Journal of Rehabilitation Medicine vol:44 issue:7 pages:547-552
Background: Few studies have investigated the long-term mortality of stroke patients admitted to rehabilitation units. This study aimed to determine the five-year mortality, and its association with baseline characteristics and functional status six months post-stroke in an inpatient stroke rehabilitation setting.
Methods: 532 stroke patients were recruited consecutively in four European rehabilitation centres. Predictors were recorded on admission. The Barthel Index was assessed at six months (BI6mths) and patients were followed-up for five years post-stroke. The risk of dying was computed using Kaplan-Meier curves and compared across three BI6mths classes (0-60, 65-90 and 95-100) (log-rank test). Significant independent predictors were determined using multivariate Cox regression analysis (hazard ratio=HR).
Results: The five-year cumulative risk of death was 29.12% (95%CI:22.86-35.38). Age (HR=1.06, 95%CI:1.04-1.09), cognitive impairment (HR=1.77, 95%CI:1.21-2.57), diabetes mellitus (HR=1.68, 95%CI:1.16-2.41), and atrial fibrillation (HR=1.52, 95%CI:1.08-2.14) were significant independent predictors of increased mortality. Hyperlipidemia (HR=0.66, 95%CI: 1.46-0.94), and higher BI6mths (HR=0.98, 95%CI:0.97-0.99) were significant independent predictors of decreased mortality. Estimated five-year survival probability was 0.85 (95%CI: 0.80-0.89) for patients in the highest BI6mths class (95-100), 0.72 (95%CI: 0.63-0.79) in the middle (65-90) and 0.50 (95%CI: 0.40-0.60) in the lowest class (0-60) (log-rank, p<0.0001).
Conclusions: Nearly a third of stroke rehabilitation patients died during the first five years following stroke. Functional status at six months was a powerful predictor of long-term mortality and should be promoted through medical interventions and rehabilitation.