Alcohol consumption and diabetes risk in a Chinese population: a Mendelian randomization analysis

Aim To assess the causality between alcohol intake, diabetes risk and related traits. Design Mendelian randomization (MR) study. Subgroup analysis, standard instrumental variable analysis and local average treatment effect (LATE) methods were applied to assess linear and non‐linear causality. Setting China. Participants A total of 4536 participants, including 721 diabetes cases. Findings Carriage of an ALDH2 rs671 A allele reduced alcohol consumption by 44.63% [95% confidence interval (CI) = –49.44%, −39.37%]. In males, additional carriage of an A allele was significantly connected to decreased diabetes risk for the overall population [odds ratio (OR) = 0.716, 95% CI = 0.567–0.904, P = 0.005] or moderate drinkers (OR = 0.564, 95% CI = 0.355–0.894, P = 0.015). In instrumental variable (IV) analysis, increasing alcohol consumption by 1.7‐fold was associated with an incidence‐rate ratio of 1.32 (95% CI = 1.06–1.67, P = 0.014) for diabetes risk, and elevated alcohol intake was causally connected to natural log‐transformed fasting, 2‐hour post‐load plasma glucose (β = 0.036, 95% CI = 0.018–0.054; β = 0.072, 95% CI = 0.035–0.108) and insulin resistance [homeostatic model assessment for IR (HOMA‐IR] (β = 0.104, 95% CI = 0.039–0.169), but was not associated with beta‐cell function (HOMA‐beta). In addition, the LATE method did not identify significant U‐shaped causality between alcohol consumption and diabetes‐related traits. In females, the effects of alcohol intake on all the outcomes were non‐significant. Conclusion Among men in China, higher alcohol intake appears to be causally associated with increased diabetes risk and worsened related traits, even for moderate drinkers. This study found no significant U‐shaped causality between alcohol consumption and diabetes‐related traits.