Fig. 7: Alteration of GDM biomarkers in the first trimester and their association with risk of GDM.

a The clustering heat map showed 11 GDM biomarkers with significant differences in the first trimester, including one saliva metabolite, nine serum metabolites, and one urine metabolite. b Analysis of content changes showed that dimethylethanolamine in saliva, and asymmetric dimethylarginine, lysylglycine, malic acid, tyramine-O-sulfate, and uridine in serum were increased in GDM in both the first and second trimesters. Glucosamine 6-phosphate, paracetamol sulfate, pyrrole-2-carboxylic acid, and S-cysteinosuccinic acid in serum, and thiodiacetic acid in urine showed different content changes between GDM and non-GDM in the first and second trimesters. c Forest plot showed correlations between GDM biomarkers in the first trimester and the risk of GDM (GDM group, n = 50 and non-GDM group, n = 50). Blue nodes represented GDM biomarkers in saliva, red nodes represented GDM biomarkers in serum, and yellow nodes represented GDM biomarkers in urine. The ORs were estimated using logistic regression, and 95% CIs were obtained via bootstrap resampling. The height of each bar represented the OR, and the error bars indicated the corresponding 95% CIs. P-values were shown to indicate the statistical significance of each association. d The early prediction model for GDM constructed by first-trimester GDM biomarkers had AUC values of 0.767 (95%CI, 0.611–0.923) in the internal test set and 0.744 (95%CI, 0.655–0.822) in the external test set. Source data were provided as a Source Data file.