Abstract

Laboratory testing of von Willebrand disease (VWD) continues to be a complicated and much debated subject. There are multiple tests recommended to assess varying aspects of von Willebrand factor (VWF) function, and no single test that identifies all VWD. In order to investigate the relative effectiveness of VWF laboratory testing, we examined a series of samples sent to a central reference laboratory for a VWD screen, which included VWF antigen (VWF:Ag), VWF ristocetin cofactor activity (VWF:RCo), factor VIII activity (FVIII), and quantitative analysis of multimer distribution. Institutional review board approval was obtained to add testing for VWF binding to collagens 3, 4, and 6 (VWF:CB3, VWF:CB4, and VWF:CB6). Although no additional clinical information was available on the subjects, we presume that some degree of bleeding symptoms were present to drive clinical ordering of the VWD screen. A total of 882 subjects had results available for all the indicated VWF assays.

With regard to VWF levels, 24% of subjects had VWF:RCo below the assay's lower limit of normal (54 IU/dL) and 16% had VWF:Ag below the assay's lower limit of normal (50 IU/dL), classified here as "Low VWF". Only 2% of subjects had either VWF:Ag or VWF:RCo < 30 IU/dL, consistent with the NHLBI's published guidelines on diagnosis of VWD. Only one subject (01.%) had an isolated defect in binding to collagen 3, defined as a VWF:CB/VWF:Ag ratio <0.5 and normal multimer distribution. However, 13 subjects (1.5%) had an isolated defect in binding collagens 4 and/or 6. Concordance between collagen 4 and 6 levels was observed for all subjects, although differences in collagen 3 vs collagens 4 and 6 were observed.

With regard to multimer distribution, 2% of subjects had an abnormal multimer distribution including loss of high molecular weight or borderline abnormal multimer structure (defined as loss of the very highest molecular weight multimers). Of the subjects with abnormal multimer distribution, 82% had either abnormal VWF:RCo/VWF:Ag or VWF:CB/VWF:Ag ratios, leaving only 0.3% of the study population with an abnormal multimer distribution that would not have been suspected without formal multimer testing. Only 1% would be considered for a diagnosis of type 2A or type 2B. However, those subjects with loss of the very highest molecular weight multimers may represent acquired abnormalities related to sample processing defects rather than true structural abnormalities. 22 subjects (2.5%) had low VWF and a discrepancy between VWF:Ag and VWF:RCo suspicious for a diagnosis of type 2M VWD.

Low VWF is a not uncommon finding in subjects undergoing evaluation for VWD. Although this data cannot confirm the degree of clinical symptoms present in the subjects under study, it is certainly consistent with the idea that specific screening for VWD is of high utility in patients presenting with bleeding complaints. A limitation of this study, however, is that laboratory data cannot be directly linked to phenotype, and it is possible that only patients at high risk of having VWD were screened and therefore the risk in the general population is much lower. Multimer defects were uncommon, and generally but not always confirmed by low collagen binding, which may be dependent upon the degree of structural abnormality, missing very subtle abnormalities. However, the use of two multimer-dependent screening tests (VWF:RCo and VWF:CB) improved the detection of multimer defects. In subjects with normal multimer distribution, defects in collagen 4 or 6 binding were far more common than defects in collagen 3 binding, and of equivalent frequency to low VWF levels. VWF collagen binding is not often included in the diagnostic workup for suspected VWD patients, but our data indicate that isolated defects in VWF binding to collagens 3 and 4 or 6 may be overlooked. Although this study did not confirm bleeding symptoms, our previously reported data indicate that subjects with VWF collagen binding defects may have increased bleeding symptoms. This data confirms the lack of a single screening test for VWD, and suggests that in order to fully exclude a diagnosis of VWD, a broad spectrum of assays is required.

Disclosures

Friedman:CSL Behring: Consultancy; Instrumentation Laboratories: Consultancy; Novo Nordisk: Consultancy; Alexion: Speakers Bureau. Flood:Baxter: Consultancy; CSL Behring: Consultancy.

Author notes

*

Asterisk with author names denotes non-ASH members.