Doppler echocardiographic evaluation of left ventricular end-diastolic pressure in patients with coronary artery disease

Abstract

Pulmonary venous flow velocity recordings have been found to be useful in complementing the information obtained from the mitral flow velocity and improving the assessment of left ventricular diastolic pressures. This study was undertaken to evaluate the accuracy of mitral flow and pulmonary venous flow variables, recorded by transthoracic Doppler echocardiography, in estimating left ventricular end-diastolic pressure (LVEDP) in 101 consecutive patients with coronary artery disease undergoing diagnostic left-sided heart catheterization. Patients were assigned to three groups according to the values of LVEDP (group 1, < or = 12 mm Hg; group 2, between 13 and 19 mm Hg; and group 3, > or = 20 mm Hg). LVEDP correlated most strongly with systolic fraction of pulmonary venous flow (r = -0.76), isovolumic relaxation time (r = -0.76), E/A ratio (r = 0.74), deceleration time of early mitral flow (r = -0.74), and mitral A wave duration/pulmonary venous A wave duration (AD/PVAD) ratio (r = -0.70) (p < 0.01 for each correlation). Discriminant analysis demonstrated that deceleration time, AD/PVAD ratio, and isovolumic relaxation time were major determinants of LVEDP, with 87.1% of patients correctly assigned to study groups; 97% of patients of group 1 and 95% of patients of group 3 were identified, whereas the accuracy in identifying the patients of group 2 was lower (41%). Deceleration times of 140 msec or less and AD/PVAD ratios of 0.9 or less were the best cutoff points in predicting an LVEDP of 20 mm Hg or greater. Multiple linear regression analysis demonstrated that the combination of mitral flow and pulmonary venous flow velocity variables provided a better estimation of LVEDP compared with that obtained from mitral flow velocity recordings alone (r = 0.88 versus 0.79; F test, 20.6). We conclude that combined analysis of mitral flow and pulmonary venous flow velocity provides, in patients with coronary artery disease, a noninvasive estimation of LVEDP with an accurate prediction of pressures of 12 mm Hg or less and 20 mm Hg or greater and less accurate prediction of intermediate values.