Diagnostic value of induced sputum in interstitial lung disease

Abstract

Background: Induced sputum (IS) has been recently used to investigate pulmonary inflammation in patients with interstitial lung disease (ILD), but still little attention has been paid to its efficacy in diagnosing sarcoidosis and other ILD. The aim of this study was to evaluate the diagnostic value of IS differential cell count and CD4+/CD8+ ratio in sarcoidosis (SA) and nonsarcoidosis ILD (NSA ILD). Material and methods: We studied prospectively newly diagnosed 59 patients: 36 SA and 23 NSA ILD [16 hypersensitivity pneumonitis (HP) and 7 idiopathic pulmonary fibrosis (IPF)]. IS was performed by inhaling a 5% NaCl solution for 4 periods of 5 minutes. Giemsa stained cytopreps were differentially counted and T-lymphocyte subsets were analyzed by flow cytometry method. The k-nearest neighbour rule (k-NN) or predictive value of CD4+/CD8+ ratio were used to discriminate between SA and NSA ILD. The variables of IS used in k-NN rule were: cells viability, total cell count, percentages of alveolar macrophages, lymphocytes, neutrophils, eosinophils, CD4+ and CD8+ subsets, and CD4+/CD8+ ratio. Results: 33 patients were able to produce an adequate sputum sample (SA - 15, HP - 11, IPF - 7). A CD4+/CD8+ ratio > 2,6 had a sensitivity of 100% and a specificity of 72% with 84% of correctly classified cases in distinguishing SA from NSA ILD. However, using k-NN rule the probability of correct classification was 79% (classification error rate - 21%). Conclusion: To distinguish SA from NSA ILD cut off CD4+/CD8+ ratio > 2,6 alone was superior to k-NN rule using all the parameters of IS. Copyright © 2011 Via Medica.