DC shocks within the coronary sinus have been abandoned because of the risk of cardiac rupture and tamponade. Catheter ablation using DC energy to electrodes straddling the ostium of the coronary sinus, when used clinically, has been reported to result in cardiac tamponade in as many as 16% of patients. A new system of energy delivery maximizes voltage while decreasing the undesirable effects caused by barotrauma. This system includes 1) a low-energy ablation power supply with a brief time-constant capacitive discharge that delivers up to 40 J and 3,000 V and 2) a low-energy ablation catheter with a contoured distal electrode. We performed in vitro and in vivo studies of this new system and compared arcing shocks with nonarcing shocks. Ablations were performed using unipolar distal shocks (D) and unipolar shocks to both electrodes made electrically common (P-D). In vitro studies were done in a large tank filled with physiological saline while recording voltage, current, and pressure. High-speed cinematography (32,000 frames/sec) of shocks of 10-40 J permitted detailed analysis of the vapor globe. Anodal shocks of less than 20 J showed no arcing or only minimal vapor globe formation. For D and P-D anodal shocks of 40 J, the diameters of the vapor globe were 31 and 22 mm, respectively, corresponding to pressure recordings of 11 and 4.9 atm. The pressure rise lasted less than 50 mu sec. In vivo studies involved 18 dogs that received nonarcing shocks (one to six shocks of 15 J) and 18 dogs that received arcing shocks (one to three shocks of 40 J). Each group was divided between D and P-D shocks; catheter ablation was performed at a mean +/- SEM distance of 2.94 +/- 0.92 cm within the coronary sinus. All dogs tolerated the procedure without cardiac rupture or tamponade. When killed 2-4 days later, the dogs had edema and hyperemia or hemorrhage in the area of the coronary sinus. We compared the effects of multiple (three to six) nonarcing shocks with the effects of one to three arcing shocks. Disruption or rupture of the coronary sinus within the epicardial fat space occurred in two of 12 dogs (17%) with multiple nonarcing shocks but in 13 of 18 dogs (72%) with arcing shocks (p less than 0.003). Occlusion of the coronary sinus occurred in two of 12 dogs (17%) with multiple nonarcing shocks and in nine of 18 dogs (50%) with arcing shocks (p less than 0.06).(ABSTRACT TRUNCATED AT 400 WORDS)