Optimal umbilical artery Doppler chart for predicting placenta-mediated fetal growth restriction

To compare the predictive accuracy of selected umbilical artery (UA) pulsatility index (PI) reference charts for outcomes associated with placenta-mediated fetal growth restriction (FGR).

This was a retrospective cohort study of individuals with a singleton pregnancy who underwent UA Doppler assessment ≥ 20 weeks' gestation between January 2012 and December 2022 at a single tertiary referral center, at which UA Doppler is measured routinely regardless of fetal size. Using 10 different UA-PI reference charts, we compared the predictive accuracy of an abnormal UA-PI (> 95^th percentile) for two primary outcomes that are considered specific and gestational-age-independent indicators of placenta-mediated FGR: (1) late-stage UA Doppler abnormalities (defined as absent or reversed end-diastolic flow); and (2) maternal vascular malperfusion (MVM) on placental pathology. We also investigated the ability of these 10 charts to predict the secondary outcome of composite adverse perinatal outcome, defined as the presence of at least one of the following: stillbirth, 5-min Apgar score < 7, UA pH < 7.1, need for neonatal resuscitation and/or admission to the neonatal intensive care unit. Generalized estimating equations were used to calculate the predictive accuracy of the UA-PI reference charts, accounting for repeated measurements within the same patient. To identify the best-performing reference chart, we ranked each chart based on its Youden index at the 95^th percentile cut-off for UA-PI for the two primary outcomes. Given the distinct phenotypes of early- and late-onset FGR, we also performed an analysis stratified by gestational age at ultrasound examination (< 32 vs ≥ 32 weeks).

A total of 15 841 patients, with 38 398 ultrasound examinations, were included in the analysis. The proportion of small-for-gestational-age (SGA) fetuses classified as FGR based on an abnormal UA-PI varied widely depending on which reference chart was applied, ranging from 2.2% to 25.7%. Similarly, the predictive accuracy of the 10 different reference charts for placenta-mediated FGR outcomes differed considerably. The predictive performance for late-stage UA Doppler abnormalities varied substantially across charts, with sensitivity ranging from 20.7% to 76.3% and specificity from 75.2% to 98.0%. Likewise, for the prediction of MVM on placental pathology, the sensitivity of the charts ranged from 6.8% to 42.0% and specificity from 77.5% to 98.6%. For most of the charts, sensitivity and specificity remained comparable between the overall cohort and gestational-age subgroups. When ranked according to overall predictive performance for the two primary outcomes using the Youden index, the UA-PI reference charts of Rahimi et al., Drukker et al. and Flatley et al. demonstrated the best overall predictive accuracy. These same three charts retained the top performance ranking for the prediction of primary study outcomes in the subgroup of cases examined < 32 weeks.

We observed substantial variation among 10 UA-PI reference charts in both the proportion of SGA fetuses classified as growth restricted and the predictive accuracy of each chart for outcomes considered specific to placenta-mediated FGR. Among the charts evaluated, those of Rahimi et al., Drukker et al. and Flately et al. demonstrated the best overall performance for predicting all three study outcomes. If confirmed in external cohorts, these findings would support the ongoing efforts to standardize the diagnosis of FGR, which is crucial for both clinical and research purposes. © 2026 The Author(s). Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.