The study consisted of 369 patients (age 62 +/- 13 years) who presented to our institution from April 1984 to April 1994 for malignant ventricular arrhythmias presenting as sustained ventricular tachycardia (VT) in 57% of patients, ventricular fibrillation in 25% of patients, and syncope due to VT in 17% of patients. Coronary artery disease was present in 74% of patients, cardiomyopathy in 19% of patients, and no evident heart disease in 7% of patients. Two hundred twenty-one patients were given drug, therapy, 47 patients underwent arrhythmia surgery, and 75 patients had an implantable cardioverter-defibrillator (ICD). During a mean follow-up of 30 months (range 1 to 101), 66 patients (18%) died from a cardiac death of which 26 (39%) were sudden. Cox regression analysis was conducted utilizing a total of 19 variables (clinical and therapeutic) in the entire population and separately in patients with coronary artery disease and cardiomyopathy. The most significant variables (multivariate analysis) of survival from cardiac mortality in the entire population were: congestive heart failure (CHF) class (p = 0.0003), ejection fraction (p = 0.02), and the use of drug therapy (p = 0.03); in patients with coronary artery disease, CHF class (p = 0.0001) and ejection fraction (p = 0.0006); and in patients with cardiomyopathy, CHF class (p = 0.009) and sustained VT on Holter monitoring (p = 0.007). Kaplan-Meier survival rates from cardiac death were: significantly lower (p = 0.005) in patients with CHF class III and IV compared with CHF class I and II (25% vs 58%, p = 0.005) with drug therapy; marginally significant (47% vs 88%, p = 0.06) from 20 to 40 months in patients with an ICD; and nonsignificant in patients who underwent arrhythmia surgery (63% vs 71%). Patients with an ICD had a better expected survival (82%) than patients who had arrhythmia surgery (69%) and drug therapy (65%). Thus, in patients with malignant ventricular arrhythmias, CHF class was the most significant independent predictor of survival from cardiac mortality over all disease substrates, and therapy influenced survival depending on the CHF class. Patients in CHF class III and IV who underwent arrhythmia surgery or had an ICD had a better expected survival than those taking drug therapy, and the negative impact of antiarrhythmic therapy was most prominent in patients with CHF class III and IV.