The first prospective randomized control trial for antiviral use in ACHBLF patients was published by Grag et al in 2011, which demonstrated that tenofovir could safely reduce short term mortality in these patients. Our study was concluded prior to the first aforementioned randomized control trial. Although antiviral treatment may provide some short term survival benefits, many patients died despite the significant reduction of HBV DNA. Thus, our data remained relevant for the understanding of disease mechanism and the future development of novel intervention. Previous studies have demonstrated that endotoxinemia and delayed clearance of LPS in the circulation resulted in the development of ACLF in alcoholic liver disease. Although Han et al proposed that LPS played an important role in ACHBLF as a secondary liver injury on top of the CHB infection in animal models. The changes of LPS levels and their roles on disease severity in patients with ACHBLF were not fully explored. Our study showed that baseline LPS levels in ACHBLF patients did not differ from those in the healthy controls. However, significant elevation in LPS levels was observed in the peak phase of ACHBLF when compared to those in the progression or remission phase. The abnormal distributions of LPS levels among different phases were statistically significant in ACHBLF. In addition, the changes in LPS levels were correlated with MELD-Na scores in the progression and the peak phase. To our knowledge, this is by far the first study in which detailed the dynamic changes of LPS levels in different phases of ACHBLF. Since MELD-Na scores were correlated with LPS levels in the progression and the peak phase, our data pointed to the direction of the secondary injury from LPS in chronic liver disease leading to liver failure, which was proposed by Han et al. in the study from animal model. Further studies with histology correlation to LPS are needed to confirm if the severity of liver injury actually is directly correlated with LPS levels in ACHBLF patients. The findings in this study also implied a possible therapeutic intervention for ACHBLF by removing LPS from the serum. Several studies done by Adachi et al observed that there was a positive correlation between the occurrence of bacterial translocation from the gut to portal system and liver dysfunction in alcoholic hepatitis. Li et al demonstrated that elevation of endotoxin levels in the circulation from translocation of gut flora occurred during acute flares in patients with chronic hepatitis. It is possible that the elevation of LPS level in CHB patients was due to bacterial translocations from the gut to portal circulation GSK1120212 resulting in endotoxemia in the early phase of ACHBLF. On the other hand, the liver dysfunction in the early stage of ACHBLF probably further induced bacterial translocation from the gut leading to higher level of endotoxemia. In addition, in patients with liver dysfunction, the uptake of endotoxin by hepatic and Kupffer cells were compromised as compared to normal physical conditions, resulting in higher circulating levels of LPS. High levels of LPS then induced the aggravations of liver injury through the LPS-MD-2/TLR4/NF-kb signal pathway and further negatively impacted on KC and hepatic clearance of endotoxin. Thus, it is expected that the peak level of LPS was observed during the peak phase of ACHBLF. In our study, the dynamic changes of LPS were paralleled with the changes of TBil and MELD-Na in different phases of ACHBLF. The changes in LPS levels were correlated with MELD-Na scores in the progression and the peak phase, further indicated that the worsen disease severity was the result of LPS induced liver injury. The MELD-Na scores were not correlated with LPS levels in the remission phase.