Cardiac resynchronization therapy Free

Abstract

The first case report introducing the concept of cardiac resynchronisation therapy (CRT) was published less than 10 years ago, opening the way to the development of the first successful non-pharmacological treatment of congestive heart failure (CHF). The now routine implantation of CRT systems is applicable to multitudes of patients as adjunctive therapy in advanced CHF. This technique has transformed the traditional concepts associated with stimulation of the heart, and is now applied not only to restore an appropriate heart rate, but also to change the process of cardiac mechanical activation. Since it must be integrated within a comprehensive and multidisciplinary CHF management program, CRT has changed the practice of experts in the field of cardiac pacing. CRT in the management of CHF was ultimately validated in 2 randomised trials. MUSTIC, the first trial, compared in a single-blind, 3 × 3 months crossover design active versus inactive biventricular stimulation in a group of patients in sinus rhythm and another group in atrial fibrillation. Both phases of the trial were completed by 48 patients, with significant positive effects conferred by CRT on the distance walked in 6 min and on peak oxygen consumption. The number of hospitalizations for management of CHF was decreased by 2/3 ( P  < 0.05), and 85% of patients preferred the atrio-biventricular over the inactive stimulation mode ( P  < 0.001). These results were amply confirmed by the parallel-design MIRACLE trial. The current indications for CRT, dignostic tools to assist in its implementation, and limitations of this new therapeutic adjunct are further discussed in this review.

Multisite cardiac pacing for haemodynamic support is celebrating 10 years of existence. The first case report introduced the concept of cardiac resynchronisation and opened the way to the development of the first non-pharmacological treatment of congestive heart failure (CHF) [1] . Over the following years, cardiomyoplasty, an attractive, though cumbersome technique, was gradually (perhaps temporarily) abandoned, and discouraging trials of traditional “dual chamber” pacing [2,,3] were replaced by biventricular multisite stimulation [4–,6] . The now routine implantation of cardiac resynchronization therapy (CRT) systems is applicable to multitudes of patients as adjunctive therapy in advanced CHF. This technique has transformed the traditional concepts associated with stimulation of the heart. It is now applied not only to restore an appropriate heart rate, but also to change the process of cardiac mechanical activation. After the successful implantation of a CRT system, the efficiency of muscular contraction is increased, confirming observations by pioneers who highlighted the importance of the sequence of mechanical activation [7,,8] .

Multisite stimulation has also modified the technical aspects of cardiac pacing. In particular, placement of the leads is no longer determined only by their quality and long-term reliability, and by the capture and sensing characteristics at the heart–electrode interface. Since the site of origin of cardiac electrical activation plays an important role in the haemodynamic outcome, the era of systematic right ventricular (RV) apical pacing is over. An optimal and stable stimulation site must be identified, perhaps at the cost of a higher capture threshold, against the teachings of traditional cardiac pacing.

Finally, since it must be integrated within a comprehensive and multidisciplinary CHF management programme, CRT has changed the practice of experts in the field of cardiac pacing, who must leave the sanctuary of the electrophysiology laboratory to enter worlds previously foreign. Both establishing the indications for CRT and programming of the pulse generators now require the assistance of cardiac imaging, and are conceivable only in the context of pharmacological therapy. Competition with other modes of treatments has forced the abandonment of ingenious “home made” techniques, brought greater scientific transparency and verification, and brought this field of cardiology into the family of medical disciplines based on evidence. With the development of multisite stimulation, cardiac pacing has entered a new era.

State of the art

The birth of cardiac stimulation for haemodynamic support

This long neglected concept emerged from disappointing attempts to improve the short- and long-term haemodynamic status based on conventional dual chamber pacing [2,,3] . By programming a short atrioventricular (AV) delay, a lengthening and optimisation of left ventricular (LV) filling was observed, along with a decrease in presystolic mitral regurgitation which may have coexisted [2,,9,,10] . This was particularly apparent in patients presenting with 1st degree AV block, whose atrial and ventricular systoles are desynchronized. In 1992, possible candidates for this type of stimulation had a wide QRS complex, functional mitral regurgitation, and an abbreviated LV filling phase. However, controlled studies failed to confirm encouraging preliminary results, or even reported adverse effects [11,,12] . These observations offered an a posteriori reinterpretation of the adverse haemodynamic effects sometimes present in patients with CHF permanently paced for AV block [13] . Finally, the unfavourable long-term results in the original series of patients [14] ended, at least for now, any further clinical research in this area.

Dyssynchronization and resynchronization

The adverse effects of RV apical pacing are explained by the anomalous ventricular electrical activation process from apex to base. This activation, which traverses areas of diseased myocardium and Purkinje network, may be the cause of contraction heterogeneity and asynchrony within the entirety of the ventricles, and of haemodynamically detrimental right–left interventricular delays. These mechanical delays were identified long ago by echocardiographic or phase-analysis angioscintigraphic studies [14–,17] .

Fig. 1 , reproduced from the first recipient of a four-chamber CRT system, in 1994 [1] , illustrates the angioscintigraphic mechanical activation abnormalities in a left anterior oblique view. The first image (1A) was obtained during spontaneous rhythm. Isochronal events are shown in various colours, allowing the measurement of total duration of contraction and the quantification of its heterogeneity. RV contraction, from base to apex, is prolonged and LV contraction is prominently heterogeneous. The high lateral LV wall and RV contract synchronously, the anterior wall and apex are delayed, while septal contraction, occurring during atrial systole, is out of phase. The overall duration of contraction, similar in both ventricles, is >50% of the cardiac cycle. The second image (1B) was obtained during conventional DDD pacing with an AV delay optimised according to Ritter's formula [18] . Both atria were paced simultaneously to eliminate the interatrial conduction delay. The RV lead was placed in the RV outflow tract, as near as possible to the base of the heart. Capture of the RV modified the contraction of both the RV and the LV. It is noteworthy that the considerable prolongation of overall ventricular contraction was due to a marked difference between the two ventricles, such that LV contraction begins when that of the RV ends. Therefore, conventional dual chamber pacing, even after programming of an AV delay optimised by biatrial stimulation during echocardiographic monitoring, induced interventricular asynchrony and aggravated pre-existent abnormalities.

Other studies were conducted throughout the 1990s in an effort to optimise LV function by stimulation from alternate ventricular sites. Victor et al reported the results of a randomised, double-blind comparison of RV apical versus RV outflow tract in 16 patients in chronic AF and complete AV block [19] . No difference was observed in LV ejection fraction (EF) and functional capacity at the end of 3 months of stimulation from either site. Likewise, in a population of patients undergoing coronary artery bypass surgery, among several short-term single- or dual-site configurations tested, including RV and LV epicardial apex, RV septal and RV outflow tract, none was found haemodynamically superior [20] . These observations suggest that electro-mechanical abnormalities cannot be corrected by merely optimising a single or dual RV stimulation site. On the other hand, synchronous RV and LV activation should allow such correction, as long as propagation of the wavefront from each electrode is preserved. The third image (1C) illustrates the changes in mechanical activation as a result of four-chamber stimulation after placement of an LV lead. Compared with the activation shown in Fig. 1B , the interventricular delay has disappeared and the intraventricular heterogeneity has diminished despite ongoing stimulation. The haemodynamic improvement, confirmed invasively, was obtained by AV resynchronization, which optimised ventricular diastolic filling without detrimental effect on systole.

To our knowledge, de Teresa et al. presented the first report of LV stimulation in 1983 [21] . In four patients with left bundle branch block undergoing surgical aortic valve replacement, an epicardial LV lead was implanted and interfaced with the ventricular channel of a dual chamber pulse generator. The atrial lead was placed in the right atrium and the AV delay programmed to allow ventricular fusion of spontaneous activation via the His bundle and activation from the stimulated site. The authors reported a 25% increase in LVEF and normalisation of sequence examined by angioscintigraphy. Additional data were presented 12 years later by Foster et al., who observed an increase in cardiac output by atrio-biventricular stimulation via temporary epicardial electrodes after withdrawal of cardio-pulmonary bypass in patients undergoing coronary artery bypass surgery [22] .

In 1996, we reported on the first series of recipients of CRT systems [23] . Prior to the implantation of a permanent system, responders were identified by temporary stimulation, which had confirmed that all biventricular configurations were superior to the conventional dual chamber configuration, regardless of the RV lead position. The study included angioscintigraphic imaging after implantation of the permanent CRT system, which confirmed a decrease in heterogeneity and duration of ventricular contraction by multisite stimulation. These observations have been amply confirmed by others, whose further contributions included an emphasis on the importance of optimal LV lead placement, on the lateral wall in particular [24–,26] .

Validation of the concept

CRT in the management of CHF was ultimately validated in two randomized trials. MUSTIC, the first trial, compared, in a single-blind, 3 × 3 months crossover design active versus inactive biventricular stimulation in a group of patients in sinus rhythm [5] and another group in atrial fibrillation (AF) [27] . Patients were then followed long-term with the stimulation mode programmed according to each patient's preference. Patients in sinus rhythm and in stable New York Heart Association (NYHA) CHF functional class III during optimal medical management had no conventional pacing indications. A QRS duration >150 ms was arbitrarily chosen as a marker of ventricular dyssynchrony. All patients received an atrio-biventricular pulse generator, interfaced with right atrial, RV and LV leads. The latter was positioned via the coronary sinus (CS) according to the method of Daubert et al. [28] . The primary study endpoint was evolution of functional capacity, measured by the 6-min hall walk. Secondary endpoints included quality of life ascertained by the Minnesota questionnaire, peak oxygen consumption, number of hospitalizations for management of CHF, patient's stimulation mode preference at the end of the crossover phase, and overall survival. The CRT system implantation success rate was 92% and, at the end of the crossover phase, 88% of LV leads were correctly functioning. The effects of CRT were confirmed by a 25 ms decrease in QRS width. Both phases of the trial were completed by 48 patients, with a 23% increase, between the active and inactive mode, in the distance walked in 6 min ( P  < 0.001), and in peak oxygen consumption ( P  < 0.03). The number of hospitalizations for management of CHF was decreased by 2/3 ( P  < 0.05), and 85% of patients preferred the atrio-biventricular over the inactive stimulation mode ( P  < 0.001).

These results, which proved the existence of a treatment effect of CRT in patients with CHF and inter- and intraventricular conduction abnormalities, were amply confirmed by the MIRACLE trial which, in its parallel design, yielded similar results [6] .

What is known and what is unknown

Though CRT has been validated, several questions remain unanswered, pertaining to the evaluation of its actual value, the precise definition of its indications, and its adjunctive therapeutic role.

Effect on survival

Despite much progress, the effects of CRT on survival remain unknown. Studies completed thus far have shown no increase in mortality compared with that observed in similar patient populations, as well as spectacular individual remissions, particularly in patients in NYHA functional class III at the time of system implantation. In the COMPANION trial, now published [29] , patients were randomly assigned to optimal medical therapy, versus CRT, versus CRT + an implantable cardioverter defibrillator (ICD). The 30-day mortality, including operative deaths, was 1.0% in the medically treated group, 1.8% in the CRT group, and 1.2% in the CRT + ICD group. At 1 year, the corresponding combined rates of deaths or hospitalizations for management of CHF were 45%, 31% and 29%, respectively. The 35% relative risk and high rate of adverse events in the medically treated group unfortunately forced the early interruption of the trial, complicating the interpretation of its results. Important, complementary information is expected from the similarly designed CARE-HF trial, which includes a considerably longer patient follow-up [30] .

Current indications for CRT

Indications for CRT based on the rhythm identified on the surface electrocardiogram (ECG) are rapidly becoming obsolete. Patients presenting in AF, i.e. without atrial contribution to cardiac output, as well as patients previously paced permanently for AV conduction disorders, benefit from CRT [31] . Furthermore, patients with AF undergoing AV junctional ablation should be considered a priori candidates for biventricular pacing to optimise their long-term haemodynamic status [32] . There is growing doubt regarding the usefulness of the ECG as a tool to identify candidates for CRT. While it is a simple test, it does not seem accurate enough to identify precisely a given candidate, gauge the quality of resynchronization, and guide the optimal lead configuration.

In its parallel design, the MIRACLE trial had identified a high percentage of non-responders or even patients whose clinical status was worsened by CRT [6] . These undesirable responses may be due to improper patient selection or to technically deficient implantation of the system. From the inception of the therapy, efforts were made to identify patient selection criteria that were both reliable and applicable on a large scale to encompass a large target population. Measurement of the QRS on the surface ECG was the simplest marker of latent dyssynchronization. A duration of 150 ms was initially chosen arbitrarily and intuitively as a cut-off value in MUSTIC and shortened to 120 ms in MIRACLE. However, a correlation between the amount of dyssynchronization and QRS duration has not been established [33,,34] . While the likelihood of a favourable haemodynamic result is greater when the QRS is >150 ms [35–,37] , the only certainty is that a QRS shorter after than before system implantation carries a good prognosis [38] . Since the QRS is currently the only marker easily applicable during implantation procedures, it is considered the reference. However, many patients have responded to CRT in the long-term despite having a QRS between 120 and 150 ms in duration. Furthermore, these measurements are inapplicable to patients with previously implanted conventional pacing systems. Finally, some investigators believe that changes in the QRS frontal axis may be of prognostic value, a marker which remains to be tested in large clinical trials.

The limitations of the surface ECG have prompted the development of other methods of identification and follow-up of candidates for CRT. Doppler echocardiography is the non-invasive method currently used, as it is easily applicable in daily practice [39] . It allows (a) imaging of dyssynchronization at three separate levels, AV, interventricular and intraventricular, (b) the identification of potential candidates for CRT, (c) the verification of proper delivery of therapy, and (d) the definition of the mechanisms of action of CRT [40] . AV dyssynchronization , identifiable only during sinus rhythm, is due to abnormal timing of the end of atrial and beginning of ventricular systole, often facilitated by prolonged AV conduction. Characterized by an abbreviated filling period, sometimes accompanied by the summation of early passive (E wave) and late active (A wave) atrial emptying, it can be normalised with a conventional dual chamber pacing system by optimizing the AV delay during Doppler echocardiographic imaging [41,,42] . The concept of interventricular dyssynchronisation , facilitated by conduction abnormalities, was behind the development of CRT [1] . The duration and temporal shifts of right and left ventricular systole can be measured by various methods, Doppler echocardiography being the simplest [39] . A shortening of these time intervals appears to predict effective resynchronization. Finally, intraventricular conduction may be heterogeneous enough to result in the coexistence of systolic and diastolic phases of different regions of the same ventricle. The ECG is currently of no diagnostic value for this type of intraventricular dyssynchronization . Its presence and its correction may both be predictive of the efficacy of CRT though it has not been compared with other indices of dyssynchronization [43–,45] . In addition, intraventricular dyssynchronization may often be implicated in mitral insufficiency correctable by CRT [40] , and may be a cause of myocardial ischaemia [46] . Since CRT improves the haemodynamic status while decreasing myocardial oxygen consumption [47] , it has been hypothesized that it improves regional myocardial perfusion by prolonging local relaxation. DESIRE, a European multicenter study, is currently examining the possible correlation between long-term clinical outcomes and quality of mechanical resynchronization.

Technological considerations

The first implantations of LV leads for CRT were epicardial and via thoracotomy. With the advent of successful endovenous techniques, this cumbersome and complicated procedure is currently reserved for patients in whom catheterization of the CS has failed. However, it may return as a routine procedure if true multipoint stimulation systems are developed. Percutaneous, transvenous implantation of the leads has become the primary technique. The LV lead may be implanted via the coronary venous network to stimulate the epicardium endovenously, or through the interventricular septum to stimulate the endocardium, though the latter approach is still experimental. Catheterization of the CS is currently the main approach to implant the LV lead, and has been the focus of most research and developmental efforts. In the past 5 years, manufacturers have designed leads specifically for this purpose, attempting to reconcile aims which are potentially conflicting. The leads have to be both as stiff as possible to enter the CS, and as thin and manoeuvrable as possible to be positioned in the smaller veins of the coronary network. Kits including preshaped guiding sheaths are now widely used to catheterize the CS ostium and perform a venogram in search of a target vein. The stimulating lead itself has an inner lumen allowing the introduction of a guide wire to stiffen or curve its contour. The guide wire may be used to catheterize the target vein, while the lead is advanced over the wire to its final destination. Close contact must be achieved between stimulating electrode and epicardium, in order to obtain the lowest possible capture threshold. Leads are generally preshaped for greater stability, and high-impedance to reduce the long-term battery drain of the pulse generator.

The <100% success rate of CRT system implantations remains a function of the morphology, distribution and calibre of the coronary venous network and target veins, as well as the experience and skills of individual implanting physicians. The complication rate is low, consisting mostly of lead dislodgement, while procedural times vary widely from a few minutes to several hours. On the other hand, intraoperative criteria of a successful implantation remain to be defined, as it is no longer satisfactory to depend on a change in QRS duration.

The type of pulse generator to interface with the lead system remains controversial. One may include an ICD, as suggested by an early decrease in mortality in the COMPANION trial [29] . On the other hand, it might be wiser to await the reverse ventricular remodelling which develops in a large percentage of patients, whose heart disease may stabilize, and focus the risk stratification efforts to identify patients at highest risk, who would benefit most from the addition of an ICD. The economical consequences of these alternate options are vastly different, particularly in view of the absence of reimbursement for simple CRT in several European countries considered wealthy.

Conclusions

CRT is a recently developed therapeutic modality which has been clinically validated. It is applicable a priori to patients who suffer from symptomatic LV systolic dysfunction despite optimal medical management, and who present with cardiac mechanical synchronization abnormalities which can be corrected or, at least, mitigated.

References