Euglycemic diabetic ketoacidosis: The paradox of delayed correction of acidosis

Objective: The effectiveness of standard treatment for diabetic ketoacidosis (DKA) in "euglycemic DKA" (EuDKA, blood glucose (BG) <= 250 mg/dL) was evaluated with respect to the time to correction of BG <= 200 mg/dL, anion gap (AG)<= 12 mmol/L, and serum bicarbonate [HCO3] >= 18 mmol/L. Methods: Data were retrieved from an electronic health record (EPIC) for "diabetic ketoacidosis." Patients were categorized by initial BG as EuDKA, middle range DKA (MrDKA, >250 < 600 mg/dL) and hyperosmolar DKA (HyperDKA >= 600 mg/dL). Results: There were 56 patients (27men, 29women; age 45.8 +/- 15.6 (SD) years. The initial 8-h insulin infusion rate (0.05 +/- 0.02, 0.09 +/- 0.03, 0.14 +/- 0.05units/kg/h, p < 0.001) and the time to correction of BG (3.4 +/- 1.9, 6.1 +/- 2.9 and 9.6 +/- 3.9 h, p < 0.001), differed for EuDKA, MrDKA and HyperDKA. There were no differences in the time to correction of AG or [HCO3]. The earlier time to correction of BG in EuDKA resulted in paradoxical longer lag times for correction of [HCO3] (p = 0.003) and AG (p = 0.048). Changes in BG, AG and [HCO3] correlated with insulin infusion rates of 0.08-0.1units/kg/h whereas in EuDKA the insulin infusion rate was 0.05 +/- 0.02 units/kg/h. Conclusion: In EuDKA, correlation analyses suggest that higher glucose and insulin infusion rates than what would be projected for the level of blood glucose are required to reverse ketoacidosis. Prospective trials are required to optimize the levels of glucose and insulin infusions in EuDKA.