Research in context
Evidence before this study
High blood eosinophil counts could predict response to inhaled corticosteroids (ICS) in patients with chronic obstructive pulmonary disease (COPD) and a history of exacerbations. However, review of the published scientific literature for clinical reports of blood eosinophil counts as a predictive biomarker in patients with COPD indicates that there is uncertainty about where a clinically relevant cutoff lies.
We searched PubMed for publications in English with the terms “chronic obstructive pulmonary disease”, “eosinophils”, and “exacerbations” over the previous 3 years (Jan 1, 2013, to Feb 29, 2016). Additionally, we supplemented this information with data from congress abstracts over this time frame. Previous clinical studies have shown that patients with a blood eosinophil count of at least 2% were less likely to experience exacerbations when ICS were given as compared with therapy with a long-acting β2-agonist (LABA). However, these studies did not consider whether the same threshold applies when two bronchodilator drugs are used, what an optimum threshold to identify responders might be, and whether this applies when ICS are withdrawn.
Based on these data, we hypothesised that patients with higher eosinophil counts would be more likely to relapse if ICS treatment was withdrawn while patients remained on a long-acting muscarinic antagonist (LAMA) plus LABA. We tested this idea using data from the WISDOM study that compared the risk of exacerbation in patients who continued treatment with LAMA/LABA/ICS with those who received LAMA/LABA alone, and stratified patients by their eosinophil count.
Added value of this study
Our study is the first to investigate and confirm that an association exists between the blood eosinophil count at study onset, ICS withdrawal, and exacerbations of COPD in the presence of a LAMA plus a LABA. For the first time, in addition to an eosinophil cut-off with a 2% threshold, we examined analysis cutoffs throughout a range of blood eosinophil distribution. Our data show that whereas the effects of ICS can be observed at lower eosinophil cutoffs (≥2%), this is driven by the response of patients with much higher blood eosinophil counts. A statistically significant difference between subgroups defined by distinct eosinophil cutoffs was seen at higher blood eosinophil counts only (≥300 cells per μL or ≥4%). We also found that there was no difference in exacerbation risk between treatments among the majority of patients with eosinophil counts of less than 4% or less than 300 cells per μL.
Implications of all the available evidence
Although we identified a subgroup of patients with high eosinophil counts at study onset in whom continuing ICS use was associated with fewer exacerbations, most patients were no more likely to exacerbate while taking two bronchodilators alone. Thus, most patients with stable severe COPD can be satisfactorily managed with appropriate inhaled bronchodilator therapy.