# Maternal BMI and Diagnostic Accuracy of Using Estimated Fetal Growth to Predict Abnormal Birthweight: Results from NICHD Fetal Growth Studies

## TL;DR

This study found that estimated fetal weight is more accurate for predicting small birthweight in overweight/obese mothers compared to normal-weight mothers.

## Contribution

The novel finding is that maternal BMI significantly affects the diagnostic accuracy of fetal weight estimation for small-for-gestational-age births but not for large-for-gestational-age births.

## Key findings

- EFW accuracy for LGA prediction was similar across BMI groups (AUC 0.77 vs 0.79).
- EFW accuracy for SGA prediction was higher in overweight/obese mothers (AUC 0.91 vs 0.84).
- EFW errors were comparable across BMI groups except for a trend toward lower errors in SGA births among overweight/obese mothers.

## Abstract

Background/Objectives: The objective of this study was to assess the diagnostic accuracy of sonographic estimated fetal weight (EFW) in predicting small (SGA)- or large-for-gestational-age (LGA) birthweight and examine whether the accuracy is associated with maternal body mass index (BMI). Methods: The participants of NICHD Fetal Growth Studies with complete data on maternal BMI (10–13.9 weeks), EFW within 14 days of delivery (18–41.3 weeks), and birthweight were included in this study. Participants were categorized as normal (BMI 18.5–24.9 kg/m2) or overweight/obese (BMI > 24.9 to 44.9 kg/m2). EFW accuracy was evaluated using area under the Receiver Operating Characteristic curves (AUCs) for SGA and LGA classification, and EFW error was analyzed across BMI groups. Results: Among 1289 women, 714 (55.4%) were in the normal BMI group. AUCs for LGA prediction were similar between BMI groups (0.77 ± 0.03 for normal vs. 0.79 ± 0.02 for overweight/obese, p = 0.593). However, for SGA, AUCs were higher in the overweight/obese group (.91 ± 0.01 vs. 0.84 ± 0.02, p = 0.004), indicating improved accuracy. EFW absolute and percent errors were comparable across BMI groups in the full, AGA, and LGA birth cohorts separately, but they trended lower (p = 0.12 and 0.15 for absolute and percent errors, respectively) in the overweight/obese group in the SGA birth cohort. Conclusions: EFW has acceptable accuracy for predicting LGA, unaffected by BMI. However, for SGA, EFW accuracy is significantly higher in the overweight/obese group, suggesting that BMI influences diagnostic performance in SGA but not LGA classification.