Early differentiation of cardioembolic from atherothrombotic cerebral infarction: a multivariate analysis.

The aim of this study was to determine factors predictive of cerebral infarction subtype from clinical data collected within 48 h of neurologic deficit. All cardioembolic (n = 231) and atherothrombotic infarctions (n = 369) included in prospective stroke registry of the Sagrat Cor-Alianza Hospital of Barcelona were analysed. Demographic characteristics, anamnestic findings, cerebrovascular risk factors and clinical data of patients with embolic stroke and patients with thrombotic infarction were compared. Predictors of stroke subtype were assessed by means of a logistic regression model based on 16 clinical variables. After multivariate analysis, atrial dysrhythmia and sudden onset to maximal deficit were significant predictors of embolic stroke, whereas hypertension, chronic obstructive pulmonary disease, diabetes, hypercholesterolemia and/or hypertriglyceridemia and age were independent predictive factors of atherothrombotic stroke. Setting a cut-off point of 0.50 for predicting mechanism of stroke on admission resulted in a sensitivity of 76%, specificity of 87% and total correct classification of 83%. Clinical features alone that are observed at stroke onset can help to distinguish cardioembolic from atherothrombotic infarctions. Copyright 1999 Lippincott Williams & Wilkins