WHO Database Flags Semaglutide for Potential Suicidal Ideation Signal — Liraglutide Shows No Signal

Significant disproportionality was detected for semaglutide-associated suicidal ideation (ROR 1.45, 95% CI 1.18-1.77). Signal strengthened in patients co-taking antidepressants (ROR 4.45) or benzodiazepines (ROR 4.07). Remained significant vs comparators: dapagliflozin (ROR 5.56), metformin (ROR 3.86), orlistat (ROR 4.24). No significant signal detected for liraglutide.

Disproportionality analysis using a case-control design in the WHO global database (VigiBase) of suspected adverse drug reactions. Reporting odds ratios (ROR) and Bayesian information components (IC) calculated for semaglutide and liraglutide vs all other medications. Sensitivity analyses included patients co-taking antidepressants/benzodiazepines and comparisons with dapagliflozin, metformin, and orlistat.

With tens of millions of people now taking semaglutide worldwide, even a small increased risk of suicidal ideation could affect thousands. This WHO database signal — while not proving causation — warrants urgent investigation given the scale of semaglutide prescribing.

This study adds to a growing safety debate around GLP-1 peptide drugs. While the signal is modest and disproportionality analysis cannot prove causation, the finding that the signal is amplified in patients already taking psychiatric medications raises important questions about monitoring patients with pre-existing mental health conditions who start semaglutide.

Disproportionality analysis cannot establish causation — only signal detection. WHO VigiBase has reporting bias (increased media attention may drive more semaglutide reports). No individual patient-level data to assess confounders. Many semaglutide users have obesity/diabetes, which are independently associated with depression. Signal was found for semaglutide but not liraglutide (both GLP-1 agonists), which is mechanistically puzzling.