Albumin dialysis with the Molecular Adsorbent Recirculating System (MARS) (Gambro, Lund, Sweden), a noncell artificial liver support device, may be beneficial in acute liver failure (ALF). To determine whether MARS improves survival in ALF. Randomized, controlled trial. (ClinicalTrials.gov: NCT00224705). 16 French liver transplantation centers. 102 patients with ALF. Conventional treatment (n = 49) or MARS with conventional treatment (n = 53), stratified according to whether paracetamol caused ALF. 6-month survival and secondary end points, including adverse events. 102 patients (mean age, 40.4 years [SD, 13]) were in the modified intention-to-treat (mITT) population. The per-protocol analysis (49 conventional, 39 MARS) included patients with at least 1 session of MARS of 5 hours or more. Six-month survival was 75.5% (95% CI, 60.8% to 86.2%) with conventional treatment and 84.9% (CI, 71.9% to 92.8%) with MARS (P = 0.28) in the mITT population and 75.5% (CI, 60.8% to 86.2%) with conventional treatment and 82.9% (CI, 65.9% to 91.9%) with MARS (P = 0.50) in the per-protocol population. In patients with paracetamol-related ALF, the 6-month survival rate was 68.4% (CI, 43.5% to 86.4%) with conventional treatment and 85.0% (CI, 61.1% to 96.0%) with MARS (P = 0.46) in the mITT population. Sixty-six of 102 patients had transplantation (41.0% among paracetamol-induced ALF; 79.4% among non-paracetamol-induced ALF) (P < 0.001). Adverse events did not significantly differ between groups. The short delay from randomization to liver transplantation (median, 16.2 hours) precludes definitive efficacy or safety evaluations. This randomized trial of MARS in patients with ALF was unable to provide definitive efficacy or safety conclusions because many patients had transplantation before administration of the intervention. Acute liver failure not caused by paracetamol was associated with greater 6-month patient survival. Assistance Publique-Hôpitaux de Paris.