The purpose of this study was to review data from published cerebrovascular studies to determine if it is possible to predict, based on clinical manifestations (eg, cervical bruit, transient ischemic attack, or stroke) of cerebrovascular atherosclerosis, the annual probability of having a stroke, as well as to determine and discuss the methods used in calculating and reporting vascular event rates. This overview analysis reveals that the annual stroke rates are as follows: for asymptomatic carotid stenosis, 1.3% (95% confidence interval [CI], 1.0 to 1.6); for transient monocular blindness, 2.2% (95% CI, 1.3 to 3.0); for transient ischemic attack, 3.7% (95% CI, 3.1 to 4.3); for minor stroke, 6.1% (95% CI, 5.7 to 6.6); and for major stroke, 9.0% (95% CI, 8.0 to 9.9). The data analyzed here suggest that a hierarchical profile of worsening clinical characteristics mirrors a hierarchical progression of increasing risk of stroke. These data support the idea that there is a clinical risk profile, in addition to the conventional atherosclerosis risk factor profile, for predicting subsequent stroke.