A. Rai, J. Marshall, S. Nandyala, M. Her, AA. Agan, TY. Huang, C. Rodriguez-Watson, A. Clary, B. Diessner, MB. Nolan, DA. Djibo, A. DeVries, K. Daniels, X. Zhang, T. Wang, I. Gantz, R. Shankar, MM. Zale, P. Ejelonu, R. Frederich, U. Masiukiewicz, et al. (2026). Post-authorization safety study to assess the risk of diabetic ketoacidosis among type 2 diabetes mellitus patients treated with ertugliflozin compared to patients treated with other antihyperglycaemic agents in a Medicare and Medicaid population. Diabetic medicine : a journal of the British Diabetic Association, 43(4), e70208. https://doi.org/10.1111/dme.70208

PubMed 41662384

[Abstract]
AIMS: To evaluate diabetic ketoacidosis (DKA) risk among new users of ertugliflozin versus sulfonylureas (SU) or thiazolidinediones (TZD) and incretin-based drugs in patients with type 2 diabetes.

METHODS: We used Medicare and Medicaid fee-for-service adjudicated claims within the Innovation in Medical Evidence and Development Surveillance network to identify three new-user cohorts: (1) ertugliflozin; (2) SU/TZD; and (3) incretin-based drugs. The outcome was a principal hospital discharge diagnosis for DKA. Adjusted hazard ratios (HRs) were estimated after 1:1 propensity score (PS) matching, separately for ertugliflozin vs. SU/TZD and vs. incretin-based drugs. Subgroup analyses were performed based on baseline insulin use.

RESULTS: After PS matching, baseline characteristics were similarly distributed in each cohort. For ertugliflozin (n = 42,907) vs. SU/TZD (n = 42,907), the incidence rates of DKA per 1000 person-years (PY) were 2.95 and 1.49, respectively. For ertugliflozin (n = 42,247) vs. incretin-based drugs (n = 42,247), the incidence rates of DKA per 1000 PY were 2.76 and 1.06, respectively. For ertugliflozin vs. SU/TZD, the HR [95% confidence interval (CI)] was 1.88 [1.17-3.02]; in non-insulin users, 2.34 [1.27-4.31]; and in insulin users, 1.17 [0.54-2.52]. For ertugliflozin vs. incretin-based drugs, the HR [95% CI] was 2.40 [1.40-4.11]; in non-insulin users, 2.84 [1.42-5.66]; and in insulin users, 1.87 [0.79-4.46].

CONCLUSIONS: Ertugliflozin was associated with a higher risk of DKA relative to comparators. HRs were higher among new users with no-concomitant insulin use than those with concomitant insulin use. Results were consistent with prior SGLT2i data and highlighted the importance of caution by both patients and physicians.