BACKGROUND: The 5-stage Chester step test assesses aerobic capacity in healthy subjects. It has not been tested in patients with COPD. OBJECTIVE: To determine the reliability of the Chester step test in patients with COPD and correlation with pulmonary function test and exercise test results. METHODS: Thirty-two patients (mean ± SD FEV(1) 46 ± 15% of predicted) undertook 2 Chester step tests and two 6-min walk tests, on different days, in random order. A subgroup of 11 patients performed incremental cycle ergometry. RESULTS: Thirty-one patients performed stage 1 of the Chester step test. Nineteen patients performed stage 2 of the Chester step test. The number of steps was highly reproducible: 66 ± 41 steps vs 68 ± 41 steps. There was no difference in heart rate or S(pO(2)) between the 2 Chester step tests at the peak of exercise or at the end of each stage. There was a significant correlation between number of steps and FEV(1) (r = 0.43, P = .02), and 6-min walk distance (r = 0.60, P = .001). Heart rate increased according to advanced stages of the Chester step test, up to 81 ± 13% of predicted. There was a significant correlation between number of steps and peak heart rate (r = 0.55, P = .001). In the 11 patients who performed the incremental cycling test there was a significant correlation between number of steps and peak work load (r = 0.69, P = .02). In the 6 patients in whom oxygen uptake could be estimated from the Chester step test, oxygen uptake was higher than that measured at the peak of the cycling test (30.8 ± 5.1 mL/kg/min vs 17.4 ± 4.5 mL/kg/min, respectively, P = .001). CONCLUSIONS: Despite being highly reproducible, the Chester step test had a very short duration in patients with COPD. The number of steps incremented in each stage seems to be too large for these patients. An adaptation of the Chester step test should be considered for patients with COPD.