The effects of gestational diabetes mellitus with maternal age between 35 and 40 years on the metabolite profiles of plasma and urine

Abstract

Background: Gestational diabetes mellitus (GDM) is defined as impaired glucose tolerance in pregnancy and without a history of diabetes mellitus. While there are limited metabolomic studies involving advanced maternal age in China, we aim to investigate the metabolomic profiling of plasma and urine in pregnancies complicated with GDM aged at 35-40 years at early and late gestation.

Methods: Twenty normal and 20 GDM pregnant participants (≥ 35 years old) were enlisted from the Complex Lipids in Mothers and Babies (CLIMB) study. Maternal plasma and urine collected at the first and third trimester were detected using gas chromatography-mass spectrometry (GC-MS).

Results: One hundred sixty-five metabolites and 192 metabolites were found in plasma and urine respectively. Urine metabolomic profiles were incapable to distinguish GDM from controls, in comparison, there were 14 and 39 significantly different plasma metabolites between the two groups in first and third trimester respectively. Especially, by integrating seven metabolites including cysteine, malonic acid, alanine, 11,14-eicosadienoic acid, stearic acid, arachidic acid, and 2-methyloctadecanoic acid using multivariant receiver operating characteristic models, we were capable of discriminating GDM from normal pregnancies with an area under curve of 0.928 at first trimester.

Conclusion: This study explores metabolomic profiles between GDM and normal pregnancies at the age of 35-40 years longitudinally. Several compounds have the potential to be biomarkers to predict GDM with advanced maternal age. Moreover, the discordant metabolome profiles between the two groups could be useful to understand the etiology of GDM with advanced maternal age.

Keywords: Advanced maternal age; Gestational diabetes mellitus; Metabolomics; Pregnancy.

Fig. 1

Heatmap showing the metabolome profiles of plasma and urine in the first and third trimester. The relative abundance of metabolites is illustrated on a log2 scale. Fold difference of metabolite concentrations compared with their corresponding controls are plotted as shades of purple (increasing levels) or yellow (decreasing levels). Only metabolites with a significant p-value (Tukey’s HSD: p < 0.05), q-value (FDR: q < 0.1) are shown

Fig. 2

ROC curve of plasma metabolites with AUC above 0.75 between GDM and normal pregnancies. Seven metabolites in the first (a) and four metabolites in the third (b) trimester. A multivariant ROC model and corresponding 95% Confidence Interval (CI) combining the seven and four metabolites are shown in the last plot

Fig. 3

ROC curve of urine metabolites with AUC above 0.75 between GDM and normal pregnancies. Three metabolites in the third trimester between the two groups. A multivariant ROC model and corresponding 95% CI combining all of the three metabolites are plotted

Fig. 4

KEGG pathway analysis. The predicated pathway activities were illustrated with log2 scale (a). The activities of the pathways are plotted by purple color (upregulation) and yellow color (downregulation). Only metabolites with a significant p-value (Tukey’s HSD: p < 0.05), q-value (FDR: q < 0.15) are shown. The different metabolites involved in the corresponding pathways (b)