The oxygen paradox refers to the abrupt release of cytoplasmic enzymes and severe cellular disruption that occurs following reoxygenation of anoxic perfused hearts. In this study, the ability of a series of oxygen-derived free radical inhibitors and scavenging agents to protect isolated perfused rat hearts from the oxygen-induced enzyme release following 30 or 60 mins of anoxic perfusion (oxygen paradox) and cumene hydroperoxide-induced injury was evaluated. Malondialdehyde (MDA) release, an indicator of lipid peroxidation, and creatine kinase (CK) release, an indicator of cellular injury, were monitored. We evaluated five agents previously reported to scavenge or inhibit the formation of oxygen free radicals. The putative hydroxyl radical scavengers dimethylthiourea (DMTU) and mannitol; catalase, an agent protective against peroxide injury; allopurinol, an inhibitor of xanthine oxidase; and albumin, a non-specific protein control, were evaluated. Coronary flow rates and myocardial temperature were continuously monitored to ensure uniform perfusion conditions. The MDA assay was carefully monitored by constructing standard curves on each experimental day. Addition of 20 microM cumene hydroperoxide to oxygenated perfused hearts caused peroxidative cell injury as evidenced by significant MDA and CK release in the coronary effluent. DMTU and catalase provided near complete protection from cumene hydroperoxide-induced cell injury but did not reduce CK release from hearts subjected to either the mild (30-min) or the severe (60-min) oxygen paradox (reoxygenation-induced injury). Allopurinol caused a significant reduction in MDA release but not CK release from oxygen paradox-injured hearts. Allopurinol and albumin had no significant effect on MDA release from cumene-hydroperoxide-injured hearts. Catalase (300 U/ml) caused a mild but not statistically significant reduction in MDA release from cumene hydroperoxide injury but did not provide protection from the oxygen paradox at either injury level. Mannitol (120 mM), in contrast to DMTU, was ineffective in reducing cumene-induced injury but showed a significant protective effect against oxygen paradox-induced damage. It is concluded that the ability of mannitol to reduce reoxygenation-induced CK release in the oxygen paradox may be due to its osmotic activity and consequent ability to prevent cellular swelling rather than its activity as an oxygen-free radical scavenger.