The airway edema that develops in guinea pigs after exposure to toluene diisocyanate (TDI) requires the presence of polymorphonuclear leukocytes (PMN). To determine whether this airway edema is mediated by the release of hydrogen peroxide from PMN, we treated animals intravenously with catalase bound to polyethylene glycol and examined the extravasation of Evans blue dye into the tracheal wall after exposure to air or 3 ppm TDI for 1 h. Catalase (25,000, 100,000, and 300,000 IU/kg) caused a dose-dependent inhibition of the TDI-induced increase in dye extravasation. However, treatment with catalase, inactivated at the peroxide binding site with 3-aminotriazole, inhibited dye extravasation after exposure to TDI as effectively as the equimolar 100,000 IU/kg dose of active catalase. Injection of polyethylene glycol alone was without effect. Dose-dependent decreases in extravascular migration of PMN and in circulating PMN also were noted after catalase treatment. These results suggest that the catalase preparations used in these studies inhibited the PMN-dependent airway edema by an effect other than hydrogen peroxide scavenging. Examination of this and other commercially available catalase preparations revealed trace concentrations of endotoxin at levels that could be responsible for the observed effects on PMN function. Treatment of animals with doses of Escherichia coli endotoxin similar to those inadvertantly administered to the catalase-treated groups (0.1 ng/kg to 100 ng/kg, intravenously) inhibited TDI-induced extravasation of Evans blue dye in a dose-dependent manner. These results suggest that contaminating endotoxin may contribute to some of the protective effects of preparations of catalase observed in previous studies of vascular permeability.