Microbiome as a Potential Diagnostic and Predictive Biomarker in Severe Alcoholic Hepatitis

Soon Sun Kim; Jung Woo Eun; Hyo Jung Cho; Do Seon Song; Chang Wook Kim; Young Seok Kim; Sung Won Lee; Yoon-Keun Kim; Jinho Yang; Jinhee Choi; Hyung Joon Yim; Jae Youn Cheong


Aliment Pharmacol Ther. 2021;53(4):540-551. 

In This Article


Through comparing the gut microbiome composition in 24 patients with SAH and 24 healthy controls, we selected 144 common taxa that showed increased or decreased abundances in patients for both bacteria and bacteria-derived EVs. Furthermore, we revealed 15 common taxa in bacteria and EVs that showed a significant pattern of recovery after rifaximin treatment, 6 of which decreased in abundance and 9 of which increased in abundance after rifaximin administration. Among them, Veillonella and Veillonella parvula group showed the most significant increases in patients with SAH, and they significantly decreased after rifaximin treatment. Moreover, Prevotella and Prevotellaceae were significantly less abundant in patients with SAH and were restored after rifaximin treatment.

Alcohol-related liver diseases are associated with bacterial overgrowth in the gut. The gut microbiota composition in patients with alcoholic liver disease shows a predominance of harmful bacteria, such as Actinobacteria and Firmicutes, and a reduction in beneficial bacteria such as Bacteroides.[28] Rifaximin has been shown to be highly effective in cirrhotic patients with hepatic encephalopathy, and could reduce the risk of hepatic encephalopathy recurrence.[29] A recent study showed that rifaximin did not cause significant changes in the gut microbiota, although modification in the microbiome-metabolome network involving several bacteria (Enterobacteriaceae, Bacteroidaceae, Veillonellaceae, Porphyromonadaceae and Rikenellaceae) was observed.[30] In terms of the functional composition of metagenomes, our study showed that the metabolic pathway is the most enriched pathway in patients with SAH, both in gut bacteria and microbe-derived EVs. However, there was no significant change in these pathways after rifaximin administration, except for the porphyrin and chlorophyll metabolism pathway, suggesting a modest functional change in the gut microbiome after rifaximin administration. Although our data failed to demonstrate a significant change to a more favourable microbiota composition during rifaximin treatment, favourable effects of rifaximin in the gut microbiome milieu could be expected. Ponziani et al[31] suggested that rifaximin might have favourable gut microbiota-modulating activity beyond its classical antibiotic activity, which could be mediated by the selection of beneficial intestinal bacteria strains such as lactobacilli. Shotgun metagenome sequencing might be required to obtain the complete functional analysis profile to elucidate the additional role of rifaximin in patients with SAH.

We found that Veillonella and Veillonella parvula group were the most significantly increased taxa in patients with SAH, and their abundances significantly decreased after rifaximin treatment. An increase in the abundance of the genus Veillonella was previously reported to be associated with alcohol consumption in patients with hepatitis B virus-related liver cirrhosis, suggesting that alcohol-related pathogenesis caused alterations in the microbial community.[32] Another recent report indicated that Veillonellaceae were enriched in patients with heavy alcohol consumption, especially in those who developed SAH.[33] Taken together, those findings suggest that Veillonella could be a biomarker or severity predictor for alcoholic hepatitis.[34]

By contrast, Prevotella was significantly less abundant in patients with SAH, and its abundance was restored after rifaximin treatment. Furthermore, treatment nonresponders showed significantly lower Prevotella abundance than did treatment responders. Previous studies on Prevotella in liver disease have reported varying results. Prevotella was decreased in duodenal aspirates of patients with alcoholic liver cirrhosis compared with that in patients with hepatitis C virus-related cirrhosis.[35]Prevotella was also decreased in faecal samples of patients with cirrhosis but not in those of alcohol drinkers without cirrhosis.[36] By contrast, Prevotella abundance was increased in the saliva and faecal samples of patients with cirrhosis.[37]Prevotella was also increased in duodenal mucosa samples of patients with hepatitis B virus-related cirrhosis.[38] The inconsistent results may be explained by the high genetic diversity of Prevotella species and their high sensitivity to diet (long-term fibre vs animal-based diet).[39,40]

The Firmicutes/Bacteroidetes ratio is associated with the morbidity and pathogenesis of several chronic diseases such as obesity and diabetes.[41–43] A recent study suggested that an increased Firmicutes/Bacteroidetes ratio was an independent risk factor for hepatitis B virus-related liver cirrhosis.[32] The increased Firmicutes/Bacteroidetes ratio in patients with SAH in the current study might be attributable to alterations in lipid metabolism and fatty acid biosynthesis.[42] However, the change in the Firmicutes/Bacteroidetes ratio should be carefully interpreted considering the limited number of patients in the pre- and post-rifaximin groups.

Microbe-derived EVs are emerging as active biomolecules that function in drug delivery to target sites or regulating host cellular responses.[10] The gut microbiota can secrete different types of EVs, including outer membrane vesicles, shedding vesicles and apoptotic bodies.[11] We found that the microbial compositions based on bacteria-derived EVs did not directly reflect those of the gut microbiota. Comprehensive studies integrating the microbiota itself and microbe-derived EVs are required to better understand the communication between the microbiota and their EVs.

The present study has a few limitations that should be acknowledged. First, we did not include active alcohol drinkers without liver disease or patients with nonsevere alcoholic hepatitis. Therefore, it is uncertain whether the SAH-related taxa we have identified are associated with the effects of alcohol itself, alcoholic steatohepatitis, or the severity of alcoholic hepatitis. As the present study was conducted as a part of an ongoing clinical trial (NCT02485106) involving patients with SAH, this limitation was inevitable. Second, follow-up faecal samples were collected only in patients who received the rifaximin treatment. Therefore, the changes in the gut microbiome after 4 weeks may be associated with the rifaximin treatment but also may be related to the improvement of liver function itself.

In conclusion, we confirmed a state of SAH-related dysbiosis and more favourable restoration after rifaximin administration in an analysis of faecal microbiomes of patients with SAH. Further external validation is required to determine whether Veillonella or Prevotella could serve as a new diagnostic biomarker or severity predictor for alcoholic hepatitis.