Identification of von Willebrand Disease Type 2N (Normandy) in Australia A Cross-Laboratory Investigation Using Different Methods

Coagulation and Transfusion Medicine / CROSS-LABORATORY INVESTIGATION OF VON WILLEBRAND DISEASE TYPE 2N Identification of von Willebrand Disease Type 2N (Normandy) in Australia A Cross-Laboratory Investigation Using Different Methods Susan E. Rodgers, MApplSc, 1 Nancy V. Lerda, 2 Emmanuel J. Favaloro, PhD, 3 Elizabeth M. Duncan, MApplSc, 1 Graeme J. Casey, 4 Diana M. Quinn, PhD, 2 Mark Hertzberg, PhD, FRCPA, FRACP, 3 and John V. Lloyd, PhD, FRACP 1 Key Words: von Willebrand disease, type 2N; Factor VIII assays; Factor VIII binding; Two-stage factor VIII Abstract We report on a cross-laboratory study of type 2N von Willebrand disease (vwd). We tested 101 selected plasma samples for factor VIII and factor VIII binding activity of von Willebrand factor (vwf). Of these plasma samples, 31 were cotested by 2 specialist centers using different detection procedures for vwf factor VIII binding: there was good agreement between results obtained by chromogenic assay and enzyme-linked immunosorbent assay. In total, 8 patients with type 2N vwd were identified. The 2-stage factor VIII assay detected a deficiency of factor VIII relative to vwf antigen in all 8 patients; the 1-stage factor VIII assay detected a relative deficiency in only 3 patients. Four patients were homozygous for the most common type 2N mutation (R854Q), 3 patients were presumed to be compound heterozygotes, and in 1 patient no type 2N mutations were identified. In this study of patients from 5 specialist centers in Australia, type 2N vwd was found in 5 families. The 2- stage factor VIII assay was more useful as a screening test than the 1-stage assay, and both vwf factor VIII binding assays were equally effective. von Willebrand disease (vwd) is a heterogeneous disorder, and several types have been described. The most common, type 1, is a quantitative autosomal dominant defect, in which there are proportional reductions in the levels of the von Willebrand protein (von Willebrand factor [vwf] antigen [Ag]), von Willebrand function (as measured by ristocetin cofactor [RCoF] activity or collagen-binding activity), and factor VIII. Type 3 vwd is a severe, autosomal recessive form of vwd, with virtually unmeasurable levels of vwf:ag, RCoF activity, collagen-binding activity, and factor VIII. Some patients with the clinical and laboratory features of mild type 1 vwd have been found to be carriers of type 3 vwd. 1 The qualitative disorders of vwd (type 2) have been classified as type 2A, 2B, 2M, and 2N, depending on details of the phenotype. 1 Type 2N vwd (vwd Normandy ) was first described in 1989 as a variant of vwd and found to be caused by a defect in the ability of vwf to bind factor VIII. 2 Because of this binding defect, factor VIII is not stable and has a reduced half-life in the circulation. In this autosomal recessive subtype, the results of measurements of vwf:ag and RCoF activity or collagen-binding activity usually are normal, but there is a reduction in the level of factor VIII. The results of these tests, therefore, mimic those of mild hemophilia A (male) or of hemophilia A carriers (female). However, type 2N vwd also has been reported in several patients who had previously been diagnosed with type 1 vwd. 3-5 In such cases, vwf:ag and RCoF activity levels have been borderline or mildly reduced, and the disproportionately lower factor VIII had been thought to be due to the variation in phenotype that occurs in type 1 vwd, even in members of the same family. 1 Where there is some doubt American Society for Clinical Pathology Am J Clin Pathol 2002;118:269-276 269

Rodgers et al / CROSS-LABORATORY INVESTIGATION OF VON WILLEBRAND DISEASE TYPE 2N about the diagnosis of mild hemophilia A or type 1 vwd, the possibility of type 2N vwd can be confirmed or excluded by measuring the factor VIII binding ability of vwf. 3,4 When the factor VIII binding ability of vwf is severely reduced (indicative of type 2N vwd), the diagnosis usually can be confirmed by DNA sequence analysis of exons 18 to 24 of the vwf gene, the region that has been shown to code for the factor VIII binding domain of vwf. 4 In 2 previous surveys of laboratories involved in testing of samples for vwd, a wide variation in the ability to detect vwd was identified. 6,7 Of particular relevance, measurement of factor VIII (by 1-stage assay) and vwf:ag, as screening assays, did not effectively detect the potential presence of type 2N vwd. 7 In addition, factor VIII measured by 2-stage assay has been found to be lower than that measured by 1-stage assay in some patients with hemophilia 8 and vwd. 9 The 2 main objectives of the present study were to compare the usefulness of the 1-stage and 2- stage factor VIII assays for the preliminary identification (screening) of patients who may have type 2N vwd and to compare the type 2N diagnostic capacity of 2 different methods to assess the factor VIII binding ability of vwf, as used by 2 specialist laboratories. In addition, we also wanted to determine whether type 2N vwd could explain the low or borderline factor VIII levels occasionally seen in patients with an inconclusive diagnosis. Materials and Methods Blood Samples Blood samples were obtained from patients for routine assays of factor VIII and vwf, into one tenth volume of a 109-mmol/L concentration of trisodium citrate anticoagulant. Platelet-poor plasma was prepared by centrifugation and stored in aliquots at 70 C. Most tests were conducted using stored plasma from previous clinical evaluations. Tests for factor VIII were conducted on plasma that had not been thawed previously. Other tests were conducted on previously unthawed plasma when possible or plasma refrozen only once or twice. Fresh samples were obtained for genetic analysis, with informed consent. Ethical approval was obtained from the Institute of Medical and Veterinary Science Royal Adelaide Hospital Ethics Committee (Adelaide, South Australia). A total of 101 patient samples were tested, selected on the basis of previous laboratory results and clinical history. These were from patients with a clinical or laboratory indication of type 2N vwd, isolated mild hemophilia A, a variant of vwd, or an inconclusive diagnosis. Twelve patients with mild hemophilia A, in whom the causative mutation had been identified previously, were included for control purposes. The remaining 89 samples were selected for testing from 5 specialist centers comprising South Australia (61 samples), New South Wales (3 samples), Queensland (6 samples), Western Australia (12 samples), and Victoria (7 samples). Of these samples, 31 were cotested by 2 specialist vwd reference laboratories. They comprised 10 samples from patients in whom there was a strong clinical suspicion of type 2N vwd, 6 from patients with vwd, 9 from patients with mild hemophilia A in whom the causative mutation had been identified (control group), 5 from patients with isolated mild hemophilia A in whom the mutation had not been determined, and 1 from a patient thought to have both hemophilia and vwd. Factor VIII Coagulation Assays Laboratory A The 2-stage factor VIII was performed by a semiautomated method using a Coag-a-Mate X2 (Organon Teknika, Durham, NC) and reagents from Diagen (Oxon, England) as previously described. 8,10 Citrated standard and test plasma samples were aluminum hydroxide adsorbed, and doubling dilutions were made in imidazole buffer containing 0.5% bovine serum albumin. Second Australian Standard for Factor VIII (Commonwealth Serum Laboratories, Melbourne, Australia) was used as the standard plasma that was diluted 1:60 to 1:960, and patient samples were tested at 3 of these dilutions. Each dilution (50 µl) was added in duplicate to a test tray. The factor VIII reagent (200 µl) was added and incubated for 12.5 minutes at 37 C before addition of the substrate plasma (100 µl) and detection of the clot. Patient results were calculated from the standard curve, using a log-log polynomial regression program. The 1-stage factor VIII assay was performed by a standard automated method using an MLA Electra 1000C (Medical Laboratory Automation, Pleasantville, NY). Coagulation Reference Plasma Normal (Realm Biomedical, Brisbane, Australia) was diluted in Owren buffer (5.88 g/l sodium diethylbarbiturate, 7.3 g/l sodium chloride, ph 7.35) in doubling dilutions from 1:10 to 1:320. Patient samples were tested at 2 dilutions, usually 1:10 and 1:20. Plasma dilutions (50 µl, in duplicate), 50 µl of factor VIII deficient plasma (Organon Teknika), and 50 µl of actin FSL activated partial thromboplastin time reagent (Dade Behring, Marburg, Germany) were incubated at 37 C for 3 minutes before addition of 50 µl of a 25-mmol/L concentration of calcium chloride and detection of the clot. Patient results were calculated from the standard curve, using a log-log polynomial regression program. 270 Am J Clin Pathol 2002;118:269-276 American Society for Clinical Pathology

Coagulation and Transfusion Medicine / ORIGINAL ARTICLE vwf Assays Laboratory A vwf:ag was assayed by enzyme-linked immunosorbent assay (ELISA), using rabbit antibodies to human vwf (DAKO, Glostrup, Denmark). RCoF activity was assayed using 1 mg/ml of ristocetin and formalin-fixed washed platelets in a PACKS-4 Platelet Aggregometer (Helena Laboratories, Beaumont, TX), according to the manufacturer s instructions. Collagen-binding activity was assayed by ELISA as previously described, 11 using collagen from ICN Biomedicals, Aurora, Ohio. Laboratory B VWF:Ag and collagen-binding activity assays were performed by ELISA as previously reported. 11,12 Factor VIII Binding Activity of vwf Laboratory A Chromogenic detection of bound factor VIII was used according to the method of Nesbitt et al 3 with some modifications, as indicated. Briefly, microtiter plates (Nunc- Immuno Module Maxisorp C12 framed, Nunc, Roskilde, Denmark) were coated with vwf antibody (rabbit immunoglobulin to human vwf (DAKO), 9 µg/ml in a 0.05-mol/L concentration of carbonate buffer, ph 9.6, to capture patient vwf. Reagent volumes of 100 µl were used for all steps except the chromogenic visualization. Pooled normal plasma used for the standard line was diluted 1:100, 1:200, 1:400, 1:800, 1:1,600, and 1:3,200. Patient samples were used at 2 dilutions in the range 1:100 to 1:800. All samples were tested in duplicate. After overnight incubation at 4 C, endogenous factor VIII was removed by 2 successive 1-hour incubations at room temperature with a 0.35-mol/L concentration of calcium chloride. Recombinant factor VIII (0.1 IU/mL; Recombinate, Baxter Healthcare, Deerfield, IL) was incubated for 2 hours at 37 C, and the amount bound was determined using a factor VIII chromogenic assay kit (Dade Behring), modified for use in microtiter plates as follows: Reagents 1 and 2 (20 µl of each) were incubated at room temperature with shaking for 10 minutes. Reagent 3 (reconstituted in 10 ml) was added (80 µl) and incubated for another 5 minutes before the reaction was stopped by incubation for 1 minute with a 1-mol/L concentration of citric acid (20 µl). Optical density was read at 410 nm using a microtiter plate reader (Dynatech, Guernsey, Channel Islands), and the standard line was drawn manually on loglinear paper. A representative standard curve is shown in Figure 1A. Binding activity of the test samples, expressed as units per milliliter, was determined by comparison with the standard line, in a way similar to that used for manual A OD (410 nm) B OD (450 nm) 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0.3 0.2 0.1 0.0 3.13 6.25 12.5 25 50 100 Factor VIII Binding (U/dL) 6.25 12.5 25 50 100 Factor VIII Binding (U/dL) Figure 1 Representative standard curves for the factor VIII binding assay. The optical density (OD) reading for the blank (where no plasma was added to the dilution buffer) for each assay is shown by the dashed line and was 0.098 for laboratory A (A) and 0.055 for laboratory B (B). Pooled normal plasma was used as the standard and assigned an arbitrary value of 1.0 U/mL. factor assay calculations. The activity of the pooled normal plasma (prepared from 40 Red Cross donors) used as the standard was assigned an arbitrary value of 1.0 U/mL. Laboratory B The ELISA as performed was technically similar to that described by Casanato et al. 13 In general, this is a modification of a standard assay of the factor VIII binding ability of vwf using sheep antihuman factor VIII (IgG-peroxidase conjugated; Cedarlane Laboratories, Hornby, Ontario) according to manufacturer s recommendations to detect American Society for Clinical Pathology Am J Clin Pathol 2002;118:269-276 271

Rodgers et al / CROSS-LABORATORY INVESTIGATION OF VON WILLEBRAND DISEASE TYPE 2N factor VIII bound to vwf (ie, instead of the chromogenic detection method), and then tetramethylbenzidine dihydrochloride substrate as previously described. 11 In brief, standard 96-well plates (EIA plates, ICN Biomedical, Sydney, Australia) first were coated with rabbit antisera to human vwf (DAKO, Sydney, Australia), as for a standard vwf:ag assay using a 1:1,000 dilution in a high ph carbonate buffer. 11 After washing (3 times) using standard wash buffer, 11 diluted plasma (1:100) was added. After further washing (3 times), the plate was incubated with a 0.4-mol/L concentration of calcium chloride to strip endogenous factor VIII bound to vwf. After rewashing (3 times), a standard concentration of recombinant factor VIII was added to permit attachment to the plate-bound vwf in all samples. After further washing (3 times), the plate was differentially exposed to either peroxidase-conjugated rabbit anti-vwf (DAKO) or peroxidase-conjugated sheep antihuman factor VIII (Cedarlane). After a final washing step (3 times), tetramethylbenzidine dihydrochloride substrate was added, and color generation was detected as for a standard vwf:ag assay. 11 All testing was performed in triplicate. The standard curve was generated using serial dilutions of pooled normal plasma (using a pool from >60 persons). A representative standard curve is shown in Figure 1B. Reference Range For this study, the reference range for the assay for the factor VIII binding ability of vwf was determined to be 60 to 245 U/dL, based on studies performed at laboratory A (n = 22). The factor VIII binding ratio (ratio of factor VIII binding ability of vwf to vwf:ag, expressed as a percentage) was determined to have a reference range of 60% to 220% (again based on studies performed at laboratory A; n = 22). Mutation Detection The presence of mutations in the factor VIII binding domain of vwf was studied by direct exon sequencing of polymerase chain reaction (PCR) products amplified from genomic DNA (isolated from peripheral blood leukocytes) using published primers. 4 Sequencing was performed using an ABI PRISM Dye Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) and an Applied Biosystems 373 or 377 Excel Automated DNA sequencer, according to the manufacturer s directions. The resulting sequences were compared with the reference sequence (Genbank accession numbers M25842-M25847) using Sequence Navigator (Applied Biosystems). (Amino acids are numbered from the translation initiation codon 14 ; nucleotides within exons are numbered from the site of transcription initiation, in accordance with the convention used in the vwf database on the World Wide Web at http://mmg2.im.med.umich.edu/vwf/nomenclature.htm). A section of exon 7 of the factor VIII gene, previously found to contain mutations causing mild hemophilia, was analyzed for patient WAR. PCR amplification was performed in a DNA cycler in a 50-µL reaction volume containing 100 ng of genomic DNA, a 0.2-mmol/L concentration of deoxynucleoside triphosphates, a 2.5-mmol/L concentration of magnesium chloride, 100 ng of each primer, and 0.5 µl of Taq DNA polymerase (Applied Biosystems), in a buffer containing a 10-mmol/L concentration of tris(hydroxymethyl)aminomethane-hydrochloride, ph 8.3. Amplification was achieved with 35 cycles of 1 minute at 94 C, 1 minute at 59 C, and 1 minute at 68 C. Primer sequences were as follows: forward primer, TTGGAATGGGCACCACTCCT; reverse primer, GGTGGGAAGAGATATGACAA. PCR products were analyzed by direct sequencing as described in the preceding text. Results A total of 8 patients were shown to have a very low factor VIII binding ratio (<30%). Results for these patients are shown in Table 1. In all patients, factor VIII by 2-stage assay was less than the reference range and less than 35% of vwf:ag. With the 1-stage assay, factor VIII was less than the reference range in 6 of 8 patients and substantially less than vwf:ag in only 3 patients. The very low factor VIII binding in these patients provides the laboratory diagnosis of type 2N vwd and explains the consequent low factor VIII detected (particularly using the 2-stage assay). Sequencing of exons 18 to 24 of the vwf gene showed that 4 of the 8 patients were homozygous for a G to A transition at nucleotide 2811 in exon 20, resulting in an arginine to glutamine substitution at amino acid 854 (R854Q). Another 3 patients (RKE, DIA, CSH) were heterozygous for the same mutation, and no other mutations were found in the known factor VIII binding region of the vwf gene. In 1 patient with very reduced factor VIII binding (PTA), no mutations were identified in exons 18 to 24 of the vwf gene. Another 3 subjects with intermediate factor VIII binding were found to be heterozygous for the mutation R854Q. These included the mother (RPA) of one of the patients with type 2N vwd (RKE) and 2 subjects from another family (NEL and her father, WAR). The results for these carriers also are shown in Table 1. Again, factor VIII determined by the 2-stage assay was markedly less than that determined by the 1-stage assay in all 3 of these subjects. The heterozygous mutation in these 3 subjects is consistent with their intermediate ratio for factor VIII binding ability of vwf, although in only one of them was it below the normal range (NEL). This was the only subject who would have been identified as a 272 Am J Clin Pathol 2002;118:269-276 American Society for Clinical Pathology

Coagulation and Transfusion Medicine / ORIGINAL ARTICLE carrier on the basis of the factor VIII binding results alone, without family studies. The disproportionately low factor VIII levels for 2 of these patients (NEL and, especially, WAR) were not explained by the factor VIII binding results. It therefore was thought likely that WAR also has mild hemophilia A and his daughter (NEL) is a carrier. This was confirmed by the identification of a mutation in exon 7 of the factor VIII gene (Ala 284 Pro), previously reported to be the causative mutation in mild hemophilia A, with factor VIII determined by 2-stage assay less than that determined by 1- stage assay. 15 In another 3 subjects from these families (Table 1), factor VIII binding ratios were normal, and no mutations were found. For another 7 patients tested, in whom the ratio for factor VIII binding ability of vwf was normal (data not shown), no mutations were found in the factor VIII binding region of the vwf gene. Comparison of results obtained from the 2 specialist laboratories are shown in Figure 2 for vwf:ag, factor VIII binding ability of vwf, and the factor VIII binding ratio. There was good agreement for vwf:ag (Figure 2A; r 2 = 0.89, slope = 1.046) and factor VIII binding ability of vwf (Figure 2B, r 2 = 0.89, slope = 1.276). There also was good concordance between the 2 laboratories for the factor VIII binding ratio (factor VIII binding ability of vwf divided by vwf:ag, expressed as a percentage, Figure 2C). Thus, all patients with very low factor VIII binding were clearly identified in both laboratories, and all control subjects with hemophilia were clearly identified as having normal factor VIII binding by both laboratories. Discussion Laboratory results obtained for patients with type 2N vwd can mimic those of mild hemophilia A, hemophilia A carriers, and vwd type 1. 1 The differential diagnosis is necessary for genetic counseling and to determine the most appropriate treatment. For treatment of type 2N vwd, a factor VIII product containing adequate amounts of normal vwf must be used so that factor VIII is bound to normal vwf:ag and is not quickly degraded in the circulation. In the present study, the first such study in the southern hemisphere, 8 patients with type 2N vwd were identified who could be distinguished from patients with other abnormalities of vwf and factor VIII. An estimate of the prevalence in Australia can be made from the patients drawn from South Australia, where most individuals with type 2N vwd are likely to have been identified. Five patients with type 2 vwd (from 4 families) were identified in this state with a population of 1.5 million. This prevalence is slightly higher than that reported in France, with 31 families identified 16 in a population of 58 million. Table 1 Results for Patients With Type 2N vwd, Carriers of Type 2N vwd, and Other Family Members * Factor VIII Assay Factor VIII Binding One-Stage Two-Stage Activity of vwf vwf:ag vwf Function Patient ID/ IU/dL Ratio IU/dL Ratio IU/dL IU/dL U/dL Ratio Genotype Family No. (45-160) (57-175) (45-150) (57-175) (55-240) (50-250) (60-245) (60-220) R854Q (R/R) Type 2N WDA/1 23 37 9 15 62 68 10 16 Q/Q RJE/1 (sister) 28 31 12 13 89 200 13 15 Q/Q CDA/1 (sister) 34 71 5 10 48 56 10 21 Q/Q BJO/2 63 43 40 27 148 91 20 14 Q/Q RKE/3 20 69 4 14 29 40 3 10 R/Q DIA/4 (brother) 46 90 12 24 51 61 10 20 R/Q CSH/4 26 130 6 30 20 35 5 25 R/Q PTA/5 32 73 15 34 44 39 7 16 R/R 2N carriers RPA/3 (mother) 60 128 33 70 47 52 34 72 R/Q NEL/6 (daughter) 36 50 14 19 72 41 35 49 R/Q WAR/6 24 19 3 2 125 144 96 77 R/Q Other family members RTR/3 (father) 83 213 56 144 39 42 41 105 R/R RPR/3 (sister) 53 126 44 105 42 39 43 102 R/R CNI/4 (daughter) 110 407 49 181 27 40 36 133 R/R Ag, antigen; Q, mutant sequence, glutamine; R, normal sequence, arginine; vwd, von Willebrand disease; vwf, von Willebrand factor. * Results from laboratory A are given. vwf function is ristocetin cofactor activity or collagen binding activity. Results for factor VIII and factor VIII binding activity of vwf are shown as absolute values (IU/dL or U/dL) and as a ratio (%) of VWF:Ag. Reference ranges are given in parentheses. Mutation analysis for the R854Q mutation at amino acid position 854 of vwf. Collagen binding result (reference range, 50-250 IU/dL). Ristocetin cofactor result (reference range, 55-210 IU/dL). Assayed in referring laboratory. American Society for Clinical Pathology Am J Clin Pathol 2002;118:269-276 273

Rodgers et al / CROSS-LABORATORY INVESTIGATION OF VON WILLEBRAND DISEASE TYPE 2N A Laboratory A (U/dL) B Laboratory A (U/dL) C Laboratory A (%) 500 400 300 200 100 0 0 100 200 300 400 500 Laboratory B (% PNP) 300 250 200 150 100 50 200 150 100 0 0 50 100 150 200 250 300 Laboratory B (% PNP) 50 0 0 50 100 150 200 Laboratory B (%) Figure 2 Comparison of results for 31 patients cotested for type 2N von Willebrand disease. A, von Willebrand factor (vwf) antigen. B, Factor VIII binding ability of vwf. C, Factor VIII binding ratio percentage, factor VIII binding activity divided by vwf antigen. PNP, pooled normal plasma. Data from the present study also confirm that the 1- stage factor VIII assay is not sensitive to the defect in type 2N vwd, having failed to detect a reduced factor VIII/vWF:Ag ratio in 5 of 8 identified cases of type 2N vwd (Table 1). Conversely, the 2-stage assay was found to be a sensitive screening test for detection of type 2N vwd, with the factor VIII/vWF:Ag ratio substantially lower than the normal range in all 8 patients (Table 1). Furthermore, the factor VIII level measured by the 2-stage assay was less than half the 1-stage result in 7 of 8 patients (Table 1). In patients with a reduced factor VIII shown by 2-stage assay, a discrepancy between the 1-stage and 2-stage results could be used as a further indication of possible type 2N vwd. In 3 subjects without a detectable factor VIII binding defect (Table 1), factor VIII measured by 2-stage assay also was less than factor VIII measured by 1-stage assay. However, in all 3, the result for the 2-stage assay was normal, and the ratio for factor VIII by the 2-stage assay/vwf:ag was well within the reference range. The lower result on the 2-stage assay in these patients (compared with the 1-stage assay) may be related to the high ratio of factor VIII by the 1- stage assay/vwf:ag in these patients, indicating that the result from the 1-stage assay may be artificially high. Alternatively, the lower result from the 2-stage assay may be due to unknown factors causing variation between results from these 2 assays. In 103 healthy donors previously tested by both assays, 8 there was general agreement between the 1- and 2-stage assay results, although there was some variation around the line of equivalence with some healthy subjects being up to 1.7-fold higher by the 1-stage assay and others up to 1.5-fold higher by the 2-stage assay. A previous study has shown a higher factor VIII result using the 1-stage assay compared with the 2-stage assay in 8 patients with variant vwd, which was thought to be due to adsorption of factor VIII by the aluminum hydroxide used for pretreatment of plasma samples in the 2-stage assay. 9 These patients later were identified as having type 2N vwd. 17,18 In another report of patients with type 2N vwd, factor VIII was higher by 1-stage assay than by chromogenic assay in which aluminum hydroxide adsorption was not used. 19 A higher result by the 1-stage assay compared with the 2-stage assay also has been reported in a proportion of patients with mild hemophilia A, but it is not due to adsorption by aluminum hydroxide in these patients. 8 The reason for the lower 2-stage assay results for factor VIII in the present report is the subject of further investigation in laboratory A. Several mutations have been reported to cause type 2N vwd. Four of the patients from the present study were shown to be homozygous for the most common type of 2N mutation, G2811A R854Q (previously identified as G2561A Arg91Gln, http://mmg2.im.med.umich.edu/vwf/nomenclature.htm). This mutation is the most common type 2N mutation 274 Am J Clin Pathol 2002;118:269-276 American Society for Clinical Pathology

Coagulation and Transfusion Medicine / ORIGINAL ARTICLE reported 20 and has been shown to be the cause of the very low factor VIII binding ability of vwf. Another 3 patients with very low factor VIII binding were heterozygous for the same mutation, with no other mutations being found. In these patients, the sequencing results are not in agreement with the factor VIII binding ratio. These patients (RKE, DIA, CSH) also are presumed to be heterozygous for a null allele (which results in no production of vwf from the vwf gene, which has no mutations in the factor VIII binding domain), similar to cases previously reported by others. 5,20 This null allele probably contains a mutation that would cause type 3 vwd if present in the homozygous form and is probably present in other family members who have low vwf:ag (RTR, RPR, CNI). One other patient (PTA) had no mutations identified in exons 18 to 24 of the vwf gene. Although these exons are the only ones to have been identified as being responsible for the factor VIII binding ability of vwf, there have been suggestions that another region also may be involved, 3 with one such example in exon 27 recently reported. 21 The factor VIII binding ratio provided a much clearer separation of the patients with type 2N vwd from other patients (Figure 2C). It also permitted a more meaningful assessment of the factor VIII binding ability of vwf than the absolute factor VIII binding activity, because it was not influenced by low levels of vwf:ag. This was demonstrated particularly by the results for the patients with low vwf:ag and low factor VIII binding ability of vwf (RTR, RPR, CNI), but a normal factor VIII binding ratio (Table 1). Similarly, in 2 patients with type 2N vwd in whom vwf:ag was much higher than usual owing to desmopressin therapy or stress and the factor VIII binding ability of vwf also was higher than expected, the factor VIII binding ratio clearly identified them both as having type 2N vwd (data not shown). Generally, results showed good agreement between the 2 specialist laboratories. Furthermore, all patients with type 2N vwd who were cotested by each laboratory were independently and appropriately identified, even though different methods were used in each laboratory. These results show better agreement than was found in recent multilaboratory surveys of other tests used for diagnosis of vwd. 6,7 This may be because the factor VIII binding results are less variable between methods than other vwf assays or because the 2 laboratories in this comparison are both reference centers for hemophilia and vwd testing and generally are more experienced with such techniques. Such concordance and effectiveness in type 2N vwd identification using diagnostic test systems also is reassuring given that use of factor VIII (measured by 1-stage assay) and vwf:ag as screening assays for type 2N vwd may not be adequate to detect the disorder. 7 We describe 8 patients with type 2N vwd identified in a cross-laboratory study using differing methods for the detection of factor VIII bound to vwf in a factor VIII binding assay. There was good agreement between the factor VIII binding ability of vwf assessed by chromogenic assay and ELISA. Seven of the patients were previously thought to have vwd, and 1 was previously thought to have hemophilia A. Patients previously classified as having vwd have been described by other authors and probably are more common than patients previously diagnosed with isolated hemophilia A. 20 This report demonstrates that, except in families in which hemophilia A has been clearly demonstrated, factor VIII binding assays should be performed on all patients in whom factor VIII coagulant activity is lower than the vwf:ag. Finally, in our study, the 1-stage factor VIII assay did not seem useful as a screening test for type 2N vwd, whereas the 2-stage factor VIII assay was effective in identifying patients in whom further study was warranted. From the 1 Haemostasis Laboratory, Haematology Division, and the 4 Molecular Pathology Division, Institute of Medical and Veterinary Science; the 2 School of Pharmaceutical, Molecular and Biomedical Sciences, University of South Australia, Adelaide; and the 3 Department of Haematology, Institute of Clinical Pathology and Medical Research, Westmead Hospital, Westmead, New South Wales, Australia. Supported by the University of South Australia and the Tied Research Fund of the Royal Adelaide Hospital. Address reprint requests to Dr Lloyd: Haematology Division, IMVS, PO Box 14, Rundle Mall, Adelaide, SA 5000, Australia. Acknowledgments: We are grateful to John Rowell, MD, of the Royal Brisbane Hospital, Queensland Health Pathology Service, and Ross Baker, MD, from the Royal Perth Hospital, Western Australia, for providing samples for assay in this project; the Oxford Haemophilia Centre for providing details of their factor VIII binding assay; the Sheffield Haemophilia Centre for providing plasma and the DNA of 2 patients with type 2N vwd; C Mazurier, PhD, for details on the vwf:factor VIII binding assay; Mark Dean, MD, from the New South Wales Red Cross Blood Bank for the donation of recombinant factor VIII; and all other clinicians and patients involved in supplying samples for this study. References 1. Sadler JE. Biochemistry and genetics of von Willebrand factor. Annu Rev Biochem. 1998;67:395-424. 2. Nishino M, Girma J-P, Rothschild C, et al. New variant of von Willebrand disease with defective binding to factor VIII. Blood. 1989;74:1591-1599. 3. Nesbitt IM, Goodeve AC, Guilliatt AM, et al. Characterisation of type 2N von Willebrand disease using phenotypic and molecular techniques. Thromb Haemost. 1996;75:959-964. 4. Schneppenheim R, Budde U, Krey S, et al. Results of a screening for von Willebrand disease type 2N in patients with suspected haemophilia A or von Willebrand disease type 1. Thromb Haemost. 1996;76:598-602. American Society for Clinical Pathology Am J Clin Pathol 2002;118:269-276 275

Rodgers et al / CROSS-LABORATORY INVESTIGATION OF VON WILLEBRAND DISEASE TYPE 2N 5. Casonato A, Gaucher C, Pontara E, et al. Type 2N von Willebrand disease due to Arg91Gln substitution and a cytosine deletion in exon 18 of the von Willebrand factor gene. Br J Haematol. 1998;103:39-41. 6. Favaloro EJ, Smith J, Petinos P, et al, for the RCPA Quality Assurance Program (QAP) in Haematology Haemostasis Scientific Advisory Panel. Laboratory testing for von Willebrand s disease: an assessment of current diagnostic practice and efficacy by means of a multi-laboratory survey. Thromb Haemost. 1999;82:1276-1282. 7. Favaloro EJ, Thom J, Baker R, for the Australasian Society for Thrombosis and Haemostasis (ASTH) Emerging Technologies Group. Assessment of current diagnostic practice and efficacy in testing for von Willebrand s disorder: results from the second Australasian multi-laboratory survey. Blood Coagul Fibrinolysis. 2000;11:729-737. 8. Duncan EM, Duncan BM, Tunbridge LJ, et al. Familial discrepancy between the one-stage and two-stage factor VIII methods in a subgroup of patients with haemophilia A. Br J Haematol. 1994;87:846-848. 9. Montgomery RR, Hathaway WE, Johnson J, et al. A variant of von Willebrand s disease with abnormal expression of factor VIII procoagulant activity. Blood. 1982;60:201-207. 10. Denson KWE, Wilkins T. Semi-automation of the two-stage factor VIII assay. Clin Lab Haematol. 1980;2:311-316. 11. 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;25:152-158. 12. 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. 2000;83:127-135. 13. Casanato A, Pontara E, Zerbinati P, et al. The evaluation of factor VIII binding activity of von Willebrand factor by means of an ELISA method: significance and practical implications. Am J Clin Pathol. 1998;109:347-352. 14. Mancuso DJ, Tuley EA, Westfield LA, et al. Structure of the gene for human von Willebrand factor. J Biol Chem. 1989;264:19514-19527. 15. Pipe SW, Saenko EL, Eickhorst AN, et al. Hemophilia A mutations associated with 1-stage/2-stage activity discrepancy disrupt protein-protein interactions within the triplicated A domains of thrombin-activated factor VIIIa. Blood. 2001;97:685-691. 16. Mazurier C, Meyer D. Factor VIII binding assay of von Willebrand factor and the diagnosis of type 2N von Willebrand disease: results of an international survey. Thromb Haemost. 1996;76:270-274. 17. Kroner PA, Friedman KD, Fahs SA, et al. Abnormal binding of factor VIII is linked with the substitution of glutamine for arginine 91 in von Willebrand factor in a variant form of von Willebrand disease. J Biol Chem. 1991;266:19146-19149. 18. Kroner PA, Foster PA, Fahs SA, et al. The defective interaction between von Willebrand factor and factor VIII in a patient with type 1 von Willebrand disease is caused by substitution of Arg19 and His54 in mature von Willebrand factor. Blood. 1996;87:1013-1021. 19. Mazurier C, Gaucher C, Jorieux S, et al. Biological effect of desmopressin in eight patients with type 2N ( Normandy ) von Willebrand disease. Br J Haematol. 1994;88:849-854. 20. Nesbitt IM, Hampton KK, Preston FE, et al. A common splice site mutation is shared by two families with different type 2N von Willebrand disease mutations. Thromb Haemost. 1999;82:1061-1064. 21. Allen S, Abuzenadah AM, Blagg JL, et al. Two novel type 2N von Willebrand disease causing mutations that result in defective factor VIII binding, multimerization, and secretion of von Willebrand factor. Blood. 2000;95:2000-2007. 276 Am J Clin Pathol 2002;118:269-276 American Society for Clinical Pathology