Level 2: Mid-level evidence
Intensive glucose control has been suggested for critically ill patients to improve outcomes. Two recent reports, a systematic review and a new randomized trial, found this treatment was associated with increased hypoglycemia and unclear benefits on mortality. A review of 26 randomized trials compared intensive glucose control with conventional glucose management in 13,567 ICU patients. There were no significant differences in overall mortality (24.7% vs. 24.9%). Intensive glucose control was associated with lower mortality in surgical ICU patients (7.4% vs. 11.8%, P <.05, NNT 23) in a meta-analysis of five trials. However, there was no difference in mortality in medical ICU patients (34.9% vs. 36.7%) or in mixed ICU patients (26.7% vs. 25.6%). Hypoglycemia was significantly increased in intensive glucose control groups (10.7% vs. 1.6%, P <.05, NNH 10). The prevalence of diabetes was not reported (CMAJ. 2009;180:821-827; available at www.cmaj.ca/cgi/content/full/180/8/821, accessed June 11, 2009).

In the Normoglycemia in Intensive Care Evaluation—Survival Using Glucose Algorithm Regulation (NICE-SUGAR) trial, 6,104 patients (20% with history of diabetes) were randomized within 24 hours of ICU admission to intensive or conventional glucose control (N Engl J Med. 2009; 360:1283-1297; available at content.nejm.org/cgi/content /full/360/13/1283, accessed June 11, 2009). In this trial, intensive control was associated with both higher mortality (27.5% vs. 24.9%, P =.02, NNH 38) and a higher rate of severe hypoglycemia (6.8% vs. 0.5% (P <.001, NNH 15). In a subgroup analysis, surgical patients had higher mortality with intensive control (odds ratio 1.31, 95% CI 1.07-1.61).