Persistently low blood eosinophils identify a high-risk phenotype in COPD

Blood eosinophil count (BEC) is used to guide COPD therapy, yet its longitudinal stability and prognostic relevance remain debated. We investigated BEC variability and the clinical implications of eosinophil trajectories. Stable COPD patients (n = 471) were prospectively recruited in Singapore from 2013 to 2025. BEC stability was assessed using paired baseline and 1-year measurements. Patients were classified as persistently low (LL), persistently high (HH), or variable (VAR) based on a 0.15 × 10⁹/L threshold across 1 year period. Associations with exacerbations and mortality were examined using multivariable regression, with findings validated in an independent cohort from China (n = 66). Among 471 patients, low baseline BEC (< 0.15 × 10⁹/L) was associated with worse lung function, higher symptom burden, and increased unadjusted mortality (p < 0.05). Paired BEC measurements (n = 176) showed moderate reproducibility (intraclass correlation coefficient = 0.423), indicating within-individual variability. Trajectory analysis revealed that compared to the LL group, both HH and VAR groups had significantly lower historical exacerbation rates. During one-year follow-up, the LL group experienced higher rates of hospitalized exacerbations. Kaplan-Meier analysis demonstrated a trend toward increased mortality in the LL group (log-rank p = 0.066). In multivariable Cox models, the VAR group had a significantly lower mortality risk compared with LL (HR = 0.465, p = 0.031), while the HH group showed a non-significant trend toward reduced mortality (HR = 0.637, p = 0.171). These patterns of higher exacerbation risk and trend toward higher mortality in the LL phenotype were consistent in the validation cohort. Persistently low BEC identifies a high-risk phenotype associated with increased exacerbations and trend toward poor prognosis, a finding corroborated in an independent cohort, supporting the value of repeated assessment.