Intensive versus subcutaneous insulin in patients with hyperacute stroke: results from the randomized INSULINFARCT...

Intensive insulin therapy (IIT) has not yet proven its efficacy on stroke prognosis or in the reduction of MRI infarct growth. The INSULINFARCT study aims at determining in patients with hyperacute stroke whether IIT, with a better control of poststroke hyperglycemia, would reduce subsequent MRI infarct growth than usual care with subcutaneous insulin.

One hundred eighty patients with MRI-proven ischemic stroke and with National Institutes of Health Stroke Scale from 5 to 25 at admission (<6 hours) were randomized to receive IIT or usual subcutaneous insulin for 24 hours. Admission hyperglycemia was not required for recruitment. Control MRI and 3-month follow-up (with functional outcome and serious adverse events) were planned. The primary objective was to detect a difference in the proportion of patients with mean capillary glucose test <7 mmol/L during 24 hours. The secondary objective was to investigate whether IIT would reduce infarct growth. The analysis was planned in intention-to-treat. Patients with >3 missing capillary glucose test were excluded (n=4).

The proportion of patients with mean capillary glucose test <7 mmol/L in the first 24 hours was higher in the IIT group (95.4% [83 of 87] versus 67.4% [60 of 89]; P<0.0001). The infarct growth was lower in the subcutaneous insulin group (median, 10.8 cm(3); 95% CI, 6.5-22.4 versus 27.9 cm(3); 14.6-40.7; 60% of increase; P=0.04). The 3-month functional outcome (45.6% [41 of 90] versus 45.6% [41 of 90]), death (15.6% [14 of 90] versus 10% [9 of 90]), and serious adverse events (38.9% [35 of 90] versus 35.6% [32 of 90]) were similar in the subcutaneous insulin and IIT group.

The IIT regimen improved glucose control in the first 24 hours of stroke but was associated with larger infarct growths. IIT cannot be recommended in hyperacute ischemic stroke. Clinical Trial Registration- URL: http://clinicaltrials.gov. Unique Identifier: NCT00472381.