Role of conventional angiography in evaluation of patients with carotid artery stenosis demonstrated by Doppler ultrasound in general practice.

BACKGROUND AND PURPOSE Previous studies have suggested that patients with carotid stenosis who are candidates for endarterectomy can be effectively identified on the basis of carotid Doppler ultrasound alone. Before widespread acceptance of this policy, the accuracy of carotid Doppler ultrasound outside selected centers and clinical trials needs to be evaluated. We performed a 12-month prospective study to evaluate the accuracy of Doppler ultrasound in identifying patients for carotid intervention in general practice settings. METHODS Each patient referred to our endovascular service for diagnostic angiography to evaluate for carotid stenosis was interviewed and examined by a neurologist. Subjects consisted of symptomatic patients with >/=50% stenosis and asymptomatic patients with >/=60% stenosis by Doppler ultrasound. Information pertaining to demographic and cerebrovascular risk factors and the results of the carotid Doppler ultrasound were recorded. The severity of stenosis on angiograms was measured with North American Symptomatic Carotid Endarterectomy Trial criteria by a blinded observer. The results of both studies were compared to determine the relative accuracy of ultrasound results. RESULTS Of 130 patients (mean age, 69+/-8.8 years) who met Doppler ultrasound criteria, 22 (17%) and 8 patients (6%) were found to have 30% to 49% or <30% stenosis by angiography, respectively. The positive predictive value of carotid Doppler ultrasound for identifying appropriate symptomatic candidates for carotid intervention (angiographic stenosis >/=50%) was 80%, with a false-positive value of 20%. The positive predictive value of carotid Doppler ultrasound for identifying appropriate asymptomatic candidates for carotid intervention (angiographic stenosis >/=60%) was 59%, with a false-positive value of 41%. Carotid endarterectomy or angioplasty and stent placement were undertaken subsequently in 60 (46%) of the patients. In 94 patients who underwent cerebral angiography alone, no complications were observed. CONCLUSIONS The present accuracy of carotid Doppler ultrasound in general practice does not justify its use as the sole basis of selecting appropriate patients for carotid intervention. Given the relatively low rate of associated morbidity with present day techniques, additional confirmatory studies such as angiography should be performed in every patient before a decision regarding intervention is made.