In paired experiments, we studied the effects of high-dose methylprednisolone on the acute pulmonary injury caused by 4 h of venous air embolization in 19 chronically instrumented, unanesthetized sheep with lung lymph fistulas. We compared the effect of methylprednisolone (30 mg/kg intravenous bolus) given before embolization, early (1 h) in the course of embolization, late (3 h) in the course of embolization, or after embolization (at the beginning of the recovery period). We measured pulmonary hemodynamics and lymph dynamics. In six sheep we also fixed lung tissue for semiquantitative histology, and in some we measured leukocyte concentrations in blood and in pulmonary lymph. Methylprednisolone did not significantly affect pulmonary hemodynamics but it largely prevented lung injury when it was given before embolization. It also lessened the degree of lung injury when it was given during embolization, although this effect became less marked as treatment was delayed. Methylprednisolone had no effect on lung injury when given after embolization was completed (4 h). We found fewer leukocytes attached to air emboli and fewer endothelial cell gaps in the lungs of sheep given methylprednisolone as prophylaxis. Leukocyte counts were lower in lung lymph and higher in the circulating blood of methylprednisolone-treated sheep. We conclude that methylprednisolone has a preventive effect on air embolism lung injury, such that its effect is greater when given earlier during the development of injury.