Trials of blood glucose control in critically ill patients have included relatively few patients who have diabetes. A randomized clinical trial was undertaken to determine whether an intervention to achieve a more liberal blood glucose range compared with usual care would reduce incident hypoglycemia in critically ill patients who have preexisting type 2 diabetes.

In this multicenter, parallel-group, open-label trial, 419 adults with type 2 diabetes who were expected to be in the ICU for at least three consecutive days were randomly assigned to either the intervention group (liberal glucose control), in which intravenous insulin was started at a blood glucose level of more than 252 mg/dL and titrated to a target range of 180 to 252 mg/dL, or the comparator group (usual care), in which intravenous insulin was started at a blood glucose level of more than 180 mg/dL and titrated to a target range of 108 to 180 mg/dL.

At least one episode of hypoglycemia occurred in 10 of the 210 patients (5%) in the intervention group compared with 38 of the 209 patients (18%) in the usual care group. Both mean blood glucose levels per patient per day and mean blood glucose levels over time were greater in the intervention group. The intervention group also had less glucose variability and less relative hypoglycemia than the usual care group. By day 90, 62 of the 210 patients (29.5%) in the intervention group and 52 of the 209 (24.9%) in the usual care group had died. Duration of the ICU admission, hospital admission, and ICU-free survival days were not significantly different in the two groups.

The study was open label, with the associated risk of bias. Other limitations were that only target blood glucose ranges were compared, and early termination of the trial was recommended. In addition, the trial was designed to detect differences in hypoglycemia, a biomarker of harm, rather than a patient-centered outcome.