SEARCH

SEARCH BY CITATION

Keywords:

  • acute-on-chronic liver failure;
  • hepatitis B;
  • hepatitis B virus DNA;
  • liver failure;
  • Model of End-stage Liver Disease;
  • survival

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute severe exacerbation of CHB: natural history and predisposing factors
  5. Other causes of ACLF in CHB
  6. Pathogenesis of ACLF in CHB
  7. Prognostic indicators in ACLF in CHB
  8. Treatment of ACLF in CHB
  9. Conclusions
  10. References

Acute-on-chronic liver failure (ACLF) in chronic hepatitis B (CHB) is most commonly caused by acute severe exacerbation of CHB. The pathophysiology of ACLF in CHB is still poorly understood. Despite the identification of important predisposing factors and prognostic markers, ACLF in CHB remains a disease associated with high mortality. The majority of studies using nucleoside analog therapy did not show any significant improvement in survival, although larger prospective studies are needed. Liver transplantation is the definitive treatment for ACLF in CHB. The challenge ahead would be prognosticating cases with favorable or unfavorable outcomes in order to streamline patients for early transplantation or for medical therapy.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute severe exacerbation of CHB: natural history and predisposing factors
  5. Other causes of ACLF in CHB
  6. Pathogenesis of ACLF in CHB
  7. Prognostic indicators in ACLF in CHB
  8. Treatment of ACLF in CHB
  9. Conclusions
  10. References

The introduction of nucleoside/nucleotide analog therapy in the last 15 years has dramatically transformed the treatment outcomes of chronic hepatitis B (CHB).1 Long-term nucleoside analog therapy, in maintaining a permanent state of virologic suppression,2 can result in reversal of fibrosis, histologic improvement,3 as well as reduce the incidence of cirrhotic complications and hepatocellular carcinoma (HCC).4,5 However, an estimated 600 000 patients are still dying annually from CHB.6 Liver-related mortality in CHB can be classified into three groups: (i) complications of chronic liver disease with the development of cirrhosis or HCC; (ii) gradual hepatic decompensation with liver failure; and (iii) acute decompensation, otherwise known as acute-on-chronic liver failure (ACLF).

The Asian Pacific Association for the Study of the Liver (APASL) was the first liver organization to publish a consensus recommendation on ACLF.7 ACLF is defined as an acute hepatic insult manifesting as jaundice and coagulopathy (international normalized ratio [INR] > 1.5), complicated within 4 weeks by ascites and/or encephalopathy in a patient with previously diagnosed or undiagnosed chronic liver disease. ACLF in CHB includes acute severe exacerbation of CHB, a disease presentation distinct from other types of chronic liver diseases by its sudden virologic surge followed by biochemical flares. Its pathophysiology and management are also very different from other causes of acute liver failure.8 ACLF in CHB should also be differentiated from acute hepatitis B, as nucleoside analog therapy for the latter is generally only reserved for fulminant cases. The different causes of ACLF in CHB, especially acute exacerbation of CHB, its prognostic indicators, and its management strategies will be discussed in this review.

Acute severe exacerbation of CHB: natural history and predisposing factors

  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute severe exacerbation of CHB: natural history and predisposing factors
  5. Other causes of ACLF in CHB
  6. Pathogenesis of ACLF in CHB
  7. Prognostic indicators in ACLF in CHB
  8. Treatment of ACLF in CHB
  9. Conclusions
  10. References

CHB is distinguished by its four disease phases: the immune tolerant phase, the immune clearance phase, the inactive carrier phase and HBeAg-negative hepatitis.9 The immune tolerant phase is characterized by its high HBV DNA levels but minimal damage in liver histology.10 Acute severe exacerbation of CHB can occur in the immune clearance phase affecting 40–50% of hepatitis B e antigen (HBeAg)-positive patients,11 and can be prolonged when there is repeated unsuccessful clearance of HBeAg. Reactivation of CHB at the HBeAg-negative phase, seen in 15–30% of HBeAg-negative patients, can also result in acute decompensation, and is occasionally associated with HBeAg seroreversion.12 There is no single cut-off HBV DNA level that can differentiate whether HBeAg-negative patients would remain inactive carriers or develop exacerbations.13

One of the first prospective studies investigating the natural history of acute exacerbations in CHB found that 15–47% of CHB patients developed exacerbations after 4 years of follow up. In patients with hepatitic flares, up to 8% of patients would develop decompensation.14 When comparing HBeAg-positive to -negative patients, the frequency of severe exacerbation was similar.15

Several predisposing factors of acute severe exacerbation had been identified, including elevated alanine aminotransferase (ALT) levels at presentation, male sex,14 and the presence of certain genotypes (Table 1). There are eight genotypes (A–H) in CHB. Asian CHB patients are predominantly infected with genotypes B and C. Although genotype C patients have an increased chance of cirrhosis and HCC,16 genotype B is associated with a higher mortality in acute severe exacerbation.17 The more aggressive nature of genotype B has also been reproduced in other studies.18 Since genotype B patients achieve HBeAg seroconversion at an earlier age,19 genotype B is likely to be more immunogenic and is associated with more immune-mediated damage.20 In the Caucasian population, genotype D was found to be more aggressive than genotype A.21

Table 1.  Predisposing factors for acute-on-chronic liver failure in chronic hepatitis B
  1. Controversial factors denoted in italics.

  2. ALT, alanine aminotransferase.

Viral factors
• Genotypes B and D
• ? Precore and core promoter mutations
Host factors
• Elevated ALT at presentation
• Male sex
• ? Acute variceal bleeding
Infection
• Superimposed acute hepatitis A and E
• Systemic infections involving the liver
External agents
• Alcoholic hepatitis
• Systemic corticosteroids and its withdrawal
• Cytotoxic therapy (e.g. rituximab)
• Hepatotoxic drugs
• Herbal indigenous medicines

Mutations in the precore and core promoter region of the hepatitis B virus genome have been suggested to be associated with acute exacerbations. In a study in which the entire precore genome was sequenced, a point mutation at the A1896 loci was associated with the development of fulminant hepatitis.22 However, further clinical studies on precore mutations reveal conflicting results.23 Several mutation sites in the core promoter region have been described to be associated with fulminant disease.18 However a study in Asia found disease activity after HBeAg seroconversion to be unrelated to the presence or absence of both precore and core promoter mutations.24 It has been suggested that the presence of precore and core promoter mutations are more relevant in differentiating acute exacerbation of CHB from acute hepatitis B,25 although further validation of this proposal is needed.

Corticosteroid therapy and its withdrawal also predisposes to acute exacerbation of CHB.26 HBV replication is enhanced during steroid therapy, with exacerbations also possible during the immunologic rebound associated with steroid withdrawal. A study among 198 patients with either asthma or chronic obstructive pulmonary disease found the incidence of CHB reactivation to be significantly higher in patients receiving systemic steroids when compared to those receiving only inhalational steroids.27 A continuous period of systemic steroids for more than 3 months and a dose of prednisolone of more than 20 mg per day were associated with an increased chance of exacerbation.

A unique presentation of acute severe exacerbation is the reactivation of “occult” CHB after immunosuppressive or cytotoxic therapy,28 especially with therapy containing rituximab.29 Mortalities had been reported when pre-emptive antiviral or close monitoring were not enforced in hepatitis B surface antigen (HBsAg)-negative but antibody to the hepatitis B core antigen (anti-HBc)-positive patients in regions with a high prevalence of CHB.

Other causes of ACLF in CHB

  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute severe exacerbation of CHB: natural history and predisposing factors
  5. Other causes of ACLF in CHB
  6. Pathogenesis of ACLF in CHB
  7. Prognostic indicators in ACLF in CHB
  8. Treatment of ACLF in CHB
  9. Conclusions
  10. References

Coexisting viral hepatitis is one important cause of ACLF in CHB. Superinfection by the hepatitis A virus (HAV) in CHB has been associated with a 5.6–29-fold increase in mortality.30 Similar findings had been observed in CHB patients infected with acute hepatitis E, contributing to 20% of acute exacerbations of CHB in regions endemic with the hepatitis E virus (HEV).31 HEV superinfection in CHB has been found to be associated with a higher mortality when compared to HAV.32 In addition, patients with compensated cirrhosis infected with HEV are at an increased risk of rapid decompensation.33

The inclusion of acute variceal bleeding as a cause of ACLF is controversial. A hemodynamic study in 144 ACLF patients found the mean hepatic venous pressure gradient (HVPG) in ACLF patients with large varices to be comparable to that of decompensated cirrhosis.34 It is possible that variceal bleeding is only an expression of elevated portal pressure due to decompensation, and not the actual event leading to ACLF.7

Other precipitants of ACLF in CHB include systemic infections involving the liver, coexisting alcoholic hepatitis, hepatotoxic drugs and herbal indigenous medicines7 (Table 1). Several herbs, including Ganoderma lucidum or lingzhi,35 have been identified as hepatotoxic. Studies in Asia found herbal indigenous medicines attributing 7.1–55% of cases diagnosed with drug-induced liver injury,36 of which patients with underlying CHB are prone to develop unfavorable outcomes.37

Pathogenesis of ACLF in CHB

  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute severe exacerbation of CHB: natural history and predisposing factors
  5. Other causes of ACLF in CHB
  6. Pathogenesis of ACLF in CHB
  7. Prognostic indicators in ACLF in CHB
  8. Treatment of ACLF in CHB
  9. Conclusions
  10. References

HBV itself is non-cytopathic under the usual circumstances, with damage to the liver mainly caused by immune-mediated processes involving both CD4 + and CD8 + T cells.38 The inflammatory processes seen in ACLF are probably non-HBV-specific, since HBV-specific CD8 + T cells are found in liver tissues with minimal histologic changes, while non-antigen-specific T cells are found in CHB histologic specimens with massive liver necroinflammation.38

The immunologic changes seen in the inflammatory process of ACLF are very similar to that of severe sepsis.39 The host immune system is affected by a depression of different cellular immune parameters, especially monocytic HLA-DR. The reduced monocyte activation is associated with an exacerbated production of anti-inflammatory cytokines, including interleukin (IL)-6, IL-10, IL-12 and interferon-gamma.39 Cytokine production is associated with activation of toll-like receptors (TLR), an important component of the innate immune system in recognizing specific pathogen-associated molecular patterns (PAMP).40 TLR2 and TLR4,41 downregulated during chronic HBV infection, show an increased expression during ACLF. Other TLR (TLR3/5/7/9/10) are also upregulated in ACLF.42

A recent study found IL-17-producing CD4 + T (Th17) cells to be involved in liver injury, particularly in CHB patients with ACLF.43 The precise mechanisms remain to be determined, although a preliminary study found an increased frequency of CpG island methylation of IL-17 in CHB patients with ACLF.44 Th-17 also has a role in liver injury in alcoholic liver disease.45 It is therefore possible that TH-17 is not specifically involved in CHB but is also involved in ACLF of different etiologies. B cells may also be involved, as shown in a study demonstrating a massive accumulation of plasma cells secreting immunoglobulin (Ig)M and IgG targeting the hepatitis B core antigen in liver tissues of CHB patients who required liver transplantation because of ACLF.46

The I27V mutation in the HBV core gene is also found to be associated with severe liver inflammation in CHB, and has a high rate of occurrence in Chinese CHB patients.47 Peripheral glucocorticoid receptor expression is also increased in CHB patients with ACLF.48 These factors may contribute to the pathogenesis of ACLF in CHB, although their exact roles remain to be determined.

The clinical development of systemic inflammatory response syndrome (SIRS) is a poor prognostic factor, and is associated with progression of encephalopathy and reduced chances of transplantation.49 In its full-blown picture, the resulting inflammatory responses in the liver and its associated cellular immune dysfunction can result in renal, circulatory and cerebral dysfunction (Fig. 1). Kidney damage in ACLF is multi-factorial, including hepatorenal syndrome, hypovolemia, infective causes and drug-related causes.50 The increased secretion of nitric oxide and tumor necrosis factor results in exaggerated vasodilatation.51 This circulatory dysfunction, together with hyperammonemia, can be associated with the elevation of intracranial pressure and cerebral edema.52 Hence, ACLF if uncontrolled could result in multi-organ failure and hence mortality.

image

Figure 1. Schematic diagram summarizing different clinical events in acute-on-chronic liver failure in chronic hepatitis B.

Download figure to PowerPoint

Prognostic indicators in ACLF in CHB

  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute severe exacerbation of CHB: natural history and predisposing factors
  5. Other causes of ACLF in CHB
  6. Pathogenesis of ACLF in CHB
  7. Prognostic indicators in ACLF in CHB
  8. Treatment of ACLF in CHB
  9. Conclusions
  10. References

The prognosis of ACLF in CHB is extremely poor, with 3-month mortality rates without liver transplantation reported to be more than 50%.53,54 An Asian study of 47 patients with severe exacerbation of CHB identified several factors associated with adverse outcomes,15 including pre-existing cirrhosis, prolonged prothrombin time (PT), elevated bilirubin, low albumin level, low platelet count and the presence of encephalopathy/ascites. The presence of neither precore nor core promoter mutations was associated with a worse prognosis. Other studies from Asia also show similar findings.55,56 These studies conclude that ALT level is not a useful prognostic indicator of mortality.

Different scoring models had been established to stratify patients based on prognosis to facilitate organ allocation for liver transplantation. The Model of End-stage Liver Disease (MELD) score has been used extensively for the allocation of donor livers worldwide,57 and has been validated for the use in CHB.58 A study comparing the MELD score with the Child–Pugh score found the MELD score to be a more objective and efficient system in predictive survival in CHB patients with ACLF.59 A subsequent increase in MELD score after initial presentation of ACLF is also useful in stratifying patients for liver transplantation.60 Serum sodium levels have been previously found to be associated with increased mortality in patients awaiting liver transplantation,61 and incorporating serum sodium into the MELD score is shown to improve the mortality prediction.62 The modified MELD-sodium score has been validated as a predictor of mortality in CHB.63 Further studies are needed to determine the MELD-sodium score in the setting of ACLF in CHB.

Other models used for predicting prognosis of ACLF in CHB have been reported (Table 2). One regression model, using the presence of hepatorenal syndrome, liver cirrhosis, positive HBeAg, low albumin and prolonged PT, was found to be superior to the MELD score in predicting 3-month mortality.53 Another model based on the presence of hepatic encephalopathy, hepatorenal syndrome, positive HBeAg, liver cirrhosis and prolonged PT was also found to be superior to both the MELD and Child–Pugh score.64 Further studies to externally validate these models would be needed before widespread use could be recommended. The King's College Criteria have been used extensively for prognostic purposes in acute liver failure,8 but have never been validated for the use of ACLF in CHB.

Table 2.  Different predictive models used in prognosticating acute-on-chronic liver failure in chronic hepatitis B
Predictive modelComponentsComments
  1. ACLF, acute-on-chronic liver failure; HBeAg, hepatitis B e antigen; MELD, Model of End-stage Liver Disease.

MELD score57–60Prothrombin timeMore objective score when compared to Child–Pugh score
Bilirubin
Creatinine
MELD-sodium score63Prothrombin timePredictor of mortality in chronic hepatitis B
BilirubinNeeds further studies in the setting of ACLF
Creatinine
Sodium
Sun et al. 200953Presence of cirrhosisFound superior to MELD when predicting 3-month mortality
Presence of hepatorenal syndrome
HBeAg statusNeeds external validation
Prothrombin time
Albumin
Zheng et al. 201164Presence of cirrhosisFound superior to MELD score and Sun et al. 2009 in predicting 3-month mortality
Presence of encephalopathy
Presence of hepatorenal syndromeNeeds external validation
HBeAg status
Prothrombin time

Treatment of ACLF in CHB

  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute severe exacerbation of CHB: natural history and predisposing factors
  5. Other causes of ACLF in CHB
  6. Pathogenesis of ACLF in CHB
  7. Prognostic indicators in ACLF in CHB
  8. Treatment of ACLF in CHB
  9. Conclusions
  10. References

All supportive measures recommended for other causes of acute liver failure hold true for CHB with ACLF,8 including antibiotics for superimposed infection, treatment for hepatic encephalopathy,52 and terlipressin and albumin for hepatorenal syndrome.65 The preliminary results of a randomized placebo-controlled trial found the administration of granulocyte-colony stimulating factor improved survival after 2 months,66 although this cannot be recommended for routine clinical practice at this juncture. Series with a small number of patients have found the use of bioartificial liver support systems to be safe and well tolerated in ACLF, and are able to bridge patients while waiting for liver transplantation or till recovery.67 However the preliminary results of a randomized controlled multicenter study in ACLF patients failed to identify any survival benefit.68

Interferon therapy is contraindicated in CHB with ACLF as it will exacerbate hepatic decompensation. Corticosteroids, based on their anti-inflammatory activity, have been used in CHB with ACLF.69 In a recent study, intravenous dexamethasone 10 mg daily was given for 5 days together with continuous lamivudine to 56 patients. When compared with controls, dexamethasone treatment was an independent factor influencing survival, with a rapid decline in serum bilirubin in the first 5 days predictive of survival.70 More studies would be needed to validate this result, especially since steroid withdrawal is an important predisposing factor of ACLF in CHB.

Medical literature on the efficacy of nucleoside analogs, while scarce, has been emerging in recent years (Table 3). Long-term lamivudine treatment in CHB is associated with improved biochemical and virologic outcomes.4,5 This also holds true for patients with severe acute exacerbation.74 However, the use of lamivudine in CHB with ACLF has been disappointing. A Japanese study compared lamivudine-treated patients with acute severe exacerbation with historical controls. Among those developing liver failure, although there was a reduction in viral load in lamivudine-treated patients, there was no difference in the incidence of 3-month mortality.56 Although a more recent study found a survival benefit in lamivudine-treated patients when compared to controls in patients with a MELD score of 30 or less,71 those treated with lamivudine still had a 3-month mortality of 50.7%. This study also identified a low pretreatment HBV DNA and a rapid decline in viral load as predictors of good outcome.

Table 3.  Efficacy of nucleoside analogs in acute-on-chronic liver failure in chronic hepatitis B
StudyTsubota et al. 200556Sun et al. 201071Cui et al. 201054Wong et al. 201172Garg et al. 201173
  • Study recruited 50 patients with acute severe exacerbation, of which 13 developed hepatic failure.

  • ETV, entecavir; HBV, hepatitis B virus; INR, international normalized ratio; LAM, lamivudine; MELD, model of end-stage liver disease; TDF, tenofovir.

LocationJapanChinaChinaHong KongIndia
DesignRetrospectiveRetrospectiveRetrospectiveProspectiveProspective
Patient cohortLAM (n = 6)LAM (n = 130)LAM (n = 34)LAM (n = 117)TDF (n = 14)
Controls (n = 7)Controls (n = 130)ETV (n = 33)ETV (n = 36)Placebo (n = 13)
Controls (n = 37)
Compensated Cirrhosis38.5% (5/13)10% (26/260)LAM: 21%
ETV: 14%
INRMedian: 4.42–4.92Mean: 2.1–3.6Mean: 2.27–2.61Mean: 1.58–1.59Median: 1.85–1.93
HBV DNA (IU/mL)Median: 8 × 107 (range: 4 × 106–7.6 × 108)56 patients < 20 000Mean: 1.1–1.7 × 105Mean: 3.9–7.3 × 106Median: 7.5 × 105–1.7 × 106 (range: 1.7 × 104–1.7 × 109)
204 patients ≥ 20 000
Mortality3-month:3-month:3-month30-day:3-month:
LAM: 50%MELD 20–30:LAM: 50%LAM: 2%TDF: 43%
Controls: 71.4%LAM: 50.7%ETV: 51.5%ETV: 11%Placebo: 85%
Controls: 75.7%Controls: 59.5%48-week:
MELD > 30:LAM: 4%
LAM: 98.0%ETV: 19%
Controls: 100%
ConclusionNo benefit with LAMLAM improves survival for MELD 20–30. No survival benefit for MELD > 30No benefit with LAM or ETVETV treatment associated with increased short-term mortalityTDF improves survival
CommentsLow pretreatment HBV DNA and rapid decline in HBV DNA predictors of favorable outcomeSurvivors had favorable long-term outcomesSmall patient number in ETV groupWeek 2 HBV DNA > 2-log reduction predictive of survival

The survival rates from the above two studies are much lower than that of lamivudine in decompensated acute hepatitis B.75 A possible explanation is the presence of substantial hepatic necrosis by the time lamivudine is started in ACLF, with synthetic function not correctable by a reduction in viral load. This hypothesis is consistent with a study showing that commencement of lamivudine before serum bilirubin reached 20 mg/dL was able to improve survival in cases of severe acute exacerbation.55

More recent studies have been performed with more potent nucleoside analogs. A study from China compared patients receiving entecavir (n = 33), lamivudine (n = 34), or no nucleoside analogs (n = 37).54 As with previous studies, both nucleoside analogs achieved significant viral suppression after 3 months, but there were no improvements in survival, with 3-month survival rates being 48.5%, 50.0% and 40.5%, respectively. A recent study from Hong Kong found entecavir (n = 36) to be associated with increased short-term mortality when compared to lamivudine (n = 117),72 of which lactic acidosis has been hypothesized as a possible cause.76 This finding needs further confirmation from other centers, especially since another study found the usage of entecavir and tenofovir (another potent nucleoside analog) in decompensated CHB to have similar outcomes.77

A prospective study from India found tenofovir to be associated with improved survival in ACLF. This study randomized patients into taking either tenofovir (n = 14) or placebo (n = 13).73 After 3 months, the tenofovir group had a survival rate of 57%, compared to 15% for placebo (P = 0.03). Another significant finding was a 2-log reduction in HBV DNA levels in 2 weeks indicating better survival. However, this study is limited by its small sample size and the unavailability of liver transplantation. Despite having a lower median INR of 1.85 at treatment commencement, the survival rate of the tenofovir-treated group was not better than that of the studies quoted above54,56,71 (Table 3).

There are two limiting factors for all nucleoside analog studies. First, the number of patients in these studies was very small. Second, there were no head-to-head comparisons between patients receiving different nucleoside analogs. The vast difference in 3-month survival rates (ranging between zero and 40.5%) found in controls suggested that different studies recruited patients with different severity of ACLF. A meta-analysis of existing studies would be useful. In addition, larger prospective studies are needed to validate the clinical efficacy of tenofovir, to clarify the survival.

The definitive treatment for ACLF in CHB is liver transplantation. A study from Hong Kong retrospectively analyzed 149 patients presenting as ACLF who received liver transplantation, of which 50 had severe exacerbation of CHB.78 The 5-year survival rates exceeded 90%. The application of different predictive models mentioned above should help in the selection of patients for liver transplantation, although further studies in this area are required. Nucleoside analog therapy after liver transplantation had been shown to be effective in achieving good long-term outcomes.79 Randomized trials comparing nucleoside analog therapy with hepatitis B immunoglobulin found the former to be convenient and cost-effective for long-term HBV prophylaxis.80–82

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute severe exacerbation of CHB: natural history and predisposing factors
  5. Other causes of ACLF in CHB
  6. Pathogenesis of ACLF in CHB
  7. Prognostic indicators in ACLF in CHB
  8. Treatment of ACLF in CHB
  9. Conclusions
  10. References

The clinical and immunological features of ACLF in CHB are distinctive from other causes of ACLF. Unlike the majority of the CHB population, patients with ACLF belong to a subgroup that does not usually respond satisfactorily to nucleoside analogs. Determining the efficacy of nucleoside analog therapy would need larger prospective studies, and this would probably require multicenter collaborative efforts to ensure patient recruitment. Liver transplantation is still considered a luxury in many developing countries, and hence it is essential to develop prognostic indicators to identify patients with poor outcomes. The clinical challenge in the future is how to incorporate different predictive models into daily clinical practice and streamline patients for either liver transplantation or medical management.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute severe exacerbation of CHB: natural history and predisposing factors
  5. Other causes of ACLF in CHB
  6. Pathogenesis of ACLF in CHB
  7. Prognostic indicators in ACLF in CHB
  8. Treatment of ACLF in CHB
  9. Conclusions
  10. References