Dual-chamber vs. single-chamber pacemaker in patients in sinus rhythm with an atrioventricular block: a nationwide cohort study

Abstract

Aims

In complete atrioventricular block (AVB) with underlying sinus rhythm, it is recommended to implant a dual-chamber pacemaker rather than a single-chamber pacemaker. However, no large-scale study has been able to demonstrate the superiority of this choice on hard clinical criteria such as morbimortality.

Methods and results

This retrospective observational study included all patients who received a primary pacemaker implantation in the indication of complete AVB with underlying sinus rhythm in France, based on the national administrative database between January 2013 and December 2022. After propensity score matching, we obtained two groups containing 19 219 patients each. The incidence of all-cause mortality was 9.22%/year for the dual-chamber pacemaker group, compared with 11.48%/year for the single-chamber pacemaker group (hazard ratio (HR) 0.807, P < 0.0001]. Similarly, there was a lower incidence of cardiovascular mortality (HR 0.766, P < 0.0001), heart failure (HR 0.908, P < 0.0001), atrial fibrillation (HR 0.778, P < 0.0001), and ischaemic stroke (HR 0.873, P = 0.008) in the dual-chamber pacemaker group than in the single-chamber pacemaker group. Regarding re-interventions and complications, there were fewer upgrades (addition of atrial lead or left ventricular lead) in the dual-chamber group (HR 0.210, P < 0.0001), but more haematomas (HR 1.179, P = 0.006) and lead repositioning (HR 1.123, P = 0.04).

Conclusion

In the indication of complete AVB with underlying sinus rhythm, our results are consistent with current recommendations to prefer implantation of a dual-chamber pacemaker rather than a single-chamber pacemaker for these patients. Implantation of a dual-chamber pacemaker is associated with a lower risk of mortality, heart failure, atrial fibrillation, and stroke during follow-up.

Graphical Abstract

Open in new tabDownload slide

Introduction

When permanent pacing is needed in patients with atrioventricular block (AVB) and with sinus rhythm, dual-chamber pacemaker (DDD) is recommended over single-chamber pacemaker (VVI).¹ When patients have reasons to avoid two leads such as limited venous access or significant comorbidities, a VVI pacemaker can be implanted.² Many studies have shown that dual-chamber pacing improves haemodynamics by restoring atrioventricular synchronization by coupling ventricular pacing to the sinus node rate.^3–6 Single-chamber pacemaker in patients with AVB and sinus rhythm may lead to a pacemaker syndrome caused by atrioventricular dyssynchrony with symptoms such as palpitations, dizziness, syncope, fatigue, or hypotension. Pacemaker syndrome is associated with a reduction in quality of life and may require a re-intervention for upgrading to DDD pacemaker.⁶

Large randomized trials, however, failed to show superiority of DDD over VVI regarding mortality and morbidity^7–9 and have not consistently shown superiority in terms of quality of life.^10,11 The largest randomized study including only AVB patients was the UKPACE trial in 2005 where 2021 patients ≥ 70 years old with high-grade AVB were randomized in VVI and DDD groups.⁸ There was no significant difference in mortality between the two groups. Similarly, it did not find any significant difference regarding atrial fibrillation (AF), stroke, or heart failure. The difference highlighted in this study was post-operative complications, which were higher for DDD. Other studies including patients with AVB and patients with sinus node disease such as Pacemaker Selection in the Elderly Investigators (1998)¹⁰ or Canadian Trial of Physiologic Pacing Investigators (2000)¹² had similar results in the AVB population. However, there has been no large analysis in the real-life comparing outcomes of patients treated with DDD or VVI pacing in these patients in a modern era. Our objective was to compare patients with DDD and VVI implanted for AVB with underlying sinus rhythm on clinical outcomes such as mortality, heart failure, AF, and ischaemic stroke using a nationwide cohort study in during the 2013–22 period.

Methods

Study design

This longitudinal cohort study was based on the national hospitalization database covering hospital care from the entire French population. All patients admitted to French hospitals were screened using the national administrative Programme de Médicalisation des Systèmes d’Information (PMSI) database, through the French data protection authority, covering hospital care for the entire French population. It includes more than 98% of the French population (67 million people) from birth (or immigration) to death (or emigration). The reliability of PMSI data has already been assessed for quality or validation by others; this database has previously been used to study patients with cardiovascular conditions including those with pacemakers and similar administrative data sets using the International Classification of Diseases (ICD) codes have previously shown satisfying reliability.^13–18

Briefly, each hospitalization is encoded in a standardized data set, which includes information about the patient (age and sex), hospital, stay (date of admission, date of discharge, and modes of discharge), pathologies, and procedures. In the PMSI system, identified diagnoses are coded according to the ICD, Tenth Revision (ICD-10). All medical procedures are recorded according to the national nomenclature, Classification Commune des Actes Médicaux (CCAM). The PMSI contains individual anonymized information on each hospitalization that is linked to create a longitudinal record of hospital stays and diagnoses for each patient. The study was conducted retrospectively, and, as patients were not involved in its conduct, there was no impact on their care. Ethical approval was not required, as all data were anonymized.

Procedures for data collection and management were approved by the Commission Nationale de l’Informatique et des Libertés, the independent National Ethical Committee protecting human rights in France, which ensures that all information is kept confidential and anonymous, in compliance with the Declaration of Helsinki.

Study population

All patients over 18 years old hospitalized in French hospitals for a first pacemaker implantation for complete AVB were identified between 1 January 2013 and 31 December 2022. Importantly, patients with a history of sinus dysfunction were not included. Patients implanted with a cardiac resynchronization therapy (CRT) or a leadless device were neither included. Patients with a history of AF were excluded. Our population was divided into two groups: those treated with single-chamber (VVI) pacemakers and those treated with DDD. Information regarding demographics, medical history, procedures and events during hospitalization or follow-up was collected using ICD-10 and CCAM. For the second analysis, we matched the patients according to their characteristics to optimize patient comparability and obtain groups that were as similar as possible in terms of their characteristics, while minimizing confounding factors.

Outcomes

Patients were followed until 31 December 2022, for the occurrence of outcomes. Clinical outcomes for the analysis were all-cause death, cardiovascular death, AF, ischaemic stroke, heart failure, and endocarditis (left- and right-sided). Device-related outcomes were upgrading from VVI to DDD, or from VVI/DDD to CRT, and also device-related complications such as pneumothorax or haemothorax, tamponade, pocket complications such as haematoma and pocket revision, lead-related complication requiring revision, and device-related infection. Information on outcomes during follow-up was obtained by analysing the PMSI codes during the follow-up for each patient until the date of each specific occurring outcome or the date of last news in the absence of outcome. All events collected are expressed as incidence per year.

Statistical analysis

Qualitative variables are described as frequencies and percentages, and quantitative variables as means ± standard deviations. The analysis for clinical outcomes during the whole follow-up in the groups of interests was performed using a Cox model and reporting hazard ratios (HRs). Owing to the non-randomized nature of the study, and considering for significant differences in baseline characteristics, propensity score matching was also used to control for potential confounders of the treatment outcome relationship. Propensity scores were calculated using logistic regression with DDD pacemaker implantation as the dependent variable. The propensity score included cardiovascular risk factors and non-cardiovascular comorbidities from baseline characteristics listed in Table 1. For each patient with DDD pacing, a propensity score-matched patient with VVI pacing (excluding leadless VVI pacing) was selected (1:1) using the one-to-one nearest neighbour method and no replacement. The incidence rates (%/year) for each outcome of interest during longer-term follow-up were estimated in both groups and were compared using HRs. Hazard ratios and two-sided 95% confidence intervals were estimated using Cox proportional hazards model for death and the model by Fine and Gray for competing risks. A P-value of <0.05 was considered statistically significant.

Results

Patients

We identified 336 420 patients with complete AVB treated with conventional VVI or DDD pacing who were included in the analyses (Figure 1). We excluded patients with a history of AF (n = 134 966). We then constituted two groups, separating those with a VVI (n = 19 481) and those with a DDD (n = 181 973). Patients with a history of sinus node disease or AF and patients with CRT or leadless pacemakers were not included in the analysis. After propensity score matching, 19 219 patients with VVI were matched 1:1 with 19 219 patients with DDD pacemaker. Mean follow-up was 2.1 ± 2.3 years (median: 1.4 years, interquartile range: 0.1–3.5 years).

Baseline characteristics of the study population

Patients’ baseline characteristics of the study population are described in Table 1, and the two groups had numerous imbalances. In the VVI group, patients were older than those in the DDD group (84.1 years old vs. 77.9 years old, P < 0.0001) and a higher prevalence of comorbidities such as heart failure (respectively 30.8% vs. 26.6%, P < 0.0001), ischaemic stroke (3.5% vs. 2.9%, P < 0.0001), poor nutrition status (17.2% vs. 8.5%, P < 0.0001), cognitive impairment (15.1% vs. 5.0%, P < 0.0001), and frailty index (10.0% vs. 8.7%, P < 0.0001). In the DDD group, patients more frequently had diabetes (24.6% vs. 20.7%, P < 0.0001) and obesity (15% vs. 10.8%, P < 0.0001), were more frequently smokers (6.6% vs. 3.9%, P < 0.0001), and had a higher prevalence of coronary artery disease compared with VVI pacemaker patients (26.2% vs. 22.0%, P < 0.0001).

Clinical outcomes and device-related outcomes in unmatched patients

Clinical outcomes in the unmatched cohort are described in Table 2. In patients treated with DDD pacemaker, the incidence of all-cause mortality was 6.17%/year compared with 11.57%/year in patients with VVI pacemaker (HR 0.680, P < 0.0001). The incidence of cardiovascular death was 1.75%/year in patients with DDD pacemaker and 3.69%/year in patients with VVI pacemaker (HR 0.645, P < 0.0001). Implantation of a DDD pacemakers was also associated with a lower incidence of heart failure compared with the VVI group (10.57%/year vs. 17.34%/year, HR 0.776, P < 0.0001) and of incident AF (7.52%/year vs. 11.65%/year, HR 0.721, P < 0.0001). Dual-chamber pacemaker implantation was also associated with a lower rate of ischaemic stroke compared with VVI pacemaker (1.36%/year vs. 2.05%/year, HR 0.810, P < 0.0001). They were also with a slightly lower incidence of endocarditis (right-sided and left-sided) (0.42%/year vs. 0.45%/year, HR 0.8, P = 0.004).

During follow-up, there were 430 patients needing DDD upgrade in the VVI group, i.e. 2.20% of the VVI population. There were no significant differences between the two groups regarding CRT update with 133 CRT upgrades for the VVI pacemaker group (0.68% of the group) and 1792 CRT upgrades for the DDD pacemaker group (0.98% of the group) (P = 0.35). Considering any upgrades, the DDD group was associated with less surgical redo with 0.40%/year vs. 1.45%/year in the VVI group with a HR 0.235 (P < 0.0001).

Therapy with DDD pacemakers was associated with a significantly higher rate of pocket complication compared with the VVI group (1.92%/year vs. 1.30%/year, HR 1.260, P < 0.0001) and lead complication requiring revision such as lead dislodgement (1.63%/year vs. 1.55%/year, HR 1.145, P = 0.002). On the other hand, they were associated with a significantly lower rate of device-related infection (0.54%/year vs. 0.68%/year, HR 0.733, P < 0.0001) and endocarditis (0.42%/year vs. 0.47%/year, HR 0.800, P = 0.004). There were no significant differences in rates of pneumothorax/haemothorax and tamponade.

Clinical outcomes and device-related outcomes in matched patients

Clinical outcomes in the matched cohort are described in Table 3. After propensity matching, patients in the DDD pacemaker group still had a lower rate of all-cause mortality (9.22%/year vs. 11.48%/year, HR 0.807, P < 0.0001), cardiovascular mortality (2.75%/year vs. 3.65%/year, HR of 0.766, P < 0.0001), heart failure (15.50%/year vs. 17.27%/year, HR 0.908, P < 0.0001), new-onset AF (9.03%/year vs. 11.67%/year, HR 0.778, P < 0.0001), and ischaemic stroke (1.76%/year vs. 2.03%/year, HR 0.873, P = 0.008) compared with VVI pacemaker group. Finally, after propensity matching, there were no significant differences between the two groups for endocarditis (right-sided and left-sided).

Dual-chamber pacemaker group was still associated with a lower rate of any upgrade to DDD or CRT (0.29%/year vs. 1.43%/year, HR 0.210, P < 0.0001). This was mainly driven by the DDD upgrade as there was no significant difference between the two groups regarding upgrade to a CRT after matching. Patients in the DDD pacemakers group had a significantly higher rate of pocket complication including haematoma compared with the VVI group (1.51%/year vs. 1.31%/year, HR 1.179, P = 0.006) and lead complication requiring revision such as lead dislodgement (1.64%/year vs. 1.54%/year, HR 1.123, P = 0.04). Finally, after propensity matching, there were no significant differences between the two groups for pneumothorax/haemothorax, device-related infection, and tamponade.

We performed interaction analyses for several subgroups of interest (age below or above 75, 85, and 90; sex; presence of hypertension; diabetes; history of coronary artery disease; history of dilated cardiomyopathy; Charlson index below or above 4; and Frailty score below or above 7) regarding all-cause death and cardiovascular death (see Supplementary material online, Tables S1 and S2). There was no clear interaction or differences regarding the lower risks for all-cause death and cardiovascular death in these subgroups. There was a marginally better prognosis with DDD vs. VVI in patients with lower Charlson score (vs. those with higher Charlson score), but this was not seen when considering lower or higher Frailty index.

We finally calculated number needed to treat as a measure for assessing the possible effectiveness of DDD vs. VVI pacing for important clinical outcomes. In the matched analysis, the average number of patients who would need to be treated with DDD pacing (instead of VVI pacing) to prevent one additional outcome would be 29 for all-cause death, 67 for cardiovascular death, 107 for incident heart failure, 34 for incident AF, and 337 for ischaemic stroke.

Discussion

Our study is to our knowledge one of the largest national retrospective studies comparing patients treated with conventional VVI or DDD pacing and analysed data from 336 420 patients with complete AVB with an underlying sinus rhythm. Our main findings in the matched population were as follows: (i) DDD group was associated with a lower rate of all-cause death, cardiovascular death, heart failure, AF, and ischaemic stroke (Figure 2); and (ii) DDD group was associated with a higher rate of lead complication and pocket complication but interestingly there were no differences regarding endocarditis and device-related infection, which are the main feared complications after such surgical revision (Figure 3). Although many physicians follow the 2021 European society of cardiology guidelines and implant DDD device in patients with AVB, data on hard clinical outcomes in this setting were lacking or did not have a high level of evidence and our results bring a scientific argument that may comfort such practice.

Baseline characteristics of the study population

Patients with sinus node disease were not included in our study as the clinical benefit of DDD (when possible) over VVI is relatively clear in such population.¹ Two main groups were analysed: VVI group (19 481 patients) and DDD group (181 973 patients). There were numerous imbalances that are described in the result section and in Table 2. Indeed, patients in the VVI group were older and had a higher rate of comorbidities. On the other hand, patients in the DDD group had higher prevalence of cardiovascular risk factors and comorbidities, notably diabetes, smoking, dyslipidaemia, obesity, alcohol consumption, sleep apnoea syndrome, and coronary artery disease. To reduce such selection bias, a propensity score matching analysis was performed including 19 219 patients in each group. After matching, the two groups had no significant differences regarding baseline characteristics. They were older with higher frailty and comorbidity index. Mean follow-up was 2.1 years, which, even if short, is in our opinion relevant in such an old and frail population. Some patients had a short follow-up but were included in the analysis as pocket complication or lead complication requiring revision use to happen in the beginning of the follow-up.

Clinical and device-related outcomes

Our study was consistent with known literature regarding yearly rate incidence of clinical outcomes and device-related complications. Before and after matching, the incidence of all-cause mortality, cardiovascular mortality, heart failure, newly diagnosed AF, and ischaemic stroke was significantly lower in the DDD group compared with the VVI group. For the first time, we were able to demonstrate significant differences on such major endpoints on a systematic and contemporary analysis of unselected patients seen at a nationwide level. The UKPACE study,⁸ one of the largest randomized trials in this field including 2021 patients, failed to find differences on hard clinical outcomes. Our study included a larger sample with 201 454 patients before matching and 38 438 patients after matching. Such number of patients may be needed to establish significant differences thanks to a high statistical power. In our opinion, the fact that our data are more recent could not in itself explain those differences as complication rates were similar to the UKPACE study.

Such apparent benefit may be driven by the role of atrioventricular synchronization in preserving cardiac function and prevention of AF as previously described by haemodynamical studies in the 80s.^3–5 It is important to mention that in the UKPACE study, the AV delay in the DDD group was willingly shortened and could maybe not enable an optimal stroke volume, which is one of the advantages of DDD pacing.² Another possibility is that for patients with DDD pacemakers, atrial lead and electrogram memories may be allowed an easier diagnosis of AF that may lead to better global healthcare management of such condition.

Historically, VVI pacing was chosen as a simple and faster way of pacing to avoid pocket/lead surgical revision and its infectious associated risk. In our study, DDD pacing was indeed associated with a significantly higher rate of pocket complication including haematoma compared with the VVI pacing and lead complication requiring revision such as lead dislodgement. In the unmatched analysis, patients in the DDD group had a significantly lower rate of device-related infection and a significantly lower rate of endocarditis. However, before matching, there were numerous imbalances regarding baseline characteristics as previously described with an older age and frailer profile for patients in the VVI group. In the outcome analyses after propensity matching, there were no more significant differences regarding device-related infection and endocarditis between the two groups. This is consistent with previously known data that did not find DDD pacemaker as an infectious risk factor.¹⁹ The divergence of survival curves for all the clinical outcomes occurred in the first year of follow-up in favour for DDD group and the differences in complications, although statistically significant, were not clinically meaningful, and above all did not translate in more infections in the DDD group.

Although there was a marginally different prognosis with DDD vs. VVI in patients with lower Charlson score (compared with those with higher Charlson score) in the interaction analyses, this was not seen when considering lower or higher Frailty index. We thus think this isolated finding may be related to random chance alone when presenting multiple statistical tests. The globally negative interaction analyses may actually reinforce our message about the better prognosis with DDD pacing compared with VVI pacing, whatever the subgroup of patients is considered. Regarding both clinical effectiveness and health system perspective, such observations in a modern setting are crucial and may support nowadays recommendations.¹

Limitations

First and foremost, the retrospective and non-randomized nature of our study obviously has associated biases. To mitigate this, we used a propensity score to match patients, a powerful statistical method for controlling confounding factors in observational studies. Even if many clinical variables were used for the propensity score matching, we cannot exclude that several other variables, especially those related to comorbidities that were not included in the model, could have had an impact in the results. However, all the available relevant variables found and coded with the ICD-10 classification were used. In addition, the large sample size of our study may contribute to the robustness and generalizability of our results, as for previous published data from our group with such methodology in cardiac pacing.¹⁸

We had no information on echocardiography or drug treatment. We also had no information on pacemaker programming, on lead position, or on percentage of ventricular pacing. We could not classify clearly paroxysmal and permanent AVB. Dual-chamber pacemaker mode with intrinsic conduction algorithm mode as recommended¹ may also have contributed to the benefit in the modern setting of this study observed in the DDD group compared with the VVI group.²⁰ Also, patients with an LVEF alteration would have been implanted with an implantable cardioverter-defibrillator or a CRT device²¹ and were not included in our study. We had no data about any conduction system pacing. However, considering the dates of implantation, such device may be marginal. Outcomes with such pacing modalities would be interesting to investigate in further studies.

Our study is based on administrative data obtained and filled manually by physicians and administrators and may therefore have information bias. However, as the coding of conditions and complications is linked to reimbursement and regularly controlled, it is expected to ensure quality and reliability. In our study, events included were only in-hospital, and we had no outpatient data. However, the main outcomes are most often managed in hospital in our (as in most other) health system. Extra-hospital death or other outcomes during follow-up may be underestimated in both groups. The study, carried out over the period 2013–22, is a representation of current trends in pacemaker device choice and recent technological improvements in pacemakers. However, the observation period may not be long enough to observe long-term events associated with each type of pacemaker. Finally, many patients with DDD pacing were excluded of the PS-matched analysis, and our study is a comparison of patients reflecting mainly VVI patients (compared with DDD patients) and not of patients reflecting DDD patients (compared with VVI patients). Thus, a general message of our work is rather that most VVI patients might have a better prognosis with DDD pacing (and not that most DDD patients would do worse with VVI pacing).

Conclusion

In this contemporary French nationwide cohort study, we found that treatment with DDD in patients with complete AVB and an underlying sinus rhythm is associated with lower rates of all-cause death, cardiovascular death, AF, heart failure, and stroke compared with VVI. Implantation of a DDD pacemaker was associated with more pocket complications such as haematoma or lead complications such as lead dislodgment, but there were no differences in terms of device-related infection or endocarditis. This study supports the nowadays recommendations on DDD pacing and the potential benefit associated with atrioventricular synchrony in cardiac pacing for these patients.

Supplementary material

Supplementary material is available at Europace online.

Funding

No funds, grants, or other support was received.

Data availability

Procedures for data collection and management were approved by the Commission Nationale de l’Informatique et des Libertés, the independent National Ethical Committee protecting human rights in France, which ensures that all information is kept confidential and anonymous, in compliance with the Declaration of Helsinki. The data and study materials will not be made available to other researchers for the purposes of reproducing the results or replicating the procedure. Because this study used data from human subjects, the data and everything pertaining to the data are governed by the French Health Agencies and cannot be made available to other researchers.

References

Author notes