GLP-1 receptor agonists linked to lower substance use disorder risk across multiple substances
In a large Veterans Affairs cohort of people with type 2 diabetes, initiation of GLP-1 receptor agonists was associated with lower risk of a broad range of substance use disorders, from alcohol and nicotine to opioids and cocaine. The signal extended beyond new diagnoses: among patients with pre-existing substance use disorders, treatment was also associated with fewer emergency department visits, fewer hospital admissions, lower mortality, fewer overdoses, and fewer episodes of suicidal ideation or attempt.
The findings come from an emulation of eight parallel new-user, active-comparator target trials using electronic health records. That design matters because it compares patients starting one glucose-lowering therapy against those starting another, rather than against no treatment, which helps reduce confounding by indication in a population with substantial medical complexity. Across both prevention and harm-reduction analyses, the direction of effect stayed consistent.
Incident substance use disorders fell across alcohol, nicotine, opioids, and other drug classes
The primary cohort included veterans with type 2 diabetes and no prior substance use disorder. Within that population, initiation of GLP-1 receptor agonists was compared with initiation of SGLT-2 inhibitors. Investigators then tracked incident alcohol, cannabis, cocaine, nicotine, opioid, and other substance use disorders, along with a composite endpoint covering all incident SUDs.
The strongest reduction appeared for alcohol-related disorders, with a hazard ratio of 0.82 and a three-year net risk difference of -5.57 per 1000 people. Opioid-related disorders also declined, with a hazard ratio of 0.75 and a net risk difference of -0.86 per 1000 people. Nicotine-related disorders fell as well, with a hazard ratio of 0.80. Cannabis, cocaine, and other SUDs showed similar though not identical magnitudes of association, and the composite endpoint for all incident SUDs reached a hazard ratio of 0.86 with a net risk difference of -6.61 per 1000 people.
One pattern stands out: the association was not confined to a single intoxicant or one narrow behavioral phenotype. It was broad.
Harm-reduction signals also emerged in veterans with established substance use disorders
The second protocol focused on veterans who already had substance use disorders at baseline. In that setting, GLP-1 receptor agonist initiation was associated with fewer SUD-related emergency department visits, fewer SUD-related hospital admissions, lower SUD-related mortality, fewer overdoses, and fewer episodes of suicidal ideation or attempt.
Emergency department visits showed a hazard ratio of 0.69, while hospital admissions carried a hazard ratio of 0.74. SUD-related mortality was reduced with a hazard ratio of 0.50, a point estimate that suggests a marked difference in fatal outcomes even in a clinically complex population. Drug overdose also declined, with a hazard ratio of 0.61. Suicidal ideation or attempt was lower as well, with a hazard ratio of 0.75 and a three-year net risk difference of -9.95 per 1000 people.
For clinicians, the practical implication is straightforward: a therapy long used for glycemic control and weight management may influence behaviors and complications that usually sit outside endocrinology. The bedside consequence could be fewer acute crises among veterans already carrying the burden of addiction.
A biologically plausible signal, but still observational
Several mechanisms could explain why a gut-brain metabolic therapy might influence substance use. GLP-1 receptor pathways intersect with reward circuitry, satiety signaling, and dopaminergic reward processing, which are central to craving and reinforcement. If a medication dampens the salience of rewarding stimuli or reduces compulsive seeking behavior, it could plausibly reduce intake across multiple substance classes rather than affecting only one drug category.
The study design strengthens confidence in the association, but it does not establish causality. Even with inverse probability weighting and active comparator emulation, residual confounding remains possible. Patients selected for GLP-1 receptor agonists may differ in ways that are difficult to fully measure, including health-seeking behavior, prescribing access, metabolic severity, obesity burden, or clinician concern about future risk.
Still, the coherence of the findings across incident disorders, adverse outcomes, and adherence analyses argues against a single isolated statistical fluke. Directionally consistent results in the adherence analyses are especially important because they suggest the observed associations were not driven solely by one-off early treatment decisions.
Why the VA cohort gives the study unusual clinical weight
The cohort was large: 606,434 US veterans with type 2 diabetes were available for assignment into the two protocols. In the primary trial, 524,817 initiators were included, with 124,001 starting a GLP-1 receptor agonist and 400,816 starting an SGLT-2 inhibitor. The second protocol included 81,617 initiators, with 16,768 in the GLP-1 receptor agonist group and 64,849 in the SGLT-2 inhibitor group.
A population this large gives the analysis enough events to evaluate several substance-specific endpoints rather than collapsing everything into a single composite. It also reflects a real-world veteran population where psychiatric comorbidity, chronic pain, nicotine dependence, and metabolic disease often overlap. Those overlaps make the observed associations clinically meaningful, because the affected patients are the same ones most likely to cycle through outpatient visits, emergency care, and hospitalization.
Clinical questions that remain open after the risk reductions
The data support a potential preventive and therapeutic role for GLP-1 receptor agonists in substance use disorders, but the study does not identify which agents, doses, or treatment durations matter most. It also does not separate the contribution of weight loss, improved glycemic control, appetite suppression, or direct neurobehavioral effects.
Randomized trials will be needed to determine whether the benefits persist outside a veteran cohort and whether the same pattern appears in women, younger adults, people without diabetes, or patients whose substance use disorder is the dominant clinical problem rather than a comorbidity. For now, the evidence points in one direction: in a large real-world diabetic population, GLP-1 receptor agonist initiation tracked with fewer incident substance use disorders and fewer serious complications in patients with established disease.