Reoxygenation injury that occurs when blood circulation is restored to previously ischemic tissues is currently discussed as a pathophysiological entity distinct from the primary anoxic injury that develops during ischemia per se. To test the hypothesis that reoxygenation injury in hepatocytes is caused by a postischemic burst of reactive oxygen species (ROS), including superoxide radicals, O2-., and hydrogen peroxide, H2O2, we performed a cytochemical study exploiting the peroxidase activity within peroxisomes as a sensitive ultrastructural detector of intracellular H2O2 generation. The osmiophilic polymer formed when tissue peroxidase is incubated with 3,3'-diaminobenzidine (DAB) and H2O2 was used as a marker for endogenous H2O2 in rat liver slices in short-term organ culture subjected to a cycle of 60-min ischemic anoxia and 30-min reoxygenation in the presence of DAB without exogenous H2O2. Peroxisomal reaction product was quantitatively evaluated in transmission electron micrographs of systematically sampled hepatocytes. Mean densities of positive peroxisomes per 1,000 micron2 (+/- SE) in liver slices subjected to various treatments were as follows: continuous anoxia (negative control) 0 +/- 0; normoxia + exogenous H2O2 (positive control) 45 +/- 12; normoxia only 26 +/- 2; ischemia-reoxygenation 13 +/- 6; ischemia-reoxygenation + xanthine oxidase inhibitor, oxypurinol 5 +/- 3; ischemia-reoxygenation + peroxidase inhibitor, aminotriazole 7 +/- 3. Endogenous H2O2 can be detected in hepatocytes by electron microscopic cytochemistry and may in part derive from xanthine oxidase, but it is not substantially increased in the postischemic state. We conclude that hepatocytes do not exhibit a postischemic burst of reactive oxygen species that could cause reoxygenation injury.