Diagnosis of von Willebrand disease: An assessment of the quality of testing in North American laboratories

Background Laboratory diagnosis of von Willebrand Disease (VWD) is complex. Reliance on laboratory testing can be problematic as different VWD screening panels, assays and methodologies can produce analytic variability in test results. Objectives To compare the degree of imprecision among the VWD assays and within the platelet binding activity (PBA) assays, to determine the consensus among the VWD assays for correct classification of sample results, and to determine consensus among laboratories' von Willebrand factor (VWF) multimer interpretations and final interpretations of the VWD panels. Patients/Methods Proficiency testing results from the North American Specialized Coagulation Laboratory Association (NASCOLA) submitted by laboratories from 2010 to 2019 for all normal, type (T) 1 VWD and T2 VWD samples were analysed. Results and Conclusions Imprecision was lowest for VWF antigen and highest for collagen binding activity (CBA) with median coefficient of variation (CV) of 12% (interquartile range (IQR) 7%) and 23% (IQR 21%) respectively. Within the VWF PBA assays, the gain-of-function mutant GP1b binding (VWF: GP1bM) methods had the least imprecision (CV 9%, IQR 10%). All assays, including the various PBA methods had excellent consensus. The majority of laboratories agreed that normal (median consensus-82%, IQR 16%) and T1 VWD (median consensus-100%, IQR 9%) samples had normal multimer distribution. Consensus among laboratories for final interpretations was excellent for normal samples (median 81%, IQR 8%), good for T1 VWD samples (median 59%, IQR 9%), and fair for T2 VWD samples (median 44%, IQR 21%). Consensus on final interpretation decreased as sample complexity increased.