Background: There is controversy regarding the use of blood eosinophil levels as a biomarker of exacerbation risk and responsiveness of patients to inhaled corticosteroids (ICS).
Methods: Patients in stable COPD with Gold Initiative for Chronic Obstructive Lung Disease airflow obstruction grades II to IV were enrolled in an observational multicenter trial. Concordance was defined as blood eosinophil values persistently lower than or persistently higher than the absolute cutoff points of 150 cells/μL and 300 cells/μL, or the percentage cutoff points of 2%, 3%, and 4%. Discordance was obtained when the blood eosinophil values varied between any two visits. ICS treatment data were recorded at one time point at the inclusion of the study.
Results: A total of 210 patients with 2,059 visits were included in the study. Seventy percent of the patients were male, and 36% were current smokers; their average age was 67.7 ± 9.4 years, and 81% were receiving ICS at the start of the study. Assessing eosinophil levels over time (median, 7 days [4; 12]), irrespective of exacerbation or hospitalization, there was a discordance of 77%, 60%, and 42% when using the 2%, 3% and 4% cutoffs, respectively. This outcome changed to 34.5%, 24%, and 17.2% discordance when only using two visits for the analysis. The discordance was similar when using absolute eosinophil values. Patients in a stable state had higher discordant values than patients with mild/moderate exacerbations. The same was seen in patients hospitalized for other illnesses compared with patients hospitalized for severe exacerbation of COPD. Discordancy was high regardless of whether patients were taking ICS at the beginning of the study period.
Conclusions: These study data suggest that blood eosinophil levels present significant variability throughout the course of COPD, and a single measurement may therefore not be a reliable predictor of ICS response.
Keywords: blood eosinophils; concordance; discordance; exacerbation.
Copyright © 2019 American College of Chest Physicians. Published by Elsevier Inc. All rights reserved.