Predictive value of blood eosinophils and exhaled nitric oxide in adults with mild asthma: a prespecified subgroup analysis of an open-label, parallel-group, randomised controlled trial

Background: Whether blood eosinophil counts and exhaled nitric oxide (FeNO) are associated with important outcomes in mild asthma is unclear. In this prespecified subgroup analysis of a previously published open-label clinical trial, we aimed to assess associations between blood eosinophil counts and FeNO with outcomes and response to asthma treatment. Methods: In the previously reported 52-week, open-label, randomised controlled trial, people with mild asthma receiving only β agonist reliever inhalers were enrolled at one of 16 clinical trials units in New Zealand, the UK, Italy, or Australia. Eligible participants were randomly assigned (1:1:1, stratified by country), to receive inhalers to take as-needed salbutamol (two inhalations of 100 μg in a pressurised metered dose inhaler), maintenance budesonide (200 μg twice per day by inhaler) plus as-needed salbutamol (two inhalations of 100 μg), or as-needed budesonide–formoterol (one inhalation of 200 μg budesonide and 6μg formoterol by inhaler). The primary outcome was the annual rates of asthma exacerbations per patient, and in this prespecified subgroup analysis, we assessed whether annual exacerbation rates in each treatment group were significantly different depending on levels of blood eosinophil count, FeNO, or a composite score of both. Analyses were done for patients with available biomarker measurements The study was registered with the Australian New Zealand Clinical Trials Registry, number ACTRN12615000999538. Findings: 675 participants were enrolled between March 17, 2016, and Aug 29, 2017, of whom 656 had results for blood eosinophil analysis and 668 had results for FeNO. Of the patients who received as-needed salbutamol, the proportion of patients having a severe exacerbation increased progressively with increasing blood eosinophil count (two [4%] of 49 participants with <0·15 × 109/L, six [6%] of 93 with 0·15 to <0·3 × 109/L, and 15 [19%] of 77 with ≥0·3 × 109/L; p=0·014). There were no significant interactions between blood eosinophil count ...