Evaluation of the Diagnostic Accuracy of Non-Specific Bronchial Provocation Tests in the Diagnosis of Asthma: A Randomized Cross-Over Study.

ABSTRACT

INTRODUCTION: The role of bronchial provocation tests in the diagnosis of asthma remains to be fully explored. We aimed to evaluate methacholine and mannitol challenge testing, and explore the factors associated with this broncoprovocation response. METHODS: Observational, cross-over, randomized trial evaluating adult cases with suspected asthma, naïve to treatment, with normal pre-bronchodilator spirometry, and negative bronchodilator test. Patients were randomized to start with methacholine or mannitol. The diagnosis of bronchial asthma was confirmed if there was a good functional and clinical response to one month with twice daily formoterol/budesonide 9/320. The diagnostic profile and the concordance were calculated. Factors associated with a positive provocation test were entered into a multivariate binomial logistic regression analysis, and classification trees were created for both tests. RESULTS: The study included 108 cases (50.0% diagnosed with asthma and 51.9% cases starting with methacholine). The percentage of cases positive to methacholine and mannitol were 30.6% and 25.0% respectively. Kappa values were 0.40 (p<0.001). The diagnostic profile for methacholine was sensitivity 59.3% and specificity 98.1%, while for mannitol it was sensitivity 48.1% and specificity 98.1%. Variables associated with a positive methacholine response included sex, atopy, FEV1, FEV1/FVC and FENO, whereas they were FEV1/FVC and FENO for mannitol. A FENO value>26ppb, FEV1≤103.3% and female sex correctly classified 78.7% of methacholine responders. FENO value>26ppb was enough to correctly classify 81.5% of mannitol responders. CONCLUSIONS: Our study confirms the diagnostic profile of methacholine and mannitol challenge tests and describes the variable associated to their positivity with new proposed cutoff values.