Get access
Clinical & Experimental Allergy

Induced sputum to assess airway inflammation: a study of reproducibility


  • This study was partially supported by the Fund for Current Research, Ministry of Health, Italy, 1994 and Royal Postgraduate Medical School, Hammersmith Hospital, London, UK.

A. Spanevello, Fondazione Salvatore Maugeri, Via Roncaccio 16, 21049 Tradate, Italy.


Background Infiltration of the airways mucosa with activated inflammatory cells appears to be a major factor in the pathogenesis of asthma and other airway diseases. Examination of sputum provides a direct method to investigate airway inflammation non-invasively.

Objectives The aim of the present study was to evaluate the reproducibility of cell counts on cytospins and fluid phase (eosinophil cationic protein, ECP) measurements in a selected portion of induced sputum. We aimed to confirm the validity of the tecnique by comparing measurements between stable asthmatics, allergic rhinithis and healthy subjects.

Methods Sputum was induced with hypertonic saline (4.5%) twice within one week in 53 stable asthmatics, 16 subjects with seasonal rhinitis (out of the pollen season), and 19 healthy subjects. Reproducibility was examined within sample (two different plugs of the same sample) between sample (two specimens of induced sputum obtained within one week) and between examiners on stable subjects taking into account sample size, number of examinations per patients and Confidence Interval (CI) of the estimates.

Results We have found that the method is highly reproducible within sample and between examiners for all types of cells and fluid phase measurements of ECP. It is reproducible between sample for eosinophils, macrophages, neutrophils and ECP, but not for lymphocytes and weakly for epithelial cells. Sputum from asthmatics, in comparison with the sputum of healthy subjects and subjects with rhinitis had higher eosinophils (asthmatics: 12.2%± 12.9, rhinitis: 0.4 ± 0.8, normals: 0.4 ± 0.7(%) and ECP (asthmatics: 827 ± 491 μg/L, rhinitis: 127 ± 82 normals: 157 ± 203). No significant differences were found between healthy subjects and subjects with rhinitis. Eosinophil counts were inversely correlated with FEV1 (r=−0.37) expressed as percentage of predicted, but not significantly correlated with PC20 methacholine (r =−0.28) or blood eosinophils (r = 0.26).

Conclusions The importance of this study is the confirmation, within important statistical guidelines for a study of reproducibility, that the methods examined are reproducible and valid.

Get access to the full text of this article