As part of the multicenter National Heart, Lung, and Blood Institute registry of patients with severe deficiency of alpha1-antitrypsin with 1,129 enrollees, this report describes measures undertaken to achieve high-quality FEV1 measurements, the rates of satisfying reproducibility and acceptability criteria, and clinical features of participants unable to achieve reproducible FEV1 values at baseline. Spirograms were performed both before and after an inhaled bronchodilator in enrollees followed up at 37 participating clinical centers. Using a reproducibility criterion of < 100 mL or 5% (whichever greater), high reproducibility rates for FEV1 measurements at baseline were observed for both prebronchodilator (95.0% of 1,090 sessions) and postbronchodilator measurements (95.7% of 1,077 sessions). Using the more recently published reproducibility criterion of < or = 200 mL, reproducibility rates were even higher. Eighty-four percent of clinical centers submitted FEV1 values that satisfied reproducibility criteria for at least 90% of spirograms. Also, the mean coefficient of variation for prebronchodilator FEV1 values measured over serial visits separated by up to 9 months was 5.6% for participants with baseline FEV1 55 to 90% predicted. This degree of reproducibility is similar to that observed in the Lung Health Study. Rates of satisfying acceptability criteria for prebronchodilator spirograms were lower, almost universally (98% of tests) due to failure to achieve end-of-test criteria (which usually required 15 s of expiration in this population with mean FEV1 = 42.6+/-29.6% [SD] predicted). Multivariate logistic regression models show that clinical correlates of failure to achieve reproducible prebronchodilator FEV1 efforts include symptoms of chronic wheeze, chronic cough, and chronic phlegm, and the degree of airflow obstruction. We conclude that highly reproducible FEV1 measurements are achievable in a population with severe airflow obstruction despite the additional challenges posed by testing in multiple centers on a variety of spirometers. Furthermore, the difficulty of satisfying end-of-test criteria in a large cohort with severe airflow obstruction did not preclude achieving high rates of reproducibility for FEV1 measurements. Finally, our study confirms prior observations that failure to achieve reproducible efforts is associated with the presence of pulmonary symptoms and the degree of airflow obstruction. Thus, excluding patients with nonreproducible FEV1 efforts from epidemiologic studies would bias results by including only healthier participants.