Diabetic ketoacidosis in patients with type 2 diabetes mellitus receiving sodium-glucose co-transporter-2 inhibitors: The Hong Kong Diabetes Study

Background Sodium-glucose cotransporter-2 inhibitors (SGLT2i) have been associated with increased risks of diabetic ketoacidosis in some studies, but not others. Also, few studies have analysed the characteristics of DKA cases nor have they developed risk models to predict DKA. Methods This was a population-based cohort study of patients with type 2 diabetes mellitus (T2DM) who were prescribed a SGLT2i between 1st January 2015 and 31st December 2020, with follow-up until 2022. Results A total of 28,814 patients (59.54% male, mean age 57.9±11.3 years old) were included. Over 159,000 person-years, 494 patients suffered from DKA and 855 patients died. Multivariable Cox regression identified seven significant predictors of incident DKA: need for financial aid, prior DKA, insulin use, acarbose use, time-weighted mean of HbA1c, fasting glucose and coefficient of variation for fasting glucose. The same predictors remained significant after excluding patients with prior DKA before SGLT2I use. Altogether, it was possible develop a multivariable score-based system, with weightings for risk stratification of incident DKA. Conclusions Significant predictors of DKA of SGLT2I use were identified with development of a score-based system for risk stratification. This enables clinicians to identify high-risk patients early and monitor them more closely. Multivariable Cox regression Risk score for DKA prediction