The effect of surgical ablation of ablation of atrioventricular accessory pathways on the incidence of atrial fibrillation in patients with Wolff-Parkinson-White syndrome was examined and the results of preoperative electrophysiologic testing were studied to determine factors predictive of outcome. Among 50 consecutive surgical cases, 19 patients were identified with a past history of at least one episode of spontaneous atrial fibrillation documented by electrocardiogram before surgery. The mean number of episodes of atrial fibrillation was 1.97/patient/year during a mean symptomatic period of 6.9 years before surgery. These patients were compared with 19 consecutive patients undergoing surgery during the same time period who had a history of only reciprocating tachycardia. Patients with atrial fibrillation had a significantly shorter antegrade accessory pathway effective refractory period (270 +/- 39 vs 330 +/- 107 msec; p less than .05) and much faster ventricular rates during induced atrial fibrillation (shortest RR interval 219 +/- 73 vs 294 +/- 60 msec, p less than .005; average RR interval 324 +/- 109 vs 405 +/- 127 msec, p less than .01). Patients with atrial fibrillation also had longer PA intervals (47 +/- 13 vs 37 +/- 7 msec; p less than .02). At preoperative electrophysiologic testing, 18 patients with atrial fibrillation had atrial fibrillation induced and 14 sustained the arrhythmia for longer than 10 min. In contrast, atrial fibrillation, although induced in 14 of 19 patients with reciprocating tachycardia, was not sustained in any. Thus electrophysiologic testing suggested that both accessory pathway properties and atrial vulnerability may predispose to atrial fibrillation.(ABSTRACT TRUNCATED AT 250 WORDS)