The Same Drug Class Behind Ozempic Is Now the Most Promising Alcohol Medication in 20 Years

Only three medications have ever been FDA-approved to treat alcohol use disorder, and the newest of them — the injectable form of naltrexone — has been on the market since 2006. If you’ve been in treatment any time in the last two decades, that’s the entire pharmacological toolkit you were offered: disulfiram, which makes you violently sick if you drink; naltrexone, which blocks the opioid receptors that make drinking feel rewarding; and acamprosate, which calms an overstimulated brain during early abstinence. Three different theories of what alcohol does to the brain, developed across more than fifty years, and nothing new since.

A drug class built to treat diabetes and obesity is turning out to work on the brain circuitry that drives alcohol use — and four separate trials are now testing it at once, faster than any new AUD mechanism has moved in a generation.

GLP-1 drugs — glucagon-like peptide-1 receptor agonists, the same class behind Ozempic and Wegovy — were designed to slow digestion and blunt appetite for people with diabetes or obesity. What researchers noticed, almost as a side effect, is that patients on these drugs reported drinking less and wanting alcohol less, independent of any weight change. That anecdotal pattern is now the basis of the most active pharmacological research pipeline addiction medicine has seen in years.

Four trials, one theory: alcohol craving runs through the same reward circuit as hunger

The VA’s CRAVE trial — “Cessation or Reduction of Alcohol consumption in Veterans” — began recruiting July 28 across 18 VA medical centers nationwide, a randomized, double-blind, placebo-controlled Phase 3 study testing weekly semaglutide injections in more than 600 veterans ages 18 to 80 with moderate-to-severe AUD, run over 24 weeks plus follow-up. It’s one of the largest AUD medication trials ever launched, and it’s happening inside the health system that treats a population with disproportionately high rates of alcohol use disorder tied to combat trauma and chronic pain.

UW Medicine in Seattle is running a parallel but distinct Phase 3 trial — one of 30 sites nationwide, the only one in the Pacific Northwest — testing a different GLP-1/GIP investigational drug called brenipatide over 14 months. Principal investigator Dr. Mark Duncan, an addiction psychiatrist, and co-investigator Mary Hatch describe the drug’s target in plain terms: the mesolimbic pathway, which is the brain’s core “this feels good, do it again” reward circuit — the same circuit both food cravings and alcohol cravings run through, which is the entire reason a diabetes drug turned out to have anything to do with drinking at all.

A smaller, already-completed pilot at the University of Colorado Anschutz gives an early read on what these larger trials might find: in 50 adults with moderate-to-severe AUD taking oral semaglutide, researchers saw fewer heavy-drinking days, lower alcohol consumption per drinking occasion, reduced daily cravings, and fewer alcohol-related problems, with high adherence and only mild side effects. Lead researcher Dr. Joseph Schnacht called the results promising for a new option “particularly for those who have not benefited from existing medications” — a real gap, since a substantial share of people who try naltrexone or acamprosate don’t respond to either. Participants in that trial also reported drinking less and, notably, using cannabis less, hinting the effect may not be alcohol-specific at all.

The most mechanistically different entry is Newleos Therapeutics’ NTX-2001, which just dosed its first participant in a Phase 1b trial run in collaboration with Yale School of Medicine. It isn’t a GLP-1 drug at all — it’s a TAAR1 partial agonist, meaning it acts on a different receptor system (trace amine-associated receptor 1) than any of the other three approaches, previously tested for safety in roughly 645 people across eight trials for schizophrenia. It’s the first time this drug class has been tested for AUD in humans, which means the current wave of research isn’t just “GLP-1s might work” — it’s multiple genuinely different mechanisms being tested against alcohol use disorder simultaneously, for the first time since the naltrexone era.

The most mechanistically different entry is Newleos Therapeutics’ NTX-2001, which just dosed its first participant in a Phase 1b trial run in collaboration with Yale School of Medicine.

Fifty years of asking what kind of problem drinking is

Drug development for alcohol use disorder has always encoded a theory of what addiction actually is. Disulfiram, approved in the early 1950s, treated it as a behavior to be punished — drink, and the drug makes you sick, full stop. Naltrexone, approved for alcohol in 1994, treated it as a reward-circuit problem — block the receptors that make drinking pleasurable, and the behavior loses its reinforcement. Acamprosate, approved in 2004, treated it as a chemical-imbalance problem — calm the glutamate overactivity that early sobriety produces, and cravings ease. Each drug reflected the neuroscience of its decade. GLP-1s reflect this one: alcohol use disorder, in this framing, isn’t a uniquely alcohol-shaped problem at all. It’s downstream of the same broad reward-and-craving machinery that governs hunger, and a drug that recalibrates that machinery for food might recalibrate it for alcohol too.

That reframing matters for anyone who has been told naltrexone or acamprosate “didn’t work for them” and concluded the problem was willpower rather than the wrong tool for their particular brain chemistry. It wasn’t. There were only ever three tools, and now there may reasonably be more within the next few years — not a moral failure, a limited toolbox.

The part these press releases don’t mention

None of this is complicated for the people who need it to be simple: promising trial data is not the same as a drug sitting on a pharmacy shelf with your insurance covering it. Semaglutide and its cousins already carry a defining fact into any future AUD approval fight — insurers restrict access to these drugs even for their existing, FDA-approved indications like diabetes and obesity, citing cost. There is no reason to assume a new off-label or newly-approved AUD indication gets a warmer reception from the same payers, and every reason, given how AUD treatment has historically been underfunded and under-covered compared to other chronic conditions, to expect a fight over coverage even if every trial reads out positive. A 340,000-person UK Biobank study presented through the American College of Cardiology this year found heavy drinkers face 24% higher all-cause mortality and 36% higher cancer mortality than light or non-drinkers — the underlying disease burden a GLP-1 for AUD would need to justify covering is not small or speculative. The clinical case for treating alcohol use disorder aggressively already exists. What’s uncertain is whether the payment system will treat a repurposed diabetes drug as addiction medicine or as a lifestyle drug the same way it currently treats Wegovy for people who aren’t diabetic.

If you’re in treatment right now and naltrexone or acamprosate hasn’t worked for you, none of these four trials will reach a pharmacy this year. But for the first time in a long time, the answer to “is there anything else” is actually yes — and the fight over who gets to take it starts well before the FDA ever rules.