Assessment of the role of small airway dysfunction in relation to exacerbation risk in patients with well controlled asthma (ATLANTIS): an observational study

Background: Recent surveys suggest that asthma remains inadequately controlled in more than 50% of adults with asthma despite guideline-based standard therapy. Small airways are often under-recognised as major sites of airway obstruction and inflammation. This might be related to lack of assessment with current tools such as impulse oscillometry, and thus under-treatment might explain inadequate control. Small airway dysfunction, which is common in adults with well controlled asthma, might represent an important biomarker of future risk of exacerbations. We aimed to investigate whether small airway dysfunction is present in patients with well controlled asthma and, if so, whether it is a risk factor for future exacerbations in this population. Methods: The observational Assessment of Small Airways Involvement in Asthma (ATLANTIS) study included 773 extensively characterised patients with asthma aged 18-65 years from 29 primary and specialty clinics in nine countries from June 30, 2014, to March 3, 2017. Patients were required to be diagnosed with asthma at least 6 months before inclusion based on evidence of airway hyper-responsiveness, bronchodilator reversibility, or peak expiratory flow variability. Patients were required to have stable asthma, defined as no asthma exacerbations and regular asthma treatment at a consistent dose for 8 weeks before baseline visits. The current analysis included patients with well controlled asthma, defined as an Asthma Control Questionnaire (ACQ-6) score of less than 0·75 at baseline. Small airway dysfunction was defined, based on deviation from predicted values of impulse oscillometry parameters, as a Z score of more than 1·645 for R5-R20 (resistance at 5 Hz minus resistance at 20 Hz) and AX (area of reactance) and a Z score of less than -1·645 for X5 (reactance at 5 Hz), with additional analyses exploring severe small airway dysfunction (Z score of 3 or -3). ATLANTIS is registered with ClinicalTrials.gov, NCT02123667. Findings: Of 773 patients, ACQ-6 assessments were available for 772 patients. Among these patients, 384 (50%) were classified as having well controlled asthma, and small airway dysfunction was present in 108 (36% [95% CI 30-41]) of 304 patients with impulse oscillometry data available for R5-R20, 89 (34% [28-42]) of 261 patients with data for AX, and 79 (26% [21-31]) of 303 patients with data for X5. In the multivariable analysis, we found that R5-R20-defined small airway dysfunction was associated with increased risk of future exacerbations, independent of age, sex, smoking status, Global Initiative for Asthma steps 4-5, previous exacerbations, percentage of predicted FEV1, ratio of residual volume to total lung capacity, and peripheral blood eosinophil count (hazard ratio [HR] 2·26 [95% CI 1·05-4·85]; p=0·038), whereas AX (2·07 [0·91-4·70]; p=0·082) and X5 (0·86 [0·33-2·21]; p=0·75) were not associated with exacerbations. For severe small airway disease, both R5-R20 (HR 2·80 [95% CI 1·26-6·26]; p=0·012) and AX (2·51 [1·04-6·04]; p=0·041) became independent predictors of future exacerbations, whereas X5 remained non-significant (0·99 [0·29-3·32]; p=0·98). Interpretations: We have addressed an undervalued trait by showing that small airway dysfunction is a common, sensitive, early independent risk biomarker for future exacerbations in adults with well controlled asthma.