Laboratory assessment of von Willebrand factor: Altered interpretation of laboratory data, and altered diagnosis of von Willebrand's disease

Three laboratory assays for von Willebrand Factor (vWF), namely, (i) a standard ''antigen'' [antisera-enzyme-linked immunosorbent assay (ELISA)-based] (vWF:Ag), (ii) a standard ristocetin-dependent platelet agglutination procedure (vWF:RCof), and (iii) a newly developed ELISA-based functional vWF collagen binding assay (vWF:CBA), were coevaluated. We (a) assessed their ability to detect VWF under different assay conditions and Cb) reevaluated normal vWF reference ranges using citrated plasma from >200 normal individuals, The following effective normal reference ranges were derived: vWF-Ag (40-200%), vWF:CBA (50-400%), and vWF:RCof (45-200%). Based on other laboratory data, caution is indicated in the interpretation of ''borderline'' or ''equivocal'' test results (e.g., 35-55%), since these may derive from normal individuals or from vWD patients. Ln addition, our study shows that selection of different assay conditions, different test plasma dilutions, or different test sample processing methods can also lead to differences in the ''apparent'' levels of detected vWF. For example, using plasma from 100 normal individuals processed as ''test samples,'' we determined that higher sample dilutions tended to provide lower (dilution factor adjusted) assay results in the case of the vWF:Ag and vWF:CBA, and higher assay results in the case of the vWF:RCof, with greater differences noted in test samples containing higher levels of vWF. The explanation for these findings relate to the assay design. Thus, in the diagnostic laboratory, the standard vWF:Ag, vWF:CBA, and vWF:RCof assays procedures are designed to provide calibration curves that are ''upward linear'' in the (vWD) clinically important region of 0-50% of normal, with a plateau tending at higher VWF levels. Using these standard methodologies, and standard plasma dilutions, high-vWF level test samples will tend to fall on or near the plateau, and results will be unreliable, with a tendency to be exaggerated the higher the vWF. These findings have particular relevance to research studies measuring VWF in patients suffering various clinical conditions (e.g., thrombotic thrombocytopenic purpura, diabetes, renal failure) where higher than normal levels of VWF may be anticipated. In these cases, laboratory-derived vWF levels could be ''misrepresented'' if the study design did not take into account the need to revise the vWF testing procedure.