Background-The vein of Marshall (VOM) is an attractive target during ablation of atrial fibrillation because of its autonomic innervation, its location anterior to the left pulmonary veins, and its drainage in the coronary sinus. Methods and Results-We studied 17 dogs. A coronary sinus venogram showed a VOM in 13, which was successfully cannulated with an angioplasty wire and balloon. In 5 dogs, electroanatomical maps of the left atrium were performed at baseline and after ethanol infusion in the VOM, which demonstrated a new crescent-shaped scar, extending from the annular left atrium toward the posterior wall and left pulmonary veins. In 4 other dogs, effective refractory periods (ERP) were measured at 3 sites in the left atrium, before and after high-frequency bilateral vagal stimulation.

The vein of Marshall (VOM) is an attractive target during ablation of atrial fibrillation due to its autonomic innervation and its location anterior to the left pulmonary veins and drainage in the coronary sinus. We studied 14 dogs. A coronary sinus venogram showed a VOM in 10, which was successfully cannulated with an angioplasty wire and a 2 mm balloon. In 5 dogs, electroanatomical (Carto) maps of the left atrium were performed at baseline and after ethanol (100%, 4 – 8 cc) was infused in the VOM, which demonstrated the creation of a new crescent-shaped scar in the left atrium, extending from the annular left atrium towards the posterior wall and left pulmonary veins. In 4 dogs, both cervical vagal trunks were isolated in the carotid sheath and cuff stimulation electrodes were attached to them. Effective refractory periods (ERP) were measured in 3 sites of the left atrium, before and after high-frequency bilateral vagal stimulation. The baseline ERP was 113.6±35.0 ms, and decreased to 82.2±25.4 ms (p<0.05) after vagal stimulation. After alcohol infusion in VOM, vagally-mediated ERP decrease was eliminated (from 108±27.2 ms to 95.6 ±16.7ms, p=NS). This elimination of vagal effects was not uniform and was limited in sites in proximity with the VOM (baseline ERP 105±18.7ms vs post vagal 98.±37.6ms, p=NS, as opposed to 106.7±27.1ms vs post vagal 73.3±19.7ms, p<0.05, in sites remote to VOM). We also tested feasibility of VOM alcohol infusion in humans: 2 patients undergoing pulmonary vein antral isolation had successful VOM cannulation: left atrial voltage maps demonstrated new scar involving the infero-posterior left atrial wall extending towards the left pulmonary veins. Retrograde alcohol infusion in the VOM achieves significant left atrial tissue ablation, abolishes local vagal responses and is feasible in humans.