Aims Early diastolic mitral annular velocity (e′) obtained by tissue Doppler imaging (TDI) is widely used to evaluate left ventricular (LV) diastolic function based on the assumption that it reflects myocardial relaxation in the long-axis direction. In this study, we aimed to determine whether or not e′ truly reflects early diastolic longitudinal myocardial relaxation, and which is the most useful for evaluating LV diastolic function among e′ measured at the interventricular-septal annulus (IS-e′), that measured at the lateral annulus (LW-e′) or their mean value (M-e′).
Methods and results IS-e′, LW-e′, and M-e′ were measured using colour TDI in 15 patients with hypertrophic cardiomyopathy, 13 patients with hypertension, and 19 control subjects. Using two-dimensional speckle-tracking imaging, early diastolic myocardial strain rates (SRE) were measured for the IS (IS-SRE), LW (LW-SRE), and entire LV myocardium (G-SRE). IS-e′ was excellently correlated with IS-SRE (r= 0.90, P< 0.001); the correlation was better than that between LW-e′ and LW-SRE (r= 0.75, P< 0.001). IS-e′ and M-e′ were well correlated with G-SRE (r= 0.88, P< 0.001 and r= 0.86, P< 0.001, respectively) and with LV early diastolic flow propagation velocity (FPV) (r= 0.77, P< 0.001 and r= 0.78, P< 0.001, respectively). The correlations of LW-e′ to G-SRE (r= 0.80, P< 0.001) and FPV (r= 0.75, P< 0.001) did not reach this level.
Conclusion IS-e′ well reflected LV longitudinal myocardial relaxation and LV diastolic function, and was found to be more useful in evaluating LV diastolic function than LW-e′.
Tissue Doppler imaging
Early diastolic mitral annular velocity
Two-dimensional speckle-tracking imaging
Left ventricular longitudinal myocardial relaxation
Early diastolic mitral annular velocity (e′) derived from tissue Doppler imaging (TDI) is widely used to evaluate left ventricular (LV) diastolic function based on the assumption that it reflects the myocardial relaxation in the long-axis direction.1–4 It can be performed easily with excellent reproducibility,1–6 and many investigators have reported the usefulness of the ratio of early diastolic transmitral flow velocity to e′ (E/e′) in estimating LV filling pressure.1,5,7,8 In this way, the value of e′ measurement has been established through many clinical and physiological studies, but it remains still unclear whether or not the mitral annular velocity truly reflects the LV longitudinal myocardial relaxation. In addition, the measurement of e′ has been performed at the interventricular septal annulus (IS-e′)2,5,7 and/or lateral annulus (LW-e′)1,9 in the apical four-chamber view; their mean value (M-e′)8,10 has also been used. However, it is not clear which of IS-e′, LW-e′, or M-e′ is the best to evaluate LV diastolic function.
Two-dimensional speckle-tracking imaging (2DSTI) is a recently developed echocardiographic technique that automatically tracks myocardial motion throughout the entire cardiac cycle on the basis of pattern-matching techniques and provides myocardial strain and strain rate.11–13 This technique can provide an accurate evaluation of myocardial longitudinal contraction and relaxation without any angle dependency and the effect of apical translation, both of which are critical limitations of TDI. In this study, we aimed to determine whether the mitral annular velocity truly reflects the LV longitudinal myocardial relaxation and which e′ measurement, IS-e′, LW-e′, or M-e′, is the most useful for evaluating LV diastolic function.
The study subjects consisted of 15 patients with hypertrophic cardiomyopathy (HCM) (11 men and 4 women, 54.3 ± 14.5 years), 13 age-matched patients with hypertension (HT) (7 men and 6 women, 52.2 ± 12.5 years), and 19 age-matched normal control subjects (11 men and 8 women, 52.6 ± 9.2 years) in whom good quality of echocardiographic images could be obtained. HCM was defined as LV hypertrophy not explained by any cardiac or systemic abnormalities and with an interventricular septal (IS) thickness/posterior wall thickness ratio in end-diastole of more than 1.3 by two-dimensional echocardiography. HT was defined as a repeatedly elevated blood pressure (>140 mmHg at systole or 90 mmHg at diastole) or having anti-hypertensive medications with a history of HT. We excluded patients with obvious systolic dysfunction whose echocardiographic LV ejection fraction was <55%, apparent pulmonary HT, or cardiac rhythm disturbances such as atrial fibrillation or artificial pacing. Patients with congestive heart failure, diabetes mellitus, coronary heart disease with apparent LV asynergy, congenital anomaly, valvular heart disease, or other systemic diseases that cause LV hypertrophy were also excluded. Furthermore, we excluded subjects in whom an ultrasound beam could not be inserted correctly from the apex in the apical four-chamber view.
Medications for patients with HT were β-blocker in 1 patient, angiotensin-converting enzyme inhibitor or angiotensin II receptor blockade in 5, calcium antagonist in 1, α-blocker in 2, diuretics in 4, anti-arrhythmic drug in 1 and nitrate in 1; those for HCM patients were β-blocker in 10 patients, angiotensin-converting enzyme inhibitor or angiotensin II receptor blockade in 5, calcium antagonist in 3, diuretics in 2, anti-arrhythmic drug in 2, and nitrate in 1. The control group consisted of healthy volunteers without any clinical or echocardiographic evidence of cardiovascular disease.
This study was approved both by the Research Ethics Committee of Hokkaido University Hospital and by the Ethics Committee of Faculty of Health Sciences in Hokkaido University.
Basic echocardiographic measurements
Echocardiography was performed using a Vivid 7 ultrasonographic machine (GE Healthcare UK Ltd, Amersham Place, Little Chalfont, Buckinghamshire, England) with an M4S transducer. The LV end-diastolic dimension and left atrial end-systolic dimension were measured from a parasternal long-axis B-mode image. The thicknesses of the IS and LV posterior wall were measured in the parasternal short-axis image at the level of the chordate tendineae at end-diastole. The LV ejection fraction was calculated by Simpson's biplane method of discs according to the American Society of Echocardiography.14 Pulsed-Doppler echocardiography was performed to measure peak early and late diastolic transmitral flow velocities (E and A, respectively, cm/s) and the early diastolic wave deceleration time (DT, ms) of the transmitral flow and to calculate the ratio of early to late transmitral flow velocities (E/A). LV isovolumic relaxation time (IRT, ms) was measured using an apical continuous-wave Doppler recording, which clearly depicted both LV inflow and outflow. LV early diastolic flow propagation velocity (FPV, cm/s) was measured using apical colour M-mode Doppler imaging of the LV inflow along with the previously reported method.15,16
Measurement of mitral annular velocity by tissue Doppler imaging
An apical four-chamber image of the colour tissue Doppler technique was acquired at a frame rate of 69.8–147.7 frames/s. Tissue Doppler digital data were stored and analysed offline using an EchoPAC PC (GE Healthcare UK Ltd) (Figure 1). We measured the peak early diastolic and late diastolic velocity (e′ and a′, respectively, cm/s) at the mitral annuli of the IS and lateral wall (LW) sides and calculated the ratio of early to late diastolic annular velocities (e′/a′). The mean value of the two-site measurements (M-e′, M-a′, and M-e′/a′) was also calculated for each study subject.
Measurement of mitral annular velocity using colour tissue Doppler imaging. In the apical four-chamber view (left panel), the peak early diastolic and peak late diastolic mitral annular velocities (e′ and a′, respectively) were measured at the interventricular septal (IS) and lateral wall (LW) side annulus (right panel).
Measurement of strain rate by two-dimensional speckle-tracking imaging
Two-dimensional harmonic images (transmit/receive 1.9/4.0 MHz) in a four-chamber view were scanned with frame rates of 49–69/s. Digital data were stored and analysed offline using an EchoPAC PC (Figure 2). The LV endocardial surface of the end-systolic frame was traced manually, and the speckle-tracking width was modified so as to cover the whole LV wall thickness to obtain a so-called ‘global’ strain rate curve. From that curve, the peak early diastolic and late diastolic longitudinal strain rates (G-SRE and G-SRA, respectively, S−1) were measured. Then, the basal IS region of interest (ROI) was extended towards the apex, and the strain rate curve of the whole IS was obtained to measure the peak early diastolic and late diastolic IS strain rates (IS-SRE and IS-SRA, respectively, S−1). In a similar manner, a strain rate curve for the whole LW was obtained, and the two peak strain rates were measured (LW-SRE and LW-SRA, S−1). In addition, the ratio of the early to late diastolic peak was calculated each for the global, interventricular, and lateral strain rates (G-SRE/SRA, IS-SRE/SRA, and LW-SRE/SRA). Success or failure of the tracking was determined chiefly based on visual inspection of the tracking motion image, referring to the results of the automatic judgement function of the software, and manual readjustments were made if needed. Subjects for whom tracking was not completely successful were excluded from the study. All strain rates were expressed as absolute values.
Measurement of myocardial strain rates using two-dimensional speckle-tracking imaging. Left ventricular global strain rates (upper panels): a region of interest for speckle-tracking analysis was set on the entire left ventricular myocardium in the apical four-chamber view (left panel). The peak early diastolic and late diastolic myocardial global strain rates (G-SRE and G-SRA, respectively) were measured (right panel). The interventricular septal strain rates (lower panels): a region of interest was set on the interventricular septum from the base to the apex (left panel), and the peak early diastolic and late diastolic interventricular septal strain rates (IS-SRE and IS-SRA, respectively) were measured (right panels).
Statistical analysis was performed using standard statistical software (Stat View version 5.0, SAS Institute, Cary, NC, USA). All numerical data were represented as means ± standard deviation. Differences among the three groups were tested by one-way analysis of variance at first, and, when a significant difference was detected, each difference between the individual two groups was tested using a post hoc test. The relationship between two parameters was examined using linear correlation and regression analysis. The significance of the difference between two correlation coefficients was calculated using Fisher's r-to-z transformation. For all statistical tests, a P–value of <0.05 was used to indicate significance.
Clinical background and echocardiographic parameters
There were no significant differences among the control, HT, and HCM groups in age, sex, heart rate, body surface area, LV end-diastolic dimension, LV ejection fraction, E, A, and E/A (Table 1). Systolic blood pressure and diastolic blood pressure were significantly greater in the HT group compared with the control or HCM group. IS thickness, LV posterior wall thickness, and IRT were significantly greater, and FPV was significantly reduced in both the HT and HCM groups compared with the control group. The left atrial dimension and DT were greater in the HCM group compared with the control group. The left atrial dimension and IS thickness were significantly greater, and FPV and LV posterior wall thickness were significantly lower in the HCM group than in the HT group. IS-e′, LW-e′, M-e′, IS-SRE, LW-SRE, and G-SRE were significantly reduced both in the HT group and in the HCM group compared with the control group. IS-SRE and G-SRE were significantly reduced in the HCM group compared with the HT group.
Clinical characteristics and echocardiographic features
Control (n= 19)
HT (n= 13)
HCM (n= 15)
52.6 ± 9.2
52.2 ± 12.5
54.3 ± 14.5
Body surface area (cm2)
1.69 ± 0.16
1.62 ± 0.23
1.69 ± 0.20
Heart rate (bpm)
62.3 ± 8.7
63.7 ± 8.6
59.6 ± 8.8
Systolic blood pressure (mmHg)
118 ± 11
132 ± 18**
115 ± 14††
Diastolic blood pressure (mmHg)
73 ± 8
82 ± 13*
67 ± 11†††
LV end-diastolic dimension (mm)
48.1 ± 2.9
46.2 ± 2.7
46.6 ± 5.8
LA end-systolic dimension (mm)
34.4 ± 4.2
37.6 ± 4.4
43.7 ± 7.0***,††
Interventricular septal thickness (mm)
9.2 ± 1.4
13.3 ± 1.3**
20.4 ± 5.5***,†††
LV posterior wall thickness (mm)
8.6 ± 1.3
11.4 ± 0.5***
10.2 ± 1.1***,††
LV ejection fraction (%)
67.3 ± 4.8
66.5 ± 4.2
70.6 ± 8.5
77.9 ± 17.2
68.5 ± 14.8
69.3 ± 20.0
65.2 ± 12.3
71.6 ± 13.7
73.0 ± 23.2
1.26 ± 0.45
0.99 ± 0.28
1.09 ± 0.66
193 ± 34
224 ± 34
261 ± 82***
77.6 ± 9.4
92.3 ± 11.4**
90.0 ± 20.6*
55.5 ± 8.7
45.5 ± 8.6**
37.5 ± 12.6***,†
7.05 ± 1.67
4.39 ± 1.67***
3.48 ± 2.32***
9.06 ± 1.84
5.88 ± 1.62***
5.21 ± 3.18***
8.06 ± 1.62
5.13 ± 1.55***
4.34 ± 2.69***
1.37 ± 0.31
1.00 ± 0.28**
0.62 ± 0.41***,††
1.71 ± 0.36
1.12 ± 0.38***
0.89 ± 0.47***
1.52 ± 0.35
1.03 ± 0.32***
0.70 ± 0.39***,†
HT, hypertension; HCM, hypertrophic cardiomyopathy; LV, left ventricle; LA, left atrium; E, peak early diastolic transmitral flow velocity; A, peak late diastolic transmitral flow velocity; E/A, ratio of E to A; DT, early diastolic wave deceleration time of the transmitral flow; FPV, LV early diastolic flow propagation velocity; IRT, LV isovolumic relaxation time; e′, peak early diastolic mitral annular velocity; IS, interventricular septum; LW, lateral wall; M-e′, the mean value of IS-e′ and LW-e′; SRE, peak early diastolic strain rate.
*P< 0.05, **P< 0.01, ***P< 0.001 vs. control.†P< 0.05, ††P< 0.01, †††P< 0.001 vs. HT-LVH.
Relationship between mitral annular velocities and strain rates of the corresponding wall
IS-e′ was significantly and well correlated with IS-SRE (r= 0.90, P< 0.001) (Figure 3). LW-e′ was significantly, but less strongly, correlated with LW-SRE (r= 0.75, P< 0.001). The correlation coefficient between IS-e′ and IS-SRE was significantly higher than that between LW-e′ and LW-SRE (P< 0.05). Significant and good correlation was also observed both between IS-e′/a′ and IS-SRE/SRA (r= 0.88, P< 0.001) and between LW-e′/a′ and LW-SRE/SRA (r= 0.84, P< 0.001).
Correlations between diastolic mitral annular velocity and myocardial strain rate of the corresponding wall. e′, peak early diastolic mitral annular velocity; e′/a′, the ratio of the e′ to the peak late diastolic mitral annular velocity; SRE, peak early diastolic strain rate; SRE/SRA, the ratio of SRE to the peak late diastolic strain rate; HCM, patients with hypertrophic cardiomyopathy; HT, patients with hypertension; IS, interventricular septum; LW, left ventricular lateral wall; N, control subjects.
Relationship between mitral annular velocities and left ventricular global strain rate
IS-e′ and M-e′ were significantly and well correlated with G-SRE (r= 0.88, P< 0.001 and r= 0.86, P< 0.001, respectively) (Figure 4). IS-e′/a′ and M-e′/a′ were also well correlated with G-SRE/SRA (r= 0.89, P< 0.001 and r= 0.90, P< 0.001, respectively). The correlations between LW-e′ and G-SRE (r= 0.80, P< 0.001) and between LW-e′/a′ and G-SRE/ SRA (r= 0.82, P< 0.001) were also good but a little weaker.
Correlations between the early diastolic myocardial ‘global’ strain rate (G-SRE) and early diastolic mitral annular velocities in the interventricular septum (IS), lateral wall (LW), and their mean (M). e′, peak early diastolic mitral annular velocity; HCM, patients with hypertrophic cardiomyopathy; HT, patients with hypertension; N, control subjects.
Relationship between mitral annular velocities and left ventricular global function
Both IS-e′ and M-e′ were well correlated with FPV (r= 0.77, P< 0.001 and r= 0.78, P< 0.001, respectively) (Figure 5) and significantly correlated with transmitral E (r= 0.48, P< 0.001 and r= 0.47, P< 0.001, respectively), DT (r= 0.49, P< 0.001 and r= 0.45, P< 0.01, respectively), and IRT (r= 0.55, P< 0.001, and r= 0.54, P< 0.001, respectively). IS-e′/a′ and M-e′/a′ were significantly and well correlated with the transmitral E/A ratio (r= 0.74, P< 0.001 and r= 0.79, P< 0.001, respectively). LW-e′ was significantly correlated with E (r= 0.44, P= 0.020), DT (r= 0.39, P< 0.05), IRT (r= 0.50, P< 0.001), and FPV (r= 0.75, P< 0.001), and LW-e′/a′ was also significantly correlated with E/A (r= 0.75, P< 0.001). However, those correlation coefficients were tended to be slightly weaker.
Correlations between left ventricular early diastolic flow propagation velocity (FPV) and early diastolic mitral annular velocities in the interventricular septum (IS), lateral wall (LW), and their mean (M). e′, peak early diastolic mitral annular velocity; HCM, patients with hypertrophic cardiomyopathy; HT, patients with hypertension; N, control subjects.
In the present study, IS-e′ was excellently correlated with IS-SRE, and LW-e′ was fairly well correlated with LW-SRE, indicating that e′ well reflects LV longitudinal myocardial relaxation. The use of e′ as a parameter of LV diastolic function seems to be based on the assumption that it reflects LV longitudinal myocardial relaxation, but this assumption has not been confirmed. Recently, Opdahl et al. reported that a good correlation was observed between early diastolic mitral annular velocity measured using TDI and myocardial relaxation assessed by sonomicrometry in mongrel dogs.17 However, no sufficient evidence for human has been provided in previous studies. The present study validated the assumption in human and provided better understanding of the property of e′ as an index for LV diastolic function.
In the recommendation for the evaluation of LV diastolic function issued by the American Society of Echocardiography, e′ plays an important role for the management of patients with heart failure.18 However, there has been no established consensus on the measurement site, that is, whether we should measure e′ at the IS annulus, lateral side annulus, or both. In the present study, the correlation between IS-e′ and IS-SRE was significantly better than that between LW-e′ and LW-SRE. In addition, IS-e′ and M-e′ were excellently correlated with G-SRE, and these correlations were slightly better than that between LW-e′ and G-SRE. These results suggest that IS-e′ is more useful in assessing LV diastolic function than LW-e′.
There have been several reports comparing IS-e′ and LW-e′. Peverill et al.19 reported that LW-e′ was greater than IS-e′ and the correlation between the two parameters was inadequate. Park et al.20 also indicated a discrepancy between the two measurements and concluded that IS-e′ tended to overestimate the LV diastolic dysfunction. However, they did not show any clear evidence to indicate the superiority of LW-e′ over IS-e′. In the present study, the correlation between IS-e′ and IS-SRE was significantly better than that between LW-e′ and LW-SRE. On the other hand, the correlation between LW-e′/a′ and LW-SRE/SRA was clearly better than that between LW-e′ and LW-SRE. These results indicate that the inadequate correlation between e′ and SRE in the LW is attributable to the angle dependency of TDI. Thus, the IS-e′ can more accurately reflect the LV longitudinal myocardial relaxation. In this study, M-e′ reflected LV diastolic function with a similar accuracy as IS-e′. It is estimated that M-e′ may be useful in the evaluation of patients with regional myocardial dysfunction such as myocardial ischaemia and regional LV hypertrophy, and in those with right ventricular disease and/or increased right ventricular pressure, which may affect the IS-e′ value.
In this study, we employed SRE measured by using 2DSTI as the standard of longitudinal myocardial relaxation. The accuracy of this method has been reported by several investigators based on comparisons with sonomicrometry and with tagging magnetic resonance imaging.13,21–23 The e′ measured using TDI is widely used as an index for global diastolic function rather than the regional one, based on the hypothesis that mitral annular motion reflects longitudinal shortening and lengthening of the myocardium from the base to the apex. Thus, we measured the so-called ‘semi-global’ SRE with ROIs in the whole IS and the whole LW for the comparisons with IS-e′ and LW-e′, respectively.
Recently, there have been several reports showing the superiority of E/SRE over E/e′ in the estimation of the LV filling pressure. Dokainish et al.24 reported that in patients with an LV ejection fraction of ≥50% and an E/e′ value from 8 to 15, E/SRE by 2DSTI was a better predictor of LV filling pressure than E/e′ by TDI. In addition, Ng et al.25 reported that E/E2D (E2D: LV myocardial velocity derived from 2DSTI) was more useful in identifying elevated LV end-diastolic pressure than E/e′. Although these results should be sufficiently appreciated, e′ measured using TDI has inevitable merits in clinical practice because it can be very easily measured with good reproducibility and there is a great amount of accumulated clinical evidence supporting its use. In addition, 2DSTI requires high-end ultrasound devices, and its analysis requires good image quality of the whole LV. Furthermore, because the 2DSTI's data analysis algorithm varies according to the ultrasound manufacturer, the criterion value remains unclear. At the present time, 2DSTI is not being administered in all routine examinations.
There are some limitations to be acknowledged in the present study. First, we employed colour TDI instead of pulsed TDI, which is more widely used in the clinical settings. To evaluate the correspondence between TDI and 2DSTI, we thought it better to perform these two measurements offline from the same image over using pulsed TDI, which requires some additional procedures during examination. Because the correspondence in the mitral annular velocity between colour TDI and pulsed TDI was reported to be excellent,26 the results of this study may be readily applicable to measurements using pulsed TDI. Secondly, e′ and SRE were studied only in the IS and LW in the four-chamber view. Analyses using more imaging planes might provide a better measuring site of e′ or a better averaged value. However, we consider that the analysis of the four-chamber view is sufficient from a practical view; this plane is very commonly used, and multiplane analysis may limit the ease of this measurement, which is the greatest merit of TDI. Thirdly, the population of this study was too small to perform subgroup analysis. Further study may be needed to evaluate differences among diseases.
In conclusion, early mitral annular velocity (e′), derived using TDI, especially measured at the septal side annulus, well reflects the LV longitudinal myocardial relaxation. The e′ measured at the septal annulus should be used for the clinical evaluation of LV diastolic function rather than the value at the lateral annulus.
Conflict of interest: none declared.
This study was partly supported by Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan, No. 19500412.
. Clinical utility of Doppler echocardiography and tissue Doppler imaging in the estimation of left ventricular filling pressures: a comparative simultaneous Doppler-catheterization study. Circulation 2000;102:1788-94.
. Optimal noninvasive assessment of left ventricular filling pressures: a comparison of tissue Doppler echocardiography and B-type natriuretic peptide in patients with pulmonary artery catheters. Circulation 2004;109:2432-9.
. Recommendations for chamber quantification: a report from the American Society of Echocardiography's Guidelines and Standards Committee and the Chamber Quantification Writing Group, developed in conjunction with the European Association of Echocardiography, a branch of the European Society of Cardiology. J Am Soc Echocardiogr 2005;18:1440-63.
. A new approach for evaluation of left ventricular diastolic function: spatial and temporal analysis of left ventricular filling flow propagation by color M-mode Doppler echocardiography. J Am Coll Cardiol 1996;27:365-71.
. Usefulness of early diastolic flow propagation velocity measured by color M-mode Doppler technique for the assessment of left ventricular diastolic function in patients with hypertrophic cardiomyopathy. J Am Soc Echocardiogr 2000;13:801-8.
. Determinants of left ventricular early-diastolic lengthening velocity: independent contributions from left ventricular relaxation, restoring forces, and lengthening load. Circulation 2009;119:2578-86.
. Regional myocardial long-axis strain and strain rate measured by different tissue Doppler and speckle tracking echocardiography methods: a comparison with tagged magnetic resonance imaging. Eur J Echocardiogr 2009;10:229-37.