The Metabolic Signature of Cardiorespiratory Fitness

Abstract

High cardiorespiratory fitness (CRF) is associated with better overall health. This study aimed to find a metabolic signature associated with CRF to identify health-promoting effects. CRF based on cardiopulmonary exercise testing, targeted and untargeted metabolomics approaches based on mass spectrometry, and clinical data from two independent cohorts of the Study of Health in Pomerania (SHIP) were used. Sex-stratified linear regression models were adjusted for age, smoking, and height to relate CRF with individual metabolites. A total of 132 (SHIP-START-2: 483 men with a median age of 58 years and 450 women with a median age of 56 years) and 118 (SHIP-TREND-0: 341 men and 371 women both with a median age of 51 years) metabolites were associated with CRF. Lipids showed bidirectional relations to CRF independent of sex. Specific subsets of sphingomyelins were positively related to CRF in men (SM (OH) C14:1, SM(OH)C22:2 SM C16:0, SM C20:2 SM(OH)C24:1) and inversely in women (SM C16:1, SM C18:0, SM C18:1). Metabolites involved in energy production (citrate and succinylcarnitine) were only associated with CRF in men. In women, xenobiotics (hippurate, stachydrine) were related to CRF. The sex-specific metabolic signature of CRF is influenced by sphingomyelins, energy substrates, and xenobiotics. The greater effect estimates seen in women may emphasize the important role of CRF in maintaining metabolic health. Future research should explore how this profile changes with different types of exercise interventions or diseases in diverse populations and how these metabolites could be implemented in primary prevention settings.

Keywords: cardiorespiratory fitness; epidemiology; exercise testing; metabolomics.

FIGURE 1

Flow chart of the study participants for SHIP‐START‐2 and SHIP‐TREND‐0.

FIGURE 2

Color‐coded corrected P values (controlling the false discovery rate (FDR) at 0.05) for the association of plasma metabolites with PeakVO₂ in men and women separately. Significant associations (FDR < 0.05) are marked with a black box. Linear regression models were adjusted for age, height and smoking status. Orange and blue shading indicate positive and inverse associations, respectively. Analyses were separately performed for SHIP‐START‐2 (S2) and SHIP‐TREND‐0 (T0). Metabolites were measured by targeted metabolomics (Biocrates). Gray indicates metabolites that were excluded in SHIP‐TREND‐0. Metabolites that were not significant in either men or women are only displayed in Figure S1.

FIGURE 3

Color‐coded corrected p values (controlling the false discovery rate (FDR) at 0.05) for the association of plasma metabolites, measured by untargeted metabolomics (Metabolon), with PeakVO₂ in SHIP‐TREND‐0 (T0) men and women separately. Significant associations (FDR < 0.05) are marked with a black box. Linear regression models were adjusted for age, height and smoking status. Orange and blue shading indicate positive and inverse associations, respectively. Unknown significant metabolites are only displayed in Figure S2.

FIGURE 4

The linear association between tyrosine, phenylalanine, carnitine, lysoPCaC18:2, lactate, and kynurenine with peakVO₂ (mL/min/kg) for men and women in SHIP‐START‐2 and SHIP‐ TREND‐0. *Metabolitelevels were log2‐transformed; beta estimated per standard deviation increase in peakVO₂ with standard error were given. FDR = false discovery rate.