Postmarket safety profile of suicide/self-injury for GLP-1 receptor agonist: a real-world pharmacovigilance analysis

Abstract

Background: Recent reports of individuals experiencing suicidal and/or self-injurious behaviors while using liraglutide and semaglutide have heightened the concerns regarding neuropsychiatric safety of Glucagon-like peptide-1 agonists (GLP-1RAs). As real-world evidence is very limited, we explored the association between GLP-1RA and suicide/self-injury by mining the FDA Adverse Event Reporting System (FAERS) database.

Methods: The FAERS database was queried from 2005 Q2 to 2023 Q2. The Reporting Odds Ratio (ROR) and Empirical Bayes Geometric Mean (EBGM) were used to conduct the disproportionality analysis.

Results: A total of 534 GLP-1RA-associated suicide/self-injury cases were reported in the FAERS during the study period. GLP-1RA did not cause a disproportionate increase in overall suicidal and self-injurious cases (ROR: 0.16, 95%CI 0.15-0.18, P < 0.001; EBGM05: 0.15). Stratified analyses found no safety signal of suicide/injury for GLP-1RA in both females and males. The ROR for suicide/self-injury with GLP-1RA was slightly elevated (ROR: 2.50, 95%CI 1.02-6.13, P = 0.05) in children, while the EBGM05 was < 2 in this population. No significant signal value was observed in other age groups. No over-reporting of suicide/self-injury was identified for GLP-1RA before or after the COVID-19 pandemic outbreak.

Conclusions: The cases of suicide or self-injury reported to FAERS do not indicate any overall safety signal attributable to GLP-1RA at this time. Subgroup analysis revealed a marginal elevation of ROR for suicide and self-injury with GLP-1RA in children, but no safety signal was detected by EBGM05 in this population. Further large-scale prospective investigations are still warranted to further confirm this finding.

Keywords: FAERS; GLP-1RA; pharmacovigilance analysis; self-injury; suicide.

Figure 1.

Flow chart of data queries within the FAERS database. FAERS, the U.S. Food and Drug Administration (FDA) Adverse Event Reporting System; Q2, the second quarter; GLP-1RA, glucagon-like peptide-1 receptor agonist; PT, preferred term; SMQ, standard MedDRA (Medical Dictionary for Regulatory Activities) query.

Figure 2.

Results of overall disproportionality analysis for GLP-1RA-associated suicide/self-injury at PT levels. (A) EBGM05s of GLP-1RA-associated suicide/self-injury. (B) RORs (95% CI) of GLP-1RA-associated suicide/self-injury. EBGM05, lower one-sided 95% confidence limit (95% CI) of the empirical Bayes geometric mean; PT, preferred term; ROR, reporting odds ratio; GLP-1RA, glucagon-like peptide-1 receptor agonist.

Figure 3.

Results of disproportionality analysis for suicidal and self-injurious reports associated with GLP-1RA at drug levels. (A) Number of suicide/self-injury and non-suicide/self-injury reports for each GLP-1RA drug. (B) EBGM05 of GLP-1RA-associated suicide/self-injury for each distinct GLP-1RA drug. (C) RORs (95% CI) of GLP-1RA-associated suicide/self-injury for each distinct GLP-1RA drug. ROR, reporting odds ratio; GLP-1RA, glucagon-like peptide-1 receptor agonist; EBGM05, lower one-sided 95% confidence limit (95% CI) of the empirical Bayes geometric mean.

Figure 4.

Results of subgroup disproportionality analysis of GLP-1RA-associated suicide/self-injury based on age, sex and reporting time. (A) EBGM05s of GLP-1RA-associated suicide/self-injury in different sex, age and reporting time groups. (B) RORs (95% CI) of GLP-1RA-associated suicide/self-injury in different sex, age and reporting time groups. ROR, reporting odds ratio; 95% CI, 95% confidence limit; GLP-1RA, glucagon-like peptide-1 receptor agonist; EBGM05, lower one-sided 95% CI of the empirical Bayes geometric mean.