OBJECTIVE Sodium-glucose cotransporter 2 inhibitors (SGLT2i) are associated with euglycemic diabetic ketoacidosis (eDKA) in patients with diabetes (DM), but the risk of euglycemic ketoacidosis (eKA) in patients without diabetes (NDM) remains uncertain. We aimed to compare the incidence and clinical characteristics of eKA between patients with and without diabetes receiving SGLT2i therapy. METHODS Retrospective chart review of all hospital admissions (N=249,517 representing 183,835 unique individuals) across 28 hospitals in a large health system over one year. Cases meeting criteria for eKA as per American Diabetes Association criteria while taking an SGLT2i were reviewed to compare DM and NDM patients. RESULTS Among 894 admissions with eKA, 59 (6.6%) were receiving SGLT2i therapy: 55 patients with diabetes (DM) and 4 patients without diabetes (NDM). NDM patients had nonsignificant trend to higher median serum bicarbonate 18.5 vs 16 mmol/L (p=0.088), and lower anion gap 16 vs 21 mEq/L (p=0.093) but significantly higher nadir bicarbonate at 48 hours of 18.5 vs 14.0 mmol/L (p=0.004). EKA was unrecognized in all NDM patients compared to 55% of DM patients (p=0.045). Based on outpatient prescriptions, the estimated annual incidence of eKA associated with hospitalization was 0.36% in DM and 0.11% in NDM SGLT2i users. CONCLUSIONS SGLT2i-associated eKA appears quite rare in NDM patients and eKA while on SGLT2i, whether in DM or NDM, represents a small fraction of all eKA admissions. NDM cases of SGLT2i-associated eKA may be less severe than in patients with DM, though a larger group of patients is needed to confirm this observation.
