Effect of Pacing Chamber and Atrioventricular Delay on Acute Systolic Function of Paced Patients With Congestive Heart Failure
Background—Previous studies of pacing therapy for dilated congestive heart failure (CHF) have not established the relative importance of pacing site, AV delay, and patient heterogeneity on outcome. These variables were compared by a novel technique that evaluated immediate changes in hemodynamic function during brief periods of atrial-synchronous ventricular pacing.
Methods and Results—Twenty-seven CHF patients with severe left ventricular (LV) systolic dysfunction and LV conduction disorder were implanted with endocardial pacing leads in the right atrium and right ventricle (RV) and an epicardial lead on the LV and instrumented with micromanometer catheters in the LV, aorta, and RV. Patients in normal sinus rhythm were stimulated in the RV, LV, or both ventricles simultaneously (BV) at preselected AV delays in a repeating 5-paced/15-nonpaced beat sequence. Maximum LV pressure derivative (LV+dP/dt) and aortic pulse pressure (PP) changed immediately at pacing onset, increasing at a patient-specific optimal AV delay in 20 patients with wide surface QRS (180±22 ms) and decreasing at short AV delays in 5 patients with narrower QRS (128±12 ms) (P<0.0001). Overall, BV and LV pacing increased LV+dP/dt and PP more than RV pacing (P<0.01), whereas LV pacing increased LV+dP/dt more than BV pacing (P<0.01).
Conclusions—In this population, CHF patients with sufficiently wide surface QRS benefit from atrial-synchronous ventricular pacing, LV stimulation is required for maximum acute benefit, and the maximum benefit at any site occurs with a patient-specific AV delay.
Treatment of congestive heart failure (CHF) with electrical stimulation has been proposed as additional therapy to be used in conjunction with pharmacological treatment.1 Although early studies reported short-term2 or long-term3 4 improvement with right ventricular (RV) pacing, subsequent studies did not reproduce these improvements with similar CHF patients.5 6 In contrast, a more uniform, beneficial result of pacing has been reported when the left ventricle (LV) has also been stimulated.7 8 The acute hemodynamic effect of atrial-triggered LV and simultaneous RV and LV (biventricular) pacing at a fixed AV delay of 100 ms has also been reported.9 In addition to issues about the pacing chamber, there is wide speculation regarding the need and method for optimizing the AV delay.2 4 5
The study objective was to investigate the effect of pacing chamber and AV delay with atrial-synchronous ventricular pacing on hemodynamic function in a specific CHF population consisting of patients in stable New York Heart Association (NYHA) class III or IV with severe LV systolic dysfunction, wide QRS complex (≥120 ms), normal sinus rhythm, and no history of arrhythmia.
Twenty-seven patients with NYHA class III or IV heart failure were enrolled in a multicenter prospective, randomized, single-blinded crossover study (Pacing Therapies for Congestive Heart Failure [PATH-CHF]). The complete PATH-CHF study inclusion and exclusion criteria, study design, and end points have been presented elsewhere.10 Main inclusion criteria were dilated cardiomyopathy of any origin; stable NYHA class III or IV with optimal medical therapy, including ACE inhibitors, nitrates, vasodilators, diuretics, digitalis, and β-blockers; a surface ECG QRS duration ≥120 ms; and a PR interval ≥150 ms. Main exclusion criteria were recent history of atrial or ventricular tachyarrhythmia; recent myocardial infarction; recent or pending coronary artery revascularization; severe valve disease; dependence on intravenous inotropes; and conventional indications for pacemaker implantation.
The protocol was approved by the ethical committees at all participating centers, as well as by the competent authorities of their respective countries. The study is monitored by a Patient Data/Safety Committee of independent reviewers. All patients provided written informed consent before participating in the study.
Acute data were collected while the patients were under general anesthesia, including fentanyl, propofol, and O2 during pacemaker implantation. Chronic pacing leads were implanted with screw-in leads (Sweet-Tip, Guidant Corp) in the right atrial appendage and RV, and an epicardial lead (model 4316, Guidant Corp, or model 4965, Medtronic) was attached to the surface of the LV via a limited thoracotomy. The LV pacing site was most commonly on the apex but could also be on midlateral segments. Two 8F dual-transducer pressure catheters (model SPC-780c, Millar Instruments) were placed to measure right atrial, RV, aortic, and LV pressures. Pressure catheters and pacing leads were connected to a custom external pacing computer (FlexStim, Guidant Corp) to acquire hemodynamic signals and execute a transient pacing protocol (FlexStim protocol) as previously described.10 The FlexStim protocol is designed to measure the immediate effects of pacing, account for local baseline shifts, and allow statistical comparison of multiple pacing combinations within individuals. Briefly, the RV, LV, or both ventricles (BV) were stimulated in a VDD mode (atrial sense followed by ventricular pacing at a predetermined AV delay) at 1 of 5 AV delays preset to percentages of the patient’s intrinsic PR interval measured with the pacing leads. Each combination of pacing chamber and AV delay was randomly repeated 5 times by pacing for 5 beats separated by 15 nonpaced beats.
Raw data were recorded on a digital tape recorder with 16-bit resolution at a 12 000-Hz sampling rate and then downsampled to 500 Hz for offline analysis. The following measurements were made automatically with custom software: aortic diastolic pressure, aortic systolic pressure (ASP), pulse pressure (PP=ASP−aortic diastolic pressure), LV pressure-derivative maximum (LV+dP/dt) and minimum (LV−dP/dt) by the method of Kass et al,11 LV end-diastolic pressure (LV EDP), RV maximum pressure derivative (RV +dP/dt), and RV systolic pressure. LV EDP was measured at the time when dP/dt reached 10% of the maximum. Reported QRS durations are the maximum of leads II, V1, and V6 measured automatically and validated manually by 2 independent observers.
Parameters were calculated for each beat of a pacing sequence as a percentage change from their average value during the immediately preceding 6 nonpaced beats (ie, local baseline). The first paced beat was always discarded, because its hemodynamic effects were systematically biased by the diastolic behavior of a preceding nonpaced beat and a transitory cycle length decrease caused by the switch to a short AV delay. In addition, beats altered by a premature ventricular complex were discarded. A 3-way ANOVA was applied to individual results to analyze differences between AV delays and pacing chambers while accounting for the ordered pacing beat number and interaction between the main effects. The AV delay and pacing chamber were considered the treatment variables and the beat number was considered a blocking factor when ANOVA results were reported. A similar 2-way ANOVA was applied to individual results at a selected optimal AV delay for each pacing chamber. We used corresponding 4-way and 3-way ANOVA tests to analyze group results by adding individuals as a treatment variable. Tukey methodology was used to evaluate multiple comparisons of the main effects in the full ANOVA models. A paired t test was used to evaluate the effects of pacing compared with no pacing. A P value of 0.0001 was considered statistically significant for ANOVA tests. A P value of 0.05 was considered statistically significant before Tukey corrections for individual results, whereas a more stringent P value of 0.01 was selected for group results owing to the larger number of sample points. Similarly, P values of 0.001 for individual results and 0.0001 for group results were considered statistically significant for the t tests. Average data are shown as mean±SD unless otherwise noted.
Table 1⇓ summarizes demographic and baseline hemodynamic data for the 27 patients enrolled in the study. Two patients (patients 1 and 22) were excluded from further analysis because LV and aortic pressure could not be recorded. All patients were able to tolerate the acute procedure without major complications.
Hemodynamic Impulse Response
Hemodynamic parameter changes during the brief pacing periods of the FlexStim protocol appeared immediately at pacing onset and diminished rapidly during the subsequent nonpacing period (Figure 1⇓). This was quantified with impulse response plots, which showed the average parameter measurement over all repetitions of a pacing combination in a patient on 5 paced beats and during the 15 nonpaced beats that followed pacing (Figure 2⇓). Typically for LV+dP/dt and PP parameters, changes from baseline occurred on the first paced beat, reached a maximum change during the next 4 paced beats, and then returned to baseline levels within 5 nonpaced beats. For all patients, recovery occurred within 10 nonpaced beats for every pacing chamber and AV delay combination. The paced-beat number had a statistical effect on parameter variation for most patients even when the first pace was discarded (ANOVA, P<0.0001).
Effects of Pacing Chamber
The statistical significance of the change in LV systolic parameters averaged over all AV delays for each pacing chamber is shown in Table 2⇓ for each patient. For the population average, pacing in any chamber significantly increased LV+dP/dt (P<0.0001), whereas BV and LV pacing significantly increased ASP and PP (P<0.0001). The selection of pacing chamber was a significant determinant of changes in all 3 LV systolic parameters (ANOVA, P<0.0001). LV+dP/dt, ASP, and PP changes with BV and LV pacing were significantly greater than the effects of RV pacing for the population; and LV pacing effects on LV+dP/dt were greater than the effects of BV pacing (P<0.01).
Population changes in other hemodynamic parameters averaged over all AV delays for each pacing chamber are shown in Table 3⇓. Changes to parameters that partially reflect LV diastolic performance were statistically significant but small and inconsistent. The LV EDP decreased, indicating lowered filling pressures, but the absolute value of LV–dP/dt also decreased, indicating slower relaxation. Average changes in RV systolic parameters were also relatively small and inconsistent.
Interpatient Variability of Pacing Chamber Effects
With the FlexStim protocol, the effects of pacing were statistically established for each individual (Table 2⇑). Most but not all patients had statistical improvement in short-term hemodynamic function with pacing at some site (P<0.001): RV, LV, or BV pacing significantly increased both LV+dP/dt and PP in 15 patients (60%), increased only LV+dP/dt in 5 patients (20%), and did not increase LV+dP/dt or PP in 5 patients (20%). Patients in this last group had significantly shorter QRS widths from surface ECG compared with other patients (180±22 versus 128±12 ms; t test, P<0.0001) and had less incidence of left bundle-branch block (20 of 20 patients versus 1 of 5 patients; χ2, P<0.0001) but were not statistically different for other demographic measures. Patients in this last group were retrospectively classified as type II, whereas all other patients were classified as type I. As shown in Figure 3⇓, all type I patients had significant improvement in LV+dP/dt or PP with pacing in the optimal chamber, and all but 2 of them had a surface QRS width >150 ms, whereas none of the type II patients had improvement in LV+dP/dt or PP, and all of them had a QRS width <150 ms. When patients were divided into type I and type II subgroups, RV-paced changes in ASP and PP reached significant levels for each subgroup (Table 2⇑). The statistical differences among pacing chambers were similar for the type I and type II subgroups, although parameter responses were in opposite directions.
Effects of AV Delay
For all patients combined and for type I and type II subgroups, AV delay was a significant determinant of changes in all LV systolic parameters (ANOVA, P<0.0001). Figure 4⇓ shows the paced change in systolic parameters as a function of AV delay for the patient subgroups. For type I patients, LV+dP/dt and PP AV delay functions were positive and unimodal, with a peak at the middle AV delay setting [0.5×(PR–30 ms)]. For type II patients, AV delay functions were negative for all parameters. For both type I and type II patients, there was significant interaction between AV delay and pacing chamber (ANOVA chamber/AV delay interaction, P<0.0001).
Interpatient Variability of AV Delay Effects
The detailed relationship between AV delay and LV+dP/dt or PP varied among individuals (ANOVA chamber/patient and AV delay/patient interactions, P<0.0001). One way to measure these differences is to compare the peaks of the AV delay functions for each patient, defined as the optimal AV delays. The optimal AV delays for the same pacing chamber and parameter varied widely among patients and often differed for PP and LV+dP/dt within an individual (examples in Figures 5⇓ and 6⇓). On average (Table 4⇓), the optimal AV delays were similar among pacing chambers for peak ASP and PP, but for the type I patients, they were significantly shorter for RV pacing and BV pacing compared with LV pacing for LV+dP/dt (paired t test, RV<LV P<0.008, BV<LV P<0.04). The optimal acute benefit attained for each patient group and pacing chamber is shown in Table 5⇓. Statistical results at the optimal AV delays were similar to those over all AV delays (Table 2⇑), except for type II patients, whose optima were near baseline at the longest AV delays.
The major finding of this study is that the acute LV systolic function of patients with moderate to severe CHF, LV systolic dysfunction, and prolonged ventricular activation is significantly and consistently improved by ventricular pacing at an optimal combination of AV delay and site, while LV diastolic and RV systolic parameter changes are relatively small and inconsistent. Patients with QRS>150 ms exhibit large positive LV systolic changes with pacing, whereas patients with QRS<150 ms exhibit predominantly negative LV systolic changes. However, even for these latter patients, LV systolic function could be optimized by individualized selection of AV delay and pacing chamber. In this patient population, which is dominated by LV conduction disorders, LV stimulation alone or in combination with RV stimulation substantially increases LV systolic benefits compared with RV stimulation alone.
Hemodynamic Impulse Response
We found that systolic parameters changed immediately with the onset and cessation of pacing, which confirms an observation made by Blanc et al,9 who also showed these changes persist after several minutes of steady-state pacing. Others have shown similar increases in PP12 and LV+dP/dt13 in CHF patients with steady-state VDD pacing. Furthermore, PP and stroke volume changes were associated with steady-state pacing in a study of CHF patients13 and have been shown to be proportional in animal studies with the FlexStim protocol.14 Thus, the effects of transient and steady-state ventricular stimulation appear to be qualitatively similar and probably arise from the same mechanism, which we surmise is a direct change of the ventricular electromechanical activation patterns. Under the conditions of our study (patients under general anesthesia and only transient pacing), the compensatory reflex response to stimulation should be small, certainly less than for steady-state stimulation of unsedated patients, and therefore the measured effects of ventricular stimulation would tend to be larger and reflect mainly the direct mechanisms.
A number of potential mechanisms to explain the positive effects of pacing in CHF patients have been discussed previously.12 These mechanisms primarily involve correcting abnormal electrical conduction15 that presumably generates mechanical dysfunction. CHF patients exhibit abnormally short or long left-side mechanical AV delays outside the normal range of 180 to 250 ms,12 resulting in increased mitral regurgitation and reduced effective preload. Similar to our findings, Nishimura et al16 and Auricchio and Salo12 showed that an appropriate AV delay is required to maximize cardiac output in CHF patients. They suggest the optimal AV delay may normalize the mechanical timing between left atrium and LV, thus reducing regurgitation and maximizing effective preload. Our observation that paced changes in PP are a positive unimodal function of AV delay is consistent with this mechanism.
CHF patients also have relatively uncoordinated contraction patterns in the LV,17 which could impair systolic function. BV pacing may improve systolic function in the short term by altering the segmental LV and interventricular septal contractile sequence in patients with depressed LV function,18 which would be expected to correlate with increased LV+dP/dt, a global measure of systolic pump performance.
There is a close relationship between abnormal contraction patterns and electrical activation disturbances, such as a wide QRS complex.19 This relationship may explain why a sufficiently wide QRS and left bundle-branch block tend to predict short-term pacing benefit. Patients may not benefit from pacing unless the conduction disorder and underlying ventricular incoordination are suitably abnormal. It may also explain why stimulation that includes the LV is far better than RV-only stimulation in a population with a high incidence of LV conduction disorders. It is our hypothesis that atrial-synchronous pacing that preexcites the ventricular wall ipsilateral to the conduction disorder restores a more normal activation pattern, thus coordinating wall motion and increasing pumping effectiveness.
Because CHF patients have heterogeneous symptoms, etiologies, and substrates, and pacing may operate through multiple mechanisms requiring individual optimization, it should not be surprising that the previously reported clinical benefit of pacing is ambiguous. Our results show that both positive and negative effects of pacing occur, at least in the short term. The outcome depends on pacing parameters and individual variables. The most important pacing parameter is pacing chamber, but AV delay significantly modulates the result. The most important variable for predicting acute pacing benefit in the present study was QRS width. If the distinction between type I and type II patients also applies to long-term pacing benefit, studies including a mixture of such patients will necessarily have mixed results. The seemingly contradictory results regarding the benefit of RV pacing3 5 6 20 might be explained by the weak and strongly patient-dependent hemodynamic effects of RV pacing and fortuitous differences in clinical patient populations.21 On the other hand, the consistently reported positive results with LV pacing7 8 9 12 13 can be understood as a consequence of the large hemodynamic response to left-sided pacing in a majority of studied patients. Thus, it would seem important to account for these individual variations in future studies of pacing therapy for CHF. Selection of the optimal pacing parameters for an individual patient can minimize the negative effects of pacing on type II patients and maximize the positive effects of pacing on type I patients, with optimized cardiac output increase being 2 to 4 times more than with suboptimal pacing.
This was an acute study performed during surgery on patients under general anesthesia, which can affect preload, afterload, and other hemodynamic variables; results may differ for unsedated patients with chronic pacing. Although this study focused on LV systolic function, the chronic response to pacing may depend also on long-term alterations of diastolic function and RV function, although they changed minimally and inconsistently during the acute protocol. The results may not extrapolate to other CHF populations, particularly those with less severe symptoms, nonsystolic dysfunction, paroxysmal or chronic atrial fibrillation, sick sinus syndrome or complete AV block, or more severe AV conduction blocks. A limitation of our protocol is that the paced-beat number has a significant effect on the hemodynamic response. The first beat in a pacing period was excluded from analysis because of expected effects of the initial AV delay shortening. However, the order effects of subsequent beats were unexpected, and although they were accounted for in the ANOVA statistical analysis, results may differ when such systematic beat-number effects are excluded. As a check of the sensitivity of the results to beat number, statistical analyses were repeated with alternative combinations of paced beats. None of our study conclusions changed for these alternative statistical treatments.
This study was supported entirely by Guidant Corporation, St. Paul, Minn, and by an unrestricted grant by Guidant to the Department of Cardiology, University Hospital Magdeburg, Germany. We are indebted to the patients who participated in this trial and to the physicians who referred their patients for study inclusion. The investigators gratefully acknowledge the nurses and colleagues at each institution of the ICU, Heart Failure Program, and Surgery without whose help and logistic support this study would not be possible. Dr Auricchio wishes to thank the CHF Research Group of Guidant for the outstanding technological support, with special thanks to Michael Hull, MS, for excellent statistical assistance.
This work was supported by a grant from Guidant Corporation.
- Received October 23, 1998.
- Revision received March 18, 1999.
- Accepted March 26, 1999.
- Copyright © 1999 by American Heart Association
Bakker PF, Meijburg H, de Jonge N, van Mechelen R, Wittkampf F, Mower M, Thomas A. Beneficial effects of biventricular pacing in congestive heart failure. Pacing Clin Electrophysiol. 1994;17:820. Abstract.
Blanc JJ, Etienne Y, Gilard M, Mansourati J, Munier S, Boschat J, Benditt DG, Lurie KG. Evaluation of different ventricular pacing sites in patients with severe heart failure: results of an acute hemodynamic study. Circulation. 1997;96:3273–3277.
Auricchio A, Stellbrink C, Block M, Bakker P, Sack S, Mortensen P, on behalf of the PATH-CHF Study Group. The Pacing Therapies for Congestive Heart Failure (PATH-CHF) Study: rationale, design and endpoints of a prospective randomized multicenter study. Am J Cardiol. 1999;83:130D–135D.
Kass DA, Maughan WL, Guo ZM, Kono A, Sunagawa K, Sagawa K. Comparative influence of load versus inotropic states on indexes of ventricular contractility: experimental and theoretical analysis based on pressure-volume relationships. Circulation. 1987;76:1422–1436.
Kass DA, Chen CH, Fetics B, Talbot M, Nevo E, Nakayama M. Ventricular function in patients with dilated cardiomyopathy is improved by VDD pacing at left but not right ventricular sites. J Am Coll Cardiol. 1998;31:31A. Abstract.
Yu Y, Ding J, Liu L, Salo R, Spinelli J, Tockman B, Pochet T, Auricchio A. Experimental validation of pulse contour methods for estimating stroke volume at pacing onset. Proc 20th Int Conf IEEE Eng Med Biol Society 1998. In press.
Vassallo JA, Cassidy DM, Marchlinski FE, Buxton AE, Waxman HL, Doherty JU, Josephson ME. Endocardial activation of left bundle branch block. Circulation. 1984;69:914–923.
Herman MV, Heinle RA, Klein MD, Gorlin R. Localized disorders in myocardial contraction: asynergy and its role in congestive heart failure. N Engl J Med. 1967;277:222–232.
Saxon LA, Kerwin WF, Cahalan MK, Kalman JM, Olgin JE, Foster E, Schiller NB, Shinbane JS, Lesh M, Merrick SH. Acute effects of intraoperative multisite ventricular pacing on left ventricular function and activation/contraction sequence in patients with depressed ventricular function. J Cardiovasc Electrophysiol. 1998;9:13–21.
Xiao HB, Roy C, Gibson DG. Nature of ventricular activation in patients with dilated cardiomyopathy: evidence for bilateral bundle branch block. Br Heart J. 1994;72:167–174.
Shinbane JS, Chu E, DeMarco T, Sobol Y, Fitzpatrick AP, Lau DM, Klinski C, Schiller N, Griffin JC, Chatterjee K. Evaluation of acute dual chamber pacing with a range of atrioventricular delays on cardiac performance in refractory heart failure. J Am Coll Cardiol. 1997;30:1295–1300.
Auricchio A, Klein H. Dual-chamber pacing in dilated cardiomyopathy: insufficient sample size, heterogeneous population and inappropriate end points may lead to erroneous conclusions. J Am Coll Cardiol. 1996;27:1548. Letter.This study evaluated left ventricular systolic function in 27 patients with severe congestive heart failure during brief periods of pacing. Patients were stimulated in the right, left, or both ventricles simultaneously over a wide range of AV delays in a repeating 5-paced/15-nonpaced beat sequence. Systolic function improved in most patients with ventricular pacing at a patient-specific AV delay. Maximum improvement in this population required left ventricular stimulation. Patients with no improvement had narrower surface QRS complexes than patients with improvement.