OUP user menu

Atrial vs. dual-chamber cardiac pacing in sinus node disease: a register-based cohort study

C. Michael Fored, Fredrik Granath, Fredrik Gadler, Paul Blomqvist, Jenny Rynder, Cecilia Linde, Anders Ekbom, Mårten Rosenqvist
DOI: http://dx.doi.org/10.1093/europace/eun118 825-831 First published online: 7 May 2008

Abstract

Aims In patients with sinus node disease, dual-chamber pacing (DDD) possibly results in adverse effects on the ventricular function. We have compared the incidence of cardiovascular morbidity and mortality in patients with sinus node disease and with atrioventricular (AV) synchronous pacemakers, DDD vs. atrial pacing (AAI).

Methods and results A nation-wide population-based cohort of 8777 patients with AAI- or DDD-mode pacemakers was followed during 12 years. The cohort was linked to national healthcare and census registers. Patients with DDD pacing and without any pre-implant admission for atrial fibrillation or flutter had an increased risk of post-implant fibrillation or flutter, in relation to corresponding AAA patients [hazard ratio (HR) = 1.30; 95% confidence interval (CI) 1.10–1.52]. A slight increase in the risk of any cardiovascular disease (HR = 1.07; CI, 1.00–1.15), and all-cause mortality (HR = 1.12; CI, 1.00–1.25), was seen among DDD patients, in relation to AAI patients, but there was no significant difference in the risk of ischaemic or unspecified stroke (HR = 1.14; CI, 0.94–1.37). Among DDD patients, the all-cause mortality did not differ from the general population [standardized mortality ratio (SMR) = 1.04; CI, 0.98–1.11]. Patients with AAI, however, had a decreased all-cause mortality risk (SMR = 0.89; CI, 0.82–0.97).

Conclusion Our results support AAI as the preferred mode of pacing in patients with sinus node disease, and a normal AV node function.

Keywords
  • Arrhythmia
  • Sinoatrial node
  • Pacing
  • Stroke
  • Epidemiology

Introduction

The choice of pacing mode in patients with sinus node disease remains controversial. Atrial pacing (AAI) has reduced atrial fibrillation, stroke, and heart failure, improving the long-term survival, in relation to ventricular pacing (VVI), in observational and small randomized studies.13 However, large trials have failed to show a survival benefit from dual-chamber pacing (DDD), in relation to VVI.46 In contrast, DDD pacing has been associated with significantly less atrial fibrillation and less heart failure hospital care.79 In a recent study, utilizing an algorithm reducing ventricular pacing during intact atrioventricular (AV) conduction, a reduction in atrial fibrillation was seen in comparison with conventional DDD pacing patients.10 These findings suggest that the absence of ventricular pacing with a preserved ventricular pattern, rather than the presence of AV synchrony, determined the outcome.

In patients with sinus node disease, the proportion of DDD has markedly increased among the implanted pacemakers during the past 10 years, especially with the introduction of rate modulation devices.11 Dual-chamber pacing with modulated pacing rate (DDDR), however, tends to provide mainly ventricular pacing and a dyssynchronous ventricular contraction pattern,12,13 with possible adverse effects on ventricular function.8,9

The aim of this study was to compare the incidence of cardiovascular morbidity and mortality in patients with sinus node disease and with AV synchronous pacemakers, with or without ventricular pacing (DDD vs. AAI). A nation-wide population-based cohort of pacemaker patients was followed during 12 years.

Methods

Setting

The study was conducted in Sweden using nation-wide population-based healthcare and census registers. Unique personal identification numbers (PINs), assigned to every Swedish resident since 1947,14 allowed record-linkage of register-data recorded prospectively with very few losses to follow-up and minimized risk of recall bias.

Study cohort

All sinus node disease patients receiving an AAI- or DDD-mode pacemaker implant between 1989 and 2002 were identified in the Swedish implantable cardioverter defibrillator and pacemaker register (n = 9563).15,16 Every hospital in Sweden performing pacemaker implantations (n = 47) reports to the register, which is directed by the Karolinska University Hospital for quality assurance in pacemaker therapy. Information on all patients receiving either a new implantation or a replacement of the generator or the lead is entered. Data collection are based on the European pacemaker patient identification card, which is used by the International Cardiac Pacing and Electrophysiology Society in world-wide surveys of cardiac pacing and implantable cardioverter defibrillator practices.17

Follow-up

The study cohort was linked to the Swedish patient register (PAR) and the Cause of Death Register through the PIN. The PAR has recorded individual-based information on all inpatient care, county-wise since 1964, and nation-wide since 1987.18,19 Each PAR record contains: (i) the patient's PIN; (ii) the date of hospital admission and discharge; (iii) one primary discharge diagnosis and up to seven additional diagnoses coded according to the International Classification of Diseases20 (ICD)-7 until 1968, ICD-8 from 1968 through 1986, ICD-9 from 1986 to 1996, and ICD-10 thereafter; and (iv) surgical procedures. The nation-wide Cause of Death Register contains information on primary and contributing causes of death and the date of death. Sinus node disease patients were followed from the date of their first pacemaker implant until 31 December 2002 or to the date of death. Primary endpoints were the first hospitalization for, or death from (i) atrial fibrillation or flutter; (ii) stroke; or (iii) any cardiovascular disease, whichever occurred first. Subjects with a hospital discharge diagnosis of stroke prior to the pacemaker implant were excluded (n = 786) [stroke (ICD-9 434, ICD-10 I63), unspecified stroke (ICD-9 436, ICD-10 I64), and haemorrhagic stroke (ICD-9 431, ICD-10 I61), atrial fibrillation or flutter (ICD-9 427.3, ICD-10 I49), and any cardiovascular disease (ICD-9 390–459, ICD-10 I00–I99)]. The total follow-up time was 30 826 patient-years. Median follow-up was 4.15 years (inter-quartile range 1.76–6.93) among patients with AAI-mode pacemakers and was 2.28 years (inter-quartile range 0.98–4.39) among DDD-mode pacemaker patients.

Statistical methods

To assess the relative risk of stroke, atrial fibrillation or flutter, or any cardiovascular disease between AAI and DDD pacing in patients with sinus node disease, we compared the incidence rates of hospitalization or death among 8777 subjects included in the Swedish implantable cardioverter defibrillator and pacemaker register [study base, AAI = 12 502 person years (PYRS); DDD = 18 324 PYRS]. The indications for pacemaker implantation for sinus node disease were tachy-brady syndrome, sinus arrest, sinus bradycardia, or unspecified sinus node disease. As the incidence of pacemaker implant and the distribution of pacing mode have varied over time, the cohort was grouped into time-periods by the year of implant (1989–93, 1994–98, and 1999–2002). During 5 years prior to the pacemaker implant, the number of hospital admissions for (i) atrial fibrillation or flutter, (ii) angina pectoris, (iii) chronic ischaemic heart disease, (iv) myocardial infarction, (v) heart failure, and (vi) diabetes were used as indicators of disease severity. Hazard ratios (HRs) and 95% confidence interval (CI) of the outcomes were calculated using the Cox proportional regression21 in (i) a basic model including sex, age, and time-period at implant and (ii) a fully adjusted model including the co-variates for pacemaker indication and the number of pre-implant hospital admissions for the diagnoses described above, in addition to the basic model variables. In addition, the fully adjusted model was used in an analysis stratifying on the presence of any pre-implant hospital admission for atrial fibrillation or flutter. Standardized mortality ratios (SMRs) were calculated using age-, gender-, and calendar year-specific expected survival estimates obtained from the Swedish Death Register. Ninety-five per cent CIs were calculated by assuming a Poisson distribution of the number of events.

The study was approved by the Stockholm regional ethical vetting board.

Results

Patient characteristics

Characteristics of patients varied between pacing modes. At the time of pacemaker implantation, patients who received DDD pacemakers were older [median age 75 years; quartile range (QR) = 67–80] than those implanted with AAI devices (median age 73 years; QR = 65–79). More women than men received AAI (60.2%, Table 1). Although the incidence of AAI device implantation remained mainly unchanged during the study period, a near 100-fold increase in annual DDD device implantation was seen (Figure 1).

Figure 1

Distribution of implanted atrial pacing and dual-chamber pacing pacemakers during the study period.

View this table:
Table 1

Pacemaker patient characteristics by pacing mode

Variable—n (%)Atrial pacing (AAI) n = 2685 (30.6%)Dual-chamber pacing (DDD) n = 6092 (69.4%)P-valueTotal n = 8777
Sex
Male1069 (39.8)3002 (49.3)4071 (46.4)
Female1616 (60.2)3090 (50.7)<0.0014706 (53.6)
Age
−49201 (7.5)261 (4.3)462 (5.3)
50–59258 (9.6)479 (7.9)737 (8.4)
60–69512 (19.1)1158 (19.0)1670 (19.0)
70–791084 (40.4)2487 (40.8)3571 (40.7)
80−630 (23.5)1707 (28.0)<0.0012337 (26.6)
Time-period
1989–93746 (27.8)399 (6.5)1145 (13.0)
1994–981193 (44.4)1787 (29.3)2980 (34.0)
1999–2002746 (27.8)3906 (64.1)<0.0014652 (53.0)
Treatment indication
Tachy-brady syndrome723 (26.9)2536 (41.6)3259 (37.1)
Sinus arrest1029 (38.3)2173 (35.7)3202 (36.5)
Bradycardia707 (26.3)1011 (16.6)1718 (19.6)
Unspecified226 (8.4)372 (6.1)<0.001598 (6.8)
Pre-implant hospital admission for atrial fibrillation or fluttera
None2369 (88.2)4829 (79.3)7198 (82.0)
1167 (6.2)581 (9.5)748 (8.5)
2+149 (5.6)682 (11.2)<0.001831 (9.5)
Pre-implant hospital admission for angina pectorisa
None2229 (83.0)4783 (78.5)7012 (79.9)
1236 (8.8)554 (9.1)790 (9.0)
2+220 (8.2)755 (12.4)<0.001975 (11.1)
Pre-implant hospital admission for chronic ischaemic heart diseasea
None2539 (94.6)5490 (90.1)8029 (91.5)
183 (3.1)318 (5.2)401 (4.6)
2+63 (2.3)284 (4.7)<0.001347 (3.9)
Pre-implant hospital admission for myocardial infarctiona
None2493 (92.8)5489 (90.1)7982 (90.9)
1160 (6.0)441 (7.2)601 (6.9)
2+32 (1.2)162 (2.7)<0.001194 (2.2)
Pre-implant hospital admission for heart failurea
None2431 (90.5)5133 (84.3)7564 (86.2)
1139 (5.2)472 (7.7)611 (7.0)
2+115 (4.3)487 (8.0)<0.001602 (6.8)
Pre-implant hospital admission for diabetesa
None2542 (94.7)5636 (92.5)8178 (93.2)
158 (2.1)163 (2.7)221 (2.5)
2+85 (3.2)293 (4.8)<0.001378 (4.3)
  • aNumber of hospital admissions during 5 years prior to the pacemaker implant.

In patients receiving DDD (41.6%), tachy-brady syndrome was the most common indication, whereas sinus arrest was more common among those treated with AAI devices (38.3%). A larger proportion of the DDD patients had been admitted to hospital for any cardiovascular disease during the 5 years before their pacemaker implantation (DDD = 52.2% vs. AAI = 40.2%, P < 0.001). Pre-implant hospital admissions for atrial fibrillation or flutter, angina pectoris, chronic ischaemic heart disease, myocardial infarction, and heart failure were significantly more common among DDD patients (Table 1).

Outcomes

Atrial fibrillation or flutter

Following pacemaker implantation, 387 patients (14%) with AAI mode and 1017 DDD patients (17%) were admitted to hospital care for atrial fibrillation or flutter. A difference in time to post-implant hospitalization for atrial fibrillation or flutter was seen, favouring the AAI patients (Figure 2A). Adjusting for age, sex, and time-period of follow-up, DDD-mode pacing was associated with an increased risk of atrial fibrillation or flutter hospitalization, in relation to AAI-mode pacing (Table 2, Model 1, HR = 1.46; 95% CI, 1.29–1.62). The risk estimate was distinctly attenuated after adjustment for pre-implant cardiac disease admissions (Table 2, Model 2, HR = 1.14; 95% CI, 1.00–1.29). In relation to AAI, DDD-mode pacing was associated with an increased risk of atrial fibrillation or flutter among patient without pre-implant hospital admission for fibrillation or flutter (HR = 1.30; 95% CI, 1.10–1.52) (Table 2).

Figure 2

(A) Time to hospitalization for atrial fibrillation or flutter. (B) Cumulative incidence of ischaemic or unspecified stroke. (C) Time to hospitalization for cardiovascular disorders. (D) Overall survival.

View this table:
Table 2

Relative risks associated with dual-chamber pacing in relation to atrial pacing

Outcome diagnosisaModel 1 HRb (95% CI)Model 2 HRc (95% CI)Pre-implant admission for fibrillation or flutter
NoYesP-valued
HRc (95% CI)HRc (95% CI)
Atrial fibrillation or flutter1.46 (1.29–1.62)1.14 (1.00–1.29)1.30 (1.10–1.52)0.92 (0.76–1.11)0.005
Ischaemic or unspecified stroke1.22 (1.02–1.48)1.14 (0.94–1.37)1.08 (0.89–1.33)1.51 (0.92–2.50)0.22
Any cardiovascular disease1.20 (1.12–1.29)1.07 (1.00–1.15)1.07 (0.99–1.15)1.10 (0.94–1.29)0.70
All-cause mortality1.17 (1.05–1.31)1.12 (1.00–1.25)1.12 (0.99–1.26)1.14 (0.84–1.56)0.89
  • Hazard ratios (HRs) and 95% confidence interval for atrial fibrillation or flutter, ischaemic or unspecified stroke, any cardiovascular disease, and all-cause mortality.

  • aICD-10 classification codes for atrial fibrillation or flutter = I49, ischaemic stroke = I63, unspecified stroke = I64, and cardiovascular disease = I00-I99.

  • bHR adjusted for sex, age, and time-period.

  • cHR adjusted for sex, age, time-period, treatment indication, and pre-implant hospital admission for atrial fibrillation or flutter, angina pectoris, ischaemic heart disease, myocardial infarction, heart failure, and diabetes.

  • dP-value for the comparison between patients with or without any pre-implant hospital admission for atrial fibrillation or flutter.

Stroke and stroke-related mortality

In total, 590 patients were admitted to hospital care or died from stroke of any type. Patients were uncommonly diagnosed with a haemorrhagic stroke (n = 41, 6.9%). Among AAI patients, 196 were admitted for hospital care of, or died (n = 12) from, ischaemic or unspecified stroke. The corresponding numbers among DDD patients were 353 events (n deaths = 10). The crude cumulative incidence of ischaemic or unspecified stroke 5 years after the pacemaker implantation was 5% among both AAI and DDD patients (Figure 2B). There was no statistically significant difference in the risk of ischaemic or unspecified stroke between pacing modes, adjusting for available important co-variates indicating a more severe cardiac disease (Table 2, Model 2, HR = 1.14; 95% CI, 0.94–1.37). No significant difference in the stroke risk between pacing modes was seen with regard to pre-implant admission for atrial fibrillation or flutter (Table 2). In relation to comparable individuals in the general population, patients with AAI- or DDD-mode pacemakers had an equal risk of dying from stroke (Table 3).

View this table:
Table 3

Standardized mortality ratios and 95% confidence intervals by pacing mode

Cause of deathaAtrial pacing (AAI)Dual-chamber pacing (DDD)
n observedSMR (95% CI)n observedSMR (95% CI)
Ischaemic or unspecified stroke360.79 (0.55–1.09)691.08 (0.84–1.37)
Haemorrhagic stroke101.04 (0.50–1.91)100.70 (0.34–1.29)
Any cardiovascular disease3400.99 (0.89–1.11)6281.29 (1.19–1.40)
All-cause mortality5900.89 (0.82–0.97)10041.04 (0.98–1.11)
All causes except cardiovascular disease2500.78 (0.69–0.89)3760.79 (0.71–0.87)
  • aICD-10 classification codes for ischaemic stroke = I63, unspecified stroke = I64, haemorrhagic stroke = I61, and cardiovascular disease = I00-I99.

Any cardiovascular disorder

Following pacemaker implantation, hospital care for any cardiovascular disorder was seen among 52% (n = 1403) of the AAI patients and among 44% (n = 2661) of the DDD patients. Time to post-implant hospitalizations for any cardiovascular disorder was similar in the AAI and DDD groups (Figure 2C). In relation to AAI-mode pacing, DDD-mode pacing was associated with a moderate risk increase of hospitalization for any cardiovascular disorder, adjusting for age, sex, time-period of follow-up, and pre-implant cardiac disease admissions (Table 2, Model 2, HR = 1.07; 95% CI, 1.00–1.15). The risk of any cardiovascular disease was equal between pacing modes regardless of pre-implant atrial fibrillation or flutter admission status (Table 2). Patients with AAI-mode pacemakers had a risk of dying from cardiovascular disease equal to that of the general population (Table 3). The SMR was moderately elevated among DDD-mode pacemaker patients (SMR = 1.29; 95% CI, 1.19–1.40).

All-cause mortality

In relation to AAI patients, the DDD patients' all-cause mortality risk was slightly increased when adjusting only for age, sex, and time-period of follow-up (Table 2, Model 1, HR = 1.17; 95% CI, 1.05–1.31). The relative risk estimate was attenuated, considering pre-implant admissions for severe cardiac diseases (Table 2, Model 2, HR = 1.12; 95% CI, 1.00–1.25). The mortality risk did not differ between pacing modes, stratifying on the presence of pre-implant admission for atrial fibrillation or flutter (Table 2). A slight difference in overall survival between pacing modes was seen with increasing time since implantation favouring AAI patients (Figure 2D). In relation to comparable individuals in the general population, sinus node disease patients receiving AAI-mode pacemakers had a significant decreased risk for all-cause mortality (SMR = 0.89; 95% CI, 0.82–0.97) (Table 3). Among DDD patients, the all-cause mortality did not differ from the general population (SMR = 1.04; 95% CI, 0.98–1.11). Standardized mortality ratio for cardiovascular disease, the most common cause of death, was increased in patients receiving DDD (Table 3). However, SMRs for all-cause mortality, except cardiovascular disease, were significantly decreased and equal between AAI and DDD patients (Table 3).

Discussion

Our register-based long-time follow-up among sinus node disease patients, comparing AAI-mode pacemaker treatment with DDD-mode pacemakers, showed a small borderline significant increase in all-cause mortality risk and in the risk of hospitalization from atrial fibrillation or flutter, or for any cardiovascular disease among DDD-mode pacemaker patients. Pacemaker patients without any pre-implant hospital admission for atrial fibrillation or flutter and receiving DDD-mode pacing had an increased risk of post-implant fibrillation or flutter, in relation to the corresponding AAI-mode patients. No statistically significant difference in the risk of hospitalization for ischaemic or unspecified stroke was seen. In relation to age- and gender-standardized individuals in the general population, sinus node disease patients with AAI mode had a decreased mortality. When excluding cardiovascular diseases from the causes of death, mortality was equally decreased for AAI- and DDD-mode patients.

Atrioventricular synchrony, whether AAI or DDD, has for long been regarded as superior to ventricular pacing. However, large trials showed no survival benefit from atrial pacing in relation to ventricular pacing,4,5,22 even with less atrial fibrillation over time in DDD pacing patients.5,7 The adverse haemodynamic effects of the altered ventricular activation mode linked to right ventricular apical pacing were demonstrated already in the early 1990s.12,13 Lately, the potential risk of excess ventricular pacing has become the focus of attention. In the MOST trial, increased ventricular pacing was associated with more hospitalizations for heart failure and atrial fibrillation.8 In the CTOPP study, atrial fibrillation was linked to pacemaker dependency, thus a greater proportion of ventricular pacing.5 In addition, the recent study by Sweeney et al.10 found that reducing ventricular pacing during DDD pacing caused less atrial fibrillation than standard DDD pacing. With the programming praxis and device technology at the time of our study, it is reasonable to assume that ventricular pacing was predominate in the DDD group, possibly contributing to the difference in atrial fibrillation found between AAI and DDD pacing among patients without any pre-implant hospital admissions for atrial fibrillation.

The incidence of ischaemic stroke in both AAI and DDD patients in our study was similar to previous observations. In the MOST trial, where patients were assigned to either DDD or ventricular inhibited pacing (VVIR), stroke occurred in 4.5% of the population.9 In the CTOPP study, the annual stroke risk was 1.1% with ventricular pacing and 1% in AV-synchronous pacing.6 There was an increased risk of ischaemic stroke in the VVI group compared with AAI in the Danish comparison, with more strokes among patients with tachy-brady syndrome.1 That study to date remains the only prospective study comparing AAI pacing with VVI showing reduced stroke and cardiovascular mortality in AAI-treated patients.1,2 Unlike the Danish trial, we excluded patients with previous strokes prior to pacemaker implantation and also analysed earlier cardiovascular hospitalizations. This may explain the lack of influence from pacing mode on stroke incidence in our study. Moreover, we did not assess the use of warfarin or antiplatelet therapy.

In our study, AAI-paced patients had better survival than individuals in the general Swedish population, comparable with regard to age and gender, whereas the standardized all-cause mortality in DDD patients did not differ from unity. In a standardized mortality analysis, co-morbidities are not considered, and the better overall survival among pacemaker patients is likely due to a selection into pacemaker treatment, in relation to the general population including other serious diseases. The difference in mortality between pacing modes disappeared when death from any cardiovascular disease was excluded. Our results show a difference in pre-implant morbidity, and AAI-mode pacemakers were selected to comparably healthier patients. It can, however, be hypothesized that this beneficial effect on morbidity and mortality by AAI pacing is linked to the absence of ventricular pacing. Acute haemodynamic detrimental effects of right ventricular apical pacing were previously substantiated by findings in patients paced on average for 2.9 years.23 Comparing DDDR with AAIR resulted in left ventricular and atrial remodelling and clearly more atrial fibrillation in the DDDR-paced group. In a meta-analysis including eight randomized trials and 35 000 patient-years of follow-up, the use of atrial-based pacing did not improve survival or reduce heart failure or cardiovascular death, but did reduce the incidence of atrial fibrillation, particularly in patients with sinus node disease, in relation to ventricular pacing.24 Only 4% (n = 282) of the 7231 patients in this meta-analysis were paced in the AAI mode. It is conceivable that the potential benefit of a maintained AV synchrony in these trials was offset by the harmful effects of unnecessary ventricular pacing.

Atrial pacing is a relatively rare mode of pacing in sinus node disease mainly because of the presumed risk of the development of a high-degree AV block in sinus node disease patients. However, the annual risk for the high-degree AV block is low at 1.8%25,26 and the risk possibly exaggerated. For this reason, only a very small proportion of patients in the large multicentre studies have been paced in the AAI mode. In CTOPP, where 42% of the 2568 patients had sinus node disease, only 6.6% were paced in the AAI mode.4 The MOST trial included 2010 sinus node disease patients and compared DDD with VVIR pacing, but no patients received an AAI pacemaker.27

Our hypothesis was that the potential hazardous effects of VVI mollified the beneficial effects of AV synchrony. Therefore, we compared AV-synchronous pacing with or without VVI (DDD vs. AAI). The use of national health care registers and the nation-wide Swedish implantable cardioverter defibrillator and pacemaker register made it possible to access all implantations for sinus node disease since 1989. We regard the data as representative of pacemaker implantation in sinus node disease patients over a 12-year period. Our findings are strengthened by the fact that patients with stroke prior to pacemaker implantation were excluded from the study and that the number of hospitalizations in the 5 years before the pacemaker implantation was considered in our analysis. The patient-years of follow-up of AAI-paced patients were substantially larger than in the previous studies.

During the study period, the use of DDD pacemakers increased enormously, especially during the years 1998–2002. A probable explanation is the conception in those years that rhythm control was preferable over rate control. Many patients with the tachy-brady syndrome might have had persistent atrial fibrillation and needed repeated cardioversions and anti-arrhythmic drug therapy. This in turn might have led to sinus bradycardia for which a DDDR pacemaker was prescribed. Moreover, sinus node remodelling causing a lower intrinsic sinus rate may also have contributed.28 In contrast, the implantation of AAI pacemakers in patients with sinus bradycardia and sinus arrest remained relatively constant over the time-period.

Our results have to be cautiously interpreted. The registers used do not include information on pre-implant cardiac function, medication, or seriousness of possible co-morbidities. However, the results seem to support AAI as the preferred mode of pacing in patients with sinus node disease, and a normal AV node function. The ongoing prospective and randomized DanPACE study may further elucidate the potential superiority of AAI pacing compared with DDD pacing.29 To date, the use of DDD pacers with algorithms that minimize VVI is to be preferred when AAI pacing is not indicated.

Funding

Funded by grants from the Swedish Heart-Lung Foundation, and from the Stockholm County Council.

Acknowledgements

The authors thank Anita Fredenson, pacemaker assistant for excellent help concerning the Swedish national pacemaker and implantable cardioverter defibrillator register data.

Conflict of interest: The Cardiology Unit at the Department of Medicine, Karolinska University Hospital receives a yearly research grant from Medtronics. C.L. is principal investigator for the REVERSE trial sponsored by Medtronics Inc., is a member of the advisory board of St Judes Medical, and has received honoraria from Medtronics Inc. and Sorin for talks. M.R. is a co-ordinator for four clinical drug studies, has been appointed to several speakers' bureaus, and is a member of the advisory boards for Boeringer-Ingelheim, AstraZeneca, and Medtronics Inc., as well as the adverse event group for implantable devices at Medtronics Inc. F.G. reports having received lecture grants from St Jude Medical and Medtronics Inc.

The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.

References

View Abstract