GLP-1 Diabetes Drugs Linked to Lower Addiction Risk in 600,000-Veteran Study

Medications prescribed for diabetes and weight loss showed unexpected benefits in reducing addiction risk and improving outcomes for people with substance use disorders, according to research published this week in the BMJ examining more than 600,000 U.S. veterans.

The study found that patients prescribed GLP-1 receptor agonists — including semaglutide (Ozempic, Wegovy), liraglutide, and tirzepatide (Mounjaro) — had a 14% lower risk of developing new substance use disorders compared to those taking a different class of diabetes medication. Among veterans who already had addiction diagnoses, those on GLP-1 drugs experienced 50% fewer deaths and 39% fewer overdoses over a three-year follow-up period.

The medications, which mimic a hormone that regulates blood sugar and appetite, are not approved by the FDA for addiction treatment. All use for substance use disorders remains off-label, undertaken at physician discretion without the regulatory oversight or evidentiary standards required for approved therapies.

Five Substances, Consistent Signals

Researchers at the U.S. Department of Veterans Affairs analyzed electronic health records spanning January 2021 through December 2023, comparing outcomes between patients newly prescribed GLP-1 drugs and those started on SGLT2 inhibitors — another established diabetes medication class that served as the comparison group.

Among the more than 600,000 veterans without prior addiction diagnoses, those taking GLP-1s showed reduced risk across every substance category examined. The risk reductions varied by substance: 25% lower for opioids, 20% lower for nicotine and cocaine, 18% lower for alcohol, and 14% lower for cannabis. These translated to one to six fewer diagnoses per 1,000 people over three years — small absolute numbers, but statistically significant patterns across multiple substances and large populations.

The research team employed target trial emulation, a methodology that structures observational data to approximate the design of a randomized controlled trial. Researchers matched patients on factors including age, weight, other health conditions, and prior substance use, attempting to isolate the effect of the medication itself.

While this approach represents current best practice for observational studies, it cannot eliminate confounding variables the way random assignment does. The findings demonstrate association — GLP-1 use correlated with better addiction outcomes — but cannot definitively prove causation. Only randomized trials, which are currently underway, can establish whether the medications directly produce these benefits or whether some unmeasured factor explains the patterns.

For the subset of veterans who entered the study with existing substance use disorder diagnoses, the differences were more pronounced. Those prescribed GLP-1 medications experienced 31% fewer emergency department visits related to their addiction, 26% fewer hospital admissions, 25% fewer incidents of suicidal thoughts or attempts, 39% fewer overdoses, and 50% fewer deaths. These outcomes translated to approximately one to ten fewer adverse events per 1,000 people over three years.

The consistency across outcomes — from relatively common events like emergency visits to rarer but more severe outcomes like death — strengthens the statistical signal, even as causality remains unproven.

Brain Chemistry, Not Just Metabolism

GLP-1 receptor agonists were developed to manage blood sugar by mimicking glucagon-like peptide-1, a hormone released in the gut after eating. The medications slow gastric emptying, signal satiety to the brain, and stimulate insulin secretion when glucose levels rise. Their effectiveness in treating type 2 diabetes and obesity is well-documented, supported by extensive clinical trial data and regulatory approval.

What has surprised researchers is the accumulating evidence that these metabolic drugs also appear to modulate brain reward pathways involved in addiction. Preclinical studies in rodent models showed that GLP-1 receptor agonists reduced alcohol consumption, decreased cocaine self-administration, and lowered cravings for various addictive substances. The mechanism seemed to involve dopamine signaling in brain regions associated with reward, motivation, and compulsive behavior.

Anecdotal reports from patients prescribed these medications for diabetes or weight loss began surfacing several years ago. Some described losing interest in alcohol without consciously trying to cut back. Others reported reduced cigarette cravings or less preoccupation with food. These observations, initially dismissed as coincidental, accumulated into patterns that prompted systematic investigation.

Small randomized trials began testing the hypothesis directly. A Phase 2 study of 48 adults with alcohol use disorder, published in JAMA Psychiatry, found that once-weekly semaglutide injections reduced drinking days and cravings over nine weeks compared to placebo. The differences were statistically significant, though the sample size limited confidence in the magnitude and durability of effects.

A systematic review published earlier this year identified 33 registered clinical trials investigating GLP-1 drugs for substance use disorders. The majority target alcohol and nicotine addiction, but growing interest extends to opioid and stimulant use disorders. Most trials remain in early phases, with results expected over the next two to four years.

The VA study adds observational data at a scale no randomized trial is likely to match. Its primary limitation is the population: 90% male, average age 65, all with type 2 diabetes, many with significant health complexity. More than half were current or former smokers. Over 40% had high cholesterol. Mental health conditions were common — 18% carried diagnoses of post-traumatic stress disorder, more than 10% had depression or anxiety.

Whether findings from this medically complex, predominantly older male veteran population generalize to women, younger adults, or people without diabetes remains uncertain. The study provides no information about concurrent addiction treatment, making it impossible to determine whether participants were receiving counseling, medication-assisted treatment for opioid use disorder, or other evidence-based interventions that might interact with GLP-1 effects.

Risks Accompany the Promise

Common side effects of GLP-1 medications include nausea, vomiting, diarrhea, constipation, and appetite suppression. For many patients, these diminish over weeks as the body adjusts. For others, they persist and become intolerable, leading to treatment discontinuation. That poses a problem in addiction care, because the VA study and other research suggest benefits fade when medication stops.

A 2026 analysis from Washington University School of Medicine found that veterans who interrupted GLP-1 treatment for as little as six months faced higher risk of heart attack, stroke, and death compared to those who continued therapy. Risk increased further with longer gaps, reaching 22% higher cardiovascular event rates after two years off medication. The mechanism likely involves the reversal of metabolic improvements — weight regain, blood sugar elevation, inflammation returning — rather than a rebound effect beyond baseline.

If similar patterns hold for addiction outcomes, discontinuation could eliminate protective effects or potentially worsen vulnerability. Treatment interruptions are common in addiction care, driven by insurance coverage changes, medication shortages, side effect intolerance, or patient ambivalence about continuing pharmacotherapy. Medications that require uninterrupted use to maintain benefits may fit poorly into the stop-start realities many patients navigate.

Rare but serious complications require monitoring. Reports of severe gastroparesis — stomach paralysis causing persistent nausea, vomiting, and inability to tolerate food — have appeared in post-marketing surveillance, though absolute incidence remains low and causality contested. Cohort studies have reported associations between semaglutide and peripheral neuropathy in diabetes patients, though whether the medication causes nerve damage or whether shared risk factors explain the correlation remains debated.

Emerging data raise questions about bone health. Some analyses link GLP-1 use to modest reductions in bone mineral density and possible increases in osteoporosis or fracture risk, particularly in older adults. Tendon injuries have appeared in patients with obesity taking these medications, possibly related to rapid weight loss or biomechanical changes.

The American Society of Anesthesiologists recommends individualized perioperative strategies for patients on GLP-1 drugs, including temporary liquid diets before procedures for some higher-risk individuals, due to delayed gastric emptying that increases aspiration risk during anesthesia.

None of these risks are hypothetical. They have appeared in clinical trials, observational cohorts, and post-marketing reports. Their frequency remains uncertain, as does the extent to which they apply to addiction treatment populations who may not have diabetes or obesity and may require different dosing strategies.

A Treatment Gap That Predates GLP-1s

The most striking number in the research may not involve GLP-1 drugs at all. An estimated 3% of people with alcohol use disorder ever receive medication proven effective in treating it. For opioid use disorder, medication-assisted treatment with methadone or buprenorphine reduces overdose mortality by more than 50% compared to abstinence-based approaches, yet fewer than one in five people with opioid dependence access these medications in any given year.

The barrier is not medication availability. FDA-approved treatments exist: naltrexone and acamprosate for alcohol, methadone and buprenorphine for opioids, alongside extensive evidence supporting behavioral therapies. These treatments are safe, effective, and supported by decades of clinical trials and real-world implementation data.

The barrier is stigma. Healthcare systems, insurance payers, criminal justice agencies, and society broadly continue treating addiction as moral failure rather than chronic health condition. People avoid seeking treatment fearing judgment, job loss, child custody consequences, or criminal prosecution. Physicians receive minimal training in addiction medicine during medical school and residency, leaving many uncomfortable prescribing medications or discussing substance use without judgment.

Treatment programs themselves sometimes perpetuate stigma, insisting on abstinence as precondition for care or framing medication-assisted treatment as "replacing one drug with another" rather than evidence-based medicine. Regulatory restrictions limit where methadone can be dispensed, requiring daily clinic visits that create logistical barriers for people maintaining employment or living in rural areas.

If GLP-1 drugs eventually prove effective for addiction through rigorous randomized trials, they will enter a healthcare landscape where existing effective treatments remain drastically underused. The medications may help some patients. They will not solve the structural stigma, regulatory barriers, and institutional resistance that prevent most people with substance use disorders from accessing any form of evidence-based care.

Researchers involved in the VA study emphasized the preliminary nature of findings and the need for randomized trials to establish causality. Clinical trials comparing GLP-1s directly against placebo and against existing addiction treatments will determine whether the medications genuinely reduce addiction risk or whether unmeasured confounding factors explain the associations.

Pharmaceutical companies have begun investing in addiction-focused trials, recognizing the potential market if regulatory approval is achieved. Novo Nordisk, manufacturer of semaglutide, has registered trials for alcohol use disorder. Other GLP-1 manufacturers are pursuing similar studies.

Whether results from randomized trials will replicate the observational findings from the VA remains uncertain. Observational studies often overestimate treatment effects when confounding variables cannot be fully controlled. The randomized evidence base for GLP-1s in addiction currently consists of small early-phase trials with short follow-up periods. Larger, longer-duration studies with diverse populations are necessary to establish efficacy, optimal dosing, treatment duration, and which patient subgroups benefit most.

The research published this week adds to a growing body of evidence suggesting GLP-1 receptor agonists may have effects beyond metabolism. It provides rationale for continued investigation. It does not provide sufficient evidence to recommend these medications for addiction treatment outside clinical trials or carefully monitored off-label use by addiction specialists.

For people living with substance use disorders, the immediate message is that effective treatments already exist. The medications and therapies proven to help — naltrexone, acamprosate, methadone, buprenorphine, cognitive behavioral therapy, contingency management, peer support — are available now, covered by most insurance, and supported by extensive evidence. The larger challenge is not developing new treatments but ensuring people can access existing ones without fear, stigma, or structural barriers that currently prevent most who could benefit from ever receiving them.