Semaglutide and depression: no significant increase detected, in very mixed evidence

Design

Systematic review and meta-analysis of trials, observational and pharmacovigilance studies

Population

Patients receiving semaglutide for type 2 diabetes, obesity or related conditions

Primary outcome

Depression, anxiety and suicidal ideation or attempt

Effect

Depression RR 1.25 (0.95–1.65); anxiety RR 1.22 (0.93–1.60); suicidality RR 1.20 (0.90–1.62); I² 92–99%

This meta-analysis pooled randomised trials, observational studies, large database analyses and pharmacovigilance disproportionality studies of patients receiving semaglutide, searching to January 2026.

The pooled risk ratio was 1.25 for depression (95% CI 0.95–1.65), 1.22 for anxiety (0.93–1.60) and 1.20 for suicidal ideation or attempt (0.90–1.62). None reached statistical significance. Heterogeneity was extreme, with I² of 92–99%, and risk of bias was substantial.

The authors frame the result as the absence of a detected risk rather than proof of safety, which is the right reading. Mixing trials with spontaneous-report studies, which measure reporting rather than incidence, makes the pooled number hard to interpret. For prescribers it supports continuing semaglutide where indicated, while asking about mood, particularly in patients with a psychiatric history.

• Ask about mood and suicidal thoughts before and during semaglutide treatment in patients with a psychiatric history.
• No significant increase in depression, anxiety or suicidality was found; do not withhold semaglutide on this basis alone.
• Report new or worsening mood symptoms on GLP-1 receptor agonists to pharmacovigilance systems such as PvPI in India.
• Distinguish mood change from the fatigue and reduced intake of rapid weight loss.

Why it matters

It does not settle the psychiatric signal, but it finds no clear increase across the evidence so far.

Don't overread it

Absence of a significant increase is not proof of safety; the confidence intervals allow a modest rise.

The statistics, in plain English

A risk ratio of 1.25 with an interval of 0.95 to 1.65 crosses 1.0, so no significant increase was shown; but the upper limit allows a risk up to 65% higher. I² near 99% means the studies disagreed almost completely, so a single pooled figure summarises them poorly.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.