LAUSR

RATIONALE Treatment of chronic obstructive pulmonary disease (COPD) with inhaled corticosteroids (ICS) is controversial because it can reduce the risk of future exacerbations of the disease at the expense of increasing the risk of pneumonia. OBJECTIVE To assess the relationship between the presence of chronic bacterial infection (CBI), reduced number of circulating eosinophils, ICS treatment and the risk of pneumonia in COPD patients. METHODS AND MEASUREMENTS Post-hoc long-term observational study of an historical cohort of 201 COPD patients (GOLD II-IV) who were carefully characterized (including airway microbiology) and followed for a median of 84 months. Results were analysed by multivariate Cox regression and network analysis. MAIN RESULTS Mean age was 70.3 years, 90.5% of patients were male, mean FEV1 was 49%, 71.6% of patients were treated with ICS, 57.2% of them had bronchiectasis and 20.9% <100 blood eosinophils/µL. Pathogenic microorganisms were isolated in 42.3% of patients (22.4% of patients fulfilled the definition of chronic bronchial infection (CBI). During follow-up, 38.8% of patients suffered ≥1 pneumonia, CBI (HR, 1.635) and <100 eosinophils/µL (HR, 1.975) being independently associated with the risk of pneumonia, particularly when both coexist (HR, 3.126). ICS treatment increased the risk of pneumonia in those patients with <100 eosinophils/µL and CBI (HR, 2.925). CONCLUSIONS Less than 100 circulating eosinophils/L combined with the presence of CBI increase the risk of pneumonia in COPD patients treated with ICS.