The purpose of this study was to evaluate the rest-retest reliability and validity of a constant-load endurance exercise test on a cycle ergometer with a workload of 75% of maximal work capacity (W(max)) in patients with moderate to severe chronic obstructive pulmonary disease (COPD). In 60 patients with COPD (FEV(1) 40 +/- 15% pred), exercise endurance time was measured with a constant-load endurance exercise tests at 75% of W(max), on two different occasions. In a subgroup of 20 patients, test-retest reliability of the measurement of end-exercise ventilatory and metabolic responses was assessed. Validity of the cycle endurance test was assessed comparing endurance time and total work performed during the cycle endurance test to peak oxygen uptake (VO(2peak)) and the 12-minute walking distance (12MWD). Test and retest assessments of cycle endurance time did not differ statistically significantly (P =.40). Highly significant intraclass correlation coefficients (ICC > or = 0.85; P <.001) were found between test and retest of assessment of endurance time as well as of end-exercise ventilatory and metabolic responses. In addition, statistically significant correlation coefficients were found between VO(2peak) and endurance time (r = 0.50; P =.001) and total work performed (r = 0.72; P <.001) during the constant-load cycle test. Significant correlation coefficients of the same magnitude were found between 12MWD and endurance time (r = 0.58; P =.001) and total work performed (r = 0.72; P <.001) during the constant-load cycle test. It is concluded that constant-load exercise testing on a cycle ergometer with a workload of 75% of maximal work capacity is a reliable and valid method to assess exercise endurance in patients with COPD.