Knott · Diabetes care 2015 · systematic review and dose-response meta-analysis of observational studies · n=38 studies (1,902,605 participants)

Alcohol Consumption and the Risk of Type 2 Diabetes: A Systematic Review and Dose-Response Meta-analysis of More Than 1.9 Million Individuals From 38 Observational Studies.

Cited 376 times in the scientific literature.

Level 3 - non-randomized controlled study

Systematic review and meta-analysis of observational studies

PubMed 26294775 · doi:10.2337/dc15-0710 · record verified 2026-08-29

What was done

Systematic review and dose-response meta-analysis of observational studies evaluating the temporal relationship between alcohol intake and incident type 2 diabetes without language or date restrictions. The authors used fractional polynomial regression to identify dose-response curves and tested interactions by sex, geographic region, and abstainer reference category definitions (noncurrent drinkers vs. lifetime never-drinkers).

What was found

Thirty-eight studies were included (1,902,605 individuals, 125,926 diabetes cases). Compared with combined abstainers, alcohol intake under 63 g/day was associated with reduced diabetes risk, peaking at 10–14 g/day with an 18% lower hazard, whereas intake above 63 g/day increased risk. However, stratified analyses revealed that this protective association was absent in Asian populations, absent in the 5 studies using a strict never-drinking reference group, and restricted primarily to women.

Why it matters

It highlights that the apparent protective effect of moderate alcohol consumption against type 2 diabetes may be an artifact of reference group bias (former drinkers classified as abstainers) and demographic confounding rather than a genuine physiological benefit.

Limits

All included data are observational and cannot establish causality. Only 5 of 38 studies separated lifetime never-drinkers from former drinkers. Alcohol intake was self-reported, exposing results to reporting and recall bias. Residual confounding from socioeconomic status, diet, and lifestyle factors could not be fully eliminated.

Cited by