GLP 1 alcohol survey finds about half of drinkers cut back on treatment

A GLP 1 alcohol survey published in JAMA Psychiatry finds that about 44 percent of US adults who were drinking when they started a GLP 1 receptor agonist say they drank less in the first three months. The share rose to about 54 percent among people who screened positive for recent heavy drinking. The study asked people to recall a change; it did not measure intake or test a drug against placebo. No molecule moves on our board, but the result adds patient voices to a question that several trials are now testing.

What the national survey measured

The JAMA Psychiatry research letter, led by Lori Uscher-Pines of the Brown University School of Public Health with senior author Haiden Huskamp of Harvard Medical School, appeared online on 14 October 2026, and Harvard publicised it this week. Co authors came from McLean Hospital, Duke, Brandeis and RAND. The GLP 1 alcohol survey drew on the RAND American Life Panel, a national group of adults who answer surveys on devices supplied to them, and the team fielded its questions in March and April 2026. Of 2,009 people who finished the first survey, 368 reported current or past GLP 1 use, and 242 said they drank alcohol when they started the drug.

Those 242 drinkers form the core of the analysis. According to the paper, 44.2 percent of them, or 106 people, said they drank less during the first three months on treatment. Most of the others reported no change, and three said they drank more. The survey did not ask how much less, so the size of any drop is unknown. The public summaries also do not split the answers by molecule, which means the finding covers the class as a whole rather than semaglutide, tirzepatide or any single brand. For readers used to trial readouts, this is a different kind of evidence: a snapshot of what users remember, gathered after the fact, with no control group and no breath or blood measures.

Heavier drinkers reported the biggest change

The survey used one screening question for unhealthy drinking: whether the person had, on any single occasion in the past year, reached the National Institutes of Health threshold for heavy drinking, which is four drinks for women and five for men. Medscape's summary of the paper puts the share of drinkers who screened positive at 37.3 percent. In that group, 54.1 percent reported drinking less, 40.2 percent reported no change and 5.7 percent, three people, said they drank more. After adjustment, people who screened positive had about 2.3 times the odds of reporting a reduction compared with those who screened negative. They were also more likely to know that GLP 1 drugs might lower the urge to drink.

In that group, 54.1 percent reported drinking less, 40.2 percent reported no change and 5.7 percent, three people, said they drank more.

That awareness point matters for reading the result. Only 10 percent of everyone surveyed said they had heard that GLP 1 drugs might reduce drinking, so most respondents were unlikely to be reporting a change they had been primed to expect. Even so, a memory of past habits can drift, and people who feel better on a drug may credit it with other changes in their lives. Huskamp called the result "pretty striking" in the Harvard release, noting that all the evidence had been anecdotal when the work began. Busch, of McLean Hospital, was more guarded. "We weren't testing GLP-1s for efficacy in reducing alcohol use. We'd need a randomized controlled trial for that," she said.

Alcohol use disorder trials hold the real answer

The authors draw a clear line between their sample and alcohol use disorder, a clinical diagnosis for drinking that causes serious harm to health and daily life. A positive answer to one heavy drinking question is not that diagnosis, so the survey says nothing direct about patients who would be candidates for treatment. The gap the team set out to study is large. In a recent national survey cited in the Harvard release, only 3 to 5 percent of people with a past alcohol use disorder reported using approved medicines for it, and only 10 percent of those seeking treatment received them. The paper also notes that about 178,000 US residents die each year from excessive alcohol use.

Controlled evidence for the peptide class is still thin. A randomized phase 2 trial of weekly semaglutide injections in adults with alcohol use disorder, published in 2025, reported lower craving and fewer drinks on drinking days than placebo, but no change in the number of days without alcohol. A phase 2 trial of oral semaglutide for the disorder appeared in the American Journal of Psychiatry in July 2026, and a recent review lists a dulaglutide trial that also found reductions against placebo. None of these is a registration program. "There are trials underway," Huskamp said. Until larger trials report, the GLP 1 alcohol survey is a signal from patients, not proof of benefit.

What the survey changes for the GLP 1 pipeline

For our board, nothing moves. The paper notes that GLP 1 receptor agonists hold FDA approval for type 2 diabetes and obesity, and no company has announced a filing for alcohol use disorder. A patient survey could not support one. What the GLP 1 alcohol survey does is put a population number on reports that had been mostly anecdotal, and show that the reported effect is stronger in people who drink more. That pattern is the one a randomized trial would need to confirm, with measured intake, a defined diagnosis, a placebo arm and enough follow up to judge relapse. Until then, any claim that a given brand treats drinking runs ahead of the data.

There are practical cautions as well. GLP 1 medicines have their own side effects, mainly digestive, and doctors prescribe them for weight and blood sugar, not for drinking. Talk about cravings can also push people toward unregulated copies sold outside pharmacies, which carry counterfeit and contamination risk; legitimate access runs through a clinician and a licensed pharmacy. Like the recent longevity signals in mice and DNA clocks, this is an off label hint that needs human trials before it changes care. Patient level data can also be noisy, as our GLP1R gene variant study coverage showed. Our explainer label follows how we score moves. This is research framing only, not medical advice.

Conclusion

The GLP 1 alcohol survey adds a large, national patient view to a question that trials are only starting to answer. About 44 percent of drinkers, and about 54 percent of those at risk, said they cut back after starting treatment, yet nobody measured how much, and respondents were not diagnosed with alcohol use disorder. The next checkpoint is randomized data on intake, craving and relapse. Until then, no GLP 1 medicine is approved for drinking, and the board stays where it was.