Objective: To measure the accuracy of anticardiolipin antibodies (aCL) in identifying a history of thrombosis among patients with systemic lupus erythematosus (SLE) or the primary antiphospholipid syndrome (PAPS).
Patients and methods: Patients with SLE or PAPS who attended our rheumatology clinic between April 1992 and March 1994 were included in a retrospective analysis of the relationship between thrombotic events and aCL. All aCL measurements were performed in the same laboratory by enzyme-linked immunosorbent assay, blinded as to the presence or absence of thrombosis. The diagnostic accuracy of IgG, IgM, and IgA aCL was quantified by means of the receiver operating characteristic area under the curve (ROC AUC) for each isotype. Stratum-specific likelihood ratios and their 95% confidence intervals were calculated to define strata with optimal discriminant power.
Results: During the period of study, aCL was measured in 117 patients (113 SLE and 4 PAPS), of whom 24 (20.5%) had experienced thrombotic events. The ROC AUC +/- the standard error for IgG aCL was 0.81 +/- 0.05, signifying an 81% accuracy in the identification of a history of thrombosis. In contrast, the accuracy of the IgM and IgA aCL was significantly lower (0.60 +/- 0.08 and 0.54 +/- 0.07, respectively, P < 0.05). Using stratum-specific likelihood ratios, we defined three strata in the IgG aCL scale that discriminate between patients with low, indeterminate, and high probabilities of thrombosis. For IgG aCL levels below 21.4 IgG phospholipid (GPL) U/mL, the posttest probability of thrombosis was 0.07, whereas for IgG aCL levels > or = 65.1 GPL U/mL, the posttest probability of thrombosis was 0.75. For IgG aCL values between 21.4 and 65.0 GPL U/mL, the probability of thrombosis was 0.20, equivalent to the entire cohort.
Conclusions: The IgG aCL determinations perform well in the identification of thrombosis in SLE or PAPS, while the IgM and IgA aCL have poor diagnostic accuracy. These findings may have implications for management of these patients.