BACKGROUND: Knowledge is insufficient of the long-term benefit of an implantable cardioverter defibrillator (ICD) after out of hospital cardiac arrest (OHCA). AIM: To describe the use and factors of importance for outcome in relation to ICD use among survivors of ventricular fibrillation (VF). METHODS: In consecutive patients discharged alive after OHCA in Gothenburg between 1988 and 2008 the long-term prognosis was followed. RESULTS: In all, there were 5443 OHCAs of which 1489 (27%) were hospitalized alive. Of those, 495 (33%) were discharged alive, of which 390 (79%) had shockable rhythm. The use of ICDs increased, but only 58 of 390 (15%) had an ICD. Among patients who received an ICD, the 2-year mortality was 2%, versus 25% of those who did not (p<0.0001). In follow-up (mean 5.5 years; maximum 10 years), the use of an ICD showed a borderline association with mortality (adjusted hazard ratio 0.49; 95% confidence interval, 024-1.01; p=0.052). Patients who had ICD were younger and had better cerebral function compared with patients without. Predictors for mortality were cerebral function at discharge, age, history of heart failure and myocardial infarction and no coronary angiography during hospitalization. CONCLUSION: Among survivors of OHCA caused by VT/VF who had ICD during hospitalization only 2% died during the subsequent 2 years. The use of ICDs was low but increasing. Factors of importance for mortality were cerebral function at the time of discharge, age, history of heart failure and myocardial infarction and no coronary angiography during hospitalization.