OBJECTIVES: In 2023, after a review concluding that methodological shortfalls are mainly responsible for reported beneficial health effects of alcohol, public health institutions recommended to abstain from drinking. Yet, reports of health benefits from moderate drinking persist. We aimed to explore whether known sources of bias and distortion were accounted for in underlying analyses. STUDY DESIGN: Systematic literature search and methodological analysis of retrieved publications. METHODS: Papers on associations between alcohol consumption and all-cause mortality and cardiovascular outcomes published between January 2023 and July 2025 were identified. Characteristics of study populations and analyses were described and acknowledgement and addressing of relevant methodological issues explored. RESULTS: Five of eight studies included reported protective effects of low or moderate alcohol intake, while three reported less positive effects. Studies showing protective effects used non-drinkers or lifetime abstainers as reference- categories known to be prone to misclassification and bias. Only one study discussed methodological issues, and none applied any method to assess or mitigate potential distortion. Reported hazard reductions of 40% for cardiovascular outcomes after minimal exposures of <1 glass/week compared to non-drinkers seem implausible. In analyses yielding less beneficial effects, non-drinking participants were excluded or Mendelian Randomization methods used to account for residual confounding. Some analyses offered leeway to report beneficial or harmful effects. CONCLUSIONS: Beneficial health effects of low alcohol consumption were largely due to methodological negligence, as misclassification, bias or residual confounding remained unaddressed. When authors selectively decide what to report as main result, their responsibility towards public health is high to prevent spreading of misleading messages.
Stolpe et al. (2026) studied this question.