Do GLP-1 Drugs Reduce Drinking? Three Trials, and They Disagree

Something is there. People report it, observational data suggested it, and a randomized trial in 2025 found semaglutide reduced how much participants with alcohol use disorder drank in a laboratory setting.^[1] But that trial ran 9 weeks with 48 people at doses below the weight-loss range. The larger trial in this field, 127 people over 26 weeks, missed its primary endpoint.^[2] No GLP-1 medicine is approved anywhere to treat drinking. If a seller has told you otherwise, they are ahead of three small trials that disagree with each other.

The trial the coverage is based on

In 2025, JAMA Psychiatry published a phase 2 trial of once-weekly semaglutide in adults with alcohol use disorder. Participants took 0.25 mg for four weeks, 0.5 mg for four weeks, and a single 1.0 mg dose at the end. The primary outcome was measured in a laboratory: how much participants drank when given the opportunity, before treatment and after.^[1]

They drank less. The effect sizes were medium to large for grams of alcohol consumed and for peak breath alcohol concentration, and the result was statistically significant.^[1] That is a real finding and it deserves the attention it got.

It also deserves its qualifiers, all of which come from the paper itself. Forty-eight people were randomized, out of 504 assessed. Treatment lasted nine weeks. The doses were the starting rungs of the ladder, not the doses anyone takes for weight. And participants were non-treatment-seeking — they were not people who had come looking for help to stop, which is the population an approved medicine would be given to.

The larger trial, which did not work

Three years earlier, a Danish group ran exenatide against placebo in 127 treatment-seeking patients with alcohol use disorder, for 26 weeks, alongside standard cognitive-behavioral therapy. The primary outcome was the number of heavy drinking days.^[2]

It did not reduce them. That is the headline result of the larger and longer trial in this field, and it is almost never mentioned in coverage of the smaller and shorter one.

Two other things did happen. Brain imaging showed exenatide blunted the response to alcohol cues in the ventral striatum and septal area, regions central to reward, and dopamine transporter availability was lower in the treated group.^[2] And in an exploratory analysis, participants with a BMI above 30 did show reduced heavy drinking days and lower total intake.

The third trial was about smoking

Dulaglutide is the third drug in this literature, and the finding attached to it comes from a predefined secondary analysis of a trial testing dulaglutide for smoking cessation. Among the 151 participants who reported drinking at baseline and completed 12 weeks, the dulaglutide group drank about 29% less than placebo.^[3]

A predefined secondary analysis is a legitimate piece of evidence and considerably better than a fishing expedition. It is still not a trial designed to answer this question, in people selected for this problem, with this as the endpoint it was powered for.

The systematic review that assembled these reached the conclusion the evidence supports: semaglutide reduced alcohol use, dulaglutide lowered intake among current drinkers, exenatide had no significant effect on heavy drinking days, and large randomized trials are needed before any of this establishes a role in treatment.^[4]

What the label says about drinking

Nothing. We read the complete Wegovy prescribing information for this — the full label text, not a summary of it. The word alcohol appears eleven times: eight are the alcohol swab in the injection instructions, and three are non-alcoholic fatty liver disease. None concerns drinking.^[5]

So there is no labeled warning to point at, and equally no labeled reassurance. Two things in these labels are worth knowing anyway if you drink. Pancreatitis is a warning on every GLP-1 label, and heavy alcohol use is among the most common causes of pancreatitis in its own right. And these drugs can cause dehydration through vomiting and diarrhea, which alcohol contributes to rather than helps. Neither is a prohibition. Both are worth a sentence with your prescriber rather than a search engine.

Why this is being sold ahead of the evidence

Because the observational signal is genuinely striking, and because a drug that reduces several appetites at once is a compelling story. Neither is a trial. Alcohol use disorder has treatments that are approved, tested in the thousands, and available now, and the honest position is that a GLP-1 is not among them.

If a telehealth seller markets a compounded product on a claim about drinking or cravings, that is marketing an unapproved use of an unapproved product, and it is the pattern we track on the FDA warning letters page. If you are worried about your drinking, that is a conversation for a clinician, and it should not wait on the next trial.

For what these drugs are approved to do and what the evidence behind that looks like, our semaglutide timeline works through the STEP trials, and the gallbladder piece covers the risk side on the same standard of evidence.