A prospective, crossover, physiologic study was performed in 10 patients with acute lung injury to assess the respective short-term effects of noninvasive pressure-support ventilation and continuous positive airway pressure. We measured breathing pattern, neuromuscular drive, inspiratory muscle effort, arterial blood gases, and dyspnea while breathing with minimal support and the equipment for measurements, with two combinations of pressure-support ventilation above positive end-expiratory pressure (10-10 and 15-5 cm H2O), and with continuous positive airway pressure (10 cm H2O). Tidal volume was increased with pressure support, and not with continuous positive airway pressure. Neuromuscular drive and inspiratory muscle effort were lower with the two pressure-support ventilation levels than with other situations (p < 0.05). Dyspnea relief was significantly better with high-level pressure-support ventilation (15-5 cm H2O; p < 0.001). Oxygenation improved when 10 cm H2O positive end-expiratory pressure was applied, alone or in combination. We conclude that, in patients with acute lung injury (1) noninvasive pressure-support ventilation combined with positive end-expiratory pressure is needed to reduce inspiratory muscle effort; (2) continuous positive airway pressure, in this setting, improves oxygenation but fails to unload the respiratory muscles; and (3) pressure-support levels of 10 and 15 cm H2O provide similar unloading but differ in their effects on dyspnea.