A randomised, controlled pilot study to investigate the potential benefit of intervention with insulin in hyperglycaemic acute ischaemic stroke patients.

Background: Hyperglycaemia on presentation with acute ischaemic stroke (AIS) is associated with poor outcome, but intervention is unproven. We investigated the safety and tolerability of one method of glycaemic control.

Methods: Patients within 24 h of AIS and plasma glucose 8-20 mmol/l were randomised to receive either rigorous glycaemic control (RC) or standard management (SM) for 48 h. RC comprised i.v. insulin at a variable rate adjusted for target glucose concentration of 5-8 mmol/l, and intravenous crystalloid. The SM group received intravenous crystalloid alone in an open-label design.

Results: Thirteen patients were randomised to RC and 12 to SM (age 75 +/- 6.2 years; 40% male; 20% lacunar stroke; time to treatment 8 +/- 6.1 h; plasma glucose 10.6 +/- 0.9 mmol/l; known diabetes 52%; NIHSS 8, range 2-28). The glucose concentration-time curve was reduced in the RC group (AUC 324 +/- 15 versus 385 +/- 28 h.mmol/l, p = 0.04). By 48 h, plasma glucose in both groups was 6.8 +/- 1.1 and 7.5 +/- 1.3 mmol/l respectively, but mean hourly insulin requirements in the RC group had dropped from 3.25 +/- 0.32 units to 1.25 +/- 0.5 units (p < 0.01). One transient episode of hypoglycaemic symptoms occurred in the RC group.

Conclusion: Glycaemic control with sliding scale insulin for 48 h is feasible and well-tolerated after AIS. Treatment after 48 h may be unnecessary.