In Helsinki in 1924, Dr. Erik von Willebrand evaluated a five-year-old girl from the Åland Islands of Finland who had a history suggestive of an inherited bleeding abnormality. Subsequent investigation revealed that 23 of 66 family members were affected by a disease process with an apparent autosomal dominant mode of inheritance that was characterized clinically by abnormal mucocutaneous bleeding.
The eponymous disease first characterized by von Willebrand 90 years ago is thought to be the world’s most common inherited bleeding disorder, with a prevalence of 1 in 100 to 1 in 10,000, depending upon the criteria used to establish the diagnosis. The diagnosis of von Willebrand disease (VWD) incorporates the following three parameters: 1) a personal history of excessive mucocutaneous bleeding, 2) laboratory studies demonstrating quantitative and/or qualitative defects of von Willebrand factor (VWF), and 3) a family history of abnormal bleeding. Classification of VWD has remained the same throughout the past decade,1 but the level of VWF required to meet the diagnostic criteria for type 1 VWD has been lowered.
While diagnosis of type 2 VWD (a group of qualitative abnormalities of VWF with an autosomal dominant mode of inheritance), and type 3 VWD (complete or near-complete absence of VWF with an autosomal recessive inheritance pattern due to homozygous or compound heterozygous mutations of VWF) is relatively straightforward, such is not the case with type 1 VWD. Defined as a mild to moderate quantitative deficiency of VWF, approximately 70 to 80 percent of VWD cases are classified as type 1.
Diagnosis of Type 1 VWD
According to the National Heart, Lung, and Blood Institute (NHLBI) Guidelines2 and a recent publication by the British Committee for Standards in Haematology,3 a definitive diagnosis of type 1 VWD requires a VWF ristocetin cofactor (VWF:RCo) and/or a VWF antigen (VWF:Ag) of < 30 IU dL-1. However, it is important to appreciate that these criteria are based on expert opinion rather than on empiric evidence. Historically, the quantitative diagnostic criterion for type 1 was less stringent, requiring only a “low” VWF concentration that in most labs meant a level < 50 IU dL-1 for either VWF:Ag or VWF:RCo. However, because the plasma concentration of VWF is a continuous variable4 and because mild bleeding symptoms are relatively common in the general population,2 the NHLBI Guidelines and the British Committee for Standards in Haematology recommend that patients classified as having type 1 VWD have levels of < 30 IU dL-1, with those having plasma concentrations between 30 and 50 IU dL-1 being classified as having “low VWF.” This proposed distinction has proven to be clinically problematic, however, because individuals with low VWF may be at risk for bleeding and may warrant desmopressin (DDAVP) or other replacement therapy; yet insurance providers may balk at covering treatment for patients who do not carry a type-specific diagnosis of VWD.
In practice, experienced clinicians deciding on whether or not to recommend prophylactic DDAVP or replacement therapy in patients with a bleeding history use VWF levels in a manner analogous to the way that practitioners use risk factors for cardiovascular disease or thrombosis when making a decision about the use of heparin prophylaxis.4 Thus, an argument can be made that drawing a “line in the sand” for diagnosis of type 1 disease by requiring VWF levels of < 30 IU dL-1 is a flawed strategy that compromises patient care. In support of this position, emerging evidence suggests that the functionally relevant VWF concentration (the concentration below which pathologic mucocutaneous bleeding is observed) lies somewhere between 30 and 50 IU dL-1. Most clinicians experienced in the management of patients with disorders of hemostasis recommend treating patients with bleeding symptoms and low VWF, but doing so without financially compromising the patient requires educating insurance providers about the need for such treatment. Another approach to managing patients with low VWF is to forego prophylaxis prior to minor surgery but to have treatment available in the event that excess bleeding is observed. This approach can also guide future management recommendations. For example, a patient rescued from excessive bleeding follow a minor surgical treatment would be a candidate for prophylactic treatment should a subsequent procedure be required.
Progress in the diagnosis of type 1 VWD during the last few years has focused on 1) improvements in laboratory testing, 2) alternative tests of platelet-VWF–binding activity that do not require ristocetin, 3) more extensive testing for VWF binding to collagen, 4) incorporating genetic testing into laboratory diagnosis, and 5) development and validation of quantitative bleeding assessment tools (BATs). VWF:RCo activity has been the mainstay of functional testing for VWF binding to its cognate platelet receptor, glycoprotein Ibα (GPIbα). This assay, which takes advantage of the capacity of the antibiotic ristocetin to induce VWF-GPIbα interactions in vitro, is particularly difficult to standardize and has a coefficient of variation as high as 50 percent. Despite this high coefficient of variation, and acknowledging that ristocetin-induced platelet agglutination is not a normal physiologic process, VWF:RCo has remained the gold standard for diagnostic laboratory testing for VWF activity. Recently, modifications aimed at enhancing the read-out of platelet aggregation using chemiluminescence and turbidimetric detection methods have been introduced in an effort to improve VWF:RCo performance measures,5,6 and a ristocetin-independent ELISA assay that uses a gain-of-function GPIbα mutant to mediated VWF binding activity has been developed.7 Another measure of activity is based on the capacity of VWF to bind collagen. Discrete binding sites on VWF for collagen III as well as collagens IV and VI have been identified, and binding activity to each collagen type can be measured independently by using VWF fragments containing the particular binding site.8 Thus, testing for VWF activity requires assays that independently measure binding to each of the collagen-binding domains and to platelet GPIb.
BATs have been developed specifically for semiquantitative assessment of mucocutaneous bleeding, and clinically validated BATs, such as one developed by the International Society on Thrombosis and Haemostasis, are being used in the diagnosis of VWD.9 BATs are useful because they assign a numerical value to a predefined bleeding parameter. Therefore the information is readily transferable, making it valuable for standardizing outcomes in multicenter clinical studies, and score totals can be used to develop algorithms to guide diagnostic testing. For example BATs have strong negative predictive value, and therefore, results can be used to avoid subjecting patients to unneeded laboratory evaluation.9 But BATs have limitations. The Zimmerman Program for the Molecular and Clinical Biology of VWD studied 685 index cases that include 323 families with either type 1 VWD or low VWF. The frequency of having an abnormal BAT was 79 percent in adult subjects with VWF:RCo/VWF:Ag < 50 IU dL-1 and 73 percent in adults with VWF:RCo/VWF:Ag < 30 IU dL-1. Thus, this tool does not distinguish low VWF from type 1 VWD (or based on clinical grounds, perhaps such a distinction does not exist). Other problems associated with using BATs include limited utility in the assessment of pediatric patients who may have had relatively few challenges to their hemostatic system and whether BATs can predict future bleeding tendencies or whether they have value in guiding management longitudinally.
Although progress has been made in understanding the functional consequences of mutations affecting VWF, studies in the European Union, United Kingdom, Canada, and the United States have reported that sequence variations in the VWF are not found in 35 percent of individuals with type 1 VWD, and some of the sequence variations identified in patients with VWD type 1 are non–disease-causing polymorphisms rather than pathogenic mutations.10 Thus, nucleotide sequencing of VWF is not a definitive approach to the diagnosis of type 1 VWD. Molecular testing is complicated technically by the fact that VWF is replicated as a partial pseudogene on chromosome 22 (VWF is located on chromosome 12), and interpretation of sequencing results must incorporate structural/functional information associated with the > 1700 VWF sequence variants that have been identified. Because of the technical and interpretative complexity of molecular analysis, nucleotide sequencing of VWF as a diagnostic tool is limited.
Compelling data indicate that genetic determinants outside of the VWF locus affect the plasma concentration of VWF. The most striking example of this quantitative effect on VWF concentration by a non-VWF locus is ABO blood type.11 Approximately 30 percent of the genetic component of the variation of plasma VWF concentration is determined by ABO blood group, with individuals with blood type O having the lowest concentration, and individuals with blood type AB having the highest concentration (the concentration in type O individuals is about 40% lower than in type AB individuals). The mechanism underlying the effect of ABO blood type on VWF concentration is incompletely understood but may involve carbohydrate processes that affect the rate at which the protein is catabolized. Other non-VWF loci affecting VWF expression have been identified.12 One such example is the gene that encodes the C-type lectin receptor, CLEC4M.13 CLEC4M is a calcium-dependent mannose-specific receptor found in endothelial cells, and polymorphisms within CLEC4M have been shown to affect binding and internalization of VWF, thereby contributing to the genetically determined variation in the plasma concentration of VWF.13
Management of VWD
Management of type 1 VWD has remained largely the same for the last 20 years. DDAVP is widely used to treat mucocutaneous bleeding and bleeding associated with minimally invasive surgical procedures.14 However, in some patients with type 1 VWD with increased clearance of VWF, referred to as type 1C VWD,15 and in type 2 VWD variants (qualitative abnormalities), DDAVP may not be effective. Moreover, for reasons that are not understood, some type 1 patients simply do not respond to DDAVP. Therefore, patients in whom DDAVP therapy is being considered should be subjected to a therapeutic challenge with measurements of VWF:Ag, VWF:RCo, and FVIII:C at baseline, and at one, two, and four hours after DDAVP administration. Other therapeutic options include the antifibrinolytics (aminocaproic or tranexamic acid) and oral contraceptives for menorrhagia in women with VWD.
For those who do not respond to the aforementioned treatment, those with certain type 2 VWD variants, those with type III disease, or those undergoing major surgical procedures, plasma-derived VWF concentrates are used as factor replacement therapy. All commercially available VWF concentrates are subject to viral inactivation manufacturing steps but differ in the methods of protein fractionation and purification. More recently, a recombinant VWF concentrate (rVWF) has been found to be safe in a prospective clinical trial, but this product is awaiting approval by the FDA.16 The rVWF product was found to differ from plasma-derived VWF concentrates in that it contained high-molecular-weight multimers, a property that might enhance its specific activity.
Since Dr. von Willebrand’s remarkable observations nearly a century ago, extraordinary progress has been made in understanding the physiology of VWF, in elucidating the pathophysiology of VWD, and in improving approaches to diagnosis and management. But challenges remain. The diverse functionality of VWF within the hemostatic system adds to the difficulty of establishing genotype-phenotype correlations in type 1 VWD that might inform both diagnosis and management. Even as we learn more about structure-function relationships, however, the view of type I VWD as a binary process may be a flawed concept, and perpetuating this idea may actually be a disservice to patients. For the foreseeable future, we need to educate clinicians and insurance providers that patients with low VWF are at risk for abnormal bleeding and that these patients benefit from therapeutic strategies aimed at reducing the risk of surgery-associated bleeding complications.
Drs. Roberts, Montgomery, and White have no relevant conflicts of interest.