BACKGROUND: The mortality of patients suffering from acute decompensated liver disease treated in the intensive care unit (ICU) varies between 50% and 100%. Previously published data suggest that liver-specific score systems are less accurate compared with the ICU-specific scoring systems acute physiology and chronic health evaluation II (APACHE II) and simplified organ failure assessment (SOFA) in predicting outcome. We hypothesized that in a Scandinavian cohort of ICU patients, APACHE II, SOFA, and simplified acute physiology score (SAPS II) were superior to predict outcome compared with the Child-Pugh score. METHODS: A single-centre retrospective cohort analysis was conducted in a university-affiliated ICU. Eighty-seven adult patients with decompensated liver alcoholic cirrhosis were admitted from January 2007 to January 2010. RESULTS: The patients were severely ill with median scores: SAPS II 60, SOFA (day 1) 11, APACHE II 31, and Child-Pugh 12. Receiver operating characteristic curves area under curve was 0.79 for APACHE II, 0.83 for SAPS II, and 0.79 for SOFA (day1) compared with 0.59 for Child-Pugh. In patients only in need of mechanical ventilation, the 90-day mortality was 76%. If respiratory failure was further complicated by shock treated with vasopressor agents, the 90-day mortality increased to 89%. Ninety-day mortality for patients in need of mechanical ventilation, vasoactive medication, and renal replacement therapy because of acute kidney injury was 93%. CONCLUSION: APACHE II, SAPS II, and SOFA were better at predicting mortality than the Child-Pugh score. With three or more organ failures, the ICU mortality was > 90%. APACHE II > 30, SAPS II > 60, and SOFA at day 1 > 12 were all associated with a mortality of > 90%. Referral criteria of patients suffering from decompensated alcoholic liver disease should be revised.