Introduction

Hepatitis B virus (HBV) infection is a major public health problem. According to World Health Organization, approximately 400 million people worldwide are affected by chronic HBV infection and roughly one million people die annually from HBV-related diseases [1]. Chronic HBV infection causes a spectrum of diseases such as chronic hepatitis B, cirrhosis, hepatocellular carcinoma, and liver failure. HBV-related acute-on-chronic liver failure (ACLF) is a clinical syndrome defined as acute hepatic insult with diagnosed or undiagnosed chronic liver disease [2]. The major liver functions, particularly synthetic function, detoxification, and metabolic regulation, are impaired to different degrees, probably resulting in life-threatening complications such as hepatic encephalopathy, hepatorenal syndrome, severe jaundice, ascites, and coagulopathy, which have a high mortality rate of 50–90%. Only a limited number of medical treatments are available for ACLF. Although liver transplantation is a life-saving treatment for ACLF, the difficulty in finding a suitable donor and the high cost hinder its extensive clinical use.

The precise mechanism underlying the liver injury caused by HBV-related ACLF and the factors contributing to the progression of liver failure remain unknown. Generally, virus factors, host factors, and their interaction determine the result of ACLF. HBV DNA replication is one of the key factors causing the progression from liver damage to liver failure. The HBV DNA level is closely associated with the risk of hepatocellular carcinoma development, and HBV DNA suppression significantly improves the prognosis of cirrhosis. Current clinical guidelines advocate oral antiviral treatment in decompensated cirrhosis and sustained HBV DNA suppression to reduce sequelae [1–5]. Four nucleotide/nucleoside analogues (NAs) (lamivudine, adefovir dipivoxil, enticavir, and telbivudine) have been used for the treatment of CHB. However, limited data are available from the Eastern and Western world on the use of antiviral treatment among patients with ACLF [6, 7]. Here, we report a single-center experience of the use of NAs in 106 patients with HBV-related ACLF. To the authors’ knowledge, this is one of the biggest single-center experiences of the use of antiviral treatment to date.

Patients

A retrospective review was carried out of adult patients with HBV-related ACLF, who visited the First Affiliated Hospital of Xi’an JiaoTong University, China between January 2008 and May 2010. In China, most patients infected HBV during early childhood.

ACLF was diagnosed according to the diagnostic criteria recommended by the Asian Pacific Association for the Study of the Liver (APASL) [2], i.e., two insults to liver are operating simultaneously: one of them being ongoing and chronic and the other being acute. The inclusion criteria were defined as follows: (1) age 18–65 years; the presence of hepatitis B surface antigen in the serum for at least 6 months; HBV DNA level above 3 log10 IU/mL; (2) recent development of increasing jaundice (a total serum bilirubin concentration of above 85 μmol/L) and a decrease in plasma prothrombin activity (PTA) (<40%) or with an international standard ratio (INR) of above 1.5; (3) recent development of complications such as hepatic encephalopathy, or abrupt and obvious increase in ascites, or spontaneous bacterial peritonitis, or hepatorenal syndrome. The exclusion criteria included the following: (1) patients who suffered from hepatitis A, hepatitis C, hepatitis D, hepatitis E, Epstein–Barr virus, or other virus infections; (2) patients with drug-induced hepatitis, Wilson’s disease, alcoholic liver disease, and autoimmune hepatitis.

The study protocol was in accordance with the ethical guidelines of the Declaration of Helsinki (1995) and was approved by the independent ethics committee at the First affiliated Hospital of Xi’an JiaoTong University, China.

Treatment schedule

A total of 2,308 patients were admitted to the Division of infectious diseases, the First affiliated Hospital of Xi’an JiaoTong University between July 2008 and June 2010. All the patients with elevated alanine aminotransferase (ALT) were detected; HBV marker and history were taken to diagnose chronic HBV infection. Of the 2,308 in-patients with hepatitis, 1,386 patients were diagnosed with chronic HBV infection; of the 1,386 patients, 168 were diagnosed with ACLF and met the inclusion criteria mentioned above; of the 168 patients, 21 were excluded according to the exclusion criteria. As a result, a total of 147 adult patients with ACLF were recruited for the study. Of the 147 patients, 42 were required to receive 0.5 mg entecavir (ETV) daily (ETV group), 30 were required to receive 100 mg lamivudine (LAM) daily (LAM group), 34 patients did not take any nucleos(t)ide analogue (non-NAs group), and the remaining 41 ACLF patients were not in the regime of this study who required others nucleos(t)ide analogue strategy, such as telbivudine, adefovir, LAM plus ADV, 1.0 mg ETV daily. All eligible patients were given standard medical treatment in the intensive care unit, including absolute rest on bed, energy supplements and vitamins, intravenous drop infusion albumin and plasma, electrolyte and acid–base equilibrium, and prevention and treatment of complications.

MELD-Na score

The serum creatinine level, INR for prothrombin time, the total serum bilirubin level, and the serum sodium concentration of each ACLF patient were recorded on the day of admission. The MELD-Na score was calculated using the following formula [8]: 3.78 × loge (bilirubin [mg/dL]) + 11.2 × loge (INR) + 9.57 × loge (creatinine [mg/dL]) + 6.43 + 1.59 × (135 − Na [mEq/L]). The maximal serum creatinine level considered in the MELD-Na score equation was 4.0 mg/dL, with the maximum and minimum Na values of 135 and 120 mEq/L, respectively.

Measurement of HBV marker, HBV DNA, and YMDD mutations

HBs-Ag, HBs-Ab, HBe-Ag, and HBe-Ab were detected by a qualitative enzyme immunoassay. Serum HBV DNA was measured by real-time polymerase chain reaction (PCR) assay. The detection limit was 1 × 103 IU/mL. HBV DNA levels of the patients were evaluated before and after the 3-week treatment. The tyrosine (Y), methionine (M), aspartate (D), and aspartate (D) motif mutant (YMDD motif mutant) were detected every 3–6 months by PCR assay during the follow-up period, with the lower limit of detection being approximately 1 × 104 IU/mL.

Follow-up

The patients were observed every 1–2 weeks for the first 3 months and every 3–6 months thereafter until October 2010. The clinical and laboratory data, adverse events, and compliance of all patients were recorded.

Statistical analysis

Quantitative data were expressed as mean ± standard deviation (SD) or median and range. One-way analysis of variance (ANOVA), t-test, or the nonparametric Mann–Whitney U test was used. A chi-squared test or Fisher’s exact test were performed for comparison of qualitative data. A P value <0.05 was considered to indicate statistical significance. Data were analyzed using the Statistical Product and Service Solution (SPSS) software version 11.5 for Windows.

Results

Clinical backgrounds of the patients

One hundred and six consecutive patients with ACLF were enrolled and divided into three groups (ETV group, LAM group, and non-NAs group), as mentioned above. Background data of the patients in each group are shown in Table 1.

Table 1 Demographic data and baseline characteristics
Full size table

There were no significant differences between the three groups in the baseline characteristics of age, gender, serum ALT, AST, ALB, GLB, TBIL, CHE, CREA, INR, MELD score, and complications. The baseline HBV DNA level was much higher in the ETV group and LAM group than that in the non-NAs group (P = 0.036).

HBV DNA level

After 3 weeks of nucleoside analogue treatment and/or supportive therapy, the HBV DNA levels were decreased, when compared with the baseline level, in the ETV group (7.04 ± 1.58 log10 IU/mL vs. 4.03 ± 2.04 log10 IU/mL, P = 0.001), the LAM group (7.25 ± 0.89 log10 IU/mL vs. 4.33 ± 2.48 log10 IU/mL, P = 0.01), and the non-NAs group (5.73 ± 0.96 log10 IU/mL vs. 4.21 ± 1.47 log10 IU/mL, P = 0.01).

A significantly high proportion of patients acquired a >2 log10 IU/mL decrease in HBV DNA in the ETV group (81.3% vs. 42.9%, P = 0.013) and LAM group (80% vs. 42.9%, P = 0.031), compared with that in the non-NAs group.

A large proportion of patients in the antiviral therapy treatment group (ETV group + LAM group) acquired undetectable HBV DNA levels (1 × 103 IU/mL), when compared with that in the non-NAs group (69.2% vs. 35.7%, P = 0.025).The rate of the undetectable HBV DNA level after 1 month of treatment was significantly higher in the ETV group than in the non-NAs group (75.0% vs. 37.7%, P = 0.014). However, no statistically significant difference was detected between the ETV and LAM groups (75.0% vs. 60.0%, P = 0.202) and between the LAM and non-NAs groups (60.0% vs. 37.7%, P = 0.148).

Short-term survival

During the first month of treatment, a total of 39 patients died, 9 (21.4%) in the ETV group, 10 (33.3%) in the LAM group, and 20 (58.8%) in the non-NAs group; 1 patient in the ETV group was transferred to the liver transplantation center for orthotopic liver transplantation. In the second month, 2 patients in the ETV group accepted liver transplantation; a total of 6 patients died, 4 in the ETV group, 1 in the LAM group, and 1 in the non-NAs group. In the third month, 2 patients died, 1 in the LAM group and 1 in the non-NAs group. In the first 3 months, a total of 3 patients received liver transplantation and all of them were from the ETV group; forty-seven patients died. The ETV and LAM groups had a similar accumulative mortality during the first 3 months of treatment (33.3% vs. 40%, χ2 = 0.568, P = 0.374). However, the non-NAs group had a significantly high mortality, compared with the ETV group (64.7% vs. 33.3%, χ2 = 7.163, P = 0.007), the LAM group (64.7% vs. 40%, χ2 = 3.906, P = 0.042), and the nucleoside analogue group (ETV group + LAM group) (64.7% vs. 36.2%, χ2 = 7.443, P = 0.006).

Prognostic factors associated with short-term mortality

We compared the pretreatment data including age, HBV-related marker, total bilirubin, liver function, creatinine, PTA, MELD-Na score, and antiviral treatment strategy between patients who expired within 3 months and those who survived more than 3 months (Table 2). Univariate analyses indicate that TBIL > 171 μM/L (P = 0.007), CHOL < 1.7 mM/L (P = 0.002), PTA < 30% (P = 0.001), without NAs treatment (P = 0.006), CREA > 146 μM/L (P = 0.004), and MELD-Na score > 25 (P = 0.001) were the significant predictors of early mortality.

Table 2 Comparisons of baseline characteristic between survivors and non-survivors after 3 months treatment
Full size table

Long-term outcome of nucleoside analogue treatment

Of the 106 patients, 47 patients died within the first 3 months of nucleoside analogue treatment, 3 patients were transferred to the liver transplantation center for orthotopic liver transplantation, 56 patients survived. All the 56 patients were followed up until death or until October 2010. The median follow-up period was 7.3 months. Recurrence was observed in a total of 6 (10.72%) patients, 4 (33.33%) in the non-NAs group, 2 (11.11%) in the LAM group after the LAM therapy was terminated by patients themselves, and 0 in the ETV group (P = 0.003). Those results indicate that the risk of recurrence may be reduced by continuous nucleoside analogue treatment. No YMDD drug-resistant mutant was detected in the survivors during the follow-up period.

Discussion

HBV infection is one of the major causes of ACLF in China. In most patients with liver diseases, ACLF is triggered by cirrhotic complication such as spontaneous bacteria peritonitis and upper GI bleeding [6, 9, 10]. In patients with chronic hepatitis B, severe acute exacerbation is another important cause of ACLF [11]. The pathogenesis of ACLF remains incomplete understood [12]. A “two-hit” hypothesis may be proposed to partially explain the pathogenesis of ACLF: the first hit is considered to be the primary injury caused by HBV, and the second one is the host’ imbalanced reaction to these initiating factors [13].

HBV replication is one of the key factors causing severe liver damage. High virus load is associated with rapidly progression and HCC [14]. Early antiviral treatment shortens and improves the symptomatic phase of infection and allows for a ready clinical and biochemical improvement [11, 15, 16]. Lamivudine, a nucleotide analogue binding to reverse transcriptase, has been proven effective in suppressing viral replication, improving histology and biochemical, and decreasing the inflammatory response. Lamivudine treatment has been shown to benefit patients with HBV-related decompensated cirrhosis and patients with ACLF in Hong Kong [17] and Taiwan [18]. However, the emergence of HBV mutant during the long-term lamivudine treatment hinders its clinical use. Enticavir is a carbocyclic analogue inhibiting HBV replication at three different steps [19]. Enticavir is more potent than lamivudine; it is effective against lamivudine-resistant mutants [20]. With the emergence of new, effective antiviral drugs with a high barrier to resistance, lamivudine is no longer recommended as a first-line treatment for chronic hepatitis B because of the high rate of drug resistance [3, 4], which prompted us to evaluate this agent in a series of patients with HBV-related ACLF.

Here, we conducted a retrospective study on 106 consecutive patients with ACLF. To the authors’ knowledge, this is one of the biggest single-center experiences of the use of antiviral treatment to date. The patients were informed by clinical physicians of the difference between nucleos(t)ide analogues and the importance of long-term treatment. The HBV DNA load was much higher in the ETV and LAM groups than in the non-NAs group (P = 0.036). After 1 month of treatment, HBV DNA decreased in all three ACLF groups, more patients achieved a >2 log10 IU/mL decrease in HBV DNA and an undetectable HBV DNA level in the ETV group than in the non-NAs group. ETV and ALM effectively decreased the HBV DNA load. It was not surprising that HBV DNA load decreases in NAs treatment group. The reason for the decreased HBV DNA load in the non-NAs group is partially explained by the continuously stimulated immune system in the process of clearing HBV and reducing the replication space (hepatocytes) in the setting of massive or sub-massive hepatic necrosis [21–23].

The term of ACLF was first used in 1995 to describe the condition in which two insults to liver are operating simultaneously: one of them being ongoing and chronic and the other being acute. ACLF was insult with the underlying cirrhotic or without (acute exacerbation). Patients with the underlying decompensated cirrhosis have a better prognosis than those with severe acute exacerbation of cirrhosis (in-hospital mortality about 75%) [24]. In our study, the 3-month survival rate in the ETV, LAM, and non-NAs groups was 66.7, 60.0, and 35.3%, respectively. The ETV and LAM groups showed a higher survival rate than the non-NAs group. However, Cui et al. [6] reported that the 3-month survival rates were 48.49, 50.00, and 40.54% in patients treated with ETV, LAM, and non-NAs, respectively, and that no significant difference in the 3-month survival rate was detected between the nucleoside analogue treatment (ETV + LAM) and non-NAs treatment, albeit a slight trend toward superiority of the nucleoside analogue treatment over the non-NAs treatment. The discordance may result from the relatively small sample sizes used in Cui’s and our studies. Relatively fewer patients with severe acute exacerbation were used in our study.

The high mortality in the non-NAs group is likely to be related to the failure of HBV eradication at an early stage. HBV continuously stimulates the immune system recruiting proinflammatory factor. Moreover, the elevated HBV DNA loads continuously activate liver NKT cells, which impair liver regeneration. A study from Germany shows that markers of liver injury were highly elevated in patients with HBV-related liver failure after only 1 week of enticavir treatment, leading to a significant decline in cell death and necrosis. We found that high total bilirubin, high serum creatinine, low CHOL, low PTA, non-nucleos(t)ide analogue treatment, and high MELD-Na score were significant predictors of early mortality, which is in consistent with a prospective study from Belgium [25].

A proportion of patients who received no nucleos(t)ide analogue treatment may repeat similar episodes of ACLF and have unfavorable prognosis even if they recover from the serious episodes. Nucleotide analogue treatment may bring long-term benefits with less recurrence. Clinicians should also be aware that HBV DNA is inhibited to an undetectable level but is not eliminated. Hepatitis flares may occur after withdrawal of treatment. Therefore, long-term treatment is recommended for patients with ACLF.

In conclusion, our study shows that short-term antiviral treatment with entecavir or lamivudine rapidly suppressed HBV replication, increased the short-term survival rate, and reduced recurrence. Early administration of long-term nucleoside analogue treatment is recommended for the prevention of recurrence of HBV-related ACLF. This study shows its weakness in that it is a retrospective review of a single-center experience of a small number of samples. A prospective, randomized, double-blind, placebo-controlled trial of NAs in ACLF patients is needed to validate the present results.