BACKGROUND AND AIM: Benzo[a]pyrene (BaP) is an endocrine-disrupting pollutant formed during incomplete combustion of organic materials. It has been recognized as a reproductive toxicant, however epidemiological evidence of its long-term effect on breast cancer (BC) is limited. Thus we evaluated associations between ambient air BaP exposure and risk of BC, overall and according to menopausal status and molecular subtypes (estrogen receptor negative/positive (ER-/ER+) and progesterone receptor negative/positive (PR-/PR+)), stage and grade of differentiation of BC in the French E3N cohort study. METHODS: Within a nested case-control study of 5,222 incident BC cases and 5,222 matched controls, annual BaP exposure was estimated using a chemistry-transport model (CHIMERE) and was assigned to the geocoded residential addresses of participants for each year during the 1990-2011 follow-up period. Multivariable conditional logistic regression models were used to estimate odds ratios (ORs) and 95% confidence intervals (CIs). RESULTS:Overall, cumulative airborne BaP exposure was significantly associated with the overall risk of BC, for each 1 interquartile range (IQR) increase in the concentration levels of BaP (1.42 ng/m3), the OR = 1.15 (95%CI: 1.04-1.27). However, by menopausal status, the significant positive association remained only in women who underwent menopausal transition (i.e. premenopausal women at inclusion who became postmenopausal at diagnosis), OR per 1 IQR = 1.20; 95%CI: 1.03-1.40. By hormone receptor status, positive associations were observed for ER+, PR+ and ER+PR+ BC, with ORs = 1.17 (95%CI: 1.04-1.32), 1.16 (95%CI: 1.01-1.33), and 1.17 (95%CI: 1.01-1.36) per 1 IQR, respectively. There was also a borderline positive association between BaP and grade 3 BC (OR per 1 IQR = 1.15(95%CI: 0.99-1.34). CONCLUSIONS:We provide evidence of increased risk of BC associated with cumulative BaP exposure, which varied according to menopausal status, hormone receptor status, and grade of differentiation of BC. Our results add further epidemiological evidence to the previous experimental studies.< Réduire