84: Metabolomic testing for gestational diabetes: utility & reproducibility

Abstract

Glucose tolerance testing for gestational diabetes (GDM) has poor reproducibility between tests and high variability by time of day. In contrast, metabolomic testing of serum has good reproducibility for multiple metabolites, including those altered in GDM. We hypothesize that metabolomics can identify GDM and maintain reproducibility over 3-7 days. Prospective observational study of women with singleton gestation undergoing GDM screening at 24- 28 wks. Two fasting plasma samples were obtained 3-7 days apart. Women were excluded for major medical illness. Exposure groups were defined as glucose tolerant (GT), defined as 1-hour, 50-g glucose challenge test <120mg/dL, and GDM, defined by Carpenter Coustan criteria on a 3-hour, 100-g glucose tolerance test. We determined a priori that 35 subjects per group were needed to detect a mean difference of 1-standard deviation with an alpha of 0.01 and 90% power between GDM and GT. Aqueous extracts of plasma were analyzed by reverse-phase nanoLC-ESI- MS/MS on a SCIEX 5600 TripleTOF in the positive ion mode. Data processing: XCMSonline (peak alignment), MetaboAnalyst (statistics) and Mummichog (pathways). Partial least squared discriminant analysis (PLS-DA) was performed to detect differences in the four groups: GDM visit 1, GDM visit 2, GT visit 1, GT visit 2. Of 65 patients with 2 samples, 30 (46%) had GDM. Women with GDM were significantly more likely to be Hispanic; GT women were mostly black (p<0.01). GDM samples were obtained at a slightly later gestational age (27.5 vs 26.0 wks, p<0.001). Overall, distinct differences were seen between GT and GDM groups. Although there were some variation in metabolomics over a 3-7 day period, several analytes of interest that can be used to distinguish GDM from GT remained stable over a 3-7 day period (Figure 1). Random forest plots demonstrate the samples separating based on GDM, not visit. Pathway analysis demonstrated differences in ascorbate, amino acid, and glycerophospholipid metabolism between GDM and GT; these pathways were not identified in differences between visits. The metabolomics profile at 24-28 wks can distinguish GDM and is reproducible over a 3-7 day period. Further exploration of the specific metabolites and the possibility of using combinations of these metabolites to diagnose GDM should be explored further.View Large Image Figure ViewerDownload Hi-res image Download (PPT)