Measurement of blood flow in the vertebral artery using colour duplex Doppler ultrasound: establishment of the reliability of selected parameters.

This study was designed to determine the reliability of the ultrasound testing procedure for evaluating vertebral artery blood flow, and to determine a robust testing protocol for future studies. Blood flow parameters were tested in ten asymptomatic subjects (mean age 33 years, standard deviation 6 years 8 months) using colour duplex Doppler imaging. Volume flow rate data at C5-6 demonstrated good reliability from a single measurement (Intraclass correlation coefficient [ICC]=0.81). Peak velocity sampled at C1-2 showed poor reliability if a single measurement was used (ICC=0.26) improving to fair levels with three measurements (ICC=0.77). Reliability for this parameter was good if five measurements were taken (ICC=0.83-0.84). Systolic/diastolic ratio measured at C5-6 showed poor reliability (ICC=0.57) if a single measurement was taken in the manner of Thiel et al. (1994). This improved to fair reliability (ICC=0.75) if the mean of three measurements was used. There was no further improvement if five measures were sampled. Sampling at C2-3 in the manner of Refshauge (1994) was found to be technically difficult and it was not possible to detect a Doppler shift in three of the ten subjects at this level. Reliability of peak velocity at C2-3 was found to be poor, regardless of whether single or multiple averaged measurements were taken (ICC=0.37-0.63). Mean (time averaged) velocity measurements at C2-3 showed poor reliability if a single measurement was taken (ICC=0.39), fair reliability if the first three measurements were averaged (ICC=0.73) and good to high reliability levels if five measurements were sampled (ICC=0.88-0.91). A review of the literature suggests that sampling volume flow rate at C5-6 and peak velocity at C1-2 represents a clinically meaningful combination of parameters to detect narrowing in the VA. The results of this current study indicate the desirability of taking a single measurement of volume flow rate at C5-6 and the mean of three measurements of peak velocity at C1-2, with the additional calculation of the standard error of measurement, if reliable results are to be achieved. Copyright 2000 Harcourt Publishers Ltd.