Objective:To test the hypothesis that, in decompensated chronic obstructive pulmonary disease (COPD), noninvasive pressure support ventilation using 70:30 helium:oxygen instead of 70:30 air:oxygen could reduce dyspnea and improve ventilatory variables, gas exchange, and hemodynamic tolerance.Design:Prospective, randomized, crossover study.Setting:Medical intensive care unit, university tertiary care center.Patients:Nineteen patients with severe COPD (forced 1-sec expiratory volume of 0.83 ± 0.3 l) hospitalized in the intensive care unit for noninvasive pressure support ventilation after initial stabilization with noninvasive pressure support for no more than 24 hrs after intensive care unit admission.Interventions:Noninvasive pressure support ventilation was administered in the following randomized crossover design: a) 45 min with air:oxygen or helium:oxygen; b) no ventilation for 45 min; and c) 45 min with air:oxygen or helium:oxygen.Measurements and Main Results:Air:oxygen and helium:oxygen decreased respiratory rate and increased tidal volume and minute ventilation. Helium:oxygen decreased inspiratory time. Both gases increased total respiratory cycle time and decreased the inspiratory/total time ratio, the reduction in the latter being significantly greater with helium:oxygen. Peak inspiratory flow rate increased more with helium:oxygen. PaO2increased with both gases, whereas PaCO2decreased more with helium:oxygen (values shown are mean ± SD) (52 ± 6 torr [6.9 ± 0.8 kPa] vs. 55 ± 8 torr [7.3 ± 1.1 kPa] and 48 ± 6 torr [6.4 ± 0.8 kPa] vs. 54 ± 7 torr [7.2 ± 0.9 kPa] for air:oxygen and helium:oxygen, respectively;p< .05). When hypercapnia was severe (PaCO2> 56 torr [7.5 kPa]), PaCO2decreased by ≥7.5 torr (1 kPa) in six of seven patients with helium:oxygen and in four of seven patients with air:oxygen (p< .01). Dyspnea score (Borg scale) decreased more with helium:oxygen than with air:oxygen (3.7 ± 1.6 vs. 4.5 ± 1.4 and 2.8 ± 1.6 vs. 4.6 ± 1.5 for air:oxygen and helium:oxygen, respectively;p< .05). Mean arterial blood pressure decreased with air:oxygen (76 ± 12 vs. 82 ± 14 mm Hg;p< .05) but remained unchanged with helium:oxygen.Conclusion:In decompensated COPD patients, noninvasive pressure support ventilation with helium:oxygen reduced dyspnea and PaCO2more than air:oxygen, modified respiratory cycle times, and did not modify systemic blood pressure. These effects could prove beneficial in COPD patients with severe acute respiratory failure and might reduce the need for endotracheal intubation.