Alcohol use: less is better. An umbrella systematic review of clinical interventions, policies, and dose–response health risks in adults
Lower alcohol intake and avoiding binge drinking reduce adult health risks.
*Umbrella systematic review; Level 2a (OCEBM).
Citation
Dionisi T, De Vita V, Di Sario G, Gasbarrini A, Addolorato G, et al. Alcohol use: less is better. European Journal of Internal Medicine. 2026;149:106837. doi:10.1016/j.ejim.2026.106837
Background
Alcohol contributes to injuries, chronic disease, early death, and health inequities. This review assessed whether “low-risk” drinking thresholds are reliable and which clinical or policy approaches reduce harm.
Patients
Adults, mainly aged 18 years or older; some global surveillance data used age 15 years or older. Excluded were opinion-only articles, animal-only studies, and very small studies unless they were key reference sources.
Intervention
Lower alcohol intake, alcohol reduction counseling, digital or clinician brief interventions, and alcohol-control policies.
Control
Higher intake, no or minimal intervention, usual care, or weaker alcohol policies.
Outcome
Alcohol-attributable death and disability, injuries, liver disease, cancer, heart and brain outcomes, drinking amount, and policy-related harms.
Follow-up Period
Searches covered January 2015 to March 2026; intervention outcomes often used 6 to 12 months.
Results
| Finding | Effect |
|---|---|
| Included evidence | 49 records addressing 37 questions |
| Any drinking before injury | Odds ratio 2.80 |
| Cirrhosis at 40 g/day | Risk ratio 9.35 in women; 2.82 in men |
| Light drinking and breast cancer | Risk ratio 1.05 |
| Primary care brief intervention | 20 g/week less alcohol at 12 months |
| Pricing and some availability policies | Generally lower consumption and harms |
g/day = grams of pure alcohol per day.
Results were synthesized narratively, not pooled across the whole umbrella review. Apparent low-dose benefits were not robust when bias was considered.
Limitations
Most included reviews had low or critically low methodological ratings. Alcohol exposure was often self-reported and inconsistently measured. Many findings were observational, so causation and exact safe thresholds remain uncertain. Policy evidence varied by setting and enforcement.
Funding
No specific funding reported; no declared competing interests.
Clinical Application
Counsel adults that less alcohol is safer; routinely screen, offer brief counseling, and refer suspected alcohol use disorder for treatment.
Discussion
Sign in to join the discussion.
In this umbrella systematic review, higher alcohol intake/risky patterns were associated with harms such as injury (OR 2.80) and light drinking with breast cancer (RR 1.05); since much evidence was observational and cannot prove causation, is the “less is better” message valid, applicable, and practice-changing? The umbrella review reported that any alcohol use within 6 hours before an injury was associated with injury compared with no alcohol use, with an OR of 2.80. Which interpretation is most appropriate?
Easy to believe most of the conclusions It’s unfortunate that the abstract, and summary, didn’t include the confidence intervals. They are available in the full article, which is not behind a paywall. I was surprised that brief interventions had such a large effect size
The brief-intervention finding is more modest than it sounds. The mean difference is −20 g of ethanol/week (95% CI −28 to −12) at approximately 12 months, accompanied by only a very small reduction in heavy episodic drinking (MD −0.08 episodes/week, 95% CI −0.14 to −0.02). A reduction of 20 g/week sounds larger than it actually is, as it represents only about 1.4 U.S. standard drinks per week. Finding Effect with 95% CI Any drinking within 6 h before injury OR 2.80 (95% CI 2.21–3.54) Cirrhosis risk at 40 g/day Women: RR 9.35 (95% CI 7.64–11.45); men: RR 2.82 (95% CI 2.53–3.14) Light drinking and breast cancer RR 1.05 (95% CI 1.04–1.07) Primary-care brief intervention 20 g/week less alcohol (95% CI 12–28 g/week less)