AIMS: Sodium-glucose cotransporter 2 (SGLT2) inhibitors increase ketoacidosis risk, limiting their use in type 1 diabetes. To better understand the pathophysiology of SGLT2 inhibitor-mediated ketoacidosis, we measured blood glucose, capillary blood and plasma β-hydroxybutyrate (BOHB) and breath acetone (BrACE) during supervised insulin withdrawal in adults with type 1 diabetes with and without dapagliflozin treatment. MATERIALS AND METHODS: Twenty adults with type 1 diabetes underwent supervised insulin withdrawal twice in a randomized crossover design: during usual care and after treatment with dapagliflozin (10 mg daily for 2 weeks plus the test day). After insulin withdrawal, capillary blood glucose, BOHB and BrACE measurements were obtained at least hourly until stopping rules were met (>
8 h elapsed, symptoms of ketosis, glucose >
400 mg/dL, BOHB >
4 mmol/L or participant request). RESULTS: The peak BOHB and BrACE values achieved during supervised insulin withdrawal were both greater with dapagliflozin than with usual care. Throughout the insulin withdrawal study, dapagliflozin treatment was associated with significantly greater BOHB and BrACE concentrations. The proportions of participants reaching BOHB >
1.5 mmol/L and >
2.5 mmol/L during supervised insulin withdrawal were greater in the dapagliflozin arm. Blood glucose reached a lower peak in the dapagliflozin arm. CONCLUSIONS: In adults with type 1 diabetes undergoing supervised insulin withdrawal, dapagliflozin treatment compared to usual care was associated with greater blood and breath ketone concentrations in the absence of significant hyperglycaemia.