Individuals with a preserved LV ejection fraction (LVEF) and no overt symptoms of heart failure according to the Framingham criteria [4] were enrolled in the present study when echocardiographic studies were performed and were found to be appropriate for analysis by TDI and 2D-STI. Subjects with atrial fibrillation or a bundle branch block as shown by electrocardiography were excluded, as were those with a < 50% LVEF or valvular heart disease. Ultimately, 105 subjects (58 males and 47 females) were included. Among the subjects were four cases of hypertension and two cases of diabetes mellitus.

In addition, the subjects were classified into three groups according to their E/E' value (i.e., the degree of diastolic impairment, Table 1). Two cut-off values were used (< 9 for the controls and > 15 for those subjects with an elevated E/E' ratio) based on previous data [5-7].

Two-dimensional and Doppler echocardiographic examinations were performed using a 3.5-MHz transducer (Vivid 7, Vingmed-General Electric, Horten, Norway). The LVEF was determined in the apical 4- and 2-chamber views using a modified version of Simpson's formula. A preserved LVEF was defined as an ejection fraction ≥ 50%. The myocardial wall thickness and chamber dimensions were measured in the M-mode; the wall-motion score index was also examined to identify subjects with regional wall motion abnormalities.

As described previously [8], pulsed-wave Doppler echocardiography was performed in the apical 4-chamber view to determine the mitral inflow profile, including the E- and A-wave velocities, E-deceleration time (DT), and E/A ratio. To obtain the peak systolic (Sm) and peak early diastolic mitral annulus velocity (Em), TDI was performed using a 1 mm to 2 mm sample volume on the septal side of the mitral annulus in the apical 4-chamber view.

At least three consecutive beats were stored, and the images were analyzed offline with the aid of a customized software package (EchoPac version 5.1.1, Vingmed-General Electric). To identify LV intraventricular dyssynchrony, the peak myocardial systolic and peak myocardial early diastolic velocities of the tissue Doppler signal were measured using the onset of the QRS complex as a reference point, wherein the basal segments were scanned just above the mitral annulus while the middle segments were scanned at the papillary muscle level. LV intraventricular systolic dyssynchrony (SDS_LV) was defined as the difference between the shortest and longest times of the peak myocardial systolic velocity among twelve segments. Likewise, LV intraventricular diastolic dyssynchrony (DDS_LV) was defined as the difference between the shortest and longest times of the peak myocardial early diastolic velocity among twelve segments.

At three levels in the apical short-axis view, speckle tracking routine grayscale images were acquired as described previously [9,10]. An end-systolic circular region of interest (ROI) was traced through the endocardial border using a point-and-click technique to adjust the tracking in all six segments. In turn, a larger concentric circle was generated automatically, which had to be adjusted manually near the epicardial border. Shifts in the location of the acoustic markers for each ROI between frames, which represented tissue movement, provided spatial and temporal data for calculating the velocity vectors wherein temporal alterations in the stable speckle patterns were identified as moving farther apart or closer together. A series of regional strain vectors were calculated based on changes in the length/initial length ratio. Myocardial thickening was assigned a positive value, whereas myocardial thinning was assigned a negative value. The software package used automatically created radial strain curves for each segment, which were used to assess radial dyssynchrony. The software package was also used to calculate LV rotation and peak LV torsion using the same grayscale images with a simple change in the analysis mode that allowed frame-by-frame calculation of the rotational angle of the stable speckles. Counterclockwise rotation independent of the LV apex was expressed as a positive value, whereas clockwise rotation was expressed as a negative value.

Radial dyssynchrony was defined as the time difference between the first and last peak values of the radial strain curves at each level (RDS_base, RDS_mid, and RDS_apex for radial dyssynchrony in the basal, mid, and apical levels in the apical short-axis view, respectively). Peak LV torsion was defined as the difference in degree between the peak apical and basal rotational degrees in the apical short-axis view.

Continuous data are presented as the mean ± standard deviation (SD) unless otherwise stated. Comparisons of the continuous data among the groups were conducted by ANOVA with post-hoc analysis and chi-square tests for comparing categorical data. The relationship between the continuous variables was examined by regression analysis; selection of the most powerful factor was performed using a forward stepwise multivariate analysis (SPSS version 12, SPSS Inc., Chicago, IL, USA). A p value ≤ 0.05 was used to define a significant result.

The mean age in the three groups increased according to the degree of increase in the E/E' ratio for the controls compared to that in the elevated E/E' ratio group (p < 0.001, Table 1). However, those subjects with an intermediate or elevated E/E' ratio were similar to the controls in terms of gender distribution, body mass index (BMI), LVEF, and QRS duration.

The LV mass and left atrial dimension (LAD) in the elevated E/E' ratio group were significantly greater than in the controls (p = 0.001 and 0.007, respectively). As shown in Table 1, the mean mitral inflow and annular velocity values differed among the groups.

The mean SDS_LV in the elevated E/E' ratio group was significantly greater than that in the control group, whereas the mean DDS_LV was not significantly different (Table 2). By univariate analysis, significant relationships were identified between the SDS_LV and DDS_LV and increases in the E/E' ratio (r = 0.405 and p < 0.001 vs. r = 0.216 and p = 0.045, respectively; Fig. 2).

The mean RDS_base value in the three groups gradually increased with unchanging significance according to the degree of increase in the E/E' ratio for the control and elevated E/E' ratio groups (p = 0.005), whereas the mean RDS_mid and RDS_apex values increased with a change in significance (Table 2). Univariate analysis revealed significant relationships between RDS_base and RDS_apex and the increase in E/E' (r = 0.298 and p = 0.002 vs. r = 0.196 and p = 0.045, respectively; Fig. 2).

The mean values for the basal and apical rotations did not differ significantly among the three groups, as was the case for LV torsion; no relationship was detected between LV torsion and the E/E' ratio.

Using the cut-off values proposed by Yu et al. [3], for SDS_LV (100 ms) and DDS_LV (113 ms), four cases of systolic dyssynchrony were identified (6.3%, one in the intermediate E/E' ratio group and three in the elevated E/E' ratio group); none of the subjects had diastolic dyssynchrony or co-existing systolic and diastolic dyssynchrony. Significant radial dyssynchrony was defined as > 130 ms by Suffoletto et al. [9]. According to this criterion, radial dyssynchrony was observed in six cases in the present study (9.4%, two in the intermediate E/E' ratio group and four in the elevated E/E' ratio group).

As advanced age can result in various physiologic changes in the myocardium, it is reasonable that the above results should be adjusted for age. After adjustment for age using a stepwise multivariate analysis, SDS_LV and DDS_LV were found to be significantly related to increases in the E/E' ratio (r = 0.405 and p < 0.001 vs. r = 0.216 and p = 0.045, respectively). RDS_base and RDS_apex were also significantly related to increases in the E/E' ratio (r = 0.298 and p = 0.002 vs. r = 0.196 and p = 0.045, respectively). A significant relationship was identified between LV torsion and age (r = 0.210 and p = 0.034) rather than E/E'.

The number of subjects was relatively small; a greater number of subjects are needed to produce a more accurate representation of the prevalence of mechanical dyssynchrony in this kind of clinical condition. The E/E' ratio was used to non-invasively estimate the LV filling pressure or PCWP because cardiac catheterization seemed too invasive for use in our subjects, who did not present with overt symptoms of heart failure.

In conclusion, systolic and diastolic dyssynchrony in subjects with isolated diastolic dysfunction but without overt symptoms of heart failure was not as common as in patients with diastolic heart failure. However, the systolic and diastolic intraventricular time delay increased with respect to increases in the E/E' ratio, an indicator of diastolic dysfunction.