Monitoring the inflammatory heterogeneity in asthma with multiple biomarkers for multidimensional endotyping

Background: Clustering of asthma patients represent attempts to group within a continuum of disease rather than clear-cut entities. Aim: Evaluate an unbiased, multidimensional approach with multiple biomarkers. Methods: We used topological data analysis to characterize 100 adult asthmatic patients based on 21 clinical, pathophysiologic, and inflammatory parameters. In order to provide further dimensions to the clusters parameters were tested for association to particular clusters Results: Six clusters were visualized according to severity (GINA) and blood eosinophilia (Figure 1). For mild asthma, there are no specific biomarkers, besides mild eosinophilia. For moderate and severe asthma, there was a clear separation between type 2 and non-type 2 asthma. Moderate eosinophilic asthma differs by the intensity of systemic type 2 biomarkers (IL-13, IL-5, eosinophil-derived neurotoxin, periostin, blood eosinophilia) and this impacts on outcomes such as corticosteroid resistance and lung function decline. For severe asthma, we observed a non-eosinophilic cluster with increased serum eotaxin and IL-8, and a mild eosinophilic cluster characterized by low dipeptidyl-peptidase 4 inhibitor, increased exacerbation rate and late-onset asthma. Conclusion: The investigation of multiple clinical and biological biomarkers demonstrates new subendotypes and extensive subgrouping within type 2 and non-type 2 asthma