Outcome-based comparison of SMFM and ISUOG definitions of fetal growth restriction

The recent international guidelines by the Society for Maternal-Fetal Medicine (SMFM) and the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG) differ in their definitions of fetal growth restriction (FGR). Our aim was to compare the performance of the two definitions in predicting neonatal small-for-gestational age (SGA) and composite adverse neonatal outcome (ANO).

This was a secondary analysis of data from a prospective study of women referred for fetal growth ultrasound examination between 26 + 0 and 36 + 6 weeks' gestation. The SMFM and ISUOG guidelines were used to define pregnancies with FGR. The SMFM definition of FGR is estimated fetal weight (EFW) or abdominal circumference (AC) < 10^th percentile. The ISUOG-FGR definition follows the Delphi consensus criteria and includes either EFW or AC < 3^rd percentile or EFW or AC < 10^th percentile combined with abnormal Doppler findings or a decrease in growth centiles. The primary outcome was the prediction of neonatal SGA, defined as birth weight < 10^th percentile, and a composite of ANO, which was defined as one or more of: Grade-III or -IV intraventricular hemorrhage, respiratory distress syndrome, neonatal death, cord blood pH < 7.1, seizures and admission to the neonatal intensive care unit. Test characteristics (sensitivity, specificity, positive predictive value (PPV), negative predictive value and positive (LR+) and negative likelihood ratios) and area under the receiver-operating-characteristics curve were determined. The association between FGR detected by each definition and selected adverse outcomes was assessed using logistic regression analysis.

Of the 1054 pregnancies that met the inclusion criteria, 137 (13.0%) and 55 (5.2%) were defined as having FGR by the SMFM and ISUOG definitions, respectively. Composite ANO and SGA neonate each occurred in 139 (13.2%) pregnancies. For the prediction of neonatal SGA, the SMFM-FGR definition had a higher sensitivity (54.7%) than did the ISUOG definition (28.8%). The ISUOG-FGR definition had higher specificity (98.4% vs 93.3%), LR+ (18.0 vs 8.2) and PPV (72.7% vs 55.5%) than did the SMFM definition for the prediction of a SGA neonate. The SMFM- and ISUOG-FGR definitions had similarly poor performance in predicting composite ANO, with sensitivities of 15.1% and 10.1%, respectively.

The SMFM definition of FGR is associated with a higher detection rate for SGA neonates but at the cost of some reduction in specificity. The ISUOG-FGR definition has a higher specificity, LR+ and PPV for the prediction of neonatal SGA. Both definitions of FGR performed poorly in predicting a composite ANO. © 2021 International Society of Ultrasound in Obstetrics and Gynecology.