ACT) AUC (ng·h/ml)Cmax (ng/ml)Tmax (h)T1/2 (h)Vd/f (l)Cl/f (l/min)ArtesunateAS alone (mean ± SD)206.4 ± 135.5231.8 ± 155.00.62 ± 0.28NANANAACT (mean ± SD)183.3 ± 146.5141.6 ± 117.50.86 ± 0.67NANANAACT-to-monotherapy ratio (%) (mean, 95% CI)36 (8–173)25 (6–105)125 (76–202)NANANA Significance (ANOVA)0.180.0570.33DihydroartemisininAS alone (mean ± SD)2,044.4 ± 564.2844.5 ± 309.41.10 ± 0.951.46 ± 0.484.89 ± 1.672.46 ± 0.86ACT (mean ± SD)1,410.5 ± 543.6446.2 ± 239.52.08 ± 1.722.20 ± 0.859.68 ± 4.163.08 ± 0.82ACT-to-monotherapy ratio (%) (mean, 95% CI)67 (51–88)51 (33–78)165 (82–334)157 (115–213)192 (133–275)122 (96–156) Significance (ANOVA)0.0080.0050.150.0080.0030.097Treatment effects generated from the ANOVA were exponentiated to express within-subject comparisons between monotherapy and combination therapy as a ratio, adjusted for period effectsAUC Area under the plasma concentration time curve, Cmax maximum concentration, Tmax time to Cmax, T1/2 elimination half-life, Vd volume of distribution, f oral bioavailability or fraction of the drug absorbed, Cl clearance, AS artesunate, SD standard deviation, CI confidence interval, NA not available (too few detectable concentrations for calculation of T1/2, Cl, Vd)Fig. 2Stick plot comparing individual patient AUCs following monotherapy and combination therapy of a dihydroartemisinin (DHA) and b desethylamodiaquine (DEAQ). (Note: volunteer A was excluded from the statistical analyses) Effect of artesunate on amodiaquine pharmacokinetic parameters Amodiaquine samples for one volunteer were missing. Figure 1b shows the mean plasma concentrations of desethylamodiaquine over time following administration of amodiaquine alone and in combination. Figure 2b illustrates the impact of artesunate on the desethylamodiaquine AUC. The volunteer (subject A) who had the highest observed AUC for desethylamodiaquine following combination therapy [29,504 ng·h/ml compared with the mean of 8,437 (95% CI 5,744–11,131) ng·h/ml] was excluded from the pharmacokinetic analyses of desethylamodiaquine, as this was an extreme outlier and influential observation (Fig. 2b; Fig. 3). Fig. 3Stick plot comparing individual patients. Day 7 concentrations of desethylamodiaquine when amodiaquine was administered as monotherapy and in combination with artesunate. (Note: volunteer A was excluded from the statistical analysis) The impact of adding artesunate to amodiaquine on the pharmacokinetic parameters of amodiaquine and desethylamodiaquine is reported for the remaining 11 volunteers. The pharmacokinetic parameters of amodiaquine were similar when administered as monotherapy or in combination with artesunate (Table 2). The desethylamodiaquine AUC is lower [ratio 65% (95% CI 46–90%); P = 0.015], Tmax shorter [ratio 60% (95% CI 45–80%); P = 0.003] and clearance more rapid [ratio 164% (95% CI 112–243%); P = 0.016] when amodiaquine was administered in combination with artesunate rather than as monotherapy (Table 2). There was a trend towards a lower day 7 concentration of desethylamodiaquine [ratio 56% (95% CI 30–104%); P = 0.064] when amodiaquine was administered in combination (Fig. 3). These ratios have been adjusted for repeated measures by subject, period and sequence effects. The Vd and Cmax were not significantly different when monotherapy and combination therapy were administered. Table 2Pharmacokinetic parameters for desethylamodiaquine (DEAQ) when amodiaquine was administered alone and with artesunate (ACT) AUC (ng·h/ml)Cmax (ng/ml)Tmax (h)[Day 7] (ng/ml)T1/2 (h)Vd/f (l)Cl/f (l/min)AmodiaquineAQ alone (mean ± SD)162.4 ± 101.429.2 ± 10.92.32 ± 1.16NA5.3 ± 4.1361.0 ± 128.369 ± 59ACT (mean ± SD)108.5 ± 56.022.7 ± 9.02.18 ± 1.61NA3.9 ± 1.2467.7 ± 180.586 ± 26ACT-to-monotherapy (%) (mean, 95% CI)77 (47–127)78 (58–103)92 (57–147)NA74 (29–189)128 (93–175)172 (84–357) Significance (ANOVA)0.270.070.680.440.110.11DesethylamodiaquineAQ alone (mean ± SD)12,041 ± 3,480268.7 ± 70.83.68 ± 1.8519.4 ± 7.3240.8 ± 146.9234.1 ± 97.5768 ± 252ACT (mean ± SD)8,437 ± 4,009301.4 ± 166.12.18 ± 1.0313.3 ± 7.3136.9 ± 83.8210.8 ± 92.91,330 ± 735ACT-to-monotherapy ratio (%) (mean, 95% CI))65 (46–90)103 (73–147)60 (45–80)56 (30–104)53 (25–111)88 (51–149)164 (112–243) Significance (ANOVA)0.0150.820.0030.0640.080.580.016Treatment effects generated from the ANOVA have been exponentiated to express within-subject comparisons between monotherapy and combination therapy as a ratio, adjusted for period and sequence effectsAUC Area under the plasma concentration time curve, Cmax maximum concentration, Tmax time to Cmax, [Day 7] day 7 concentration, T1/2 elimination half-life, Vd apparent volume of distribution, f oral bioavailability or fraction of the drug absorbed, Cl clearance, AQ amodiaquine, SD standard deviation, CI confidence interval, NA not available (as too few patients had quantifiable amodiaquine concentrations) Safety Twenty-nine adverse events were reported over the course of the study by 10 (67%) of the 15 volunteers. The frequency of adverse events was similar across all three treatment arms (P > 0.10): 4/15 (27%) following artesunate alone, 8/15 (53%) following amodiaquine alone, and 5/15 (33%) following the combination. All adverse events were consistent with the product information available, resolved spontaneously and, except for the transaminitis (see below) were mild or moderate in intensity. A case of asymptomatic, prolonged, severe transaminitis (NCI grade 3, AST 5–20× upper limit of normal) that developed after phase 3 of the study has been previously published [7]. This was considered to be probably related to amodiaquine, rather than artesunate, before the pharmacokinetic results were available. The subsequent finding that this volunteer had the highest measured desethylamodiaquine AUC following the administration of artesunate plus amodiaquine [29,504 ng·h/ml compared with the combination therapy mean of 8,437 (95% CI 5,744–11,131) ng·h/ml] strengthened this assessment. There were no liver function abnormalities detected in any other volunteers. No renal or electrolyte abnormalities were detected in any volunteers. Significant haematological changes were confined to the white cell counts. During phase 2 of the study, two volunteers (one following amodiaquine alone, one following the combination) developed asymptomatic, NCI grade 1 leucopaenia (3 to <4 × 109/L) and either a grade 1 (1.5 to <2 ×109/L) or grade 2 (1 to <1.5 × 109/L) neutropaenia. None of the other observed changes in haematological parameters were outside the normal range. There were no significant changes following treatment seen on the ECGs; the mean (95% CI) QTc interval was 398 (390–406) ms at screening, 401 (393–409) ms following artesunate alone, 400 (385–414) ms following amodiaquine alone and 412 (400–424) ms following the combination. In two patients, the prolongation following treatment with amodiaquine or artesunate plus amodiaquine was considered of borderline clinical significance. Discussion The pharmacokinetic analyses in this randomised, cross-over study showed statistically significant pharmacokinetic interactions resulting in reductions in the AUC of both dihydroartemisinin and desethylamodiaquine when artesunate and amodiaquine were given in combination to healthy volunteers. Of further concern is that one healthy volunteer failed to reach quantifiable concentration of both artesunate and dihydroartemisinin throughout the initial 12 h. There is insufficient evidence currently available to explain the basis of these interactions. The clinical significance of our findings is unclear because we studied healthy volunteers. In one study comparing pharmacokinetic parameters of dihydroartemisinin in patients and volunteers, patients had a significantly higher Cmax (1,948 vs. 1,192 nmol/l) and AUC0–12 (4,024 vs. 1,763 nmol·h/l), while their Vd and Cl were both lower. The Tmax and terminal elimination half-life of the patients were similar to those of the volunteers [19]. Desethylamodiaquine pharmacokinetic parameters in malaria-infected adults in Africa [20] and children with acute malaria in Papua New Guinea [21] were similar to those of normal volunteers. In artemisinin-based combination therapy, the artemisinin is given to kill most of the parasites quickly, while the longer-acting partner drug (e.g. amodiaquine) is given to kill the residual parasites. For an antimalarial with a long elimination half-life like desethylamodiaquine, the key pharmacokinetic parameter for parasite killing is the AUC, which is a useful predictor of therapeutic response as it captures both the drug concentration and the duration of exposure [22, 23]. For such drugs the antimalarial concentration on day 7 is strongly correlated with AUC, and thus treatment response [23]. An association of desethylamodiaquine concentration on day 7 and amodiaquine treatment response has been found in children with uncomplicated falciparum malaria [18]. In our study, the AUC (and probably the day 7 desethylamodiaquine concentration) was reduced following the artemisinin-based combination; this may result in reduced parasite killing, particularly in areas where there is reduced amodiaquine sensitivity. Although there is a linear relationship between amodiaquine dose and the desethylamodiaquine AUC [24], the wide inter-individual variability in desethylamodiaquine levels (five-fold in our study), and the consequent potential for drug-induced toxicity, could preclude increasing the amodiaquine dose to compensate for the lower desethylamodiaquine AUC observed following co-administration with artesunate. Extensive clinical trial data show that artesunate plus amodiaquine is more effective (at clearing parasitaemia and preventing recrudescence) than amodiaquine alone [1, 6]. Therefore, if reductions in the Cmax and AUC of desethylamodiaquine and dihydroartemisinin occur in patients, the negative effect on parasite killing may become clinically insignificant due to the additive (or possibly synergistic) parasiticidal effects of the combination. However, this may only apply to areas where the level of amodiaquine resistance is low. Although clinically significant resistance to the artemisinins has not yet been documented, amodiaquine resistance is well documented and this is probably the principal explanation for low artesunate plus amodiaquine cure rates observed in a number of African countries [1, 6]. Despite our small sample size, we detected the well-known amodiaquine-related adverse drug reactions. Clinically important adverse events were reported by a quarter of these volunteers, even though all were healthy and only two doses were administered 3 weeks apart (rather than the daily administration for 3 days recommended for malaria treatment). When the artesunate plus amodiaquine combination becomes widely used, monitoring of liver function tests and haematological parameters are warranted to define these risks in malaria patients, particularly in patients at special risk such as those at the extremes of age, with HIV co-infection or malnutrition [1]. However, the feasibility of this recommendation outside of the research setting is of concern. To conclude, the total exposure to the main active metabolites of both artesunate and amodiaquine was significantly reduced when administered in combination to healthy African volunteers; this might be important clinically. However, because cure rates with this combination are generally higher than with amodiaquine monotherapy, artesunate and amodiaquine could remain in the armamentarium of drugs used to combat falciparum malaria, provided efficacy continues to be monitored and adequate safety precautions can be taken. Further pharmacokinetic research on artesunate plus amodiaquine, when administered concurrently or as a fixed dose combination, is urgently required in patients with malaria to establish the extent and clinical significance of these pharmacokinetic interactions.	[ "pharmacokinetics", "tolerability", "artesunate", "amodiaquine", "healthy normal volunteers" ]	[ "P", "P", "P", "P", "P" ]
Virchows_Arch-3-1-1888720	Elevated expression of cyclooxygenase-2 is a negative prognostic factor for overall survival in intrahepatic cholangiocarcinoma	The production of prostaglandins is regulated by cyclooxygenases (COXs), which also have a role in tumour development and progression in various human malignancies, including cholangiocarcinoma. Limited information is available of the correlation of COX-2 protein expression and prognosis in intrahepatic cholangiocarcinoma (ICC). The aim of the present study was to determine the clinical significance of COX-2 expression in ICC. In addition the correlation of COX-2 expression and apoptosis/proliferation was analysed. COX-2 expression was determined immunohistochemically in 62 resected ICCs. Proliferation was assessed using Ki67-immunohistochemistry, and apoptosis was measured with the TdT-mediated dUTP nick-end-labelling technique. COX-2 was identified as an independent prognostic factor (P = 0.028) in resected ICC by survival analysis. High levels of COX-2 expression were found to be associated both with reduced apoptosis and increased proliferation of tumour cells. This study demonstrates the independent prognostic value of the COX-2 expression in resected ICC, thus, offering a potential additional adjuvant therapeutic approach with COX-2 inhibitors. Introduction Cholangiocarcinoma (CCC) is composed of cells resembling those of bile ducts. CCC arises either from the extrahepatic duct including the hilar bifurcation or from the intrahepatic ducts. Intrahepatic cholangiocarcinoma (ICC) is less frequent than extrahepatic bile duct carcinoma but second among primary malignant liver tumours after hepatocellular carcinoma. In Western countries, ICC accounts for about 10% of the primary liver malignancies with increasing incidence in the last two decades [24]. The identification of patients with most aggressive ICC might help to optimise surgical treatment and avoid unnecessary surgical therapy. Therefore, it is desirable to identify molecular markers that individually predict tumour behaviour and may facilitate an individualized therapy. Among the potential prognostic parameters, cyclooxygenase (COX)-2 is of particular interest, as it may also offer the option of treatment with nonsteroidal anti-inflammatory drugs (NSAIDs) [15]. COXs regulate the synthesis of prostaglandins and are, thus, the major target of NSAIDs. Its two isoforms (COX-1 and COX-2) have different expression patterns, with COX-1 being expressed in a broad variety of tissues. COX-2 has been shown to participate in tumour development and progression [11, 21]. Both elevated COX-2 messenger RNA (mRNA) and protein levels were found to be raised in ICC [3, 9]. There exist only few studies dealing with the potential prognostic value of the COX-2 expression. A small study with only 24 cases of CCC could not prove a significant association with survival; however, this study did not discriminate between extra- and intrahepatic cholangiocarcinoma [12]. Another study on extrahepatic cholangiocarcinoma failed to demonstrate a correlation of the COX-2 expression and the patients’ outcome in CCC [16]. In the present study, COX-2 expression was determined by immunohistochemistry in one of the largest contemporary series of consecutive resected ICC. The findings were correlated with a broad range of clinicopathological features. Recent observations suggest that COX-2 overexpression may cause a prostaglandin E2-mediated inhibition of apoptosis in CCC. Moreover, COX-2 overexpression was shown to increase cell growth by the activation of E series of prostaglandins (EP receptors) [8, 31]. Thus, apoptosis and proliferation was determined in this series to monitor a putative imbalance between cell death and cell proliferation induced by COX-2 protein overexpression [26]. Materials and methods Between August 1998 and August 2006, a total of 62 patients with a mean age of 58 ± 11.5 years were available for this study. The study comprised consecutive patients who underwent surgery for liver resection. Patients solely undergoing an explorative laparatomy without subsequent resection were excluded from the study. Patients with hilar cholangiocarcinoma, gallbladder carcinoma, or mixed hepato/-CCC and liver cirrhosis were also excluded from the study. The diagnosis of ICC was based on histology by the examination of the resected liver specimen. The tumours were classified according to the pathologic TNM (pTNM) system (sixth edition) [28]. Detailed clinical data was available including preoperative therapy, operative details, and pathological findings including surgical radicability, tumour staging, and clinical follow-up. One patient suffered from primary sclerosing cholangitis without cirrhosis; hepatolithiasis was not present in any case. Data were completed by August 2006, and minimum follow-up was every six months or until death. The median length of the follow-up was 12 months. Immunohistochemistry Immunohistochemistry was performed with an automated staining device (Dako Autostainer, Glostrup, Denmark). Cyclooxygenase-2 In this study, a monoclonal rabbit anti-human COX-2 antibody (DCS, Hamburg, Germany) was used. Immunohistochemistry was performed on 5-μm-thick paraffin, and antigen retrieval was carried out with 0.01-M citrate buffer at pH 6.1 for 20 min in a hot water bath (95°C). The primary antibody was incubated for 30 min at 1:250 dilution. Antibody demonstration was achieved using the commercially available anti-mouse IgG detection kit (EnVision, DakoCytomation, Carpenteria, CA, USA) The replacement of the primary antibodies by mouse immunoglobin served as negative controls. Positive controls (colorectal carcinoma) were included in each staining series. In ICC, COX-2 was scored according to the amount of positive stained tumour cells. One complete tumour slide was examined for specific cytoplasmic COX-2 immunostaining. If none or less than 10% of the tumour cells showed specific COX-2 immunostaining regardless of staining intensity, the case was classified as negative. The cases with 11–50% of the positively stained tumour cells were classified as moderately positive and tumours with more than 50% stained tumour cells as strongly positive. Ki67 immunostaining and TdT-mediated dUTP nick-end labelling Ki67 immunohistochemistry was performed on 5-μm-thick paraffin sections. Dewaxed and rehydrated sections were incubated with hydrogen peroxide to block endogenous peroxidase. After the antigen retrieval in a hot water bath, the prediluted monoclonal anti-Ki67 antibody (Biogenex, San Ramon, USA) was incubated for 30 min; antibody demonstration was performed with the commercially available anti-mouse IgG detection kit (EnVision, DakoCytomation). The replacement of the primary antibodies by mouse immunoglobin served as negative controls. The growth fraction was defined as the percentage of Ki67-positive, randomly chosen nuclei per 600 tumour cells. In situ DNA fragmentation was established using the terminal desoxyribonucleotide transferase TdT-mediated dUTP nick-end-labelling technique (TUNEL) in paraffin-embedded sections. We used ApoTag™ plus peroxidase in situ apoptosis detection kit (Intergen). The staining procedures were performed according the manufacturer’s recommendations. The percentage of the stained apoptotic tumour cells per 600 randomly chosen tumour cells was calculated. To avoid miscounting of the necrotic cells, corresponding H&E sections were examined. Statistical analysis COX-2 immunostaining was assessed by two of the authors (K.J.S., H.R.) in a blind-trial fashion without knowledge of the clinical outcome. In case of disagreement, slides were re-evaluated by both investigators until agreement was reached. All data were converted to a PC and statistically analysed using SPSS version 12 for Windows Statistical Package for Social Sciences, Chicago, IL, USA. The kappa coefficient of the inter-observer agreement was calculated for the semiquantitative COX-2 immunostaining results; the interpretation of the kappa value was performed using the commonly cited scale of Landis and Koch [19]. The interobserver agreement of COX-2 immunostaining was substantial (kappa = 0.73). Relationships between ordinal parameters were investigated using the two-tailed χ2 analysis (or the Fisher’s exact test where patient numbers were small). The relationship between categorical data (e.g. COX-2) and numeric data (e.g. number on Ki-67 positive tumour cells) was determined using the ANOVA test. The overall survival (OS) curves were estimated using the Kaplan–Meier method, and any differences in the survival curves were compared by the log-rank test. For multivariate analysis, the COX regression model was used. Overall, 95% confidence intervals were used throughout. Results COX-2 immunohistochemistry The COX-2 immunostaining was located in the cytoplasm of epithelial cells. The plasma cells exhibited a positive cytoplasmic immunostaining, serving as internal control. Noncancerous intrahepatic bile duct epithelial cells next to invasive cancer showed frequently a positive COX-2 immunoreactivity, whereas no COX-2 staining could be detected distant from the tumour (Fig. 1). Hepatocytes near and distant from ICC revealed a positive staining in all cases. In all, 12 (19.4%) tumours were classified as negative, 25 (40.3%) as moderately positive, and 25 (40.3%) as strongly positive. Epithelial tumour cells exhibited a broad range of COX-2 immunoreactivity with a complete lack of staining or a specific and strong granular cytoplasmic COX-2 immunostaining (Fig. 2). COX-2 staining was not associated with any of the investigated clinicopathological parameters (Table 1). Fig. 1Immunostaining for COX-2 in noncancerous intrahepatic bile duct epithelial cells. Whereas portal bile ducts distant from the ICC consistently lacked COX-2 immunostaining (asterisk, left side), bile duct epithelial cells (asterisk, right side) adjacent to ICC (arrow) exhibited cytoplasmic immunoreactivity of varying intensity. Notice the COX-2 protein expression in normal hepatocytes. Original magnification, ×400Fig. 2Representative COX-2 immunostaining results in ICC. Missing COX-2 staining (left) in contrast to strong cytoplasmic COX-2 immunoreactivity in ICC tumour cells (right). Original magnification, 400×. Inset: positive control (colorectal carcinoma) with strong immunostainingTable 1COX-2 expression, demographic characteristics, and clinicopathological characteristics in 62 patients of the resected ICC AllCOX-2 negativeCOX-2 moderateCOX-2 strongP values n (%)6212 (19.4)25 (40.3)25 (40.3)Gender (male/female)7/59/1610/150.422Mean age at diagnosis (y±SD)58.8 ± 11.552 ± 12.160.4 ± 9.8 60.6 ± 11.80.067Grading0.612 G1 (n%)2 (3.2)0 (0)2 (100)0 (0) G2 (n%)41 (66.1)7 (17.1)15 (36.6)19 (46.3) G3 (n%)19 (30.6)5 (26.3)8 (42.1)6 (31.6)UICC stage grouping0.277 I166 (37.5)3 (18.8)7 (43.8) II30 (0)1 (33.3)2 (66.7) IIIA,B,C385 (13.2)19 (50)14 (36.8) IV51 (20)2 (40)2 (40)Staging0.390 T1 (n,%)217 (33.3)5 (23.8)9 (42.9) T2 (n,%)70 (0)4 (57.1)3 (42.9) T3 (n,%)284 (14.3)13 (46.4)11 (39.3) T4 (n,%)6 1 (16.7)3 (50)2 (33.3)Lymph vessel infiltration0.121 L0477 (14.7)18 (38.3)22 (46.8) L1155 (33.3)7 (46.7)3 (20)Blood vessel infiltration0.916 V0398 (20.5)15 (38.5)16 (41) V1234 (17.4)10 (43.5)9 (39.1)Nodal statusa0.281 03710 (27)15 (40.5)12 (32.4) 1212 (9.5)10 (47.6)9 (42.9)Distant metastasis0.999 No5711 (19.3)23 (40.4)23 (40.4) Yes51 (20)2 (40)2 (40)Solitary tumourb327 (21.9)11 (34.4)14 (43.8)0.545Multifocal tumour295 (17.2)14 (48.3)10 (34.5)Tumour size 8.18.75 ± 4.3 8.3 ± 2.97.4 ± 3.60.506Resection status0.627 R0399 (32.1)15 (38.5)15 (38.5) R1/2c233 (13)10 (43.5)10 (43.5)P values were calculated using chi-square analysis and ANOVA for continuous variablesaData available for 58 patients; bdata available for 61 patients; c20 cases were classified R1, only 3 were R2 Clinical outcome by COX-2 expression OS dependent on COX-2 expression was analysed using Kaplan–Meier survival (n = 62; p = 0.036; Fig. 3). Patients with a strong COX-2 expression exhibited a significantly decreased OS, whereas patients classified as COX-2 negative revealed the most favourable OS. The unfavourable prognostic effect of tumours with elevated COX-2 protein expression remained in the subgroup of patients with primarily R0-resected ICC (n = 39; p = 0.038). Fig. 3Kaplan–Meier survival curves in resected intrahepatic cholangiocarcinoma in relation to different COX-2 expression. Tumours with elevated COX-2 levels exhibit a significantly decreased overall survival (p = 0.036, Log Rank test) The parallel univariate survival analysis showed a significant association of the resection status (p < 0.001), multifocal tumour growth (p = 0.005), higher International Union Against Cancer (UICC) stage (p = 0.012), and vascular invasion (p = 0.011) with a reduced OS. Table 2 shows the mean OS for all patients in the various subgroups. Table 2Mean overall survival in relation to different patient subgroupsSubgroupsMean overall survival in all ICC patients (months)All patients16.5R0-resected patients19.7R1-resected patients11.4R2-resected patients9.3COX-negative patients22.8Moderate COX-2 expression18.7Strong COX-2 expression11.4 To clarify the independent prognostic value of the COX-2 expression in the patients with resected ICC, a multivariate analysis of the relevant parameters was performed. The COX regression analysis revealed the R-classification to be the best prognostic factor for OS followed by the COX-2 expression (Table 3). Table 3Multivariate COX regression analysis for overall survival in 62 patients with resected intrahepatic cholangiocarcinoma HR (95% CI) P valuespTstage I–II/III–IV1.63 (0.44–6.02)0.465UICC I–II, III–IV1.20 (0.62–2.35)0.589Multifocality, solitary/multifocal1.26 (0.41–3.89)0.689Resection status, R0/R1, R2a3.08 (1.27–7.44)0.012Vascular invasion, V0/V11.98 (0.88–4.41)0.095COX-2, negative/moderate/strong1.07 (1.07–3.49)0.028aTwenty cases were classified R1, only three were R2 COX-2 and apoptosis/proliferation There were significant lower mean numbers of apoptotic tumour cells per 600 tumour nuclei in tumours with strong COX-2 expression in contrast to tumours with moderate or negative COX-2 immunoreactivity (Table 4). In contrast, the proliferative activity increased in tumours with high levels of COX-2, reflected by a higher mean value of proliferating tumour cells. However, due to a substantial overlap of values, statistical significance was not reached. Table 4Apoptosis (TUNEL) and proliferative activity (Ki67) in relation to COX-2 expression All (mean value/SD)COX-2 negativeCOX-2 moderate/strongP valuesApoptosis (percentage of positive tumour cells)2.17 ± 2.403.84 ± 4.041.76 ± 1.730.028Growth fraction (percentage of Ki67 positive tumour cells)9.74 ± 7.378.55 ± 8.419.99 ± 7.340.486P values were calculated using ANOVA for continuous variables Discussion CCC is a devastating cancer of the hepatic biliary tract. The assessment of the individual prognosis of ICC patients may be helpful to optimise decision making for surgical treatment and avoid unnecessary surgical therapy. At present, surgical therapy remains the only curative treatment for ICC; thus, novel therapeutic strategies for this aggressive cancer are necessary. This study on a large series of 62 patients with resected ICC showed the independent prognostic value of immunohistochemically determined COX-2 protein expression, thus, offering a potential additional adjuvant therapeutic approach with COX-2 inhibitors. Elevated COX-2 expression was associated with a marked reduced OS in multivariate analysis. The main advantage of our study is the large number of patients (n = 62) in combination with a high homogeneity of this series composed of consecutively resected ICC. A previous study on extrahepatic cholangiocarcinoma failed to demonstrate a relation of COX-2 expression with clinical outcome [16]. This suggests (genetic) differences depending on the site of origin within the biliary tract. Mutations of the RAS and TP53 genes are the most common genetic abnormalities identified both in ICC and cancer of extrahepatic bile duct (BDC) [13, 14, 16, 20, 29]. However, recent studies disclosed genetic differences regarding p53 protein overexpression among carcinomas of the proximal (ICC) and distal bile ducts (BDC), indicating different patterns of the inactivation of tumour-suppressor genes [1]. Because cell line studies suggest COX-2 both as an effector and inhibitor of the tumour suppressor p53 or DNA-damaging agents [4, 7], different genetic patterns in ICC and BDC, with varying induction of COX-2 gene/protein expression, might explain the diverse clinical impact of COX-2 in ICC and BDC. A small study on 24 patients with CCC failed to show a prognostic significance of the COX-2 expression [12]. However, the power of this study is limited due to two reasons: (1) The authors did not discriminate between extra- and intrahepatic cholangiocarcinoma, and (2) the statistical power of a small series with 24 patients is very low. However, Javle et al. [12] demonstrated in their study that the median survival time of patients with low COX-2 expression was more than twice as high than that of patients with high COX-2 expression. This finding is in accordance with our results identifying COX-2 overexpression as a predictor of reduced OS. Javle et al. found COX-2 to be expressed in all 24 ICC, whereas in our study, only 81% of ICC exhibited an either moderate or strong COX-2 expression. Kim et al. [16] found COX-2 to be overexpressed in only 52% of the examined extrahepatic cholangiocarcinomas. These discrepancies are likely to be caused by differences in antigen retrieval and varying scoring systems, emphasizing the relevance of standardized protocols and immunoscores. Previous studies suggest COX-2 overexpression as an early carcinogenetic event in the human biliary tract and pointed out that COX-2 expression was highest in more differentiated CCC, whereas COX-2 expression decreased with the loss of differentiation [5]. Our study did not reveal a significantly increased COX-2 expression in tumours with higher differentiation. Nevertheless, a detailed analysis of our data rather supports this notion. The group of well and moderately differentiated tumours exhibited the highest amount of cancers classified as strongly COX-2 positive (44.2%) in contrast to ICC with poor tumour differentiation (31.6%). On the other hand, we found the lowest levels of COX-2 expression to be more frequent in poorly differentiated tumours (26.3%) in contrast to well or moderately differentiated ICC (16.5%). The study by Kim et al. [16] on 102 extrahepatic cholangiocarcinoma describes no significant differences in terms of tumour differentiation. In the light of these varying data, one should be careful not to draw premature conclusions about the differential expression COX-2 protein in ICC. The clinical conditions associated with ICC include parasitic liver fluke infestations, hepatolithiasis, which is frequently observed in clonorchiasis [10], nonbiliary cirrhosis, and primary sclerosing cholangitis [30]. It would be interesting to compare the levels of COX-2 expression in ICC with or without these established associated clinical risk factors. However, our series is not suitable to answer this question because all cases with liver cirrhosis, regardless of its origin, were excluded from our study to maintain the high homogeneity of our cohort and rule out cirrhosis- related influence on patient survival. As a consequence, this series comprises only one patient with associated primary sclerosing cholangitis. Patients with liver fluke infestations and associated hepatolithiasis were not present in this series because our cohort is composed from patients living in western Europe. Interestingly, we found the up-regulation of COX-2 protein expression not only in cancerous epithelial cells but also in noncancerous epithelial cells adjacent to invasive cancer. Noncancerous epithelial cells distant from the ICC lacked COX-2 immunostaining. This staining pattern is in accordance with the results of Hayashi et al. [9]. The elevated expression of COX-2 in bile duct epithelial cells adjacent to invasive cancer is not surprising, as inflammation often accompanies invasive cancer and inflammatory cytokines such as IL-1α [22], IL-1β [2], or TNF-α are believed to induce COX-2 gene at the transcriptional level [23]. We observed a constant positive specific immunostaining in normal hepatocytes regardless of the distance to the ICC. This finding is in agreement with previous studies documenting elevated COX-2 protein and mRNA expression in nontumourous liver [3, 17]. Against the background of NSAIDs, recent studies have evaluated selective COX- inhibitors for their effect on CCC cells cell growth and invasion in vitro and in nude mice [6, 18, 27, 32, 33]. The treatment with COX-2 inhibitors resulted in induced apoptosis and inhibited proliferation. The mechanisms by which COX-2 contributes to the poor prognosis in variety of human cancers have not been completely elucidated yet. COX-2 has been shown to contribute to both tumourigenesis and the malignant phenotype of tumour cells by different mechanisms, including (1) the inhibition of apoptosis by the activation of the PI3-Kinase/Akt pathway and increased production of PGE2, leading to increased expression of Bcl-2 and attenuation of nitric oxide signalling, (2) increased angiogenesis by increased PGE2 production with subsequent vascular endothelial growth factor production, (3) increased invasiveness by the overexpression of CD44, and (4) increased cell growth by the activation of EP receptors [8, 31]. The above mentioned effect of COX-2 overexpression on apoptosis and proliferation can be confirmed in our study. COX-2 over-expressing ICC revealed a significantly decreased apoptosis and a higher mean value of proliferating tumour cells, although the latter finding did not reach statistical significance. This finding suggests an impaired balance between cell loss and cell gain, resulting in a shift towards tumour net tumour growth due to increased apoptosis. The observed decrease of apoptosis in COX-2 overexpressing ICC might be due to the PGE2-mediated induction of the antiapoptotic protein Bcl-2 and increased NFκB transcriptional activity, which is a key antiapoptotic mediator [25, 26]. In conclusion, this study demonstrates the independent prognostic value of immunohistochemical COX-2 protein expression in resected ICC, thus, offering a potential additional adjuvant therapeutic approach with COX-2 inhibitors and facilitating an optimised therapeutic strategy. Our results suggest that COX-2 overexpression causes a shift towards increased tumour cell proliferation and decreased apoptosis contributing to the unfavourable clinical course.	[ "immunohistochem", "cancer", "mol path" ]	[ "P", "P", "U" ]
Mcgill_J_Med-11-1-2322923	Asian wasp envenomation and acute renal failure: a report of two cases	Acute renal failure is an unusual complication of wasp stings. We report two cases of renal failure after multiple wasp stings (Vespa affinis). Both patients had evidence of intravascular haemolysis, hepatic dysfunction, oligo-anuria and azotaemia and required dialysis. The first patient had severe hemolysis, rhabdomyolysis, pigment and venom nephropathy and died on the 8th day in hospital. The second patient, who recovered completely in 3 weeks time with steroid and antihistaminic therapy, had interstitial nephritis. Although acute renal failure after wasp stings is typically caused by acute tubular necrosis (ATN) in the setting of haemolysis or rhabdomyolysis, in some patients, acute renal failure may result from a direct nephrotoxic effect or acute interstitial nephritis from a hypersensitivity reaction. INTRODUCTION In the Pokhara Valley, Nepal, there are many unpublished cases of wasp poisoning which take a heavy death toll annually. Wasp stings usually cause local allergic reactions but can sometimes lead to intravascular haemolysis, rhabdomyolysis, thrombocytopenia, acute tubular necrosis, acute hepatic injury (1) and even myocardial infarction (2) in addition to various respiratory and neurological (3) manifestations. Death from wasp envenomation is a rare event and results from acute renal failure (ARF) involving various mechanisms. Although ARF after wasp stings is typically caused by ATN in the setting of haemolysis or rhabdomyolysis, in some patients, renal failure may result from a direct nephrotoxicity of wasp venom or acute interstitial nephritis from a hypersensitivity reaction. We here we report two cases of acute renal failure after wasp stings (Vespa affinis). CASE 1 A 55-year-old farmer who had been collecting fodder from a jungle was admitted with dyspnea, hoarse voice and myalgia within 8 hours of being attacked by several wasps. The patient was given intravenous saline, oxygen, salbutamol (β2-adrenergic receptor agonist), chlorpheniramine (antihistamine), cyproheptadine (antihistamine), prednisolone (corticosteroid) and ranitidine (histamine H2-receptor antagonist). He also received fluid, mannitol and furosemide. By the next morning he had haemoptysis and had produced 400 ml of dark urine. On examination, the patient was drowsy, pale, icteric and cyanosed and had approximately one hundred and fifty red and swollen sting marks all over the body. Systemic examination revealed polyphonic wheezes and crepitations at the base of the right lung. Investigations are shown in Table 1. The onset of the oliguric phase was at 12 hours postenvenomation. Chest radiograph showed right basal consolidation. He died on the 8th day following admission despite aggressive therapy with medication, blood transfusions, assisted ventilation, and 16 cycles of dialysis. CASE 2 A 40-year-old forest guard was attacked by a swarm of wasps. He presented with approximately twenty-five sting marks in exposed areas of face, throat, hands and legs [Fig. 1]. The patient was treated in the primary health centre with rubbing of saliva and papaya slices over the sting marks and referred to teaching hospital almost 24 hours after being stung. He developed anuria and had not passed any urine the previous night. On examination, the patient had a rapid pulse, unrecordable blood pressure, icterus, an urticarial rash in exposed parts of the face, legs and hands, facial puffiness, and a swollen left ankle and right knee joint. The rest of the physical examination was unremarkable. Investigations are shown in Table 1. He was treated with oxygen, salbutamol, chlorpheniramine, cyproheptadine, prednisolone and ranitidine. His hepatic and renal function improved gradually with fluid challenge, furosemide, mannitol, bicarbonate infusion, dopamine, and 12 cycles of dialysis in 3 weeks’ time. OBSERVATION DISCUSSION Wasp stings are well-known causes of toxic and hypersensitivity reactions. Direct toxicity is rare, but has been reported in cases when a very large amount of venom is injected. Immediate hypersensitivity reactions, such as bronchospasm in the first case and an urticarial rash in the second, are known to occur. In our report, both reactions responded to steroids and antihistaminics. The second patient had swollen joints indicating serum sickness-like reaction in a sensitized individual. Wasp venom contains toxic melittin, apamine, phospholipases A1, hyaluronidase, acid phosphatase, histamine, and degranulating peptide mastoparan (4). These components have direct and indirect cytotoxic (hepatic, renal and myocyte membrane), hemolytic, neurotoxic and vasoactive properties, which can cause intravascular haemolysis and rhabdomyolysis (5, 6). Wasp venom can cause ARF by several mechanisms, which include ATN, acute interstitial nephritis, pigment nephropathy resulting from rhabdomyolysis (myoglobinuria) or intravascular haemolysis (haemoglobinuria) and hypotension caused by an anaphylactic reaction (7, 8). Previously rhabdomyolysis and renal ischemia were thought to be main causes of nephropathy. Sakhuja et al had postulated that direct toxic injury could be one of the possible mechanisms of ARF following wasp poisoning (9). Many cases of rhabdomyolysis-associated ARF have been published, but those due to wasp stings are rare. The wasp venom has deleterious effect on renal tubules and glomeruli (albuminuria, haematuria and ARF), red blood cells (haemolysis, reticulocytosis, unconjugated hyperbilirubinaemia), muscles (rhabdomyolysis, elevated creatinine phosphokinase and lactate dehydrogenase, myoglobinuria) and liver (elevated transaminases, hypoalbuminaemia and prolonged prothrombin time) (10). Kularatne et al had described similar multi-organ failure with high mortality following wasp poisoning owing to direct toxic effect (11). In the first case we presented, the patient had myalgia, indicating muscle injury as evidenced by elevated CPK, LDH and AST and myoglobinuria (728 ng/ml). He also had intravascular haemolysis and haemoglobinuria. Toxic pigments might have caused nephropathy resulting in ARF. The alternative mechanism of ARF postulated was direct nephrotoxicity by massive wasp venom. Zhang R et al. (12) reported for the first time that acute tubulointerstitial nephritis could lead to ARF in wasp sting cases. In the second case we present, the patient had eosinophiluria, indicating interstitial nephritis. He recovered fully with the mentioned treatment. He did not report taking any medication which might have had nephrotoxic side-effects, and no other causes of ATN could be found. Ultrasound abdomen was unremarkable. Kidney biopsy revealed proximal peritubular necrosis and eosinophilic infiltration. Hence it can be hypothesized that the ATN was caused by a hypersensitivity reaction to the wasp venom. CONCLUSION Wasp stings pose a great environmental hazard in Nepal and early recognition of anaphylactic shock, hepatic or renal dysfunction, rhabdomyolysis or haemolysis and rapid transport to hospital are essential steps of management to avoid fatalities. ARF due to toxic or pigment nephropathy and tubulointerstitial nephritis should be considered in any oliguric and azotemic patient following wasp attack.	[ "wasp envenomation", "acute renal failure", "rhabdomyolysis", "interstitial nephritis", "nepal" ]	[ "P", "P", "P", "P", "P" ]
Mol_Biochem_Parasitol-1-5-1885993	Kinetoplastid PPEF phosphatases: Dual acylated proteins expressed in the endomembrane system of Leishmania	Bioinformatic analyses have been used to identify potential downstream targets of the essential enzyme N-myristoyl transferase in the TriTryp species, Leishmania major, Trypanosoma brucei and Trypanosoma cruzi. These database searches predict ∼60 putative N-myristoylated proteins with high confidence, including both previously characterised and novel molecules. One of the latter is an N-myristoylated protein phosphatase which has high sequence similarity to the Protein Phosphatase with EF-Hand (PPEF) proteins identified in sensory cells of higher eukaryotes. In L. major and T. brucei, the PPEF-like phosphatases are encoded by single-copy genes and are constitutively expressed in all parasite life cycle stages. The N-terminus of LmPPEF is a substrate for N-myristoyl transferase and is also palmitoylated in vivo. The wild type protein has been localised to the endocytic system by immunofluorescence. The catalytic and fused C-terminal domains of the kinetoplastid and other eukaryotic PPEFs share high sequence similarity, but unlike their higher eukaryotic relatives, the C-terminal parasite EF-hand domains are degenerate and do not bind calcium. 1 Introduction Protein phosphorylation and dephosphorylation are critical processes in a variety of cellular mechanisms for the detection, transmission, and integration of intra- and extra-cellular signals. In eukaryotes, the extensively studied PPP family of serine/threonine protein phosphatases function in cellular processes as diverse as regulation of the cell-cycle, RNA splicing and T cell activation [1]. PPPs have been divided into three subfamilies, commonly referred to as the ‘Classical’ PPP phosphatases; PP1, PP2A and PP2B [1]. Novel phosphatases that cannot be categorised into these subfamilies include the retinal degeneration C protein (RdgC) from Drosophila melanogaster [2]. RdgC homologues are found in a number of eukaryotic species, including Homo sapiens, Mus musculus, Gallus gallus and Caenorhabditis elegans [3,4] but have not been identified in fungi, yeast or plants to date. These novel phosphatases, distinguished by several putative EF-hand motifs within a fused C-terminal domain, have subsequently been termed PPEFs or Protein Phosphatases with EF-Hands [3]. Interestingly, PPEFs exhibit a much narrower tissue distribution than classical PPPs, being restricted to the central nervous system or primary sensory structures in all metazoans studied to date. Thus, Dm RdgC has been principally localised to photoreceptors and the mushroom bodies of the central brain [2,5] while C. elegans CePPEF is highly enriched in primary sensory neurones [6]. In situ hybridisation and immunostaining have also localised mammalian PPEF isoforms to sensory structures such as the inner ear, dorsal root ganglia, embryonic brainstem nuclei, photoreceptors and pinealocytes [3,4]. These findings suggest that the PPEFs have conserved functions in diverse sensory systems and may have a role in development in mammals. The substrates of the RdgC/PPEF phosphatases, however, remain elusive. The domain architecture of RdgC/PPEF homologues is highly conserved, consisting of three fused domains (Fig. 1A). The N-terminal regulatory domain often contains myristoyl/palmitoyl acylation motifs and a downstream conserved IQ (isoleucine-glutamine) calmodulin binding domain. The catalytic domain contains several RdgC/PPEF specific sequences/motifs together with insertions of unknown function. The C-terminal domain contains at least two EF-hand motifs that have been shown to bind calcium in HsPPEF-1 and CePPEF [6,7] and one or more degenerate EF-hand like motifs. Here, we describe the characterisation of PPEF-like genes in the diverse lower eukaryotes Leishmania and Trypanosoma, the sole members of this phosphatase family in these parasites. The kinetoplastid PPEFs were identified following a genome-wide search for N-myristoylated proteins, carried out to identify downstream targets of the essential gene N-myristoyl transferase (NMT) in Leishmania major, Trypanosoma brucei and Trypanosoma cruzi. We show that LmPPEF and TbPPEF are substrates for NMT in vivo and that LmPPEF (and probably TbPPEF) can also be palmitoylated in vivo. Using immunofluorescence, LmPPEF has been localised to the endocytic system of Leishmania parasites, with some accumulation at the flagellar pocket. This location requires downstream regions of the protein in addition to the unique acylated N-terminus. Unlike other members of the RdgC/PPEF family, the EF-hand domains within the C-terminus of LmPPEF are degenerate and do not bind calcium under the experimental conditions used here. 2 Materials and methods 2.1 PCR amplification and sub cloning The 2862-bp LmPPEF open reading frame (ORF) was amplified from cosmid 1567.3 (gift from Al Ivens) using Pfu DNA polymerase (Promega) at 64 °C annealing temperature and the primers LmPPEFFor (5′-ATGGGGTGTGACTCATCCAAG-3′) and LmPPEFRev (5′-TTAGCGACTAGTGCCGAGGC-3′). The amplified LmPPEF ORF was cloned into pPCR-Script AMP SK(+) (Stratagene) to generate pLmPPEF. 236-bp and 1056-bp fragments from the 3′ end of the LmPPEF ORF (nucleotides 2154–2862 and 1806–2862, respectively) were amplified at 60 °C annealing temperature using primers LmPPEF-Cterm1For (5′-GACGATcatatgCGCATCTGGTAC-3′) and LmPPEF-Cterm1Rev (-5′-TGGCggatccTCTAGCCCTTA-3′) or primers LmPPEF-Cterm2For (5′-ATTAATTTcatatgCAGGTGGTGAGTCTA-3′) and LmPPEF-Cterm2Rev (5′-AATAggatccTTAGCGACTAGTGCC-3′). Cloning sites are shown in lower case. The PCR fragments were digested with NdeI/BamHI and cloned into pET-15b (Novagen) generating pLmPPEF-Cterm1 and pLmPPEF-Cterm2, respectively. The 2775-bp TbPPEF ORF was amplified from T. brucei genomic DNA at 59 °C annealing temperature, using primers TbPPEFFor (5′-CTTACGTTTccatggGTTGCTC-3′) and TbPPEFRev (5′-CCTCCcTcgagatCTCTCACAAA-3′), digested with NcoI/XhoI and cloned into pET-33b, generating pTbPPEF. The recombinant TbPPEF protein expressed from this plasmid has an N-terminal myristoylation motif (MGCSTSK). 2.2 Parasite culture, membrane fractionation and nucleic acid analysis L. major Friedlin parasites (MHOM/IL/80/Friedlin) were cultured, nucleic acids extracted and DNA/RNA blotting and hybridisation carried out as previously described [8]. For membrane fractionation, mid-log phase parasites (5 × 107) were lysed by sonication on ice in either PBS alone, PBS plus 1 mM CaCl2 or PBS plus 1 mM EGTA. Undisrupted cells were removed by two centrifugation steps (500 × g, 4 °C, 10 min). Cell lysates were separated into membrane and cytosolic fractions by ultra centrifugation (100,000 × g, 4 °C, 1 h). Following separation, membrane fractions were washed twice in PBS and proteins from both fractions analysed by SDS-PAGE and immunoblotting as described [8]. 2.3 Antibody production and immunoblotting Expression of N-terminally His-tagged recombinant LmPPEF-Cterm1 was induced by isopropyl-β-d-thiogalactopyranoside (IPTG) in Escherichia coli Rosetta (DE3) pLysS (Novagen). Cells were subsequently lysed in 6 M Gu-HCl prior to affinity chromatography using Talon Ni2+-nitrilotriacetic acid-agarose (Ni-NTA; BD Biosciences). Eluted protein was precipitated using 10% trichloroacetic acid, air dried and used for immunisation and generation of rabbit polyclonal antiserum (Eurogentech). Partial purification of LmPPEF-specific polyclonal antibodies was carried out using ammonium sulphate precipitation as described [9], followed by affinity purification against purified recombinant LmPPEF-Cterm1 as described [10]. Parasites were lysed in SDS-PAGE gel loading buffer, and immunoblotted as above with purified LmPPEF antiserum (abSK2031, 1:200 dilution), anti-NMT (abSK805, 1:2000 [8]), peroxidase anti-peroxidase (PAP) complex (P-2026, 1:2000, Sigma), or anti-GFP (ab290, 1:10,000, Abcam). Immune complexes were detected using an ECL kit (Amersham Biosciences). 2.4 L. major episomal expression constructs and parasite transfection A 111-bp fragment from the 5′ end of the LmPPEF ORF (nucleotides1–111) was amplified from pLmPPEF at 58 °C annealing temperature using primers Lm37WT-GFPFor (5′-TAAAggatccATGGGGTGTGACTC-3′) and Lm37WT-GFPRev (5′-TTATAgatatcGCTACAAGTGCGTCG-3′). The fragment was digested with BamHI/EcoRV and cloned into pX-GFP [11], generating pLm37WT-GFP. Plasmids pLm37G2A-GFP, pLm37C3S-GFP and pLm37G/A,C/S-GFP were generated as above, using forward primers Lm37G2A-GFPFor (5′-TAAAggatccATGGCGTGTGACTC-3′), Lm37C3S-GFPFor (5′-TAAAggatccATGGGGTCTGACTC-3′) and Lm37G/A,C/S-GFPFor (5′-TAAAggatccATGGCGTCTGACTC-3′), respectively, and the reverse primer Lm37WT-GFPRev. The 2862-bp LmPPEF ORF was amplified from pLmPPEF at 60 °C annealing temperature using primers LmPPEF-TAPFor (5′-ATTAATTTcatatgGGGTGTGACTCAT-3′) and LmPPEF-TAPRev (5′-ATAtctagaCTTGCGGCTAGTGCC-3′), digested with NdeI/XbaI and cloned into the TAP vector pGL893 (gift from Sebastion Besteiro), generating pLmPPEF-TAP. The recombinant LmPPEF-TAP protein expressed from this plasmid has an N-myristoylation motif (MGCDSSK). All constructs used in this study are shown in Fig. 1B. Mid-log phase L. major were electroporated with 20–50 μg of either pLm37WT-GFP, pLm37G2A-GFP, pLm37C3S-GFP, pLm37G/A,C/S-GFP or pLmPPEF-TAP as described [11] and cultures subsequently grown in media supplemented with 1 mg/ml G418 (Life Technologies, Inc.). 2.5 Metabolic labelling and immunoprecipitation Mid-log phase L. major promastigotes were metabolically labelled as previously described [11]. Cells were lysed for 1 h at 4 °C in lysis buffer (PBS containing 50 mM Tris, pH 7.5, 150 mM NaCl, 5 mM EDTA, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 100 μg/ml leupeptin, 500 μg/ml pepstatin, 198 μg/ml 1,10 phenanthroline and 25 μg/ml E64). The lysates were pre-cleared by incubation for 1 h at 4 °C with protein A-coupled Sepharose (Amersham Biosciences). Labelled proteins were then recovered from the supernatant by incubation with either anti-LmPPEF or anti-GFP antibodies overnight at 4 °C. After a second protein A-coupled Sepharose incubation, the beads were collected by centrifugation, washed twice in lysis buffer and proteins removed by boiling in SDS-PAGE gel loading buffer, prior to separation by SDS-PAGE. Detection of radiolabelling was improved using EN3HANCE™ Autoradiography Enhancer (Kodak). DTT was omitted from the loading buffer for separation of [9,10-3H] palmitate-labelled proteins. 2.6 Calcium mobility shift assay This assay was carried out as described [12]. In brief, proteins were lysed in SDS-PAGE gel loading buffer and separated by SDS-PAGE using either 5 mM CaCl2 or 5 mM EGTA in both the stacking and resolving gels. Separated proteins were Coomassie-stained or analysed by immunoblotting. 2.7 N-Myristoylation co-expression assay This assay was performed as described [8,13]. In brief, E. coli BL21(DE3) pLysS cells were co-transformed with pTbPPEF and either pNMT [8] or pTbNMT [14]. Expression of recombinant TbPPEF and NMT protein was induced by IPTG in the presence of [3H]-myristate (Amersham Biosciences) and, following SDS-PAGE, radiolabelled proteins were detected by autoradiography. 2.8 Fluorescence microscopy For indirect immunofluorescence, L. major promastigotes were fixed and processed as previously described [15]. Cells were stained prior to fixation by incubation in 10 μM FM4-64 FX (Invitrogen) in serum-free medium at 26 °C for 10 min. For direct immunofluorescence, parasites were fixed in 4% (w/v) paraformaldehyde at room temperature for 15 min, washed in PBS and immobilised on poly-l-Lysine coated coverslips (4 × 106 cells per cover slip). Slides were mounted using Vectashield plus 4′,6′-diamidino-2-phenylindole (DAPI; Vecta Laboratories). 3 Results 3.1 The kinetoplastid N-myristoylomes N-Myristoylation describes the co-translational addition of the C14:0 fatty acid myristate to the amino-terminal glycine residue of a subset of eukaryotic and viral proteins. Catalysed by the enzyme myristoyl-CoA:protein N-myristoyltransferase (NMT), N-myristoylation often plays a role in targeting proteins to membrane locations and can additionally be involved in mediating protein–protein interactions and stabilising protein structure. We have previously shown that NMT is essential for viability in kinetoplastid parasites [8] while myristate analogues (non-specific inhibitors of NMT) are lethal to L. major promastigotes and bloodstream T. brucei [8,16,17]. A range of more specific NMT inhibitory compounds has recently been tested and several shown to inhibit T. brucei NMT activity in vitro, with toxicity to bloodstream T. brucei at low μM concentrations [18]. However, the cellular processes contributing to lethality in NMT-depleted cells are as yet unknown, although identification of the essential downstream targets of this enzyme may be crucial to the successful development of NMT as a potential drug target. Few kinetoplastid N-myristoylated proteins have been studied in depth to date—the Leishmania HASPB and SMP-1 proteins and the T. brucei ARL1 protein are exceptions [11,14,19]. We used several strategies, all based on the well characterised eukaryotic N-myristoylation consensus motif, to search the L. major, T. brucei and T. cruzi genome datasets for putative N-myristoylated proteins. The first publicly available prediction algorithm for these proteins, the PDOC00008 PROSITE myristoylation signature [20], is known to generate high numbers of false positive as well as false negative predictions. The more recent NMT Predictor program was developed using a set of ‘positive’ amino acid sequences for profile training, substantially reducing the number of false results generated and extending the N-myristoylation motif from 6 to 17 amino acids [21,22]. Further refinement has come with the Myristoylator program, based on a neural network model and incorporating an additional ‘negative’ training set of amino acids [23]. We firstly used the ‘relaxed’ version of PROSITE PDOC00008 derived from the NMT Predictor, G-{EDRKHPFYW}-x(2)-[STAGCNDEF]-{P} (described in [21]), to probe the kinetoplastid genome datasets (http://www.genedb.org/), thereby reducing the >8000 predicted ORFs to ∼250–300 potentially N-myristoylated proteins. The Myristoylator prediction algorithm was then applied to this refined data set, generating a final list of predicted N-myristoylated proteins within high confidence boundaries (scores: 0.85–1; Table 1; Tables S1–S3, supplementary data). For the three genomes analysed, 62 N-myristoylated proteins were predicted in L. major, 62 in T. brucei and 123 in T. cruzi, with the latter higher figure a consequence of the presence of two different haplotypes in the T. cruzi CL Brener genome sequence [24]. Predicted N-myristoylated proteins contribute 0.5–0.8% of the proteomes of other eukaryotes studied to date [22,25], and the kinetoplastid data fall well within this range (Table 1). Further analysis of the N-termini of the high confidence kinetoplastid proteins revealed a strong preference (>70% of the total) for serine at position 6 in all three species, with threonine used in a further ∼8% of proteins. In Saccharomyces cerevisiae, serine or threonine in these positions allows stabilisation of the peptide–NMT complex through hydrogen bonding with Asp417, Gly418 and His221 in the ScNMT crystal structure [21]. Of the putative N-myristoylated proteins identified with high confidence, 8 were specific to L. major, including the dual acylated HASPB [11] and the ADP-ribosylation factor LmARF3 [14], with a similar number of targets specific to T. brucei and T. cruzi (Tables S1–S3, supplementary data). Other previously characterised N-myristoylated proteins included the calpain-like proteins, T. brucei CAP 5.5 [26] and L. major SMP-1 [19], while a number of putative signal transduction pathway proteins were also identified, including protein kinases and phosphatases (Table 2). However, the largest group of kinetoplastid N-myristoylated proteins was that currently characterised as of “unknown function”. A majority of the proteins within this group are conserved between the three Tri-Tryp species. 3.2 Kinetoplastid PPEF-like protein phosphatases Analyses of the kinetoplastid genomes have identified a range of kinases and protein phosphatases, some possessing novel motifs and insertions suggesting possible structural and functional differences from their mammalian homologues [24,27]. Of the phosphatases, only a few of the classical types have been characterised to date: protein phosphatase 2C (PP2C) from L. chagasi [28], PP1 from T. cruzi [29], PP1, PP2A and PP5 from T. brucei [30]. The N-myristoylome analysis described above identified a conserved serine/threonine protein PPEF-like phosphatase (hereafter called PPEF) in all three kinetoplastid species: LmjF12.0660 in L. major, Tb927.1.4050 in T. brucei, Tc00.1047053506529.380/Tc00.1047053510889.80 in T. cruzi. These kinetoplastid PPEFs are of similar size and composition: 954 aa, pI 8.0, 109.0 kDa in L. major; 925 aa, pI 7.7, 105.7 kDa in T. brucei; 923 aa, pI 7.7, 104.8 kDa in T. cruzi. The phylogenetic relationship between these three proteins, the RdgC/PP5 family and the classical phosphatase groups was investigated by multiple sequence alignment (using CLUSTAL W [31]) and generation of an unrooted tree (using the Tree View software, version 1.5.2 [32]). This analysis confirmed that the trypanosomatid PPEFs are more closely related to the RdgC/PPEF subfamily than to the classical PPP subfamilies, PP1, PP2A or PP2B, and are also separate from the PP5/PPT and PP7 subfamilies (Fig. 2A). Alignments revealed sequence similarity between the kinetoplastid PPEFs and other members of the RdgC/PPEF subfamily (Fig. S1, supplementary data), with the overall percentage of amino acid identity between representative proteins of this group shown in Fig. 2B. While the kinetoplastid proteins share 60–70% identity with each other, they are only ∼20% conserved when compared to their higher eukaryotic relatives. The similarity between these proteins is more apparent, however, when comparing their domain organisation, both at the primary sequence level (Fig. 1A; Fig. S2, supplementary data) and by homology modelling of the catalytic domain of LmPPEF with that of mammalian PP1 for which a high resolution structure is available [33] (Fig. S1, supplementary data). The central catalytic domains of the kinetoplastid and mammalian PPEFs, together with D. melanogaster RdgC, share significant sequence similarity (42%), including several RdgC/PP5 specific catalytic motifs [1]. Kinetoplastid PPEFs also share RdgC/PPEF specific mutations found in the conserved SAPNYC motif (common to all PPP phosphatases) that is found within the β12/β13 loop of the catalytic domain (Fig. S2, supplementary data). The first of these, a Pro to Ser substitution at position 3, is found in all members of the Rdg/PP5 family including the kinetoplastid proteins and at least two other protozoan phosphatases, PP1 from T. cruzi [29] and PfPPJ, a novel protein phosphatase from Plasmodium falciparum [34]. The second substitution within the SAPNYC motif, replacement of the Cys-6 residue by either Tyr, Asp or Asn, is restricted to the RdgC/PPEF sub-family. Overall, the level of conservation within the catalytic domain of LmPPEF, including those residues acting as metal ligands, suggest that this protein and its kinetoplastid orthologues are functional phosphatases (Fig. S1, supplementary data). At their C-termini, the kinetoplastid PPEFs are more closely related to the C-termini of the RdgC/PPEFs than to any other calcium binding proteins or EF-hand domain-containing enzymes [35]. This observation supports the hypothesis that the catalytic domain of the ancestral form of RdgcC/PPEF fused with an EF-hand Ca2+-binding protein prior to the acquisition of the N-terminal domain [1]. At their N-termini, there is little similarity between the kinetoplastid PPEFs and the RdgC/PPEFs except for the presence of putative N-myristoylation and palmitoylation motifs, which are found in some but not all members of the RdgC/PPEF phosphatase family. 3.3 Genomic organisation and expression of the L. major and T. brucei PPEFs The Artemis annotation tool [36] was used to examine the level of gene synteny in the regions flanking the PPEF genes in the L. major and T. brucei genomes. LmPPEF is found on chromosome 12 in L. major while TbPPEF is found on chromosome 1 in T. brucei (http://www.genedb.org/). The chromosomal regions surrounding these genes have been analysed previously as part of a wider study, revealing high conservation of gene order [37]. Thus the kinetoplastid PPEFs, including TcPPEF, can be classified as true orthologues. To confirm the in silico gene analysis, genomic DNA blotting and hybridisation were used to show that both LmPPEF and TbPPEF are present as single copy genes in the L. major and T. brucei genomes, respectively (Fig. 3A and B). RNA expression from these genes in different parasite life cycle stages was initially demonstrated using RT-PCR (Fig. 3Ci and Di). RNA blotting and hybridisation were then used to confirm constitutive expression of single PPEF transcripts (of 4.3 and 4.8 kb, respectively) in insect and mammalian stages of L. major and T. brucei (Fig. 3Cii and Dii). Expression of the LmPPEF protein was analysed by immunoblotting, using affinity-purified antibodies raised against the C-terminal 236 residues of the recombinant protein (expressed from construct LmPPEF-Cterm1, Fig. 1B). These antibodies recognised a single polypeptide band of 109 kDa on immunoblots of whole L. major parasite lysates, correlating with the size of the deduced open reading frame of gene LmjF12.0660 (Fig. 4A). Whole cell lysates from L. major procyclics, metacyclics and amastigotes were then analysed and a single 109 kDa protein detected in all three life cycle stages (Fig. 4B). However, no signal was detected in T. brucei procyclic or bloodstream form parasite lysates using the same anti-LmPPEF (data not shown). This suggests that this C-terminal antibody does not cross react with TbPPEF (despite the 51% amino acid conservation in this domain) and/or that TbPPEF is expressed at low levels during the parasite life cycle. 3.4 LmPPEF is not a calcium-binding protein Given the degeneracy of EF-hand motifs in the C-terminal domains of the kinetoplastid PPEFs, and the absence of N-terminal IQ calmodulin binding motifs, we investigated whether LmPPEF can bind calcium, either as a recombinant or wild type protein, in a mobility shift assay. In this analysis, target and control proteins were separated by electrophoresis through denaturing gels in the presence of either Ca2+ or EGTA. Fig. 5A shows the positive (calreticulin) and negative (BSA) control proteins after separation: the migration of BSA was similar in both gel types whereas the migration of calreticulin was significantly altered on electrophoresis through Ca2+ as compared to EGTA, with >50% showing a mobility shift. To analyse the behaviour of LmPPEF in this assay, the C-terminal 351 residues, containing all three EF-hand like motifs, were expressed (from construct LmPPEF-Cterm 2) as a 40 kDa N-terminally His-tagged protein in L. major. Comparing the mobility of this protein and wild type LmPPEF in whole cell lysates, in the presence of Ca2+ or EGTA, revealed no significant differences in migration, suggesting that LmPPEF does not bind calcium (Fig. 5B). Cell fractionation of total L. major proteins was also carried out, to analyse the relative distribution of LmPPEF between membrane and cytoplasm and to investigate whether this distribution was altered in the presence or absence of calcium. Wild type parasites were lysed in either PBS alone, PBS plus 1 mM CaCl2 or PBS plus 1 mM EGTA and separated by electrophoresis. Cell fractionation carried out in PBS alone revealed that LmPPEF is predominantly membrane-associated, with ∼20% of the protein detected in the soluble or cytoplasmic fraction (Fig. 5B). The presence or absence of calcium did not alter this membrane versus cytoplasmic distribution, again suggesting that the degenerate EF-hands in LmPPEF are unlikely to bind calcium. A similar conclusion emerged from 45Ca overlay experiments with both wild type and recombinant protein (data not shown). 3.5 The N-termini of recombinant and wild type kinetoplastid PPEFs are substrates for acylation in vivo Given the presence of N-myristoylation motifs together with cysteine residues predicted to be palmitoylation sites at the N-termini of kinetoplastid PPEFs (Tables S1–S3, supplementary data), we firstly confirmed that these proteins could act as templates for NMT in an E. coli co-expression assay [8]. In these experiments, plasmids expressing NMT and TbPPEF as substrate were co-expressed in the presence of [3H]-myristoyl CoA and incorporation of radioactivity into myristoylated product detected by autoradiography following SDS-PAGE (Fig. 6A). Using NMTs from both L. major and T. brucei with TbPPEF, these results demonstrate radiolabelling of a ∼106 kDa product in the presence of each enzyme, indicative of the transfer of myristate to TbPPEF (as previously demonstrated for HASPA and TbARL1) [8,14]. While the amount of radiolabelled product appears greater with LmNMT rather than TbNMT in this experiment, this is likely due to an artefact of loading (as indicated by the higher amount of NMT-myristoyl CoA binary complex loaded on to the gel) rather than substrate specificity. To demonstrate acylation in parasites in vivo, metabolic labelling experiments were carried out with L. major transgenic lines expressing the first 37 N-terminal residues of PPEF fused with GFP (Lm37WT-GFP, Fig. 1Bii) or a mutated version, in which the Gly in position 2, essential for N-myristoylation, is substituted with Ala (Lm37G2A-GFP, Fig. 1Bii). These parasites express equivalent amounts of the two fusion proteins, as indicated by immunoblotting with anti-GFP and anti-NMT (Fig. 6B). Following radiolabelling with [3H]-myristate or [3H]-palmitate, the fusion proteins were immunoprecipitated using anti-GFP and subjected to SDS-PAGE and autoradiography. This analysis detected a [3H]-myristate-labelled protein of 35 kDa, corresponding to the predicted molecular mass of the Lm37WT-GFP protein, which was absent from parasites expressing Lm37G2A-GFP lacking the residue required for N-myristoylation (Fig. 6C). Similarly, autoradiography of parasites labelled with [3H]-palmitate revealed a single radiolabelled protein of the same molecular mass, corresponding to [3H]-palmitate-labelled Lm37WT-GFP. This protein was again absent in parasites expressing Lm37G2A-GFP. These data correlate with similar experiments with the dual acylated HASPB protein, in which loss of the residue required for N-myristoylation also prevented palmitoylation, due to mislocalisation of the protein within the cell [11]. The experiments in Fig. 6A and C confirm that TbPPEF and LmPPEF can be acylated in vivo, both in a heterologous cell system and in live parasites. To confirm that endogenous LmPPEF is N-myristoylated in vivo, anti-LmPPEF was used in immunoprecipitations from parasite lysates following metabolic labelling with [3H]-myristate. As shown by autoradiography in Fig. 6D, a radiolabelled protein of ∼109 kDa, corresponding to the predicted molecular mass of LmPPEF, was detected in the lysate from wild type L. major parasites, while no proteins were detected following precipitation with protein A beads alone. Radiolabelled parasites expressing LmPPEF-TAP (Fig. 1Bi) were also subjected to immunoprecipitation with anti-LmPPEF, generating two detectable [3H]-myristate-labelled proteins in the lysate: the smaller corresponding to the 109 kDa native LmPPEF and the larger weaker ∼129 kDa band corresponding to the predicted size of the LmPPEF-TAP fusion protein. 3.6 LmPPEF is localised to the endomembrane system of L. major Immunofluorescence microscopy with the affinity-purified LmPPEF antibody was used to investigate the location of LmPPEF in procyclic L. major (Fig. 7A). Overlay of the fluorescent signal with DAPI (for identification of the nucleus and kinetoplast) indicated a punctate distribution of signal within the cytosol, reminiscent of the endocytic system, together with some concentration in the region of the flagellar pocket. These observations are consistent with the predominantly membrane association of LmPPEF revealed by cellular fractionation (Fig. 5). Counter-staining with anti-LCB2, which recognises an ER-resident sub-unit of serine palmitoyltransferase [38] showed no overlap with PPEF (Fig. 7B) while labelling with the endocytic lipophilic tracer, FM4-64, demonstrated some co-localisation with the flagellar pocket and endosomes (Fig. 7C). PPEF staining was excluded from the nucleus, plasma membrane and flagellum. We attempted to express and analyse GFP-fusions with either full-length wild type LmPPEF or the N-terminal 473 residues alone (lacking the C-terminal EF-hands). However, these constructs did not express after transfection into L. major, probably due to the large sizes of the recombinant proteins generated. As an alternative approach, given that the N-terminus of LmPPEF is a target for acylation (Fig. 6), the N-terminal GFP transgenic parasite lines (see Fig. 1B) were used to investigate the role of these modifications in intracellular localisation of the fusion proteins. Counter-staining with FM4-64 was also utilized to detect the flagellar pocket in these analyses. Transfected parasites expressing the first 37 residues of LmPPEF fused to GFP (Lm37WT-GFP) targeted the fusion protein predominantly to the flagellum and the flagellar pocket, with weak staining only in the cell body (Fig. 7D). Loss of the N-myristoylation site by mutation of Gly to Ala (Lm37G2A-GFP) resulted in homogeneous distribution of cytosolic fluorescence throughout the parasite (Fig. 7E), in a pattern almost identical to the localisation of GFP alone (data not shown). Separate mutation of the putative palmitoylation site alone, by substituting Cys-3 with Ser (Lm37C3S-GFP), resulted in concentrated signal in a region adjacent to the flagellar pocket most likely to be the Golgi (Fig. 7F). Expression of the double mutant Lm37G2A, C3S-GFP, lacking both acylation sites, caused the fusion protein to remain localised within the cytosol (Fig. 7G). These observations correlate with those previously observed using similar N-terminal mutations of the L. major HASPB protein [11], suggesting that the wild type fusion protein (Lm37WT-GFP) requires N-myristoylation in the cytosol to reach the Golgi region, where palmitoylation further modifies the protein for trafficking to other locations, predominantly the flagellum. Clearly, the localisations of these GFP fusion proteins are distinct from the endocytic location of wild type LmPPEF, indicating that other regions and/or signals within this large protein are important for subcellular localisation. More generally, these data suggest that N-terminal dual acylation might act as a primary signal in targeting proteins to the flagellum in Leishmania, as exemplified by HASPB, SMP-1 and PPEF [11,19 and this study]. 4 Discussion In this paper, as part of a wider study of the downstream targets of NMT, we describe the PPEF-like protein phosphatases that are encoded by single copy genes in three kinetoplastid species. These phosphatases were identified following in silico genome analysis that predicted a subset of kinetoplastid proteins as “high confidence” substrates for NMT. Within this subset were several proteins already confirmed as N-myristoylated in L. major or T. brucei, together with a number of species-specific molecules and a larger group of proteins found in all three species, many of no known function. We predict that all of these “high confidence” putative substrates are acylated in vivo but do not exclude other “medium confidence” proteins as candidates for NMT modification. LmPPEF and TbPPEF, studied in detail here, are constitutively expressed, membrane-associated acylated proteins. LmPPEF is found in the endocytic system and the flagellar pocket. Conservation of key features and residues within the catalytic domains, which have been modelled against the mammalian PP1 catalytic region, suggest that these proteins are active enzymes (although this has not yet been demonstrated biochemically). Unlike these conserved central regions, however, divergence in the N- and C-terminal domains has led to loss of the IQ calmodulin binding motif and degeneration of the EF-hands, features that characterise the higher eukaryotic PPEFs [39,40]. Thus, there is no evidence that the kinetoplastid proteins are regulated by cytoplasmic Ca2+ levels and/or calmodulin, in contrast to the demonstrated interactions of both human PPEF and Drosophila RdgC with these regulators [39,40]. The analysis in Fig. 5 would support this conclusion. Although PPEF-like proteins have been characterised in a number of species, the functions of these unusual phosphatases are not well understood. The Drosophila RdgC protein is implicated in dephosphorylation of rhodopsin, a G protein-coupled receptor (GPCR) that initiates vertebrate and invertebrate phototransduction. However, mutant mice with targeted disruptions in each of their two PPEF genes show no retinal degeneration and normal rhodopsin dephosphorylation kinetics, suggesting interspecies functional differences despite high similarities in protein sequence [41]. Completion of the Tri-Tryp genome projects has confirmed earlier predictions that these parasites have ‘stream-lined’ signal transduction mechanisms as compared to their higher eukaryotic multicellular counterparts [42]. The lack of several classes of signalling molecules, including serpentine receptors, heterotrimeric G proteins and most classes of catalytic receptors, contrasts with the presence of a large and diverse family of kinase and phosphatases, suggesting complex cellular interactions [43]. While the conservation and expression of PPEF genes in the Kinetoplastida strongly suggest a functional role for these unusual phosphatases, it is clear that this is unlikely to be similar to that demonstrated for the RdgC protein. We have perturbed TbPPEF expression by inducible RNAi in both bloodstream and procyclic stages of T. brucei, which results in a partial growth defect under normal culture conditions (data not shown). However, it cannot be discounted in these experiments that a more extreme phenotype is masked by sufficient residual expression to support enzyme activity. Functional analysis of conditional gene deletion mutants will be required, together with substrate identification, to delineate roles for the kinetoplastid PPEFs in parasite viability.	[ "bioinformatics", "n-myristoylation", "protein phosphatases", "palmitoylation", "ppef, protein phosphatase with ef-hands", "nmt, n-myristoyl transferase", "bsf, bloodstream form", "pcf, procyclic form" ]	[ "P", "P", "P", "P", "R", "R", "M", "M" ]
Hum_Reprod-1-1-2387221	A distinct cohort of the TGFβ superfamily members expressed in human endometrium regulate decidualization	BACKGROUND Successful blastocyst implantation requires the differentiation of human endometrial stromal cells (HESC), a process known as decidualization. Activin A, a transforming growth factor β (TGFβ) superfamily member, enhances HESC decidualization and localizes to decidual cells in human endometrium. Other TGFβ superfamily members, including BMP2, BMP4, BMP7, GDF5, GDF8, GDF11, TGFβs and Nodal, may also play a role during decidualization. This study aimed to identify these TGFβ family members in human endometrium, and to determine whether they are involved in human decidualization. IntroductionThroughout each menstrual cycle the human endometrium undergoes morphological and physiological changes, in preparation for pregnancy should the cycle include conception (Loke et al., 1995). An absolute requirement for successful implantation is the differentiation and proliferation of endometrial stromal cells into enlarged, phenotypically different decidual cells, a process termed endometrial decidualization (Tang et al., 1994; Salamonsen et al., 2003). This process is initiated near the spiral arterioles (Bell, 1991), and occurs under the influence of progesterone (Psychoyos, 1973), although cAMP appears to be essential for priming the endometrial stromal cells to the actions of progesterone; hence both pathways are absolutely required for decidualization (Gellersen and Brosens, 2003). It is now clear that the decidualization response is mediated by a complex array of bioactive molecules, including IL-11, prostaglandin E2, relaxin and activins (Frank et al., 1994; Robb et al., 1998; Jones et al., 2002; Dimitriadis et al., 2005).The transforming growth factor β (TGFβ) superfamily is a large family of proteins that encompasses the sub-families TGFβs, activins, bone morphogenetic proteins (BMPs) and growth differentiation factors (GDFs). Overall, this superfamily of proteins exhibits functional diversity, with biological roles in cell differentiation, proliferation, apoptosis and tissue remodelling, all consistent with various reproductive processes (Jones et al., 2006).Activins consist of two β subunits, βA and βB, that homo/heterodimerize to form activin A, activin B and activin AB, respectively. The activin β subunits are produced in the glandular epithelium during the proliferative and secretory phases, but in stromal cells only following decidualization during the mid-late secretory phase (Otani et al., 1998; Jones et al., 2000; Mylonas et al., 2004). High levels of dimeric activin A are secreted by cAMP-treated endometrial stromal cells. Addition of activin A to decidualizing cells in vitro significantly increased the secretion of prolactin (PRL) and IGFBP-1 (decidual cell markers), suggesting activin A drives decidualization (Jones et al., 2002; Tierney and Giudice, 2004). This decidualization response can be neutralized by follistatin, the naturally occurring activin antagonist (Jones et al., 2002; Tierney and Giudice, 2004). In addition to its tight regulation of activin (Shimonaka et al., 1991; Schneyer et al., 1994), follistatin can also bind and regulate other TGFβ members including the BMPs and GDFs (Table I).Table I.Inhibitors of TGFβ superfamily: ligand–receptor blocking.LigandFollistatinActivin-M108ASB431542Binds ligands and blocks association with type II receptorsActivin type II receptor (ActRIIA/B) antagonistInhibitor of TGFβ and activin type I receptorsActivin A++++ (Schneyer et al., 2003)++ (Harrison et al., 2004)+++ (Inman et al., 2002)Activin B++ (Schneyer et al., 2003)+++a+++ (Inman et al., 2002)GDF 8++ (Amthor et al., 2004)+++a+++ (Inman et al., 2002)GDF11++ (Gamer et al., 1999)+++a+++ (Inman et al., 2002)BMP2+ (Iemura et al., 1998)−−BMP4+ (Iemura et al., 1998; Glister et al., 2004)−−BMP7+ (Iemura et al., 1998; Glister et al., 2004)(−/+)−GDF5(+)−−Nodal−(+++)+++ (Inman et al., 2002)TGFβ1−−+++ (Inman et al., 2002)TGFβ2−−+++ (Inman et al., 2002)TGFβ3−−+++ (Inman et al., 2002)aC.A. Harrison (unpublished observations); (), anticipated activity.BMPs have been identified in the rodent uterus (Cunningham et al., 1995; Ozkaynak et al., 1997; Fitzpatrick et al., 1998; Zhao et al., 1999; Ying and Zhao, 2000; Erickson et al., 2004; Lee et al., 2007; Li et al., 2007). In pregnant mice the temporospatial localization of BMPs 2, 4, 6 and 7 suggested roles during pregnancy (Ying and Zhao, 2000). BMP2 is present within the decidual area at implantation sites (Ying and Zhao, 2000; Li et al., 2007) and plays a role in decidualization in vitro (Li et al., 2007). Importantly, conditional ablation of uterine BMP2 in mice resulted in ineffective decidualization and disrupted pregnancy outcome (Lee et al., 2007). In a human model of endometrial stromal cell (HESC) decidualization in vitro, BMP2 mRNA increased during decidualization and addition of BMP2 to cultures accelerated PRL mRNA expression (a measure of the extent of decidualization). BMP2 protein has not been examined in vivo in human endometrial tissue. The only GDFs identified in the uterus are GDF9 and GDF10, detected by northern analysis in both human and mouse (Cunningham et al., 1995; Fitzpatrick et al., 1998; Zhao et al., 1999): no studies have shown GDF proteins.TGFβ isoforms (1–3), which cannot bind follistatin, have also been localized in human endometrial stroma (Jones et al., 2006), although it is not clear whether any isoform is specific to decidualized cells or increases as the cells decidualize. The TGFβs transmit signal through the TGFβ type I, II and III receptors, all of which are present on endometrial stromal cells (Chegini et al., 1994; Dumont et al., 1995). Microarray studies have demonstrated increases in both TGFβ1 and TGFβ2 mRNA with HESC decidualization in vitro (Popovici et al., 2000; Tierney et al., 2003), while in women treated with medroxyprogesterone acetate (MPA) for 10 days, stromal mRNA and protein increased for TGFβ3 but not TGFβ1 (Reis et al., 2002). Whether the TGFβs play a functional role during human decidualization remains unknown.Given that the TGFβ superfamily of proteins are such common and crucial differentiative and proliferative factors, it is surprising that very few members of the family, including the GDFs and BMPs, have been examined in the human endometrium, particularly in the mid-secretory phase when differentiation of stromal cells is initiated. The high level of redundancy between TGFβ family members and the promiscuity of receptor–ligand interactions would suggest that it is likely that a number of other family members are involved in decidualization and in the other processes occurring in this highly dynamic tissue.In this study, the presence of BMPs (BMP2, BMP4, BMP7), GDFs (GDF5, GDF8/Myostatin, GDF11), Nodal and TGFβs were examined in vivo by immunohistochemistry in secretory phase endometrial tissue, and in vitro for mRNA expression in both non-decidualized and decidualized HESC. Inhibitors with different specificities were administered to HESC to elucidate whether activin is the major family member driving decidualization or whether other family groupings (BMPs, GDFs, TGFβs) might equally contribute to the process. As these are secreted factors, individual ligands were measured in conditioned medium from non-decidualized and decidualized HESC to assess whether they increased with decidualization. The effect of the secreted ligands on decidualization was also examined. It is important to identify which TGFβ members are important during decidualization since this process is critical for implantation and the establishment of pregnancy. Materials and MethodsTissue collectionsEndometrial biopsies (n = 18) were collected by dilatation and curettage from fertile women who were scheduled for tubal ligation or were undergoing testing for tubal patency. Tissues were assessed by a pathologist and had no obvious endometrial pathology. The women had no steroid treatment or other medication for at least 2–3 months before the collection of tissue. Written and informed consent was obtained from all women participating in the study, and the protocols were approved by Monash Medical Centre Human Ethics Committee.ImmunohistochemistryImmunohistochemical analysis was performed using a total of nine endometrial tissue biopsies from fertile women, confirmed by Noyes criteria (Noyes et al., 1975) as mid-late secretory phase tissue (POD 6–12). For each of the antibodies, activin βA and βB, BMP2, BMP4, BMP7, GDF5, GDF8, GDF11, Nodal and TGFβ1, n = 4–5 different tissue biopsies were used. For immunolocalization in first trimester placental tissue (kindly provided by Professor Euan Wallace, Obstetrics and Gynaecology, Monash University, Melbourne, Australia), n = 2 different biopsies were used per antibody. Briefly, 5 µm sections of formalin-fixed, paraffin-embedded tissues were dewaxed and rehydrated. For each antibody an antigen retrieval step was required, which involved microwave exposure or trypsin digestion (0.01% in CaCl2for 10 min at 37°C). Endogenous hydrogen peroxidase activity was quenched using 3% H2O2 in dH2O for 10 min at room temperature. Non-specific binding was prevented by pre-incubation of tissue sections with a non-immune block [5% fetal calf serum (FCS), 2% normal human serum in 0.1% Tween/Tris-buffered saline (TBS) with addition of 10% normal horse serum for BMP2, BMP4, BMP7, GDF5, GDF8; 10% normal swine serum for Nodal; normal goat serum for activin βA and βB, GDF11, TGFβ1]. Primary antibodies were against activin βA and activin βB (400 and 600 µg/ml, respectively; both gifts provided by W. Vale, Salk Institute, La Jolla, CA, USA); BMP2, BMP4, BMP7, GDF5, GDF8, Nodal (200 µg/ml; Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA); GDF11 (1 mg/ml; Orbigen, San Diego, CA, USA); TGFβ1 (μg/ml; Abcam, Cambridge, UK). Specificity for these antibodies has been previously published (Nadiri et al., 2004; Kiyono and Shibuya, 2003; Mendler et al., 2000; McPherron et al., 1999; Garba and Relinger, 2001). For GDF5 and Nodal, the specificity has been shown by western blot analysis (unpublished observation). Antibodies were diluted to 2–4 µg/ml in the non-immune block and applied overnight (18–20 h) at 4°C. A non-immune goat/rabbit/mouse IgG (R&D Systems Incorporated, Minneapolis, MN, USA) was used at a matching concentration to the primary antibody and was included for each tissue. After stringent washing with high salt TBS and 0.6% Tween 20 (BioRad Laboratories, Hercules, CA, USA), detection of positive binding was performed by the sequential application of the appropriate secondary antibody diluted in non-immune block and avidin–biotin–peroxidase conjugate (Dako, Glostrup, Denmark). This was followed by the substrate diaminobenzidine (Dako) for between 3 and 5 min. Sections were counterstained with Harris' hematoxylin (Sigma Chemical Company, St Louis, MO, USA), dehydrated, and mounted from Histosol with DPX mounting medium (BDH Laboratory Supplies, Poole, UK). Immunostaining was analysed by two independent observers for staining intensity and heterogeneity in each of the endometrial compartments (glandular and luminal epithelium; stroma, including decidualized stromal cells; vasculature).Stromal cell isolation and cultureTissue biopsies ranging from Days 10–21 were used for in vitro decidualization studies. HESC were isolated from tissue by enzymatic digestion and filtration as described previously (Dimitriadis et al., 2002). Briefly, tissue was finely minced with scissors and digested by bacterial collagenase type III (Worthington Biochemical Corporation, Freehold, NJ, USA) at a concentration of 45 IU/ml, in the presence of 3.5 µg/ml deoxyribonuclease (DNase; Boehringer Mannheim Biochemica, Mannheim, Germany). After 30–45 min of agitation at 37°C, the digested tissue was filtered sequentially through 45 and 10 µm nylon filters to remove glands. Cell suspensions were centrifuged at 1500 rpm for 5 min. Cells were plated at a density of 5 × 106 in a 75 cm2 flask and grown to confluency in DMEM and Ham's F12 medium (1:1) (DMEM/F12) (Trace Biosciences, Sydney, Australia), supplemented with 1% PSF (penicillin, streptomycin and fungizone) (Commonwealth Serum Laboratories, Melbourne, Australia) and 10% charcoal stripped FCS (csFCS) (Thermo Scientific, Maple Plain, USA). Approximately 30 min after plating, medium containing non-attached epithelial cells was removed and replaced with fresh medium. This results in a >97% pure stromal cell culture (Dimitriadis et al., 2002).In vitro decidualizationConfluent HESC were rinsed with phosphate-buffered saline, trypsinized and replated into 24-well plates using DMEM/F12 and 10% csFCS. Once confluent, HESC were washed with DMEM/F12 and the medium replaced with a serum-free medium containing DMEM/F12 and a serum-free mix (SFM) including transferrin (10 µg/ml; Sigma), sodium selenite (25 ng/ml; Sigma), linoleic acid (10 nmol/L; Sigma), bovine serum albumin (0.1%; Sigma) and insulin (5 µg/ml; Actrapid, Novo-Nordisk Pharmaceuticals Pty Ltd, Sydney, Australia) for 48 h prior to treatment addition. HESC were decidualized by two distinct methods as previously described (Dimitriadis et al., 2005). For the inhibitor, mRNA expression and protein secretion studies, 0.5 mM cAMP (Sigma) was added to medium for 4–6 days. Alternatively, when exogenous proteins were added, cells were decidualized with E2 (10−8 M; Sigma) and MPA (10−7M; Sigma) for 8 days. All media were replaced every 48–72 h.RNA extraction and cDNA synthesisTotal RNA was extracted from non-decidualized and decidualized HESC from two separate biopsies at Day 4–6 using the RNeasy Minikit (Qiagen Sciences, Germantown, MD, USA), according to the manufacturer's instructions. RNA from positive control tissues was extracted by homogenization in Trizol reagent (Qiagen Sciences, Clifton Hill, Victoria, Australia), according to the manufacturer's instructions, with the exception of an additional chloroform extraction step to minimize carryover of phenol into the precipitate. The control tissues used were: term placenta (activin βA, βB, GDF11, TGFβ1, -2, -3); pregnant mouse endometrium (BMP2); mature rat ovary (BMP4, GDF5); cycling mouse endometrium (BMP7); mouse heart (GDF8); JEG3 cells (Nodal). All samples were treated with RNase-free DNase (Ambion, Austin, TX, USA) to remove the possibility of genomic DNA contamination. RNA samples were then analysed by spectrophotometry to determine RNA concentration, yield and purity. Total RNA (1 µg) was reverse transcribed at 46°C for 1.5 h in 20 µl reaction mixture using 100 ng random hexanucleotide primers and 6 IU AMV reverse transcriptase (Roche, Castle Hill, Australia) in the presence of cDNA synthesis buffer (Roche), 1 mmol/l dNTPs (Roche), 10 mmol/l dithiothreitol (Roche), 10 IU ribonuclease inhibitor (RNasin; Promega, Annandale, Australia). The resultant cDNA mixtures were heated at 95°C for 3 min before storage at −20°C. Negative controls were performed by omission of reverse transcriptase. Triplicate RNA samples were reverse transcribed in triplicate for each condition (non-decidualized and decidualized) with the successful conversion to cDNA monitored by 18S expression (data not shown).RT–PCRMessenger RNA expression for all named TGFβ superfamily ligand members was determined using a conventional PCR block cycler (Hybaid, Middlesex, UK). All ligands (excluding Nodal), used a 1 µl aliquot of RT product, to be amplified in a total volume of 40 µl using 4 µl of RT single strength PCR buffer (10 mmol/l Tris–HCl, 1.5 mmol/l MgCl2, 50 mmol/l KCl, pH 8.3; Roche), 2.5 mmol/l dNTPs (Gibco, Melbourne, Australia), 0.5 pmol/µl sense and antisense primers (Sigma Genosys Australia Pty Ltd, Castle Hill, Australia) and 2.5 IU Taq DNA polymerase (Roche). For Nodal, 1 µl of RT was amplified in a total of 50 µl using the KOD-Taq PCR kit (Bioron, Germany), which included 10× PCR KOD Hot Start buffer, 2 mM dNTPs, 0.5 pmol/μl primers, 2 mM MgSO4 and 2.5 IU Taq DNA polymerase (Roche). For all ligands, the PCR was performed in three stages as follows: the first stage involved 94°C for 5 min, x C for 1 min, where x is the annealing temperature for the individual primer pairs (see Supplementary data) and 72°C for 3 min; the second stage involved 35–40 cycles of 94°C for 1 min, x C for 1 min, and 72°C for 1 min; and the final stage was 72°C for 7 min. PCR products including positive controls were analysed by electrophoresis on a 2% agarose gel (Roche) and stained with ethidium bromide. Bands of interest were excised from the gel, purified (DNA purification kit, Qiagen) and directly sequenced to confirm their identity.Inhibitor experimentsFor examination of the effect of inhibitors on decidualization, cells (from n = 10 biopsies) were plated at a density of 2.5 × 105/well in 24-well plates until confluent, and then equilibrated in serum-free medium containing SFM for 48 h. This was designated Day 0 and first day of treatments. HESC were exposed to varying doses of either the inhibitor Activin-M108A (M108A) (0.39 1.56, 6.25 and 25 nM) (Harrison et al., 2004, 2006) or the inhibitor SB431542 (1.25, 2.5, 5 and 10 µM) (TOCRIS Bioscience, Northpoint, UK). Following 1 h incubation with either inhibitor, 0.5 mM cAMP (Sigma) was added as decidualizing stimulus. Each treatment was performed in triplicate wells of a 24-well plate and the experiments ran for 4–6 days with media collection and replenishment (including inhibitors) every 48–72 h. In each case, the final medium collection represented the final 48 h. Each experiment included a medium-only and a cAMP-only control. Cells were photographed for morphological analysis and cells from triplicate wells were pooled for trypan blue cell exclusion to check cell viability at the end of each experiment. Five separate cultures were conducted for each inhibitor.TGFβ superfamily assaysCulture medium from triplicate wells from three tissue biopsies (n = 3 separate decidualization experiments) were collected, pooled and concentrated 5-fold to measure BMP2 (ELISA; R&D Systems), BMP4, BMP7, TGFβ1 and TGFβ2 (ELISA; Ray BioTech., Norcross, GA, USA) according to the manufacturer's instructions. Mean sensitivities of the assays were: BMP4, 15 pg/ml, BMP7, 10 pg/ml, TGFβ1, 80 pg/ml, TGFβ2, 15 pg/ml, with intra-assay CV <10% and inter-assay CV <12% for each assay. Mean sensitivity for BMP2 was 11 pg/ml and intra- and inter-assay variabilities were 2.6 and 6.3%, respectively. TGFβ1 and TGFβ2 ELISAs measured only the activated protein form and required activation steps prior to assay. Dimeric activin A secretion from stromal cells was measured by Activin A Immunofluorometric Assay (IFMA) (Harrison et al., 2006). The working range of the assay is 0.03–3 ng/well, with a sensitivity of 0.03 ng/well. All assays were read at 450 nm.Addition of exogenous BMP2 and TGFβ1 during decidualizationHESC from three separate cultures were decidualized using E2 and MPA as described above, in the presence of varying doses of either recombinant human (rh) BMP2 (5, 50, 500 ng/ml; R&D Systems) or TGFβ1 0.5, 5, 50 ng/ml; PeproTech Inc., NJ, USA) added every 48 h with media and decidualizing stimulus replenishment. At Day 8, cells were morphologically assessed and conditioned media were collected for PRL measurement.PRL and protein assaysPRL production by HESC was assayed in duplicate by ELISA (Bioclone Australia Pty Ltd, Sydney, Australia) to determine the extent of decidualization (Dimitriadis et al., 2002). Media collected at the end of the experiment were concentrated 5-fold for PRL measurement. A quality control sample (culture medium from a single endometrial cell culture) was included in every assay. The lower detection limit of the assay was 50 mIU/l. The inter- and intra-assay variablilities were 5.3 and 3.0%, respectively. PRL concentrations (mIU/l) were corrected for the amount of protein (μg/μl) determined using the Bradford reagent.Statistical analysisData were expressed as mean ± SEM. Statistical significance for inhibitor studies was determined following confirmation of normal distribution, by one-way ANOVA followed by Dunnett's multiple comparisons test. Data from ELISA and recombinant protein studies were analysed by Student's t-test. A P-value of <0.05 was considered statistically significant. ResultsLocalization of TGFβ superfamily ligands in human endometrium and first trimester deciduaAll TGFβ superfamily members studied, except Nodal, were detected in mid-late secretory endometrium and first trimester decidua. As previously described (Jones et al., 2000), activin A and activin B were localized to the morphologically distinct decidual cells (Fig. 1A and B), glandular and luminal epithelial cells and some leukocytes (data not shown). Some glands had punctate staining for BMP2 (Fig. 1C), but BMP2 was most strongly stained in decidual cells in mid-late secretory endometrium (Fig. 1D). In first trimester decidua, BMP2 protein was detected in both the decidual cells (Fig. 1E, inset) and glandular epithelium, while the vessels were devoid of stain (Fig. 1E). GDF5 protein was observed in the decidualized stromal cells (Fig. 1F and G) and was very low in the glands (Fig. 1F) and luminal epithelium (data not shown). Similarly, GDF5 was very strongly expressed by the decidua of first trimester placenta, as well as the glandular epithelium (Fig. 1H, inset). TGFβ1 was present in decidualized stromal cells (Fig. 1I and J), with minimal staining evident in the glandular epithelium (Fig. 1I). As previously described (Graham et al., 1992), TGFβ1 was detected in decidual cells and extravillous trophoblasts of first trimester placenta (Fig. 1K). BMP4 was not detected in the glandular or luminal epithelium, but was present in the cytoplasm of both non-decidualized and decidualized stromal cells (Fig. 1L and M). However, in first trimester placenta, BMP4 was reduced in the decidual cells, but strongly produced by the glandular epithelium (Fig. 1N, inset). BMP7 was detected in the stroma and not in the glands in mid-late secretory endometrium (Fig. 1O). Stromal staining for BMP7 did not appear to be decidual cell specific, however, in highly decidualized tissue stronger ‘vesicle staining’ was evident (Fig. 1P). Similar but faint punctate staining for BMP7 was detected around spiral arterioles of first trimester placenta, but not specifically in decidual cells (Fig. 1Q, inset). BMP7 was also strongly localized to glandular and luminal epithelium (Fig. 1Q). GDF8 immunostaining in mid-late secretory endometrium showed minimal glandular and luminal epithelial staining, and intense stromal cell cytoplasmic staining, in both non-decidualized and decidualized cells (Fig. 1R). GDF11 immunolocalized to glandular epithelium and endothelial cells, but was very low in the stroma (Fig. 1S). GDF8 and GDF11 in first trimester placenta showed similar staining patterns to BMP4 (data not shown). No immunoreactive Nodal was detected in late secretory endometrium (Fig. 1T), although the immunostaining protocol was verified by positive staining in term placenta (not shown).Figure 1:Localization of TGFβ superfamily members during decidualization in mid-late secretory endometrium and first trimester decidua.Photomicrographs are representatives of immunostaining for activin A (A), activin B (B), BMP2 (C–E), GDF5 (F–H), TGFβ1 (I–K), BMP4 (L–N), BMP7 (O–Q), GDF8 (R), GDF11 (S), Nodal (T). In mid-late secretory endometrium; scale bar = 100 µm (C, F, I, L, O, R, S); scale bar = 10 µm (D, G, J, M, P, T). In first trimester decidua; scale bar = 200 µm (E, H, K, N, Q); inserts show higher power images of same tissue (E, H, K, N, Q). Negative controls for each antibody are shown in inserts (C, F, I, L, O, R, S, T). Arrows highlight decidualized (D), non-decidualized (ND) stromal cells and glands (G).TGFβ superfamily mRNA expression in non-decidualized and decidualized HESCBy RT–PCR, mRNA expression was evident for activin βA, activin βB, BMP2, BMP4, BMP7, GDF5, GDF8, GDF11, TGFβ1, TGFβ2 and TGFβ3 in both non-decidualized and decidualized HESC (Fig. 2). Nodal mRNA was undetectable in both. As conventional RT–PCR was used, only the presence/absence of each gene was assessed. Positive control PCR products (Fig. 2; see supplementary data) confirmed sequence identity and no signals were detected when reverse transcriptase was omitted (Fig. 2).Figure 2:mRNA expression for TGFβ superfamily members by decidualized and non-decidualized HESC.HESC were either non-decidualized or decidualized with cAMP for 7 days. Total RNA was extracted from cells for RT–PCR. Representative products demonstrating Activin βA, Activin βB, BMP2, BMP4, BMP7, GDF5, GDF8, GDF11, TGFβ1, TGFβ2, TGFβ3 and Nodal mRNA are shown in non-decidualized (ND) and cAMP-decidualized (D) HESC; +ve=positive control; −RT=negative control. Data shown is from a single experiment, which is representative of 2 independent culture experiments.Activin-M108A and SB431542 significantly decreases PRL secretion during HESC decidualizationTo establish whether HESC decidualization could be blocked or reduced, the newly developed activin type II receptor antagonist, Activin-M108A or the type I receptor kinase inhibitor, SB431542 were used in an in vitro decidualization model with PRL secretion, a marker of decidualization, as the end-point. These inhibitors affect activin and various other TGFβ family members (see Table I). Cells maintained in medium alone (non-decidualized) showed non-detectable PRL levels over the final 48 h of culture (Fig. 3). In contrast, cells treated with cAMP secreted PRL (mean 372.6 ± 145.5 mIU/l), and morphologically changed from elongated spindle-shaped cells to typical enlarged polygonal cells demonstrating successful decidualization, as previously described (Dimitriadis et al., 2002). Stromal cells cultured with cAMP and Activin-M108A showed a significant dose-dependent decrease in PRL secretion at 0.39, 1.56 and 6.25 nM with no further decrease at 25 nM (Fig. 3A). At the 25 nM dose, a 60% reduction in PRL secretion was seen compared with cAMP-treated cells without inhibitor, which was similar to decidualization reduction seen with follistatin. Similarly cells co-cultured with cAMP and SB431542 showed a dose-dependent decrease in PRL secretion. Addition of SB431542 at a concentration of 1.25 µM significantly reduced PRL secretion compared with cAMP-only treated cells (Fig. 3B). Further significant decreases were observed with doses ranging from 2.5–10 µM (Fig. 3B). At 10 µM, SB431542 reduced PRL secretion by 77% compared with cAMP alone (Fig. 3B). Higher doses were not tested. When either inhibitor (Activin-M108A or SB431542) was administered to cells in the absence of cAMP, PRL was not detected in culture medium. In addition, cell viability was tested using the trypan blue exclusion method at the end of the experiment and was found to not differ between treatments with medium-only, cAMP and either inhibitor. Microscopic visualization indicated cell morphology was not altered by inhibitor use, suggesting this effect was not due to toxicity to cells.Figure 3:Activin-M108A and SB43 154 inhibitors significantly decrease HESC decidualization.Confluent stromal cells were cultured for 4–5 days with cAMP, with medium changes every 2–3 days. (A) M108A was added at doses 0.39, 1.5, 6.25 and 25 nM to cAMP-treated cultures. (B) SB431542 was added at doses 1.25, 2.5, 5 and 10 µM. Controls (not decidualized; ND) and cAMP (decidualized; D). PRL secretion from cells was used as a marker of decidualization, and values were corrected for total protein. All data (mean ± SEM) is expressed as % change from control (100%). Results are combined data from triplicate wells of five independent experiments for each inhibitor. P < 0.05 and P < 0.01 compared with cAMP alone.Secretion of activin A, TGFβ1, BMP2 and BMP4, but not TGFβ2 or BMP7 from non-decidualized and decidualized HESC in vitroTGFβ ligands are active as secreted proteins. Therefore we measured BMP2, BMP4, BMP7, TGFβ1, TGFβ2 from three different cultures by ELISA and activin A by IFMA (n = 1) in serum-free culture medium. There was a 4-fold increase in dimeric activin A secreted from decidualized (1.9 ng/106cells) compared with non-decidualized cells (0.43 ng/106cells), supporting previous published data from our laboratory (Jones et al., 2002). Non-decidualized and decidualized HESC secreted BMP2, TGFβ1 and BMP4, each in the range of 5–20 pg/106 cells (Fig. 4). BMP7 and TGFβ2 levels were below the sensitivity of the assays (data not shown). Decidualized stromal cells secreted significantly higher amounts of BMP2 and TGFβ1 compared with non-decidualized cells (TGFβ1 20.6 ± 3.8 versus 14 ± 3.1 pg/106cells; BMP2 17.2 ± 2.32 versus 8.44 ± 1.5 pg/106cells) (Fig. 4). BMP4 secretion was increased slightly with decidualization (6.9 ± 2.17 versus 10.1 ± 3.45 pg/106cells), but this was not significant (Fig. 4). Western blot analysis was performed for GDF5, GDF8, GDF11 and TGFβ3 using 100–150 µg of cell lysates from non-decidualized and decidualized HESC; however, these proteins were below the level of detection. Positive controls (GDF5—rat heart and human term placenta; TGFβ3—human term placenta; GDF8 and GDF11—recombinant human myostatin/GDF8 and recombinant human GDF11, respectively; data not shown) validated the technique.Figure 4:BMP2, TGFβ1 and BMP4 protein secretion by HESC during decidualization.BMP2, BMP4 and TGFβ1 protein were measured in culture medium from non-decidualized (dark) and decidualized (grey) cells at Day 4 of culture. Results are combined data from n = 3 independent culture experiments and data (mean ± SEM) is represented as % difference from control (non-decidualized cells; defined as 100%). P < 0.05 and *P < 0.01 compared to control.Exogenous BMP2 and TGFβ1 increases HESC decidualization in vitroTo assess whether addition of rhBMP2 and/or rhTGFβ1 to HESC enhances decidualization in vitro, cells were decidualized with E2 and MPA under serum-free conditions for 8 days and PRL secretion measured. At Day 8, PRL secretion from non-decidualized cells was minimal but was increased when cells were decidualized (Fig. 5). When rhBMP2 (5, 50 or 500 ng/ml) was added to decidualizing HESC, PRL secretion was significantly increased at the 50 ng/ml dose compared with the decidualized control (P < 0.05). The addition of rhTGFβ1 at 0.5, 5 or 50 ng/ml to decidualizing HESC, however this was only significant at the 0.5 ng/ml dose (P < 0.01) compared with decidualized cells alone.Due to the variability of endometrial biopsies only one dose for each BMP2 and TGFβ1 has resulted in a significant promotion of decidualisation. Also as the rhTGFβ1 is known to be quite potent, this result could be explained by the fact that at the higher doses it might be negatively affecting the response, for example receptor and downstream signalling disturbances.Figure 5:Effect of exogenous BMP2 and TGFβ1 on PRL secretion from HESC.Cells were untreated or treated with E2 (10–8 M), MPA (10–7 M) alone or combined with recombinant human (rh) BMP2 (5–500 ng/ml) (A), rhTGFβ1 (0.5–50 ng/ml) (B) for 8 day. PRL was measured in cultured medium from the last 48 h of treatment. Data are represented as percent fold change from control (decidualized; D) defined as 100%. ND, non-decidualized HESC. Results are combined data from triplicate wells of three independent cultures. P < 0.05 and **P < 0.01 compared with D. DiscussionThis is the first study to identify BMP2, BMP4, BMP7, GDF5, GDF8 and GDF11 protein in secretory phase human endometrium and show the mRNA expression of these ligands in cultured HESC. In addition, the current study has immunolocalized BMP2, TGFβ1 and GDF5 protein to decidualized stromal cells, and provided further evidence for a role for secreted activin, BMP2 and TGFβ1 in decidualization through the use of novel inhibitors and a well-established ex vivo decidualization model.Decidualization of endometrial stromal cells is a pivotal event in the preparation for blastocyst implantation. Although this process is governed by hormonal regulation, it is becomingly increasingly evident that a multitude of factors, including cytokines and growth factors, are essential in ensuring this process occurs successfully such that endometrial stromal cells are differentiated into the implantation-favourable decidual cells. The identification of BMP2, TGFβ1 and GDF5 protein in mid-secretory human endometrial tissue, together with the increased secretion of both BMP2 and TGFβ1 from stromal cells following decidualization, suggests these TGFβ family members are hormonally regulated and important for decidualization.The finding that BMP2 secretion from HESC increases as they decidualize in vitro and that addition of rhBMP2 further drives decidualization as determined by secretion of PRL, extends the recent study (Li et al., 2007), which showed an increase of BMP2 mRNA with decidualization in a similar model and of PRL mRNA following treatment of HESC with rhBMP2. This is important given that translation and secretion do not necessarily follow increases in mRNA. The inhibitor M108A which acts primarily on those ligands acting through the activin type 11 receptor (Table I) is not known to block the action of BMP2 which acts through the BMP type II receptor (BMPRII). Therefore, the inability of M108A to fully block the decidualization process suggests that factors other than activins and GDFs are likely to be important to this process: BMP2 is a likely candidate.Interestingly in mice, uterine-specific conditional ablation of BMP2 resulted in uterine stromal disruption and an inability to undergo normal decidual reactions and pregnancy progression (Lee et al., 2007). Our demonstration of BMP2 in human endometrium in vivo with the strongest immunostaining in decidualized cells in the mid-late secretory phase and in the decidua of early pregnancy, confirms the biological relevance in women. Given this data in conjunction with our functional studies of human decidualization, there appears to be little doubt that BMP2 is an important mediator of both mouse and human decidualization.This study also demonstrates, for the first time, the presence of GDF5 in human endometrium and first trimester placenta where it is intensely and specifically stained in the decidual cells. GDF5 is best known for its role in early chondrogenesis and joint formation, being involved in inducing cartilage differentiation, growth and maturation (Mikic, 2004), as well as bone formation and angiogenesis (Yamashita et al., 1997). Differentiation and angiogenesis are the main events of decidualization. Unfortunately, assays for GDF5 were not available to measure the secretion from decidualized cells. The decrease in decidualization seen following administration of M108A was not likely to be due to GDF5 blockade, as this ligand has a relatively low, if any, affinity for the activin type II receptor. Future studies should address whether GDF5 has a functional role in decidualization, whether it be in promoting differentiation or in the formation of associated blood vessels to support the blastocyst in early pregnancy and throughout gestation.Although the TGFβs have been previously detected in the human endometrium, the information regarding their role during decidualization is contentious, as studies report contrasting effects on PRL production by endometrial stroma and during decidualization in vitro (Kubota et al., 1997; Kim et al., 2005). We showed here that under serum-free conditions, TGFβ1 secretion increases during cAMP-induced decidualization in vitro. Microarray studies have likewise shown increased TGFβ1 and TGFβ2 mRNA when cAMP or progesterone is used as decidual stimuli (Popovici et al., 2000; Tierney et al., 2003). The protein for both isoforms localizes to stromal cells in human endometrium throughout the menstrual cycle (Gold et al., 1994; Godkin and Dore, 1998), although decidualized stromal cells were not shown. In contrast, the present study identified TGFβ1 protein in decidualized stromal cells in human endometrium, and its secretion increased in cAMP-decidualized HESC. Administration of SB431542, which effectively blocks the TGFβs as well as activins by blocking their structurally similar ALK5 and ALK4 receptors (Inman et al., 2002), resulted in a 77% reduction in decidualization, a greater decrease than that seen with M108A (this study) and follistatin (Tierney and Giudice, 2004). This, along with our demonstration that addition of TGFβ1 during progesterone-induced decidualization enhances decidualization, provides convincing evidence to suggest TGFβ1 is important during decidualization. The level of activated TGFβ2 was undetectable by ELISA in cAMP-decidualized endometrial stromal cells, and therefore this ligand is unlikely to contribute to the decidual response.The identification of BMP4, BMP7, GDF8 and GDF11 during the mid-secretory phase is also an important finding. In addition to decidualization, there are many remodelling processes occurring in the endometrium during this time. Both BMP4 and GDF8 protein showed intense immunolocalization to the cytoplasm of the stroma, irrespective of whether or not the cells were decidualized. This was verified for BMP4 by the minimal difference in secretion levels between non-decidualized and decidualized cells in vitro. BMP4 mRNA expression has been localized to the vascular endothelial cells in pregnant mice (Ying and Zhao, 2000), and in the non-pregnant rat uterus, where it is constitutively expressed throughout the cycle (Erickson et al., 2004). This differs to the localization of BMP4 protein shown here in human endometrium. It may be that the mRNA localized to blood vessels in the rodent studies is not translated into functional protein.GDF8 (more commonly referred to as Myostatin) is a negative regulator of skeletal muscle, but more recently has been shown to be involved in glucose metabolism (McPherron and Lee, 2002), particularly in placenta (Mitchell et al., 2006). This study verified the presence of GDF8 protein in first trimester placenta, and showed for the first time its presence in secretory phase endometrium and in HESC. Given the lack of specific decidual staining it is unlikely that the response from M108A inhibitor was due to blocking GDF8 in HESC, even though the predominant receptor for GDF8 is the activin type II receptor. The presence of GDF8 in non-muscular tissue and organs is rare and future experiments should determine whether it has a role in endometrial biology as it does throughout gestation in the placenta.Vesicular staining for BMP7 (or Osteogenic protein-1; OP-1) was evident in highly decidualized cells during the secretory phase of human endometrium with minimal staining in selected glandular and luminal epithelial cells. OP-1 has been identified in the uterine epithelium of non-pregnant mice and rat uterus (Ozkaynak et al., 1997; Erickson et al., 2004), with localization remaining unchanged across the cycle. In pregnant mice, however, BMP7 mRNA was detected in the decidualizing stromal cells surrounding the blastocyst with no staining in the epithelial cells (Ying and Zhao, 2000). In the present study, BMP7 secretion from HESC was undetectable by ELISA, making it difficult to assess whether it may have a functional role during decidualization.In conclusion, this study has identified that a wide range of TGFβ ligands are present in mid-late secretory human endometrium but of these, in addition to activins, only BMP2, GDF5 and TGFβ1 increase in stromal cells as they undergo decidualization. BMP2 and TGFβ1, along with activins, are important in driving the decidualization process. Functions for the remaining ligands remain to be determined but they are likely to participate in the extensive tissue remodelling that occurs in this highly dynamic tissue as it prepares for blastocyst implantation and pregnancy. Supplementary materialSupplementary material is available at HUMREP Journal online FundingThis work was funded by the National Health and Medical Research Council of Australia (#241000, #388901) and a Prince Henry's Institute Postgraduate Scholarship. Supplementary Material [Supplementary Data]	[ "human endometrium", "decidualization", "implantation", "growth factors (activins, bmp, tgfβ)" ]	[ "P", "P", "P", "R" ]
Urol_Res-3-1-2082064	The analgesic effect of inhalational Entonox for extracorporeal shock wave lithotripsy	Extracorporeal shock wave lithotripsy (ESWL) is a non-invasive procedure that allows urinary stones to be fragmented using acoustic shock waves. The impact of the shock waves causes transient stinging pain at the entry site as well as deep visceral discomfort, requiring analgesia during the procedure. The objective of this study was to compare the clinical efficacy of Entonox and pethidine for pain relief during outpatient ESWL. We randomized 150 outpatients undergoing elective ESWL into three groups of 50 patients, each group receiving inhalational Entonox, intravenous pethidine, or inhalational compressed air during ESWL. Quantitative evaluation of pain was performed according to a visual analogue scale (VAS), before and after the intervention. Analysis of variance (ANOVA) and paired t tests were used to compare VAS scores in the three groups, before and after the intervention. Entonox and pethidine decreased the pain score significantly, while compressed air did not. There was no significant difference between pain relief by Entonox and pethidine. This study demonstrates for the first time that inhalational Entonox is an effective analgesic regimen for ESWL. Entonox can be regarded as an appropriate alternative to analgesics like opioids in relieving pain during ESWL. Introduction Since 1980, extracorporeal shock wave lithotripsy (ESWL) has become the first-line treatment for most urinary stones in adults and children [1]. However, the vast majority of patients do not tolerate the procedure without analgesia or sedation. Several monitored anesthesia care techniques have been used to provide sedation and analgesia [2–5]. Some of the analgesic drugs administered for ESWL, however, carry the risk of respiratory depression, delayed discharge, and/or unplanned hospital admission [6]. Entonox is a mixture of 50% nitrous oxide and 50% oxygen. Inhalation produces analgesia without loss of consciousness [7]. Self-administration of Entonox as an analgesic has been widely used for many years, particularly in obstetric practice [8], and by paramedics for prehospital care [9]. Entonox has a good safety record, with no serious side effects recorded for intermittent use [7], and is rapidly cleared from the circulation by exhalation [10]. The aim of this study was to compare the clinical efficacy of inhalational Entonox and pethidine for pain relief during outpatient ESWL. Methods A total of 150 patients (97 men and 53 women) with stones located in the renal pelvicalyceal system were prospectively randomized to receive Entonox (prefixed equimolar nitrous oxide and oxygen mixture), pethidine, or compressed air for sedoanalgesia during lithotripsy with a third generation electromagnetic lithotripter (Dornier Compact Delta magneto lithotripter). After institutional review board approval, written informed consent was obtained during the anesthesia consultation, which was performed at least 48 h before ESWL. During this consultation, the method of pain assessment was explained to the patients. Patients were asked to rate their pain on a scale from 0 (no pain) to 10 (worst possible pain). Patients were included in the study if they were older than 15 years of age, had pelvicalyceal stones, and understood the pain scoring. Patients were excluded from the trial if they had any of the following: (1) serum creatinine > 200 μmol/l; (2) history of chronic use of analgesics and/or sedatives; (3) allergy to any of the study medications; or (4) history of middle ear surgery within the previous month. Just before entering the lithotripsy room, the patients were randomly assigned to one of the three treatment groups (Entonox, pethidine, 1 mg/kg intravenously over 10 min, or compressed air, each to be provided if the patient asked for pain relief during ESWL) by opening a sealed envelope. All patients were assessed by a single physician who was blinded to the patient group assignment. All patients received no premedication and they were told not to eat or drink for at least 4 h prior to the procedure. The control group received compressed air, followed by Entonox gas if pain relief was considered inadequate by the patient. Pain was assessed before and 2 min after receiving pain medication by using a visual analogue scale (VAS) with two anchor points, zero denoting no pain and ten for the worst pain the patient had ever experienced. Statistical analysis Data are expressed as mean (±SD) and numbers (percentages). The paired t test was used to analyze VAS score before and after medication. Comparison of means (VAS decrease) in the three groups was performed using one-way ANOVA. Comparison of percentages was performed using Chi-square analysis. All statistical analyses were performed using SPSS® software, version 9.0, for Windows. Results throughout the text and tables are presented as mean ± SD unless otherwise specified, and statistical significance was defined as P < 0.05. Results The three study groups were comparable with respect to demographic data, history of urinary calculi and ESWL, and location of calculi. There were no statistically significant differences between the three groups (Table 1). All patients in the three groups required analgesia. Rescue medication in the form of pethidine was administered in only one case in the Entonox group after recording VAS scores because the analgesia was inadequate. Table 1Demographic data and ESWL variables in three different groupsGroupsEntonoxPethidineControlAge (years)43.76 ± 11.8945 ± 13.4643.9 ± 14.21Weight (kg)75 ± 278 ± 379 ± 2Height (cm)165 ± 10167 ± 12164 ± 8Gender (M/F) (n)30/2034/1633/17Stone location (%) Renal807884 Ureteral202015 Both021Number of shocks delivered2,010 ± 3202,000 ± 5602,500 ± 450Maximum voltage (kV) used4.04 ± 1.154.14 ± 0.754.51 ± 1.00Past history of urinary calculi (%)707276Past history of ESWL (%)424636Data are expressed as mean ± SD, numbers (n), or percentages (%) No ESWL procedure had to be prematurely terminated because of inadequate analgesia. A patient in the Entonox group had mild nausea that subsided spontaneously. The other patients did not develop any complications. Mean ± SD of VAS scores before and after intervention are shown in Table 2. Statistical analysis showed significant decrease (Table 2) in pain severity in patients receiving Entonox or pethidine. In contrast, patients receiving compressed air had no such change. Table 2Mean ± SD of VAS scores and P-value of comparing scores before and after Entonox, pethidine, and compressed air administration in different groupsVAS scoresEntonoxPethidineCompressed airEntonox in the compressed air groupBefore administration5.76 ± 1.926.04 ± 1.543.6 ± 1.14.96 ± 1.42After administration3.73 ± 2.134.11 ± 1.693.48 ± 1.33.8 ± 1.74P-value0.0010.0010.4260.001In control group, after testing with compressed air, we used Entonox. As can be seen in this group, like Entonox group, VAS score decreased significantly While VAS scores after receiving either Entonox or pethidine were significantly different from those in the control group (P = 0.001) after receiving compressed air, statistical comparison showed no significant difference between post-Entonox or post-pethidine VAS scores (P = 0.5), suggesting that inhalational Entonox may be as effective as intravenous pethidine in alleviating ESWL-associated pain. No significant difference was found in VAS scores of male and female patients in each group (Fig. 1). Fig. 1VAS scores in different groups before and after administration of Entonox, pethidine, and compressed air Discussion Extracorporeal shock wave lithotripsy (ESWL) has become a valuable asset to the urologist and greatly benefits stone patients. In a short period of time, ESWL has completely changed the management of urinary stone disease and has almost entirely supplanted open surgical and most endourologic approaches [11]. Today, third-generation mobile electromagnetic lithotripters yield an average of 80% stone-free rate for kidney and ureteral calculi [12]. As ESWL is frequently carried out on an outpatient basis, it is crucial to provide adequate analgesia with minimal adverse effects [13]. Chaussy and Thuroff [14] demonstrated that analgesia requirements in ESWL depend on lithotripter, stone location, age, gender, and number of shock waves. The pathogenesis of pain in ESWL is still not clearly known. Whether it is due to cutaneous or deep visceral afferent stimulation is debatable [15]. The pain is presumably due to cavitation-mediated stimulation of nerve fibers. The intensity of pain perceived during ESWL depends on the energy level of shock waves passing through the tissues [16]. Although the development of new generation lithotripters has decreased pain during ESWL, some form of sedation and anesthesia may still be required to provide adequate patient comfort and effective treatment [15]. Any technique for alleviating ESWL-associated discomfort/anxiety should ideally be reliable in its effect, of rapid onset and short duration of action, free of adverse events, cost-effective, and easily administered [17]. Opioids are the most commonly used analgesics during ESWL [18]. Despite their effectiveness, their use may be complicated by central nervous system (CNS) or respiratory depression, circulatory failure, or gastrointestinal problems [19]. Several studies have compared different local and systemic analgesia regimens for pain relief during ESWL, but this, to our knowledge, is the first randomized clinical trial to evaluate the efficacy of Entonox in the treatment of ESWL-associated pain. Nitrous oxide has been used for pain relief during childbirth since the 1930s and was initially delivered at a concentration of 50% in air, producing an effective analgesic mixture. In 1961, Mike Tunstall premixed 50% nitrous oxide with oxygen, and called it Entonox, which was stored as a compressed gas mixture in cylinders. Since then, the convenience of Entonox has made it a successful and popular analgesic for labour [20]. Entonox has also proved effective in providing pain relief in other specialties [21, 22]. Our randomized clinical trial demonstrated that Entonox and pethidine were equally effective in providing analgesia during ESWL and that they were superior to compressed air. Entonox fulfills many of the criteria for an ideal analgesic during ESWL [17] and, compared to intravenously administered pethidine, provided comparable analgesia with less CNS, circulatory, and respiratory depression or nausea and vomiting. Entonox provides rapid and effective analgesia without heavy sedation and leads to adequate patient relaxation and cooperation. The effect of Entonox was of short duration, allowing the patients to leave the ESWL unit without the need for a long recovery period [23]. Entonox appeared to be associated with few, minor, and short-lived adverse effects such as nausea and vomiting, dizziness, dry mouth (breathing dry gas), buzzing in the ears, and rarely, pins and needles or numbness, dreams or drowsiness, ranging from 0 to 30% [24]. Nitrous oxide–oxygen inhalation may thus provide a valuable alternative to conventional analgesia regimens during ESWL. In addition to short-acting parenteral sedative narcotics, e.g., alfentanil, midazolam, and propofol, topical agents such as EMLA cream, lidocaine, prilocaine and piroxicam have been used to minimize pain during SWL [25]. Further studies are, therefore, warranted to compare the efficacy and safety of Entonox with those of other analgesic regimens. Conclusions Entonox can provide easy, rapid, and adequate pain relief for patients undergoing ESWL, and is associated with few and minimal side effects.	[ "entonox", "extracorporeal shock wave lithotripsy", "pethidine", "pain relief" ]	[ "P", "P", "P", "P" ]
Immunome_Res-1-_-1312312	IMGT, the international ImMunoGeneTics information system®: a standardized approach for immunogenetics and immunoinformatics	IMGT, the international ImMunoGeneTics information system®, was created in 1989 by the Laboratoire d'ImmunoGénétique Moléculaire (LIGM) (Université Montpellier II and CNRS) at Montpellier, France. IMGT is a high quality integrated knowledge resource specialized in immunoglobulins (IG), T cell receptors (TR), major histocompatibility complex (MHC) of human and other vertebrates, and related proteins of the immune system (RPI) of any species which belong to the immunoglobulin superfamily (IgSF) and to the MHC superfamily (MhcSF). IMGT consists of five databases, ten on-line tools and more than 8,000 HTML pages of Web resources. IMGT provides a common access to standardized data from genome, genetics, proteome and three-dimensional structures. The accuracy and the consistency of IMGT data are based on IMGT-ONTOLOGY, a semantic specification of terms to be used in immunogenetics and immunoinformatics. IMGT-ONTOLOGY comprises six main concepts: IDENTIFICATION, CLASSIFICATION, DESCRIPTION, NUMEROTATION, ORIENTATION and OBTENTION. Based on these concepts, the controlled vocabulary and the annotation rules necessary for the immunogenetics data identification, classification, description and numbering and for the management of IMGT knowledge are defined in the IMGT Scientific chart. IMGT is the international reference in immunogenetics and immunoinformatics for medical research (repertoire analysis of the IG antibody sites and of the TR recognition sites in autoimmune and infectious diseases, AIDS, leukemias, lymphomas, myelomas), veterinary research (IG and TR repertoires in farm and wild life species), genome diversity and genome evolution studies of the adaptive immune responses, biotechnology related to antibody engineering (single chain Fragment variable (scFv), phage displays, combinatorial libraries, chimeric, humanized and human antibodies), diagnostics (detection and follow up of residual diseases) and therapeutical approaches (grafts, immunotherapy, vaccinology). IMGT is freely available at . Introduction IMGT, the international ImMunoGeneTics information system®[1,2], was created in 1989, by Marie-Paule Lefranc, at the Laboratoire d'ImmunoGénétique Moléculaire (LIGM) (Université Montpellier II and CNRS) at Montpellier, France, in order to standardize and manage the complexity of the immunogenetics data. Fifteen years later, IMGT is the international reference in immunogenetics and immunoinformatics, and provides a high quality integrated knowledge resource, specialized in the immunoglobulins (IG) and T cell receptors (TR), major histocompatibility complex (MHC) of human and other vertebrates, and related proteins of the immune systems (RPI) of any species which belong to the immunoglobulin superfamily (IgSF) and to the MHC superfamily (MhcSF)[3-13]. The number of potential protein forms of the antigen receptors, IG and TR, is almost unlimited. The potential repertoire of each individual is estimated to comprise about 1012 different IG (or antibodies) and TR, and the limiting factor is only the number of B and T cells that an organism is genetically programmed to produce. This huge diversity is inherent to the particularly complex and unique molecular synthesis and genetics of the antigen receptor chains. This includes biological mechanisms such as DNA molecular rearrangements in multiple loci (three for IG and four for TR in humans) located on different chromosomes (four in humans), nucleotide deletions and insertions at the rearrangement junctions (or N-diversity), and somatic hypermutations in the IG loci (see FactsBooks[3,4] for review). Although IMGT was initially implemented for the IG, TR and MHC of human and other vertebrates [6], data and knowledge management standardization, based on the IMGT unique numbering [14-19], has now been extended to the IgSF [15-17,20-22] and MhcSF [18,23,24] of any species. Thus, standardization in IMGT contributed to data enhancement of the system and new expertised data concepts were readily incorporated. IMGT, the international ImMunoGeneTics information system® consists of five databases, ten on-line tools and Web resources [1,2]. Databases include sequence databases (IMGT/LIGM-DB, IMGT/PRIMER-DB and IMGT/MHC-DB), one genome database (IMGT/GENE-DB) and one three-dimensional (3D) structure database (IMGT/3Dstructure-DB) [1,2] (Figure 1). Interactive tools are provided for sequence analysis (IMGT/V-QUEST, IMGT/JunctionAnalysis, IMGT/Allele-Align, IMGT/PhyloGene), genome analysis (IMGT/LocusView, IMGT/GeneView, IMGT/GeneSearch, IMGT/CloneSearch and IMGT/GeneInfo) and 3D structure analysis (IMGT/StructuralQuery) [1,2] (Figure 1). Web resources ("IMGT Marie-Paule page") comprise more than 8,000 HTML pages of synthesis [IMGT Repertoire (for IG and TR, MHC, RPI)], knowledge [IMGT Scientific chart, IMGT Education (IMGT Lexique, Aide-mémoire, Tutorials, Questions and answers), IMGT Medical page, IMGT Veterinary page, IMGT Biotechnology page, IMGT Index], and external links [IMGT Immunoinformatics page, IMGT Bloc-notes (Interesting links, etc.) and IMGT other accesses (SRS, BLAST, etc.)] [2]. Despite the heterogeneity of these different components, all data in the IMGT information system are expertly annotated. The accuracy, the consistency and the integration of the IMGT data, as well as the coherence between the different IMGT components (databases, tools and Web resources) are based on IMGT-ONTOLOGY[5], which provides a semantic specification of the terms to be used in immunogenetics and immunoinformatics. IMGT-ONTOLOGY, the first ontology in the domain, has allowed the management of knowledge in immunogenetics [2,25] and provided standardization for immunogenetics data from genome, genetics, proteome and 3D structures [3-13]. IMGT-ONTOLOGY concepts are available, for the biologists and IMGT users, in the IMGT Scientific chart[2], and for the computing scientists, in IMGT-ML which uses XML (eXtensible Markup Language) Schema [26]. Figure 1 IMGT, the international ImMunoGeneTics information system® . Databases and tools for sequences, genes and structures are in green, yellow and blue, respectively. The IMGT Repertoire and other Web resources are not shown. Interactions in the genetics, genomics and structural approaches are represented with dotted, continuous and broken lines, respectively. IMGT-ONTOLOGY concepts and IMGT Scientific chart rules The IMGT Scientific chart[2] comprises the controlled vocabulary and the annotation rules necessary for the immunogenetics data identification, description, classification and numbering and for knowledge management in the IMGT information system. Standardized keywords, labels and annotation rules, standardized IG and TR gene nomenclature, the IMGT unique numbering, and standardized origin/methodology were defined, respectively, based on the six main concepts of IMGT-ONTOLOGY: IDENTIFICATION, CLASSIFICATION, DESCRIPTION, NUMEROTATION, ORIENTATION and OBTENTION[2,5] (Table 1). The IMGT Scientific chart is available as a section of the IMGT Web resources (IMGT Marie-Paule page). Examples of IMGT expertised data concepts derived from the IMGT Scientific chart rules are shown in Table 1. Table 1 IMGT-ONTOLOGY concepts, IMGT Scientific chart rules and examples of IMGT expertised data concepts. IMGT-ONTOLOGY main concepts 5 IMGT Scientific chart rules [2] Examples of IMGT expertised data concepts [2] IDENTIFICATION Standardized keywords [5] Species, molecule type, receptor type, chain type, gene type, structure, functionality, specificity CLASSIFICATION Reference sequencesStandardized IG and TR gene nomenclature (group, subgroup, gene, allele) [5] Nomenclature of the human IG and TR genes (entry in 1999 in GDB, HGNC [27] and LocusLink at NCBI) [3, 4]Alignment of alleles [3, 4]Nomenclature of the IG and TR genes of all vertebrate species DESCRIPTION Standardized labels and annotations [5] Core (V-, D-, J-, C-REGION) Prototypes [5]Labels for sequencesLabels for 2D and 3D structures NUMEROTATION IMGT unique numbering [14-18] for: V- and V-LIKE-DOMAINs [16]C- and C-LIKE-DOMAINs [17]G- and G-LIKE-DOMAINs [18] Protein displaysIMGT Colliers de Perles [19]FR-IMGT and CDR-IMGT delimitations [16]Structural loops and beta strands delimitations [16, 17] ORIENTATION Orientation of genomic instances relative to each other Chromosome orientation Locus orientation Gene orientation DNA strand orientation OBTENTION Standardized originStandardized methodology [2] The IMGT Scientific chart rules, based on the IMGT-ONTOLOGY concepts [5], are used in the three major IMGT biological approaches, genomics, genetics and structural approaches [2], and corresponding data (Genes, Sequences, 3D structures) are available in the IMGT components (databases, tools and Web resources) [1,7-13]. IMGT sequence databases, tools and Web resources IMGT sequence databases, tools and Web resources correspond to the IMGT genetics approach that refers to the study of genes in relation with their polymorphisms, mutations, expression, specificity and evolution (Table 2). The IMGT sequence knowledge management and the IMGT genetics approach heavily rely on the DESCRIPTION concept (and particularly on the V-REGION, D-REGION, J-REGION and C-REGION core concepts for the IG and TR), on the CLASSIFICATION concept (gene and allele concepts) and on the NUMEROTATION concept (IMGT unique numbering [14-18]). Table 2 The IMGT sequence databases, sequence analysis tools and Web resources IMGT sequence databases [1] IMGT sequence analysis tools [1] IMGT Repertoire"Proteins and alleles" section [2](2) IMGT/LIGM-DB [7]IMGT/PRIMER-DB [1]IMGT/MHC-DB [28] IMGT/V-QUEST [10]IMGT/JunctionAnalysis [11]IMGT/Allele-AlignIMGT/PhyloGene [12]IMGT/Automat [29, 30] (1) Alignments of alleles IG and TR [3, 4]Alignments of alleles RPI [22]Protein displays IG and TR [3, 4, 16, 17]Protein displays MHC [18]Protein displays RPI [16-18, 21]Tables of alleles IG and TRTables of alleles RPI [22, 24]Allotypes Isotypes, etc. (1) IMGT/Automat [29, 30] is an integrated internal IMGT Java tool which automatically performs the annotation of rearranged cDNA sequences that represent the half of the IMGT/LIGM-DB content. So far 7,418 human and mouse IG and TR cDNA sequences have been automatically annotated by the IMGT/Automat tool, with annotations being as reliable and accurate as those provided by a human annotator. (2) IMGT publications from the IMGT Repertoire "Proteins and alleles" section are available as pdf in IMGT Locus in Focus , in IMGT Index (see also [2]). IMGT sequence databases IMGT/LIGM-DB IMGT/LIGM-DB is the comprehensive IMGT database of IG and TR nucleotide sequences from human and other vertebrate species, with translation for fully annotated sequences [7]. It was created in 1989 by LIGM (Montpellier, France), and is on the Web since July 1995 [6]. In August 2005, IMGT/LIGM-DB contained more than 96,500 sequences of 150 vertebrate species [7]. The unique source of data for IMGT/LIGM-DB is EMBL, which shares data with the other two generalist databases GenBank and DNA DataBank of Japan (DDBJ). Based on expert analysis, specific detailed annotations are added to IMGT flat files. The annotation procedure includes the IDENTIFICATION of the sequences, the CLASSIFICATION of the IG and TR genes and alleles, and the DESCRIPTION of all IG and TR specific and constitutive motifs within the nucleotide sequences. The Web interface allows searches according to immunogenetic specific criteria and is easy to use without any knowledge in a computing language. Selection is displayed at the top of the resulting sequences pages, so the users can check their own queries. Users have the possibility to modify their request or consult the results with a choice of nine possibilities. The IMGT/LIGM-DB annotations (gene and allele name assignment, labels) allow data retrieval not only from IMGT/LIGM-DB, but also from other IMGT databases. Thus, the IMGT/LIGM-DB accession numbers of the cDNA expressed sequences for each human and mouse IG and TR gene are available, with direct links to IMGT/LIGM-DB, in the IMGT/GENE-DB entries. IMGT/LIGM-DB data are also distributed by anonymous FTP servers at CINES and EBI and from many Sequence Retrieval System (SRS) sites . IMGT/LIGM-DB can be searched by BLAST or FASTA on different servers (EBI, IGH, INFOBIOGEN, Institut Pasteur, etc.). IMGT/PRIMER-DB IMGT/PRIMER-DB[1] is the IMGT oligonucleotide primer database for IG and TR, created by LIGM, Montpellier in collaboration with EUROGENTEC S.A., Belgium, on the Web since February 2002. In August 2005, IMGT/PRIMER-DB contained 1,827 entries. IMGT/PRIMER-DB provides standardized information on oligonucleotides (or Primers) and combinations of primers (Sets, Couples) for IG and TR. These primers are useful for combinatorial library constructions, scFv, phage display or microarray technologies. The IMGT Primer cards are linked to the IMGT/LIGM-DB flat files, IMGT Colliers de Perles and IMGT Alignments of alleles (IMGT Repertoire) of the IMGT/LIGM-DB reference sequence used for the primer description. IMGT/MHC-DB IMGT/MHC-DB[28] comprises databases hosted at EBI and includes a database of human MHC allele sequences or IMGT/MHC-HLA, developed by Cancer Research UK and maintained by ANRI, London, UK, on the Web since December 1998, and a database of MHC sequences from non human primates IMGT/MHC-NHP, curated by BPRC, The Netherlands, on the Web since April 2002. IMGT sequence analysis tools The IMGT sequence analysis tools comprise IMGT/V-QUEST[10], for the identification of the V, D and J genes and of their mutations, IMGT/JunctionAnalysis[11] for the analysis of the V-J and V-D-J junctions which confer the antigen receptor specificity, IMGT/Allele-Align for the detection of polymorphisms, and IMGT/PhyloGene[12] for gene evolution analyses. IMGT/V-QUEST IMGT/V-QUEST (V-QUEry and STandardization) is an integrated software for IG and TR [10]. This tool, easy to use, analyses an input IG or TR germline or rearranged variable nucleotide sequence. The IMGT/V-QUEST results comprise the identification of the V, D and J genes and alleles and the nucleotide alignments by comparison with sequences from the IMGT reference directory, the FR-IMGT and CDR-IMGT delimitations based on the IMGT unique numbering, the translation of the input sequence, the display of nucleotide and amino acid mutations compared to the closest IMGT reference sequence, the identification of the JUNCTION and results from IMGT/JunctionAnalysis (default option), and the two-dimensional (2D) IMGT Collier de Perles representation of the V-REGION [10] ("IMGT/V-QUEST output" in IMGT/V-QUEST Documentation). IMGT/JunctionAnalysis IMGT/JunctionAnalysis[11] is a tool, complementary to IMGT/V-QUEST, which provides a thorough analysis of the V-J and V-D-J junction of IG and TR rearranged genes. IMGT/JunctionAnalysis identifies the D-GENEs and alleles involved in the IGH, TRB and TRD V-D-J rearrangements by comparison with the IMGT reference directory, and delimits precisely the P, N and D regions [11] ("IMGT/JunctionAnalysis output results" in IMGT/JunctionAnalysis Documentation). Several hundreds of junction sequences can be analysed simultaneously. IMGT/Allele-Align IMGT/Allele-Align is used for the detection of polymorphisms. It allows the comparison of two alleles highlighting the nucleotide and amino acid differences. IMGT/PhyloGene IMGT/PhyloGene[12] is an easy to use tool for phylogenetic analysis of variable region (V-REGION) and constant domain (C-DOMAIN) sequences. This tool is particularly useful in developmental and comparative immunology. The users can analyse their own sequences by comparing with the IMGT standardized reference sequences for human and mouse IG and TR [12] (IMGT/PhyloGene Documentation). IMGT sequence Web resources The IMGT sequence Web resources are compiled in the IMGT Repertoire "Proteins and alleles" section that include Alignments of alleles, Proteins displays, Tables of alleles, Allotypes, Isotypes, etc. (Table 2). Standardized IMGT criteria for amino acid sequence analysis are described in [31]. IMGT gene databases, tools and Web resources IMGT gene databases, tools and Web resources correspond to the IMGT genomics approach that refers to the studies of the genes within their loci and on their chromosome [2] (Table 3). Table 3 The IMGT gene database, genome analysis tools and Web resources IMGT genome database [1] IMGT genome analysis tools [1] IMGT Repertoire"Locus and genes" section [2] (1) IMGT/GENE-DB [8] IMGT/LocusViewIMGT/GeneViewIMGT/GeneSearchIMGT/CloneSearchIMGT/GeneInfo [13] Chromosomal localizations [3, 4]Locus representations [3, 4]Locus descriptionGene exon/intron organizationGene exon/intron splicing sitesGene tablesPotential germline repertoiresLists of genesCorrespondence between nomenclatures [3, 4] (1) IMGT Web resources (IMGT Marie-Paule page) also include IMGT Index, IMGT Education (IMGT Lexique, Aide-mémoire, Tutorials, Questions and answers), The IMGT Medical page, The IMGT Veterinary page, The IMGT Biotechnology page, The IMGT Immunoinformatics page, IMGT Bloc-notes (Interesting links, etc.) [2] which are not detailed in this paper. (2) IMGT publications from the IMGT Repertoire "Locus and genes" section are available as pdf in IMGT Locus in Focus , in IMGT Index (see also [2]). IMGT/GENE-DB, the IMGT gene database Genomic data are managed in IMGT/GENE-DB, which is the comprehensive IMGT genome database [8]. IMGT/GENE-DB, created by LIGM (Montpellier, France) is on the Web since January 2003. In August 2005, IMGT/GENE-DB contained 1,377 genes and 2,207 alleles (673 IG and TR genes and 1,209 alleles from Homo sapiens, and 704 IG and TR genes and 998 alleles from Mus musculus, Mus cookii, Mus pahari, Mus spretus, Mus saxicola, Mus minutoïdes). All the human and mouse IG and TR genes are available in IMGT/GENE-DB. Based on the IMGT CLASSIFICATION concept, all the human IMGT gene names [3,4] were approved by the Human Genome Organisation (HUGO) Nomenclature Committee HGNC in 1999 [27], and entered in IMGT/GENE-DB [8], Genome DataBase GDB (Canada) [32], LocusLink and Entrez Gene at NCBI (USA) [33], and GeneCards [34]. Reciprocal links exist between IMGT/GENE-DB, and the generalist nomenclature (HGNC Genew) and genome databases (GDB, LocusLink and Entrez at NCBI, and GeneCards). All the mouse IG and TR gene names with IMGT reference sequences were provided by IMGT to HGNC and to the Mouse Genome Database (MGD) [35] in July 2002. Queries in IMGT/GENE-DB can be performed according to IG and TR gene classification criteria and IMGT reference sequences have been defined for each allele of each gene based on one or, whenever possible, several of the following criteria: germline sequence, first sequence published, longest sequence, mapped sequence [2]. IMGT/GENE-DB interacts dynamically with IMGT/LIGM-DB [7] to download and display gene-related sequence data. As an example ans as mentioned earlier, the IMGT/GENE-DB entries provide the IMGT/LIGM-DB accession numbers of the IG and TR cDNA sequences which contain a given V, D, J or C gene. This is the first example of an interaction between IMGT databases using the CLASSIFICATION concept. IMGT gene analysis tools The IMGT gene analysis tools comprise IMGT/LocusView, IMGT/GeneView, IMGT/GeneSearch, IMGT/CloneSearch and IMGT/GeneInfo. IMGT/LocusView and IMGT/GeneView manage the locus organization and the gene location and provide the display of physical maps for the human IG, TR and MHC loci and for the mouse TRA/TRD locus. IMGT/LocusView allows to view genes in a locus and to zoom on a given area. IMGT/GeneView allows to view a given gene in a locus. IMGT/GeneSearch allows to search for genes in a locus based on IMGT gene names, functionality or localization on the chromosome. IMGT/CloneSearch provides information on the clones that were used to build the locus contigs displayed in IMGT/LocusView (accession numbers are from IMGT/LIGM-DB, gene names from IMGT/GENE-DB, and clone position and orientation, and overlapping clones from IMGT/LocusView). IMGT/GeneInfo[13] provides and displays information on the potential TR rearrangements in human and mouse. IMGT gene Web resources The IMGT gene Web resources are compiled in the IMGT Repertoire "Locus and genes" section that includes Chromosomal localizations, Locus representations, Locus description, Gene exon/intron organization, Gene exon/intron splicing sites, Gene tables, Potential germline repertoires, the complete lists of human and mouse IG and TR genes, and the correspondences between nomenclatures [3,4] (Table 3). The IMGT Repertoire "Probes and RFLP" section provides additional data on gene insertion/deletion. IMGT structure database, tool and Web resources The IMGT structural approach refers to the study of the 2D and 3D structures of the IG, TR, MHC and RPI, and to the antigen or ligand binding characteristics in relation with the protein functions, polymorphisms and evolution (Table 4). The structural approach relies on the CLASSIFICATION concept (IMGT gene and allele names), DESCRIPTION concept (receptor and chain description, domain delimitations), and NUMEROTATION concept (amino acid positions according to the IMGT unique numbering [14-18]). Table 4 IMGT structure database, analysis tool and Web resources IMGT structural database [1] IMGT structural analysis tool [1] IMGT Repertoire"2D and 3D structures" section [2] IMGT/3D structure-DB [15] IMGT/StructuralQuery [15] 2D Colliers de Perles IG and TR [3, 4, 16, 17, 19] (1)2D Colliers de Perles MHC [18, 36]2D Colliers de Perles RPI [16-18, 21, 22, 24, 37]IMGT classes for amino acid characteristics [31]IMGT Colliers de Perles reference profiles [31]3D representations (1) (1) Cover of the Nucleic Acids Research 1999 database issue Structural and functional domains of the IG and TR chains comprise the variable domain or V-DOMAIN (9-strand beta-sandwich) which corresponds to the V-J-REGION or V-D-J-REGION and is encoded by two or three genes [3,4], the constant domain or C-DOMAIN (7-strand beta-sandwich), and, for the MHC chains, the groove domain or G-DOMAIN (4 beta-strand and one alpha-helix). A uniform numbering system for IG and TR V-DOMAINs of all vertebrate species has been established to facilitate sequence comparison and cross-referencing between experiments from different laboratories whatever the antigen receptor (IG or TR), the chain type, or the species [14-16]. In the IMGT unique numbering, conserved amino acids from frameworks always have the same number whatever the IG or TR variable sequence, and whatever the species they come from. As examples: Cysteine 23 (in FR1-IMGT), Tryptophan 41 (in FR2-IMGT), hydrophobic amino acid 89 and Cysteine 104 (in FR3-IMGT) (Figure 2). This numbering has been applied with success to all the sequences belonging to the V-set of the IgSF [20], including non-rearranging sequences in vertebrates (human CD4, Xenopus CTXg1, etc.) and in invertebrates (drosophila amalgam, drosophila fasciclin II, etc.) [15,16,21]. The IMGT unique numbering, initially defined for the V-DOMAINs of the IG and TR and for the V-LIKE-DOMAINs of IgSF proteins other than IG and TR, has been extended to the C-DOMAINs of the IG and TR (Figure 2B), and to the C-LIKE-DOMAINs of IgSF proteins other than IG and TR [17]. An IMGT unique numbering has also been implemented for the groove domain (G-DOMAIN) of the MHC class I and II chains (Figure 3), and for the G-LIKE-DOMAINs of MhcSF proteins other than MHC [18]. Figure 2 IMGT Colliers de Perles of a V-DOMAIN (A) and of a C-DOMAIN (B) (code PDB 1mcd in IMGT/3Dstructure-DB [9]). IMGT Colliers de Perles are shown on one layer (on the left hand side) and on two layers with hydrogen bonds (on the right hand side). (A) The IMGT Collier de Perles of a V-DOMAIN is based on the IMGT unique numbering for V-DOMAIN and V-LIKE-DOMAIN [16]. The CDR-IMGT are limited by amino acids shown in squares, which belong to the neighbouring FR-IMGT. The CDR3-IMGT extends from position 105 to position 117. CDR-IMGT regions are colored as follows on the IMGT site: CDR1-IMGT (blue), CDR2-IMGT (bright green), CDR3-IMGT (dark green) and hydrogen bonds are shown as green lines. (B) The IMGT Collier de Perles of a C-DOMAIN is based on the IMGT unique numbering for C-DOMAIN and C-LIKE-DOMAIN [17]. Amino acids are shown in the one-letter abbreviation. Arrows indicate the direction of the beta strands that form the two beta sheets of the immunoglobulin fold [3, 4]. Hatched circles correspond to missing positions according to the IMGT unique numbering [16, 17]. In the IMGT Collier de Perles on the IMGT Web site hydrophobic amino acids (hydropathy index with positive value) and Tryptophan (W) found at a given position in more than 50 % of analysed IG and TR sequences are shown in blue, and all Proline (P) are shown in yellow. Figure 3 IMGT Colliers de Perles of the two G-DOMAINs of MHC class I (A) and of MHC class II (B) proteins (codes PDB 1bd2 and 1aqd, respectively, in IMGT/3Dstructure-DB [9]). The IMGT Collier de Perles of a G-DOMAIN is based on the IMGT unique numbering for G-DOMAIN and G-LIKE-DOMAIN [18]. (A) The two MHC-I G-DOMAINs, G-ALPHA1 (top) and G-ALPHA2 (bottom), form the groove of the MHC class I chain (I-ALPHA). (B) The two MHC-II G-DOMAINs, G-ALPHA (top) of the MHC class II alpha chain (II-ALPHA) and G-BETA (bottom) of the MHC class II beta chain (II-BETA), form the groove of the MHC class II protein [36]. Amino acids are shown in the one-letter abbreviation. Hatched circles correspond to missing positions according to the IMGT unique numbering [18]. Positions in colour correspond to the IMGT contact sites provided, for each peptide/MHC 3D structure, in IMGT/3Dstructure-DB [36]. IMGT/3Dstructure-DB, the IMGT 3D structure database IMGT/3Dstructure-DB is the IMGT 3D structure database, created by LIGM, and on the Web since November 2001 [9]. In August 2005, IMGT/3Dstructure-DB contained 946 atomic coordinate files. IMGT/3Dstructure-DB comprises IG, TR, MHC and RPI with known 3D structures [9,36,37]. Coordinate files extracted from the Protein Data Bank (PDB) [38] are renumbered according to the standardized IMGT unique numbering [16-18]. The IMGT/3Dstructure-DB card provides, on-line, the complete information for each IMGT/3Dstructure-DB entry. The IMGT/3Dstructure-DB card shows a summary table and a menu that gives access to five sections: "Chain details", "Contact analysis", "Visualization with Jmol", "Renumbered file" and "References and links". The "Chain details" section provides chain description, IMGT gene and allele names, IMGT chain and domain labels, domain delimitations, amino acid positions according to the IMGT unique numbering, IMGT Colliers de Perles [16-19]. The "Contact analysis" section provides contact types and categories between domains (in IMGT/3Dstructure-DB Domain contacts) and atom contacts at the residue and position level (in IMGT/3Dstructure-DB Residue@Position contacts) [37]. (IMGT/3Dstructure-DB Documentation). The "Renumbered file" section downloadable provides renumbered IMGT/3Dstructure-DB flat files. IMGT/StructuralQuery tool The IMGT/StructuralQuery tool [9] analyses the interactions of the residues of the antigen receptors IG and TR, MHC, RPI, antigens and ligands. The contacts are described per domain (intra- and inter-domain contacts) and annotated in term of IMGT labels (chains, domain), positions (IMGT unique numbering), backbone or side-chain implication [37]. IMGT/StructuralQuery allows to retrieve the IMGT/3Dstructure-DB entries, based on specific structural characteristics: phi and psi angles, accessible surface area (ASA), amino acid type, distance in angstrom between amino acids, CDR-IMGT lengths. IMGT structure Web resources The IMGT stucture Web resources are compiled in the IMGT Repertoire "2D and 3D structures" section which includes 2D representations or IMGT Colliers de Perles [16-19], 3D representations, FR-IMGT and CDR-IMGT lengths [16], amino acid chemical characteristics profiles [31], etc. In order to appropriately analyse the amino acid resemblances and differences between IG, TR, MHC and RPI chains, eleven IMGT classes were defined for the 'chemical characteristics' amino acid properties and used to set up IMGT Colliers de Perles reference profiles [31]. The IMGT Colliers de Perles reference profiles allow to easily compare amino acid properties at each position whatever the domain, the chain, the receptor or the species. The IG and TR variable and constant domains represent a privileged situation for the analysis of amino acid properties in relation with 3D structures, by the conservation of their 3D structure despite divergent amino acid sequences, and by the considerable amount of genomic (IMGT Repertoire), structural (IMGT/3Dstructure-DB) and functional data available. These data are not only useful to study mutations and allele polymorphisms, but are also needed to establish correlations between amino acids in the protein sequences and 3D structures and to determine amino acids potentially involved in the immunogenicity. Conclusion In order to allow any IMGT component to be automatically queried and to achieve a higher level of interoperability inside the IMGT information system and with other information systems, our current objectives include the modelling of the three major IMGT biological approaches, genomics, genetics and structural approaches, the analysis of the IMGT components (databases, tools and Web resources) in relation with the concepts, and the development of Web services [2]. They are the first steps towards the implementation of IMGT-Choreography [2], which corresponds to the process of complex immunogenetics knowledge [25] and to the connection of treatments performed by the IMGT component Web services. IMGT-Choreography has for goal to combine and join the IMGT database queries and analysis tools. In order to keep only significant approaches, a rigorous analysis of the scientific standards [3,4], of the biologist requests and of the clinician needs [39-42] has been undertaken in the three main biological approaches: genomics, genetics and structural approaches. The design of IMGT-Choreography and the creation of dynamic interactions between the IMGT databases and tools, using the Web services and IMGT-ML, represent novel and major developments of IMGT, the international reference in immunogenetics and immunoinformatics. IMGT-Choreography enhances the dynamic interactions between the IMGT components to answer complex biological and clinical requests. Since July 1995, IMGT has been available on the Web at . IMGT has an exceptional response with more than 140,000 requests a month. The information is of much value to clinicians and biological scientists in general. IMGT databases, tools and Web resources are extensively queried and used by scientists from both academic and industrial laboratories, from very diverse research domains: (i) fundamental and medical research (repertoire analysis of the IG antibody sites and of the TR recognition sites in normal and pathological situations such as autoimmune diseases, infectious diseases, AIDS, leukemias, lymphomas, myelomas), (ii) veterinary research (IG and TR repertoires in farm and wild life species), (iii) genome diversity and genome evolution studies of the adaptive immune responses, (iv) structural evolution of the IgSF and MhcSF proteins, (v) biotechnology related to antibody engineering (single chain Fragment variable (scFv), phage displays, combinatorial libraries, chimeric, humanized and human antibodies), (vi) diagnostics (clonalities, detection and follow up of residual diseases) and (vii) therapeutical approaches (grafts, immunotherapy, vaccinology). Citing IMGT If you use IMGT databases, tools and/or Web resources, please cite [1] and this paper as references, and quote the IMGT Home page URL address, .	[ "imgt", "immunogenetics", "information system", "immunoinformatics", "knowledge resource", "immunoglobulin", "t cell receptor", "mhc", "superfamily", "database", "three-dimensional", "ontology", "annotation", "antibody", "3d structure", "collier de perles", "polymorphism", "hla" ]	[ "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "P", "U" ]
Qual_Life_Res-3-1-1915610	Evaluating the discriminatory power of EQ-5D, HUI2 and HUI3 in a US general population survey using Shannon’s indices	Objectives To compare quantitatively the discriminatory power of the EQ-5D, HUI2 and HUI3 in terms of absolute and relative informativity, using Shannon’s indices. Introduction The need for assessing health-related quality of life (HRQL) has brought forth hundreds of HRQL instruments, both generic and disease-specific [1, 2]. Generic instruments fall into two main categories: (1) preference-based health classification systems, and (2) non-preference based measures, sometimes referred to as health profile or psychometric measures [1, 3, 4]. Preference-based classification systems, also referred to as multi-attribute utility instruments (MAUIs) are standardized health state classifications that can be used to obtain a single summary index (utility score) or so-called preference weight for different health states. At the core of any MAUI is a classification system consisting of multiple attributes (dimensions) with ordered levels for each dimension. Most MAUIs are generic and aim to cover the full spectrum of disease and disability. MAUIs are widely used as measures of health outcome and are applied in clinical and economic evaluation (to calculate QALYs) and in population health surveys. Three widely used MAUIs are the EQ-5D, the Health Utilities Index Mark 2 (HUI2) and the Health Utilities Index Mark 3 (HUI3) [5–7]. All three instruments have shown acceptable psychometric properties as established by conventional measures [8, 9]. Feasibility, reliability, validity and responsiveness are important measurement properties in MAUIs, just as they are in non-preference based HRQL and health status measures such as the SF-36. However, these properties may be operationalized differently in MAUIs compared to non-preference based measures [4, 10, 11]. An underlying property to the concepts of reliability, validity and responsiveness is the ability of an instrument to discriminate between (‘true’) different levels of health. This requires a MAUI to define the full range of potential health states, and to be sensitive over this range. A necessary measurement property for any health status measure (including MAUIs) is the ability to discriminate among people at a single point in time. This property is sometimes referred to as sensitivity or, more accurately: “discriminatory power” [12–14]. Guyatt et al. (1992) proposed a reliability coefficient as a suitable statistic to express discriminatory power [15]. Reliability essentially reflects two different concepts: (1) consistency, e.g. between raters (inter-rater reliability) or over time (test-retest reliability), and (2) discriminatory power: the ability of an instrument to discriminate among people [16]. We propose Shannon’s indices of informativity as suitable measures that solely reflect discriminatory power [17]. Discriminatory power of MAUIs is usually investigated in an informal and partial manner by examining the frequency distributions, e.g. for floor or ceiling effects [12, 13, 18–20]. Shannon’s indices are suitable to assess discriminatory power in MAUIs for two reasons: first, they are theoretically based and second, they incorporate the frequency distribution across all categories of a MAUI’s health status classification system (not just the highest and lowest categories, as is the case with ceiling and floor effects). Our aim is to investigate the discriminatory power of the EQ-5D, HUI2 and HUI3 in a general population sample, as expressed by Shannon’s indices. Informativity was assessed separately by dimension and by MAUI as a whole. Methods Data A publicly available dataset was used (at http://www.ahrq.gov/rice/), resulting from the US EQ-5D valuation study [21, 22]. Collected data consisted of self-completed EQ-5D and HUI2/3 data from a sample of the general adult US population, with an over-sampling of Hispanics and non-Hispanic Blacks. The HUI2/3 data were collected using a standardized 15-item questionnaire, from which HUI2 and HUI3 health profiles were extracted using available recoding algorithms [23]. Only the responses of 3,691 respondents who had no missing data on any of the three instruments were included in this study (91.2% of the total number of respondents). Instruments The EQ-5D descriptive system consists of 5 dimensions (items) with 3 levels each, logically defining 243 unique health states (permutations). The HUI2 was originally developed to assess outcomes in survivors of cancer in childhood and contains 6 dimensions (excluding the original HUI2 dimension of fertility) with 4–5 levels per dimension. The HUI3, originally developed for a general population health survey in Canada, has 8 dimensions with 5–6 levels per dimension. The HUI2 and HUI3 descriptive systems define 8,000 and 972,000 unique health states, respectively [6]. Table 1 compares the 5 dimensions common to at least two of the classification systems: Mobility/Ambulation; Anxiety/Depression/Emotion; Pain/Discomfort (EQ-5D; HUI2; HUI3); Self-Care (EQ-5D; HUI2); and Cognition (HUI2; HUI3).Table 1Level descriptions for common dimensions between EQ-5D, HUI2 and HUI3EQ-5DHUI2HUI3MobilityMobilityAmbulationNo problems in walking aboutAble to walk, bend, lift, jump, and run normally for ageAble to walk around the neighbourhood without difficulty, and without walking equipmentSome problems in walking aboutWalks, bends, lifts, jumps, or runs with some limitations but does not require helpAble to walk around the neighbourhood with difficulty; but does not require walking equipment or the help of another personConfined to bedRequires mechanical equipment (such as canes, crutches, braces, or wheelchair) to walk or get around independentlyAble to walk around the neighbourhood with walking equipment, but without the help of another personRequires the help of another person to walk or get around and requires mechanical equipment as wellAble to walk only short distances with walking equipment, and requires a wheelchair to get around the neighbourhoodUnable to control or use arms and legsUnable to walk alone, even with walking equipment. Able to walk short distances with the help of another person, and requires a wheelchair to get around the neighbourhoodCannot walk at allSelf-careSelf-careNo problems with self-careEats, bathes, dresses, and uses the toilet normally for ageSome problems washing or dressing selfEats, bathes, dresses, or uses the toilet independently with difficultyUnable to wash or dress selfRequires mechanical equipment to eat, bathe, dress, or use the toilet independentlyRequires the help of another person to eat, bathe, dress, or use the toiletPain/DiscomfortPainPainNo pain or discomfortFree of pain and discomfortFree of pain and discomfortModerate pain or discomfortOccasional pain. Discomfort relieved by non-prescription drugs or self-control activity without disruption of normal activitiesMild to moderate pain that prevents no activitiesExtreme pain or discomfortFrequent pain. Discomfort relieved by oral medicines with occasional disruption of normal activitiesModerate pain that prevents a few activitiesFrequent pain; frequent disruption of normalactivities. Discomfort requires prescription narcotics for reliefModerate to severe pain that prevents some activitiesSevere pain. Pain not relieved by drugs and constantly disrupts normal activitiesSevere pain that prevents most activitiesAnxiety/DepressionEmotionEmotionNot anxious or depressedGenerally happy and free from worryHappy and interested in lifeModerately anxious or depressedOccasionally fretful, angry, irritable, anxious, depressed, or suffering "night terrors"Somewhat happyExtremely anxious or depressedOften fretful, angry, irritable, anxious, depressed, or suffering "night terrors"Somewhat unhappyAlmost always fretful, angry, irritable, anxious, depressedVery unhappyExtremely fretful, angry, irritable, anxious, or depressed usually requiring hospitalization or psychiatric institutional careSo unhappy that life is not worthwhile CognitionCognitionLearns and remembers school work normally for ageAble to remember most things, think clearly and solve day to day problemsLearns and remembers school work more slowly than classmates as judged by parents and/or teachersAble to remember most things, but have a little difficulty when trying to think and solve day to day problemsLearns and remembers very slowly and usually requires special educational assistanceSomewhat forgetful, but able to think clearly and solve day to day problemsUnable to learn and rememberSomewhat forgetful, and have a little difficulty when trying to think or solve day to day problemsVery forgetful, and have great difficulty when trying to think or solve day to day problemsUnable to remember anything at all, and unable to think or solve day to day problems Shannon’s indices: background and properties The Shannon index, named after Claude Shannon who is considered to be the founder of information theory, was initially developed to separate noise from information carrying signals in telecommunication systems [17]. The Shannon index is also known as the Shannon–Weaver index because of Warren Weaver’s contribution to Shannon’s original paper, and as the Shannon–Wiener index named after Norbert Wiener who independently developed a concept similar to Shannon’s [24, 25]. The Shannon index has been applied in a variety of fields, ranging from ecology (as a measure of biodiversity) to psychology, record linkage and molecular biology (genetic diversity) [26–30]. In information theory, the information of a signal is distinguished from the meaning or the semantic content of a signal. Rather, the information is quantified and is identified with uncertainty. Informativity is dependent on the number of classes (e.g. bits or response options) and the distribution of the observations (the ‘signal’) among classes. For classifications, this implies that if one would want to develop a useful (informative) distinction between, say, European countries, distinguishing between Scandinavian and non-Scandinavian countries would be far less informative than distinguishing between Northern, Western, Eastern and Southern European countries. Note that the latter classification not only contains more categories but the countries are also more evenly distributed among categories. The Shannon index is defined as: where H′ represents the absolute amount of informativity captured, C is the number of possible categories (levels or permutations in this study), and pi = ni/N, the proportion of observations in the ith category (i = 1,...,C), where ni is the observed number of scores (responses) in category i and N is the total sample size [17]. Any log base can be used, as long as one is consistent. Using log base 2, as did Shannon, allows the interpretation of the resulting units as bits per individual. The higher the index H′ is, the more information is captured by the system. In case of a homogeneous (rectangular) distribution, i.e. ratings are evenly distributed among categories (pi = p* for all i), the optimal amount of information is captured and H′ has reached its maximum (H′max) which equals log2C. If the number of categories (C) is increased, H′max increases accordingly but H′ will only increase if the newly added categories are actually used. The variance of the Shannon index is defined as [31]: Accordingly, standard errors and 95% confidence intervals can be calculated. The Shannon index combines the absolute information content as expressed by the number of categories with the extent to which the information is evenly spread over these categories. Shannon’s Evenness index (J′) exclusively reflects the latter component, i.e. the rectangularity of a distribution. This measure was first proposed by Lloyd and Ghelardi [32]; Shannon already referred to it as relative entropy and Pielou termed the concept ‘evenness’ [17, 33]. Shannon’s Evenness index (J′) is defined as: J′ = H′/H′max, which expresses the use of the system (H′) given its potential (H′max). Shannon’s index H′ can be considered as an expression of the absolute informativity of a system whereas Shannon’s Evenness index J′ expresses the relative informativity of a system or ‘evenness’ of a distribution, regardless the number of categories. Two alternative measures of (bio)diversity are the Simpson and the Brillouin index. We used the Shannon index, since the Brillouin index is dependent on sample size and the Simpson index gives very little weight to categories that are rarely occupied [26, 34, 35]. Shannon indices applied to MAUIs The basic characteristics of Shannon’s indices which make them suitable to reflect discriminatory power have been documented and are explained as follows. In an item where a response option has a very high (or low) endorsement, e.g. p is over 0.95 (or under 0.05), one learns very little because one can predict with more than 95% certainty what the answer will be. In other words, there is very little information being transmitted. Conversely, the maximum amount of information (uncertainty) is being transmitted when, in an item with two response options, p is 0.50 for each response option. As described above, this characteristic of an even distribution underlies the Shannon indices. In case of an even distribution, the item (dimension) is being most efficiently used, which means that the discriminant ability of the level descriptors is maximal. The Shannon indices can be calculated by dimension separately or by MAUI as a whole. To calculate Shannon’s indices by dimension, levels are treated as categories, so C represents the number of levels (L), pi is the proportion of responses of the ith level, and H′max equals log2L. Suppose the EQ-5D Mobility dimension is scored by 10 respondents: no problems (n = 6), some problems (n = 3) and confined to bed (n = 1). Shannon’s index for Mobility is calculated as H′ = –((0.6 log2 0.6) + (0.3 log2 0.3) + (0.1 log2 0.1)) = 1.30 and H′max = log23 = 1.58, so J′ = 1.30/1.58 = 0.82. Figure 1 illustrates the difference between absolute and relative informativity (H′, evenness J′) relative to the number of levels (L) in a series of hypothetical health classification systems designed to describe the same underlying dimension. For illustrative purposes we consider only one dimension. Figure 1a shows two distributions of responses corresponding to two different classification systems, both of which have 3 levels; one system results in a skewed distribution while the other results in a rectangular distribution. Assuming these responses are obtained within the same population, the system that yields the rectangular distribution is superior in discriminating between patients and the Shannon indices have both reached their maximum values. Figure 1b illustrates the concept of relative informativity. The left panel shows the same skewed distribution as depicted in Figure 1a, the right panel shows the same distribution of responses but now as it results from a 5 level classification system in which levels 2 and 4 are unused. Absolute informativity (Shannon’s H′) remains unchanged but J′ decreases, expressing lower relative informativity. Clearly, adding 2 extra levels that do not represent anyone in the population (no individual shifts from a current level to any of the new levels) does not lead to a gain in absolute informativity (H′) while the potential of a 5 level system is underutilized, compared to a 3 level system, which is expressed by a lower J′. So why not use just the Shannon Evenness index? Figure 1c shows the added value of absolute informativity (the H′ index). If the 3 and 5 level systems both yield rectangular distributions, evenness J′ will be the same but obviously H′ increases since the 5 level system is much more refined in discriminating between patients.Fig. 1Examples of Absolute Informativity (H′) and Relative Informativity (J′) with Skewed and Rectangular Distributions in a 3 Level System and 5 Level System To calculate Shannon’s indices by instrument as a whole, permutations are treated as unique categories (e.g. 243 categories for EQ-5D), so C is the number of permutations (Pmax), pi is the proportion of the ith permutation, and H′max now equals log2Pmax. Since the number of observations in our study (N = 3,691) is lower than the number of theoretically possible permutations in HUI2 (8,000) and HUI3 (972,000), maximum informativity (H′max) in HUI2 and HUI3, and consequently maximum relative informativity J′ cannot be reached a priori. Therefore, Shannon’s indices by MAUI as a whole were calculated using an estimation approach. Assuming that the current sample is representative, subsamples of the original set of observed health states were drawn in order to estimate the number of different health states in hypothetical populations of 1, 10, and 100 million respondents, by means of extrapolation. This procedure was repeated for different proportions of the population in relation to the number of health states (e.g. 11 different EQ-5D health states accounted for a 90% proportion of the respondents), in order to estimate the shape of the frequency distribution in the hypothetical populations of 1, 10, and 100 million respondents. Finally, Shannon’s H′ and J′ could be calculated (details can be obtained from the authors). Results The mean age of the respondents was 42.9 years (range: 18.0–99.3 years), with 42.2% of the respondents being male. White (non-Hispanic) respondents were 1,435 (38.9%), non-Hispanic blacks were 1,018 (27.6%) and Hispanic were 1,100 (29.8%). Table 2 shows the frequencies of responses to the EQ-5D, HUI2 and HUI3 dimensions. The dominant response was ‘no problems’ (level 1) for all dimensions in all instruments, with a proportion larger than 90% for 1 out of 5 dimensions in EQ-5D (Self-Care), 1 out of 6 in HUI2 (Self-Care) and 3 out of 8 in HUI3 (Hearing, Speech, Dexterity). In all EQ-5D and HUI2 dimensions, frequencies decreased with increasing level severity. In the HUI3 Cognition dimension however, more respondents reported problems at level 3 (17.9%) and level 4 (7.4%) than at level 2 (4.1%). Although small, these differences also occurred in the HUI3 Vision and HUI3 Hearing dimensions.Table 2Frequency distribution (%) of responses to the EQ-5D, HUI2 and HUI3 instruments (N = 3,691) Level 1Level 2Level 3Level 4Level 5Level 6EQ-5DMobility82.1717.530.30–––Self care95.584.010.41–––Usual activities84.8813.571.54–––Pain/Discomfort61.2834.714.01–––Anxiety/Depression73.8623.572.57–––HUI2Sensation44.5443.5410.761.16––Mobility87.248.483.600.680.00–Emotion69.2027.851.820.650.49–Cognition68.3629.941.630.08––Self-care96.642.950.190.22––Pain48.1740.947.102.980.81–HUI3Vision48.5047.871.002.470.030.14Hearing94.580.921.521.650.301.03Speech92.684.822.030.430.03–Ambulation87.248.482.551.060.510.16Dexterity92.445.820.790.700.140.11Emotion72.5022.323.741.160.27–Cognition68.364.1517.857.372.190.08Pain49.3433.7311.464.011.46– Figure 2 shows absolute informativity (Shannon’s H′) and relative informativity (Shannon’s Evenness J′) of the common dimensions among the three instruments. Absolute informativity (H′) was highest for HUI3 in all common dimensions, with largest differences between HUI3 and the other two instruments in the dimensions Pain/Discomfort (0.52 compared to EQ-5D; 0.15 compared to HUI2) and Cognition (0.41 compared to HUI2).Fig. 2The Shannon Index (H′) and the Shannon Evenness Index for the common dimensions between EQ-5D, HUI2 and HUI3: comparison by dimension. NA = not available; Confidence intervals had an average range of 0.0012 (H′) and of 0.00081 (J′) Relative informativity (J′) was highest for EQ-5D in all common dimensions, with largest differences with the other two instruments in the dimensions Mobility/Ambulation (0.14 compared to HUI2; 0.16 compared to HUI3) and Anxiety/Depression/Emotion (0.14 compared to HUI2; 0.13 compared to HUI3). Table 3 shows Shannon’s indices by classification system as a whole. The EQ-5D, HUI2 and HUI3 descriptive systems distinguished 91, 322, and 694 observed different unique health states, accounting for 37.4%, 4.0%, and 0.07% of all possible permutations, respectively. The estimation procedure indicated that absolute informativity was highest for HUI3 (range 10.96–13.36), followed by HUI2 (range 8.57–9.48), and lowest for EQ-5D (range 6.24–6.41). Relative informativity was highest in EQ-5D (range 0.79–0.81), followed by HUI2 (range 0.66–0.73), and lowest for HUI3 (range 0.55–0.67).Table 3Shannon’s index (H′) and Shannon’s evenness index (J′) for EQ-5D, HUI2, and HUI3: Comparison by instrument EQ-5DHUI2HUI3Pmax (permutations)2438000972,000Observed health states91322694H′max7.9212.9719.89EstimationH′J′H′J′H′J′N = 1,000,0006.240.798.570.6610.960.55N = 10,000,0006.370.809.120.7012.290.62N = 100,000,0006.410.819.480.7313.360.67 Discussion We compared the discriminatory power of the EQ-5D, HUI2 and HUI3 in the general population, using Shannon’s indices of absolute and relative informativity, for each dimension separately and by MAUI as a whole. As might be expected in a general population sample, most respondents reported no problems on all dimensions and there were fewer responses with increasing level severity. An exception is HUI3 Cognition, where respondents reported more problems on levels 3 and 4 than on level 2. This is probably due to the fact that this dimension is not unidimensional, and levels 2 and 3 are conceptualized parallel rather than ordinal. That is, HUI3 Cognition level 2 focuses on problems in thinking and problem solving, level 3 addresses problems in remembering, whereas level 4 combines the problems mentioned in levels 2 and 3. Absolute informativity by dimension was highest for the HUI3 descriptive system. EQ-5D appears to underperform in the Pain/Discomfort dimension. Moreover, EQ-5D appears to miss a considerable ‘amount’ of disability: 61.3% of the population indicated to have no problems on EQ-5D, against 48.2% on HUI2 and 49.3% on HUI3 (Table 2). Shannon’s H′ ‘translated’ this difference adequately (Figure 2). Apparently, for this population, the EQ-5D would benefit from more levels on the Pain/Discomfort dimension. Regarding the Cognition dimension, the difference in absolute informativity between HUI2 and HUI3 might be explained by the 2 extra levels in HUI3, but the higher J′ value in HUI3 suggests an alternative contributive factor. One explanation may be that HUI3 Cognition is not unidimensional and more sensitive to mild problems (levels 2–4) than HUI2 Cognition (level 2). Another explanation could be that the difference is due to currently suboptimal recoding algorithms. For relative informativity by dimension, the EQ-5D descriptive system showed superior results in Mobility/Ambulation, Self-Care and Anxiety/Depression/Emotion. The large differences in Mobility/Ambulation could be due to a relatively large leap in the grading of the level descriptions in HUI3 Ambulation, where the difference between level 1 (‘without difficulty’) and level 2 (‘with difficulty’) can be considered disproportionately large in a 6 level dimension. The same leap from level 1 (‘normal’) to level 2 (‘with difficulty’) occurs in HUI2 Self-Care. We found that the 3 level EQ-5D Self Care outperformed the 4 level HUI2 Self-Care in both absolute and relative informativity (Fig. 2), which is probably due to the severe grading of level 2 in HUI2. The difference in relative informativity between EQ-5D and the HUI instruments in Anxiety/Depression/Emotion is probably due to the 2 extra levels in HUI2 and HUI3 that are rarely endorsed. Overall, performance in terms of informativity of EQ-5D, HUI2 and HUI3 of the common dimensions varies over dimensions. The Pain/Discomfort dimension of EQ-5D, but perhaps also other dimensions, might benefit from an extension to 4 or 5 levels. HUI2 and HUI3 might benefit from more sensitive grading terms in their level descriptions, especially the ‘threshold’ level 2, in Ambulation (HUI3) and Self-Care (HUI2). When assessing informativity by instrument, HUI3 shows the best results on absolute informativity but the lowest on relative informativity while EQ-5D shows highest relative informativity and lowest absolute informativity. HUI2 seems to be the optimal compromise. The importance of differences in the Shannon indices ultimately requires empirical evidence over a wider range of populations, conditions and instruments, including evidence on discriminant validity. As Shannon’s indices are new in the field of health status measurement, some methodological issues need to be addressed, taking into account that their principal focus is on classifications with mutually exclusive categories, rather than conventional (health status) measures which by design contain multiple partially overlapping items. The Shannon indices share some properties with reliability coefficients. Like reliability indices, they express discriminatory power. Furthermore, they are also non-dimensional, i.e. they have no relation to the content, meaning or clinical relevance of what the instrument aims to measure, which make them suitable for comparability, between instruments as well as between populations. However, reliability reflects two different concepts: discriminatory power as such, and consistency, e.g. consistency between raters (inter-rater reliability) or consistency over time (test-retest reliability). This requires a repeated measurement (repetition ‘over raters’ or over time) which introduces an error component in case of a difference among the repeated measurements. Shannon’s indices solely reflect discriminatory power, and need only a single measurement. Furthermore, the Shannon indices are non-parametric measures and therefore highly suitable for nominal or ordinal measurement scales. Since Shannon’s indices have no dimension and are independent of any external standard, a rectangular distribution is always the ideal from the perspective of informativity. When comparing the discriminatory power of similar dimensions of different MAUIs, rectangularity is always optimal as it reflects which MAUI is the most sensitive in discriminating between different persons in that particular population. This implies that one MAUI cannot be superior in varying populations (e.g. a general population and a diseased population sample). Furthermore, MAUIs are bound to score rather low on discriminatory power in a general population sample, as the extreme categories, which have to be included for coverage of the full spectrum of diseases, will not be endorsed frequently. Previously, the common approach to investigate discriminatory power was examining the frequency distributions of responses, e.g. for ceiling or floor effects. A comprehensive, formal measure to express discriminatory power such as Shannon’s indices seems clearly superior to such a ‘face-value’ method. Furthermore, when the number of categories is large (e.g. when comparing MAUIs as a whole), it becomes very difficult to make a sound comparison by just looking at the distributions. We have demonstrated the use of the Shannon indices to compare the discriminatory power of different MAUIs, to show which instrument is more sensitive in differentiating between levels of health in the population at hand. But they may also be used to guide the development of new, or optimization of existing MAUIs, by helping determine how many levels are efficient for each dimension. This is a particularly relevant consideration for MAUIs, since adding extra levels in a descriptive system makes it increasingly complex, and the derivation of a robust set of preference weights more challenging. Apart from MAUIs, the Shannon indices can also be used in a wide range of other classifications in the medical domain (e.g. the Karnofsky scale, the Spitzer QL index) and in the clinical domain (e.g. the APGAR score, the Child-Pugh classification). A practical weakness of the Shannon approach is that when the sample size is exceeded by the total number of health states described by all permutations across all dimensions of a MAUI, informativity (for the instrument as a whole) has to be estimated. This implies that using the Shannon Evenness index by instrument is not very practical when a health classification system has a large number of permutations as was the case in HUI3 (972,000 permutations). This however is not a disadvantage of the Shannon methodology per se, but also a matter of classification design (overload of dimensions with detailed response options producing an excessive amount of ‘empty’ permutations), or a practical problem (excessive data collection). From a clinical or psychometric perspective it may seem tempting to extend any MAUI with extra levels or dimensions as it provides more clinically relevant detail generally and improves reliability. But Shannon’s indices reveal that this may not always be a prudent approach. Increasing the number of levels per dimension (or permutations in the entire system) will probably result in higher H′ values but J′ values are likely to drop, as in fact our results for HUI3 indicate. This raises the question where the balance between H′ and J′ is optimal as more categories require more extensive subsequent studies to derive utility functions for the associated classification system. How the Shannon indices will behave in a different population, such as patient populations, remains to be investigated. So far, Shannon’s indices proved to be useful in showing weaknesses of level gradings used in EQ-5D, HUI2 and HUI3, and offers leads for improvement, establishing their practical psychometric value.	[ "psychometrics", "population health", "health status", "methodology", "health-related quality-of-life" ]	[ "P", "P", "P", "P", "M" ]
Arch_Orthop_Trauma_Surg-3-1-1914284	Poor accuracy of freehand cup positioning during total hip arthroplasty	Several studies have demonstrated a correlation between the acetabular cup position and the risk of dislocation, wear and range of motion after total hip arthroplasty. The present study was designed to evaluate the accuracy of the surgeon’s estimated position of the cup after freehand placement in total hip replacement. Peroperative estimated abduction and anteversion of 200 acetabular components (placed by three orthopaedic surgeons and nine residents) were compared with measured outcomes (according to Pradhan) on postoperative radiographs. Cups were placed in 49.7° (SD 6.7) of abduction and 16.0° (SD 8.1) of anteversion. Estimation of placement was 46.3° (SD 4.3) of abduction and 14.6° (SD 5.9) of anteversion. Of more interest is the fact that for the orthopaedic surgeons the mean inaccuracy of estimation was 4.1° (SD 3.9) for abduction and 5.2° (SD 4.5) for anteversion and for their residents this was respectively, 6.3° (SD 4.6) and 5.7° (SD 5.0). Significant differences were found between orthopaedic surgeons and residents for inaccuracy of estimation for abduction, not for anteversion. Body mass index, sex, (un)cemented fixation and surgical approach (anterolateral or posterolateral) were not significant factors. Based upon the inaccuracy of estimation, the group’s chance on future cup placement within Lewinnek’s safe zone (5–25° anteversion and 30–50° abduction) is 82.7 and 85.2% for anteversion and abduction separately. When both parameters are combined, the chance of accurate placement is only 70.5%. The chance of placement of the acetabular component within 5° of an intended position, for both abduction and anteversion is 21.5% this percentage decreases to just 2.9% when the tolerated error is 1°. There is a tendency to underestimate both abduction and anteversion. Orthopaedic surgeons are superior to their residents in estimating abduction of the acetabular component. The results of this study indicate that freehand placement of the acetabular component is not a reliable method. Introduction Acetabular cup position after total hip arthroplasty is correlated with the risk of dislocation, wear and range of motion [1–6]. Lewinnek et al. [7] described a safe range (5–25° anteversion and 30–50° abduction) to position the cup. Within this range the dislocation rate was 1.5% and outside this range 6.1%. Although the position of the cup is important for the prognosis and function of the hip, most surgeons place the cup without any specific guidance devices. During surgery the surgeon estimates the position of the acetabular component and decides if it meets the desired orientation before securing it. In this study the accuracy of the surgeons peroperative estimation of the position of the cup is evaluated. Materials and methods All patients undergoing primary total hip arthroplasty were enrolled in the study. Patient characteristics as name, sex, age, operated side and the body mass index were recorded. All cooperating surgeons (three orthopaedic surgeons and nine residents) received a list on which the following data concerning the operation were recorded: fixation (cemented or uncemented), cup model, surgical approach, complications and the peroperative estimated anteversion and abduction of the acetabular component. Two days after the operation two standardised X-rays were made. One anteroposterior X-ray was taken with the beam centered over the hip, the second was a plain AP radiograph centred on the symphisis, showing both hips. Radiographic cup anteversion was measured on the first radiograph, according to Pradhan [8], and cup abduction was assessed on the second by measuring the angle between the teardrop line and the line bisecting the opening of the acetabular cup. The combination of both radiographs was used to determine if a cup was in the ante- or retroverted position, as described by Flabeck et al. [9]. For both anteversion and abduction, the inaccuracy of estimation was determined by calculating the difference between peroperative estimation and postoperative X-ray measurements. Mean values and standard deviations were calculated for measured cup position, estimated cup position and inaccuracy of estimation. A comparison of these values was assessed between orthopaedic surgeons and residents, using student’s t test (P < 0.05 assumed as significant) and a multivariate linear regression analysis was applied on the data in order to investigate which factors influenced the inaccuracy of estimation. Results Two hundred cups in 194 patients, placed between June 2003 and May 2005, were included in the study. There were 55 males and 139 females. At time of operation the mean age of the patients was 72.4 years (34–92) with a mean body mass index of 27.5 (16.6–38.1). The cups were placed by 12 different surgeons. Eighty-five cups (42.5%) were placed by three orthopaedic surgeons and 115 (57.5%) by 9 of their residents, always under supervision of one of the surgeons. One hundred and fifty-seven (78.5%) cups had a cemented and 43 (21.5%) an uncemented fixation; 89 (44.5%) cups were placed in the patients’ right hip and 111 (55.5%) in the left. Two surgical approaches were used. The anterolateral approach according to Mallory was used in 57 cups (28.5%), supervised by one orthopaedic surgeon and the postero-lateral approach by the other two surgeons in 143 cups (71.5%). The residents used the approach of their supervising surgeon. For the anterolateral approach according to Mallory, mean placement of the cup was in 11.8° anteversion and 47.1° abduction. Positioning of the cup differed significantly when a posterolateral approach was applied: 17.7° anteversion and 50.7° abduction (both P < 0.001). Comparing both anterolateral and posterolateral approaches, no significant differences were found for inaccuracy of estimation for either anteversion (4.9° vs. 5.5°, respectively) or abduction (5.3° vs. 6.1°). Table 1 shows relevant data of estimations and the measurements of the acetabular component by both orthopaedic surgeons and residents. None of the cups was placed in retroversion. Table 1Mean estimated and measured anteversion and abductionCups NAnteversion (degrees)DifferenceMean inaccuracy of estimationEstimatedMeasuredIa Orthopedic surgeon8514.514.7−0.25.2 (SD 4.5) Residents11514.716.9−2.25.7 (SD 5.0) Total20014.6 (SD 5.9)16.0 (SD 8.1)−1.45.5 Significance––––Cups NAbductionDifferenceMean inaccuracy of estimationEstimatedMeasuredIb Orthopedic surgeon8547.248.6−1.44.09 (SD 3.9) Residents11545.650.5−4.96.28 (SD 4.6) Total20046.3 (SD 4.3)49.7 (SD 6.7)−3.45.4 Significance0.0010.0430.0000.001 From the data from which Fig. 1 was constructed, it was derived that 129 cups (64.5%) were placed within 5° of the estimated abduction, for anteversion this was the case in 122 cups (61%). There is a tendency to underestimate both abduction and anteversion. Fig. 1Results of the distribution of the difference between the estimated and measured values (degrees) for abduction (a) and anteversion (b) of the acetabular components (N) Table 2 demonstrates the results of the chance for cup placement according to Lewinnek (5–25° anteversion and 30–50° abduction), based upon inaccuracy of estimation for the orthopaedic surgeons, their residents and the group in total. Virtual intended cup placement was set at 15° anteversion and 40° abduction. Only 56.5% of the cups were placed in this safe zone according to their measured abduction and anteversion. Table 3 shows the percentage for the group and orthopaedic surgeons and residents separately, for cup placement within different ranges (20, 10, 5, 2.5 and 1°) of an intended cup position X. Table 2Results of the chance for cup placement within the limits according to Lewinnek (5–25° anteversion and 30–50° abduction) for the orthopedic surgeons, their residents and the entire group. Intended cup placement was virtually set on 15° anteversion and 40° abductionAnteversion (%)Abduction (%)Placement according to Lewinnek (%)Orthopedic surgeons85.793.780.3Residents80.578.963.5Group82.785.270.5Table 3Deviation in % of cup placement in between 20°, 10°, 5°, 2.5° and 1° of position X for the entire group (orthopaedic surgeons and residents): (a), the orthopaedic surgeons: (b) and the residents: (c)X ± 20 (%)X ± 10 (%)X ± 5 (%)X ± 2.5 (%)X ± 1 (%)(a) Group Abduction10085.246.82616.4 Anteversion99.982.74626.717.5 Combined99.970.521.56.92.9(b) Surgeon Abduction10093.759.33421.2 Anteversion99.985.748.327.517.6 Combined99.980.328.69.43.7(c) Residents Abduction99.878.93920.712.7 Anteversion99.880.544.526.217.4 Combined99.663.517.45.42.2 A multivariate regression analysis was applied in order to identify any factors that might be responsible for the inaccuracy of the estimation. A significant result was found for abduction and anteversion, concerning age of the patient and if the patient was operated by an orthopaedic surgeon or resident. Other factors like body mass index, sex, operated side, (un)cemented fixation of the acetabular component, model of the cup and the surgical approach did not reveal any significant differences. Discussion Malposition of the acetabular cup is probably the most important factor for dislocation of a total hip prosthesis. Therefore it is essential that the surgeon has maximum control over the position of the socket during the operation. Free hand positioning with the patient in a standardized position and a cup positioner with the patient and floor as reference is the routine method. Specific mechanical alignment guides were designed to add precision, as do navigation systems in the concept of CT-guided computer-assisted surgery. Surprisingly, there are very limited data on the precision rate of free hand positioning of the acetabular shell. The aim of our prospective study was to determine the accuracy of the free hand technique. Only then the presently introduced systems can be tested against this standard. The term “free hand” is not used uniformly in the literature. Either it is referred to pure manual positioning or with the aid of the acetabular cup impactor-positioner provided with the implant. In the present study, the latter definition was applied. One comparable study is that of Saxler et al. [10] who showed, in a retrospective CT-controlled design, that 27 out of their 105 cups (25.7%) were placed within the safe zone of Lewinnek. A second study by DiGioa et al. [11] with a specific mechanical acetabular alignment guide (with A-frame) where the results were controlled peroperatively with a hip navigation system in 74 hips aimed at 45° of abduction and 20° of flexion. With their specific guide only 22% of the cups were placed within Lewinneks safe zone. Compared with these two studies our prospective data demonstrate an accuracy within the safe zone of 70.5%. There is an elegant in vitro study by Jolles et al. [12] comparing free hand (without the help of any guide), mechanical alignment guide and computer-assisted cup placement. Onehunderd and fifty acetabular implants were placed in ten identical models of the pelvis (covered with artificial soft tissue of soft cast and foam) by ten surgeons. The mean accuracy for anteversion was 8° (5.0–10.5) for free hand with cup positioner and 4° (3.0–5.5) for abduction; with computer navigation this was 1.5° (1.0–2.0) and 2.5° (2.0–3.5), respectively. In the present in vivo study the data were 5.5° for anteversion and 5.4° for abduction with 200 acetabular shells. Jolles et al. state that their in vitro conditions of ideal approach and anatomic relations and perfect placement of the patient would favour the free hand positioning as computer-assisted surgery devices are only slightly sensitive to modifications of these parameters and inaccuracy of the freehand will be enhanced greatly in vivo. Our study proves this statement not to be the case as results are quite comparable. Without doubt it is essential to actually place what is aimed for. Our present study shows that the 70.5% positioned within the safe zone is remarkably good for free hand placement compared with literature. But when the target is reset from the wide safe zone of Lewinnek to for instance within 5° or even 1° of error from the judged position for both abduction and anteversion, only 21.5 and 2.9% of our sockets are placed within this narrow definition of safe zone. We believe that the power of our study comes from the distinctive set up where the recorded data are a reflection of the surgeon’s perception of cup placement. To our opinion this is superior then to allow for a wide range or set a predefined target. At the actual point of cup placement a surgeon is also committed to or governed by the patient’s anatomy, which will influence the actual position. For instance when less anteversion is accepted in the socket; the stem will be given some additional anteversion. Another important item as stipulated by DiGioia et al. [11] is the actual position of the patient in the lateral decubitus position. They demonstrated that the mean difference of pelvic orientation on the operating table with the desired position during acetabular alignment was 18° in version and 3° in abduction. This effect is not completely eliminated with the use of computer navigation as these systems are influenced by pelvic tilt [13]. Considering the estimation of abduction, significant results were found in favour of the orthopaedic surgeons compared to their residents. A learning curve as a result of experience is probably responsible for this difference. As to surgical approach, acetabular components placed posterolaterally were measured to have a significantly higher-degree of anteversion and abduction. This is obviously related with the intention to minimize the change of a dislocation for both approaches. Although both approaches take a different anatomical route to the acetabulum, visualization and presentation of the acetabulum do not seem to be important factors. The conclusion is that an anterolateral or posterolateral approach is not a distinguishing factor in the accuracy of cup placement in our series. An important potential flaw in this investigation is the use of planar radiographs for the evaluation of true abduction and especially anteversion. Variations in pelvic flexion-extension during imaging are responsible of variations in flexion up to −26° to +10° [14]. This study concludes that radiographic measurements are not a reliable method to evaluate cup orientation, especially flexion or anteversion alignment [14]. Pradhan, however, stated that his method, as used in our study, proved to be reliable in an in vitro model in which the calculated anteversion on the X-ray was compared with the true (known) cup anteversion. In our study special attention was given for optimal positioning of the patients pelvis to minimize rotation and tilt before the X-ray was taken. From a practical point of view it is preferable to determine the actual cup position with planar radiographs rather then with CT as in most clinics the availability and the costs of a CT scan combined with the high radiation dose for the patient are reasons for not using this instrument as a routine practice. From this prospective study it is concluded that an accuracy of 70.5% for placement of the socket with an acetabular shell imactor-positioner within Lewinnek’s safe zone and of only 21.5% within an error of 5° for both abduction and anteversion is unacceptable. The strength of our research is that literature is extremely scarce on studies as our prospective study on free hand positioning without special aiming devices. The results of the present study appear convincing but limitations are the high-number of participating surgeons, different surgical approaches (anterolateral and posterolateral) and two methods of cup fixation (uncemented and cemented). Multivariate analysis on these factors revealed only differences between the achievements of surgeons and residents. The consequence of our conclusion that free hand cup positioning is an unreliable method is that attention now should be focussed on the results of a randomized trial comparing the different techniques of cup positioning.	[ "accuracy", "freehand cup positioning", "total hip arthroplasty" ]	[ "P", "P", "P" ]
Exp_Brain_Res-4-1-2335293	The latency for correcting a movement depends on the visual attribute that defines the target	Neurons in different cortical visual areas respond to different visual attributes with different latencies. How does this affect the on-line control of our actions? We studied hand movements directed toward targets that could be distinguished from other objects by luminance, size, orientation, color, shape or texture. In some trials, the target changed places with one of the other objects at the onset of the hand’s movement. We determined the latency for correcting the movement of the hand in the direction of the new target location. We show that subjects can correct their movements at short latency for all attributes, but that responses for the attributes color, form and texture (that are relevant for recognizing the object) are 50 ms slower than for the attributes luminance, orientation and size. This dichotomy corresponds to both to the distinction between magno-cellular and parvo-cellular pathways and to a dorsal–ventral distinction. The latency also differed systematically between subjects, independent of their reaction time. Introduction Neurons in different cortical visual areas respond to different visual attributes with different latencies (Livingstone and Hubel 1987, 1988; Schmolesky et al. 1998; Bullier 2001). How do such differences affect the way in which visual information guides our movements? Two main streams of visual processing have been identified within the brain; each specialized for processing a different kind of information (Trevarthen 1968; Mishkin et al. 1983; Goodale and Milner 1992; Ungerleider and Haxby 1994). A ventral stream in which information travels from V1 to the inferotemporal cortex is crucial for the identification of objects, whereas a dorsal stream passing through the posterior parietal region plays a major role in spatial vision. A widely accepted explanation for this distinction is that the dorsal stream is specialized for quickly processing spatial information to control action, whereas the slower, detailed visual analysis of other attributes that are important for recognizing objects takes place in the ventral pathway (DeYoe and Van Essen 1988; Goodale and Milner 1992; Tanne et al. 1995; Desmurget et al. 1999; Rossetti et al. 2000; Lee and van Donkelaar 2002). This view implies that not all information can be used to quickly adjust hand movements. We tested this prediction. One of the problems with comparing experiments that investigate whether various kinds of visual information can contribute to fast control of actions is that investigators use different experimental paradigms for different visual attributes, which themselves may give different results for a single attribute (e.g., compare Pisella et al. 1998; Schmidt 2002; Brenner and Smeets 2004; Cressman et al. 2006; White et al. 2006 for the role of color in guiding hand movements). It would therefore be very useful to systematically investigate the latency of responses to various visual attributes with a single paradigm. We therefore compared responses to changes in target location for hand movements directed toward targets defined by attributes that are normally relevant for goal-directed actions (luminance-contrast, size and orientation) with ones toward targets defined by attributes that are normally more relevant for object recognition (color, shape and texture). In all cases, there were three possible target positions. The target was at one of them. Other objects (references) occupied the other two positions. In half of the trials, the target switched locations with one of the references just after the hand started moving (Fig. 1a, b). Fig. 1Schematic overview of the experiment. a The participant holds a pen at the starting position when a target (dark square in this example) and two other objects (references, the brighter squares in this example) appeared. b In half of the trials, the target switches to one of the other locations just after movement onset. c Ten paths of one participant’s hand (P6) for a target that remained at the leftmost position in the “luminance (390%)” condition. d All ten paths of the same participant’s hand in the condition in which the target jumped from the leftmost to the rightmost position at movement onset An issue that cannot be ignored is how conspicuous the distinction between target and reference is within each attribute; i.e., how easily the target can be distinguished from the references. When comparing responses to targets defined by the same attribute, the response becomes faster if the relevant contrast is increased (Brenner and Smeets 2003). We therefore have to show that any differences in latency that we find cannot be simply caused by differences in conspicuousness. The question is how to evaluate conspicuousness independently from the response latency. In previous studies, this issue was resolved by equating a critical stimulus parameter (e.g. cone contrast, White et al. 2006), or by equating the stimuli in terms of an independent perceptual judgment (e.g., how conspicuous the targets are judged to be, Brenner and Smeets 2003). The first approach is not feasible in our current experiment because it is not at all clear how stimulus parameters for different attributes could be matched. The second approach is also problematic, because the visual processing for the perceptual judgment task might differ from that for the real experiment. We will therefore develop a new method to deal with differences in conspicuousness. This method, which will be developed in the results section, is based on examining responses to various targets that are all distinguished on the basis of the same attribute, but that differ in conspicuousness. We will show that beside leading to shorter latencies, more conspicuous targets also lead to shorter reaction times, higher velocities, and a steeper slope of the responses. We will then compare latencies for targets defined by different attributes when they give rise to comparable reaction times, velocities, and response slopes. Differences in latency after such matching cannot be caused by differences in conspicuousness. Materials and methods Participants Twelve subjects took part in this study. Three of them were the authors and the others were colleagues and friends who were unaware of the hypothesis that was being tested. All of them had normal or corrected to normal vision. Two of the participants were left-handed. Each participant made 120 pointing movements in each of 13 different sessions. Six of the participants also made 120 pointing movements in each of 7 additional sessions in which various luminance contrasts were used. Each participant performed the sessions in a different random order. The ethics committee of the Faculty of Human Movement Sciences approved the study. Apparatus and stimuli Participants sat behind an A2-sized graphic tablet (Digitizer II, Wacom Ltd, Tokyo, Japan) and viewed a projection surface via a semitransparent mirror that was placed above this graphic tablet. The images were back-projected from above the projection surface. The distance between the mirror and the projection surface was identical to that between the mirror and the surface of the tablet, so that the projected image appeared to be at the surface of the tablet. Lamps underneath the mirror ensured that the participants could see their hands. The resolution of the projected image was 1,024 by 768 pixels, with 1 pixel corresponding to about 0.5 mm. The position of the pen was determined every 5 ms (200 Hz). At the near edge of the screen was a starting point (a black, 1 cm diameter disk with a luminance of 2.3 cd/m2 as measured with a Minolta LS-110 luminance meter). At a distance of 25 cm from this point in the sagittal direction, three objects were projected next to each other on a white background (35 × 45 cm; 28.0 cd/m2). The distance between the centers of the objects was 3 cm (Fig. 1a). One of the three objects served as the target; this object differed from the other two objects (the references). We varied the attribute in which the target differed from the references (and the magnitude of this difference) between conditions (Fig. 2). Each condition was tested in a separate session. In the control condition, the reference objects were equal to the background so they were invisible: only the dark gray target square (1 cm sides, 5.1 cd/m2) was visible. Fig. 2Schematic representations of the stimuli in the 20 conditions (depicted with the target on the left). The luminance of both the target and the references is 5.1 cd/m² and they are both 1 cm2 gray squares unless otherwise mentioned (see “relevant details”). a The 13 conditions that were performed by all participants. Red lettering indicates that the responses depend on attributes that are normally needed to recognize objects, and blue lettering indicates that all items change. b The seven additional luminance conditions that were performed by six participants (numbers between brackets are the percentage that the references are brighter than the target) In the luminance (390%) condition, the target differed from the references in luminance only (for further details about this and subsequent conditions see Fig. 2). There were seven additional conditions with different luminance contrasts between the target and the references. These were used to determine how responses depend on how conspicuous the distinction is between the target and the references. In the three orientation conditions, the target and the references were rectangles (0.67 × 1.50 cm, 5.1 cd/m2). The target rectangle was orientated at an angle of 45°, 60° or 90° with respect to the parallel references. In the three size conditions the target squares had different sizes. The reference squares had surface areas of 1 cm2, as did all the objects in the other conditions. In the shape session the target was a circle, and the references were squares. In the color condition, the target was a green square (5.1 cd/m2) and the references were two red squares (5.1 cd/m2). In the texture condition, the target square was filled with ten 1-mm wide stripes, alternating between dark (2.3 cd/m2) and bright (7.9 cd/m2) ones. The average luminance of the stripes was equal to the luminance of the references (5.1 cd/m2), to ensure that participants had to analyze the texture to recognize the target. During pilot-experiments, some of the participants had the impression (especially in the orientation and the shape conditions) that they were not responding to a difference between the target and the references, but to the change at the new target position. To investigate whether this was really an issue we performed two control conditions in which we masked such effects by also changing aspects of the references when the target changed its position. In a masked orientation condition, the references both changed their orientation by 90° when the target jumped to a new location. In a masked shape condition, the reference squares increased in size at the same moment that the target circle changed position. Procedure Participants started each trial by placing the pen on the starting point. After a random time (about 2 s), the target and references appeared (Fig. 1a). The task was to reach the target with the pen and stop there as soon as possible. The sum of the movement time and the reaction time was to be minimized. Participants were shown which object was the target (and which the references) before they started each session. Within each session of 120 trials, the target appeared 40 times at each of the 3 positions. In half of the trials the target and the references remained at their initial position (unperturbed trials, Fig. 1c). In the other half of the trials the target changed its position to one of the two reference positions about 62 ms after the subject’s finger reached a velocity threshold of 0.02 m/s (perturbed trials, Fig. 1b). At the same moment the reference (that had been at that position) changed its position to where the target had been. Whenever the target changed position participants had to adjust their hand movement to reach the new target location (Fig. 1d). If a movement ended on the target within 1.2 s of the target first appearing, the participant was rewarded with auditory feedback. The 120 trials (three initial positions, change in position or not, two positions to which to change, ten replications) were presented in random order. Since different positions to which to change are irrelevant when there is no change, there were actually 20 trials in which the target remained at each initial position. Analysis Velocities were calculated for the interval between every two measurements by dividing the displacement of the tip of the pen by the 5 ms between the measurements. The beginning of the movement was defined as the first position after the tangential velocity reached 0.02 m/s. The end of the movement was defined as the first position after the tangential velocity fell below 0.02 m/s. To evaluate the corrections, we only used the lateral component of the velocity (parallel to the displacement of the target). In order to isolate the responses to changes in target position, we first synchronized all the measurements relative to the moment that the target changed—or would have changed—position. We then separately averaged the lateral velocity for each combination of initial and final target position. We defined the lateral velocity in the direction of the position change as being positive; when the position did not change, we defined the direction in which it would have changed as being positive. We characterize the response to a change in target position by the additional lateral velocity: the difference in the lateral component of the velocity between the trials in which the target changed from a certain initial position to a certain final position and the ones in which it remained at the same initial position. To obtain a single response per attribute for each participant we averaged the additional lateral velocity across the six combinations of initial and final target positions. We used these curves to determine the latencies of the responses. To estimate the latencies of the responses we approximated the initial responses by straight lines (dotted lines in Fig. 3). We used a robust method (one that is not too sensitive to the exact shape of the response near its onset) to approximate the initial response: we determined the amplitude of the peak in the average additional lateral velocity and then determined the points at which the additional lateral velocity reached 25 and 75% of this value. We consider a line through these two points as a good and robust approximation of the initial response. We considered the intersection of this line with a line at an additional lateral velocity of zero to give a reliable estimate of the latency of the reaction to the change in target position. We evaluated the influence of the conditions on the slopes of the above-mentioned lines and on their intersection points with the zero-velocity line (across and participants) using paired t-tests. Fig. 3Mean additional lateral velocity for each attribute as a function of the time after the perturbation. The mean additional lateral velocity is the mean difference between the lateral velocity on the trials in which the target did and did not change position. To estimate the latency of the response, a line through the points at 25 and 75% of the peak velocity (see dotted lines) was extrapolated to intersect the line representing an additional velocity of 0 (black horizontal line). The time at the point of intersection was considered to be the response latency. a Averages of six participants for the luminance contrast conditions. b Averages of 12 participants for the conditions in which they all participated The slope of the line through the 25 and 75% points of the average additional lateral velocity does not only depend on the intensity of the response on individual trials (lower acceleration results in a shallower slope). If the response does not always occur at the same time, averaging will result in an average response that has a lower peak velocity, shallower slope and longer duration than the responses in individual trials. To evaluate the shape of the response curve without the influence of variability in timing, we used a second way of synchronizing trials before determining the additional lateral velocity. To estimate the intensity of individual responses for each condition and participant, we synchronized the lateral velocity curves of all perturbed trials at the peak lateral velocity in the direction of the new target position (irrespective of when the target position changed), and produced an additional lateral velocity curve for each subject by averaging across replications and perturbation directions. We averaged these synchronized additional lateral velocity curves across positions and across participants to investigate whether there are systematic differences between the intensity of the responses for different attributes. We averaged the curves across conditions rather than participants to investigate whether there are systematic differences between the intensity of different participants’ responses. In order to determine whether the response was proportional to the size of the perturbation, we also averaged the curves separately for all pairs of positions in which the target jumped 3 cm, and all pairs of positions in which the target jumped 6 cm. Results Overview of responses On average, it took participants 365 ms to react to the appearance of the target (and the two references). The average movement time (time until participants stopped their movement on the target) was another 275 ms. Figure 1 shows one participant’s paths in one condition with the target initially positioned on the left. When the target stayed at the initial target position (Fig. 1c), the hand moved to the target along a slightly curved path (note that the lateral deviations are exaggerated due to the unequal scaling of the figure). When the target jumped at the onset of the movement (Fig. 1d), participants obviously initially moved along similar paths to those seen in Fig. 1c, but approximately half way to the target the path curved toward the new target position. Figure 3 shows the additional lateral velocity for the various conditions (summarized in Fig. 2), averaged over all subjects. Each curve represents the difference in lateral velocity between trials in which the target did and did not change position (see “materials and methods” for details). The dotted lines are drawn through the points at which the mean additional lateral velocity curves reached 25 and 75% of their peak values. We consider the points at which these lines intersected the line at an additional lateral velocity of zero as the latencies of the responses. We used targets that only differed in luminance from the references to evaluate effects of conspicuousness (for half of our subjects). We found that lower luminance contrast (a less conspicuous target) results in both a longer latency and a shallower slope (Fig. 3a). The conclusion that we can draw from the main conditions depicted in Fig. 3b, is that both the timing and the slopes of the curves differ between conditions. Figure 3 shows an analysis over responses averaged over subjects. The rest of the analysis is done on the basis of determining slopes and latencies for each subject (and condition) separately. Dealing with conspicuousness It is clear from Fig. 3a that the latencies are larger for less conspicuous targets. In order to compare the latencies of responses to targets defined by different attributes, we have to take into account differences in conspicuousness. Figure 4a, b and c shows how not only the latency of the response varies with conspicuousness, but also the reaction time of the initial movements toward less conspicuous targets are longer (Fig. 4a), the slope of the response to a perturbation is shallower (Fig. 4b), and both the peak velocity of the initial movement toward the target and that of the correction are slower (Fig. 4c). All three plots yield significant correlations (P < 0.05), and the data are distributed along the fitted line with deviations that can be expected on the basis of the error bars. Fig. 4Comparing the responses across conditions and subjects. a–c Relationship between various movement parameters for the luminance conditions. The latency and slope refer to the correction (Fig. 3). Points are averages across subjects with standard errors. Dashed lines are linear fits through these points. d–f Similar graphs for each of the conditions in which all subjects participated. Symbols in red are the conditions based on attributes for object recognition; symbols in blue are for conditions in which the change is masked by changing all items. g–i Data for individual subjects, averaged across conditions. The brightness of the symbol indicates the average response latency of the subject (bright is short latency) Since several measures co-vary with the conspicuousness (i.e. the luminance contrast) of the target, we can use these measures to evaluate whether latency differences can be due to differences in conspicuousness. Except for the latency (the variable we are interested in), the slope is the variable that varies most clearly with conspicuousness. This is therefore the best measure to use to evaluate whether effects of condition on latency are caused by a difference in conspicuousness, followed by the reaction time (Fig. 4a) and response velocity (Fig. 4c). The correlation of the conspicuousness with the response velocity (Fig. 4c) is less clear than its correlation with the slope (Fig. 4b). This suggests that a shallower slope of the average response is not only caused by the response being less vigorous, but is also largely caused by more variability in the latency. To verify the effect of variability in latency on the slope of the response, we removed the effect of variability in the latency by averaging the responses after synchronizing them at the moment of peak lateral velocity. The responses resulting from this averaging method have almost the same slope for all conditions: 282 m/s2 with a coefficient of variation of 0.13 (see Fig. 5a). The slopes of the mean additional lateral velocity curves (Fig. 3) are not only much shallower (on average 187 m/s2), but are also much more variable between conditions (coefficient of variation 0.28). We conclude that the differences between the slopes of the mean additional lateral velocity curves are mainly caused by differences in the variability in timing between trials (rather than by differences in the intensity of the response). Thus a more conspicuous target results in less variable (Fig. 4b) as well as more vigorous (Fig. 4c) responses. Fig. 5Magnitudes of the responses as determined by averaging all trials after synchronizing at their peak lateral velocity. a Each curve indicates a condition; color coding as in Fig. 4. The similarity between the curves indicates that the intensity of the response did not differ between conditions. b Responses of subjects. The brightness of the curve indicates the subject’s average latency (bright is short latency). c Comparing the responses for the two jump amplitudes. The additional lateral velocity for all trials in which the target jumped 6 cm was twice as large as that for the trials in which the target jumped 3 cm Different attributes, different latencies? Figure 4d and e shows the same relationships as Fig. 4a, b for the main experimental conditions (Fig. 2a). The dashed lines reproduce the relationship that we found for the luminance conditions (Fig. 4a, b). Participants responded to the various size and orientation conditions with a relationship between latency and conspicuousness that was similar to that for the responses to luminance-defined targets. The targets defined by the attributes color, shape and texture were not less conspicuous than the other targets (reaction times and slopes are similar), but subjects take exceptionally long to react to changes in their position (t-test, P < 0.05). These responses were about 50 ms longer than the ones indicated by the dashed lines. The movements toward the targets defined by color, shape, or texture had similar peak velocities, both of the movements themselves and of the corrections, as the movements toward other targets (Fig. 4f). For two attributes (orientation and shape) we constructed a “masked” condition, in which not only the target changed location, but also the references changed. This prevents subjects from using the change itself as an indicator of the new target location. The latency for these two conditions is 210 ms, which is longer than for the two corresponding original conditions (paired t-test, P < 0.05). The masking decreased the velocity of the movements and of the corrections (Fig. 4f), but did not change the conspicuousness of the targets, either in terms of the reaction time (Fig. 4d) or of the slope of the average response (Fig. 4e). So the increase in latency due to the masking is not caused by masking making the targets less conspicuous, but by requiring additional processing, presumably for a renewed selection of the target. Due to the blocked design, subjects could counteract the effects of their later responses by slowing down their movements (low values for the masked targets in Fig. 4f). Subjects and amplitudes Before discussing how to interpret the results, we want to answer two additional questions. The first is, how does the variability between participants contribute to the results? To find this out, we plotted the data averaged over the conditions for each of the participants separately (Figs. 4g–i, 5b). We coded subjects according to the latency of their adjustments: bright curves and symbols correspond to subjects with short latencies. The intensity of the response differed much more between subjects (SD = 91 m/s2, Fig. 5b) than between conditions (SD = 37 m/s2, Fig. 5a), with no apparent correlation between the peak velocity and latency (Fig. 4i: brighter symbols are not concentrated in one region of the figure). All subjects show more or less the same pattern of latencies as we found for the average data (not shown). This was so for the different attributes as well as for the different levels of conspicuity. There is quite some variability between the subjects, but subjects that are slow in one condition are generally slow in the other conditions as well. The result is that subjects differ systematically in their latency, without any correlation with their reaction time (Fig. 4g) or with the slope of their response (Fig. 4h). This performance difference cannot easily be explained by factors such as age, visual acuity or body size. The second question is whether the similarity of the curves in Fig. 5a means that subjects have a pre-programmed response to a perturbation that is used irrespective of the precise visual information. To answer this question, we compared the responses for the two amplitudes of the perturbation. When the target jumped 6 cm from its initial location (from one extreme position to the other), participants responded about twice as strongly as when the target jumped 3 cm (Fig. 5c). Thus, the response is proportional to the position change, and therefore clearly tailored to the visual information rather than that pre-programmed and triggered by the change. Discussion By considering the relationship between response latency and measures of conspicuousness (reaction time and variability in the latency), we can conclude that participants reacted with a similar latency to differences in orientation and size than to corresponding differences in luminance. The attributes color, shape and texture gave rise to approximately 50 ms longer latencies than did differences in luminance, orientation and size that were equally conspicuous. Note that these latencies are still 150 ms faster than the original reaction times. We found a similar order of fast and more slowly processed attributes across participants, which supports our idea that we can use this method to measure systematic differences between attributes in terms of the time that it takes to process them, even if participants respond differently (see different velocity profiles in Fig. 5b). An important question in this kind of research is whether the apparent response to a particular attribute could actually be a response to a small difference in another attribute. For instance, in the color condition, participants could respond to small differences in luminance instead of to differences in color between the target and the references, since the luminance of the target and the references was not equated for individual participants. They clearly do not, because if they had done so, the response would be similar to that for one of the small luminance contrast conditions, which is clearly not the case (the slope of the response is much steeper for color). As mentioned earlier, it takes about as long to respond to changes in position defined by target size as to those defined by target luminance. We cannot reject the possibility that the reactions to the difference in surface area between the target and the references in the size condition were actually responses to the average luminance within an area larger than that of the target itself. To be sure that this was not the case we would have had to vary the luminance of the targets and references, which we did not do in the present study. Desmurget et al. (1999), showed that applying transcranial magnetic stimulation over the left posterior parietal cortex disrupted corrections of reaching movements of the right hand after a perturbation in target location in four of their five subjects, thereby demonstrating that the dorsal pathway is necessary for fast corrections of the hand. The fact that participants are able to respond very quickly to changes in target location when the target is defined by the attributes orientation, size and luminance (but not color, texture or shape) is therefore an indication for dorsal processing of orientation, size and luminance (but not of color, texture and shape). This division of attributes is consistent with the distinction between pathways for “what” and “where” (Mishkin et al. 1983; Ungerleider and Haxby 1994). The fact that subjects were still able to correct their movements to targets defined by color, texture or shape (albeit at a longer delay) implies that these attributes also reach the parietal cortex, although at a longer latency. This is in line with studies using electrophysiology (Murata et al. 2000; Sakata et al. 2005), positron emission tomography (Faillenot et al. 1997; Vidnyanszky et al. 2000) and functional magnetic resonance imaging (Oliver and Thompson-Schill 2003) that all suggest that the parietal cortex is involved in at least some aspects of discriminating shape. These studies were, however, not concerned with the exact timing of the activity. Why are responses for targets defined by shape relatively slow? Assuming that all the responses that we measured are indeed mediated by the posterior parietal cortex, the dichotomy in latencies could be explained in two ways. It could be that the parietal cortex mediates the responses to some targets after initial processing of the attributes that are relevant for the “what” of the object (like color, shape and texture (Mishkin et al. 1983) in the ventral pathway. This indirect pathway is likely to result in longer processing times than the direct dorsal pathway for the “where” attributes. Alternatively, the dichotomy may arise earlier, with the distinction between the magnocellular retinogeniculate pathway that is faster and more sensitive to luminance contrast, but is color-blind and has a low spatial resolution, and the parvocellular pathway that is color selective and has a higher spatial resolution, but is slower and less sensitive to luminance contrast (Livingstone and Hubel 1987, 1988). This distinction corresponds nicely with the dichotomy in the latencies that we found. Assuming that both the magnocellular and the parvocellular pathways provide input to the dorsal pathway could therefore account for the different latencies that we found for different attributes (Brenner and Smeets 2003), without requiring input from the ventral pathway. Although we cannot decide between these two lines of explanation, the two masked conditions indicate what is essential for the very fast responses. Being able to identify the target on the basis of “where” attributes (size, orientation, luminance) is necessary for fast responses. However this is not sufficient, because the latency of responses to the change in location of a target defined by orientation increases by more than 50 ms if the references change their orientation. Masking changes prevents responses to the location of the change. Presumably, the target has to be found (identified) again, which requires ventral processing (according to the first explanation). The conclusion that identifying targets by their color takes 50 ms longer than doing so on the basis of luminance seems to be in conflict with earlier results where we reported that subjects could respond within 120 ms to color information (Brenner and Smeets 2004). In that study, we compared two conditions: one with one reference object of a different color and one with no reference object (similar to the control condition in the present experiment). The main difference between the responses in these conditions was the larger variation in latency for the condition with one reference object (a more shallow slope). It seems as if the distribution of responses consisted of a small peak at short latency, and a larger peak 50 ms later (Fig. 3 in Brenner and Smeets 2004). Perhaps, the few very fast responses in the condition with the reference object were related to the very simple design of that experiment: the position to which the target could change was known in advance, so detecting the change in color at the original target location was enough to start an adequate correction. In the present experiment, the location of the new target position had to be determined on the basis of color information. That directly responding to a transient can lead to exceptional fast responses is evident from the fact that masking the jump of the oriented target by orientation changes of the references caused an increase in latency of 50 ms. A last issue to discuss is the variability between subjects. Some subjects responded consistently later to the perturbations than others (Fig. 4g, h). The difference between the slowest and fastest subject was 70 ms; larger than the difference between the responses to targets defined by color and luminance. Such a large variation is in line with the remarkable finding that stimulating the posterior parietal cortex directly after movement onset disrupts adjustments in some subjects, but not in others (Desmurget et al. 1999). We can draw two conclusions from these variations in latency. The first is that since none of the subjects had particular complaints of clumsiness, the exact value of the latency is apparently not critical for performing every day tasks. The second is that in situations in which short latencies are important, such as when playing table tennis or boxing, some participants may perform better than others could ever hope to, due to differences between their minimal visuomotor latencies. This latter implication (and in particular whether a shorter latency can be achieved by training) is open for experimental verification.	[ "hand movement", "color", "shape", "visual processing", "dorsal stream", "target perturbation", "double step" ]	[ "P", "P", "P", "P", "P", "R", "U" ]
J_Struct_Funct_Genomics-3-1-2140095	The scientific impact of the Structural Genomics Consortium: a protein family and ligand-centered approach to medically-relevant human proteins	As many of the structural genomics centers have ended their first phase of operation, it is a good point to evaluate the scientific impact of this endeavour. The Structural Genomics Consortium (SGC), operating from three centers across the Atlantic, investigates human proteins involved in disease processes and proteins from Plasmodium falciparum and related organisms. We present here some of the scientific output of the Oxford node of the SGC, where the target areas include protein kinases, phosphatases, oxidoreductases and other metabolic enzymes, as well as signal transduction proteins. The SGC has aimed to achieve extensive coverage of human gene families with a focus on protein–ligand interactions. The methods employed for effective protein expression, crystallization and structure determination by X-ray crystallography are summarized. In addition to the cumulative impact of accelerated delivery of protein structures, we demonstrate how family coverage, generic screening methodology, and the availability of abundant purified protein samples, allow a level of discovery that is difficult to achieve otherwise. The contribution of NMR to structure determination and protein characterization is discussed. To make this information available to a wide scientific audience, a new tool for disseminating annotated structural information was created that also represents an interactive platform allowing for a continuous update of the annotation by the scientific community. Introduction The long-term goal of structural genomics (SG) has been ambitiously defined as “to make three-dimensional atomic level structures of most proteins easily obtainable from knowledge of their corresponding DNA sequences” (http://www.nigms.nih.gov/Initiatives/PSI.htm). Long before this goal is achieved, the multiple specialized SG projects are expected to have a significant impact on many aspects of the biological sciences. The most readily apparent contribution of SG is the rapid expansion in the number of available protein structures, derived at a reduced cost because of the efficiency of specialized centers. Proper target selection is critical to ensure that the structures solved by SG centers are indeed valuable to the research and industrial community, either because of the intrinsic interest of the proteins investigated, or because of the improved mapping of the protein structure universe, providing homologous structural models. A second important contribution of SG projects for the scientific community is the development of methods for efficient protein production and structure determination, which could be adopted in smaller research laboratories to improve productivity. Other scientific deliverables of structural genomics derive from the scale and nature of the operations, and include comparative studies on members of protein families, identifying determinants of specificity, deriving general rules, and improving the capability to predict protein structure and function from gene sequences. The Structural Genomics Consortium (SGC), operating in the Universities of Oxford and Toronto and the Karolinska Institute, was initiated in 2003 to address needs of industrial and academic pharmaceutical research. The SGC investigates human and apicomplexan proteins; the targets are selected based on their potential as drug targets or involvement in disease processes. Technologically, the SGC focuses on interaction of proteins with small molecules (ligands, inhibitors, substrates and co-factors), and on coverage of protein families. This report provides several examples of the impact of research undertaken at the Oxford node of the SGC, including methodology for high-throughput structure determination, generic means for ligand screening, selected examples of insight from specific structures, insights from family coverage, and the possibilities resulting from the availability of large numbers of purified protein samples. The other SGC nodes share the core technologies but investigate non-overlapping target areas. Finally, the scientific impact depends on dissemination of structural data. We describe a new platform for distribution of annotated protein structures, which aims at making this data more meaningful to an audience beyond the usual users of the PDB. Methodology Protein production Method adaptation and development for structural genomics involved a change of mindset, no less than developments in instrumentation, chemistry and computer software. Industrialization of protein production––applied to a huge variety of proteins with very divergent chemical properties––is not straightforward. Yet, extensive work in several SG centres have led to a convergence to core procedures, which are widely applicable, and often sufficient to generate purified proteins, crystals and structures (Table 1). Where the core protocol fails, additional steps (e.g., further purification, crystal optimization), or alternative methods (e.g., different cloning vectors) are applied. Table 1Core protocols employed at the SGC1. Source of DNA1. Sequence-verified cDNA clone collections.2. Synthetic DNA.3. RT-PCR, site-directed mutagenesis.4. Genomic (microbial).2. CloningLigation-independent cloning.Recombinase-based cloning (e.g., Gateway, InFusion).3. Expression vectors and hostsT7 promoters, controlled by Lac repressor.Bacterial vectorsN-terminal hexahistidine tag, cleavable by specific proteases (TEV, Thrombin, C3).Host strains based on BL21(DE3), often expressing rare-codon tRNAs or chaperone proteins.4. Eukaryotic expressionBacoluvirus-infected insect cells.5. Protein expressionRich media, grow at 37°C to mid-log, then induce at low temperature with IPTG.OR: Similar protocol using minimal medium for Selenomethionine or isotopic labelling.6. PurificationTwo-step purification: Affinity chromatography, Gel filtration, all in high-salt buffers (0.5 M NaCl). Optional: tag cleavage and re-purification.7. Ligand and buffer screeningThermal denaturation assays are used to screen purified proteins against 1–103 small molecules and several buffer compositions, to identify stabilizing conditions and potential ligands.8. CrystalliationInitial coarse screens (2–4 × 96 conditions; 3 protein concentrations each). Vapour diffusion, sitting drops, imaged by robots but scoring done by humans.Include ligands identified from screening or biochemical knowledge to promote crystallization.Follow-up screens and crystal optimization.9. Data collection and structure determinationManual or robotic screening of crystals for diffraction properties; data collection in rotating anode or synchrotron sources.Phasing: Molecular replacement (95%), experimental phasing using SeMet derivatives, and MIR. Several features of this protocol have been optimized to capture a large portion of target proteins. Gene clones have been predominantly obtained from public and commercial cDNA libraries. However, gene synthesis may become the method of choice, allowing to optimize codon frequency, restriction sites, and mRNA structure and to introduce site-directed mutations. Ligation-independent cloning is a generic, high-throughput process that can be uniformly applied regardless of the target gene or the cloning vector. Short N-terminal fusion tags, including a hexahistidine sequence and a specific protease cleavage site, are almost universally used. It has been widely documented, that larger fusion tags (e.g., GST, thioredoxin, MBP) can enhance solubility of proteins that are not soluble when expressed with a short peptide tag. However, such fusion proteins have not been widely used in the SGC, since removal of the tag often leads to loss of solubility. The standard purification protocol is designed to be widely applicable, and experience has shown that it results in effective purification of a large fraction of proteins solubly expressed in E. coli. A protein presented for crystallization must be homogeneous in composition, post-translational modification and oligomeric state; the presence of protein aggregates may be especially detrimental to subsequent crystallization. Affinity purification of highly-expressed proteins eliminates most other proteins, while gel filtration effectively separates different oligomeric forms of the protein and removes protein aggregates, which may otherwise promote irreversible aggregation of the protein preparation. The use of high salt concentration (typically, 0.5 M NaCl) throughout the purification process seems to reduce protein aggregation and non-specific binding of protein contaminants. Tag cleavage followed by another passage through the affinity column provides a further generic and highly effective purification step, which removes other proteins that bind adventitiously to the first affinity column. The generic purification procedure has provided in the majority of cases protein of sufficient purity to achieve crystallization. In most other cases, the generic procedure could be followed by polishing and protein modification steps to achieve homogeneous preparations. The greatest barrier to production of human proteins in bacteria is recovery of soluble protein. Less than 15% of protein targets yielded detectable levels of soluble protein when tested as full-length constructs in the SGC, while more than 80% were expressed as insoluble aggregates. The key to achieving higher success rates has been the parallel production of large numbers of truncated constructs, often containing a compact protein domain. Construct design is initially based on domain boundary analysis, using a number of bioinformatic tools; 3–4 endpoints are designated around each of the predicted termini of the domain, resulting in 9–16 constructs. We have consistently found that this approach results in a 4-fold increase in the number of targets that can be produced as soluble proteins; a similar impact has been seen on the production of diffracting crystals, which can be dramatically affected by minute changes in protein termini. Although not rigorously tested, it is presumed that a protein construct that is inherently well-behaved (little tendency to aggregate or denature) will be less dependent on specialized conditions for expression and purification, and may crystallize in a wider range of conditions. Crystallization, crystal screening and data collection For successful crystallization of a given target, the SGC’s phase I operation appears to have confirmed that the most important driver for success is to explore protein diversity at the crystallization stage. One major form of variation was discussed above, namely testing multiple constructs of the target. Equally effective has been setting up co-crystallization with multiple ligands, along with varying protein concentration in the primary crystallization screens. At the same time, it appears not to be vital to explore chemical space extensively for any given protein preparation; instead, the primary goal of the initial (coarse) screen can be to identify which preparations are “crystallizable”, and a limited set of coarse screen conditions (∼200) generally seems sufficient. Practically, this requires only two 96-well crystallization plates, and by setting up three drops per condition, at different protein-well ratios (in Greiner 3-drop plates), the protein concentration is simultaneously varied. The conditions themselves are derived from those found to be most successful in other high-throughput initiatives [1–3], although according to this “crystallizability” philosophy, the exact composition is probably not important. Naturally, coarse screens do not always yield high-quality crystals that can produce a dataset; however, the SGC operation does not rely on these crystals showing up in coarse screens, and a good optimization infrastructure is in place. In practice, this diversity exploration leads to large numbers of parallel crystallization experiments, presenting a logistical challenge which, at this scale, can only be met with an efficient robotics and IT infrastructure. For the automation, the SGC has been able to exploit the devices developed on the back of the first wave of structural genomics initiatives, and our investment has been less in developing the machines, than in integrating them and implementing experimental best practices. Particular examples: by minimizing sample requirements with nanolitre crystallization, the available protein can be used in more experiments. The large numbers of drops thereby produced (1.5 million/year) would be practically impossible to view by eye under the microscope, whereas automatic drop imaging on a fixed schedule allows images to be reviewed at leisure at the desk. Automation has also played an important role in crystal characterization. An automatic sample changer has been used for initial characterization of diffraction quality of a vast number of crystals. This allows to rank the crystals for more careful data collection, especially at the synchrotron, and to direct further efforts at crystal optimization. A significant saver of upstream efforts has been to exploit each crystal’s diffraction as efficiently as possible, even those traditionally considered to be marginal or problematic. Marginal diffractors would include crystals that are “very small” (<40 μm in longest dimension), twinned, or have streaky or anisotropic diffraction. The latter cases generally require the undivided attention of experienced crystallographers. Small crystals require an excellent X-ray beam: the PXII beamline of the Swiss Light Source synchrotron provides a beam which is reliably small but also well-aligned and very stable. Most efficient use of the beamline relied on pre-screening all crystals at the laboratory source for thorough work prioritization; real-time data processing during data collection; and close attention to radiation damage of crystals. It has been crucial to have experienced crystallographers on site. Adherence to these good practices has been highly productive: of datasets collected on 24-hour trips to SLS, 66% were used for final structures, while 90% of all depositions relied on synchrotron data. The ability to extract useful data from marginal crystals has been especially productive in combination with the protein/ligand diversity approach of the SGC, as a significant fraction of structures (>50%) could be derived from crystals emerging from the primary screens, saving the need for further optimization. Phasing and structure solution Due to the family-based approach, for most SGC targets a homologous structure is already known, and most structures (>95%) can be phased by molecular replacement (MR). While this saves significant experimental efforts upstream compared to experimental phasing, by eliminating the need for selenomethionine-derived protein or heavy atom soaks, we find this does not actually save time overall, because starting phases from MR are heavily phase biased. Removing the bias has required many iterations of careful and incremental model building and refinement by experienced crystallographers who can see the danger signs of a poorly-refined model, and know how to deal with it [4, 5]. The final step, namely finalizing and depositing the model, is in fact a frequent stalling point, not only in high-throughput contexts. The reason is that the final model is not merely a result that can be trivially read off a few measurements, but instead is an interpretation of often rather noisy data, with a lot of detail that is easy to miss, where individual errors influence the clarity in all areas. Moreover, poor model definition affects biologically interesting parts of a structure, and interpreting it becomes a matter of judgment and using in orthogonal information. Indeed, the “final” model is as much scientific hypothesis as result, and depositing the model means signing off on the hypothesis––which is why it has traditionally been a bottleneck in structural genomics efforts. The SGC has used a peer proofreading system combined with strict timelines to counteract the problem: before deposition, the structure is reviewed by another crystallographer for errors or alternative interpretations, and comments passed back to the original refiner. The intention is threefold: First, to introduce quality control on the final output. Second, the refiner does not feel compelled to spend excessive time on the model to flush out the final errors, since she knows it will be checked. Third, by mixing up refiners and proofreaders, over time this should lead to common interpretations of marginal modeling decisions. The timelines depend on situation and difficulty, but typically allow two weeks for refinement, a day for proofreading, and two further days for deposition. This approach has made it possible to deposit novel structures at a considerable rate (6 each month from a team of 6 dedicated and 4–5 occasional crystallographers) without compromising quality. Information infrastructure An efficient laboratory information management system (LIMS) has been vital to manage not only target tracking, but also capturing and integrating where possible information generated from robotics, as well as capturing human assessments of experimental outcomes, where these could be entered via a client (e.g., scoring of crystallization images). Fortuitously, the solution we settled on, BeeHive from Molsoft (http://www.molsoft.com/beehive.html), is in essence an extremely intuitive database query tool that enables even inexperienced users to extract information relevant to their current work––including the simplification of data entry. This is a weak point of many LIMS solutions, whose focus often evolves around data entry but have very inflexible retrieval mechanisms. This has proved to be a powerful means of communication between all persons involved in a project, allowing immediate and error-free retrieval of “hard” information (e.g., protein sequence, ligand and buffer conditions and project history), as well as evaluation and prioritization of crystals and of concurrent projects. Protein characterization and ligand screening One of the major challenges in structural genomics is identifying the function and evaluating the functional integrity of the proteins. Examining the physical state of a protein––by methods such as analytical ultracentrifugation, chromatography or dynamic light scattering––is valuable in assessing the prospects for crystallization. In contrast, specific activity assays need to be tailored for each protein class, and may be impractical or impossible when the activity of the protein is not known. We have implemented a generic screen, based on the increase in thermal stability of a protein upon ligand binding. The fluorescent readout is based on monitoring of protein unfolding using a hydrophobicity-sensing dye. Differential Scanning Fluorimetry (DSF) assays [6–9] are ideal for screening a large number of compounds for binding to each target protein. Significantly, the shift in Tm (the unfolding transition midpoint) measured by this method is comparable to measurements obtained by differential scanning calorimetry (DSC), the well-established standard method for thermal shift measurements. In selected cases, a direct correlation between Tm shift and binding constants has been observed [8, 10]. Several advantages have been derived from this capability: First, the identification of relatively strong interacting molecules out of several hundreds of candidates. As detailed below, the compounds discovered in this manner are then included in crystallization experiments; in many cases, only protein–ligand complexes yielded diffracting crystals. Secondly, the reactivity profiles provide data on binding selectivity of the protein active site, which is the most crucial information for drug design; we have often followed up the results from ligand screens by analyzing the structures of several protein–ligand complexes. In parallel, the properties of the protein–ligand interactions are studied by biophysical methods and by enzyme inhibition studies. Third, such screens have allowed us to identify ligands or substrates of proteins with unknown function (sometimes termed “de-orphanizing”). Finally, DSF-based screens can be expanded to explore other conditions, such as buffer composition that enhance the stability of a protein. These conditions may then be introduced to improve the outcome of protein purification and crystallization [8]. The limited scale of protein production and other limitations on resources do not allow a full-scale screen as done in the pharmaceutical industry (105 compounds). Rather, we have assembled smaller family-specific compound libraries (10–103 compounds each), which can reasonably be tested against available amounts of protein (∼200 μg for 100 assays). The compound libraries are based on the scientific and patent literature; the chemical structure of prospective compounds is used to search an in-house compilation of vendor databases to identify potential sources. Acquisition of desired compounds is not trivial: not all published compounds, even those appearing in vendor catalogues, are actually available when required; alternative vendors, or collaborative sources may then be accessed. With continuous updating based on current literature and our own experimental results, these libraries have allowed to derive binding profiles and new insights on ligand specificity. SGC target and biology area selection: relevance for the treatment of human diseases For any structural genomic organisation target selection is an important consideration as it can have a major impact on the procedures that are implemented during the process of structure determination. There are a number of approaches applied by different structural genomics projects to select targets for structural analysis such as blanket coverage of an organism’s genome, targets with potential novel folds, percentage cut off based on sequence identity or total coverage of selected protein families. The SGC has opted for the family-based approach with an emphasis on protein families whose members are important in human health, disease and are potentially druggable. From our point of view, the main advantages of this approach are 2-fold. Firstly, the methods and procedures identified for one family member can be applied to another family member improving everything from expression, solubility, stability, and purification, to crystallisation and structure determination. Secondly, analysis of the structures from all family members can reveal additional significant information such as ligand binding site specificity, conformational dynamics, understanding of aberrant behaviour of specific family members or the converse revealing common structural properties within all family members. The availability of high resolution structures constitutes the foundation for structure-guided drug discovery projects. In recent years SG has significantly increased the number of human protein structures available for structure-based design projects [11]. In particular, protein family focused efforts originating from high-throughput structural biology projects have contributed to the structural description of a number of members from human protein families and thus provided valuable structural and chemical information for the design of bioactive compounds. In addition, established expression and crystallization conditions have been used to generate essential reagents, methodologies and technologies which have facilitated research projects in academia and drug discovery programs in industry. The SGC has focused on providing protein structures to support drug development and understanding of the structural determinants for human disease. Of 160 unique targets deposited by the SGC (in phase 1), clear disease relevance has been established for 70% and a further 18% are likely to be involved in at least one disease. This pattern holds true for all the human protein families the SGC is working on. The following sections provide an overview of the three distinct biological areas selected at the Oxford site of the SGC. Biology area I: Structural Genomics of human metabolic enzymes Selection of metabolic enzymes as biological target area at the SGC was based on two distinct features: they are fundamentally involved in a multitude of human diseases, including cardiovascular, metabolic diseases or cancer, and in addition several enzymes constitute possible drug targets. Emphasis has been given to certain metabolic enzyme families such as oxidoreductases (mostly short-chain dehydrogenases/reductases (SDR), medium-chain dehydrogenases/reductases (MDR), long-chain dehydrogenases/reductases, aldehyde dehydrogenases (ALDH), aldo keto reductases (AKR) and 2′oxoglutarate dependent oxygenases (2OGs). In addition, pathways of importance, e.g., in lipid or amino acid metabolism were selected with a distribution of about 1:1 between oxidoreductases and other metabolic enzymes. The target list comprises about 300 metabolic enzymes, and after three years of operation, >60 unique novel structures have been solved. Three points of importance are highlighted in this review: structural characterization of enzymes shown to be causative of metabolic inherited diseases, structure determination of drug discovery targets in metabolic diseases such as metabolic syndrome or osteoporosis, and structure-guided “de-orphanization” of insufficiently characterized human gene products or even entire pathways. Structural basis of inherited metabolic diseases Genetic defects in enzymes involved in metabolic pathways such as amino acid or lipid catabolism are causative of a whole spectrum of symptoms, including dysmorphologies, mental retardation, neuropathies or life threatening situations like fasting induced hypoglycemia [12, 13]. Understanding of molecular causes and possible interventions of inherited metabolic diseases requires besides biochemical and clinical management a structural template for explanation of mutational effects. Thus far the focus has been to a large extent on oxidoreductases in the area of metabolic diseases. Associated disorders comprise electron transfer reactions for energy production (e.g., mitochondrial myopathies), oxidative and reductive roles in the metabolism of amino acids (e.g., hyperprolinemia or branched-chain hydroxyacyl CoA dehydrogenase defects), fatty acids (e.g., inborn errors in α- and β-oxidation of short-, medium- or long-chain fatty acid metabolites), cofactors (e.g., phenylketonuria type 2), hormones (e.g., male pseudohermaphroditism or adrenal hyperplasia), mediators (e.g., congestive heart failure) and lipids (e.g., inborn errors in cholesterol synthesis, CHILD syndrome, Smitz-Opitz Laemmli syndrome as examples). The impact of the structural approach is illustrated by the successful structure determination of phytanoyl-CoA hydroxylase [14], the major molecular cause of Refsum disease, a peroxisomal disorder with severe neurological symptoms. The structure provides a framework to interpret the majority of the disease causing polymorphic alleles, and we were able to map those to changes in the active site, around the Fe2+ and 2-oxoglutarate binding sites in this 2OG enzyme [14]. Metabolic enzymes as drug targets Oxidoreductions at specific positions of lipid hormones such as steroids selectively alter nuclear receptor binding properties. Therefore, inhibition of dehydrogenases/reductases carrying out these reactions selectively influences cellular hormone levels and transcriptional responses. This concept has recently found great attention with the development of specific inhibitors against 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) as a novel drug target in diabetes and obesity [15–18]. Similar drug development efforts are underway regulating androgen or estrogen levels through specific modulation of distinct hydroxysteroid dehydrogenases (17β- and 3α-HSDs in cancer, inflammation, osteoporosis, ageing, and autoimmune diseases). We determined the structure of human 11β-HSD1 in complex with a clinically relevant inhibitor, carbenoxolone (Wu et al., unpublished) and have provided a platform for drug development efforts. Other hydroxysteroid dehydrogenase structures comprise 17β-HSDs such as types 4, 8, 10, 11 and a novel type 14 (see below), necessary for determination of off-target activities of compounds directed against type 1 and 3 17β-HSDs. Other targets of pharmaceutical relevance successfully pursued are farnesyl diphosphate synthase (FDPS) and geranylgeranyl diphosphate synthase (GGPS), which are critical in synthesis of isoprenoids necessary for covalent modification of GTPases involved in cell signalling and survival. Crystal structures of FDPS complexed with nitrogen-containing bisphosphonates currently used for osteoporosis therapy allowed a molecular mechanism of action to be postulated for these drugs [19] (Fig. 1). Furthermore, several prokaryotic and parasitic dehydrogenases have been identified as novel targets for antibiotic and antiparasite drug development, and thus allow synchronization with the SGC Toronto efforts, where an apicomplexan/protozoan SG program has been established. Thus, structure determination of related human enzymes will facilitate structure aided drug design and allow virtual and focused screening efforts in this emerging disease area. Fig. 1Bisphosphonate binding to human farnesyl diphosphate synthase. Electron density is shown in green around the clinically used inhibitor risedronate Deorphanization of metabolic enzymes and pathways A significant proportion of the metabolic enzymes targeted were at the time of structure determination devoid of assigned activity or function. High throughput protein production, structure determination and functional characterization allowed “deorphanization” of unknown enzymes. We employed ligand screening, enzyme activity assays, expression and subcellular localization data, as well as structure determination combined with docking analysis to describe novel human enzymes. In the absence of co-crystal structures, interpretation of results from biochemical assays and compound screening was rationalized by in silico docking of potential ligands into the active site of the orphan structures. Analysis of the different docking poses was correlated with experimental results, allowing direct visualization of the putative protein–ligand complex. In this manner we determined a novel 17β-HSD14 [20], possibly involved in cancer, and a novel type-2 R-hydroxybutyrate dehydrogenase, involved in ketone body utilization [21]. Further emphasis was given on novel pathways such as mitochondrial fatty acid synthesis. This recently discovered pathway is important in the synthesis of lipoic acid, essential for mitochondrial function. Thus far we have determined three distinct enzymes of this metabolic route, namely the malonyl transferase (2c2n), ketoacyl synthase (2c9h) and the enoyl-ACP reductase (1zsy). These structures represent the only higher eukaryotic structures thus far available for this pathway. The data will be instrumental to compare to the multidomain type I fatty acid synthase, where we recently solved the structure of the malonyl/acyl transferase domain (2jfk, 2jfd). This cytosolic enzyme is involved in production of endogenous fatty acids and lipids, and is discussed as potential target in metabolic diseases and cancer. Biology area II: Structural Genomics of transmembrane receptor signalling pathways Complete coverage of the14-3-3 protein family A human protein family that the SGC has completed the structure determination of all members is the 14-3-3 family. This family consists of seven members (β, ε, η, γ, σ, τ, and ζ) of which σ [22, 23], τ [24] and ζ [25] structures were previously determined. This protein family plays a central role in many fundamental cellular roles such as cell cycle control, apotosis, protein trafficking, signal transduction and stress response [26–28]. Before the structural completion of the 14-3-3 family most of the structural studies utilised 14-3-3ζ which provided details of the conserved peptide binding site [25], the primary peptide interaction [29, 30] and secondary target domain interactions [31]. As all of these structures displayed similar overall conformations, structurally it was proposed that 14-3-3s behaved as “molecular anvils” in that their overall structure remained unchanged whether in the apo-form or bound to their target protein [32]. The structure determination of the remaining members allowed for a family-wide comparative study that revealed another story with a major emphasis on the flexibility of 14-3-3s [33]. This was most obviously with the apo-form of 14-3-3β in which one of the subunits was in a similar conformation to all other 14-3-3 structure while the opposing monomer displayed a more open conformation for the peptide binding groove (Fig. 2). Fig. 2The flexibility of the 14-3-3 is illustrated by the superimposition of 14-3-3β (blue) with 14-3-3η (orange). The monomer conformations of both isoforms are essentially identical on the left hand side. However, the beta monomer on the right side has a more open peptide binding groove and flexibility at the dimeric interface Additional flexibility of 14-3-3 proteins was observed when all of the family members were superimposed against one subunit. It became instantly clear that the position of the second subunit varied between the different 14-3-3 isoforms [33]. This is achieved through the N-terminal helices that make up the dimeric interface sliding over one another (Fig. 2). The significance of the interface flexibility is that it allows for the widening or shortening of the distance between the two peptide binding grooves hence allowing a 14-3-3 to accommodate structures of varying shapes and sizes. As 14-3-3 are known to have bind hundreds of partners [34–36] this interface flexibility would provide the necessary structural adaptability to accommodate the wide structural range of target proteins. As all of the human 14-3-3 structures are now known they allow for a detailed bioinformatic analysis of the 14-3-3 family. This approach identified common protein–protein interaction patches at the subunit interfaces plus two additional non-specific protein interaction sites that would attract and bind the globular structured regions of the target protein thus providing a mechanism by which the 14-3-3s can initially attract and then bind a wide range of structurally diverse target proteins [33]. Another more numerous protein–protein interaction family that was targeted by the SGC are the PDZ domains which have been implicated in the regulation of drug transporters [37] and involved in the clustering, targeting and localisation of the target proteins [38]. These domains bind mostly to C-terminal peptides that fall into two classes: class I peptides are –(Ser/Thr)–X–Φ–COO− while class II peptides are –Φ–X–Φ–COO− where X represents any amino acid and Φ represents any hydrophobic residue [39, 40]. PDZ domains Initial attempts at structure determination of 18 unique human PDZ domains resulted in a successful outcome for only 3 of these targets. To improve our success rate we took advantage of the family based approach and generated new expression clones of the remaining 16 targets with generic class I and II PDZ binding peptides attached to the C-terminus of each domain. The idea was for these peptides to bind adjacent PDZ domains initiating protein–protein interactions and thus crystal nucleation. As such the linker between the predicted end of the PDZ domain and the C-terminal peptide was varied from 2 to 6 amino acids allowing for flexibility but restraining the distance between adjacent domains [41]. Using this approach we have now solved 11 of the remaining 15 targets many of which have thrown up new details regarding peptide selectivity and structural adaptability of the PDZ domain when bound with a peptide. As expected for most of these domains the peptide interaction was similar to the standard configuration [42, 43] in that the side-chain of the C-terminal hydrophobic residue (position 0) was bound in a conserved hydrophobic pocket and that the peptide’s -2 position Ser/Thr coordinates the His side chain from the αB helix. However, there were a number of surprises of which the biggest was for MPDZ@3 in which a class II mode of binding was observed for a class I peptide which involved a translation of the αB helix (Fig. 6a of [41]). Biology area III: Structural Genomics of human protein kinases Kinases play an essential role in most (if not all) signalling pathways and dysregulation has often been linked to disease. Several successful inhibitors developed to target kinases have shown that members of this large protein family are excellent targets for the development of drugs. Currently protein kinases constitute about 25% of presently pursued drug targets in industry [44–47]. There are 518 identified human protein kinases constituting 1.7% of all human genes, which have been grouped into 10 families [48]. Despite the large number of members and their involvement in large variety of pathways, evidence points to a common single ancestral protein. As a result, the structural features as well as key regulatory elements and catalytic mechanism of phosphate transfer are all well conserved. High resolution structures are therefore essential for the rational design of potent and selective inhibitors. Before the contribution of SG efforts, the progression of publicly available kinase structures was linear with only 38 human kinase structures publicly available in 2004 [46]. Currently, 21 novel human kinases structures have been released by the SGC (19 from Oxford), which started to target this protein class in 2004. This increased the number of unique human kinase catalytic domain structures available in the pdb (http://www.pdb.org/pdb/home/home.do) to 93 by the end of 2006. Many structures, released by SG, were only distantly related catalytic domain structures previously known and in some cases provided the first structural information for a subfamily. Thus, these structures significantly enriched the coverage of the three dimensional structure description of the kinome. Among the structures where the SGC determined the first representative structure of a family were: the NEK (“never in mitosis”/NIMA) family member NEK2, the CDC2 like kinases family member CLK1 and CLK3 as well as the first structure of a NAK (Numb-associated kinases) kinase MPSK1. These kinases are quite diverse in terms of primary structure and it is therefore not surprising that many novel structural features have been discovered. For instance, a novel activation loop architecture characterized by a large helical insert has been discovered in the structure of MPSK1, the structures of CLK1 and CLK3 revealed a family conserved antiparallel beta sheet flanking the kinase hinge region, and the structure of NEK2 identified a short helix following the activation segment DFG motif that may be explored for the development of specific inhibitors [49]. Kinases are extremely flexible proteins that may adopt a number of distinct catalytically active or inactive conformations during their catalytic cycle, upon activation by phosphorylation, or by binding of a regulatory protein, and consequently a number of clinically successful inhibitors have been developed to target specifically the inactive state of kinases [50]. For example the anti-leukaemia drug Imatinib binds selectively to the inactive state cABL characterized by an outward conformation of the DFG motif, a conserved tripeptide motif that ligates Mg2+ ions [51, 52]. It is not clear to date how many kinases are able to adopt this conformation, which makes development of these so-called type II inhibitors possible. In general, these are characterized by largely improved specificity. For the development of conventional inhibitors that target the active state of kinases information about the plasticity of the catalytic domain greatly facilitates the rational design of inhibitors. Consequently it is desirable that several structures of the same target in complex with different ligands are available. Also here the structural information content regarding ligand binding was significantly increased during the last three years by SG. In 2004, only 38 human kinases had a structure available in the public domain and only 12 publicly available structures contained non-adenosine chemotypes [46]. From the 19 structures of kinase catalytic domains released by our laboratory, 16 were determined in the presence of an non-adenosine kinase inhibitor and several structures were determined in complex with more than one inhibitor scaffold (Table 2, and Fig. 3, showing PAK5 apo/inhibitor). Table 2Protein kinase structures determined by SGCNamePDB IDResolution [Å]Inhibitor nameDisease linkFamilyCLK11Z571.70HymenialdisinePot. GeneticeCMGCCLK3c2EU91.53nonePot. GeneticeCMGCCK1γ12CMW1.75Compound 52CK1CK1γ22C472.405-IodotubercidinGeneticCK1CK1γ3d2CHL1.95Triazolodiamine 1CancerCK1ERK32I6L2.25noneCancerCMGCASK12CLQ2.30StaurosporineInflammation, CVfSTENEK22JAV2.10SU11652CancerOther-NEKPAK4a2CDZ2.40Cdk1 InhibitorCancerSTEPAK52F571.80Cdk1 InhibitorPot. CancerSTEPAK62C301.60noneCancerSTEPIM1b1XWS1.80BIM I, HB1Cancer, InflammationCAMKPIM22IWI2.80HB1Cancer, InflammationCAMKSLKc2J512.10Triazolodiamine 1Pot. CancereSTEMPSK12BUJ2.60StaurosporinePot. CancereOther-NAKSTK102J7T2.0SU11274Not knownSTEDAPK32J902.0Pyridone 6Cancer, InflammationCAMKCAMK1G2JAM1.7SU11652Not knownCAMKCAMK1D2JC62.5GSK inhibitor XIIIGeneticCAMKaPAK4 also deposited as an apo-structure in two different spacegroups: 2BVA, 2J01bStructures with different inhibitors and substrate: PIM1: 2BIK, 2BZH,2BZI, 2BZJ, 2BZK, 2C3I, 2BIL, 2J2I; SLK: 2JA0cCLK3 also deposited as phosphorylated protein: pdb-code: 2EXE. Detailed description of structures of targets solved in Oxford is available in form of “iSee” datapacks freely downloadable on: http://www.sgc.ox.ac.uk/structures/KIN.htmldCK1γ3 with different inhibitors: 2CHL, 2IZR, 2IZS, 2IZU, 2IZTeA formal link to the disease has not been established so far but is likelyfCardiovascular diseaseFig. 3Superimposition of apo-PAK5 (cyan) and the PAK5 purine complex (orange), highlighting the decomposed movements of the glycine-rich loop (flapping) and the αC helix (swinging) [53] In addition, the SGC has supported development of entirely new inhibitor classes exemplified by co-crystal structures with Ruthenium-half sandwich complexes. These stable organometallic compounds are extremely potent inhibitors for PIM1 kinases [54]. The co-crystal structure of three inhibitors of this class showed that the inert metal centre in this scaffold functions as a hypervalent carbon, allowing it to occupy the binding pocket efficiently with excellent shape complementarity. Contributions of NMR to Structural Genomics NMR as a complementary method to crystallography for protein structure determination The NMR spectroscopy can play an important role in structural genomics, providing complementary information to that obtained from X-ray crystallography. Importantly for large-scale structural genomics projects, NMR provides an alternative route to solving the high resolution, three-dimensional structures of proteins that prove refractory to crystallization. We were able to use NMR to solve the structures of a number of relatively small protein domains (∼20 kDa) in which the domain contained at least one flexible region. The RGS domains from the regulator of G-protein signalling proteins, RGS3, RGS10, RGS14, RGS18, RGS20 were all very good examples of this. Multiple constructs of these were designed, which expressed to high yield in stable, highly soluble form yet did not yield high quality crystals despite many months of concerted effort. The domains were therefore expressed as uniformly 15N-labelled proteins using standard growth methods in E. coli, and their 15N-HSQC spectra were recorded to assess the feasibility of structure determination by NMR. In all cases, excellent spectral dispersion was observed and we were able to obtain almost complete assignment of the protein resonances. We have since deposited the high resolution NMR structures of three RGS domains in the PDB and the resonance assignments of four RGS domains in the BioMagResBank (BMRB). The structures and assignments of two further non-crystallizing domains (Spred2 EVH1 domain and JARID1CA Bright/ARID domain) have also been deposited, and those of several other non-crystallizing domains ‘rescued’ by NMR are currently underway (Table 3). Table 3Deposited NMR structures and assignmentsGenePDB depositionResonance assignment depositionRGS3–BMRB-15178RGS102I59BMRB-7272RGS142JNUBMRB-15128RGS182OWIBMRB-7106SPRED22JP2BMRB-5939JARID1CA2JRZBMRB-15348 NMR as an assessment tool for the feasibility of structure determination Further examples where NMR has proven useful as a rescue strategy include particular families of signalling domains which have a known tendency to be partially unfolded in their unliganded states. Some examples include certain WW domain [55, 56]. We successfully identified peptide binding partners for a WW-tandem construct using the SPOTs screening technique [57, 58] following which the most strongly binding peptides were synthesized on a large scale for NMR measurements. Although the 15N-HSQC spectra of this pair of tandem domains in isolation were very unhopeful, the spectra of their complexes showed significant improvements in signal dispersion, indicating that in the complexed form, the protein was better folded. At this point, the protein entered our NMR structure determination pipeline. The recording of a quick 15N-HSQC spectrum has in several cases allowed us to rescue protein constructs with promising but borderline behaviour, for example, proteins showing good signal dispersion but low-medium levels of aggregation. Far from abandoning these constructs, we took these constructs ‘back to the drawing board’ and made rational construct improvements with the help of bio-informatic tools. Successfully re-designed constructs were then re-screened for fold quality by 15N-HSQC. After 2–3 iterations of this procedure, it was often possible to refine promising constructs sufficiently for structure determination. For example, Fig. 4 shows the stepwise improvement in the spectral properties of a hopeful, though initially problematic DNA-recognition domain from the oxygenase protein, JARID1CA. The NMR structure is now deposited (PDB code: 2JRZ). In all of the above cases, a quick 15N-HSQC showed immediately whether the structure determination of a protein, having to failed to crystallize, should be pursued or abandoned, hence reducing unnecessary attrition in the structure determination pipeline. Fig. 4Visible improvement in quality of 15N-HSQC spectra over two rounds of iterative construct re-design for the JARID1CA Bright/ARID domain. The leftmost (initial) construct shows potential. The structure of the final construct on the far right was determined by NMR (PDB code: 2JRZ) The study of protein dynamics by NMR The use of NMR to study the rotational correlation times and internal dynamics of the proteins offers good explanations as to why crystallization sometimes fails even for well-folded proteins. In all of the proteins we rescued by NMR, 15N heteronuclear NOE and 15N T1, T2 relaxation data revealed regions of internal mobility within the proteins, which would have hindered long-range order and impaired or prevented efficient crystal packing. A striking example was the case of the RGS domain from RGS10, in which NMR relaxation data confirmed true local mobility in a region of the domain which not only lacked in NMR restraints, but also showed no electron density in the crystal structure of the complex of RGS10 with G-alpha-i3 (PDB 2IHB). Comparison of mobility in RGS domains from different branches of the phylogenetic tree leads to clues about their specificity and helps to guide further investigations. In some cases, the 15N T1 and T2 data have also identified partial dimerization in proteins that fail to crystallize, thus explaining the latter. NMR relaxation data were in each case confirmed by analytical ultracentrifugation (AUC). The combined information allowed us to decide whether these proteins should be highlighted as candidates for structure determination by NMR and to judge the best conditions under which they should be studied. Future and outlook The future role that NMR will play in structural genomics will depend heavily on the continued development and implementation of new, faster methods of data acquisition, processing, resonance- and NOE-assignment and structure determination and refinement. These topics have been covered extensively in other reviews; for a concise summary see [59] and references therein. The potential time gains that could be gained from these methods make high throughput structure determination by NMR a realistic possibility for the future. Structural bioinformatics and rationalisation of experimental results A crystal structure of a protein in absence of ligand or substrate may not always provide insight on reaction mechinasms or specificity. Ideally, such information can be derived from additional structures with bound ligands. In the absence of such co-crystals, interpretation of results from biochemical assays and compound screening is more speculative. However, these results can be rationalised with in silico docking of potential ligands into the active site of unliganded protein structures. An example illustrating this point is the analysis of the DHRS10 structure [20]. Analysis of the different docking poses can be correlated with experimental results, allowing direct visualisation of the putative protein–ligand complex. With these results, further modifications of the enzyme can be suggested more reliably, allowing a faster progress towards the complete elucidation of the mechanistics. Dissemination of structural genomics data and knowledge Structural genomics produces a wealth of information of different types: DNA and protein seqeuences, biochemical information, coordinates of crystal structures, and structural annotation. This information is deposited in one or more public databases, predominantly the PDB, in addition to publication in journals. This form of data distribution does not adequately disseminate the full information to a wide scientific audience. The first issue is the fragmentation of data between different formats. A user may have to read text information in a journal paper, which may include a few two-dimensional Figures; then download a PDB structure file and image with a separate application; and then perform analysis and alignment of data from, say, SNP database using alignment software. The second issue is that non-structural biologists do not routinely access PDB files, especially of structures that were not published in pubmed-indexed journals. We have approached this challenge by developing a new intuitive dissemination concept in conjunction with Molsoft LLC (San Diego, CA) [60]. This concept, (which we denoted iSee) integrates all the information associated with any given target solved by SGC into a small, self-contained file, annotated by the authors (Fig. 5). The file not only allows the direct visualisation of text information, but also offers an interactive visualisation feature fully integrated to the structural data being presented. At any stage, the annotation written by the expert can be coupled with an interactive molecular graphics scene. Transition between each anotated viewpoint is fully animated on-the-fly, to convey a sense of three-dimensionality which is vital for the user to grasp the spatial relationship between different features on a structure. Fig. 5Screenshot of iSee datapack. The annotation text (top left panel) includes links (blue text), which lead to structural images focused at areas of interest, simultaneously accessing other types of information (sequence alignment, small molecule formulae, etc.) Each of these files (called an iSee datapack), as well as the software needed to visualise them (ICM-Browser) are available for free download from our website (http://www.sgc.ox.ac.uk/iSee). We also maintain and curate each of these files by revising each datapack quarterly to ensure that all the recently disclosed information is added (either by ourselves through follow-up experiments or by external collaborators working on the same targets). Each of the datapacks has a built-in automated updating function that can be executed on user’s request.	[ "protein kinase", "reductase", "high-throughput", "dehydrogenase", "14-3-3", "pdz", "binding specificity", "protein crystallography" ]	[ "P", "P", "P", "P", "P", "P", "R", "R" ]
Matern_Child_Health_J-2-2-1592250	Strategies to Reduce Alcohol-Exposed Pregnancies	Introduction Prenatal alcohol exposure remains a leading preventable cause of birth defects and developmental disabilities in the United States, with the prevalence of alcohol consumption by women in childbearing age remaining high and unchanged over time (Fig. 1). This paper provides a summary of current knowledge and information on recognition and prevention of an alcohol-exposed pregnancy, including overall alcohol exposure burden among women of child-bearing age, diagnostic criteria for fetal alcohol syndrome (FAS), and recommendations for screening, assessment and interventions to reduce alcohol exposed pregnancies. Burden of alcohol exposure in pregnancy Most women reduce their alcohol use substantially when they realize they are pregnant, but significant numbers continue to drink at levels that can be hazardous to the fetus [1, 2]. Of particular concern are the group of sexually active women who are not planning to become pregnant (about half of all pregnancies in the US), but do so and continue alcohol use during the early stages of embryonic and fetal development [1]. For many women, pregnancy recognition does not occur until the 6th week of gestation [3]. According to national survey data, in 2002 about 8% of women aged 18 to 44 years were sexually active, fertile, not using any form of birth control, and at risk of becoming pregnant [1]. This group of women also report high rates of binge drinking, with 1 in 5 in the age category of 18 to 24 years reporting multiple episodes of binge drinking, and are therefore at risk for an alcohol-exposed pregnancy. Fig. 1Weighted percentage of alcohol consumption for non-pregnant women aged 18 to 44 years during the previous 30 days, BRFSS, United States. Data were not collected in 1994, 1996, 1998, and 2000. Frequent refers to = 7 drinks/week or binge. Binge drinking refers to = 5 drinks on one occasion Adverse fetal outcomes Prenatal alcohol exposure can result in a wide range of adverse outcomes including fetal alcohol syndrome (FAS), although not all infants exposed to alcohol in utero develop the same severity of effects as do others. The term fetal alcohol spectrum disorder (FASD) has been introduced to describe the range of physical, mental and behavioral effects that can occur in an individual exposed to alcohol in utero. Because diagnostic criteria are not available for all conditions along the spectrum, there are no prevalence rates available for the full spectrum. Prevalence rates of FAS range from 0.3 to 2.0 cases per 1,000 live births [4] depending on the methodology used, and the sub-populations assessed. Reported rates are higher among infants born to women who are American Indian/Alaska Native or African American, unmarried, smokers, have low incomes, and have a history of previous drug use or mental health conditions [5, 6]. In an effort to promote the more complete recognition of FAS, guidelines for identifying and referring persons with FAS have been collaboratively developed and disseminated recently [4] (www.cdc.gov/ncbddd/fas/documents/fas_guidelines_accessible.pdf). Approaches to reducing alcohol-exposed pregnancies A 1996 report from the Institute of Medicine addressing prevention of FAS recommended implementation of preventative actions at multiple levels including individual, group, and universal levels [7]. To date, most of the effective prevention strategies identified have focused on the individual level. In 2004 the U.S. Preventive Services Task Force (USPSTF) released a report recommending screening and behavioral interventions to reduce alcohol misuse in adults in primary care settings [8], www.preventiveservices.ahrq.gov. The report concluded that effective interventions include brief counseling comprised of feedback regarding screening and assessment information, advice, goal-settings and follow-up assistance. The report further cited complementary practices of motivational interviewing, assessing readiness to change, and use of the 5 A’s behavioral counseling framework of assess, advice, agree, assist and arrange. Validated screening instruments are available for screening pregnant and non-pregnant women of reproductive age including the T-ACE, TWEAK, and AUDIT [9]. More information on these instruments is available at the following website: www.nih.gov/publications/Assessing/Alcohol/Index.htm. Primary healthcare providers can play a pivotal role in identifying women of reproductive age (during pregnancy and before pregnancy occurs) who are at high risk for an alcohol-exposed pregnancy and providing them with advice, counseling, and referral as appropriate. This strategy has widespread support among professional organizations including the American College of Obstetricians and Gynecologist (ACOG), and the American Academy of Pediatrics (AAP), as well the U.S. Office of the Surgeon General, and the U.S. Department of Health and Human Services. A useful example of an intervention targeting groups of women using many of the components described in the recommendation of the USPSTF is found in a CDC sponsored study, Project CHOICES, a feasibility pilot study that targeted non-pregnant women at high risk for an alcohol-exposed pregnancy [10]. Project CHOICES was conducted in diverse sub-populations of women determined to have higher proportions of individuals at risk for an alcohol-exposed pregnancy. Compared to the overall estimated 2% of childbearing women at risk for an alcohol-exposed pregnancy, the Project CHOICES intervention settings (e.g. a jail, alcohol and drug treatment facilities, and primary care clinics) had an overall rate of 12.5% of women at risk for an alcohol-exposed pregnancy. A unique aspect of the intervention is that it focused not only on reducing risk drinking, but also addressed pregnancy postponement as a route for avoiding an alcohol-exposed pregnancy. The intervention consisted of 4 brief motivational interventions sessions and 1 consultation visit to a family planning provider. All participants were at risk for an alcohol-exposed pregnancy in that they were sexually active, fertile, risky drinkers and not taking effective measures to avoid pregnancy. At the 6 month follow-up assessment, 68.5% were at reduced risk because they had either changed their risk drinking, instituted effective contraception, or both. Subsequently, the Project CHOICES Research Group completed a randomized controlled trail to test the efficacy of the intervention with a report of the study findings currently underway. Universal level interventions have not received recent attention with the exceptions of the 2005 release of U.S. Surgeon General’s Advisory on Alcohol Use and Pregnancy, http://www.hhs.gov/surgeongeneral/pressrelease/sg0222205.html. This advisory drew attention to the continuing problem of FASD and the continuing high rates of alcohol use, including binge drinking, among childbearing aged women in the U. S. The impact of the advisory has provided support for those seeking to inform and educate the public healthcare system overall to this important public health concern. Summary In addition to the adverse effects of alcohol on the fetus, alcohol use can leads to other adverse effects on the health of women including reproductive health conditions that are not addressed in this brief. Though the approaches mentioned here are primarily aimed at reduction of prenatal alcohol-related pregnancies, any reduction in hazardous alcohol consumption among women will add to improvements in the general health of women. Multiple federal and non-fedral health agencies and organizations recommend that pregnant women and those planning a pregnancy should abstain from alcohol use. Although assessment and interventions are valuable tools to recognize and address alcohol use and secondary pregnancy outcomes, they are underutilized in primary care settings. To reduce the burden of alcohol exposed pregnancies and alcohol’s impact on the health of families, it is critical for physicians and health care providers to consistently screen childbearing-aged women for alcohol use with validated screening tools which can be embedded in the patient screening protocol. Tools are now available for pediatricians and child health care providers to enhance early recognition of FASD and reduce secondary conditions that often accompany physical maturation [4]. Brief clinician-delivered behavioral interventions to women and their partners, counseling regarding effective contraceptive options when not planning a pregnancy, and improving access to such services for those who are unable reduce their alcohol intake can help more women reduce their risk for an alcohol exposed pregnancy [11]. Referral to needed social services should complement health services in order to achieve maximum benefits of primary care-based attempts to reduce alcohol-exposed pregnancies. Given the levels of alcohol use among childbearing aged women, primary care-based individual level assessment and intervention continues to remain a critical prevention strategy for reducing alcohol-exposed pregnancies.	[ "alcohol", "pregnancy", "preconception care" ]	[ "P", "P", "M" ]
Skeletal_Radiol-4-1-2226078	Radiofrequency ablation of osseous metastases for the palliation of pain	A number of different methods have been proposed for pain relief in cancer patients with bone metastases, each with different indications, contraindications and complications (systemic analgesics, bisphosphonates, antitumor chemotherapy, radiotherapy, systemic radio-isotopes, local surgery and vertebroplasty). The ideal treatment has to be fast, safe, effective and tolerable for the patient. CT-guided radiofrequency (RF) ablation may fulfill these criteria. Our experience in the treatment of 30 patients (34 lesions) with painful bone metastases using RF ablation was assessed. There was a significant decrease in the mean past-24-h Brief Pain Inventory (BPI) score for worst pain, for average pain and for pain interference during daily life (4.7, 4.8 and 5.3 units respectively) 4 and 8 weeks after treatment. There was a marked decrease (3 out of 30 patients 4 and 8 weeks after treatment) in the use of analgesics. CT-guided RF ablation appears to be effective for treatment of painful bone metastases. Introduction Painful bone metastases are a common cause of morbidity in patients with metastatic cancer, especially when combined with possible neural compression and pathologic fractures. Several solid cancers are associated with bone involvement, most often, prostate and breast. Thirty to seventy percent of cancer patients develop bone metastases [1]. They indicate widespread disease. Treatment of local disease may reduce the pain of these patients who, in most cases, have a life expectancy of months. Such treatment must be fast, safe, effective and tolerable. A number of treatment methods are available that have variable success and complications. Radiation therapy is the preferred treatment in this setting, but other modalities such as chemotherapy, hormonal therapy, radiopharmaceutical therapy and surgery—alone or in combination with non-steroid anti-inflammatory drugs (NSAIDs), opioids and adjuvant drugs—are used for pain palliation [1–3]. Radiofrequency (RF) ablation is a relatively new method for the treatment of painful bone metastases. Previously, tumour ablation was performed with percutaneous ethanol injection under CT guidance [4]. Administration of 95% ethanol was described in 25 terminally ill cancer patients with 27 bone lesions who had been unsuccessfully treated by radiation therapy and/or chemotherapy. Radiofrequency ablation has been employed for the treatment of hepatocellular carcinoma (HCC), liver metastases, renal and lung tumours, as well as for the treatment of osteoid osteoma, for which it has become the treatment of choice [1–3]. Competing methods include chemical ablation (with ethanol or acetic acid) and thermal therapies, such as with laser, microwave, ultrasound and cryoablation [5]. The aim of this study was to demonstrate the effectiveness of RF ablation of bone metastases using CT guidance. Materials and methods Thirty patients were retrospectively identified. There were 19 men and 11 women. Their ages were between 47 and 91 years (mean ± standard deviation [SD]: 66.53 ± 10.56 years). The patients had bone metastases, which were treated with RF ablation under CT guidance, at our hospital, over a period of 4 years. All treated lesions were osteolytic with a combination of bone destruction and a soft tissue mass. In 26 there was a solitary lesion, and in 4 patients there were two such lesions, resulting in a total number of 34 metastases. Bone metastases were diagnosed by bone scintigraphy and spiral CT. The diagnosis was confirmed with a core biopsy obtained at the beginning of the procedure. Their topographical distribution and the originating primary malignancies are presented in Table 1. In our study the most common treated metastases originated from colon cancer, which was probably related to the patient population treated at the oncology department of our hospital. Table 1Bone metastasis classification, with regard to the primary malignant lesion and the site of the skeleton involvedSite of primary neoplasm originnSite of metastasisnColon13Pelvis15Breast7Ribs6Prostate2Sacrum5Lung4Femur3Renal2Spine3Thyroid1Scapula2Skin melanoma1Tibia 1Total30Total34 Lesion diameter was between 1 and 14 cm (mean ± SD: 3.9 ± 2.6 cm). For sizes over 3 cm, two or more electrode placements were needed (with a maximum of five). Previously obtained imaging examinations were evaluated for lesion characteristics and feasibility of electrode positioning and ablation. Lesions located in proximity to the spinal cord and major nerves (less than 1 cm) were excluded from RF treatment. Patient selection criteria are summarised in Table 2. The study was in accordance with the ethical principles of the Helsinki Declaration and informed consent was obtained in each case. Table 2Criteria for the selection of patients to undergo radiofrequency ablationPatient selection criteriaBrief Pain Inventory (BPI) score above 4Lesions not responding to chemotherapy and/or radiation therapy (completion of therapy at least 3 weeks before the radiofrequency ablation session)Chemotherapy-associated complications that halted this treatmentLesions adjacent to structures sensitive to irradiationPatients with life expectancy greater than 2 months who were not eligible for surgical treatment Patients who preferred this treatment over the other alternatives Physical examination was performed by the oncologist and in collaboration with the radiologist performing the ablation. Pain was assessed with the Brief Pain Inventory (BPI) The use of analgesics was recorded the day before the procedure. Before the procedure blood cell count and blood clotting analysis were performed. Minimal requirements were: platelet (PLT) count >50,000/ml (normal range, 150,000–350,000/ml); prothrombin time (PT), international normalised ratio (INR) <1.3 (normal range, 0.8–1.2); and partial thromboplastin time (PTT) <34 s (normal range, 25–34 s). The procedure was performed under conscious sedation (administration of 3 mg of bromazepam PO and 50 mg of pethidine hydrochloric acid intramurally, 45 min prior to the procedure) and was trained in regular breathing and breath-holding (suspended respiration) before the procedure. He/she was placed in the appropriate position (prone, supine, or lateral, depending on the site of the lesion) and a scan of the desired area with a 5-mm slice thickness was performed, using a Picker 5000® (Philips Medical Systems, Amsterdam, The Netherlands). At least one of two staff radiologists with extensive experience in biopsies and tumour ablations was involved in all ablations. The lesion’s exact location and depth, in relation to the overlying skin, was determined on CT. The skin was then prepared with povidone iodine (10%) solution. Local anaesthesia (15 ml of 2% lidocaine hydrochloride solution) was administered. Radiofrequency ablation was performed with a RITA Model 1500® electrosurgical generator (RITA Medical Systems, Mountain View, CA, USA) and a seven-array, 2- to 3-cm multitined electrode for lesions smaller than 3 cm (20 out of 34), or a nine-array multitined electrode for larger lesions (14 out of 34). The electrode tip was inserted to approximately 1 cm from the centre of the target. The electrodes were then deployed slowly, taking into account the need to ablate the lesion–bone interface. The net ablation time was ~15 min at an energy level of 90–110 W, with the target goal temperature set to 80–110°C. During the procedure the infusion port of the electrode was flushed with a 2% lidocaine hydrochloride solution in order to reduce patient discomfort and to decrease tissue overheating and vaporisation. The number of electrode placements, individual (per electrode) and total ablation times, the total energy delivered to the target and the lesion temperatures achieved were recorded. After each session a dual-phase spiral CT examination with intravenous contrast medium was performed in order to assess response, as confirmed by low lesion attenuation values and lack of contrast enhancement. Patients were hospitalised and observed for 24 h. Analgesics were administered if required. Before patient discharge the pain was re-evaluated with the BPI score. Post-ablation assessment was completed with telephone interview after one, four and eight week. The BPI score and the use of analgesics were recorded again. Results For lesions smaller than 3 cm (20 out of 34), one placement was adequate, while for the remaining 14 out of 34 cases of lesions that were larger than 3 cm, two or more placements were required; one lesion sized 14 cm required five placements, accomplished in two sessions (since more deployments are required in larger lesions). The total procedure time ranged between 33 and 65 min (mean ± SD: 42 ± 11 min; Table 3). There were no complications and post-treatment CT revealed a good response, as confirmed by low lesion attenuation values and a lack of contrast medium enhancement, consistent with necrosis. Post-procedural CT did not demonstrate any major complications (such as haemorrhage, thrombosis of neighbouring veins, or skin burns). Table 3Lesion characteristics andtreatmentNumber of lesionsSize of lesion (cm)Number of electrode placementsTime of radiofrequency energy deposition (min)15<31553157427352816391841019410114515Total 34 Eleven patients reported no pain reduction during the first 24 h after the procedure and were treated with analgesics (opioids or an opioid/NSAID combination). Nineteen of the 30 patients reported early pain reduction. In none of the patients was increased pain reported. Prior to the procedure, the mean past-24-h BPI score for worst pain was 8.3 (on a numerical rating scale where 0 indicates no pain, and 10 indicates worst pain imaginable), mean pain was 6.8, and mean pain interference with daily life 7.5. These scores were reduced to 7.4, 4.7 and 6.5 24 h after the procedure, dropped to 4.9, 3.2 and 4.0 after 1 week, to 3.6, 2.00 and 2.2 after 4 weeks, and to 2.1, 1.4 and 1.7 after 8 weeks respectively. These results revealed a marked decrease in pain with subsequent improvement in the life quality for all participants since the first week post-treatment that lasted throughout the 8-week follow-up (Fig. 1). For all time points, the mean past-24-hour BPI score for worst pain, for average pain and for pain interference in daily life improved in comparison to preprocedural symptoms (p < 0.001, paired t test; Figs. 2, 3, 4). Fig. 1Mean Brief Pain Inventory (BPI) scores over time for patients treated with radiofrequency ablation. a Worst pain. b Average pain. c Interference of pain in everyday lifeFig. 2Computed tomography scan a during and b after the radiofrequency session with the patient in a supine position. The electrode is deployed inside the metastatic lesion of the left acetabulum (metastasis from breast cancer). There is no enhancement after intravenous contrast media administration. Before the radiofrequency session, the average pain score was 7. During the first 24 h after radiofrequency, it was 4 and 1, 4 and 8 weeks later the average pain scores were 3, 2 and 0 respectivelyFig. 3Computed tomography scan images at the level of the sacrum show a, b two different electrode placements within a soft tissue mass involving the sacrum and right iliac bone (metastasis from thyroid cancer). Before the radiofrequency session, the average pain score was 7. During the first 24 h after radiofrequency, it was 4 and 1, 4, and 8 weeks later the average pain scores were 3, 2 and 1 respectivelyFig. 4Computed tomography scan image during the electrode placement inside a a metastatic lesion involving a rib (the primary neoplasm originates in the lung). Immediately after the radiofrequency session the contrast enhancement CT scan revealed a hypodense area (necrosis) inside the lesion. Before the radiofrequency session, the average pain score was 6. During the first 24 h after RF, it was 4, and 1, 4 and 8 weeks later the average pain scores were 2, 2 and 1 respectively Prior to RF ablation 27 out of 30 patients received opioids or an opioid/NSAID combination. The remaining 3 patients received NSAIDs. One week after treatment 5 out of 30 patients were treated with a combination of NSAID/low-dose opioids. Six out of thirty used NSAIDs. After 4 and 8 weeks only 3 out of 30 patients received any medication NSAIDs. One patient died during the 8-week follow-up for reasons not related to RF ablation. Discussion In patients with cancer, pain originating from bone metastases can be difficult to treat. A number of treatment options are available, including NSAIDs, opioids, and adjuvant drugs medications, radiation therapy, chemotherapy, hormonal therapy, radiopharmaceutical therapy, surgery and vertebroplasty. Medication represents the first line of treatment. NSAIDs and adjuvant drugs represent basic medication, potentially followed by NSAID/low-dose opioid combinations, and finally increasing the opioid dose. Radiation therapy is another treatment option that may also be employed in pathologic or impending fractures [6]. Approximately 70% of patients undergoing radiation therapy will experience pain relief after between 2–3 days and up to 4 weeks after treatment. However, radiation therapy may also cause complications, mostly from damage of adjacent soft tissues [6]. Chemotherapy and radiopharmaceutical therapy are the only systemic methods of treatment that deal with even small foci of metastatic cells. However, not all metastases are sensitive to chemotherapeutic agents. Chemotherapy is often not well tolerated and is associated with complications. Radiopharmaceutical therapy can be more useful in treating patients with multifocal bone metastases. It has been reported that radiopharmaceuticals proved efficient in pain palliation mostly in bone metastases from breast, prostate and perhaps small cell lung cancer. As in the case of chemotherapy, all agents have advantages and possible side effects. Radiopharmaceutical agents vary with regard to the analgesic efficacy, duration of pain palliation, ability to repeat treatments, toxicity and expense [7]. The term “ablation” refers to the local destruction of the tumour by the means of application of either chemical agents (ethanol, acetic acid), or local deposition of some form of energy (radiofrequency, laser, microwave, ultrasound and cryoablation). Image-guided RF ablation is currently used for the treatment of various tumours with good results. According to preliminary results by Dupuy et al. [8], RF ablation can provide palliative treatment for patients with painful osseous metastases. Later on, Callstrom et al. [9] reported results after treating 12 patients and concluded that this modality provides an effective and safe alternative method of pain palliation in patients with osteolytic metastases. A multicentre study involving 43 patients with painful osseous metastases was carried out by Goetz et al. [10] and showed again significant reduction of pain and decrease in the use of opioids, with only minor complications. The proposed mechanisms by which RF ablation decreases pain may involve: pain transmission inhibition by destroying sensory nerve fibres in the periosteum and bone cortex; reduction of lesion volume with decreased stimulation of sensory nerve fibres; destruction of tumour cells that are producing nerve-stimulating cytokines (tumour necrosis factor-alpha [TNF-α], interleukins, etc.) and inhibition of osteoclast activity [11, 12]. In our patients, we observed a considerable reduction of pain and improvement of the quality of life, as measured by the BPI score. A decrease in the use of analgesic medications was notable in our series and possibly greater than others reported in previous studies [9, 10]. The reduction of the procedural time is limited by both the time needed to achieve the optimal target temperature, and the size of the lesion, because more than one deployment is required in larger bone metastases. Although a few patients reported mild discomfort during the ablation, none of the sessions was forced to stop owing to considerable patient distress. There were none of the possible adverse effects, including infection, haemorrhage, neurological complications, skin burns, or the so-called post-ablation syndrome (low-grade fevers ≤100°F [37.8°C]), myalgias, and malaise for up to 1 week after the procedure). Pain reduction was fast and occurred within the first 24 h for some and during the first week in the majority of the patients. This appears to be a fairly well-tolerated procedure and the combination of conscious sedation and local anaesthesia is adequate for its needs. In the spine, RF ablation may be contraindicated due to the close relationship with the spinal cord and nerve root. Vertebroplasty may be used for pain relief and stabilisation of osteolytic lesions. Pain relief occurs within hours or days (mean 24 h) of the procedure, sometimes after a transient worsening of pain [13]. Mechanical or thermal damage to the adjacent soft tissue from needle positioning or cement leakage are the potential complications. In our study, there were no lesions threatening the stability of the spine. There are two conflicting studies in the literature concerning the use of RF ablation in spinal metastases. One of them has demonstrated that the presence of cancellous or cortical bone between the lesion and the spinal canal can provide adequate safety for the procedure [14, 15]. Another study in an in vivo animal model, with the use of magnetic resonance imaging (MRI) and pathological evaluation, has demonstrated that the placement of the electrode against the posterior vertebral body wall resulted in damage of the spinal cord [16]. In most series, lesions within 1 cm of the spinal cord, lesions involving the posterior wall and lesions with cortical bone destruction with involvement of soft tissue were considered to represent contraindications to treatment [17]. All spinal metastases treated in our series involved the anterior part of the vertebral body. The follow-up period for this study was 8 weeks, a period that we believed was sufficient to demonstrate that RFA provides effective palliation. There is, however, a need for randomised prospective studies, to evaluate the method and to compare it with other treatment modalities, such as radiation therapy. Continued follow-up is warranted to determine the long-term efficacy of this interventional approach. In conclusion, imaged-guided RF ablation of painful bone metastases is promising. It appears to be effective, safe and well tolerated by patients.	[ "radiofrequency ablation", "osseous metastases", "pain", "minimally invasive treatment" ]	[ "P", "P", "P", "M" ]

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.