Abstract Rationale Severe asthma is a significant challenge in pediatrics. Convenient methods for phenotyping asthma using peripheral markers correlate poorly to immunologic signatures in the airspaces. Analysis of immune cells in bronchoalveolar lavage (BAL) specimens can target treatments; eosinophilia is associated with the T2HIGH phenotype, whereas neutrophilia is associated with Th1 and Th17 phenotypes. Little is known about the most prevalent granulocyte pattern in children with asthma, paucigranulocytic; specifically, whether it is driven by macrophages, lymphocytes, or intrinsic smooth muscle dysfunction. Previous reports suggest that the paucigranulocytic landscape resulted from corticosteroid- induced eosinophil apoptosis. This study aims to examine clinical factors, pathogen effects, and lung immune cell landscapes as they inform this phenotype. Methods 728 children with severe wheeze/asthma underwent diagnostic bronchoscopy and BAL for pathogen analysis and differential cell counts to investigate treatment failure despite high-dose corticosteroid treatment. 54% (n = 391) had BAL granulocytosis (eosinophils ≥1% and/or neutrophils ≥ 6%) and 46% (n = 337) were paucigranulocytic (eosinophils = 0, neutrophils 6%). Paucigranulocytic BAL was the response outcome in a parsimonious multivariable logistic model that included non-colinear factors with robust odds informing the outcome identified by univariate screening. Results Paucigranulocytic is an established BAL phenotype by preschool (43%) with stable prevalence (50-55%) from school-age to adolescence. The logistic model included six factors with good fit based on a deviance/degree of freedom = 1.03, classification accuracy of 75%, and area under the sensitivity/1-specificity curve = 0.83. The most important clinical factors with odds ratios 1 are absent pathogens (adjusted odds ratio 2.98, p 0.001), and BAL macrophage percentage (1.04, p 0.001). Factors with lower odds are high-dose corticosteroid treatment (0.94, p = 0.72), BAL lymphocyte percentage (0.92, p 0.001), sensitization to ≥ 4 allergens (0.64, p = 0.04), and blood eosinophilia (0.41, p = 0.006). In a second model limited to children with BAL pathogens, the paucigranulocytic signature fell from 46% to 28% prevalence, however, BAL macrophages endured as a significant factor favoring (odds ratio 1.04, p 0.001) paucigranulocytic BAL. Conclusions Unexpectedly, in contrast to past views, treatment with high-dose corticosteroids did not raise but significantly lowered the odds of paucigranulocytic BAL in children with severe wheeze/asthma. These results instead support a distinct phenotype which is notable for absent pathogens, lower blood markers of T2 inflammation, and a robust lung macrophage signature. The phenotype is less prevalent but can be found in children with respiratory infection. Further studies into how pathogens alter immune landscapes and lung macrophage lineages are needed. This abstract is funded by: None
Baldwin et al. (Fri,) studied this question.