Stroke units versus general medical wards, I: twelve- and eighteen-month survival: a randomized, controlled trial.

BACKGROUND AND PURPOSE The long-term effect on survival of treatment in stroke units is still under debate. The hypothesis that a stroke unit with short length of stay increases 1-year and 18-month survival rates was tested in this study. METHODS A quasi-randomized, controlled study was undertaken among 802 patients > or =60 years old admitted to the Central Hospital of Akershus in Norway with a diagnosis of stroke between January 1, 1993, and February 1, 1995. All patients with onset of symptoms <24 hours before admittance were included and enrolled and were followed until death or to the end of the observation 18 months after stroke. Patients were allocated to a stroke unit (n=364) or a general medical ward (n=438). RESULTS Case fatality within the first 10 days was 8.2% among patients in the stroke unit and 15.1% among patients in the general medical ward (P=.0019). One-year survival among patients treated in the stroke unit was 70.6% and in the general medical wards 64.6% (P=.026); 18-month survival rates were 65.1% and 58.0%, respectively (P=.021). Among patients with cerebral hemorrhage, 10-day case fatality was 24.5% and 51.6% (P=.004) in favor of the stroke unit. CONCLUSIONS Stroke units increase survival rates among stroke patients compared with general medical wards. The effect on survival occurs early after the stroke and sustains during at least 18 months of observation.