Despite intensive study, the relationship between oral contraception (OC) and breast cancer remains unclear. OCs contain a potent synthetic estrogen (ethinyl estradiol) but lower endogenous estradiol levels, and ethinyl estradiol is a weak progenitor of semiquinones, catechol estrogens capable of damaging DNA. NAD(P)H:quinone oxoreductase (NQO1) stabilizes semiquinones, thus potentially preventing genetic damage from catechol estrogens, and the NQO1 C609T polymorphism seems functionally relevant. Using data from the Shanghai Breast Cancer Study, a population-based case-control study, we investigated the relationships between OC use (20% ever using), breast cancer, and NQO1 (C/C 31% and C/T + T/T 69%) among 1,039 cases and 1,121 controls. Breast cancer was not significantly associated with NQO1 genotype. There was a significant protective association between OC after age 30 years and premenopausal breast cancer [odds ratio (OR) 0.51, 95% confidence interval (95% CI) 0.29-0.89] primarily with the NQO1 T allele (C/C OR 0.76, 95% CI 0.31-1.82; C/T + T/T OR 0.38, 95% CI 0.18-0.80; P for interaction = 0.19). The association between premenopausal breast cancer and OCs significantly differed with NQO1 genotype when using OCs for >18 months (C/C OR 2.34, 95% CI 0.92-5.99; C/T + T/T OR 0.69, 95% CI 0.38-1.25; P for interaction = 0.02). Among women with the C/C genotype, postmenopausal breast cancer was significantly associated with ever-using OCs (C/C OR 2.01, 95% CI 1.08-3.74; C/T + T/T OR 0.72, 95% CI 0.49-1.05; P for interaction < 0.01). This crossover was stronger with OC use prior to age 30 years (C/C OR 3.00, 95% CI 1.43-6.25; C/T or T/T OR 0.49, 95% CI 0.29-0.81; P for interaction < 0.01). Our results require confirmation but suggest that the OC and breast cancer association depends on the ability to invoke protection from catechol estrogens.