Coagulation and Transfusion Medicine

Transcription

1 Coagulation and Transfusion Medicine / BLOOD COLLECTION TUBES AND PT Influence of Blood Collection Systems on the Prothrombin Time and International Sensitivity Index Determined With Human and Rabbit Thromboplastin Reagents Anton M.H.P. van den Besselaar, PhD, 1 Martha M.C.L. Hoekstra, 1 Evelina Witteveen, 1 Jan H. Didden, 2 and Felix J.M. van der Meer, MD, PhD 1 Key Words: Prothrombin time; International sensitivity index; International normalized ratio; Blood collection tube DOI: /NW194EAMD4WMEHJD Abstract Three brands of blood collection tubes were studied for their influence on the prothrombin time (PT) and international sensitivity index (ISI) for 5 commercial thromboplastin reagents. With all reagents, PTs were shorter in Vacutainer (Becton Dickinson Vacutainer Systems, Plymouth, England) samples than in S- Monovette (Sarstedt, Nümbrecht, Germany) or Venosafe (Terumo Europe, Leuven, Belgium) samples. ISI values were higher with Vacutainer samples than with S- Monovette or Venosafe samples. The ISI differences between the tubes were small for Thromborel-S (2.1%; Dade Behring, Marburg, Germany) and Hepato Quick (1.1%; Diagnostica Stago, Asnières, France; Roche Diagnostics Nederland, Almere, the Netherlands) but greater for Neoplastin Plus (5.5%; Diagnostica Stago; Roche Diagnostics Nederland), Simplastin HTF (8.3%; biomérieux, Durham, NC), and Innovin (8.8%; Dade Behring). The PT and ISI differences between the tubes could be explained mostly by the effect of magnesium ion contamination in the sodium citrate solutions. When PT ratios were transformed into international normalized ratios (INRs) using crossover ISI (ie, samples collected with one type of tube and ISI determined with another collection system for the PT reagent), the differences in mean INRs could be approximately 10%. For ISI calibration of reference thromboplastins, blood collection tubes should be used with minimal divalent metal ion contamination of the citrate solution. Monitoring of vitamin K antagonists (eg, warfarin, phenprocoumon, and acenocoumarol) is performed primarily with the prothrombin time (PT) test, ie, the tissue factor induced coagulation time. Although warfarin, phenprocoumon, and acenocoumarol are drugs with different pharmacokinetic profiles, they all act by inhibiting the vitamin K dependent carboxylation of clotting factors II, VII, IX, and X at a postribosomal stage in the liver, leading to a buildup of biologically inactive precursor clotting proteins. The result of the PT test should be reported as the international normalized ratio (INR) using the international sensitivity index (ISI) of the PT test system. 1 The ISI of the PT test system depends on the combination of the thromboplastin reagent and the coagulation instrument. The PT is also influenced by preanalytic conditions such as the sodium citrate concentration in the blood collection system. 2-5 Furthermore, the PT can be shortened by magnesium ions contaminating the sodium citrate solution in some commercial blood collection systems. 6-8 Previous studies on the influence of magnesium ions on the PT were performed primarily with Innovin (Dade Behring, Marburg, Germany), ie, recombinant human thromboplastin. In one study, we compared 2 commercial brands of blood collection tubes, Vacutainer (Becton Dickinson Vacutainer Systems, Plymouth, England) and Venoject II (Terumo Europe, Leuven, Belgium), using another brand of recombinant human thromboplastin. 8 Recently, a new brand of blood collection tubes (Venosafe, Terumo Europe) was introduced. The purpose of the present study was to compare Venosafe with 2 other blood collection systems, Vacutainer and S-Monovette (Sarstedt, Nümbrecht, Germany) tubes, by assessment of the influence of these 3 brands on the PT and ISI. Five commercial thromboplastin reagents were used in the study. These represent different types 724 Am J Clin Pathol 2007;127: DOI: /NW194EAMD4WMEHJD

2 Coagulation and Transfusion Medicine / ORIGINAL ARTICLE of reagents: (1) plain rabbit brain thromboplastin (Neoplastin Plus, manufactured by Diagnostica Stago, Asnières, France, and provided by Roche Diagnostics Nederland, Almere, the Netherlands); (2) rabbit brain combined with adsorbed plasma (Hepato Quick, Diagnostica Stago; Roche Diagnostics Nederland); (3) thromboplastin derived from cultured human cells (Simplastin HTF, biomérieux, Durham, NC); (4) thromboplastin extracted from human placenta (Thromborel-S, Dade Behring); and (5) recombinant human thromboplastin (Innovin). In the present study, we showed that the blood collection tube had a statistically significant influence on the ISI of these reagents. A second purpose of the study was to assess the influence of crossover ISI on the INR calculation, ie, for samples collected with one type of tube and using the ISI determined with a different type of tube. Materials and Methods Blood Collection Tubes For the simultaneous comparison of the 3 blood collection systems, the following lots were used: 4.5-mL Vacutainer blood collection tubes made of siliconized glass and containing mol/l of buffered sodium citrate, lot ; Venosafe blood collection tubes made of polyethylene terephthalate and containing mol/l of buffered sodium citrate, lot (provided at no charge by Terumo Europe); and plastic S-Monovette blood collection tubes containing mol/l of sodium citrate, lot For primary ISI calibration of the thromboplastin reagents, different lots of the aforementioned brands were used. Thromboplastin Reagents Innovin (lot ) and Thromborel-S (lot ) were provided at no charge by Dade Behring, Simplastin HTF (lot ) by biomérieux, and Hepato Quick (lot ) and Neoplastin Plus (lot ) by Diagnostica Stago and Roche Diagnostics Nederland. ISI calibration of these reagents was performed by our laboratory using the international reference preparations for thromboplastin human recombinant (rtf/95), thromboplastin bovine combined (OBT/79), and thromboplastin rabbit plain (ERM-AD149). The first 2 preparations were obtained from the World Health Organization (WHO) laboratory for biological standards (Amsterdam, the Netherlands) and the last from the Institute for Reference Materials and Measurements (Geel, Belgium). ISI calibration was performed according to the WHO guidelines. 1 Instruments The following coagulation instruments were used: Sysmex CA-1500 (Toa Medical Electronics, Kobe, Japan), Coag-A- Mate MAX (biomérieux), and STA-R (Diagnostica Stago). Primary ISI Calibration Each thromboplastin reagent was used with one of the aforementioned instruments. First, each combination of reagent and instrument was calibrated with the appropriate international reference preparation. Innovin, Thromborel-S, and Simplastin HTF were calibrated with rtf/95, Hepato Quick was calibrated with OBT/79, and Neoplastin Plus with ERM-AD149. These primary ISI calibrations were not performed at the same time and not with the same set of freshly collected test samples. For each primary calibration, only 1 blood collection system was used, ie, the system routinely used by our anticoagulant clinic at that time. Simplastin HTF was calibrated in March 2004, and Innovin, Thromborel-S, and Hepato Quick in July Until September 2005, all blood specimens for primary ISI calibration were collected in Vacutainer tubes. Then, the routine blood collection system was changed to S-Monovette. Finally, primary calibration of Neoplastin Plus took place in November Comparison of Blood Collection Tubes The comparative study of the 3 blood collection systems was performed after the primary calibrations on 3 days in December On each day, venous blood samples of patients treated with vitamin K antagonists and nontreated control subjects were collected with the 3 aforementioned blood collection tubes from the same venipuncture in each subject. The order of filling of the tubes was changed from day to day. Nine volumes of blood were mixed with one volume of sodium citrate solution. The samples were transported to the laboratory and centrifuged at a force of 2,500g for 15 minutes at a controlled room temperature. Plasma was transferred into plastic tubes that were capped and stored at room temperature until testing. Single PT determinations were performed with each PT system. The 3 samples from each subject were tested in succession before the next subject s samples were analyzed. On each working day, the procedure was completed within 6 hours from the start of blood collection. The total numbers of patients and control subjects included in the study were 64 and 20, respectively. The patients were receiving long-term treatment with phenprocoumon or acenocoumarol. Two patient samples were excluded because the INRs were greater than 4.5. Statistical Methods ISI calibration was performed using orthogonal regression analysis as described previously. 1 For the simultaneous comparison of blood collection systems, differences in ISI between the blood collection systems were assessed by the method of Poggio et al. 9 Differences in PT and the PT ratio were assessed by using the t test on paired observations. Am J Clin Pathol 2007;127: DOI: /NW194EAMD4WMEHJD 725

3 van den Besselaar et al / BLOOD COLLECTION TUBES AND PT Calcium and Magnesium Measurements Calcium and magnesium measurements were made by using routine laboratory colorimetric methods on a Modular Analytics P800 System (Roche Diagnostics Nederland). Three tubes of each lot of blood collection tubes were analyzed. Results Primary ISI Calibrations The results of the primary ISI calibrations are shown in Table 1. The coefficients of variation of the regression line slopes were less than 3%, in accordance with the WHO requirements. 1 Effect of Blood Collection System on ISI The mean PTs obtained with the 3 blood collection tubes are shown in Table 2. For each thromboplastin, the mean patients PTs with the Vacutainer samples were significantly shorter than the corresponding values obtained with the Venosafe and S-Monovette tubes. The differences between Venosafe and S-Monovette were also significant, but not for Thromborel-S. Furthermore, the PT ratios (ie, the patients PT divided by the mean normal PT) calculated for the Vacutainer samples were significantly lower than those for the Venosafe and S-Monovette samples. The ISI values from the primary calibration were used to calculate the ISI for the other blood collection systems in the Table 1 Primary ISI Calibration of Prothrombin Time Systems * Reagent Instrument IRP Blood Collection System ISI CV of Slope (%) Simplastin HTF MAX rtf/95 Vacutainer Innovin CA-1500 rtf/95 Vacutainer Thromborel-S CA-1500 rtf/95 Vacutainer Hepato Quick STA-R OBT/79 Vacutainer Neoplastin Plus STA-R AD149 S-Monovette AD149, thromboplastin rabbit plain; CV, coefficient of variation; IRP, international reference preparation; ISI, international sensitivity index; OBT/79, thromboplastin bovine combined; rtf/95, thromboplastin human recombinant. * Manufacturers and suppliers were as follows: reagents: Hepato Quick (lot ) and Neoplastin Plus (lot ) manufactured by Diagnostica Stago (Asnières, France) and provided by Roche Diagnostics Nederland, Almere, the Netherlands; Innovin (lot ) and Thromborel-S (lot ), Dade Behring, Marburg, Germany; and Simplastin HTF (lot ), biomérieux, Durham, NC; instruments: Sysmex CA-1500, Toa Medical Electronics, Kobe, Japan; Coag-A-Mate MAX, biomérieux; and STA-R, Diagnostica Stago; and blood collection systems: S-Monovette, Sarstedt, Nümbrecht, Germany; and Vacutainer, Becton Dickinson Vacutainer Systems, Plymouth, England. Table 2 Mean PT for Patients and Healthy Subjects and ISI for Different Measurement and Blood Collection Systems * Mean PT (s) PT System/Blood Patients Healthy Subjects Mean PT Ratio Collection System (n = 62) (n = 20) (n = 62) ISI CV of Slope (%) Innovin/Sysmex CA-1500 Vacutainer Venosafe S-Monovette Thromborel-S/Sysmex CA-1500 Vacutainer Venosafe S-Monovette Simplastin HTF/MAX Vacutainer Venosafe S-Monovette Hepato Quick/STA-R Vacutainer Venosafe S-Monovette Neoplastin Plus/STA-R Vacutainer Venosafe S-Monovette CV, coefficient of variation; ISI, international sensitivity index; PT, prothrombin time. * Venosafe, Terumo Europe, Leuven, Belgium. For additional proprietary information, see Table 1. The CV of the calibration line slope was calculated for each blood collection system by comparison with the blood collection system used in the primary calibration (see also Figure 1). Difference with Vacutainer significant at the 1% level. Difference with Venosafe significant at the 1% level. Difference with Vacutainer significant at the 5% level. 726 Am J Clin Pathol 2007;127: DOI: /NW194EAMD4WMEHJD

4 Coagulation and Transfusion Medicine / ORIGINAL ARTICLE comparative study. For example, results obtained with Venosafe samples for Simplastin HTF were plotted against those obtained with Vacutainer samples Figure 1. The slope of the orthogonal regression line was multiplied with the ISI for the Vacutainer samples to obtain the ISI for Simplastin HTF and Venosafe samples. Figure 2 shows the test of statistical significance of the ISI difference between Vacutainer and Venosafe. The ISI values for each thromboplastin and blood collection system are given in Table 2. Relatively small differences in ISI were observed for Thromborel-S (2.2%) and Hepato Quick (1.1%) between Vacutainer and Venosafe tubes. For Neoplastin Plus, the difference in ISI was 5.5% between Vacutainer and S- Monovette tubes. The greatest differences in ISI were observed for Innovin (8.8%) and Simplastin HTF (8.3%) between Vacutainer and Venosafe tubes. Nearly all ISI differences were significant, except the differences between Venosafe and S- Monovette for Hepato Quick and Thromborel-S. Effect of Crossover ISI on INR To study the effect of an inappropriate ISI on the calculation of the INR, we used the ISI obtained with one blood collection system for the samples tested with the other 2 blood collection systems. As expected, the differences observed for Innovin and Simplastin HTF were the greatest Table 3. For Innovin, the mean INRs of all patient samples collected with Vacutainer tubes and with Venosafe were 2.68 and 2.94 if the ISI for Vacutainer tubes was used for the calculation. When the mean INRs obtained with different PT reagent systems were compared, the greatest difference was observed between Neoplastin Plus (INR 3.19) and Innovin (INR 2.46). Calcium and Magnesium Measurements Calcium and magnesium concentrations measured in the citrate solutions are shown in Table 4. Magnesium was not greater than 0.08 mmol/l in the S-Monovette and not greater than 0.15 mmol/l in the Venosafe, but its concentration in the Vacutainer tubes was at least 1.4 mmol/l. Discussion The present study demonstrates that commercial sodium citrate blood collection tubes have a statistically significant influence on the PT and ISI. With all reagents, the mean PTs of the coumarin-treated patients were shorter in Vacutainer samples than in Venosafe or S-Monovette samples (Table 2). Similar significant differences were observed with the PT Ln PT Vacutainer Ln PT Vacutainer Ln PT Venosafe Ln PT Venosafe Average Ln PT Vacutainer and Venosafe Figure 1 Scatter plot of fresh normal and coumarinanticoagulated plasma samples collected with 2 different evacuated tubes. The natural logarithms (Ln) of the prothrombin times (PT) determined with Simplastin HTF (biomérieux, Durham, NC) are plotted. The regression line for all data is shown. The coefficient of variation of the slope of the regression line was 0.5% (see also Table 2). Linear R 2 = Vacutainer, Becton Dickinson Vacutainer Systems, Plymouth, England; Venosafe, Terumo Europe, Leuven, Belgium. Figure 2 Difference plot of the natural logarithms (Ln) of the prothrombin times (PT) determined with Simplastin HTF (biomérieux, Durham, NC). The difference between Vacutainer (Becton Dickinson Vacutainer Systems, Plymouth, England) and Venosafe (Venosafe, Terumo Europe, Leuven, Belgium) is plotted against the mean value of each subject. The regression line for all data is shown. The correlation coefficient is 0.82, indicating that the slope of the line relating Ln(PT, Venosafe) and Ln(PT, Vacutainer) is significantly different from 1.0. Hence, the international sensitivity index difference between Venosafe and Vacutainer is significant at the 1% level. Linear R 2 = Am J Clin Pathol 2007;127: DOI: /NW194EAMD4WMEHJD 727

5 van den Besselaar et al / BLOOD COLLECTION TUBES AND PT ratios. The magnitude of the effect depends on the thromboplastin reagent. Some reagents such as Innovin and Simplastin HTF were more sensitive than others such as Thromborel-S and Hepato Quick. With Innovin and Simplastin HTF, the difference in ISI between Vacutainer and Venosafe was approximately 8%. Most likely, the short PT and high ISI obtained with the Vacutainer samples were caused by the relatively high concentration of magnesium ions contaminating the citrate solution (Table 4). Previous studies have shown that magnesium ions can shorten the PT, particularly if the magnesium concentration in the citrate solution is greater than 1.0 mmol/l. 7 The shortening of the PT was relatively small with Thromborel-S and Hepato Quick. With the Hepato Quick reagent, the plasma sample is diluted 21-fold in the final reaction mixture, whereas with the other reagents, only 3-fold dilution of the sample is achieved. The magnitude of the effect of magnesium ions in the Hepato Quick reaction mixture is, therefore, much less than in the other reagent mixtures. The relative insensitivity of Thromborel-S to the effect of magnesium ions is not completely understood. This reagent might contain a substance neutralizing the effect of magnesium ions. Some of the differences in PT and ISI between Venosafe and S-Monovette tubes were statistically significant but were smaller than those between Vacutainer and Venosafe or S- Monovette (Table 2). We cannot provide an explanation for these minor differences. Apparently, there were interfering factors other than the influence of magnesium ions. Our findings are in contrast with those reported by Fiebig and coworkers 10 who reported that PTs derived from blood Table 4 Calcium and Magnesium Ions Measured in the Sodium Citrate Solutions in Three Blood Collection Systems * Calcium Magnesium Blood Collection System (mmol/l) (mmol/l) Vacutainer lot S-Monovette lot Not detected Venosafe lot Not detected * The range of measured concentrations is given. Venosafe, Terumo Europe, Leuven, Belgium. For additional proprietary information, see Table 1. samples collected in 2 common types of plastic collection tubes (2.7-mL Vacutainer Plus and 3.5 ml Vacuette [Greiner Bio-One North America, Monroe, NC]) were shorter compared with paired samples collected in standard glass tubes (4.5-mL Monoject Blue Stopper, Tyco Kendall Medical, Mansfield, MA). In another study, no statistically significant differences were observed between glass (4.5-mL BD Vacutainer) and plastic (2.7-mL BD Vacutainer Plus) PT test results. 11 In contrast, Tripodi and coworkers 12 observed differences in ISI between siliconized glass and plastic BD Vacutainer tubes amounting to 5.7%. Proteolytic activation of factor VII resulting in shortening of the PT may be induced by contact with a glass surface. 13 Surface activation can be prevented by siliconization of the glass. 14 It is, therefore, unlikely that the shorter clotting times observed in the siliconized Vacutainer tubes were caused by surface activation. Table 3 Effect of Blood Collection System on INR in 62 Samples * Mean INR Calculated With PT System/Blood Collection System Mean PT Ratio Vacutainer ISI Venosafe ISI S-Monovette ISI Innovin/Sysmex CA-1500 Vacutainer Venosafe S-Monovette Thromborel-S/Sysmex CA-1500 Vacutainer Venosafe S-Monovette Simplastin HTF/MAX Vacutainer Venosafe S-Monovette Hepato Quick/STA-R Vacutainer Venosafe S-Monovette Neoplastin Plus/STA-R Vacutainer Venosafe S-Monovette INR, international normalized ratio; ISI, international sensitivity index; PT, prothrombin time. * The ISI values shown in Table 2 were used for the calculations. Venosafe, Terumo Europe, Leuven, Belgium. For additional proprietary information, see Table Am J Clin Pathol 2007;127: DOI: /NW194EAMD4WMEHJD

6 Coagulation and Transfusion Medicine / ORIGINAL ARTICLE We studied the effect of using an inappropriate ISI (crossover ISI) for the calculation of INR. If the manufacturer of a thromboplastin reagent determined the ISI with one type of blood collection tube and this ISI is used by clinical laboratories for samples collected with a different type of tube, there may be a bias in the calculated INRs. The differences in mean INR observed in the 3 blood collection tubes could be approximately 10% when the crossover ISI was used for a given thromboplastin-instrument combination (Table 3). The 10% difference in mean INR is at the border of clinical relevance. A more serious situation would arise if the manufacturer used one type of blood collection system for ISI calibration of reagent lot A but another type for calibration of lot B. In that case, the difference in ISI in combination with different blood collection systems used by clinical laboratories could lead to INR bias greater than 10% (Table 3). Furthermore, in some comparisons between thromboplastin reagent brands, the differences were greater than 10%, eg, Innovin-Vacutainer (mean, INR 2.68) compared with Neoplastin Plus Monovette (mean, INR 3.19). Part of the differences in the mean INR are caused by analytic errors such as ISI bias between international standards and imprecision of the calibration procedure itself. 15,16 It has been suggested that INR differences not greater than 10% do not seriously interfere with oral anticoagulant dosage regulation. 15 The present study showed that the total error in the mean INR was increased by the influence of different blood collection systems. The total error would be reduced if the influence of the blood collection systems could be eliminated. Although the influence of these blood collection tubes on a single ISI calibration is probably not of clinical importance, the combination of multiple systematic errors in successive calibration steps may lead to important INR differences. It should be realized that the calibration of thromboplastin reagents is a hierarchical system in which international reference standards are at the top of the hierarchy. 1 Systematic errors in successive ISI calibrations may accumulate to a level that could be clinically relevant. Therefore, we recommend that multicenter calibrations for the establishment of international reference standards be performed using blood collection systems with low levels of magnesium contamination. From the 1 Hemostasis and Thrombosis Research Center, Department of Hematology, and 2 Department of Clinical Chemistry, Leiden University Medical Center, Leiden, the Netherlands. Supported by the Federation of the Netherlands Thrombosis Services. Address reprint requests to Dr van den Besselaar: Hemostasis and Thrombosis Research Center, Dept of Hematology, C2-R, Leiden University Medical Center, PO Box 9600, 2300 RC Leiden, the Netherlands. Acknowledgments: We thank N. van Tilburg and V. van Marion for technical assistance. References 1. WHO Expert Committee on Biological Standardization. Guidelines for thromboplastins and plasma used to control oral anticoagulant therapy. World Health Organ Tech Rep Ser. 1999;889: Duncan EM, Casey CR, Duncan BM, et al. Effect of concentration of trisodium citrate anticoagulant on calculation of the international normalised ratio and the international sensitivity index of thromboplastin. Thromb Haemost. 1994;72: Adcock DM, Kressin DC, Marlar RA. Effect of 3.2% vs 3.8% sodium citrate concentration on routine coagulation testing. Am J Clin Pathol. 1997;107: Danielson CFM, Davis K, Jones G, et al. Effect of citrate concentration in specimen collection tubes on the international normalized ratio. Arch Pathol Lab Med. 1997;121: Chantarangkul V, Tripodi A, Clerici M, et al. Assessment of the influence of citrate concentration on the international normalized ratio (INR) determined with twelve reagentinstrument combinations. Thromb Haemost. 1998;80: van den Besselaar AMHP, Van Dam W, Sturk A, et al. Prothrombin time ratio is reduced by magnesium contamination in evacuated blood collection tubes. Thromb Haemost. 2001;85: van den Besselaar AMHP, Witteveen E, Meeuwisse-Braun J, et al. The influence of exogenous magnesium chloride on the apparent INR determined with human, rabbit, and bovine thromboplastin reagents. Thromb Haemost. 2003;89: van den Besselaar AMHP, Rutten WPF, Witteveen E. Effect of magnesium contamination in evacuated blood collection tubes on the prothrombin time test and ISI calibration using recombinant human thromboplastin and different types of coagulometer. Thromb Res. 2005;115: Poggio M, van den Besselaar AMHP, van der Velde EA, et al. The effect of some instruments for prothrombin time testing on the international sensitivity index (ISI) of two rabbit tissue thromboplastin reagents. Thromb Haemost. 1989;62: Fiebig EW, Etzell JE, Ng VL. Clinically relevant differences in prothrombin time and INR values related to blood sample collection in plastic vs glass tubes. Am J Clin Pathol. 2005;124: Kratz A, Stanganelli N, Van Cott EM. A comparison of glass and plastic blood collection tubes for routine and specialized coagulation assays: a comprehensive study. Arch Pathol Lab Med. 2006;130: Tripodi A, Chantarangkul V, Bressi C, et al. How to evaluate the influence of blood collection systems on the international sensitivity index: protocol applied to two new evacuated tubes and eight coagulometer/thromboplastin combinations. Thromb Res. 2003;108: Seligsohn U, Østerud B, Brown SF, et al. Activation of human factor VII in plasma and in purified systems: roles of activated factor IX, kallikrein, and activated factor XII. J Clin Invest. 1979;64: van den Besselaar AMHP, van Halem-Visser LP, Loeliger EA. The use of evacuated tubes for blood collection in oral anticoagulant control. Thromb Haemost. 1983;50: Poller L, Keown M, Chauhan N, et al. European Concerted Action on Anticoagulation: a multicentre calibration study of WHO international reference preparations for thromboplastin, rabbit (RBT/90) and human (rtf/95). J Clin Pathol. 2005;58: Chantarangkul V, van den Besselaar AMHP, Witteveen E, et al. International collaborative study for the calibration of a proposed international standard for thromboplastin, rabbit, plain. J Thromb Haemost. 2006;4: Am J Clin Pathol 2007;127: DOI: /NW194EAMD4WMEHJD 729