Recent reviews reach conflicting conclusions on breast cancer risk after spontaneous or induced abortion. E3N is a large-scale cohort study collecting detailed information on environmental and reproductive factors. We investigated the relation between breast cancer and a history of induced and/or spontaneous abortion, using the data from the 100,000 women aged 40-65 at entrance in 1990. Among them, over 2,600 new invasive breast cancers had been diagnosed by June 2000. Multivariate analysis, adjusted for known potential confounders, showed no association between a history of induced abortion and breast cancer risk either in the whole population (relative risk [RR] = 0.91, 95% confidence interval [CI] 0.82-0.99) or in subgroups defined by parity or by menopausal status. Overall, the association between spontaneous abortion and breast cancer was not significant (RR = 1.05, 95% CI 0.95-1.15). However, there is a suggestion of increased risk with increased number of miscarriages (RR = 1.20, 95% CI 0.92-1.56 after 3 or more). Moreover, an interaction with menopausal status was observed. In premenopause, the risk decreased with increasing number of spontaneous abortions, whereas it increased in postmenopause. Among nulliparous and parous women, the relative risk estimates were respectively equal to 1.16 (95% CI 1.04-1.30, p trend < 0.0008) and 1.14 (95% CI 1.01-1.28, p trend = 0.005). Premenopausal breast cancer, on the other hand, appeared to be less frequent in women who had had repeated miscarriages. We conclude that there is no relationship between breast cancer and induced abortion but that an association with spontaneous abortion is possible and may depend on menopausal status.