SEARCH

SEARCH BY CITATION

Keywords:

  • bleeding score;
  • blood group;
  • co-segregation;
  • linkage analysis;
  • type 1;
  • von Willebrand disease

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. References

Summary. Background: von Willebrand disease (VWD) type 1 is a congenital bleeding disorder caused by genetic defects in the von Willebrand factor (VWF) gene and characterized by a reduction of structurally normal VWF. The diagnosis of type 1 VWD is difficult because of clinical and laboratory variability. Furthermore, inconsistency of linkage between type 1 VWD and the VWF locus has been reported. Objectives: To estimate the proportion of type 1 VWD that is linked to the VWF gene. Patients and methods: Type 1 VWD families and healthy control individuals were recruited. An extensive questionnaire on bleeding symptoms was completed and phenotypic tests were performed. Linkage between VWF gene haplotypes and the diagnosis of type 1 VWD, the plasma levels of VWF and the severity of bleeding symptoms was analyzed. Results: Segregation analysis in 143 families diagnosed with type 1 VWD fitted a model of autosomal dominant inheritance. Linkage analysis under heterogeneity resulted in a summed lod score of 23.2 with an estimated proportion of linkage of 0.70. After exclusion of families with abnormal multimer patterns the linkage proportion was 0.46. LOD scores and linkage proportions were higher in families with more severe phenotypes and with phenotypes suggestive of qualitative VWF defects. About 40% of the total variation of VWF antigen could be attributed to the VWF gene. Conclusions: We conclude that the diagnosis of type 1 VWD is linked to the VWF gene in about 70% of families, however after exclusion of qualitative defects this is about 50%.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. References

von Willebrand factor (VWF) is a multimeric glycoprotein synthesized by endothelial cells and megakaryocytes. VWF mediates the adhesion of platelets to sites of vascular damage by binding to platelet membrane glycoproteins and to exposed vascular subendothelium [1]. von Willebrand disease (VWD) is defined as a congenital bleeding disorder caused by genetic defects in the VWF gene [1]. Those defects result in quantitatively reduced VWF levels (types 1 and 3 VWD) or qualitatively abnormal variants of VWF (type 2 VWD) [2]. Type 1 is characterized by mild reduction of structurally normal VWF, whereas type 3 VWD displays a nearly undetectable level of plasma VWF. The qualitative variant type 2 VWD may be subdivided depending on the specific functional or structural defect of VWF [2]. Provisional criteria for the diagnosis of type 1 VWD have been published only recently on behalf of the ISTH-SSC on VWF [3].

The prevalence of VWD has been estimated to be present in up to 1% by population screening [4,5]. However, the prevalence of patients with clinically significant bleeding seems much lower (about 1 per 20 000) [1,6]. The diagnosis of type 1 VWD, which comprises about 50–75% of all VWD cases, is often difficult because of incomplete penetrance, considerable variability of VWF levels, low specificity of bleeding symptoms, and low sensitivity of diagnostic tests. The issue is complicated even more by the dependence of VWF levels on ABO blood group, exercise, age, sex, ethnicity, menstrual cycle and use of oral contraceptives [7–9]. It may therefore be difficult to distinguish patients with type 1 VWD from healthy individuals with VWF levels at the lower end of the normal distribution curve. Sadler [6] recently calculated that VWD may even be diagnosed falsely positive in 0.4% of the population. Two papers showed in a limited number of families that a personal and family bleeding history and persistently low VWF levels, the usual criteria of type 1 VWD, may not co-segregate with genetic markers in the VWF gene [10,11]. So, it could very well be that among the patients there are in fact individuals who have low VWF levels not related to defects in the VWF gene.

The aim of the study reported here was to reliably estimate in a large cohort the proportion of families diagnosed with type 1 VWD that is linked to the VWF gene locus. Linkage analysis between VWF gene haplotypes and the phenotypic diagnosis of type 1 VWD, the plasma levels of VWF and the severity of bleeding symptoms was performed.

Patients and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. References

Study design

The ‘Molecular and Clinical Markers for the Diagnosis and Management of Type 1 von Willebrand disease Study (MCMDM-1VWD)’ is an EU funded multicenter survey on type 1 VWD. See Ref. no. [12] for further details on study design and recruitment (see also http://www.shef.ac.uk/euvwd; accessed 19 December 2005). The aims of the MCMDM-1VWD study were to determine the value of clinical, phenotypic and molecular markers for the diagnosis of type 1 VWD. Twelve centers from nine European countries participated. All centers planned to recruit 10–15 families with type 1 VWD. The study aimed at recruiting the full spectrum of type 1 VWD families with the diagnosis based on routine laboratory testing, including also milder cases of type 1 VWD to mimic the clinical situation as much as possible and not to have bias toward the more severe, highly penetrant type 1 VWD. Patients who had been diagnosed historically by each of the centers as having type 1 VWD were eligible for inclusion [12]. A family met the inclusion criteria when it consisted of at least two affected individuals. We aimed to include as many affected and unaffected family members as possible and to include families that consisted of at least two generations. All family members were interviewed, an extensive questionnaire on bleeding symptoms was completed, a family tree was obtained, and venous blood samples were taken at the time of enrolment into the study.

Each center recruited about 100 healthy control individuals defined as subjects that never sought medical attention for bleeding symptoms. In two of the centers the recruited controls also completed the questionnaire on bleeding symptoms. Venous blood samples were obtained from each enrolled control individual.

The local ethical review committees of all centers approved the study protocol. Informed consent was obtained from all individuals upon recruitment by signing a consent form in the individual's own language.

Bleeding score

The bleeding symptoms were recorded by a personal interview with a clinician using a standardized questionnaire adapted from a previous investigation [13]. Twelve bleeding symptoms were investigated. Each symptom was given a grade of hemorrhagic severity of −1 (did not bleed when challenged by two surgeries, tooth extractions or deliveries) to +4 (required transfusion or replacement therapy). The summation of the score of all bleeding symptoms resulted in a quantitative measure of bleeding severity. Total possible scores were −3 to +45. Details of the bleeding score are reported by Tosetto et al. [12]. The questionnaire and scoring system are available at http://www.shef.ac.uk/euvwd/bleed_score.htm; (accessed 19 December 2005). Based on the bleeding scores of 195 healthy control individuals a score higher than three was considered abnormal.

Coagulation studies

On all family members a bleeding time was performed using the center's usual method, either a sterile microlancette (Ivy) or a commercial Simplate II device (Simplate II; General Diagnostics, Morris Plains, NJ, USA). Platelet counts were performed on EDTA-anticoagulated blood. Bleeding time and platelet counts were not performed in healthy controls. Venous blood was anticoagulated with 3.8% sodium citrate (1:10, v:v) and centrifuged at 2500 × g for 15 min. Plasma was aliquoted and stored at −80°C and cells were kept for extraction of DNA.

All plasma samples were analyzed for VWF antigen (VWF:Ag) by ELISA, VWF ristocetin cofactor activity (VWF:RCo) by aggregometry using formalin-fixed platelets and factor VIII activity (FVIII:C). These assays were performed locally in the recruiting centers, but VWF:Ag and VWF:RCo were also confirmed by one central laboratory and are reported here (Vicenza, Italy) [4]. Results are expressed as international units (IU) with reference to a normal plasma pool calibrated against the fourth WHO international standard FVIII/VWF plasma (97/586) (National Institute for Biological Standards and Control, Potters Bar, UK). Multimer profiles were performed in all family members by one central laboratory (Hamburg, Germany) [14].

STR genotyping

Genomic DNA was extracted from peripheral blood leukocytes according to standard procedures in each of the centers and stored at −20°C. For linkage analysis performed in Leiden, three highly polymorphic short tandem repeat (STR) polymorphisms were selected: one (GT)n repeat within the promoter region of the VWF gene (VWP) [15] and two (TCTA)n repeats within intron 40 of the VWF gene (VNTR2 [16] and VNTR3 [17]). DNA fragments containing these STRs were amplified by PCR using fluorescent oligonucleotide primers (Invitrogen Corporation, Carlsbad, CA, USA) (Table 1). Of each primer set one was modified by 5-labeling with a specific phosphoramidite dye to discriminate between different PCR products. The other primer of each primer set was extended at the 5-end with the sequence GTGTCTT to promote PCR product adenylation and thereby ensuring accurate genotyping [18,19]. One hundred nanograms of genomic DNA was used as a template. Each PCR mixture of 25 μL consisted of PCR Buffer II (Perkin-Elmer, Boston, MA, USA) and contained in addition 25 pmol of each oligonucleotide primer, 0.2 mm of each deoxynucleotide (dNTP) (Larova, Teltow, Germany), 1.5 mm MgCl2 (VWP) or 2 mm MgCl2 (VNTR2 and VNTR3), 0.75 units of AmpliTaq® Gold DNA polymerase (Perkin-Elmer). The thermal cycling parameters were: denaturing for 10 min at 95 °C; followed by 33 cycles of 1 min denaturing at 94°C, 1 min annealing at 59°C, and 1 min extension at 72 °C. The PCR reaction was finalized with 45 min extension at 60 °C.

Table 1.  Oligonucleotide primer sequences
PrimerSequence (5′[RIGHTWARDS ARROW] 3′)5′ modificationLocation*
  1. f, forward primer; r, reverse primer.

  2. *Numbering according to Mancuso et al. [32].

  3. Phosphoramidite dye (Invitrogen Corporation).

VWPfAAT GGG AGG CTG GGA AGA AGA AD2 dye1/1490
VWPrGTA AAA TGG TGA AGG TGG GGA GTGTGTCTT1/1665
VNTR2fTGT ACC TAG TTA TCT ATC CTGD4 dye31/2215
VNTR2rGTG ATG ATG ATG GAG ACA GAGGTGTCTT31/2380
VNTR3fCCC TAG TGG ATG ATA AGA ATA ATCGTGTCTT31/1640
VNTR3rGGA CAG ATG ATA AAT ACA TAG GAT GGA TGGD3 dye31/1793

The PCR fragment length analysis was performed on a Beckman Coulter CEQTM 2000 DNA Analysis System (Beckman Coulter, Fullerton, CA, USA) by means of capillary gel electrophoresis. For VWP, 11 alleles were detected ranging from 181 to 203 bp, for VNTR2 eight alleles ranging from 166 to 194 bp, and for VNTR3 10 alleles ranging from 139 to 175 bp. The frequency distribution of the alleles was similar to previously reported frequencies [15–17].

The ABO blood group genotypes were determined in Sheffield by analysis of three single nucleotide polymorphisms (SNPs) in the ABO glycosyltransferase gene using the 5′nuclease TaqMan allelic discrimination assay (ABI TaqMan 7900; Applied Biosystems) [20,21]. The SNP 261 delG/G discriminates between O/non-O, the SNP 930 A/G between B/non-B and SNP 467 T/C between A2/non-A2. By this combination of SNPs the four alleles A1, A2, B and O were assigned [22].

Linkage and statistical analyses

For segregation and linkage analysis a binary disease trait, affected or unaffected, was used. Different phenotypic definitions were used for separate analyses: (i) affected status as indicated by the recruiting center (the composite of personal and family bleeding symptoms and current and historic VWF parameters), further referred to as ‘clinical practice diagnosis’; (ii) affected status defined by VWF:Ag or VWF:RCo below 2.5 percentile of the ABO blood group specific normal distribution in combination with a bleeding score higher than three (measurements at time of recruitment), further referred to as ‘stringent diagnosis’; (iii) affected status defined by a bleeding score higher than three, further referred to as ‘bleeding diathesis’. For subgroup analyses these diagnosis groups were further divided: families with all family members having normal multimer pattern versus families with individuals with abnormal multimer patterns; families with an index case with VWF:RCo/VWF:Ag ratio <0.70 vs. ≥0.70; families with an index case with VWF:Ag ≤30 IU dL−1 vs. >30 IU dL−1.

Segregation analysis was performed using the module SEGREG of the program S.A.G.E. version 5.0 [23]. The likelihood was corrected for ascertainment through the index case. The parameters obtained from the segregation analysis were used as input in the parametric linkage analysis. As it was expected that only a fraction of the sibships is linked to the VWF gene locus, linkage analysis was performed according to Faraway's admixture test for the detection of linkage under heterogeneity [24]. Linkage analysis was performed using LODLINK of S.A.G.E. version 5.0 [23]. The haplotypes formed by the three STRs, which are located at the same locus, were used as single alleles.

The heritability of the phenotypic variance of the VWF levels and the quantitative bleeding score that could be attributed to the VWF gene locus was analyzed by the pedigree-wide inverse regression analysis method [25] as implemented in Merlin–Regress (Merlin version 0.10.2; Center for Statistical Genetics, University of Michigan, Ann Arbor, MI, USA). The estimated heritability of VWF (the proportion of variance attributed to genetic factors) reported in the literature ranges from 0.30 to 0.75 [26–29]. We used a heritability of 0.50 in the model. The analysis was performed after adjustment of VWF levels for the covariates ABO blood group, age and sex. Standardized residuals were calculated using the regression coefficients obtained by multiple linear regression of the covariates in the healthy control population. For the analysis of heritability the three STRs were analyzed separately since there were too many different haplotypes to be handled by the computer program. SPSS for Windows version 10.0.7 (SPSS Inc., Chicago, IL, USA) was used for additional statistical analyses.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. References

Patients, families and controls

A total number of 744 individuals from 154 families with type 1 VWD were recruited, including 154 index cases, 278 affected family members and 312 unaffected family members. Eleven families were excluded from further analysis in the present study: seven families because of deviation from the enrolment criteria (only one affected family member); three families because reduced VWF levels could not be demonstrated in any family member; and one family because of suspicion of a sampling error (no maternal allele was present in the offspring). The remaining 143 families included for further analysis comprised a total number of 709 individuals, including 143 index cases, 272 affected and 294 unaffected family members. The average family size was 5 (range two to 12). A total of 1166 normal controls were recruited. The average number of controls recruited per center was 97 (range 73–105). Basic characteristics are summarized in Table 2. A representative pedigree is shown in Fig. 1.

Table 2.  Characteristics of included families and normal controls
 Index case, n = 143Affected family member, n = 272Unaffected family member, n = 294Normal control, n = 1166
  1. *Determined in a subset of 195 normal controls.

Male (%)37.144.5 51.0 40.5
Median age, years (range)40 (1–80)32 (2–91) 41 (3–90) 40 (8–100)
Median FVIII:C, IU dL−1 (25–75 percentile)56 (28–77)63 (34–88)100 (78–133)108 (88–133)
Median VWF:Ag, IU dL−1 (25–75 percentile)33 (19–49)36 (21–54.5) 92 (69.5–118) 97 (75.3–121)
Median VWF:RCo, IU dL−1 (25–75 percentile)34 (10–49)35 (12–54) 86.5 (65–112) 94 (75–117)
VWF:RCo/VWF:Ag ratio (25–75 percentile) 0.92 (0.63–1.10) 0.92 (0.61–1.13)  0.94 (0.81–1.16)  1.0 (0.85–1.17)
VWF:RCo/VWF:Ag ratio ≥0.70 (%)72.770.5 90.9 92.8
Multimers normal (%)60.661.1 99.6
Bleeding score (25–75 percentile) 9 (5–13) 4 (2–8)  0 (−1–1)−1.0 (−1–0)*
Bleeding score >3 (%)82.558.8  9.2  0.5*
Blood group O (%)67.959.3 49.1 37.9
image

Figure 1. Representative pedigree of a type 1 VWD family showing complete co-segregation of the disease phenotype (filled symbols) with a single VWF gene haplotype (191, 159, 174, allele sizes in bp of VWP, VNTR3 and VNTR2, respectively). Males are indicated by squares, females by circles. Affected individuals are represented by filled symbols and an arrow indicates the index case. The bars illustrate identical haplotypes.

Download figure to PowerPoint

Segregation and linkage analysis

Segregation analysis (Table 3) was performed using the three different phenotypic definitions: the clinical practice diagnosis, the stringent diagnosis, and the bleeding diathesis. For the stringent diagnosis the blood group specific distribution of VWF:Ag and VWF:RCo was determined in the normal controls. For blood group O the 2.5 percentile cut off levels were VWF:Ag below 44.4 IU dL−1 or VWF:RCo below 43.0 IU dL−1 and for blood group non-O the cut off levels were VWF:Ag below 54.1 IU dL−1 or VWF:RCo below 54.0 IU dL−1.

Table 3.  Segregation and linkage analysis
Phenotypic definition subgroupsnSegregation analysis*Linkage analysis
ModelPenetranceAllele frequencyLOD scoreRecombination fractionLinkage proportion (95% CI)
  1. AD, autosomal dominant; CI, confidence interval.

  2. *Inheritance model, penetrance and allele frequency are derived by segregation analysis from the data of this study and these parameters were used as input in the parametric linkage analysis.

  3. Number of families and in brackets number of affected/unaffected individuals per analysis.

Clinical practice diagnosis143 (415/294)AD93%0.000523.20.01820.70 (0.39–1.0)
 normal multimers84 (236/157)   5.690.00.46 (0.25–0.66)
 abnormal multimers59 (179/137)   20.40.03751.0 –
 VWF:RCo/VWF:Ag ≥0.70104 (301/205)   12.90.00.57 (0.40–0.74)
 VWF:RCo/VWF:Ag <0.7039 (114/89)   11.60.04621.0 –
 VWF:Ag ≤30 IU dL−163 (197/129)   17.40.04721.0 –
 VWF:Ag >30 IU dL−180 (218/165)   8.410.00.51 (0.32–0.70)
Stringent diagnosis143 (216/493)AD84%0.081.870.00.28 (0.06–0.49)
 normal multimers84 (88/305)   00.5
 abnormal multimers59 (128/188)   4.160.00.62 (0.31–0.94)
 VWF:RCo/VWF:Ag ≥0.70104 (134/372)   0.290.00.13 (0–0.38)
 VWF:RCo/VWF:Ag <0.7039 (82/121)   2.280.00.56 (0.17–0.96)
 VWF:Ag ≤30 IU dL−163 (129/197)   2.510.00.47 (0.15–0.79)
 VWF:Ag >30 IU dL−180 (87/296)   00.5
Bleeding diathesis143 (305/404)AD91%0.071.830.00.22 (0.05–0.40)
 normal multimers84 (164/229)   0.060.3150.32 (0–1)
 abnormal multimers59 (141/175)   2.340.00.36 (0.10–0.61)
 VWF:RCo/VWF:Ag ≥0.70104 (214/292)   00.5
 VWF:RCo/VWF:Ag <0.7039 (91/112)   2.350.11980.76 (0–1)
 VWF:Ag ≤30 IU dL−163 (140/186)   1.990.08020.46 (0–1)
 VWF:Ag >30 IU dL−180 (165/218)   0.210.00.12 (0–0.37)

Parametric linkage analysis for the clinical practice diagnosis resulted in a highly significant summed LOD score of 23.2 for the 143 families included. The estimated proportion of the families in which type 1 VWD is linked to the VWF gene is 0.70. When the stringent diagnosis was considered the summed LOD score dropped to 1.87 and when the bleeding diathesis was considered the summed LOD score was 1.83. The LOD scores for the different diagnosis groups are difficult to compare as the LOD scores are dependent on the number of informative families. As is shown in Table 3 the number of affected individuals in the stringent diagnosis group is nearly half of the clinical practice diagnosis. Subgroup analyses show that for all diagnosis groups the LOD scores and linkage proportions were higher in the families with the more severe phenotypes (VWF:Ag ≤30 IU dL−1) and the phenotypes suggestive of qualitative VWF defects (abnormal multimers or VWF:RCo/VWF:Ag ratio <0.70).

In the healthy control population, multiple linear regression analysis was performed to obtain the regression coefficients used to adjust VWF levels in the VWD families for the influence of ABO blood group, age and sex. About 16% of the total variation in VWF:Ag and 10% of the total variation in VWF:RCo is explained by ABO blood group (O vs. non-O), age and sex (R2 = 0.159 and R2 = 0.094, respectively). After adjustment of VWF levels for the influence of ABO blood group (O vs. non-O), age and sex, the heritability of the remaining phenotypic variance that could be attributed to the VWF locus was 48% for VWF:Ag and 53% for VWF:RCo (Table 4). Thus, 40% (0.48 × 84%) of the total variation in VWF:Ag levels and 48% (0.53 × 90%) of the total variation in VWF:RCo in these families with type 1 VWD can be explained by the VWF locus. Heritability of the variance of the bleeding score in the VWD families that could be attributed to the VWF gene was 2% (Table 4).

Table 4.  Heritability attributable to VWF gene
 STRHeritability (95% CI)LOD score
  1. CI, Confidence interval.

  2. *Adjusted for ABO blood group (O vs. non-O), age and sex.

VWF:Ag*VWP0.485 (0.395–0.575)24.0
VNTR30.477 (0.387–0.567)23.3
VNTR20.478 (0.388–0.568)23.4
VWF:RCo*VWP0.535 (0.429–0.641)21.5
VNTR30.527 (0.421–0.633)20.8
VNTR20.527 (0.421–0.633)20.8
Bleeding scoreVWP0.017 (0.009–0.025)3.8
VNTR30.015 (0.007–0.023)3.2
VNTR20.016 (0.008–0.024)3.3

Phenotypic differences between families with complete and incomplete co-segregation

Based on the co-segregation of VWF haplotypes and the clinical practice diagnosis the families were, under the assumption of autosomal dominant inheritance, categorized as ‘complete co-segregation’ (pedigrees with no phenocopies and fully penetrant; where a phenocopy is an individual diagnosed with VWD, but who does not share the same VWF allele as other diagnosed individuals), ‘incomplete co-segregation’ (pedigrees with either phenocopies or non-penetrances), or ‘non-informative’ (usually due to small pedigree size). In the 143 families co-segregation was complete in 63, incomplete in 43 and non-informative in 37 pedigrees. Thus of the informative pedigrees, 59% (63 of 106) show complete co-segregation and 41% (43 of 106) incomplete co-segregation.

Co-segregation is associated with the severity of the phenotype. When families are categorized by VWF:Ag level in the index case, the families of index cases with the mildest phenotype are four times less likely to show complete co-segregation compared to those with the most severe phenotype. This pattern was even more dramatic for VWF:RCo (Table 5). When comparing the VWF:Ag levels (all blood groups) between families with complete co-segregation and families with incomplete co-segregation it was evident that the index cases and affected family members from families with complete co-segregation have lower mean VWF:Ag levels (30 and 31 IU dL−1, respectively) than index cases and affected family members from families with incomplete co-segregation (42 and 48 IU dL−1; P < 0.01 and P < 0.001, respectively). This indicates that in general the patients from families with complete co-segregation are more severely affected. The unaffected family members from families with complete co-segregation have mean VWF:Ag levels similar to normal controls (105 and 100 IU dL−1, respectively, P = 0.10), whereas unaffected family members from families with incomplete co-segregation have lower VWF:Ag levels than normal controls (88 and 100 IU dL−1, respectively, P < 0.001). This indicates that other factors than the VWF gene determine reduction of the VWF levels in families with incomplete co-segregation. When we performed this comparison per blood group (O vs. non-O), we found that this effect was restricted mainly to blood group O (Fig. 2). The individuals with blood group O and incomplete co-segregation have intermediate phenotypes (higher levels in affected and lower levels in unaffected individuals) compared with complete co-segregation. This indicates that the VWF:Ag levels and the diagnosis of type 1 VWD in individuals with incomplete co-segregation are determined to a large extent by blood group O. The results for VWF:RCo were similar to VWF:Ag.

Table 5.  Association between co-segregation of the ‘clinical practice diagnosis’ and categories of VWF in index cases
VWF level in index caseComplete co-segregationIncomplete co-segregationOR (95% CI)
  1. OR, odds ratio; CI, confidence interval; adjusted for blood group (O vs. non-O).

  2. *Reference category.

VWF:Ag (IU dL−1)
 0–151561*
 16–301890.73 (0.20–2.68)
 31–4518100.67 (0.19–2.32)
 >4512180.24 (0.07–0.82)
VWF:RCo (IU dL−1)
 0–152991*
 16–301050.61 (0.15–2.54)
 31–4517120.47 (0.16–1.38)
 >457170.13 (0.04–0.42)
image

Figure 2. Differences in distribution of VWF:Ag levels between individuals from families with complete co-segregation, families with incomplete co-segregation, and normal controls categorized by blood group. The boxplots indicate the median, interquartile range and extreme values. IC, index cases; AFM, affected family members; UFM, unaffected family members, and N, normal controls.

Download figure to PowerPoint

Blood group O is more prevalent in patients than in normal controls (Table 2) and more prevalent in individuals from families with incomplete co-segregation than in individuals from families with complete co-segregation (Fig. 3). Blood group O is a risk factor for being diagnosed with type 1 VWD (OR 3.5, 95% CI 2.4–5.0) and this is even more obvious in families with incomplete co-segregation (OR 5.2, 95% CI 2.6–10.8) (Table 6).

image

Figure 3. Prevalence of blood group O among individuals with complete co-segregation vs. incomplete co-segregation. The prevalence in normal controls was 38% (dashed line).

Download figure to PowerPoint

Table 6.  Blood group O as a risk factor for VWD
 Index caseNormal controlOR (95% CI)
  1. OR, odds ratio; CI, confidence interval.

  2. *ABO blood group missing in one index case of a non-informative family, one index case of a family with complete co-segregation, and one index case of a family with incomplete co-segregation. Data based on ‘clinical practice diagnosis’.

  3. Reference category.

All families*
 blood group non-O456861
 blood group O954193.5 (2.4–5.0)
Complete co-segregation
 blood group non-O276861
 blood group O354192.1 (1.3–3.6)
Incomplete co-segregation
 blood group non-O106861
 blood group O324195.2 (2.6–10.8)

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. References

The diagnosis of type 1 VWD is often difficult [7–9], the prevalence of 1% in the general population seems overestimated [4,5], and a large proportion of the diagnoses of type 1 VWD may be falsely positive [6]. When the diagnosis of type 1 VWD is falsely positive or when other genetic loci are involved, no linkage to the VWF gene is to be expected. We have investigated the proportion of type 1 VWD that is linked to the VWF gene in a large cohort of families throughout Europe to reliably estimate the proportion of linkage.

When the type 1 VWD diagnosis made by the clinicians of the recruiting centers was considered, parametric linkage analysis showed a lod score of 23.2 with an estimated linkage proportion of 0.70. Detailed phenotypic analyses revealed that a number of subjects had abnormal multimer patterns. When families with and without abnormal multimers were analyzed separately, it was clear that the subgroup with abnormal multimers showed the highest LOD score and a linkage proportion of 1.0, whereas in the group with all normal multimers the linkage proportion was 0.46. Although the normal multimer group probably best reflects proper type 1 cases, the multimer abnormalities that we observed were often minimal. We expect that in current clinical practice many of those minimally abnormal multimers go unnoticed and that those families are labeled type 1 VWD. Using the VWF:RCo/VWF:Ag ratio as indicator of qualitative abnormalities gave similar results (Table 3). Thus, the clinical practice diagnosis of type 1 VWD is not linked to the VWF gene in 30% of families and after exclusion of qualitative VWF defects this could be around 50%. For this study we selected on heredity by only including families with at least two affected individuals per family. However, in clinical practice one often encounters sporadic cases of type 1 VWD and the presumption is justified that among those patients the proportion of linkage to the VWF gene may even be lower. This non-linkage may be due to misdiagnosis, the coincidence of bleeding and low VWF by chance, locus heterogeneity, environmental factors, a de novo mutation, or recessive inheritance in some families.

When the stringent diagnosis was considered the LOD score was much lower (Table 3), which is explained by a loss of the number of affected individuals using these stringent criteria, leading to many non-informative families. A limitation of the stringent diagnosis is the fact that it is based on a single measurement, whereas the clinical practice diagnosis was based on historic and family data as well. Nevertheless, subgroup analyses confirmed higher LOD score and linkage proportion for qualitative VWF defects and more severe phenotypes just as for the clinical practice diagnosis.

The proportion of linkage we found in this study is comparable with the paper by Casana et al. [11], but higher than observed in the study by Castaman et al. [10]. However, in the latter report the patients were ascertained through a population-based study and not through referral, which may explain a lower proportion of linkage in the study by Castaman et al. [10].

The use of cut off levels and a binary disease trait (affected vs. unaffected) has lower statistical power than quantitative traits (liability) and therefore we have analyzed the two major diagnostic criteria (VWF levels and bleeding score) as quantitative traits. We found that the VWF locus contributes considerably to the variability in VWF levels: 40% of the total variance of VWF:Ag and 48% of VWF:RCo is explained by the VWF locus. In previous studies it has been shown that the VWF gene has limited or no influence on the normal variation of VWF levels [30,31]. The major contribution of the VWF locus to the plasma levels of VWF in this cohort of type 1 VWD families is therefore most likely due to VWF gene mutations and not to polymorphic variability. The VWF locus has only a limited contribution to the severity of bleeding symptoms (Table 4), which may reflect the low specificity and high prevalence of bleeding symptoms in the general population [6].

We confirmed that the diagnosis of type 1 VWD is influenced to a large extent by the presence of blood group O (Figs 2, 3 and Table 6). Carrying blood group O increases the risk for being diagnosed with type 1 VWD by 3.5. Even in families with complete co-segregation, where the disease is primarily determined by the VWF gene, the prevalence of blood group O is higher than in the general population and thus blood group O is an additional risk factor for type 1 VWD (OR 2.1). This indicates that the absolute level of VWF may be more important than a blood group specific cut off level. However, in the cases from the families with incomplete co-segregation blood group O has a major contribution to the phenotype (nearly 80% of those index cases carry blood group O). Finally, our data show that the more severe the phenotype, the more likely it is that the disease is linked to the VWF gene (Tables 3 and 5).

We conclude that among patients referred to specialized centers and diagnosed with type 1 VWD about 70% of the families show linkage between the disease phenotype and the VWF gene, however after exclusion of qualitative defects this is about 50%. In sporadic cases of VWD this may be lower. Even though it has been suggested that most diagnoses of type 1 VWD may be false positives [6], this study shows that, with the current diagnostic procedures, patients are identified of whom the majority are expected to have defects at the VWF locus.

Addendum

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. References

Contributions of authors

J. Eikenboom2−7, V. Van Marion2−7, H. Putter4−7, A. Goodeve1−3,6,7, F. Rodeghiero1−3,6,7, G. Castaman2,3,6,7, A.B. Federici2,3,6,7, J. Batlle2,7, D. Meyer2,7, C. Mazurier2,7, J. Goudemand2,7, R. Schneppenheim2,6,7, U. Budde2,7, J. Ingerslev2,7, Z. Vorlova2,7, D. Habart2,7, L. Holmberg2,7, S. Lethagen2,7, J. Pasi2,7, F. Hill2,7, I. Peake1−3,6,7

1Study initiation and coordination; 2Study design, data collection and performing laboratory analyses; 3Analysis and interpretation of results; 4Statistical analysis; 5Lead authors of initial manuscript; 6Revisions of draft manuscripts; 7Review and approval of final manuscript.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. References

The European Community under the Fifth Framework Programme (QLG1-CT-2000-00387) and the Van den Tol Foundation (to J. Eikenboom) financially supported this study.

Some of the results of this paper were obtained by using the program package S.A.G.E., which is supported by a U.S. Public Health Service Resource Grant (RR03655) from the National Center for Research Resources.

We would like to thank E. Jennings, M. Makris, H. Powell, M. Walker, L. Marsden, M. Hashemi, A. Al-Buhairan, S. Joyce, A. Bowyer, A. Tosetto, L. Baronciani, E. Fressinaud, A.S. Ribba, A. Stephanian, L. Hilbert, C. Caron, E. Gomez, J. Lambert, F. Oyen, T. Obser, K. Will, E. Drewke, J. Suttnar, J. Dudlova, C. Hallden, C. Watson, J. Warren, S. Mughal, W. Lester, A. Guilliatt, S. Enayat, G. Surdhar, and P. Short for their contributions to this study.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. References
  • 1
    Sadler JE. Biochemistry and genetics of von Willebrand factor. Annu Rev Biochem 1998; 67: 395424.
  • 2
    Sadler JE. A revised classification of von Willebrand disease. Thromb Haemost 1994; 71: 5205.
  • 3
    Sadler JE, Rodeghiero F. Provisional criteria for the diagnosis of VWD type 1. J Thromb Haemost 2005; 3: 7757.
  • 4
    Rodeghiero F, Castaman G, Dini E. Epidemiological investigation of the prevalence of von Willebrand's disease. Blood 1987; 69: 4549.
  • 5
    Werner EJ, Broxson EH, Tucker EL, Giroux DS, Shults J, Abshire TC. Prevalence of von Willebrand disease in children: a multiethnic study. J Pediatr 1993; 123: 8938.
  • 6
    Sadler JE. Von Willebrand disease type 1: a diagnosis in search of a disease. Blood 2003; 101: 208993.
  • 7
    Gill JC, Endres-Brooks J, Bauer PJ, Marks WJ Jr, Montgomery R. The effect of ABO blood group on the diagnosis of von Willebrand disease. Blood 1987; 69: 16915.
  • 8
    Wahlberg TB, Blomback M, Magnusson D. Influence of sex, blood group, secretor character, smoking habits, acetylsalicylic acid, oral contraceptives, fasting and general health state on blood coagulation variables in randomly selected young adults. Haemostasis 1984; 14: 3129.
  • 9
    Kadir RA, Economides DL, Sabin CA, Owens D, Lee CA. Variations in coagulation factors in women: effects of age, ethnicity, menstrual cycle and combined oral contraceptive. Thromb Haemost 1999; 82: 145661.
  • 10
    Castaman G, Eikenboom JCJ, Bertina RM, Rodeghiero F. Inconsistency of association between type 1 von Willebrand disease phenotype and genotype in families identified in an epidemiological investigation. Thromb Haemost 1999; 82: 106570.
  • 11
    Casana P, Martinez F, Haya S, Espinos C, Aznar JA. Significant linkage and non-linkage of type 1 von Willebrand disease to the von Willebrand factor gene. Br J Haematol 2001; 115: 692700.
  • 12
    Tosetto A, Rodeghiero F, Castaman G, Goodeve A, Federici AB, Batlle J, Meyer D, Fressinaud E, Mazurier C, Goudemand J, Eikenboom J, Schneppenheim R, Budde U, Ingerslev J, Vorlova Z, Habart D, Holmbeg L, Lethagen S, Pasi J, Hill F, Peake I. A quantitative analysis of bleeding symptoms in type 1 von Willebrand disease: results from a multicenter European study (MCMDM-1 VWD). J Thromb Haemost 2006; 4: 76673.
  • 13
    Rodeghiero F, Castaman G, Tosetto A, Battle J, Baudo F, Cappelletti A, Casana P, De Bosch N, Eikenboom JCJ, Federici AB, Lethagen S, Linari S, Srivastava A. The discriminant power of bleeding history for the diagnosis of von Willebrand disease type 1: an international, multicenter study. J Thromb Haemost 2005; 3: 261926.
  • 14
    Schneppenheim R, Plendl H, Budde U. Luminography-an alternative assay for detection of von Willebrand factor multimers. Thromb Haemost 1988; 60: 1336.
  • 15
    Zhang ZP, Deng LP, Blombäck M, Anvret M. Dinucleotide repeat polymorphism in the promotor region of the human von Willebrand factor gene (vWF gene). Hum Mol Genet 1992; 1: 780.
  • 16
    Van Amstel HK, Reitsma PH. Tetranucleotide repeat polymorphism in the vWF gene. Nucleic Acids Res 1990; 18: 4957.
  • 17
    Kimpton C, Walton A, Gill P. A further tetranucleotide repeat polymorphism in the vWF gene. Hum Mol Genet 1992; 1: 287.
  • 18
    Clark JM. Novel non-templated nucleotide addition-reactions catalyzed by procaryotic and eucaryotic DNA-polymerases. Nucleic Acids Res 1988; 16: 967786.
  • 19
    Brownstein MJ, Carpten JD, Smith JR. Modulation of non-templated nucleotide addition by tag DNA polymerase: primer modifications that facilitate genotyping. Biotechniques 1996; 20: 10046.
  • 20
    Livak KJ. Allelic discrimination using fluorogenic probes and the 5nuclease assay. Genet Anal 1999; 14: 1439.
  • 21
    De Kok JB, Wiegerinck ETG, Giesendorf BAJ, Swinkels DW. Rapid genotyping of single nucleotide polymorphisms using novel minor groove binding DNA oligonucleotides (MGB probes). Human Mutation 2002; 19: 5549.
  • 22
    Yip SP. Single-tube multiplex PCR-SSCP analysis distinguishes 7 common ABO alleles and readily identifies new alleles. Blood 2000; 95: 148792.
  • 23
    S.A.G.E. Statistical Analysis for Genetic Epidemiology. Cork, Ireland: Computer program package available from Statistical Solutions Ltd, 2004.
  • 24
    Faraway JJ. Distribution of the admixture test for the detection of linkage under heterogeneity. Genet Epidemiol 1993; 10: 7583.
  • 25
    Sham PC, Purcell S, Cherny SS, Abecasis GR. Powerful regression-based quantitative-trait linkage analysis of general pedigrees. Am J Hum Genet 2002; 71: 23853.
  • 26
    Souto JC, Almasy L, Borrell M, Gari M, Martinez E, Mateo J, Stone WH, Blangero J, Fontcuberta J. Genetic determinants of hemostasis phenotypes in Spanish families. Circulation 2000; 101: 154651.
  • 27
    Rosendaal FR, Bovill EG. Heritability of clotting factors and the revival of the prothrombotic state. Lancet 2002; 359: 6389.
  • 28
    Orstavik KH, Magnus P, Reisner H, Berg K, Graham JB, Nance W. Factor VIII and factor IX in a twin population. Evidence for a major effect of ABO locus on factor VIII level. Am J Hum Genet 1998; 37: 89101.
  • 29
    De Lange M, Snieder H, Ariens RA, Spector TD, Grant PJ. The genetics of haemostasis: a twin study. Lancet 2001; 357: 1015.
  • 30
    Souto JC, Almasy L, Soria JM, Buil A, Stone W, Lathrop M, Blangero J, Fontcuberta J. Genome-wide linkage analysis of von Willebrand factor plasma levels: results from the GAIT project. Thromb Haemost 2003; 89: 46874.
  • 31
    De Visser MC, Sandkuijl LA, Lensen RP, Vos HL, Rosendaal FR, Bertina RM. Linkage analysis of factor VIII and von Willebrand factor loci as quantitative trait loci. J Thromb Haemost 2003; 1: 17716.
  • 32
    Mancuso DJ, Tuley EA, Westfield LA, Worrall NK, Shelton-Inloes B, Sorace JM, Alevy YG, Sadler JE. Structure of the gene for human von Willebrand factor. J Biol Chem 1989; 264: 1951427.