Endocrine Abstracts

Background: Using small for gestational age (SGA) as a marker for fetal growth restriction (FGR), studies link an adverse intrauterine environment to cardiometabolic risk markers in childhood. Focusing on 3–6 year old children, where the majority were born following pregnancies at greater risk of suboptimal fetal growth (SFG) but only a minority were born SGA, cardiometabolic risk markers were measured and blood samples collected for metabolomic analysis. Nuclear magnetic resonance (NMR) data previously implicated the arginine–nitric oxide pathway, alongside higher childhood systolic blood pressure (SBP). Liquid chromatography mass spectroscopy (LCMS), a more sensitive technique, has now been applied to these specimens.

Aims: 1) Determine whether differences exist in the LCMS metabolome of this cohort. 2) Use a supervised approach based on quartiles of weight trajectory to establish differentially expressed metabolites (DEMs) between groups.

Methods: Fetal and childhood growth trajectories were divided into quartiles, and cardiometabolic differences established (BSPED 2022, OC 6.2). These included SBP, HDL (both P<0.05) and serum insulin (P=0.08). K means clustering, a machine learning approach, was used to establish whether underlying differences in the LCMS metabolome exist. Cardiometabolic differences and DEMs were examined between clusters. As a supervised approach, DEMs were identified between weight trajectory quartiles and used to examine underlying metabolic pathways using specialist software, Metaboanalyst.

Results: Underlying differences in the LCMS metabolome separated participants into two groups. These differed in sum of skinfold thicknesses (P=0.03) and brachial augmentation index (P=0.04), a measure of arterial stiffness. Although 130 pathways were identified, none were significant after multiple testing correction. Supervised analyses identified pathways supporting previous findings (nitrogen metabolism, P=0.01, arginine/proline metabolism, P=0.04), as well as those relating to HDL/LDL-cholesterol (vitamin B3, P=0.04) and insulin resistance (carnitine shuttle, P=0.04).

Conclusions: Differences in the underlying metabolome exist, relating to cardiometabolic risk markers in childhood. Supervised analysis of the LCMS metabolome supported involvement of the arginine–nitric oxide pathway in relation to SBP, and revealed others involved in lipid and glucose metabolism. Integration with gene expression (transcriptome) markers within this cohort could reveal early-life markers of raised cholesterol and insulin resistance.