Efficacy and safety of intensive insulin therapy for critically ill neurologic patients: a meta-analysis.

摘要

BACKGROUND: : Whether intensive insulin therapy (IIT) may improve clinical outcomes for patients admitted to intensive care units, especially critically ill neurologic patients, is still debated. In the present study, we performed a meta-analysis of literature comparing the efficacy and safety of IIT and conventional insulin therapy (CIT) for critically ill neurologic patients in terms of mortality, infection rate, neurologic outcome, and hypoglycemia. METHODS: : We searched for published reports of studies of randomized control trials (up to March 10, 2011) of patients admitted to neurologic intensive care units and investigated an IIT (target of blood glucose control <120 mg/dL) with a control of CIT. Data were abstracted by a standardized protocol. RESULTS: : We retrieved reports of five studies involving 924 patients. The risk of mortality, infection rate, and neurologic outcome did not differ with IIT or CIT. However, the incidence of hypoglycemic episodes was significantly higher with IIT than CIT (78.8% vs. 48.9%), with a relative risk of 2.62 (95% confidence interval [CI]: 1.07-6.43; p < 0.04). CONCLUSIONS: : As compared with CIT, IIT may not benefit critically ill neurologic patients in terms of mortality, infection rate, or neurologic outcome and in fact may be associated with increased hypoglycemic complications. Therefore, IIT cannot be recommended over conventional control for critical neurologic disease, but further study is warranted.