Performance of the ultrasound examination in the early and late third trimester for the prediction of birth weight deviations.

OBJECTIVES: This study aimed to define the optimal gestational age in the third trimester, early (30-33 weeks + 6 days) versus late (34-37 weeks), for performing an ultrasound examination for fetal biometry to predict birth weight deviations: small for gestational age (SGA ≤ 5th centile) and large for gestational age (LGA ≥ 95th centile) neonates.
METHODS: We used an observational cross-sectional study in uncomplicated singleton pregnancies that had a third trimester ultrasound for fetal biometry and umbilical and middle cerebral fetal Doppler studies. Estimated fetal weight and fetal Doppler parameters were the examined variables for the prediction of SGA and LGA.
RESULTS: Three thousand six hundred ninety women had an early examination, and 2288 women had a late one. For a screen-positive rate of 10%, estimated fetal weight achieved 58% and 53.4% sensitivity for the prediction of SGA [area under the curve (AUC) = 0.8578, p < 0.001] and LGA (AUC = 0.8547, p < 0.001), respectively, by the early examination. Accordingly, the sensitivities significantly increased to 75.2% and 63.2% for the prediction of SGA (AUC = 0.9074, p < 0.001) and LGA (AUC = 0.8782, p < 0.001), respectively, by the late examination. The inclusion of the Doppler indices did not improve the predictive models.
CONCLUSIONS: A late third trimester ultrasound was superior in the prediction of SGA and LGA, and this improvement was more pronounced for the prediction of SGA.