OUP user menu

★ editor's choice

A comparison of single-lead atrial pacing with dual-chamber pacing in sick sinus syndrome

Jens Cosedis Nielsen, Poul Erik B. Thomsen, Søren Højberg, Mogens Møller, Thomas Vesterlund, Dorthe Dalsgaard, Leif S. Mortensen, Tonny Nielsen, Mogens Asklund, Elsebeth V. Friis, Per D. Christensen, Erik H. Simonsen, Ulrik H. Eriksen, Gunnar V.H. Jensen, Jesper H. Svendsen, William D. Toff, Jeffrey S. Healey, Henning R. Andersen
DOI: http://dx.doi.org/10.1093/eurheartj/ehr022 686-696 First published online: 7 February 2011

Abstract

Aims In patients with sick sinus syndrome, bradycardia can be treated with a single-lead pacemaker or a dual-chamber pacemaker. Previous trials have revealed that pacing modes preserving atrio-ventricular synchrony are superior to single-lead ventricular pacing, but it remains unclear if there is any difference between single-lead atrial pacing (AAIR) and dual-chamber pacing (DDDR).

Methods and results We randomly assigned 1415 patients referred for first pacemaker implantation to AAIR (n = 707) or DDDR (n = 708) pacing and followed them for a mean of 5.4 ± 2.6 years. The primary outcome was death from any cause. Secondary outcomes included paroxysmal and chronic atrial fibrillation, stroke, heart failure, and need for pacemaker reoperation. In the AAIR group, 209 patients (29.6%) died during follow-up vs. 193 patients (27.3%) in the DDDR group, hazard ratio (HR) 1.06, 95% confidence interval (CI) 0.88–1.29, P = 0.53. Paroxysmal atrial fibrillation was observed in 201 patients (28.4%) in the AAIR group vs. 163 patients (23.0%) in the DDDR group, HR 1.27, 95% CI 1.03–1.56, P = 0.024. A total of 240 patients underwent one or more pacemaker reoperations during follow-up, 156 (22.1%) in the AAIR group vs. 84 (11.9%) in the DDDR group (HR 1.99, 95% CI 1.53–2.59, P < 0.001). The incidence of chronic atrial fibrillation, stroke, and heart failure did not differ between treatment groups.

Conclusion In patients with sick sinus syndrome, there is no statistically significant difference in death from any cause between AAIR pacing and DDDR pacing. AAIR pacing is associated with a higher incidence of paroxysmal atrial fibrillation and a two-fold increased risk of pacemaker reoperation. These findings support the routine use of DDDR pacing in these patients. Clinical Trial Registration: URL http://www.clinicaltrials.gov. Unique identifier: NCT00236158.

  • Cardiac pacing
  • Sick sinus syndrome
  • Single-lead atrial pacing
  • Dual-chamber pacing
  • Randomized controlled trial
  • Outcomes

Introduction

In patients with sick sinus syndrome, bradycardia should be treated with either a single-lead atrial pacemaker (AAIR) or a dual-chamber pacemaker (DDDR). AAIR pacing necessitates conduction over the atrio-ventricular node, which preserves the normal ventricular contraction pattern, but it does not protect against bradycardia if atrio-ventricular block subsequently develops. In contrast, DDDR pacing protects against bradycardia in case of atrio-ventricular block, but it usually causes some degree of unnecessary pacing in the right ventricle, which changes the electrical activation and contraction pattern of the ventricles. This may result in ventricular remodelling, with decreased left ventricular ejection fraction and left atrial dilatation,13 which has been associated with an increased risk of atrial fibrillation and heart failure.47 Therefore, we hypothesized that AAIR pacing was superior to DDDR pacing in patients with sick sinus syndrome, and we aimed to test that in a large-scale randomized trial.

Methods

Study design

We randomly assigned patients with sick sinus syndrome to AAIR pacing or DDDR pacing. From 10 March 1999 to 30 June 2008, we enrolled patients from all Danish pacemaker centres and from selected centres in UK and Canada, and followed them until 15 September 2009. The trial was initiated by investigators from all Danish pacemaker centres, and the trial design and the pacemaker programming were agreed on after advice from an international Advisory Board (Appendix). The trial was conducted in accordance with the Helsinki Declaration and approved by the regional Ethics Committee and the Danish Data Protection Agency. All patients gave their written informed consent before inclusion.

Criteria for eligibility

All patients referred for first pacemaker implantation were evaluated for inclusion. The criteria for inclusion were: symptomatic bradycardia; documented sino-atrial block or sinus-arrest with pauses >2 s or sinus bradycardia <40bpm for more than 1 min while awake; PR interval ≤0.22 s if aged 18–70 years or PR interval ≤0.26 s if aged ≥70 years; and QRS width <0.12 s. The main exclusion criteria were: atrio-ventricular block; bundle branch block; long-standing persistent atrial fibrillation (>12 months); atrial fibrillation with ventricular rate <40bpm for ≥1 min or pauses >3 s; a positive test for carotid sinus hypersensitivity; planned cardiac surgery; or a life-expectancy shorter than 1 year (Figure 1). Documented paroxysmal atrial fibrillation was not an exclusion criterion. Randomization by sealed envelope was performed before pacemaker implantation.

Figure 1

Flow diagram: screening, exclusion, randomization, and follow-up of patients. At last follow-up, four patients in the AAIR group and two patients in the DDDR group had a DDDR pacemaker programmed with features automatically prolonging or eliminating the atrioventricular interval to withhold ventricular pacing.

Implantation and programming of pacemakers

A bipolar lead was implanted in the right atrium, and in patients randomized to DDDR pacing an additional lead was implanted in the right ventricle. An atrial pacing test was performed at 100bpm in all patients and 1:1 atrio-ventricular conduction was required for implantation of an AAIR pacemaker. In patients randomized to AAIR pacing demonstrating atrio-ventricular block when paced at 100bpm, a ventricular lead and a DDDR pacemaker were implanted. Patients with atrial fibrillation during implantation either underwent DC cardioversion or received a pacemaker system without an atrial pacing test at the implanter's discretion.

The rate adaptive function was activated in all pacemakers and programmed with a lower rate of 60bpm and an upper rate of 130bpm. In patients with DDDR pacemakers, the paced atrio-ventricular interval was programmed to 140–220 ms according to a pre-specified algorithm: the paced atrio-ventricular interval was initially programmed to a value 10% longer than either the interval measured from the atrial pacing spike to start of the conducted QRS complex at 60bpm or the PR interval if the sinus rate was faster than 60bpm. If ventricular pacing occurred with this programming, the paced atrio-ventricular interval was gradually increased in steps of 20 ms until ventricular pacing ceased or until a maximum of 220 ms was reached. If ventricular pacing still occurred at a programmed interval of 220 ms, the paced atrio-ventricular interval was shortened to a length of 140–160 ms, and the atrio-ventricular hysteresis function was activated to allow automatic search for intrinsic atrio-ventricular conduction with an atrio-ventricular interval of 220 ms. The atrio-ventricular interval after sensed atrial beats was set 20–30 ms shorter than the paced interval, and automatic shortening of the atrio-ventricular interval was allowed during rate increases. The maximum tracking rate was individualized and the mode switch function was activated.

Patient follow-up

Follow-up took place after 3 months and again every year after implantation up to 10 years. At each planned follow-up visit, a printout was made of the pacemaker memory data accumulated since the previous resetting of the memory. The percentage of atrial and ventricular pacing at each follow-up was calculated using the number of paced and the number of sensed beats. The mean pacing percentage throughout the total follow-up period was computed for each patient by calculating the mean of these values. A 12-lead ECG was recorded and the atrial electrogram was obtained from the pacemaker telemetry to document the atrial rhythm. Furthermore, information regarding secondary outcomes and medication was collected. At the end of each planned follow-up visit, the correct programming of the atrio-ventricular delay was ensured according to the pre-specified algorithm, and the pacemaker memory was reset. Once every month, new deaths were identified by checking the study database against the Danish Civil Registration System and supplementary information regarding deceased patients was collected from hospitals and general practitioners.

Study outcomes

The primary outcome was death from any cause. Secondary outcomes included paroxysmal atrial fibrillation; chronic atrial fibrillation; stroke; peripheral embolism; heart failure; and need for pacemaker re-operation. Paroxysmal atrial fibrillation was defined as the first diagnosis of atrial fibrillation at a planned follow-up visit and chronic atrial fibrillation was defined as atrial fibrillation at two consecutive follow-up visits and at all subsequent follow-up visit. First cardioversion for atrial fibrillation after hospital discharge was recorded. Stroke was defined as: the sudden development of focal neurological symptoms lasting more than 24 h. A diagnosis of peripheral embolism required confirmation at autopsy, intra-operatively (during embolectomy), or at angiography. New York Heart Association (NYHA) functional class, use of diuretics, and hospitalization for heart failure were used as indicators of heart failure. Need for pacemaker reoperation was decided by the physician in charge of follow-up. The investigators were asked to only change the pacing mode from AAIR to DDDR pacing in cases of high-grade atrio-ventricular block or documented symptomatic atrio-ventricular block of the Wenckebach type. The incidental finding of a low Wenckebach block point at a follow-up visit was not an indication for change of pacing mode. A Clinical Event Committee (Appendix), which was unaware of the assigned pacing mode adjudicated stroke and thrombo-embolic events.

Statistical plan and analysis

In an earlier trial comparing AAIR pacing and single-lead ventricular pacing in patients with sick sinus syndrome, mortality was significantly higher in the single-lead ventricular group.5 We assumed that the relative difference in mortality between AAIR pacing and DDDR pacing would be half the difference observed between AAIR pacing and single-lead ventricular pacing. Therefore, the study was planned to include 1900 patients followed for a mean of 5.5 years to identify a 6% absolute difference (32 vs. 26%) in death from any cause between treatment groups, with a power of 80% and an overall α = 0.05. An independent Safety and Ethical Committee (Appendix) monitored two planned interim analyses performed after one-third and two-thirds of the expected number of deaths. The first interim analysis was planned with a conservative α = 0.001, and subsequently with α = 0.016 and α = 0.044 at the second interim and final analysis, respectively. According to the study plan, the trial would only be stopped early in case of a significant result at an interim analysis. An unplanned interim analysis with respect to the occurrence of atrial fibrillation was performed at the request of the Safety and Ethical Committee after publication of the Search AV Extension and Managed Ventricular Pacing for Promoting Atrioventricular Conduction (SAVE PACe) Trial in 2007.7 Due to the increasing use of dual-chamber pacemakers with new features prolonging or eliminating the atrio-ventricular interval in order to minimize ventricular pacing in patients with sick sinus syndrome, which were not permitted in the trial, the recruitment rate decreased in several Danish centres from 2005. Therefore, collaboration was established with pacemaker centres in Canada and the UK with the expectation that ∼400 patients from these countries would be enrolled during a 2-year period. Recruitment was commenced in the UK in May 2007 and in Canada in July 2007. By June 2008, a total of 31 patients had been enrolled between the two countries and it was clear that the planned enrolment of 400 patients could not be achieved. Consequently, it was decided to stop enrolment on 30 June 2008. After advice from the Safety and Ethical Committee, the investigators decided to stop follow-up on 15 September 2009.

Fewer than the planned 1900 patients were included in the study. From the second planned interim analysis, it could be foreseen that no significant difference could be reached with respect to the primary outcome even with the planned 1900 patients.

Treatment groups were compared on an intention-to-treat basis. Statistical tests were two-tailed, and P < 0.05 was considered significant. Mean ± SD are reported for continuous data. Cumulative event rates were calculated by the Kaplan–Meier method and the log-rank test was used for comparison between groups. Relative risk was expressed as hazard ratio (HR) with 95% confidence interval (CI). Pre-specified supplemental analyses adjusting for baseline characteristics were done using Cox proportional-hazards regression analysis. Categorical variables were compared by means of the Pearson's chi-square test and continuous variables by means of the Mann–Whitney test. Post-hoc, a supplementary on-treatment analysis was done comparing the outcomes ‘death from any cause’ and ‘paroxysmal atrial fibrillation’ between patients randomized to and treated with single-lead atrial pacemakers and dual-chamber pacemakers, respectively, during the whole study period.

The interim analysis design was a hybrid of a Pocock design and an O'Brien-Fleming design, and was performed using the EaSt software package, version 2.0 (Cytel Software). Data management was done using SIR/DBMS and SIR/FORMS (SIR Database Software), and statistical analysis was performed using SPSS/PASW version 18, BMDP release 8.1, and STATA version 11.

Results

Study population

After screening 18 553 patients referred for first pacemaker implantation, a total of 1415 patients were included and underwent pacemaker implantation (Table 1). At implantation, 93.4% of the patients randomized to AAIR pacing and 98.9% of the patients randomized to DDDR pacing were implanted as assigned (Figure 1). The mean duration of follow-up until death or end of study was 5.4 ± 2.6 years. The mean percentage of atrial pacing was 58 ± 29% in the AAIR group and 59 ± 31% in the DDDR group (P = 0.52). The mean percentage of ventricular pacing was 65 ± 33% in the DDDR group. In the AAIR group, 585 patients (83%) never had a ventricular lead. In the remaining 122 patients, a ventricular lead was implanted at the first operation or at some point during follow-up. Pacemaker memory data were recorded in 103 of these 122 patients, and showed a mean of 53 ± 35% of ventricular pacing. At the end of follow-up, the mean programmed maximum paced atrio-ventricular delay in the dual-chamber group was 225 ± 39 ms.

View this table:
Table 1

Baseline characteristics of the patients

CharacteristicAAIR (n = 707)DDDR (n= 708)P-value
Female gender, n (%)472 (66.8)441 (62.3)0.08
Age, years (mean ± SD)73.5 ± 11.272.4 ± 11.40.054
Body mass index (mean ± SD)26 (4)26 (4)0.62
Prior history of atrial fibrillation, n (%)303 (42.9)318 (44.9)0.44
Hypertension, n (%)241 (34.1)239 (33.8)0.90
Previous myocardial infarction, n (%)94 (13.3)90(12.7)0.74
Diabetes, n (%)68 (9.6)72 (10.2)0.73
Previous transient cerebral ischaemia, n (%)35 (5.0)37 (5.2)0.81
Previous stroke, n (%)61 (8.6)53 (7.5)0.43
Left ventricular ejection fraction reduced (< 50%), n (%)59 (10.6)54 (9.5)0.55
Left ventricular end-diastolic diameter in mm, mean ± SD47.7 ± 7.347.8 ± 7.30.45
Left atrial diameter in mm, mean ± SD39.3 ± 6.538.8 ± 6.40.23
Symptoms before pacemaker, n (%)
 Syncope359 (50.8)349 (49.3)0.58
 Dizzy spells597 (84.4)587 (82.9)0.44
 Heart failure86 (12.2)79 (11.2)0.56
 ≥2 of the above three symptoms317 (44.8)291 (41.1)0.16
Medication at randomization, n (%)
 Anticoagulation108 (15.3)89 (12.6)0.14
 Aspirin369 (52.2)361 (51.1)0.67
 Sotalol43 (6.1)44 (6.2)0.91
 Beta-blocker other than sotalol159 (22.5)132 (18.7)0.08
 Calcium-channel blocker137 (19.4)142 (20.1)0.75
 Digoxin73 (10.3)62 (8.8)0.32
 Amiodarone25 (3.5)24 (3.4)0.88
 Class I antiarrhythmics14 (2.0)20 (2.8)0.30
 Angiotensin-converting-enzyme inhibitors160 (22.6)170 (24.0)0.53
 Diuretics304 (43.0)263 (37.2)0.03
NYHA class, n (%)0.33
 I503 (71.4)522 (73.9)
 II172 (24.4)158 (22.4)
 III29 (4.1)24 (3.4)
 IV02 (0.3)
Wenckebach block point ≥100 bpm (%)611 (94.1)581 (91.6)0.08
Treated as randomized660 (93.4)700 (98.9)<0.001
  • The data were not complete for the following parameters: body mass index (AAIR: n = 646, DDDR: n = 637), left ventricular ejection fraction (AAIR: n = 558, DDDR: n = 569), left ventricular end-diastolic diameter (AAIR: n = 479, DDDR: n = 507), left atrial diameter (AAIR: n = 506, DDDR: n = 517), medications except calcium channel blocker (AAIR: n = 707, DDDR: n = 707), NYHA class (AAIR: n = 704, DDDR: n = 706), and Wenckebach block point ≥100 bpm (AAIR: n = 649, DDDR: n = 634).

Death from any cause

In the AAIR group, 209 patients (29.6%) died during follow-up vs. 193 patients (27.3%) in the DDDR group, HR 1.06; 95% CI 0.88–1.29, P = 0.53. Kaplan–Meier plots are shown in Figure 2. Patients randomized to DDDR pacing were slightly younger and, after correcting for differences in baseline variables, the adjusted mortality HR was 0.94, 95% CI 0.77–1.14, P = 0.52 for AAIR pacing vs. DDDR pacing (Table 2). The incidence of death from any cause was similar in all predefined subgroups (Figure 3). Variables significantly associated with death from any cause in multivariate analysis, including treatment group, were age (HR per year 1.093, 95% CI 1.079–1.107, P < 0.001), left ventricular ejection fraction below 50% (HR 1.82, 95% CI 1.33–2.48, P < 0.001), and prior myocardial infarction (HR 1.51, 95% CI 1.15–1.98, P = 0.003).

View this table:
Table 2

Clinical outcomes: unadjusted and adjusted hazard ratios

AAIR, n (%)DDDR, n (%)Unadjusted HRUnadjusted 95% CIP-value
Death from any cause209 (29.6)193 (27.3)1.060.88–1.290.53
Paroxysmal atrial fibrillation201 (28.4)163 (23.0)1.271.03–1.560.024
Chronic atrial fibrillation79 (11.2)76 (10.7)1.020.74–1.390.93
Stroke39 (5.5)34 (4.8)1.130.72–1.800.59
Pacemaker reoperation156 (22.1)84 (11.9)1.991.53–2.59<0.001
Adjusted HRAdjusted 95% CIP-value
Death from any cause0.940.77–1.140.52
Paroxysmal atrial fibrillation1.241.01–1.520.042
Chronic atrial fibrillation1.010.74–1.390.93
Stroke1.110.70–1.770.65
Pacemaker reoperation2.001.54–2.61<0.001
  • Death from any cause, paroxysmal atrial fibrillation, and chronic atrial fibrillation were adjusted for age, gender, prior history of atrial fibrillation, prior myocardial infarction, left ventricular ejection fraction lower than 50%, and hypertension. Stroke was adjusted for age, gender, prior history of atrial fibrillation, hypertension, and prior stroke, whereas pacemaker reoperation was adjusted for age, gender, prior history of atrial fibrillation, prior myocardial infarction, left ventricular ejection lower than 50%. Paroxysmal atrial fibrillation was defined as the first diagnosis of atrial fibrillation at a planned follow-up visit.

  • CI, confidence interval.

Figure 2

Time-to-event curves for the primary outcome and selected secondary outcomes. Atrial fibrillation denotes paroxysmal atrial fibrillation defined as: the first diagnosis of atrial fibrillation at a planned follow-up visit. Unadjusted P-values (log-rank test) are shown.

Figure 3

Hazard ratios for death from any cause are shown for different subgroups. CI, confidence interval; LVEF, left ventricular ejection fraction; MI, myocardial infarction; NYHA, New York Heart Association functional class. Smallest P-value for interaction was 0.45.

Atrial fibrillation

Paroxysmal atrial fibrillation was observed more frequently in the AAIR group than in the DDDR group, HR 1.27, 95% CI 1.03–1.56, P = 0.024, also after correction for baseline variables (Table 2, Figure 2). The incidence of chronic atrial fibrillation did not differ between the groups (HR 1.02, 95% CI 0.74–1.39, P = 0.93) (Table 2). The strongest predictor of paroxysmal atrial fibrillation was prior history of atrial fibrillation (HR 3.23, 95% CI 2.59–4.03, P < 0.001). First cardioversion for atrial fibrillation was done in 32 patients in the AAIR group vs. 20 patients in the DDDR group (HR 1.60, 95% CI 0.92–2.80, P = 0.10).

In the subgroup analysis, three significant interactions and one borderline significant interaction were observed: the incidence of paroxysmal atrial fibrillation was less with DDDR pacing in the subgroups of patients without a prior history of atrial fibrillation, with a longer PQ interval, a higher body-mass index, and a dilated left atrium at baseline (Figure 4). P-values for all interactions were <0.05 except for PQ interval with P = 0.084.

Figure 4

Hazard ratios for paroxysmal atrial fibrillation are shown for different subgroups. CI, confidence interval; LVEF, left ventricular ejection fraction; MI, myocardial infarction; BMI, body mass index. P-values for interaction were <0.05 for history of atrial fibrillation, BMI, and left atrial diameter, P = 0.084 for PQ interval, all other P-values for interaction were greater than 0.34.

Stroke and peripheral embolism

Stroke was diagnosed in 39 patients (5.5%) in the AAIR group and 34 patients (4.8%) in the DDDR group (HR for AAIR group 1.13, 95% CI 0.72–1.80, P = 0.59) (Figure 2, Table 2). The only predictors of stroke in a multivariate analysis were age (HR per year 1.037, 95% CI 1.012–1.062, P = 0.004) and gender (HR for women 0.56, 95% CI 0.35–0.91, P = 0.02). Peripheral embolism was diagnosed in one patient in the AAIR group and in two patients in the DDDR group.

Heart failure

NYHA class did not differ between groups at inclusion (Table 1) or at last follow-up, where the number of patients in class I/II/III/IV were 341/260/61/4 in the AAIR group vs. 364/231/67/4 in the DDDR group, P = 0.43. The proportion of patients treated with diuretics at last follow-up was 324/692 (46.8%) in the AAIR group vs. 328/695 (47.2%) in the DDDR group, P = 0.89. Hospitalization for heart failure occurred in 27 patients in the AAIR group vs. 28 patients in the DDDR group (P = 0.90).

Pacemaker reoperations

A total of 240 patients underwent one or more pacemaker reoperations during follow-up, 156 (22.1%) in the AAIR group vs. 84 (11.9%) in the DDDR group (HR 1.99, 95% CI 1.53–2.59, P < 0.001) (Table 3).

View this table:
Table 3

Number of patients undergoing re-operation for various indications

IndicationAAIR, n (%)DDDR, n (%)P-value
Battery depletion59 (8.3)42 (5.9)0.09
Need for surgical change of mode of pacing66 (9.3)4 (0.6)<0.001
Lead complications37 (5.2)30 (4.2)0.42
Surgical or mechanical complications10 (1.4)7 (1.0)0.52
Infection3 (0.4)3 (0.4)0.98
Skin erosion1 (0.1)3 (0.4)0.31
Device failure2 (0.3)2 (0.3)0.99
  • n is the number of patients undergoing reoperation because of each specific indication. Some patients underwent more than one pacemaker reoperation, for the same or for different indications.

Drug treatment during follow-up

Comparing drug treatment at last follow-up between the two groups revealed no significant differences: beta-blocker: 281/692 (41%) in the AAIR group vs. 274/695 (39%) in the DDDR group, P = 0.65; sotalol: 27/692 (4%) vs. 28/695 (4%), P = 0.90; calcium-channel blocker: 162/692 (23%) vs. 174/695 (25%), P = 0.48; amiodarone: 37/692 (5%) vs. 37/695 (5%), P = 0.99; class I antiarrhythmics: 13/692 (2%) vs. 11/694 (2%), P = 0.68; and anticoagulation: 170/692 (25%) vs. 175/695 (25%), P = 0.72.

On-treatment analysis

A total of 584 patients in the AAIR group and 639 patients in the DDDR group were treated as randomized throughout the entire study period. Comparing these two groups, we found no difference in death from all causes (HR for AAIR pacing 1.01, 95% CI 0.82–1.24, P = 0.93), whereas the difference in atrial fibrillation was more marked than in the intention-to-treat analysis (HR for AAIR pacing 1.42, 95% CI 1.12–1.81, P = 0.0038), also after adjusting for age, gender, and prior history of atrial fibrillation (HR 1.36, 95% CI 1.07–1.73, P = 0.012).

Discussion

The present trial is the first large multicentre trial to compare AAIR pacing with DDDR pacing in patients with sick sinus syndrome. The primary result was that no statistically significant difference was observed in survival between AAIR pacing and DDDR pacing. AAIR pacing was associated with a higher incidence of paroxysmal atrial fibrillation, and a two-fold increase in the risk of pacemaker reoperation during long-term follow-up. These findings suggest that DDDR pacing programmed with a moderately prolonged atrio-ventricular interval may be a better choice of pacing mode than AAIR pacing for patients with sick sinus syndrome.

More than a decade ago, a small randomized trial showed better overall survival with AAIR pacing than with single-lead ventricular pacing.5 Since the current trial was initiated, three large trials comparing single-lead ventricular and dual-chamber pacing in different patient populations have been published.810 In these trials, pacing mode had no influence on overall mortality in patients with sinus node disease or atrio-ventricular block.11 The current study is the first large trial to demonstrate that the presence of a ventricular lead, with some degree of ventricular pacing in the DDDR mode, as implemented in this trial, has no adverse effect on survival in patients with sick sinus syndrome. These findings support the contention that pacing mode has no significant influence on survival in patients with sick sinus syndrome and preserved left ventricular function.

Unexpectedly, we found a lower incidence of paroxysmal atrial fibrillation in the DDDR group than in the AAIR group. Modern DDDR pacemakers can report time in mode-switch as a very sensitive measure of atrial fibrillation, but no comparable data are available from AAIR pacemakers. Although the methods used in this study were likely to underdiagnose paroxysmal atrial fibrillation, they were systematic and the same in both treatment arms and should therefore yield an unbiased estimate of the influence of pacing mode on the development of atrial fibrillation. Furthermore, given the era in which this study was initiated (the 1990s) and the fact that pacemakers from every manufacturer were allowed, we feel that the methods to detect atrial fibrillation in our trial could not have been improved. Supporting the finding of more atrial fibrillation in the AAIR group, we observed a trend towards more patients undergoing their first cardioversion for atrial fibrillation in the AAIR group.

How can a higher incidence of atrial fibrillation with AAIR pacing be explained? The benefit of preserving atrio-ventricular synchrony in avoiding atrial fibrillation is well established from large randomized trials.8,9 In the current trial, the reason for excess atrial fibrillation in the AAIR group may be the prolonged atrio-ventricular conduction that is often observed with atrial pacing. Prolonged atrio-ventricular conduction reduces ventricular preload and causes mitral regurgitation.12 This accords with the finding that the benefit of DDDR pacing on atrial fibrillation in the present trial was restricted to the subgroup with longer PQ interval at baseline. It also accords with the results found in a large community-based cohort study, where prolonged atrio-ventricular conduction was associated with development of atrial fibrillation.13 Furthermore, in patients with implantable defibrillators, prolonged PR interval was found to be associated with a worse prognosis with atrial pacing.14 The programming of a lower rate of 60bpm can be criticized, as atrio-ventricular conduction often prolongs during atrial pacing.15 However, this is the same lower rate as was used in other large trials in sick sinus syndrome7,9 and the rate was the same in the two groups, as was the frequency of atrial pacing.

In sick sinus syndrome, DDDR pacing with a very short atrioventricular interval and more than 99% ventricular pacing has been reported to increase the incidence of atrial fibrillation, when compared with DDDR pacing with automated features to minimize ventricular pacing by prolonging or eliminating the atrioventricular interval.7 Post hoc analysis of data from the Mode Selection Trial in Sinus-Node Dysfunction indicated an association between ventricular pacing and atrial fibrillation.6 In contrast to these findings, we found less atrial fibrillation with DDDR pacing and a mean of 65% ventricular pacing when compared with AAIR pacing. The most likely explanation of this disparity is the difference in pacemaker programming. Our patients all had preserved atrio-ventricular conduction at baseline. To promote intrinsic atrioventricular conduction in the DDDR group, we programmed a moderately prolonged atrio-ventricular interval, with a maximum of 220 ms. This probably resulted in less ventricular desynchronization. The present findings suggest that DDDR pacing with a moderately prolonged atrio-ventricular interval may be a good alternative to DDDR pacemakers with automated features to minimize ventricular pacing for patients with sick sinus syndrome. However, no direct comparison has been made between these two pacing modes. The algorithms used to minimize ventricular pacing can, in rare cases, cause inadvertent bradycardia16 and pose a risk of potentially lethal bradycardia-related tachyarrhythmias.1719

The subgroup analysis on paroxysmal atrial fibrillation indicated that atrial fibrillation was significantly less frequent with DDDR pacing in the group of patients without a history of atrial fibrillation at time of pacemaker implantation, but not in the group who had a prior history of atrial fibrillation. Similar results were reported from the Mode Selection Trial in Sinus-Node Dysfunction, where a significant reduction in atrial fibrillation with DDDR pacing when compared with single-lead ventricular pacing was observed solely in the subgroup of patients without a history of atrial fibrillation.9 These results support that pacing mode selection is less important with respect to later development of atrial fibrillation in patients who already have a history of prior atrial fibrillation, and therefore are at the higher risk of later atrial fibrillation. In contrast, pacing mode selection seems to be more important for later occurrence of atrial fibrillation in patients who never had atrial fibrillation. However, these results were generated in subgroup analysis, and therefore should be interpreted cautiously.

In patients with severely compromised left ventricular function, ventricular pacing promotes heart failure, most likely due to ventricular desynchronization.20 In patients with preserved left ventricular function, DDDR pacing with a short atrioventricular interval and a high percentage of ventricular pacing also results in a reduced left ventricular ejection fraction.2,21,22 Analysis of data from the Mode Selection Trial in Sinus-Node Dysfunction showed that patients without heart failure who had normal ejection fraction and normal QRS duration at baseline tolerated atrioventricular synchronous ventricular pacing for years.23 In accordance with these findings, we found no difference in hospitalizations for heart failure, New York Heart Association functional class, or use of diuretics between treatment groups after a mean of 5.4 years of follow-up.

The risk of undergoing pacemaker reoperation was twice as high in the AAIR group when compared with the DDDR group. This difference was mainly due to an increased need for a change of pacing mode in the AAIR group, which occurred in 9.3% of the patients, or a mean of 1.7% per year. The need to change the pacing mode in 1.7% of patients per year in the AAIR group strengthens the case for DDDR pacing as the preferred initial pacing mode. Any pacemaker reoperation is associated with a 2% risk of device infection, with a potential need to extract the complete pacemaker system.24,25 These findings suggest that when a patient with an AAIR pacemaker is referred for a generator replacement or other reoperation, consideration should be given to implanting a DDDR pacing system to prevent future unplanned reoperations for a system change.

Reoperation because of battery depletion also occurred slightly more frequent in the AAIR group. This may reflect that many DDDR pacemakers are larger with greater battery capacity than smaller single chamber devices. However, the difference was not statistically significant and the finding may be incidental.

Only 1415 of the planned 1900 patients were included in the present trial, and this is a limitation that has to be taken into account when interpreting the findings. Because of the growing opinion in the pacing community that pacing the right ventricle was harmful, it became increasingly difficult to include patients in the trial as physicians chose to implant a DDDR pacemaker with features prolonging or eliminating the atrioventricular interval instead.26 Owing to the inclusion of fewer patients than initially planned, the power of the study was attenuated and the possibility of a small difference in all-cause mortality between DDDR pacing and AAIR pacing cannot be ruled out with certainty. However, the trial does include data from 1415 patients followed for a mean of 5.4 years, a total of more than 7600 years of follow-up, and no patients were lost for follow-up. From the second planned interim analysis, it could be foreseen that no significant difference could be reached with respect to the primary outcome even with enrolment of the planned 1900 patients.

Conclusion

In patients with sick sinus syndrome, there is no statistically significant difference in death from any cause between AAIR pacing and DDDR pacing programmed with a moderately prolonged atrioventricular interval. AAIR pacing is associated with a higher incidence of paroxysmal atrial fibrillation and a two-fold increased risk of pacemaker reoperation. These findings support the routine use of DDDR pacing in these patients.

Funding

The DANPACE trial was funded by unrestricted grants from Medtronic, St Jude Medical, Boston Scientific, Ela Medical, Pfizer, and The Danish Heart Foundation (10-04-R78-A2954-22779).

Conflict of interest: J.C.N. and J.H.S. have received consultant honoraries and speakers fees from Medtronic, St Jude Medical, and Biotronik. L.S.M. is an employee of UNI-C, and has been paid consultants fees for his participation in designing the study, taking care of data management and statistical analysis in the study, being a member of the study data monitoring board, and reviewing the manuscript. W.D.T. has received a grant from Medtronic for follow-up of patients enrolled in a clinical trial of cardiac resynchronization therapy. J.S.H. reports receiving a research grant from Boston for conduct of the SIMPLE trial—a 2500 patient study of implantable defibrillators; consulting fees and consultant honoraries from St Jude Medical; and speakers’ fees from Boston Scientific and St Jude Medical. The other authors report no conflicts.

Appendix

Investigators from Denmark (numbers of patients included): Henning Rud Andersen (co-chairman) and Jens Cosedis Nielsen (co-chairman), Aarhus University Hospital, Skejby (337); Poul-Erik Bloch-Thomsen, Gentofte Hospital (180); Søren Højberg, Bispebjerg Hospital (121); Mogens Møller, Odense University Hospital (114); Thomas Vesterlund, Aalborg Hospital (111); Dorthe Dalsgaard, Herning Hospital (108); Tonny Nielsen, Esbjerg Hospital (77); Mogens Asklund, Kolding Hospital (72); Elsebeth Vibeke Friis, Haderslev Hospital (70); Per Dahl Christensen, Viborg Hospital (56); Erik Hertel Simonsen, Hillerød Hospital (47); Ulrik Hedegaard Eriksen, Vejle Hospital (39); Gunnar Vagn Hagemann Jensen, Roskilde Hospital (28); Jesper Hastrup Svendsen, Rigshospitalet (24).

Investigators from the UK (numbers of patients included): William D. Toff (UK coordinating investigator), J. Douglas Skehan, Kieran Brack, Glenfield Hospital, Leicester (8); Craig Barr, Andreas Tselios, Nicola Gordon, Russells Hall Hospital, Dudley (6); John Cleland, Andrew Clark, Sarah Hurren, Castle Hill Hospital, East Cottingham (3); David McEneaney, Andrew Moriarty, Anne Mackin, Craigavon Area Hospital, Craigavon (2); Arif Ahsan, Jane Burton, Ruth Oliver, Nottingham City Hospital (2); Barry Kneale, Lynda Huggins, Worthing Hospital (2).

Investigator from Canada (number of patients included): Jeffrey S. Healey, Hamilton (8).

International Advisory Board: USA members: Victor Parsonnet, S. Serge Barold, Seymour Furman†, David L Hayes, Gervasio A Lamas, Paul A Levine, Melvin M Scheinman. UK members: A John Camm, Richard Sutton, William D Toff. From Canada: Stuart J Connolly. From France: Jacques Mugica†.

Safety and Ethical Committee: Kristian Thygesen (chairman), Denmark; David L Hayes, USA; Lukas Kappenberger, Switzerland; Hans Schüller, Sweden & Leif Spange Mortensen (datamanagement and statistics), Denmark.

Clinical Event Committee: Jørgen Videbæk (chairman), Kenneth Egstrup, Henning Bagger, all Denmark.

References

View Abstract