Prediction of in-hospital mortality after first-ever stroke: the Lausanne Stroke Registry.

We aimed to study in-hospital mortality after a first-ever stroke (brain infarction or parenchymatous hemorrhage) and to determine its predictors using easily obtainable variables. The main outcome measure was vital status at hospital discharge. Clinical features and type of stroke, with a particular emphasis on age, stroke topography and presumed causes of stroke, were studied in 3362 consecutive patients from the Lausanne Stroke Registry. Overall mortality was 4.8%. Brain hemorrhage mortality was 14.4% (48/333) and brain infarction mortality was 3.70% (112/3029). Localizations with high mortality included infratentorial (17.5%) and deep hemispheric (15.9%) territories for brain hemorrhage and, for brain infarction, multiple localizations in the posterior circulation (18.4%) and large middle cerebral artery territory (15.5%). Presumed causes of stroke associated with high mortality included saccular aneurysm (58.3%) and hypertensive arteriopathy (13.0%) for brain hemorrhage and, for brain infarction, dissection (10.4%), arteritis (8.3%), hematologic conditions (6.7%) and coexisting arterial and cardiac sources of embolism (5.2%). Multivariate logistic analysis showed that impaired consciousness on admission and limb weakness were good predictors of mortality for brain hemorrhage, while impaired consciousness and the cumulative effect of progressive worsening, limb weakness, left ventricular hypertrophy, past history of cardiac arrhythmia and previous transient ischemic attack were predictors of mortality for brain infarction. Age was not an independent predictor of stroke mortality, but for brain infarction the number of cumulative factors considered in the model increased with age. Our study shows that several factors associated with death risk are available during the first few hours after onset of stroke. Age alone is not critical, although its interaction with other factors should be considered.