von Willebrand disease update: diagnostic and treatment dilemmas
The authors stated that they had no interests which might be perceived as posing a conflict or bias.
Abstract
Summary. Although von Willebrand disease (VWD) is now well-described, many facets of diagnosis and management continue to be debated. The diagnosis of type 1 disease can be difficult but recent genetic analyses help to distinguish many factors which can influence von Willebrand factor (VWF) levels and bleeding phenotype. Type 2 disease (functional abnormalities) includes a particularly interesting group of disorders with faulty binding between VWF and FVIIIC (Normandy) where treatment methods need careful consideration. Type 3 VWD is the most severe form of VWD and a new international study is underway to examine the use of prophylaxis.
Introduction
In 1926, the Finnish physician Erik Adolf von Willebrand first described the inherited bleeding disorder with clinical features distinguishing it from haemophilia and which is now known as von Willebrand disease (VWD) [1]. Erik von Willebrand′s first case was a girl, named Hjördis, aged 5 years with marked and recurrent bleeding. Both her mother and father belonged to families with histories of bleeding. Hjördis was the ninth of 11 children, of whom seven had experienced bleeding symptoms. Four of her sisters had died from uncontrolled bleeding at an early age. What differentiated this bleeding disorder from classical haemophilia was that it appeared typically not to be associated with muscle and joint bleeding and it affected both women and men. Hjördis herself experienced several severe episodes of bleeding from the nose and lips and following tooth extractions, as well as bleeding from the ankle. At the age of 3 years, she had severe bleeding from a deep wound in her upper lip and required hospitalization for 10 weeks. At the age of 13 years, she bled to death from menstrual bleeding. The clinical description that Erik von Willebrand presented in his original paper, based on this large family living on the island of Föglö in the Åland archipelago in the Baltic Sea, is still valid. These early descriptions defined what we now know is the most severe form of VWD, type 3.
VWD is the most common inherited bleeding disease in humans. Recent population-based studies suggest that approximately one in 1000 subjects experience clinical symptoms as a result of this condition.
Since the original observations by von Willebrand, the disease has been extensively studied and it was shown in the mid 1950s that the impaired haemostasis was because of lack or an abnormality of a plasma factor – the von Willebrand factor (VWF) – present both in normal and in haemophilia A patients. The first steps toward effective replacement therapy were taken [2] and today the available treatment options have broadly expanded [3–5]. Thus, desmopressin as well as antifibrinolytics are widely used in milder forms of VWD whereas factor VIII concentrates containing VWF are used, mainly as on demand treatment, in patients not responsive to desmopressin, i.e. those with type 3 and many who have type 2 VWD.
von Willebrand factor (VWF) pathology in patients with VWD involves either a reduction in the amount of functionally normal VWF (type 1) or the production of dysfunctional protein that fails to participate effectively in either of the two physiological roles of VWF, mediating platelet adhesion and acting as a carrier protein for factor VIII in plasma (type 2). Type 3 is defined by absence of VWF and will be further discussed below.
Type 1 von Willebrand disease – advances in molecular genetics: disease or bleeding risk? David Lillicrap
The phenotypic diagnosis of type 1 von Willebrand disease
Type 1 VWD represents the most common form of this disease, accounting for approximately 80% of cases in most surveys. The diagnosis of type 1 VWD requires consideration of three factors: documentation of a clinical history of excessive mucocutaneous bleeding, a series of coagulation test results consistent with a reduction of normal VWF levels and finally a family history of the condition. Despite what appear to be clearly defined diagnostic criteria for this condition, definitive type 1 VWD characterization can be extremely challenging and all three diagnostic components can be confusing.
While mucocutaneous bleeding symptoms are often clear in reproductive aged women, in males and young children, in whom haemostatic challenges may not have occurred, false negative bleeding histories may be obtained. Second, while family histories of excessive bleeding may be found in some patients, the incomplete penetrance and variable expressivity of type 1 VWD often complicates the confirmation of a family history. Finally, the acute phase nature of VWF and various environmental influences that contribute to VWF levels often challenge the straightforward interpretation of quantitative VWF assays. Overall, even after several temporally distinct sets of VWF testing, the diagnosis of type 1 VWD may still be unresolved.
The genetics of type 1 von Willebrand disease: introduction
The gene that encodes VWF is located on the short arm of chromosome 12 and spans 172 kb of genomic DNA. The gene has 52 exons, 12 of which (23–34) are duplicated with 3% variation in a partial VWF pseudogene on chromosome 22. While several positive and negative transcriptional regulatory elements have been characterized in the 5′ flanking region of the gene, the mechanisms regulating expression of VWF in vivo remain poorly defined.
Pattern of genetic transmission of type 1 von Willebrand disease
From the evaluation of many kindred over several decades, type 1 VWD has been classified as a dominantly transmitted trait. However, it is also clear that the phenotype shows significant variability with both incomplete penetrance and variable expressivity having been well documented.
Recent studies of type 1 von Willebrand disease molecular pathology
Despite being the most prevalent form of VWD, until very recently, little was known about the molecular genetic pathology responsible for this quantitative trait. Now, studies in approximately 300 type 1 VWD families from nine European countries and Canada have begun to generate the first systematic knowledge relating to this subject.
As discussed above, this condition is highly variable at the phenotypic level and initial genetic linkage studies in both the European and Canadian study cohorts have indicated that there is also very likely to be significant genetic locus heterogeneity underlying the condition [6,7]. In these studies, in which the family structures were variably informative for linkage analysis, between 40–60% of families showed no evidence of linkage to the VWF gene, thus implying that mutations in other genes involved in VWF biosynthesis, metabolism and clearance are likely to be responsible for some cases of this quantitative trait.
Subsequent to these linkage results, mutational analysis in these same study populations has further emphasized the genetic heterogeneity underlying type 1 VWD. In three distinct patient cohorts in nine European countries, the UK and Canada, candidate VWF gene mutations have been found in between 59 and 70% of index cases with the disease [8–10]. All of these studies had examined the VWF coding region, splice sites and proximal promoter sequence. Once again, these results imply that mutations at genetic loci other than the VWF gene may well be playing a causative role in a substantial number of type 1 VWD cases, although mutations embedded in the VWF introns, or affecting distant transcriptional regulatory elements remain possible pathogenetic explanations that require further evaluation.
von Willebrand factor gene mutations in type 1 von Willebrand disease
The types of VWF gene mutations described to date in patients with type 1 VWD represent a heterogeneous array of changes that span the entire length of the gene, from the promoter to the C-terminal domain of the protein. Missense mutations predominate (60% of mutations) and approximately 10% of patients each have putative transcriptional mutations, small VWF gene insertions or deletions or splicing mutations. In many instances, the definitive proof that the documented missense substitutions and promoter variations are pathogenic remains to be established. In significant numbers of patients, more than a single candidate sequence variation has been documented. Finally, some mutations, such as the Y1584C variant have been found at relatively high frequency (10–20%) in all three study populations. However, even with this most common variant, the pathogenetic mechanisms operating are still debated [11–13].
Summary of current status and future genetic studies of type 1 von Willebrand disease
Type 1 VWD represents an excellent model of a complex quantitative trait. Preliminary data confirms that the phenotypic heterogeneity, incomplete penetrance and variable expressivity seen in this condition are associated with significant allelic and genetic locus heterogeneity. There is growing evidence that more severe cases of type 1 VWD (VWF levels <25%) are caused by highly heritable mutations in the VWF gene. In contrast, more moderate phenotypes are likely to show incomplete penetrance and to be due to a combination of effects from mutations in the VWF gene, from genetic influences derived from known genetic modifiers such as the ABO blood group locus and to contributions from additional genetic modifiers yet to be defined. The magnitude of these effects and the numbers of influential loci may make subsequent analysis challenging.
Despite the fact that significant progress has been made in the past 5 years in terms of an initial genetic characterization of type 1 VWD, these early studies have emphasized that much remains to be learnt about the pathogenetic basis of this condition. In the meantime, the extreme mutational heterogeneity associated with the disease and relative lack of knowledge concerning pathogenetic mechanisms precludes the use of molecular genetic studies as a diagnostic aid.
Diagnosis and management of type 2N VWD Jenny Goudemand
VWF has two main functions: (i) it promotes the adhesion of platelets to sub endothelium at the site of small vessel injury and (ii) it protects FVIII from inactivation and rapid catabolism by forming a non-covalent complex with FVIII. VWD type 2 includes qualitative defects of VWF that impair platelet adhesion of FVIII binding. In type 2N VWD variants, the affinity of VWF to FVIII is markedly decreased resulting in decreased levels of FVIII in the circulation without disturbing primary haemostasis. This subtype was first characterized in France more than 15 years ago in patients with normal bleeding time (BT), normal or subnormal VWF levels and reduced FVIII levels [14,15]. It was called type 2N (‘N’ for Normandy) according to the birthplace of the first patient described [16].
This recessively inherited subtype is caused by mutations in the D′ to D3 domains which modulate the VWF binding site to FVIII. Patients can be homozygous for FVIII binding missense mutations, or are compound heterozygous either for two different FVIII binding mutations or for a FVIII binding mutation plus a VWF null mutation. About 20 different type 2N mutations have been described. Most are located in exons 18–20 affecting the FVIII domain. A few mutations have been identified in exons 21–27 of the VWF gene affecting amino acids located outside the FVIII binding site. The R854Q mutation is the most frequently reported mutation in type 2N VWD. In the French INSERM network study, 47% of unrelated patients with type 2N VWD are homozygous for this mutation and 43% are compound heterozygous with this mutation on one allele [17]. The R854Q mutation was found in 0.92% of the normal population of the Netherlands [18] and 0.31% in Euro-Brazilians [19].
Type 2N VWD is suspected when the FVIII:C level is disproportionately decreased compared with levels of VWF:Ag and VWF:RCo. FVIII:C usually ranges from 5 to 40 IU dL−1 although some have more severe FVIII:C deficiency (1–2 IU dL−1) [20]. The FVIII:C levels correlate with the mutation. The most common mutation (R854Q) is associated with moderate decrease in FVIII:C (around 20 IU dL−1) whereas lower FVIII:C levels (1–10 IU dL−1) are observed with other mutations at codons 791 or 816 [21]. The most severe defect (1–2 IU dL−1) is observed with the E787K mutation which markedly decreases the VWF binding capacity to FVIII [20]. Usually plasma VWF:Ag levels are normal or subnormal depending on the ABO blood group [22] and genotype of the patient (i.e. presence of a silent allele) [23]. As a consequence, the FVIII:C to VWF:Ag ratio is reduced (<0.5) in all the patients with type 2N VWD. Type 2N may be misdiagnosed as moderate and mild haemophilia A [24,25]. The diagnosis depends mainly on the measurement of the affinity of VWF to FVIII (VWF:FVIIIB) which is markedly decreased. The original assay is a solid phase immunoassay [15] but several modifications have enabled simplification and even automation of the assay [26–29]. In France, the test is now carried out using only commercially available reagents [30]. The bleeding time and VWF multimeric structure is usually normal although some mutations are associated with a decrease in the proportion of high molecular weight multimers or the presence of ultra large multimers [31–33].
The clinical expression depends on the FVIII:C levels. Bleeding may occur after trauma or surgery. Spontaneous bleeding manifestations are generally mild (ecchymosis, epistaxis, excessive bruising). Haemarthrosis and gastrointestinal bleeding are rare and observed mainly in patients with low FVIII:C levels (<5 IU dL−1) [14,24]. Rare patients with a low VWF:Ag level or abnormal VWF multimeric pattern may exhibit additional mucous membrane bleeding characteristic of primary haemostasis defects. Uneventful pregnancies and delivery were reported in patients with the R854Q mutation [34,35] and in a patient with the Y795C mutation in whom the FVIII:C level rose from 13 IU dL−1 to only 22 IU dL−1 at the 22nd gestational week [32]. However, there is a bleeding risk in those patients in whom the FVIII:C levels remain low at the end of gestation [36].
Treatment
Desmopressin is the first choice treatment provided it produces and maintains adequate haemostatic FVIII:C levels. However, response to desmopressin is highly variable depending on the mutation. Patients with the R854Q mutation have a sustained and therapeutically useful FVIII increase after desmopressin whereas patients with other mutations (e.g. R816W, C1060R) have only a short and insufficient increase in FVIII [3,37]. It is thus especially important to perform a desmopressin trial for these patients.
When desmopressin is insufficient or contraindicated, concentrates containing VWF are used, but not those containing FVIII alone. A patient with a low FVIII:C level (6 IU dL−1) who received a single infusion of an immunopurified FVIII concentrate (with no VWF) demonstrated a very short FVIII:C half-life (approximately 2.5 h), much shorter than that measured in haemophilia A patients treated with the same product. By contrast, another patient from the same family received a single injection of a VWF concentrate almost devoid of FVIII [24] and displayed a delayed but sustained rise in FVIII:C characteristic of the FVIII:C response observed in patients with severe VWD [38]. Although there are some anecdotal reports in the literature of patients with type 2N VWD treated with recombinant FVIII [39], replacement therapy with VWF is the only way to efficiently stabilize the FVIII:C activity because of the absence of endogenous VWF.
In conclusion, the possibility of type 2N VWD should be considered in patients with a reduced FVIII level and a reduced FVIII/VWF:Ag ratio when there is no clear historical familial data suggesting the presence of X-linked haemophilia A. The VWF:FVIIIB assay is the quickest way to distinguish between mild or moderate haemophilia A and type 2N VWD. A correct diagnosis is important for genetic counselling and clinical management of the patients as (i) desmopressin may have highly variable effects depending on the causal mutations and (ii) VWF concentrates represent the correct replacement therapy when desmopressin is inefficient or contraindicated.
von Willebrand disease type 3 – contemporary treatment dilemmasErik Berntorp
The primary symptom of VWD is mucosal bleeding, but haemophilia-like joint bleeds may also occur in severe VWD. Chronic morbidity is common where joint disease and/or frequent mucosal bleeds impact quality of life [40]. The rationale for long-term prophylaxis in patients that bleed frequently is obvious, and in Sweden, a cohort of patients (n = 37) has been successfully treated with prophylaxis for a median of 11 years (range, 2–45 years) [41]. The majority have type 3 VWD. Indications for prophylaxis included joint bleeds, bleeds from nose and mouth and gastrointestinal tract, and menorrhagia. The mean treatment dose is 24 units factor VIII per kg body weight given one to three times weekly. Prophylaxis has dramatically reduced the annual number of bleeds. Notably, subjects who began prophylaxis at an early age because of mucosal bleeds never developed joint problems [42].
A VWD Prophylaxis Network (vWD PN) has recently been established [43]. This network is an international study group which will examine the role of prophylaxis in clinically severe VWD. A feasibility survey was performed in 74 centers in Europe and North America. Of 6208 patients, 102 (74.5% type 3, 17.6% type 2 and 7.8% type 1) were using prophylaxis. Patients with type 3 were significantly more likely to be treated with prophylaxis in Europe than in North America. The most common triggers for prophylaxis were joint bleeds (40%), epistaxis/oral (23%) and gastrointestinal (GI) bleeding (14%), and menorrhagia (5%). A VWD International Prophylaxis Study is underway, with the goal of establishing optimal treatment regimens for the most common bleeding indications through prospective and retrospective data collection. Health economics and quality-of-life assessment will also be examined. In the prospective study patients with
- 1
severe type 1 VWD (i.e. ≤20% VWF:RCo and/or ≤20% FVIII and non-responsive to desmopressin);
- 2
type 2, desmopressin non-responsive or type 2B; and
- 3
people with type 3, will be enrolled if they meet defined criteria for frequency or severity of joint bleeding, epistaxis, GI bleeding, or menorrhagia.
Participants will undergo an escalation from one to three dose levels of VWD product, based on responsiveness to treatment, similar to the design of the Canadian prophylaxis trial in hemophilia [44] and starting with 50 U of VWF:RCo kg−1 once weekly, escalating to twice weekly and three times per week. The retrospective studies include examination of the effect of prophylaxis on bleeding frequency and the natural history of GI bleeding.
We believe that this comprehensive study, as well as other planned initiatives, will identify and implement optimal prophylaxis regimens for VWD, especially among those with type 3, who are the most severely affected by the disease.
