Background: Asthma is a heterogeneous disease, and phenotyping can facilitate understanding of disease pathogenesis and direct appropriate asthma treatment. This nationwide cohort study aimed to phenotype asthma patients in Japan and identify potential biomarkers to classify the phenotypes.
Methods: Adult asthma patients (n = 1925) from 27 national hospitals in Japan were enrolled and divided into Global Initiative for Asthma (GINA) steps 4 or 5 (GINA 4, 5) and GINA Steps 1, 2, or 3 (GINA 1-3) for therapy. Clinical data and questionnaires were collected. Biomarker levels among GINA 4, 5 patients were measured. Ward's minimum variance hierarchical clustering method and tree analysis were performed for phenotyping. Analysis of variance, the Kruskal-Wallis, and chi-square tests were used to compare cluster differences.
Results: The following five clusters were identified: 1) late-onset, old, less-atopic; 2) late-onset, old, eosinophilic, low FEV1; 3) early-onset, long-duration, atopic, poorly controlled; 4) early-onset, young, female-dominant, atopic; and 5) female-dominant, T1/T2-mixed, most severe. Age of onset, disease duration, blood eosinophils and neutrophils, asthma control questionnaire Sum 6, number of controllers, FEV1, body mass index (BMI), and hypertension were the phenotype-classifying variables determined by tree analysis that assigned 79.5% to the appropriate cluster. Among the cytokines measured, IL-1RA, YKL40/CHI3L1, IP-10/CXCL10, RANTES/CCL5, and TIMP-1 were useful biomarkers for classifying GINA 4, 5 phenotypes.
Conclusions: Five distinct phenotypes were identified for moderate to severe asthma and may be classified using clinical and molecular variables (Registered in UMIN-CTR; UMIN000027776.)
View full abstract