Comparison of blood and sputum analysis in evaluation of eosinophilic COPD

<b>Rationale:</b> The use of eosinophils (eos) as a biomarker in management of COPD has been validated but a standardized approach for the interpretation of sampling methods has not been established. Appreciating that there are differences between sputum and blood eosinophil count (BEC) analysis, we sought to compare the two sampling methods. <b>Methods:</b> In an interim analysis of an ongoing observational study at UCLA, 108 subjects with stable moderate/severe COPD and 14 individuals having exacerbations were evaluated on 2 occasions over a 3-month interval, each time completing extensive clinical and biomarker characterization. <b>Results:</b> Agreement between sputum and blood eos values was poor (κ=0.05). In contrast to sputum analysis, BEC was more consistently captured (94.9 v 48.7%), was more repeatable (76.7%, κ=0.53 for BEC v 72.7%, κ=0.29 for sputum) and showed better cell count correlation between cytology and flow cytometry (ρ=0.89 v ρ=0.06, respectively). Using cutoff values ≥150 cells/mL for BEC, and ≥2% for sputum, we classified more individuals with eosinophilic COPD using blood (58.8%) v sputum (31.7%). High sputum eos were associated with history of exacerbations, lower FEV₁, and symptoms in COPD individuals, but this was not observed with high BEC (Figure 1). <b>Conclusions:</b> Routine testing for sputum eos is less practical in daily management of COPD but, compared to BEC cutoff ≥150, sputum Eos ≥2% may better distinguish eosinophilic COPD.