OBJECTIVE: To explore the effect of smoking and smokeless tobacco, 'snus', on the risk of type 2 diabetes. DESIGN: Population-based cross-sectional and prospective follow-up study in northern Sweden. SUBJECTS: A total of 3384 men, aged 25-74 years, who participated in the MONICA study in 1986, 1990, 1994 or 1999, 1170 of whom had an oral glucose tolerance test. In 1999, 1757 men from previous cohorts returned for re-examination. Main outcome measures. We compared the prevalence of type 2 diabetes or pathological glucose tolerance (PGT) amongst tobacco users to that of nonusers at entry into the study and at follow-up, using odds ratios. RESULTS: Compared with never users, the age-adjusted risk of prevalent clinically diagnosed diabetes for ever smokers was 1.88 (CI 1.17-3.0) and for smokers 1.74 (0.94-3.2). Corresponding odds ratios for snus users were 1.34 (0.65-2.7) and 1.18 (0.48-2.9). We found no increased risk of prevalent PGT in snus users or smokers. Former smokers and snus users had an insignificantly increased risk for PGT. Compared with nonusers, the age-adjusted risk of developing clinically diagnosed diabetes during follow-up was 4.63 (1.37-16) in consistent exclusive smokers, 3.20 (1.16-8.8) in ex-smokers and no cases in consistent snus users. The risk of PGT during follow-up was not increased in consistent tobacco users but evident, although not statistically significant, in those who quit snus during the follow-up period, 1.85 (0.60-5.7). Adjustment for physical activity and alcohol consumption did not change the major findings. CONCLUSIONS: The risk of diabetes for snus users was not significantly increased. Smoking was associated with prevalent and incident cases of diabetes. Ex-tobacco users tended towards more PGT.