Evaluating Performance of Contemporary and Historical von Willebrand Factor (VWF) Assays in the Laboratory Identification of von Willebrand Disease (VWD): The Australasian Experience.

Abstract

von Willebrand disease (VWD) is a common bleeding disorder that arises from deficiency and/or defects of von Willebrand factor (VWF). Appropriate diagnosis of VWD, including differential identification of qualitative (types 2A, 2B, 2M, 2N VWD) versus quantitative (types 1 and 3 VWD) defects remains problematic but has important management implications, given differential therapy. Complete assessment for VWD in a patient with a bleeding history requires comprehensive test panels, including VWF activity and antigen. We describe the Australasian experience, using data from the Royal College of Pathologists of Australasia (RCPA) Quality Assurance Program (QAP) related to VWF testing in their VWD test module. The RCPAQAP has been providing samples for VWF testing since 1998, representing 25 years of proficiency testing related to VWD diagnosis. A total of 109 samples have been dispatched to participants over these years, with current assessment involving dispatches of two samples (=4 samples) per year. Samples have represented all types of VWD, as well as normal or other samples, including acquired von Willebrand syndrome and plasma VWF concentrates as used in VWD therapy. Different VWF assays and activity/antigen ratios show different utility in VWD and type identification. In the past 9 years of data capture, a total of 166 errors were identified from a total of 1,839 interpretations, representing a base error rate of 9.0%. Identification errors were highest for type 2 VWD samples (15.3%), intermediate for type 1 VWD samples (7.5%), and lowest for normal samples (2.4%). Errors can be linked to assay limitations, including assay variability and low-level VWF detection limits, as well as laboratory issues (including test result misinterpretation, which accounts for approximately 40% of all errors for type 2 VWD). For test-associated errors, VWF:RCo and VWF:GPIbM were associated with the highest variability and error rate, which was up to 10x higher than that using VWF:CB. As a test group, chemiluminescence-based procedures were associated with lowest inter-laboratory variability, best low-level VWF detection (down to <1 U/dL), and least errors overall. These findings inform on reasons behind high rates of errors associated with VWD diagnosis, with some assays and methodologies performing substantially better than others.