Development and internal validation of early- and mid-pregnancy prediction models for macrosomia in nulliparous singleton pregnancies
- 1Eindhoven University of Technology
Abstract
Purpose: Macrosomia is associated with adverse maternal and neonatal outcomes, but early risk stratification remains challenging. Most existing prediction approaches rely on later-pregnancy ultrasound measurements, limiting opportunities for earlier preventive counseling and monitoring. Methods: In this secondary analysis of the prospective nuMoM2b cohort in the United States, we included 6,371 nulliparous singleton pregnancies selected for available periconceptional diet and ultrasound data at Visits 2 and 3. We developed visit-updated multivariable logistic regression models for birthweight >4000 g and >4500 g using routinely available maternal characteristics, anthropometry, periconceptional diet, and serial ultrasound measurements, with sensitivity analyses using large-for-gestational-age birthweight. Model performance was assessed by repeated stratified cross-validation and compared with World Health Organization fetal growth chart-based ultrasound percentile benchmarks, emphasizing the estimated fetal weight percentile as the most clinically familiar proxy. Results: Birthweight >4000 g and >4500 g occurred in 7.9% and 1.2% of pregnancies, respectively. Models updated with later pregnancy information achieved the best discrimination, with an area under the receiver operating characteristic curve of 0.73 for birthweight >4000 g and 0.82 for birthweight >4500 g. In this selected analytic cohort, the early-pregnancy model based on maternal data alone provided moderate discrimination (0.63 and 0.72, respectively), although lower than the strongest third-trimester WHO fetal growth chart-based ultrasound percentile benchmarks. Conclusion: Within this selected cohort, routinely available early-pregnancy maternal characteristics and anthropometry provided moderate risk stratification for macrosomia, and later ultrasound measurements improved performance further. Estimates for birthweight >4500 g require caution given the 77 events. External validation is required before clinical implementation.
Related
- Development and validation of an explainable machine learning-based risk prediction model for macrosomia in han Chinese women with singleton cephalic pregnancies
- Can maternal characteristics improve due date prediction without ultrasound? Development and internal validation of a machine-learning prediction model in 600 000 pregnancies in Sweden
- Risk factors for post-PCI restenosis in patients with coronary heart disease and the development and validation of a nomogram prediction model
- Development and internal validation of an exploratory clinical prediction model for MRI-defined white matter injury in moderate-to-late preterm infants: a single-center retrospective study
- Development and internal validation of a machine learning prediction model for postoperative hemorrhage following modified radical mastectomy for breast cancer
BibTeX
@misc{wu2026development,
title = {{Development and internal validation of early- and mid-pregnancy prediction models for macrosomia in nulliparous singleton pregnancies}},
author = {Y. Wu and M. van der Ven and E. Peri and S. Asvadi and M. B. van der Hout-van der Jagt and S. G. Oei and M. Mischi and X. Long},
howpublished = {medRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.29.26366704},
url = {https://www.medrxiv.org/content/10.64898/2026.10.29.26366704v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - Development and internal validation of early- and mid-pregnancy prediction models for macrosomia in nulliparous singleton pregnancies
AU - Y. Wu
AU - M. van der Ven
AU - E. Peri
AU - S. Asvadi
AU - M. B. van der Hout-van der Jagt
AU - S. G. Oei
AU - M. Mischi
AU - X. Long
PY - 2026
DA - 2026/10/29
PB - medRxiv
DO - 10.64898/2026.10.29.26366704
UR - https://www.medrxiv.org/content/10.64898/2026.10.29.26366704v1
AB - Purpose: Macrosomia is associated with adverse maternal and neonatal outcomes, but early risk stratification remains challenging. Most existing prediction approaches rely on later-pregnancy ultrasound measurements, limiting opportunities for earlier preventive counseling and monitoring. Methods: In this secondary analysis of the prospective nuMoM2b cohort in the United States, we included 6,371 nulliparous singleton pregnancies selected for available periconceptional diet and ultrasound data at Visits 2 and 3. We developed visit-updated multivariable logistic regression models for birthweight >4000 g and >4500 g using routinely available maternal characteristics, anthropometry, periconceptional diet, and serial ultrasound measurements, with sensitivity analyses using large-for-gestational-age birthweight. Model performance was assessed by repeated stratified cross-validation and compared with World Health Organization fetal growth chart-based ultrasound percentile benchmarks, emphasizing the estimated fetal weight percentile as the most clinically familiar proxy. Results: Birthweight >4000 g and >4500 g occurred in 7.9% and 1.2% of pregnancies, respectively. Models updated with later pregnancy information achieved the best discrimination, with an area under the receiver operating characteristic curve of 0.73 for birthweight >4000 g and 0.82 for birthweight >4500 g. In this selected analytic cohort, the early-pregnancy model based on maternal data alone provided moderate discrimination (0.63 and 0.72, respectively), although lower than the strongest third-trimester WHO fetal growth chart-based ultrasound percentile benchmarks. Conclusion: Within this selected cohort, routinely available early-pregnancy maternal characteristics and anthropometry provided moderate risk stratification for macrosomia, and later ultrasound measurements improved performance further. Estimates for birthweight >4500 g require caution given the 77 events. External validation is required before clinical implementation.
ER -