Many animal studies have shown that high peak inspiratory pressures (PIP) during mechanical ventilation can induce acute lung injury with hyaline membranes. Since 1984 we have limited PIP in patients with ARDS by reducing tidal volume, allowing spontaneous breathing with SIMV and disregarding hypercapnia. Since 1987 50 patients with severe ARDS with a "lung injury score" greater than or equal to 2.5 and a mean PaO2/FiO2 ratio of 94 were managed in this manner. The mean maximum PaCO2 was 62 mmHg, the highest being 129 mmHg. The hospital mortality was significantly lower than that predicted by Apache II (16% vs. 39.6%, chi 2 = 11.64, p less than 0.001). Only one death was due to respiratory failure, caused by pneumocystis pneumonia. 10 patients had a "ventilator score" greater than 80, which has previously predicted 100% mortality from respiratory failure. Only 2 died, neither from respiratory failure. There was no significant difference in lung injury score, ventilator score, PaO2/FiO2 or maximum PaCO2 between survivors and non-survivors. We suggest that this ventilatory management may substantially reduce mortality in ARDS, particularly from respiratory failure.