The objective of this study was to test the hypothesis that cytotoxic oxygen metabolites participate in lytic cardiac cell damage, detected as creatine kinase release, upon reoxygenation of hypoxic, isolated buffer-perfused hearts (oxygen paradox). Perfusate additives included: superoxide dismutase (30 mg/l); catalase (2 mg/l); deferoxamine (0.5 mM); and allopurinol (1 mM). Creatine kinase release upon reoxygenation was reduced, to levels not significantly different from nonhypoxic controls, by adding either catalase, allopurinol or deferoxamine to the buffer during hypoxia. Reduced creatine kinase leakage was not accompanied by parallel preservation of ventricular function or coronary vascular resistance. Administration of catalase during hypoxia was superior to administering it only during reoxygenation. Treatment with catalase during both hypoxia and reoxygenation provided no more protection than administration only during hypoxia. The data suggest that an important component of hypoxia-induced cardiac cell damage is due primarily to hydrogen peroxide, which may then form hydroxyl radical. Superoxide anion plays an important role as a precursor of these species, but added superoxide dismutase alone did not significantly reduce creatine kinase loss. The data also suggest that damage resulting in creatine kinase release upon reoxygenation occurs during oxygen deprivation, and it is mediated in part by cytotoxic oxygen metabolites.