Metabolomic Signatures May Aid COPD Diagnosis and Prognosis

Machine learning models based on targeted plasma metabolomics distinguished chronic obstructive pulmonary disease (COPD), early-stage disease, and clinical subtypes and predicted several longitudinal outcomes in a 2-cohort study published as an Article in Press in Nature Communications.

Investigators analyzed 1,344 participants in 2 cohorts. Both cohorts were prospective. The discovery cohort included 787 participants (321 with COPD and 466 without COPD); 651 completed 3 years of follow-up. The external validation cohort included 557 participants (372 with COPD and 185 without COPD) and provided 1-year follow-up data. Targeted liquid chromatography-mass spectrometry quantified 206 plasma metabolites. Diagnostic models combined selected metabolites with age, sex, body mass index, and smoking status.

The model for established COPD used 12 metabolites plus clinical variables and achieved an area under the receiver operating characteristic curve (AUROC) of 0.883 (95% CI, 0.839-0.928) in the discovery-cohort test set and 0.813 (95% CI, 0.776-0.849) in the validation cohort. A 27-metabolite model with clinical variables showed more modest discrimination for early COPD, with AUROCs of 0.750 and 0.638, respectively; validation accuracy was 63.8%.

For subtype classification, combined metabolite-clinical models achieved validation AUROCs of 0.849 for small-airway disease and 0.760 for emphysema-predominant COPD. In the discovery cohort, an 11-metabolite model predicted rapid lung-function decline over 3 years with an AUROC of 0.768 (95% CI, 0.696-0.831). A 17-metabolite exacerbation model achieved an AUROC of 0.775 (95% CI, 0.711-0.836) and identified 69.4% of cases at a 0.5 cutoff.

The authors noted that targeted metabolomics limited metabolome coverage, some metabolite effects differed between cohorts, and follow-up was relatively short. The analysis was limited to participants of Chinese ancestry, and the authors characterized the metabolomic analysis as cross-sectional. They called for longer follow-up, larger cohorts, and validation in more diverse populations to determine the broader clinical utility of the signatures.

Source: Li C, Yao J, Wu F, et al. Plasma metabolomic signatures enable the diagnosis and prognosis of chronic obstructive pulmonary disease. Nat Commun. 2026. doi:10.1038/s41467-026-77040-y