A prospective two-centre cohort study of 2,444 hospitalized asthma-COPD overlap (ACO) exacerbation patients used unsupervised k-means clustering to identify inflammatory phenotypes and assess their association with in-hospital adverse outcomes (ICU admission, invasive ventilation, in-hospital mortality).
Three phenotypes emerged: neutrophil-dominant/hypoeosinophilic (T1), eosinophilic (T2), and eosinophil-neutrophil balanced (T3). T1 had the highest composite adverse outcome rate (24.2%; p < 0.001). Key predictors were absolute neutrophil count (β = 0.38), D-dimer (β = 0.25), and absolute lymphocyte count (β = −0.41); a nomogram achieved AUC 0.775 (training) and 0.766 (validation).
Two-centre design may limit generalizability; inflammatory markers were collected at a single acute timepoint without longitudinal tracking; cluster stability across broader populations is unconfirmed.
On admission for an ACO exacerbation, check a simple CBC with differential, CRP, and D-dimer — a neutrophil-dominant, hypoeosinophilic pattern flags the highest-risk patients who may need early escalation of care. Eosinophilic phenotype patients appear to carry lower in-hospital risk and may warrant a distinct management approach.
Explore related topics
