The Ozempic-for-Drinking Trial Just Published Real Numbers. The Headline Effect Missed Its Own Primary Target.

On July 28, the Department of Veterans Affairs began recruiting patients at multiple VA medical centers for a new trial of GLP-1 medication for alcohol use disorder — the agency’s own announcement frames it as chasing one of the most promising leads in addiction pharmacology in a decade. One day later, the trial that lead is actually built on published its full numbers, and the numbers are messier and more interesting than the momentum around them suggests.

The study, out of the University of Colorado Anschutz Medical Campus and published July 29 in the American Journal of Psychiatry, was a Phase 2, randomized, double-blind, placebo-controlled trial of oral semaglutide — the same drug class behind Ozempic and Wegovy, in pill rather than injectable form — in 50 treatment-seeking adults with moderate-to-severe alcohol use disorder. Participants got an escalating dose (3mg daily for four weeks, then 7mg daily for four more) or a placebo, over eight weeks total, while researchers tracked both a controlled lab measure (how strongly the sight and smell of alcohol triggered craving in a clinic setting) and real-world drinking behavior logged day by day.

The drug worked on real life and missed its own lab test — which tells you the effect is more about drinking behavior than about turning down a craving switch in the brain.

What actually moved

Compared to placebo, participants on oral semaglutide had significantly fewer heavy drinking days, drank less per drinking day, reported lower craving in their day-to-day environment (not the lab), fewer cannabis use days, and fewer alcohol-related consequences overall — the kind of composite measure that captures things like missed work, arguments, and health scares tied to drinking. CU Anschutz’s own summary and independent coverage from The American Journal of Managed Care both frame this as a genuine positive signal, consistent with a growing body of GLP-1 research spanning multiple substances.

What didn’t move, and why that’s the finding worth reading twice

Here’s the part that got buried in most of the coverage: the trial’s pre-registered primary endpoint — craving measured in a controlled cue-reactivity lab session, where researchers show participants alcohol-related cues and measure the physiological and self-reported urge to drink — did not reach statistical significance. Neither did drinks-per-calendar-day, a stricter measure than the “drinking days” metric that did hit significance. Only the secondary, naturalistic measures — logged in participants’ real environments over the following weeks — showed a clear effect, as reported by News-Medical.

That’s not a footnote. It’s a hint about the mechanism. A drug that fails to blunt craving in a lab but succeeds at reducing actual drinking in daily life isn’t necessarily quieting a neural craving circuit the way naltrexone or acamprosate are designed to. It may be doing something closer to what GLP-1 drugs do in the gut and appetite system more broadly — changing satiety, reward timing, or how rewarding a drink feels once it’s already in hand, rather than how badly someone wants one before they’ve started. The researchers themselves flagged this, cautioning that endpoint selection and the specific measurement window mattered enormously to what showed up as significant, and calling for larger, longer confirmatory trials before anyone treats this as settled pharmacology.

The caveat that has to travel with every headline about this drug

Semaglutide is not FDA-approved for alcohol use disorder. This trial, like the VA’s new one, is investigational — a research protocol, not a prescribing pathway, and a doctor cannot currently write “semaglutide for AUD” the way they can write naltrexone or acamprosate. The sample size here was 50 people at a single academic center; that’s enough to justify the VA scaling up a bigger trial, and nowhere near enough to justify a patient assuming their own Ozempic prescription (written for diabetes or weight loss) is doing double duty on their drinking. If you’re taking a GLP-1 drug for another condition and have also noticed you’re drinking less, that’s a real and increasingly well-documented phenomenon — but it is not yet a treatment plan, and nobody should stop an existing, FDA-approved AUD medication to wait for one.

What this study actually earns is more specific and more useful than a headline: a plausible, still-unproven secondary pathway by which a drug built for something else might change drinking behavior without touching the craving circuitry addiction medicine has spent decades trying to target directly. The VA is now the best-positioned institution in the country to find out whether that pathway holds up outside a 50-person study — and whether it holds up specifically in veterans, a population with alcohol use disorder rates and treatment-access barriers distinct from the general population this first trial recruited from.