We developed an in vivo intact canine model to study pulmonary ischemia-reperfusion (IR) injury. The surgical approach simulates that of unilateral lung transplantation but is free of technical difficulties and other factors related to lung preservation. Serial measurements of regional pulmonary blood flow (rPBF), extravascular density (EVD), and transcapillary protein flux were made with the quantitative imaging technique of positron emission tomography. Eleven experimental and six control animals were studied. After 2 h of warm ischemia followed by reperfusion, no significant change occurred in rPBF despite significantly increased EVD, which was greater on the ischemic than on the nonischemic side. Protein flux, measured as a rate constant, was also greater on the ischemic than on the nonischemic side (median 181 x 10(-4)/min, range 104-619, vs. median 90, range 33-132) immediately after reperfusion. Both sides were also significantly different from control values (median 37, range 21-57). On both sides, protein flux decreased over time and at 5 h after reperfusion was not different from that of controls. Data from the control animals showed that these findings in the experimental animals were not due to surgical technique, deterioration in the surgical preparation, or hyperperfusion of the nonischemic lung. Thus IR injury of one lung can lead to similar, but less severe, injury in the contralateral lung. Because injury in the nonischemic lung develops only after reperfusion of the ischemic lung, injury to the nonischemic lung is probably humorally mediated. The model is a useful and relevant method for studying the physiological consequences of pulmonary IR injury.