Metabolic Dynamics and Prediction of Gestational Age and Time to Delivery in Pregnant Women

Pregnancy is a vital time period for both mothers and fetuses, and obstetric care can greatly increase positive outcomes of mothers and children. Accurate prediction of gestational age is one of the many important components of care during pregnancy. However, there are only 2 main methods in dating a pregnancy-early ultrasound and last menstrual period. While there are markers in maternal blood that can be used to estimate gestational age, there is currently no precise way to do so. Therefore, the goal of this study is to identify an accurate and cost-effective way to estimate gestational age through blood metabolites. In this study, pregnancy-related metabolites and metabolic pathways were identified to create a picture of the metabolic changes that take place during pregnancy and the postpartum period. With this, the researchers built a metabolic clock to predict gestational age and determine unique pregnancy variations undetectable by ultrasound that could potentially affect delivery timing. A cohort of 38 Danish pregnant women were enrolled in this single-center study spanning multiple years. Weekly blood draws were performed on 30 of these women starting at 5 weeks' gestation and ending during the postpartum period. Of these, 21 women were assigned to a discovery cohort, and 9 were assigned to a validation cohort. In total, 784 samples were analyzed in 2 separate years from these 30 women. There were an additional 8 women assigned to a second validation cohort, and their samples were analyzed 3 years after the discovery cohort. The 784 samples were analyzed using liquid chromatography-mass spectrometry for untargeted metabolomics. In all, 9651 metabolic features were identified across the different samples, and 4995 features (51.7%) appeared altered during pregnancy and/or the postpartum period. This suggests that numerous metabolic changes take place during pregnancy. Through this study, the researchers found that metabolites with unidirectional behaviors increased throughout pregnancy and reached their maximum before labor. To better understand the pregnancy-related metabolites, 952 metabolites were mapped to 687 compounds. Significance analysis for microarrays was applied to investigate the quantity of the compound and reported gestational age at the time the blood was collected. Of the 687 annotated compounds, 460 were associated with pregnancy (67.0%; false discovery rate <0.05). Overall, the results suggest that human metabolites change on a scheduled, systematic level throughout pregnancy. Correlation analyses were performed on the top 68 pregnancy-related compounds in order to identify the functional groups of metabolites that alter during pregnancy. In these compounds, metabolites that were increased or decreased clustered together in 3 main groups: steroid clusters, lipid clusters, and nonlipid clusters. The steroid clusters positively correlated with almost all upregulated metabolites. Comparatively, all downregulated metabolites negatively correlated with this cluster. Taken together, this suggests that steroid hormones are important components of the metabolome during pregnancy. In the lipid cluster, the quantity of most long-chain fatty acids decreased after birth. In terms of the nonlipid cluster, metabolites were regularly elevated throughout pregnancy. Overall, the functional metabolite groups varied in a very programmed way during pregnancy. This is true in many components of the adrenal cortex gland, gonad, and placenta. Furthermore, women with a delayed metabolic clock had a more delayed natural labor onset and delivery compared to the due date based on ultrasound. The results of this study suggest that testing metabolites during pregnancy may eventually benefit pregnant women by predicting accurate delivery timing.