No abstract

Objective: To compare the validity of different methods for the assessment of bronchial hyperresponsiveness used by different centers. Methods: Case series of 648 subjects referred to six pulmonary centers, all with a history of shortness of breath without airway obstruction, without use of medication that might influence the tests and without viral infections during the previous two weeks. All subjects answered a questionnaire of recent symptoms and underwent bronchial challenge with a chemical stimulus according to each center's protocol. Analysis was performed by receiver operating characteristic (ROC) plots using the questionnaire's answers as the gold standard. Diagnostic test sensitivities at the cut-offs for bronchial hyperresponsiveness indicated by each center were compared. Results: ROC plots showed poor validity of all tests, i.e. both acceptable sensitivity and specificity were not observed with any test. There was no obvious difference of the slope of the ROC plots between the different centers. However, maximal sensitivity differed considerably: for ªwheeze during the previous 12 monthsº, sensitivity at each center's cut-off for the definition of bronchial hyperresponsiveness varied between 0.35 and 0.73. The choice of the question used as the standard had little influence on test validity. Conclusion: Although some of the differences between centers may be explained by subject characteristics, the large differences of the test sensitivities are unacceptable and underscore the need for standardization of these tests, primarily with respect to sufficient sensitivity.