The long-term performance of epimyocardial pacing leads in children is well established, but few studies have analyzed the performance in adults. This issue has clinical relevance in view of the increased use of epimyocardial leads with implantable cardioverter defibrillator and antitachycardia pacing systems. We analyzed 93 epimyocardial pacing "systems" (121 leads: 65 unipolar, 28 bipolar) in adult patients (age 57 +/- 16 years), implanted since January 1980. Two different models were studied: Medtronic 4951 "Stab-on" (n = 35) and Medtronic 6917/6917A "Screw-in" (n = 58). A control group was created by randomly matching each epimyocardial system with two endocardial leads, according to age and year of implant. Epimyocardial and endocardial leads were followed-up for 44 +/- 35 and 43 +/- 35 months, respectively (P = NS). Freedom from failure for epimyocardial leads was 0.91 (95% Confidence Interval [95% CI] = 0.82 to 0.96) at 5 years, and 0.91 (95% CI = 0.69 to 0.98) at 10 years. No difference was found between the two analyzed models. Freedom from failure for endocardial leads was 0.97 (95% CI = 0.93 to 0.99) and 0.90 (95% CI = 0.61 to 0.97) at 5 and 10 years, respectively. Epimyocardial leads had a significantly poorer short-term survival than endocardial leads, secondarily to earlier "technique related" failures (P = 0.03; relative risk 3.0; Wilcoxon test). However, overall long-term performance was similar to endocardial leads. Epimyocardial pacing leads, meticulously implanted and tested, have a long-term performance similar to endocardial pacing leads.