This study was designed to identify characteristics that might be predictors of failure of surgical treatment alone (endocardial resection) for sustained ventricular tachycardia. Thirty-three consecutive patients with sustained ventricular tachycardia were studied by standard techniques preoperatively, intraoperatively, and 7 to 36 days postoperatively. Standard endocardial resection was guided by intraoperative mapping in all patients. Adjuvant cryoablation was used in areas that were not accessible to excision. Patients were divided into two groups on the basis of the results of the postoperative electrophysiologic study. Group I (14) were patients who still had ventricular tachycardia (failure) and Group II (19) were those who did not have ventricular tachycardia (success). On the basis of the postoperative electrophysiologic testing, the time from myocardial infarction to surgical treatment (less than 3 months) was a powerful predictor of failure of operation alone to prevent ventricular tachycardia (p less than 0.01). This may indicate a different mechanism of ventricular tachycardia in this group of patients. Another possible predictor of surgical failure was three-vessel disease. The site of origin of ventricular tachycardia, the use of cryoablation, the number of morphologies, and the amount of tissue resected were not significant predictors of success or failure. The result of the postoperative electrophysiologic study was also a strong prognostic predictor of subsequent arrhythmias.