It is generally accepted that Doppler echocardiography cannot provide unequivocal evidence of diastolic dysfunction in HFNEF. The E/A ratio, IVRT, DT, and pulmonary vein Doppler16 characterizing flow across the mitral valve do not allow direct measurement of LV relaxation, stiffness, or filling pressure.15,17 Several authors have demonstrated that conventional Doppler is accurate in patients with a reduced EF but not in those with normal EF.16,35 We also found only a weak correlation between IVRT and in our study population, in agreement with others.36,17 Furthermore, a short DT indicates increased LA pressure in patients with systolic37 or pseudonormal and restrictive diastolic heart failure.15 In contrast, patients with a mild diastolic dysfunction have a prolonged DT and therefore reduced K. Thus, with regard to their biphasic response to increasing diastolic dysfunction, the interpretation of E/A or DT in diagnosing diastolic dysfunction is rather complicated. Although the mitral inflow parameters DT and K correlated with b (dP/dV), simple analysis of E/A or DT was limited in at least our study population, which was patients with impaired mitral flow. This limitation is further underscored by our finding that these parameters were not significantly related to ß, known to be a relatively load-independent parameter. If we had used the mitral flow Doppler as the only technique for detecting diastolic dysfunction in HFNEF, 70% of patients would have been correctly identified. Although all 5 patients with pseudonormal mitral flow pattern had been identified correctly in our study, indicating that the mitral Doppler is a helpful diagnostic tool in cases of more severe diastolic dysfunction, we conclude that it is of only limited value in the diagnosis of early diastolic dysfunction in HFNEF. We observed that the duration of atrial reverse flow was significantly prolonged, an early sign of disturbed mitral inflow resulting from increased LV stiffness. However, the limited diagnostic accuracy and signal quality limited the practical use of pulmonary vein Doppler in our study. A combined analysis based on E/A and IVRT, DT, or Ar–A duration improved the accuracy of the mitral flow Doppler method, but only to a moderate degree. In summary, filling pattern analysis from mitral flow Doppler measurement alone is found to be more complicated and limited to detecting early diastolic dysfunction in patients with HFNEF.
Tissue Doppler Imaging
TDI proved to be more accurate than conventional Doppler for detecting impaired diastolic function in patients with HFNEF.29,36 In general, we found that the lateral annular velocities were more closely related to the LV relaxation and compliance indexes as determined by PV-loop analysis than the septal annular velocities (Table 3). Thus, only the lateral velocities are taken into consideration in the following discussion. With regard to impaired LV relaxation, we confirmed its relation to the early diastolic mitral annular velocity (E’)18 and TE-E’.30 However, the latter did not correlate with the filling pressure, as previously suggested.30 We found that the TDI indexes E’lat and E’/A’lat correlated more closely with LV stiffness than any conventional echocardiography index. Similarly, the dimensionless E/E’ index, introduced recently as an echocardiographic measure of LA pressure and LV filling,2,16,35,38,39 showed the best correlation with indexes of diastolic parameters obtained by PV-loop measurements. In our study, patients with HFNEF and E/E’lat >8 had a significantly increased LV stiffness (Figure 1). Both E/E’lat >8 and E’/A’lat <1 detected HFNEF patients with diastolic abnormalities equally well, but E’/A’lat showed lower sensitivity, yielding more false-negative results than E/E’lat. Because we did not perform PV-loop and echocardiographic investigations simultaneously, however, we found only rather moderate correlations in our small study population. Nevertheless, from a clinical point of view, the key question is whether the echocardiographic method used allows reliable detection of the correct diagnosis. In contrast to Doppler echocardiography, TDI detected diastolic dysfunction in 81% (35 of 43) and the E/E’lat index in 86% (37 of 43) of our patients with HFNEF. Three additional patients with HFNEF were identified by adding E’/A’ to E/E’lat, raising the detection rate to 93% (40 of 43). In contrast, the additional application of conventional Doppler indexes did not considerably improve the diagnostic accuracy.
Some recent studies40–42 reported reduced regional systolic peak velocities in patients with HFNEF and impaired systolic reserve,43 suggesting that systolic function also is impaired. Our study has not confirmed this finding. In addition, our invasive catheter measurements showed that global systolic function and contractility of the patients with HFNEF were not impaired under basal condition, in agreement with others.1,19,44,45 Compensatory capacities and/or systolic reserve in our relatively young study population may have contributed to a limited difference in systolic parameters. However, further studies under stress conditions are needed to further clarify the role of systolic function in patients with HFNEF.
In summary, in clinically stable patients at rest presenting with reduced exercise capacity in whom the diagnosis of diastolic dysfunction was proven by conductance catheter analysis, single indexes of conventional Doppler echocardiography were insufficient or inferior compared with TDI parameters in detecting the correct diagnosis. Although the diagnostic accuracy improved after several indexes of the mitral and pulmonary venous flow analysis were added, we do not recommend their use as isolated method for investigating diastolic function, which is in agreement with the latest consensus statement of the Heart failure and Echocardiography Association of the European Society of Cardiology.46 In contrast to flow Doppler, TDI parameter showed better linear correlation with diastolic parameters and provided a simple means of diagnosing diastolic dysfunction. Accordingly, TDI was a more reliable technique to identify early disturbances of both LV relaxation and stiffness. However, although the LV filling index E/E’lat showed a similar sensitivity but higher specificity than E’/A’lat in detecting diastolic dysfunction, we recommend the use of E/E’lat in both clinical diagnostic routine and scientific studies to investigate diastolic function in patients with HFNEF.