Middle cerebral artery median peak systolic velocity validation: effect of measurement technique

We sought to validate center-specific published medians and estimate the effects of sonologist and Doppler measurement techniques on middle cerebral artery (MCA) peak systolic velocity (PSV) values. We studied 154 gravidas with normal singletons who underwent MCA PSV measurement at 18 to 35 weeks' gestation by one of three experienced sonologists. Pregnancies complicated by a known fetal anomaly (structural or aneuploidy), amniotic fluid volume disturbance, intrauterine growth restriction, multiple gestation, or isoimmunization were excluded. MCA PSV was measured using both manual caliper and auto-trace techniques. Regression models of log-transformed PSV values and gestational age were developed. Although auto-trace medians were significantly lower than those obtained with manual calipers ( P < 0.0001), they more closely approximated published medians used in clinical practice. Minimal intersonologist differences (maximum mean difference <3 cm/s) were statistically significant ( P < 0.01). Compared with manual caliper, auto-trace measurement yielded significantly lower medians. However, center-specific medians obtained by our sonologists using auto-trace more closely approximated published standards. Estimated interobserver variability suggested that different sonologists may utilize the same median values. We suggest that centers that utilize Doppler velocimetry for the prediction of fetal anemia examine their measurement protocol and consider formal confirmation of their own center-specific median values.