Abstract
BACKGROUND & AIMS: The efficacy of renal replacement therapy (RRT) in critically ill patients with cirrhosis remains dubious. We aimed to assess the impact of RRT on these patients' outcomes. METHODS: Multicenter retrospective cohort study including adult patients with cirrhosis admitted to intensive care units at University of Alberta Hospital (Edmonton, Canada) and Northwestern Memorial Hospital (Chicago, US) from January 2010 to December 2017. Primary exposure was receipt of RRT on ICU days 1 to 3. Fine and Gray multivariable regression with competing endpoints, in-hospital liver transplant (LT) and mortality, was performed. RESULTS: Among 898 patients, median (IQR) age was 57 (49-64) years and 539 (60.0%) were males. RRT on days 1 to 3 was used in 249 (27.7%) patients. Patients on RRT on days 1 to 3 had higher CLIF-C-ACLF scores on days 1 (61 vs. 55, p < 0.001) and 3 (59 vs. 50; p < 0.001) than others. During the hospital stay, 97 (10.8%) patients were transplanted and 296 (33.0%) died. Following adjustment for aetiology, number of extra-renal organ failures, and year of inclusion, using mortality as competing event, RRT on days 1 to 3 was associated with higher hazard of LT (HR (95% CI) = 1.54 (1.02-2.32); p = 0.039). Conversely, using LT as competing event, RRT on days 1 to 3 was not associated with mortality (HR (95% CI) = 1.15 (0.90-1.47); p = 0.25). CONCLUSIONS: Among critically ill patients with cirrhosis, early RRT, while offered more often to the sickest patients, was associated with a higher likelihood of receiving LT, but not with mortality.