Eleven studies (n=12,342, range 70 to 7,152) were included in the review: seven randomised controlled trials (RCTs), one non-randomised controlled trial, one consecutive-series historically controlled trial, one observational study and one cohort study.
Six studies assessed intensive care unit mortality and eight studies assessed hospital mortality. One study found both significantly lower intensive care unit mortality (after adjusting for repeated interim analyses) and significantly lower hospital mortality in the intensively treated group compared with those who received conventional glucose control; all other studies found no significant difference in either outcome. There were no significant differences between treatment and control in 28-day (one study), 30-day (one study) and 90-day (two studies) mortality.
Six studies assessed septicaemia and one reported a significantly lower rate of septicaemia in the intensively treated group than in the control group; five studies reported no significant difference. Six studies reported data on organ dysfunction and three of these reported significantly lower rates in the intensively treated group than in the control group.
Ten studies reported data on hypoglycaemia and in five of these the rate of hypoglycaemia was significantly higher in the intensively treated group than in the control group.
All 11 studies reported lower mean blood glucose levels in the intensively treated group than in the control group, but in one study the difference did not reach statistical significance.