Asthma Phenotype Progression in Children and Adolescents: A Single Center Experience

Authors:
Open Access
Publisher website
Google Scholar
Abstract
Background: Asthma phenotype has implications for presentation and treatment selection. Methods: This was a retrospective study of patients attending a pediatric pulmonary clinic (2009–2015) with EHR data available for phenotyping (allergic, eosinophilic/non‐allergic or non‐allergic/non‐eosinophilic). Patients were classified by age. A multinomial regression model examined factors associated with phenotype. A logistic regression model examined factors associated with a change in phenotype (present/absent). Results: There were 2042 patients included. Approximately 44% of participants were allergic, 17% eosinophilic/non‐allergic and 39% non‐allergic/non‐eosinophilic. Among children 0–4 years, age of asthma onset (aOR 1.42, 95% CI 1.21, 1.65) and mean exacerbations (aOR 1.26 95% CI 1.09, 1.44) were associated with higher odds of an allergic phenotype (vs. non‐allergic/non‐eosinophilic). Among those aged 5–11 and 12–21, male sex and mild obstruction were positively associated, and age was negatively associated, with an allergic phenotype. Moderate/severe obstruction on spirometry (aOR 6.67, 95% CI 2.00, 22.29) and smoke exposure (aOR 1.77, 95% CI 1.17, 2.67) were also positively associated with an allergic phenotype in patients 12–21 years. Factors positively associated with a change in phenotype included index phenotype (eosinophilic/non‐allergic: aOR 5.31, 95% CI 3.46, 8.14) and mean exacerbations (aOR 1.34, 95% CI 1.16, 1.55) ( n = 1379). The allergic phenotype was most stable (> 95% predicted probability of remaining allergic); an eosinophilic/non‐allergic phenotype had the greatest predicted probability of change (48%). Conclusions: Factors associated with asthma phenotype vary by age. The allergic phenotype was the most stable, while an eosinophilic/non‐allergic phenotype was most prone to change.
Funding Information
  • College of Medicine, University of Cincinnati

This publication has 42 references indexed in Scilit: