Management of hepatitis B-related cirrhosis in the era of effective antiviral therapy

Chronic hepatitis B affects more than 248 million people worldwide and is one of the leading causes of cirrhosis and hepatocellular carcinoma (HCC), particularly in Asia and Africa.1, 2 In untreated patients with compensated cirrhosis, the annual incidence of HCC is as high as 2-3%.3 Nevertheless, a landmark clinical trial in patients with advanced fibrosis or cirrhosis confirmed that the first generation antiviral drug lamivudine could reduce the risk of disease progression and HCC.4 The finding was subsequently confirmed by several observational studies and meta-analyses.5 However, lamivudine is associated with a 70% risk of drug resistance and is no longer recommended. Current guidelines recommend entecavir and tenofovir as first-line agents due to their superior antiviral activity and high genetic barrier to resistance. Since then, a number of observational studies with untreated historical cohorts have shown that entecavir treatment is associated with a lower risk of HCC development in patients with cirrhosis.6-8 Another study also showed that tenofovir-treated patients had a lower incidence of HCC than was predicted by HCC risk scores.9

In this month's Liver International, Su et al.10 report the results of the Cirrhosis Taiwanese Entecavir Multicenter (C-TEAM) Study, which included 1315 cirrhotic chronic hepatitis B patients treated with entecavir for a median of 4 years. This was compared with an untreated historical cohort of 503 patients (median follow-up 6 years) at one of the participating centres. This is by far the largest cirrhotic treatment cohort. It could therefore provide a more precise estimate on the impact of entecavir treatment on HCC. Other cirrhotic complications could also be studied in detail.

Overall, HCC developed in 119 of 1315 (9.0%) entecavir-treated patients and 121 of 503 (24.1%) untreated patients (adjusted hazard ratio 0.40). The protective effect was seen across a variety of cirrhotic complications except hepatic encephalopathy. The adjusted hazard ratios for variceal bleeding, spontaneous bacterial peritonitis, liver-related mortality and all-cause mortality were 0.38, 0.06, 0.14 and 0.15 respectively. A similar trend was observed after propensity score matching. Among entecavir-treated cirrhotic patients, older age, male sex, positive hepatitis B e antigen and baseline alpha-foetoprotein ≥7 ng/mL predicted HCC development. In contrast, a serum hepatitis B virus (HBV) DNA level of <20 IU/mL at 1 year of entecavir treatment was associated with a 38% reduction in HCC incidence.

This large cohort study provides compelling evidence to support the efficacy of entecavir in reducing hard clinical endpoints in patients with hepatitis B-related cirrhosis. While it is true that patients in the entecavir cohort were older and had more advanced disease as evidenced by higher Child–Pugh score and total bilirubin and lower albumin and platelet count, those represent negative bias against entecavir treatment and do not affect the validity of the results. Like any studies involving historical cohorts, changes in clinical practice over time might have affected the results. For example, increased HCC and varices surveillance, better HCC treatment and primary and secondary prophylaxis against spontaneous bacterial peritonitis probably contributed to a lower all-cause and liver-related mortality in the later entecavir-treated cohort.

How would the latest evidence change our practice? Current guidelines already recommend antiviral therapy in patients with hepatitis B-related cirrhosis. The role of antiviral therapy in patients with HBV DNA <2000 IU/mL remains controversial, but this as well as previous studies have found incomplete viral suppression as a residual risk factor of HCC,7, 10 suggesting that low detectable HBV DNA level may still have an impact. In a retrospective cohort of 385 Korean patients with compensated cirrhosis, 8.8% of those with low detectable HBV DNA (<2000 IU/mL) developed HCC in 5 years, and antiviral therapy might still reduce HCC risk.11

What is more interesting, however, is how we should manage patients with hepatitis B-related cirrhosis in the era of effective antiviral therapy. In patients with compensated cirrhosis, the main management includes antiviral therapy, HCC surveillance by regular abdominal ultrasonography and varices screening. As the prescription of antiviral therapy is no longer controversial, the latter two deserve more discussion.

Although good quality trial data are lacking, numerous observational studies have shown that patients having regular HCC surveillance have smaller tumours that are more amenable to curative treatment.12 However, HCC surveillance is costly and labour-intensive and may not be affordable in high endemic areas. As antiviral therapy can reduce the risk of HCC and even reverse cirrhosis in up to 70% of patients with compensated cirrhosis,13 it is timely to ask if antiviral therapy can reduce HCC risk to such an extent that surveillance is no longer required. Unfortunately, in this study by Su and colleauges,10 the annual incidence of HCC remained significant at 2.4%, a level still warranting regular surveillance. Nevertheless, cirrhosis takes years to reverse. The duration of this study (4 years of entecavir treatment) is too short to answer this question. Future studies should revisit the incidence of HCC in patients who have been treated with antiviral drugs for a longer period.

A related question is how one may diagnose reversal of cirrhosis confidently. Liver biopsy is invasive and cannot be performed repeatedly outside research setting. Generic serum tests such as the aspartate aminotransferase (AST)-to-platelet ratio index (APRI) and the FIB-4 index are unreliable in patients on antiviral therapy and may miss the diagnosis of cirrhosis.14 For example, the APRI is calculated from AST and platelet count. The serum AST level can decrease rapidly with antiviral therapy, resulting in false impression of fibrosis regression. Liver stiffness measurement by transient elastography is another non-invasive test of fibrosis and has been shown to predict HCC development in a dose-dependent manner.15 Similar to serum tests, however, liver stiffness also decreases after antiviral therapy even in patients with static fibrosis stage.16 This is because hepatic necroinflammation also contributes to liver stiffness. Although data are lacking for other modalities, the same phenomenon should similarly affect other physical measurements of liver stiffness or elasticity such as acoustic radiation force impulse, shear wave elastography and magnetic resonance elastography.17 Before we can identify a fibrosis biomarker that is unaffected by hepatic necroinflammation, it is premature to discuss scaling down HCC surveillance programs.

Variceal bleeding is a complication of portal hypertension associated with high morbidity and mortality. The latest Baveno VI consensus guidelines recommend screening by upper gastrointestinal endoscopy in patients with a liver stiffness ≥20 kPa and/or a platelet count of less than 150 × 10⁹/L because those not fulfilling either criterion have a very low risk of harbouring large varices and developing variceal bleeding.18 Patients not requiring endoscopy should repeat liver stiffness measurement and clinical assessment annually. There are, however, reasons to believe that annual assessments are unnecessary in patients on hepatitis B treatment. In a prospective cohort study of 107 Italian patients with compensated hepatitis B-related cirrhosis, only one patient had progression of varices and six developed de novo varices over 12 years, among whom six were due to lamivudine resistance or HCC.19 If we believe that patients on antiviral therapy can achieve fibrosis regression and rarely have disease progression, perhaps a one-off assessment by liver stiffness measurement and platelet count would suffice. Low risk patients will probably continue to have low risk, and only HCC surveillance is needed. This notion deserves validation in prospective studies and can potentially simplify the management of numerous patients.

On the other hand, while the almost elimination of spontaneous bacterial peritonitis in entecavir-treated patients in this study by Su et al. is interesting,10 the small number of patients with decompensated cirrhosis precludes a reliable conclusion about this subgroup. Likewise, the lack of protection against hepatic encephalopathy is surprising. This may be attributable to different thresholds of diagnosis in the prospective and historical cohorts. Minimal and mild hepatic encephalopathy might have been recorded in the entecavir group but would be difficult to ascertain retrospectively. A more granular reporting of the severity of hepatic encephalopathy would be informative.

Merely 20 years ago, we were still dreaming of an antiviral drug for chronic hepatitis B that is orally administered, potent, free of side effects and having a high genetic barrier to resistance. Now that entecavir and tenofovir have become standard treatments, it is easy to forget the tremendous progress the field has made through the concerted efforts of many people. What we should not forget, however, is that the drugs are only good if they are taken by patients in need. In some parts of the world, stigmatization still prevents patients from seeking medical care, and misconceptions among doctors result in the use of suboptimal agents for suboptimal durations.20 It is heartening to learn from this study by Su and colleagues that the antiviral drugs are doing what we wished for.10 We should put that knowledge into practice.

Conflict of Interest

Vincent Wong has served as an advisor for AbbVie, Gilead Sciences, Janssen and Perspectum. He has also served as a speaker for AbbVie, Echosens, Gilead Sciences and Merck Sharp & Dohme.