A phase II randomised controlled trial of intensive insulin therapy in general intensive care patients

To determine the safety and efficacy of an intensive insulin regimen compared with a conventional insulin regimen in general intensive care unit patients.

A phase II, randomised controlled trial was conducted in 70 critically ill patients in a closed multidisciplinary ICU of a university-affiliated tertiary hospital. We assessed patient characteristics at baseline. Trial process measures included number of blood glucose measurements per day and number in target range, type and quantity of caloric intake, patient outcome and insulin dosing. The primary outcome was the median blood glucose concentration. Secondary outcome measures were incidence of hypoglycaemia (blood glucose level < 2.2 mmol/L), clinical sequelae of hypoglycaemia and hospital mortality.

Thirty-five patients were randomised to each of the two groups. More blood glucose samples were taken per day in the intensive insulin group (16 versus 9), but the number of samples in the normoglycaemic range was 48.5%, compared with 79.8% within the target glucose range in the conventional insulin group. The median (interquartile range) blood glucose concentrations in the intensive and conventional insulin therapy groups were 5.4 (5.1-5.7) mmol/L and 7.9 (7.2-9.0) mmol/L, respectively (difference, 2.5 mmol/L; P < 0.0001). Five patients (14.3%) in the intensive insulin therapy group became hypoglycaemic versus none in the conventional insulin therapy group. There were no detected clinical sequelae of hypoglycaemia.

The intensive insulin regimen was effective in achieving the target blood glucose concentration, with clear separation from the conventional insulin regimen. Although the incidence of hypoglycaemia was increased, there was no detectable harm.