914-P: Association of Sodium–Glucose Cotransporter 2 Inhibitors with Risk of Diabetic Ketoacidosis among Hospitalized Patients—A Multicentre Cohort Study

Importance: Sodium glucose co-transporter-2 (SGLT-2) inhibitors are being used among hospitalized patients, but there is no real-world empiric data on the risk of diabetic ketoacidosis (DKA) in-hospital. Objective: To assess the risk of DKA with SGLT-2 inhibitor use during hospital stay.

Design, Setting, Participants: We conducted a multi-centre cohort study of hospitalized patients in 19 hospitals in Canada between January 1, 2015 and December 31, 2021. We included patients over age 18 years with type 2 diabetes mellitus who received an SGLT-2 inhibitor or a dipeptidyl peptidase-4 (DPP-4) inhibitor in hospital. Exposure: SGLT-2 inhibitor

Comparator: DPP4-inhibitor

Main Outcomes and Measures: The primary outcome was the risk of DKA defined using American Diabetes Association (ADA) DKA criteria (pH <7.30, bicarbonate less than 18 mmol/l and positive ketones). Patients were followed from the date of admission to hospital to date of discharge or occurrence of DKA.

Results: We identified 11,098 patients who received an SGLT-2 inhibitor and 67,429 who received a DPP-4 inhibitor. The average age was 67.7 years (SD: 18.3 years), and 47.7% were women. In the matched population, the risk of DKA was 0.06% (N=6 events) over 5.2 hospital days with DPP-4 inhibitors and 0.19% (N=20 events) over 5.4 hospital days with SGLT-2 inhibitors. The relative risk was 3.33 (95% CI 1.33 - 8.30) for SGLT-2 inhibitors compared to DPP-4 inhibitors.

Conclusion and Relevance: Among hospitalized patients, the risk of DKA with use of SGLT-2 inhibitors was three-fold higher compared to DPP-4 inhibitors, but the absolute risk of DKA was low.