The 2019 Global initiative for chronic Obstructive Lung Disease (GOLD) strategy document recommends that pharmacological therapy for chronic obstructive pulmonary disease (COPD) should aim to reduce symptoms and the frequency and severity of exacerbations,1 which are key contributing factors to the high clinical and economic burden of the disease.2, 3, 4
Inhaled corticosteroids (ICS) reduce the rate of exacerbations (ie, the number of exacerbations per patient year) in patients with COPD and a history of exacerbations.5, 6, 7, 8 However, the heterogeneity of COPD leads to variability between individuals in the magnitude of ICS benefit, and the potential for drug-related adverse events such as pneumonia.9 Post-hoc and prespecified analyses of randomised controlled trials have shown that higher blood eosinophil counts are associated with greater ICS treatment effects.10, 11, 12, 13 A pooled analysis14 of clinical trials, which compared ICS and long-acting β2-agonist (LABA) combination therapy with LABA monotherapy, modelled this relationship with data from 4528 patients; beneficial effects of ICS on exacerbation reduction were apparent at blood eosinophil counts of more than 100 cells per μL, with greater effects observed at higher blood eosinophil counts.14 Dichotomising the eosinophil count into high and low on the basis of an arbitrary threshold10, 11 cannot fully describe the continuous nature of the relationship between blood eosinophils and ICS effect.14 The increasing effect of ICS at higher blood eosinophil counts has also been observed in other analyses of ICS/LABA versus LABA clinical trials involving multiple eosinophil subgroups.11, 12
Research in context
Evidence before this study
Therapies containing inhaled corticosteroids (ICS) reduce exacerbation rates (ie, number of exacerbations per patient per year) in patients with chronic obstructive pulmonary disease (COPD). Studies have shown that higher blood eosinophil counts are associated with a greater treatment benefit with ICS. We searched PubMed using the terms “inhaled corticosteroids”, “blood eosinophil count”, “exacerbation”, and “COPD” to identify studies published from Jan 1, 2008, onwards that investigated the association between blood eosinophil counts and response to ICS-containing therapies in COPD. The relationship between ICS and blood eosinophil count has been modelled in various post-hoc analyses, with the greatest effects of ICS on reduction of COPD exacerbations shown at higher eosinophil counts. The majority of these analyses focused on the response of patients to ICS and long-acting β2-agonist (LABA) therapy versus LABA monotherapy using dichotomised eosinophil counts. However, additional evidence is required to understand the relationship between blood eosinophil count as a continuous variable and the effect of ICS with both single-inhaler triple therapy and dual inhaled therapies in patients with COPD, and to determine whether this relationship is modified by other factors known to modulate the effects of ICS, such as cigarette smoking.
Added value of this study
The large size of the population in the IMPACT trial, which compared once-daily single-inhaler triple therapy (fluticasone furoate, umeclidinium, and vilanterol) with dual inhaled therapy (fluticasone furoate and vilanterol, and umeclidinium and vilanterol), allows modelling of the relationship between baseline blood eosinophil count as a continuous variable and the effect of ICS on COPD exacerbations and other outcomes. These analyses showed that greater ICS treatment effects were associated with increased blood eosinophil counts. This was seen in both former and current smokers, with larger benefit seen in former smokers.
Implications of all the available evidence
To our knowledge, this is the first analysis to use data modelling of single-inhaler triple therapy or ICS-LABA versus long-acting muscarinic antagonist–LABA to show that, in a population at high risk of exacerbations, the response to ICS-containing therapy can be predicted with blood eosinophil counts and smoking status. The continuous nature of the relationship between increasing blood eosinophil count and reduction in exacerbations with use of ICS is apparent. Assessment of blood eosinophils should, therefore, be used as part of a precision medicine approach to optimise the use of ICS within combination therapies in patients with COPD.
Recent clinical trials have shown the benefits of single-inhaler triple therapy versus the dual combination treatments ICS–LABA and long-acting muscarinic antagonists (LAMA)–LABA on exacerbation prevention.15, 16, 17, 18 Data modelling is needed to understand the effects of blood eosinophils on response to triple therapy versus LAMA–LABA, and ICS–LABA versus LAMA–LABA. Such data would support the use of blood eosinophil counts to aid decision-making regarding the use of these treatments in clinical practice. The InforMing the PAthway of COPD Treatment (IMPACT) trial18, 19 was designed to establish the relative benefits of the single-inhaler triple therapy containing fluticasone furoate–umeclidinium–vilanterol (ICS–LAMA–LABA) compared with both fluticasone furoate–vilanterol (ICS–LABA) and umeclidinium–vilanterol (LAMA–LABA) in patients with moderate-to-very-severe COPD and at risk of exacerbation. In IMPACT, 4877 (47%) of patients had at least two moderate exacerbations and 2671 (26%) had at least one severe exacerbation in the year before enrolment.18 The IMPACT results have been reported elsewhere,18 and show that single-inhaler triple therapy reduced moderate or severe COPD exacerbations, improved lung function, and improved quality of life compared with either dual combination therapy. There was also a reduction in the risk of on-treatment mortality with single-inhaler triple therapy versus umeclidinium–vilanterol.18 In a pre-specified analysis, a greater reduction in annual rate of moderate and severe exacerbations was reported with single-inhaler triple therapy versus umeclidinium–vilanterol in patients with at least 150 eosinophils per μL, compared with those with less than 150 eosinophils per μL.
Studies comparing single-inhaler triple therapy with LAMA–LABA, and ICS–LABA with LAMA–LABA, have used different bronchodilator molecules within the dual bronchodilator compared with the ICS-containing combination.17, 20 Furthermore, different inhaler devices and regimens (once vs twice a day) have been compared, which can introduce confounding. A strength of the IMPACT trial is the use of the same bronchodilator molecules, inhaler devices, doses, and regimen for all treatments, meaning that any differences observed cannot be attributed to these variables.18
We report an analysis of the IMPACT trial that modelled the relationship of blood eosinophil counts with ICS therapy across multiple clinical outcomes including exacerbations, lung function, and health status. It has previously been reported that current smoking modifies the relationship between blood eosinophil counts and ICS effect, with a greater ICS effect observed in former smokers.14, 21, 22 We therefore also investigated the effect of current or former smoking on this relationship in post-hoc analyses.