Prediction of Small-For-Gestational-Age Neonates at 33-39 Weeks’ Gestation in China: Logistic Regression Modeling of The Contributions of Second- and Third-Trimester Ultrasound Data and Maternal Factors

Objectives: A screening model for prediction of small-for-gestational-age (SGA) neonates (SGAp) was established by logistic regression using ultrasound data and maternal factors (MF). We aimed to evaluate the ability of SGAp as well as abdominal circumference (AC) and estimated fetal weight (EFW) measurements to predict SGA neonates at 33-39 weeks’ gestation. Methods: This retrospective study evaluated 5298 singleton pregnancies that had involved three ultrasound examinations at 21 +0 –27 +6 , 28 +0 –32 +6 , and 33 +0 –39 +6 weeks. All ultrasound data were transformed to MoM values (multiple of the median). Multivariate logistic regression was used to analyze the correlation between SGA status and various variables (ultrasound data and MF) during pregnancy to build the SGAp model. EFW was calculated according to the Hadlock formula at 33-39 weeks of gestation. The predictive performance of SGAp, AC MoM value at 33 +0 –39 +6 weeks (AC-M), EFW MoM value (EFW-M), and EFW-M plus MF was evaluated using ROC curves, the detection rates (DRs) at false-positive rates (FPRs) of 5% and 10%, and the FPRs at DRs of 85%, 90%, and 95%. Results: The AUCs of SGAp, AC-M, EFW-M, and EFW-M plus MF for SGA neonates screening were 0.933 (95%CI: 0.916-0.950), 0.906 (95%CI: 0.887-0.925), 0.920 (95%CI: 0.903-0.936), and 0.925 (95%CI: 0.909-0.941), respectively. The DRs of SGAp, AC-M, EFW-M, and EFW-M plus MF for SGA neonates were 80.4%, 69.6%, 73.8% and 74.3% at 10% FPR. The AUCs of SGAp, AC-M, EFW-M, and EFW-M plus MF 0.950 (95%CI: 0.932-0.967), 0.929 (95%CI: 0.909-0.949), 0.938 (95%CI: 0.921-0.956) and 0.941 (95%CI: 0.924-0.958), respectively for screening SGA neonates delivered within 2 weeks after the assessment. The DRs for these births increased to 85.8%, 75.8%, 80.0%, and 82.5%, respectively. Conclusion: The rational use of ultrasound data can significantly improve the prediction of SGA statuses.