The relationship between alcohol consumption and breast cancer risk was investigated using data from a co-operative case-control study conducted in Italy between 1991 and 1994 on 2569 incident, histologically confirmed breast cancer cases and 2588 controls in hospital for acute, non-neoplastic, non-hormone related conditions. Overall, 915 (38%) cases and 1048 (43%) controls were abstainers. Compared with them, the odds ratio (OR), adjusted only for age, was 1.31 (95% confidence interval (CI) 1.13-1.53) for drinkers and became 1.39 (95% CI 1.(1)21-1.60) after correction for measurement error. The multivariate OR was 1.21 for drinkers of < or = 5.87 g/day and 1.23, 1.19, 1.21, 1.41 for drinkers of 5.88-13.40, 13.41-24.55, 24.56-27.60, > 27.60 g/day, respectively. The trend in risk was significant (chi 2 = 12.28, P < 0.0005). The association was apparently stronger in premenopausal women (OR = 1.80 for > 27.60 g/day). Considering the different types of alcoholic beverages (wine, beer, digestives, grappa and other spirits), a significant direct trend in breast cancer risk was seen for wine with an OR of 1.27 (95% CI 1.06-1.53) for the category > 26.34 g/day. The ORs were also above unity for beer, grappa, digestives and spirits drinkers. No appreciable interaction was observed between alcohol drinking and body mass index, smoking, or any other covariate considered. Thus, the present data, based on a validated alcohol consumption questionnaire and on a population characterised by a relatively high alcohol consumption in women, confirmed that alcohol drinking is moderately related to breast cancer risk. If causal, this association could explain 12% (95% CI, 5-19%) of breast cancers in Italy, thus representing one of the major avoidable risk factor for breast cancer.