A large observational study of over 600,000 U.S. veterans with type 2 diabetes has found that GLP-1 receptor agonists, medications such as semaglutide (Ozempic, Wegovy) and tirzepatide (Mounjaro), are associated with substantially lower risks of developing substance use disorders and fewer severe outcomes among those who already have them.
Published in The BMJ on March 4, 2026, the research, led by clinical epidemiologist Dr. Ziyad Al-Aly of Washington University School of Medicine in St. Louis and the VA Saint Louis Health Care System, compared veterans starting GLP-1 drugs with those beginning SGLT2 inhibitors (another class of diabetes medications, such as empagliflozin or dapagliflozin). The analysis used a “target trial emulation” approach to mimic randomized trial conditions as closely as possible in real-world data.
Key Findings on Prevention
Among more than 500,000 veterans without a prior substance use disorder, those prescribed GLP-1 drugs showed a 14% lower overall risk of developing any new substance use disorder over up to three years. Specific reductions included:
- Alcohol use disorder: 18% lower risk
- Cannabis use disorder: 14% lower
- Cocaine use disorder: 20% lower
- Nicotine use disorder: 20% lower (some reports note up to 26%)
- Opioid use disorder: 25% lower
These translated to roughly 6–7 fewer new diagnoses per 1,000 people over the study period.
Benefits for Those with Existing Disorders
For the approximately 81,000 veterans with pre-existing substance use disorders, GLP-1 use correlated with better outcomes across multiple serious events:
- 31% fewer substance-related emergency department visits
- 26% fewer hospital admissions
- 39% fewer overdoses
- 25% fewer suicidal thoughts or attempts
- 50% fewer substance-related deaths
This amounted to about 10–12 fewer serious events per 1,000 people over three years, including notably fewer fatalities.
The consistency across diverse substances such as alcohol, opioids, cocaine, nicotine, and cannabis suggests GLP-1 drugs may target a shared biological pathway in the brain’s reward system. GLP-1 receptors are present in the mesolimbic system, which governs motivation, reward, and cravings.
By modulating dopamine signaling, these medications appear to dampen the “craving” or “drug noise” that drives addictive behaviors, similar to how they reduce “food noise” in people managing obesity.
Animal studies have long supported this mechanism, showing reduced self-administration of alcohol, cocaine, and nicotine in rodents and primates. Human anecdotes from patients and smaller observational reports have echoed these effects, with some individuals losing interest in alcohol or cigarettes after starting the drugs for diabetes or weight loss.
Important Limitations and Context
The cohort was predominantly male (around 90%), with an average age of 65, and many had complex health profiles, including high rates of smoking history, cardiovascular issues, and mental health conditions like PTSD, depression, and anxiety.
Results may not fully apply to women, younger adults, or those without diabetes. The study shows strong associations but cannot prove causation due to its observational nature—unmeasured factors could influence outcomes.
Existing treatments for substance use disorders, such as naltrexone for alcohol or buprenorphine for opioids, remain effective but underused, often due to stigma and access barriers. GLP-1 drugs are not approved for addiction treatment, and prescribing them solely for that purpose lacks sufficient evidence.
Ongoing randomized controlled trials, including a large VA-led phase 3 study of semaglutide for alcohol use disorder in veterans (NCT07218354), will provide clearer insights into causality, long-term effects, and potential risks like rebound cravings upon discontinuation or impacts on general motivation.
For patients with type 2 diabetes who also face substance use challenges, these findings offer an additional consideration when choosing treatments. As millions already take GLP-1 medications, their broader public health implications could be significant if confirmed.
This emerging evidence highlights a potential paradigm shift: addressing addiction not substance-by-substance but through a common neural vulnerability.