International study to compare antigen-specific methods used for the measurement of antiplatelet autoantibodies

Authors


Dr Peter Berchtold Department of Medicine, University Hospital, Inselspital, 3010 Bern. Switzerland.

Abstract

Platelet-associated and plasma autoantibodies against platelet glycoproteins (GP) have been demonstrated in patients with autoimmune thrombocytopenia (AITP) using various methods. Eight laboratories in seven countries participated in this international study to evaluate the interlaboratory agreement using glycoprotein-specific immunoassays for these autoantibodies. The participating laboratories received blind samples of frozen washed platelets and plasma from 22 normal donors and 22 AITP patients. Platelet-associated and plasma autoantibodies against GPIIb–IIIa and GPIb–IX were measured by MAIPA, immunobead assay or modified antigen capture assay. Of the control samples, 96.0% and 97.2% of all results for platelet-associated and plasma autoantibodies to GPIIb–IIIa/GPIb–IX, respectively, were negative. The mean variation coefficient of the control samples of platelet-associated and plasma autoantibodies was 89.5% (range 11.1–272.9%) and 46.5% (range 21.0–78.0%), respectively. In 20/22 patient samples, platelet-associated autoantibodies to either glycoprotein were noted by at least two laboratories. The mean degree of agreement in these samples was 74.0%. There was a significant correlation in the individual antibody measurements between all laboratories (Kendall coefficient of concordance 0.60 and 0.38, P < 0.001; Spearman rank order test, range of correlation coefficient 52.3–94.0% and 42.2–85.0%, P < 0.05, for anti-GPIIb–IIIa and anti-GPIb–IX, respectively). In contrast, plasma autoantibodies to either glycoprotein were noted by at least two laboratories in only 13/22 patient samples. Moreover, the degree of agreement was poor (50.1%) and a significant correlation was noted between only six pairs of laboratories. We conclude that methods used in this study yield good interlaboratory agreement in measuring platelet-associated autoantibodies against GPIIb–IIIa and GPIb–IX. In contrast, poor agreement was found in detecting plasma autoantibodies to the same glycoproteins.

Ancillary