The role of pulmonary veins vs. autonomic ganglia in different experimental substrates of canine atrial fibrillation

Aims Pulmonary vein (PV)-encircling ablation, which is effective in suppressing atrial fibrillation (AF), damages autonomic ganglia near the PV ostia. This study examined the effects of PV isolation (PVI) vs. peri-PV ganglionic plexus ablation (GPA) in two discrete canine AF models: ventricular tachypacing (240 bpm, 2 weeks)-induced congestive heart failure (CHF), and atrial tachypacing (400 bpm, 1 week)-induced atrial tachycardia remodeling (ATR). Methods and results All PVs were isolated with an epicardial radiofrequency clamp in nine CHF and eight ATR dogs. Peri-PV ganglionic plexi (identified by bradycardic responses to high-frequency stimulation) were ablated in six CHF and five ATR dogs with an epicardial radiofrequency-ablation pen. Electrophysiologic measurements, including 240-electrode AF mapping, were obtained and dominant frequencies (DFs) determined. Atrial growth associated protein-43 (GAP43) and neurofilament-M (NF-M) expression were determined immunohistologically. In CHF, neither PVI nor GPA affected AF duration, DF or the already low AF vulnerability. In ATR, PVI reduced AF vulnerability (75 +/- 6% to 55 +/- 11%, P < 0.05) but did not alter AF duration or DF. In contrast, GPA prolonged atrial refractory period and decreased AF vulnerability (75 +/- 8 to 30 +/- 10%, P, 0.05), AF duration (617 +/- 246 to 39 +/- 23 s, ** P < 0.01), and DF (11.4 +/- 0.6 to 8.6 +/- 0.3** Hz, left atrium) in ATR dogs. Both GAP-43 and NF-M expression were decreased in CHF (by 63.1** and 60.0%**) and increased in ATR (by 65.5** and 92.1%, P < 0.001) compared with control. Conclusions PVs play a minor role in experimental AF due to CHF or ATR, but autonomic ganglia are important in AF related to ATR. Differential neural remodelling may contribute to varying effects of GPA in discrete AF substrates.