Stroke in patients with cancer: incidence and etiology.

OBJECTIVE To assess the incidence and type of strokes in patients with cancer at Memorial Sloan-Kettering Cancer Center. METHODS Retrospective review of all ischemic strokes diagnosed by a neurologist and confirmed by neuroimaging between February 1997 and April 2001 was conducted. Age, gender, cancer diagnosis and stage, and vascular risk factors were recorded. NIH Stroke Scale and modified Rankin Scale scores were calculated retrospectively. Stroke etiology was assigned independently by two neurologists using the Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria. RESULTS Ninety-six patients with a confirmed stroke were identified. The median age was 67, and 61.5% were men. The distribution of vascular risk factors was comparable with that seen in large stroke trials. Lung cancer (30%) was the most common primary tumor followed by brain and prostate cancer (9% each). Strokes were embolic in 52 (54%) and nonembolic in 44 (46%). Eleven of 12 tested patients had an elevated D-dimer level, but in only 3 patients could a definitive diagnosis of nonbacterial thrombotic endocarditis be made. The median survival was 4.5 months (95% CI 2.8 to 9.5) from the diagnosis of stroke; 25% of patients died within 30 days. Treatment had no effect on survival. CONCLUSIONS Embolic strokes are the commonest cause of stroke in patients with cancer, due partially to hypercoagulability, whereas atherosclerosis accounted for only 22% of stroke in this population. Outcome was primarily determined by the underlying malignancy and the patient's neurologic condition.