Exercise-related dyspnea is the main symptom of chronic obstructive pulmonary disease (COPD), yet its relation to lung function deterioration is weak. The aim of this study was to evaluate the relation between the patients' usual level of dyspnea and dyspnea caused by a maximum cardiopulmonary stress test or a 6-minute walking test.Thirty-six consecutive patients with stable COPD (age 66 7 years post-bronchodilator FEV1 47 14% of predicted) were studied. In addition to full baseline function testing, all patients underwent stress testing on a cycle ergometer and a 6-minute walking test in a corridor 50 m long. Exercise-induced dyspnea was assessed by the patient on a Borg scale before beginning and after completing each test. Chronic dyspnea during activities of daily living was quantified on the Medical Research Council (MRC) scale.The MRC value was only weakly related to percent of predicted FEV1 (r = 0.34, p = 0.04). Parameters obtained during exercise tests that were associated with the MRC were SaO2 at the end of the 6-minute walking test (r = 0.49, p = 0.004) and change in dyspnea on the Borg scale during the 6-minute walking test (deltaBorg-6mWT, r = 0.54, p = 0.0008) and during the stress test (r = 0.35, p = 0.04). Multiple regression analysis, with the MRC result as the dependent variable, showed that deltaBorg-6mWT and SaO2 at the end of the walking test explained 29% of the variance.Severity of chronic dyspnea in COPD patients assessed on the MRC scale is more related to dyspnea triggered by the walking test than with dyspnea induced by cycle ergometer stress testing.