Exploring pathological mechanism and potential diagnosis biomarker‐Screening Gestational diabetes mellitus differential metabolites : A cross‐sectional cohort study

To study the metabonomics differences between pregnant women with GDM in the third trimester and those in the normal group. Screening a group of highly efficient and sensitive markers for GDM and validated the previously published early metabolic markers of GDM METHODS: A cross-sectional cohort study based on UPLC-MS/MS untargeted metabolomics analysis of serum samples were collected from 59 with GDM and 59 normal pregnant women.121 metabolites were detected, and 27 significantly were identified differential metabolites in GDM and control. The combination of 27 metabolic peaks had the area under curve(AUC)values of 0.90, 0.92, and 0.93 in the prediction models using support vector machine, PLS, and random forest, respectively. Finally, five metabolite biomarkers were selected to construct logistic regression models: l-valine, hypoxanthine, eicosapentaenoic acid 2-amino-1,3,4-octadecanotriol(2-AO), and choline. The AUC value of these metabolites was 0.769 between the GDM group and the control group.The discovery of a group of differential metabolites in pregnant women with GDM in the third trimester and normal pregnant women may facilitate the study of the pathological mechanism of GDM, and it is possible to find an efficient and sensitive alternative GDM detection method.