105-OR: Microbiome-Derived Metabolite Signatures in Early- to Midpregnancy for Risk of Gestational Diabetes (GDM)

Pre-diagnostic disturbances in gut microbiome and metabolome have been associated with an increased risk of diabetes in non-pregnant populations. However, such prospective studies in pregnant women are scant. We aimed to develop and validate microbiome-derived metabolite predictive markers in early to mid-pregnancy for GDM. We designed a nested case-control study (91 GDM, 180 non-GDM; discovery set) and a random subsample (42 GDM, 372 non-GDM; validation set 1) from the PETALS cohort, and a case-control study from the GLOW trial (35 GDM, 70 non-GDM; validation set 2) . We collected fasting serum untargeted metabolomics data at gestational weeks (GW) 10-13 and 16-by gas chromatography/time-of-flight mass spectrometry (TOF-MS) , liquid chromatography (LC) /quadrupole TOF-MS, and hydrophilic interaction LC/quadrupole TOF-MS. Among 1167 annotated metabolites, 193 microbiome-derived metabolites were identified from the Virtual Metabolic Human Database. Multivariate enrichment analysis examined metabolite-GDM associations. Ten-fold cross-validated LASSO prediction models identified metabolite signatures for GDM. At GW 10-13, the branched-chain amino acids cluster was marginally, positively associated with GDM risk (FDR = 0.073) . At GW 16-19, the ceramides, pyridines, pyrimidine nucleosides, and unsaturated fatty acids clusters were significantly, positively associated with GDM risk (all FDR <0.05) . A 38-metabolite panel at GW 10-13 outperformed the model using conventional risk factors including fasting glucose (discovery AUC: 0.915 vs. 0.743; validation 1: 0.865 vs. 0.731; validation 2: 0.970 vs. 0.742; P <0.01) . Similar results were observed for a 21-metabolite panel at GW 16- (discovery AUC: 0.842 vs. 0.760; validation 1: 0.840 vs. 0.774; P <0.01) . Panels of microbiome-derived metabolites in early to mid-pregnancy can predict GDM risk beyond conventional risk factors, suggesting the role of microbiome-metabolome-host interactions in GDM.