Intraobserver reproducibility of Doppler measurements of uterine artery blood flow velocity in premenopausal women

OBJECTIVE: To determine the intraobserver repeatability of Doppler measurements of uterine artery blood flow velocity and the contribution of various factors to within-subject variance. DESIGN: Seventeen healthy premenopausal women underwent vaginal Doppler ultrasound examination of the uterine artery by the same observer. Three measurements were taken at each of three sites: 1) the currently recommended sampling site; 2) the ascending branch of the uterine artery at a level between the lower and middle third of the corpus uteri; 3) 1.5 cm lateral to the recommended sampling site. Three measurements were taken at each site. For each measurement, three uniform consecutive cardiac cycles were analyzed. Peak systolic velocity, time-averaged maximum velocity, and pulsatility index were calculated. Each Doppler shift spectrum was analyzed twice. Thus, for each women, 18 measurement results per sampling site were obtained. Analysis of variance was used. RESULTS: The effect of sampling site on measurements of peak systolic velocity and time-averaged maximum velocity was non-significant, but pulsatility index values obtained at the distal sampling site were slightly higher than those obtained at the other sites (P = 0.01). Repetition accounted for most of the within-subject variance. Averaging the results of the three repeat measurements yielded increased intraclass correlation coefficients: 0.79-0.89 for peak systolic velocity, 0.80-0.92 for time-averaged maximum velocity and 0.86-0.93 for pulsatility index. CONCLUSION: As the effect of repetition on the results of Doppler measurements of uterine artery blood flow velocity is large, the average of several repeat measurements should be used to enhance measurement reproducibility. However, it is not worth doing more than one analysis of a Doppler shift spectrum, and it is not worth analyzing more than one cardiac cycle per spectrum