Published Online: May 23, 2013
Sputum induction is a minimally invasive means of assessing inflammation in the lower airways. Identification of sputum eosinophilic or neutrophilic phenotypes has proven successful in predicting asthma exacerbations and response to therapy. However, sputum induction and processing requires technical experience and skills that are not available in most clinics. This aspect has driven the search for and use of more readily accessible biomarkers as surrogates for predicting airway inflammation, for management of more severe or difficult-to-treat asthma, and as response endpoints in clinical trials. Nevertheless, the use of surrogate biomarkers requires clear recognition and understanding of any limitations.
In an original article in The Journal of Allergy & Clinical Immunology (JACI), Hastie and colleagues have analyzed the accuracy and potential predictive capability of several biomarkers which have associations with eosinophilic and neutrophilic sputum inflammation. Data were derived from a comprehensively characterized group of 328 subjects with a broad spectrum of asthma severity, enrolled in the NHLBI Severe Asthma Research Program at Wake Forest School of Medicine. Various biomarkers were examined individually and in combination in more complex models to evaluate the potential usefulness of these surrogates to accurately predict sputum eosinophilia and neutrophilia.
The authors’ hypotheses tested whether the variables FENO, serum IgE, or blood eosinophils accurately predict sputum eosinophil percentages, and whether age, reduced FEV1 levels, and blood neutrophils accurately predict sputum neutrophil percentages. Although there were significant positive associations for sputum eosinophil percentages with blood eosinophil counts, FENO levels, and IgE levels, maximum sensitivity and specificity criteria misclassified substantial numbers of subjects with both higher and lower sputum eosinophil counts. Similarly, despite significant associations for sputum neutrophils with age and FEV1 levels, maximum sensitivity and specificity criteria misclassified substantial numbers of subjects with both higher and lower neutrophil percentages. More complex models of combined variables did not improve prediction and correctly assigned only 69% of subjects for eosinophilic phenotype, 64% of subjects for neutrophilic phenotype, and 41% of subjects for both phenotypes.
These observations have important implications for studies relying on surrogate biomarkers instead of directly evaluating airway inflammation, and response to therapeutic intervention, by sputum induction and analysis. One third of subjects potentially misclassified could have major effects on the ability of a clinical trial to demonstrate significant differences between groups, as well as resulting in incorrect inferences for use in the appropriately responsive patient with asthma. Further investigation of novel biomarkers may provide more readily available measures of airway inflammation, but for now, there are no good alternatives for the analyses of sputum cell counts.
The Journal of Allergy and Clinical Immunology (JACI) is an official scientific journal of the AAAAI, and is the most-cited journal in the field of allergy and clinical immunology.