Background: Intensive insulin therapy (IIT) is the standard of care in the ICU, but precise implementation of insulin protocols has been difficult in clinical practice. The authors' objective was to quantify adherence to an IIT protocol in a practice setting, and to describe how adherence impacts overall blood glucose (BG) control.
Methods: A retrospective analysis of a cohort of critically ill patients treated with IIT was performed. Protocol adherence was evaluated by assessing the timing of BG measurements. Each measurement was categorized according to the time from the previous reading: early (<1 hour), on time (1-3 hours), and late (>3 hours). Outcome measures included mean and median BG for each time category as well as the proportion of values within the target range.
Results: In 1106 trauma and surgical ICU patients, 54,139 measurements were available for analysis. The overall mean BG (116 mg/dL) was near the target (80-110 mg/dL), but only 46% of values were within this range. There were 45,806 (86%) measurements on time, 2749 (5%) early, and 4478 (9%) were late. BG values of late measurements were less likely to be within range (34% vs 46% for on time measurements, P<.001). Of late measurements, 19% were >200 mg/dL, 13% were 150-200 mg/dL, and 16% were <60 mg/dL.
Conclusions: IIT is difficult to implement precisely in a complex ICU environment. Measurement timing impacts overall BG control, with late measurements more often associated with severe hyperglycemic (BG>150 mg/dL) and hypoglycemic (BG<60 mg/dL) episodes.