GLP-1 Drugs and Addiction: Can Ozempic Curb Alcohol, Opioids, and More?

A landmark study of 606,000 veterans found GLP-1 drugs reduced new substance use disorders by 14-25% and overdose deaths by 50%. From alcohol to opioids to cannabis, the reward-pathway science is catching up to what patients have been reporting for years.

Research Digest9 min readApril 14, 2026

GLP-1 Drugs and Addiction: Can Ozempic Curb Alcohol, Opioids, and More?

The Anecdotes That Launched a Thousand Studies

It started with patient reports. People prescribed semaglutide or tirzepatide for weight loss or diabetes began telling their doctors something unexpected: they'd lost interest in drinking. The glass of wine with dinner they used to crave just didn't appeal anymore. Smokers found themselves reaching for cigarettes less. Some reported reduced compulsive shopping and gambling urges. These weren't isolated stories. Clinicians across the country noticed the pattern, and researchers took notice. The question shifted from "is this real?" to "how does this work, and how far does it extend?" The answer, emerging from a rapid wave of preclinical and clinical studies, is that GLP-1 receptor agonists appear to modulate the same dopamine reward pathways that drive addiction — and the effect may be far broader than anyone initially expected.

The Landmark Veterans Study: 606,000 Patients, Dramatic Results

The most compelling evidence to date comes from a massive cohort study of 606,434 US veterans with type 2 diabetes, published in The BMJ in March 2026. Researchers compared veterans prescribed GLP-1 receptor agonists to those prescribed SGLT2 inhibitors (a different class of diabetes medication) over up to 3 years. The results were striking across every substance category: **Prevention of new substance use disorders:** - Alcohol use disorder: 18% reduced risk - Cannabis use disorder: 14% reduced risk - Cocaine use disorder: 20% reduced risk - Nicotine use disorder: 20% reduced risk - Opioid use disorder: 25% reduced risk - Any substance use disorder: 14% reduced risk overall **Reduction in serious harms among those with existing disorders:** - Emergency department visits: 30% fewer - Hospitalizations: 25% fewer - Overdose events: 40% fewer - Drug-related deaths: 50% fewer The numbers translate to 7 fewer new substance use disorder diagnoses and 12 fewer serious harm events per 1,000 GLP-1 users. Senior author Ziyad Al-Aly described the mechanism as "blunting that craving that pulls people toward whatever they're addicted to" — the drugs appear to "quiet the roar of addiction" similarly to how they reduce food preoccupation.

Alcohol: The Furthest Along

Alcohol use disorder has the most advanced clinical evidence. A Phase 2 randomized, double-blind trial published in JAMA Psychiatry in April 2025 tested low-dose semaglutide in 48 adults with AUD over 9 weeks. Semaglutide significantly reduced: - Laboratory alcohol consumption (medium-to-large effect size) - Drinks per drinking day - Weekly alcohol cravings - Heavy drinking days over time A secondary finding was equally striking: among smokers in the trial, semaglutide also reduced cigarettes per day — even though that wasn't what the study was testing. Separately, a massive Swedish observational study of 227,866 individuals with AUD (published in JAMA Psychiatry, January 2025) found that semaglutide users had a 36% lower risk of AUD-related hospitalization, and liraglutide users had a 28% lower risk — both outperforming officially approved AUD medications. Phase 3 trials evaluating semaglutide for AUD are now underway, and Eli Lilly's once-monthly brenipatide is in Phase 3 specifically for alcohol use disorder with approximately 1,100 participants enrolled.

Opioids: Early But Promising

The opioid evidence is earlier-stage but potentially more impactful given the ongoing overdose crisis. A small study of 20 patients with opioid use disorder found that liraglutide reduced opioid cravings by 40% over just three weeks. The effect was notable for two reasons: it appeared at the lowest dose tested, and it was most effective in the afternoon and evening — the highest-risk periods for relapse. An NIH-funded Phase 2 trial (NCT06548490) is now testing semaglutide in 200 participants with opioid use disorder across three sites, evaluating it as an add-on to existing medication-assisted treatment (buprenorphine or methadone). The primary endpoint is opioid abstinence measured by urine tests over 12 weeks. The veterans study data — showing a 25% reduced risk of new opioid use disorder and 40% fewer overdose events among GLP-1 users — provides strong population-level support for pursuing this indication aggressively.

Cannabis, Nicotine, and Beyond

The scope of GLP-1's reward-modulating effects continues to expand. **Cannabis:** A retrospective cohort study of over 105 million patient records found semaglutide was associated with a 44% lower risk of new cannabis use disorder diagnoses in the obesity cohort and a 60% lower risk in the diabetes cohort. **Nicotine:** Preclinical research shows GLP-1 receptor activation in the medial habenular pathway makes nicotine's effects aversive rather than rewarding, reducing drug-taking behavior. The semaglutide AUD trial incidentally found significant reductions in cigarettes per day among smokers. The veterans study showed 20% lower risk of new nicotine use disorder. **Cocaine:** The veterans data showed a 20% risk reduction for cocaine use disorder. Earlier preclinical work with exendin-4 (a GLP-1 agonist) demonstrated reduced cocaine self-administration and cocaine-induced dopamine release in the nucleus accumbens. **Compulsive behaviors:** Physicians have reported patients on GLP-1 medications spontaneously losing interest in compulsive shopping and gambling, though formal studies on behavioral addictions are still in early stages.

The Dopamine Reward Mechanism

The unifying explanation is that GLP-1 receptors aren't just in the gut and pancreas — they're expressed throughout the brain's reward circuitry, including the ventral tegmental area (VTA) and nucleus accumbens, the core structures driving addiction. GLP-1 receptor activation modulates reward in several ways: **Dopamine suppression:** GLP-1 agonists dampen phasic dopamine signaling in reward circuits through alpha-MSH-driven GABAergic inhibition of the VTA. This reduces the "rush" associated with addictive substances without eliminating normal pleasure. **Glutamate modulation:** GLP-1 affects presynaptic AMPA/kainate receptors, altering glutamatergic neurotransmission that reinforces reward-seeking behavior. **Top-down control:** GIP receptor activation (relevant for dual agonists like tirzepatide and brenipatide) promotes PI3K/AKT/mTOR-mediated synaptic plasticity, potentially strengthening prefrontal cortex control over impulsive behavior. This is why the effect appears substance-agnostic. GLP-1 drugs don't block a specific substance's receptor (like naltrexone blocks opioid receptors). They modulate the reward system itself — the common pathway that all addictive substances and behaviors hijack.

What This Means Right Now

As of April 2026, no GLP-1 drug is FDA-approved for any addiction indication. But the research trajectory is clear and accelerating. **What's happening now:** - Semaglutide Phase 3 trials for AUD are enrolling, with results expected late 2026 - Brenipatide Phase 3 for AUD has ~1,100 participants enrolled - Semaglutide Phase 2 for opioid use disorder is underway (200 participants) - The NIAAA is running one of the most closely watched addiction trials in a decade **What patients should know:** - Clinicians can already prescribe GLP-1s off-label for patients who also qualify for approved indications (obesity, type 2 diabetes) - Self-reported reductions in alcohol and substance cravings are common but not universal - GI side effects (nausea, vomiting) can be significant, and in the opioid liraglutide study contributed to a high dropout rate - The once-monthly dosing of brenipatide could be particularly important for addiction populations where treatment adherence is a challenge **What we don't yet know:** - Optimal dosing for addiction (may differ from metabolic indications) - Whether effects persist long-term or require ongoing treatment - How GLP-1s compare head-to-head with existing addiction medications (naltrexone, acamprosate, buprenorphine) - Whether the effect extends to all patients or specific subpopulations The convergence of a massive observational study, multiple randomized trials, robust preclinical mechanisms, and widespread patient reports makes this one of the most promising therapeutic developments in addiction medicine in decades.

Key Findings

A 606,434-patient veterans study found GLP-1 drugs reduced new substance use disorders by 14-25% and drug-related deaths by 50% compared to other diabetes medications
Low-dose semaglutide significantly reduced alcohol consumption, cravings, and heavy drinking days in a Phase 2 RCT (JAMA Psychiatry, 2025)
Semaglutide users had 36% lower AUD hospitalization risk in a 227,866-patient Swedish observational study — outperforming approved AUD medications
Liraglutide reduced opioid cravings by 40% in a small study, with effects appearing at the lowest dose and during high-relapse evening hours
Semaglutide was associated with 44-60% lower risk of cannabis use disorder diagnoses across obesity and diabetes cohorts (105 million patient records)
The mechanism is substance-agnostic: GLP-1 receptors in the brain's reward circuitry modulate dopamine signaling, reducing cravings regardless of the specific substance

Limitations & Caveats

The veterans study is observational — it cannot prove GLP-1 drugs caused the reductions, only that they were associated with them
The semaglutide AUD randomized trial was small (48 participants) and short (9 weeks) — Phase 3 results are pending
Most data comes from patients prescribed GLP-1s for metabolic conditions, not addiction — there may be selection bias
Optimal addiction-specific dosing, duration, and patient selection criteria are unknown
No GLP-1 drug is FDA-approved for any addiction indication — all use for addiction is off-label
Long-term effects on reward processing and whether benefits persist after discontinuation remain unstudied