The ganglionated plexus: The upstream triggers of atrial fibrillation

Abstract

The ganglionated plexuses (GP) are dense epicardial nerves that are implicated in atrial fibrillation (AF). They can be functionally located from the endocardium using high frequency stimulation (HFS) which can locate distinct GP that trigger atrial ectopy/AF (ET-GP) or atrioventricular (AV) dissociating (AVD-GP). Our aim was to map and understand the histological, anatomical and functional properties of the different types of GP and ablate them with or without pulmonary vein isolation (PVI) in patients with AF. We hypothesised that ablating these specific GP sites is feasible, and prevents AF. Firstly, to investigate this, we mapped for AVD-GP and ET-GP using HFS in the left atrium of patients with AF. An automated process was used to merge and transform all patient maps onto one reference left atrial shell. A probability density function was applied at each tested site, including GP and negative HFS response sites, to create a probability distribution atlas of AVD-GP and ET-GP. There were distinct anatomical regions according to each GP sub-type, and ET-GP had preponderance to the PV ostia, roof, and mid-anterior wall. These are the areas that would usually be targeted with circumferential PVI. Therefore, a prospective, randomised, controlled study was performed (GANGLIA-AF) which assessed ET-GP ablation without PVI and PVI alone in patients with paroxysmal AF. Patients were followed-up for 12 months with multiple 48hr Holter monitors. The primary endpoint was any documented atrial arrhythmia >30s, and the secondary endpoints included complications and redo ablations. This showed that there was no statistically significant difference in AF prevention between the two arms, however the GP ablation arm required less ablation on average than the PVI arm. We also performed a smaller pilot study of redo AF ablation patients, assessing for feasibility and safety of GP ablation in addition to redo PVI. The same follow-up and endpoint criteria were used as in the GANGLIA-AF study. Some patients had permanent PVI, and non-PV triggers of AF were identifiable with HFS. We also developed a custom-built high frequency stimulator (Tau-20) that was used to identify ectopy-triggering (ET) sites in Langendorff-perfused porcine hearts. We were able to replicate the HFS responses used in the clinical setting in the porcine atria. Transmural cross-sectional dissections were taken from ET and non-ET sites, and the tissues were stained for parasympathetic and sympathetic nerves using immunohistochemistry methods. This showed that the mean density of nerves was greater in ET sites compare to non-ET sites. The Tau-20 has been successfully trialled in humans in the clinical setting, and with further improvements, it may replace the old Grass S88 stimulator for future GP ablation cases. In conclusion, ET-GP are upstream triggers of AF that can be ablated without PVI to prevent paroxysmal AF. GP ablation can be achieved with less RF energy than PVI implying a more specific technique on mechanistic grounds. The cross-over rate and clinical outcomes for GP ablation needs further improvement but GANGLIA-AF provides evidence that GP ablation may be an alternative or an adjunct technique to PVI. In addition, our ex-vivo evidence of increased nerve density at ET sites may account for the differential functional response of ET-GP stimulation in the clinical setting.