Article Text

Download PDFPDF
Original research
Interaction of bacterial metagenome and virome in patients with cirrhosis and hepatic encephalopathy
  1. Jasmohan S Bajaj1,
  2. Masoumeh Sikaroodi2,
  3. Amirhossein Shamsaddini2,
  4. Zachariah Henseler3,
  5. Tasha Santiago-Rodriguez3,
  6. Chathur Acharya1,
  7. Andrew Fagan1,
  8. Phillip B Hylemon1,
  9. Michael Fuchs1,
  10. Edith Gavis1,
  11. Tonya Ward3,
  12. Dan Knights3,4,5,
  13. Patrick M Gillevet2
  1. 1 Gastroenterology, Hepatology and Nutrition, Virginia Commonwealth University and Central Virginia Veterans Healthcare System, Richmond, Virginia, USA
  2. 2 Microbiome Analysis Center, George Mason University, Manassas, Virginia, USA
  3. 3 Diversigen, New Brighton, Minnesota, USA
  4. 4 Department of Computer Science and Engineering, U, University of Minnesota, Minneapolis, MN, USA
  5. 5 Minnesota Biotechnology Institute, University of Minnesota, Minneapolis, MN, USA
  1. Correspondence to Dr Jasmohan S Bajaj, Gastroenterology, Hepatology and Nutrition, Virginia Commonwealth University Health System, Richmond, Virginia, USA; jasmohan{at}


Objective Altered bacterial composition is associated with disease progression in cirrhosis but the role of virome, especially phages, is unclear.

Design Cross-sectional and pre/post rifaximin cohorts were enrolled. Cross-sectional: controls and cirrhotic outpatients (compensated, on lactulose (Cirr-L), on rifaximin (Cirr-LR)) were included and followed for 90-day hospitalisations. Pre/post: compensated cirrhotics underwent stool collection pre/post 8 weeks of rifaximin. Stool metagenomics for bacteria and phages and their correlation networks were analysed in controls versus cirrhosis, within cirrhotics, hospitalised/not and pre/post rifaximin.

Results Cross-sectional: 40 controls and 163 cirrhotics (63 compensated, 43 Cirr-L, 57 Cirr-LR) were enrolled. Cirr-L/LR groups were similar on model for end-stage liver disease (MELD) score but Cirr-L developed greater hospitalisations versus Cirr-LR (56% vs 30%, p=0.008). Bacterial alpha/beta diversity worsened from controls through Cirr-LR. While phage alpha diversity was similar, beta diversity was different between groups. Autochthonous bacteria linked negatively, pathobionts linked positively with MELD but only modest phage-MELD correlations were seen. Phage–bacterial correlation network complexity was highest in controls, lowest in Cirr-L and increased in Cirr-LR. Microviridae and Faecalibacterium phages were linked with autochthonous bacteria in Cirr-LR, but not Cirr-L hospitalised patients had greater pathobionts, lower commensal bacteria and phages focused on Streptococcus, Lactococcus and Myoviridae. Pre/post: No changes in alpha/beta diversity of phages or bacteria were seen postrifaximin. Phage–bacterial linkages centred around urease-producing Streptococcus species collapsed postrifaximin.

Conclusion Unlike bacteria, faecal phages are sparsely linked with cirrhosis characteristics and 90-day outcomes. Phage and bacterial linkages centred on urease-producing, ammonia-generating Streptococcus species were affected by disease progression and rifaximin therapy and were altered in patients who experienced 90-day hospitalisations.

  • cirrhosis
  • intestinal microbiology
  • hepatic encephalopathy
  • liver

Data availability statement

Data are presented in supplement to the extent allowed by our institutional review boards

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Data availability statement

Data are presented in supplement to the extent allowed by our institutional review boards

View Full Text


  • Presented at Portions of this manuscript will be presented as an oral presentation for ILC 2020, which is now virtual.

  • Correction notice This article has been corrected since it published Online First. Author name for Tasha Santiago-Rodriguez has been corrected.

  • Contributors JSB conceptualised this and was involved at each step. MS was involved in sample processing and analysis, AF, CA, MF, EG were involved in recruitment and study conduct. PBH was involved in critical revision. ZH, TW, DK and TS-R were involved in metagenomics and bioinformatics. PMG and AS were also involved in bioinformatics analysis.

  • Funding Partly supported by VA Merit Review I0CX00176, NCATS R21TR002024 and R21TR003095,AHRQ RO1HS025412 and an investigator-initiated grant from Bausch Health.

  • Competing interests DK, EH, TSR and TW are employees of Diversigen, where the metagnomic and virome analysis was performed.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.