Intra- and inter-observer reproducibility and generalizability of first trimester uterine artery pulsatility index by transabdominal and transvaginal ultrasound.

OBJECTIVES: The primary aim of the study was to assess intra-observer and inter-observer reproducibility and generalizability (general reliability) of first trimester Doppler measurements of uterine arteries (UtA) performed both transabdominally (TA) and transvaginally (TV). Secondary aims were to investigate whether maternal BMI and acquisition modality (transabdominal or transvaginal) affect feasibility and reliability of UtA Doppler evaluation.

METHODS: Singleton pregnancies between 11(+0) and 13(+6) weeks underwent TA and TV Doppler measurements of UtA prospectively, blindly and independently by Fetal Medicine Foundation-accredited operators. The more experienced operator performed all the scans to assess intra-observer reproducibility. Inter-observer reproducibility and generalizability was evaluated among different operators and different ultrasound equipment. Intra- and inter-rater reliability was assessed by Concordance Correlation Coefficient and Intraclass Correlation Coefficient; agreement was estimated with Limits of Agreement and reported graphically with Bland-Altman plots. Correlation between BMI and feasibility and accuracy of UtA measurements were studied with Spearman's correlation coefficient. Differences between TA and TV measurements were evaluated by Wilcoxon test.

RESULTS: A total of 101 women were enrolled. The intra-observer reproducibility for TA and TV measurements ranged from 0.87 to 0.94, respectively. The inter-observer reproducibility varied from 0.57 (for TA measurements by the less experienced operator) to 0.95 (for TV measurements by the more experienced operator), and it was independent of the ultrasound equipment used. Spearman's correlation coefficient failed to demonstrate a correlation between BMI and time to get the first set of TA-TV measurements, or between BMI and TA-TV reproducibility. In four patients, the operators could not obtain the TA measurements due to high BMI. In our population the UtA pulsatility index (PI) values did not differ significantly between the TA or TV approach.

CONCLUSIONS: UtA PI measurements obtained by experienced operators show a moderate intra- and inter-observer reproducibility. The TV approach might be used when TA measurements fail in obese women, especially since specific reference ranges for TV measurements may not be necessary. © 2016 John Wiley & Sons, Ltd.