Symptoms, risk of future exacerbations, and response to long-term macrolide treatment in bronchiectasis: an observational study

Background: Previous studies have suggested that daily symptoms are a marker of bronchiectasis disease activity and could therefore identify patients at increased risk of exacerbation. However, international bronchiectasis guidelines recommend long-term macrolide treatment only in patients with three or more exacerbations per year. We aimed to investigate if symptoms independently predict future exacerbations and therefore identify additional responders to long-term macrolide treatment. Methods: We used data from the EMBARC registry, a multicentre international bronchiectasis database. Baseline symptoms were evaluated with the quality-of-life bronchiectasis questionnaire respiratory symptoms score (QoL-B-RSS), followed-up for at least 1 year, and were related to the future risk of exacerbations. We subsequently conducted a post-hoc pooled analysis of three randomised controlled trials of macrolides (ie, BLESS, BAT, and EMBRACE) in 341 participants with bronchiectasis to determine if baseline symptoms were associated with response to long-term macrolide treatment, using a negative binomial regression model. Findings: 9466 patients from the 19 324 patients included in the EMBARC registry had available QoL-B-RSS assessment at baseline and 1-year follow-up. The median age was 68 years (IQR 58–74), 5763 (60·9%) were female, and 3703 (39·1%) were male. The median Bronchiectasis Severity Index score was 7 (4–10) and Pseudomonas aeruginosa was present in the sputum of 2041 (21·6%) patients within 12-months of baseline. Previous exacerbations (rate ratio (RR) for every additional exacerbation 1·11, 95% CI 1·10–1·12; p<0·0001) and symptoms (RR for every 10 points lower QoL-B-RSS 1·10, 1·09–1·11; p<0·0001) were identified as independent risk factors for future exacerbations. The number of exacerbations during 1-year of follow-up was similar between patients with three or more exacerbations at baseline and average symptom scores (QoL-B-RSS 60–70; RR 1·58, 95% CI 1·48–1·69) and the group with no previous exacerbations but high symptom scores (RR 1·55, 1·41–1·70). The same pattern was observed in the post-hoc analysis of randomised controlled trials, both in the macrolide and placebo groups. The number-needed-to-treat to prevent exacerbations with long-term macrolide therapy was similar for patients selected based on frequent exacerbations (1·45, 95% CI 1·08–2·24) and in those with few previous exacerbations, but high symptom scores 1·43 (1·06–2·18). Interpretation: Our results suggest that symptoms are an independent risk factor for future exacerbations in bronchiectasis. Patients who are highly symptomatic derive a similar benefit from macrolide treatment as patients with a high baseline exacerbation frequency. Funding: EU, European Federation of Pharmaceutical Industries and the Associations Innovative Medicines Initiative Inhaled Antibiotics in Bronchiectasis and Cystic Fibrosis Consortium, European Respiratory Society.