Author + information
- Received March 10, 2003
- Revision received June 24, 2003
- Accepted October 21, 2003
- Published online April 7, 2004.
- ↵*Reprint requests and correspondence:
Dr. Mandeep R. Mehra, Ochsner Clinic Foundation, Cardiology Department, 1514 Jefferson Highway, New Orleans, Louisiana 70121, USA.
The device era in heart failure has been heralded by successes in the realm of pacemakers, implantable defibrillators, and ventricular assist devices. In particular, the concept of cardiac resynchronization therapy, which seeks to optimize ventricular contractility by decreasing areas of focal dyssynchrony, is gaining wide acceptance. Recent trials of cardiac resynchronization therapy have suggested that this treatment modality yields benefits that are reflected in improved functional capacity, reversal of ventricular modeling, and decreased hospitalizations. Cardiac resynchronization device therapy exerts a substantial placebo effect, with evidence of improved functional capacity and quality-of-life parameters in almost half of those in the control group, probably as a result of device implantation. Furthermore, analysis of the different trials suggests heterogeneity of response (differences in magnitude of observed benefit between trials presumably enrolling similar heart-failure populations) and a large non-responder rate (no improvement in functional capacity and well-being). The appropriate approach to resynchronization must include much more than simple characteristics of device implantation. We must detect the presence and precise location of mechanical dyssynchrony and be able to find the technical location and place the pacing leads in the appropriate position. Finally, we must be able to show evidence for sustained improvement in ventricular dyssynchrony. Thus, the current approach to resynchronization represents a “best guess” approach to achieving resynchronization by observation of surrogate responses. Continued investigation to determine optimal approaches for achieving a beneficial clinical response is essential to ensure that cardiac resynchronization therapy is offered to those most likely to benefit.
Validation of therapeutic strategies in heart failure (HF) is proceeding at a rapid pace but with divergent outcomes. On one hand, the neurohormonal model has faced a major obstacle, with accumulating evidence suggesting that additive pharmacologic treatments are unlikely to yield incremental benefits beyond those achieved by antagonism of the renin-angiotensin-aldosterone and adrenergic systems (1). On the other hand, trials examining device therapy in HF that address electrical and mechanical aberrations by the application of pacemakers, implantable defibrillators, and ventricular assist devices are demonstrating incremental success (2). In particular, the concept of cardiac resynchronization therapy (CRT), which seeks to optimize ventricular contractility by decreasing areas of focal dyssynchrony, is gaining wide clinical acceptance (3). Although the scientific evidence supporting this avenue of therapy is compelling, translation of this approach to the clinical realm requires some caution. We wish to review the pitfalls of currently available data, emphasize the complexity involved in using electrical conduction defects as a surrogate for mechanical dyssnchrony, and discuss technical aspects, including pacing site selection and programming of optimal settings.
Do the clinical trials conclusively support the usefulness of CRT?
Several recent clinical trials have provided support for the usefulness of cardiac resynchronization therapy using biventricular pacing (4–7). Recent randomized clinical trials of cardiac resynchronization therapy have suggested that application of this treatment modality in severe systolic HF despite optimal drug therapy yields benefits that result in improved functional capacity, reversal of ventricular remodeling, and decreased hospitalizations (4–7). Indeed, a recent meta-analysis of these trials has even suggested decreased deaths from progressive HF as a consequence of cardiac resynchronization (8). More recently, the COMPANION (Comparison of medical therapy, pacing, and defibrillation in HF) trial results were made available (9). This trial enrolled patients with New York Heart Association (NYHA) functional class III or IV symptoms of HF despite maximized medical therapy (left ventricular [LV] ejection fraction ≤35%, LV end-diastolic diameter ≥6 cm). Other inclusion criteria included a QRS duration >120 ms and a PR interval >150 ms. The trial had three treatment arms: one of five patients was to receive optimal pharmacologic therapy, two of five were to receive optimal pharmacologic therapy plus biventricular pacing, and the remaining two of the five were to receive biventricular pacing, plus backup implantable cardioverter-defibrillator therapy. In contrast to previous trials, this study was powered to evaluate a primary end point of combined all-cause mortality and hospitalization. Peak oxygen consumption was the other (functional) primary end point. Secondary end points included all-cause mortality, cardiac mortality, exercise performance, NYHA functional class, quality of life, and the 6-min walk test distance. Data were analyzed using an intention-to-treat approach. A total of 1,520 patients were randomized, and 1,080 patients were implanted with a CRT pacer or defibrillator (CRT/D). A total of 118 patients failed initial implant (88% implant success for CRT and 92% for CRT/D). Left ventricular lead dislodgment was seen in 2% and 2.5% in the CRT and CRT/D groups, respectively. As compared to patients treated with medical therapy only, there was a statistically significant event rate reduction in the primary combined end point of total hospitalization and total mortality at one year in the CRT and CRT/D groups (odds ratio 0.82, p = 0.05 and 0.81, p = 0.015, respectively), as well as in the combined end point of hospitalization for CHF and death (odds ratio 0.64 and 0.60, respectively, p = 0.05). Mortality at 1 year decreased nonsignificantly by 24% (p = 0.121, ns) in the CRT group, and significantly by 43% (p = 0.002) in the CRT/D group.
Although these lines of evidence are encouraging, several limitations of this database deserve mention. First, placement of the device exerts a substantial placebo effect, with evidence of improved functional capacity and quality-of-life parameters in those randomized to the control group, probably as a result of device implantation. Second, most of these investigations randomized patients only after initial successful device implantation, thereby limiting the usefulness of the final results through a lack of a “true” intent-to-treat study design. Finally, an analysis of the benefits demonstrated among the different trials suggests remarkable heterogeneity of response (differences in magnitude of observed benefit between trials presumably enrolling similar HF populations) and a large non-responder rate (no improvement in functional capacity and well-being). Although the current clinical trial evidence definitely supports the “proof of concept” that CRT is of benefit, translating this therapeutic approach to general practice might be fraught with a number of technical obstacles that are unlikely to allow accurate replication of observed within-trial effects to the population at large.
Implications of the placebo response to CRT
Randomized controlled trials that include a placebo group represent the current standard in clinical trial design. As with drug therapy, surgical devices are well known to evoke a significant placebo response in patients. Thus, the magnitude of benefit due to an intervention is typically possible by noting the “placebo-subtracted” efficacy (intervention group minus control group benefits). A critical analysis of the placebo-subtracted benefits of CRT on functional capacity and quality-of-life parameters suggest significant heterogeneity (differences in benefit varying by almost 50%) in different trials that presumably enroll similar patient populations that receive the same CRT device. Although variable benefits occur with medications as well, device therapy needs to be held to a higher standard because of the additional variable of technical skills of the implanting physician and device complications. The observed placebo response to CRT varies between 38% to nearly 50% in various trials, whereas the placebo-subtracted improvement in NYHA functional capacity (by ≥1 functional class) is in the realm of 15% to 30% (5,6). Interestingly, the heterogeneity of benefit among the different trials cannot be simply explained by criteria of patient enrollment nor by the device used to achieve CRT. The implication of this observation is that the large majority of the observed functional improvements in responders to CRT is in fact dependent on the placebo response that is evoked by the implantation and not necessarily by the actual process of resynchronization.
Other lines of evidence pointing to the heterogeneity of response are also forthcoming. Similarly, although the primary end point in one trial of an improvement in 6-min walk test was significantly achieved, a separate trial failed to demonstrate any benefit in that particular parameter (5,6). Heart failure hospitalizations were noted to decrease with resynchronization therapy in one trial but not in others (8). The meta-analysis by Bradley et al. (8)that evaluated combined results from four randomized trials (number of enrolled patients ranged from 58 to 554 per trial, n = 1,634) of cardiac resynchronization therapy suggested that deaths from progressive HF were decreased, yet no benefit on all-cause mortality was noted and not enough information on all-cause hospitalizations was available to generate an adequate composite analysis. The fact that the magnitude of benefit is not consistent across the spectrum of clinical trials is disturbing. Because the benefit of cardiac resynchronization therapy must likely ensue for achieving mechanical synchronicity, it is possible that restoration of mechanical synchrony is not a consistent event. Why might this be so?
Where is the resynchronization in CRT?
Although the technique of multisite pacing is most often referred to as resynchronization therapy, the typical process involves selection of patients based on HF severity, surface electrocardiographic QRS widening, and evidence of underlying cardiac enlargement with consequent systolic dysfunction. This is then followed by the technical implant via standardized techniques, and success of the procedure is heralded by completion of device placement. The appropriate approach to resynchronization must include much more than these simple characteristics. First, we must be able to demonstrate the presence and precise location of mechanical dyssynchrony. Then, we must be able to technically locate and position the pacing leads in the appropriate position so as to improve the dyssynchronous contractility. Finally, we must be able to show evidence for sustained improvement in ventricular dyssynchrony. Thus, the current approach to resynchronization represents a “best guess” approach to achieving resynchronization by observation of surrogate responses.
One of the major technical limitations to achieving success with resynchronization therapy lies in the anatomic variability of the coronary sinus anatomy as well as the experience of the implanting physician (10,11). In this regard, coronary sinus anatomy often requires delineation by venography and carries up to a 6% risk of venous dissection or perforation as reported in the trials that presumably were conducted by experienced investigators who underwent supervision and implant training (4–7,12). Furthermore, despite the experience levels of the trial investigators, the rate of unsuccessful implants ranged from 8% to 13% (4–7). One can imagine, therefore, that if these findings were to be translated to general practice, the net benefit noted in responder rates would necessarily be decreased by the magnitude of unsuccessful implants and technical complications.
Does a wide QRS always denote dyssynchrony?
The recognition that surface electrical conduction defects represented by a widened QRS complex are associated with mechanical asynchrony is well accepted (3). In particular, this electrocardiographic finding has been linked with an independent impact on cardiac mortality, and the severity of this abnormality tends to correlate with worse outcome (13). However, the ability of a wide QRS interval to consistently predict the presence, magnitude, and location of mechanical asynchrony is questionable. Thus, benefits of resynchronization therapy are predicted neither by the severity of QRS widening nor by the magnitude of decrease in QRS width after resynchronization (14). An important investigation by Aurrichio et al. (15)compared pacing sites in relation to percent change in rate of change of pressure over time (dP/dt). Increases in systolic function were significantly greater at mid-lateral LV sites compared to any other region that could be accessed via the coronary sinus. In addition, pacing at apical or mid-anterior sites increased LV +dP/dt significantly more than pacing at the base. It has also been noted that the maximum short-term benefit at any site occurs with customizing the atrioventricular delay to an individual. If indeed an individualized atrioventricular delay is essential for achieving at least some of the proposed benefits of cardiac resynchronization therapy, one could infer that the presence of underlying atrial fibrillation could create a situation wherein it might be difficult to maximize synchronicity. The implications of this finding are important because a substantial number of patients with severe HF present with concomitant atrial fibrillation and most studies of cardiac resynchronization therapy have excluded this population (4–7). Another provocative study by Leclercq et al. (16)provided evidence that LV pacing alone may lengthen the QRS interval but yet resynchronize the mechanical function of the ventricle in a similar fashion, as does biventricular pacing. Left ventricular pacing starts with focal lateral wall contraction that advances slowly, with prominent contraction appearing next in the septum. The investigators speculate that the mechanism for this inherent electromechanical paradox might arise from the creation of a focal area of slowly contracting muscle in the mid-lateral wall.
Predictors of response to CRT
In evaluating the relationship of baseline ventricular configuration and outcome from resynchronization therapy, Pitzalis et al. (17)suggested that the mere use of underlying LV volume and ejection fraction might not be an adequate predictor of response to resynchronization therapy, and a baseline measure of mechanical contractile dyssynchrony is needed to identify those patients most likely to benefit. In this study, the investigators used septal-to-posterior wall-motion delay as an echocardiographic parameter of mechanical asynchrony and showed that only patients with prolongation in this measured index achieved reversal of LV remodeling. Using tissue Doppler imaging, Sogaard et al. (18)have convincingly demonstrated that the site and degree of mechanical asynchrony can vary from patient to patient and in particular is influenced by the underlying etiology of disease, whether ischemic or non-ischemic. Thus, it is confirmed that a QRS width on the electrocardiogram represents at best a crude measure of underlying mechanical asynchrony, cannot be expected to provide an adequate means of predicting response to resynchronization therapy, and definitely cannot be used to guide appropriate lead placement.
The definition of a successful implant therefore should not be narrowed to merely accepting device implantation, but must include demonstration and identification of the precise area of ventricular asynchrony, followed by positioning the coronary sinus catheter in the optimal location to tackle the mechanical dyssynchronous contractility.
Extrapolation of CRT to mild HF
As with any therapeutic application, the urge to extrapolate potential for benefit to less morbidly ill populations exists but must be resisted. It has been observed that even trial investigators tend to over classify severity of functional impairment in HF and a gap exists in the intended severity of illness and final population enrolled. Thus, trials of beta-blockers that purported to represent severe HF were noted to have enrolled populations in whom the observed mortality rates in the control groups were far less than those that would be expected based on the enrollment criteria of severity (19,20). Furthermore, the tendency to extrapolate evidence to less ill patients with HF might be harmful. In this regard, a recent investigation by Bozkurt et al. (21)provided evidence for a rising rate of inappropriate prescriptions for spironolactone in less severe forms of HF, with consequent increases in complication rates beyond those reported in the clinical trials with this agent. Unlike drug therapy, which can be followed and the offending drug stopped before potential harm, the clinical cost of devices is typically front-loaded, with reported mortality rates ranging from 0.3% to 2.1% and other major complications including vascular complications (6%), infection (1%), lead dislodgement or loss of capture (7%), and technically unsuccessful implants (8% to 13%) (4–7).
In view of this, it is appropriate to consider resynchronization therapy in patients with severe systolic HF in spite of structured use of optimal pharmacologic therapy. Ideally, objective measures of reduced functional capacity on either a 6-min walk or cardiopulmonary exercise stress test should be used to confirm the limitation in exercise capacity. This would ensure the application of such therapy to a patient population in whom the risk-benefit ratio of device therapy might be more justifiable. Thus, the decision to perform cardiac resynchronization therapy should not be taken lightly, and patients should be clearly informed of the potential risks and heterogeneous response expectation.
One of the most important advances in CRT must accrue from diminishing the implantation risks and redefining “successful” implantation to include correlation of the site and magnitude of mechanical asynchrony. As more information regarding mechanisms of benefit from resynchronization accumulate and patient selection evolves (22,23), further large-scale, adequately powered clinical trials will be needed to assess effects on morbidity and mortality as well as functional outcomes in patients with less severe HF. Finally, continued investigation to improve pacing and sensing capabilities of the devices and to determine optimal approaches for achieving a beneficial clinical response are essential to ensure that CRT is offered to those most likely to benefit.
☆ Dr. Lynne W. Stevenson acted as Guest Editor for this article.
- cardiac resynchronization therapy
- cardiac resynchronization therapy pacer or defibrillator
- heart failure
- left ventricle/ventricular
- New York Heart Association
- Received March 10, 2003.
- Revision received June 24, 2003.
- Accepted October 21, 2003.
- American College of Cardiology Foundation
- Mehra M.R.,
- Uber P.A.,
- Francis G.S.
- McMurray J.,
- Pfeffer M.A.
- Tavazzi L.
- Multicenter InSync ICD Randomized Clinical Evaluation (MIRACLE ICD) Trial Investigators,
- Young J.B.,
- Abraham W.T.,
- Smith A.L.,
- et al.
- Summary of Safety and Effectiveness: Guidant CONTAK CD CRT-D System [report] U.S. Food and Drug Administration Web site. Available at: http://www.fda.gov/cdrh/pdf/p010012.html. Accessed June 4, 2003
- ↵Bristow MR, Saxon LA, Boehmer J, et al., for the COMPANION Investigators. Cardiac resynchronization therapy (CRT) reduces hospitalizations and CRT an implantable defibrillator (CRT/D) reduces mortality in chronic heart failure: preliminary results of the COMPANION Trial. ACC Late Breaking Trials, 52nd Annual Scientific Session, Chicago, IL, March 30 to April 2, 2003.
- Baldasseroni S.,
- Opasich C.,
- Gorini M.,
- et al.
- Auricchio A.,
- Stellbrink C.,
- Block M.,
- et al.
- Leclercq C.,
- Faris O.,
- Tunin R.,
- et al.
- Pitzalis M.V.,
- Iacoviello M.,
- Romito R.,
- et al.
- Sogaard P.,
- Egeblad H.,
- Pedersen A.K.,
- et al.
- The Cardiac Insufficiency Bisoprolol Study II (CIBIS-II): a randomised trial. Lancet 1999;353:9–13.
- Bozkurt B.,
- Agoston I.,
- Knowlton A.A.
- Breithardt O.A.,
- Sinha A.M.,
- Schwammenthal E.,
- et al.
- Nowak B.,
- Sinha A.M.,
- Schaefer W.M.,
- et al.