von Willebrand disease and heavy menstrual bleeding: when and how to test

Visual Abstract

Abstract

von Willebrand disease (VWD) is a common inherited bleeding disorder caused by von Willebrand factor (VWF) deficiency and is an important cause of heavy menstrual bleeding in young patients. A clinical evaluation using standardized bleeding scores helps determine when screening hemostatic testing is indicated by identifying patients with a moderate or high probability of an inherited bleeding disorder. The diagnosis of VWD is made when VWF levels are under 30 IU/dL or between 30 and 50 IU/dL when there is a positive bleeding history. Activity levels above 100 IU/dL have a high negative predictive value. Multiple factors, including stress from acute bleeding and anemia, pregnancy, and medications, can affect VWF levels, hence testing for VWD is best performed when a person's health is at its baseline level, although this is not always possible in clinical practice. Variation in assay methodologies measuring VWF activity can have a significant impact on the diagnostic evaluation, and it is important for clinicians to be familiar with the limitations of the assay used by their local or reference laboratory. Genetic testing can be useful in establishing the VWD subtype and providing accurate reproductive counseling but is not required to make a diagnosis.

Learning Objectives

• Apply the results of the clinical and laboratory evaluation to the diagnostic approach of patients with HMB in whom VWD is suspected

• Compare the differences in principle and methodology of laboratory assays used in the evaluation of patients with VWD

CLINICAL CASE

A 15-year-old patient (they/them) was diagnosed with iron-deficiency anemia after presenting with fatigue. They have had heavy menstrual bleeding (HMB) since menarche with a pictorial blood loss assessment chart score over 100, multiple bruises over 1 cm in exposed areas, and 2 to 3 nosebleeds a year (at least 1 requiring packing). There is no family history of excessive bleeding. An estrogen-containing combined oral contraceptive pill (OCP) and oral iron are started, and they are referred to hematology for evaluation for a possible bleeding disorder.

Introduction

von Willebrand disease (VWD) is a common bleeding disorder resulting from the quantitative or qualitative deficiency of von Willebrand factor (VWF). Affected patients present with mucocutaneous bleeding of variable severity, and HMB can be the first and only bleeding manifestation, especially in the adolescent population. Clinically significant bleeding is the common presentation in all inherited and acquired bleeding disorders, and in many cases, such as that of VWD, the specificity of the diagnosis is based on a unique laboratory phenotype that defines the presence of VWF deficiency and subclassifies it into different subtypes (Table 1).

Table 1.

VWD subtypes, pathophysiology, and inheritance pattern

Defect type	Subtype	Pathophysiology	Inheritance
Quantitative defect	1	Partial quantitative deficiencya,b	Autosomal dominant, variable penetrance
	3	Complete quantitative deficiency	Autosomal recessive
Qualitative defect (protein dysfunction)	2A	Decreased VWF-dependent adhesion due to loss of HMWM	Autosomal dominant
	2B	Increased affinity of VWF for platelet GPIb	Autosomal dominant
	2M	Decreased VWF-dependent adhesion to collagen or platelets without loss of HMWM	Autosomal dominant
	2N	Decreased binding of VWF to FVIII	Autosomal recessive

^a Includes subtype 1C characterized by increased clearance.

^b Compound heterozygosity type 1/2N (biallelic variants, 1 quantitative and 1 affecting FVIII on the other allele) can lead to low VWF antigen and activity (with a conserved ratio) and low FVIII that is out of proportion to the VWF antigen.

The clinical probability of VWD in patients presenting with bleeding

The success of the diagnostic evaluation of patients with a suspected bleeding disorder, especially VWD, is dependent on the clinical evaluation, through which the pretest probability of the clinical diagnosis is established. HMB can be the first and only manifestation of VWD in young patients, but in many, due to the limited knowledge of clinicians and the social, cultural, and economic aspects, VWD may go unrecognized and the diagnosis delayed more than a decade. Standardized bleeding scores, such as the International Society for Thrombosis and Haemostasis Bleeding Assessment Tool (ISTH-BAT), are the pillar of this clinical evaluation, and their use is recommended over nonstandardized clinical assessment.¹ Other tools specifically designed to evaluate the severity of HMB, such as the pictorial blood loss assessment chart, are available and should be used when HMB is part of the clinical presentation. Patients with self-reported heavy menses who do not meet the criteria for HMB using standardized scores and who have no other bleeding manifestations have a low probability of a bleeding disorder and, in general, should not be tested.

A positive bleeding score (ISTH-BAT ≥4 in adult males, ≥6 in adult females, and ≥3 in children) identifies patients with an intermediate or high probability of VWD (positive predictive value  =  0.2, negative predictive value  =  1).² It is common in clinical practice to encounter patients who have had VWF laboratory testing before a bleeding history has been obtained, and completing this is essential to interpreting the results and establishing the next steps.

Patients with a positive family history of VWD

Some individuals present for evaluation due to a family history of VWD. When a first-degree relative is affected, their pretest probability is not low, and additional testing is typically considered, even if their bleeding score is negative.¹ Due to the variable penetrance of VWD and the effect of other parameters on VWF levels, there can be carriers of disease-causing VWF alleles (in particular those associated with quantitative deficiency) who do not have bleeding and may have normal VWF levels.³

CLINICAL CASE (continued)

With an ISTH-BAT score of 7 (3 points for epistaxis requiring cautery, 1 point for cutaneous bleeding with 5 or more bruises ≥1  cm in exposed areas, and 3 points for HMB since menarche and lasting over 12 months), this patient has a moderate to high probability of having an inherited bleeding disorder and warrants additional laboratory testing. Even if their presentation was only HMB, inherited bleeding disorders are present in around 66% of young patients presenting with this symptom, with 26% of those having VWD.⁴ Therefore, additional laboratory testing would still be appropriate.

When and how to test

Testing in the setting of conditions that can increase VWF levels can lead to a missed diagnosis of VWD. Physical stress (from acute bleeding, anemia, acute illness, or a traumatic blood draw), psychological stress, and medications (estrogen-containing drugs) can increase VWF levels. Blood group O, medications (valproic acid and ciprofloxacin), and hypothyroidism can decrease VWF (Figure 1).⁵

Figure 1.

Factors that affect VWF plasma levels.

In adolescents presenting with acute HMB, the mean VWF antigen and activity measured at presentation was on average 45% higher compared to follow-up. The change was most significant in those with elevated factor VIII:c (FVIII:c), supporting its role as a marker of acute stress. While 70% of patients with VWD presenting with HMB will be identified at the time of initial testing,⁶ the diagnosis of VWD may be overlooked in a large proportion of patients, and retesting is important, especially in those with borderline results.

Estrogens increase several procoagulant proteins and decrease the activity of natural anticoagulants. These changes are most pronounced during pregnancy but can also be seen with the use of medications containing estrogen at high doses. In pregnancy beyond the first trimester, VWD plasma studies are typically used for delivery planning and management of bleeding risk, but for the purpose of confirming the diagnosis of VWD, baseline (nonpregnant) levels are required.¹ The currently available OCPs have estrogen concentrations of 30  µg or less, and at this level their effect on VWF levels is not significant, so it is not necessary or practical to stop OCPs for diagnostic purposes only.⁷ Changes in VWF levels related to the different phases of the menstrual cycle have been reported in some studies but likely result from other factors such as stress and not from hormonal changes.⁸

Screening laboratory evaluation and interpretation

The initial laboratory evaluation of patients with a suspected bleeding disorder includes screening coagulation studies: platelet count, prothrombin time, activated thromboplastin time, and fibrinogen activity or thrombin time, as well as VWF antigen, VWF activity, and FVIII activity (Table 2). The 2021 guidelines for the diagnosis of VWD recommend using newer assays that measure the platelet-binding activity of VWF (GP1bM, GP1bR) over the ristocetin cofactor (VWF:RCo).¹ However, in many centers in the United States and around the world, only the VWF:RCo is available.

Table 2.

Frequently used abbreviations in the context of testing for VWD

CLIA	Chemiluminescence immunoassay	VWF:Ag	von Willebrand factor antigen
FVIII:C	Coagulation factor VIII activity	VWF:CB	von Willebrand factor collagen binding
HMWM	High-molecular-weight multimers	VWF:FVIIIB	von Willebrand factor binding to FVIII
GPIb	Platelet glycoprotein Ib	VWF:RCo	von Willebrand factor activity using ristocetin as cofactor
RIPA	Ristocetin-induced platelet aggregation	VWF:GPIbM	von Willebrand factor activity by glycoprotein Ib binding using recombinant mutated GPIb
VWD	von Willebrand disease	VWF:GPIbR	von Willebrand factor activity by glycoprotein Ib binding using recombinant GPIb and ristocetin
VWF	von Willebrand factor	VWFpp	von Willebrand disease propeptide

The goal of the screening evaluation is to rule out other causes of bleeding and establish whether a VWF deficiency (qualitative or qualitative) could be responsible for the bleeding phenotype.⁹ Ruling out other causes is important in establishing the diagnosis of VWD in patients who do not have severely decreased VWF levels. Evaluation for iron deficiency (with or without anemia) should also be performed in all patients presenting with HMB.

Step 1: Is the VWF activity low?

A VWF antigen or activity level under 30 IU/dL can be diagnostic of VWD. The antigen or the activity level alone does not provide sufficient information to establish whether the VWF deficiency is quantitative or qualitative. The reliability of the activity measurement depends greatly on the methodology used. Biological differences can also lead to significant discrepancies between assays. The principles of the methodologies available to measure VWF activity (VWF:Ab, VWF:RCo, GP1bM, Gp1bR) and important clinical correlates are presented in Table 3. Clinicians should be familiar with the characteristics of the methodology available locally and know when to request testing with alternate methodology.

Table 3.

Testing principle, advantages, and disadvantages of commercially available assays to evaluate VWF activity

	Assay principle	Advantages	Disadvantages
VWF:Ab	Measures binding of a monoclonal antibody to the VWF A1 domain epitope	Not affected by VWF exon 28 benign variantsa	• Not a true functional assay • Misses the type 2M VWF c.3971G>C, p.Gly1324Ala variant
VWF:RCo	Measures the ability of VWF in patient plasma to agglutinate donor platelets in the presence of ristocetin	Broadly available	• Highest coefficient of variation • Limited lower level of quantificationc • Affected by VWF exon 28 benign variants • Can miss type 2B patients, especially those with the VWF c.3916C>T, p.Arg1306Trp variant
VWF:GPIbM	Measures the spontaneous binding of VWF to a gain-of-function mutant GPIba fragment	Not affected by VWF exon 28 benign variants Low variability	• Can miss some type 2B patients • Can overestimate activity in patients with very low VWF:Ag
VWF:GPIbR	Measures the ristocetin-induced binding of VWF to a recombinant wild-type GPIb fragment	Least variable and lowest level of detectionb	May be affected by VWF exon 28 benign variantsd

^a VWF exon 28 benign variants affecting ristocetin sensitivity include VWF c.4138A>G, p.Ile1380Val (I1380V); VWF c.4304A>G, p.Asn1435Ser (N1435S); and VWF c.4414G>C, p.Asp1472His (D1472H).

^b When performed by chemiluminescence immunoassay.

^c Lowest sensitivity activity assay. Does not allow for accurate subtype classification in 18% of patients.¹⁰

^d The effect of these variants has been hypothesized to be similar to VWF:RCo; however, a large study comparing activity assays did not confirm this finding.

The lower limit of detection of the different activity assays can affect the clinician's ability to distinguish between the two types of quantitative deficiencies (type 1 and 3) when the VWF:Ag is very low (typically under 5 IU/dL) as well as precisely calculate the ratio of activity to antigen.¹⁰

Variability in the operating characteristics of different activity assays (Table 3) leads to discrepant classification of the VWD subtype in up to 20% of patients.¹⁰ In terms of screening testing, this variability is important to consider in patients who have an activity result close to the 30 IU/dL or 50 IU/dL threshold, which can define whether they meet diagnostic criteria for VWD. Interference by antibodies (heterophilic, rheumatoid factor, or human antimouse antibodies) can also lead to variation in results and can be present in over 10% of people.¹¹

An activity between 30 and 50 IU/dL can still be consistent with VWD in the context of a qualitative defect (type 2), acquired VWD, or a quantitative defect providing there is a positive bleeding history, and alternate causes of bleeding have been ruled out. Patients with a mild quantitative deficiency but no clinically significant bleeding have decreased VWF, which is a risk factor for bleeding but not considered a bleeding disorder.¹²

Step 2: Is the VWF activity concordant or discordant with the VWF antigen?

A ratio of VWF activity to VWF antigen over 0.7 is consistent with a quantitative defect when the VWF activity is under 50 IU/dL. Concordant antigen and activity above 50 IU/dL are typically not consistent with VWD, providing the sample reflects a baseline level (untreated and in a period of good health). The correlation of the antigen and activity is important because in functional defects (type 2) the antigen can be normal.

A ratio of VWF activity to VWF antigen under 0.7 is characteristic of qualitative defects (type 2, except 2N), and follow-up studies with VWF collagen binding (VWF:CB), VWF multimer analysis, platelet count, and additional specific tests will help determine the exact subtype (Table 4). An abnormal ratio can also be seen in individuals with quantitative VWD or without VWD who carry benign VWF variants in exon 28 that affect sensitivity to ristocetin when ristocetin-dependent activity assays are used. These include the VWF c.4414G>C, Asp1472His (D1472H), which can be present in up to 40% of people and leads to a decrease in VWF:RCo by around 25%.¹³ In these cases, repeat testing to verify the ratio with a different activity assay, ideally one that is platelet dependent and does not use ristocetin, is recommended.

Table 4.

Assay principle and interpretation of available test results for the subclassification of VWD types

	Principle	Interpretation of results
VWF multimers	Two-dimensional immunoelectrophoresis followed by labeling and autoradiography	• All sizes present: normal, type 1, 2M, 2N • Absent: type 3 • Absence of large and intermediate forms: type 2A • Absence of large forms: 2B
Collagen binding (VWF:CB)	Functional assay measuring the ability of VWF to bind collagen I and III	• Decreased proportional to VWF:Ag in type 1 • Absent in type 3 • Low with decreased ratio to the VWF:Ag in type 2A and 2B • Normal or decreased proportional to VWF:Ag
Ristocetin-induced platelet aggregation	Measures the ability of patient platelets to agglutinate in the presence of multiple concentrations of ristocetin	• Present agglutination at low doses of ristocetin and increased at standard doses in type 2Ba
Factor VIII binding (VWF:FVIIIB)	Measures VWF activity to bind FVIII	•Low in type 2N

^a Indistinguishable from platelet-type VWD caused by gain-of-function variants in GPIb.

The ratio of VWF activity to VWF antigen cannot be calculated when the activity reported is under the lower limit of detection of the methodology used. The ratios are also less reliable in the context of very low activity and antigen since small variations in the absolute numbers can modify the ratio significantly and move it across the established thresholds.

Step 3: Is the FVIII activity concordant or discordant with the VWF antigen?

A low FVIII:C with a FVIII:c/VWF:Ag under 0.7 could indicate the presence of type 2N VWD (which is characterized by a normal VWF activity:antigen ratio), but an FVIII lower than the VWF antigen can also be seen in patients with mild hemophilia A. Studying VWF binding to FVIII (VWF:FVIIIB), or genetic testing, is useful in distinguishing these disorders (Table 4).¹

CLINICAL CASE (continued)

Screening laboratory tests were obtained during the visit with hematology and showed normal prothrombin time, activated thromboplastin time, and fibrinogen activity. The hemoglobin level was 9.2  g/dL with an mean corpuscular volume of 71 fL, and the platelet count was 160,000/uL. VWF antigen (VWF:Ag) was 61 IU/dL, VWF activity by ristocetin cofactor (VWF:RCo) was 35 IU/dL, and FVIII activity (FVIII:c) was 65 IU/dL.

The activity by ristocetin cofactor is low enough that in the context of the negative additional hemostatic screen it could be consistent with VWD. However, an exon 28 VWF polymorphism could be affecting the in vitro activity and the ratio of VWF: RCo/VWF:Ag. Does this patient have type 2 VWD? Quantitative VWD with VWF D1472H? A different bleeding disorder?

Obtaining additional specialized plasma testing

Additional specialized testing is used to further characterize the VWD subtype in a patient who already has a diagnosis of VWD. The subtype is important because it guides disease management and influences the reproductive risk for a specific individual (Table 4). The availability of these assays vary, but most of them are performed exclusively at reference laboratories (Figure 2).

Figure 2.

High-yield clinical principles for testing in patients with HMB in whom VWD is suspected.

CLINICAL CASE (continued)

The patient remained on OCPs and oral iron, and 3 months later, once the anemia had improved, they were retested with a non-ristocetin–based activity. Results showed a VWF activity of 52 IU/dL, GP1bM activity of 30 IU/dL, FVIII:c of 60%, and the absence of high-molecular-weight multimers, consistent with type 2A or 2B.

When to retest

Retesting can be considered in patients in whom the initial set of laboratory studies does not reflect their true baseline. An activity over 100 IU/mL (untreated) has a very high negative predictive value for VWD and typically does not require retesting.¹⁴ Patients initially tested during an acute bleeding episode, severe anemia, an acute illness, or pregnancy and who have activities over 50 IU/dL should be retested in a period of wellness.

It is also important to note that VWF rises with age. Up to one-third of patients who receive a diagnosis of VWD or low VWF in childhood normalize their VWF levels as adults. The likelihood of normalization is highest in patients with VWF activities between 30 and 50 IU/dL.¹⁵

Genetic testing

One of the greatest utilities of genetic testing in patients with VWD is confirmation of the specific subtype, especially when specialized plasma or platelet assays are not available or have discordant or unexpected results. Genetic testing should be performed in the context of adequate genetic counseling and clear testing goals.

Targeted sequencing of VWF exon 28 is recommended to confirm type 2B in patients with a low VWF activity to antigen ratio and low VWF:CB/VWF:Ag or abnormal multimers, and targeted sequencing of VWF exons 17-21 and 24-27 in patients with suspected 2N VWD is useful when VWF:FVIIIB is not available.¹ Genetic testing can also help identify variants associated with increased clearance or confirm a genetic etiology when the clinical presentation has overlap with acquired VWD or a combination of subtypes. Identification of the causative variant(s) is important to provide appropriate counseling of reproductive risk for individuals with VWD and their families.

The negative predictive value of genetic testing is low. With multiple factors outside the VWF gene affecting levels, a large proportion of patients with quantitative defects do not have an identifiable disease-causing variant.¹⁶ Large deletions and duplications are a known cause of VWD and can be missed by sequencing assays.¹⁷ Deep intronic variants, complex structural rearrangements, and mobile genomic elements also cause inherited disorders and are not detected by routine methods.

CLINICAL CASE (continued)

Targeted sequencing of VWF exon 28 was performed and confirmed the presence of the heterozygous pathogenic VWF c.3917G>A, p.Arg1306Gln variant, which is associated with type 2B VWD.

Conflict-of-interest disclosure

Juliana Perez Botero: no competing financial interests to declare.

Off-label drug use

Juliana Perez Botero: Nothing to disclose.