STUDY OBJECTIVES: To elucidate whether a simple walking aid may improve physical performance in COPD patients with chronic respiratory insufficiency who usually carry their own heavy oxygen canister. DESIGN: Randomized crossover trial. SETTING: Physiopathology laboratory of three rehabilitation centers. PATIENTS AND INTERVENTIONS: We studied 60 stable COPD patients (mean age, 70.6 +/- 7.9 years; FEV(1), 44.8 +/- 14.3% of predicted [+/- SD]) with chronic respiratory insufficiency who randomly performed, on 2 consecutive days, a standardized 6-min walking test using two different modalities: a full-weight oxygen canister transported using a small wheeled cart and pulled by the patient (Aid modality) or full-weight oxygen canister carried on the patient's shoulder (No-Aid modality). MEASUREMENTS AND RESULTS: The distance walked, peak effort dyspnea, and leg fatigue scores as primary outcomes, and other cardiorespiratory parameters as secondary outcomes were recorded during both tests. A significant difference (p < 0.05) between the two tests occurred for all the measured outcomes in favor of the Aid modality. Most importantly, significant changes for distance (+ 43 m, p < 0.001), peak effort dyspnea (- 2.0 points, p < 0.001), leg fatigue (- 1.4 points, p < 0.001), as well as for mean and nadir oxygen saturation and heart rate with the Aid modality (but not with the No-Aid modality) were recorded in the subgroup of patients walking < 300 m at baseline. CONCLUSIONS: This study suggests that a simple walking aid may be helpful in COPD patients receiving long-term oxygen therapy, particularly in those with lower residual exercise capacity.