GLP-1s and Alcohol Use: An Emerging Signal

A recurring story has followed GLP-1 drugs since they became widely used: people on them report that they simply want to drink less. The desire for a glass of wine fades, the second drink stops appealing, the habit loosens. When enough people independently describe the same thing, it stops being noise and starts being a signal worth taking seriously. The harder question is how seriously.

Why the idea is biologically plausible

GLP-1 receptor agonists act on the brain, not just the gut, and some of where they act overlaps with the circuitry of reward and craving. Alcohol, food, and other rewards run partly through shared pathways. If a drug can dampen the reward pull of food, it is at least reasonable that it might soften the pull of alcohol too.

Animal studies have pointed this direction for a while. Rodents given GLP-1 drugs tend to consume less alcohol in several models. Animal data does not transfer cleanly to humans, but it makes the human anecdotes more interesting rather than less.

The plausible mechanism matters because it means the reports are not obviously coincidence. But plausibility is a reason to study something, not proof that it works.

What the human evidence actually shows

Here is where calm is required. The human data is early and mostly not the kind that settles questions. But one signal is now a genuine randomized trial rather than a story. In 2025, JAMA Psychiatry published a phase 2, double-blind, randomized trial by Hendershot and colleagues testing low-dose semaglutide in adults with alcohol use disorder.

The study was small — 48 participants, randomized to semaglutide (titrated to 1.0 mg) or placebo over nine weeks — and the participants were not seeking treatment, so the results are a starting point, not a verdict. Within those limits, it met its primary outcome:

• Laboratory alcohol self-administration fell with medium-to-large effect sizes — grams of alcohol consumed (β −0.48, P = .01) and peak breath alcohol concentration (β −0.46, P = .03).
• Drinks per drinking day were reduced (β −0.41, P = .04), as was alcohol craving (β −0.39, P = .01), and heavy drinking days declined more over time on semaglutide.
• Some outcomes did not reach significance, including drinks per calendar day (P = .17) and the proportion of drinking versus abstinent days (P = .89).

In short: real signal, real limits. Larger observational analyses of prescription databases point the same direction, but cannot prove cause, since people who lose weight change many habits at once.

What would change the picture

For this to become an established use rather than an intriguing signal, the field needs larger, properly randomized trials measuring real drinking outcomes over months, not nine weeks. The authors themselves frame their result as justification to run exactly those trials. Until they read out, the honest framing is “promising lead,” not “proven treatment.”

A note on tone

Alcohol use disorder is serious, and effective options are limited, which is exactly why a possible new angle attracts attention and, sometimes, hype. It would be a mistake to either dismiss the signal or to oversell it to people who need real help. Both errors do harm.

The takeaway

The link between GLP-1 drugs and reduced drinking is one of the more genuinely interesting findings to emerge from this drug class. It is biologically plausible, supported by animal work, and now backed by at least one small randomized human trial showing reduced craving and drinking with semaglutide. It is also not yet established as a treatment — that trial had 48 people and ran nine weeks. If you are taking a GLP-1 and notice less interest in alcohol, that is consistent with what others describe and with the early data. If you are hoping these drugs are a proven treatment for problem drinking, the evidence is not there yet, and the larger trials that could provide it are still in progress.

Sources

• Once-Weekly Semaglutide in Adults With Alcohol Use Disorder: A Randomized Clinical Trial (JAMA Psychiatry, 2025)