Estimating due date in twin pregnancy by second- and third-trimester ultrasound

To determine the optimal sonographic dating approach for twin gestations presenting to antenatal care beyond the first trimester.

This was a retrospective cohort study of twin pregnancies receiving care at the University of North Carolina between November 2017 and August 2023 for which the gestational age had been established definitively by first-trimester ultrasound or in-vitro fertilization. All ultrasound studies with complete biometry conducted ≥ 14 weeks' gestation were included. We assessed the accuracy of gestational-age dating using the biometric data of the smaller twin, the larger twin and an average of both cotwins ('average twin'), by comparing estimates with the previously established gestational age ('ground truth').

Our analysis dataset consisted of 596 twin pregnancies. The mean absolute error ± standard error (SE) was 5.8 ± 0.1 days for the smaller twin, 4.4 ± 0.1 days for the average twin and 4.1 ± 0.1 days for the larger twin. The mean error ± SE showed that all three measurements underestimated gestational age compared to the ground truth, with the larger twin exhibiting the least bias (smaller twin, -4.5 ± 0.1 days; average twin, -2.3 ± 0.1 days; larger twin, 0.0 ± 0.1 days). These results persisted when the dataset was restricted to one ultrasound scan per pregnancy and when stratified by chorionicity.

In the second and third trimesters, gestational-age dating in twin pregnancies is more accurate and less biased when using the biometry of the larger twin compared with that of the smaller twin or an average of the two twins. © 2025 International Society of Ultrasound in Obstetrics and Gynecology.