Premise

Hepatocellular carcinoma (HCC) is generally, but not always, seen on a background of cirrhosis. Since antiviral therapy with nucleoside(tide) analogues (NAs) slows or prevents cirrhosis and reduces the incidence of HCC, its benefits are obvious. Indeed, it is widely assumed that withholding antiviral therapy until the appearance of clinically active liver disease and fibrotic lesions is an adequate standard of care. Guidelines from international liver associations have mainly recommended that antiviral therapy be initiated if patients show signs of progressive liver disease. However, this approach has several problems. First, it requires monitoring of hepatitis B virus (HBV) carriers over many years to determine when to begin therapy. For cultural and practical reasons, many carriers who could be helped by current treatment guidelines are missed.1 Second, although strictly following treatment guidelines may delay the appearance of HCC, there is no proof that the delay is permanent.2 Third, the guidelines assume that HBV infection is largely harmless in HBV carriers until many decades have passed and active liver disease is readily apparent, an assumption for which there are no solid data1 and which, from experience with other chronic infections (eg, HIV), may be incorrect. Fourth, this assumption fails to consider extensive evidence from animal and human studies of other forms of chronic liver injury, including chemical carcinogenesis, which show that clonal hepatocyte repopulation is a major risk factor for HCC4 and therefore may be important in the aetiology of HBV-associated HCC, even in early stages of infection, before the appearance of raised alanine aminotransferase (ALT) activity and fibrotic or cirrhotic lesions. Clonal hepatocyte repopulation is a genetic narrowing of the hepatocyte population that occurs in response to either endogenous or exogenous injury, with sensitive hepatocytes dying and being replaced by clonal proliferation of rare but resistant hepatocytes that may evolve to form HCCs. Although not so far studied in the early stages of HBV infection, data on late stages of infection, with or without cirrhosis (an HCC risk factor), are most easily interpreted by assuming that clonal hepatocyte population also occurs in HBV carriers. Studies of clonal hepatocyte expansion in chimpanzees and humans with chronic hepatitis B support this idea, as do studies of woodchucks chronically infected with woodchuck hepatitis virus.5 In late stages of infection, cirrhosis is a major source: at least half of cirrhotic nodules are clonal (eg, Paradis et al6). This probably represents an adaptation of rare hepatocytes to survive when there are extreme alterations in hepatic architecture. We suggest that an unappreciated cause of clonal hepatocyte repopulation occurs in non-cirrhotic liver as well. Immune killing of infected hepatocytes is the strongest known pressure on the infected hepatocyte population in the non-cirrhotic liver and, analogous to cirrhosis, should lead to emergence of HBV-resistant hepatocytes that are able, in this example, to avoid immune killing. Indeed, most analyses of long-term carriers suggest that 50% or more of hepatocytes no longer support HBV infection and/or support much reduced levels of replication (see Mason et al7 for review)—that is, the chronically infected liver appears highly modified from the virus naïve liver. Therefore, although it may seem paradoxical at first glance, any reduction in HBV titres in HBV carriers may warrant initiation of antiviral therapy, even if biopsy does not reveal active hepatitis.

Background

Chronic HBV infection is typically divided into four sequential phases8 9: (1) an immunotolerance phase lasting two or three decades or more; (2) an immunoreactive phase with bouts of hepatitis over a period of months or years as the immune system attempts to clear the infection; (3) a low replicative phase in which very little virus is found in the serum (usually <10⁴ copies per ml or <2000 IU/ml) or liver, and the disease is generally inactive10; (4) sometimes a re-activation phase, in which virus titres rebound along with immune-mediated liver disease.

The immunoreactive phase typically leads to loss of hepatitis B e antigen (HBeAg), a marker of active replication, and the appearance of hepatitis B e antibody, defining the low replicative phase or inactive carrier state. However, in some patients, loss of HBeAg is due to a viral mutation(s), not complete control of virus replication, and liver disease may remain active, or may re-activate if the mutants emerge after seroconversion. Fibrosis and cirrhosis generally progress during the immunoreactive phase, although progression to cirrhosis is not inevitable.

The REVEAL-HBV studies10 on the natural history of HBV infection were carried out on patients 30 years of age or older, mostly in or past the immunoreactive phase, and showed a direct correlation between virus titres in the serum and risk of HCC. In particular, the lowest risk in a 12-year follow-up was seen in patients that had titres of <10⁴ per ml at study entry, with an increasing risk seen in patients with virus titres between 10⁵ and 10⁶ copies per ml and >10⁶ per ml. These are highly important studies and justify the need to reduce virus load in HBV carriers in order to reduce the risk of HCC. However, these results are often misunderstood as well, because virus titres above 10⁶ per ml were not quantified. Thus it needs to be remembered that the very high virus titres often seen in immunotolerant carriers (>10⁹ per ml) are not generally considered an HCC risk factor. Rather, it is virus titres that are >10⁴ but below some undefined upper limit that reflect HCC risk because they also reflect cumulative immune damage to the infected liver.11

Current treatment guidelines recommend delaying therapy until patients show clear signs of active liver disease extending over several months, including persistent increases in ALT activity and, when biopsy specimens are available, evidence of inflammation and/or fibrosis.9 12–15 These guidelines, rigorously applied, should identify patients entering the immunoreactive phase, when synergy with the host response can maximise therapeutic outcomes of antiviral therapy, hopefully with HBeAg and, ideally, HBsAg seroconversion. Application of the guidelines reduces the rate of progression to HCC, and can block the progression of fibrosis and cirrhosis.

Options for antiviral therapy of HBV carriers include interferon α (or pegylated interferon) and NA inhibitors of viral DNA synthesis. Interferon therapy enhances the antiviral immune response and may induce virus clearance, most effectively during the immunoreactive phase of infection.16 It is 2–3-fold less effective in the immunotolerance phase of infection.17 NAs may also assist virus loss during the immunoreactive phase, but will suppress virus replication at any stage of infection. An important placebo control trial was carried out with lamivudine, a cytosine analogue and the first NA approved by the Food and Drug Administration for treatment of HBV, to determine its efficacy on clinical end points. Patients enrolled for this study had significant liver disease, with an Ishak fibrosis score18 of 4 or greater on a scale of 6. A >50% reduction in liver disease progression including HCC was found after ∼36 months of therapy,19 at which point all patients were offered access to lamivudine. This study was a first proof of concept that antiviral therapy of chronic hepatitis B even at late stages can decrease the major complications of chronic infection.19 Other studies suggested a trend for a lower incidence of HCC in patients treated with lamivudine for chronic hepatitis compared with those treated at the stage of cirrhosis.20

It is important to remember, however, that the HCC incidence in these chronic hepatitis B patients treated with NAs was significantly decreased but not eliminated.2 21 A meta-analysis of 10 years worth of such data found that lamivudine therapy (in the absence of resistance) reduced the HCC risk only 2–4-fold.22 Many of these patients were well past the immunotolerance phase, and a strict adherence to treatment guidelines may have reduced the risk even further. However, strict adherence probably requires a level of clinical monitoring, public awareness and case ascertainment that will be difficult to achieve as long as young adults think they are not yet at risk and possibly in need of treatment to reduce the incidence of cirrhosis and HCC later in life. Another important concern is that HCC risk factors in HBV carriers are not well understood. Current thinking favours the notion that the HCC risk begins, in the vast majority of cases, with the immunoreactive phase, but there is no proof that it does not begin much earlier.

Nevertheless, there are several reasons why the guidelines have not been changed to include younger, immunotolerant, carriers. First, a major deficiency of lamivudine, the first and still widely used nucleoside, is a low-resistance barrier, up to 75% of patients developing resistant virus after 5 years of therapy.23 Treatment of immunotolerant patients with lamivudine is particularly ineffective, with emergence of drug-resistant HBV and the absence of HBeAg seroconversion.24 Second, even as NAs with better resistance profiles have become available, there has been a paucity of information on the possible long-term toxicity of combined therapy and a reluctance to combine drugs to which resistance may then develop, limiting the options for salvage therapy. The consequences of lifetime administration of the two NAs with the highest barrier to resistance, tenofovir (an adenosine analogue) and entecavir (a guanosine analogue), are not known. Safety has been excellent after 3 years of therapy with tenofovir25 and 5 years with entecavir.26 Third, patients and public health systems are reluctant to accept the financial burden and compliance needed to treat an infection, especially when it is still asymptomatic, although this has proven to be effective in the case of HIV infection. Fourth, not all patients will develop cirrhosis and HCC and it is not currently possible to determine, during the immunotolerant phase, the 25–50% who are at risk in the long term; thus, for half the patients, treatment can be avoided by following current recommendations, the caveat being that delay may considerably increase the risk of cirrhosis and/or HCC in the remainder who are at risk.

Fifth, and in our view the most important for those who might be inclined to seek medical advice, there is the belief that significant disease progression does not occur in the immunotolerant phase. This is true in the short term, by definition, if disease progression is defined by persistently raised ALT activity and/or biopsy-confirmed fibrosis or HCC. Liver biopsy is unlikely to take place without raised ALT activity, and ALT measurements can be unreliable in practice, especially in defining, in a typical hospital setting, values that are at the high end of normal or only slightly increased.27 Thus the problem of detecting early disease progression, and the additional perception that the vast majority of young carriers are not at risk, may lead to delayed diagnosis of many carriers until, or if, disease becomes symptomatic.1

Thus, from the physician's point of view, the absence of detectable liver disease in the immunotolerant phase (inflammation, fibrosis, cirrhosis) is taken as evidence that the infection is not currently harmful and therefore treatment is unnecessary. If the issue is progression to fibrosis and cirrhosis, then we agree that this interpretation makes sense. However, if the end point is HCC, the belief has no experimental support and, at best, can be considered an untested hypothesis. Furthermore, the REVEAL studies also showed that inactive carriers of HBV with viraemia levels <10 000 copies/ml and normal ALT activity still have a substantial risk of developing HCC,11 28 indicating that the absence of detectable liver disease by the currently recognised criteria does not preclude a risk of HCC development at the individual level.

Using chemical carcinogenesis as a model, current thinking focuses on promotion (liver regeneration), which would certainly be most evident although not exclusive to the immunoclearance phase of HBV infection, largely while ignoring initiation, which might occur at a significant rate at any time during the infection, including the immunotolerance phase. If the notion that the immunotolerance phase of infection is largely harmless is incorrect, many unnecessary deaths may occur that might be prevented by earlier antiviral intervention. Indeed, considering its long duration, initiation and promotion may both be significant during the immunotolerant phase, increasing the risk of HCC later in life even in the absence of cirrhosis.

Evidence for damage to the hepatocyte population during the immunotolerant phase of infection

Although quite rare, even young children may develop symptomatic liver disease, with raised ALT activity, chronic active hepatitis, cirrhosis and hepatocellular carcinoma. More commonly, infection is asymptomatic. Nonetheless, liver histology is not normal. In early studies it was found that the majority of asymptomatic immunotolerant children under the age of 10–12 have either minimal chronic hepatitis or, more commonly, non-specific reactive hepatitis,3 29 30 although ALT activity is typically in the normal range. A more recent longitudinal study of a small group (n=7) from ∼7 to ∼25 years of age revealed that ALT activity gradually increased starting around 15 years of age, reaching the high end of normal, if not above, by the final time point.31 Evidence for an increase in nucleotide diversity, particularly in HBV structural genes, was also detected between ∼15 and ∼25 years of age, suggesting immune pressure on the virus. It is also important to note that, in adults with persistently normal ALT activity, significant liver fibrosis was found in association with older age, leading to a recommendation for liver biopsy of patients older than 40 years.32 In addition, virus titres in immunotolerant patients are often considered high. As noted above, this is a view that can be misleading. The concept originated, in part, because their titres exceeded a lower cut-off (∼10⁴ copies per ml or 2000 IU/ml) below which active disease is not usually detected. However, when viewed against the hepatic potential for HBV production, titres are often unexpectedly low. The fully infected liver can yield 10⁹ to 10¹⁰ viruses per ml of serum, a level of production that would be expected to persist if infection were benign and the host were truly immunotolerant. Virus titres in adolescent and young adult carriers in the immunotolerant phase of infection tend to be lower, ranging from 10⁷ to 10⁹ copies per ml.31 33 34 ALT values at the high end of normal and ‘low virus’ titres (eg, <10⁸ per ml) would both point, independently, to accumulating hepatocyte damage (ie, changes in the hepatocyte population) during the immunotolerant phase of infection.

It is unlikely that declining virus titres during the immunotolerant phase are due directly to release of antiviral cytokines by cytotoxic T lymphocytes. An alternative is that a low but persistent immune destruction of infected hepatocytes, by low-level cytotoxic T lymphocyte infiltration, leads over time to an adaptive response of the liver, with the clonal emergence of hepatocytes that are resistant to HBV infection35 or produce reduced amounts of virus. This is suggested by many studies over the past 40 years, including studies of infected children, using HBcAg staining of liver biopsy tissue (eg, Hsu et al30). While these conclusions may reflect a limited sensitivity of the immunostaining for HBcAg in some studies,36 other studies examined the extent of hepatocyte infection in adult liver, generally from patients with active liver disease, cirrhosis and/or HCC.7 36–38 In these samples, it again seems that a substantial fraction of hepatocytes are not infected, or at least not producing significant amounts of viral nucleic acids or proteins, despite the fact that the patients are still viraemic. The idea that these virus-negative hepatocytes have a clonal origin is suggested by their appearance as foci in chimpanzees and woodchucks, in which the extent of hepatocyte infection was still high.35 39 The likelihood that this change occurs at the level of individual hepatocytes is supported by the observations that virus antigen negative hepatocytes may be adjacent to strongly positive cells. In addition, the relatively low viraemia typifying such patients, often >100-fold below the hepatic capacity, cannot be explained entirely by declines in the fraction of infected hepatocytes. A major additional factor may be reduced virus production by the hepatocytes that remain infected,38 which might include a number of adaptive factors including hyper-responsiveness to antiviral cytokines. It appears that these changes in the hepatocyte population may begin in the immunotolerant phase of infection, as suggested by the relatively low-level viraemia (eg, 10⁸ copies or less) in many immunotolerant children and young adults.

Finally, a considerable degree of clonal hepatocyte expansion is found in both non-cirrhotic40 and cirrhotic liver6 41 42 (T Tu, A R Jilbert et al (University of Adelaide), personal communication) of patients in the late stages of infection, as well as in chimpanzees chronically infected with HBV.39 Plausibly, in non-cirrhotic liver this may involve selective survival and emergence of mutant or epigenetically altered hepatocytes that are resistant to HBV infection, as well as hepatocytes that produce highly reduced amounts of virus.38 Clonal hepatocyte repopulation is an HCC risk factor.4 It is not known if this occurs during the immunotolerant phase, although we think any persistently low virus titres, in the immunotolerant phase, point in that direction.

These considerations, taken together, suggest to us that it is time to consider earlier treatment intervention in patients with chronic HBV infection. A change in treatment practices is much more feasible than it was even a few years ago, as much better drugs have become available with a better antiviral potency and a higher barrier to resistance. In our view, in an ideal world, it would be best to initiate NA treatment in all immunotolerant patients. This would also circumvent one major problem with the current guidelines, which require a high level of involvement of the asymptomatic patient with the clinician to determine when to begin therapy. However, a more conservative approach that is one step beyond the current guidelines would be to propose therapy in all patients with persistently high-normal ALT activity, or with normal ALT activity who show relatively low levels of viraemia (eg, >10⁴ but ≤10⁸ copies per ml), including patients in their 20s, not just those beyond 40 years of age.32 Where biopsy specimens are available, attempts should be made to establish hepatocyte infection levels. A significant fraction of HBcAg-negative hepatocytes and clustering of HBsAg-positive hepatocytes,43 together with virus levels of 10⁸ per ml or less (but still >10⁴), would suggest a high level of accumulated hepatocyte damage/change even in the absence of other indicators of histological change, and treatment would seem strongly warranted.

In conclusion, although more studies are required to better understand the immunotolerant phase of chronic HBV infection, there are already indications, from clinical studies, that the infection is not totally benign. Therefore earlier treatment intervention may be beneficial in preventing disease progression in the long term. While the benefits of early intervention may take a decade or more to evaluate, with a broader availability of NAs with a high barrier to resistance, it seems to be time to begin an approach that may benefit those at risk, before irreversible liver damage takes place. This would also provide an opportunity to search for factors (biomarkers) that identify immunotolerant carriers without a long-term risk of cirrhosis or HCC. Clinical studies addressing these questions seem to be highly desirable in the very near future.