The reliability of color duplex ultrasound in diagnosing total carotid artery occlusion

Abstract

Background and purpose: Color duplex ultrasound has been advocated as an alternative to arteriography before carotid endarterectomy. However, one limitation of color duplex ultrasound is that it sometimes fails to differentiate high-grade stenosis from total carotid occlusion. This study was done to determine (1) the accuracy of carotid duplex ultrasound in diagnosing total carotid occlusion, and (2) when angiography is necessary.

Patient population and methods: Carotid duplex ultrasound and angiography results were compared for 520 carotid arteries, and 103 of these had a duplex diagnosis of total carotid occlusion or suspected almost total-to-total occlusion. The diagnosis of total carotid occlusion was primarily based on the absence of flow in the internal carotid artery as visualized on B-mode imaging for at least 1 inch beyond the bifurcation (optimal study). If the internal carotid artery was not optimally seen beyond the bifurcation, but secondary criteria were present, such as dampening of the common carotid signal and internalization of the external carotid artery, a diagnosis of suspected subtotal to total occlusion was made (limited study).

Results: In the optimal studies, 91 arteries had total carotid occlusions and of these, 87 were confirmed by angiography. The accuracy of carotid duplex ultrasound in diagnosing total carotid occlusion was 97% with a positive predictive value of 96%, negative predictive value of 98%, sensitivity of 91%, and specificity of 99%. Twelve arteries were diagnosed as suspected subtotal to total occlusion (limited studies), and of these, three were occluded on angiography, eight had stenoses ranging from 90% to 99%, and one had 80% stenosis.

Conclusions: A carotid duplex ultrasound study is an acceptable method for predicting total carotid occlusion when the study is optimal, and angiography is unnecessary in asymptomatic patients. Angiography is recommended for patients who are surgical candidates with a limited duplex study.