Detection of Diastolic Dysfunction

Abstract & Commentary

By Michael H. Crawford, MD, Professor of Medicine, and Chief of Clinical Cardiology, at the University of California, San Francisco. Dr. Crawford is on the speaker's bureau for Pfizer.

Source: Kasner M, et al. Utility of Doppler Echocardiography and Tissue Doppler Imaging in the Estimation of Diastolic Functions in Heart Failure With Normal Ejection Fraction; A comparative Doppler-Conductance Catheterization Study. Circulation 2007; 116: 637-647.

Heart failure with normal left ventricular (LV) systolic function is thought to be due to diastolic dysfunction in most cases, but confirming LV diastolic dysfunction by echo Doppler techniques is often challenging. Thus, Kasner and associates studies 43 patients with clinical heart failure who had normal LV ejection fraction (EF) and confirmed LV diastolic dysfunction by conductance LV catheter measurements of the isovolumetric relaxation time constant; LV end diastolic pressure; LV stiffness constant and LV stiffness. They were compared to 12 controls with chest pain, normal LVEF and no symptoms or signs of heart failure. Patients with atrial fibrillation, valve disease, significant coronary disease or lung disease were excluded. The objective of the study was to see which echo Doppler parameter best predicted diastolic dysfunction measured invasively.

Results: All patients studied had normal LV volumes and EF, but the heart failure patients had increased LV mass index and left atrial volume as compared to the controls. In 70% of the heart failure patients some abnormality of standard mitral inflow velocity measures of diastolic function were found. Similar results were obtained when the pulmonary vein flow was examined, but only 72% had an adequate signal for analysis. Tissue Doppler indices at the lateral mitral annulus were abnormal in 81% of the heart failure patients and E/E' was >8 in 86%. Also, LV end-diastolic pressure was best predicted by E/E'. The authors concluded that E/E' was the best echo Doppler parameter for the detection of diastolic dysfunction in heart failure patients with normal EF.


The non-invasive diagnosis of LV diastolic dysfunction by standard mitral inflow velocity parameters is complicated. There are numerous influences on these values that need to be considered. The biggest problem is their lack of specificity for diagnosing mild diastolic dysfunction in older individuals. Abnormal relaxation (A>E) becomes more prevalent with age and is not necessarily indicative of diastolic dysfunction.

Assessment of pulmonary venous flow parameters did not help the diagnostic accuracy for at least 2 reasons: Young people have "abnormal" patterns (D>S); and adequate pulmonary venous signals can only be obtained in 70% of individuals. Tissue Doppler lateral mitral annulus velocity parameters alone increase sensitivity for detecting diastolic function abnormalities about 10% over standard mitral Doppler velocity and pulmonary vein flow measures, but specificity is an issue. Combining the standard Doppler E wave peak velocity with the tissue Doppler equivalent E' results in the dimensionless ratio E/E', which has the highest sensitivity for detecting diastolic dysfunction of any parameter (86%) and a specificity of 100% in this study at a cut-off of 8. Also, an E/E' >8 was highly correlated with an increased LV end-diastolic pressure. Combining other parameters with E/E' did not increase accuracy in this study.

Given these results should we rely on E/E' alone to make the diagnosis of diastolic dysfunction in patients with heart failure and normal systolic function? Unfortunately, no one echo Doppler measure is perfect. E/E' should be the keystone of the diagnosis with the caveat that the negative predictive value is less than the positive predictive value. So in those with a high E/E', LV end-diastolic pressure is highly likely to be elevated, but perhaps not always due to diastolic dysfunction. Extreme fluid overload, constrictive pericarditis and other unusual conditions could raise LV diastolic pressures independent of the diastolic properties of the LV. When E/E' is normal some cases of mild diastolic dysfunction could be missed. In these cases evaluation of the other echo Doppler parameters may be useful; especially pulmonary vein flow in an older individual, if it can be obtained. It is always satisfying when all the echo Doppler and tissue Doppler parameters point to the diagnosis, but not infrequently this doesn't happen. In such cases more reliance should be placed on E/E' with the recognition that some mild cases may be missed. Of course, this means that tissue Doppler imaging needs to be obtained in anyone suspected of having heart failure.