Since COPD patients respond differently to PLB, in line with previous authors we divided our sample in
IMPROVER and NONIMPROVER aiming at identifying the profile of individuals that can
benefit from this technique. We used a 25% increase in endurance time with PLB as a threshold to
include patients in the IMPROVER subgroup, because this threshold was larger than the test–retest
coefficient of variation for endurance time in COPD patients that underwent endurance test with the
same load used in this study.