Clinical Implications of Low Absolute Blood Eosinophil Count in the SPIROMICS COPD Cohort

Background: The Global Initiative for Chronic Obstructive Lung Disease (GOLD) considers blood eosinophil counts < 100 cells/μL (BEC_≤100) in people with COPD to predict poor inhaled corticosteroid (ICS) responsiveness. However, the BEC_≤100 phenotype is inadequately characterized, especially in advanced COPD.

Research question: Are there differences between GOLD group D patients with high BEC and those with low BEC regarding baseline characteristics and longitudinal outcomes?

Study design and methods: We used multivariable mixed models and logistic regression to contrast clinical characteristics and outcomes of BEC_≤100 vs BEC > 100 (BEC₁₀₀₊) in all subjects with COPD (n = 1,414) and GOLD group D subjects (n = 185) not receiving ICS.

Results: We identified n = 485 with BEC_≤100 (n = 61 GOLD group D) and n = 929 people with BEC₁₀₀₊ (n = 124 GOLD group D). BEC_≤100 status was stable at 6 weeks and approximately 52 weeks (intraclass correlations of 0.78 and 0.71, respectively). Compared with BEC₁₀₀₊, BEC_≤100 comprised more women, with greater current smoking, and less frequent childhood asthma. Among all analyzed participants, the two BEC-defined subsets showed similar rates of lung function decline (mean slope, BEC_≤100 vs BEC₁₀₀₊, -50 vs -39 mL/y; P = .140), exacerbations (0.40 vs 0.36/y; P = .098), subsequent ICS initiation (2.5% vs 4.4%; P = .071), and mortality (7.8% vs 8.4%; P = .715). However, in GOLD group D, people with BEC_≤100 showed higher exacerbation rates within 365 days of enrollment (0.62 vs 0.33/y; P = .002) and total follow-up (1.16 vs 0.83/y; P = .014). They also had greater lung function decline (mean slope of -68 mL/y vs -23 mL/y; P = .036) and had greater emphysema at baseline (voxels < 950 Hounsfield units at total lung capacity of 7.46% vs 4.61%; P = .029).

Interpretation: In non-ICS-treated GOLD group D COPD, people with BEC_≤100 had more baseline emphysema, prospective exacerbations, and lung function decline. Our analysis has identified a particularly vulnerable subpopulation of people with COPD, suggesting the need for studies focused specifically on their therapeutic treatment.

Clinical trial registration: ClinicalTrials.gov; No.: NCT01969344; URL: www.

Clinicaltrials: gov.