AbstractBackground Alcohol is a major modifiable cause of morbidity, premature mortality and health inequalities, yet evidence informing "low-risk" thresholds and prevention strategies is fragmented. Methods Umbrella systematic review conducted according to PRISMA 2020 (protocol on OSF). PubMed/MEDLINE and Scopus were searched (Jan 2015–Mar 2026). An overlap-management approach selected an anchor synthesis per research question (Q1-Q37); supporting records were retained for triangulation. Quality appraisal used design-appropriate tools. Synthesis was narrative. Results Of 14,991 records, 49 were included (46 systematic reviews/meta-analyses, 2 WHO documents, 1 cross-sectional study) covering 37 pre-specified questions. Across most outcomes, higher intake and riskier patterns were associated with higher risk, with harms evident at levels often labelled 'moderate'. Any drinking increased injury odds (OR 2.80). Dose-response evidence showed steep gradients for cirrhosis (RR 9.35 in women and 2.82 in men at 40 g/day) and small but measurable increases in selected cancers at light drinking (e.g., breast cancer RR 1.05). In primary care, brief interventions reduced consumption at 12 months by -20 g/week. Pricing measures and some availability restrictions were directionally associated with lower consumption and harms, whereas evidence for other policy levers was more heterogeneous. Conclusions Overall evidence favoured lower alcohol intake and avoidance of heavy episodic drinking, although confidence varied by endpoint and was limited for several questions by the quality of the available syntheses. Apparent low-dose benefits were not robust to bias-aware analyses. These findings support a pragmatic counselling and policy message of "less is better" rather than a universal safe threshold.
Dionisi et al. (Sun,) studied this question.