Background: Tight glycemic control in a mixed surgical intensive care unit patient population has been associated with improved survival. We postulated targeted therapy to achieve glucose levels <140 mg/dL would reduce infectious complications and mortality in trauma patients admitted to the intensive care unit (ICU).
Methods: Adult trauma patients admitted to our American College of Surgeons Level I Trauma Center ICU from July 2004 through June 30, 2006 were studied. Insulin therapy was instituted for ICU patients admitted after July 1, 2005 with glucose >140 mg/dL. Data on infections and all glucose values were collected. Multivariate analysis adjusting for age, Injury Severity Score, Glasgow Coma Scale Score, admit blood pressure, and intubation status was performed.
Results: Five thirty-one ICU patients were admitted with a mean Injury Severity Score of 23 +/- 13 and mean age of 45 years +/- 19 years. The admission, mean, and maximum glucoses were 141, 129, and 192 respectively. In multivariate analyses, increases in all three glucose values were associated with a significantly higher mortality, with the best model achieved using mean glucose with a receiver operating curve of 0.90. For mean glucose categories of >200 mg/dL, 141 mg/dL to 200 mg/dL, and </=140 mg/dL, the mortality was 40%, 20%, and 3.3%, respectively. Higher glucose levels were not associated with increased rates of infection after risk adjustment. After July 1, 2005, the use of insulin drips rose from 13% to 22% (p = 0.01), and the number of glucose checks per patient in the ICU rose from 27 to 43 (p < 0.02), and the percent of ICU patients with all glucose values less than 140 mg/dL rose from 59% to 78%.
Conclusion: Higher glucose levels were significantly associated with increased risk of fatal outcome in trauma patients. Hyperglycemia was not an independent predictor of infectious complications. Despite the increased use of insulin drips and the higher number of glucose checks after adopting a stricter insulin treatment protocol, ICU outcomes remained unchanged.