Introduction: The main result of economic evaluation studies of healthcare technologies is the incremental cost-outcome-ratio, a measure expressing the additional costs in relation to the additional outcome, e.g. the incremental costs per life-year gained or the costs per quality-adjusted life-year (QALY) gained. In this paper, the economic evaluation of a high-priced innovative medical device, the artificial liver support system MARS (Molecular Adsorbent Recirculating System) for patients with an acute-on-chronic liver failure (ACLF) and a history of alcohol induced liver disease is outlined. Methods: In a controlled clinical cohort study all patients with ACLF admitted to a large university hospital between 1999 and 2002 were included and prospectively followed up over 3 years. Survival, healthrelated quality of life and all relevant direct medical resource uses were measured. All costs were calculated from a healthcare system’s perspective and standardized to EUR 2002. Results: The direct medical costs of the initial inpatient treatment were 31,539 EUR in the MARS group compared to 7,543 EUR in controls. Important predictive factors were MARS (mean cost per patient treated 14,631) and renal dialysis. Mean direct costs during the follow-up were 8,493 EUR for patients treated with MARS and 5,194 EUR in controls. The patients treated with MARS had a higher mean survival time and showed a significantly higher survival probability using a cox-regression model. The healthrelated quality of life of the surviving patients was nearly constant (EQ-5D 70%) over the 3-year follow-up. The incremental costs per life-year gained over a time horizon of 3 years were 31,448 EUR, the incremental costs per QALY 47,171 EUR. Conclusions: The study results overall show an acceptable cost-effectiveness of MARS from a healthcare system’s perspective, and, keeping the limits of the present scientific evidence in mind, the technology seems to be promising from a medical as well as an economic point of view. The costs for the acute inpatient hospital treatment of ACLFpatients with MARS are four times higher compared to standard treatment. As long as there is no regular complete reimbursement for MARS, hospitals have almost no incentive to use it and the implementation of the technology might be delayed.