In patients with symptoms of heart failure after mitral valve replacement, identification of a stenosed prosthesis may be difficult. Twelve such patients were evaluated, presenting at a mean of 8.4 years after mitral valve replacement (four mechanical, eight porcine). Transvalvular pressure gradients were obtained using both indirect (pulmonary capillary wedge) and direct (transseptal catheterization) measurements of left atrial pressure. In all 12 patients, the diastolic gradient across the prosthetic valve was overestimated when pulmonary wedge rather than transseptal measurements were used. Calculated mitral valve prosthetic area was underestimated by the pulmonary wedge determinations. These findings may be caused by either the phase delay of the pulmonary wedge V wave relative to the transseptal V wave, resulting in a higher diastolic mean left atrial pressure, or the faulty wedge determinations in the setting of pulmonary hypertension, or both. In patients being considered for repeat mitral valve replacement because of prosthetic valve stenosis, transseptal catheterization allows for more accurate determination of prosthetic valve area and more accurately defines the need for repeat mitral valve surgery.