Association between alcohol consumption and incidence of impaired insulin secretion and insulin resistance in Japanese: The Saku study

•The risk of developing impaired insulin secretion increases with increase in alcohol consumption.•The risk of developing insulin resistance increases with increase in alcohol consumption.•There is no sex difference in the associations between alcohol consumption and the risk of insulin action. To investigate the effect of alcohol consumption on impaired insulin secretion and insulin resistance in Japanese. The participants in this 5-year cohort study were 2100 Japanese aged 30–74 years without type 2 diabetes mellitus, impaired insulin secretion, or insulin resistance who underwent a medical checkup including 75-g OGTT between April 2008 and March 2009 at Saku Central Hospital. Alcohol consumption was categorized as follows: non-drinker (0 g/week), light drinker (1–139 g/week in men and 1–69 g/week in women), moderate drinker (140–274 g/week in men and 70–139 g/week in women) and heavy drinker (≥275 g/week in men and ≥140 g/week in women). The hazard ratios (HRs) and 95% CIs among light to heavy drinkers for incidence of impaired insulin secretion (insulinogenic index ≤51.7) and insulin resistance (HOMA-IR ≥ 2.5), detected by an OGTT at the time of a follow-up medical checkup before the end of March 2014, were estimated by multivariable adjusted Cox proportional hazard models as reference values for non-drinkers. There were 708 cases of impaired insulin secretion and 191 cases of insulin resistance. The HRs (95% CIs) for impaired insulin secretion in light, moderate and heavy drinkers were 1.16 (0.96–1.40), 1.35 (1.07–1.70) and 1.64 (1.24–2.16), respectively (P for trend