Introduction: Epidemiologic studies suggest that heavy alcohol drinking is associated with unhealthy profile, whereas its light to moderate intake is associated with reduced hazard of total mortality as compared with lifetime abstention, a result largely driven by a reduction in cardiovascular mortality. Up to date, there is no agreement on a possible safe threshold for alcohol consumption in relation with cancer risk, especially breast cancer (BC) in women. Hypothesis: We assessed the hypothesis that the association of alcohol consumption with development of BC is mostly dependent on the average daily volume consumed. Methods: A prospective investigation was conducted on 11,394 women (age ≥35y) from the Moli-sani study (apparently free from any cancer at baseline; 2005-2010). The alcohol volume consumed in the year before enrolment was classified as: lifetime or former or occasional drinkers and current drinkers who drank 1-12, 12.1-24, and > 24 g/d. Incident cases of BC were validated through medical and histological records. Multivariable hazard ratio (HR) and 95% confidence interval (CI) for BC were estimated using Cox proportional hazard models and multiple imputation for missing data on covariates. Results: During a median follow-up of 13.2 years, 296 BC events occurred. In comparison with lifetime abstention, intake of alcohol up to 12 g/d was not associated with BC, while consumption >12 g/d had higher hazard of BC (Table). These findings remained consistent after the exclusion of BC events occurred during the first 12 months of follow-up or stratifying by BC subtypes (Table). Conclusions: Drinking up to one unit of alcohol (12 g) per day was not associated with either a protection or an increased BC risk in women. In contrast, drinking more than 12 g/d of alcohol had hazardous impact on BC incidence. These findings contribute to the current literature debate, suggesting that there is indeed an alcohol threshold below which alcohol drinking would not affect breast cancer risk.