Microbiome-driven clusters in severe asthma derived from induced sputum: identification and stability over time

Introduction: Asthmatics have been reported to have airway microbial dysbiosis [Chung JACI 2017]. We aimed to detect severe asthma phenotypes using induced sputum microbiome profiles and to assess cluster-wise stability after 12-18 months of prospective follow-up. Methods: Induced sputum samples were collected from a subset of the U-BIOPRED adult severe asthmatics. 16s rRNA sequencing and metagenomics were used to characterize the sputum microbiome. Unsupervised hierarchical clustering was performed by Bray-Curtis β-diversity measure of 16s microbiome data. The clustering was validated using partition around medoids, topological data analysis, consensus cluster distribution and bootstrapping. Sputum samples collected after 12-18 months were used to assess patient cluster migration. Results: Analysis of sputum samples of 100 severe asthmatics (median age: 55 years, 42% male) revealed two microbiome-driven clusters. The clusters were significantly different in: age of asthma onset, residential location, smoking status, percentage of sputum neutrophils, and spirometry (p-values Conclusion: Induced sputum microbiome identified two distinct severe asthma phenotypes. The relative stability of microbiome-driven phenotypic clusters indicates its suitability for diagnostic and therapeutic precision medicine approaches.