SUN-183 Prolonged Glycosuria after Canagliflozin Discontinuation in a Patient with Euglycemic Diabetic Ketoacidosis

Background: Euglycemic diabetic ketoacidosis (DKA) is an uncommon side effect of sodium-glucose cotransporter 2 (SGLT2) inhibitors. Prolonged glycosuria has been reported after discontinuation of SGLT2 inhibitors in patients treated for euglycemic DKA. Clinical Case: A 52-year old woman with type 2 DM, obesity (BMI 36 kg/m2) and tongue cancer was admitted for elective hemi-glossectomy. Type 2 DM was treated with Metformin and Glimepiride for about 6 years. She started taking Canagliflozin 4 days prior to the surgery because of uncontrolled T2DM (A1C 11.2%). She had never had an episode of DKA prior to the surgery. Canagliflozin was discontinued 2 days prior to the surgery. Her DM was treated with intravenous insulin infusion intraoperatively then subcutaneous insulin injections postoperatively. On postoperative day 2, she was found to have anion gap metabolic acidosis (PH 7.22, n 7.35 – 7.45; Anion gap 17 mmol/L, n < 16 mmol/L; bicarbonate 13 mmol/L, n 21 - 33 mmol/L) with a normal lactic acid level and an elevated beta-hydroxybutyrate (BHB) level (4.18 mmol/L, n < 0.27 mmol/L). Blood glucose was relatively controlled (120-206 mg/dL) so she was diagnosed with euglycemic DKA. It was successfully treated with intravenous insulin and glucose infusion. Because recurrent ketosis was reported previously, her urine glucose and serum BHB levels were monitored while she was treated with insulin infusion. In spite of relatively controlled serum glucose levels (120-209 mg/dL) on insulin drip, her urine glucose levels remained inappropriately high (≥ 500 mg/dL) for at least 9 days after Canagliflozin discontinuation. Quantified random urine glucose was 2646 mg/dL 4 days after the last dose of canagliflozin, and 246 mg/dL 10 days after the drug was discontinued. Her kidney function and urine output were normal during the hospital stay. Conclusion: Similar to previous case reports of SGLT2 inhibitors related euglycemic DKA, glycosuria persisted for at least 9 days after SGLT2 inhibitor discontinuation in this case despite very brief exposure. The duration of glycosuria was far longer than expected based on medication half-life. This raises the question whether there are genetic factors influencing SGLT2 inhibitor-induced glycosuria or SGLT2 inhibitor-related ketosis that need to be answered in further studies.