Heterogeneity of Obese Asthma: A Hierarchical Cluster Analysis From the Australasian Severe Asthma Network.

Abstract

Clinical heterogeneity exists within obese asthma (OA) traditionally defined by body mass index-based measurements of obesity, which can either underestimate or overestimate adiposity. To explore potential OA subtypes and verify future outcomes of the identified clusters. We applied the new definition of obesity introduced by the European Association for the Study of Obesity to identify participants with OA from the Australasian Severe Asthma Network and observed them in a 12-month prospective cohort study. We explored potential OA clusters using a hierarchical cluster analysis and assessed clinical outcomes including asthma exacerbations (AEs) and clinical remission (CR) in the identified clusters. Additionally, we performed a decision tree analysis to validate cluster assignments in another separate dataset. Cluster analysis of 244 subjects with OA identified three clusters. Cluster 1 was younger mild OA with high skeletal muscle mass; cluster 2 was male smokers predominant uncontrolled OA with multimorbidity and eosinophilic inflammation; and cluster 3 was older female nonsmokers predominant OA with high visceral fat area and neutrophilic inflammation. Clusters 2 and 3 had higher risk of moderate to severe AEs (odds ratio [OR] = 5.60, 95% CI, 2.42-12.94; and OR = 3.86, 95% CI, 1.74-8.58) and were less likely to achieve two-component CR (OR = 0.18, 95% CI, 0.08-0.41; and OR = 0.26, 95% CI, 0.12-0.57). Decision tree analysis validated the clustering results in another dataset with an accuracy of 88.93%. We identified three OA clusters with distinct clinical characteristics and differential future outcomes including AE and CR. These findings may provide clinical implications for the targeted management of OA.