Rationale: Though the inflammatory process in COPD is predominantly driven by macrophages, neutrophils and lymphocytes, recently blood and sputum eosinophils have received attention, specifically as predictors of response to inhaled and oral corticosteroids. Most data with blood eosinophil counts is cross-sectional, which does not inform us about reproducibility, changes over time and with exacerbations.
Methods: In a longitudinal study at a VA Medical Center from 1994 to 2014, patients with COPD were seen for monthly clinical evaluations. Sputum and serum samples were obtained monthly and at exacerbation. Blood eosinophil counts were obtained from the electronic medical records for the period of time that each patient participated in the study, Mean, absolute and percent, eosinophil count for each patient were calculated, and patients were grouped using cut-offs for absolute eosinophil counts of < 200 to 2 200, < 300 to 300, < 400 to 400 and % eosinophils of < 2 to <3 to <4 to ?.4 and <6 to .-?6. 'Low' and 'High' eosinophil groups were compared for participant demographics, spirometry, smoking exposure and exacerbation frequency with appropriate statistical methods. We defined significant changes in blood eosinophils counts by either a 2-fold or a 2 standard deviation increase or decrease from the mean value of each patient.
Results : In the 182 study participants, defined with the various cut-offs, ‘Low’ and ‘High’ eosinophil groups did not differ in age, body mass index and pack years of smoking (p>0.05 for all). Patients with ‘low’ eosinophils defined as <200 or <300 had lower FEV1 (p=0.0014, p=0.016, respectively). Patients with <2% eosinophils also had lower FEV1% (p=0.0019). There was no difference in exacerbation frequency between the 'low; and 'high' eosinophil groups. There were 186 events of significant change in blood eosinophils in 72 patients, 44 events by change of 2 standard deviations and 142 with the 2-fold criteria.
Conclusions: Increasing severity of airflow obstruction in COPD is associated with lower blood eosinophils. Exacerbation frequency in our cohort was not related to blood eosinophil counts. Variation in blood eosinophils occurs often, and that could affect their usefulness as biomarkers to make treatment decisions. We are analyzing whether this variation was related to medications and exacerbations.