A causal relationship between alcohol intake and type 2 diabetes mellitus: A two-sample Mendelian randomization study

Abstract

Background and aimsWe investigated whether alcohol intake has a causal relationship with type 2 diabetes mellitus (T2DM) risk in adults of the Korean Genomic Epidemiology Study using two-sample Mendelian randomization (MR) analysis. Methods and resultsDaily alcohol intake was calculated based on the type, average amount, and frequency of alcohol consumption for six months before the interview. The participants were divided into low- and high-alcohol intake of 20 g/day. After adjusting for the covariates related to T2DM, the independent genetic variants (instrumental variables) related to alcohol intake were explored by GWAS analysis in a city hospital-based cohort (n = 58,701). SNPs with a significant level of p-value <5 × 10−8 and linkage disequilibrium of r2 < 0.001 were retrieved. MR methods were used to analyze the causality between alcohol intake and the T2DM risk, and the heterogeneity and leave-one-out sensitivity analyses were conducted in Ansan/Ansung plus rural cohorts (n = 13,598). High alcohol intake increased T2DM risk when the inverse-variance weighted (P = 0.012) and weighted median (P = 0.034) methods were used, but not when the MR-Egger method was used. No significant heterogeneity and horizontal pleiotropy between alcohol intake and T2DM were detected. A single genetic variant did not affect the causal association in a leave-one-out sensitivity analysis. ConclusionThis study supports that heavy alcohol intake appears to be causally associated with T2DM risk.