The actuarial survival of 60 consecutive recipients of the implanted cardioverter defibrillator (ICD) were compared with 120 matched concurrent medically treated patients using a case-control design. All ICD patients and controls presented with either sustained ventricular tachycardia or ventricular fibrillation. Controls were matched to ICD recipients according to 5 variables: age, left ventricular ejection fraction, arrhythmia at presentation, underlying heart disease and drug therapy status. Mean ages were 58 and 59 years in ICD patients and controls, and the average ejection fractions were 36 and 35%. Coronary artery disease was present in 75 and 79% of ICD patients and controls, respectively. During follow-up, sudden deaths were fewer in ICD recipients than in controls (5 vs 10%, p less than 0.01). At 1 and 3 years, actuarial survival was 0.89 vs 0.72 and 0.65 vs 0.49 for ICD recipients and controls. The 5-year actuarial survival curves were significantly different by the Cox proportional hazards model (p less than 0.05). It is concluded that in this retrospective case-control study, the use of the ICD in the management of patients at risk for sudden death results in improved probability of survival.