Liberal glucose control worked well for ICU patients with diabetes, study finds

ICU patients who were given insulin only when glucose reached 252 mg/dL had slightly less hypoglycemia and similar creatinine levels, white cell counts, and mortality rates to those treated with insulin at 180 mg/dL.

ICU patients with diabetes who were treated under a policy of liberal glucose control had slightly less hypoglycemia and similar outcomes to patients receiving conventional glucose control.

The prospective, open-label study included 700 patients with diabetes treated in a 22-bed mixed ICU at a tertiary hospital in Australia. A control population of 350 consecutive patients who had received conventional glucose control (insulin given when glucose >10 mmol/L [180 mg/dL] with a target of 6 to 10 mmol/L [108 to 180 mg/dL]) was compared to 350 consecutive patients admitted later and treated under a liberal glucose protocol (insulin given when glucose >14 mmol/L [252 mg/dL] with a target of 10 to 14 mmol/L [180 to 252 mg/dL]).

The median time-weighted average blood glucose was significantly higher in the liberal group than in the conventional group (11.0 mmol/L [198 mg/dL] vs. 9.6 mmol/L [173 mg/dL]; P<0.001). More conventional patients received insulin in the ICU (53.7% vs. 37.7%; P<0.001). The overall rate of hypoglycemia was 6.6% in the liberal group compared to 8.6% in the conventional patients (P=0.32), and this difference was greater among the 314 patients with HbA1c levels of 7% or greater (4.1% vs. 9.6%; P=0.053).

Otherwise, the groups did not show any significant differences: The trajectories of creatinine and white cell count were similar, and multivariable analyses showed no independent association between glucose control and mortality, duration of mechanical ventilation, or ICU-free days to day 30. The study also did not identify any signals of harm from liberal control in any patient subgroup. Results were published by Critical Care Medicine on March 6.

Many of the previous studies comparing liberal and conventional glucose control in the ICU have not focused specifically on patients with diabetes, the study authors said. However, prior research has noted that patients with elevated glucose in the ICU and HbA1c of 7% or greater had lower mortality than similar patients with HbA1c under 7%. “In their aggregate, these studies suggest that ‘permissive hyperglycemia’ in ICU patients with diabetes, who are likely adapted to some degree of chronic hyperglycemia, is safe. Our study, the first large interventional comparison between liberal and conventional targets, is consistent with such findings,” the authors wrote.

They did note several limitations to the study, including that it was not a randomized controlled trial and that it was conducted at a single center. Still, they concluded that “the findings support further cautious investigation of liberal glycemic control in ICU patients with diabetes.”