LAUSR

I was very interested to see the recent publication on diagnosis of type 2M von Willebrand disease (VWD) in this jounal.1 This manuscript investigated 52 patients with 2M VWD, and in particular reported on the correlation between genotype and phenotype; for example, finding different phenotype patterns based on whether mutations were found in the von Willebrand factor (VWF) A1 vs A3 domains. As such, the manuscript certainly advances the field. This Commentary reflects on some recent data that seem to be omitted from the manuscript. First, a recent paper has been published, similarly finding differences in VWF phenotype/genotype correlations for type 2 VWD according to the affected VWF domain but seems to be omitted from the reference citations.2 It is possible that this paper was not in press at the time of their study. Nevertheless, Woods et al. reported on the differences between and among 2A and 2M VWD according to whether cases reflected variants in the A1 domain (2x 2A, 36x 2M) or the A2 domain (29x 2A, 39x 2M). In brief, in type 2M with diseasecausing variants (DCVs) in the VWFA1 domain, VWF collagen binding (CB), using type 1 collagen, to antigen (Ag) ratio (CB1/Ag) was normal, but with DCVs in the VWFA2 domain, CB1/Ag was low. There was also a higher frequency of major bleeding in VWD 2M patients with DCVs in the VWFA2 domain than those with DCVs in the VWFA1 domain, which the authors proposed could be a summative effect of abnormal C1B/Ag, on top of the reduced VWF platelet glycoprotein Ib (GPIb) binding. In silico modeling also suggested that DCVs impairing the VWFA2 domain somehow modulate collagen binding to the VWFA3 domain. Another comment on the data provided in the earlier report1 is around the different utility of different “platelet GPIbbinding” assays for phenotypic assessment of 2M VWD. In their report, although a single VWF:CB assay was used to interrogate VWF collagen binding in all 2M VWD cases,1 three different assays were used to interrogate what the authors called VWF “activity” (VWF:Act). Thus, ristocetin cofactor (VWF:RCo) was used in 35 patients, a monoclonalbased assay (VWF:Ab) was used in nine patients, a gainoffunction “mutated” GPIb (VWF:GPIbM) assay in three patients, and a recombinant GPIb (VWF:GPIbR) in five patients.1 However, the patients were more or less treated as a homogeneous group for such VWF:Act, and only composite data were reported in the main manuscript. The authors “justified” this on the basis that all these assays essentially reflected assays able to identify “VWFdependent platelet adhesion.” Even in the Supplementary file, the data are not differentially reported according to the different VWF:Act assays. The problem here is that these assays have been shown to have different sensitivity to type 2 VWD, inclusive of 2M VWD. For example, in one study that assessed assay sensitivity to 2M VWD in a crosslaboratory setting,3 the VWF:Ab assay in particular had poor sensitivity to two different 2M VWD cases assessed in some 60 laboratories. I have provided the differential test patterns as summarized for these cases in Figure 1 to highlight the problem. Note that these are historical data, including results as available from 2012 to 2015. At this time, VWF:Ab use in our geography was higher than VWF:GPIbR, but similar to VWF:GPIbM. VWF:GPIbR and VWF:GPIbM use has significantly exceeded that of VWF:Ab in subsequent years.4 Nevertheless, these data clearly show different patterns between different “VWF:Act” assays in two different cases of type 2M VWD (Figure 1). Both cases (3965A>c = His1322Pro and 3974C>T = Ser1325Phe [homozygous]) represent VWF A1 variants, and the findings from Maas et at.,1 for A1 variants in their case series, reflected low relative “VWF:Act/Ag” (median 0.32 [interquartile range 0.24– 0.48]) but normal VWF:CB/Ag (0.80 [0.65– 1.01]). In the two cases shown in Figure 1, VWF:CB/Ag ratios were (as “expected”) normal, and VWF:Act/Ag ratios were generally low; however, VWF:Ab/Ag ratios were the least low, especially for the 3974C>T = Ser1325Phe variant, where ratios were generally normal. VWF:Ab levels are also less useful that either of VWF:RCo, VWF:GPIbR, or VWF:GPIbM in samples with loss of high molecular weight (HMW) VWF (e.g., type 2A, 2B VWD).4– 6 In HMWdeficient samples, VWF:RCo/Ag, VWF:GPIbR/Ag, and VWF:GPIbM/ Ag ratios tend to be lower than VWF:Ab/Ag ratios, and thus VWF:RCo, VWF:GPIbR, or VWF:GPIbM tend to be more effective than VWF:Ab for identification of such samples. Differences in “VWF:Act/Ag”’ ratios in type 1 vs 2 (HMW deficient) VWD, and thus relative utility to discriminate type 1 vs 2 (HMW deficient) VWD, have also been recently highlighted in a recent commentary7 on the American Society of Hematology, International Society on Thrombosis and Haemostasis, National Hemophilia Foundation, and World Federation of Hemophilia 2021 guidelines on the diagnosis of VWD.8