Abnormal pulmonary venous flow velocity patterns are present in multiple cardiac disease states, but the determinants of pulmonary venous flow velocity have not been fully elucidated. To determine the relative importance of several proposed factors that could influence pulmonary venous flow, anatomic, hemodynamic, and Doppler mitral and pulmonary venous flow velocity data were compared in 50 consecutive patients undergoing cardiac catheterization for clinical reasons. Pulmonary venous diastolic flow velocity was most strongly related to left ventricular isovolumetric relaxation time (r = -0.59), left ventricular end-diastolic pressure (LVEDP, r = 0.50), left atrial minimum volume (r = 0.50), mitral deceleration time (r = -0.50), and early diastolic mitral flow velocity (r = 0.45). Pulmonary venous systolic flow velocity related best with left atrial minimum volume (r = -0.45) and left atrial ejection fraction (r = 0.44). Pulmonary venous systolic flow velocity integral also showed the strongest relation with left atrial minimum volume (r = -0.48). Relations between pulmonary venous flow velocity and velocity time integrals and other variables were sometimes significantly different, apparently caused in part to differences in heart rate. These results suggest that pulmonary venous diastolic flow velocity is influenced by the same factors that influence early left ventricular diastolic filling; pulmonary venous systolic flow velocity relates best to left atrial volume and atrial ejection fraction and does not relate to left ventricular ejection fraction. Future studies analyzing pulmonary venous flow velocity variables should include data on both peak velocities and velocity time integrals as well as left atrial size and function.