To determine outcomes of implantable cardioverter-defibrillator (ICD) therapy in a uniform population of survivors of sudden cardiac death, we used epicardial defibrillation lead systems to study 300 patients with coronary artery disease (CAD) presenting exclusively with ventricular fibrillation (VF) unassociated with acute myocardial infarction. Operative (30-day) mortality, 2.7% overall, was lower (0.6%) in patients with ejection fractions (EF) > or = 0.30. Over a median follow-up of 1.9 years, cumulative actuarial shock incidence was similar in patients who underwent concomitant coronary artery bypass graft (CABG) surgery (38%) and in those who did not. The 2-year cumulative actuarial incidences of any or appropriate shocks were 65% and 38%, respectively. Sudden death survival at 2 years was 92.5% and 99.3% for patients with EFs < or = 0.30 and > 0.30, respectively. The total mortality rate was similar in shocked and in unshocked patients. Multivariate analysis identified EF and female gender as significant predictors of any and appropriate shock occurrence (all p values < or = 0.05) and EF as a significant predictor of sudden, cardiac, and total mortality (all p values < 0.03). We conclude that in CAD patients presenting exclusively with VF unassociated with acute myocardial infarction and treated with thoracotomy-requiring ICD therapy: (1) operative (30-day) mortality is minimal for patients with an EF > or = 0.30; (2) device use is high and sudden death rates low regardless of concomitant CABG; (3) low EF is a significant predictor of cumulative shock occurrence and mortality (sudden, cardiac, and total); (4) female gender may be a predictor of shock occurrence; and (5) similar mortalities and low sudden-death rates in shocked and nonshocked ICD patients imply that ICD therapy improves survival in shocked patients to a level observed in comparable patients in whom ventricular tachyarrhythmia does not recur.