Abstract
von Willebrand disease (VWD) is a very common inherited bleeding disorder. The current phenotypic classification of VWD variants includes disorders of both quantitative and qualitative defects in von Willebrand Factor (VWF) that determines optimal treatment of patients. Current phenotype determination is multimodal, cumbersome, performed by only a few specialized laboratories, and may delay the definitive diagnosis necessary in proper selection of therapy. We have developed an ELISA-based strip assay that is capable of rapid determination of relative qualitative and quantitative VWF functionality to correctly assign phenotypic variants of VWD.
136 VWD plasma samples from the Zimmerman PPG were analyzed on a new ELISA based platform. In single, individual wells this assay measures relative values of VWF:Ag (antigen), VWF:IbCo (Ib cofactor, no ristocetin), VWF:RCo (ristocetin cofactor), VWF:F8B (binding to FVIII), VWF:CB3 (binding to collagen III), and VWF:pp (propeptide) in comparison to a 30% normal control standard in a single ELISA strip assay. The study included 22 type 1 VWD, 32 type 1C VWD, 18 type 2A VWD, 23 type 2B VWD, 20 type 2M VWD, 7 type 2N VWD, 4 type 3 VWD, and 10 hemophilia A subjects. Each sample was run in single wells for each assay and optical densities (OD) were compared to the OD of a 30% standard control plasma and a 100% VWF:Ag control. The standard ELISA plate read time was 30 minutes and full assay can be accomplished in 3 hours. Two of the coauthors were blinded as to the Zimmerman PPG VWD phenotypes of test samples. Using the ELISA strip results, phenotype assignment was determined and then compared to the unblinded Zimmerman PPG VWD diagnosis. Further statistical analysis of VWF functional profile relationships was performed using the Mann-Whitney test and ROC analysis, and can quantify the ability to identify these phenotypes.
VWF functional profiles based on visually observed ratio relationships correctly assigned VWD phenotypic variant on first attempt in 122 of 136 subjects (89.7%). Repeat testing of the 14 incorrectly assigned subjects along with 11 random, correctly assigned subjects for a validation check, accurately re-assigned 9 of 14 previously incorrect phenotypes, suggesting initial plate to plate variability since all ELISA plates were made fresh for each run. Previously correctly assigned subjects, 11 of 11, remained correctly assigned. Comparing specific phenotypes revealed VWF:IbCo/VWF:Ag is good at separating type 1C from 2A; ROC area under the curve 0.875, with an optimal ratio threshold 0.649 (sensitivity 0.969, specificity 0.611, p<0.001) and excellent at separating type 2B from type 1; ROC area under the curve 0.978 with a ratio threshold 1.268 (sensitivity 1.000, specificity 0.864, p<0.001), but was poor at separating type 2A from 2M; ROC area under the curve 0.622. VWF:F8B/VWF:Ag was excellent at separating hemophilia A from type 2N; ROC area under the curve 1.000 with a ratio threshold 0.865 (sensitivity 1.000, specificity 1.000, p<0.001). VWF:CB3/VWF:Ag was excellent at separating type 2M from 2A; ROC area under the curve 0.961 with a ratio threshold 0.757 (sensitivity 0.950, specificity 0.889, p<0.001) and excellent at separating type 1C from 2A; ROC area under the curve 0.922 with a ratio threshold 0.597 (sensitivity 1.000, specificity 0.778, p<0.001).
The rapid VWF functional screening assay is able to discriminate VWD phenotypic variants with good sensitivity and specificity, and may facilitate the rapid laboratory assignment of VWD phenotypes once a subject with low VWF is identified.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.
This icon denotes a clinically relevant abstract
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal