Although arm activity is poorly tolerated by patients with COPD, the ventilatory response to arm elevation alone is not well understood. We therefore studied the ventilatory response to arm elevation using a customized arm support sling to eliminate the effect of an increase in metabolic activity that might be attributable to independent arm elevation and used leg exercise to increase metabolic activity. During arm elevation at rest, there was a significant decrease in vital capacity (180 ml) and a small decrease in functional residual capacity (120 ml) as measured by body plethysmography. Minute ventilation was unchanged. When supported arm elevation (SAE) was compared with the control arm position (CAP), minute ventilation was unchanged although the pattern of breathing became more rapid and shallow (mean +/- SD, SAE vs CAP: fb = 17.9 +/- 5.3 vs 16.2 +/- 4.8 breaths.min-1; VT = 533 +/- 126 vs 579 +/- 142 ml; p < 0.05). During steady-state leg exercise, the increase in VO2, VCO2 and VE did not differ between SAE and CAP; however, both fb and VT changed toward a more rapid, shallow pattern of breathing (SAE vs CAP: fb = 24.3 +/- 3.0 vs 22.8 +/- 3.5 breaths.min-1; VT = 990 +/- 293 vs 1,081 +/- 309 ml; p < 0.05). During unsupported arm elevation VO2, VCO2, and VE, and fb were significantly greater than during the CAP. Approaches that train arm muscles and strategies that either support arm muscles or allow for frequent rests during upper arm activity may improve the endurance and the quality of life for COPD patients.