LAUSR

Objective: To investigate the association between alcohol use and incidence of type 2 diabetes mellitus (T2DM) in the middle-aged and elderly male population. Methods: All participants were from Dongfeng-Tongji cohort, 27 009 retired employees from Dongfeng Motor Corporation in Hubei Province were enrolled in the Dongfeng-Tongji cohort baseline survey in 2008. In baseline study, information of alcohol use and other covariates were collected by semi-structured questionnaire and all participants completed physical examination including the test of fasting glucose and blood lipid levels. A total of 6 784 male participants from Dongfeng-Tongji cohort who were without diagnosis of diabetes, coronary heart disease, stroke, or cancer in baseline study were enrolled in this study. We completed the first follow-up in 2013 and the outcome of disease or death was retrieved based on health-care medical records according to the unique medical insurance ID. Cox proportional hazard regression model was used to estimate the association between alcohol use and incidence of type 2 diabetes mellitus (T2DM), by drinking features and patterns. Results: Out of the 6 784 participants, 3 541 participant were defined as non-alcohol drinkers and there were 15 852.2 person-years of follow-up; among which 270 new cases of T2DM were diagnosed withthe crude incidence density of non-alcohol drinkers at 1 703.2/100 000 person-years. The other 3 243 subjects were classified as alcohol drinkers and there were 14 509.8 person-years of follow-up; and among which 258 new cases of T2DM were diagnosed, with the crude incidence density of T2DM at 1 778.1/100 000 person-years. Multivariate COX proportional hazard regression model indicated that there was no significantly increased risk of T2DM incidence between alcohol drinkers and non-alcohol drinkers(HR(95% CI): 1.09 (0.91- 1.30)). However, participants who averagely consumed >20 g/d or>7 times/week had a significantly increased risk of T2DM compared with non-alcohol drinkers, and the value of HR(95%CI) was 1.27 (1.02- 1.58) and 1.35 (1.00- 1.83), respectively. Among men who consumed alcoholic beverages more than 7 times/week, HR (95%CI) for T2DM incidence in the subjects who consumed 0.01 to 40 g and > 40 g once a time were 1.48 (1.05- 2.09) and 1.27 (0.80- 2.10), respectively. Conclusion: Although we found no relationship between alcohol use and T2DM incidence overall, alcohol use more than 20 g/d or more than 7 times/week would increase the risk of T2DM.