Aims/introduction: Previous meta-analyses identified an inverse association of total alcohol consumption with the risk of type 2 diabetes. The current study further explored the relationship between specific types of alcoholic beverage and the incidence of type 2 diabetes.
Materials and methods: A search of PubMed, Embase and Cochrane Library databases from January 1966 to February 2016 was carried out for prospective cohort studies that assessed the effects of specific types of alcoholic beverage on the risk of type 2 diabetes. The pooled relative risks with 95% confidence interval were calculated using random- or fixed-effect models when appropriate.
Results: A total of 13 prospective studies were included in this meta-analysis, with 397,296 study participants and 20,641 cases of type 2 diabetes. Relative to no or rare alcohol consumption, wine consumption was associated with a significant reduction of the risk of type 2 diabetes, with the pooled relative risks of 0.85, whereas beer or spirits consumption led to a slight trend of decreasing risk of type 2 diabetes (relative risk 0.96, 0.95, respectively). Further dose-response analysis showed a U-shaped relationship between all three alcohol types and type 2 diabetes. Additionally, the peak risk reduction emerged at 20-30 g/day for wine and beer, and at 7-15 g/day for spirits, with a decrease of 20, 9 and 5%, respectively.
Conclusions: Compared with beer or spirits, wine was associated with a more significant decreased risk of type 2 diabetes. The present study showed that wine might be more helpful for protection against type 2 diabetes than beer or spirits.
Keywords: Meta-analysis; Specific alcohol consumption; Type 2 diabetes.
© 2016 The Authors. Journal of Diabetes Investigation published by Asian Association for the Study of Diabetes (AASD) and John Wiley & Sons Australia, Ltd.