OBJECTIVE Pulsed Doppler measurement of pulmonary venous flow (PVF) in the left superior pulmonary vein has been suggested as a noninvasive method to evaluate pulmonary capillary wedge pressure (PCWP). In previous studies, PVF was measured at end-expiration, and it is unknown to what extent PVF is affected by the respiratory cycle. It is hypothesized that phasic variations of PVF during mechanical ventilation may be used to estimate PCWP. METHODS Prospective clinical study. METHODS Tertiary care university hospital. METHODS Thirty patients undergoing elective cardiac surgery. METHODS At multiple intervals during the surgery, the PVF was measured with transesophageal pulsed Doppler echocardiography, and measurements of PCWP and airway pressure were simultaneously obtained. RESULTS Components of PVF evaluated were the systolic (X), diastolic (Y), and atrial (Z) waves with their velocity-time integrals (VTI). The systolic fraction (SF = VTI X/[VTI X + VTI Y]) and respiratory variations of each component of PVF were determined and compared with PCWP. There was a greater respiratory variation of the X wave (X expiratory-X inspiratory/X expiratory) in patients with PCWP < 18 mmHg than in patients with PCWP > or = 18 mmHg (0.19 +/- 0.19 v 0.14 +/- 0.13, respectively, p < 0.01). PVF components measured at end-expiration that related best with PCWP were the X/Y peak velocities (r = -0.53), VTI X/VTI Y ratio (r = -0.42), and the SF (r = -0.49). When measured during end-inspiration, the relation of the X/Y ratio, VTI X/VTI Y, and SF with the PCWP were r = -0.54, r = -0.41, and r = -0.50, respectively. CONCLUSIONS It has been documented that PVF velocity is influenced by the respiratory cycle during mechanical ventilation in patients undergoing cardiac surgery, and the magnitude of this variation is influenced by PCWP. However, it is not actually possible to predict PCWP accurately using these findings. Further studies are needed in which preload is varied acutely to confirm the usefulness of the results.