Author + information
- Received February 12, 1998
- Revision received May 27, 1998
- Accepted August 6, 1998
- Published online December 1, 1998.
- Philippe Delfaut, MDa,
- Sanjeev Saksena, MD, FACCa,1,*,
- Atul Prakash, MD, MRCPa and
- Ryszard B. Krol, MD, PhDa
- ↵*Address for correspondence: Dr. Sanjeev Saksena, 55 Essex Street, Suite 3-2, Millburn, New Jersey 07041
Objectives. An initial crossover study comparing dual- and single-site right atrial pacing was performed followed by a long-term efficacy and safety evaluation of dual-site right atrial pacing in patients with drug-refractory atrial fibrillation (AF). Also examined was the efficacy of two single-site right atrial pacing modes (high right atrium and coronary sinus ostium) and the long-term need for cardioversion, antithrombotic and antiarrhythmic drug therapies during dual-site atrial pacing.
Methods. Thirty consecutive patients with drug-refractory symptomatic AF and documented primary or drug-induced bradycardia were implanted with a dual chamber rate-responsive pacemaker and two atrial leads. Single-site atrial pacing was performed at the high right atrium or the coronary sinus ostium. Continuous atrial pacing was maintained.
Results. Mean arrhythmia-free intervals increased from 9 ± 10 days in the control period preceding implant to 143 ± 110 days (p < 0.0001) in single-site right atrial pacing and 195 ± 96 days in dual-site right atrial pacing (p < 0.005 versus single-site pacing and p < 0.0001 versus control). Dual-site right atrial pacing significantly increased the proportion of patients free of AF recurrence (89%) as compared to single-site right atrial pacing (62%, p = 0.02). High right atrial pacing and coronary sinus ostial pacing had similar efficacy for AF prevention. Effective rhythm control was achieved in 86% of patients during dual right atrial pacing. Seventy-eight percent of patients at 1 year and 56% at 3 years remained free of symptomatic AF. The need for cardioversion was reduced after pacemaker implant (p < 0.05) and antithrombotic therapy was reduced (p < 0.06) without any thromboembolic event. Coronary sinus ostial lead dislodgement was not observed after discharge.
Conclusions. Atrial pacing in combination with antiarrhythmic drugs eliminates or markedly reduces recurrent AF. Prevention of AF is enhanced by dual-site right atrial pacing. High right atrial and coronary sinus ostial pacing do not differ in efficacy. Dual-site right atrial pacing is safe, achieves long-term rhythm control in most patients, decreases the need for cardioversion, and antithrombotic therapy can be selectively reduced.
Prevention of atrial fibrillation (AF) has become a major therapeutic challenge with increasing knowledge of the morbidity and mortality related to this widely prevalent arrhythmia (1). Numerous studies (2–4)have shown incomplete or even ineffective rhythm control with antiarrhythmic drug therapy when used alone for AF prevention. Currently, new non-pharmacologic methods such as multisite atrial pacing, catheter ablation, and direct atrial surgery are under active development for treatment of drug-refractory patients with AF. Atrial pacing has been shown to reduce AF recurrence rates as compared to ventricular pacing in retrospective (5–7)and prospective studies (8,9). Recently, it has been suggested that methods such as dual-site right atrial pacing (10)and biatrial resynchronization (11)have an incremental benefit relative to both high right atrial pacing and antiarrhythmic drug therapies on short-term follow-up. Acute electrophysiologic studies (12)have suggested that reduction in atrial conduction delay and modification of dispersion in atrial refractoriness are important mechanisms of improved AF prevention by multisite atrial pacing. Single-site atrial pacing may also be of value in arrhythmia prevention (10). The optimal pacing site for single-site pacing is still unknown, but lately, coronary sinus locations have been advocated (13). With the known proarrhythmic risk of some antiarrhythmic agents, the safety of such combination therapy is also unclear (2,14–16).
The long-term results of dual-site right atrial pacing in patients with drug-refractory AF have not been reported. Effective long-term AF prevention has important implications for antithrombotic therapy, antiarrhythmic drug and cardioversion therapies in this arrhythmia. In this article, the long-term results of a prospective study employing atrial pacing for AF prevention is reported. Firstly, a sequential crossover design was used to compare the efficacy of dual-site right atrial pacing with single-site right atrial pacing in patients with drug-refractory AF. Secondly, we also evaluated the efficacy of two single-site right atrial pacing methods, namely high right atrial pacing and coronary sinus ostial pacing to address the possibility of an optimal single right atrial pacing site. Finally, we examined the feasibility of long-term rhythm control with dual-site right atrial pacing after the completion of the crossover phases to address the need for continued antithrombotic therapy, antiarrhythmic and cardioversion therapies with this pacing mode.
The patient selection, study protocol, methods of implantation, and pacing system for our pilot study have been previously described in detail elsewhere (10). Briefly, in consecutive consenting patients with drug-refractory symptomatic AF or flutter and a documented primary or drug-induced bradycardia, we implanted a DDDR pacemaker generator (Medtronic Elite 7086 or Thera 7960i dual chamber rate responsive pacemaker, Medtronic Inc., Minneapolis, Minnesota) with dual atrial leads (Medtronic models 4568-53 and 4068-58) and a right ventricular lead (Medtronic model 5034-58). The two pacing leads were fixed in the right atrium, one at the high right atrium and the other just outside the coronary sinus ostium. An active screw-in fixation lead was invariably used at the coronary sinus ostium. The atrial leads were connected via a Y-connector to the atrial output of the pacemaker. The cathode was the high right atrial lead in the first 16 patients and the coronary sinus ostium lead in the next 14 patients. The ventricular lead was positioned at the right ventricular apex and connected to the ventricular output of the pacemaker. The pacemaker was programmed in the DDDR mode with a lower rate between 80–90 bpm; an individual rate response and the sensor threshold was selected to obtain continuous overdrive atrial pacing. Unipolar atrial pacing allowed single-site right atrial pacing from the cathodal lead only and bipolar atrial pacing established dual-site right atrial pacing from both cathodal and anodal electrodes with a biphasic P wave morphology in leads II, III, and aVF. The atrial sensing threshold was usually set at 0.5 mV. At the time of an atrial high rate event detection (>175 bpm), the atrial electrogram intervals were stored in the pacemaker memory up to a maximum of eight consecutive events. The pacemaker datalogs also stored atrial electrograms for one recorded event. Paced and sensed events were categorized continuously in device memory and also for the stored events. The percentage of atrial pacing was perioperatively reviewed to ensure that the objective of continuous atrial pacing was achieved. Continuous atrial overdrive pacing was considered acceptable if the percentage of paced atrial events was >80% of all atrial events. If this percentage of atrial pacing was not reached, the pacemaker was reprogrammed with a more aggressive atrial pacing prescription and/or the drug regimen was modified by either increasing the dose or changing the drug.
Study protocol and follow-up
The study protocol consisted of an initial prospective sequential crossover comparison of dual-site right atrial pacing and single-site right atrial pacing efficacy in prevention of AF. The patients were successively paced in these two modes initially for 90-day periods and then 180-day periods or if a symptomatic AF recurrence was noted (Fig. 1). At the crossover study completion, the patients’ pacemakers were definitively reprogrammed to the dual-site right atrial pacing mode to study long-term rhythm control.
During follow-up, all patients were followed with periodic clinic visits, accompanied by electrocardiogram (ECG) recordings, at intervals of three months or less with their cardiologist and/or their primary physician. Pacemaker clinic visits were scheduled at 1 week postoperatively, 1 month, 3 months, and then at 3-month intervals. ECGs and interrogation of the pacemaker datalogs were performed at each clinic visit. Twenty-four hour ambulatory ECG recordings were performed at 1 month and 4 months if the patient had symptoms potentially suggestive of arrhythmia recurrence, and also before antiarrhythmic drug and antithrombotic treatment withdrawal. Transtelephonic and event monitoring were also performed in symptomatic patients who also subsequently underwent a 12-lead and an ambulatory ECG.
After pacemaker implantation, all patients were usually placed on a previously ineffective antiarrhythmic drug treatment. The daily drug dose was reduced or the antiarrhythmic drug treatment was withdrawn if the patient had no arrhythmia recurrence for a minimum period of one year. Antithrombotic treatment was instituted in accordance with previously established guidelines (17). The choice of warfarin or aspirin was based on age, associated risk factors, and the presence or absence of contraindications to each therapy. Antithrombotic treatment was withheld in patients with contraindications to this therapy. If no asymptomatic or symptomatic AF episode was documented for a period of one year as determined from pacemaker data logs, periodic Holter monitoring, clinic visits and event recordings in symptomatic patients, antithrombotic treatment therapy was evaluated for potential withdrawal.
Continuous variables are expressed as mean ± SD and were compared by two-tailed paired or the unpaired Student ttest. Subgroup comparison of categorical variables were performed using Fisher’s exact test. The actuarial graphs of the primary endpoints, i.e., arrhythmia-free intervals, and time to first and subsequent AF recurrence are presented using the Kaplan-Meier life table method for actuarial analysis. For purposes of analysis of arrhythmia-free intervals in single- and dual-site right atrial pacing during the sequential crossover study periods, the two segments in each of two pacing modes were combined for a maximal arrhythmia-free interval of 270 days. The discontinuous nature of the two time periods is eliminated in the actuarial analysis. The Log Rank test was also applied for comparison of these curves. For all comparisons, the difference was considered significant at a level of 5%.
Thirty patients (14 men and 16 women) with a mean age of 69 ± 12 years were enrolled in the study and followed up for 28 ± 11 months (range 0–41). All patients had recurrent, refractory and symptomatic AF or atrial flutter of 42 ± 50 months duration with a documented bradycardia (≤45 bpm). The primary indication for cardiac pacing was sick sinus syndrome in 8 (27%) patients, conduction system disease in 6 (20%), drug-induced bradycardia (≤45 bpm) in 11 (37%) and neurocardiogenic syncope with a bradycardic mechanism in 5 (17%) patients; thus, 16 of 30 (53%) patients lacked a primary persistent bradycardia, coexisting at the time of appearance of spontaneous AF or atrial flutter. Spontaneous AF episodes when documented also confirmed the absence of a preceding bradycardia in these patients.
Twenty-six patients had AF. It was paroxysmal in 21 and chronic in 5. Six of these AF patients also had associated paroxysmal atrial flutter. Three study patients had paroxysmal atrial flutter alone and one patient had a combination of paroxysmal atrial flutter and atrial tachycardia. Their mean left atrial diameter was 39 ± 7 mm and mean left ventricular ejection fraction was 46 ± 11%. Twenty-two (73%) patients presented with cardiovascular disease, 12 patients had coronary artery disease, 5 patients had hypertensive heart disease, 2 patients had dilated cardiomyopathy, 1 patient had valvular heart disease, 1 patient had hypertrophic cardiomyopathy and 1 patient had congenital heart disease. Before pacemaker system implantation, pharmacologic treatment of AF had failed to prevent AF or flutter in all patients with a mean of 3.6 ± 1.7 antiarrhythmic drugs. Five patients have failed preceding attempts of atrial flutter and atrial tachycardia ablation. One patient withdrew from the study immediately after pacemaker implantation due to severe heart failure and entered a cardiac transplant program. This patient was censored from the analysis at that point. Four patients declined crossover to single-site atrial pacing after institution of dual-site right atrial pacing during the course of the study.
Efficacy of atrial pacing in AF prevention during crossover trial
Fig. 2shows the percent freedom from any AF or flutter recurrences in the study population in the three-month period before institution of pacing (the control period on antiarrhythmic drug therapy alone) and after institution of atrial pacing for both pacing modes during the entire duration of the pilot crossover study. Before pacing, all patients had experienced very frequent AF with at least two relapses within the prior 90 days. No patient was AF eventfree after 41 days in the control period. In the single-site right atrial pacing mode, 62% of patients were free of AF recurrence during nine months of combined follow-up in study phases 2 and 4 (p < 0.0001 versus control). Dual-site right atrial pacing significantly increased the proportion of patients free of AF recurrence in a similar follow-up in combined phases 1 and 3 to 89% (p = 0.02 versus single-site right atrial pacing and p < 0.0001 versus control). The mean arrhythmia-free interval (Fig. 3)for the entire study population was 9 ± 10 days (range 1–41 days) before pacing and increased significantly to 143 ± 110 days (range 1–270 days) in single-site right atrial pacing (p < 0.0001) and to 195 ± 96 days (range 25–270 days) in dual-site right atrial pacing (p < 0.005 versus single-site right atrial pacing and p < 0.0001 versus control). Four patients died during the course of the pilot crossover study (renal failure in 2 patients, malignancy in 1 patient, and respiratory failure in 1 patient).
Ten of 25 (40%) patients had one or more AF recurrences in single-site right atrial pacing. Five of 29 (17%) patients had AF recurrences in dual-site right atrial pacing (p = 0.06 versus single-site pacing) and all 4 patients who crossed over also had a recurrence in single-site right pacing. Three patients with paroxysmal AF and one patient with chronic AF lost rhythm control in phase 3 of the crossover study. In one patient, this was due to withdrawal of all antiarrhythmic drugs due to multiple drug intolerance. These four patients remained in chronic AF subsequently.
The relative risk for AF recurrences compared to the control period was significantly lower during dual-site right atrial pacing (0.05, 95% confidence interval [CI] 0.02–0.09), and single-site atrial pacing (0.16, 95% CI 0.05–0.20). Comparing dual- to single-site atrial pacing, the relative risk of AF recurrence was significantly lower (0.25, 95% CI 0.08–0.80).
Efficacy of individual right atrial pacing sites during single-site pacing in the crossover study
Both coronary sinus ostial pacing and high right atrial pacing (Fig. 4)significantly increased the proportion of patients free of AF recurrences at nine months when compared to the control period (53%, 71% versus 0% respectively, p < 0.0001). There was no significant difference in efficacy in AF prevention between coronary sinus ostial pacing and high right atrial pacing (p = 0.36). The mean arrhythmia-free interval (Fig. 5)compared to the control period was also significantly increased by coronary sinus ostial pacing (145 ± 125 days versus 5 ± 2 days) and high right atrial pacing as compared to the control period (142 ± 102 days versus 14 ± 13 days, p < 0.001). The mean arrhythmia-free interval was not significantly different between high right atrial and coronary sinus ostial pacing (p = 0.94). Five of 16 patients had a recurrence of AF in high right atrial pacing and 5 of the 14 remaining patients had recurrent AF in coronary sinus ostial pacing.
Comparing high right atrial to coronary sinus ostial pacing, the relative risk was not significantly lower (0.55, 95% CI 0.13–2.05); however, coronary sinus ostial pacing had significantly lower risk of AF recurrence compared to the control period (0.25, 95% CI 0.05–0.42) as did high right atrial pacing (0.11, 95% CI 0.02–0.18).
Long-term rhythm control during dual-site right atrial pacing
After the initial crossover study period which extended from 6 to 18 months, the follow-up during dual-site right atrial pacing in 21 remaining patients has ranged from 25 up to 41 months. Fourteen (67%) patients did not experience any recurrent AF during dual-site right atrial pacing in the long-term study period. One of the seven patients with AF recurrence had been without AF for 3.5 years and did have a recurrence due to loss of atrial pacing at pacemaker generator end-of-life when the device reverted to demand ventricular pacing. Rhythm control or return to atrial paced rhythm was reestablished after reinstitution of dual-site pacing and drug therapy. All six remaining patients have had only one AF recurrence which was due to drug withdrawal secondary to intolerance in one patient. The overall mean arrhythmia-free interval for the study population at last follow-up was 513 ± 360 days.
Arrhythmia control in combined crossover and long-term study
Fig. 6shows the actuarial freedom from AF recurrence and rhythm control achieved up to 3.5 years of follow-up in the two segments of our study. The proportion of patients remaining free of AF recurrence was 78% at 1 year, 63% at 2 years, and 56% at 3 years. Four (13%) patients progressed to or remained in chronic AF. At last follow-up, 25 of 29 patients in whom a paced rhythm could be established remained atrially paced. Atrial pacing could not be initially established for any significant period in the one (3%) remaining patient due to intractable heart failure and severe mitral regurgitation. This patient was referred to a heart transplant program and censored from the study. The presence or absence of symptomatic primary bradycardia in this patient population did not influence the ability to achieve rhythm control (p = 0.13); however, the extent of arrhythmia-free intervals was significantly greater (p < 0.02) in patients with primary bradycardia (>1,000 days) than in those without it (p < 0.05).
Pattern of recurrence of AF during combined single- and dual-atrial pacing
The time course and frequency of AF recurrences in patients demonstrating recurrent symptomatic AF events is shown in Fig. 7. There is an increased event rate (0.77 events/patient over a six-month interval) during the first six months. The event rate in phase 1 (dual-site right atrial pacing) was 0.23 and in phase 2 (single-site right atrial pacing) was 0.54. This observed rate was thus in part potentially attributable to the crossover into phase 2 (high right atrial pacing) since subsequently a much lower rate of events (0.3 events/patient) was observed in the subsequent six-month periods which were predominantly in dual-site right atrial pacing. In this regard, no further change in frequency is obvious after the initial six months.
The potential etiologies for recurrent AF in both pacing modes was analyzed. Of a total of 13 symptomatic and/or self-terminating AF recurrences observed in the seven patients during the pilot and long-term dual-site right atrial pacing phases, seven were related to either antiarrhythmic drug treatment modification (2 recurrences), heart failure (2 recurrences), pneumonia (1 recurrence), loss of atrial pacing (1 recurrence), or combination of heart failure and drug modification (1 patient). No causative or precipitating factor could be identified for the remaining six recurrences. During the two pilot phases of single-site right atrial pacing, 10 recurrences occurred in 10 patients. Two (20%) recurrences were attributed to a drug withdrawal (intolerance) and drug dose reduction of drug (intolerance). No etiology nor precipitating factor could be documented for the eight (80%) other recurrences. Long-term single-site pacing was not performed by study design.
Antiarrhythmic drug treatment
Fig. 8shows the distribution of antiarrhythmic drugs given to the patients immediately after implant and at last follow-up. There was a decrease in class 1 drug use because of intolerance (two patients) or inefficacy (five patients). One additional patient continues on a class 1 drug at last follow-up after being switched from a class 3 drug for drug intolerance. The number of patients initially and finally treated with class 3 drugs remained equal; however, out of the nine patients initially on class 3 drugs, only four patients remained on the same drug and dosage. This treatment was stopped in three patients, changed for another class 3 drug in one patient, and changed for a class 1 drug in one patient. Four patients were switched over to a class 3 drug. There was little overall change in class 2 and 4 drug use during the entire period of treatment. In 14 patients, we employed a combination of class 2 or 4 drugs with class 1 or 3 drugs at implant; 9 patients continued on the combination at last follow-up. The number of patients without any antiarrhythmic drug treatment has increased slightly. At last follow-up, there were five patients on no antiarrhythmic drug which included four patients who were not commenced on a drug at pacing system implant. The four patients initially untreated with drug therapy did not experience any AF recurrence.
Cardioversion therapy after dual-site right atrial pacing
Twenty-three (77%) patients in the study had cardioversion performed prior to pacing system implant in the year preceding pacemaker implant. The subsequent need for direct current cardioversion over the course of the study period is shown in Fig. 9. All symptomatic AF recurrences are also depicted. As can be seen, often self-terminating AF recurrences were observed. Patients required direct current or drug cardioversion during the follow-up period. The overall need for cardioversion is very modest and somewhat higher in the initial six months paralleling the recurrent AF pattern, reflecting in part the crossover to single-site pacing in phase 2. After six months, a relatively low monthly event rate is noted reflecting the infrequent recurrences in patients who exhibit this arrhythmia after dual-site atrial pacing.
During the study, there has been no embolic event observed clinically nor evidence of any cerebrovascular accidents. At implant, 23 (79%) patients were on antithrombotic treatment. Fourteen were on warfarin and nine on aspirin. Six (21%) patients were not on antithrombotic therapy because of contraindications in three patients, noncompliance in one patient, and complications (perioperative hematomas) in two patients. These hematomas subsided with conservative management and did not require evacuation. At last follow-up, there is a significant modification of the antithrombotic treatment. Thirteen patients (45%) who had at least one documented recurrence are still on antithrombotic therapy. Antithrombotic therapy was not being employed in 16 (55%) patients with 10 patients having initially initiated therapy being withdrawn. The mean time to withdrawal of antithrombotic treatment was 19 ± 13 months. As compared to implant period, there is a significant reduction in antithrombotic treatment usage with dual-site right atrial pacing. Warfarin has been replaced by aspirin in two patients. Eleven patients who had withdrawal of initially instituted antithrombotic therapy have been followed for 9–27 months. They have remained without such therapy in ten patients while it has been replaced with aspirin in one patient.
These were related to device implant. There was a single intraoperative coronary sinus lead dislodgement which was inadvertently located just inside the ostium and was immediately repositioned outside the ostium. There has been no lead dislodgement after hospital discharge. One patient had a pneumothorax at implant requiring evacuation. There were two late complications (pocket infection; pocket twitch due to disruption of adhesive insulation of Y-connector) which required system explant or revision.
Single-site atrial pacing has been shown to be associated with a reduced risk of recurrent AF in patients with sick sinus syndrome with and without manifest AF (5–7). Recently reported prospective studies in this population have confirmed this observation (8,9). The ability of atrial pacing to prevent AF recurrences has not been well understood nor has the quantitative extent of benefit been characterized. Furthermore, whether this benefit is a direct effect on the atrial arrhythmia or mediated by prevention of associated bradycardia has been unclear. The extension of this benefit to populations with AF alone is also unknown. While we and others have suggested increased benefit with dual-site atrial pacing, the long-term efficacy and relative benefits of this new pacing method vis-a-vis traditional single-site atrial pacing remain undefined. Whether the efficacy of single-site atrial pacing can be enhanced by new atrial pacing sites is also conjectural but has been suggested (13). Finally, it is critical to know whether this pacing preventative strategy could be combined with atrial defibrillation in implantable devices, and reduce the need for drug (antiarrhythmic or antithrombotic), cardioversion or defibrillation therapy. Our prospective study with its initial sequential crossover design between single- and dual-site right atrial pacing and subsequent long-term outcome in dual-site right atrial pacing provides insights and information relative to all these issues. The use of two single atrial sites was performed sequentially since the safety and efficacy of long-term coronary sinus ostialpacing were unknown at the time of study commencement.
AF prevention with atrial pacing
In our population with very frequent recurrent AF with an average arrhythmia-free interval of nine days, both single- and dual-site right atrial pacing, most often in combination with antiarrhythmic drug therapy, demonstrated clear benefit in prevention of recurrent AF (14). Elimination or a marked reduction in recurrent AF was achieved in 25 of 29 patients. In single-site right atrial pacing, 62% of patients had elimination of recurrent AF during a nine-month follow-up with a significantly higher proportion (82%) being arrhythmia-free in dual-site right atrial pacing. The remaining patients had a marked reduction in recurrence rates with very infrequent events (1–3)over a mean follow-up of >2 years. The mean arrhythmia-free interval increased to 143 and 195 days, respectively, during a follow-up period censored at 270 days. Clearly, with unlimited follow-up, this would be significantly longer. This is apparent during long-term follow-up with dual-site right atrial pacing when the arrhythmia-free interval has increased to a mean value of 513 ± 360 days. This observation would argue that the preventative benefits of dual-site right atrial pacing are long-term in nature.
There is clear evidence from this study to suggest that there is an incremental benefit in AF prevention with dual-site right atrial pacing over the single-site pacing methods used in the study. Five additional patients became arrhythmia-free in the former mode and the mean arrhythmia-free interval increased by 36% in the first nine months of follow-up alone. The former translates into possible benefits with potential for reduction in antiarrhythmic drug therapy and even antithrombotic therapy while the latter would reduce the need for cardioversion therapies. In patients with recurrent AF, the frequency of recurrences and intervals between them also increased with dual-site atrial pacing. Both single-site atrial pacing modes had comparable efficacy, suggesting that an ideal individual site was not present between these two choices. Biatrial pacing has also been shown to be effective in prevention of atypical left atrial flutter and AF (11). This has been thought to be due to similar mechanisms as in our patients but uses a triggered pacing algorithm.
The incidence of subsequent chronic AF in patients with paroxysmal arrhythmia at study entry was 3 of 25 (12%) patients during the study period up to 3.5 years for an annual incidence of 5%. In prior reports of single-site atrial pacing in patients with sick sinus syndrome and history of AF, the estimated incidence was >20% annually (15). While the populations are not exactly identical, these data do suggest potential further benefit with dual-site pacing.
Patient selection for preventative dual-site right atrial pacing
This benefit of atrial pacing in prevention of AF has been seen in patients with and without manifest bradyarrhythmias at the time of AF emergence. This would suggest that this beneficial effect is a direct action on the atrial arrhythmia rather than mediated by bradycardia prevention. In fact, we have not documented a single episode of recurrent spontaneous AF before or after pacing which is preceded by bradycardia. We would suggest that patients with recurrent paroxysmal or persistent AF can be candidates for this treatment modality. Four of 5 patients with chronic AF achieved rhythm control; further study is needed in this group. The presence of cardiac disease was noted in most of our patients. Lone AF patients were a relatively uncommon minority; thus, patients with organic heart disease and AF are candidates for atrial pacing therapies. Modest enlargement of the left atrium and depressed left ventricular systolic function do not contraindicate this procedure. Intractable heart failure or severe mitral regurgitation have not been associated with AF suppression and should be exclusions for patient selection at this time. Patients with primary bradycardia achieved longer arrhythmia-free intervals (>1,000 days) than those without it (p < 0.05).
Impact of atrial pacing on concomitant drug and electrical therapies
Atrial pacing established rhythm control in 82% of patients during this long-term study. Reduction in antiarrhythmic drug therapy (particularly class 1 agents) was indeed achieved in a significant minority (16%) of patients; however, a major limitation was the need for drug therapy changes for intolerance or inefficacy, most often to a class 3 drug. Our current policy is now to start with such class 3 agents, especially in view of the risk of ventricular proarrhythmia in patients with coronary heart disease and left ventricular dysfunction (15).
Antithrombotic therapies could be eliminated in ten patients, fully 33% of the study group. This was in accordance with clinical guidelines operative at the time of the study (17). While the long-term (>5 years) impact of this reduction cannot be judged from this study in terms of efficacy and safety, the absence of embolic events in this study is hypothesis generating for the future.
The need for cardioversion therapies in this population is eliminated or markedly reduced based on the clinical course of these patients. Seventy-seven percent of all patients had had one or more cardioversion therapies in the year prior to the implantation of the pacing system. Nineteen patients never required cardioversion after system implant; however, only 13 patients had recurrent AF with 44% being self-terminating in nature. Cardioversion was necessary in ten patients with three patients requiring more than one (maximum three) cardioversion. For patients with an average arrhythmia-free interval of <10 days, this frequency over a course of several years is remarkably low. Our data are compelling evidence that prevention of AF by atrial pacing reduces the need for direct current or other cardioversion therapy.
Role of atrial pacing in implantable atrial defibrillation devices
Based on our data, a role for dual-site and single-site right atrial pacing can be defined in a rhythm management device. Prototype devices have lacked this capability and shock frequency during follow-up has been variable (18). Dual- and single-site atrial pacing can reduce the need for shock therapy for cardioversion and defibrillation in such devices by reducing the frequency of recurrent atrial flutter and AF. It could be envisaged that such pacing therapies could be invariably programmed as a first line of treatment for prevention. Algorithms such as continuous atrial pacing could reduce the need for high rate pacing and avoid slow heart rates (19). Such devices could eliminate the Y-connector requirement by reconfiguring the header to allow three ports and avoid additional hardware and its problems.
Mechanism of AF prevention and remodelling of atrial myocardium
The mechanism of AF prevention with dual-site atrial pacing can only be surmised from our data. Continuous pacing can suppress triggering premature atrial beats as well as modify the conduction delay and recovery of excitability in specific atrial regions. Our current studies suggest that premature atrial beats for spontaneous AF often arise in the crista terminalis, interatrial septum and superior left atrium (20). Maximal conduction delay for right atrial premature beats exists in the septum, His bundle and coronary sinus ostial regions (18). Initiation of spontaneous AF is often in proximity to the site of atrial premature contractions. Finally, specific pacing modes may alter (reduce or exaggerate) conduction delays in these regions (21). Dual-site right atrial and biatrial pacing may exert the most favorable effects of many pacing modes in minimizing delay in these atrial regions (22).
Long-term benefit with elimination of AF recurrences may be ascribed to continuous electrical modification by overdrive pacing or atrial electrical remodelling with progressive rhythm control or both mechanisms. Based on our data, we would suggest both mechanisms are operative. Individual patients who were not paced for even brief periods could have AF recurrence, suggesting the need for continuous electrical substrate modification for effective prevention. Patients with very frequent or chronic AF without long periods in sinus rhythm would have all progressed to chronic AF without the possibility of electrical remodelling. This may in part explain the declining frequency of AF recurrences after the first six months.
Safety and study limitations
The safety of dual-site right atrial pacing is established from our data. Complications observed were conventional for atrial pacing without evidence of increased problems related to the additional atrial lead. Coronary sinus lead dislodgement after hospital discharge was not observed. This is in contrast to distal coronary sinus placement of pacing and defibrillation leads with reported rates of 8–20% (23,24). Our study does not establish the long-term rhythm control for single-site atrial pacing but this is available in some other reports (7,8). The memory in device datalogs was limited to eight events and could be saturated by high density-sensed events. As such, the total AF event rate could not be accurately assessed.
In this prospective study, atrial pacing in combination with antiarrhythmic drug therapy can eliminate or markedly reduce recurrent AF in patients with drug-refractory AF. While neither right atrial site used had superior efficacy for single-site pacing, dual-site right atrial pacing had significant incremental benefit demonstrable in the crossover study. Long-term rhythm control is effectively and safely achieved with rare progression to chronic AF, particularly in paroxysmal or persistent AF. Infrequent cardioversion may be needed in a minority of patients for this purpose. Antithrombotic therapy could be withdrawn in one third of patients without adverse embolic events in the follow-up period of 2–4 years. Dual-site atrial pacing should be an integral element of implantable arrhythmia management devices for AF incorporating bradycardia and antitachycardia pacing, cardioversion and defibrillation. Future directions in technology could include elimination of Y-connectors and availability of continuous atrial pacing algorithms as well as antitachycardia pacing and defibrillation for patients with infrequent AF recurrences in a given patient.
↵1 Dr. Saksena is or has been a consultant and/or investigator for Medtronic, Inc. (Minneapolis, Minnesota), Cardiac Pacemakers, Inc. (St. Paul, Minnesota), Intermedics, Inc. (Angleton, Texas), and Angeion, Inc. (Minneapolis, Minnesota).
☆ Dr. Delfaut is a research fellow of the Electrophysiology Research Foundation from the University of Lille, Lille, France and is supported in part by grants from the University of Lille, Medtronic Europe, and ELA-Medical, Paris, France. Drs. Krol and Prakash are coinvestigators for research protocols funded by Medtronic, Inc., Angeion Inc., and Cardiac Pacemakers, Inc.
- atrial fibrillation
- confidence interval
- Received February 12, 1998.
- Revision received May 27, 1998.
- Accepted August 6, 1998.
- American College of Cardiology
- Coplen S.E.,
- Antman F.M.,
- Berlin J.A.,
- Hewitt P.,
- Chalmers T.C.
- Juul-Moller S.,
- Edvardsson N.,
- Rehnqvist-Ahlberg N.
- Sgarbossa E.B.,
- Pinski S.L.,
- Maloney J.D.,
- et al.
- Andersen H.R.,
- Nielsen J.C.,
- Thomsen P.E.B.,
- et al.
- Saksena S.,
- Prakash A.,
- Hill M.,
- et al.
- Daubert C.,
- Mabo P.H.,
- Berder V.,
- Gras D.,
- Leclercq C.
- Prakash A.,
- Saksena S.,
- Hill M.,
- et al.
- Papageorgiou P.,
- Anselme F.,
- Kirchhof C.J.,
- et al.
- Garrigue S.,
- Cazeau S.,
- Gencel L.,
- Jais P.,
- Haissaguerre M.,
- Clementy J.
- Wyse D.G.,
- Morganroth J.,
- Ledingham R.,
- et al.
- Ricci R.,
- Azzolini P.,
- Puglisi A.,
- et al.
- Prakash A.,
- Saksena S.,
- Delfaut P.,
- Krol R.,
- Mathew P.
- Giorgberidze I.,
- Saksena S.,
- Prakash A.,
- et al.
- Prakash A.,
- Giorgberidze I.,
- Hill M.,
- et al.
- Daubert J.C.,
- Leclercq C.,
- Pavin D.,
- Mabo P.
- Bardy G.H.,
- Allen M.D.,
- Mehra R.,
- et al.