Sudden death is one of the two main causes of mortality in congestive heart failure. Implanted cardioverter defibrillator (ICD) is an expensive but highly effective treatment in preventing sudden death. The gain of primary prophylactic ICD in preventing sudden death in heart failure has not been clearly established. OBJECTIVE: Compare the cost-effectiveness of prophylactic ICD with conventional treatment for preventing sudden death. METHODS: A lifetime decision model was built. The perspective is societal. The target population is U.S. HF patients, aged 60, with NYHA functional Class II and III. Estimates of cost, utility and probabilities are taken from literature, clinical experts, CMS fee schedule payment, and the Bureau of Labor Statistics. In all cases, we assume that ICD is effective in preventing all sudden death, and the ICD would be reimplanted at nineth year. In our base case, we assume that total annual mortality rate is 20%, of which sudden death accounts for 40%; the utility of ICD is 10% less during the 1st year after implantation, and reverts back to pre-implantation level in the 2nd year. We did a one-way sensitivity analyses on all model parameters. RESULTS: the lifetime cost is $117,095 for patients with prophylactic ICD and $24,709 for patients with conventional treatment in 2002; the QALYs gained were 2.9088 and 1.9045 respectively. The CE ratio was $91,990 per QALY saved. We failed to show that ICD is cost-effectiveness under any plausible scenario if we use $50,000 per quality-adjusted-life-year saved as the cut-off point. CE ratio is sensitive to the utility at the second and subsequent years after ICD implantation, and the proportion of sudden death in all HF-related deaths. CONCLUSIONS: Using conventional cost effectiveness benchmarks of $50,000 per life year saved, it is unlikely that ICD would be cost-effectiveness in preventing sudden death compared to conventional treatment for heart failure patients. Future research should focus on patient utility with and without ICD.