Randomized controlled trials have demonstrated that cardiac resynchronization therapy (CRT) can improve exercise capacity, left ventricular (LV) function, and survival in patients with advanced heart failure (HF) and wide QRS durations. Despite these promising results, up to one third of patients are nonresponders when selected on the basis of QRS duration. Observational studies have shown that the assessment of mechanical dyssynchrony can be used to identify CRT responders. However, the detection of mechanical dyssynchrony usually depends on time delay measurements, which are basically limited because residual contractility is not taken into account. Patients with mechanical dyssynchrony may not respond to CRT because myocardial segments are scarred and thus lack residual contractility that can be recruited by CRT. This phenomenon is particularly evident in patients who have scarred myocardium in the early or delayed activated segments.

To overcome these limitations, various indices of inefficient deformation, such as the systolic rebound stretch (SRS) in the septal wall (SRS _sept ), strain delay index (SDI), and internal stretch fraction, have been proposed for better prediction of response to CRT by taking the phase and strain amplitude into account rather than by simply quantifying the time delay in mechanical events. SRS _sept and SDI are dedicated indices for derivation and validation because of their relatively large populations and their excellent prediction power for CRT response and prognosis. In addition, both indices are obtained from longitudinal deformation, which is believed to be more feasible and reproducible than radial or circumferential deformation.

Previous studies have identified four clinical characteristics associated with a favorable response to CRT: female sex, left bundle branch block (LBBB), QRS duration ≥ 150 msec, and nonischemic etiology of HF. We hypothesized that CRT candidates with these favorable characteristics would show greater amounts of dyssynchrony-related inefficient deformation than those without and that the differences in inefficient deformation between patients with and without favorable characteristics, and between responders and nonresponders, could be revealed by SRS _sept or SDI assessment. In the present study, SRS _sept , SDI, and time delay indices were calculated in CRT candidates using longitudinal strain assessed by two-dimensional speckle-tracking echocardiography and compared between patients with and without favorable characteristics. In addition, the incremental value of baseline SRS _sept and SDI over clinical characteristics for predicting reverse remodeling was also evaluated.

Methods

Study Population

One hundred thirty-six consecutive patients with HF with sinus rhythm, LV ejection fractions (EFs) ≤ 35%, and QRS durations ≥ 120 msec were enrolled prospectively. Seventeen patients with technically unsatisfactory echocardiographic images for speckle-tracking longitudinal strain analysis were excluded, and thus, 119 patients were analyzed. These patients were divided into two groups on the basis of sex (female vs male), QRS morphology (LBBB vs non-LBBB), QRS duration (≥150 vs < 150 msec), and etiology of HF (ischemic vs nonischemic). Because of our national health care insurance policy, CRT was limited to patients who had LBBB, LV EFs ≤ 35%, and New York Heart Association class III or IV HF. Among the 78 patients with LBBB, 45 patients had received CRT. Two CRT patients did not complete 6-month follow-up, two patients are waiting for CRT, and eight patients refused CRT. Twenty-three patients with LBBB did not meet the criteria for CRT because of mild symptoms of HF. Response to CRT was evaluated in 43 patients because these patients had paired echocardiograms at baseline and 6 months after CRT. Reverse remodeling was defined as a decrease in LV end-systolic volume ≥ 15% at 6-month follow-up. The study protocol was approved by the local institutional review board, and informed consent was obtained from each participant.

Echocardiography

The study protocol included detailed transthoracic echocardiography performed using a commercially available system (Vivid 7; GE Vingmed Ultrasound AS, Horten, Norway). LV EFs were assessed in the apical two-chamber and four-chamber views using the biplane Simpson’s method. Two-dimensional longitudinal strain of the left ventricle was analyzed from the apical four-chamber, two-chamber, and long-axis views using commercial software (EchoPAC version 7.0; GE Vingmed Ultrasound AS). The echocardiographic images were recorded with a frame rate of 50 to 80 Hz. A single beat was analyzed for each time period, and values from three cardiac cycles were averaged to obtain each index. The software automatically tracked the image speckle and produced the longitudinal strain curves in six regional segments from each apical view, respectively. The timings of mitral valve closure and aortic valve closure were derived from Doppler flow recordings over the mitral and aortic valves. Deformation events were temporally aligned with these cardiac intervals through the electrocardiographic traces, with the onset of QRS taken as the zero reference. Systolic phase was defined as the period between mitral valve closure and aortic valve closure.

Longitudinal Dyssynchrony Analysis

Time to peak longitudinal strain was measured to derive the septal-to-lateral (SL) delay and the 12-segment (basal and mid LV) standard deviation of time to peak strain (ϵ _peak ). Within the systolic period, the deformation curve was separated into shortening (negative slope) and stretching (positive slope). SRS was defined as the cumulative amount of systolic stretching after prematurely terminated shortening. SRS was determined at the septum (SRS _sept ; Figure 1 A) or averaged over the left ventricle (SRS _LV ). SDI calculates the sum of the absolute differences between ϵ _peak and end-systolic strain across 16 myocardial segments on the basis of three apical views ( Figure 1 B). For a segment that exhibited positive strain or biphasic strain with a peak positive strain greater than the maximal absolute negative strain, the term (ϵ _peak −end-systolic strain) was entered as zero for the calculation of SDI. LV dyssynchrony was defined using the previously reported cutoff values (the values for predicting reverse remodeling are shown in parentheses): SL delay (130 msec), SRS _sept (4.7%), and SDI (25%).

Measurement of SRS _sept and SDI. For the sake of clarity, only representative curves of septal and lateral walls are displayed. (A) Within the systolic period, the deformation curve is classified as STC (red) , systolic prestretch (SPS) (green) , and SRS (blue) . SRS _sept is defined as the cumulative amount of systolic stretching after prematurely terminated shortening at the septum. (B) SDI is the sum of differences between ϵ _peak and end-systolic strain (ϵ _ES ) across 16 myocardial segments.

CRT Procedure

The pacing leads were positioned at the right ventricular apex or mid septum, at the right atrial appendage, and in the posterior or posterolateral branch of the coronary vein. The atrioventricular interval was optimized using the established method to ensure adequate LV filling. No adjustments were made to the interventricular interval before 6-month follow-up.

Statistical Analysis

Data are expressed as mean ± SD or as median (interquartile range) for continuous variables and as absolute frequencies and relative percentages for categorical variables. Unpaired t tests or Mann-Whitney tests were used to compare between patients. Receiver operating characteristics curves were generated to determine the predictive values of clinical characteristics alone (model 1) and clinical characteristics combined with SL delay, SDI, or SRS _sept (models 2, 3, and 4, respectively) with regard to LV reverse remodeling. The statistical significance of differences between the areas under the curves was tested using the method proposed by Hanley and McNeil. Intraobserver and interobserver variability were assessed in 45 randomly selected patients. For all tests, P values < .05 were considered statistically significant. Statistical analyses were performed using SPSS version 18.0 (SPSS, Inc., Chicago, IL).