Eosinophil cationic protein concentration in saliva does not correlate with eosinophil cationic protein concentration in sputum.

compared to L-IS (n 18, p 0.02) but no difference in total cell count; all other cell types, including epithelial and squamous cells, were not significantly different between the two collections. E-IS and BAL eosinophils were significantly correlated (n 20, 0.72, p 0.0004) in all groups and in asthma alone (n 10, 0.82, p 0.004); L-IS showed weaker correlations for percentage of eosinophils with BAL ( 0.51, p 0.05 in all groups; 0.81, p 0.049 in asthma). There were no significant correlations between either E-IS or L-IS with BAL for other cell types, and no significant correlations of E-IS or L-IS inflammatory cells with any submucosal cell types (lymphocytes-CD3, macrophages-CD68, neutrophil, or activated eosinophils-EG2 ). E-IS and to a lesser extent L-IS were highly correlated with BAL for eosinophils in all groups and in asthma, but did not correlate with endobronchial biopsy submucosal inflammatory cell counts. This may be attributed to the fact that sputum and BAL sample the airway lumen, whereas the cellular inflammation in the airway submucosa may be distinct. IS could be used to monitor luminal eosinophilic inflammation but probably not mucosal inflammation in asthma.