Recent case-control studies indicate that alcohol increases the risk of oral premalignant lesions (OPL) among tobacco users, but the independent association between alcohol and OPL remains unclear. We prospectively evaluated the association between alcohol consumption and the incidence of OPL. Participants were 41,458 men in the Health Professionals Follow-up Study. Alcohol consumption was assessed every 4 years using validated food frequency questionnaires. We confirmed clinically or histopathologically diagnosed OPL events occurring between 1986 and 2002 by medical record review (193 cases). Multivariate-adjusted relative risks of OPL were calculated from Cox proportional hazards models. With detailed control for tobacco and other variables, multivariate relative risks (95% confidence intervals) were 1.7 (0.9-3.2) for drinkers of 0.1 to 14.9 g/d, 2.9 (1.5-5.6) for 15 to 29.9 g/d, and 2.5 (1.3-5.1) for > or =30 g/d, compared with nondrinkers. Approximately one additional drink per day (12.5 g) was associated with a 22% increase in risk (P < 0.001). The associations did not vary by beverage type, frequency, or consumption with meals. Results were similar when restricted to cases of oral epithelial dysplasia. Alcohol increased OPL risk in never-users of tobacco as well as in past or current users. An interaction between alcohol and tobacco was apparent by their more-than-additive joint effects. Alcohol is an independent risk factor for OPL, regardless of beverage type or drinking pattern. Recommendations to reduce alcohol intake have the potential to reduce incidence of OPL in nonsmokers and smokers alike.