Objective: Differentiating between asthma and chronic obstructive pulmonary disease (COPD) and treating them correctly is challenging because their symptoms overlap and current diagnostic methods are inadequate. We explored the potential of integrating breath metabolomics with sputum inflammatory phenotyping to enhance the discrimination and characterization of respiratory diseases.
Methods: Exhaled breath samples were gathered from 74 participants (35 with asthma, 39 with COPD), and sputum samples were examined for eosinophil and neutrophil counts. Orthogonal partial least squares-discriminant analysis (OPLS-DA), logistic regression, and principal component analysis (PCA) were used for the identification of significant volatile organic compounds (VOCs) that discriminate between the diseases, and to determine their relationship to disease state and inflammatory classifications.
Results: Breath metabolomics effectively distinct asthma from COPD with 93.2% accuracy, identifying 10 important VOC biomarkers. Significantly, predicting sputum eosinophilia and neutrophilia phenotypes using established thresholds was more accurate (80.6%-90.5%, AUROC 0.67-0.91) compared to directly differentiating the diseases, where each inflammatory subtype presented its own VOC profile. Specific biomarkers, such as allyl methyl sulfide, epizonarene, and camphor, were identified as potential indicators of disease status and inflammatory subtypes.
Conclusion: Using breath metabolomics in conjunction with sputum inflammatory phenotyping shows promise in effectively distinguishing between asthma and COPD, as well as identifying inflammatory subtypes, which can aid in creating personalized treatment strategies.
Keywords: Asthma; breath metabolomics; chronic obstructive pulmonary disease; personalized medicine; sputum inflammatory phenotyping.