The efficacy and safety of glucose control algorithms in intensive care: a pilot study of the Survival Using Glucose Algorithm Regulation (SUGAR) trial.

INTRODUCTION The benefits, harms and feasibility of intensive insulin therapy in critically ill patients remain unclear. Several single center studies have attempted to demonstrate the benefit of intensive insulin therapy in critically ill patients with variable results. OBJECTIVES We conducted a pilot randomized trial to assess the feasibility, safety and clinical outcomes of preprinted glucose management algorithms before the initiation of a large multicenter trial. PATIENTS AND METHODS Within 48 hours of admission to the intensive care unit, we randomized mechanically ventilated patients to either the "high" group (target serum glucose concentration 9-11 mmol/l) or the "low" group (target serum glucose concentration 5-7 mmol/l). To assess feasibility we measured the time to reach target glucose range, time in target range, morning glucose concentrations, average daily glucose concentrations, and number of crossovers. To assess safety, we measured the number of hypoglycemic events (serum glucose <2.2 mmol/l), and other serious adverse events such as cardiac arrests and seizures. RESULTS Sixty-eight patients were enrolled (35 in the high group and 33 in the low group). During the first week, the median proportions of time spent in the target range were 35.7% and 53.0% for the high and low groups, respectlively (p = 0.0001). Morning glucose concentrations were 8.3 +/-1.6 mmol/l and 6.2 -/+1.2 mmol/l. One (2.9%) and 8 (24.2%) episodes of hypoglycemia (<2.2 mmol/l) occurred in the high and low groups, reflecting 0.002 and 0.03 hypoglycemic events per patient-day, respectlively. CONCLUSIONS This pilot trial of intensive insulin therapy identified numerous challenges that helped in the preparation of an international multicenter randomized trial of intensive insulin therapy to evaluate benefits and harms.