Neuropsychiatric adverse events of SGLT2 inhibitors: a pharmacovigilance analysis of the FDA adverse event reporting system database

Objective To systematically analyze the occurrence features, temporal trends and risk signals of neuropsychiatric adverse events (AEs) linked to three sodium-glucose co-transporter 2 (SGLT2) inhibitors (canagliflozin, dapagliflozin, empagliflozin), and provide evidence for clinical safe medication. Methods We extracted primary suspected (PS) only SGLT2 inhibitor-related reports from the FAERS database (2015Q4–2025Q3) after standardized deduplication. Descriptive analysis, dual disproportionality analysis ( ROR and IC ), time-stratified sensitivity analysis, and indication-based subgroup stratification were conducted. Results A total of 141,555 eligible AE reports were included, with 11,071 neuropsychiatric AEs (7.82%). Annual trends differed; males and 45–65-year-olds were high-risk. Diabetic neuropathy was a class-wide signal; each drug had unique signals. Time-stratified analysis confirmed signal robustness. Central nervous system vascular disorders were a cross-indication core signal. Conclusion The three SGLT2 inhibitors have non-negligible neuropsychiatric risks with population/indication differences and class/drug-specific signals. All findings represent exploratory safety signals and do not confirm causal relationships between SGLT2 inhibitors and neuropsychiatric adverse events. Limitations include no indication confounding adjustment or multiple testing correction, as well as inherent flaws of FAERS such as reporting bias, underreporting, and outcome heterogeneity. Targeted monitoring is warranted for clinical awareness, but changes in clinical practice require further confirmation in large observational cohorts or prospective studies, especially for elderly/non-diabetic patients on dapagliflozin.