We studied pulmonary resistance as a surgical criterion for atriopulmonary shunt. We created a model of experimental pulmonary hypertension by establishing a systemic-pulmonary shunt in 11 dogs. Two to 3 months after the shunt operation, total pulmonary resistance was calculated before (7.24 +/- 1.54 U . m2) and after (3.50 +/- 1.54 U . m2) ligation of the shunt. An atriopulmonary anastomosis technique was then performed and the hemodynamic status of the dogs during the first 2 postoperative hours was evaluated. Pathologic study of the lungs disclosed no arterial lesions. From our experimental work, we conclude that the surgical criterion for establishing the indication for this operation should be pulmonary resistance, which is conditioned by the state of the intrapulmonary vessels. We propose a procedure for determining real pulmonary resistance in the course of preoperative catheterization.