Tight blood glucose control: what is the evidence?

In two large, randomized studies, maintenance of normoglycemia with intensive insulin therapy largely prevented morbidity and reduced mortality of critically ill patients. Recently, questions have been raised about the efficacy and safety of this therapy. These issues are systematically addressed and discussed with the evidence available from these and other studies. The available studies show that an absolute reduction in risk of hospital death of 3% to 4% is to be expected from intensive insulin therapy in an intention-to-treat analysis. Future studies designed to confirm a statistically significant survival benefit should be adequately powered, with inclusion of >or=5,000 patients. When patients are treated with intensive insulin therapy for >3 days, the absolute reduction in the risk of death increases to approximately 8%. Insulin therapy also improves morbidity, more so when continued for >3 days. Strict blood glucose control to normoglycemia (<110 mg/dL) is required to obtain the most clinical benefit, but this inherently increases the risk of hypoglycemia. It remains unclear whether short hypoglycemic episodes are truly harmful for these patients. In conclusion, demonstration of the clinical benefits of intensive insulin therapy depends on the quality of blood glucose control and the statistical power of the studies.