GLP-1 drugs tied to fewer addictions and overdose deaths in large veterans study

BMJ Group

Can a diabetes drug prevent addiction? The question sounds implausible, but a study of more than 600,000 U.S. veterans published in The BMJ provides the most comprehensive evidence yet that it might.

GLP-1 receptor agonists, the class of medications that includes semaglutide, liraglutide, and dulaglutide, were originally developed for type 2 diabetes and later adopted for weight management. Patients began reporting that they had lost interest in drinking or smoking after starting the drugs. Scattered observational studies found associations with lower rates of alcohol and cannabis disorders, fewer opioid overdoses, and reduced alcohol-related hospitalizations.

But those studies examined one substance at a time. This new analysis asked a more ambitious question: do GLP-1 drugs reduce the risk of substance use disorders across the board?

Two questions, one cohort

Researchers based in Saint Louis, Missouri used Department of Veterans Affairs data to study 606,434 veterans with type 2 diabetes. They split the population into two analyses. In the first, 524,817 veterans without any history of substance use disorders were tracked for up to three years to see who developed new addictions. In the second, 81,617 veterans with pre-existing substance use disorders were monitored for serious adverse outcomes.

In both analyses, veterans who started GLP-1 drugs were compared with those who started SGLT2 inhibitors, a different class of diabetes medication that does not target GLP-1 receptors. The researchers used an emulated target trial design, a rigorous statistical approach that mimics the structure of a randomized trial using observational data.

Consistent reductions across every substance

Among veterans with no prior substance use disorders, GLP-1 use was associated with an overall 14% reduction in risk. The reductions held for each substance individually: 18% for alcohol, 14% for cannabis, 20% for cocaine, 20% for nicotine, and 25% for opioids. In practical terms, that meant roughly 1 to 6 fewer cases per 1,000 people over three years.

Among those with existing addictions, the results were more dramatic. GLP-1 use was tied to 31% fewer substance-related emergency department visits, 26% fewer hospitalizations, 50% fewer drug-related deaths, 39% fewer overdoses, and 25% fewer instances of suicidal ideation or attempt. That translated to about 1 to 10 fewer events per 1,000 people over three years.

Limitations the authors acknowledge

The study was conducted within the VA healthcare system, which skews toward older men. While subgroup analyses showed consistent effects in women, the sample was small. Socioeconomic status, lifestyle factors, and other unmeasured variables could confound the results. People who choose GLP-1 medications, or whose doctors prescribe them, may differ in systematic ways from those prescribed SGLT2 inhibitors.

The comparison group itself is an open question. Whether SGLT2 inhibitors have any independent effect, positive or negative, on substance use behavior is unknown. If they happened to increase addiction risk, the relative benefit of GLP-1 drugs would be inflated.

And this remains observational evidence. No matter how sophisticated the statistical design, it cannot provide the causal certainty of a randomized controlled trial.

From observation to intervention

In a linked editorial, Fares Qeadan of Loyola University Chicago writes that the study strengthens the case for GLP-1 receptor agonists influencing substance-related outcomes in real-world practice. He suggests that when GLP-1 drugs are clinically indicated for metabolic reasons, the potential benefits for addiction could factor into prescribing decisions.

But he also emphasizes that evidence-based addiction treatments remain the preferred approach and that the challenge now is translating the observational signal into clinical trials while ensuring equitable access. GLP-1 medications are expensive and subject to supply constraints. If they do prove effective against addiction, ensuring they reach the populations most affected by substance use disorders, who often face the greatest barriers to healthcare access, will be essential.

The biological hypothesis is that GLP-1 receptors in brain reward circuits may dampen craving at a fundamental level, not specific to any one substance but operating on the shared neurobiology of wanting. If that hypothesis survives controlled testing, it would represent a new category of addiction treatment: one that targets the common mechanism rather than the individual drug.