Neuropsychiatric adverse event reporting with GLP-1-based therapies: a post-marketing pharmacovigilance analysis using the WHO global individual case safety report database (VigiBase)

The source, as retrieved

Abstract

[OBJECTIVE] Concerns regarding neuropsychiatric adverse events (NPAEs), including depression and suicidality, have emerged with the expanding use of GLP-1-based therapies across metabolic and weight-management indications. This study aimed to characterize neuropsychiatric adverse-event reporting patterns for GLP-1-based therapies in VigiBase using disproportionality analysis. [METHODS] Individual case safety reports entered into the WHO global pharmacovigilance database (VigiBase) from inception through 31 January 2025 were examined. Disproportionality analysis using the information components and the reporting odds ratio was conducted to identify putative safety signals. The full VigiBase reporting background during the study period was used as the comparator. We described the drug-event pairs, mapped signal overlap across agents, evaluated the relationship between reporting frequency and signal magnitude, and conducted stratified analyses by sex and age. [RESULTS] Among 899 280 GLP-1RAs-related reports, 19 811 (2.2%) related NPAEs comprised 273 preferred terms and generated 69 positive signals. Semaglutide and liraglutide presented the broadest and most potent signal portfolios, driven chiefly by binge eating (semaglutide, IC₀₂₅ = 3.379; ROR₀₂₅ = 14.483), and aversion (liraglutide, IC₀₂₅ = 3.365; ROR₀₂₅ = 20.512). In contrast, albiglutide was essentially signal negative, and dulaglutide and tirzepatide generated narrower spectra. In addition, semaglutide was positively associated with two "suicide/self-injury" AEs: "depression suicidal" (ROR₀₂₅ = 4.822) and "suicidal ideation" (ROR₀₂₅ = 2.643). [CONCLUSIONS] This study identified several potential drug safety signals. However, the estimates were unadjusted and may be influenced by confounding by indication, channelling, notoriety bias, and event competition. Eating-related signals should be interpreted cautiously because they may overlap with the treated population and appetite-related pharmacology of these agents. Given the spontaneous reporting nature of VigiBase, these findings should be interpreted as potential reporting signals rather than evidence of causality. Further well-designed observational studies and prospective clinical investigations are needed to clarify whether GLP-1-based therapies are associated with increased neuropsychiatric risk and to identify patient-level risk modifiers.

Where this record came from