Inhaled Corticosteroid Escalation After a Single Moderate Exacerbation in COPD : Differential Benefit by Exacerbation Subtype

BACKGROUND AND OBJECTIVE: In chronic obstructive pulmonary disease (COPD), evidence supporting escalation to inhaled corticosteroid (ICS)-containing therapy after single moderate exacerbation in patients receiving dual bronchodilator therapy remains inconsistent. We aimed to evaluate whether the clinical benefit of ICS escalation differs according to the treatment phenotype of moderate exacerbations. METHODS: Using the Korean National Health Insurance Service database, we identified ICS-naïve patients with COPD initiating long-acting muscarinic antagonist (LAMA)/long-acting β2-agonist (LABA) therapy. Sequential target trial emulations were conducted among patients who experienced one moderate exacerbation while on LAMA/LABA. Exacerbations were classified as antibiotic-treated (without systemic corticosteroids) or systemic corticosteroid-treated (with or without antibiotics). Outcomes were compared between ICS escalation and LAMA/LABA continuation groups using propensity score matching. RESULTS: Among 94,938 patients who initiated LAMA/LABA therapy, treatment persistence remained high, with over 60% maintaining dual therapy at 6 years, despite a marked decline in exacerbation-free survival. In matched analyses, ICS escalation after single antibiotic-treated exacerbations was not associated with reduced subsequent exacerbation risk. In contrast, among patients with steroid-treated moderate exacerbations, continuation of LAMA/LABA was associated with a higher risk of exacerbations (HR 1.18, 95% CI 1.12-1.23), compared with escalation to ICS-containing therapy. Findings were consistent across subgroups defined by exacerbation history. CONCLUSION: In this large real-world COPD cohort, the benefit of ICS escalation after single moderate exacerbation varied by exacerbation treatment phenotype. Steroid-treated events were associated with a reduced subsequent exacerbation with ICS escalation, whereas antibiotic-treated events did not confer additional benefit from ICS, supporting a more precise approach to ICS escalation beyond exacerbation frequency alone.