Uterine artery pulsatility index in the first trimester: assessment of intersonographer and intersampling site measurement differences.

Objective: To determine intersonographer, intersampling site pulsatility index differences in the ascending branch of the uterine artery (UtA-PI) and their effect on detection rates (DR) for early onset preeclampsia (PE).

Methods: A prospective observational study was conducted including 52 women with singleton viable pregnancy at 11-13 weeks' gestation. Consecutive bilateral UtA-PI measurements were performed by two sonographers. Both sonographers hold the Fetal Medicine Foundation (FMF) uterine artery Doppler assessment competency certificates. Sonographer "A" underwent mentorship-based specialist training at the FMF; whilst sonographer "B" is a fetal maternal specialist who was deemed competent to measure UtA-PI based on completion of the FMF online course. Both sonographers were unaware of each other's UtA-PI and peak systolic velocity (PSV) measurements throughout the study. UtA-PI was measured by sonographer "A" at 1, 2 and 3 cm distally from the internal os. UtA-PI minimum ("Low-PI") and mean ("Mean-PI") were determined. Intraclass correlation (ICC), Bland-Altman analysis and Wilcoxon signed rank test were performed to determine bias, 95% limits of agreement (LOA) for intersonographer and intersampling site differences. Simulation studies were performed to determine the effect on early onset PE screening DR.

Results: (1) Intersite assessment indicated that UtA-PI and PSV decreased by 7-8% per centimeter relative to the measurement taken at the internal os; (2) Sonographer "B" UtA-PI measurements were significantly lower than those of sonographer "A" for "Low-PI" (p = .001), "Mean-PI" (p = .002) and PSV (p = .004) determined by Wilcoxon signed rank test. The mean reduction in "Low-PI", "Mean-PI" and PSV of sonographer "B" relative to sonographer "A" were 14.04%, 11.09% and 10.99%, respectively; (3) Measurements taken by sonographer "B" at the level of the internal os were comparable to measurements taken by sonographer "A" at 2 cm distal to the internal os (low-PI: p = .98, Mean-PI: p = .49 and PSV: p = .24); (4) Between sonographer ICC for UtA-PI was asymmetrical strong (left ICC = 0.72, 95%CI: 0.51-0.84) to fair (right ICC = 0.38, 95%CI: -0.08-0.64); and (5) The 14% mean intersonographer difference in lowest UtA-PI would have resulted in an 7% difference in PE screening performance.

Conclusions: The measurement of UtA-PI is sampling site dependent with the potential for significant intersonographer differences despite the availability of a prescriptive measurement protocol. This is an important observation as it implies that sonographer "B" inadvertently measured the UtA-PI at a distal site, not at the level of internal os, compared to those measured by sonographer "A", resulting in a lower DR for early onset PE.