In a series of 33 patients with accessory pathways, 26 had successful catheter ablation (fulguration [23 patients] or modification [3 patients]) of their accessory pathway conduction, and could be considered as a clinical success. One hundred thirteen single discharge or double discharge shocks were delivered, and each shock was studied to reveal which parameters were important to predict the success or failure of catheter ablation. Double discharge shocks resulted in successful accessory pathway modification or ablation twice as often as single discharge shocks (32% vs 16%). This effect was more pronounced in left lateral accessory pathways (48% vs 4%). Shocks in the electrophysiologically defined ventricular zone were more likely to be successful (33%) than shocks delivered in the atrial zone (14%), irrespective of accessory pathway location. The presence of a probable Kent potential was the parameter most strongly associated with success. The parameter most strongly associated with failure, with a 100% negative predictive value, was the absence of earliest activation recorded on the ablating catheter prior to shock delivery. An AV interval of < 60 msec significantly divided the successful from the unsuccessful shocks (P = 0.01). The VA interval during orthodromic reciprocating tachycardia or right ventricular stimulation did not allow for significant division into successful and unsuccessful attempts in this relatively short series. VA intervals, when longer, were predictive of failure but, when shorter, had low positive predictive value. Mean follow-up in 25 successful patients was 15 +/- 6 months. All patients did well in the follow-up period. Neither those patients with ablation nor modification of the accessory pathway had recurrent episodes of tachycardia or required pharmacological treatment for control of arrhythmias.