Antibiotics With or Without Rifaximin for Acute Hepatic Encephalopathy in Critically Ill Patients With Cirrhosis: A Double-Blind, Randomized Controlled (ARiE) Trial

Abstract

INTRODUCTION: Critically ill patients with cirrhosis admitted to the intensive care unit (ICU) are usually on broad-spectrum antibiotics because of suspected infection or as a hospital protocol. It is unclear if additional rifaximin has any synergistic effect with broad-spectrum antibiotics in ICU patients with acute overt hepatic encephalopathy (HE). METHODS: In this double-blind trial, patients with overt HE admitted to ICU were randomized to receive antibiotics (ab) alone or antibiotics with rifaximin (ab + r). Resolution (or 2 grade reduction) of HE, time to resolution of HE, in-hospital mortality, nosocomial infection, and changes in endotoxin levels were compared between the 2 groups. A subgroup analysis of patients with decompensated cirrhosis and acute-on-chronic liver failure was performed. RESULTS: Baseline characteristics and severity scores were similar among both groups (92 in each group). Carbapenems and cephalosporin with beta-lactamase inhibitors were the most commonly used ab. On Kaplan-Meier analysis, 44.6% (41/92; 95% confidence interval [CI], 32–70.5) in ab-only arm and 46.7% (43/92; 95% CI, 33.8–63) in ab + r arm achieved the primary objective (P = 0.84).Time to achieve the primary objective (3.65 ± 1.82 days and 4.11 ± 2.01 days; P = 0.27) and in-hospital mortality were similar among both groups (62% vs 50%; P = 0.13). Seven percent and 13% in the ab and ab + r groups developed nosocomial infections (P = 0.21). Endotoxin levels were unaffected by rifaximin. Rifaximin led to lower in-hospital mortality (hazard ratio: 0.39 [95% CI, 0.2–0.76]) in patients with decompensated cirrhosis but not in patients with acute-on-chronic liver failure (hazard ratio: 0.99 [95% CI, 0.6–1.63]) because of reduced nosocomial infections. DISCUSSION: Reversal of overt HE in those on ab was comparable with those on ab + r.