Background: Hyperglycemia has a detrimental effect in several acute neurological critical illnesses. No consensus exists on the optimal management of hyperglycemia in spontaneous intracerebral hemorrhage (sICH). Our aim was to determine whether blood glucose (BG) would predict 30-day mortality in sICH.
Methods: All patients with a well-defined diagnosis of sICH admitted into 24 h in three primary referred centers were included in this prospective observational follow-up study. Patients had extensive monitoring of BG values and those with BG values >8.29 mmol/l (150 mg/dl) received a variable intravenous insulin dose to maintain BG concentrations during the first 72 h after sICH between 3.32 and 8.29 mmol/l (60-150 mg/dl) using pre-specified insulin dosing schedule protocol.
Results: Between January 1, 2002, and December 31, 2003, 295 consecutive patients (mean +/- SD age 66 +/- 12 years) were prospectively included. A 1.0 mmol/l (18 mg/dl) increase in the BG concentration at admission was associated with a 33% mortality increase (OR: 1.33; 95%CI: 1.22-1.46; P < 0.0001). Adjusting for demographics, risk factors, stroke severity, and surgery there was no change in the increased risk. During the first 12 h after sICH, the insulin treatment protocol was enabling to reduce mortality (OR: 1.36, 95%CI: 1.14-1.61; P = 0.0005, per 1 IU increase) while thereafter this association was greatly attenuated and not more significant.
Conclusions: Hyperglycemia is a common condition after sICH and may worsen prognosis. Very early insulin therapy apparently does not improve prognosis. These results raise concern about routine clinical practice implementation of this intervention without any evidence from randomized trials.