A randomized, placebo-controlled trial just tested semaglutide against alcohol use disorder for the first time in treatment-seeking patients. The results were bigger than researchers expected - and the drug isn't approved for this yet.
Only three medications are FDA-approved for alcohol use disorder (AUD) - disulfiram, acamprosate, and naltrexone - and none work especially well. GLP-1 receptor agonists were never designed for addiction, but the same brain reward circuits they dampen for food cravings also process alcohol craving. Registry studies of people already on GLP-1s for diabetes or obesity had hinted at lower rates of alcohol-related diagnoses, and a small 2025 trial of low-dose semaglutide in non-treatment-seeking drinkers showed reduced consumption in a lab setting.
What was missing was a real randomized trial in people who were actually seeking treatment for AUD. That trial - run out of Copenhagen University Hospital with NIH's addiction and alcohol institutes as co-authors - published its results in The Lancet on May 2, 2026.
108 adults with moderate-to-severe AUD and a BMI of 30 or higher were randomized 1:1 to once-weekly subcutaneous semaglutide (titrated to 2.4 mg, the same dose as Wegovy) or placebo injections, for 26 weeks. Everyone - both arms - also got up to ten sessions of cognitive behavioral therapy, so this tested semaglutide as an add-on, not a replacement for therapy.
The primary outcome was the change in "heavy drinking days" - a validated measure - from baseline to week 26. Alcohol use was cross-checked against phosphatidyl ethanol, a blood biomarker that can't be self-reported around, which is a meaningfully strong design choice for a trial that otherwise relies on self-report.
Heavy drinking days fell 41.1 percentage points in the semaglutide group vs. 26.4 points with placebo - a 13.7-point treatment difference (p=0.0015). Craving scores, total alcohol consumed, and drinks per drinking day all improved significantly more on semaglutide, and the liver/alcohol biomarker GGT dropped more than twice as much.
Sources: Klausen et al., The Lancet, May 2026 (DOI: 10.1016/S0140-6736(26)00305-3) | NIH Research Matters summary.
GI side effects were common and clearly dose-related to the drug itself: nausea in 57% of the semaglutide group vs. 7% on placebo, constipation 35% vs. 17%, reflux 28% vs. 2%. Most were mild-to-moderate and transient - the pattern that shows up in every semaglutide trial - and only one serious adverse event (a hospitalization for abdominal pain) was recorded in the semaglutide arm across the full 26 weeks.
This is one trial, not an approval. It was single-center, 108 people, and every participant had both AUD and obesity (BMI >=30) - the authors are explicit that results may not generalize to people with AUD who aren't overweight or obese. Semaglutide is not FDA-approved to treat alcohol use disorder, and the researchers themselves call for larger, more diverse trials before this becomes a treatment recommendation rather than a promising signal.
This article summarizes a single published clinical trial for general education. It is not medical advice. Semaglutide is not approved for alcohol use disorder, and no one should start or adjust any medication for drinking without talking to a licensed prescriber.