Diabetic Ketoacidosis Risk in Sodium-Glucose Cotransporter 2 Inhibitors versus Glucagon-Like Peptide-1 Receptor Agonists Initiators with CKD Stages 3–4 and Type 2 Diabetes

KEY POINTS: The risk of diabetic ketoacidosis associated with sodium-glucose cotransporter 2 inhibitors among patients with CKD stages 3-4 and type 2 diabetes is not well established. We found that the overall incidence of hospitalization with diabetic ketoacidosis remained low among sodium-glucose cotransporter 2 inhibitors and glucagon-like peptide-1 receptor agonists initiators with CKD stages 3-4 and type 2 diabetes. sodium-glucose cotransporter 2 inhibitors was associated with a 40% higher risk of hospitalization with diabetic ketoacidosis compared with glucagon-like peptide-1 receptor agonists initiation. BACKGROUND: sodium-glucose cotransporter 2 inhibitors (SGLT2i) and glucagon-like peptide-1 receptor agonists (GLP-1 RA) are effective in managing CKD and type 2 diabetes (T2D). However, there are concerns about an increased risk of diabetic ketoacidosis (DKA) associated with the use of SGLT2i, particularly in patients with CKD. METHODS: The is a population-based, new-user, active comparator cohort study comparing SGLT2i versus GLP-1 RA initiators using three US health care databases: Optum's deidentified Clinformatics Data Mart (2014-2024), Merative MarketScan (2014-2022), and Medicare Fee-for-Service (2014-2020). The exposure is initiation of either SGLT2i or GLP-1 RA, defined as a new prescription without prior use in 365 days. The primary outcome was hospitalization with DKA identified using inpatient International Classification of Diseases, Ninth/Tenth Revision codes. Incidence rates, rate differences, and hazard ratios were calculated. RESULTS: After 1:1 propensity score matching, the study population included 143,858 patients, 71,929 in the SGLT2i arm, and 71,929 in the GLP-1 RA arm. The mean age (SD) was 71.28 (8.16) years, and 48.8% were female. Patients initiating SGLT2i had higher risk of hospitalization with DKA compared with GLP-1 RA initiators (incidence rate 4.37 versus 3.13 events per 1000 person-years; rate differences 1.23 [95% confidence interval, 0.54 to 1.92]; hazard ratios 1.40 [95% confidence interval, 1.16 to 1.68]). The number needed to harm was 813 per 1000 person-years for SGLT2i compared with GLP-1 RA initiation. Results remained consistent in subgroup analyses stratified by age (younger than 70 versus 70 years or older), sex, body mass index (