BACKGROUND Perioperative stroke is a devastating complication after cardiac surgery. In an attempt to minimize this complication, many cardiac surgeons routinely preoperatively order carotid artery duplex scans to assess for significant carotid stenosis. We hypothesize that the routine screening of preoperative cardiac surgery patients with carotid artery duplex scans detects few patients who would benefit from carotid intervention or that a significant carotid stenosis reliably predicts stroke risk after cardiac surgery. METHODS A retrospective review identified 1,499 patients who underwent cardiac surgical procedures between July 1999 and September 2010. Data collected included patient demographics, comorbidities, history of previous stroke, preoperative carotid artery duplex scan results, location of postoperative stroke, and details of carotid endarterectomy (CEA) procedures before, in conjunction with, or after cardiac surgery. Statistical methods included univariate analysis and Fisher's exact test. RESULTS Twenty-six perioperative strokes were identified (1.7%). In the 21 postoperative stroke patients for whom there is complete carotid artery duplex scan data, 3 patients had a hemodynamically significant lesion (>70%) and 1 patient underwent unilateral carotid CEA for bilateral disease. Postoperative strokes occurred in the anterior cerebral circulation (69.2%), posterior cerebral circulation (15.4%), or both (15.4%). Patient comorbidities, preoperative carotid artery duplex scan screening velocities, or types of cardiac surgical procedure were not predictive for stroke. Thirteen patients (0.86%) underwent CEA before, in conjunction with, or after cardiac surgery. Two of these patients had symptomatic disease, 1 of whom underwent CEA before and the other after his cardiac surgery. Of the 11 asymptomatic patients, 2 underwent CEA before, 3 concurrently, and 6 after cardiac surgery. Left main disease (≥50% stenosis), previous stroke, and peripheral vascular disease were found to be statistically significant predictors of carotid revascularization. A cost analysis of universal screening resulted in an estimated net cost of $378,918 during the study period. CONCLUSIONS The majority of postoperative strokes after cardiac surgery are not related to extracranial carotid artery disease and they are not predicted by preoperative carotid artery duplex scan screening. Consequently, universal carotid artery duplex scan screening cannot be recommended and a selective approach should be adopted.