Prospective evaluation of new duplex criteria to identify 70% internal carotid artery stenosis.

PURPOSE Large multicenter trials (North American Symptomatic Carotid Endarterectomy Trial, European Carotid Surgery Trial) have documented the benefits of carotid endarterectomy for treating symptomatic patients with >or=70% stenosis of the internal carotid artery. Although color-flow duplex scanning has become the preferred method for noninvasive assessment of internal carotid artery disease, no criteria have been generally accepted to identify this subset of patients. We previously reported a retrospective series to establish such criteria. This study details our results when these criteria were applied prospectively. METHODS Carotid color-flow duplex scans were compared with arteriograms in 457 patients who underwent both studies. Criteria for >or=70% internal carotid artery stenosis were peak systolic velocity >130 cm/sec and end-diastolic velocity >100 cm/sec. Internal carotid arteries with peak systolic velocity <40 cm/sec in which only a trickle of flow could be detected were classified as preocclusive lesions (95% to 99% stenosis). Arteriographic stenosis was determined by comparing the diameter of the internal carotid artery at the site of maximal stenosis to the diameter of the normal distal internal carotid artery. RESULTS Internal carotid artery stenosis of >or=70% was detected with a sensitivity of 87%, specificity of 97% positive predictive value of 89%, negative predictive value of 96%, and overall accuracy of 95%. Eighty-seven percent of 70% to 99% stenoses were correctly identified. False-positive errors (n=10) were attributed to contralateral internal carotid artery occlusion or high-grade (>90%) stenosis (n=5) and to interpreter error (n=1); no explanation was apparent in the other four. Eleven of 12 false-negative examinations occurred in patients with 70% to 80% internal carotid artery stenosis. CONCLUSIONS In our laboratories, prospective application of the above velocity criteria identified internal carotid artery stenosis of >or=70% with a reasonably high degree of accuracy. Errors occurred when stenoses were borderline and in patients with severe contralateral disease. With suitably modified velocity criteria, color-flow duplex scanning remains the most reliable noninvasive method for identifying symptomatic patients who are candidates for carotid endarterectomy.