Antibodies against prothrombin are detected by enzyme immunoassays (EIA) in sera of patients with antiphospholipid syndrome (APS). However, there are two methods for antiprothrombin EIA; one that uses high binding plates (aPT-A), and another that utilizes phosphatidylserine bound plates (aPS/PT). We aimed to evaluate and compare aPT-A and aPS/PT in a clinical setting. We performed EIA for anti-PT, anti-PS/PT, IgG, and IgM anticardiolipin antibodies (aCL), and IgG beta2-glycoprotein I-dependent aCL (abeta2GPI/CL) with serum samples from 139 systemic lupus erythematosus (SLE) patients (16 with history of at least one thrombotic episode) and 148 controls. We observed that: (1) although titers of anti-PT and anti-PS/PT were significantly related with each other (P < 0.0001, rho = 0.548), titer of anti-PT and anti-PS/PT differed greatly in some samples; (2) odds ratio and 95% confidence interval for each assay was 3.556 (1.221-10.355) for aPT-A, 4.591 (1.555-15.560) for aPS/PT, 4.204 (1.250-14.148) for IgG aCL, 1.809 (0.354-9.232) for IgM aCL, and 7.246 (2.391-21.966) for abeta2GPI/CL. We conclude that, while all EIA performed in this study except IgM aCL are of potential value in assessing the risk of thrombosis, aPS/PT and abeta2GPI/CL seemed to be highly valuable in clinical practice, and that autoantibodies detected by anti-PT and anti-PS/PT are not completely identical.