We evaluated factors associated with fetal losses in patients with severe lupus in a nested case-control study. We assessed separately 73 pregnancies that occurred in 46 women from a cohort of 633 Systemic Lupus Erythematosus (SLE) patients. They had at least one pregnancy after SLE diagnosis, one or more of our severity criteria and all had taken immunosuppressive drugs. Included data were related to disease severity, anti-phospholipid syndrome (APS), anticardiolipin antibodies (a-CL ab), and drugs received during pregnancy. Cases were pregnancies with fetal wastage; controls were pregnancies with live-born children. The mean age at pregnancy was 26.6 +/- 4.5 years. Cases had longer disease duration, 6.1 +/- 3.5 years vs 4.5 +/- 4.3 of controls (p = 0.02); higher prevalence of renal involvement, hemolysis and recurrent venous thrombosis (p < 0.05); they also tended to have a greater prevalence of a-CL ab, and previous fetal losses (p = 0.06). Cases used azathioprine more frequently than controls (p = 0.04). Univariate analysis showed an association of renal involvement, hemolytic anemia, azathioprine or cyclophosphamide prescription during pregnancy, previous fetal losses and APS with fetal wastage. Immunosuppressive drugs and the APS remained significant in the multivariate analysis (p = 0.05; F = 0.01). Factors related with fetal losses in women with severe SLE were: longer disease duration, ingestion of immunosuppressive drugs during pregnancy and any related manifestation of APS. We did not find macroscopic malformations in live-children of women that took azathioprine during pregnancy.