The effect of contralateral disease on carotid Doppler frequency.

Continuous-wave Doppler frequency analysis is one standard of noninvasive evaluation for stenotic carotid disease. Interpretation is most commonly based on frequency recorded at the site of stenosis. If flow velocity is increased because of contralateral disease, this criterion may overstate the degree of stenosis. To investigate this, three centers contributed 167 patients to a study of carotid Doppler frequency analysis. Each patient underwent bilateral carotid studies by a standard protocol with use of continuous-wave 5 MHz Doppler probe. There were 309 patent and 25 occluded internal carotids seen on subsequent angiographs. Each patient side was categorized by the severity of ipsilateral disease (less than 45%, 45% to 70%, 71% to 99%, occlusion) and then subgrouped by contralateral disease. For all carotid arteries, the optimal peak frequency to detect 45% or more stenosis was 5500 Hz, with an overall accuracy of 92.2%. Within the three classifications of ipsilateral stenosis, a trend was noted that related increasing peak frequency to greater degrees of contralateral disease. This was most apparent when the contralateral internal carotid artery was occluded. Of all false-positive errors, 43% were accounted for by this subset that represented 15% of the total population. For the 25 patients with occlusion opposite a stenosis, the 5500 Hz cutoff of Doppler peak frequency yielded only 76% accuracy; an ROC curve showed that a cutoff value of 8500 Hz improved overall accuracy to 92% for these patients. By applying two cutoff criteria--5500 Hz for those with patent contralateral internal carotids and 8500 Hz for those in whom a contralateral Doppler signal was not detected--the overall accuracy for all patients improved to 93.3%. It is concluded that contralateral occlusion affects Doppler peak frequency by increasing the velocity in the companion carotid artery, whereas severe contralateral stenosis did not result in a consistent pattern of increased velocity.