Randomized trial · 2026

Semaglutide for the treatment of cognitive dysfunction in major depressive disorder: A randomized clinical trial.

S Badulescu, H Gill, H Shah, R Brudner, L Phan, JD Di Vincenzo, A Tabassum, M Armanyous. Med (New York, N.Y.).

BACKGROUND: Evidence suggests that glucagon-like peptide 1 receptor agonists (GLP-1RAs) might have pro-cognitive effects. No prior study has evaluated the efficacy and safety of a GLP-1RA for the treatment of cognitive dysfunction in adults with major depressive disorder (MDD) in a randomized clinical trial. METHODS: This was a 16-week, randomized, double-blind, placebo-controlled, parallel-group trial (NCT04466345). Eligible adults met DSM-5-defined criteria for MDD, exhibited pre-treatment evidence of cognitive impairment, and were overweight/obese. Patients were randomized (1:1) to receive an adjunctive placebo or 14 mg oral semaglutide. The primary outcome was an executive function composite score comprising the digit symbol substitution test, the Stroop test, and the n-back test. Secondary outcomes included a global cognition composite score, measures of functioning, depressive symptom severity, suicidality, and body weight. FINDINGS: 72 participants were randomized to oral semaglutide (n = 35) or placebo (n = 37). Semaglutide did not improve executive function (adjusted Z score difference [semaglutide - placebo]: 0.32, 95% confidence interval [CI]: -0.92 to 1.58, p = 0.60). Preplanned secondary analysis showed treatment effects for global cognition (2.39, 95% CI: 0.19 to 4.60, p = 0.03) and body weight (kg) (adjusted mean difference -6.03, 95% CI: -8.76 to -3.29, p < 0.001). Treatment did not affect depressive symptom severity or the frequency of suicidal ideation. Gastrointestinal side effects were common in the semaglutide group, with no serious adverse events. CONCLUSION: Semaglutide did not improve executive function; results from secondary analyses suggested effects on specific domains of cognition. Semaglutide was safe for patients with MDD. FUNDING: This work was supported by the Physicians' Services Incorporated Foundation.

Open sourcedoi:10.1016/j.medj.2025.100916PMID 41218611