GLP and Alcohol Craving: Three Trials, One Unresolved Question

Three separate clinical trials have now tested whether a GLP receptor agonist can reduce drinking in people with alcohol use disorder. They don’t agree with each other, and that disagreement is more informative than a single clean result would be.

Trial one: a real brain signal, an unclear behavioral one

The earliest test used exenatide, an older GLP receptor agonist, in a randomized, double-blind, placebo-controlled design in people with alcohol dependence. Researchers used functional MRI to track cue reactivity, how strongly the brain’s reward circuitry responded to alcohol-related cues, and found a real effect: exenatide attenuated that reactivity in the ventral striatum and septal area compared to placebo.

The catch is that the brain signal didn’t translate cleanly into a behavioral one. Heavy drinking days and total alcohol intake weren’t significantly different between the exenatide and placebo groups overall. The exception was a subgroup of participants who also had obesity, where a clearer effect on drinking did show up. A reduced brain response to alcohol cues is a real finding. It’s not the same as a reduced drinking outcome, and this trial found one without reliably finding the other.

Trial two: same drug, drinking as the primary outcome

A second randomized, placebo-controlled trial tested exenatide again, this time with drinking behavior itself as the primary measure rather than a secondary one built around brain imaging. The results were, again, mixed rather than a clean confirmation of benefit, consistent with the pattern from the first trial: a real biological signal that doesn’t reliably show up as a straightforward reduction in drinking across an unselected population.

Trial three: a different drug, a clearer result

The most recent trial, published in JAMA Psychiatry in 2025, tested semaglutide rather than exenatide, in a randomized, double-blind, placebo-controlled design. This one found a more consistent signal: participants on semaglutide consumed less alcohol in a laboratory self-administration task, reported fewer drinks per drinking day, and reported reduced craving compared to placebo.

Why three trials in the same drug class gave three different pictures

Here’s the honest read. Exenatide, tested twice, produced a real neural signal but an inconsistent behavioral one. Semaglutide, tested once, produced a clearer result across multiple outcome measures. That’s not necessarily a contradiction; different GLP drugs have different receptor binding profiles, different half-lives, and different dosing regimens, and it’s entirely possible that semaglutide’s specific pharmacology matters here in a way exenatide’s doesn’t. It’s also possible that a single positive trial needs replication before treating it as settled.

Multiple additional semaglutide trials for alcohol use disorder are currently registered and underway, which is itself a sign the field considers this an open question worth more data, not a closed one. Until more of those trials report, the responsible summary is that this class of drugs shows real promise for alcohol use disorder specifically with semaglutide, and a real but unresolved signal with the older GLP drugs tested before it.

References

Antonsen, K., et al. “Does Glucagon-Like Peptide-1 (GLP-1) Receptor Agonist Stimulation Reduce Alcohol Intake in Patients with Alcohol Dependence…” BMJ Open, vol. 8, no. 7, 2018, e019562.

Hendershot, Christian S., et al. “Once-Weekly Semaglutide in Adults with Alcohol Use Disorder: A Randomized Clinical Trial.” JAMA Psychiatry, 2025.

Klausen, M. K., et al. “Exenatide Once Weekly for Alcohol Use Disorder Investigated in a Randomized, Placebo-Controlled Clinical Trial.” JCI Insight, vol. 7, no. 19, 2022, e159863.