GLP-1 RAs and Substance Use Disorders: Findings from a Veteran Cohort

In a large U.S. cohort of veterans with type 2 diabetes, initiation of GLP‑1 receptor agonists was associated with a lower risk of developing several new substance use disorders (SUDs) compared with initiation of SGLT2 inhibitors.

The analysis was an observational target-trial emulation using electronic health record data rather than a randomized comparison, with results presented as hazard ratios and three-year net risk differences per 1,000 individuals. Investigators also reported associations among veterans with preexisting SUD for several SUD-related acute care outcomes and self-harm or overdose outcomes during follow-up.

The investigators described a VA-based cohort assembled from electronic health records and emulated multiple parallel trials comparing new initiators of GLP‑1RAs with new initiators of SGLT2 inhibitors, with follow-up being up to three years. The primary approach used inverse probability–weighted Cox models to estimate hazard ratios and three-year net risk differences per 1,000 individuals.

For incident diagnoses, GLP‑1RA initiation versus SGLT2 inhibitor initiation was associated with a lower hazard for the composite new‑onset SUD outcome (HR 0.86), with a three-year net risk difference of −6.61 per 1,000 individuals. There were also lower estimated hazards for selected substance-specific outcomes, including alcohol use disorder (HR 0.82; NRD −5.57 per 1,000) and opioid use disorder (HR 0.75; NRD −0.86 per 1,000).

Among veterans with a preexisting SUD, GLP‑1RA initiation was associated with lower hazards and negative net risk differences for SUD‑related ED visits (HR 0.69; NRD −8.92 per 1,000), SUD-related hospitalizations (HR 0.74; NRD −6.23), overdose events (HR 0.61; NRD −1.49), suicidal ideation (HR 0.75; NRD −9.95), and SUD-related mortality (HR 0.50; NRD −1.52).

Key Takeaways:

• An observational, EHR-based target-trial emulation compared new initiators of GLP‑1RAs versus SGLT2 inhibitors among U.S. veterans with type 2 diabetes.
• GLP‑1RA initiation was associated with a lower incidence of a composite new-onset SUD outcome and selected substance-specific SUD diagnoses over up to three years.
• Among veterans with preexisting SUD, GLP‑1RA initiation was associated with lower hazards for several SUD-related acute care, overdose, and suicidal ideation outcomes, alongside author-noted observational limitations.