UNLABELLED: Induced sputum is a useful noninvasive method for assessing parenchymal diseases. This retrospective study investigated its potential application in combination with functional parameters to differentiate sarcoidosis from non-sarcoid interstitial lung disease (NSA-ILD), especially when bronchoscopy is clinically contraindicated. All 120 study patients (67 sarcoidosis and 53 NSA-ILD) underwent both bronchoalveolar lavage (flexible fiberoptic video bronchoscope; Pentax, Japan) and induced sputum testing (3% NaCl, selecting plugs method, 300 cells differentially counted in Giemsa stained cytopreps). CD4/CD8 subsets were identified by a fluorescence-activator cell sorter. All patients underwent high-resolution computerized tomography and 103 of 120 underwent transbronchial biopsy. Multivariate logistic regression models were applied to the data to predict the probability of having the sarcoidosis as a function of the explanatory variables: Model I contained demographic and induced sputum data, and Model II included demographic data and combined sputum and pulmonary function test results. The area under the curve was 0.899 for induced sputum parameters alone and 0.914 for induced sputum and pulmonary function parameters. CONCLUSION: The results derived from the combination of noninvasive induced sputum approach can be used as predictors with high specificity and sensitivity in the differential diagnosis of sarcoidosis.
UNLABELLED: Induced sputum is a useful noninvasive method for assessing parenchymal diseases. This retrospective study investigated its potential application in combination with functional parameters to differentiate sarcoidosis from non-sarcoid interstitial lung disease (NSA-ILD), especially when bronchoscopy is clinically contraindicated. All 120 study patients (67 sarcoidosis and 53 NSA-ILD) underwent both bronchoalveolar lavage (flexible fiberoptic video bronchoscope; Pentax, Japan) and induced sputum testing (3% NaCl, selecting plugs method, 300 cells differentially counted in Giemsa stained cytopreps). CD4/CD8 subsets were identified by a fluorescence-activator cell sorter. All patients underwent high-resolution computerized tomography and 103 of 120 underwent transbronchial biopsy. Multivariate logistic regression models were applied to the data to predict the probability of having the sarcoidosis as a function of the explanatory variables: Model I contained demographic and induced sputum data, and Model II included demographic data and combined sputum and pulmonary function test results. The area under the curve was 0.899 for induced sputum parameters alone and 0.914 for induced sputum and pulmonary function parameters. CONCLUSION: The results derived from the combination of noninvasive induced sputum approach can be used as predictors with high specificity and sensitivity in the differential diagnosis of sarcoidosis.