Statistics from Altmetric.com
The evolution of liver disease caused by chronic viral hepatitis, in particular the development of cirrhosis and hepatocellular carcinoma (HCC), is slow. As a consequence, clinicians have sought short-term, surrogate markers of disease progression as measures of treatment response.
In chronic hepatitis B virus (HBV) infection, suppression of serum HBV DNA below limits of detection and/or hepatitis B ‘e’ antigen (HBeAg) seroconversion have become the immediate goals of therapy. However, the pursuit of these treatment goals should not detract from the real treatment aims—the prevention of premature mortality from liver disease and/or the cure of chronic infection indicated by permanent hepatitis B surface antigen (HBsAg) seroconversion. Lying between these extremes in the hierarchy of goals would be the prevention of histological or clinical disease progression and, for patients with advanced disease, the induction of disease regression.
As experience with nucleos(t)ide analogues (NAs) grows and more effective treatments become available, the attainment of these goals in the treatment hierarchy has been achieved at increasingly higher levels. With the current generation of antivirals, the suppression of viral load below the limit of detection after 1 year of treatment is achievable in the great majority of patients.1–3 HBeAg seroconversion may be achieved in 30% of HBeAg positive patients after 4 years of treatment.4 Increasingly, loss of HBsAg is observed among patients who start NA treatment in the HBeAg positive phase of disease.5 HBsAg loss has also been observed in a small number of patients with HBeAg negative disease after treatment withdrawal.6 7
There seems little doubt that current therapy for HBV infection reduces the risk of progression from chronic hepatitis to cirrhosis and decompensated liver disease. After 5 years of treatment with consistent suppression of HBV DNA, histological reversal of cirrhosis has been demonstrated.8 The question remains, however, whether treatment with NAs actually reduces the risk of premature mortality. Because of the time scales involved, there are few studies capable of addressing this question. An advantage has been clearly demonstrated in patients with decompensated cirrhosis due to HBV, who, untreated, have a 1-year survival rate of 15% but where treatment with NAs can restore liver function within 3–6 months.9 Reversal of severe histological injury such as cirrhosis certainly suggests that mortality from decompensation of end-stage chronic HBV may be reduced or at least delayed. However, whether treatment with NAs improves mortality through a reduction in the incidence of HCC remains to be seen.
In 2004, Liaw and colleagues published a prospective randomised, placebo-controlled trial of lamivudine in patients with chronic HBV infection and advanced fibrosis or cirrhosis. They reported that the rate of progression of liver disease in the lamivudine-treated group was reduced.10 The primary end point of this study was time to disease progression, where progression included development of any form of decompensation, development of HCC, or death. The trial was terminated early by the data safety and monitoring committee after a median of 32 months follow-up when the primary end point was achieved with p<0.001. At this stage, the cumulative incidence of HCC was 3.9% in the lamivudine-treated group and 7.4% in the placebo-treated controls, which, despite the large number of patients enrolled in the study, only just achieved statistical significance at p=0.047.
While it is probably true that many of us in the hepatology community accepted this study as evidence that NAs are effective at reducing the incidence of HCC in chronic HBV, clinical experience has undermined our confidence and led us to question the evidence because some of our patients still develop HCC while receiving treatment. In Gut (see page 1109), a paper by Papatheodoridis and colleagues reports a retrospective analysis of the incidence of HCC in a cohort of 818 Greek patients with chronic HBV infection treated with lamivudine with subsequent modification of drug treatment to achieve virological control, followed for a median of 4.7 years.11 The cumulative incidence of HCC at 5 years remains high (5.6%) with no apparent reduction in the annual incidence over time. The cumulative incidence of HCC was 9.2% in compensated cirrhotics and 19.3% in decompensated cirrhotics, which is comparable to the incidence expected in this group of patients based on published rates in untreated European patient groups.12 Additional concern is raised by the finding that virological control did not appear to impact on disease progression.
There are clearly a number of differences between the Liaw study and that currently published by Papatheodoridis and colleagues. These include the design of the study, age of patients, viral genotype, host genetic background, and environmental variables. In a recent meta-analysis of 21 studies in which the incidence rates of HCC were compared between patients treated with NAs and untreated patients, the authors concluded that treatment did indeed reduce the risk of developing HCC.13 However, most of the studies included were designed to answer a question other than whether NAs prevent HCC, and could therefore be regarded as methodologically flawed. Doubt over the efficacy of treatment to prevent the development of HCC has also been demonstrated in a meta-analysis of the effect of interferon (IFN)-based therapy. Here, significantly lower rates of hepatic events were reported among IFN-treated patients compared with controls, but most of the effect was in reducing the risk of decompensation rather than protection against the development of HCC.14 Although the conclusion is that IFN therapy does reduce the risk of HCC, there is significant heterogeneity between studies.
One possible explanation for the lack of therapeutic efficacy in the Greek cohort is the high proportion (>25%) of cirrhotics. It is possible that, in patients with cirrhosis, the molecular and cellular events in the oncogenic pathway are already too far advanced for antiviral therapy to influence the future development of HCC. However, in two previous studies on cirrhotic patient cohorts, antiviral therapy was shown to reduce the incidence of HCC.15 16
In order to demonstrate definitively that NAs are effective in reducing HCC risk, it would be necessary to conduct a long-term, placebo-controlled trial. Such a study would no longer be considered ethical. In clinical practice, it may take longer than the 5 years of the Papatheodoridis study for the full beneficial effects to be seen. Longer-term cohort studies are therefore required and reliable historical control data are needed for comparison. In the mean time, patients with chronic HBV infection will continue to require careful surveillance for HCC even when they are receiving antiviral therapy.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.