RESUMO
The diagnostic value of isocapnic hyperventilation of cold air (IHCA) is not fully established. All 342 adult patients in whom IHCA had been performed because of a clinical suspicion of asthma between 1992 and 1994 were analysed retrospectively in the authors' hospital. In addition, 26 healthy subjects were recruited. According to strict criteria, the patients were divided into asthmatics and symptomatic non-asthmatics. For the calculations of sensitivity, specificity and accuracy, the symptomatic non-asthmatic group served as a control. The post-test probability of asthma after IHCA was determined for all the possible pre-test probabilities by applying Bayes' theorem. A linear regression model was used to investigate the factors associated with the reactivity to IHCA. A single 4-min IHCA and skin prick tests were performed in the healthy subjects. Of the 287 patients in the final analysis, 113 were defined as asthmatics and 174 as symptomatic non-asthmatics. The accuracy was highest using a 9.0% fall in forced expiratory volume in 1 s (FEV1) as a cut-off value; the specificity was then 86.8% and the sensitivity 31.9%. The authors found IHCA to be a useful diagnostic test only if the pre-test probability of asthma is between 0.30 and 0.56. The positive final diagnostic gain of IHCA is 22% at its best, but the negative gain is negligible for all possible pre-test probabilities. Factors associated with reactivity to IHCA were young age and, to a lesser extent, a history of cold-weather-associated respiratory symptoms and pre-challenge bronchial obstruction. If a rigid cut-off value for a positive response is used in all age groups, the specificity of IHCA is good but the sensitivity is unacceptably low in adults. The diagnostic value of IHCA might increase if age is taken into account when defining the cut-off value.
