Abstract
BACKGROUND: Chronic obstructive pulmonary disease (COPD) phenotyping is an approach for developing tailored therapies. The eosinophilic phenotype is associated with exacerbation risk and response to specific treatments. This study evaluates the relationship between sputum and blood eosinophilia, hypothesizing that sputum eosinophil percentage (SpE%) better reflects disease severity and exacerbation risk than blood eosinophil counts (BEC). METHODS: This was a single-center, prospective observational cohort with 107 participants aged 40-80 with clinically diagnosed COPD. Participants completed spirometry, a 6-minute walk test, and questionnaires, and blood and sputum samples were provided at baseline and 3 months. BEC and SpE% were measured via routine complete blood counts and flow cytometric analyses (fluorescence-activated cell sorting [FACS]). Eosinophilic phenotype thresholds were defined as BEC≥300 cells/μL and SpE%≥2%, and associations with clinical characteristics and outcomes were investigated. RESULTS: Adequate sputum specimens were obtained less frequently than blood (60.7% versus 98%). SpE% showed poor repeatability (interclass coefficient 0.36) and poor correlation with FACS (Spearman's 𝜌=0.008, p=0.58). Conversely, BEC showed higher repeatability (𝜌=0.67, p<0.01) and better correlation with FACS (𝜌=0.74, p<0.01). More participants were classified as eosinophilic COPD by sputum (33.3%) than by blood (19.6%). BEC values were poorly correlated with SpE% (𝜌=0.13, P=0.39), and sputum and blood-based diagnostic criteria showed poor agreement (64.5%, Cohen's 𝜅 0.10). High SpE%, but not high BEC, was associated with lower forced expiratory volume in 1 second percentage predicted. CONCLUSIONS: In stable COPD patients, BEC and SpE% did not correlate well, and blood- and sputum-based diagnostic criteria identified different individuals. Defining eosinophilic COPD requires a better understanding of the bio-compartment sampled, testing methods, and cut-off values used.