MULTIVARIATE-ANALYSIS OF PROGNOSTIC VARIABLES PREDICTING LONG-TERM MORTALITY AFTER ISCHEMIC STROKE

It was the aim of the present study to determine prognostic factors concerning the long-term mortality of patients with ischaemic stroke. 829 patients were hospitalized for ischaemic stroke between 1976 and 1985; using questionnaires, all of them were asked about the course and outcome. The return rate of the questionnaires was 63.6%. Consequently, the results of 447 patients formed the basis of the study. 137 patients had died during the observation period of 62.5 .+-. 29.9 months (range: 14-133 months). Several variables were analysed in respect of their prognostic relevance for long-term mortality: sex, age, vascular area involved, number of vascular areas clinically involved, number of stroken, and severity of clinical symptoms. The clinical symptoms (motor deficits, sensory deficits, speech disorders, visual field defects and organic mental syndrome) were graded semiquantitatively. Additionally, the single results were added up to a "total score". Using univariate analysis, age, number of vascular areas involved, motor deficits, and organic mental syndrome had a significant influence on long-term mortality. In patients with the lesion localized in the carotid territory, with more than one ischaemic event, or with high total scores, a tendency towards an unfavourable outcome was observed. However, multivariate analysis showed that age, number of vascular areas involved and an organic mental syndrome can serve as significant and independent predictors of long-term mortality. Applying the logistic regression formula, it is possible to predict the outcome (long-term mortality) in 78% of patients.