The Medication That Could Help 14.5 Million People — and That Fewer Than 5% of Them Ever Receive

Jillian is 38 and lives outside Cleveland. She started drinking heavily in her late twenties — a glass of wine with dinner that became a bottle, then the second bottle she hid in the bathroom cabinet. By 35 she had tried Alcoholics Anonymous three times. Each time she relapsed within months. The meetings helped her feel less alone, she said, but they didn’t address what happened in the hours between midnight and 3 a.m. when the craving arrived like weather — specific, physical, not a choice. No one at those meetings talked about medication. Her primary care physician, who knew about her drinking, prescribed a referral to an outpatient program. The outpatient program offered group therapy. Jillian eventually found a psychiatrist who mentioned naltrexone almost as an aside. She’s been taking it for two years. She drinks occasionally now, not compulsively.

She is not an outlier. She is statistically typical of how people with alcohol use disorder eventually find their way to a medication that has been FDA-approved since 1994, if they find their way to it at all.

The number that should make everyone angry

Alcohol use disorder affects approximately 14.5 million Americans. It kills roughly 178,000 of them every year — more than opioids in many years, more than car accidents, more than most diseases that generate their own awareness campaigns and charitable foundations. The toll is almost completely invisible because it accumulates in ones and twos across hospital wards and emergency departments and quiet deaths at home that get coded as liver failure, or heart disease, or falls.

Of those 14.5 million Americans, fewer than 8% receive any treatment in a given year. Of those who do receive treatment, only about 2% receive an FDA-approved medication. Naltrexone has been available since 1994. Acamprosate since 2004. Disulfiram since the 1950s. The body of evidence supporting naltrexone, specifically, for reducing heavy drinking days and preventing relapse is robust — multiple randomized controlled trials, consistent effect sizes, a mechanism that is well understood.

Fewer than 3% of primary care physicians who treat patients with AUD prescribe naltrexone for it. A 2022 survey published in PubMed found that many physicians simply didn’t know the drug was indicated, or believed patients needed to be enrolled in a dedicated addiction program before receiving a prescription. Some had confused the contraindications — naltrexone cannot be prescribed to someone actively dependent on opioids, because it precipitates acute withdrawal — with a broader clinical caution that doesn’t apply to AUD without comorbid opioid use.

The gap between what medicine knows and what medicine does is not abstract here. It is measured in 178,000 deaths a year.

What cravings are, and why medication matters

The behavioral model of alcohol use disorder — the model that underlies most AA-based and traditional counseling approaches — treats craving as a psychological event that willpower and social support can overcome. The neurobiological model treats it as a brain state produced by specific receptor activity, as legible and as involuntary as hunger. Both models have clinical utility. The problem is that the behavioral model has historically crowded out the neurobiological one in treatment settings, not because the evidence supports that prioritization, but because of history, culture, and the specific stigma that attaches to medication for addiction.

Alcohol activates the brain’s reward system through multiple pathways — primarily by increasing dopamine release in the nucleus accumbens and by binding to GABA receptors in a way that produces sedation and disinhibition. Over time, with heavy use, the brain downregulates dopamine production in response to alcohol and upregulates excitatory glutamate signaling to compensate. The result is a brain that is dysregulated at baseline — producing less reward, more anxiety, more craving — and that registers alcohol not primarily as pleasurable but as necessary to feel normal. Withdrawal from this state is medically serious. Seizures and delirium tremens are not metaphors.

Over time, with heavy use, the brain downregulates dopamine production in response to alcohol and upregulates excitatory glutamate signaling to compensate.

Naltrexone works by blocking the opioid receptors that mediate part of alcohol’s rewarding effect — specifically the endorphin release that contributes to the “buzz.” When someone on naltrexone drinks, the neurobiological reward is blunted. The drink doesn’t hit the same way. Over time, this changes the learned association: the cue that previously triggered strong craving produces a weaker signal. This is not willpower. It is pharmacology working on the same circuitry that created the dependence.

Acamprosate stabilizes the glutamate system that alcohol’s chronic suppression has left in a hyperexcitable state, reducing the anxiety and discomfort of early recovery that drive relapse. Disulfiram creates an aversive reaction if alcohol is consumed. These are different mechanisms, different patients, different moments in the recovery trajectory — which is the clinical argument for having all three available and prescribed.

Why primary care doesn’t prescribe them

The answer has layers. The practical one first: a primary care appointment is 15 to 20 minutes. Starting a patient on naltrexone for AUD requires screening for opioid dependence (the contraindication), reviewing liver function, discussing expectations and common side effects, arranging follow-up, and having a frank conversation about the treatment model. Most PCPs have not done this before. There is no established workflow for it in most practices, no staff support, no prompts in the electronic health record. The path of least resistance is to refer to a specialist or a program — and most programs don’t use medication either.

The cultural layer is the one that nobody talks about in clinical settings: stigma. Not the patient’s stigma, though that exists too, but the physician’s. A survey of primary care physicians on naltrexone prescribing practices found that a significant minority believed medications for addiction were somehow less legitimate than the “real” behavioral work of recovery. This belief has no scientific basis. It is the residue of decades of treatment culture that positioned medications as a crutch, recovery as a spiritual process, and medication-assisted treatment as a sign of insufficient commitment to sobriety.

That cultural frame is most embedded in the twelve-step world and in treatment programs that grew up around twelve-step principles. It has migrated into primary care in exactly the way that implicit assumptions about addiction as a character problem have migrated into medicine broadly: quietly, below the level of explicit policy, resistant to the obvious clinical counter-evidence.

The GLP-1 moment

The May 2026 Lancet trial of semaglutide for AUD with comorbid obesity represents a different kind of opening than another naltrexone awareness campaign would.

Semaglutide is culturally legible in a way that naltrexone isn’t. Patients know about it. Their doctors know about it. When the mechanism of action for reducing heavy drinking is framed as “GLP-1 receptors in the brain’s reward circuitry” rather than “you need addiction medicine,” something shifts in the conversation. The drug has cultural permission that naltrexone was never given, in part because nobody involved in its development was trying to navigate the stigma of addiction treatment.

This is not a small thing. The health communication argument for GLP-1s in AUD is potentially as important as the clinical argument — if patients start asking their PCPs about semaglutide for drinking, primary care physicians may engage with AUD pharmacotherapy in a way the last three decades of naltrexone advocacy have not managed to produce.

The trial’s limitations are real. Only patients with comorbid obesity. Only 26 weeks. Only with concurrent CBT. Those constraints matter enormously and shouldn’t be glossed over in the enthusiasm. But the signal is there, published in the most prominent medical journal in the world, generating mainstream coverage, reaching people who wouldn’t have searched for “naltrexone AUD.”

What would actually change things

Jillian’s path to naltrexone required a level of persistence and navigation sophistication that most people in active alcohol use disorder don’t have. She had a stable job, insurance, a psychiatrist she could afford. She happened to see a provider who knew what to ask about. None of these things should be prerequisites for accessing a medication that has been available at most pharmacies for thirty-two years.

She happened to see a provider who knew what to ask about.

The structural changes that would actually move the treatment gap are straightforward to describe: train primary care physicians on AUD pharmacotherapy, build the prescribing workflow into standard practice, remove the implicit stigma that positions medication as a concession rather than a clinical tool. Add telehealth pathways that make initiating naltrexone as easy as initiating an antidepressant. Stop referring people with AUD exclusively to programs that don’t use medication.

The barrier to these changes isn’t knowledge. It’s culture and will. The knowledge has been available since 1994. The culture hasn’t caught up.

---

Related: The May 2026 Lancet trial of semaglutide for AUD, explained · Find treatment options in Arizona at Rize Recovery