RT Journal Article SR Electronic T1 Plasma lipidomics profile in pregnancy and gestational diabetes risk: a prospective study in a multiracial/ethnic cohort JF BMJ Open Diabetes Research & Care JO BMJ Open Diab Res Care FD American Diabetes Association SP e001551 DO 10.1136/bmjdrc-2020-001551 VO 9 IS 1 A1 Mohammad L Rahman A1 Yen-Chen A Feng A1 Oliver Fiehn A1 Paul S Albert A1 Michael Y Tsai A1 Yeyi Zhu A1 Xiaobin Wang A1 Fasil Tekola-Ayele A1 Liming Liang A1 Cuilin Zhang YR 2021 UL http://drc.bmj.com/content/9/1/e001551.abstract AB Introduction Disruption of lipid metabolism is implicated in gestational diabetes (GDM). However, prospective studies on lipidomics and GDM risk in race/ethnically diverse populations are sparse. Here, we aimed to (1) identify lipid networks in early pregnancy to mid-pregnancy that are associated with subsequent GDM risk and (2) examine the associations of lipid networks with glycemic biomarkers to understand the underlying mechanisms.Research design and methods This study included 107 GDM cases confirmed using the Carpenter and Coustan criteria and 214 non-GDM matched controls from the National Institute of Child Health and Human Development Fetal Growth Studies-Singleton cohort, untargeted lipidomics data of 420 metabolites (328 annotated and 92 unannotated), and information on glycemic biomarkers in maternal plasma at visit 0 (10–14 weeks) and visit 1 (15–26 weeks). We constructed lipid networks using weighted correlation network analysis technique. We examined prospective associations of lipid networks and individual lipids with GDM risk using linear mixed effect models. Furthermore, we calculated Pearson’s partial correlation for GDM-related lipid networks and individual lipids with plasma glucose, insulin, C-peptide and glycated hemoglobin at both study visits.Results Lipid networks primarily characterized by elevated plasma diglycerides and short, saturated/low unsaturated triglycerides and lower plasma cholesteryl esters, sphingomyelins and phosphatidylcholines were associated with higher risk of developing GDM (false discovery rate (FDR) <0.05). Among individual lipids, 58 metabolites at visit 0 and 96 metabolites at visit 1 (40 metabolites at both time points) significantly differed between women who developed GDM and who did not (FDR <0.05). Furthermore, GDM-related lipid networks and individual lipids showed consistent correlations with maternal glycemic markers particularly in early pregnancy at visit 0.Conclusions Plasma lipid metabolites in early pregnancy both individually and interactively in distinct networks were associated with subsequent GDM risk in race/ethnically diverse US women. Future research is warranted to assess lipid metabolites as etiologic markers of GDM.