In some studies, use of vitamin supplements has been inversely associated with the risk of several chronic diseases, but little is known about whether vitamin use affects the risk of diabetes mellitus. Using data from the National Health and Nutrition Examination Survey I Epidemiologic Follow-up Study, the author examined whether vitamin use was related to diabetes incidence in a cohort of United States adults aged 25--74 years. In the analytic sample of 9,573 participants, 1,010 participants developed diabetes mellitus during about 20 years of follow-up. A smaller percentage of participants with incident diabetes (21.4%) reported using vitamins during the previous month at baseline compared with participants who remained free of this disease (33.5%) (p < 0.001). After multiple adjustment, the hazard ratios for participants using vitamin supplements were 0.76 (95% confidence interval (CI): 0.63, 0.93) for all participants, 0.70 (95% CI: 0.54, 0.92) for men, and 0.84 (95% CI: 0.64, 1.11) for women. Sex did not modify the association between vitamin use and diabetes incidence. Whether specific vitamins or other factors closely correlated with vitamin use account for this observation is unclear.