SUN-178 A Not So "Sweet" DKA: A Case of Euglycemic Diabetic Ketoacidosis Caused by Dapagliflozin/Liraglutide Combination

Background: Combined use of Glucagon-like peptide-1 receptor agonists (GLP-1RA) & sodium-glucose cotransporter 2 inhibitors (SGLT2i) may provide particular advantages for treating type 2 diabetes mellitus (DM2). Both promote weight loss and have favorable effects on cardiovascular outcomes while minimizing hypoglycemic events, however, in some circumstances their combination might result in previously unexpected side effects. We describe a case where initiation of a GLP1RA appeared to induce euglycemic diabetic ketoacidosis (euDKA) in a patient taking an SGLT2i. Clinical Case: A 52-year-old woman with hyperlipidemia and DM2 presented with 3 days of nausea & vomiting. Approximately 1 week prior she started liraglutide due to suboptimal glycemic control on dapagliflozin (HbA1c 8%). She reported no fevers, chills, cough, dysuria, diarrhea or sick contacts. She had headache without meningeal signs. Laboratory findings were significant for leukocytosis (15.400/µL), low bicarbonate (12 mmol/L), hyperglycemia (206 mg/dL), low lactate (0.9 mmol/L) and glucosuria/ketonuria (+4). Venous blood gas revealed metabolic acidosis with pH 7.18, pCO(2) 37, HCO(3) 13.4. Based on these findings, the patient was diagnosed with euDKA due to dapagliflozin in the setting of liraglutide associated nausea. Insulin drip and IV fluids were initiated with rapid resolution of the metabolic disturbance. No infectious process was identified. Prior to starting dapagliflozin 10 months ago, she was on metformin, long acting and pre-meal insulin. She had ketoacidosis 8 months prior when liraglutide was first prescribed, but at the time it was attributed to infection in the setting of dapagliflozin, liraglutide and metformin use, and metformin and liraglutide were stopped, but dapagliflozin was continued. Conclusion: EuDKA is defined as diabetic ketoacidosis without marked hyperglycemia, generally <300 mg/dL. The association between euDKA and SGLT2i is presumably due to increased urinary excretion of glucose with diminished glycogen stores, compounded by increased ketone production and impaired excretion. GLP-1RA can augment glucose stimulated insulin secretion but if they produce significant and prolonged nausea and vomiting, ketogenesis may occur. In this case initiation of liraglutide seemed to be the factor that precipitated euDKA on our patient. Clinicians must be aware of the association between SGLT2i and euDKA- an entity with clinically distinct presentation. Recognizing risk factors, such as marked β-cell insufficiency, prolonged starvation, illness, will be key to early detection, treatment and recommendations to suspend SGLT2i use. Combined use of SGLT2i and GLP-1RA requires additional research, as well as identifying characteristics of the patients prone to euDKA.