GLP-1 Medications and
Alcohol Use Disorder: Human Studies

Alcohol use disorder and “drinking less” are not one outcome

Alcohol use disorder (AUD) is a medical condition involving impaired ability to stop or control alcohol use despite harmful social, work, or health consequences. A diagnosis and its severity require a clinical assessment; a craving score, an app tally, or a medication response cannot diagnose AUD.^[7]

Human GLP-1 studies have measured different things. Heavy drinking days count days above a study-defined threshold. Drinks per drinking day measures amount on the days a person drinks. Drinks per calendar day includes non-drinking days, while the number of drinking days measures frequency. One early trial also used a laboratory task in which participants could choose alcohol under controlled conditions. These outcomes can move in different directions, so “reduced drinking” is not a complete description of a study result.^{[2] [4]}

GLP-1 receptor agonists have FDA-approved uses for specified conditions such as type 2 diabetes and chronic weight management, but a current Wegovy label does not list AUD. NIAAA identifies naltrexone, acamprosate, and disulfiram as the three medications currently approved in the United States for AUD. As of September 23, 2026, no GLP-1 medication is FDA-approved to treat AUD.^{[7] [8]}

The randomized evidence is encouraging, mixed, and still early

Randomized results apply to the product, population, setting, duration and endpoints tested—not automatically to an entire medication class.

Exenatide: the overall primary endpoint was not positive

The first dedicated randomized AUD trial enrolled 127 treatment-seeking patients. Weekly 2 mg exenatide or placebo was given for 26 weeks with standard cognitive behavioral therapy. Exenatide did not significantly reduce the primary outcome, heavy drinking days, in the overall group; gastrointestinal events were the main adverse events. Exploratory findings for heavy drinking days and total intake in the BMI-above-30 subgroup are a research signal, not proof for every person with AUD, a body-size selection rule, or a class effect.^[1]

Two small semaglutide studies found different endpoint patterns

A U.S. phase 2 trial assigned 48 non-treatment-seeking adults with AUD to low-dose injectable semaglutide or placebo for nine weeks. Its primary outcome was a laboratory self-administration task. Semaglutide was associated with less alcohol consumed, lower peak breath alcohol concentration, fewer drinks per drinking day, lower weekly craving, and a greater reduction in heavy drinking over time. It did not affect average drinks per calendar day or number of drinking days. A small laboratory study in people not seeking treatment cannot answer what will happen in usual care or over the long term.^[2]

A later U.S. trial tested oral semaglutide in 50 treatment-seeking adults with moderate-to-severe AUD for eight weeks. Its prespecified primary cue-elicited-craving endpoint and drinks per day were not significantly different from placebo; heavy drinking days, drinks per drinking day, and naturalistic craving were lower. The modest, mostly White non-Hispanic sample had overweight or obesity, and the study lacked a biological alcohol-consumption measure. These limits leave duration, dose, durability, and generalizability unknown.^[4]

The 26-week semaglutide-and-CBT trial is the clearest randomized signal so far

The 2026 single-center SEMALCO trial randomized 108 treatment-seeking adults with moderate-to-severe AUD and obesity—54 to once-weekly semaglutide and 54 to placebo—for 26 weeks. All were offered up to 10 standardized cognitive behavioral therapy sessions, and 88 completed the intervention. It tested a monitored regimen within behavioral care, not medication as a stand-alone replacement.^[3]

Heavy drinking days declined 41.1 percentage points from baseline with semaglutide and 26.4 with placebo; the estimated difference was −13.7 percentage points (95% CI, −22.0 to −5.4; p=0.0015). Gastrointestinal symptoms were more common with semaglutide, and five participants discontinued because of side effects: four in the semaglutide group and one in the placebo group. The trial required BMI of at least 30 and excluded people with diabetes, pancreatitis, alcohol-withdrawal seizures, current AUD medicine use, and other conditions. It cannot establish results without obesity, years of safety, or durability after stopping treatment.^[3]

Observational studies are a signal, not proof that GLP-1 treatment caused the outcome

Two large health-record studies support the reason randomized trials were pursued, but neither measures drinking as trials do. Denmark’s study followed 38,454 new GLP-1 users and 49,222 new DPP-4-inhibitor users; its outcome was hospital AUD contact or purchase of alcohol-dependence medicine. In the first three months, GLP-1 use was associated with lower event risk (hazard ratio 0.46; 95% CI, 0.24–0.86).^[5]

A U.S. electronic-health-record study associated semaglutide with 50%–56% lower 12-month risk of incident and recurrent AUD diagnoses than other anti-obesity medicines among 83,825 patients with obesity, with a type 2 diabetes replication analysis. These coded-record associations do not prove less alcohol consumed, abstinence, or a medicine-caused effect. Prescribing, access, documentation, other care, and unmeasured health factors can remain different after matching.^[6]

Established care and withdrawal safety should not wait for future trials

NIAAA describes evidence-based AUD care as potentially including behavioral treatments, FDA-approved medication, and mutual-support groups. The setting and plan require professional assessment; a semaglutide research result does not choose them for a reader.^[7]

A clinician or alcohol-treatment service can assess symptoms, withdrawal risk, current medicines, and evidence-based options. NIAAA’s Alcohol Treatment Navigator is one resource. For related background, see what semaglutide is, GLP-1 real-world effectiveness, GLP-1 medications versus bariatric surgery, and treatment information.

Trial medicines and FDA-approved products are not compounded-product evidence

The trials used defined study medicines and monitored protocols. Current Wegovy labeling describes product-specific uses, contraindications, warnings, and adverse effects; it does not establish an AUD indication. Do not start, stop, substitute, or dose-adjust a GLP-1 medicine from a research summary.^[8]

FDA says compounded GLP-1 drugs are not FDA approved or reviewed for safety, effectiveness, or quality before marketing. No trial here establishes a compounded preparation’s quality, bioequivalence, dose behavior, safety, or alcohol outcomes. It should not be described as “the same as” a trial medicine or FDA-approved finished product.^[9]

Frequently asked questions