Semaglutide Showed No Neuropsychiatric Harm and May Protect Against Cognitive Decline and Nicotine Addiction

Semaglutide was not associated with increased risk of any of 22 neurological or psychiatric outcomes over 12 months compared to three other diabetes medications. Instead, it was associated with significantly reduced risk of cognitive deficit (HR 0.72, 95% CI 0.64–0.80 vs sitagliptin; HR 0.72, 95% CI 0.63–0.81 vs glipizide), dementia (HR 0.52, 95% CI 0.40–0.68 vs sitagliptin), and nicotine misuse (HR 0.72, 95% CI 0.61–0.85 vs glipizide; HR 0.77, 95% CI 0.65–0.90 vs empagliflozin). No differences were observed in negative control outcomes, strengthening confidence in the findings.

Retrospective cohort study using TriNetX US Collaborative Network electronic health records covering over 100 million patients. Three propensity-score matched cohorts (1:1) compared semaglutide users with type 2 diabetes to users of sitagliptin, empagliflozin, or glipizide prescribed between December 2017 and May 2021. Cox regression assessed 22 neurological and psychiatric outcomes over 12 months. Negative control outcomes were used to detect unmeasured confounding. Multiple-testing correction was applied.

Concerns about suicidality and other neuropsychiatric side effects have dogged GLP-1 receptor agonists. This large-scale study not only found no increased neuropsychiatric risk with semaglutide but discovered potential protective effects on cognition and addiction — findings that could expand semaglutide's therapeutic applications if confirmed in clinical trials.

This study addresses one of the most debated questions in current medicine: are GLP-1 drugs safe for the brain? The finding that semaglutide may actually protect against cognitive decline aligns with emerging preclinical evidence that GLP-1 receptors in the brain play roles in neuroprotection. Clinical trials are now underway testing semaglutide specifically for Alzheimer's disease and addiction disorders.

Retrospective observational design cannot prove causation. The study relied on electronic health records, which may undercount diagnoses not coded in routine care. The exploratory nature meant no pre-registered protocol. The cohort was U.S.-based and may not generalize globally. Some comparisons lost significance after multiple-testing correction.