Background Delirium prevalence is high in critical care settings. We examined the incidence, risk factors, and outcome of delirium in a medical intensive care unit (MICU) with a particular focus on liver diseases. We analyzed this patient population in terms of delirium risk prediction and differentiation between delirium and hepatic encephalopathy. Methods We conducted an observational study and included 164 consecutive patients admitted to an MICU of a university hospital. Patients were assessed for delirium using the Confusion Assessment Method for ICUs and the Richmond Agitation-Sedation Scale (RASS). On admission and at the onset of delirium Sequential Organ Failure Assessment (SOFA) score was determined. A population of patients with liver disease was compared to a population with gastrointestinal diseases. In the population with liver diseases, hepatic encephalopathy was graded according to the West Haven classification. We analyzed the incidence, subtype, predisposing, precipitating, and health-care setting-related factors, treatment, outcome of delirium and the association between delirium and hepatic encephalopathy in patients with liver diseases. Results The incidence of delirium was 32.5% (n = 53). Univariable binary regression analyses adjusted by the Holm-Bonferroni method showed that the development of delirium was significantly determined by 10 risk factors: Alcohol abuse (p = 0.016), severity of disease (Simplified Acute Physiology Score (SAPS) II, p = 0.016), liver diseases (p = 0.030) and sepsis (p = 0.016) compared to the control group (gastrointestinal (GI) diseases and others), increased sodium (p = 0.016), creatinine (p = 0.030), urea (p = 0.032) or bilirubin (p = 0.042), decreased hemoglobin (p = 0.016), and mechanical ventilation (p = 0.016). Of note, we identified liver diseases as a novel and relevant risk factor for delirium. Hepatic encephalopathy was not a risk factor for delirium. Delirium and hepatic encephalopathy are both life-threatening but clearly distinct conditions. The median SOFA score for patients with delirium at delirium onset was significantly higher than the SOFA score of all patients at admission (p = 0.008). Patients with delirium had five times longer ICU stays (p = 0.004) and three times higher in-hospital mortality (p = 0.036). Patients with delirium were five times more likely to be transferred to an intensive medical rehabilitation unit for post-intensive care (p = 0.020). Treatment costs per case were more than five times higher in patients with delirium than in patients without delirium (p = 0.004). Conclusions The 10 risk factors identified in this study should be assessed upon admission to ICU for effective detection, prevention, and treatment of delirium. Liver diseases are a novel risk factor for delirium with a level of significance comparable to sepsis as an established risk factor. Of note, in patients with liver diseases delirium and hepatic encephalopathy should be recognized as distinct entities to initiate appropriate treatment. Therefore, we propose a new algorithm for efficient diagnosis, characterization, and treatment of altered mental status in the ICU. This algorithm integrates the 10 risk factor prediction-model for delirium and prompts grading of the severity of hepatic encephalopathy using the West Haven classification if liver disease is present or newly diagnosed.