Fetal outcome in systemic lupus erythematosus (SLE) was retrospectively analysed in 242 pregnancies in 112 unselected patients, and the outcome was compared with that of 417 pregnancies in 192 control women matched for age, parity and socio-economic status. Relative risk for fetal loss after the diagnosis of SLE was 2.5 (95% confidence interval (CI), 1.4-4.5), for prematurity 5.8 (3.2-10.5) and for intra-uterine growth retardation (IUGR) 8.6 (3.0-24.3). Fetal outcome of pregnancy in patients with pre-existing stable lupus nephritis was no worse than in other SLE pregnancies. Relations of three lupus anticoagulant (LA) assays and tree anticardiolipin (aCL) enzyme-linked immunosorbent assays to fetal outcome were studied. Patients positive by any LA assay had a previous fetal loss more often than patients negative by all LA assays (odds ratio 3.4; 95% CI, 1.3-9.0; P = 0.01). Of the 41 patients whose antiphospholipid antibody (aPL) tests were all negative, five (12%) had a history of fetal loss (16% in controls). As a group, aCL was more sensitive for fetal loss than LA (64% vs 50%), but LA was more specific (77% vs 52%). Combinations of one aCL assay with one LA assay had a 41-73% sensitivity and a 64-73% specificity for a history of fetal loss. aPL did not correlate to prematurity or fetal growth retardation. In conclusion, fetal loss in SLE is 2.5 times more prevalent than in the normal population. The presence of LA indicates a high risk for fetal loss, and the absence of aPL is an indication of a favorable pregnancy outcome.(ABSTRACT TRUNCATED AT 250 WORDS)