Dual-chamber implantable cardioverter defibrillators reduce clinical adverse events related to atrial fibrillation when compared with single-chamber defibrillators: a subanalysis of the DATAS trial

Abstract

Aims

The aim of the present analysis of the DATAS study was to compare the impact of dual- vs. single-chamber defibrillators on atrial fibrillation (AF) occurrence and AF-related clinical events in patients with Class I indication for implantable cardioverter defibrillators (ICDs) and no indication for dual-chamber pacing.

Methods and results

Three hundred and thirty-four patients were randomized, through a centralized assignment, to single-chamber ICDs, dual-chamber ICDs programmed as single-chamber ICDs, and dual ICDs with full diagnostics and AF prevention and therapy capabilities. The latter two groups in the first 8 months of the study, when the study design was that of a randomized parallel trial, were compared in the present analysis. The primary endpoint was composed by the following AF-related clinical events: permanent AF, AF-related hospitalizations, cardiac-embolic events, and inappropriate ICD shocks due to AF misclassification. Two hundred and twenty-three patients were available for this analysis, of whom 111 in the single-chamber-simulated group and 112 in the dual-chamber true group. Atrial fibrillation-related composite endpoint raw incidence was 9 of 111 (8.1%) in the single-chamber group vs. 1 of 112 (0.9%) in the dual-chamber group ( P = 0.0098 by Fisher’s exact test). Single-chamber ICDs were associated with a significantly higher risk to develop the AF-related composite endpoint by Cox regression analysis (hazard ratio 8.25, 95% CI 1.03–65.96, P = 0.047) and by the Kaplan–Meier survival analysis (log-rank test, P = 0.047).

Conclusion

Dual-chamber ICDs compared with single-chamber ICDs reduced the incidence of an endpoint composed by permanent AF, AF-related hospitalizations, and ICD shocks deemed inappropriate due to AF misclassification.

Introduction

Atrial fibrillation (AF) is common in implantable cardioverter defibrillator (ICD) patients with a prevalence ranging between 3% and 25%. ^1–5 Atrial fibrillation is associated with increasing mortality and morbidity. ^{6 , 7}

New generation dual defibrillators feature sophisticated preventive algorithms aimed to keep rhythm control, to suppress premature atrial complexes, to prevent short–long cycle, and to prevent early recurrences of atrial tachyarrhythmias. ⁸ Many current ICDs also feature atrial anti-tachycardia pacing (ATP) therapies, ^{9 , 10} which have been demonstrated to be effective in early interrupting ∼50% of atrial tachyarrhythmias. In case of persistent AF, device-delivered atrial shock ^{11 , 12} is a documented effective tool and may represent a cost-effective approach, by reducing the need for external cardioversion and hospitalizations. ^13–15

Even if atrial ATP algorithms and therapies have been associated with a good capability to interrupt atrial tachyarrhythmias, a significant impact of these therapeutic strategies on clinical outcomes has not been shown yet. ¹⁶

Recently, the DATAS study ¹⁷ in patients with conventional single-chamber ICD indication randomly compared single- and dual-chamber ICDs. Dual-chamber ICDs were associated with improvement of a score calculated as a weighted sum of the incidence of clinical outcomes comprising all-cause mortality, invasive intervention, hospitalization for cardiovascular causes, inappropriate shocks, and sustained symptomatic AF longer than 48 h.

We performed a post hoc analysis of the DATAS trial in order to compare, in a randomized parallel design study, the impact of a dual- vs. a single-chamber defibrillator programming on AF occurrence and AF-related clinical outcomes in ICD patients.

Methods

Study design

DATAS study results have been already published. ¹⁷ In brief, it is a prospective, multicentre, randomized, open-labelled study in patients with Class I indication for single-chamber ICDs and no indication for atrial-based pacing. The study was designed to compare single-chamber ICDs (‘SC true arm’), dual-chamber ICDs initially programmed as single-chamber ICDs (‘SC-simulated arm’), and dual-chamber ICDs initially programmed as a dual-chamber ICDs with full diagnostics and AF prevention and therapy capabilities (‘DC true arm’). The randomization procedure was centralized balancing the sample size of all three arms over the different centres. Study flow chart is shown in Figure  1 .

Patients were followed up at 1, 4, 8, 9, 13, and 17 months. At 8-month follow-up, ‘SC-simulated arm’ and ‘DC true arm’ were crossed over.

After study completion and publication of the manuscript about the primary objective, we selected the two study arms in which dual-chamber devices and AF diagnostic capabilities were available (grey box of Figure  1 ) and we analysed the clinical and diagnostic endpoints occurred in the first 8 months of the study, when the study design was that of a randomized parallel trial. The primary endpoint was composed by the following AF-related clinical events: permanent AF, defined as AF presence in two consecutive follow-up visits, AF-related hospitalization, defined as hospitalizations for which primary diagnosis of hospital admission was AF, and cardio-embolic events and inappropriate ICD shocks, defined as two or more ICD-detected episodes with inappropriate shocks due to AF misclassification. Secondary endpoints were device diagnostics data such as AF burden, frequency, and duration, in order to also evaluate asymptomatic AF episodes, and atrial shock and atrial ATP efficacy.

Implantable cardioverter defibrillator devices and programming

Dual-chamber devices used in the trial included Jewel AF (model 7250; Medtronic Inc., Minneapolis, MN, USA) and GEM III AT (model 7276; Medtronic Inc.).

These devices feature the dual-chamber algorithm (PR Logic), which discriminates between supraventricular and ventricular tachycardias, and detects AT/AF accurately and continuously. ¹⁸ Implantable cardioverter defibrillator devices used in this investigation had automatic atrial tiered anti-tachycardia therapies and were programmed to store data from atrial arrhythmic events.

Device programming is shown in Table  1 , for the two randomized arms.

Statistical analysis

Summary data were expressed as means ± SD or percentages of patients. Differences between mean data were compared by a ‘ t ’ test, for gaussian variables, and by Mann–Whitney or Wilcoxon non-parametric test, for non-gaussian variables, respectively, for independent or paired samples.

Differences in proportions were compared by a χ² analysis or Fisher’s exact test, as adequate.

Only atrial arrhythmia episodes longer than 5 min were considered in the statistical analysis, in order to discard inappropriate detections due to premature atrial contraction runs or spurious events. ¹⁹

The number of patients free from clinical endpoints such as AF recurrences or AF composite endpoint was studied by calculating the Kaplan–Meier curves that show the cumulative percentage of freedom during the follow-up. To compare differences between two or more Kaplan–Meier curves, the log-rank test was used.

Cox regression analysis was performed to evaluate predictors of single and composite endpoints occurrence.

A value of P < 0.05 for two-sided comparisons was considered significant.

For the statistical analysis, SPSS software (SPSS Inc., Chicago, IL, USA) was used.

Results

One thousand and sixteen patients were screened for study participation and 334 patients were enrolled in the three arms of the study ( Figure  1 ). In particular, 112 patients were enrolled in the dual-chamber ICD arm and 111 patients in the single-chamber-simulated ICD arm which represent the arms evaluated in this analysis.

Patients characteristics are described in Table  2 . No characteristics were significantly different between studied groups.

Primary endpoint

In the single-chamber group, two patients developed permanent AF, eight patients were hospitalized due to AF (mean duration = 3.3 ± 4.3 days), and two patients suffered inappropriate shocks due to AF misclassification; in the dual-chamber group, only one patient was hospitalized due to AF. No cardiac-embolic events were observed in the whole study population.

Three patients in the single-chamber ICD group reached more than one primary endpoint. Therefore, AF-related composite endpoint raw incidence was 9 of 111 (8.1%) in the single-chamber group vs. 1 of 112 (0.9%) in the dual-chamber group ( P = 0.0098 by Fisher exact test). Cox regression analysis showed that single-chamber ICDs were associated with a significantly higher risk to develop the AF-related composite endpoint (hazard ratio 8.25, 95% CI 1.03–65.96, P = 0.047) in comparison with dual-chamber ICDs. Also Kaplan–Meier analysis showed that single-chamber ICDs were associated with a significantly higher (log-rank test, P = 0.047) incidence of AF-related composite endpoint in comparison with dual-chamber ICDs, as shown in Figure  2 .

Secondary endpoints

Device diagnostics

Twenty-two of 223 (10%) patients had AF episodes during the observation period. No differences were observed in the incidence of AF as a function of device type—12 single-chamber ICD patients and 10 dual-chamber ICD patients (log-rank test, P = 0.443).

Atrial fibrillation incidence was higher in patients with baseline AF history, regardless of device type, as shown in Figure  3 .

Median AF duration was 81 min (inter-quartile range between 24 and 1161 min) in the single-chamber ICD group and 28 min (inter-quartile range between 10 and 163 min) in the dual-chamber ICD group ( P = 0.039).

Mean AF burden, in the 22 patients with AF recurrences, was 49 ± 125 min per day in the single-chamber ICD group vs. 8 ± 11 min per day in the dual-chamber ICD group ( P = 0.21). Median (and 25th–75th quartile) values for AF burden were 4.4 (0.2–9.3) min/day in the single-chamber ICD group and 1.0 (0.2–4.2) min/day in the dual-chamber ICD group.

Mean AF frequency was 2 ± 2 episodes/month in the single-chamber ICD group vs. 1 ± 2 episodes/month in the dual-chamber ICD group ( P = 0.52).

Mean atrial arrhythmia cycle length was 240 ± 55 ms. No differences were observed in single- vs. dual-chamber ICD groups.

Mean percentage of ventricular pacing, obtained from the device counters at each follow-up visit, was 40 ± 35% in the dual-chamber ICD group and 6 ± 13% in the single-chamber ICD group ( P < 0.05). Mean percentage of atrial pacing in the dual-chamber group was 35%.

Atrial therapies efficacy

Overall, 200 atrial tachyarrhythmia episodes were treated in the dual defibrillator group: early delivery of ATP therapies (three sequences of burst, ramp, or 50 Hz burst) was effective in interrupting 128 of 200 (64%) episodes. In the remaining 72 episodes, 52 episodes self-terminated, whereas 20 episodes, in eight patients, lasted >2 h, and were treated by atrial cardioversion between 3:00–5:00 a.m. with a termination efficacy of 18 of 20 (90%) episodes.

No adverse events were reported as a consequence of atrial therapies.

Discussion

Main study findings

The main finding of our study is that dual-chamber ICDs compared with single-chamber ICDs reduce the incidence of a clinical endpoint composed by permanent AF, AF-related hospitalizations, cardiac-embolic events, and ICD shocks deemed inappropriate due to AF misclassification.

Dual-chamber ICD recipients were also associated with a lower AF duration.

These findings should be regarded as clinically relevant taking into account the fact that statistical significance was reached, despite only a minority of patients suffered from AF.

Mechanisms and explanation of study results

Mechanisms that may explain these findings are ATP capability to terminate atrial tachycardia and atrial flutter and atrial shock efficacy in persistent AF episodes. ^8–15 These mechanisms would be confirmed by the significant reduction of median AF duration, found in our analysis. Other possible mechanisms are the maintenance of AV synchrony in patients requiring pacing, ^{20 , 21} prevention of sinus bradicardia, and capability to discriminate supraventricular rhythms, thus reducing AF-induced inappropriate therapies. ^{22 , 23}

Device diagnostic endpoints, such as AF burden and AF frequency, did not show significant differences in the two study groups. Neutral results were somehow expected due to the high variability of these quantities both among different patients and in the same patient as a function of time, as shown by recent observational studies. ^{24 , 25} Recently, Gillis et al . ²⁶ evaluating patients with sinus node dysfunction and paroxysmal AF have shown that ATP therapies are associated with a reduction in AF burden but only in a subset of patients with ATP efficacy >60%. In our trial, AF duration was significantly shorter in dual ICD arm, which can have clinical relevance by itself.

High efficacy and safety of atrial ATP and shock therapies ^8–15 have been confirmed in our trial, being ATP and shock efficacy equal to 64% and 90%, respectively. Atrial cardioversion was mandatory; it was programmed at night in order to minimize patient discomfort. This strategy was successful since no adverse events were reported as a consequence of atrial therapies and no reprogramming was done for atrial cardioversion due to patients complaints.

In dual-chamber ICD group, the ventricular pacing percentage was 40%. This finding was unexpected due to the fact that we programmed quite long AV intervals—200 SAV and 230 PAV. Interpretation for such a high ventricular pacing percentage may be the fact that in patients with long PR intervals, AV delay could not be programmed longer than 250 ms in the devices used in the study. We also observed the presence of fusion beats. Negative effects of ventricular pacing on ventricular function, heart failure, and AF recurrences have been shown in DAVID study ²⁷ and in subanalysis from randomized clinical trials. ^{28 , 29} In our series, incidence of AF and AF burden was not significantly different in the two groups and did not correlate with ventricular pacing percentage which was higher in the dual-chamber ICD group. That could be judged as surprising. A possible explanation could be related to the beneficial effects of prevention algorithms as well as of ATP therapies. Furthermore, many paced ventricular beats were actually fusion beats that are less impairing ventricular activation. Recently, the SAVE PACe trial ³⁰ has shown that algorithms designed to minimize ventricular pacing are associated with lower risk to develop persistent AF in sinus node disease patients wearing dual-chamber pacemakers. Superiority of dual-chamber ICDs, in terms of lower incidence of permanent AF and AF-related hospitalizations, was obtained in our study despite a significantly higher ventricular pacing percentage. This observation may induce the hypothesis that the clinical difference between dual- and single-chamber ICDs would have been even larger by exploiting algorithms recently designed to minimize right ventricular pacing.

Clinical implications

Guidelines for primary prevention of sudden death are changing the population of patients wearing an ICD: a higher proportion of ICD patients suffer from congestive heart failure and left ventricular dysfunction and a population with AF prevalence comprised between 25% and 50%. ^31–33 The overall frequency with which AF occur in heart failure patients is even higher: among 931 Framingham ³⁴ participants diagnosed with HF, 223 (24%) had history of AF and another 159 (17%) had AF subsequently, bringing the total proportion of HF patients with AF to 41%. Similarly, in a recent study, ³⁵ which followed for a median period of 13 months a population of 516 patients wearing pacemakers for cardiac resynchronization therapy via biventricular pacing, investigators measured a baseline prevalence of 51.7% and a proportion of patients with AF at sometime as high as 72%. In our study, designed in 2000, the majority of enrolled patients were implanted in secondary prevention of sudden death and their ejection fraction was significantly higher than that found in primary prevention trials. ² We could speculate that the benefits of dual defibrillators with detection algorithms and optimization of therapies devoted to prevent or terminate atrial tachyarrhythmias may be even higher in patient population with heart failure implanted in primary prevention who are at higher risk for AF. Furthermore, we excluded patients requiring atrial pacing, a population theoretically more likely to get advantage from atrial pacing. ^{8 , 16}

Limitations of the study

Per protocol, the parallel phase of the study was limited to 8 months. Therefore, we were not able to characterize the differences in long-term AF-related endpoints.

Conclusions

Dual-chamber ICDs compared with single-chamber ICDs reduced the incidence of a clinical endpoint composed by permanent AF, AF-related hospitalizations, and ICD shocks deemed inappropriate due to AF misclassification.

Funding

Medtronic Inc. supplied a grant to support the study.

Acknowledgements

This work was supported by a grant from Medtronic Inc. (Minneapolis, MN, USA) which, however, had no role in the interpretation of study results.

Conflict of interest: J.A. has received honoraria from Medtronic, Guidant (now Boston Scientific), Johnson & Johnson and St Jude Medical for lectures, and has served as a consultant for Johnson & Johnson. C.W. has received honoraria from Medtronic and St Jude Medical for lectures. A.Q. is currently conducting research sponsored by Medtronic and has served as a paid consultant for Medtronic and Boston Scientific. R.P.R. and M.S. have minor consultancy fees from Medtronic and St Jude Medical. X.N. is an employee of Medtronic Iberica. T.D. and A.G. are employees of Medtronic Italia.

Appendix

DATAS Investigators and study centres: B. Lüderitz, J. Schwab, T. Lewalter, R. Schimpf, University Hospital, Bonn, Germany; M.S., R.P.R., C. Pignalberi, M. Russo, San Filippo Neri, Rome, Italy; P. Hanrath, Ch. Stellbrink, K. Mischke, R. Koos, University Hospital RWTH, Aachen, Germany; J. Brugada, L. Mont, M. Matas, H. Clinic i Provincial, Barcelona, Spain; J. Gill, R. Simon, A. Rinaldi, N. Gall, St Thomas’ Hospital, London, UK; M. Glikson, Sheba Medical Center, Tel-Hashomer, Israel; J. Roda, S. Villalba, V. Palanca, J. Belchi, H. General Universitario,Valencia, Spain; C. Muto, M. Canciello, G. Carreras, B. Tuccillo, Loreto Mare Hospital, Naples, Italy; A. Arenal, E. Gonzalez-Torrecillas, F. Atienza, H. Gregorio Marañon, Madrid, Spain; M. Borggrefe, S. Spehl, 1st Department of Medicine Cardiology, University Hospital Mannheim, Mannheim, Germany; J.L. Merino, R. Peinado, H. La Paz, Madrid, Spain; J.C. Rodriguez, O. Medina, J. García, H. Insular de Gran Canaria, Las Palmas, Spain; F. Morgado, Santa Cruz, Lisbon, Portugal; I. Lozano,J. Toquero, R. Arroyo, H. Puerta de Hierro, Madrid, Spain; J.M. Ormaetxe, M. Arkotxa, H. de Basurto, Bilbao, Spain; G. Steinbeck, E. Hoffman, S. Janko, U. Dorwarth, Ludwig-Maximilian-University Hospital, München, Germany; M. Geist, V. Turkisher, Wolfson Medical Center, Holon, Israel; P. Della Bella, G. Fassini, C. Carbucicchio, F. Giraldi, Centro Cardiologico Monzino, Milano, Italy; P. Golino, M. Viscusi, F. Mascia, Hospedale Civile, Caserta, Italy; L. Tercedor, M. Alvarez, H. Virgen de las Nieves, Granada, Spain; J.G. Martinez, A. Ibañez, H. General Universitario, Alicante, Spain; A. Moya, E. Rodriguez, C. Alonso, H. Valle Hebron, Barcelona, Spain; M. Lopez Gil, J. Sanz, H. 12 Octubre, Madrid, Spain; R. Garcia-Civera, R. Ruiz, S. Morell, R. SanJuan, H. Clinico Universitario, Valencia, Spain; A. García-Alberola, J. Martinez, J.J. Sanchez, H. Virgen de la Arrixaca, Murcia, Spain; M. Manz, D.Burkhardt, A. Markewitz, Krankenhaus Marienhof, Koblenz, Germany; E. Castellanos, L. Rodriguez-Padial, H. Virgen de la Salud, Toledo, Spain; M. Sassara, A. Achilli, E. Scabbia, Civile Hospital, Viterbo, Italy; J. Olagüe, J.E. Pareja, M.J. Sancho-Tello, H. La Fe, Valencia, Spain; S. Hohnloser, G. Grönefeld, Johann Wolfgang Goethe University, Frankfurt, Germany; T. Fuchs, Assaf Harofe Medical Center, Tzerifin, Israel; W. Jung, N. Schwick, B. Roggenbuck-Schwilk, Klinikum Villingen-Schwenningen, Villingen, Germany; B. Lemke, T. Lawo, T. Deneke, S. Holt, BG Kliniken Bergmannsheil, Bochum, Germany; G. Baumann, H. Bondke, M. Claus, Campus Charite Mitte, Berlin, Germany; A. Maresta, S. Silvani, D. Cornacchia, E. Tampieri, Civile Hospital, Ravenna, Italy; and J.J. Manzano, A. Medina, E. Caballero, F. Wangüemert, H. General Dr Negrín, Las Palmas, Spain.

DATAS Adverse Event Committee: J.A. Garcia Robles, Madrid, Spain; T. Korte, Hannover, Germany; X. Viñolas, Barcelona, Spain; and M.Z.-B., Genova, Italy.

References