The role of spontaneous and voluntary abortions was assessed in a case-control study conducted in Milan, northern Italy on 953 cases of epithelial ovarian cancer (median age, 54) and 2500 control subjects (median age, 52) in hospital for acute diseases other than malignant, gynecological, or hormonal disorders. Overall the inverse relationship between total number of incomplete pregnancies and ovarian cancer risk was statistically significant, estimated relative risks (RRs) being, respectively, 0.9 (95% confidence interval (CI), 0.7-1.1) for one abortion and 0.8 (95% CI, 0.6-1.0) for two or more as compared to none. Furthermore, the protections afforded by spontaneous and voluntary abortions were well comparable (RRs 0.7 and 0.8 for > or = 2 spontaneous and voluntary abortions, respectively, versus none). While the protective effect of incomplete pregnancy was not explicable in terms of other characteristics of women nor significantly different across strata of reproductive factors or oral contraceptive use, it seemed more marked in women whose ovarian cancer occurred below age 50. Etiologically, these findings suggest that interrupted pregnancy per se and not predisposition to spontaneous abortion affects ovarian cancer risk.