Presented at the 41st Annual International Society for Pharmacoepidemiology Conference
Background: Use of sodium-glucose cotransporter 2 inhibitors (SGLT2i) including ertugliflozin for type 2 diabetes mellitus (T2DM) has grown in the past decade, accompanied by efforts to understand the associated diabetic ketoacidosis (DKA) risk for this drug class.
Objectives: To evaluate DKA risk among new users of ertugliflozin relative to 1) sulfonylureas (SU) or thiazolidinediones (TZD), and 2) incretin-based drugs (dipeptidyl peptidase 4 inhibitors or glucagon-like peptide-1 receptor agonists).
Methods: This retrospective cohort study used Medicare and Medicaid fee-for-service adjudicated claims in the Innovation in Medical Evidence and Development Surveillance network. Three new user cohorts were identified: (1) ertugliflozin; (2) SU/TZD; and (3) incretin-based drugs. Eligibility included: initiation of the cohort-defining drug(s) between 7/1/2018-12/31/2021 (index exposure); age ≥18 at initiation date (index date); no use of cohort defining drug or non-ertugliflozin SGLT2i in the 6 months before the index date; no DKA ever before the index date; no initiation of comparator or first ever insulin use on the index date; and ≥1 T2DM diagnosis without type 1 or gestational diabetes before or on the index date. Follow-up ended at the earliest of index exposure stoppage; comparator or non-ertugliflozin SGLT2i use; first ever insulin use; disenrollment; data end; or death. The outcome was principal hospital discharge diagnosis of DKA. Adjusted hazard ratios (aHRs) were estimated using Cox proportional hazards models after 1:1 propensity score matching (PSM), separately for ertugliflozin vs. SU/TZD and vs. incretin-based drugs. Subgroup analyses were performed based on baseline insulin use.
Results: For ertugliflozin vs. SU/TZD, there were 42,288 new users of ertugliflozin with 41 DKA events (2.93/ 1,000 person-years [PYs]); and 835,324 new users of SU/TZD with 612 events (1.04 / 1,000 PYs). For ertugliflozin vs. incretin-based drugs, there were 41,407 new users of ertugliflozin with 37 DKA events (2.76/1,000 PYs); and 789,956 new users of incretin-based drugs with 613 events (1.28/1,000 PYs). After PSM, characteristics were similarly distributed in each cohort (mean age 53, female sex 55%, insulin use 26%). For ertugliflozin vs. SU/TZD, the aHR [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 aHR [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].
Conclusion: Ertugliflozin was associated with a higher risk of DKA relative to comparator drugs. The results were consistent with prior SGLT2i data and highlight the importance of caution by both patients and physicians.
