Aims To investigate the association between alcohol consumption and impaired glucose tolerance and Type 2 diabetes mellitus. Methods A population-based cross-sectional study consisting of 3128 Swedish men, aged 35-56 years. Oral glucose tolerance testing identified 55 cases of Type 2 diabetes and 172 cases of impaired glucose tolerance. Information on alcohol consumption, family history of diabetes, smoking and physical activity was obtained by questionnaire. Results After adjustment for family history, smoking, physical activity and body mass index, the odds ratio of diabetes was 2.1 (95% confidence interval (CI) 1.0-4.5) in men with high consumption of alcohol (corresponding to over 12 drinks per week) and 0.7 (0.3-1.8) in moderate consumers (7-12 drinks), compared to occasional drinkers. For impaired glucose tolerance, the corresponding odds ratios were 0.7 (0.5-1.1) and 0.6 (0.4-1.0), respectively. Separate analyses for type of beverage indicated that high consumers of beer, spirits and wine had an odds ratio for diabetes of 2.9 (1.2-6.9), 3.3 (1.4-7.8) and 1.2 (0.5-2.7), respectively. Conclusions The results indicated that high consumption of alcohol increases the occurrence of Type 2 diabetes and that this may primarily concern consumption of beer and spirits. For impaired glucose tolerance, regular alcohol consumption was associated with a reduced prevalence, particularly at moderate levels.