Background and purpose: Factors that determine the benefit of stroke units (SU) are unknown. The aim of our study was to analyze whether semi-intensive monitoring during the acute phase of stroke reduces mortality and dependency at long term.
Methods: We studied patients with an ischemic stroke or intracerebral hemorrhage, consecutively admitted to our SU within 24 h of symptoms onset. Based on bed availability, patients were allocated to either a conventional care stroke unit (C-SU, n = 209) or a semi-intensive stroke unit (SI-SU, n = 321) with continuous monitoring of cardiac, respiratory, metabolic and neurological functions during the first 72 h. Both groups were treated following the same medical and nursing protocols. Criteria for exclusion were patients with stupor/coma, previously dependent (Barthel score <85) and with TIA. Using logistic regression models, we analyzed the influence of semi-intensive care on mortality and dependency at one year.
Results: Baseline characteristics were similar between patients admitted to the SI-SU and the C-SU, except for a higher frequency of more severe stroke and intracerebral hemorrhage in the SI-SU. Twenty-six percent of patients in the SI-SU and 4% in the C-SU were randomized in acute clinical trials (p < 0.01), and 61% and 39% were seen by a neurologist in less than 6 h from the onset of symptoms (p < 0.01). At 1 year, mortality and combined mortality and dependency were not significantly different between the two groups. However, due to the presence of a significant interaction between the type of unit and stroke severity, the OR of mortality for SI-SU allocation was 0.19 (95% CI, 0.07-0.54) in patients with severe stroke (CSS < or =4), whereas it was 0.64 (95% CI, 0.37-1.11) in those with mild-to-moderate stroke.
Conclusions: This study suggests that semi-intensive monitoring in a stroke unit reduces mortality at 1 year in patients with severe stroke, with no influence over dependency.