Current controversies in the diagnosis and management of von Willebrand disease

Abstract

Von Willebrand disease (VWD) is the most common inherited bleeding disorder in the world. The spectrum of VWD spans quantitative and qualitative deficiencies of von Willebrand factor (VWF), a platelet adhesive protein. It manifests primarily as mucocutaneous bleeding, but severely affected patients may suffer soft tissue bleeding and hemarthroses. There is disagreement in the multiple guidelines published regarding diagnosis, especially of type 1 VWD, which also remains the most opaque with respect to molecular characterization. Treatment with desmopressin (DDAVP) is most effective in type 1 VWD, but regimens are not standardized. It is not clear which type 2 VWD patients with qualitative deficiencies can be treated with DDAVP and which ones should receive VWF concentrates. No guidelines stipulate which patients might benefit from prophylactic VWF infusions and how they should be dosed. These are some current controversies in VWD that are discussed in this review.

Introduction

Von Willebrand disease (VWD) is the most common inherited bleeding disorder. It is an autosomal disease of quantitative (type 1 or 3) or qualitative (type 2 variants) abnormalities of von Willebrand factor (VWF.) It is manifest primarily as mucocutaneous bleeding or hemorrhage after invasive procedures, and less commonly by soft tissue and joint bleeding in the more severe types [Lillicrap, 2013].

There are clinical guidelines published to facilitate the diagnosis and management of the various types of VWD [Nichols et al. 2008; Castaman et al. 2013; Laffan et al. 2014]. But despite these guidelines, there remain controversies in diagnosis and in treatment. Diagnosing type 1 VWD, especially mildly affected persons, still produces spirited debate [Quiroga et al. 2014; Rodeghiero, 2014]. Is there any consensus on the diagnosis of mild type 1 VWD? Is having a low VWF equivalent to having a bleeding disorder? Common management of VWD depends upon increasing the circulating concentration of functional VWF, either endogenous or exogenous. Desmopressin acetate (DDAVP) increases endogenous levels but can have important side effects, some of which are dependent upon the dosing regimen [Svensson et al. 2014]. Is there an optimal DDAVP dose for each patient in each situation? Which patients with type 2 VWD will demonstrate a rise in VWF function after DDAVP? Plasma-derived VWF concentrates will raise functional VWF levels. Patients with severe disease may benefit from prophylactic treatment with concentrates yet only a small percentage of severely affected patients are on this regimen. How should one select patients for this intensive therapy? These are a few of the current controversies in von Willebrand disease.

Diagnosis of VWD

The diagnosis of VWD is based upon the patient’s bleeding history since childhood, a positive family history of bleeding, and a reduced VWF activity level such as the VWF:ristocetin cofactor (VWF:RCo.) Questionnaires such as the International Society on Thrombosis and Haemostasis-Bleeding Assessment Tool (ISTH-BAT) allow one to obtain a standardized history of bleeding events [Rodeghiero et al. 2010]. Scores >4 for men, >6 for women and >3 for children suggest that the patient’s bleeding justifies measurement of VWF levels [Elbatarny et al. 2014].

Once personal and family bleeding histories have been established, laboratory tests allow one to classify further the type of VWD. Types 1 (partial deficiency; autosomal dominant) and 3 (complete deficiency; autosomal recessive) are quantitative deficiencies of VWF, with type 1 having normal function of a decreased amount of protein, VWF antigen (VWF:Ag) that is present. In contrast, type 2 disease consists of malfunctioning VWF further divided into 4 subtypes, 3 of which are autosomal dominant and characterized by low activity to antigen ratios <0.6 for VWF:RCo to VWF:Ag [Nichols et al. 2009]. Type 2A encompasses loss of VWF high molecular weight multimers (HMWM) and decreased binding to the platelet receptor GPIb. It typically has low activity as measured by a VWF collagen binding (VWF:CB) to VWF:Ag ratio <0.6. Type 2B describes a gain-of-function mutation increasing the binding of VWF to GPIb and subsequent loss of the HMWMs. This increased binding and clearance can also cause thrombocytopenia. Type 2N, inherited as an autosomal recessive trait, produces only a decrease in the activity level of factor VIII (FVIII:C) with typically normal VWF:Ag and VWF:RCo, and normal VWF:RCo to VWF:Ag ratio. It is distinguishable from mild FVIII deficiency (hemophilia A) by a decrease in the VWF binding for FVIII and gene studies to locate the mutation to the VWF gene instead of the FVIII gene. Type 2M typically has normal VWF:Ag and low VWF:RCo, but continued presence of HMWMs on multimer analysis. The VWF:CB to VWF:Ag ratio is more often normal in type 2M, but some subtypes have very low collagen binding and thus the ratio is not a reliable discriminator between type 2A and type 2M [Nichols et al. 2009].

How do I classify the patient with borderline VWF levels?

Patients with a personal and family bleeding history and VWF levels below 30 IU/dl fulfill any guidelines’ criteria for a diagnosis of VWD. However, how to label the patient with VWF:RCo 30–50 IU/dl is much less clear. By the definition of 2 standard deviations below the mean value being abnormal, 2.5% of the population will have levels below 50 IU/dl and yet the vast majority of these people will not have any abnormal bleeding and hence the important recommendation of not ordering these tests on asymptomatic patients. VWF levels between 30–50 IU/dl correlate poorly with VWF gene defects. In type 1 VWD patients, a mutation in the VWF gene is found in up to 74% [Flood, 2014]. Of the type 1 VWD patients in whom a mutation is found, most have a missense mutation with a dominant negative effect that increases clearance or decreases the production of VWF, but deletions, frameshift, non-sense and other VWF null mutations are found rarely [Collins et al. 2008]. Other genetic mutations outside of the VWF gene may influence VWF levels. The ABO blood type of the patient influences VWF:Ag levels, with type O patients having approximately 30% lower VWF that any other blood types [Gill et al. 1987]. However, when family members are tested, only 28% of the VWF variance is inherited and of that amount, the ABO gene accounts for 30–40% [Souto et al. 2003; Vossen et al. 2004]. Correlation of bleeding symptoms with VWF:RCo becomes less robust with levels 30–50 IU/dl [Goodeve et al. 2007].

A recent study laid bare the discrepancies in numbers of patients assigned a diagnosis of type 1 VWD depending upon whose criteria for diagnosis was employed. Of 4298 patient samples tested with four VWD parameters, type 1 VWD was diagnosed in 2.8% using the most strict criteria and up to 8.3% with the most liberal criteria – a nearly three-fold difference [Quiroga et al. 2014]. An Italian group examined 93 patients who were referred for evaluation due to bleeding or abnormal coagulation tests and had VWF levels of 30–60 IU/dl. When these patients were evaluated with second level tests including multimer analysis, VWF content in platelets, ristocetin-induced platelet aggregation, and a collagen-binding assay to confirm the presence of a VWD diagnosis, the main predictors of the positive results confirmed by these tests in 48% of the original 93 were initial VWF:RCo level, female sex and non-O blood type. In this small sample, all patients with VWF:RCo <40 IU/dl confirmed positive on further testing [Bucciarelli et al. 2015]. There is a lack of consensus on the criteria for type 1 VWD. See Table 1 for comparison of different guidelines.

Table 1.

Diagnostic criteria for the diagnosis of type 1 von Willebrand disease.

Diagnosis	NHLBI criteria [Nichols et al. 2008]	ISTH-SSC [Sadler and Rodgehiero, 2005]	European group on VWD [Castaman et al. 2013]	Zimmerman program for molecular and clinical biology of VWD [Montgomery et al. 2013]
Type 1 VWD	VWF:RCo <30 IU/dl and/or VWF:Ag <30 IU/dl	VWF:RCo <2 SDs below the population and ABO-type mean and VWF:Ag <2 SDs below the population and ABO-type mean On >2 determinations	VWF:RCo <40 IU/dl Or VWF:CB <40 IU/dl	VWF:RCo <40 IU/dl or VWF:Ag <40 IU/dl
‘low VWF’	VWF:RCo or VWF:Ag 30–50 IU/dl	No difference in lab tests; only patients without personal bleeding history	No such defined category	VWF:RCo 40 IU/dl to lower end of normal range

ISTH-SSC, International Society on Thrombosis and Haemostatis Scientific and Standardization Committee; IU, international unit; NHLBI, National Heart, Lung and Blood Institute; VWD, von Willebrand disease; VWF, von Willebrand factor; VWF:Ag, VWF antigen; VWF:RCo, VWF:ristocetin cofactor.

What is the difference between type 1 VWD and ‘low VWF’?

If 2.5% of the population by definition (2 standard deviations below the mean) has <50% VWF activity and surveyed patients will commonly report that they have minor bleeding symptoms [Nichols et al. 2008], there will be overlap of those two populations merely by chance alone. We also know from family studies that less than a third of variations in VWF are found to be heritable [Souto et al. 2003]. Further complicating the family history is incomplete penetrance and variable expressivity in VWD. Levels of VWF between 30–50 IU/dl have little bleeding predictive value in previously asymptomatic patients. That is the main reason asymptomatic patients should not be tested for VWF. Past bleeding history is a much better predictor of future bleeding [Tosetto et al. 2006]. A personal bleeding history, as well as a family history of abnormal hemorrhage, coupled with significantly low VWF levels are the crucial elements to diagnose someone as having true disease.

Because of the difficulty in diagnosis of type 1 VWD, some have proposed the concept of ‘low VWF’ for those persons with VWF:RCo between 30 and 50 IU/dl but no significant personal or family history of bleeding [Sadler, 2009]. This is seen as a ‘risk factor’ for bleeding in a manner similar to hypertension and hypercholesterolemia being risk factors for heart disease. What is not clear is what one is supposed to do with this ‘risk factor’. Do you mitigate it with interventions to raise VWF such as one would treat hypertension to lower the risk of heart disease? If so, how is that different from the treatment/prophylaxis one would employ for diagnosed type 1 VWD? The concept remains controversial.

Treatment of VWD

Desmopressin

Desmopressin (1-desamino-8-D-arginine vasopressin) is an analogue of the human antidiuretic hormone (ADH). It has been used as a treatment for VWD since the 1970s [Mannucci et al. 1977]. DDAVP binds and activates vasopressin V2 receptors, which leads to VWF and factor VIII (FVIII) secretion from Weibel-Palade bodies inside endothelial cells into the circulation [Kaufmann et al. 2000]. Definitions of adequate response vary and will depend in part upon the severity of the hemostatic challenge. At a minimum, the VWF:RCo and FVIII:C levels should increase at least 2–3 fold and to levels >30% and preferably >50% to promote hemostasis for invasive procedures. Confirmation of this response is crucial and obtained by performing a trial with a standard DDAVP dose administered intravenously. Baseline, 1 and 4 hours post-infusion levels of VWF:Ag, VWF:RCo, and FVIII:C are drawn to document the patient’s responsiveness. Possible immediate complications of DDAVP include flushing, hypotension/hypertension, gastrointestinal upset and headache. Because of its ADH-mimicking properties, DDAVP can cause hyponatremia which can lead to seizures. Limiting free water intake will help prevent symptomatic low sodium levels. Tachyphylaxis can appear if doses are less than 24 hours apart [Vicente et al. 1993].

What is the best dosing regimen of DDAVP to minimize side effects and maximize response?

Standard dosing recommendations for DDAVP for years have been 0.3 µg/kg intravenously or subcutaneously outside of the USA [Nichols et al. 2008]. Informal surveys of practices at different treatment centers yield regimen descriptions ranging from multiple sequential doses at 12 hourly intervals to no more than 2 total doses at least 24 hours apart. The controversies are over whether or not significant tachyphylaxis occurs with repeated dosing at frequent intervals and if the likelihood of complications increases with higher doses. Is there a maximum dose one should never exceed? As populations become increasingly obese, is the weight-based regimen safe or even necessary to obtain the desired rise in VWF:RCo and FVIII:C?

Tachyphylaxis with DDAVP infusions in patients with VWD was reported in a 1992 study, noting a 30% drop in response after the second dose but not a further drop after the third and fourth doses given at 24 hour intervals [Mannucci et al. 1992]. Even in normal persons without VWD, the response drops dramatically with DDAVP 0.3 µg/kg intravenously given at 12 hourly intervals. After an initial doubling of mean VWF:Ag from 104% to 218%, further dosing of DDAVP with 12 hours between doses yielded only a 12% increase at either 12 or 24 hours after the initial dose [Vicente et al. 1993]. Dosing at intervals less than 24 hours would seem to have all the risks with little incremental benefit.

Of the potential complications of desmopressin, hyponatremia is the most concerning. In a 2014 retrospective review of children receiving DDAVP 0.3 µg/kg IV to promote hemostasis in various disorders (VWD and/or qualitative platelet defects), serum sodium (Na) levels fell below 135 mEq/l in 72% of patients at 24 hours after dosing and 10.3% had levels below 130 mEq/l. Of the 107 patients, 12.1% had their second DDAVP dose withheld due to hyponatremia. All of these patients were fluid restricted in a hospital. After a single dose of DDAVP, one child developed generalized seizures associated with a Na level of 124 mEq/l. This prompted the authors to consider inpatient observation for 20–24 hour, monitoring Na levels, and restricting free water in all patients receiving this medication for hemostasis [Sharma and Stein, 2014]. To decrease the risk of clinically significant hyponatremia, most centers recommend fluid restrictions based upon the patient’s weight [Neff and Sidinio, 2014]. This is particularly pertinent to perioperative hyponatremia according to a study of children undergoing adenotonsillectomy where degree of hyponatremia was related to the volume of perioperative fluid resuscitation [Davidson et al. 2011].

Desmopressin doses are weight-based but many treatment centers cap the maximum dose they will give to a patient at 20–25 µg. A Canadian treatment center began using only 15 µg DDAVP which was a standard vial size available to them. A retrospective report of the responses of VWD patients to this fixed dose administered subcutaneously showed they were comparable with previously published VWF levels obtained with the standard weight-based algorithm. The average patient weight was 69 kg (range 50–165 kg). Complete response (VWF:RCo >50%) at 1 hour post dose was obtained in 82.5% of type 1 VWD, partial response in 12.5% and no response in 5%. The median VWF:RCo level post-DDAVP was 91 IU/dl (range 15–223) [Siew et al. 2014]. The study did not measure Na levels nor did it report on efficacy, similar to other studies that have reported only laboratory results.

In summary, DDAVP at 24 hour intervals may reduce the risk of side effects without a significant compromise in VWF activity/FVIII:C levels. Most treatment centers will not prescribe more than three doses at this frequency. Standard intravenous and subcutaneous DDAVP doses are 0.3 µg/kg, but a fixed dose of 15 µg for patients over 50 kg may provide acceptable hemostatic levels of VWF:RCo and FVIII:C with less risk of serious side effects. Trials documenting its efficacy in a clinical setting are necessary.

Can I treat type 2 VWD with DDAVP?

DDAVP is the treatment of choice for most patients with type 1 VWD where the defect is quantitative, but is it effective for type 2 VWD patients who have functional defects in their VWF? Surprisingly little published data exist on this topic. A multicenter European study to examine the laboratory response of severe type 2 VWD patients to DDAVP showed mixed results. To be included in this study patients had to have at least one of the following: bleeding time >15 min, FVIII:C <20%, or VWF:RCo <10%. There were 40 type 2 VWD patients included, with only 18% responding as defined by an increase in VWF:RCo and FVIII:C by 3-fold or more plus achieving at least 30% activity and a bleeding time less than 12 min. By subtype the responses were: type 2A 1/15 (7%); type 2M 3/21 (14%); and type 2N 3/4 (75%) [Federici et al. 2004]. Genotypes were more helpful that phenotypes in types 2A and 2N at predicting response to DDAVP. The investigators noted that the trial was not designed to examine efficacy with a clinical bleeding challenge, only laboratory response [Federici et al. 2004].

DDAVP response specifically in children has been reported by Schneppenheim and colleagues in 28 children with type 2 VWD. They documented complete responses in 61%, partial responses in 7%, and no response was found in 32%. They also reported only laboratory test changes, not assessments of clinical effectiveness [Schneppenheim et al. 2009]. Revel-Vilk and colleagues tested five children with type 2A and found only one responder [Revel-Vilk et al. 2003].

When adequate VWF:RCo response is proved with a DDAVP trial, it is generally assumed that it can be used for the prevention and treatment of most bleeding episodes despite the lack of comparative trials that document this effectiveness. There are several caveats to that assumption. Although some responses have been seen in type 2 patients other than 2A or 2M, they are usually either too small or not sufficiently sustained to be effective. With type 2N, one must be aware that without normal VWF to stabilize the FVIII, FVIII:C may wane rapidly and thus not prove hemostatic despite an initial robust response [Mazurier et al. 2001]. DDAVP is usually considered contraindicated in type 2B patients as it may promote thrombocytopenia. The intranasal product, Stimate®, warns in the package insert [Behring, 2010] against its use in patients with type 2B VWD fearing induction of platelet aggregation. Despite this caution, DDAVP has been used successfully in some type 2B patients, specifically those with the p.Arg1315Cys mutation that does not tend to develop severe thrombocytopenia [Schneppenheim, 2011]. DDAVP is not recommended for patients under age 2 regardless of type due to poor response [Revel-Vilk et al. 2003].

In summary, DDAVP may be used to treat or prophylax against bleeding in a minority of patients with type 2A and 2M, and perhaps some 2N patients in certain situations, if the patient has been shown to respond adequately in a DDAVP trial. Despite the rare reports of efficacy, most treatment centers will not use DDAVP in type 2B patients.

VWF concentrates

In VWD patients with severe quantitative or qualitative deficiency that exhibit inadequate response to DDAVP and for type 3 patients, who by definition, have no VWF to mobilize, infusion with VWF concentrates is the mainstay of treatment. All currently licensed VWF concentrates in all countries are plasma-derived, virally inactivated formulations with varying ratios of VWF to FVIII ranging from approximately 1:1 to 2.4:1 [Neff and Sidinio, 2014]. VWF concentrates are used for treatment of active bleeding or the prophylaxis of bleeding with invasive procedures. There is no current recommendation of any VWF product for routine prophylaxis against recurrent hemorrhage such as is common in hemophilia.

Should I use VWF concentrate prophylaxis for severe VWD patients?

Patients who do not respond adequately to desmopressin or are scheduled for major surgery are given VWF clotting factor concentrates. All patients with type 3 disease and most with type 2 and severe type 1 disease require VWF concentrates. These products are plasma-derived human factors containing VWF and FVIII with an excellent safety record utilizing purification technology. In the United States these products are dosed in RCo units and in Europe by FVIII:C units. Both VWF:RCo and FVIII:C levels are expected to rise with infusions. With repeated dosing, the FVIII:C levels may rise higher than desired due to stabilization of endogenous FVIII by the infused VWF. In type 3 patients the immediate availability of FVIII is necessary for treatment of acute hemorrhage or emergency surgery as FVIII:C levels will not rise for 6–8 hours if pure VWF is infused and one must await native FVIII binding to infused VWF.

With the success of prophylactic factor infusions for severe hemophilia A, some centers have prescribed prophylactic VWF concentrates to those patients with severe clinical forms of VWD. The Medical and Scientific Advisory Committee (MASAC) of the National Hemophilia Foundation recommends routine prophylaxis in young boys with severe hemophilia A or B [MASAC, 2007], but is silent on whether or not patients with severe types of VWD should be treated in a similar fashion [MASAC, 2010]. In a survey of type 3 VWD treatment at US centers, only 13.9% of these patients were on prophylaxis despite 73% of center staff stating that prophylaxis was very or extremely helpful [Sumner and Williams, 2004].

In 2006, published results of prophylaxis in a Swedish cohort of 37 patients, 28 of whom were type 3, revealed its success at reducing the occurrence of epistaxis and gastrointestinal bleeding and at preventing joint disease if treatment was begun at an early age [Berntorp, 2006]. The von Willebrand Disease Prophylaxis Network (VWD PN) reported data from 20 centers in 10 countries on 61 subjects. Reasons for prophylaxis need were recurrent epistaxis and joint bleeding seen more often in younger patients, and menorrhagia or gastrointestinal bleeding, seen more commonly in older patients. The VWD PN report was comprised of 34 patients with type 3, 5 patients with type 1, 10 patients with type 2A, 8 patients with type 2B, and 2 patients with type 2M. Infusions of 40–60 RCo IU/kg VWF concentrates 1–3 times weekly reduced joint bleeding by nearly 90%. It also reduced the number of episodes of epistaxis and the intensity of menorrhagia. In the older populations, it reduced gastrointestinal bleeding which required higher factor doses closer to 60 RCo IU/kg of VWF concentrate for improvement. Antibodies to VWF were reported, but there were no thrombotic complications with regular VWF infusions [Abshire et al. 2013]. A recent update of this study reported continued success with a total of 105 patients. A total of 10 of these patients are enrolled in a prospective study to help ascertain the most appropriate dosing regimen using a dose-escalation strategy [Holm et al. 2015].

Italian centers reported their experience with long-term prophylaxis in 12% of factor concentrate - using VWD patients who had received 40 FVIII IU/kg of high or intermediate purity VWF concentrates 2–3 times weekly for recurrent bleeds. Prophylaxis stopped bleeding events in eight of the 11 patients treated and decreased hospitalizations or the number of red cell transfusions in the others [Federici et al. 2005]. A standardized bleeding score higher than 10 may be a means of identifying patients with the highest risk of spontaneous hemorrhage that would benefit from a prophylaxis regimen [Federici et al. 2014]. A German cohort reported in 2011 included 32 VWD patients with a different mix of subtypes: four type 1; 15 type 2; and 13 type 3 patients. With 40 RCo IU/kg 2–4 times weekly, they had dramatic drops in their bleeding scores and hemoglobin levels rose to the normal range. An inhibitor developed in one type 3 patient for a prevalence of 3.1% in that population studied [Halimeh et al. 2011]. Inhibitors in VWD are not common with reported prevalence rates of 5.8–9.5%, so it is not unexpected that a study of that size would have one develop. They typically occur in type 3 patients [James et al. 2013]. It remains to be seen whether or not the new recombinant human VWF will prove as efficacious with prophylaxis without FVIII. Initial safety and tolerability studies are encouraging [Mannucci et al. 2013].

In summary, it is clear that prophylactic infusions improve patient outcomes and quality of life for those most severely affected. Many questions remain to be answered. At what age should prophylaxis start? What are the patient selection criteria for those who will likely benefit from factor concentrate prophylaxis? What is the optimal dosing regimen? Further research is needed and a prospective trial is ongoing in the VWD PN [Holm et al. 2015].

Conclusion

We still lack the appropriate tools to determine reliably to what extent a patient is at risk for hemorrhage due to low VWF. If guidelines were harmonized with respect to diagnostic criteria, then research into treatments for each type might be more applicable to similar patients worldwide. DDAVP is the primary therapy for most patients with this disease and dosing regimens should be safe and effective. It may be possible to administer a fixed DDAVP dose to most adults every 24 hours to decrease risks yet still be effective. Studies are in progress to identify those VWD patients who are likely to suffer recurrent hemorrhage and thus be candidates for prophylaxis. Prevention of bleeding can decrease long-term physical damage and improve quality of life. These are all important controversial issues, the resolution of which can improve care for patients with VWD.