Antiphospholipid antibody ELISAs: survey on the performance of clinical laboratories assessed by using lyophilized affinity-purified IgG with anticardiolipin and anti-beta2-Glycoprotein I activity.

Lupus anticoagulant (LA) and anticardiolipin (aCL) antibodies are the classical tests used to diagnose the antiphospholipid syndrome (APS). Unfortunately, since these are nonspecific and standardization is lacking, the results of laboratory work-ups upon which diagnosis are made are often misleading. The performance of clinical laboratories in detecting LA using lyophilised affinity purified immunoglobulin has been previously reported. The same material was used to investigate the inter-laboratory variability of aCL and anti-beta(2)-Glycoprotein I (beta(2)-GPI) antibody measurements. Laboratories were asked to test normal plasma spiked with purified IgG or distilled water in order to obtain 3 samples positive for aCL and anti-beta(2)-GPI at different antibody concentration (A, B and C) and 3 samples of normal plasma. Thirty-five laboratories participated and interpreted their test results. All performed an ELISA for IgG aCL antibodies, while 17 also tested samples using IgG anti-beta(2)-GPI antibody ELISA. Sensitivity and specificity were calculated on the basis of the responses provided by each laboratory. Overall, 99/105 samples were correctly interpreted as positive and 97/101 as negative for the presence of IgG aCL, corresponding to a sensitivity and specificity of 94% and 96%, respectively. Likewise, 46/51 samples were correctly defined as positive and 50/51 as negative for the presence of IgG anti-beta(2)-GPI corresponding to a sensitivity and specificity of 90% and 98%, respectively. A wide variability in results pertaining to the positive samples was found for aCL-ELISA (coefficient of variation of 79%, 59%, and 53% for samples A, B, and C, respectively) as well as for abeta(2)-GPI-ELISA (coefficient of variation of 85%, 95%, and 50% for samples A, B, and C, respectively). This was confirmed when the analysis was restricted to those centres using the same commercial kit. Median antibody concentrations reported by centres for positive samples were consistent with the prolongation of coagulation tests assessing lupus anticoagulant (LA). Among these, dRVVT showed a good sensitivity and linear correlation with aCL antibody concentration. In conclusion, on the whole this survey found correct interpretation of positive and negative samples by both ELISAs. Nonetheless the high variability of reported data remains a major problem that only a consensus on the part of laboratories and manufacturers to utilize standard, uniform materials and procedures can hope to overcome.