Abstract
Preeclampsia (PE) and fetal growth restriction (FGR) complicate 5-10% of pregnancies and are major causes of maternal and fetal morbidity and mortality. Here we demonstrate that measuring circulating cell-free RNAs (cfRNAs) from maternal plasma can accurately predict pregnancies complicated by the combination of PE and FGR. We investigated 751 maternal plasma samples from 195 pregnant women (39 cases; 156 non-cases). We developed machine learning models from our discovery cohort (15 cases; 60 non-cases) and evaluated their predictive performances internally (24 cases; 96 controls) and externally (40 cases; 73 non-cases). We found circulating leptin (LEP) and pappalysin2 (PAPPA2) cfRNAs are the strongest cfRNA predictors of complicated pregnancies, each with an area under the receiver operating characteristic curve (AUC) of ~0.82. Using an external validation dataset of women with established PE, the combination of LEP and PAPPA2 had an AUC ~0.951. Our findings show that cfRNAs can predict complications of human pregnancy.