Pacing for Heart Failure: Selection of Patients, Techniques and Benefits

Chronic heart failure is a syndrome characterized by left ventricular dysfunction, reduced exercise tolerance, impaired quality of life, and a markedly shortened life-expectancy 1,2. A substantial proportion of patients with heart failure have electrical conduction disturbances. Recent data have shown that patients with chronic heart failure and conduction defects have a higher risk for cardiovascular events and a worse outcome 3,4. Current pharmacological therapy administered to heart failure patients is, however, unable to directly correct the electrical conduction abnormalities present in these patients 2,5–8. Many patho-physiological mechanisms could be advocated for the reported increase in adverse events observed for patients with chronic heart failure and electrical conduction defects. The prolongation of the PR interval in normal hearts is associated with a primary loss of the atrial contribution to ventricular mechanics 9,10. However, in dilated cardiomyopathy patients, this prolongation can also have significant effects on diastolic filling 11, facilitating diastolic regurgitation 11. The deleterious hemodynamic consequences of a prolonged PR interval are more evident for patients with severely depressed ventricular function. In these patients we can observe severe hemodynamic deterioration due to a prolonged mechanical atrio-ventricular interval, even in the presence of a normal PR due to the presence of large electromechanical delays 12. Intra- and interventricular conduction delays, such as left bundle branch block prolong the pre-ejection time 13, reduce the global and regional ejection fraction, negatively affect the contraction and relaxation times, and prolong the duration of a pre-existing mitral regurgitation. The mechanical consequences of abnormal activation can be easily assessed by echocardiography; the M-mode echocardiography shows that in the right-bundle branch block, the onset of right-sided atrioventricular ring motion is delayed, and that the left is delayed in the left bundle branch block 14. In addition, prolongation of ventricular depolarization, in particular the left bundle branch block generates discoordinate contraction, altering regional workload and stress 15. The region of early activation contracts against minimal load, but wastes its energy, as no effective pressure develops since much of the rest of the cavity is still non-activated. The late depolarized region contracts against an already stiffened portion of the wall (i.e. the early-activated part) increasing its stress. If there is dyskinesis of the earlier depolarized region by the late activated region, then additional cardiac work is wasted-internally transferring ejection from one portion of the chamber to another. Changes in regional load might contribute to myocyte damage, including cell necrosis or apoptosis and replacement fibrosis. The results of small, uncontrolled studies 16,17 suggested that the restoration of properly timed atrial and ventricular contractions by implantation of a conventional right-ventricular dual-chamber pacemaker was beneficial in the treatment of patients with chronic heart failure. These studies were conducted on patients who failed to respond to aggressive pharmacological therapy. These preliminary results, however, were not reproduced when better-controlled studies were conducted 18,19. On the other hand, more consistent and positive data have been presented when the left ventricle has been paced alone 20 or simultaneously paced with the right one (biventricular stimulation) 21–24. Some of the reasons that might be responsible for this discrepant findings are: patient selection; lack of clearly defined inclusion and exclusion criteria; statistical design; technique used for guiding the selection of the atrioventricular delay pacing site; and the study end-points 25. The recognition that conduction disturbances might be responsible for a degraded coordination of the atrio-ventricular, and intra-ventricular contractions — and that this lack of coordination might play a key role for worsening left ventricular function — led to the development of a prospective randomized trial to test the effect of applying pacing therapy to patients with moderate to severe chronic heart failure. The PAcing THerapies in Chronic Heart Failure (PATH-CHF) study includes patients who had symptoms of moderate to severe heart failure. All the patients included in the trial had a New York Heart Association Functional Classification of III (for at least 6 months) or IV despite treatment at the maximum tolerated dosage with diuretics, an angiotensin-converting enzyme inhibitor, digitalis, vasodilators or nitrates and β-blockers. All the patients had dilated cardiomyopathy of any etiology, heart rate higher than 55 bpm, QRS complex duration wider than 120 ms in at least two surface-ECG leads (maximum of leads II, V1, V6) and a PQ interval longer than 150 ms. Patients with conventional indications for pacemaker implantation were excluded from the study. A complete list of exclusion criteria has been previously published 26. The PATH-CHF Study consisted of an acute testing phase using the FLEXSTIMTM system during the implantation procedure, and of a randomized crossover protocol which tests two different pacing modes (univentricular or biventricular pacing) with a no-pacing (wash-out) period in-between. The randomization procedure included two arms. In the first arm we applied atrial triggered univentricular pacing for 4 weeks (right or left ventricle). The other received atrial triggered biventricular pacing for 4 weeks (simultaneous right and left ventricle). At the end of the 4-week pacing period, the pacemakers were programmed off in all patients for the next 4 weeks. After these 4 weeks of no-pacing therapy, the patients were crossed over to the other pacing mode, and pacing therapy was restarted again for another 4 weeks. The study protocol was completed after 12 weeks. All the included patients are going to be followed up for 1 year. The rationale for the FLEXSTIMTM acute protocol was to measure the cardiac response to pacing 27. In fact, a controlled study of pacing chamber and AV delay is technically difficult because of the large number of treatment combinations. The analysis of the acute hemodynamic effects of pacing using the standard methodology (e.g. cardiac output determined by thermo-dilution or by Doppler examination 28) is time consuming and may not provide sufficient data for meaningful statistical comparisons of the different pacing sites and AV delays in a single patient. This is due to the fact that these techniques can have measurement errors that are in the same order of magnitude than the changes that we are trying to detect 29. Therefore, they force the use of group statistics to try to get around this technical limitation. Nevertheless, group statistics will also be misleading when there are large interactions between individuals and pacing variables, preventing the elimination of the measurement errors present in the above mentioned techniques by the use of population averages 30. In addition, normal physiological oscillations in the baseline during the study period will also affect the measurements and need to be corrected 31. The unknown periodicity of these baseline changes, makes this source of error very difficult to eliminate. The ‘FLEXSTIM™ protocol’ was specifically designed to address the issues mentioned above. It consists of periods of pacing in VDD mode (i.e. atrial sense followed by a ventricular pace after an atrio-ventricular delay) separated by periods of no pacing in normal sinus rhythm in a 5-paced beat/15-non-paced beat duty cycle. When pacing and no pacing sequences were interrupted by a premature ventricular contraction (PVC) additional beats were added to allow the PVC beat and surrounding beats to be removed from the analysis. Pacing has been performed in the right ventricle (RV), left ventricle (LV), or right and left ventricles (biventricular) simultaneously (BV) at one of five different AV delays. The five AV delays were equally distributed between 0 ms and 30 ms less than the patient’s intrinsic PR interval as measured by the intracardiac electrograms. Each treatment combination of pacing chamber and AV delay has been repeated five times in random order for each patient. The entire sequence summed to 1500 beats, which, at an average heart rate of 75 bpm, required 20 min. The major finding of the acute testing during the PATH-CHF study was that the acute left ventricular systolic function of most patients with moderate to severe CHF and left ventricular systolic dysfunction was significantly improved acutely by paced ventricular pre-excitation at an optimal combination of AV delay and site (Fig. 1). Left ventricular diastolic and right ventricular systolic function changed minimally 27. In this patient population, which is dominated by left ventricular conduction disorders, left ventricular stimulation alone or in combination with right ventricular stimulation substantially increases the hemodynamic benefits when compared to right ventricular stimulation alone. In addition, two patients groups have been retrospectively identified, that have opposite acute pacing responses 27. Curiously, these two groups are characterized by different surface QRS widths: patients with sufficiently wide QRS (>150 msec) exhibit large positive hemodynamic changes with pacing (Type I), while patients with narrower QRS (5%, only a few patients increase LV+dP/dt >20% with RV pacing. If these acute differences are predictive of chronic pacing benefit, the seemingly contradictory results of Hochleitner et al. 16, Auricchio et al. 9, Gold et al. 18 and Linde et al. 19 on the benefit of RV pacing might be explained by fortuitous differences in clinical patient populations. On the other hand, the consistently reported positive chronic results that have been obtained when pacing the left ventricle may be related to the large positive hemodynamic response to left sided pacing that we have observed in a majority of the PATH-CHF patient population. Given that CHF patients have heterogeneous symptoms, etiologies, and substrates, and also that pacing may operate through multiple mechanisms (that in turn may require individual optimization), we should not be surprised by the conflicting chronic results that have been reported up to now. The data of the PATH-CHF study show that pacing can produce both positive and negative hemodynamic effects, at least acutely. The outcome seems to depend on the pacing site and AV delay, but most significantly on the individual patient. The most important individual variable for predicting acute pacing benefit in this study seems to be the QRS width. If the distinction between Type I and Type II patients also applies to chronic pacing benefit, studies including a mixture of such patients may produce mixed results. Thus, it would seem important to account for these individual variations in future studies of pacing therapy for CHF. By selecting the optimal pacing parameters for each individual patient, the negative effects of pacing Type II patients can be minimized. And the positive effects of pacing on the Type I patients can be maximized, with optimized cardiac output increase being two to four times larger than with sub-optimal pacing. The primary end-points of the PATH-CHF trial were to compare the relative benefit of pacing vs. no pacing using objective measures of: maximum oxygen consumption, oxygen consumption at anaerobic threshold, and changes in a 6-min walk distance. In addition, changes in functional class as assessed by the NYHA classification, in the quality-of-life measured according to the Minnesota Living with Congestive Heart Failure questionnaire, and the modifications in left ventricular ejection fraction were recorded. Another objective of this study was to compare the effects of uni-ventricular and BV pacing on the already listed parameters. The PATH-CHF study is still an ongoing trial; a total of seven active centers enrolled 42 patients to date. The study is expected to be completed by the end of 1998 and the first quarter of 1999. Preliminary data of the PATH-CHF study indicates that the benefit achieved during the pacing period are of similar extension to what have been recently reported by Gras et al. 33. Preliminary data of the PATH-CHF study indicates that significant benefit occurs during the pacing periods, these results have been reported elsewhere 26. The magnitude of the observed changes are similar to the ones reported by the uncontrolled InSync trial 33. This latter study showed that 12 weeks of continuous biventricular pacing was effective in prolonging the walked distance as well as in improving the functional class and the feeling-of-well being in the vast majority of the patients. In conclusion, theoretical data, acute results and clinical observations now join force to suggest that amelioration of signs and symptoms of chronic heart failure may occur in selected patients with moderate to severe chronic heart failure after implantation of an appropriately programmed pacemaker. Furthermore, early data indicate that the addition of sequential pacing therapy to conventional optimal medical therapy might be associated with prolongation of exercise capacity, a reduction of norepinephrine plasma level, 34 improvement of functional class and feeling of well being among patients with moderate to severe chronic heart failure and electrocardiographic documented conduction defects. The final results from three prospective randomized ongoing studies, the PATH-CHF, VIGOR CHF and VENTAK CHF trials will help in the understanding of this potential new therapeutic alternative. But, even if the preliminary positive effects of pacing are validated, there is no doubt that a survival trial will be still needed to unequivocally demonstrate the long-term efficacy of this non-pharmacological approach. If the results of the PATH-CHF trial are positive, at this point in time, it should be considered that pacing therapy only applies to those patients who fulfill the described inclusion criteria of the prospective randomized trials, and further experiences are needed for other patient population. Our acute data also seems to support that, unless an acute optimization is performed before implant, a minimum QRS width of 150 ms should be required 35.