Intensive insulin therapy and mortality in critically ill patients

Introduction Intensive insulin therapy ( IIT) with tight glycemic control may reduce mortality and morbidity in critically ill patients and has been widely adopted in practice throughout the world. However, there is only one randomized controlled trial showing unequivocal benefit to this approach and that study population was dominated by post- cardiac surgery patients. We aimed to determine the association between IIT and mortality in a mixed population of critically ill patients. Methods We conducted a cohort study comparing three consecutive time periods before and after IIT protocol implementation in a Level 1 trauma center: period I ( no protocol); period II, target glucose 80 to 130 mg/ dL; and period III, target glucose 80 to 110 mg/ dL. Subjects were 10,456 patients admitted to intensive care units ( ICUs) between 1 March 2001 and 28 February 2005. The main study endpoints were ICU and hospital mortality, Sequential Organ Failure Assessment score, and occurrence of hypoglycemia. Multivariable regression analysis was used to evaluate mortality and organ dysfunction during periods II and III relative to period I. Results Insulin administration increased over time ( 9% period I, 25% period II, and 42% period III). Nonetheless, patients in period III had a tendency toward higher adjusted hospital mortality ( odds ratio [ OR] 1.15, 95% confidence interval [ CI] 0.98, 1.35) than patients in period I. Excess hospital mortality in period III was present primarily in patients with an ICU length of stay of 3 days or less ( OR 1.47, 95% CI 1.11, 1.93 There was an approximately fourfold increase in the incidence of hypoglycemia from periods I to III. Conclusion A policy of IIT in a group of ICUs from a single institution was not associated with a decrease in hospital mortality. These results, combined with the findings from several recent randomized trials, suggest that further study is needed prior to widespread implementation of IIT in critically ill patients.