OBJECTIVES: The majority of patients who get an implantable cardioverter defibrillator (ICD) for primary prevention of sudden cardiac death (SCD) never need to receive appropriate ICD therapies. Better risk-stratifying tools are thus needed to improve the cost-effectiveness of device implantation. The purpose of this study was to assess the cost-effectiveness of using delayed enhancement cardiac magnetic resonance (CMR) to refine patient selection for ICD therapy in primary prevention. METHODS: A Markov model was developed to assess long-term survival, quality-adjusted life years (QALYs), and lifetime costs for a cohort of patients with reduced left ventricular ejection fraction without previous ventricular arrhythmias. Two different strategies were compared: A) implant an ICD in all patients (“ICD for all”); and B) perform CMR and implant an ICD only in patients with delayed enhancement exceeding 5% of myocardial mass (“CMR-guided ICD”). Clinical and economic inputs were derived from European representative literature. RESULTS: The "ICD for all" strategy led to a total cost of 68,818€ and 7.0 QALY per patient, implying an incremental cost of 7,372€ and allowing a gain of 0.1 QALY when compared to the strategy “CMR-guided ICD”. Consequently, the cost-utility ratio is 77,943€ per QALY. Probabilistic sensitivity analysis showed that, assuming a willingness to pay of 50,000€ per QALY, the "ICD for all" strategy would be cost-effective in 31 % of the simulations. CONCLUSIONS: Results suggest that implanting ICD in selected patients allows to use this device in those that benefit most, improving the cost-effectiveness of ICD implantation for the primary prevention of sudden cardiac death.