OBJECTIVE This study was initiated to test the hypothesis that fetal hemodynamic changes observed under pulsatile flow bypass might be related to the release of endothelium-derived relaxing factor through oscillating shear stress. METHODS Normothermic bypass was instituted in utero in 21 preterm fetal lambs for a 1-hour period through the right atrium and main pulmonary artery. Ultrasonic flowmeters were positioned around the descending aorta and the umbilical artery. The circuit consisted of an oxygenator and a pump set to either continuous flow (n = 7) or pulsatile flow (n = 7) and adjusted to maintain a fetal main arterial pressure of 50 mm Hg. In seven other animals, endothelium-derived relaxing factor was blocked by a continuous infusion of N omega-nitro-L-arginine after 30 minutes of pulsatile flow. RESULTS During the first 30 minutes of bypass, pump flows were significantly lower in the continuous-flow group than in the pulsatile-flow or blocked-flow groups (respectively, 612 +/- 144, 907 +/- 153 and 987 +/- 228 ml/min), with similar changes in aortic and umbilical flows. Systemic vascular resistances were significantly lower in the pulsatile-flow and blocked-flow groups than in the continuous-flow group (550 +/- 106 vs 821 +/- 212 dynes/sec/cm-5). However, after blockade of endothelium-derived relaxing factor, resistances increased gradually in the blocked-flow group to reach the level of that of the continuous-flow group at the end of bypass (943 +/- 77 vs 556 +/- 143 dynes/sec/cm-5 in the pulsatile-flow group). CONCLUSIONS Blockade of endothelium-derived relaxing factor after 30 minutes of pulsatile flow returns fetal hemodynamics to continuous flow conditions. The specific inhibitor of endothelium-derived relaxing factor used in this experiment suggests that nitric oxide may be released by fetal endothelium during pulsatile bypass.