Evaluation of a Subcutaneous Insulin Protocol for Management of Diabetic Ketoacidosis in the Emergency Department

Diabetic ketoacidosis (DKA) is a common, yet critical complication seen in type 1 and type 2 diabetes. Treatment traditionally begins with an insulin bolus of 0.1 units/kg followed by a continuous intravenous (IV) infusion at 0.1 units/kg/hr. In 2024, the American Diabetes Association updated its guidelines to include subcutaneous (SQ) insulin as a treatment option for uncomplicated mild-to-moderate DKA. This quality improvement (QI) project was performed to determine the efficacy and safety of Northwestern Medicine’s (NM) SQ insulin protocol for mild-to-moderate DKA.

This single-center, retrospective QI project included patients treated in the emergency department (ED) at Northwestern Memorial Hospital from November 5, 2023 to August 31, 2025. Eligibility included ≥18 years of age with a blood glucose > 200 mg/dL, pH ≤ 7.3, bicarbonate ≤ 18 mg/dL, and received a dose of insulin lispro SQ and insulin glargine SQ per protocol. Individuals were excluded if IV insulin was administered initially or did not meet defining criteria of DKA. Primary outcome was intensive care unit (ICU) admission due to initiation of IV insulin infusion. Secondary outcomes included time to DKA resolution, hypoglycemia, and disposition. Descriptive statistics were utilized to analyze data.

Among the 28 patients included, the mean age was 48.4 years, 64.3% (n= 18) had mild DKA, the median initial lispro dose was 0.19 units/kg, and the median initial glargine dose was 0.2 units/kg. ICU admission due to initiation of IV insulin infusion occurred in 7.1% (n=2) of patients. The median time to DKA resolution was 16.3 hours (IQR 7.5-36.5) and hypoglycemia occurred in 10.7% (n= 3) of patients. Regarding disposition from the ED, 7.1% (n=2) went to the ICU, 14.3% (n=4) were discharged home, and 78.6% (n=22) were admitted to the floor.

Among patients treated with SQ insulin for the management of DKA in the ED, the majority were admitted to the floor, and several were discharged from the ED. This project demonstrates that the use of SQ insulin for mild-to moderate DKA is a safe and effective method.