Finding the sweet spot: Identification of optimal glucose levels in critically injured patients

Background: Conflicting data exist regarding optimal glycemic control in critically ill trauma patients. We therefore compared glucose parameters and outcomes among three different glycemic control regimens in a single trauma intensive care unit (ICU), hypothesizing that a moderate regimen would yield optimal avoidance of hyper- and hypoglycemia with equivalent outcomes when compared with a more aggressive approach. Methods: We retrospectively reviewed 1,422 trauma patients with at least 3-day ICU stay and five glucose measurements from May 2001 to January 2010, spanning three nonoverlapping, sequential glucose control protocols: "relaxed," "aggressive," and "moderate." For each, we extracted mean blood glucose, hypoglycemic and hyperglycemic event frequency, and glucose variability and investigated their association with outcomes. Results: Mortality was associated with elevated mean glucose (135.6 mg/dL vs. 126.2 mg/dL), more frequent hypoglycemic (2.67 ± 7 vs. 1.28 ± 5) and hyperglycemic (30.6 ± 28 vs. 16.0 ± 22 per 100 patient-ICU days) events, and higher glucose variability (37.1 ± 20 vs. 29.4 ± 20; all p < 0.001). Regression identified hyperglycemic episodes (p < 0.05) as an independent predictor of mortality. The "moderate" regimen had rare hyperglycemia, low glucose variability, and intermediate mean blood glucose range and frequency of hypoglycemia. Multiorgan failure and mortality did not differ between groups. Conclusions: Hyperglycemic events (glucose >180 mg/dL) most strongly predicted mortality. Of glucose control protocols analyzed, the "moderate" protocol had fewest hyperglycemic events. As outcomes were otherwise equivalent between "moderate" and "aggressive" protocols, we conclude that hyperglycemia can be safely avoided using a moderate glycemic control protocol without inducing hypoglycemia.