Fifty years of cardiac pacing: the dark side of the moon?

Abstract

Fifty years after its introduction, cardiac pacing has evolved from an experimental medical treatment to an expanding field in today's cardiology. Only recently there is accumulating evidence that prolonged stimulation of the right ventricular apex is associated with clinically significant adverse effects. In this commentary, the potential adverse effects are summarised and potential modifications in contemporary pacing are discussed. (Neth Heart J 2008;16(Suppl1):S12-S14.)

Fifty years after its introduction, cardiac pacing for symptomatic bradycardia can look back on its impressive past. It is one of those rare medical therapies that has changed the lives of numerous patients faced with a disease with high morbidity and mortality to a prognosis which almost equals that of the normal population. Associated with a steep increase in life expectancy in the industrial world during the last half century, the number of device implants has been steadily growing and is expected to grow even further. Because of the clear benefits of cardiac pacing in patients with symptomatic bradycardia, potential harmful effects of pacing have only recently been recognised. This commentary will discuss the evidence of the adverse effects of cardiac pacing and will provide potential future modifications.

Adverse effects of cardiac pacing

More than 80 years ago, in 1925, Wiggers showed that right ventricular (RV) pacing in mammals was associated with a marked reduction in left ventricular (LV) pump function.³ Right ventricular apical pacing bypasses the His-Purkinje system resulting in a left bundle branch block like pattern on the surface ECG. More importantly, LV contraction is altered and dyssynchrony may be induced. Subsequent studies showed that pacing results in LV remodeling with asymmetric hypertrophy and dilatation, mitral re-gurgitation, reduced myocardial perfusion and reduced ejection fraction.^4,5

Many studies on pacing in patients with sinus node dysfunction have focussed on the potential benefits of atrial based pacing (imposing normal His-Purkinje conduction) and ventricular based pacing.^6,7 Large randomised trials have demonstrated that mortality is not influenced by pacing mode. In the large MOde Selection Trial (MOST) randomising 2010 patients with sinus node dysfunction to VVIR or DDDR pacing mode no differences were seen with respect to stroke-free survival during a six-year follow-up. In patients assigned to dual chamber pacing, the risk of atrial fibrillation was lower and heart-failure scores were better but this did not translate into a difference in rates of hospitalisation for heart failure and/or death or stroke.⁷ However, Sweeney and co-workers convincingly demonstrated that not pacing mode but cumulative percentage right ventricular pacing (as determined by stored pacemaker data) was a major determinant of outcome.⁸ A sustained and increasing incidence of hospitalisation for heart failure was found in DDDR patients with >40% cumulative ventricular pacing (as compared with ≤40% pacing) and in WTR patients with >80% cumulative pacing (as compared with ≤80% pacing) with a hazard risk of 2.5 in both patient groups. A similar highly significant pattern was seen with respect to onset of atrial fibrillation (AF). More recently, the DAVID trial provided additional evidence that right ventricular pacing was associated with an increased risk of death and heart failure hospitalisation in patients with an impaired LV function selected for ICD therapy.⁹ In a retrospective analysis, Steinberg et al. reported on behalf of the MADITII investigators in a similar group of patients with impaired LV function and defibrillator therapy: the cumulative probability of new/worsened heart failure or death during three years of follow-up was seen in 50% of patients with >50% ventricular pacing whereas this endpoint was met in 20% of patients with ≤50% pacing (p<0.001).¹⁰ Thus, despite of its bright side, prolonged ventricular pacing may constitute the dark side of the moon: you cannot see it and it might be dangerous.

Can pacing-associated effects be prevented?

Four goals have been defined to optimise pacemaker therapy, particularly in patents with sinus node disease: (1) prevent symptomatic bradycardia, (2) provide chrono tropic competence when necessary, (3) maintain AV synchrony and (4) maintain normal ventricular activation. Prevention of bradycardia and maintenance of chronotropic competence is now provided with all DDDRpacemakers. Since very long AV delays (>250 msec) interfere with optimal upper-rate limits and AF recognition, long AV delays should be avoided. Moreover, static long AV intervals are not very effective to allow normal ventricular contraction in a substantial number of patients. Algorithms that used automatic AV interval search to promote normal conduction result in a modest reduction in cumulative ventricular pacing.¹⁷ Recently, a new algorithm was introduced that aggressively imposed normal AV conduction by automatically switching between the single and dual chamber pacing mode.¹⁷ In a recent prospective randomised study in 1065 patients with sinus node dysfunction, a significant 40% reduction in the development of persistent AF was observed using this algorithm as compared with patients assigned to conventional DDDR pacing.¹³

To prevent the deleterious effects of long-term right ventricular pacing, alternative pacing sites have been studied. Although a number of these studies reported a better haemodynamic response at alternative pacing sites, these results could not be reproduced by others. The most impressive results were reported by Deshmukh and colleagues with direct His-bundle pacing in patients with dilated cardiomyopathy and chronic atrial fibrillation. After a follow-up of two years a remarkable absolute increase in ejection fraction of 11% was demonstrated with significant reductions in both end-systolic and end-diastolic dimensions. This technique remains challenging with a relatively high proportion of failures to implant the lead.¹⁴ In a quantitative review, including nine studies on alternative right ventricular pacing sites, a modest but significant haemodynamic benefit was observed as compared with conventional pacing.¹⁵ These studies, however, were hampered by a number of factors including duration of follow-up, assessment of haemodynamic response and a non-uniform definition of alternative pacing position. A significant contribution in this respect was made by Giudici and Karpawich's proposal, suggesting anatomic definitions of alternative pacing sites.¹⁶

The most appropriate conclusion that can be drawn from all these studies is that there is no such thing as ‘one site fits all’. This was elegantly demonstrated in a recent study by Lieberman and co-workers. They compared LV systolic function (using pressure-volume loops) during RV pacing, LV pacing and biventricular pacing in patients with and without LV dysfunction and narrow QRS complexes. Both in patients with and those without LV dysfunction, RV apical pacing impaired LV haemodynamics. Biventricular pacing and LV pacing did not impair LV haemodynamics in patients with normal LV function and improved haemodynamics in patients with poor LV function. Most importantly, however, was the finding that optimised RV pacing (defined as the alternative position with the highest stroke volume) could attenuate the detrimental effects of RV apical pacing in both patients with normal and those with impaired LV function.¹⁷ Thus RV lead position definitely plays a role and an individually, targeted approach is more likely to result in optimal haemodynamic response in patients selected for cardiac pacing. More evidence on the variable response in haemodynamics during RV apical pacing was provided in a recent study by Varma and co-workers. They measured LV activation delay in response to RV apical pacing in patients with normal and abnormal LV function. RV apical pacing in patients with a normal LV function resulted in an increase in QRS duration to 151 msec and delayed inferolateral activation to a mean of 70 msec. These delays mirror closely those with intrinsic conduction and left bundle branch block. In addition, the effects of RV apical pacing were much greater in patients with impaired LV function. The most remarkable finding, however, was that the effect of RV apical pacing varied considerably in patients with normal LV function.¹⁸ This finding may explain why some patients do remarkably well with prolonged RV pacing, whereas others deteriorate much earlier. It is conceivable that patients with a significant LV activation delay may be at risk of developing LV dysfunction whereas patients without or with minor delayed LV activation may have preserved LV function.

In summary the optimal pacing modality in patients with symptomatic bradycardia is far from settled. There is evidence that prolonged RV apical pacing is not only bad for ‘your heart but also bad for your health’. Using appropriate algorithms to prevent unnecessary RV pacing is likely to prevent the detrimental effects of pacing. Although the debate on alternative RV pacing sites will still continue, there is a little bit of light on the dark side of the moon.