Background: Glycemic control is crucial in the intensive care setting to reduce morbidity and adverse outcomes.1 The debate over methods for achieving this control highlights the risks of hypoglycemia and unclear management parameters.2 Continuous glucose monitoring software, like Glucommander, helps manage blood sugar safely and prevents hypoglycemic events.3,4 Generally, maintaining euglycemia with insulin improves outcomes in critically ill patients.5 The standard for managing diabetic ketoacidosis (DKA) involves transitioning from intravenous (IV) to subcutaneous (SQ) insulin once DKA resolves.6 Recent pediatric data indicate that earlier glargine administration may facilitate faster resolution of DKA.6 A quality improvement initiative redesigned protocols for meal delivery, blood glucose monitoring, and insulin administration, achieved an 8% improvement in glycemic control and faster insulin administration.7Enhancing care for critically ill patients requires addressing delays in administration of glargine during the transition from continuous drip. Each institution faces unique challenges in similar initiatives. At a university-affiliated hospital, delays were previously caused by the central pharmacy’s lengthy distribution process. Relocating glargine to the ICU pyxis distribution system, is proposed to reduce time from order to administration without causing adverse glycemic outcomes.
Methods: A retrospective review used Glucommander and Powerchart EMR data for ICU patients aged 18 and older with DKA transitioning from IV insulin to glargine. Key data points included IV drip start time, time of glargine order and administration, and instances of hypoglycemia (≤69 & ≤54) and hyperglycemia (≥180 & ≥300) within five days of transitioning from the IV insulin drip. Two reviewers obtained and validated the data. The study compared two cohorts: one before and one after relocating glargine from the hospital’s central pharmacy to the ICU pyxis. Data analysis employed box and whisker plots to identify and remove outliers greater than 1.5 times the interquartile range (IQR) beyond the 75th percentile for both cohorts and those that were impossibly short.
Results: After removing outliers for both cohorts, samples size was N = 40 for each. The time from order to administration of glargine was calculated for pre- and post-intervention cohorts as 79.5 min (95% CI ± 13 min) and 43.2 minutes (95% CI ± 10 min) respectively. There was a statistically significant difference between these groups (p< 0.0001) with a mean difference of 36.3 min (95% CI ± 16.3 min). The mean time from IV insulin start time to glargine administration for pre- and post-intervention cohorts was 1420.8 min. (95% CI ± 241.5 min) and 1335.9 min. (95% CI ± 185.6 min) respectively with no significant difference between the groups (p-value = 0.6265). Additionally, no significant differences were observed for the number of adverse glycemic events following transition subcutaneous insulin.
Conclusions: Relocating glargine from the central pharmacy to the ICU Pyxis system significantly reduced the time from glargine order to administration, without increasing the risk of hyper- or hypoglycemia after transitioning from IV insulin.