- Email: [email protected]
Vagal responses induced by endocardial left atrial autonomic ganglion stimulation before and after pulmonary vein antrum isolation for atrial fibrillation
Vagal responses induced by endocardial left atrial autonomic ganglion stimulation before and after pulmonary vein antrum isolation for atrial fibrillation Atul Verma, MD,a Walid I. Saliba, MD, Dhanumjaya Lakkireddy, MD, J. David Burkhardt, MD, Jennifer E. Cummings, MD, Oussama M. Wazni, MD, William A. Belden, MD, Sergio Thal, MD, Robert A. Schweikert, MD, David O. Martin, MD, Patrick J. Tchou, MD, Andrea Natale, MD From the Cleveland Clinic Foundation, Section of Electrophysiology, Cleveland, Ohio. BACKGROUND Elimination of vagal inputs into the left atrium (LA) may be necessary for successful catheter ablation of atrial fibrillation (AF). These vagal inputs are clustered in autonomic ganglia (AG) that are close to the pulmonary vein antrum (PVA) borders, but whether standard intracardiac echocardiography (ICE)guided PVA isolation (PVAI) affects these inputs is unknown. OBJECTIVE The purpose of this study was to assess whether standard ICE-guided PVAI affects vagal responses induced by endocardial AG stimulation in the LA. METHODS Twenty consecutive patients undergoing first-time PVAI (group 1) and 20 consecutive patients undergoing repeat PVAI for AF recurrence (group 2) were enrolled in the study. Before ablation, electrical stimulation (20 Hz, pulse duration 10 ms, voltage range 12–20 V) was performed through an 8-mm-tip ablation catheter. Based on prior data, regions around all four PVA borders were carefully mapped and stimulated to localize AG inputs. A positive stimulated vagal response was defined as atrioventricular (AV) block, asystole, or increase in mean RR interval by ⬎50%. Locations of positive vagal responses were recorded wth biplane fluoroscopy and CARTO. All patients then underwent standard ICE-guided PVAI by an operator blinded to the locations of vagal responses. Stimulation of the AG locations was then repeated postablation.
Introduction Ever since Armour et al1 reported on the anatomy of the intrinsic cardiac nervous system in humans, it has been suggested that autonomic inputs from ganglionated plexi that surround the heart may contribute to both the initiation and maintenance of atrial fibrillation (AF). High-frequency stimulation of epicardial autonomic plexi can induce triggered activity from the pulmonary veins (PVs)2 and also affect atrial refractory periods to provide a substrate for the a Atul Verma is supported by a fellowship award from the Heart and Stroke Foundation of Canada and is currently with Southlake Regional Health Centre, Newmarket, Ontario, Canada. Address reprint requests and correspondence: Andrea Natale, M.D., Co-Section Head of Pacing and Electrophysiology, Co-Chairman Center for Atrial Fibrillation, Cleveland Clinic Foundation, Desk F 15, 9500 Euclid Avenue, Cleveland, OH 44195. E-mail address: [email protected]. (Received March 1, 2007; accepted April 29, 2007.)
RESULTS Patients (age 54 ⫾ 11 years, 30% female, ejection fraction 54% ⫾ 7%) had a history of paroxysmal (75%) and persistent (25%) AF. In group 1, vagal responses were induced in all 20 patients around a mean of 3.8 ⫾ 0.4 PVAs per patient. The most common response was asystole (53%), mean RR slowing ⬎50% (28%), and AV block (20%). Postablation, vagal responses could no longer be induced in all 20 patients. A diminished response was induced (RR slowing ⬍50%) in 2/20 patients around one PVA each. In group 2, vagal responses were not induced in any of the 20 repeat patients. Stimulation capture postablation was confirmed because transient, nonsustained (⬍30 seconds) AF or atrial flutter was induced in all 40 patients with stimulation, whether vagal responses were induced or not. CONCLUSIONS Standard ICE-guided PVAI eliminates vagal responses induced by AG stimulation. Responses are not seen in patients presenting for repeat PVAI, despite clinical recurrence of AF. KEYWORDS Vagal response; Atrial fibrillation; Ablation; Autonomic; Ganglion (Heart Rhythm 2007;4:1177–1182) © 2007 Heart Rhythm Society. All rights reserved.
conversion of PV firing into sustained AF.3,4 More specifically, increased vagal tone has been shown to be a trigger for AF in a subset of patients.5 Enhanced vagal tone can increase the inducibility of AF,6 and elimination of vagal inputs may prevent AF recurrence in both animal and patient models of vagal AF.7,8 Radiofrequency (RF) ablation has emerged as an effective therapy for patients with symptomatic AF. The traditional approach of ablation has been to eliminate all of the triggers for AF by isolating the PVs.9 Recent data has suggested that identification and ablation of autonomic ganglia (AG) during PV isolation may improve long-term success.10 However, it is unknown whether these regions are already modified during standard PV antrum isolation (PVAI). AG responses are also unknown in patients who have AF recurrence postablation. Thus, the purpose of this study was (1) to assess the effects of standard PVAI guided
1547-5271/$ -see front matter © 2007 Heart Rhythm Society. All rights reserved.
doi:10.1016/j.hrthm.2007.04.023
1178 by intracardiac echocardiography (ICE) on vagal responses induced by AG stimulation and (2) to assess vagal responses induced by AG stimulation in patients with AF recurrence after failed PVAI.
Methods Patient population Twenty consecutive patients undergoing first-time PVAI for symptomatic, drug-refractory AF (group 1) and twenty consecutive patients undergoing repeat-PVAI for late AF recurrences beyond 2 months (group 2) were studied. Only patients with symptomatic paroxysmal or persistent AF were included. All patients had AF that was refractory to at least one antiarrhythmic medication. Patients with prior open-heart cardiac surgery or permanent AF were excluded from the study. All patients gave written informed consent before the stimulation and ablation procedures. Collection of patient data was performed in accordance with institutional ethics guidelines. AG stimulation protocol All patients’ procedures were done under conscious sedation with intravenous fentanyl and midazolam. For group 1 patients, AG stimulation was performed before ablation and also after ablation. In group 2 patients, AG stimulation was only performed before ablation. High-frequency stimulation of the AG stimulation was performed from the tip of an 8-mm ablation catheter (Celcius DS, Biosense Webster, Diamond Bar, CA) using a Grass stimulator model S88 with an isolation unit model SIU5 (Grass Technologies, West Warwick, RI). Stimulation was performed at 20 Hz (cycle length 50 ms) with a pulse duration of 10 ms. Voltage delivery was 12–20 V, and stimulation was applied for up to 10 seconds to assess for response. With the level of sedation used, most patients experienced mild to moderate discomfort during stimulation, but none experienced pain severe enough to stop performing stimulation. Mapping for AG stimulation was performed according to anatomic locations published elsewhere.10 In particular, the regions anterior to the right superior pulmonary vein (PV) (anterior right AG), inferior to the right inferior PV (inferior right AG), superior and medial to the left superior PV (superior left AG), and inferior to the left inferior PV (inferior left AG) were focused on. However, stimulation was not limited to these sites, and stimulation was also performed within all PV antra, along the atrial septum, and along the inferior left atrium (LA) if responses were not found in the typical locations listed above. Stimulation was not routinely performed in the superior vena cava. A positive vagal response achieved by AG stimulation was defined as transient ventricular asystole, atrioventricular (AV) block, or an increase in mean RR interval by ⬎50% during AF. Sites exhibiting positive vagal responses were tagged on a three-dimensional electroanatomic map of the LA created using the CARTO system (Biosense Webster). Sites were also recorded fluoroscopically by saving cine images in two planes.
Heart Rhythm, Vol 4, No 9, September 2007 PVAI procedure Once AG stimulation was completed and the locations of positive vagal responses were tagged on the electroanatomic map, patients underwent standard ICE-guided PVAI by an operator who was blinded to the locations of the identified AG stimulation. In other words, no attempt was made to target AG locations during ablation beyond what is normally ablated during a typical, standard ICE-guided PVAI, the technique of which is described in detail elsewhere.9 Briefly, a 10-F 64-element phased-array ultrasound imaging catheter (Siemens AG, Malvern, PA) was positioned in the right atrium via the left femoral vein. Under ICE guidance, a decapolar circular (Lasso, Biosense Webster, Diamond Bar, California) mapping catheter and an 8-mm-tip (Celcius DS) ablation catheter (Biosense Webster) were advanced into the LA via two transseptal punctures. ICE was used to define the PV antra and guide sequential placement of the Lasso catheter in all positions surrounding (and outside of) each PV. RF ablation was performed wherever PV potentials were recorded around the PV antrum. RF energy output was limited to a maximum of 70 W and 50°C and was titrated according to microbubble formation detected by ICE. RF isolation of the PV was considered complete when all PV potentials surrounding the vein were abolished, as recorded by the Lasso during sinus rhythm or coronary sinus pacing. All four PVs and the superior vena cava were isolated in every patient. No mitral annular ablation lines were performed. The patient was systemically anticoagulated with intravenous heparin to maintain an activated clotting time (ACT) of 350 – 400 seconds. If the patient was in AF that did not terminate during or after the procedure, an external electrical cardioversion was performed. Warfarin was discontinued 48 hours before the procedure. All antiarrhythmic drugs except amiodarone (which was discontinued 4 –5 months before) were discontinued at least 5 half-lives before the ablation procedure.
Follow-up All patients were discharged home the day after the procedure. Postprocedure, patients continued anticoagulation with warfarin to maintain an INR of 2.0 –3.0. In all patients, antiarrhythmic medications were continued for 2 months postablation and were chosen from one of sotalol, propafenone, flecainide, or dofetilide. Amiodarone was not used postablation. Antiarrhythmic medications were discontinued in all patients after 2 months. Procedural success was defined as a lack of AF recurrence beyond 2 months post-PVAI off antiarrhythmic medication. Patients were followed in the outpatient clinic at 3, 6, and 12 months post-PVAI. A rhythm transmitter was used in all patients to monitor for arrhythmic events for the first 3– 4 months after ablation. Patients were encouraged to transmit at least three random transmissions per week while asymptomatic. Twelve-lead electrocardiogram (ECG) and 24-hour Holter recording were done routinely in all patients immediately after the procedure and at the 3-, 6-, and 12-month follow-up visits. A routine computed tomography scan of
Verma et al
Vagal Responses Pre- and Post-PVAI
1179
Figure 1 Recording of surface ECGs, intracardiac electrograms, and arterial line tracing during high-frequency stimulation around the left superior AG in a patient undergoing PVAI. Stimulation at this site resulted in ventricular asystole as indicated by the arterial pressure tracing labeled “P1 ART.” The stimulation artifact on the surface ECGs indicates the duration of high-frequency stimulation performed. The asystolic response continued even after stimulation was stopped but recovered within about 5 seconds.
the LA was also performed at 3 months to screen for PV stenosis.
Statistical analysis All data are reported as a mean ⫾ standard deviation for continuous variables and number of subjects (%) for categorical variables unless otherwise indicated. Continuous variables were compared using the independent samples Student’s t-test, and categorical variables were compared using the 2 test. P ⬍.05 was considered statistically significant for all statistical determinations. All analysis was performed using SPSS software version 11.0 (SPSS, Chicago).
Results Baseline characteristics Of the 40 patients studied in groups 1 and 2, most were male (n ⫽ 2, 70%) with an average age of 54 ⫾ 11 years. AF was paroxysmal in 30 (75%) patients and persistent in 10 (25%)—14/20 (70%) of group 1 patients and 16/20 (80%) group 2 patients had paroxysmal AF. Mean AF duration preablation was 5.9 ⫾ 4.0 years. Patients had failed 2.2 ⫾ 0.6 antiarrhythmic medications before ablation. Structural heart disease was present in 14 (35%) patients and included hypertensive heart disease (n ⫽ 8), coronary artery disease (n ⫽ 4), and valvular heart disease (n ⫽ 2). The mean pre-PVAI ejection fraction was 53% ⫾ 7%, and LA size from transthoracic echo (parasternal long-axis view) was 4.1 ⫾ 0.9 cm. There were no significant differences in the characteristics of patients between groups 1 and 2.
Induction of vagal responses preablation (group 1) In the 20 group 1 patients (undergoing first-time ablation), AG stimulation was performed at 4 ⫾ 3 sites in each of the four AG regions described in the Methods section. A positive vagal response was seen in at least three of these regions in all 20 patients. A typical positive vagal response is depicted in Figure 1. Positive vagal responses were seen in all four AG regions in only 15/20 (75%) patients. In the remaining five patients, stimulation was performed outside of the standard four regions to look for more positive vagal responses. In two of these five patients, one positive vagal response was found outside of the standard regions: in one patient, a positive response occurred on the septal aspect of the right inferior PV, and in another, a positive response occurred on the posterior roof of the right superior PV (RSPV). In the other three of five patients, no more vagal responses were found. Thus, 17/20 (85%) patients had four positive sites and 3/20 (15%) had three positive sites (mean 3.8 ⫾ 0.4 positive vagal responses per patient). Figures 2A and 2B show typical locations for vagal responses in two patients. Of the total 77 positive response sites found in the 20 group 1 patients, 41 demonstrated transient asystole (53%), 21 mean RR slowing ⬎50% (27%), and 15 transient AV block (19%). Mapping for AG sites by performing stimulation took a mean of 44 ⫾ 20 minutes per patient (range 33–52 minutes). Induction of vagal responses postablation (group 1) All 20 group 1 patients underwent standard ICE-guided PVAI by an operator blinded to the sites of positive vagal
1180
Heart Rhythm, Vol 4, No 9, September 2007
Figure 2 Fluoroscopy and three-dimensional electroanatomic images showing the catheter position during stimulation of the left superior AG (A) and the right inferior AG (B) in a patient undergoing PVAI. In both images, the fluoroscopic image is shown on the left with a tracing of where the pulmonary vein is located in relation to the catheter position. The electroanatomic image is seen on the right with a circle and arrow indicating the corresponding fluoroscopic location to the location on the map.
responses. All four PV antra were successfully isolated in all patients. Mean RF ablation time was 51 ⫾ 30 minutes with a mean power of 45 ⫾ 10 W. Postablation, AG stimulation was performed in the same locations where positive responses were found preablation for each patient. However, positive vagal responses were not induced in any of the 20 patients. In two patients, diminished vagal responses were seen in one site each where there was RR interval slowing, but not ⬎50% from baseline. When maps of the ablations were examined postablation, the standard PVAI lesion set covered all of the positive vagal response sites in all 20 patients.
Interestingly, vagal responses during ablation were only noted in two (10%) patients. In one patient, vagal response (transient asystole) occurred while ablating along the left superior PV (LSPV)-LA appendage ridge (AG stimulationinduced response occurred in the posterior roof of LSPV), and in the other, extreme RR slowing occurred while ablating in the posterior roof of the RSPV (AG stimulationinduced response on anterior aspect of RSPV). In both cases, the location of the ablation-induced response was quite distant from the locations of the AG stimulationinduced responses. Preablation stimulation in areas of close proximity to the sites where vagal response occurred during
Verma et al
Vagal Responses Pre- and Post-PVAI
ablation in these two patients did not result in vagal response. After initial ablation, 16/20 (80%) group 1 patients remained arrhythmia-free after a mean follow-up of 10 ⫾ 3 months. Four of 20 patients did have clinical AF recurrence.
1181
Localizing AG sites for ablation
Induction of vagal responses in repeat patients (group 2) Group 2 patients had their repeat ablations performed 4 ⫾ 2 months after the initial ablation procedure. Nineteen (95%) of the 20 patients presented with documented AF as the recurrent arrhythmia on ECG, Holter, or rhythm transmitter. Only 1/20 (5%) presented with documented atypical atrial flutter as the recurrent arrhythmia. Before repeat ablation, AG stimulation was performed at 5 ⫾ 3 sites in each of the four AG regions for all patients. No positive vagal responses were induced. Further AG stimulation was performed at 22 ⫾ 7 sites outside of the four AG regions in the PV antra, LA septum, and inferior LA. Still, no positive vagal responses were induced. Local capture of the high-frequency stimulation was confirmed because AG stimulation induced transient, nonsustained (⬍30 seconds) AF or atrial flutter (AFL) in all tested sites, despite the lack of positive vagal response. During the repeat ablation, no vagal responses were seen in any of the group 2 patients. No vagal responses were seen during these patients’ initial (first) ablations either. In the 20 group 2 patients, reconnection of conduction between at least one PV antrum and the LA was found. Reisolation of the PV antra was performed without additional linear lesions in all group 2 patients (mean RF time 32 ⫾ 16 minutes). After 12 ⫾ 3 months of follow-up, 18/20 (90%) of group 2 patients have remained arrhythmia-free.
Preliminary data have suggested that searching for AG sites and targeting them for ablation in addition to conventional PVAI may significantly improve overall outcome.10 In this study, we found that identification of AG sites was possible in all of the included patients and that AG sites occurred in relatively consistent locations from patient to patient. However, AG stimulation added a significant amount of time to the procedure (mean 44 minutes), added patient discomfort, and required the use of specialized stimulation equipment. Importantly, this study showed that AG sites were already ablated in the course of a standard ICE-guided PVAI performed by an operator blinded to the locations of the AG. Another study has also reported that in most patients undergoing circumferential LA ablation, the sites where AGs have been described are encompassed in the standard lesion set.11 Thus, searching for AGs may be redundant and may unnecessarily add length and discomfort to the ablation procedure. Vagal responses during ablation are rare, and they are not a good surrogate for localizing or ablating AG sites. We found that ablation-induced vagal responses occurred in only two patients, and when they happened, they were not in the same location as an AG. These findings are consistent with other reports.12,13 Furthermore, preablation stimulation performed at sites in close proximity to the ablation-induced vagal responses failed to induce a vagal response. Vagal responses during ablation may be related to pain during ablation since both patients who had vagal responses during ablation also complained of pain at the time. There are also data to suggest that ablation outcome does not differ in patients who exhibit a vagal response during ablation compared with those who do not.14
Discussion
Relationship of AG ablation to outcome
Main findings This study demonstrates that in patients undergoing a firsttime, ICE-guided PVAI procedure, vagal responses can be easily and reliably induced by AG stimulation in predictable anatomical locations in the LA. More importantly, these vagal responses are always eliminated by the standard PVAI lesion set, without specifically targeting these sites. This observation suggests that the success of PVAI in curing AF may, in part, be due to the elimination of these AG inputs in the LA. Alternatively, elimination of these inputs may be unrelated to AF cure postablation. That is because this study also demonstrates that patients who have previously undergone PVAI may continue to have recurrent AF, despite having absent vagal responses at the time of repeat ablation. Finally, we also observed that vagal responses induced during ablation do not occur often, in spite of elimination of autonomic inputs, and have little anatomic relationship to AG sites identified by AG stimulation. This study is the first to examine the relationship of AG sites to standard lesion sets during PVAI and also to look at the results of AG stimulation in a repeat population.
It has long been known that stimulation of autonomic nerves innervating the heart could induce AF.15 Thus, ever since Armour et al1 first described the existence of AGs around the heart, there has been interest in the relationship of these sites to the pathophysiology of AF. Stimulation of AGs has been shown to be involved in both the initiation and maintenance of AF in experimental animal models.2,3 Furthermore, catheter ablation of AGs can prevent vagal AF in a canine model.7 The fact that the standard PVAI lesion set encompasses AG sites in all patients may also suggest that the mechanism by which PVAI works is a result of autonomic denervation. Based on these observations, it may seem promising that selective ablation of AG sites alone may successfully control AF. However, this study found that almost all patients presenting for repeat ablation continued to have AF in spite of having absent vagal responses. This calls into question the relationship of AG site ablation to the elimination of AF. It is possible that postablation AF is caused by a different mechanism than baseline AF. For example, ablation-induced scars may serve as a substrate for macroreentrant tachycardias (flutters) that may degenerate into AF. However, we have previously
1182 shown that the incidence of flutter after ICE-guided PVAI is less than 3%,16 and in this study, 95% of the group 2 patients had AF (not flutter) documented as their recurrent arrhythmia. Furthermore, a study of AG ablation in patients undergoing first-time ablation for vagal AF showed that only two (28%) of seven patients remained free of AF postablation, while three (100%) of three patients who underwent PV isolation were cured.8 Lemery et al12 also showed that in two of 14 patients where AG sites were definitely not ablated, AF was successfully cured by ablating only the trigger sites. In the present study, none of the repeat patients had recurrent vagal responses, but all had recurrent PV-LA conduction, and elimination of this conduction resulted in a high rate of long-term success. Scherlag et al10 reported that addition of AG ablation to PV antral isolation significantly improved outcome. However, the fact that the AG sites were not already encompassed in the PVAI lesion set suggests that the set was not as extensive as that employed by our group and others. It is therefore possible that when they added extra ablation to cover the AGs, the lesion set may have been extended to isolate more of the PV antra, and this is why the combined ablation was more successful. Extensive ablation of the PV antra may also change autonomic tone in patients where no AGs can be identified,8,17 suggesting that the mechanism may be unrelated to the AG sites.
Clinical implications The data from this study question the need to precisely identify AG sites during AF ablation if a standard PVAI lesion set already encompasses most, if not all, of these sites. Although it may be attractive to limit the total amount of ablation performed by targeting only the AG sites, the fact that repeat patients can continue to have recurrent AF in the absence of positive vagal responses calls into question such a limited strategy. Clearly, further study into the utility of AG site localization and ablation is required before such a strategy can be widely adopted.
Study limitations Only a small number of patients were studied in each group, although the fact that the results were the same in nearly all of the patients in each group suggests that the findings would be similar in a larger study. Furthermore, we only assessed the location of AG sites by performing endocardial high-frequency stimulation. It is possible that more AG sites may have been identified had epicardial stimulation been performed. However, the two largest previous reports employed only endocardial stimulation,10,12 and in the one report that performed epicardial stimulation,8 AG sites were found in similar locations. Finally, since we did not test a strategy of AG ablation alone and compare it with PV antral isolation, this study cannot make any definitive conclusions about the relative role that AG inputs play in the pathophysiology of human AF versus the role of PV antral triggers.
Heart Rhythm, Vol 4, No 9, September 2007
Conclusions Standard ICE-guided PVAI eliminates vagal responses induced by AG stimulation. Vagal responses during ablation do not correlate with vagal responses induced by AG stimulation. Vagal responses with AG stimulation are not seen in patients presenting for repeat PVAI despite clinical recurrence of AF.
References 1. Armour JA, Murphy DA, Yuan BX, Macdonald S, Hopkins DA. Gross and microscopic anatomy of the human intrinsic cardiac nervous system. Anat Rec 1997;247:289 –298. 2. Patterson E, Po SS, Scherlag BJ, Lazzara R. Triggered firing in pulmonary veins initiated by in vitro autonomic nerve stimulation. Heart Rhythm 2005;2:624 – 631. 3. Liu L, Nattel S. Differing sympathetic and vagal effects on atrial fibrillation in dogs: role of refractoriness heterogeneity. Am J Physiol 1997;273:H805– 816. 4. Po SS, Li Y, Tang D, Liu H, Geng N, Jackman WM, Scherlag B, Lazzara R, Patterson E. Rapid and stable re-entry within the pulmonary vein as a mechanism initiating paroxysmal atrial fibrillation. J Am Coll Cardiol 2005;45:1871– 1877. 5. Coumel P, Attuel P, Lavallee J, Flammang D, Leclercq JF, Slama R. [The atrial arrhythmia syndrome of vagal origin]. Archives des maladies du coeur et des vaisseaux. 1978;71:645– 656. 6. Sharifov OF, Zaitsev AV, Rosenshtraukh LV, Kaliadin AY, Beloshapko GG, Yushmanova AV, Schuessler RB, Boineau JP. Spatial distribution and frequency dependence of arrhythmogenic vagal effects in canine atria. J Cardiovasc Electrophys 2000;11:1029 –1042. 7. Schauerte P, Scherlag BJ, Pitha J, Scherlag MA, Reynolds D, Lazzara R, Jackman WM. Catheter ablation of cardiac autonomic nerves for prevention of vagal atrial fibrillation. Circulation 2000;102:2774 –2780. 8. Scanavacca M, Pisani CF, Hachul D, Lara S, Hardy C, Darrieux F, Trombetta I, Negrao CE, Sosa E. Selective atrial vagal denervation guided by evoked vagal reflex to treat patients with paroxysmal atrial fibrillation. Circulation 2006;114: 876 – 885. 9. Verma A, Marrouche NF, Natale A. Pulmonary vein antrum isolation: intracardiac echocardiography-guided technique. J Cardiovasc Electrophys 2004;15: 1335–1340. 10. Scherlag BJ, Nakagawa H, Jackman WM, Yamanashi WS, Patterson E, Po S, Lazzara R. Electrical stimulation to identify neural elements on the heart: their role in atrial fibrillation. J Interv Card Electrophysiol 2005;13(Suppl 1):37– 42. 11. Pappone C, Santinelli V, Manguso F, Vicedomini G, Gugliotta F, Augello G, Mazzone P, Tortoriello V, Landoni G, Zangrillo A, Lang C, Tomita T, Mesas C, Mastella E, Alfieri O. Pulmonary vein denervation enhances long-term benefit after circumferential ablation for paroxysmal atrial fibrillation. Circulation 2004; 109:327–334. 12. Lemery R, Birnie D, Tang AS, Green M, Gollob M. Feasibility study of endocardial mapping of ganglionated plexuses during catheter ablation of atrial fibrillation. Heart Rhythm 2006;3:387–396. 13. Haissaguerre M, Jais P, Shah DC, Takahashi A, Hocini M, Quiniou G, Garrigue S, Le Mouroux A, Le Metayer P, Clementy J. Spontaneous initiation of atrial fibrillation by ectopic beats originating in the pulmonary veins. N Eng J Med 1998;339:659 – 666. 14. Takahashi Y, Jais P, Hocini M, Sanders P, Rotter M, Rostock T, Hsu LF, Sacher F, Clementy J, Haissaguerre M. Shortening of fibrillatory cycle length in the pulmonary vein during vagal excitation. J Am Coll Cardiol 2006;47:774 –780. 15. Armour JA, Hageman GR, Randall WC. Arrhythmias induced by local cardiac nerve stimulation. Am J Physiol 1972;223:1068 –1075. 16. Cummings JE, Schweikert R, Saliba W, Hao S, Martin DO, Marrouche NF, Burkhardt JD, Kilicaslan F, Verma A, Beheiry S, Belden W, Natale A. Left atrial flutter following pulmonary vein antrum isolation with radiofrequency energy: linear lesions or repeat isolation. J Cardiovasc Electrophysiol 2005;16:293–297. 17. Hsieh MH, Chiou CW, Wen ZC, Wu CH, Tai CT, Tsai CF, Ding YA, Chang MS, Chen SA. Alterations of heart rate variability after radiofrequency catheter ablation of focal atrial fibrillation originating from pulmonary veins. Circulation 1999;100:2237–2243.
Introduction Ever since Armour et al1 reported on the anatomy of the intrinsic cardiac nervous system in humans, it has been suggested that autonomic inputs from ganglionated plexi that surround the heart may contribute to both the initiation and maintenance of atrial fibrillation (AF). High-frequency stimulation of epicardial autonomic plexi can induce triggered activity from the pulmonary veins (PVs)2 and also affect atrial refractory periods to provide a substrate for the a Atul Verma is supported by a fellowship award from the Heart and Stroke Foundation of Canada and is currently with Southlake Regional Health Centre, Newmarket, Ontario, Canada. Address reprint requests and correspondence: Andrea Natale, M.D., Co-Section Head of Pacing and Electrophysiology, Co-Chairman Center for Atrial Fibrillation, Cleveland Clinic Foundation, Desk F 15, 9500 Euclid Avenue, Cleveland, OH 44195. E-mail address: [email protected]. (Received March 1, 2007; accepted April 29, 2007.)
RESULTS Patients (age 54 ⫾ 11 years, 30% female, ejection fraction 54% ⫾ 7%) had a history of paroxysmal (75%) and persistent (25%) AF. In group 1, vagal responses were induced in all 20 patients around a mean of 3.8 ⫾ 0.4 PVAs per patient. The most common response was asystole (53%), mean RR slowing ⬎50% (28%), and AV block (20%). Postablation, vagal responses could no longer be induced in all 20 patients. A diminished response was induced (RR slowing ⬍50%) in 2/20 patients around one PVA each. In group 2, vagal responses were not induced in any of the 20 repeat patients. Stimulation capture postablation was confirmed because transient, nonsustained (⬍30 seconds) AF or atrial flutter was induced in all 40 patients with stimulation, whether vagal responses were induced or not. CONCLUSIONS Standard ICE-guided PVAI eliminates vagal responses induced by AG stimulation. Responses are not seen in patients presenting for repeat PVAI, despite clinical recurrence of AF. KEYWORDS Vagal response; Atrial fibrillation; Ablation; Autonomic; Ganglion (Heart Rhythm 2007;4:1177–1182) © 2007 Heart Rhythm Society. All rights reserved.
conversion of PV firing into sustained AF.3,4 More specifically, increased vagal tone has been shown to be a trigger for AF in a subset of patients.5 Enhanced vagal tone can increase the inducibility of AF,6 and elimination of vagal inputs may prevent AF recurrence in both animal and patient models of vagal AF.7,8 Radiofrequency (RF) ablation has emerged as an effective therapy for patients with symptomatic AF. The traditional approach of ablation has been to eliminate all of the triggers for AF by isolating the PVs.9 Recent data has suggested that identification and ablation of autonomic ganglia (AG) during PV isolation may improve long-term success.10 However, it is unknown whether these regions are already modified during standard PV antrum isolation (PVAI). AG responses are also unknown in patients who have AF recurrence postablation. Thus, the purpose of this study was (1) to assess the effects of standard PVAI guided
1547-5271/$ -see front matter © 2007 Heart Rhythm Society. All rights reserved.
doi:10.1016/j.hrthm.2007.04.023
1178 by intracardiac echocardiography (ICE) on vagal responses induced by AG stimulation and (2) to assess vagal responses induced by AG stimulation in patients with AF recurrence after failed PVAI.
Methods Patient population Twenty consecutive patients undergoing first-time PVAI for symptomatic, drug-refractory AF (group 1) and twenty consecutive patients undergoing repeat-PVAI for late AF recurrences beyond 2 months (group 2) were studied. Only patients with symptomatic paroxysmal or persistent AF were included. All patients had AF that was refractory to at least one antiarrhythmic medication. Patients with prior open-heart cardiac surgery or permanent AF were excluded from the study. All patients gave written informed consent before the stimulation and ablation procedures. Collection of patient data was performed in accordance with institutional ethics guidelines. AG stimulation protocol All patients’ procedures were done under conscious sedation with intravenous fentanyl and midazolam. For group 1 patients, AG stimulation was performed before ablation and also after ablation. In group 2 patients, AG stimulation was only performed before ablation. High-frequency stimulation of the AG stimulation was performed from the tip of an 8-mm ablation catheter (Celcius DS, Biosense Webster, Diamond Bar, CA) using a Grass stimulator model S88 with an isolation unit model SIU5 (Grass Technologies, West Warwick, RI). Stimulation was performed at 20 Hz (cycle length 50 ms) with a pulse duration of 10 ms. Voltage delivery was 12–20 V, and stimulation was applied for up to 10 seconds to assess for response. With the level of sedation used, most patients experienced mild to moderate discomfort during stimulation, but none experienced pain severe enough to stop performing stimulation. Mapping for AG stimulation was performed according to anatomic locations published elsewhere.10 In particular, the regions anterior to the right superior pulmonary vein (PV) (anterior right AG), inferior to the right inferior PV (inferior right AG), superior and medial to the left superior PV (superior left AG), and inferior to the left inferior PV (inferior left AG) were focused on. However, stimulation was not limited to these sites, and stimulation was also performed within all PV antra, along the atrial septum, and along the inferior left atrium (LA) if responses were not found in the typical locations listed above. Stimulation was not routinely performed in the superior vena cava. A positive vagal response achieved by AG stimulation was defined as transient ventricular asystole, atrioventricular (AV) block, or an increase in mean RR interval by ⬎50% during AF. Sites exhibiting positive vagal responses were tagged on a three-dimensional electroanatomic map of the LA created using the CARTO system (Biosense Webster). Sites were also recorded fluoroscopically by saving cine images in two planes.
Heart Rhythm, Vol 4, No 9, September 2007 PVAI procedure Once AG stimulation was completed and the locations of positive vagal responses were tagged on the electroanatomic map, patients underwent standard ICE-guided PVAI by an operator who was blinded to the locations of the identified AG stimulation. In other words, no attempt was made to target AG locations during ablation beyond what is normally ablated during a typical, standard ICE-guided PVAI, the technique of which is described in detail elsewhere.9 Briefly, a 10-F 64-element phased-array ultrasound imaging catheter (Siemens AG, Malvern, PA) was positioned in the right atrium via the left femoral vein. Under ICE guidance, a decapolar circular (Lasso, Biosense Webster, Diamond Bar, California) mapping catheter and an 8-mm-tip (Celcius DS) ablation catheter (Biosense Webster) were advanced into the LA via two transseptal punctures. ICE was used to define the PV antra and guide sequential placement of the Lasso catheter in all positions surrounding (and outside of) each PV. RF ablation was performed wherever PV potentials were recorded around the PV antrum. RF energy output was limited to a maximum of 70 W and 50°C and was titrated according to microbubble formation detected by ICE. RF isolation of the PV was considered complete when all PV potentials surrounding the vein were abolished, as recorded by the Lasso during sinus rhythm or coronary sinus pacing. All four PVs and the superior vena cava were isolated in every patient. No mitral annular ablation lines were performed. The patient was systemically anticoagulated with intravenous heparin to maintain an activated clotting time (ACT) of 350 – 400 seconds. If the patient was in AF that did not terminate during or after the procedure, an external electrical cardioversion was performed. Warfarin was discontinued 48 hours before the procedure. All antiarrhythmic drugs except amiodarone (which was discontinued 4 –5 months before) were discontinued at least 5 half-lives before the ablation procedure.
Follow-up All patients were discharged home the day after the procedure. Postprocedure, patients continued anticoagulation with warfarin to maintain an INR of 2.0 –3.0. In all patients, antiarrhythmic medications were continued for 2 months postablation and were chosen from one of sotalol, propafenone, flecainide, or dofetilide. Amiodarone was not used postablation. Antiarrhythmic medications were discontinued in all patients after 2 months. Procedural success was defined as a lack of AF recurrence beyond 2 months post-PVAI off antiarrhythmic medication. Patients were followed in the outpatient clinic at 3, 6, and 12 months post-PVAI. A rhythm transmitter was used in all patients to monitor for arrhythmic events for the first 3– 4 months after ablation. Patients were encouraged to transmit at least three random transmissions per week while asymptomatic. Twelve-lead electrocardiogram (ECG) and 24-hour Holter recording were done routinely in all patients immediately after the procedure and at the 3-, 6-, and 12-month follow-up visits. A routine computed tomography scan of
Verma et al
Vagal Responses Pre- and Post-PVAI
1179
Figure 1 Recording of surface ECGs, intracardiac electrograms, and arterial line tracing during high-frequency stimulation around the left superior AG in a patient undergoing PVAI. Stimulation at this site resulted in ventricular asystole as indicated by the arterial pressure tracing labeled “P1 ART.” The stimulation artifact on the surface ECGs indicates the duration of high-frequency stimulation performed. The asystolic response continued even after stimulation was stopped but recovered within about 5 seconds.
the LA was also performed at 3 months to screen for PV stenosis.
Statistical analysis All data are reported as a mean ⫾ standard deviation for continuous variables and number of subjects (%) for categorical variables unless otherwise indicated. Continuous variables were compared using the independent samples Student’s t-test, and categorical variables were compared using the 2 test. P ⬍.05 was considered statistically significant for all statistical determinations. All analysis was performed using SPSS software version 11.0 (SPSS, Chicago).
Results Baseline characteristics Of the 40 patients studied in groups 1 and 2, most were male (n ⫽ 2, 70%) with an average age of 54 ⫾ 11 years. AF was paroxysmal in 30 (75%) patients and persistent in 10 (25%)—14/20 (70%) of group 1 patients and 16/20 (80%) group 2 patients had paroxysmal AF. Mean AF duration preablation was 5.9 ⫾ 4.0 years. Patients had failed 2.2 ⫾ 0.6 antiarrhythmic medications before ablation. Structural heart disease was present in 14 (35%) patients and included hypertensive heart disease (n ⫽ 8), coronary artery disease (n ⫽ 4), and valvular heart disease (n ⫽ 2). The mean pre-PVAI ejection fraction was 53% ⫾ 7%, and LA size from transthoracic echo (parasternal long-axis view) was 4.1 ⫾ 0.9 cm. There were no significant differences in the characteristics of patients between groups 1 and 2.
Induction of vagal responses preablation (group 1) In the 20 group 1 patients (undergoing first-time ablation), AG stimulation was performed at 4 ⫾ 3 sites in each of the four AG regions described in the Methods section. A positive vagal response was seen in at least three of these regions in all 20 patients. A typical positive vagal response is depicted in Figure 1. Positive vagal responses were seen in all four AG regions in only 15/20 (75%) patients. In the remaining five patients, stimulation was performed outside of the standard four regions to look for more positive vagal responses. In two of these five patients, one positive vagal response was found outside of the standard regions: in one patient, a positive response occurred on the septal aspect of the right inferior PV, and in another, a positive response occurred on the posterior roof of the right superior PV (RSPV). In the other three of five patients, no more vagal responses were found. Thus, 17/20 (85%) patients had four positive sites and 3/20 (15%) had three positive sites (mean 3.8 ⫾ 0.4 positive vagal responses per patient). Figures 2A and 2B show typical locations for vagal responses in two patients. Of the total 77 positive response sites found in the 20 group 1 patients, 41 demonstrated transient asystole (53%), 21 mean RR slowing ⬎50% (27%), and 15 transient AV block (19%). Mapping for AG sites by performing stimulation took a mean of 44 ⫾ 20 minutes per patient (range 33–52 minutes). Induction of vagal responses postablation (group 1) All 20 group 1 patients underwent standard ICE-guided PVAI by an operator blinded to the sites of positive vagal
1180
Heart Rhythm, Vol 4, No 9, September 2007
Figure 2 Fluoroscopy and three-dimensional electroanatomic images showing the catheter position during stimulation of the left superior AG (A) and the right inferior AG (B) in a patient undergoing PVAI. In both images, the fluoroscopic image is shown on the left with a tracing of where the pulmonary vein is located in relation to the catheter position. The electroanatomic image is seen on the right with a circle and arrow indicating the corresponding fluoroscopic location to the location on the map.
responses. All four PV antra were successfully isolated in all patients. Mean RF ablation time was 51 ⫾ 30 minutes with a mean power of 45 ⫾ 10 W. Postablation, AG stimulation was performed in the same locations where positive responses were found preablation for each patient. However, positive vagal responses were not induced in any of the 20 patients. In two patients, diminished vagal responses were seen in one site each where there was RR interval slowing, but not ⬎50% from baseline. When maps of the ablations were examined postablation, the standard PVAI lesion set covered all of the positive vagal response sites in all 20 patients.
Interestingly, vagal responses during ablation were only noted in two (10%) patients. In one patient, vagal response (transient asystole) occurred while ablating along the left superior PV (LSPV)-LA appendage ridge (AG stimulationinduced response occurred in the posterior roof of LSPV), and in the other, extreme RR slowing occurred while ablating in the posterior roof of the RSPV (AG stimulationinduced response on anterior aspect of RSPV). In both cases, the location of the ablation-induced response was quite distant from the locations of the AG stimulationinduced responses. Preablation stimulation in areas of close proximity to the sites where vagal response occurred during
Verma et al
Vagal Responses Pre- and Post-PVAI
ablation in these two patients did not result in vagal response. After initial ablation, 16/20 (80%) group 1 patients remained arrhythmia-free after a mean follow-up of 10 ⫾ 3 months. Four of 20 patients did have clinical AF recurrence.
1181
Localizing AG sites for ablation
Induction of vagal responses in repeat patients (group 2) Group 2 patients had their repeat ablations performed 4 ⫾ 2 months after the initial ablation procedure. Nineteen (95%) of the 20 patients presented with documented AF as the recurrent arrhythmia on ECG, Holter, or rhythm transmitter. Only 1/20 (5%) presented with documented atypical atrial flutter as the recurrent arrhythmia. Before repeat ablation, AG stimulation was performed at 5 ⫾ 3 sites in each of the four AG regions for all patients. No positive vagal responses were induced. Further AG stimulation was performed at 22 ⫾ 7 sites outside of the four AG regions in the PV antra, LA septum, and inferior LA. Still, no positive vagal responses were induced. Local capture of the high-frequency stimulation was confirmed because AG stimulation induced transient, nonsustained (⬍30 seconds) AF or atrial flutter (AFL) in all tested sites, despite the lack of positive vagal response. During the repeat ablation, no vagal responses were seen in any of the group 2 patients. No vagal responses were seen during these patients’ initial (first) ablations either. In the 20 group 2 patients, reconnection of conduction between at least one PV antrum and the LA was found. Reisolation of the PV antra was performed without additional linear lesions in all group 2 patients (mean RF time 32 ⫾ 16 minutes). After 12 ⫾ 3 months of follow-up, 18/20 (90%) of group 2 patients have remained arrhythmia-free.
Preliminary data have suggested that searching for AG sites and targeting them for ablation in addition to conventional PVAI may significantly improve overall outcome.10 In this study, we found that identification of AG sites was possible in all of the included patients and that AG sites occurred in relatively consistent locations from patient to patient. However, AG stimulation added a significant amount of time to the procedure (mean 44 minutes), added patient discomfort, and required the use of specialized stimulation equipment. Importantly, this study showed that AG sites were already ablated in the course of a standard ICE-guided PVAI performed by an operator blinded to the locations of the AG. Another study has also reported that in most patients undergoing circumferential LA ablation, the sites where AGs have been described are encompassed in the standard lesion set.11 Thus, searching for AGs may be redundant and may unnecessarily add length and discomfort to the ablation procedure. Vagal responses during ablation are rare, and they are not a good surrogate for localizing or ablating AG sites. We found that ablation-induced vagal responses occurred in only two patients, and when they happened, they were not in the same location as an AG. These findings are consistent with other reports.12,13 Furthermore, preablation stimulation performed at sites in close proximity to the ablation-induced vagal responses failed to induce a vagal response. Vagal responses during ablation may be related to pain during ablation since both patients who had vagal responses during ablation also complained of pain at the time. There are also data to suggest that ablation outcome does not differ in patients who exhibit a vagal response during ablation compared with those who do not.14
Discussion
Relationship of AG ablation to outcome
Main findings This study demonstrates that in patients undergoing a firsttime, ICE-guided PVAI procedure, vagal responses can be easily and reliably induced by AG stimulation in predictable anatomical locations in the LA. More importantly, these vagal responses are always eliminated by the standard PVAI lesion set, without specifically targeting these sites. This observation suggests that the success of PVAI in curing AF may, in part, be due to the elimination of these AG inputs in the LA. Alternatively, elimination of these inputs may be unrelated to AF cure postablation. That is because this study also demonstrates that patients who have previously undergone PVAI may continue to have recurrent AF, despite having absent vagal responses at the time of repeat ablation. Finally, we also observed that vagal responses induced during ablation do not occur often, in spite of elimination of autonomic inputs, and have little anatomic relationship to AG sites identified by AG stimulation. This study is the first to examine the relationship of AG sites to standard lesion sets during PVAI and also to look at the results of AG stimulation in a repeat population.
It has long been known that stimulation of autonomic nerves innervating the heart could induce AF.15 Thus, ever since Armour et al1 first described the existence of AGs around the heart, there has been interest in the relationship of these sites to the pathophysiology of AF. Stimulation of AGs has been shown to be involved in both the initiation and maintenance of AF in experimental animal models.2,3 Furthermore, catheter ablation of AGs can prevent vagal AF in a canine model.7 The fact that the standard PVAI lesion set encompasses AG sites in all patients may also suggest that the mechanism by which PVAI works is a result of autonomic denervation. Based on these observations, it may seem promising that selective ablation of AG sites alone may successfully control AF. However, this study found that almost all patients presenting for repeat ablation continued to have AF in spite of having absent vagal responses. This calls into question the relationship of AG site ablation to the elimination of AF. It is possible that postablation AF is caused by a different mechanism than baseline AF. For example, ablation-induced scars may serve as a substrate for macroreentrant tachycardias (flutters) that may degenerate into AF. However, we have previously
1182 shown that the incidence of flutter after ICE-guided PVAI is less than 3%,16 and in this study, 95% of the group 2 patients had AF (not flutter) documented as their recurrent arrhythmia. Furthermore, a study of AG ablation in patients undergoing first-time ablation for vagal AF showed that only two (28%) of seven patients remained free of AF postablation, while three (100%) of three patients who underwent PV isolation were cured.8 Lemery et al12 also showed that in two of 14 patients where AG sites were definitely not ablated, AF was successfully cured by ablating only the trigger sites. In the present study, none of the repeat patients had recurrent vagal responses, but all had recurrent PV-LA conduction, and elimination of this conduction resulted in a high rate of long-term success. Scherlag et al10 reported that addition of AG ablation to PV antral isolation significantly improved outcome. However, the fact that the AG sites were not already encompassed in the PVAI lesion set suggests that the set was not as extensive as that employed by our group and others. It is therefore possible that when they added extra ablation to cover the AGs, the lesion set may have been extended to isolate more of the PV antra, and this is why the combined ablation was more successful. Extensive ablation of the PV antra may also change autonomic tone in patients where no AGs can be identified,8,17 suggesting that the mechanism may be unrelated to the AG sites.
Clinical implications The data from this study question the need to precisely identify AG sites during AF ablation if a standard PVAI lesion set already encompasses most, if not all, of these sites. Although it may be attractive to limit the total amount of ablation performed by targeting only the AG sites, the fact that repeat patients can continue to have recurrent AF in the absence of positive vagal responses calls into question such a limited strategy. Clearly, further study into the utility of AG site localization and ablation is required before such a strategy can be widely adopted.
Study limitations Only a small number of patients were studied in each group, although the fact that the results were the same in nearly all of the patients in each group suggests that the findings would be similar in a larger study. Furthermore, we only assessed the location of AG sites by performing endocardial high-frequency stimulation. It is possible that more AG sites may have been identified had epicardial stimulation been performed. However, the two largest previous reports employed only endocardial stimulation,10,12 and in the one report that performed epicardial stimulation,8 AG sites were found in similar locations. Finally, since we did not test a strategy of AG ablation alone and compare it with PV antral isolation, this study cannot make any definitive conclusions about the relative role that AG inputs play in the pathophysiology of human AF versus the role of PV antral triggers.
Heart Rhythm, Vol 4, No 9, September 2007
Conclusions Standard ICE-guided PVAI eliminates vagal responses induced by AG stimulation. Vagal responses during ablation do not correlate with vagal responses induced by AG stimulation. Vagal responses with AG stimulation are not seen in patients presenting for repeat PVAI despite clinical recurrence of AF.
References 1. Armour JA, Murphy DA, Yuan BX, Macdonald S, Hopkins DA. Gross and microscopic anatomy of the human intrinsic cardiac nervous system. Anat Rec 1997;247:289 –298. 2. Patterson E, Po SS, Scherlag BJ, Lazzara R. Triggered firing in pulmonary veins initiated by in vitro autonomic nerve stimulation. Heart Rhythm 2005;2:624 – 631. 3. Liu L, Nattel S. Differing sympathetic and vagal effects on atrial fibrillation in dogs: role of refractoriness heterogeneity. Am J Physiol 1997;273:H805– 816. 4. Po SS, Li Y, Tang D, Liu H, Geng N, Jackman WM, Scherlag B, Lazzara R, Patterson E. Rapid and stable re-entry within the pulmonary vein as a mechanism initiating paroxysmal atrial fibrillation. J Am Coll Cardiol 2005;45:1871– 1877. 5. Coumel P, Attuel P, Lavallee J, Flammang D, Leclercq JF, Slama R. [The atrial arrhythmia syndrome of vagal origin]. Archives des maladies du coeur et des vaisseaux. 1978;71:645– 656. 6. Sharifov OF, Zaitsev AV, Rosenshtraukh LV, Kaliadin AY, Beloshapko GG, Yushmanova AV, Schuessler RB, Boineau JP. Spatial distribution and frequency dependence of arrhythmogenic vagal effects in canine atria. J Cardiovasc Electrophys 2000;11:1029 –1042. 7. Schauerte P, Scherlag BJ, Pitha J, Scherlag MA, Reynolds D, Lazzara R, Jackman WM. Catheter ablation of cardiac autonomic nerves for prevention of vagal atrial fibrillation. Circulation 2000;102:2774 –2780. 8. Scanavacca M, Pisani CF, Hachul D, Lara S, Hardy C, Darrieux F, Trombetta I, Negrao CE, Sosa E. Selective atrial vagal denervation guided by evoked vagal reflex to treat patients with paroxysmal atrial fibrillation. Circulation 2006;114: 876 – 885. 9. Verma A, Marrouche NF, Natale A. Pulmonary vein antrum isolation: intracardiac echocardiography-guided technique. J Cardiovasc Electrophys 2004;15: 1335–1340. 10. Scherlag BJ, Nakagawa H, Jackman WM, Yamanashi WS, Patterson E, Po S, Lazzara R. Electrical stimulation to identify neural elements on the heart: their role in atrial fibrillation. J Interv Card Electrophysiol 2005;13(Suppl 1):37– 42. 11. Pappone C, Santinelli V, Manguso F, Vicedomini G, Gugliotta F, Augello G, Mazzone P, Tortoriello V, Landoni G, Zangrillo A, Lang C, Tomita T, Mesas C, Mastella E, Alfieri O. Pulmonary vein denervation enhances long-term benefit after circumferential ablation for paroxysmal atrial fibrillation. Circulation 2004; 109:327–334. 12. Lemery R, Birnie D, Tang AS, Green M, Gollob M. Feasibility study of endocardial mapping of ganglionated plexuses during catheter ablation of atrial fibrillation. Heart Rhythm 2006;3:387–396. 13. Haissaguerre M, Jais P, Shah DC, Takahashi A, Hocini M, Quiniou G, Garrigue S, Le Mouroux A, Le Metayer P, Clementy J. Spontaneous initiation of atrial fibrillation by ectopic beats originating in the pulmonary veins. N Eng J Med 1998;339:659 – 666. 14. Takahashi Y, Jais P, Hocini M, Sanders P, Rotter M, Rostock T, Hsu LF, Sacher F, Clementy J, Haissaguerre M. Shortening of fibrillatory cycle length in the pulmonary vein during vagal excitation. J Am Coll Cardiol 2006;47:774 –780. 15. Armour JA, Hageman GR, Randall WC. Arrhythmias induced by local cardiac nerve stimulation. Am J Physiol 1972;223:1068 –1075. 16. Cummings JE, Schweikert R, Saliba W, Hao S, Martin DO, Marrouche NF, Burkhardt JD, Kilicaslan F, Verma A, Beheiry S, Belden W, Natale A. Left atrial flutter following pulmonary vein antrum isolation with radiofrequency energy: linear lesions or repeat isolation. J Cardiovasc Electrophysiol 2005;16:293–297. 17. Hsieh MH, Chiou CW, Wen ZC, Wu CH, Tai CT, Tsai CF, Ding YA, Chang MS, Chen SA. Alterations of heart rate variability after radiofrequency catheter ablation of focal atrial fibrillation originating from pulmonary veins. Circulation 1999;100:2237–2243.