Evaluation of Insulin Dispensing Practices on Time to Administration for Hyperkalemia in the ICU

Introduction: Hyperkalemia is a life-threatening condition often requiring rapid treatment with intravenous (IV) regular insulin. In April 2025, one of the towers at our institution changed how IV insulin for hyperkalemia was dispensed, from one-time syringe doses dispensed from the central pharmacy to prepared 5 unit syringes stocked in the intensive care unit (ICU) automated dispensing cabinets (ADCs). The other hospital tower, however, has only stocked insulin vials in the ICU ADC to be drawn up at the bedside. This quality improvement project aimed to compare how dispensing methods affected time to insulin administration and safety outcomes for the treatment of hyperkalemia in the ICU.


Methods: This retrospective cohort study included adult ICU patients admitted between 3/3/2025 to 6/21/2025 who received IV regular insulin for hyperkalemia. The three groups included insulin vials from the ADC, one-time doses from central pharmacy, and ADC stocked syringes.


Results: This study included 44 patients with 78 total insulin orders, 26 orders per group. Median age ranged from 59 to 61 years old, and 29 patients (66%) were male. The primary outcome of median time from insulin order placement to administration was -4 minutes for the vial group (IQR -19.8 to 18.8), -1 minute for the central pharmacy group (IQR -7.8 to 23.8), and -8 minutes for the stocked syringe group (IQR -17 to 3.5). For secondary outcomes, median time from potassium result to insulin administration was 2.1 hours in the vial group (IQR 1.4 to 3.1), 2.2 hours in the central pharmacy group (IQR 1.7 to 2.9), and 1.9 hours in the stocked syringe group (IQR 1.4 to 2.4). Re-administration of insulin occurred in 6 patients (46%) in the vial group, 2 (13%) in the central pharmacy group, and 5 (33%) in the stocked syringe group. Hypoglycemia occurred after 3 insulin administrations (12%) in the vial group, 4 (15%) in the central pharmacy group, and 3 (12%) in the stocked syringe group.


Conclusions: No clinically significant time differences were found between groups, though the stocked syringe group showed slight reductions from order and potassium result to insulin administration. This suggests stocked syringes may shorten time to administration with similar rates of hypoglycemia to the other groups, justifying this dispensing practice change.