Introduction The intrinsic cardiac autonomic nervous system (ANS) is implicated in atrial fibrillation (AF), but the mechanism of ANS activation contributing to the initiation and maintenance of AF in humans remains unclear. In animal studies, stimulation of fat pads containing ganglionated plexi (GP) induces heart rate slowing, spontaneous premature depolarisations and facilitates AF induction. Acetylcholine applied to pulmonary vein (PV) preparations shortens refractory periods, promoting rapid stable re-entry within PV. We hypothesise that activation of the ANS by endocardial GP high frequency stimulation (HFS) causes ectopic atrial or PV activity, which may initiate AF and shortens local atrial fibrillation cycle lengths (AFCL) and promotes dispersion of AFCL within the PV and atria in humans.
Methods Patients undergoing ablation for AF were recruited. Two protocols were used. The first was used in patients in sinus rhythm, in whom short bursts of HFS were delivered from the mapping catheter during the local atrial refractory period at GP sites adjacent to each PV during fixed rate atrial pacing (synchronised HFS): episodes of ectopy were identified from catheters placed in the PV, coronary sinus (CS) and high right atrium (HRA). A negative ectopic response was defined as lack of ectopy after 15 bursts of HFS. The second protocol was used in patients in AF: HFS was applied continuously to GP sites to obtain a vagal response (>100% prolongation in the RR interval), and spatiotemporal changes in AFCL (averaged over sequential 10 cycles) were calculated from PV, CS, HRA and the site of HFS.
Results A total of 23 patients was studied (table). In 12 patients, synchronised HFS caused ectopy in 91 (81%) of 112 episodes, inducing AF greater than 30 s in 16 (14%) cases (fig 1). In 48% of 91 cases, earliest ectopic activity was recorded in PV adjacent to the site of HFS. During AF in 15 patients, GP sites could be identified from a vagal response in 14/15 left upper PV, 8/12 right upper PV, 9/9 right lower PV and 1/1 left lower PV GP sites. A vagal response was observed in 128 of 193 continuous HFS episodes, resulting in a sequential gradient of AFCL shortening from the PV (21% reduction, 167 ± 28 to 132 ± 24 ms, p<0.0001) to the PV–atrial junction (9% reduction, 173 ± 21 to 157 ± 19 ms, p<0.0001) to the rest of the atrium (fig 2). Without a vagal response, no dispersion of AFCL shortening was observed. With a vagal response, the reduction in AFCL was greater and lasted longer in the PV adjacent to the site of HFS (21% reduction, lasting ∼4.5 s) compared with the PV–atrial junction (actual site of HFS) (9% reduction, lasting ∼1.5 s).
Conclusion Activation of the intrinsic ANS results in ectopic activity, which may initiate AF, and promotes the dispersion of AFCL across the PV and atria with the maximal shortening of AFCL within the PV. The ANS may play an important role in the initiation of triggers within the PV or atria, and may contribute to maintenance of drivers within the PV, which sustains AF.