Long-term outcomes after ICU admission in critically ill patients with liver cirrhosis: An Australian state-wide cohort study.

BACKGROUND

Patients with liver cirrhosis requiring intensive care unit (ICU) admission have a high in-hospital mortality, yet long-term mortality and predictors of mortality in survivors remain unknown.

METHODS

All patients with liver cirrhosis admitted to 27 ICUs in Victoria, Australia, between 2007 and 2018 reported to the Australian and New Zealand Intensive Care Society Adult Patient Database were included. Poisson regression and Cox regression were used to explore factors associated with in-hospital mortality and all-cause 12-month mortality. Liver transplantation was considered a censoring event.

RESULTS

A total of 5725 cirrhosis patients (3565 patient-years) were included. All-cause 12-month mortality was 43.8% (95% CI 42.5-45.1) and reduced over time (HR 0.979 per year, 95% CI 0.967-0.991). On multivariable analysis, factors associated with in-hospital death (risk ratio-RR, 95% CI) included ACLF (acute-on-chronic liver failure) (1.73, 1.57-1.90), decompensated cirrhosis (1.31, 1.13-1.51), ALBI (albumin-bilirubin) grade 3 versus 2 (1.65, 1.45-1.87), APACHE-III ICU admission diagnosis of sepsis (1.43 vs. upper gastrointestinal bleeding, 1.24-1.66) or liver failure (1.81 vs. upper gastrointestinal bleeding, 1.55-2.12) and older age. On multivariable analysis of 4068 transplant-free hospital survivors, 12-month mortality was influenced by (HR, 95% CI) liver disease severity (1.37 ALBI grade 3 vs. 2, 1.15-1.64) and decompensated cirrhosis (1.25, 1.06-1.49) rather than ACLF (0.88, 0.75-1.03).

CONCLUSIONS

Long-term mortality in liver cirrhosis patients admitted to the ICU is substantial but has decreased over time. ACLF confers a higher risk of in-hospital but not long-term death in hospital survivors. Cirrhosis severity and decompensation increase the risk of in-hospital and long-term mortality.