The purpose of the cohort study reported here was to investigate the association between oral contraceptive use and risk of benign breast disease (BBD), overall and by histological subtype, within the 56,537 women in the Cana-dian National Breast Screening Study (NBSS) who completed self-administered lifestyle and dietary questionnaires. The NBSS is a randomized controlled trial of screening for breast cancer in women aged 40-59 at recruitment. Cases were the 2,116 women in the dietary cohort who were diagnosed with biopsy-confirmed incident BBD. For comparative purposes, a subcohort consisting of a random sample of 5,681 women (including 197 subjects with incident BBD) was selected from the full dietary cohort. After exclusions for various reasons, the analyses were based on 2,116 cases and 5,338 non-cases. There was an inverse association between use of oral contraceptives and risk of all types of BBD combined. The reduction in risk was confined largely to proliferative forms of BBD (BPED), and in particular, to those forms of BPED without histological atypia, in whom there was a progressive reduction in risk with increasing duration of use (the IRR (95% CI) for use of more than 7 years was 0.64 (0.47-0.87)); risk of BPED with atypia was increased somewhat in association with oral contraceptive use (the IRR (95% CI) for use of more than 7 years was 1.43 (0.68-3.01)), but not in a dose-dependent manner. The results were similar when examined separately in the screened and control arms of the NBSS and for screen-detected and interval-detected BPED. There is some evidence for an inverse association between use of oral contraceptives and risk of benign breast disease (BBD) (Goehring and Morabia, 1997). Given that women with BBD are at increased risk of subsequent breast cancer (Bodian, 1993), it might be anticipated that use of oral contraceptives would also be associated with reduced risk of breast cancer. However, the accumulated epidemiological evidence does not support this prediction. Indeed, recent data suggest that current use of oral contraceptives , or use within the last 10 years, is actually associated with a small increase in breast cancer risk (Collaborative group on hormonal factors in breast cancer, 1996). In an attempt to reconcile these apparently contradictory observations, Hsieh et al. (1984) suggested that use of oral contraceptives might suppress the symptoms of BBD and thereby decrease the likelihood of its detection, without influencing the underlying BBD pathology, while Cole (1977) drew attention to the difference in breast cancer risk according to histological subtype of BBD, and postulated that oral contraceptives might be associated with reduced risk of those forms of BBD not associated with increased breast cancer risk, but with increased risk of those forms of BBD associated with increased risk of breast cancer. To date, however, investigation of the former of these possibilities has been limited to adjustment of estimates of association for breast self-examination and indices of medi...