RMD Open. 2025 May 8;11(2):e005278. doi: 10.1136/rmdopen-2024-005278.
ABSTRACT
OBJECTIVES: To identify specific metabolomic profiles associated with gout flares in people with gout.
METHODS: Participants with gout were sequentially enrolled. In cross-sectional analysis, data were analysed according to the presence of gout flare (acute group) or absence of gout flare (intercritical group) at the time of enrolment. Participants in the intercritical group were prospectively followed and analysed according to the development of gout flares (recurrent flare group) or no gout flare (no flare group) over 1 year. Relative abundances of metabolites in serum obtained at the baseline visit were measured by untargeted liquid chromatography-mass spectrometry. Risk of incident flare was analysed using least absolute shrinkage and selection operator (LASSO)-Cox regression and time-receiver operating characteristic (ROC). Machine learning models were performed to identify biomarkers in cross-sectional and longitudinal analysis, which was further optimised using quantitative targeted metabolomics in an independent validation cohort.
RESULTS: Participants in the acute and intercritical groups showed distinct metabolic profiles, including carbohydrate, lipid and nucleotide metabolism. Many metabolites were associated with recurrent gout flare in the prospective analysis. The metabolic risk score with six LASSO-derived metabolites, including 5-methoxytryptamine, differentiated well for gout flare risk, yielding an area under the ROC curve (AUC) of 0.82 (95% CI 0.74 to 0.90). Machine learning models achieved an AUC of 0.828 for comparison between the acute and intercritical groups. For the prediction of recurrent flare, AUC reached 0.807-0.867 with combined metabolites and clinical measurements.
CONCLUSIONS: Metabolic reprogramming differentiates between the acute and intercritical stages of gout, and implicated metabolites may serve as biomarkers for future gout flares.
PMID:40345707 | DOI:10.1136/rmdopen-2024-005278