A Veterans Affairs study of more than 606,000 patients found that GLP-1 receptor agonists — the class of drugs that includes semaglutide, sold as Ozempic and Wegovy — reduced the risk of developing any substance use disorder by 14%, cut overdose deaths by 39%, and reduced overall mortality by 50% in patients with existing addictions. The findings, published in The BMJ, represent the largest real-world evidence that a single drug class works across all major substance types: alcohol, opioids, cocaine, nicotine, and cannabis. [1]

The mechanism targets a common brain reward pathway. GLP-1 receptors in the central nervous system modulate dopaminergic responses to substances of abuse — the same pathway that regulates appetite and satiety. Preclinical research has shown that GLP-1 signaling curbs craving across multiple substance types, suggesting the effect is not substance-specific but circuit-level. [2]

The VA study compared patients who initiated GLP-1 therapy against those who initiated SGLT-2 inhibitors, another diabetes drug class. Participants starting GLP-1s were 14% less likely to develop a substance use disorder over three years. Among those with pre-existing addictions, GLP-1 use was associated with reduced emergency room visits, hospitalizations, overdoses, and suicidal ideation. [1]

The paper's May 14 edition covered a Lancet randomized controlled trial showing GLP-1 cut heavy drinking days by 14 points. That trial was alcohol-specific. The VA study is larger, broader, and includes hard endpoints — overdose and death — that change the cost-benefit calculus for insurers and employers. A 50% mortality reduction in a population of veterans with substance use disorders is a clinical finding that demands a policy response. [3]

The demographic-winter thread has tracked GLP-1 as a population-health variable. This study makes it an addiction-medicine variable. The question the paper flagged in May — whether any insurer, employer, or addiction-medicine specialty body reacts to GLP-1/AUD evidence as population-health policy — now has a VA-scale data set to react to.

The VA population is predominantly male, older, and veteran — a demographic that limits generalizability to the broader U.S. population. But the mechanism is neurobiological, not demographic. If GLP-1 drugs reduce addiction through a shared brain circuit, the effect should be present across populations. The question is whether insurers expand formulary access based on addiction-medicine evidence, or whether the addiction indication remains a secondary benefit of a weight-loss drug.

The 50% mortality reduction is the number that changes the conversation. Addiction kills. A drug that cuts mortality in half across all substance classes is not a side effect. It is the finding.

-- NORA WHITFIELD, Chicago