The influence of induced or spontaneous abortion on breast cancer risk has been the subject of numerous epidemiologic studies over the past decades and has recently received heightened attention. Here, we review the evidence to understand better the apparent inconsistencies among studies. We considered possible biases in data collection, presentation, and analysis that could create spurious associations or obscure real relations. A particularly important issue is the sensitive nature of abortion, which is a pervasive problem that could affect the validity of many studies. Also, an incomplete pregnancy deprives the woman of a potentially protective full-term pregnancy and therefore may appear harmful when compared with the experience of a woman who did carry to term. The dual effect of parity on breast cancer risk-short-term risk increase and long-term protection-adds another dimension of complexity to the interpretation of abortion studies. Long-term influences of abortion may have been insufficiently captured in some studies, as follow-up time was too short. Studies to date are inadequate to infer with confidence the relation between induced or spontaneous abortion and breast cancer risk, but it appears that any such relation is likely to be small or nonexistent.