Gut Microbiota: Its Role in Hepatic Encephalopathy

Abstract

Ammonia, a key factor in the pathogenesis of hepatic encephalopathy (HE), is predominantly derived from urea breakdown by urease producing large intestinal bacteria and from small intestine and kidneys, where the enzyme glutaminases releases ammonia from circulating glutamine. Non-culture techniques like pyrosequencing of bacterial 16S ribosomal ribonucleic acid are used to characterize fecal microbiota. Fecal microbiota in patients with cirrhosis have been shown to alter with increasing Child-Turcotte-Pugh (CTP) and Model for End stage Liver Disease (MELD) scores, and with development of covert or overt HE. Cirrhosis dysbiosis ratio (CDR), the ratio of autochthonous/good bacteria (e.g. Lachnospiraceae, Ruminococcaceae and Clostridiales) to non-autochthonous/pathogenic bacteria (e.g. Enterobacteriaceae and Streptococcaceae), is significantly higher in controls and patients with compensated cirrhosis than patients with decompensated cirrhosis. Although their stool microbiota do not differ, sigmoid colonic mucosal microbiota in liver cirrhosis patients with and without HE, are different. Linkage of pathogenic colonic mucosal bacteria with poor cognition and inflammation suggests that important processes at the mucosal interface, such as bacterial translocation and immune dysfunction, are involved in the pathogenesis of HE. Fecal microbiome composition does not change significantly when HE is treated with lactulose or when HE recurs after lactulose withdrawal. Despite improving cognition and endotoxemia as well as shifting positive correlation of pathogenic bacteria with metabolites, linked to ammonia, aromatic amino acids and oxidative stress, to a negative correlation, rifaximin changes gut microbiome composition only modestly. These observations suggest that the beneficial effects of lactulose and rifaximin could be associated with a change in microbial metabolic function as well as an improvement in dysbiosis.

Hepatic encephalopathy (HE) is brain dysfunction caused by liver insufficiency and/or porto-systemic shunts. It manifests as a wide spectrum of neurological or psychiatric abnormalities ranging from subclinical alterations to coma.¹ For decades the pathogenesis of HE has been believed to be related to high levels of ammonia resulting from decreased detoxification in liver as a result of liver failure and/or due to the presence of porto-systemic shunts in patients with cirrhosis.² Recently, it has been shown that inflammation and oxidative and nitrosative stress play a role in the development of HE. There is evidence that HE is linked to alterations in gut microbiota and their by-products such as amino acid metabolites (indoles, oxindoles), endotoxins, etc. These factors, superimposed on a background of leaky intestinal barrier and immune dysfunction, are involved in the pathogenesis of HE.³ Thus, modulation of gut microbiota using lactulose, probiotics and nonabsorbable antibiotics is likely to play a role in the management of HE. Insight into the gut–liver axis as well as influence of gut microbiota on liver and vice versa is of utmost importance to understand the role of gut microbiota in pathogenesis of HE.

Gut–liver axis

Gut and liver share a close relationship. The liver, which receives 70% of its blood supply from the gut through the portal venous system, is significantly affected by the gut and its contents. The liver also influences intestinal functions through several mechanisms in physiological and pathological conditions. The gut and the liver have a pivotal role in absorption and metabolism of various nutrients and drugs. Abnormal bile acid homeostasis may lead to diarrhea and bacterial overgrowth. Cirrhosis contributes to bacterial overgrowth by decreasing gastrointestinal motility which leads to an increased risk of infections including spontaneous bacterial peritonitis (SBP).

Gut microbiota

The term microbiota is used to describe the complete population of microorganisms that inhabit the body in various locations including the gut. It numbers approximately 10¹⁴ microbial cells, and include bacteria, viruses, protozoa, etc.^4–6 It performs vital functions related to nutrition and metabolism, including food processing, digestion of complex carbohydrates and vitamin synthesis.^7,8 It also secretes a number of biologically active compounds which perform various functions like inhibition of pathogens and metabolism of toxic compounds including ammonia.⁸

Changes in gut microbiota in health

Earlier culture and biochemical typing were the gold standards for the identification of bacterial species; however for last two decades culture-independent techniques have revolutionized knowledge of the gut microbiota. These techniques are based on sequence divergences of the small subunit ribosomal ribonucleic acid (16S rRNA) and are able to demonstrate the microbial diversity of the gut microbiota, providing qualitative as well as quantitative information on bacterial species and changes in the gut microbiota in health and disease.^9,10 Changes have been described in composition of the human gut microbiota from early infancy to old age as well as due to the effect of environmental influences. Gastrointestinal tract is sterile at birth, when colonization of the gut takes place with the mother's vaginal canal flora and thereafter from the surrounding environment. The diversity of microbiota depends upon factors like mode of delivery (vaginal birth versus assisted delivery), gestational age, diet (breast versus formula milk), antibiotics, hygiene and sanitation.^11,12 The gut microbiota of newborn babies has relative dominance of the phyla Proteobacteria and Actinobacteria; later, the microbiota becomes more diverse with relative dominance of Firmicutes and Bacteroides, which are typical of adult microbiota.^7,13 Recently, three different genera ‘enterotypes’, enterotype 1 (Bacteroides), enterotype 2 (Prevotella) and enterotype 3 (Ruminococcus), have been described in the adult gut microbiome from different continents. Their dominance is likely to be independent of factors like sex, age, race and body mass index.¹⁴ Once maturity is reached, the microbiota remains stable until old age when changes can be seen, likely due to changes in diet and digestive metabolism.^15,16

Changes in gut function and microbiota in cirrhosis of liver

There are multiple mechanisms which are involved in defective gut functions and altered microbiota in patients with cirrhosis. These include impaired small intestinal motility, increased intestinal permeability, impaired antimicrobial defense and small intestine bacterial overgrowth (SIBO). Additionally, decreased bile acids, due to decreased synthesis and defective enterohepatic circulation, contribute to altered gut microbiota.^17,18

Gut Motility

Evidence suggests that gastrointestinal motility is delayed in patients with liver cirrhosis.^19–22 Several mechanisms have been proposed, including bowel wall edema, autonomic dysfunction, altered concentration of intestinal active peptides and neurotransmitters, and altered intestinal myoelectrical activity.^21,23,24 The abnormal antroduodenojejunal pressure wave leads to increased risk of SIBO in patients with portal hypertension.²⁵ In addition, HE itself may influence small bowel transit, as transit time has been shown to improve following treatment of HE.²⁶

Intestinal Permeability and Impaired Antimicrobial Defense

In cirrhosis, changes in intestinal tight junctional proteins have been described; though the pathophysiology is not clear, alcohol metabolites and proinflammatory cytokines have been postulated to result in leaky intestines.⁴ In addition, impaired antimicrobial defense mechanisms, involving intestinal Paneth cells, contribute to the development of bacterial translocation in cirrhosis.²⁷

Small Intestinal Bacterial Overgrowth

There is a very high prevalence of SIBO (35–61%) in patients with cirrhosis.^21,22,28,29 The pathogenesis of SIBO in these patients is postulated to include impaired intestinal motility, reduced gastric acid secretion, luminal IgA deficiency and malnutrition.³⁰ SIBO correlates with the severity of chronic liver disease, is shown to be linked to minimal and overt HE and confers increased risk for development of SBP through bacterial translocation across the gut.^21,22,31,32

Gut Microbiota in Cirrhosis

In prior studies, culture-based techniques were used for the characterization of gut flora in cirrhosis. Chen et al³³ were the first to use pyrosequencing of the 16S rRNA V3 region and real-time quantitative polymerase chain reaction (PCR) to characterize the fecal microbiota of patients with cirrhosis. They demonstrated that, when compared with controls, in patients with cirrhosis the proportion of phylum Bacteroidetes was significantly reduced, whereas Proteobacteria and Fusobacteria were abundant. At the family level Enterobacteriaceae, Veillonellaceae, and Streptococcaceae were increased and Lachnospiraceae was less prevalent. They observed a positive correlation of Child-Turcotte-Pugh (CTP) score with Streptococcaceae and a negative correlation with Lachnospiraceae^33,34 and also showed that fecal microbiome differed significantly between patients with cirrhosis and controls. They demonstrated a significantly greater abundance of Lachnospiraceae and Ruminococcaceae and reduction of Enterobacteriaceae, Fusobacteriaceae, Alcaligenaceae, Lactobacillaceae and Leuconostocaceae families in control group when compared with cirrhotic patients. Model for end stage liver disease (MELD) score was positively linked with the taxon Enterobacteriaceae and negatively with Ruminococcaceae.³⁴ In a recent study by Qin et al from China, the authors constructed a gene catalogue from 98 Chinese patients with cirrhosis and 83 healthy controls using Illumina-based metagenomic sequencing.³⁵ This catalogue was then compared with three other available gut microbial catalogues: MetaHIT (European), US National Institutes of Health Human Microbiome Project (HMP) and Type 2 diabetes. Bacteroidetes and Firmicutes were the most common microbes in both cirrhosis and healthy controls. Patients with cirrhosis had fewer Bacteroidetes and higher proportions of Proteobacteria and Fusobacteria compared to controls; a higher proportion of Veillonella, Streptococcus, Clostridium and Prevotella and a lower proportion of Bacteroides were seen. On the basis of only 15 microbial genes as biomarkers, a highly accurate patient discrimination index was created and validated on an independent cohort.³⁵

Ammonia, inflammation and gut flora in hepatic encephalopathy

Ammonia has been known to play a central role in the pathogenesis of HE since the 1890s.^2,36 The intervening century has seen refinement of our understanding of ammonia production, handling and disposal, specific mechanisms by which it produces hepatic encephalopathy, how it interacts with other pathogenic mechanisms and how it can be manipulated to therapeutic advantage. Urease producing bacteria exist in abundance in the gut of “ureolytic” animals. Urease is a bacterial enzyme that catalyzes the hydrolysis of urea to carbamate and ammonia. Urease producing bacteria are frequently gram negative Enterobacteriaceae, but may be anaerobes or gram positive bacteria.³⁷ Since hyperammonemia in cirrhosis has been thought to be derived predominantly from urea breakdown by urease producing bacteria, most of the treatments for HE have targeted these ammonia producing colonic bacteria.^38,39 However, lately, focus is shifting from the large to the small intestine and kidneys as the major sources of ammonia production. Glutamine deamidation by glutaminase seems to be the main source of ammonia generation in patients with cirrhosis.³⁹

Inter-organ Ammonia Trafficking

In the physiological state, glutamine is a crucial source of energy especially for small intestine. The intestine takes up glutamine from the blood-stream and breaks it down by the enzyme glutaminase, which contributes to gut-derived ammonia production. Ammonia detoxification in the liver occurs mainly in periportal hepatocytes via the Krebs urea cycle (ornithine cycle), leading to urea formation. Normally this is a very efficient process and very little ammonia escapes into the circulation. Any ammonia escaping detoxification in the periportal hepatocytes is usually trapped in the perivenous hepatocytes, where it is incorporated into glutamine by the glutamine synthetase reaction and released into circulation. Urea produced by the liver enters the systemic venous circulation via the hepatic veins and is excreted by the kidneys in the urine. Kidneys also contribute to formation of ammonia by metabolizing circulating glutamine by glutaminase. The ammonia thus generated mainly enters the systemic circulation to reach the liver, with only 30% being excreted into urine.^40,41 Skeletal muscle has low glutamine synthetase activity, but by virtue of its mass, it is an important glutamine synthesizing organ. It is involved in the regulation of blood ammonia level by incorporating it in to glutamine.⁴²

The physiologic balance of ammonia production and clearance is disrupted at multiple levels in patients with cirrhosis, resulting in hyperammonemia. Evidence suggests that patients with cirrhosis harbor more urease-active bacteria in the gut than controls, and that this leads directly to increased intestinal hydrolysis of urea and absorption of nitrogenous products.³⁸ Altered small intestinal motility frequently accompanies cirrhosis and likely exacerbates this problem.^21,26 Further, high levels of portal ammonia result in markedly increased systemic ammonia levels because of impaired hepatic processing of ammonia and shunting of portal blood. Perivenous hepatocytes, which harbor a high affinity, low capacity glutamine synthetase system, also contribute to increased glutamine levels. Glutamine subsequently undergoes degradation by the kidney which secretes a larger fraction of ammonia (70%) in urine.⁴¹

When liver fails to clear the ammonia generated, other organs, such as brain and muscle, are forced to adapt to a situation of high systemic ammonia levels, with glutamine synthesis being the most important alternative detoxification pathway. However, this pathway does not contribute to net nitrogen removal from systemic circulation, with glutamine only acting as a non-toxic nitrogen carrier.⁴¹ In the brain, ammonia is handled primarily by astrocytes, leading to increased glutamine synthesis, which produces a cascade of neurochemical events eventually leading to brain dysfunction.⁴³ Despite evidence of increased ammonia uptake by skeletal muscles in acute and chronic liver failure, studies show conflicting results regarding glutamine efflux.^42,44 Cirrhotics with significant muscle wasting are more likely to exhibit hyperammonemia, but two recent studies evaluating a link between malnutrition and HE showed equivocal results.^45,46

Linkage of Gut Flora with Systemic Inflammation and Cognition in Hepatic Encephalopathy

In the setting of impaired intestinal motility, SIBO, intestinal barrier dysfunction and systemic inflammation, altered gut flora and their by-products such as ammonia, amino acid metabolites (indoles, oxindoles), endotoxins, etc. play an important role in the pathogenesis of HE. In a study, presence of SIBO was linked to delayed intestinal transit and multivariate analysis demonstrated that SIBO was the only factor associated with MHE.²¹ SIBO, along with impaired intestinal barrier integrity, results in increased bacterial translocation and release of endotoxins (lipopolysaccharides, flagellin, peptidoglycan, and microbial nucleic acids) in circulation.^21,22,47 These bacterial products are also known as ‘pathogen-associated molecular patterns (PAMPs)’. Interaction of PAMPs with Toll-like receptors, which are specific pattern recognition receptors on Kupffer cells, leads to the activation of immune response and systemic inflammation.⁴⁸ Wright et al demonstrated that intra-peritoneal administration of lipopolysaccharide to cirrhotic rats induced pre-coma and exacerbated cytotoxic edema because of the synergistic effect of hyperammonemia and the endotoxemia induced inflammatory response.⁴⁹ There is human evidence now that systemic inflammation causes development and exacerbation of the symptoms of HE in patients with cirrhosis. Shawcross et al demonstrated that inflammation is an important determinant of the presence and severity of MHE in patients with cirrhosis. There is a greater deterioration in neuropsychological function following induced hyperammonemia in patients with cirrhosis who have more severe inflammation or infection.^36,50,51 They also showed that there is an association of infection and systemic inflammatory response syndrome (SIRS) with severe HE but not with ammonia. However the presence or absence of infection and SIRS did not determine survival.⁵² Proinflammatory cytokines like tumor necrosis factor (TNF)-α, interleukin (IL)-1β and IL-6 modulate the cerebral effect of ammonia by acting synergistically with ammonia in producing cognitive impairment in patients with cirrhosis and HE.⁵³ Now there is evidence for a role of neuroinflammation per se in liver failure. This evidence includes activation of microglia and increased in situ synthesis of the proinflammatory cytokines TNF-α, IL-1β and IL-6. In addition there is increased liver–brain signaling which includes direct effects of systemic proinflammatory molecules, recruitment of monocytes after microglial activation, and altered permeability of the blood–brain barrier.⁵⁴ However the exact mechanism by which inflammation causes brain edema and HE is not clear.

Recently Zang et al and Bajaj et al conducted studies in which they used a novel multitag pyrosequencing technique to characterize the microbiome from the fecal and colonic mucosal samples in patients with cirrhosis with and without HE and healthy controls.^35,55–57 Zang et al⁵⁷ found over-representation of two bacterial families, Streptococcaceae and Veillonellaceae, in cirrhotic patients with and without MHE as compared with controls. They also showed that gut urease containing bacteria Streptococcus salivarius was absent in controls but was present in cirrhotics with or without MHE. The abundance of S. salivarius was significantly higher in MHE than non-MHE patients, and its count positively correlated with ammonia levels in patients with MHE.⁵⁷

Bajaj and coworkers performed cognition testing, and inflammatory cytokines and endotoxin analysis in cirrhotics with and without HE.³⁵ They found altered fecal flora (higher Veillonellaceae), poor cognition, endotoxemia, and inflammation in patients with HE when compared with those without HE. However, there were no significant differences in the other microbiome families between HE and no HE groups. This study, for the first time, also demonstrated a direct positive correlation between Porphyromonadaceae and Alcaligeneceae and poor cognitive function in cirrhotics. Alcaligeneceae degrade urea to produce ammonia, which likely explains the association of Alcaligeneceae with poor cognitive function. They also demonstrated that Enterobacteriaceae, Fusobacteriaceae, and Veillonellaceae were positively, and Ruminococcaceae negatively, related to inflammation. Network analysis comparison showed robust correlations, only in the HE group, between the microbiome, cognition and inflammatory cytokines.³⁵ In another study, Bajaj et al⁵⁵ compared sigmoid mucosal microbiome of patients with cirrhosis with HE versus no HE. They showed lower Roseburea and higher Enterococcus, Villonella, Megasphaera and Burkholderia among sigmoid colonic mucosal microbiota in HE group when compared to no HE group. On the contrary, there were no significant differences between stool microbiome among these two groups. Among cirrhotics, they also demonstrated that autochthonous genera—Blautia, Fecalibacterium, Roseburia, and Dorea in sigmoid biopsies were associated with good cognition and decreased inflammatory markers, but pathogenic species like Enterococcus, Villonella, Megasphaera and Burkholderia were abundant in patients with HE, and were linked with inflammation and poor cognition.⁵⁵ It was hypothesized that in patients with cirrhosis, decrease in bile acids, antibacterial peptides and mucin in the colon and immune dysfunction at mucosal interface allow for growth and adherence of pathogenic bacteria resulting in increased bacterial translocation and constitutes an important process in development in HE.

In a recent study, Bajaj et al⁵⁶ studied stool microbiota in controls and age-matched compensated/decompensated/hospitalized cirrhotics. They calculated the ratio of autochthonous to non-autochthonous taxa [‘the cirrhosis dysbiosis ratio (CDR)’], with a low number indicating dysbiosis. Apart from comparing controls and cirrhotic sub-groups, authors also analyzed changes in CDR after decompensation using (1) longitudinal comparison in patients before and after development of HE and (2) longitudinal comparison in cohorts of hospitalized infected and uninfected cirrhotics, followed for 30 days. CDR was significantly higher in controls (2.05) than compensated (0.89), decompensated (0.66) and inpatients with cirrhosis (0.32) and negatively correlated with serum endotoxin levels. In patients studied before and after HE development, dysbiosis occurred post-HE (CDR: 1.2 to 0.42) and microbiota were significantly different between infected and uninfected patients with cirrhosis at baseline. Also, low CDR was associated with death and organ failures within 30 days.⁵⁶

Modulation of gut microbiota improves hepatic encephalopathy

Now that gut microbiota are clearly implicated in the development of HE, their modulation by various agents provides an opportunity to treat covert and overt HE. Successful modulation of gut microbiota leading to improvement in HE strengthens the belief that derangement in microbiota is certainly an important factor in development of HE.

Prebiotics, Probiotics and Synbiotics

Prebiotics, probiotics and synbiotics modulate gut microbiota and may exhibit efficacy in MHE and HE by various mechanisms including a decrease in counts of pathogenic bacteria, decreased bacterial urease activity and reduced ammonia absorption by decreasing luminal pH. They may decrease endotoxemia, inflammation and uptake of toxins like indoles, oxindoles, phenols, mercaptans, etc. Despite a number of trials showing the efficacy of prebiotics, probiotics and synbiotics, their role in the management of HE is inconclusive and, currently, they cannot be recommended.⁵⁸ In a recent study, the safety and tolerability of probiotic Lactobacillus GG strain (LGG) was evaluated in patients with MHE in a phase I randomized controlled trial (RCT). Patients were randomized to receive either LGG or placebo for 8 weeks. Endotoxin levels, systemic inflammation, fecal microbiome and serum and urine metabolome were analyzed at baseline and at 8 weeks and correlations between groups were analyzed using correlation network. The authors demonstrated that endotoxin and TNF-α level were decreased and improvement in CDR was observed with treatment only in the LGG group. Improvement in dysbiosis was associated with beneficial changes in metabolites like amino acids (lower amino-malonate, methionine, higher ammonia-detoxification products: benzoate, hydroxylamine), vitamin (lower urine riboflavin and ascorbate), and secondary bile acids (lower deoxycholate). There was also a beneficial change in correlations between bacterial taxa and metabolites in LGG group for Enterobacteriaceae and with hippurate/asparagine/glutamate. However, there was no change in cognition as the study was not powered for efficacy.⁵⁹ We have demonstrated that over a 6-month period, treatment with probiotic significantly reduced the risk of hospitalization due to HE and significantly reduced CTP and MELD scores. Treatment with probiotic was also associated with improvements in rates of SIRS, plasma indole and in brain natriuretic peptide, renin and aldosterone levels.⁶⁰

Lactulose

Lactulose, used as standard therapy in HE, works by altering gut flora to decrease ammonia production and absorption.⁶¹ Among the many actions of lactulose an important one is its role as a “prebiotic”, causing increased growth of endogenous bacteria that are potentially beneficial to the host like Lactobacilli, thereby indirectly reducing the strength of potentially more harmful urease producing bacteria.⁶²

Bajaj et al demonstrated that, despite lactulose treatment in those patients who developed HE, there was an increase in dysbiosis, with a lower CDR and relative abundance of gram negative non-autochthonous bacteria (Enterobacteriaceae, Bacteroidaceae).⁵⁶ This is in contrast to earlier culture-based studies which showed increased autochthonous bacteria (Lactobacillaceae) after administration of lactulose in patients with cirrhosis.⁶²

Lactulose withdrawal did not exert a very significant effect on the composition of the fecal microbiome except for the reduction of Faecalibacterium species, suggesting that changes in gut bacterial function rather than a change in the microbiome composition may be responsible for the effects of lactulose.³⁵ In another study HE patients underwent characterization of their phenotype (cognition, inflammatory cytokines, in vivo brain MR spectroscopy), gut microbiome and urine and serum metabolome analysis while on lactulose and on days 2, 14 and 30 post-withdrawal. When patients with and without recurrent HE post-withdrawal were compared, brain MRS findings consistent with low grade brain edema (increased glutamine/glutamate and decreased myo-inositol) were reported, with a relatively minor change in fecal microbiome (reduction in abundance of stool Faecalibacterium sp). HE recurrence was associated with altered choline metabolism by gut microbiome, resulting in low urine tri-methylamine oxide, high urine glycine and high serum choline, dimethylglycine, creatinine, which play an important role in development of HE.⁶³

Rifaximin

Rifaximin, a poorly absorbable synthetic antibiotic, modulates gut microbiota, is safe, is associated with low risk of bacterial resistance and is effective in both covert and overt HE.⁶⁴ The effect of rifaximin therapy on the metabiome, i.e. the interaction between the phenome (cognition, liver disease severity and endotoxin level), microbiome and metabolome, was evaluated in patients with cirrhosis and MHE. There was no significant microbial change after treatment with rifaximin, apart from a modest decrease in Veillonellaceae and increase in Eubacteriaceae. A significant improvement in cognition, reduction in endotoxemia and a significant increase in serum long-chain fatty acids were seen after rifaximin therapy. A significant linkage of pathogenic bacterial taxa was demonstrated with the metabolites, especially those linked to ammonia, aromatic amino acids and oxidative stress, which shifted from a positive correlation to a negative correlation after rifaximin therapy, reflecting changes in bacterial metabolic function. These results suggest that the mechanism of action of rifaximin that lead to cognitive improvement could be associated with changing microbiota-associated metabolic function rather than just changing the numbers of beneficial or harmful bacteria.⁶⁵

Later, in addition to re-affirming the beneficial effect of rifaximin on reducing endotoxemia, the same group demonstrated that rifaximin improves working memory performance on an N-back task and inhibitory control in MHE patients.⁶⁶ MHE patients underwent functional MRI (fMRI), diffusion tensor imaging and MRS. On fMRI, under scanner, two tasks - N-back (outcome: correct responses) and inhibitory control tests (outcomes: lure inhibition) were performed. On N-back fMRI analysis, there was significantly higher subcortical and left parietal operculum (LPO) activation after rifaximin therapy for 8 weeks. Decrease in fronto-parietal activation required for inhibiting lures, including LPO during inhibitory control test, compared to pre-rifaximin values, suggested that lesser neuronal recruitment is required after rifaximin to achieve the same control of the lure response. While there was a reduction in fractional anisotropy, suggesting improved neuronal integrity with rifaximin, no changes in mean diffusivity or MRS metabolites were observed.⁶⁶ However, in advanced MRI based studies, lactulose therapy in MHE patients has been shown to increase choline and decrease glutamine/glutamate and also to increase mean diffusivity but not fractional anisotropy, suggesting reversible low grade brain edema.^67,68 This may point to a difference in mechanism of action of rifaximin compared to lactulose that may only be apparent on brain imaging, but, clearly, more studies are needed before definite conclusions can be drawn.

In conclusion, ammonia has an important role in pathogenesis of HE and systemic inflammation modulates its effect. Degradation of circulating glutamine contributes to the intestinal production of ammonia, in addition to its production by gut bacteria. Alterations in gut microbiota as well as related metabolomes are now well characterized in cirrhosis. These changes are linked to endotoxemia, systemic inflammation and impaired cognition, in the setting of impaired gut motility, SIBO, and increased gut permeability. Recent RCTs, which utilized multitag pyrosequencing technique to characterize the microbiome from the fecal and colonic mucosal samples in patients with cirrhosis, suggested that probiotics, lactulose and gut specific antibiotic like rifaximin cause improvement in dysbiosis and HE. Such improvement relates to both changes in microbiota-associated metabolic function and improvement in dysbiosis.

Conflicts of interest

All authors have none to declare.