Metabolomic markers of fatigue: Association between circulating metabolome and fatigue in women with chronic widespread pain

BACKGROUND: Fatigue is a sensation of unbearable tiredness that frequently accompanies chronic widespread musculoskeletal pain (CWP) and inflammatory joint disease. Its mechanisms are poorly understood and there is a lack of effective biomarkers for diagnosis and onset prediction. We studied the circulating metabolome in a population sample characterised for CWP to identify biomarkers showing specificity for fatigue. MATERIAL AND METHODS: Untargeted metabolomic profiling was conducted on fasting plasma and serum samples of 1106 females with and without CWP from the TwinsUK cohort. Linear mixed-effects models accounting for covariates were used to determine relationships between fatigue and metabolites. Receiver operating curve (ROC)-analysis was used to determine predictive value of metabolites for fatigue. RESULTS: While no association between fatigue and metabolites was identified in twins without CWP (n = 711), in participants with CWP (n = 395), levels of eicosapentaenoate (EPA) ω-3 fatty acid were significantly reduced in those with fatigue (β = − 0.452 ± 0.116; p = 1.2 × 10(− 4)). A significant association between fatigue and two other metabolites also emerged when BMI was excluded from the model: 3-carboxy-4-methyl-5-propyl-2-furanpropanoate (CMPF), and C-glycosyltryptophan (p = 1.5 × 10(− 4) and p = 3.1 × 10(− 4), respectively). ROC analysis has identified a combination of 15 circulating metabolites with good predictive potential for fatigue in CWP (AUC = 75%; 95% CI 69–80%). CONCLUSION: The results of this agnostic metabolomics screening show that fatigue is metabolically distinct from CWP, and is associated with a decrease in circulating levels of EPA. Our panel of circulating metabolites provides the starting point for a diagnostic test for fatigue in CWP.