Clinical Subtypes of Neutrophilic Asthma: A Cluster Analysis From Australasian Severe Asthma Network

Background Clinical heterogeneity may exist within asthma subtypes defined by inflammatory markers. However, the heterogeneity of neutrophilic asthma (NA) remains largely unexplored. Objective To explore potential clusters and the stability of NA. Methods Participants with NA from the Australasian Severe Asthma Network underwent a multidimensional assessment. They were then asked to participate in a 12-month longitudinal cohort study. We explored potential clusters using a hierarchical cluster analysis and validated the differential future risk of asthma exacerbations in the identified clusters. A decision tree analysis was developed to predict cluster assignments. Finally, the stability of prespecified clusters was examined within 1 month. Results Three clusters were identified in 149 patients with NA. Cluster 1 (n = 99; 66.4%) was characterized by female-predominant nonsmokers with well-controlled NA, cluster 2 (n = 16; 10.7%) by individuals with comorbid anxiety/depressive symptoms with poorly controlled NA, and cluster 3 by older male smokers with late-onset NA. Cluster 2 had a greater proportion of participants with severe exacerbations (P = .005), hospitalization (P = .010), and unscheduled visits (P = .013) and a higher number of emergency room visits (P = .039) than that of the other two clusters. The decision tree assigned 92.6% of participants correctly. Most participants (87.5%; n = 7) in cluster 2 had a stable NA phenotype, whereas participants of clusters 1 and 3 had variable phenotypes. Conclusions We identified three clinical clusters of NA, in which cluster 2 represents an uncontrolled and stable NA subtype with an elevated risk of exacerbations. These findings have clinical implications for the management of NA. Clinical heterogeneity may exist within asthma subtypes defined by inflammatory markers. However, the heterogeneity of neutrophilic asthma (NA) remains largely unexplored. To explore potential clusters and the stability of NA. Participants with NA from the Australasian Severe Asthma Network underwent a multidimensional assessment. They were then asked to participate in a 12-month longitudinal cohort study. We explored potential clusters using a hierarchical cluster analysis and validated the differential future risk of asthma exacerbations in the identified clusters. A decision tree analysis was developed to predict cluster assignments. Finally, the stability of prespecified clusters was examined within 1 month. Three clusters were identified in 149 patients with NA. Cluster 1 (n = 99; 66.4%) was characterized by female-predominant nonsmokers with well-controlled NA, cluster 2 (n = 16; 10.7%) by individuals with comorbid anxiety/depressive symptoms with poorly controlled NA, and cluster 3 by older male smokers with late-onset NA. Cluster 2 had a greater proportion of participants with severe exacerbations (P = .005), hospitalization (P = .010), and unscheduled visits (P = .013) and a higher number of emergency room visits (P = .039) than that of the other two clusters. The decision tree assigned 92.6% of participants correctly. Most participants (87.5%; n = 7) in cluster 2 had a stable NA phenotype, whereas participants of clusters 1 and 3 had variable phenotypes. We identified three clinical clusters of NA, in which cluster 2 represents an uncontrolled and stable NA subtype with an elevated risk of exacerbations. These findings have clinical implications for the management of NA.