Rapid progress in cystic fibrosis research affords the possibility of halting the progress of the lung disease. We used data from 215 patients who had sputum cultures negative for Burkholderia cepacia, at least one outpatient pulmonary function test during 1990, and at least one test a year later to estimate the number of subjects and study duration required to demonstrate that a hypothetical treatment reduces the rate of decline of forced expiratory volume in 1 s (FEV1) to zero. Mean rate of decline of FEV1 (percent predicted) was about 2% predicted per year. Variability decreases with increasing time of observation. For a 1-y study, with alpha = 0.05 and beta = 0.20, over 550 patients must complete the study in each group to show that a treatment halts pulmonary decline. For a 2-y study, 86 subjects in each group are required, and for 4 y, 65. Increasing the number of data points used to determine the rate of decline of FEV1 had only small effect on sample size. Use of pulmonary function data collected at regular intervals for research purposes did not alter these conclusions. Higher initial FEV1 was associated with a greater rate of decline, and among patients with initial FEV1 > 60% predicted, younger subjects had a faster decline than did older subjects. Thus, fewer subjects will be required to detect a complete halt in progression of lung disease if the patients are young and have mild pulmonary disease.