Clinical Predictors of Intensive Care Unit Transfer in Admitted Patients with Cirrhosis

• Published in: Digestive diseases and sciences Vol. 68; no. 6; pp. 2344 - 2359
• Main Authors: Dukewich, Matthew, Liu, Chung-Heng, Weinberg, Ethan M, Mahmud, Nadim, Reddy, K. Rajender
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.06.2023; Springer Nature B.V
• Subjects: Biochemistry; Gastroenterology; Hepatology; Hospital Mortality; Humans; Hyponatremia; Hyponatremia - complications; Intensive Care Units; Liver cirrhosis; Liver Cirrhosis - complications; Liver Cirrhosis - diagnosis; Liver Cirrhosis - therapy; Liver diseases; Medicine; Medicine & Public Health; Mortality; Oncology; Original Article; Prognosis; Respiratory failure; Retrospective Studies; Transplant Surgery; Decompensated liver disease; Critical illness; Intensive care unit care; End-stage liver disease
• ISSN: 0163-2116; 1573-2568
• DOI: 10.1007/s10620-023-07856-x

Background Patients with cirrhosis are at high risk of mortality after organ failure that requires ICU care. There have been attempts to predict which patients are at highest risk, with some success found in adapting liver disease-specific scoring systems with clinical variables commonly associated with critical illness. However, the clinical factors predictive of which patients with cirrhosis are most at-risk of needing ICU level care are unknown. Aims Our study set out to better understand which clinical variables were associated with need for ICU care in patients with cirrhosis. Methods Retrospective analysis of admitted patients with cirrhosis at single tertiary care center. Results Patients with cirrhosis admitted to our center were categorized into three groups: those without ICU transfer, those admitted to the ICU directly from the emergency department (ED), and those admitted to the ICU from the medicine floor. These groups differed in mortality at 30 days (3.5% vs. 15% vs. 25%, P  < 0.001) and at subsequent intervals up to 1 year. These groups differed in indication for ICU transfer, with GI bleed, hemorrhagic shock, hepatic encephalopathy, and hyponatremia occurring more in the ED-to-ICU group, while respiratory failure was more common in the floor-to-ICU group. In multivariable analysis, factors associated with ICU transfer included worsened kidney function, anemia, hyponatremia, leukocytosis, and the decision to obtain a lactate level. Similar analysis with only floor-to-ICU patients found that ICU transfer was associated with hypoalbuminemia, hyponatremia, hypotension, and SIRS score. Conclusion Our study found significant differences in mortality among three distinct groups of patients with cirrhosis. A risk factor model for ICU transfer found that variables both specific and nonspecific to liver disease were associated with ICU transfer, with between-group differences supporting the idea of different clinical phenotypes and suggesting factors that should be considered in early triage and assessment of hospitalized patients with cirrhosis. Graphical Abstract