Overview
- In a BMJ cohort analysis of 606,434 U.S. veterans with type 2 diabetes, initiating GLP-1 receptor agonists was compared with starting SGLT-2 inhibitors using target-trial emulation and inverse probability–weighted Cox models over up to three years.
- Among veterans without prior substance use disorders, GLP-1 initiation was associated with lower incidence across alcohol, cannabis, cocaine, nicotine and opioid disorders, with a composite hazard ratio of 0.86 versus SGLT-2s.
- Absolute risks declined by only a few cases per 1,000 over three years, including an estimated net reduction of 5.57 per 1,000 for alcohol use disorder and 6.61 per 1,000 for the composite of new-onset SUDs.
- In veterans with existing SUD, GLP-1 use was linked to fewer SUD-related emergency visits (HR 0.69), hospitalizations (0.74), overdoses (0.61), suicidal ideation (0.75) and SUD-related mortality (0.50).
- Findings were consistent across multiple sensitivity analyses, though authors note potential residual confounding and limited generalizability to non-VA populations and call for randomized trials, as preclinical data suggest plausible effects on brain reward pathways.