In patients with a history suggestive of asthma, diagnosis is usually confirmed by spirometry with bronchodilator response (BDR) or confirmatory methacholine challenge testing (MCT).We examined the proportion of participants with negative BDR testing who had a positive MCT (and its predictors), and characteristics of MCT, including effects of controller medication tapering and temporal variability (and predictors of MCT result change); and concordance between MCT and pulmonologist asthma diagnosis.Adults with self-reported physician-diagnosed asthma were recruited by random-digit dialing across Canada. Subjects performed spirometry with BDR testing and returned for MCT if testing was non-diagnostic for asthma. Subjects on controllers underwent medication tapering with serial MCTs over 3-6 weeks. Subjects with a negative MCT (PC20 > 8 mg/mL) off medications were examined by a pulmonologist and had serial MCTs after 6 and 12 months.Of 500 subjects (50.5 +/- 16.6 years old, 68.0% female) with a negative BDR test for asthma, 215 (43.0%) had a positive MCT. Subjects with pre-bronchodilator airflow limitation were more likely to have a positive MCT (odds ratio 1.90; 95% confidence interval 1.17-3.04). MCT converted from negative to positive with medication tapering in 18/94 (19.1%) participants, and spontaneously over time in 25/165 (15.2%) participants. Of 231 subjects with negative MCT, 28 (12.1%) subsequently received an asthma diagnosis from a pulmonologist.In subjects with a self-reported physician diagnosis of asthma, absence of bronchodilator reversibility had a negative predictive value of only 57% to exclude asthma. A finding of spirometric airflow limitation significantly increased chances of asthma. MCT results varied with medication taper and over time, and pulmonologists were sometimes prepared to give a clinical diagnosis of asthma despite negative MCT. Correspondingly, in patients for whom a high clinical suspicion of asthma exists, repeat testing appears to be warranted.