In ‘real world’ patients with COPD, exacerbation history, and not blood eosinophils, is the most reliable predictor of future exacerbations

INTRODUCTION: There is an interest in the role of blood eosinophils for predicting inhaled corticosteroid (ICS) response in chronic obstructive pulmonary disease (COPD). Most data are from interventional clinical studies; data from unselected real-world populations may help better inform treatment decisions. DACCORD is a non-interventional real-world study. Cohort 3 recruited patients with COPD who had received triple therapy for ≥ 6 months; prior to entry patients either continued triple therapy, or switched to a long-acting muscarinic antagonist/long-acting beta(2)-agonist (LABA/LAMA), and were followed for 12 months. METHODS: For these post-hoc analyses, patients were divided into four groups based on exacerbation history and baseline blood eosinophil count (< 100 vs. > 300 cells/µL). Exacerbation rates were calculated overall and for the two treatments. RESULTS: Among the 430 patients in the current analyses, the largest groups had low exacerbation history with high (44.2%) or low eosinophils (36.7%). Most patients did not exacerbate during follow-up (68.8% overall; 83.2% and 63.7% with LABA/LAMA and triple therapy). The highest exacerbation rates were in groups with high exacerbation history, differing significantly in the overall analyses from those with low exacerbation history (matched by eosinophil count); rates did not differ when grouped by eosinophil count (matched by exacerbation history). CONCLUSIONS: Although most patients in these analyses did not exacerbate during follow-up, whereas exacerbation history is a predictor of future exacerbations, blood eosinophil count is not. This suggests that although eosinophil count may help to guide ICS initiation, this is less of a consideration when ‘stepping-down’ from triple therapy to a LABA/LAMA