Many of the current approaches to atrial fibrillation (AF) ablation involve ablation of regions where autonomic ganglia have been found. These approaches include the pulmonary vein (PV) isolation approach, trigger elimination, and the mapping and ablation of complex fractionated atrial electrograms, all of which involve modification of the left atrial tissue substrate. The purpose of this review is to examine the research that supports the role of the ganglia in AF initiation and then to look at catheter-based and surgical approaches to AF ablation that target the autonomic sites.

Historical Reports and Findings in the Animal Lab
       The first suggestion that paroxysmal atrial fibrillation (PAF) could be induced from parasympathetic or vagal ganglia came from Coumel in 1989.¹ In this work, two forms of PAF were proposed: one being an adrenergic form that occurred with exercise or stress, and the second form that was more likely to emerge during sleep, a vagal form.

Figure 1.

Left side pulmonary veins and location of ganglionic plexi (GP) (solid circles). Notice location of the Ligament of Marshall (LOM). (2007 Atricure, Inc. Adapted from Oklahoma Health Sciences)

       Subsequently, the autonomic nerve distribution of the heart was mapped, and these nerves were found to largely exist around the great vessels.² In a 1997 canine study, Chiou et al found the nerves converged into fat pads found around the SVC-aortic junction. Ablation of these regions caused denervation in the atria but not in the ventricles.³ Chen et al⁴ reported that postganglionic parasympathetic neurons were primarily found in three fat pads: RPV-A junction, which also connects to the SA node; the junction of the IVC and LA, which connects to the AV node; and the SVC and aortic root.
       The search for a canine model of PAF induction includes the work of Scherlag et al from Oklahoma.⁵ In this study, continuous delivery of a high-frequency stimulus to the left pulmonary artery caused marked slowing of the atrial rate. The authors speculated that parasympathetic stimulation of the pre-ganglionic fibers from the left vagosympathetic tract was a mechanism for induction of the AF form, which initiates with bradycardia. In another canine study from Oklahoma that looked at the initiation of PAF from PVs, parasympathomimetic agents were injected into a fat pad at the base of the PV. The results resembled PAF that is seen clinically. The authors concluded that hyperactive autonomic ganglia may be a critical element in AF induction from the PVs.⁶
       A number of studies have examined the characteristics of atrial ganglia. Chevalier et al reported that nerves were more abundant in the proximal PV than distal PV and more abundant in epicardium than in the endocardium. However, adrenergic and cholinergic were not distinguished from each other.⁷ Tan et al (2007) analyzed nerve densities of the PV-atrial junction and found that adrenergic and cholinergic densities were highest in the LA within 5 mm from the PV-LA junction; densities were also higher in the superior aspect of the left superior PV, anterosuperior aspect of the right superior PV, and inferior aspects of both inferior PVs. Densities were also found to be higher in epicardial than endocardial portions of the muscle tissue. Sympathetic and parasympathetic nerves were found to exist in basically the same locations.⁸
       Authors of a surgical study reported on the histology of atrial tissue retrieved during procedures from patients with persistent AF versus those with sinus rhythm. Levels of atrial sympathetic innervation were found to be greater in the persistent AF group.⁹
       Yuan et al reported that the high-frequency stimulation of a right or left ganglionated plexus would induce fractionated atrial potentials at contralateral pulmonary vein/ganglionated plexi areas.¹⁰ This finding suggested the existence of an interconnection within the atrial autonomic system. In a canine study from Hou et al, functional interactions between ganglionated plexi in the intrinsic cardiac autonomic system were found through both ablation and stimulation evidence.¹¹ The authors believed that all areas were interconnected and that one or more plexi serve as an ‘integration center.’
       The effects of denervation on the heart in the long term are being studied by many investigators. Pappone (2004) measured heart rate variability in patients up to a year post denervation and found no reversal of changes.¹² A 2006 canine study of the long-term effects of fat pad ablation on the atria and AF included ablation of the right pulmonary vein fat pad and the IVC-LA fat pad.¹³ Denervation effects reversed in four weeks. Reports from the heart failure/transplant literature cite the occurrence of a reinnervation phenomenon in the denervated hearts of transplant patients.¹⁴ One must ask if this same result could occur for denervated ablation patients.
       This cumulative knowledge about the location of ganglia has led to a number of graphical representations. Examples are seen in Figures 1 and 2. The drawing of the left side (Figure 1) identifies the ganglia as solid circles. The Ligament of Marshall (LOM) can also be seen on the left. Figure 2 shows the location of the ganglia on the right side.

Figure 2.

Right side pulmonary veins and location of ganglionic plexi (GP) (solid circles). (2007 Atricure, Inc. Adapted from Oklahoma Health Sciences)

Catheter Ablation Strategies
       Clinical catheter ablation strategies have taken three major directions. The first has been the ablation of triggers for atrial fibrillation initiation. This approach to a large extent is represented by the mapping of individual PVs and ablating individual spots.¹⁵ The second approach has involved placing a circumferential lesion around the PVs, causing electrical isolation of the veins.¹⁵ The isolation approach has, for many centers, included multiple lesions within the left and sometimes right atrium. In Milan, Dr. Carlo Pappone¹² discovered during post-procedure follow-up on patients who had received extensive circumferential and isolation lesions that those with a documented autonomic denervation had less AF than those without. In addition, he identified a subgroup of patients in whom there was documented hypotension and bradycardia during RF to the posterior wall of LA near PV-LA junction. For these patients, there was a 99% success rate for AF elimination. According to Pappone, circumferential PV ablation is the most effective approach as it addresses all of the current AF initiation theories by incorporating trigger isolation, substrate modification, and ablation of the local vagal innervation.¹⁶
       In 2004, Nademanee et al first reported a third treatment approach in which mapping and ablation of complex fractionated atrial electrograms were performed.¹⁷ These fractionated electrograms were found in consistent locations, and it was assumed that these were areas of slow conduction and shortened refractoriness where AF wavelets could reenter rapidly. Lemery et al reported on 14 patients in whom high-frequency stimulation and mapping of the ganglionated plexi were performed. This was followed by complex fractionated atrial electrogram-guided ablation.¹⁸ The majority of lesions were placed in antral areas of PVs circumferentially, over and adjacent to ganglionated plexi regions. The authors found that most plexi seemed to be located in the path of routine AF ablation lesions.
       Scherlag et al from Oklahoma reported on a group of 60 AF patients who were paroxysmal or persistent.¹⁹ Twenty-seven patients received PV antrum ablation alone; 33 underwent antrum ablation plus localization and ablation of ganglionated plexi (GP). The ten-month follow-up revealed that the addition of the GP ablation increased success rate from 70% to 91%. Another report from Oklahoma found that sites of complex fractionated atrial electrograms corresponded to GP sites.²⁰
       The LOM has also been a target both in catheter ablation approaches and in surgical approaches to AF ablation. The LOM is an epicardial vestigial fold that is found in the location of what was the embryonic left SVC. The area contains nerves, a vein, and muscle tracts. The LOM contains sympathetic nerve trunks and ganglia in addition to being parasympathetically innervated. In 2006, Kurotabi et al reported on a series of 100 patients undergoing ablation for AF.²¹ The anatomy of the vein of Marshall was studied in all. There was a significant increase in the incidence of PV foci in LSPV in those with a well-developed vein of Marshall. The authors reported that the ends of the vein of Marshall branches were good markers for seeking out additional PV foci.

Minimally-Invasive Surgical Approach
       A variety of minimally-invasive surgical approaches have been reported for the treatment of AF. This review includes just a few.
       Garrido et al (2004) reported on the use of microwave (AFx, Fremont, California) and a surgical robot (da Vinci surgical system, Intuitive Surgical, Inc., Sunnyvale, California) in a minimally-invasive AF procedure.²² This procedure included the amputation or suturing excision of the left atrial appendage. Ad et al reported on a series of 72 patients who underwent a minimally-invasive AF procedure with cryosurgery.²³ These procedures were conducted with an arrested heart (cardioplegia). The lesion set included a circumferential line around the left atrial appendage and excision, a right atrial appendage excision, circumferential lesions around the pulmonary veins, and multiple endocardial lesions.
       Dr. Randall Wolf of Cincinnati has pioneered in the beating heart a minimally-invasive surgical approach with the use of a bipolar clamping mechanism that is placed across the PV opening.^24-25 This ablation clamp (AtriCure, Inc., West Chester, Ohio) applies pressure along the intended ablation line and creates a contiguous transmural ablation lesion. The Wolf Minimaze is an epicardial bipolar RF ablation that includes isolation of the pulmonary veins, ablation of ganglionated plexi, removal of the Ligament of Marshall, and removal of the left atrial appendage. The procedure requires two separate surgical positionings and two preps as the patient is turned first to one side and then the other, in order to create working ports for insertion of the fiberoptic camera, ablation clamp and other surgical tools. The approach is to first test the GPs with high-frequency stimulation to see if a vagal response can be evoked. If so, these GPs are ablated and then retested for successful denervation. After this testing and ablation is completed, the ablation clamp is placed around the cuff area at the base of the PVs and a transmural lesion is created.
       The Wolf Minimaze has evolved significantly in the hands of Dr. James Longoria at Sutter General in Sacramento. Dr. Longoria is performing this procedure with the patient in the supine position, requiring one prep. After completion of the right side, he proceeds immediately to the left side to begin while his assistant closes on the right. This has increased the efficiency and decreased the time required for the OR usage by at least an hour. Dr. Longoria has also been able to reduce the number of atriotomies required, reducing the patient’s pain post-operatively. Dr. Longoria is using this procedure for both paroxysmal and chronic patients with failed catheter ablation. For the chronic patients, he adds an intercaval line and a connecting roof line. These patients go on to EP follow-up study where the electrophysiologist will study fractionated potentials and ablate any remnants, making this a combined EP and surgical procedure with great curative potential.

Conclusion
       A cautionary note comes from Alessie and Schotten, who point out that multiple factors contribute to AF, such as acute atrial stretch, oxidative stress, atrial ischemia, inflammatory response, electrical remodeling, pathologic changes in atrial architecture and disturbances in atrial conduction.²⁶ They remind us that all of these factors should be considered in treatment, not just the nerves.
       Ablation of the ganglionated plexi for treatment of atrial fibrillation is certainly a new and exciting therapeutic direction. It seems there is much to learn about anatomy, therapeutic approach, and outcomes. Addressing the issue of approach, Kwaku et al stated “…it remains to be demonstrated whether such systematic targeting of GPs is more readily achieved with an epicardial or endocardial approach.”²⁷ The data is being collected, and we all look forward to the next chapter.