Inhaled nitric oxide (NO) causes selective pulmonary vasodilation and improves gas exchange in acute lung failure. In experimental pulmonary hypertension, we compared the influence of the aerosolized vasodilatory prostaglandins (PG) PGI2 and PGE1 on vascular tone and gas exchange to that of infused prostanoids (PGI2, PGE1) and inhaled NO. An increase of pulmonary artery pressure (Ppa) from 8 to approximately 34 mmHg was provoked by continuous infusion of U-46619 (thromboxane A2 (TxA2) analogue) in blood-free perfused rabbit lungs. This was accompanied by formation of moderate lung oedema and severe ventilation-perfusion (V'/Q') mismatch, with predominance of shunt flow (>50%, assessed by the multiple inert gas elimination technique). When standardized to reduce the Pps by approximately 10 mmHg, inhaled NO (200 ppm), aerosolized PGI2 (4 ng x kg(-1) x min(-1)) and nebulized PGE1 (8 ng x kg(-1) x min(-1)) all reduced both pre- and postcapillary vascular resistance, but did not affect formation of lung oedema. All inhalative agents improved the V'/Q' mismatch and reduced shunt flow, the rank order of this capacity being NO > PGI2 > PGE1. In contrast, lowering of Ppa by intravascular administration of PGI2 and PGE1 did not improve gas exchange. "Supratherapeutic" doses of inhaled vasodilators in control lungs (400 ppm NO, 30 ng x kg(-1) x min(-1) of PGI2 or PGE1) did not provoke vascular leakage or affect the physiological V'/Q' matching. We conclude that aerosolization of prostaglandins I2 and E1 is as effective as inhalation of nitric oxide in relieving pulmonary hypertension. When administered via this route instead of being infused intravascularly, the prostanoids are capable of improving ventilation-perfusion matching, suggesting selective vasodilation in well-ventilated lung areas.