Altered adolescents obesity metabolism is associated with hypertension: a UPLC-MS-based untargeted metabolomics study

Abstract

ObjectiveThis study aimed to explore the relationship between the plasma metabolites of adolescent obesity and hypertension and whether metabolite alterations had a mediating effort between adolescent obesity and hypertension.MethodsWe applied untargeted ultra-performance liquid chromatography–mass spectrometry (UPLC-MS) to detect the plasma metabolomic profiles of 105 adolescents. All participants were selected randomly based on a previous cross-sectional study. An orthogonal partial least squares- discriminant analysis (OPLS-DA), followed by univariate statistics and enrichment analysis, was used to identify differential metabolites. Using logistic regression for variable selection, an obesity-related metabolite score (OMS, OMS=∑k=1nβnmetabolite n) was constructed from the metabolites identified, and hypertension risk was estimated.ResultsIn our study, based on P< 0.05, variable importance in projection (VIP) > 1.0, and impact value > 0.1, we identified a total of 12 differential metabolites. Significantly altered metabolic pathways were the sphingolipid metabolism, purine metabolism, pyrimidine metabolism, phospholipid metabolism, steroid hormone biosynthesis, tryptophan, tyrosine, and phenylalanine biosynthesis. The logistic regression selection resulted in a four-metabolite score (thymidine, sphingomyelin (SM) d40:1, 4-hydroxyestradiol, and L-lysinamide), which was positively associated with hypertension risk (odds ratio: 7.79; 95% confidence interval: 2.13, 28.47; for the quintile 4 compared with quartile 1 of OMS) after multivariable adjustment.ConclusionsThe OMS constructed from four differential metabolites was used to predict the risk of hypertension in adolescents. These findings could provide sensitive biomarkers for the early recognition of hypertension in adolescents with obesity.

Targeted blood metabolomics in infants with bronchopulmonary dysplasia

“…Soft Independent Modelling by Class Analogy (v.14.1; Umetrics, Sweden) was utilized for multivariate data analysis. Additionally, orthogonal partial least squaresdiscriminant analysis (OPLS-DA) was employed to enhance classi cation separation, streamline the dataset, and pinpoint potential biomarkers 10 . Model quality was assessed using two key parameters: R2Ycum (goodness of t) and cumulative Q2 (goodness of prediction).…”

“…Model quality was assessed using two key parameters: R2Ycum (goodness of t) and cumulative Q2 (goodness of prediction). A threshold of 0.5 is commonly recognized in model classi cation to distinguish between strong (Q2cum ≥ 0.5) and weak (Q2cum < 0.5) predictive capabilities 10 .…”