Reproducibility of first- and second-trimester uterine artery pulsatility index measured by transvaginal and transabdominal ultrasound

Objectives To examine and compare the reproducibility of measurement of first-and second-trimester uterine artery pulsatility index (UtA-PI) using transabdominal (TAS) and transvaginal (TVS) ultrasound. Methods This was an observational study including women >= 18 years of age, with a singleton pregnancy, in the first trimester (between 11+0 and 13+6 weeks' gestation) or second trimester (between 20 and 26 weeks' gestation). UtA-PI and angle of insonation were assessed by two observers (one with 15 and the other with 3 years of experience) using both TAS and TVS. The more experienced observer performed two scans alternated by the other observer. The acquisitions were completely independent and the observers were blinded to each other and to their own measurements. Reproducibility of the measurements by TVS and TAS was assessed using the concordance correlation coefficients (CCCs), intraclass correlation coefficients (ICCs) and limits of agreement (LoA). Results We analyzed data from 97 women in the first trimester and 96 in the second trimester. The mean +/- SD UtA-PI was significantly higher when measured using TVS, compared with TAS, in both the first (1.60 +/- 0.49 vs 1.52 +/- 0.63, respectively; P= 0.03) and second (1.07 +/- 0.33 vs 0.96 +/- 0.32, respectively; P<0.001) trimesters. The median angle of insonation was significantly lower when using TVS, compared with TAS, in both the first (8.0 degrees (interquartile range (IQR), 2.5-16.3 degrees) vs 12.5 degrees (IQR, 2.5-20.0 degrees), respectively; P= 0.04) and second (10.0 degrees (IQR, 4.5-16.5 degrees) vs 17.5 degrees (IQR, 5.0-27.9 degrees), respectively; P< 0.001) trimesters. Both ultrasound techniques had similar reproducibility: the intraobserver CCC ranged from 0.93 to 0.95 and the interobserver CCC ranged from 0.81 to 0.86; and the ICCs of both techniques were highly comparable (the intraobserver LoA was approximately +/- 20-30% and the interobserver LoA was approximately +/- 30-40%). Conclusions When measuring UtA-PI, assessment by TVS provides higher values and better insonation angle compared with TAS. The reproducibility of the ultrasound methods in both first and second trimesters of pregnancy was comparable and should not be considered as good. Future studies examining technical improvements with the aim of increasing the reproducibility of this technique should be encouraged. Copyright (C) 2014 ISUOG. Published by John Wiley & Sons Ltd.