Key points are not available for this paper at this time.
Abstract Early prediction of pre-eclampsia (PE) is crucial for timely intervention and medical monitoring. The accuracy of existing prediction models is limited, especially in the Chinese population. Here, we conducted a retrospective cohort analysis of 3,772 pregnancies from eight hospitals across China. Using ultra-high performance liquid chromatography-mass spectrometry (UHPLC-MS) and enzyme-linked immunoassay (ELISA) techniques, a novel biomarker IGFBP1 was identified in maternal plasma samples. Furthermore, white blood cell (WBC), platelet (PLT), monocyte count (MO#), gamma-glutamyl transferase (GGT), high-density lipoprotein cholesterol (HDL-C), aspartate aminotransferase (AS)/alanine aminotransferase (AL), and uric acid (UA) were systemically evaluated as indicators from 90 routine laboratory tests. Machine learning model incorporating maternal factors, protein biomarkers, and laboratory indicators outperforming existing prediction model and validated in an external cohort (EPE: AUC 0.95, sensitivity 92.86%, specificity 90% and LPE: AUC 0.84, sensitivity 55.93%, specificity 90%). Those results suggest our study provide a novel protein biomarker and a valuable prediction strategy for early prediction and management of PE in the obstetric clinic.
Qi et al. (Wed,) studied this question.