Predictors of stroke mortality in elderly people from the general population. The CArdiovascular STudy in the ELderly.

Stroke occurs particularly frequently in elderly people and, being more often disabling than fatal, entails a high social burden. The predictors of stroke mortality have been identified in 3282 subjects aged > or = 65 years, taking part in the CArdiovascular STudy in the ELderly (CASTEL), a population-based study performed in Northeast Italy. Historical and clinical data, blood tests and 14-year fatal events were recorded. Continuous items were divided into quintiles and, for each quintile, adjusted relative risk (RR) with 95% confidence intervals [CI] was derived from multivariate Cox analysis. Age, historical stroke (RR: 5.2; 95% CI: 3.18-8.6) and coronary artery disease (RR: 1.38; CI: 1.18-2.1), atrial fibrillation (RR: 2.40; CI: 1.42-4.0), arterial hypertension (RR: 1.33; CI: 1.15-1.76), systolic blood pressure > or = 163 mmHg (RR: 1.84; CI: 1.20-2.59), pulse pressure > or = 74 mmHg (RR: 1.50; CI: 1.13-2.40), cigarette smoking (RR: 1.60; CI: 1.03-2.47), electrocardiographic left ventricular hypertrophy (RR: 1.72; CI: 1.10-2.61), impaired glucose tolerance (IGT, RR: 1.83; CI: 1.10-3.0), uric acid (UA) > 0.38 mmol/l (RR: 1.61; CI: 1.14-2.10), serum potassium > or = 5 mEq/l (RR: 1.70; CI: 1.24-2.50) and serum sodium < or = 139 mEql/l (RR: 1.34; 1.10-2.10) increased the risk of stroke. In the CASTEL, stroke was the first cardiovascular cause of death. Some independent predictors usually unrelated to stroke mortality (namely pulse pressure, pre-diabetic IGT, UA and blood electrolytes disorders) have been identified.