Background
Selective fetal growth restriction (sFGR) is an extreme complication that significantly increases the risk of perinatal mortality and long-term adverse neurological outcomes in offspring, affecting approximately 15% of monochorionic diamniotic (MCDA) twin pregnancies. The lack of longitudinal cohort studies hinders the early prediction and intervention of sFGR.
Methods
We constructed a prospective longitudinal cohort study of sFGR, and quantified 25 key metabolites in 337 samples from maternal plasma in the first, second, and third trimester and from cord plasma. In particular, our study examined fetal growth and brain injury data from ultrasonography and used the Ages and Stages Questionnaire-third edition subscale (ASQ-3) to evaluate the long-term neurocognitive behavioral development of infants aged 2–3 years. Furthermore, we correlated metabolite levels with ultrasound data, including physical development and brain injury indicators, and ASQ-3 data using Spearman’s-based correlation tests. In addition, special combinations of differential metabolites were used to construct predictive models for the occurrence of sFGR and fetal brain injury.
Results
Our findings revealed various dynamic patterns for these metabolites during pregnancy and a maximum of differential metabolites between sFGR and MCDA in the second trimester (n = 8). The combination of l-phenylalanine, l-leucine, and l-isoleucine in the second trimester, which were closely related to fetal growth indicators, was highly predictive of sFGR occurrence (area under the curve [AUC]: 0.878). The combination of l-serine, l-histidine, and l-arginine in the first trimester and creatinine in the second trimester was correlated with long-term neurocognitive behavioral development and showed the capacity to identify fetal brain injury with high accuracy (AUC: 0.94).
Conclusions
The performance of maternal plasma metabolites from the first and second trimester is superior to those from the third trimester and cord plasma in discerning sFGR and fetal brain injury. These metabolites may serve as useful biomarkers for early prediction and promising targets for early intervention in clinical settings.