Fifty-one consecutive patients underwent modified catheter-mediated direct-current ablation of accessory pathways. Energy was delivered through the distal pair of electrodes (dual electrode configuration) of a 6F quadripolar catheter to the internal surface of the right or left atrioventricular (AV) annulus. In an attempt to prevent the later resumption of accessory pathway conduction, one additional shock was given after the initial successful interruption of accessory pathways. A mean of 2.7 shocks with cumulative energy of 453 +/- 32 Joules/patient interrupted the accessory pathways in 47 patients and modified the accessory pathway conduction in 2 patients. Forty-eight patients were asymptomatic and free of any antiarrhythmic agents with a follow-up ranging from 3-20 months (mean 12 +/- 1 months), without early or late serious complications (AV block or tamponade). Conduction characteristics, concealed or manifest, and recording of accessory pathway activity did not affect the outcome. Mean cumulative energy and number of applications of energy to achieve a successful outcome were lower in patients with concealed (376 +/- 31 Joules, 2.4 +/- 0.2 shocks) than manifest accessory pathways (516 +/- 50 Joules, 2.9 +/- 0.2 shocks). At the successful ablation sites, the mean shortest retrograde ventriculoatrial interval during orthodromic reentrant tachycardia (VA') was 80 +/- 3 msec (78% had VA' less than 90 msec) and was not different between concealed and manifest accessory pathways; the mean shortest antegrade AV interval was 47 +/- 3 msec in manifest preexcitation; the mean ratio of atrial to ventricular wave amplitude was not significantly different between left-sided (0.8 +/- 0.1) and right-sided (1.1 +/- 0.2) accessory pathways (p greater than 0.05). A successful outcome was achieved in 94% of 51 patients. This procedure is relatively safe and effective, regardless of the location of the accessory pathway.