Validity of Revised Doppler Echocardiographic Algorithms and Composite Clinical and Angiographic Data in Diagnosis of Diastolic Dysfunction

Abstract

Commonly used echocardiographic indices for grading diastolic function predicated on mitral inflow Doppler analysis have a poor diagnostic concordance and discriminatory value. Even when combined with other indices, significant overlap prevents a single group assignment for many subjects. We tested the relative validity of echocardiographic and clinical algorithms for grading diastolic function in patients undergoing cardiac catheterization. Patients (n = 115), had echocardiograms immediately prior to measuring left ventricular (LV) diastolic (pre-A, mean, end-diastolic) pressures. Diastolic function was classified into the traditional four stages, and into three stages using a new classification that obviates the pseudonormal class. Summative clinical and angiographic data were used in a standardized fashion to classify each patient according to the probability for abnormal diastolic function. Measured LV diastolic pressure in each patient was compared with expected diastolic pressures based on the clinical and echocardiographic classifications. The group means of the diastolic pressures were identical in patients stratified by four-stage or three-stage echocardiographic classifications, indicating that both classifications schemes are interchangeable. When severe diastolic dysfunction is diagnosed by the three-stage classification, 88% and 12%, respectively, were clinically classified as high and intermediate probability, and the mean LV pre-A pressures was >12 mmHg (P < 0.005). Conversely, the mean LV pre-A pressure in the clinical low probability or echocardiographic normal groups was <11 mmHg. Use of a standardized clinical algorithm to define the probability of diastolic function identifies patients with elevated LV filing pressure to the same extent as echocardiographic methods.