The purpose of this study was to determine the correlation between different intervals in the 12-min walk test, to determine which of the intervals best correlated with maximal oxygen intake (VO2max) and maximal CO2 expelled (VCO2max) and to determine the degree of correlation between changes in the VO2max and changes in the walk test and spirometry. Nine elderly (age, 67 +/- 4 years) patients with moderate COPD (FEV1, 1.32 +/- 0.28 L) who were ventilatory limited were seen 6 times over a 14-week period. At each visit they performed spirometry, a 12-min walk, and a symptom-limited maximal exercise test. During the 12-min walk the distances traversed in each 2-min interval were recorded. A Borg score was obtained at the end of each 2 min. The distances that the patients walked in each 2-min interval were very similar, but the mean Borg score became higher the longer the patient walked. The Borg score after 2 min was 1.64 +/- 1.15, while the Borg score after 12 min was 5.70 +/- 3.59. The correlation coefficients between the distance walked at various intervals and the oxygen consumption (VO2)/kg (approximately 0.65) and the VCO2/kg (approximately 0.52) were very similar. However, changes in the VO2/kg were more closely correlated with changes in the 12-min walk (r = 0.72), than with changes in the 6-min walk (r = 0.64), the 4-min walk (r = 0.59), or the 2-min walk (r = 0.53). This latter observation suggests that the 12-min walk may be preferable to tests in which the patient walks 4 or 6 min for documenting changes in the exercise capabilities.