Abstract

Background: Minimal hepatic encephalopathy (MHE) is associated with poor outcomes but treatment strategies are limited. Rifamycin SV MMX (RiVM) is a novel rifampin derivative which a non-absorbable antibiotic with maximal impact in the colon. Aim: Evaluate impact of RiVM on microbiome, safety & gut-brain axis in an RCT.

Methods: We performed a phase 2 placebo-controlled, double-blind RCT under FDA IND. We randomized cirrhosis outpts with MHE (PHES or Stroop) 1:1 into RiVM or placebo 600 mg BID (1200 mg) BID for 30 days with 7 day post-drug f/u. There were 4 visits; baseline, day 7, 15 & 30. Primary outcome was stool microbial change (cirrhosis dysbiosis ratio CDR, high=good) in rifamycin vs placebo through 16SrRNA sequencing from baseline to day 30 (end). CDR is the ratio of Lachnospiraceae + Ruminococcaceae + Veillonellaceae to Enterobacteriaceae + Bacteroidaceae. Secondary outcomes were gut-brain (cognition, serum ammonia, optional brain MR spectroscopy, MRS), inflammatory (stool calprotectin), PROs (SIP: total, physical, psychosocial, high=worse) and handgrip strength. Comparisons between/within gps & delta (Δ Post minus Pre) values were compared.

Results: 58 pts were screened; 8 had overt HE, 11 screen failed due to no MHE on testing, 9 were not interested. Ultimately 30 pts were enrolled (15/gp), who completed the study without any safety concerns, including the post-drug visit with good adherence. Groups were largely equivalent on baseline but ammonia & SIP scores were higher in RiVM vs placebo (Fig B). 7 RiVM and 11 placebo-assigned pts agreed & were eligible for the optional brain MRS.
Microbiota: CDR decreased in RiVM pts due to ↓Lachnospiraceae & Ruminococcaceae, although Bacteroidaceae↑. There was ↓α-diversity & significant β-diversity change with clustering of post-RiVM vs pre & post-RiVM vs post-placebo (Fig D/E).
Labs: No change in MELD-Na but ammonia & calprotectin decreased in RiVM vs baseline and Δ ammonia was higher in RiVM (Fig B); no change in placebo. Cognition and Brain MRS: Although serial dotting, which tests for psychomotor speed improved in RiVM, no other changes were seen within/between gps. Brain Glutathione ↑with RiVM & decreased in placebo (p=0.03) on brain MRS but remaining metabolites (choline, myoinositol, glutamate/glutamine) remained similar. PROs: ΔPhysical SIP and handgrip were higher indicating improved strength & better physical QOL with RiVM vs placebo.

Conclusion: In this phase 2 double-blind, placebo-controlled RCT of rifamycin SVMMX in patients with cirrhosis and MHE, we found no safety concerns. RiVM Rx resulted in lowered gut microbial α-diversity and cirrhosis dysbiosis ratio. RiVM therapy was associated with reduction in blood ammonia and improved physical function and handgrip. There was also a reduction in brain oxidative stress with RiVM but no change in cognitive testing. RiVM, with predominant colonic action, may have important gut-brain axis modulatory impact in cirrhosis and MHE.