The technique of meta-analysis was used to derive typical estimates of the risk of congenital malformations from oral contraceptive (OC) exposure in early pregnancy. Several definitions were used for non-exposure to reduce bias from this source. Congenital heart defects and limb reduction defects were analyzed separately. The typical relative risk from the 12 prospective studies analyzed was 0.99 (95% confidence interval 0.83, 1.19) for all malformations. The definition of non-exposure did not materially change this estimate. The typical risk for congenital heart defects was 1.06 (95% confidence interval 0.72, 1.56) and for limb reduction defects, 1.04 (95% confidence interval 0.30, 3.55). This lack of an association between OCs and birth defects in prospective studies agrees with the results of most of the better-designed case-control studies.