One Drug Class, Every Addiction: The 600,000-Patient Study That Changed the Conversation

The Question Nobody Had Asked

Prior to this study, the evidence linking GLP-1 medications to reduced addiction was accumulating — but in silos. One study found semaglutide associated with lower alcohol use disorder risk. Another linked it to reduced opioid overdose. A third showed benefits for cannabis use disorder. Each examined one substance at a time.

Dr. Ziyad Al-Aly, a clinical epidemiologist at WashU Medicine and the study's senior author, asked the broader question: do GLP-1s work against addiction across the board? And can they reduce the most serious consequences — overdose and death — in people who already have substance use disorders?

The answer, based on the largest and most comprehensive dataset ever assembled on this question, appears to be yes on both counts.

How the Study Worked

The team analyzed electronic health records from 606,434 U.S. veterans with type 2 diabetes. They compared those who started a GLP-1 receptor agonist (most commonly semaglutide, liraglutide, or dulaglutide) against those who started an SGLT2 inhibitor — a different class of diabetes medication that also has metabolic benefits but doesn't act on the brain's reward pathways.

The participants were split into two cohorts. The first group — 524,817 people without a pre-existing substance use disorder — was tracked to see who developed new substance use disorders. The second group — 81,617 people who already had a substance use disorder — was monitored for serious harmful outcomes: emergency department visits, hospitalizations, overdoses, deaths, and suicidal ideation.

Preventing New Addictions

Among people without prior substance use disorders, GLP-1 use was associated with a 14% reduced risk of developing any substance use disorder compared to those on non-GLP-1 diabetes medications. The reductions were consistent across every substance category:

In absolute terms, this translated to seven fewer new substance use disorder diagnoses per 1,000 GLP-1 users over the study period. At population scale — with millions of Americans now taking GLP-1 medications — the public health implications are significant.

Reducing Harm in Existing Addictions

The findings were even more striking among veterans who already had substance use disorders. Over three years of follow-up, GLP-1 use was associated with substantially fewer serious adverse outcomes:

This translated to 12 fewer serious harm events per 1,000 GLP-1 users. For context, there are currently no medications that have demonstrated this kind of cross-substance harm reduction.

"Quieting the Roar of Addiction"

"In addiction medicine, a lot of treatments target just one thing — for example, a nicotine patch helps with smoking, but not alcohol — but there is no medication that works across addictive substances, let alone all of them. The revelation about GLP-1 medication is that it really works against all major substances, and it works uniformly, not because it acts against alcohol or opioids or nicotine specifically, but because it is likely acting against the craving itself." — Dr. Ziyad Al-Aly, senior author, WashU Medicine

Al-Aly draws a direct parallel to the well-documented "food noise" phenomenon — where GLP-1 users report a quieting of the persistent preoccupation with food that drives overeating. What this study suggests, he argues, is something broader: these drugs may also quiet what he calls "drug noise," the relentless craving that drives addiction across substances.

The biological basis for this is that GLP-1 receptors are expressed in key brain regions involved in reward processing — particularly the ventral tegmental area (VTA) and nucleus accumbens. These are the same circuits hijacked by addictive substances. By modulating dopamine signaling in these areas, GLP-1 agonists may reduce the "incentive salience" of rewards across the board — whether that reward is food, alcohol, nicotine, or opioids.

Important Caveats

This is an observational study, not a randomized controlled trial. That means it can identify associations but cannot definitively prove that GLP-1 medications caused the reductions in substance use disorders. Confounding factors — differences between the types of patients prescribed GLP-1s versus SGLT2 inhibitors — could play a role, even though the researchers used propensity-score matching to account for this.

The study population was U.S. veterans with type 2 diabetes, a group that is disproportionately male and older. Whether these findings generalize to younger populations, women, or people without diabetes remains to be confirmed.

The researchers are clear about the implications: these findings support the case for randomized clinical trials to test GLP-1 medications as standalone treatments for addiction — trials specifically designed to measure effects on overdose and drug-related death.