44 Percent of GLP-1 Users Who Drank Say They Drank Less. The Survey Cannot Say the Drug Did It.

What is actually known about GLP-1 drugs and drinking comes mostly from one small randomized trial of 48 people. A research letter published online October 14 in JAMA Psychiatry adds a different kind of evidence: what a nationally representative sample of 2,009 US adults says happened to their drinking after they started one of the drugs.

The survey ran in March and April 2026 on RAND's American Life Panel, a probability-based national panel. Of 2,857 invited panelists, 2,009 completed it, a 70.3 percent response rate. In all, 368 respondents reported current or past use of a GLP-1 receptor agonist. Of those, 126 were not drinking when they started and were left out, which left 242 who drank at the time.

The team, led by Lori Uscher-Pines and Ateev Mehrotra of the Brown University School of Public Health, asked how their drinking changed in the first three months. Among the 242, 44.2 percent said they drank less, 52.6 percent saw no change, and 3.3 percent drank more. These are weighted figures.

Who reported cutting back

Unhealthy drinking was the only factor tied to reporting a reduction. After adjusting for sex, age, income, race and ethnicity, and self-rated health, the adjusted odds ratio was 2.30 (95% CI, 1.20 to 4.42). No demographic variable showed a significant link.

The definition matters. The screen was the NIAAA single-question measure, which counts one or more heavy-drinking days in the past year. By that standard 37.3 percent of the entire sample screened positive. The label covers a wide range of drinkers, not only people with an alcohol use disorder.

The randomized trial is smaller and less tidy

The best controlled evidence is a phase 2 trial published in JAMA Psychiatry in April 2025, led by Christian Hendershot, whose affiliations include the Keck School of Medicine at the University of Southern California. It randomized 48 adults with alcohol use disorder who were not seeking treatment to nine weeks of weekly semaglutide injections or placebo. About 71 percent were women.

In a laboratory drinking session after treatment, the semaglutide group consumed less alcohol (beta -0.48, 95% CI -0.85 to -0.11). Weekly craving fell, and so did drinks per drinking day. But average drinks per calendar day and the number of drinking days did not change. The authors called the result initial prospective evidence and said larger trials were justified.

The two studies do not line up neatly. The survey reports that nearly half of drinkers felt they drank less. The trial, measuring actual consumption over nine weeks, found a narrower effect: fewer drinks on the days people drank, with no drop in how often they drank.

What a recall survey cannot show

The survey authors list their own limits: self-reported alcohol use, retrospective recall of changes after starting the drug, and the risk of recall and social desirability bias. The answer choices ran from much more to much less, so the 44.2 percent likely folds in small shifts that were never counted in drinks.

There is also no comparison group. People who start a GLP-1 drug are often changing other things at the same time, and a survey of users alone cannot separate the drug from those changes. The letter also does not establish how long any reduction lasted past the first three months.

An awareness gap, and a policy question

Only 9.7 percent of respondents had heard that GLP-1 drugs may reduce the urge to drink. For comparison, 79.3 percent knew of the weight-loss use and 45.7 percent knew of blood glucose control. Among unhealthy drinkers the awareness figure was 13.3 percent, against 7.6 percent for others.

The authors write that awareness of a potential effect on alcohol use "was not yet widespread." They add that rising demand could be hard to meet without policy changes, such as easing insurance barriers or expanding telehealth prescribing. The work was funded by two NIAAA grants. The funder had no role in the study, and co-author Haiden Huskamp serves on the NIAAA National Advisory Council.