Corticosteroids are widely used as therapy for the adult respiratory distress syndrome (ARDS) without proof of efficacy. We conducted a prospective, randomized, double-blind, placebo-controlled trial of methylprednisolone therapy in 99 patients with refractory hypoxemia, diffuse bilateral infiltrates on chest radiography and absence of congestive heart failure documented by pulmonary-artery catheterization. The causes of ARDS included sepsis (27 percent), aspiration pneumonia (18 percent), pancreatitis (4 percent), shock (2 percent), fat emboli (1 percent), and miscellaneous causes or more than one cause (42 percent). Fifty patients received methylprednisolone (30 mg per kilogram of body weight every six hours for 24 hours), and 49 received placebo according to the same schedule. Serial measurements were made of pulmonary shunting, the ratio of partial pressure of arterial oxygen to partial pressure of alveolar oxygen, the chest radiograph severity score, total thoracic compliance, and pulmonary-artery pressure. We observed no statistical differences between groups in these characteristics upon entry or during the five days after entry. Forty-five days after entry there were no differences between the methylprednisolone and placebo groups in mortality (respectively, 30 of 50 [60 percent; 95 percent confidence interval, 46 to 74] and 31 of 49 [63 percent; 95 percent confidence interval, 49 to 77]; P = 0.74) or in the reversal of ARDS (18 of 50 [36 percent] vs. 19 of 49 [39 percent]; P = 0.77). However, the relatively wide confidence intervals in the mortality data make it impossible to exclude a small effect of treatment. Infectious complications were similar in the methylprednisolone group (8 of 50 [16 percent]) and the placebo group (5 of 49 [10 percent]; P = 0.60). Our data suggest that in patients with established ARDS due to sepsis, aspiration, or a mixed cause, high-dose methylprednisolone does not affect outcome.