Potential Laboratory Misdiagnosis of Hemophilia and von Willebrand Disorder Owing to Cold Activation of Blood Samples for Testing

Transcription

1 Coagulation and Transfusion Medicine / MISDIAGNOSIS OF VON WILLEBRAND DISORDER AND HEMOPHILIA Potential Laboratory Misdiagnosis of Hemophilia and von Willebrand Disorder Owing to Cold Activation of Blood Samples for Testing Emmanuel J. Favaloro, PhD, Soma Soltani, and Jane McDonald Key Words: von Willebrand factor; vwf; von Willebrand disorder; von Willebrand disease; vwd; Hemophilia; Hemostasis testing; Diagnostic practice; Laboratory misdiagnosis; Test artifacts DOI: 1.139/E4947DG48TVY19C2 Abstract To assess the potential for misdiagnosis of von Willebrand disorder (vwd) and hemophilia A while following current National Committee for Clinical Laboratory Standards (NCCLS) guidelines and consequent to a poorly recognized cold-activation phenomenon, we processed 39 normal citrateanticoagulated samples by standard procedures (reference) or stored at low (~4 C) or ambient (~22 C) temperature for 3.5 hours before centrifugation and processing. Samples were tested in parallel for several hemostasis factors, including von Willebrand factor (vwf). Similar results were obtained for all samples for factors II, V, VII, IX, X, XI, and XII. For factor VIII (FVIII) and vwf, only samples stored at ambient temperature had results comparable to reference sample results. In most cases, low temperature storage led to much lower results. Taking the lower reference limit as 5%, most would have been defined as abnormal, and a misdiagnosis of vwd or hemophilia A could easily arise. ABO classification and age were associated with FVIII and vwf levels, but neither was associated conclusively with relative loss of plasma FVIII coagulant and vwf caused by the cold-activation phenomenon. We advise laboratories following current NCCLS guidelines not to store or transport whole blood samples for FVIII and vwf testing at 2 C to 4 C because of the risk of misdiagnosing vwd or hemophilia A. Storage and transport at ambient temperature seem acceptable and provide results comparable to freshly centrifuged samples. von Willebrand disorder (or disease [vwd]) is the most common inherited bleeding ailment. 1-3 Patients with vwd have defects in or reduced levels of von Willebrand factor (vwf), an adhesive plasma protein essential for primary hemostasis. 1-4 vwd is a heterogeneous disorder subtyped using clinical and laboratory criteria, 1-3 and the diagnostic process typically requires a panel of tests. 1,5-7 These typically include assessment of factor VIII coagulant (FVIII:C) activity, vwf antigen (vwf:ag), and functional vwf activity using the ristocetin cofactor and/or the collagen binding activity (vwf:cb) assays. 1,5-7 While vwd is a common form of hemophilia, hemophilia more usually is ascribed to discrete deficiencies or defects in various coagulation factors. For example, hemophilia A is the term used to describe a deficiency or defect in factor VIII, and hemophilia B is the term used to describe a deficiency or defect in factor IX. The misdiagnosis of hemophilia or vwd, for any reason, is a concern for all diagnostic laboratories and treating physicians. A lifetime label of either disease carries personal, health, and career and insurance costs and limitations that must be avoided if possible. Potential issues relating to test results at borderline or equivocal cutoff points are particularly relevant because they frequently arise. For example, our laboratory has highlighted the potential for misdiagnosis of vwd or hemophilia A following inappropriate testing of filtered plasma, 8,9 in which levels of vwf and FVIII:C are reduced compared with nonfiltered plasma. This situation might occur when the clinician or laboratory is following up cases of prolonged routine coagulation test times, perhaps with vague clinical histories, to exclude vwd, hemophilia A, and the presence of lupus anticoagulant. Inadvertent or inappropriate testing of vwf and FVIII:C using filtered plasma might occur Downloaded from by guest on 3 September Am J Clin Pathol 24;122: DOI: 1.139/E4947DG48TVY19C2

2 Coagulation and Transfusion Medicine / ORIGINAL ARTICLE because filtered plasma is an appropriate specimen type for assessment of lupus anticoagulant. Recently, our laboratory published studies 1,11 that identified that processing of whole blood samples according to National Committee for Clinical Laboratory Standards (NCCLS) guidelines current at that time 12 also might lead to the potential for a misdiagnosis of vwd, apparently due to a cold-activation phenomenon. The NCCLS guidelines (third edition) recommended that coagulation specimens be kept at 2 to 4 C or 18 to 24 C until centrifuged and tested (within 4 hours of collection). 12 The first pilot study provided evidence that short-term storage at the low temperature led to the potential misdiagnosis of vwd in a proportion of normal samples and to the potential misclassification of type 1 vwd (quantitative defect) as type 2 vwd (qualitative defect). 1 The second study 11 reported on a cross-laboratory (n = 44) evaluation that involved similar short-term storage of a single whole blood sample according to the NCCLS guidelines. 12 That study 11 used a blood group O normal sample known to be susceptible to the cold-activation phenomenon (as identified in the original study 1 ). This sample was differentially processed according to NCCLS guidelines, keeping half the whole blood tubes at the low temperature (~4 C) and the other half at the ambient temperature (~22 C) for 3.5 hours before centrifugation and processing. Samples were otherwise treated identically. For the sample stored at ambient temperature, multilaboratory test results aggregated at the bottom end of the normal reference range. It is important to note that most laboratories interpreted their test results as normal or equivocal/borderline. For the sample stored at low temperature, multilaboratory test results tended to be below 5%. Of significance, no laboratory interpreted test results as normal, despite the sample having been derived from a healthy person and processed according to NCCLS guidelines. Indeed, 16 (41%) of 39 laboratories identified this sample as being type 1 vwd and 16 (41%) identified it as being type 2 vwd. In the present study, we reinvestigated the impact of refrigerated whole blood samples (ie, the cold-activation phenomenon) as a potential for misdiagnosis of vwd and hemophilia A, as well as other clotting factors, using a large number of normal samples and also evaluated the influence of donor age and blood group. Materials and Methods Sample Preparation and Sample Testing The 39 samples used in the present study all were derived from healthy blood donors, with the characteristics identified in Table 1, and collected into standard (3.2% sodium citrate) Table 1 Characteristics of Test Sample Population All Men Women (n = 39) (n = 2) (n = 19) Median age 45 (18-68) 42 (19-68) 49 (18-6) (range), y Blood group O A B AB anticoagulant tubes (Greiner Vacuette, catalog ; Interpath, Sydney, Australia). Following receipt, 3 aliquots were differentially processed. The first aliquot was centrifuged immediately (2,g, 15 minutes) and processed by our standard laboratory practice (ie, plasma separated immediately from centrifuged sample and aliquots frozen at 8 C). These samples provided the reference sample. The second aliquot from each sample was kept at low temperature (~4 C) and the third aliquot at ambient temperature (~22 C) for 3.5 hours before centrifugation and processing, but otherwise, samples were treated identically to the reference samples. Sample Testing Plasma samples subsequently were thawed and tested in parallel using our standard laboratory test procedures for factor assays and vwf testing We included the following assays in our assessment: factors II, V, VII, VIII, IX, X, XI, and XII, vwf:ag, and vwf:cb. For testing factor levels, we used an ACL-3R instrument (IL-Coulter, Sydney, Australia). Factordeficient plasma was obtained from Dade-Behring, Sydney, Australia (for factors VII, VIII, IX, X, XI, and XII); Helena Laboratories, Melbourne, Australia (for factor II); and Stago via Bayer Diagnostics, Sydney, Australia (for factor V). Inhouse derived pooled normal plasma (PNP; containing >4 individual normal plasma samples) and commercial reference plasma (Verify Reference Plasma, BioMerieux via Vital Diagnostics, Sydney, Australia) were used in each assay run as reference control and calibration materials, respectively. vwf:ag was performed as previously described using rabbit antihuman vwf antibodies (DAKO, Sydney, Australia). vwf:cb was performed as previously described using ICN collagen (catalog ; ICN, Sydney, Australia; collagen used at a final concentration of 15 µg/ml to coat enzyme-linked immunosorbent assay wells). For both assays, the in-house PNP and a commercial reference plasma (Specialty Assayed Reference Plasma, Helena Laboratories) were used in each assay run as calibration and reference control materials, respectively. Additional assay control samples included a commercial low level sample (pathologic control sample; Specialty Assayed Control-2, Helena Laboratories), a vwf-deficient plasma sample (Biopool International via Abacus Diagnostics, Brisbane, Downloaded from by guest on 3 September 218 Am J Clin Pathol 24;122: DOI: 1.139/E4947DG48TVY19C2 687

3 Favaloro et al / MISDIAGNOSIS OF VON WILLEBRAND DISORDER AND HEMOPHILIA Australia), and an in-house derived cryosupernatant sample. For this study, the PNP was defined to contain 1% of all coagulation factors and 1% of vwf. Statistical Analysis Data were analyzed using traditional statistical methods and GraphPad Prism software (Prism 2.C, GraphPad Software, San Diego, CA). Results Summary assay results are shown in Figure 1 and Figure 2. Similar assay results were obtained for all test samples for factors II, V, VII, IX, X, XI, and XII (ie, reference sample and samples stored at low [~4 C] and ambient [~22 C] temperatures). In contrast, for FVIII and vwf testing, only samples stored at ambient temperature provided test results comparable to those of the reference samples. Samples stored at low temperature provided much lower test results in most cases. Indeed, taking the lower reference limit of assay results as 5%, most could be defined as abnormal (ie, <5%). Accordingly, a potential for misdiagnosis of vwd or hemophilia A could arise easily. For vwf, the observed effect seemed to be slightly greater with the functional vwf assay (ie, vwf:cb) than with the nonfunctional vwf assay (ie, vwf:ag). Consequently, testing of most samples was accompanied by a rise in the vwf:ag/vwf:cb ratio (Figure 2). In 2 cases, the derived D E F G H Figure 1 Temperature storage treatment study results for factors II (A), V (B), VII (C), VIII (D), IX (E), X (F), XI (G), and XII (H). Results are shown for reference samples (Ref), samples stored at ambient temperature (22 C), and samples stored at low temperature (4 C). There were no significant differences between the levels detected for reference samples vs storage treatment at ambient or low temperature for factors II, V, VII, IX, X, XI, and XII (P >.1 for each comparison; 2-tailed Mann- Whitney test). There was no significant difference between the levels detected for reference samples vs storage treatment at ambient temperature for factor VIII (P =.4155). In contrast, there was a significant difference between the levels detected for reference samples vs storage treatment at low temperature for factor VIII (P <.1). Downloaded from by guest on 3 September Am J Clin Pathol 24;122: DOI: 1.139/E4947DG48TVY19C2

4 Coagulation and Transfusion Medicine / ORIGINAL ARTICLE ratio was more than 2., and this often is used to define a vwf functional discordance (ie, identify a qualitative vwd defect such as type 2 vwd). Because the ABO blood group system has been described as a determinant of plasma FVIII and vwf levels, we also assessed the influence of this system on the cold-activation phenomenon. Figure 3A shows the relationship between ABO and plasma levels of FVIII and vwf in our study group. There was a significant difference between levels of plasma FVIII and vwf as detected in samples from the O blood group vs those from non-o blood groups (ie, for each study parameter, FVIII:C, vwf:ag, and vwf:cb). However, there was no relationship between the ABO blood group and the relative loss of plasma FVIII:C and vwf caused by cold activation Figure 3B. Interestingly, there was a differing tendency in terms of the absolute loss of plasma FVIII:C and vwf Figure 3C. Because increasing age also has been described as a determinant of plasma FVIII and vwf levels, we similarly assessed the influence of donor age on the cold-activation phenomenon. Figure 4 shows the relationship between donor age and plasma levels of FVIII and vwf in our study group. There was a small but statistically significant association between levels of plasma vwf and donor age. However, there was no relationship between donor age and the relative loss of plasma FVIII:C and vwf caused by cold activation Figure Figure 2 Temperature storage treatment study results for von Willebrand factor (vwf) parameters (A, vwf antigen [Ag]; B, vwf collagen binding activity [CB]; and C, vwf:ag/vwf:cb ratio). Data are shown for reference samples (Ref), samples stored at ambient temperature (22 C), and samples stored at low temperature (4 C). There were no significant differences between the levels detected for reference samples vs storage treatment at ambient temperature for any parameter (ie, vwf:ag, vwf:cb, Ag/CB; P >.8 for each comparison; 2-tailed Mann-Whitney test). In contrast, there were significant differences between the levels detected for reference samples vs storage treatment at low temperature for each parameter (P <.1 in each case) n-o O n-o O n-o O FVIII:C vwf:ag vwf:cb 5 n-o O n-o O n-o O FVIII:C vwf:ag vwf:cb % Absolute Loss vwf:ag/vwf:cb Ratio 5 n-o O n-o O n-o O FVIII:C vwf:ag vwf:cb Figure 3 Relationship between ABO blood group and study parameters. Data are for people with blood group O vs people with non-o blood group for each study parameter (ie, factor VIII coagulant [FVIII:C], von Willebrand factor [vwf] antigen [Ag], and vwf collagen binding activity [CB]). Error bars represent 95% confidence intervals. A, Relationship between ABO and plasma levels of FVIII:C, vwf:ag, and vwf:cb. O blood group data were statistically different (lower) from non-o blood group data (P <.5 for each comparison; 2-tailed Mann-Whitney test). B, Relationship between ABO and percentage of relative loss (calculated as [(Reference Level 4 C Level)/Reference Level] 1) of plasma FVIII:C, vwf:ag, and vwf:cb. Data for the O blood group were not statistically different from data for the non-o blood groups (P >.6 for each comparison; 2-tailed Mann-Whitney test). C, Relationship between ABO and absolute percentage of loss (calculated as Reference Level 4 C Level) of plasma FVIII:C, vwf:ag, and vwf:cb. O blood group were data statistically different from data for the non-o blood groups for vwf:ag (P =.195) and approached significance for vwf:cb (P =.722). Data were not significantly different for FVIII:C (P =.3993). Downloaded from by guest on 3 September 218 Am J Clin Pathol 24;122: DOI: 1.139/E4947DG48TVY19C2 689

5 Favaloro et al / MISDIAGNOSIS OF VON WILLEBRAND DISORDER AND HEMOPHILIA Discussion The misdiagnosis of hemophilia or vwd, for any reason, is a concern for all diagnostic laboratories and treating physicians. A lifetime label of either disease carries personal, health, and career and insurance costs and limitations that must be avoided if possible. We reassessed the potential for misdiagnosis of vwd or hemophilia A consequent to a poorly recognized cold-activation phenomenon using a large number of normal samples. In most test cases, whole blood samples held at low temperature before centrifugation led to a detectable loss of FVIII:C and vwf and to the potential for misdiagnosis of these samples as vwd or hemophilia A, depending on the testing undertaken (ie, vwf studies ordered or only FVIII ordered, respectively). This finding, therefore, identifies a cause of possible misdiagnosis in addition to those previously reported 8,9 and extends the findings of previous reports regarding the cold-activation phenomenon. 1,11 A potential was shown for misdiagnosis or misclassification of vwd using a small number of samples processed according to the thencurrent NCCLS guidelines, 12 which recommended that coagulation specimens be kept at 2 to 4 C or 18 to 24 C until centrifuged and tested (within 4 hours of collection) The NCCLS guidelines recently have been updated (fourth edition). 18 One of the modifications indicates that Specimens for PT [prothrombin time] assays uncentrifuged or centrifuged with plasma remaining on top of the cells in an unopened tube kept at 18 to 24 C should be tested within 24 hours from time of specimen collection. Storage at 2 to 4 C may result in cold activation of factor VII and therefore alter PT results. There is as yet no modification to the other test guidelines noting any other effect of cold activation. We recently brought the results of previous published findings to the attention of the NCCLS. The present results further document the need for additional modification of the guidelines to identify the potential for cold-activation effects on FVIII and vwf. Some time ago, others also noted cold-activation effects on vwf, 19,2 but findings seem not to have been followed up. The reason for the cold-activation effect remains unknown, although other substances have been implicated, 1,11 likely vwf proteases and/or reductases, in the phenomenon. Thus, whole blood exposure to cold probably activates the platelet or other cellular release of vwf-cleaving proteases (such as ADAMTS13) or vwf reductases (such as thrombospondin), which in turn degrade plasma vwf. In support of Figure 4 Relationship between donor age and factor VIII and von Willebrand factor (vwf). A, Relationship between factor VIII coagulant and donor age (P =.634; linear regression model). B, Relationship between vwf antigen and donor age (P =.346; linear regression model). C, Relationship between vwf collagen binding activity and donor age (P =.64; linear regression model). 5 5 Downloaded from by guest on 3 September 218 Figure 5 Relationship between donor age and relative loss (calculated as [(Reference Level 4 C Level)/Reference Level] 1) of plasma factor VIII and von Willebrand factor (vwf) following cold activation. A, Relationship between factor VIII coagulant and relative loss (P =.4157; linear regression model). B, Relationship between vwf antigen and relative loss (P =.6921; linear regression model). C, Relationship between vwf collagen binding activity and relative loss (P =.6341; linear regression model). 69 Am J Clin Pathol 24;122: DOI: 1.139/E4947DG48TVY19C2

6 Coagulation and Transfusion Medicine / ORIGINAL ARTICLE this, the decrease in levels of vwf:cb (a marker of highmolecular-weight vwf) seems to be higher than that of vwf:ag. The fall in the plasma FVIII:C level also is likely to be due to the loss of vwf, although a more direct effect on FVIII:C cannot be discounted. Because the ABO blood group system has been described as a determinant of plasma FVIII and vwf levels, we also assessed the influence of the ABO system on the cold-activation phenomenon. This is particularly relevant in view of the association described between vwf proteolysis by ADAMTS13 and the ABO group. 27 In support of previous findings, we also found a significant difference between levels of plasma FVIII and vwf as detected in samples from the O blood group vs samples from non-o blood groups (Figure 3A). We also determined a differing tendency in terms of the absolute loss of plasma FVIII:C and vwf following cold storage (Figure 3C). However, there was no relationship between the ABO blood group and the relative loss of plasma FVIII:C and vwf caused by cold storage, as shown in Figure 3B. Thus, the differing tendency in terms of the absolute loss is likely to be related simply to the ABO group influence on plasma vwf and FVIII (ie, starting levels or amounts of vwf and FVIII), rather than an influence on vwf and FVIII proteolysis per se. In any case, it is likely that a larger number of test samples would be required to more definitively answer this question. Because increasing age also has been described as a determinant of plasma FVIII and vwf levels, 28 we assessed the influence of donor age on the cold-activation phenomenon. Again, in support of previous findings, 28 there was a small but statistically significant association between levels of plasma vwf and donor age (Figure 4). However, there was no relationship between donor age and the relative loss of plasma FVIII:C and vwf caused by cold activation (Figure 5). We advise laboratories following current NCCLS guidelines and testing for FVIII and vwf not to store or transport whole blood samples refrigerated at 2 C to 4 C because of the potential risk of misdiagnosing vwd or hemophilia A. Storage and transport at ambient temperature (eg, 22 C) seem to be acceptable and provide test results comparable to those for freshly centrifuged samples. While cold activation did not influence results of other factor assays, storage and transport at ambient temperature for these also would be acceptable, given that no significant loss of any factor was identified in the present study under these conditions. Study findings are summarized in Table 2. Recommendations We make the following suggestions for modification of the NCCLS guidelines, 18 as additional points under section 5.2 Storage: Table 2 Cold Activation Effects on Hemostasis Factors Factor II V VII VIII IX X XI XII von Willebrand Affected by Cold Activation? Yes Yes Specimens for FVIII and von Willebrand factor (vwf) testing uncentrifuged in an unopened tube kept at 18 C to 24 C should be centrifuged and processed within 4 hours from the time of collection. Storage at 2 C to 4 C might result in the loss of FVIII and vwf and, therefore, adversely affect a diagnosis of hemophilia or von Willebrand disorder (vwd). Specimen processing indicates centrifugation and immediate testing or separation of plasma from the cell pellet for freezing and later testing. Specimens for other factor testing (factors II, V, VII, IX, X, XI, and XII) uncentrifuged in an unopened tube kept at 2 C to 4 C or 18 C to 24 C should be centrifuged and processed within 4 hours from the time of collection. Specimen processing indicates centrifugation and immediate testing or separation of plasma from the cell pellet for freezing and later testing. In addition, the last point in section 5.2, 18 Specimens for other assays (eg, thrombin time, protein C, factor V, factor VIII) kept at 2 to 4 C or 18 to 24 C, should be centrifuged and tested within 4 hours from time of collection should delete the reference to factors V and VIII. From the Department of Hematology, Institute of Clinical Pathology and Medical Research, Westmead Hospital, Westmead, Australia. Address reprint requests to Dr Favaloro: Dept of Haematology, Institute of Clinical Pathology and Medical Research (ICPMR), WSAHS, Westmead, NSW, 2145, Australia. References 1. Sadler JE, Mannucci PM, Berntorp E, et al. Impact, diagnosis and treatment of von Willebrand disease. Thromb Haemost. 2;84: Sadler JE, Gralnick HR. Commentary: a new classification for von Willebrand disease. Blood. 1994;84: Sadler JE, for the Subcommittee on von Willebrand Factor of the Scientific and Standardisation Committee of the International Society on Thrombosis and Haemostasis. A revised classification of von Willebrand disease. Thromb Haemost. 1994;71: Ruggeri ZM. Structure and function of von Willebrand factor. Thromb Haemost. 1999;82: Downloaded from by guest on 3 September 218 Am J Clin Pathol 24;122: DOI: 1.139/E4947DG48TVY19C2 691

7 Favaloro et al / MISDIAGNOSIS OF VON WILLEBRAND DISORDER AND HEMOPHILIA 5. Favaloro EJ. Appropriate laboratory assessment as a critical facet in the proper diagnosis and sub-classification of von Willebrand s disorder. Baillieres Best Pract Res Clin Haematol. 21;14: Favaloro EJ. Laboratory assessment as a critical component of the appropriate diagnosis and sub-classification of von Willebrand s disease. Blood Rev. 1999;13: Favaloro EJ. von Willebrand factor (vwf) collagen binding (activity) assay (vwf:cba) in the diagnosis of von Willebrand s disorder (vwd): a 15-year journey. Semin Thromb Hemost. 22;28: Favaloro EJ, Mohammed A, Coombs R, et al. Filtered plasma as a potential cause of clinical misdiagnosis: inappropriate testing in a haematology laboratory. Br J Biomed Sci. 1995;52: Favaloro EJ, Mehrabani PA, Koutts J. Laboratory assessment of von Willebrand factor: altered interpretation of laboratory data, and altered diagnosis of von Willebrand s disease, as influenced by the use of different vwf assays and assay conditions. Clin Appl Thromb Hemost. 1997;3: Favaloro EJ, Nair SC, Forsyth CJ. Collection and transport of samples for laboratory testing in von Willebrand s disease (vwd): time for a reappraisal [letter]? Thromb Haemost. 21;86: Favaloro EJ, Bonar R, Sioufi J, et al, on behalf of the RCPA QAP in Haematology Haemostasis Committee. Laboratory diagnosis of von Willebrand disorder: current practice in the Southern Hemisphere. Am J Clin Pathol. 23;119: Arkin CF, Bowue EJW, Carroll JJ, et al. Collection, Transport, and Processing of Blood Specimens for Coagulation Testing and General Performance of Coagulation Assays; Approved Guideline. 3rd ed. Wayne, PA: National Committee for Clinical Laboratory Standards; Mohammed A, Mehrabani PA, Coombs R, et al. The MDA- 18 coagulation analyser: a laboratory evaluation. Pathology. 1997;29: Favaloro EJ, Grispo L, Dinale A, et al. von Willebrand s disease: laboratory investigation using an improved functional assay for von Willebrand factor. Pathology. 1993;: Favaloro EJ, Koutts J. Laboratory assays for von Willebrand factor: relative contribution to the diagnosis of von Willebrand s disease. Pathology. 1997;29: Favaloro EJ. Collagen binding assay for von Willebrand factor (vwf:cba): detection of von Willebrand s disease (vwd), and discrimination of vwd subtypes, depends on collagen source. Thromb Haemost. 2;83: Favaloro EJ, Henniker A, Facey D, et al. Discrimination of von Willebrand s disease (vwd) subtypes: direct comparison of von Willebrand factor:collagen binding activity/assay (vwf:cba) with monoclonal antibody (MAB) based ELISA vwf-detection systems. Thromb Haemost. 2;84: Arkin CF, Adcock DM, Ernst DJ, et al. Collection, Transport, and Processing of Blood Specimens for Testing Plasma-Based Coagulation Assays; Approved Guideline. 4th ed. Wayne, PA: National Committee for Clinical Laboratory Standards; Farrugia A, Street A, Douglas S, et al. Stabilization of von Willebrand factor in banked blood by leucocyte depletion. Transfus Med. 1993;3: Lloyd JV, Tunbridge LJ, Rodgers SE, et al. Destruction of ristocetin cofactor by coagulation at 4 C. Thromb Res. 1985;37: Obert B, Tout H, Veyradier A, et al. Estimation of the von Willebrand factor cleaving protease in plasma using monoclonal antibodies to vwf. Thromb Haemost. 1999;82: Gerritsen HE, Turecek PL, Schwarz HP, et al. Assay of von Willebrand factor (vwf)-cleaving protease based on decreased collagen binding affinity of degraded vwf: a tool for the diagnosis of thrombotic thrombocytopenic purpura (TTP). Thromb Haemost. 1999;82: Xie L, Chesterman CN, Hogg PJ. Reduction of von Willebrand factor by endothelial cells. Thromb Haemost. 2;84: O Donnell J, Laffan MA. The relationship between ABO histo-blood group, factor VIII and von Willebrand factor. Transfus Med. 21;11: Haley E, Babar N, Ritter C, et al. Effect of ABO blood group on the collagen-binding assay for von Willebrand factor. Am J Hematol. 22;71: Miller CH, Haff E, Platt SJ, et al. Measurement of von Willebrand factor activity: relative effects of ABO blood type and race. J Thromb Haemost. 23;1: Bowen DJ. An influence of ABO blood group on the rate of proteolysis of von Willebrand factor by ADAMTS13. J Thromb Haemost. 23;1: Rodeghiero F. von Willebrand disease: still an intriguing disorder in the era of molecular medicine. Haemophilia. 22;8: Downloaded from by guest on 3 September Am J Clin Pathol 24;122: DOI: 1.139/E4947DG48TVY19C2