How repeat measurements affect the mean diameter of the umbilical vein and the ductus venosus.

Abstract

Blood flow measurement in the fetus is susceptible to substantial error. Increased numbers of measurements can improve the precision. To quantify this, we assessed how the number of ultrasound measurements affected the calculation of individual mean diameters of fetal veins. A study of the umbilical vein (UV) and ductus venosus (DV) was carried out in 167 pregnancies in a clinical setting. The UV diameter was measured 2-13 times in 163 participants and the DV for the same number of times in 158. A one-way analysis of variance gave mean within-subject SD of 0.23 mm for the UV (diameter 2.0-8.0 mm) and 0.17 mm for the DV (diameter 0.3-2.3 mm). The 95% limit for the within-subject mean diameter calculated from two, four, six and ten measurements of the UV were 0.31, 0.22, 0.18 and 0.14 mm, respectively. For the DV the corresponding results were 0.23, 0.16, 0.13 and 0.10 mm. The study shows how increasing the number of measurements to calculate a mean diameter substantially improves the precision of vascular cross-section assessment. The limitation of the method is expressed in the exponential increase in numbers needed to achieve further improvement.