Baseline Blood Eosinophil Count and 28-Day in-Hospital Mortality in ICU-Admitted Patients with AECOPD: A Cohort Study with External Validation.

BACKGROUND: Blood eosinophil count is widely used to guide corticosteroid decisions in stable chronic obstructive pulmonary disease (COPD) and non-ICU acute exacerbations of COPD (AECOPD), but its prognostic value and interpretation in ICU-admitted AECOPD remain uncertain. METHODS: We conducted a retrospective cohort study using MIMIC-IV as the primary cohort and eICU-CRD for external validation. Adults with AECOPD were included if ICU admission occurred within 24 h of hospital admission and baseline eosinophil count was available within a predefined peri-ICU window. The primary outcome was 28-day in-hospital mortality after ICU admission. Baseline eosinophil count was analyzed per 100 cells/µL increase in prespecified multivariable Cox models. RESULTS: The analytic cohorts included 645 patients from MIMIC-IV and 2503 from eICU-CRD. Twenty-eight-day in-hospital mortality was 14.3% and 6.7%, respectively. In MIMIC-IV, higher baseline eosinophil count was associated with a lower hazard of death (adjusted hazard ratio [HR] per 100 cells/µL: 0.58, 95% confidence interval [CI]: 0.39-0.85; P = 0.005). A similar association was observed in eICU-CRD (adjusted HR: 0.82, 95% CI: 0.72-0.93; P = 0.002). Categorical, competing-risk, and age-restricted analyses were broadly consistent with the primary findings, and restricted cubic splines did not identify a clear threshold. CONCLUSION: Among ICU-admitted patients with AECOPD who had clinically available baseline eosinophil measurements, higher baseline eosinophil count was associated with lower 28-day in-hospital mortality.