Background
Left ventricular (LV) twist mechanics are a promising, sensitive tool for assessing pathophysiologic changes in patients with systolic heart failure. Although LV twist is known to be load dependent in healthy volunteers, this has not been examined in patients with “long-standing” dilated cardiomyopathy (DCM). The aim of this study was to determine whether LV twist remains load dependent in the setting of long-standing, nonischemic DCM.
Methods
Thirty-four patients with DCM with baseline LV ejection fractions (LVEFs) < 40% and 13 subjects with preserved LVEFs (≥50%) were enrolled. After baseline measurements, pneumatic compression of the lower extremities (Pcom) was used to increase LV afterload. Subsequently, sublingual nitroglycerin (SL-NG) was administered to modify preload. Conventional echocardiographic parameters, LV end-systolic wall stress, net LV twist angle, and apex-to-base-rotation delay (ABRD) were assessed under each condition.
Results
In patients with DCM, although LV end-systolic wall stress significantly increased under Pcom (196.9 ± 64.9 g/m 2 at baseline vs 231.8 ± 78.9 g/m 2 under Pcom, P < .017) and decreased after SL-NG application (231.8 ± 78.9 g/m 2 under Pcom vs 197.4 ± 67.4 g/m 2 after SL-NG, P < .017), net LV twist angle and ABRD showed no significant changes depending on LV loading condition (for LV twist, 7.63 ± 4.47° at baseline vs 7.03 ± 4.13° under Pcom vs 7.35 ± 4.36° after SL-NG, P = 0.65; for ABRD, 16.56 ± 13.81% at baseline vs 17.19 ± 14.81% under Pcom vs 15.95 ± 13.27% after SL-NG, P = .53). Careful examination of individual patient data revealed that LV twist was load independent when patients had LV twist < 12°. ABRD was also found to be load independent, but only in patients with LVEFs < 34%. In contrast, LV twist and ABRD were load dependent in patients with preserved LVEFs.
Conclusions
LV twist and its component, ABRD, had relatively load insensitive properties in patients with long-standing DCM and can be used in future clinical trials as load-independent indexes of LV dyssynchrony.
Left ventricular (LV) dyssynchronous movement is not infrequent in patients with nonischemic dilated cardiomyopathy (DCM) and is closely associated with LV systolic function impairment and a dismal long-term prognosis. Restoration of LV synchronous contraction by cardiac resynchronization therapy (CRT) has shown improvement in quality of life, functional status, and morbidity and mortality in patients with drug-refractory heart failure. However, although promising overall, about 30% of patients do not favorably respond to this sophisticated and expensive treatment, suggesting that current selection criteria need to be improved. To meet this requirement, continued efforts have been made by using a variety of echocardiographic LV dyssynchrony indexes (LVdys) to identify CRT responders more effectively before device implantation. Thus far, most mechanical dyssynchrony analyses have been based on a variety of echocardiographic techniques, which largely use “one” direction of LV movement (i.e., longitudinal or radial). This choice of orientation is determined primarily by operator convenience regarding the accessibility of the echocardiographic window, and thus these methods are not considered comprehensive enough for LVdys assessment.
The advent of speckle-tracking echocardiography (STE) allows LV twist mechanics to be investigated in various populations, and LV twist is increasingly being viewed as a comprehensive measure of LV mechanics. Because LV twist is defined as the net difference at isochronal time points between the basal and apical rotation angles, its magnitude is inherently dependent on how synchronously the LV apex rotates in relation to the LV base, and it has been quantitatively assessed and described as apex-to-base rotation delay. LV twist is also strictly dependent on the electromechanical activation sequence, as has been demonstrated by different pacing models and by measurements in patients with preexcitation before and after ablation. Thus, attempts to expand the use of LV twist and apex-to-base rotation delay to the CRT field have been recently made.
Heart failure is considered a “dynamic” condition, because LV loading status can be altered rapidly by a variety of medications that are frequently used to reduce patient discomfort. For example, vasodilators such as nitrates and diuretics, which are commonly used to reduce LV load, can significantly affect LV volume and thus loading status. Prior studies have indicated that LV loading status exerts a significant influence on echocardiographic LVdys obtained by tissue velocity imaging or STE-derived radial strain analysis. To be more reliably used, any suggested indicator should be clinically tested regarding whether it is load dependent or not, before it is adopted as an indicator of LVdys in a multicenter clinical trial (such as the Predictors of Response to Cardiac Resynchronization Therapy [PROSPECT] study). Furthermore, recent efforts to use LVdys as a surrogate marker of prognosis in various populations have increased the importance of identifying a relatively load independent marker of LVdys.
Therefore, we sought to determine whether LV twist and apex-to-base rotation delay are significantly altered by modifications of LV loading status in patients with chronic, nonischemic DCM and histories of heart failure.
Methods
Study Population
Thirty-nine consecutive patients diagnosed with nonischemic DCM were recruited. All patients had undergone invasive coronary angiography and were confirmed to be free of significant coronary artery disease (defined as luminal stenosis > 50% in any epicardial coronary artery). All patients were free of any congenital heart disease, valvular heart disease (more than mild in degree), or arrhythmias, including atrial fibrillation. Five of the 39 patients were excluded from the analyses because of poor echocardiographic image quality (two patients), poor speckle-tracking (two patients), and LV enlargement precluding acquisition of the whole left ventricle in one screen at an acceptable frame rate (one patient). Ultimately, 34 patients remained in the final analyses, constituting the subjects of the present study. For comparison, 15 subjects with preserved LV ejection fraction (LVEFs), defined as ≥50%, served as controls, of whom two were excluded because of poor echocardiographic image quality for STE in one and withdrawal of consent from one. Computed tomographic coronary angiography revealed no evidence of epicardial coronary stenosis in these 15 control subjects. All patients and control subjects were fully informed about the study, and written informed consent was obtained from all participants before enrollment. The study protocol was approved by the institutional review board of our hospital.
Sample Size Calculation for Patients with DCM
The number of the study population was estimated beforehand on the basis of our previous echocardiographic experiences. The study sample size was calculated assuming an α error of 0.017 (after Bonferroni’s correction) and a β error of 0.10, with statistical power of 90%. We assumed that a change of 1.5° in net LV twist angle could be found, with a standard deviation of 2.25°. With this assumption, a total of 34 patients were necessary. When we supposed that 15% of patients were likely to be excluded because of difficulty in the application of STE and other unexpected reasons, the estimated cohort size required was 39 for the present study.
Study Protocol
Baseline transthoracic echocardiograms were obtained using commercially available equipment (Vivid 7; GE Medical Systems, Milwaukee, WI) in the left lateral decubitus position. During routine echocardiographic examinations, we decided whether patients were suitable for inclusion on the basis of two-dimensional echocardiographic image quality. Patients with technically inappropriate image quality were excluded at this stage.
Blood pressure (BP) and heart rate (HR) were measured initially after a 20-min resting period. Routine standard echocardiographic examinations included measurements of LV end-diastolic wall thicknesses, LV end-diastolic volume (LVEDV) and LV end-systolic volume (LVESV) and LVEF using the modified biplane Simpson’s method, pulsed-wave Doppler examination of mitral inflow, and pulsed-wave Doppler tissue imaging at the medial mitral annulus. From the mitral inflow Doppler signals, early transmitral inflow velocity, late transmitral inflow velocity, and the deceleration time of early transmitral inflow velocity were obtained with the sample volume placed between the tips of both mitral leaflets. Standard M-mode and two-dimensional images were acquired during end-expiratory breath hold and stored in cine loop format from three consecutive beats. After acquisition of the baseline transthoracic echocardiographic images, pneumatic trousers without the bladder for compression of the lower abdomen were put on the patient. At the beginning of the study, a specially designed compressor inflated the pneumatic trousers to a pressure of 100 mm Hg on both lower extremities, and this pressure was maintained throughout the examination. Echocardiographic images for STE began to be scanned 3 min after pneumatic compression of the lower extremities (Pcom). BP and HR were again measured under Pcom. Pcom could not be applied to three patients, because of previous amputation of one leg for the surgical treatment of osteosarcoma in two and the development of sudden dyspnea about 1 min after Pcom application in one. After finishing the procurement of echocardiographic images under Pcom, pneumatic trousers were deflated and removed. Five minutes after the removal of pneumatic trousers, one tablet (0.6 mg) of sublingual nitroglycerin (SL-NG) was administered. SL-NG administration was considered effective when BP dropped below baseline value 3 min after SL-NG administration. In case of high BP even after one tablet of SL-NG, another tablet was administered. After ensuring that target BP had been achieved, acquisition of echocardiographic images was commenced. SL-NG could not be administered to one patient in whom baseline systolic BP was only 90 mm Hg.
Because systemic vascular resistance was reported to be an unreliable index of LV afterload, we calculated LV end-systolic wall stress (LVESWS) as a reliable representative index of LV afterload, in an attempt to quantify the effects of either Pcom or SL-NG administration on LV myocardium. LVESWS was calculated using the following formula :
LVESWS = P es ( D es H es ( 1 + H es / D es ) ) 0.34 ,
Speckle-Tracking Echocardiographic Protocol
After standard routine echocardiography, scanning of apical and basal short-axis planes was performed using an M3S probe without a dual-focusing tool to quantify basal and apical LV rotation. Frame rate (range, 70–100 frames/sec) and probe frequency (range, 1.7–2.0 MHz) were adjusted during end-expiratory breath hold for optimal image acquisition. Sector width and image depth were optimized to maintain an adequate frame rate without losing two-dimensional image quality. Basal and apical levels were defined as the point of the tips of mitral valve leaflets and the point just proximal to the level with LV luminal obliteration at the end-systolic period, respectively. Care was taken not to obtain oblique LV short-axis images and to obtain short-axis images with the most circular geometry possible. Three consecutive heartbeats were digitally stored in cine loop format and were analyzed offline to obtain reliable values. Image analysis was performed by one independent cardiologist, who was blinded to the status of image acquisition, using a customized dedicated software package (EchoPAC version 108.1.12 for PC; GE Medical Systems, Milwaukee, WI).
Image Analysis
A region of interest was manually traced on the endocardial cavity interface using a point-and-click approach on an end-systolic single frame. Then, an automated tracking system followed the endocardium throughout the cardiac cycle. The validity of tracking was verified by a reliability parameter offered by the system and was again visually confirmed. Further adjustment of the region of interest was performed, if necessary, to ensure that all of the myocardial regions were included.
Curves of averaged basal and apical rotation in six myocardial segments and their net difference at each time point were directly generated by EchoPAC and transported to an Excel 2007 worksheet (Microsoft Corporation, Redmond, WA) for further analyses.
To assess synchronous rotation between the LV base and apex, the time interval from Q-wave onset on the electrocardiogram to peak rotation (the Q-peak rotation interval) was obtained from basal and apical short-axis images ( Figure 1 ). The difference in Q-peak rotation interval between the LV base and apex is referred to as the apex-to-base rotation delay. To avoid errors derived from interindividual HR differences, time points were normalized to the duration of the whole RR interval, and thus Q-peak rotation interval at the LV base or apex is expressed as a percentage.
