In this study we investigated the patterns of pulmonary venous flow in children with functional single ventricles to obtain a better understanding of the determinants of transpulmonary blood flow. Sixty-eight patients with functional single ventricles and aortopulmonary shunt (n = 34, group I), or superior cavopulmonary connection (n = 34, group II) underwent transesophageal Doppler echocardiographic assessment of flow in the left upper pulmonary vein before undergoing the next stage of surgery. Twelve patients from group II also underwent simultaneous evaluation of superior vena caval flow. Biphasic forward pulmonary venous flow was noted in 62 patients in sinus rhythm (S wave in systole, D wave in diastole); in 6 patients with junctional rhythm, significant early systolic reversal of flow was present. Both the S- and D-wave velocity-time integrals (VTI) were greater in group I than in group II (S(VTI) 9.9 +/- 4.2 vs 8.0 +/- 2.6, p = 0.02; D(VTI) 8.0 +/- 3.5 vs 4.2 +/- 2.6, p <0.001). In both groups, pulmonary venous flow was predominantly systolic; however, the proportion of flow during ventricular systole was significantly greater in group II than in group I (S(VTI)/D(VTI) group II: 2.4 +/- 1.5; group I 1.4 +/- 0.5, p = 0.001; percent systolic fraction of pulmonary venous flow group II = 67%, group I = 56%, p <0.001). Analysis of superior vena caval flow in group II revealed a single predominant wave with onset at early systole and peak in late systole at a mean of 150 ms after the pulmonary venous S-wave peak. Our data suggest that ventricular systole (i.e., atrial relaxation, atrioventricular valve descent) asserts great influence on transpulmonary blood flow in the functional single ventricle.