Randomized, controlled trial of insulin for acute poststroke hyperglycemia

Abstract

Poststroke hyperglycemia is common and is associated with increased risk of death and dependence, but appropriate management remains uncertain. Glucose potassium insulin (GKI) infusion did not benefit patients with moderate poststroke hyperglycemia in a recent trial. Using magnetic resonance imaging (MRI), previous studies identified a relationship between recruitment of ischemic tissue to the final infarct and hyperglycemia, possibly mediated by brain lactic acidosis. We undertook a randomized, placebo-controlled trial of GKI infusion in patients with blood glucose >126mg/dl (7mmol/l) within 24 hours of ischemic stroke. The primary endpoint was infarct growth on MRI between baseline and day 7. Brain lactate concentrations were measured with magnetic resonance spectroscopy. Forty patients were randomized, 15 to saline and 25 to GKI infusions of different durations. Capillary blood glucose concentrations were lowered significantly from 6 to 12 hours after GKI initiation. There was no significant difference on any measure of infarct growth between the GKI and saline groups. In a secondary analysis, GKI was associated with significantly greater infarct growth in patients with complete intracranial vessel occlusion compared with controls (p = 0.011 for group-vessel status interaction). Brain lactate levels increased in control subjects, but were significantly lower with GKI infusion. Predominantly asymptomatic hypoglycemia occurred in 76% of GKI-treated subjects. GKI infusion within 24 hours of stroke lowered blood glucose and attenuated an increase in brain lactate, but did not affect cerebral infarct growth. Exploratory analysis found that GKI infusion was associated with greater infarct growth in patients with persistent arterial occlusion, and with a high incidence of asymptomatic hypoglycemia.