Accuracy and repeatability of spirometry measurements are essential to obtain reliable efficacy data in randomized asthma clinical trials. We report our experience with a centralized spirometry quality assurance program that we implemented in our phase III asthma trials. Six asthma trials of 4 to 21 weeks in duration were conducted at 232 clinical centers in 31 countries. Approximately 23,100 prebronchodilator and 13,700 postbronchodilator spirometry tests were collected from 2523 adult and 336 pediatric asthmatic patients. The program used a standard spirometer (the Renaissance spirometry system) with maneuver quality messages and automated quality grading of the spirometry tests. Each clinical center transmitted spirometry data weekly to a central database, where uniform monitoring of data quality was performed and feedback was provided in weekly quality reports. Seventy-nine percent of all patients performed spirometry sessions with quality that either met or exceeded American Thoracic Society standards and improved over time. Good-quality spirometry was associated with (1) less severe asthma; (2) active treatment; (3) infrequent nocturnal awakenings; (4) age above 15 years; and (5) low body weight. Maneuver-induced bronchospasm was rare. Good-quality spirometry was observed in multicenter asthma clinical trials that employed a standard spirometer and continuous monitoring. Both within- and between-patient variability decreased. Spirometry quality improved with time as study participants and technicians gained experience.