Abstract
BACKGROUND: Acute exacerbations profoundly influence asthma prognosis, yet large-scale data linking baseline blood eosinophil counts to both short-term and long-term exacerbation risk remain limited. We therefore examined this association in one of the largest nationwide cohorts. METHODS: Using Taiwan's asthma pay-for-performance database linked to nationwide claims, we conducted a retrospective cohort study of adults with asthma (2015-2020). Baseline peak eosinophil counts defined two groups: high (≥300 cells/µL) and low (<300 cells/µL). Propensity-score matching (1:1) yielded two balanced subcohorts. Multivariable Cox models estimated adjusted HRs (aHRs) for moderate (outpatient, steroid-treated) and severe (Emergency Department visit or hospitalisation) exacerbations within 1 year and over the entire follow-up, controlling for age, sex, Charlson comorbidity score and propensity score. RESULTS: Among 407 725 and 961 268 asthma patients in the high and low eosinophil groups, respectively, matched subgroups of 50 360 patients each were analysed. The high eosinophil group had higher rates of severe acute exacerbations (SAEs) within 1 year (14.67 vs 10.950 per 100 person-years) with an adjusted HR of 1.392. Similar trends were observed for moderate acute exacerbations (AEs) (HR 1.548) and all AEs (HR 1.373) within 1 year. These associations persisted after adjustment for age, gender, Charlson comorbidity score and propensity score. CONCLUSIONS: In this first nationwide, propensity-matched cohort exceeding 100 000 adults, elevated blood eosinophil counts independently predicted both short-term and long-term moderate-to-severe asthma exacerbations across all age and sex strata. Peak eosinophil measurement thus offers a practical biomarker for early identification of high-risk patients who may benefit from intensified anti-eosinophilic or inhaled corticosteroid/long-acting beta agonist-based strategies.