Exhaled H2O2 is considered an indicator of lung inflammatory and oxidative stress. Moreover, H2O2 may be involved in signal transduction processes. It is not fully elucidated to what extent (i) H2O2 escapes from the intravascular compartment, and (ii) pulmonary H2O2 generation and nasopharyngeal H2O2 generation contribute to exhaled H2O2. We investigated H2O2 concentrations in breath condensate from isolated buffer-perfused and ventilated rabbit lungs, and from both intubated and spontaneously breathing rabbits with a horseradish peroxidase/2',7'dichlorofluorescin assay. For the perfused lungs, a H2O2 concentration of 58 +/- 19 nM was found. Addition of H2O2 to the buffer fluid resulted in only minute appearance in the exhaled air (<0.001%). Levels of exhaled H2O2 in intubated rabbits and perfused lungs were virtually identical. Nearly ten-fold higher levels were detected in spontaneously breathing rabbits. Decreasing the inspired oxygen concentration from 21% to 1% resulted in a tendency toward decreased H2O2 exhalation in perfused lungs. In contrast, phorbol-12-myristate-13-acetate (PMA) prompted a approximately 4-fold increase in H2O2 exhalation. We conclude that the horseradish peroxidase/2',7'dichlorofluorescin assay is a feasible technique to measure H2O2 in exhaled breath condensate in rabbits. When collecting exhaled air via the tracheal tube, the signal represents pulmonary H2O2 generation with the contribution of the remaining body being negligible.