The molecular phenotype of severe asthma in children

Although the clinical attributes of severe asthma in children have been well described, the differentiating features of the lower airway inflammatory response are less understood.

We sought to discriminate severe from moderate asthma in children by applying linear discriminant analysis, a supervised method of high-dimensional data reduction, to cytokines and chemokines measured in the bronchoalveolar lavage (BAL) fluid and alveolar macrophage (AM) lysate.

Bronchoalveolar lavage fluid was available from 53 children with asthma (severe asthma, n = 31) undergoing bronchoscopy for clinical indications and 30 nonsmoking adults. Twenty-three cytokines and chemokines were measured by using bead-based multiplex assays. Linear discriminant analyses of the BAL fluid and AM analytes were performed to develop predictive models of severe asthma in children.

Although univariate analysis of single analytes did not differentiate severe from moderate asthma in children, linear discriminant analyses allowed for near complete separation of the moderate and severe asthmatic groups. Significant correlations were also noted between several of the AM and BAL analytes measured. In the BAL fluid, IL-13 and IL-6 differentiated subjects with asthma from controls, whereas growth-related oncogene (CXCL1), RANTES (CCL5), IL-12, IFN-gamma, and IL-10 best characterized severe versus moderate asthma in children. In the AM lysate, IL-6 was the strongest discriminator of all the groups.

Severe asthma in children is characterized by a distinct airway molecular phenotype that does not have a clear T(H)1 or T(H)2 pattern. Improved classification of children with severe asthma may assist with the development of targeted therapeutics for this group of children who are difficult to treat.