Author + information
- Received September 11, 2001
- Revision received April 17, 2002
- Accepted May 20, 2002
- Published online August 21, 2002.
- Peter Søgaard, MD, DMSc*,* (, )
- Henrik Egeblad, MD, DMSc, FESC*,
- W.Yong Kim, MD, PhD*,
- Henrik K. Jensen, MD, PhD*,
- Anders K. Pedersen, MD, DMSc*,
- Bent Ø. Kristensen, MD, DMSc* and
- Peter T. Mortensen, MD*
- ↵*Reprint requests and correspondence:
Dr. Peter Søgaard, Department of Cardiology, Aarhus University Hospital, DK-8200 Aarhus North, Denmark.
Objectives We sought to evaluate the long-term impact of cardiac resynchronization therapy (CRT) on left ventricular (LV) performance and remodeling using three-dimensional echocardiography and tissue Doppler imaging (TDI).
Background Three-dimensional echocardiography and TDI allow rapid and accurate evaluation of LV volumes and performance.
Methods Twenty-five consecutive patients with severe heart failure and bundle branch block who underwent biventricular pacemaker implantation were included. Before and after implantation of the pacemaker, three-dimensional echocardiography and TDI were performed. These examinations were repeated at outpatient visits every six months.
Results Five patients (20%) died during one-year follow-up. In the remaining 20 patients, significant reductions in LV end-diastolic volume and LV end-systolic volume of 9.6 ± 14% and 16.5 ± 15%, respectively (p < 0.01), could be demonstrated during long-term follow-up. Accordingly, LV ejection fraction increased by 21.7 ± 18% (p < 0.01). According to a newly developed TDI technique—tissue tracking—all regional myocardial segments improved their longitudinal systolic shortening (p < 0.01). The extent of the LV base displaying delayed longitudinal contraction, as detected by TDI before pacemaker implantation, predicted long-term efficacy of CRT. The QRS duration failed to predict resynchronization efficacy.
Conclusions Cardiac resynchronization significantly improved LV function and reversed LV remodeling during long-term follow-up. Patients likely to benefit from CRT can be identified by TDI before implantation of a biventricular pacemaker.
According to the findings at cardiac catheterization, cardiac resynchronization therapy (CRT) by biventricular pacing acutely improves left ventricular (LV) function in patients with severe heart failure and bundle branch block (1,2). Reports on the use of noninvasive techniques for the short- and long-term evaluation of CRT efficacy are limited thus far (3–5). However, radioisotope techniques have a low temporal and spatial resolution, making them less suitable to quantify regional LV synchrony. Recently introduced echocardiographic techniques, such as tissue Doppler imaging (TDI) and three-dimensional echocardiography, enable accurate quantification of regional and global LV function and volumes (6–10). From digitally recorded TDI loops of one or more heart beats containing velocity data from the entire myocardium, two new TDI modalities can be derived: tissue tracking (TT) and strain rate (SR) analysis (11). Tissue tracking visualizes the longitudinal motion amplitude in each myocardial segment during systole (Figs. 1 and 2), ⇓ and SR analysis can be used to determine whether this motion represents contraction or is merely passive (Fig. 3). We found TT and SR useful for the evaluation of regional myocardial pathophysiology before and after cardiac resynchronization and, therefore included TT and SR analysis in our follow-up study.
Using TDI and three-dimensional echocardiography (5,6), we have recently reported on the immediate benefit of CRT in terms of myocardial synchrony and performance, resulting in the early reduction in operating LV volumes (5). We also quantitated the degree of LV mechanical asynchrony by the extent of myocardium at the LV base displaying delayed longitudinal contraction (DLC) (i.e., contraction after closure of the aortic valve). Using multivariate analysis, we found that the number of segments with DLC (at the LV base) was the only co-variate to predict short-term CRT efficacy (5).
At present, the long-term efficacy of CRT in terms of LV performance is poorly elucidated. Because of the observed short-term effect on LV function and volumes, we hypothesized that long-term therapy would further improve LV systolic performance and reverse LV remodeling. These aspects, as well as potential predictors of long-term CRT benefit, are addressed in the present report.
The patient group comprised 25 consecutive patients included in a recently published study on the short-term effect of CRT (5). Despite receiving stable (>6 months), contemporary medication for heart failure, all patients were in New York Heart Association (NYHA) functional class III or IV. All were in sinus rhythm and had bundle branch block with a QRS width above 120 ms.
Pacemaker leads were inserted by the transvenous route. Three pacing leads were implanted: one in the right atrium, another in the high septum or outflow tract of the right ventricle (RV) and one in the coronary sinus (Medtronic 2187 or 10512, Minneapolis, Minnesota); this last lead was positioned on the LV free wall through a coronary sinus tributary. The RV and LV pacing leads were positioned to ensure that the sensed local intracardiac activation during sinus rhythm was measured as early for the RV lead and as late for the LV lead as possible, according to the QRS complex on the surface electrocardiogram. The pacing leads were connected to a dual-chamber biventricular pacemaker (Medtronic, InSync), with nonprogrammable simultaneous pacing of the two ventricular leads. The biventricular pacemaker was programmed in DDD mode. Adjustment of the atrioventricular delay was performed to ensure the longest possible atrioventricular filling time evaluated from pulsed Doppler analysis of transmitral LV filling.
We have previously reported in detail on the three-dimensional echocardiographic and TDI protocol (5–7). We have also established our coefficient of variation of TDI measures of velocities and asynchrony to be 4% to 6% (5), comparable to values reported by other investigators (9). Three-dimensional echocardiographic reproducibility was previously reported in depth (6,7). In brief, three-dimensional echocardiography and TDI were performed before implantation of the pacemaker, immediately after implantation and every six months after implantation. One observer, who had no knowledge of the clinical data of the patients, performed all three-dimensional echocardiographic and TDI analyses.
Three-dimensional echocardiography was performed during end-expiratory apnea within one breath hold, using electrocardiogram-triggered co-axial rotation from the apical window with 30° intervals between the scanning planes. The resulting six digital cine loops were transferred to a computer for off-line analysis (Echo-Pac Software, GE-Vingmed Ultrasound, Horten, Norway). In each view, the endocardial borders were drawn manually in end diastole and end systole for LV volume measurement.
Tissue Doppler imaging was acquired as digital loops during one heart beat in the apical four-chamber, two-chamber and long-axis views (5). Analysis of regional peak contraction velocities toward the apex and the duration of contraction (Fig. 1, showing a normal subject) was performed in each of 16 segments of the LV model of the American Society of Echocardiography (12).
A TT image in systole was derived from the TDI loop in each of the apical views (Fig. 1, lower left panel; Fig. 2A, lower left panel). If DLC was present, another TT image was recorded in diastole to visualize the extent of myocardium with DLC (Fig. 2B, lower left panel). The number of segments displaying DLC at the base of the LV (5) was recorded and expressed as the percentage of the total circumference of the LV base. Myocardial segments showing motion toward the apex were explored with SR analysis. Motion toward the apex after closure of the aortic valve was only registered as DLC if negative SR documented that the motion reflected true shortening. Figure 3 shows the SR analysis.
Figure 4 shows the development in myocardial performance, displayed as TT images obtained during long-term follow-up in one patient. Again using the 16-segment LV model, the average motion amplitude toward the apex in systole was estimated for each segment, and averages were calculated for the six segments at the LV base, six mid-ventricular segments and four apical segments. The global systolic contraction amplitude (GSCA) index was calculated as the average amplitude of all 16 segments.
The paired Student t test was used for comparison of three-dimensional echocardiographic and TDI measurements at baseline and during CRT. The QRS width and DLC extent were correlated to long-term improvement in LV performance by univariate linear regression analysis. In addition, linear regression analysis was used for the evaluation of changes in GSCA and percent LV ejection fraction (LVEF) during follow-up. A value of p < 0.05 was considered statistically significant. Data are presented as the mean value ± SD.
Five (20%) of the 25 patients died during follow-up, before the first outpatient examination. All of the patients who died had ischemic heart disease. The QRS duration (179 ± 23 ms) was comparable to that of the remaining study group. However, these patients displayed only limited DLC at the LV base (20 ± 12% of the LV base circumference), and their immediate response to CRT, in terms of improvement in LVEF, was minor (20.1 ± 6% to 21.9 ± 7%). The remaining 20 patients were followed for a mean period of 12.6 months (range 6 to 18 months). The pre-implantation demographic and clinical variables of these patients are given in Table 1.
At the end of follow-up, significant reductions in the LV end-diastolic volume and LV end-systolic volume of 9.6 ± 14% and 16.9 ± 15%, respectively, as compared with hospital discharge, were demonstrated (p < 0.01). This was associated with a significant increase in LVEF of 21.7 ± 18% (p < 0.01) (Table 2).
By means of TT, it was documented that the average motion amplitude toward the apex during systole improved significantly throughout the entire LV (Table 2). Figure 4 demonstrates the use of TT to visualize changes in motion amplitude toward the apex in systole. The calculated improvement in GSCA from hospital discharge to the latest follow-up visit for each patient was significantly correlated to the improvement in LVEF (r = 0.83, p < 0.01). The percentage of the LV base circumference displaying DLC was reduced from 18.7 ± 7% at discharge to 8.1 ± 8% at one year (p < 0.05).
There was improvement in NYHA functional class, from 3.5 ± 0.4 at baseline to 2.2 ± 0.5 at the completion of follow-up (p < 0.01). No significant change in medication, such as angiotensin-converting enzyme inhibitors, diuretics, or beta-blockers, was performed during follow-up.
Predictors of long-term efficacy
The percentage of the LV base circumference exhibiting DLC, as measured at baseline, predicted long-term improvement in LVEF and GSCA (t = 3.1, p < 0.01 and t = 3.3, p < 0.01, respectively) (Fig. 5). The QRS width did not predict long-term CRT improvement in LVEF and GSCA (t = 1.0, p = NS and t = 1.7, p = NS, respectively).
Cardiac resynchronization therapy appears to be a promising therapeutic option in patients with bundle branch block and severe heart failure. In this study, we documented long-term improvement in LV systolic performance and a parallel reduction in LV volumes (i.e., reversed LV remodeling). The latter effect may be of particular interest, as it might be associated with an improved prognosis in patients with ischemic cardiomyopathy, where LV dilation has emerged as a powerful predictor of mortality (13). In patients with idiopathic dilated cardiomyopathy, however, the prognostic determinants may seem more complex, but the presence of a recruitable LV contractile reserve seems to predict a more favorable outcome (14). However, the potential impact of CRT on mortality awaits larger randomized trials.
Potential mechanisms underlying CRT benefit
Several mechanisms could underlie the benefit of CRT on LV function and remodeling. If LV wall stress fails to decline during systole due to ailing LV systolic function, overall LV wall stress will be persistently elevated, causing progressive remodeling of the LV with continuous dilation. Therefore, by improving global systolic LV function, long-term CRT may reduce LV wall stress, a reduction that may, in turn, contribute to reversed LV remodeling. In addition, DLC (i.e., contraction during diastole representing mechanical LV asynchrony) not only reflects a loss of systolic force, it also causes inappropriate energy consumption and increased LV wall stress in early diastole. This would presumably interfere with regional myocardial perfusion that could, in turn, further reduce LV systolic performance and aggravate remodeling. Moreover, because CRT reduces the number of segments displaying DLC, pacing not only improves systolic function by recruiting this contractile reserve, it also facilitates diastolic filling of the LV. Interestingly, the improvement in LV systolic performance during CRT is not accompanied by an increase in myocardial oxygen consumption (15), presumably because the reversion of LV remodeling, as documented in the present study, reduces myocardial oxygen demand. Finally, the majority of patients with end-stage heart failure and bundle branch block have secondary mitral valve regurgitation. Thus, reduced mitral regurgitation during CRT (16) could also contribute to the observed reversed LV remodeling.
During short-term CRT, both LV volumes and systolic function improve acutely (5,16). In our study of the immediate-term effect of CRT, the improvement in systolic function appeared to be confined to the base of the LV, whereas the mid-ventricular and apical areas of the ventricle did not seem to be resynchronized (5). We therefore proposed the hypothesis that multiple LV leads might be required to obtain resynchronization of more extended regions of the LV (5). However, during long-term CRT, both the mid-ventricular and the more distal part of the LV eventually showed improved systolic performance (Table 2). The absolute improvement in LVEF from baseline of the 20 patients was 11% (23.8% to 34.9%) before implantation to one-year follow-up. The long-term effect of CRT after hospital discharge was responsible for more than 50% of the total increase in LVEF (Tables 1 and 2). Thus far, two ventricular leads appear to be sufficient, although resynchronization of the mid and apical parts of the LV seems to be a delayed phenomenon.
Selection of responders to CRT
Our observation of improvement in LV performance corresponds with the recently published Multisite Stimulation in Cardiomyopathies trial (MUSIC) showing a benefit of NYHA functional class and physical endurance (17). In that study, patient inclusion was based on a QRS duration above 150 ms, and previous studies have emphasized the predictive value of the QRS duration in relation to an improvement in LV function during CRT (18). Our data, however, do not support the use of the QRS duration as the primary selection criterion. Indeed, it appears that the detection of mechanical asynchrony, as reflected by DLC, is a superior predictor of CRT efficacy. This is further supported by the observations in patients who died during follow-up. These patients had QRS widths comparable to those of survivors, but they had less extensive mechanical asynchrony (i.e., areas with DLC that could potentially be resynchronized into systole). Therefore, it seems that pre-implantation TDI screening could improve on the selection of candidates prone to benefit from CRT, a hypothesis that is currently being tested in a substudy in the Cardiac Resynchronization in Heart Failure study (CARE-HF) (19).
Echocardiography in the evaluation of CRT
In the present study, we applied two new TDI modalities. One advantage of TT is the “at-a-glance” visualization of the extent of myocardium with DLC (Fig. 2B), although more conventional TDI may also be useful (5). Tissue tracking also facilitates the demonstration of improvement in regional and global longitudinal contraction (Fig. 4), and it was shown that the improvement in overall longitudinal contraction reflected the improvement in LVEF. In patients with bundle branch block, the LV commonly exhibits awkward rocking throughout the heart cycle, and it is often impossible to decide whether motion reflects fiber shortening or merely passive stretching. In this setting, SR analysis (Fig. 3) is a valuable tool to document when DLC represents contraction and thus a potential contractile “reserve” that could be resynchronized into systole.
The LVEF measurements by three-dimensional echocardiography show excellent reproducibility, and according to our most recent studies, the patient group required in a two-dimensional echocardiographic study should be four times larger, equivalent to 80 patients, to achieve identical statistical power (6).
Cardiac resynchronization therapy immediately improves LV function. However, according to three-dimensional echocardiography, the gain is even greater on a long-term (one-year) basis, when the improvement includes reversion of LV remodeling. Tissue Doppler imaging seems to add important pathophysiologic information on the degree, location and reduction of mechanical LV asynchrony during continuous biventricular pacing and on regional and global LV function. Finally, TDI may result in the better selection of potential responders to CRT.
- cardiac resynchronization therapy
- delayed longitudinal contraction
- global systolic contraction amplitude
- left ventricle/ventricular
- left ventricular ejection fraction
- New York Heart Association
- right ventricle/ventricular
- strain rate
- tissue Doppler imaging
- tissue tracking
- Received September 11, 2001.
- Revision received April 17, 2002.
- Accepted May 20, 2002.
- American College of Cardiology Foundation
- Leclercq C.,
- Cazeau S.,
- Le Breton H.,
- et al.
- Kass D.A.,
- Chen C.H.,
- Curry C.,
- et al.
- Kerwin W.F.,
- Botvinick E.H.,
- O’Connell J.W.,
- et al.
- Kim W.Y.,
- Sogaard P.Kristensen B.Ø,
- Egeblad H.
- Hatle L.,
- Sutherland G.R.
- Brodin L.Å.,
- Gabella M.,
- van der Linden J.,
- et al.
- Schiller N.B.,
- Shah P.M.,
- Crawford M.,
- et al.
- Norris R.M.,
- White H.D.,
- Cross D.B.,
- et al.
- Ramani T.M.,
- Longo M.D.,
- Cadariu A.R.,
- et al.
- Nelson G.S.,
- Berger R.D.,
- Fetics B.J.,
- et al.
- Kim W.Y.,
- Søgaard P.,
- Mortensen P.T.,
- et al.