Methacholine tests are widely used as a diagnostic aid for asthma. Their strength has been reputed to be the high sensitivity and very infrequent occurrence of false-negative test results (ie, high negative predictive value). There are 2 commonly used methods that have been outlined by the American Thoracic Society. These methods were thought to give equivalent results. However, in 3 investigations in which we have compared the 2 methods, we have demonstrated a marked lack of comparability. In subjects with borderline to mild airway responsiveness (tidal breathing, methacholine PC20 >2 mg/mL), the 5 deep inhalations required of the dosimeter method produce marked bronchoprotection in some subjects with asthma. The result of this bronchoprotection is that in 55 subjects with asthma, 50% of those whose tidal breathing PC20 value was greater than 2 mg/mL and 25% of the total had negative methacholine challenge results. This indicates that the standardized dosimeter method has an unacceptable loss of diagnostic sensitivity. We recommend that the dosimeter method not be performed as outlined by the American Thoracic Society and that methacholine should be administered by means of submaximal inhalations or tidal breathing.