Author + information
- Received August 16, 2004
- Revision received October 12, 2004
- Accepted January 11, 2005
- Published online May 3, 2005.
- Alain Bernheim, MD,
- Peter Ammann, MD,
- Christian Sticherling, MD,
- Peter Burger, MD,
- Beat Schaer, MD,
- Hans Peter Brunner-La Rocca, MD,
- Jens Eckstein, MD,
- Stephanie Kiencke, MD,
- Christoph Kaiser, MD,
- Andre Linka, MD,
- Peter Buser, MD,
- Matthias Pfisterer, MD and
- Stefan Osswald, MD⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Stefan Osswald, Cardiac Unit, University Hospital, CH 4031-Basel, Switzerland.
Objectives We aimed to compare the hemodynamic effects of right-atrial-paced (DDD) and right-atrial-sensed (VDD) biventricular paced rhythm on cardiac resynchronization therapy (CRT).
Background Cardiac resynchronization therapy improves hemodynamics in patients with severe heart failure and left ventricular (LV) dyssynchrony. However, the impact of active right atrial pacing on resynchronization therapy is unknown.
Methods Seventeen CRT patients were studied 10 months (range: 1 to 46 months) after implantation. At baseline, the programmed atrioventricular delay was optimized by timing LV contraction properly at the end of atrial contraction. In both modes the acute hemodynamic effects were assessed by multiple Doppler echocardiographic parameters.
Results Compared to DDD pacing, VDD pacing resulted in much better improvement of intraventricular dyssynchrony assessed by the septal-to-posterior wall motion delay (VDD 106 ± 83 ms vs. DDD 145 ± 95 ms; p = 0.001), whereas the interventricular mechanical delay (difference between onset of pulmonary and aortic outflow) did not differ (VDD 20 ± 21 ms vs. DDD 18 ± 17 ms; p = NS). Furthermore, VDD pacing significantly prolonged the rate-corrected LV filling period (VDD 458 ± 123 ms vs. DDD 371 ± 94 ms; p = 0.0001) and improved the myocardial performance index (VDD 0.60 ± 0.18 vs. DDD 0.71 ± 0.23; p < 0.01).
Conclusions Our findings suggest that avoidance of right atrial pacing results in a higher degree of LV resynchronization, in a substantial prolongation of the LV filling period, and in an improved myocardial performance. Thus, the VDD mode seems to be superior to the DDD mode in CRT patients.
In patients suffering from severe systolic heart failure, abnormal electrical conduction contributes to asynchrony between right and left ventricular (LV) contraction and discordant wall motion within the LV (1–3). Left-bundle-branch-block-type conduction delay results in early activation of the LV septal wall and delayed activation of the LV free wall (intraventricular delay) (4–6). This intraventricular dyssynchrony impedes the mechanical systolic function of the LV (7). Additionally, the inhomogeneous course of contraction within the LV leads to a delayed onset of relaxation, which results in shortening of diastole and an impaired LV filling, which further aggravates LV dysfunction (8–13). Cardiac resynchronization therapy (CRT) has been shown to resynchronize contraction and to improve hemodynamics and clinical outcome (14–17).
To a large part, the beneficial effect of CRT depends on an appropriate atrioventricular (AV) timing delay (14,15). If the atrioventricular timing delay (AVD) is set too long, pacing-induced resynchronization of the LV lateral wall may be lost. Conversely, exceedingly short AV intervals may cause premature valve closure and result in an interruption of transmitral LV inflow during atrial contraction. Consequently, LV filling is diminished, and atrial pressure rises. The impact of right atrial pacing on these parameters is unknown.
By using the VDD pacing mode, both atria are activated via the intrinsic conduction system. In contrast, pacing of the right atrial appendage in the routinely used DDD mode leads to a delayed electrical and mechanical activation of the left atrium (18). The resulting delay of the left atrial contraction may compromise transmitral inflow during late diastole and impede LV preload compared to intrinsic left atrial activation in the VDD mode. This effect may counteract the beneficial effects of premature LV stimulation required for proper resynchronization of both ventricles.
To test this hypothesis, we analyzed the hemodynamic effects of right atrial pacing using Doppler echocardiographic assessment of the LV filling and LV systolic function in CRT patients with chronic heart failure due to LV systolic dysfunction, normal sinus rhythm, and left-bundle-branch-type intraventricular conduction delay.
Seventeen consecutive patients with an atriobiventricular pacemaker were enrolled in the study. Indications for CRT were severe heart failure due to dilated or ischemic cardiomyopathy, New York Heart Association dyspnea class ≥III, an LV ejection fraction of <35%, and presence of a left bundle branch block with a QRS interval on the surface electrocardiogram (ECG) ≥150 ms. Exclusion criteria were unstable cardiac condition and permanent atrial fibrillation. All patients gave written informed consent for participation in the study.
Eleven patients were implanted with a CRT defibrillator, and the remaining patients received a CRT pacemaker system. The operation was performed under conscientious sedation using the standard technique (19). All atrial leads were placed in the right atrial appendage. All right ventricular leads were positioned in the right ventricular apex. The coronary sinus lead was positioned in the middle cardiac vein in 1 patient, the posterior/left marginal veins in 15 patients, and in the anterior interventricular vein (diagonal branch) in 1 patient.
All patients underwent clinical examination and received a 12-lead ECG during biventricular-paced rhythm. Subsequently, transthoracic echocardiography was performed. At baseline, the programmed AVD was optimized by Doppler measurements ensuring a proper beginning of mechanical LV contraction at the end of the atrial contraction (1,20). The device was then randomly programmed into the VDD and DDD mode, respectively, and all Doppler echocardiographic measurements were repeatedly obtained in both pacing modes. In the DDD mode, a standard additional AVD of 40 ms was programmed (“pace compensation”) in order to compensate the delay of left atrial activation induced by right-atrial-appendage pacing (18,21). Additionally, in a subset of patients, the pace compensation was turned off to investigate its influence in DDD pacing. In order to ensure continuous right atrial pacing in the DDD mode, the pacing rate was set 10 beats/min above the intrinsic atrial rate. In both pacing modes, monodimensional and two-dimensional echocardiographic evaluations as well as Doppler measurements were obtained.
Transthoracic echocardiography was performed using a standard echocardiographic Doppler system (Toshiba Aplio 80, Tokyo, Japan). Patients were examined in the left lateral recumbent position. Standard parasternal long- and short-axis as well as apical views were used for data acquisition. Time velocity integral in the left ventricular outflow tract (LVOT-TVI) was acquired from the apical long-axis view by placing the pulsed-wave Doppler at the center of the LVOT and measured by tracing the Doppler spectral profile. Left ventricular ejection time (ET) was calculated by measuring the time from aortic valve opening to aortic valve closure. Isovolumic contraction time (ICT), the time from mitral valve closure to aortic valve opening, and isovolumic relaxation time (IVRT), the interval from aortic valve closure to mitral valve opening, were measured by simultaneous Doppler recording of the aortic and mitral flow. For assessment of the LV filling phase, pulsed-wave Doppler was used with the Doppler sample volume set at the tip of the mitral valve leaflets. From the recorded Doppler pattern, the LV filling time (E + A dur) was measured, and LV filling delay was calculated from the delay between the beginning of the Q-wave on the surface ECG and the onset of the E-wave (E-delay). All systolic and diastolic time intervals were adjusted for differences in heart rate ([time intervalDDDin ms] × [RR intervalVDD/RR intervalDDD]).
Intraventricular asynchrony was assessed by two methods: 1) M-mode echocardiographic imaging was used in the short-axis view at the papillary muscle level. The difference of the maximal inward motion of the septum and of the maximal displacement of the left posterior wall was used to calculate the intraventricular delay on the midventricular level (septal-to-posterior wall motion delay [SPWMD]) (22); and 2) tissue Doppler imaging (TDI) was performed from the apical four-chamber view placing the sample in the lateral and septal mitral annulus, respectively. The time from the onset of the Q-wave on the surface ECG to the onset of the regional systolic motion was measured in the basal segments of the interventricular septum and the left lateral wall and the difference calculated to assess TDI-derived intraventricular delay on the mitral annular level.
Interventricular electromechanical delay was evaluated using the delay between the onsets of pulmonary and aortic outflow, calculated as the difference between the time from the onset of the Q-wave on the surface ECG to simultaneously recorded onset of aortic flow and the time from the beginning of the Q-wave to the onset of pulmonary artery flow (22).
Myocardial performance index was calculated as the sum of ICT and IVRT divided by the ET (23).
Echocardiographic data were digitally stored and independently analyzed by two investigators with an interobserver variability of 7%. Reported is the average of the two.
Data are presented as the mean value ± one SD or as frequencies (percentage). The method of Kolmogorov and Smirnov was used to test if measured data were distributed nomally. In case of normally distributed data, the paired student ttest was used for calculations. In case of a non-normal distribution, the Wilcoxon signed rank test was performed. A p value of <0.05 was considered statistically significant. All calculations were performed with a commercially available statistical package (Statview 5.0, SAS Institute Inc., Cary, North Carolina).
Baseline clinical characteristics of the patients enrolled in the study are depicted in Table 1.Seventeen patients with a mean age of 64 ± 10 years were included. Twelve patients were male. Seven patients suffered from coronary artery disease and 10 from dilated cardiomyopathy. Eleven patients (65%) were in New York Heart Association funtional class III before device implantation; the other six (35%) were in New York Heart Association functional class IV. The mean QRS duration of the left bundle branch block without pacing was 179 ± 21 ms; the LV ejection fraction was 18 ± 6% before CRT. The intrinsic inter- and intraventricular delays were 49 ± 20 ms and 181 ± 98 ms, respectively. The mean ventricular cycle length was 904 ± 161 ms in the VDD group and 763 ± 100 ms in the DDD group.
Acute effects of VDD and DDD pacing on Doppler echocardiographic parameters
After a median of 10 months (range 1 to 46 months), an optimized AV delay of 111 ± 34 ms (range 40 to 150 ms) was determined echocardiographically in the VDD pacing mode as previously described.
Differences between VDD and DDD pacing with respect to systolic and diastolic time intervals are depicted in Table 2.Isovolumic contraction time, systolic ET, and IVRT were significantly shorter in VDD as compared to DDD mode even after adjustment to changes in heart rate. Shortening of these time intervals resulted in an earlier onset of the E-wave (E-delay: VDD 526 ± 55 ms vs. DDD 638 ± 114 ms; p < 0.0001). Importantly, as a consequence, the rate-adjusted LV filling time was significantly prolonged with VDD compared to DDD pacing, contributing by 50 ± 6% to the cardiac cycle in the VDD compared to 41 ± 8% in the DDD mode (p = 0.0002) (Figs. 1 and 2).⇓⇓
In a subset of five patients (29%), measurements were repeated without pace compensation in the DDD mode. When pace compensation was inactivated, LV filling times remained significantly shorter than in the VDD mode (VDD 483 ± 97 ms vs. DDD pace compensation “off” 442 ± 103 ms; p = 0.048). There was a nonsignificant difference in the LV filling period in the DDD mode with and without pace compensation (DDD pace compensation “off” 442 ± 103 ms vs. DDD pace compensation “on” 407 ± 85 ms; p = 0.08).
In five patients (29%) a complete fusion of early diastolic and atrial filling was observed in the DDD mode, whereas in the VDD mode no patients showed a fusion between the E- and A-wave (Figs. 3A and 3B).This finding precluded quantitative E/A ratio comparisons between the modes, because the A-wave could not be differentiated in all patients in the DDD mode.
Myocardial performance index was significantly lower with VDD pacing, indicating an amelioration of LV function in this mode (VDD 0.60 ± 0.18 vs. DDD 0.71 ± 0.23; p = 0.0044). Furthermore, a beneficial effect of VDD pacing was observed on LVOT-TVI (VDD 21.3 ± 3.7 cm vs. DDD 18.3 ± 3.3 cm, p < 0.0001).
While the interventricular delay did not differ between both modes (VDD 20 ± 21 ms vs. DDD 18 ± 17 ms; p = NS), we found significant differences in the intraventricular delays in favor of the VDD mode when assessed by the SPWMD on the midventricular level (VDD 106 ± 83 ms vs. DDD 145 ± 95 ms; p = 0.001) and by TDI-derived measures of the interventricular septum and the left lateral wall performed on the mitral annular level (VDD 38 ± 39 ms vs. DDD 83 ± 52 ms; p = 0.03).
The main finding of our study is that right atrial pacing abolished some of the beneficial effects of CRT. This was reflected by a worsening of LV synchronization and a highly significant reduction of LV filling times in the DDD compared to the VDD mode. Complete fusion of early (passive) and late (atrial) transmitral inflow was a common observation during active right atrial pacing. Overall, cardiac performance was significantly better in the VDD pacing mode with avoidance of right atrial pacing.
Hemodynamic consequences of conduction abnormalities
Early activation of the septal wall and delayed inward motion of the LV free wall caused by intraventricular conduction delay worsens systolic dysfunction in chronic heart failure (24). Apart from deteriorating systolic LV function, these conduction abnormalities also add to diastolic dysfunction. The LV filling period is shortened because the mitral valve is kept closed at the time of left atrial contraction by a delayed intraventricular pressure decay, which is caused by the prolonged LV lateral wall contraction. These mechanisms were extensively explored in previous studies in patients with heart failure due to dilated cardiomyopathy. The investigators found that the LV filling time was significantly diminished, when left bundle branch block or a first-degree AV block were present (11,12). Furthermore, intraventricular conduction delays not only contribute to delayed LV activation but also to slowed and delayed muscle relaxation, which further aggravates passive LV filling. As a consequence, the overall duration of LV diastole is shortened significantly (8,12,24,25).
The fact that early LV filling (represented by the E-wave in transmitral Doppler) is delayed, but atrial contraction and the resulting late filling phase (corresponding to the A-wave) are not, results in a “single-phase” transmitral diastolic flow pattern due to fusion of the two diastolic flow components (Fig. 3A). This further compromises LV filling and the preload of the LV as a consequence of a decreased left atrial contraction efficacy (8,11,12). In patients with critical heart failure, where appropriate LV filling may be crucial to maintain hemodynamic stability via the Frank-Starling mechanism, this fusion phenomenon is likely to have a deleterious effect on cardiac performance (26).
Effect of biventricular pacing on LV performance
The influence of CRT on LV filling has previously been investigated (27,28). Compared to intrinsic rhythm, active biventricular pacing resulted in a significantly increased LV filling time. This finding was mainly attributed to a shortening of the ICT. This effect on ICT was obtained by an optimization of the left-sided AV delay. Furthermore, the increase in diastolic filling time gained by advancing LV systole led to a better separation of the early (passive) and the late (atrial) filling period (27,28).
Our data comparing VDD versus DDD biventricular pacing also showed a significant difference in ICT between the two modes with a prolongation during DDD pacing. However, the shortening of ICT achieved in the VDD mode was substantially smaller than the observed additional prolongation of LV diastole. This suggests that the shortening of ICT was not the only component of the cardiac cycle that contributed to the improved LV filling observed during VDD pacing. In fact, we found that the ET as well as the IVRT were also shorter in the VDD compared to the DDD mode (Figs. 1A and 1B).
Suggested mechanisms for the observed adverse effects of DDD pacing
There are basically two mechanisms by which the adverse hemodynamic effects of DDD pacing may be explained. First, by atrial pacing via the implanted lead in the right atrial appendage, activation of the left atrium is artificially delayed by at least 60 to 100 ms. This is shifting the A-wave towards the onset of systole, which results in an A-wave cutoff and decreased LV filling (Fig. 4A).Second, activation of an AVD “pace compensation” of 40 ms in the DDD mode, which was basically programmed to prevent the above mechanism and to preserve left atrial contribution to LV filling, resulted in some loss in LV synchronization due to fusion activation of the LV between intrinsic- and biventricular-paced rhythm. This effect also accounted for less beneficial hemodynamic effects of CRT in the DDD mode (18,20). This hypothesis is supported by the observed prolongation of intraventricular delay during DDD pacing, although the interventricular delays remained similar in the two pacing modes.
Given the two competing mechanisms, the main dilemma of DDD pacing using right atrial appendage pacing cannot be solved; pacing from the right atrial appendage inevitably delays left atrial activation, which has to be compensated one or the other way. Therefore, either way, LV inflow time will be shortened if the right atrium is being paced; if “pace compensation” is inactivated, atrial contribution to LV filling will be significantly impaired (A-wave cutoff) with the benefit of fully preserved LV synchronization. If “pace compensation” is activated, the onset of passive mitral inflow (E-wave) will be delayed with the hemodynamic benefit of preserved atrial contribution to LV filling (Figs. 4A and 4B). As a consequence, LV filling time remains always shorter in the DDD mode when the right atrium is paced from the appendage, whether a pace compensation is programmed or not. However, other right atrial lead placements (e.g., in a septal position) may not be associated with such a delay of left atrial activation and, therefore, might be preferable if DDD pacing is envisioned (29).
As suggested by the present acute study, by active right atrial appendage pacing, the previously reported amelioration of LV filling and myocardial performance achieved by biventricular pacing (27,28) may substantially be diminished. Based on our findings, it can be speculated that an important part of the beneficial hemodynamic effects achievable with CRT may be abolished by the inappropriate use of DDD pacing.
One limitation of our study is that we only tested the different pacing modes in the acute setting. Hence, the chronic effects on LV performance cannot be predicted by our study. This needs further evaluation in the long-term.
Furthermore, all measurements were obtained at rest. We did not study the response to exercise, but it may be expected that the complex timing and flow interrelations of LV diastole are influenced in a similar way under exercise conditions. Finally, the potential drawbacks of VDD programming, such as possible occurrence of symptomatic sinus bradycardia or, as recently reported, a possible increase in atrial arrhythmias (30) remains to be considered. Therefore, for a subgroup of patients (i.e., those with relevant sinus bradycardia), DDD pacing still may be indicated. However, one should be aware that prolongation of the programmed AVD by programming any kind of automatic “pace-compensation” feature might have deleterious effects on the degree of achievable LV resynchronization. In order to resolve these issues and to prove the clinical implications of our findings, further studies with clinical cross-over design comparing the two pacing modes over time are warranted.
Clinical implications and conclusions
Our data provide strong evidence that intrinsic atrial activation is important to improve diastolic filling and is superior to right atrial paced rhythm for maximization of the beneficial effects of CRT. Based on our results, assessed in an acute setting, the VDD mode seems to be the superior pacing mode in CRT patients. However, long-term clinical follow-up trials are needed to confirm a lasting hemodynamic benefit.
Drs. Bernheim and Ammann contributed equally to this work.
- Abbreviations and acronyms
- atrioventricular delay
- cardiac resynchronization therapy
- ejection time
- isovolumic contraction time
- isovolumic relaxation time
- left ventricle/ventricular
- time velocity integral in the left ventricular outflow tract
- septal-to-posterior wall motion delay
- tissue Doppler imaging
- Received August 16, 2004.
- Revision received October 12, 2004.
- Accepted January 11, 2005.
- American College of Cardiology Foundation
- Kass D.A.
- Murkofsky R.L.,
- Dangas G.,
- Diamond J.A.,
- et al.
- Park R.C.,
- Little W.C.,
- O’Rourke R.A.
- Xiao H.B.,
- Brecker S.J.,
- Gibson D.G.
- Xiao H.B.,
- Lee C.H.,
- Gibson D.G.
- Nishimura R.A.,
- Hayes D.L.,
- Holmes D.R. Jr..,
- et al.
- Shinbane J.S.,
- Chu E.,
- DeMarco T.,
- et al.
- Auricchio A.,
- Stellbrink C.,
- Block M.,
- et al.
- Kass D.A.,
- Chen C.H.,
- Curry C.,
- et al.
- Leclercq C.,
- Cazeau S.,
- Le Breton H.,
- et al.
- Nelson G.S.,
- Berger R.D.,
- Fetics B.J.,
- et al.
- Cha Y.M.,
- Nishimura R.A.,
- Hayes D.L.
- Janosik D.L.,
- Pearson A.C.,
- Buckingham T.A.,
- et al.
- Pitzalis M.V.,
- Iacoviello M.,
- Romito R.,
- et al.
- Grines C.L.,
- Bashore T.M.,
- Boudoulas H.,
- et al.
- Gillebert T.C.,
- Sys S.U.,
- Brutsaert D.L.
- Holubarsch C.,
- Ruf T.,
- Goldstein D.J.,
- et al.
- Yu C.M.,
- Chau E.,
- Sanderson J.E.,
- et al.
- St John Sutton M.G.,
- Plappert T.,
- Abraham W.T.,
- et al.
- Knight B.P.,
- Desai A.,
- Coman J.,
- et al.