A Weight-Loss Drug Is Beating Alcoholism's Two Boring Pills, and the VA Just Bet On It

Ozempic’s maker isn’t running this trial. The Department of Veterans Affairs is, with its own money and its own protocol, dosing 600 veterans at 18 medical centers with a drug that has zero FDA approval for what they’re testing it on.

That’s the sentence worth sitting with. Recruitment opened July 28 for CRAVE, a Phase 3 trial putting semaglutide — a GLP-1, a hormone-mimicking drug class built for diabetes and weight loss that also happens to turn down the brain’s reward circuit — head-to-head against placebo in veterans with moderate-to-severe alcohol use disorder. Weekly injections, 24 weeks, ages 18 to 80.

A drug built to shrink stomachs is turning out to shrink the urge to drink, and addiction medicine hasn’t had news this good in decades.

The VA isn’t chasing a hunch. In April, The Lancet published a randomized controlled trial — the gold-standard, placebo-versus-drug design — out of Copenhagen University Hospital: 108 people with obesity and alcohol use disorder, split into semaglutide-plus-therapy or placebo-plus-therapy, for 26 weeks. The semaglutide group cut heavy-drinking days 41 percentage points from baseline; placebo only got people 26 points. That 13.7-point gap held up statistically (p=0.0015), and it wasn’t just people self-reporting fewer drinks — blood biomarkers for recent alcohol use dropped right along with it. Lead researcher Anders Fink-Jensen called it “strong therapeutic benefits” in patients who’d already sought treatment and were still stuck.

Run the number the FDA actually cares about — how many people you’d need to dose before one extra person gets meaningfully better — and semaglutide landed at 4.3. Naltrexone and acamprosate, the only two drugs approved for alcohol use disorder since the 1990s, run 7 or worse, and barely anyone gets offered them: NIAAA director George Koob says plainly that the existing drugs “are vastly underutilized.” NIDA’s Nora Volkow went further, calling this the moment GLP-1s’ addiction potential is turning “into reality.”

The mechanism is not mysterious. Alcohol works by flooding the reward pathway, the dopamine circuit that fires when something feels good and pushes you to repeat it. GLP-1s appear to turn that same circuit down, which is why people already taking Ozempic or Wegovy for diabetes or weight loss show up in meta-analyses with lower rates of alcohol and substance use disorder than people on other diabetes drugs.

None of that makes semaglutide an alcohol drug today. It carries no FDA approval for any addiction use, the VA trial won’t read out for years, and a weekly injection running several hundred dollars a month, with insurers still fighting over weight-loss coverage, will not reach the person who needs it most on a Tuesday night. A trial is not a treatment, and Rize isn’t telling anyone to ask their doctor about Ozempic for their drinking. But thirty years of exactly two underused pills is a low bar, and a drug that clears it by that much on a real placebo comparison doesn’t get to sit in the “promising” pile. It gets to be the reason the next five years of alcohol treatment look different than the last thirty.