The national overdose data tells a story that keeps getting cited in the wrong direction. Drug overdose deaths fell 26 percent from 2023 to 2024 — the largest single-year decline on record. The number everyone uses: 79,384 people died. That is a genuine improvement. It is also roughly the same as the number of Americans who died from alcohol-related causes in that same year. About 105,400 people died from alcohol-induced causes in 2022 — up roughly 30 percent from pre-pandemic levels — and the trend has not reversed. Alcohol kills, year in and year out, approximately as many Americans as all illicit drugs combined. You wouldn’t know it from the policy attention.
What compounds the disparity is the treatment gap. Among all the substances that destroy health and end lives in the United States, alcohol may have the largest gap between the number of people who need care and the number who get it. Approximately 28.9 million Americans met the diagnostic criteria for alcohol use disorder in 2023, according to SAMHSA. About 7 percent of them received any treatment. That is 2 million people. The other 26.9 million were on their own.
This piece is a state-of-the-field dispatch on alcohol use disorder treatment in mid-2026: what the existing options are, what a landmark new clinical trial found, what the pipeline holds, and why the gap persists even as the science improves.
Three drugs. Thirty years. And why most prescribers don’t use them.
The FDA has approved three medications for alcohol use disorder: naltrexone (oral and extended-release injectable), acamprosate, and disulfiram. Naltrexone was first approved in 1994. Acamprosate followed in 2004. None of the three were developed in the last two decades.
Efficacy comparisons are imprecise because trials use different endpoints, but the standard metric — number needed to treat (NNT), the number of patients you’d need to treat to get one additional positive outcome compared to placebo — is roughly seven to eight for naltrexone and nine to twelve for acamprosate. These are real effects, meaningful in clinical terms, worth prescribing. But they are modest effects. For every ten people started on naltrexone, you’d expect one or two additional people to reduce their drinking substantially, compared to placebo. Many patients don’t respond. Many others stop taking the medication before it has a chance to work.
Disulfiram (Antabuse) is a different animal — it doesn’t reduce craving or rewire reward circuitry, it creates a physical aversion reaction to alcohol consumption. It works if patients take it. The compliance problem is severe. Many patients, by the time disulfiram seems like a good idea, are not in a place where they reliably take medication.
The deeper problem: most physicians don’t prescribe any of them. In primary care settings — where people with AUD are most likely to have any provider contact — estimates suggest fewer than 20 percent of physicians routinely offer medication for alcohol use disorder. Surveys of prescribers cite multiple reasons: inadequate training, discomfort managing addiction, belief that patients won’t take the medications, and lack of time. The result is that the medications exist, they work, they are covered by most insurance, and they reach a small fraction of the people who need them. The treatment gap is not primarily a scientific failure. It is a delivery failure.
The Lancet study: a better number
On May 2, 2026, The Lancet published the first large randomized controlled trial of once-weekly semaglutide — the active ingredient in Ozempic and Wegovy — in people seeking treatment for alcohol use disorder who also had obesity. The study was led by Dr. Anders Fink-Jensen at Copenhagen University Hospital, in collaboration with NIH researchers, and enrolled 108 participants who received either semaglutide or placebo for 26 weeks, alongside standard cognitive behavioral therapy.
The headline finding: semaglutide produced significantly larger decreases in heavy drinking days, total monthly alcohol consumption, drinks per drinking day, self-reported craving, and measures of harmful alcohol use. The effect was not subtle. The NNT for semaglutide in this trial was 4.3 — roughly twice as good as the NNT for naltrexone or acamprosate.
The NNT for semaglutide in this trial was 4.3 — roughly twice as good as the NNT for naltrexone or acamprosate.
“We’re beginning to see some of that potential for GLP-1s to treat drug addiction turn into reality,” Dr. Nora Volkow, outgoing Director of the National Institute on Drug Abuse, told NIH reporters after the study’s publication.
The biological mechanism is credible. GLP-1 receptors are expressed in the brain’s reward and motivation circuitry — specifically in the ventral tegmental area and nucleus accumbens, the same structures that amplify craving signals across most substances. Animal studies have shown that GLP-1 agonists reduce alcohol consumption in rodent models. The Fink-Jensen trial is the most rigorous human evidence to date that this mechanism translates into clinical benefit. It won’t be the last study. But it’s the one that changes the conversation.
What’s in the pipeline
Two other candidates are in active clinical trials for AUD, and both reflect the same therapeutic hypothesis: that a new generation of metabolic medicines is likely to have psychiatric applications that the original obesity indication didn’t predict.
Pemvidutide, developed by Altimmune, is a dual GLP-1/glucagon receptor agonist — the glucagon component adds a specific hepatic benefit (reducing liver fat, which is a major comorbidity in people with severe AUD) on top of the GLP-1 craving-reduction effect. The FDA granted pemvidutide Fast Track designation for AUD in May 2025. The RECLAIM trial (NCT06987513) is currently enrolling roughly 100 patients in a Phase 2 randomized controlled trial of 24 weeks. If the trial reads out in late 2026 as currently projected, it could accelerate a Phase 3 filing.
Brenipatide, a Eli Lilly compound, is further along. A Phase 3 trial (NCT07219966) is now enrolling participants with moderate-to-severe AUD, with last update on ClinicalTrials.gov dated June 3, 2026. Eli Lilly’s Phase 3 commitment represents a substantial investment — and a signal that at least one major pharmaceutical company believes the AUD indication is commercially viable, not just clinically interesting.
The insurance wall
Here is where the pipeline hits a structural problem that the science cannot solve on its own.
Semaglutide is not approved for AUD. It is approved for obesity (Wegovy) and type 2 diabetes (Ozempic). An off-label prescription is legal, and many addiction medicine specialists will write one. Insurance coverage is a different matter. Plans that cover semaglutide for obesity frequently do not cover it for AUD — because there is no FDA-approved indication, no specific coverage code, and no established prior authorization pathway. A patient whose physician is willing to prescribe semaglutide for alcohol use disorder may face a $1,200 per month out-of-pocket cost without coverage. That cost is, for most people with AUD, prohibitive.
This gap will narrow only with FDA approval for the indication. Pemvidutide and brenipatide, if they succeed in trials, would seek approval specifically for AUD — which would establish a coverage pathway. Semaglutide could also be studied in a Phase 3 trial sufficient to support an AUD indication; whether Novo Nordisk invests in that trial, given how strong the obesity revenues already are, remains an open question.
Why the gap persists regardless of the science
The 93 percent who never get treatment for AUD are not a random sample of people with the disorder. They are more likely to be people who do not see their relationship with alcohol as a medical problem, more likely to be people who have tried to cut back before and couldn’t, more likely to be people who feel shame about how much they drink, and more likely to be people who lack the insurance, transportation, and time to access a specialty clinic.
These are not barriers that GLP-1 receptor agonists, on their own, overcome. A monthly injection that requires an obesity or diabetes diagnosis to get covered doesn’t reach the person who drinks a bottle of wine every night for twenty years and has never mentioned it to a doctor. A Phase 3 trial doesn’t help someone who lives in a county with no addiction medicine specialist.
A Phase 3 trial doesn’t help someone who lives in a county with no addiction medicine specialist.
What could close the gap, at least partially: primary care physicians who screen for alcohol use disorder routinely (they currently don’t), pharmacists authorized to prescribe naltrexone directly (currently happening in a handful of states), telehealth models that bring AUD treatment into people’s homes rather than waiting for them to walk through a clinic door, and — eventually — GLP-1 approvals for AUD that create a clear insurance pathway for the most efficacious treatment option the field has ever had.
The science is ahead of the delivery system. That is the central fact of AUD treatment in 2026. The Lancet data gives providers a better tool. Getting that tool to the 26 million people who need it and aren’t getting anything — that’s the part of the problem that no trial solves.
Browse the Rize substance center for alcohol for treatment resources and facility matching.
Filed Under
treatmenttrendsscienceAcamprosateThe Treatment Gap
Keep up with the reporting.
One email each morning with the stories that put days like this in context.