Adverse event reporting signals for SGLT-2 inhibitor–metformin combination therapy: a disproportionality analysis of the FAERS database

Objective This study aimed to evaluate the real-world safety profile of combination therapy with a sodium-glucose cotransporter-2 inhibitor (SGLT-2i) and metformin relative to monotherapy, utilizing the U.S. Food and Drug Administration Adverse Event Reporting System (FAERS) database. Methods We extracted adverse event (AE) reports from the FAERS database from the first quarter (Q1) of 2013 to the fourth quarter (Q4) of 2025 in which the combined regimen (ertugliflozin, dapagliflozin, empagliflozin, or canagliflozin plus metformin) was designated as the primary suspect. Disproportionality analyses were performed using the Reporting Odds Ratio (ROR), Proportional Reporting Ratio (PRR), Bayesian Confidence Propagation Neural Network (BCPNN), and Multi-item Gamma Poisson Shrinker (MGPS) algorithms. Additionally, we examined the time-to-onset distribution and the clinical outcomes of the documented AEs. Results A total of 23,407 combination-therapy reports were included. At the System Organ Class (SOC) level, ROR-based signal detection identified pronounced disproportionalities in metabolism and nutrition disorders (ROR = 9.00), renal and urinary disorders (ROR = 3.91), reproductive system and breast disorders (ROR = 2.59), and infections and infestations (ROR = 2.49). At the Preferred Term (PT) level, the five most frequently reported events cross-validated by all four algorithms were diabetic ketoacidosis (n = 3,522), weight decreased (n = 1,294), blood glucose increased (n = 1,226), acute kidney injury (n = 1,153), and euglycaemic diabetic ketoacidosis (euDKA, n = 1,114). Notably, euDKA demonstrated the highest signal intensity (ROR = 335.43), with the point estimate for the combination group being numerically higher than that observed for SGLT-2i monotherapy (ROR = 185.24). Regarding documented outcomes, hospitalisation or prolonged hospitalisation constituted the most prevalent serious outcome (33.7%), followed by life-threatening events (7.6%) and death (2.7%). In the sub-cohort with definitive time-to-onset data, 39.56% of AEs occurred within the first 30 days following treatment initiation. Conclusion This pharmacovigilance analysis of the FAERS database detected significant disproportionality signals for SGLT-2i/metformin combination therapy, particularly within metabolic/nutritional, renal/urinary, and infectious disease categories. The euDKA signal appeared prominent in descriptive comparisons. The predominance of hospitalisation-related events, coupled with the substantial concentration of AEs within the initial 30-day window, highlights the early treatment phase as a potentially critical monitoring period. These exploratory findings warrant rigorous validation through prospective cohort studies.