Step-down of inhaled corticosteroids in non-eosinophilic asthma: A prospective trial in real life

Background: While non-eosinophilic asthmatics are usually considered poorly responsive to inhaled corticosteroids (ICSs), studies assessing a step-down of ICS in this specific population are currently lacking. Objectives: To assess the proportion of non-eosinophilic asthmatics in whom ICS may be withdrawn without any clinical degradation and to determine the predictive markers of a failure to stop treatment with ICS. Methods: This prospective study was completed by 36 non-eosinophilic asthmatics, defined by sputum eosinophils <3% and blood eosinophils <400/mu L. In these patients, whichever the baseline asthma control level, the dose of ICS was gradually reduced every 3 months until they met the failure criteria or successfully discontinued ICS for 6 months. The failure criteria were an ACQ score >= 1.5 with an increase from baseline >0.5 or a number of severe exacerbations during the study which was greater than the number during the year prior to the baseline visit. Receiver-operating characteristic (ROC) curves were constructed to assess predictors of a failure to stop ICS. This study is registered with ClinicalTrials.gov, number NCT02169323. Results: In 14 patients (39%), ICSs were completely withdrawn, and in 10 further patients (28%), ICS were stepped-down to a reduced ICS dose without any deterioration of asthma control and exacerbation rate. Baseline predictors of a failure to stop ICS were a greater age (area under ROC curve [ROC AUC] and [95% CI]: 0.77 [0.62-0.93]) and elevated blood eosinophils (ROC AUC [95% CI]: 0.77 [0.61-0.93]). After the first step-down of ICS, the best predictor was an elevated blood eosinophil count (ROC AUC [95% CI]: 0.85 [0.72-0.99]). Conclusions & Clinical Relevance: Withdrawing or reducing the dose of ICS is feasible in two-thirds of non-eosinophilic asthmatics irrespective of baseline asthma control. An elevated blood eosinophil count may predict the failure to stop ICS.