OBJECTIVE The purpose of this study was to analyze the mechanisms associated with lung injury after cardiopulmonary bypass and to propose strategies of prevention. METHODS Thirty-two neonatal piglets underwent 90 minutes of hypothermic cardiopulmonary bypass without aortic cross-clamping. Five experimental groups were defined: group I had standard cardiopulmonary bypass (control), group II received continuous low-flow lung perfusion during cardiopulmonary bypass, group III treatment was similar to that of group I with maintenance of ventilation, group IV received pneumoplegia, and group V received nitric oxide ventilation (30 ppm) after cardiopulmonary bypass. Data drawn from hemodynamic and gas exchange values and muscular and pulmonary tissular levels of adenosine triphosphate (in micromoles per gram) and myeloperoxidase (in international units per 100 mg) were used for comparisons before and 30 and 60 minutes after cardiopulmonary bypass. Pulmonary and systemic vascular endothelial functions were assessed in vitro after cardiopulmonary bypass on isolated rings of pulmonary and iliac arteries. RESULTS Pulmonary vascular resistance index, cardiac index, and oxygen tension were better preserved in groups II, IV, and V. All groups disclosed a significant decrease in lung adenosine triphosphate levels and an increase in myeloperoxidase activity whereas these levels stayed within pre-cardiopulmonary bypass ranges in muscular beds. Endothelium-dependent relaxation was preserved in systemic arteries but was strongly affected in pulmonary arteries after cardiopulmonary bypass. None of the methods that aimed to protect the pulmonary vascular bed demonstrated any preservation of pulmonary endothelial function. CONCLUSIONS Cardiopulmonary bypass results in ischemia-reperfusion injury of the pulmonary vascular bed. Lung protection by continuous perfusion, pneumoplegia, or nitric oxide ventilation can prevent hemodynamic alterations after cardiopulmonary bypass but failed to prevent any of the biochemical disturbances.