The purpose of this study was to determine the variability in maximal oxygen consumption (VO2 max) determined from repeated exercise tests in patients with chronic air-flow obstruction (CAO). Three incremental maximal cycle ergometer tests were performed in each of 11 CAO patients who were familiar with such testing. Two tests (Test 1, Test 2) were carried out on the same day, separated by a 60-min rest period, and 1 (Test 3) was performed on a consecutive day. Group mean values for VO2 max were: 1.313 +/- 0.259, 1.311 +/- 0.281, 1.306 +/- 0.288 L/min, for Tests 1 to 3, respectively. These nearly identical values did not differ significantly. There was no systematic "fatigue" or "learning" effect from test to test. Other mean measurements obtained at maximal exercise were likewise not significantly different among the 3 tests. For tests performed on the same day, the mean of the absolute values of the individual patient VO2 max differences (delta VO2 max) was 53 +/- 30 ml. The delta VO2 max was less than 6% in 10 of the 11 patients and less than 10% in the remaining patient. For tests performed on consecutive days, delta VO2 max was 93 +/- 81 ml; delta VO2 max was less than 6% in 6 patients and less than 10% in 9 patients. From this study we concluded that repeated maximal exercise testing is highly reproducible for groups of CAO patients, although some individual patient variability is seen. Individual patient differences in delta VO2 max are less when the 2 tests are performed on the same day than when duplicate testing is performed on consecutive days.