Standardization of carotid ultrasound: a hemodynamic method to normalize for interindividual and interequipment variability.

Abstract

Accurate carotid Doppler examination is an important issue in the light of large endarterectomy trials, but recommended cutoff values for detection of >70% stenosis vary widely. Standardization of diagnostic criteria should consider patient variation and instrument variability. We prospectively analyzed various Doppler parameters in 44 patients undergoing carotid angiography to evaluate whether normalization through individual reference measurements from the common carotid artery or the distal internal carotid artery could improve accuracy. For assessment of interindividual and interequipment variability, we performed repeated measurements of 40 carotid arteries in 21 patients. Two color-coded duplex ultrasound systems were compared for machine variability estimation: Hewlett Packard SONOS 2500 and ATL Ultramark 9 HDI. Intrastenotic divided by distally recorded mean blood flow velocity (mean velocity ratio) showed the closest correlation with angiography: R2=0.93. Mean velocity ratio >5 was 97% sensitive and 98% specific for detection of >70% carotid stenosis. Intrastenotic blood flow velocities were significantly different between the 2 duplex systems (0.22+/-0.16 versus 0.17+/-0. 11 m/s; P<0.001), whereas mean velocity ratio values did not differ significantly. Interobserver variation expressed as 95% CI for predicted stenosis between 2 observers was 13.6% (peak systolic velocity) and 15.4% (mean velocity ratio). A mean velocity ratio using distal reference measurement in the internal carotid artery can normalize for interindividual and interequipment variability.