Scientific Overview

Discovery and Identification of VWD/VWF

The patient who led to the discovery of a hereditary bleeding disorder that we now call VWD was a 5-year-old girl who lived on the Åland Islands and was brought to Deaconess Hospital in Helsinki, Finland, in 1924 to be seen by Dr. Erik von Willebrand.¹⁰ He ultimately assessed 66 members of her family and reported in 1926 that this was a previously undescribed bleeding disorder that differed from hemophilia and exhibited (1) mucocutaneous bleeding, (2) autosomal inheritance rather than being linked to the X chromosome, (3) prolonged bleeding times by the Duke method (ear lobe bleeding time), and (4) normal clotting time. Not only did he recognize the autosomal inheritance pattern, but he recognized that bleeding symptoms were greater in children and in women of childbearing age. He subsequently found that blood transfusions were useful not only to correct the anemia but also to control bleeding.

In the 1950s, it became clear that a "plasma factor," antihemophilic factor (FVIII), was decreased in these persons and that Cohn fraction I-0 could correct both the plasma deficiency of FVIII and the prolonged bleeding time. For the first time, the factor causing the long bleeding time was called "von Willebrand factor." As cryoprecipitate and commercial FVIII concentrates were developed, it was recognized that both VWF and "antihemophilic factor" (FVIII) purified together.

When immunoassays were developed, persons who had VWD (in contrast to those who had hemophilia A) were found to have reduced "factor VIII-related antigen" (FVIIIR:Ag), which we now refer to as VWF:Ag. Characterization of the proteins revealed that FVIII was the clotting protein deficient in hemophilia A, and VWF was a separate "FVIII carrier protein" that resulted in the cofractionation of both proteins in commercial concentrates. Furthermore, a deficiency of VWF resulted in increased FVIII clearance because of the reduced carrier protein, VWF. Since the 1980s, molecular and cellular studies have defined hemophilia A and VWD more precisely. Persons who had VWD had a normal FVIII gene on the X chromosome, and some were found to have an abnormal VWF gene on chromosome 12. Variant forms of VWF were recognized in the 1970s, and we now recognize that these variations are the result of synthesis of an abnormal protein. Gene sequencing identified many of these persons as having a VWF gene mutation. The genetic causes of milder forms of low VWF are still under investigation, and these forms may not always be caused by an abnormal VWF gene. In addition, there are acquired disorders that may result in reduced or dysfunctional VWF (see section on "Acquired von Willebrand Syndrome" [AVWS]). Table 2 contains a synopsis of VWF designations, functions, and assays. Table 3 contains abbreviations used throughout this document.

Table 2. Synopsis of VWF Designations, Properties, and Assays

*See Table 3. Nomenclature and Abbreviations.

The VWF Protein and Its Functions In Vivo

VWF is synthesized in two cell types. In the vascular endothelium, VWF is synthesized and subsequently stored in secretory granules (Weibel-Palade bodies) from which it can be released by stress or drugs such as desmopressin (DDAVP, 1-desamino-8-D-arginine vasopressin), a synthetic analog of vasopressin. VWF is also synthesized in bone marrow megakaryocytes where it is stored in platelet alpha-granules from which it is released following platelet activation. DDAVP does not release platelet VWF.

VWF is a protein that is assembled from identical subunits into linear strings of varying size referred to as multimers. These multimers can be >20 million daltons in mass and >2 micrometers in length. The complex cellular processing consists of dimerization in the endoplasmic reticulum (ER), glycosylation in the ER and Golgi, multimerization in the Golgi, and packaging into storage granules. The latter two processes are under the control of the VWF propeptide (VWFpp), which is cleaved from VWF at the time of storage. VWF that is released acutely into the circulation is accompanied by a parallel rise in FVIII, but it is still not entirely clear whether this protein–protein association first occurs within the endothelial cell.^11,12

In plasma, the FVIII–VWF complex circulates as a loosely coiled protein complex that does not interact strongly with platelets or endothelial cells under basal conditions. When vascular injury occurs, VWF becomes tethered to the exposed subendothelium (collagen, etc.). The high fluid shear rates that occur in the microcirculation appear to induce a conformational change in multimeric VWF that causes platelets to adhere, become activated, and then aggregate so as to present an activated platelet phospholipid surface. This facilitates clotting that is, in part, regulated by FVIII. Because of the specific characteristics of hemostasis and fibrinolysis on mucosal surfaces, symptoms in VWD are often greater in these tissues.

Plasma VWF is primarily derived from endothelial synthesis. Platelet and endothelial cell VWF are released locally following cellular activation where this VWF participates in the developing hemostatic plug or thrombus (see Figure 1).

Plasma VWF has a half-life of approximately 12 hours (range 9–15 hours). VWF is present as very large multimers that are subjected to physiologic degradation by the metalloprotease ADAMTS13 (A Disintegrin-like And Metalloprotease domain [reprolysin type] with Thrombospondin type I motifs). Deficiency of ADAMTS13 is associated with the pathologic microangiopathy of thrombotic thrombocytopenic purpura (TTP). The most common variant forms of type 2A VWD are characterized by increased VWF susceptibility to ADAMTS13.

Table 3. Nomenclature and Abbreviations

*These abbreviations (for FVIII and VWF and all their properties) are defined in Marder VJ, Mannucci PM, Firkin BG, Hoyer LW, Meyer D. Standard nomenclature for factor VIII and von Willebrand factor: a recommendation by the International Committee on Thrombosis and Haemostasis. Thromb Haemost 1985 Dec;54(4):871–872; Mazurier C, Rodeghiero F. Recommended abbreviations for von Willebrand Factor and its activities. Thromb Haemost 2001 Aug;86(2):712.

Factors that affect levels of plasma VWF include age, race, ABO and Lewis blood groups, epinephrine, inflammatory mediators, and endocrine hormones (particularly those associated with the menstrual cycle and pregnancy). VWF is increased during pregnancy (a three- to fivefold elevation over the woman's baseline by the third trimester), with aging, and with acute stress or inflammation. Africans and African Americans have higher average levels of VWF than the Caucasian population.^13,14 VWF is reduced by hypothyroidism and rarely by autoantibodies to VWF. The rate of VWF synthesis probably is not affected by blood group; however, the survival of VWF appears to be reduced in individuals who have type O blood. In fact, ABO blood group substance has been identified on VWF.

The Genetics of VWDM

Since the 1980s, molecular and cellular studies have defined hemophilia A and VWD more precisely. Persons who have severe VWD have a normal FVIII gene on the X chromosome, and some are found to have an abnormal VWF gene on chromosome 12. The VWF gene is located near the tip of the short arm of chromosome 12, at 12p13.3.¹⁵ It spans approximately 178 kb of DNA and contains 52 exons.¹⁶ Intron–exon boundaries tend to delimit structural domains in the protein, and introns often occur at similar positions within the gene segments that encode homologous domains. Thus, the structure of the VWF gene reflects the mosaic nature of the protein (Figure 2).

A partial, unprocessed VWF pseudogene is located at chromosome 22q11.2.¹⁷ This pseudogene spans approximately 25 kb of DNA and corresponds to exons 23–34 and part of the adjacent introns of the VWF gene.¹⁸ This segment of the gene encodes domains A1A2A3, which contain binding sites for platelet glycoprotein Ib (GPIb) and collagen, as well as the site cleaved by ADAMTS13. The VWF pseudogene and gene have diverged 3.1 percent in DNA sequence, consistent with a relatively recent origin of the pseudogene by partial gene duplication.¹⁸ This pseudogene is found in humans and great apes (bonobo, chimpanzee, gorilla, orangutan) but not in more distantly related primates.¹⁹ The VWF pseudogene complicates the detection of VWF gene mutations because polymerase chain reactions (PCRs) can inadvertently amplify segments from either or both loci, but this difficulty can be overcome by careful design of gene-specific PCR primers.¹⁸

Figure 1. VWF and Normal Hemostasis

A cross-sectioned blood vessel shows stages of hemostasis. Top, VWF is the carrier protein for blood clotting factor VIII (FVIII). Under normal conditions VWF does not interact with platelets or the blood vessel wall that is covered with endothelial cells. Middle left, following vascular injury, VWF adheres to the exposed subendothelial matrix. Middle right, after VWF is uncoiled by local shear forces, platelets adhere to the altered VWF and these platelets undergo activation and recruit other platelets to this injury site. Bottom left, the activated and aggregated platelets alter their membrane phospholipids exposing phosphatidylserine, and this activated platelet surface binds clotting factors from circulating blood and initiates blood clotting on this surface where fibrin is locally deposited. Bottom right, the combination of clotting and platelet aggregation and adhesion forms a platelet-fibrin plug, which results in the cessation of bleeding. The extent of the clotting is carefully regulated by natural anticoagulants. Subsequently, thrombolysis initiates tissue repair and ultimately the vessel may be re-endothelialized and blood flow maintained.
Note: Used by permission of R.R. Montgomery.

The VWF pseudogene may occasionally serve as a reservoir of mutations that can be introduced into the VWF locus. For example, some silent and some potentially pathogenic mutations have been identified in exons 27 and 28 of the VWF gene of persons who have VWD. These same sequence variations occur consecutively in the VWF pseudogene and might have been transferred to the VWF by gene conversion.^20-22 The segments involved in the potential gene conversion events are relatively short, from a minimum of 7 nucleotides²⁰ to a maximum of 385 nucleotides.²² The frequency of these potential interchromosomal exchanges is unknown.

The spectrum of VWF gene mutations that cause VWD is similar to that of many other human genetic diseases and includes large deletions, frameshifts from small insertions or deletions, splice-site mutations, nonsense mutations causing premature termination of translation, and missense mutations affecting single amino acid residues. A database of VWF mutations and polymorphisms has been compiled for the International Society on Thrombosis and Haemostasis (ISTH)^23,24 and is maintained for online access at the University of Sheffield (http://www.shef.ac.uk/vwf/index.html). Mutations causing VWD have been identified throughout the VWF gene. In contrast to hemophilia A, in which a single major gene rearrangement causes a large fraction of severe disease, no such recurring mutation is common in VWD. There is a good correlation between the location of mutations in the VWF gene and the subtype of VWD, as discussed in more detail in "Classification of VWD Subtypes." In selected families, this information can facilitate the search for VWF mutations by DNA sequencing.

Figure 2. Structure and Domains of VWF

The von Willebrand factor (VWF) protein sequence (amino acid 1–2813) is aligned with the cDNA sequence (nucleic acid 1–8439). The VWF signal peptide is the first 22 aa, the propeptide (VWFpp) aa 23–763, and mature VWF aa 764–2800. Type 2 mutations are primarily located in specific domains (regions) along the VWF protein. Types 2A, 2B, and 2M VWF mutations are primarily located within exon 28 that encodes for the A1 and A2 domains of VWF. The two different types of 2A are those that have increased proteolysis (2A²) and those with abnormal multimer synthesis (2A¹). Type 2N mutations are located within the D′ and D3 domains. Ligands that bind to certain VWF domains are identified, including FVIII, heparin, GPIb (platelet glycoprotein Ib complex), collagen, and GPIIb/IIIa (platelet glycoprotein IIb/IIIa complex that binds to the RGD [arginine-glycine-aspartate] amino acid sequence in VWF).
Note: Used by permission of R.R. Montgomery.

Classification of VWD Subtypes

VWD is classified on the basis of criteria developed by the VWF Subcommittee of the ISTH, first published in 1994 and revised in 2006 (Table 4).^25,26

The classification was intended to be clinically relevant to the treatment of VWD. Diagnostic categories were defined that encompassed distinct pathophysiologic mechanisms and correlated with the response to treatment with DDAVP or blood products. The classification was designed to be conceptually independent of specific laboratory testing procedures, although most of the VWD subtypes could be assigned by using tests that were widely available. The 1994 classification reserved the designation of VWD for disorders caused by mutations within the VWF gene,²⁵ but this criterion has been dropped from the 2006 classification²⁶ because in practice it is verifiable for only a small fraction of patients.

VWD is classified into three major categories: partial quantitative deficiency (type 1), qualitative deficiency (type 2), and total deficiency (type 3). Type 2 VWD is divided further into four variants (2A, 2B, 2M, 2N) on the basis of details of the phenotype. Before the publication of the 1994 revised classification of VWD,²⁵ VWD subtypes were classified using Roman numerals (types I, II, and III), generally corresponding to types 1, 2, and 3 in the 1994 classification, and within type II several subtypes existed (designated by adding sequential letters of the alphabet; i.e., II-A through II-I). Most of the latter VWD variants were amalgamated as type 2A in the 1994 classification, with the exception of type 2B (formerly II-B) for which a separate new classification was created. In addition, a new subtype (2M) was created to include variants with decreased platelet dependent function (VWF:RCo) but no significant decrease of higher molecular weight VWF multimers (which may or may not have other aberrant structure), with "M" representing "multimer." Subtype 2N VWD was defined, with "N" representing "Normandy" where the first individuals were identified, with decreased FVIII due to VWF defects of FVIII binding.

Type 1 VWD affects approximately 75 percent of symptomatic persons who have VWD (see Castaman et al., 2003 for a review).²⁷ Almost all of the remaining persons are divided among the four type 2 variants, and the partitioning among them varies considerably among centers. In France, for example, patients’ distribution was reported to be 30 percent type 2A, 28 percent type 2B, 8 percent type 2M (or unclassified), and 34 percent type 2N.²⁸ In Bonn, Germany, the distribution was reported to be 74 percent type 2A, 10 percent type 2B, 13 percent type 2M, and 3.5 percent type 2N.²⁹ Table 5 summarizes information about inheritance, prevalence, and bleeding propensity in persons who have different types of VWD.

Table 4. Classification of VWD

Note: VWD types are defined as described in Sadler JE, Budde U, Eikenboom JC, Favaloro EJ, Hill FG, Holmberg L, Ingerslev J, Lee CA, Lillicrap D, Mannucci PM, et al. Update on the pathophysiology and classification of von Willebrand disease: a report of the Subcommittee on von Willebrand Factor. J Thromb Haemost 2006 Oct;4(10):2103–2114.

Table 5. Inheritance, Prevalence, and Bleeding Propensity in Patients Who Have VWD

The prevalence of type 3 VWD in the population is not known precisely but has been estimated (per million population) as: 0.55 for Italy,³⁰ 1.38 for North America,³¹ 3.12 for Sweden,³⁰ and 3.2 for Israel.³² The prevalence may be as high as 6 per million where consanguinity is common.¹

Type 1 VWD

Type 1 VWD is found in persons who have partial quantitative deficiency of VWF. The level of VWF in plasma is low, and the remaining VWF mediates platelet adhesion normally and binds FVIII normally. Laboratory evaluation shows concordant decreases in VWF protein concentration (VWF:Ag) and assays of VWF function (VWF:RCo). Levels of blood clotting FVIII usually parallel VWF and may be reduced secondary to reduced VWF. Usually, in type 1 VWD, the FVIII/VWF:Ag ratio is 1.5–2.0. In most persons who have type 1 VWD, this results in FVIII being normal, or mildly decreased, and not reduced as much as the VWF. VWF multimer gels show no significant decrease in large VWF multimers.²⁵ The laboratory evaluation of VWD is discussed in the "Diagnosis and Evaluation" section.

The spectrum of mutations occurring in VWD type 1 has been described extensively in two major studies.^33,34 Particularly severe, highly penetrant forms of type 1 VWD may be caused by dominant VWF mutations that interfere with the intracellular transport of dimeric proVWF^35-39 or that promote the rapid clearance of VWF from the circulation.^38,40,41 Persons who have such mutations usually have VWF levels <20 IU/dL.^33,34 Most of the mutations characterized to date cause single amino acid substitutions in domain D3.^35-37,39,42 One mutation associated with rapid clearance has been reported in domain D4.³⁸

Increased clearance of VWF from the circulation in type 1 VWD may account for the exaggerated but unexpectedly brief responses to DDAVP observed in some patients. Consequently, better data on the prevalence of increased clearance could affect the approach to diagnosing type 1 VWD and the choice of treatment for bleeding.

A diagnosis of type 1 VWD is harder to establish when the VWF level is not markedly low but instead is near the lower end of the normal range. Type 1 VWD lacks a qualitative criterion by which it can be recognized and instead relies only on quantitative decrements of protein concentration and function. VWF levels in the healthy population span a wide range of values. The mean level of plasma VWF is 100 IU/dL, and approximately 95 percent of plasma VWF levels lie between 50 and 200 IU/dL.^43,44 Because mild bleeding symptoms are very common in the healthy population, the association of bleeding symptoms with a moderately low VWF level may be coincidental.⁴⁵ The conceptual and practical issues associated with the evaluation of moderately low VWF levels are discussed more completely later in this section. (See "Type 1 VWD Versus Low VWF: VWF Level as a Risk Factor for Bleeding.")

Type 2 VWD

The clinical features of several type 2 VWD variants are distinct from those of type 1 VWD, and they can have strikingly distinct and specific therapeutic needs. As a consequence, the medical care of patients who have type 2 VWD benefits from the participation of a hematologist who has expertise in hemostasis. Bleeding symptoms in type 2 VWD are often thought to be more severe than in type 1 VWD, although this impression needs to be evaluated in suitable clinical studies.

Type 2A VWD refers to qualitative variants in which VWF-dependent platelet adhesion is decreased because the proportion of large VWF multimers is decreased. Levels of VWF:Ag and FVIII may be normal or modestly decreased, but VWF function is abnormal as shown by markedly decreased VWF:RCo.⁴⁶ Type 2A VWD may be caused by mutations that interfere with the assembly or secretion of large multimers or by mutations that increase the susceptibility of VWF multimers to proteolytic degradation in the circulation.^47-49 The deficit of large multimers predisposes persons to bleed.

The location of type 2A VWD mutations sometimes can be inferred from high-resolution VWF multimer gels. For example, mutations that primarily reduce multimer assembly lead to the secretion of multimers that are too small to engage platelets effectively and therefore are relatively resistant to proteolysis by ADAMTS13. Homozygous mutations in the propeptide impair multimer assembly in the Golgi and give rise to a characteristic "clean" pattern of small multimers that lack the satellite bands usually associated with proteolysis (see "Diagnosis and Evaluation"); this pattern was initially described as "type IIC" VWD.^50-52 Heterozygous mutations in the cystine knot (CK) domain can impair dimerization of proVWF in the ER and cause a recognizable multimer pattern originally referred to as "type IID."^53,54 A mixture of monomers and dimers arrives in the Golgi, where the incorporation of monomers at the end of a multimer prevents further elongation. As a result, the secreted small multimers contain minor species with an odd number of subunits that appear as faint bands between the usual species that contain an even number of subunits. Heterozygous mutations in cysteine residues of the D3 domain also can impair multimer assembly, but these mutations often also produce an indistinct or "smeary" multimer pattern referred to as "type IIE."^55,56

In contrast to mutations that primarily affect multimer assembly, mutations within or near the A2 domain of VWF cause type 2A VWD that is associated with markedly increased proteolysis of the VWF subunits⁵⁶ (see Figure 2). These mutations apparently interfere with the folding of the A2 domain and make the Tyr1605–Met1606 bond accessible to ADAMTS13 even in the absence of increased fluid shear stress. Two subgroups of this pattern have been distinguished: group I mutations enhance proteolysis by ADAMTS13 and also impair multimer assembly, whereas group II mutations enhance proteolysis without decreasing the assembly of large VWF multimers.⁴⁹ Computer modeling of domain A2 suggests that group I mutations affect both assembly and proteolysis, because group I mutations have a more disruptive effect on the folding of domain A2 than do group II mutations.⁵⁷

Type 2B VWD is caused by mutations that pathologically increase platelet–VWF binding, which leads to the proteolytic degradation and depletion of large, functional VWF multimers.^56,58 Circulating platelets also are coated with mutant VWF, which may prevent the platelets from adhering at sites of injury.⁵⁹

Although laboratory results for type 2B VWD may be similar to those in type 2A or type 2M VWD, patients who have type 2B VWD typically have thrombocytopenia that is exacerbated by surgery, pregnancy, or other stress.^60-62 The thrombocytopenia probably is caused by reversible sequestration of VWF–platelet aggregates in the microcirculation. These aggregates are dissolved by the action of ADAMTS13 on VWF, causing the characteristic decrease of large VWF multimers and the prominent satellite banding pattern that indicates increased proteolytic degradation.^63,64 The diagnosis of type 2B VWD depends on finding abnormally increased ristocetin induced platelet aggregation (RIPA) at low concentrations of ristocetin.

Type 2B VWD mutations occur within or adjacent to VWF domain A1,^23,55,65-68 which changes conformation when it binds to platelet GPIb.⁶⁹ The mutations appear to enhance platelet binding by stabilizing the bound conformation of domain A1.

Type 2M VWD includes variants with decreased VWF-dependent platelet adhesion that is not caused by the absence of high-molecular-weight VWF multimers. Instead, type 2M VWD mutations reduce the interaction of VWF with platelet GPIb or with connective tissue and do not substantially impair multimer assembly. Screening laboratory results in type 2M VWD and type 2A VWD are similar, and the distinction between them depends on multimer gel electrophoresis.⁶⁷

Mutations in type 2M VWD have been identified in domain A1 (see Figure 2), where they interfere with binding to platelet GPIb.^{23,55,67,70-72} One family has been reported in which a mutation in VWF domain A3 reduces VWF binding to collagen, thereby reducing platelet adhesion and possibly causing type 2M VWD.⁷³

Type 2N VWD is caused by VWF mutations that impair binding to FVIII, lowering FVIII levels so that type 2N VWD masquerades as an autosomal recessive form of hemophilia A.^74-76 In typical cases, the FVIII level is less than 10 percent, with a normal VWF:Ag and VWF:RCo. Discrimination from hemophilia A may require assays of FVIII–VWF binding.^77,78

Most mutations that cause type 2N VWD occur within the FVIII binding site of VWF (see Figure 2), which lies between residues Ser764 and Arg1035 and spans domain D′ and part of domain D3.^23,79,80 The most common mutation, Arg854Gln, has a relatively mild effect on FVIII binding and tends to cause a less severe type 2N VWD phenotype.⁷⁷ Some mutations in the D3 domain C-terminal of Arg1035 can reduce FVIII binding,^81-83 presumably through an indirect effect on the structure or accessibility of the binding site.

Type 3 VWD

Type 3 VWD is characterized by undetectable VWF protein and activity, and FVIII levels usually are very low (1–9 IU/dL).^84-86 Nonsense and frameshift mutations commonly cause type 3 VWD, although large deletions, splice-site mutations, and missense mutations also can do so. Mutations are distributed throughout the VWF gene, and most are unique to the family in which they were first identified.^23,87,88

A small fraction of patients who have type 3 VWD develop alloantibodies to VWF in response to the transfusion of plasma products. These antibodies have been reported in 2.6–9.5 percent of patients who have type 3 VWD, as determined by physician surveys or screening.^85,89 The true incidence is uncertain, however, because of unavoidable selection bias in these studies. Anti-VWF alloantibodies can inhibit the hemostatic effect of blood-product therapy and also may cause life-threatening allergic reactions.^85,90 Large deletions in the VWF gene may predispose patients to this complication.⁸⁹

VWD Classification, General Issues

The principal difficulties in using the current VWD classification concern how to define the boundaries between the various subtypes through laboratory testing. In addition, some mutations have pleiotropic effects on VWF structure and function, and some persons are compound heterozygous for mutations that cause VWD by different mechanisms. This heterogeneity can produce complex phenotypes that are difficult to categorize. Clinical studies of the relationship between VWD genotype and clinical phenotype would be helpful to improve the management of patients with the different subtypes of VWD.

The distinction between quantitative (type 1) and qualitative (type 2) defects depends on the ability to recognize discrepancies among VWF assay results,^80,91 as discussed in "Diagnosis and Evaluation." Similarly, distinguishing between type 2A and type 2M VWD requires multimer gel analysis. Standards need to be established for using laboratory tests to make these important distinctions.

The example of Vicenza VWD illustrates some of these problems. Vicenza VWD was first described as a variant of VWD in which the level of plasma VWF is usually <15 IU/dL and the VWF multimers are even larger than normal, like the ultralarge multimers characteristic of platelet VWF.⁹² The low level of VWF in plasma in Vicenza VWD appears to be explained by the effect of a specific mutation, Arg1205His, that promotes clearance of VWF from the circulation about fivefold more rapidly than normal.⁴¹ Because the newly synthesized multimers have less opportunity to be cleaved by ADAMTS13 before they are cleared, accelerated clearance alone may account for the increased multimer size in Vicenza VWD.⁹³ Whether Vicenza VWD is classified under type 1 VWD or type 2M VWD depends on the interpretation of laboratory test results. The abnormally large multimers and very low RIPA values have led some investigators to prefer the designation of type 2M VWD.⁹⁴ However, the VWF:RCo/VWF: Ag ratio typically is normal, and large VWF multimers are not decreased relative to smaller multimers, so that other investigators have classified Vicenza VWD under type 1 VWD.⁴¹ Regardless of how this variant is classified, the markedly shortened half-life of plasma VWF in Vicenza VWD is a key fact that, depending on the clinical circumstance, may dictate whether the patient should receive treatment with DDAVP or FVIII/VWF concentrates.

Type 1 VWD Versus Low VWF: VWF Level as a Risk Factor for Bleeding

Persons who have very low VWF levels, <20 IU/dL, are likely to have VWF gene mutations, significant bleeding symptoms, and a strongly positive family history.^{33,34,37,95-99} Diagnosing such persons as having type 1 VWD seems appropriate because they may benefit from changes in lifestyle and from specific treatments to prevent or control bleeding. Identification of affected family members also may be useful, and genetic counseling is simplified when the pattern of inheritance is straightforward.

On the other hand, VWF levels of 30–50 IU/dL, just below the usual normal range (50–200 IU/dL), pose problems for diagnosis and treatment. Among the total U.S. population of approximately 300 million, VWF levels <50 IU/dL are expected in about 7.5 million persons, who therefore would be at risk for a diagnosis of type 1 VWD. Because of the strong influence of ABO blood group on VWF level,⁴³ about 80 percent of U.S. residents who have low VWF also have blood type O. Furthermore, moderately low VWF levels and bleeding symptoms generally are not coinherited within families and are not strongly associated with intragenic VWF mutations.^100-102 In a recent Canadian study of 155 families who had type 1 VWD, the proportion showing linkage to the VWF locus was just 41 percent.⁹⁸ In a similar European study, linkage to the VWF locus depended on the severity of the phenotype. If plasma levels of VWF were <30 IU/dL, linkage was consistently observed, but if levels of VWF were >30 IU/dL, the proportion of linkage was only 51 percent.⁹⁷ Furthermore, bleeding symptoms were not significantly linked to the VWF gene in these families.⁹⁷

Family studies suggest that 25–32 percent of the variance in plasma VWF is heritable.^103,104 Twin studies have reported greater heritability of 66–75 percent,^105,106 although these values may be overestimates because of shared environmental factors.^104,107 Therefore, it appears that, at least in the healthy population, a substantial fraction of the variation in VWF level is not heritable.

Few genes have been identified that contribute to the limited heritability of VWF level. The major genetic influence on VWF level is ABO blood group, which is thought to account for 20–30 percent of its heritable variance.^13,106,108 The mean VWF level for blood type O is 75 U/dL, which is 25–35 U/dL lower than other ABO types, and 95 percent of VWF levels for type O blood donors are between 36 and 157 U/dL.⁴³ The Secretor locus has a smaller effect. Secretor-null persons have VWF levels slightly lower than Secretors.¹⁰⁹

Table 6. Bleeding and VWF Level in Type 3 VWD Heterozygotes

An effect of the VWF locus has been difficult to discern by linkage analysis. One study suggested that 20 percent of the variance in VWF levels is attributable to the VWF gene,¹⁰⁸ whereas another study could not demonstrate such a relationship.¹¹⁰

In sum, known genetic factors account for a minority of the heritable variation in VWF level, and moderately low VWF levels (30–50 IU/dL) do not show consistent linkage to the VWF locus.^{97,98,100,101} The diagnosis and management of VWD would be facilitated by better knowledge of how inherited and environmental factors influence the plasma concentration of VWF.

The attribution of bleeding to a low VWF level can be difficult because mild bleeding symptoms are very common, as discussed in the section on "Diagnosis and Evaluation" and the risk of bleeding is only modestly increased for persons who have moderately decreased VWF levels.⁴⁵ For example, in the course of investigating patients who have type 3 VWD, approximately 190 obligate heterozygous relatives have had bleeding histories obtained and VWF levels measured (see Table 6). The geometric mean VWF level was 47 IU/dL,⁴⁵ with a range (±2 SD) of 16–140 IU/dL. Among 117 persons who had VWF <50 IU/dL, 31 (26 percent) had bleeding symptoms. Among 74 persons who had VWF >50 IU/dL, 10 (14 percent) had bleeding symptoms. Therefore, the relative risk of bleeding was 1.9 (P = 0.046, Fisher's exact test) for persons who had low VWF. There was a trend for an increased frequency of bleeding symptoms at the lowest VWF levels: among 31 persons who had VWF levels <30 IU/dL, 12 (39 percent) had symptoms. Bleeding was mild and consisted of epistaxis, bruising, menorrhagia, and bleeding after tooth extraction. The one person who experienced postoperative bleeding had a VWF level >50 IU/dL.¹¹³

The management of bleeding associated with VWF deficiency would be facilitated by better understanding of the heritability of low VWF levels (in the range of 20–50 IU/dL), their association with intragenic VWF mutations, and their interactions with other modifiers of bleeding risk. Such data could provide a foundation for treating VWF level as a biomarker for a moderate risk of bleeding, much as high blood pressure and high cholesterol are treated as biomarkers for cardiovascular disease (CVD) risk.

Acquired von Willebrand Syndrome

Acquired von Willebrand syndrome (AVWS) refers to defects in VWF concentration, structure, or function that are not inherited directly but are consequences of other medical disorders. Laboratory findings in AVWS are similar to those in VWD and may include decreased values for VWF:Ag, VWF:RCo, or FVIII. The VWF multimer distribution may be normal, but the distribution often shows a decrease in large multimers similar to that seen in type 2A VWD.^117,118 AVWS usually is caused by one of three mechanisms: autoimmune clearance or inhibition of VWF, increased shear-induced proteolysis of VWF, or increased binding of VWF to platelets or other cell surfaces. Autoimmune mechanisms may cause AVWS in association with lymphoproliferative diseases, monoclonal gammopathies, systemic lupus erythematosis, other autoimmune disorders, and some cancers. Autoantibodies to VWF have been detected in less than 20 percent of patients in whom they have been sought, suggesting that the methods for antibody detection may not be sufficiently sensitive or that AVWS in these settings may not always have an autoimmune basis.

Pathologic increases in fluid shear stress can occur with cardiovascular lesions, such as ventricular septal defect and aortic stenosis, or with primary pulmonary hypertension. The increased shear stress can increase the proteolysis of VWF by ADAMTS13 enough to deplete large VWF multimers and thereby produce a bleeding diathesis that resembles type 2A VWD. The VWF multimer distribution improves if the underlying cardiovascular condition is treated successfully.^117-122

Increased binding to cell surfaces, particularly platelets, also can consume large VWF multimers. An inverse relationship exists between the platelet count and VWF multimer size, probably because increased encounters with platelets promote increased cleavage of VWF by ADAMTS13. This mechanism probably accounts for AVWS associated with myeloproliferative disorders; reduction of the platelet count can restore a normal VWF multimer distribution.^123-125 In rare instances, VWF has been reported to bind GPIb that was expressed ectopically on tumor cells.^118,126

AVWS has been described in hypothyroidism caused by nonimmune mechanism.¹²⁷ Several drugs have been associated with AVWS; those most commonly reported include valproic acid, ciprofloxacin, griseofulvin, and hydroxyethyl starch.^117,118

AVWS occurs in a variety of conditions, but other clinical features may direct attention away from this potential cause of bleeding. More studies are needed to determine the incidence of AVWS and to define its contribution to bleeding in the many diseases and conditions with which it is associated.

Prothrombotic Clinical Issues and VWF in Persons Who Do Not Have VWD

Whether elevation of VWF is prothrombotic has been the subject of several investigations. Both arterial and venous thrombotic disorders have been studied.

Open-heart surgery. Hemostatic activation after open-heart surgery has been suggested as a mechanism of increased risk of postoperative thrombosis in this setting. A randomized trial comparing coronary artery surgery with or without cardiopulmonary bypass ("off-pump") found a consistent and equivalent rise in VWF:Ag levels at 1–4 postoperative days in the two groups,¹²⁸ suggesting that the surgery itself, rather than cardiopulmonary bypass, was responsible for the rise in VWF. There is no direct evidence that the postoperative rise in VWF contributes to the risk of thrombosis after cardiac surgery.

Coronary artery disease. Three large prospective studies of subjects without evidence of ischemic heart disease at entry have shown, by univariate analysis, a significant association of VWF:Ag level at entry with subsequent ischemic coronary events.^129-131 However, the association remained significant by multivariate analysis in only one subset of subjects in these studies,¹²⁹ a finding that could have occurred by chance. These findings suggest that the association of VWF with incidence of coronary ischemic events is relatively weak and may not be directly causal.

Thrombosis associated with atrial fibrillation. A prospective study of vascular events in subjects with atrial fibrillation found, by univariate analysis, a significant association of VWF:Ag level with subsequent stroke or vascular events. The association with vascular events remained significant with multivariate analysis.¹³²

Thrombotic thrombocytopenic purpura (TTP). The hereditary deficiency or acquired inhibition of a VWF-cleaving protease, ADAMTS13, is associated with the survival in plasma of ultralarge VWF multimers, which are involved in the propensity to development of platelet-rich thrombi in the microvasculature of individuals who have TTP.^133,134

Deep vein thrombosis (DVT). In a case-control study of 301 patients, evaluated at least 3 months after cessation of anticoagulation treatment for a first episode of DVT, plasma levels of VWF:Ag and FVIII activity were related to risk of DVT, according to univariate analysis. In multivariate analysis, the relation of VWF level with risk of DVT was not significant after adjustment for FVIII levels.¹³⁵

Back to Top
Back to Table of Contents