Etiology of motor or sensory stroke: a prospective study of the predictive value of clinical and radiological features.

The diagnostic utility of clinical and radiological features for distinguishing penetrating artery disease from other causes of stroke has been questioned. To address this issue, we prospectively evaluated more than 40 features in 85 patients with pure motor, sensorimotor, or pure sensory syndromes. Nonischemic causes were present in 4 patients. The causes of ischemic stroke in the other 81 patients were penetrating artery disease (32 patients), large artery occlusive disease (17), cardioembolism (12), other causes (8), and undetermined (12). Of the features evaluated, frequent transient ischemic attacks (greater than or equal to 1/day), transient ischemic attacks occurring only within 1 week of stroke, pure motor hemiparesis (similar involvement of face, arm, and leg), pure sensory stroke, and round or oval infarction were associated (p less than 0.05) with penetrating artery disease. However, only pure sensory stroke involving two or more regions of the body and pure motor hemiparesis associated with subcortical infarction of less than 1.5 cm had positive predictive values exceeding 90% for penetrating artery disease. Carotid bruit ipsilateral to an anterior circulation infarction, monoparesis, and striatocapsular infarction had positive predictive values exceeding 90% for causes other than penetrating artery disease; however, half of the striatocapsular infarctions were of undetermined cause. These data suggest that certain clinical and radiological features are useful for distinguishing penetrating artery disease from other causes of stroke.