We investigated the relationship between left atrial (LA) and pulmonary capillary wedge pressure (PCW) in order to define the clinical settings in which PCW may be used to approximate LA pressure and to determine the cause and significance of difference between LA and PCW pressures. Nineteen patients who at cardiac catheterization had LA, PCW, and left ventricular (LV) pressures recorded, had mitral valve gradients and areas determined. Mean PCW and LA pressures correlated well (r = 0.94). Phasic PCW consistently overestimated the MV gradient and underestimated the MV area compared to LA pressure, (8 +/- 4 versus 4 +/- 3 mm Hg, p less than 0.001, 1.3 +/- 0.3 versus 1.6 +/- 0.3 cm2, p less than 0.005, respectively). Three patients had abnormal MV prosthesis function assessed by PCW pressure but not by LA pressure. Diastolic MV gradients between PCW and LV were caused or increased by a slowed y descent in the PCW tracing. Using the PCW pressure may falsely elevate MV gradients and falsely reduce MV areas and lead to incorrect clinical action.