Abstract P202: Non-targeted Metabolomics Study Identify Multiple Metabolites Associated With Body Mass Index

Abstract

Background: Non-targeted metabolomics technology provides a powerful tool to examine the complex etiology of obesity and the underlying mechanisms for its health consequences. The current study aimed to examine the associations of metabolites, quantified by non-targeted metabolomics technology, with body mass index (BMI) in the Bogalusa Heart Study. Methods: The study included 1,261 participants (825 White and 436 Blacks) aged 34-58 years. General linear models were used to examine the associations of BMI with the 1,202 metabolites that passed rigorous quality control measures, adjusted for age, sex, smoking, drinking, education, and total physical activity in Blacks and Whites, separately. Weighted correlation network analysis (WGCNA) was used to build metabolites modules according to pair-wise correlations (signed positive correlation and unsigned correlation regardless of correlation direction) among all metabolites; partial correlation was used to assess the link between each module and BMI. Results: Six-eight metabolites showed Bonferroni-corrected (P < 4.16E-5) associations with BMI in both Blacks and Whites. The most significant metabolite was glutamate (P = 1.52E-12 in Blacks and P = 1.47E-37 in Whites). Among the 46 significant metabolites with known identities, a majority were involved in pathways of lipids (22 metabolites) and amino acids (13 metabolites). Fifteen unsigned metabolites modules were identified and six of the modules showed significant correlation with BMI (absolute r = 0.11-0.14); ten signed metabolites modules were identified and five of them were significantly correlated with BMI (absolute r = 0.14-0.31). Conclusion: We have identified multiple metabolites robustly associated with BMI, which have novel biological implications for obesity and its health consequences.