Introduction
Since the inception of cardiac pacing, the favoured site has been the right ventricular apex (RVA). The RVA is popular as it is usually straightforward, provides stable lead position and is associated with few complications.
A critical re-evaluation of this practice is occurring, however. There is accumulating evidence that RVA pacing causes adverse effects on left ventricular (LV) function,1,2-4 and is associated with an increased risk of atrial fibrillation (AF)5,6 and death.7
This review will explore what is known about the mechanisms underlying the adverse effects of RVA pacing and discuss potential solutions to the problem.
Evidence for Adverse Clinical Effects of RVA Pacing
Multiple studies have shown adverse effects of RVA pacing. Left bundle branch block (LBBB), the electrocardiographic consequence of RVA pacing, has been demonstrated to be an independent predictor of cardiac morbidity and mortality in those with structural heart disease.8
Long-term follow-up of younger patients, usually with structurally normal hearts, has demonstrated deterioration in cardiac function. Twenty-four patients (mean age 19.5 years) paced from the RVA and followed for a mean of 9.5 years, had poorer LV function than unpaced controls,3 with a correlation between paced QRS duration and degree of LV dysfunction. A similar study showed lower resting cardiac output (CO) and poorer exercise performance in paced patients compared with controls.2
A registry study of patients with conventional pacing indications corroborates this data. After >10 years of RVA pacing, the prevalence of heart failure (HF) increases from 24-38% and of AF from 26-45%.9
Evidence is also available from comparisons of atrial versus ventricular pacing (VP). In the Danish study,7 patients with sinus node disease (SND) had improved survival and reduced HF when AAI rather than VVI paced. To eliminate concerns over loss of AV synchrony during VVI pacing, the Danish II study compared AAI and DDD (using RVA pacing) in a similar cohort of patients.5 There was a 30% reduction in overall mortality in the AAI paced group. Although this was not statistically significant, there were significant reductions in HF and AF.
The Mode Selection Trial (MOST) sub-study6 highlighted the importance of cumulative RV apical pacing time (%VP) on the development of AF and HF. In this study, patients with SND were randomised to VVIR/DDDR RVA pacing. The risk of HF hospitalisation and AF were directly related to the %VP (see Figure 1). Patients with >40% VP had a 2.6-fold increase in HF and there was a 1% increase in AF incidence for every 1% increase in %VP (see Figure 2).
The MOST data was confirmed by the larger Madit II sub-study10 and the Dual Chamber and VVI Implantable Defibrillator (DAVID) study.4 Clinical HF incidence was higher in these studies than in the MOST sub-study, suggesting that the negative effects of RVA pacing are more pronounced in patients with pre-existing LV dysfunction.
The accumulation of data relating to RVA pacing led the American Heart Association (AHA) to publish an advisory in 2005, which stated 'For patients who need a dual chamber pacemaker, efforts should be made to minimise the amount of VP when atrioventricular conduction is intactÔÇÖ.11
Mechanistic Insights
Normal cardiac electrical activation occurs via specialised conduction tissue (His-Purkinje system), where impulse propagation from the His bundle to the entire ventricular myocardium takes <55ms.12 Pacing ventricular myocardium results instead in activation via ventricular muscle. In patients paced from the RVA, ventricular activation frequently takes >150ms.13 Additionally, ventricular activation occurs from apex to base with late activation of the lateral LV wall.14 LBBB arising from RVA pacing causes slow, dyssynchronous LV activation with adverse haemodynamic consequences;3,15 typically LV impairment,14 increased myocardial oxygen demand16 and local changes in myocardial perfusion.17,18
The initial reduction in LV ejection fraction (LVEF) due to pacing-induced LBBB is compounded in the longer term by LV remodeling.19 Ventricular dilatation and asymmetric LV hypertrophy LVH occur20,21 and at the tissue level, there is localised cellular degeneration, myofibrillar disarray and disorganisation of mitochondria.22
In crossover studies, patients have significantly higher cardiac output,23 better LVEF,23 lower pulmonary wedge pressure24 and higher maximal oxygen uptake (V02Max)25 when paced in AAI mode than in DDD mode.
Strategies to Avoid RVA Pacing
Strategies to avoid deleterious effects may be divided into two broad categories: avoiding unnecessary VP, and pacing the ventricle from a site other than the RVA. For patients with intact AV conduction, manipulation of pacing modes may reduce %VP. In patients in whom VP is inevitable (e.g. AF with bradycardia, complete heart block), alternate pacing sites may provide a solution (see Table 1).
Alternative Pacing Modes to Reduce RV Pacing
AAI(R) Pacing
This is the most straightforward way to reduce VP. 37% of patients26 are paced for sinus node disease alone, making them suitable for AAI(R) pacing. A concern of AAI(R) pacing is that the patient is unprotected should they develop AV block. Estimates of the incidence of AV block range to 4.5% per annum, but the mean figure is only 0.6% in a meta-analysis.27 Overall the rate is likely to be acceptably low with careful patient selection but clinicians seem unwilling to leave the patient without the 'protection' of a ventricular lead-rates of AAI(R) pacing range from less than 1-12%.
DDD(R) Pacing with Long AV Delay
A long AV delay theoretically yields functional AAI(R) pacing while protecting from AV block.5,28 In practice this causes significant difficulties in achieving proper pacing function at faster heart rates. Furthermore, programming the AV delay to above the resting PR interval still results in 88% of patients with %VP >50% due to PR prolongation on exercise.29 Programming a long, fixed AV delay of 300ms still leaves half of patients with %VP >40%.5
Automatic AV Interval Prolongation Algorithms
Several pacemaker manufacturers have developed algorithms that periodically extend the AV interval to establish whether intrinsic AV conduction occurs. They reduce the percentage of RVA pacing to 15-20%,30 a considerable improvement on manually programmed AV delays, but still higher than is desirable.
New Pacing Programming Modalities
The expectation that every atrial event is followed by a ventricular event within a relatively short time-frame is inherent in conventional dual chamber pacing algorithms. Marked first-degree AV block or 'physiological' second-degree AV block (e.g. Wenkebach during sleep) result in V pacing and an increased %VP.
Some device manufacturers have developed algorithms which operate in AAI(R) mode, permitting first degree or Mobitz type-I heart block to occur. AAISafeR31 (ELA) and Managed VP32 (MVP-Medtronic) permit a significant reduction in %VP in patients with intermittent high-grade AV block or persistent first-degree AV block. These devices operate by 'mode switching' from AAI pacing to DDD pacing if high-grade AV block develops. They periodically check for intrinsic AV conduction and, if appropriate, revert to AAI mode. These algorithms reduce %VP to <1% in selected patients,32 but clinical outcome data, specifically reductions in AF and HF, are not yet available. Several on-going clinical trials will hopefully address this issue; the MVP trial will compare MVP-based pacing with DDDR pacing32 and SAVE PACe will compare MVP with an older AV prolonging algorithm.33
Alternative Pacing Sites
Although pacing algorithms other than DDD(R) may reduce %VP in selected patients, VP is sometimes unavoidable. Alternative site pacing has been developed in an attempt to overcome the deleterious effects of RVA pacing. Potential sites include RV septal or outflow tract pacing, direct His bundle pacing and bi-VP.
RV Septal Pacing
Pacing from a high midseptal or RV outflow site activates the ventricles in a more physiological manner. While activation does not occur via the His Purkinje system (as in His bundle pacing) or as rapidly as with bi-VP, it avoids their complexities. Initial concerns over increased lead instability and higher pacing thresholds have proved unfounded.34 Most studies have examined the acute haemodynamic effects of RV septal pacing. Fewer have assessed longer-term effects on LV function. To date, only one study has examined effects on clinical HF.
Acute Haemodynamic Effects of RV Septal Versus RV Apical Pacing
RV septal pacing normalises QRS axis and causes earlier LV activation35 than RVA pacing but may not shorten QRS duration.1 Three studies have demonstrated shortening of the QRS duration19,36,37 whereas others have shown no difference.38-45
Studies evaluating the haemodynamic effects of septal pacing have largely shown no acute improvement in cardiac index (CI) or LV function,38-40,43,45,46 but two have demonstrated superior LV systolic function compared with RVA pacing.44,47 Three studies48,49,50 showed improvement in CI of around 20%.
Trial data regarding the acute haemodynamic effects of RV septal pacing is conflicting. Overall, there is proven non-inferiority of RV septal pacing over RVA pacing but no clear acute haemodynamic advantage.
Longer-term Studies of RV Septal Pacing
RVA pacing promotes ventricular remodelling in the longer-term;2-7 RV septal pacing may prevent this.5 studies have prospectively assessed RV septal versus RVA pacing in the medium- to long-term.19,36,37,41,42 Only two demonstrated any benefit of RV septal over RVA pacing. Mera et al.36 found in a randomised study of 12 patients with AF that RV septal pacing was associated with higher resting EF (RV septum 51% versus RVA 43%, p<0.01) although exercise EF was similar in both groups.
Tse et al.19 randomised 24 patients with complete heart block to VP of the RV septum or RVA. They found that RV septal pacing was associated with a shorter QRS duration and improved LVEF (RV septum 56% versus RVA 47%, p<0.001). Importantly, this study had an extended follow-up period of 18 months; at six months there was no difference between the groups. The time taken for RV apical pacing induced LV dysfunction to develop may explain negative results in shorter studies.
Randomised Studies with Clinical Outcome Data
The only published randomised study utilising clinical outcome data and comparing RV septal with RVA pacing is the ROVA study.37 This assessed quality of life (QOL) after three months of RVA or RV septal pacing in patients with HF, chronic AF and LVEF <40%. Although RV septal pacing shortened QRS duration, it did not improve QOL. The relatively short follow-up in this study may explain the negative result.
On-going Studies
It is noteworthy that <400 patients have been studied in clinical trials of RV septal versus RVA pacing, of which <200 have been involved in chronic studies.1 Although currently published data does not support routine RV septal pacing, there are several on-going studies of alternative site pacing that will provide more definitive data. These include the BRIGHT study of multisite pacing,34 the Biopace study of RVA pacing versus CRT,51 Block HF, DAVID II and INTRINSIC.
Direct His Pacing
Direct pacing of the His bundle is an attractive concept as ventricular activation should be entirely physiological. The practicalities of the procedure mean that it has not gained widespread acceptance. Problems often occur with lead positioning and stability as well as high pacing thresholds.13 Even in the largest series, direct His pacing is achieved in only 70%.52
Many patients have conducting system disease distal to the His bundle and in this situation, His pacing does not overcome AV block.
Bi-VP
Although there is plentiful evidence that bi-VP is beneficial in patients with impaired LV function, refractory HF and LBBB,53-55 there is much less data in patients without these indications. Left VP alone and bi-VP lead to better acute haemo-dynamics than RVA pacing56,57 but still induce dyssynchrony58 and reduce LV function59 compared with intrinsic conduction. Only those patients with impaired LV function undergoing AV node ablation for poorly controlled AF have been shown to benefit from bi-VP.60
Conclusion
There is very strong evidence that RVA pacing has negative effects on acute haemodynamics, ventricular remodelling and clinical outcomes. Alternative strategies have been developed and are being embraced by the pacing community, but clinical outcome data to support these strategies is scarce. The results of on-going larger and longer-term clinical trials will be pivotal in providing the basis for a rational approach to pacing in the future.