BACKGROUND Two catheter electrode systems were compared for delivering radiofrequency current for ablation of the atrioventricular junction. Seventeen patients with drug-resistant supraventricular tachyarrhythmias were studied. RESULTS A 6F or 7F catheter with six or eight standard electrodes (1.25 mm wide, 2.5-mm spacing) was used in the first seven patients (group 1). A 7F quadripolar catheter with a large-tip electrode (4 mm long; surface area, 27 mm2) was used in the final 10 patients (group 2). Both ablation catheters were positioned to record a large atrial potential and a small but sharp His bundle potential from the distal bipolar electrode pair. Radiofrequency current was applied between a large skin electrode on the left posterior chest and either 1) each individual electrode on the standard-tip electrode catheter at 40 V (group 1) or 2) the large-tip electrode at 50-60 V (group 2). Radiofrequency current was limited to 40 V in group patients because of the strong potential for an early impedance rise when higher voltage is applied through standard electrodes. Complete atrioventricular block was achieved in six of seven group 1 patients and all 10 group 2 patients. A junctional escape rhythm followed ablation in five or six group 1 patients (mean cycle length, 1,066 +/- 162 msec) and eight of 10 group 2 patients (mean cycle length, 1,281 +/- 231 msec). Atrioventricular block was produced in a mean of 4.7 +/- 4.6 radiofrequency current applications delivered over a period of 42 +/- 45 minutes using the large-tip electrode (group 2) compared with 46 +/- 22 applications using standard electrodes (15.9 +/- 10.2 applications delivered through the standard-tip electrode) over a period of 147 +/- 59 minutes (group 1). For the application producing atrioventricular block, the large-tip electrode used higher voltage (58 +/- 17 versus 38 +/- 5 V, p less than 0.03) and had lower impedance (103 +/- 22 versus 148 +/- 40 omega, p less than 0.01), resulting in greater power (33.0 +/- 13.0 versus 10.2 +/- 0.6 W, p less than 0.003) and shorter time to block (8 +/- 3 versus 22 +/- 3 seconds, p less than 0.001). Current delivery through standard electrodes was limited by an impedance rise occurring 7 +/- 7 seconds after the onset of one or more radiofrequency current applications at 10 +/- 1 W in six of seven patients. Using the large-tip electrode, an impedance rise occurred in five of 10 patients, but at 25 +/- 10 W and after 21 +/- 9 seconds. Atrioventricular block occurred before the impedance rise in three of these five patients. Complete atrioventricular block persisted in 15 of 16 patients at a mean follow-up of 8.7 months. Atrioventricular conduction returned at 1 month in one group 2 patient and was successfully ablated by a second procedure. Three group 1 patients died 0.5-2 months after ablation, and a fourth patient underwent cardiac transplantation after 10 months. Pathological examination of the heart in two of these patients showed necrosis of the atrioventricular node and origin of the His bundle, without injury to the middle or distal His bundle. All 10 group 2 patients are alive and subjectively improved after ablation. CONCLUSIONS We conclude that catheter-delivered radiofrequency current effectively produces complete atrioventricular block (94%) without requiring general anesthesia or the risk of ventricular dysfunction or cardiac perforation. The large-tip electrode allows a threefold increase in delivered power and markedly decreases the number of pulses and time required to produce atrioventricular block.