Gut flora clearly matters for the pathogenesis of hepatic encephalopathy (HE). Indeed, lactulose, the most commonly used standard therapy, works in part by altering gut flora to decrease ammonia production and absorption.1 Moreover, evidence suggests that the latter effects occur because lactulose serves as a “prebiotic” ingredient,2 encouraging the growth of endogenous bacteria that resemble those found in many probiotics.3, 4 Thus, the concept that therapeutic manipulation of the endogenous gut bacterial flora (gut flora therapy) by prebiotics, probiotics, or a combination of the two (synbiotics) might work for HE, as reflected in this issue,5 is not novel. Indeed, study co-author Riordan touted the idea and cited previous human trials in a review seven years ago.6 Furthermore, gut flora therapy has additional potentially beneficial effects that other therapies for HE lack.7 Preliminary work suggests that such effects help to explain why gut-directed therapies also improve uremia in hemodialysis patients.8 Thus, the conceptual foundation for gut therapy of HE is ripe.
The present study by Liu et al.5 is a single center, controlled trial in a relatively homogenous study population (mostly hepatitis B cirrhotics) with thoughtful exclusion criteria. The authors use a single psychometric test, the number connection test (NCT), and a simple neurophysiologic test, brainstem auditory evoked potential (BAEP), to monitor changes in participants taking either placebo, fermentable fiber, or a synbiotic preparation (fermentable fiber plus a probiotic compound). Participants were evaluated with these tests, plus serum and stool analyses, at study entry, after a 1-month treatment period, and again after a 2-week washout period. By designing such a relatively basic study, the authors risked the possibility of incorrectly failing to identify a true therapeutic difference between the treatment and placebo groups (type II error). Thankfully for gut flora enthusiasts, however, impressive and exciting improvements in HE with both synbiotic therapy and fiber alone are reported.
One should note that clinical trials in HE are not straightforward. The ideal therapeutic trial has yet to be conducted. According to a recent consensus statement, criticisms that apply, to some degree, to all HE trials include “the large spectrum of clinical conditions summarized under this term, the definition of study endpoints, the treatment of control groups, and the methods used to quantify therapeutic effects.”9 Further, despite the expert panel's attempt to provide guidance for HE trials, controversies and difficulties persist.10
Therefore, like all previous research on HE, there are important concerns about the present study design that deserve commentary. It is notable, for example, that while the authors describe a straight randomization scheme, the groups were matched almost perfectly for age, gender, etiology of cirrhosis, and Child-Pugh class despite the relatively small numbers of subjects enrolled. A single psychometric test (NCT) which relies primarily on psychomotor speed and visuo-spatial orientation was used to define and monitor HE. This approach is contrary to the recommendations of many experts who recommend using a more complete psychometric battery.11 Indeed, the fact that only a small fraction of the subjects (7%) were found to have HE by both NCT and BAEP should, itself, raise concern about the sensitivity and specificity of the NCT. Finally, treatment crossover after washout and subsequent re-evaluation of study endpoints would also have strengthened the trial. Given these “nuts-and-bolts” concerns about trial design, the present study by Liu et al. is notable as an HE treatment trial mostly because it worked.
On the other hand, a major strength of the present study includes detailed stool analyses for bacterial colony counts and pH. Importantly, the analyses include matched controls. This effort not only lends biological plausibility to the treatment success, but also advances our understanding of how the putative therapy actually changes the gut flora by providing data on eight classes of bacteria. The detailed bacteriologic analysis provided by Liu et al. should be requisite for human probiotic studies and, hopefully, will serve to better elucidate mechanisms of action.
In light of this major accomplishment, it is somewhat disappointing that the authors neither explain nor reference their choice of the probiotics that comprise “Cocktail 2000” or the composition of the fermentable fiber (although co-author Bengmark is extensively published on these topics12, 13). Since there are many candidate prebiotic ingredients and probiotic bacteria (as well as a bewildering number of possible combinations), such information is critical for identifying the optimal composition and dose schedule of gut flora therapy. Presently, Lactobacillus and Bifidobacteria species are common in commercial probiotic preparations due, at least in part, to their traditional role in fermentable dairy products such as yogurt. Nevertheless, in response to rising consumer interest, numerous companies are developing an abundance of products containing a wide variety of constituent bacteria and mixtures, resulting in many competing claims.14 Patients, skeptics, and regulators deserve no less than a precise scientific rationale for study treatment choices.
Focusing on such issues, however, misses the study's greater excitement: altering the gut flora may improve not only HE, but also liver disease. The authors find improvement not only in HE outcomes, but also ALT and Child-Pugh class. Of almost equal interest, similar benefits were noted following either synbiotic or fermentable fiber. These results are impressive, and to our knowledge, the most compelling demonstration of such efficacy in humans.
The concept that gut flora matters for the pathogenesis of liver disease generally is also not novel. An extensive body of literature, mostly derived from animal studies, highlight the concept that gut-derived bacteria and their products injure the liver and can cause systemic illness.15, 16 The notion that gut flora-based therapy, specifically probiotics, might impact and disrupt the pathogenesis of liver disease, however, is relatively more recent. Interestingly, probiotics have been demonstrated to be effective for alcoholic17 and nonalcoholic18 steatohepatitis in their respective murine models. Much essential research is ongoing to elucidate the basic mechanisms of action of this interaction, and a sampling of very recent work is presented in Table 1. The biological systems under study are extremely complex. The human gut contains >1011 bacteria consisting of >400 separate bacteria interacting with a dynamic GI lumen that is home to a large portion of the body's resident and circulating lymphocytes; modern molecular techniques are only beginning to unravel these complex relationships.19, 20 Additionally, given the inherent differences between structure and metabolism of murine and human gastrointestinal tracts, the relevance of animal models is uncertain. Nevertheless, recent progress has been impressive.
|Author/Country||Study Model/Probiotic||Study Endpoint|
|Chiva24/Spain||Cirrhotic rats/Lactobacillus johnsonii La1 + antioxidant||Treatment prevents bacterial translocation and endotoxemia|
|Bauer25/Spain||Cirrhotic rats/Lactobacillus stain GG||Treatment failed to prevent bacterial translocation|
|Adawi26/Sweden||Acute liver injury rats/various lactobacilli and a Bifidobacterium stain studied separately||Some efficacy in bacterial translocation and hepatocellular damage; effect varies by bacterial strain|
|Horie27/Sweden||Human flora associated mouse model/Strept, Clostrid, Baciullus spp||Treatment decreases DNA adduct formation in colon epithelium, but not liver or kidney|
|Riordan28/Australia||Cirrhotic patients/synbiotic compound||Treatment upregulates TLR2 expression on peripheral blood monocytes|
The study by Liu et al. merits attention as a big step forward in the translation of such research to humans. Their finding that gut therapy improved hepatic inflammation and overall liver function in cirrhotic patients (mostly with viral hepatitis) has implications for the pathogenesis of liver disease, as well as its clinical management. Certainly, this provocative work needs to be corroborated by larger studies. But if additional research shows similar results, one immediately wonders … Could gut flora therapy reduce inflammation in other forms of hepatitis? Might it prevent other complications of cirrhosis where bacteria or their products are implicated (e.g. peritonitis, variceal hemorrhage)? Also, given the multiple GI derangements in patients with cirrhosis, would gut flora therapy be applicable to patients with lesser degrees of hepatic impairment? Indeed, the potential research agenda is vast (Figure 1).
Finally, from the standpoint of human subjects research, the finding that fermentable fiber alone demonstrated therapeutic benefits merits attention. Since patients with cirrhosis are considered to be immunosuppressed,21 they may be more vulnerable to real or theoretical risks of ingesting living bacteria. While probiotics are thought to be extremely safe by many experts,22 few have been rigorously evaluated in open label safety studies. As the real biological effects of probiotics are increasingly appreciated, it is incumbent upon investigators to carefully monitor risk. This topic is of particular interest to American investigators due to the lively ongoing debate about the proper regulation of substances widely available as “dietary supplements.”23 The U.S. Food and Drug Administration holds any supplement given to improve health (including probiotics) to the same regulatory standards as more conventional pharmaceuticals. However, since fermentable fiber and other “prebiotics” contain no living organisms, some safety and regulatory concerns are obviated.
In conclusion, Liu et al. have made a major contribution to the application to gut flora therapy to humans with liver disease. We expect this research will stimulate further interest in the study of gut flora therapy and the “gut-liver” axis, because the liver does, indeed, care about the gut.