There are two potential purposes for cardiac evaluation in patients with cerebrovascular disease: to identify possible cardioembolic pathophysiology for ischemic symptoms and to identify concomitant coronary artery disease. Both have important implications for patient prognosis and treatment, and testing therefore appears to be warranted. On the other hand, the cost conservation movement in medicine dictates that physicians limit unnecessary, costly, possibly risky testing when the diagnostic yield is low. For example, the overall yield of cardiac testing in "usual stroke patients" who have no suggestive history or findings on examination, chest X-ray, or electrocardiogram is less than 10% and may not be indicated routinely. Conversely, young patients with stroke of unknown cause are likely to benefit from aggressive cardiac testing. Many reported series and clinical trials have demonstrated that patients with cerebrovascular disease are more likely to die in follow-up from cardiovascular than from cerebrovascular causes. This risk is best defined and may be highest in patients with carotid disease, in whom the 5-year cardiac mortality rate may be as high as 40 to 50%. Studies have shown that such patients are also likely to have abnormal tests for cardiac ischemia, even when a history of cardiovascular events or symptoms or electrocardiographic abnormalities are lacking. These results, combined with further investigations into which cerebrovascular patients are at highest risk for cardiovascular disease and what testing best identifies underlying, treatable cardiovascular disease, are needed to direct the care and improve the cardiovascular prognosis of patients with cerebrovascular disease.