Objective. To determine the impact of pregnancy on systemic lupus erythematosus (SLE) outcome. Methods. SLE patients, age ≥16 yrs, disease duration ≤5 yrs at enrolment in LUMINA, a multiethnic cohort (Hispanics, African-Americans and Caucasians), were studied. The first pregnancy after SLE diagnosis was examined. A good pregnancy outcome was a full-term delivery; an adverse outcome was a miscarriage, abortion, premature birth or stillbirth. Dependent variables were disease activity (Systemic Lupus Activity Measure-Revised, SLAM-R) and damage accrual [Systemic Lupus International Collaborating Clinics (SLICC) Damage Index, SDI]. Differences in these variables between the visit immediately prior to, and the first visit after, pregnancy and their relationship with pregnancy outcome were examined. Damage accrual due to pregnancy exposure was examined by a case-crossover design. Results. Sixty-three SLE women from all ethnic groups were included. The mean (s.d.) age and disease duration at pregnancy outcome were 27.6 (6.5) yrs and 18.3 (22.5) months, respectively. Adverse pregnancy outcomes occurred in 76.2% women. The SLAM-R and SDI scores were statistically different after pregnancy (P = 0.050 and P < 0.001, respectively); the SDI score was independent of pregnancy outcome but strongly associated with pregnancy duration (P = 0.006), disease activity (P = 0.001), damage prior to pregnancy (P < 0.001) and total disease duration (P = 0.039) by multivariable analyses. Exposure to pregnancy itself did not impact on damage accrual in the case-crossover analyses of 142 patients (17 pregnancy exposures) (OR = 1.25; 95% CI 0.336-4.655; P = 0.480). Conclusions. Pregnancy duration, total disease duration, disease activity and damage immediately prior to pregnancy decisively impact on damage accrual after pregnancy in patients with SLE. © 2006 Oxford University Press.