Retrospective cohort study on absolute risks of diabetic ketoacidosis

Retrospective cohort study using patient records data from Canada and the UK between 2013 and 2018 compared the incidences of diabetic ketoacidosis in type 2 diabetes patients receiving SGLT2 (sodium glucose co-transporter 2) inhibitors vs. DPP4 (dipeptidyl peptidase 4) inhibitors.

Each of the 208,757 adult T2DM patients initiating an SGLT2 inhibitor was matched to a recipient of a DPP4 inhibitor (164,032 unique DPP4 inhibitor recipients owing to matching with replacement) using timeconditional propensity scores to form the two study cohorts.

Each participating site conducted the analyses (including descriptive analyses and Cox proportional hazards models) independently according to a common analytical protocol. Results from each site were further pooled by using random-effects models. Additional secondary analyses were performed stratifying by molecule, patient characteristics etc. as well as sensitivity analyses.

During a total of 370,454 person-years of follow-up, 521 patients were hospitalised with diabetic ketoacidosis, representing an overall incident rate of 1.41 per 1,000 person-years; with a higher incident rate in the SGLT2 vs. DPP4 exposure groups (2.03 vs. 0.75 per 1,000 person-years). Higher incidence rates were observed in all 3 available SGLT2 molecules suggesting a class effect that SGLT2 inhibitors are associated with higher risks for diabetic ketoacidosis compared with DPP4 inhibitors.

This study gives us the best estimate for the absolute risk of diabetic ketoacidosis to help us counsel our patients. The absolute risk is relatively small with an excess crude rate of 1.2 per 1000 patient years.