OUP user menu

Prevalence and determinants of left ventricular systolic dyssynchrony in patients with normal ejection fraction received right ventricular apical pacing: a real-time three-dimensional echocardiographic study

Fang Fang, Joseph Yat-Sun Chan, Gabriel Wai-Kwok Yip, Jun-Min Xie, Qing Zhang, Jeffrey Wing-Hong Fung, Yat-Yin Lam, Cheuk-Man Yu
DOI: http://dx.doi.org/10.1093/ejechocard/jep171 109-118 First published online: 20 November 2009


Aims Right ventricular apical (RVA) pacing may induce mechanical dyssynchrony. However, its impact on patients with normal ejection fraction (EF) is not fully understood. This study examined the prevalence and predictors of RVA pacing-induced systolic dyssynchrony by real-time three-dimensional echocardiography (RT3DE), and evaluated its impact on left ventricular (LV) function.

Methods and results Ninety-three patients with sinus node dysfunction and normal EF (>50%) received RVA-based dual-chamber pacing were assessed by RT3DE during RVA pacing (V-pace) and intrinsic conduction (V-sense). Systolic dyssynchrony was evaluated using the standard deviation of the time to minimal regional volume of 16 LV segments (Tmsv-16SD), and a cutoff value of 16 ms was determined from 93 normal controls. Systolic dyssynchrony was induced in 49.5% of patients at V-pace with significant increase in LV end-systolic volume (LVESV), decrease in EF, and worsening of Tmsv-16SD (all P < 0.001). Furthermore, patients who developed dyssynchrony had larger LVESV (P < 0.001), lower EF (P < 0.001) at V-pace mode, and higher cumulative percentage of RVA pacing in the past 6 months (P < 0.001) than those without systolic dyssynchrony. In multivariate logistic regression analysis, independent predictors of developing LV systolic dyssynchrony during V-pace included a low normal EF at V-sense, pre-existing LV hypertrophy, and cumulative RVA pacing >40% in the past 6 months.

Conclusion For patients with preserved EF received RVA pacing, half of them would develop systolic dyssynchrony which was associated with EF deterioration and LV enlargement. A low normal EF, a high cumulative percentage of RVA pacing, and pre-existing LV hypertrophy were predictors of developing dyssynchrony.

  • Right ventricular pacing
  • Dyssynchrony
  • Real-time three-dimensional echocardiography


Right ventricular-based pacing is currently the commonest route of ventricular lead implantation during permanent pacemaker implantation for patients with bradycardia. Among various pacing sites, right ventricular apical (RVA) pacing is commonly adopted by virtue of its easy accessibility and a low acute complication rate. In recent years, however, it has been suggested that RVA pacing may exert deleterious effect on left ventricular (LV) systolic function, even in subjects without pre-existing systolic dysfunction.16 In the MOST trial, the cumulative RVA pacing >40% conferred a 2.6-fold increased risk of heart failure hospitalization when compared with patients with less pacing despite preserved atrioventricular synchrony.7 The DAVID trial further substantiated that high cumulative percentage of right ventricular pacing is particular harmful in patients received implantable cardioverter defibrillators with deleterious effect on LV ejection fraction (LVEF).8 In patients with congenital complete heart block and atrial fibrillation who received atrioventricular nodal ablation followed by RVA pacing, there was reduction of LVEF and worsening of functional class when compared with normal controls and tissue Doppler imaging revealed a higher degree of intra- and inter-ventricular dyssynchrony.4,9 These findings suggested that electromechanical delay as a result of RVA pacing-induced left bundle branch block (LBBB) pattern may result in systolic dyssynchrony which, in turn, may play a pivotal role in deterioration of LV systolic function. In heart failure population, tissue Doppler imaging studies observed that intra-ventricular dyssynchrony occurred in ∼70% of patients with a wide QRS who are primarily caused by LBBB.1012 However, the prevalence of dyssynchrony as a result of RVA-induced LBBB in the setting of preserved LV systolic function is not known. In patients with normal LVEF, the impact of RVA pacing in the induction of mechanical dyssynchrony could be different from those with poor LVEF. Furthermore, in the setting of normal LVEF, factors contributing to the occurrence of systolic dyssynchrony have not been determined. Recently, real-time three-dimensional echocardiography (RT3DE) has been validated to be an accurate and sensitive tool in the assessment of LV volume and LVEF.13 The offline analysis of 3D full-volume data has also been confirmed useful in the assessment of systolic dyssynchrony, and its accuracy has been validated when compared with tissue Doppler velocity imaging.14,15 Therefore, the objectives of the study were to determine the prevalence of RVA pacing-induced systolic dyssynchrony in patients with normal LVEF; to investigate whether this will result in reduction of systolic function; and to explore the determinants of developing systolic dyssynchrony.


Study population

This cross-sectional study enrolled 93 patients with sinus node dysfunction who had been implanted with dual chamber pacemakers for a minimum of 6 months. The mean pacemaker implantation period was 57 ± 47 months (range: 6 months to 13 years). Patients were selected on the basis of having a RVA-based dual chamber pacing system and availability of ventricular intrinsic sensing by programming the atrioventricular interval. These patients had normal LVEF of >50%. The exclusion criteria included atrial lead implanted at site other than right atrial appendage, ventricular lead not placed in RV apex, complete heart block who are pacing-dependent, history of clinical heart failure or acute coronary syndrome, LV systolic dysfunction with LVEF <50%, atrial fibrillation during echocardiographic examination, significant valvular dysfunction, and other cardiac conditions which are not safe to program the pacemaker to ventricular pacing-off mode temporarily, as well as cardiac disease with regional wall motion abnormalities or impaired cardiac systolic function. Subjects were also excluded from the analysis if it was not possible to obtain adequate 3D images. A control group of 93 healthy volunteers were studied. They had no history of systemic or chronic illness, and had normal physical examination, ECG as well as echocardiographic examination. The study protocol was approved by Institutional Review Board and written informed consents were obtained from all patients.

Study protocol

Echocardiography with RT3DE was performed at rest during intrinsic ventricular conduction (V-sense) and RVA pacing at DDDR mode (V-pace) in a random order. For examination during V-sense, atrioventricular interval was prolonged progressively to ensure intrinsic conduction without fusion. When programmed to V-pace, the atrioventricular interval was progressively shortened until right ventricular pacing occurred without fusion, which was confirmed by the LBBB pattern on ECG, and the AV interval was then kept at least 20 ms shorter than the intrinsic atrioventricular delay. The atrial pacing rate was set at a low value (typically ≤50 bpm) to allow intrinsic rate with atrial sensing unless patients were dependent on atrial pacing. The heart rate was therefore maintained in the same value during V-sense and V-pace. After pacemaker programming, patients were allowed to rest for at least 15 min before echocardiographic examinations were repeated. The control group was studied during sinus rhythm without pacing intervention.


Standard echocardiography with RT3DE was performed (iE33, Phillips, Andover, MA, USA). LF diastolic function was assessed by transmitral flow pattern by pulse-wave Doppler echocardiography. LV mass was estimated by the Devereux-modified American Society of Echocardiographic cube formula and was divided by the body surface area to derive the LV mass index.16 LV hypertrophy was defined as LV mass index >134 g/m2 for men and >110 g/m2 for woman.16

RT3DE was performed at the apical four-chamber view with different pacing mode by acquiring pyramidal full volume images of LV with a matrix-array transducer (X3-1, 1.9/3.8 MHz). The image was adjusted to optimize the orthogonal 2D and then 3D image qualities with modified gain setting and compression controls as well as depth and lateral gain compensation to optimize full volume acquisition. Patients were instructed to hold the breath to minimize artefacts induced by breathing during full volume acquisition which was triggered to the R wave on the ECG of every cardiac cycle, resulting in a total acquisition time of 4 heart beats. Frame rate was kept at ≥20 Hz when sector width was optimized. LV full volume images with clear endocardial border were stored digitally and transferred to the work station for off-line analysis.14,15,1720

Quantitative analysis of RT3DE

Quantitative analysis of RT3DE images was performed offline in a blinded fashion by dedicated software (Q-Lab 6.0, Philips, Andover, MA, USA). First, from the automatically cut planes that consisted of non-foreshortened end-diastolic apical two- and four-chamber views, five anatomic points were manually defined which included two points to identify the mitral valve annulus in each of the two apical views and one point to identify the apex in either view. A detection procedure was followed by the software automatically to trace the LV endocardial border according to the preset mathematic model. The same procedure was repeated in the end-systolic frame which was identified as the frame just ahead of mitral valve closure. The surface detection was edited manually if tracing was suboptimal. Thus, global LV volumes with time–volume curves were derived which included global and segmental LV volumes. The regional LV volumetric data were used to derive global LVEF. The global volume was then divided into 17 standard myocardial segments as defined by the American Society of Echocardiography (six basal, six middle, and five apical segments). Systolic dyssynchrony indices were calculated as the standard deviation of time interval between R wave and minimal regional volume for 16 segments (Tmsv-16SD) (six basal, six middle, and four apical segments, excluding the apical cap, segment 17).14,15 LV systolic dyssynchrony could be also analysed with parametric imaging, a bull eye mapping of regional contraction timings with colour coding of measured Tmsv in each region, in the software based on the time–volume data. Timing reference, defined by the global Tmsv which varies from patient to patient, is coded in green. Early segments are coded in blue, whereas late segments are coded in red/yellow (Figure 1). Inter- and intra-observer variability for measurement of LVEF and systolic dyssynchrony index derived from RT3DE were obtained in 15 randomly selected patients by two independent blinded investigators and by the same observer at two different occasions within 1 week. The inter- and intra-observer variability for LVEF was 6.7 and 5.0%, respectively, and the corresponding figures for Tmsv-16SD were 6.5 and 4.4%, respectively.

Figure 1

Illustration of left ventricular (LV) systolic dyssynchrony induced by RVA pacing in parametric imaging. LV regional timing is colour-coded in a polar plot with the timing reference defined by global Tmsv coded as green colour. Blue-coded segments are early contraction region, and red-coded segments are delayed contraction regions. (A) During V-sense, the homogenous green colour in the bull eye presentation suggests synchronous contraction with the Tmsv-16SD 6 ms. (B) When programmed to V-pacing, Tmsv-16SD increased to 39 ms and the basal- and mid-lateral segments were the late regions with the red colour.

Statistical analysis

Data were analysed using a statistical software program (SPSS for windows, Version 11.5, SPSS Inc., Chicago, IL, USA). Paired sample t-test and independent t-test were used to compare the difference in mean for parametric data as appropriate. Categorical variables were analysed with Pearson Chi-square test. Analysis of covariance was used with general linear model to examine the effect of age on the dependent variables. Univariate followed by multivariate logistic regression analyses were used to determine potentially independent predictors of LV systolic dyssynchrony during V-pace mode from the list of clinical and echocardiographic covariates. Data were expressed as mean ± SD and a P-value <0.05 was considered statistically significant.


Demographic parameters in patient group

There was no difference in gender distribution (male/female: 27/66 vs. 32/61, χ2 = 0.62, P = NS) and heart rate (61 ± 9 vs. 63 ± 8 bpm, P = NS) between the patient group and normal controls. The age was observed to be older in the patient group than in controls (68.7 ± 11.2 vs. 53.4 ± 10.1 years, P < 0.001). However, age was not a confounder of systolic asynchrony according to analysis of covariance, and hence unadjusted P-values were presented. In the patient group, hypertension was present in 61.3%, diabetes mellitus in 19.4%, and coronary heart disease in 25.8% of the population. For medications, beta-blockers were prescribed in 43%, angiotensin-converting enzyme inhibitors or angiotensin receptor blockers in 17%, statin in 27%, and calcium channel antagonist in 30% of the patients. The mean QRS duration at V-sense was 95 ± 17 ms.

Comparison between V-sense and V-pace in the patient group

Table 1 shows the comparison between V-sense and V-pace in the patient group. During V-pace, the QRS duration was prolonged dramatically, while the LV end-systolic volume (LVESV) was increased significantly resulting in the reduction of LVEF and LV stroke volume when compared with the V-sense mode (all P < 0.001) (Figure 2). Furthermore, systolic dyssynchrony was increased as reflected by the worsening of Tmsv-16SD (P < 0.001). There was no change in LVEDV or parameters of diastolic function (Table 1).

Figure 2

Global volumetric curves derived from all 17 left ventricular segments in a patient showing that the left ventricular end-systolic volume (ESV) increased significantly with reduction of ejection fraction (EF) and stroke volume (SV), but no change in left ventricular end-diastolic volume (EDV) when compared between V-sense (A) and V-pace mode (B).

View this table:
Table 1

Comparison of echocardiographic parameters between V-sense and V-pace modes in patients received dual chamber pacing

QRS duration, ms95 ± 17162 ± 34<0.001
Heart rate, bpm61 ± 962 ± 7NS
LVEDV, mL59.7 ± 14.460.0 ± 13.4NS
LVESV, mL23.0 ± 6.025.0 ± 6.6<0.001
LV stroke volume, mL36.7 ± 10.235.1 ± 10.00.001
LVEF, %61.3 ± 6.258.5 ± 7.7<0.001
Tmsv-16SD, ms11.6 ± 5.919.8 ± 8.4<0.001
Transmitral E velocity, m/s0.70 ± 0.200.70 ± 0.20NS
Transmitral A velocity, m/s0.79 ± 0.240.79 ± 0.22NS
E/A ratio0.94 ± 0.360.95 ± 0.42NS
Deceleration time of E wave, ms224 ± 50222 ± 61NS
IVRT, ms106 ± 27105 ± 32NS
Left atrial diameter, mm34 ± 635 ± 6NS
  • IVRT, isovolumic relaxation time; LVEDV, left ventricular end-diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end-systolic volume; Tmsv-16SD, standard deviation of time to minimum regional volume for all 16 segments.

Prevalence of RVA pacing-induced systolic dyssynchrony during V-pace

A cutoff value of Tmsv-16SD of >16 ms was used to define the presence of systolic dyssynchrony which was derived from +2SD above the mean value of the 93 normal controls. Patients were further stratified into two groups according to whether they had developed significant systolic dyssynchrony when programmed to V-pace. There were 46 (49.5%) patients who developed systolic dyssynchrony during V-pace. This group of patients was found to have enlargement of LVESV and reduction of LVEF and stroke volume when programmed from V-sense to V-pace (Table 2), but not in the group who did not develop V-pace induced dyssynchrony. As a result, the amplitude of adverse decrease in LVEF (−5.2 ± 2.6 vs. −1.1 ± 3.0%, P < 0.001), increase in LVESV (2.9 ± 2.4 vs. 1.0 ± 2.1 mL, P < 0.001), and decrease in LV stroke volume (−3.1 ± 3.2 vs. −0.2 ± 5.4 mL, P = 0.003) were significantly greater in patients who developed significant dyssynchrony when programmed from V-sense to V-pace mode (Table 2). At V-sense, patients with pacing-induced systolic dyssynchrony also had significantly larger LVESV (P = 0.001), lower LVEF (P < 0.001) as well as lower LV stroke volume (P = 0.007) when compared with those who did not develop systolic dyssynchrony (Table 2) (Figures 3 and 4).

Figure 3

A patient with induction of left ventricular systolic dyssynchrony during V-pace. (A) Left ventricle contraction was synchronous during V-sense as reflected by the congregated points of minimal regional volume (red triangles). Systolic dyssynchrony index was derived from the standard deviation of the time interval between R wave and minimal regional volume of 16 segments (Tmsv-16SD) which was 10 ms. (B) The same patient showed deterioration of LV dyssynchrony during V-pace with more scattered timings to minimal regional volume. The Tmsv-16SD was prolonged to 38 ms.

Figure 4

A patient without evidence of developing LV systolic dyssynchrony during V-pace when compared between V-sense (A) and V-pace modes (B). Regional volumetric curves measured by RT3DE for both pacing modes illustrated that the points of minimal regional volume congregated at end-systolic period. The Tmsv-16SD was 8 ms during V-sense and 10 ms during V-pace.

View this table:
Table 2

Comparison of echocardiographic parameters in patients with and without developed systolic dyssynchrony during right ventricular apical pacing

No dyssynchrony group (n = 47)Dyssynchrony group (n = 46)P-value
Sex, male/female11/3616/30NS
Age, year66 ± 1071 ± 120.03
Body surface area, m21.58 ± 0.141.59 ± 0.14NS
V-pace in last 6 months, %24 ± 2554 ± 38<0.001
QRS duration at V-sense, ms95 ± 1696 ± 19NS
QRS duration at V-pace, ms168 ± 34*156 ± 34*NS
QRS duration change, ms73 ± 3360 ± 35NS
LVEDV at V-sense, mL60.6 ± 15.858.9 ± 12.8NS
LVEDV at V-pace, mL61.3 ± 14.759.5 ± 12.1NS
LVEDV change, mL0.7 ± 6.5-0.2 ± 4.7NS
LVESV at V-sense, mL21.0 ± 6.224.9 ± 5.30.001
LVESV at V-pace, mL22.0 ± 6.127.9 ± 5.8*<0.001
LVESV change, mL1.0 ± 2.12.9 ± 2.4<0.001
LV stroke volume at V-sense, mL39.5 ± 10.533.9 ± 9.20.007
LV stroke volume at V-pace, mL39.3 ± 9.530.8 ± 8.7*<0.001
LV stroke volume change, mL−0.2 ± 5.4−3.1 ± 3.20.003
LVEF at V-sense, %65.3 ± 4.157.1 ± 5.1<0.001
LVEF at V-pace, %64.2 ± 4.052.6 ± 6.0*<0.001
LVEF change, %−1.1 ± 3.0−5.2 ± 2.6<0.001
Left atrial diameter at V-sense, mm34 ± 634 ± 6NS
Left atrial diameter at V-pace, mm34 ± 635 ± 5NS
  • *P < 0.001; P < 0.01; P < 0.05 between V-sense and V-pace in each subgroup.

  • Abbreviations as in Table 1.

When compared between V-sense and V-pace mode, LVEF decreased to <50% in 11 out of 46 (23.4%) patients in those who developed systolic dyssynchrony, but in none of the 47 patients without developed systolic dyssynchrony (χ2 = 12.46, P < 0.001). Figure 1 shows the parametric imaging of a patient who developed systolic dyssynchrony during V-pace, while Figure 5 shows another patient who did not. Interestingly, the group with pacing-induced dyssynchrony had a higher percentage of cumulative ventricular pacing in the past 6 months than those without pacing-induced dyssynchrony (54 ± 38 vs. 24 ± 25%, P < 0.001). There was no difference in QRS duration for both V-sense and V-pace in the two groups (Table 2).

Figure 5

Parametric imaging in the same patient as Figure 4 without evidence of developing left ventricular systolic dyssynchrony when programmed from V-sense (A) to V-pace mode (B). The bull eye presentation reflected the homogenous green colour code in the whole left ventricle.

Determinants of developing LV systolic dyssynchrony during V-pace

In view of the fact that significant systolic dyssynchrony during V-pace only occurred in half of the patients, potential determinants of dyssynchrony during V-pace were examined by using univariate and multivariate logistic regression analyses. Potential clinical characteristics and echocardiographic determinants of developing systolic dyssynchrony during V-pace were examined. In the univariate model, there were no differences in gender, QRS duration at V-sense, and prevalence of hypertension, diabetes and ischaemic heart diseases between those with and without developing systolic dyssynchrony. However, an older age, a lower LVEF at V-sense, a higher cumulative percentage of pacing in the past 6 months (defined as >40%), as well as pre-existing LV hypertrophy were found correlated significantly with more severe systolic dyssynchrony during V-pace (all P < 0.05) (Table 3). In the multivariate logistic regression analysis, independent determinants of systolic dyssynchrony during V-pace included pre-existing LV hypertrophy (P = 0.03), a lower LVEF at V-sense (P < 0.001), and cumulative amount of RVA pacing >40% (P = 0.005) (Table 3).

View this table:
Table 3

Univariate and multivariate logistic regression analyses to examine for significant determinant(s) of systolic dyssynchrony using Tmsv-16SD>16 ms at V-pace as the dependent covariate

ParameterUnivariate modelMultivariate model
β (95% confidence interval)P-valueβ (95% confidence interval)P-value
Age, years1.043 (1.003–1.085)0.0341.058 (0.993–1.128)NS
Sex, M/F1.745 (0.704–4.327)NS1.336 (0.254–7.340)NS
QRS duration, ms1.001 (0.978–1.025)NS0.953 (0.903–1.005)NS
Cumulative ventricular pacing in the recent 6 months >40%5.474 (2.138–14.012)<0.0019.748 (1.957–48.547)0.005
Hypertension0.966 (0.419–2.224)NS0.992 (0.225–4.387)NS
Diabetes1.029 (0.407–2.606)NS2.337 (0.319–17.135)NS
Ischaemic heart disease1.354 (0481–3.809)NS0.679 (0.097–4.772)NS
LVEF, %0.713 (0.624–0.813)<0.0010.726 (0.615–0.844)<0.001
Pre-existing LV hypertrophy5.318 (2.045–13.830)<0.0018.187 (1.1233–54.387)0.035
  • Abbreviations as in Table 1.


The present study consisted of a reasonably large population of patients with preserved LV systolic function received dual chamber pacing. The important observation was that only half of them would develop LV systolic dyssynchrony during RVA pacing. Furthermore, this group of patients, but not those without developing systolic dyssynchrony, was associated with significant decrease in LVEF and stroke volume as well as increase in LVESV when programmed from V-sense to V-pace mode. We also identified independent determinants of developing systolic dyssynchrony, which included a low normal LVEF at V-sense mode, LV hypertrophy as well as a high cumulative amount of ventricular pacing in the past 6 months of >40%.

The prevalence of LV systolic dyssynchrony during RVA pacing

In patients with heart failure, a wide QRS complex (typically a LBBB pattern) has well been demonstrated to cause LV electromechanical delay.1012,15 As a result, there is heterogeneous electrical activation with functional regions of block within the LV rendering the occurrence of regional dispersion of contraction, i.e. systolic mechanical dyssynchrony. Despite our understanding of electromechanical delay, not all patients with wide QRS complex developed systolic dyssynchrony. Previous studies observed that in patients with heart failure due to systolic dysfunction, systolic dyssynchrony occurred in between 62 and 89% in those with prolonged QRS duration of >120 ms.1012,15 This variation is likely related to the different aetiology of the dyssynchrony.

For patients with preserved systolic function, the commonest cause of LBBB is iatrogenic in relation to pacing therapy for bradycardia, commonly due to sinus node dysfunction and advanced atrioventricular block. Whether RVA pacing-induced LBBB will have the same impact on the development of electromechanical delay in those patients with normal LV systolic function remains to be addressed. Although two small studies observed that systolic dyssynchrony could be induced by RVA pacing, it was not reported how prevalent the problem existed.21,22 In the current study, systolic dyssynchrony was observed by RT3DE to occur in only half of patients received RVA pacing when LVEF was preserved. This is in contrast to a previous RT3DE study that reported a prevalence of 62% (18 out of 29 patients) in heart failure group with LVEF <50%.15 However, our result corroborates the finding by Pastore et al.23 using tissue Doppler imaging that LV dyssynchrony was exhibited in 44.9% of patients with normal LVEF after acute RVA pacing. Therefore, it appears that patients with normal systolic function are more resistant to develop systolic dyssynchrony than those with impaired LVEF. Further studies are needed to address whether the diseased myocardium with LV systolic dysfunction is more vulnerable to develop electromechanical delay when they developed coexisting prolongation of QRS duration. Another potential explanation is that RVA pacing-induced LBBB per se may have a lower propensity of developing mechanical dyssynchrony than those with cardiomyopathy-related LBBB. In the presence of systolic dysfunction, the myocardial structural damages (e.g. scarring, interstitial fibrosis, myocardial filamentary, and mitochrondrial dysfunction), functional changes (e.g. receptors and secondary messengers), and adverse haemodynamic are likely to impose deleterious effect in the development of mechanical dyssynchrony, in addition to the impact of electrical dyssynchrony.

Previous studies observed that in heart failure patients with LBBB, two patterns of electrical conduction were present. The first and more common pattern is ‘regional’ conductional delay (usually occurring in the LV anterior wall) which results in heterogeneity of electrical activation and delay in excitation of the LV free wall. The second pattern is a ‘homogenous’ conduction delay within the LV in which the ventricle is activated en bloc albeit globally delayed. In the homogenous pattern, regional functional delay will be absent. When correlated with tissue Doppler imaging, only the former pattern showed evidence of systolic dyssynchrony, but not the latter one.24,25 Therefore, further studies are needed to determine whether the change in prevalence of these two conduction patterns may contribute to the relatively lower prevalence of systolic dyssynchrony in RVA pacing-induced LBBB.

Although the deleterious effect of RVA pacing has been described,17,9 the mechanism, however, is not fully understood. Dyssynchronous patterns of right ventricular and LV contraction and relaxation4,9,26,27 may result in misdistribution of stress and strain leading to asymmetrical hypertrophy28,29 and worsened LV function. Previous studies have observed reversible LV histological abnormalities30 and alteration of perfusion31,32 in patients with long-term RVA pacing. Although the acute deleterious effect of RVA pacing on LVESV and LVEF is relatively small, the chronic damage could be progressive and cumulative which lead to clinical dysfunction of the heart, unless preventive measures are adopted.5,6,8,33

Determinants of systolic dyssynchrony during RVA pacing

Not every patient subjected to RVA pacing will develop systolic dyssynchrony. Therefore, it is prudent to search for potential factors of the condition. The current study identified three independent determinants, namely the low normal LVEF during V-pace, pre-existing LV hypertrophy, as well as the cumulative amount of RVA pacing >40% in the past 6 months. The first factor indicates that when cardiac function is at the lower side of the normal, subclinical dysfunction already exists which may predispose to the development of systolic dyssynchrony through the aforementioned reasons. This is compatible with previous observation that when cardiac dysfunction has developed, patients are even vulnerable to develop dyssynchrony.15 For hypertensive patients with LV hypertrophy, a previous study had reported that the tendency of developing systolic dyssynchrony would be higher.34 Another intriguing finding is that patients who developed systolic dyssynchrony at V-pace had a higher cumulative percentage of RVA pacing in the past 6 months. Therefore, the longer the patients are exposed to RVA pacing, the higher likelihood that electromechanical will develop leading to impairment of LV systolic function.

Clinical implications

As RVA pacing may induce systolic dyssynchrony in only half of the patients which in turn will impair LV contractile function, this pacing modality may render the development of heart failure, in particular in those with low normal LVEF, had pre-existing LV hypertrophy, or when they are more pacing dependent.5,6,8,33 Therefore, it is worthwhile to assess the vulnerability of developing systolic dyssynchrony in patients who are receiving pacemaker therapy, in particular in those with a high degree of ventricular pacing dependency, e.g. advanced atrioventricular block. If during acute testing patients already develop significant systolic dyssynchrony by RT3DE or possibly other echocardiographic tools, it might be worthwhile to consider the use of special pacing algorithm to minimize RVA pacing;35 or alternatively using other pacing modality such as right ventricular outflow track pacing, LV pacing, or biventricular pacing to minimize dyssynchrony.3638 Nevertheless, the acute haemodynamic effects may not be translated into the effects at the long term; a prospective study is therefore warranted to demonstrate the outcome of chronic RVA pacing.


In patients with preserved EF, up to half of patients developed systolic dyssynchrony during RVA pacing. This will result in reduction of EF and LV enlargement. Furthermore, a low normal EF, pre-existing LV hypertrophy, and a high cumulative RVA pacing >40% in the past 6 months were independent predictors of developing dyssynchrony. This might implicate on the choice of alternative pacing site or pacing algorithm in vulnerable patients.

Conflict of interest: none declared.


This study was supported by an earmark research grant from Research Grants Council of Hong Kong, RGC Ref No. CUHK4485/05M.


View Abstract