Ozempic Cut Alcohol Consumption by 70 Percent in the First Rigorous Clinical Trial

Semaglutide — the drug sold as Ozempic and Wegovy that reshaped the conversation about obesity medicine — just completed its first rigorous randomized controlled trial for alcohol use disorder. The results, published May 2 in The Lancet, are the kind that make researchers double-check their data and then quietly start changing their clinical practice before the FDA gets around to it.

The study enrolled 108 treatment-seeking patients who had both alcohol use disorder and obesity (BMI ≥30). Half received weekly subcutaneous semaglutide for 26 weeks; half received placebo. Both groups participated in standard cognitive behavioral therapy. The primary endpoint was total alcohol consumption, measured across the trial period. By week 26, patients in the semaglutide group had reduced their total monthly alcohol consumption by roughly 70 percent from baseline. The placebo group also improved — CBT alone is not nothing — but the reduction in the semaglutide arm was “significantly larger,” as the authors put it, across every measure: drinks per drinking day, frequency of heavy drinking days, self-reported craving, and validated measures of alcohol-related harm.

Four serious adverse events were reported across the whole trial — one in the semaglutide group, three in placebo. All resolved. The safety profile was consistent with what has been observed in obesity and diabetes trials: primarily GI side effects (nausea, vomiting) that tend to diminish as the dose escalates.

Why this was expected, and why the evidence still matters

The hypothesis that GLP-1 receptor agonists might reduce alcohol consumption is not new. It comes from animal studies going back more than a decade showing that GLP-1 receptors are expressed not just in the gut and pancreas — where they regulate appetite and glucose — but in the brain’s reward circuitry, specifically the nucleus accumbens and ventral tegmental area. These are the structures where substances produce the dopamine surges that drive compulsive use. In rodent models, semaglutide and similar drugs reduced alcohol self-administration significantly. The leap from rodent to human is not always clean, but in this case the mechanism made biological sense: if you dull the reward signal in the circuits that alcohol lights up, you reduce the pull toward drinking.

What had been missing was the human evidence at clinical standards. GLP-1 drugs had been prescribed off-label for alcohol use disorder by physicians paying attention to their patients’ self-reports — “I just don’t want to drink as much,” patients on Ozempic said, unprompted, in one diabetes clinic after another. Anecdote is not data. This trial is data.

The NIH has separately reported that adding weekly GLP-1 to CBT further reduces heavy drinking relative to CBT alone, corroborating the Lancet findings from a different research group. A target trial emulation published as a preprint on medRxiv in June 2025 found that real-world patients with AUD who initiated a GLP-1 receptor agonist had fewer alcohol-related hospitalizations compared to matched patients who initiated naltrexone or acamprosate. That’s not a randomized trial, but real-world comparative effectiveness data from a large health system is a meaningful second signal.

What the study cannot tell you

The Lancet trial has three limitations worth naming directly, because they determine who this research currently applies to.

First, all participants had a BMI of 30 or higher. The GLP-1 mechanism may well work in people with AUD who don’t have obesity — the reward-circuitry pathway doesn’t require excess adiposity — but this trial can’t confirm that. Alcohol use disorder is distributed across body types, and the 50 percent of people with AUD who do not have obesity are not yet part of this evidence base.

Second, the study population was predominantly white. Given what is known about metabolic differences in GLP-1 drug response across racial groups, and given that alcohol mortality disproportionately affects American Indian and Alaska Native populations (at five times the rate of white Americans), the applicability of these findings to the populations most burdened by alcohol-related death is uncertain.

Third, the trial followed patients for 26 weeks. Alcohol use disorder is a chronic condition. We don’t know whether the gains at 26 weeks are sustained, whether patients require ongoing semaglutide to maintain them, or what happens to drinking behavior if the drug is discontinued.

What it means now, while the FDA takes its time

No semaglutide product is FDA-approved for alcohol use disorder. The FDA approval process for a new indication takes years. Off-label prescribing, however, is legal and common in addiction medicine — and based on the Lancet data, it is not unreasonable for a clinician managing a patient with both AUD and obesity to discuss semaglutide as part of a comprehensive treatment plan, particularly when first-line medications (naltrexone, acamprosate, disulfiram) have failed or are poorly tolerated.

The more significant medium-term implication is for how treatment systems think about the overlap between metabolic disease and substance use disorder. For years, SUD treatment and obesity medicine operated in separate clinical siloes. The GLP-1 evidence — not just in alcohol, but in preliminary data for opioids, stimulants, and nicotine — is forcing a reckoning with the shared neurobiology. What if the reason so many people with SUD struggle with weight is not purely lifestyle, but partially that the same reward dysregulation that drives compulsive substance use also drives compulsive eating? And what if addressing the neurobiological substrate helps with both?

The GLP-1 evidence — not just in alcohol, but in preliminary data for opioids, stimulants, and nicotine — is forcing a reckoning with the shared neurobiology.

That’s still a research question, not a clinical protocol. But for treatment programs trying to figure out what a holistic approach looks like in 2026, the Lancet trial is a data point that belongs in the conversation.

For context on Arizona treatment facilities offering evidence-based AUD therapies, see Rize Recovery’s substance guide for alcohol.