Viral hepatitis and hepatocellular carcinoma

Hepatitis B virus (HBV) infection is a major public health problem. It is estimated that two billion people have been infected worldwide and 360 million suffer from chronic HBV infection [12]. Over 520,000 die each year, 50,000 from acute hepatitis B, 470,000 from cirrhosis and liver cancer. In South-East Asia hepatitis B is mostly acquired perinatally from an infected mother. In sub-Saharan Africa, it is mostly acquired in early childhood by horizontal infection, whereas in Northwestern Europe, North America and Australia infection is mainly through sexual contact or needle sharing among injecting drug users, with a peak incidence in the 15–25 age group [12]. Infection acquired perinatally and in early childhood is usually asymptomatic, becoming chronic in 90 and 30% of cases, respectively. In adults, infection resolves in >95% with loss of serum HBsAg and the appearance of anti-HBs. Chronic infection is characterized by the persistence of HBsAg for more than 6 months. Acute hepatitis B usually results in complete recovery with little if any risk of HCC. In cases with persistent HBV infection, HBV is one of the most important risk factors for HCC.

Chronic HBV infection presents as one of three potentially successive phases: immune tolerant, immune active and low- or non-replicative. In the immune tolerant phase, serum HBsAg and HBeAg are detectable, serum HBV DNA levels are high, serum aminotransferases are normal or minimally elevated. In the immune active phase, serum HBV DNA levels decrease and serum aminotransferase levels increase. Flares of aminotransferases may be observed, in some patients these flares are followed by HBeAg-anti-HBe seroconversion. Following this conversion, in the low- or non-replicative phase the HBV replication persists but at a very low level suppressed by the host immune response. HBV DNA in serum is undetectable by conventional, non-PCR based techniques. This phase is also called the 'inactive carrier state'. It may lead to resolution of HBV infection where HBsAg becomes undetectable and anti-HBs is detected, anti-HBc staying positive as sign of contact with the virus. Recently it has been reported that HBV DNA can persist in the serum and liver tissue even after negativation of HBsAg [13]. Recent advances in molecular technology have allowed the isolation of HBV variants that either cannot produce HBeAg or produce it less efficiently, based on precore stop codon mutation and mutations in the core promoter region respectively. In patients with HBV variants, progressive liver damage occurs in parallel with relatively high levels of viremia. In perinatally infected people, the immunotolerant phase lasts till the age of 15–35 years, after which hepatitis flares may occur, leading eventually to viral remission. In patients infected during later childhood or adulthood, there is no immunotolerant phase.

Most studies on the risk of developing HCC in chronic HBV infection have been performed in the Far East. Here, most patients acquired the HBV infection as newborn infants [14]. It has been noted that the probability of acquiring HCC increases with severity of liver disease. The annual risk of HCC is 0.5% for asymptomatic HBsAg carriers and 0.8% for patients with chronic hepatitis B [15]. Patients with HBV-cirrhosis have a 1000 times higher risk of developing HCC compared to a HBsAg negative control group [16]. The incidence of HCC in compensated cirrhosis due to HBV from Asia was 2.7%. In Japan, the mean interval between the time of initial infection with HBV and the occurrence of HCC is 50 years. As most people here are infected at birth, HBV related liver cirrhosis usually develops in patients in their 40's and HCC in their 50's [17].

Few adequate studies have been performed in the West to address the issue of the incidence of HCC in persons who are positive for HBsAg. Most of the studies in Western countries have included small numbers of HBsAg positive patients and/or have not specifically analyzed the group of HBsAg carriers. There is also lack of uniformity in the timing of initiation of follow-up monitoring. In a cohort of 350 Western European patients with compensated cirrhosis followed for a mean period of 6 years, the 5-year cumulative incidence of HCC was 6% [18, 19]. The incidence was 2.2% in a series of 179 untreated Caucasian patients [19, 20]. In a retrospective analysis of cirrhotic European patients with HBV infection, the 5-year incidence of HCC was 9% irrespective of HBeAg or HBV DNA status at the time of diagnosis of cirrhosis [21].

The hepatitis B replication status seems to play an important role in determining the risk of development of HCC [22‐24]. A recent study found that whereas the relative risk of HCC among men with HBsAg alone was 9.6 compared to those without HBsAg, the risk increased to 60.2 when they were positive for both HBsAg and HBeAg [23]. Another analysis showed that the level of HBV DNA is a prognostic marker for HBV-related HCC and that HCC patients with a less favorable course appear to either clear the virus poorly or to have a greater level of virus production [24]. It was recently demonstrated that positivity for anti-HBc alone in absence of HBsAg and anti-HCV is not rare in Japanese patients with HCC, which may indicate that HBV virus might be involved in so-called non-B HCC [25].

The entire nucleotide sequences of HBV genomes have been classified into 8 genotypes (A-H), with predominance of genotypes A and D in Western countries, and B and C in Southeast Asia and the Far East [26‐29]. Several studies from the Far East evaluated the association between distinct genotypes and severity of liver disease. Genotype C was shown to be associated with the development of liver cirrhosis and HCC in Taiwan [30], China [31] and Japan [32], whereas genotype B was shown rarely to be associated with the development of HCC in China and Japan. In contrast, in Taiwan genotype B is the predominant type in patients with HCC who are younger than 35 years [30]. Another study from Taiwan showed that patients with genotype C had a greater tumor recurrence rate after curative resection of HCC compared with those with genotype B [33]. It was also shown that the likelihood of presence of T1762/A1764 mutations in the basal core promoter parallels the progression of liver disease, and that this mutation is found more frequent in HBV genotype C than B patients [34]. PreS deletions were shown to be more frequent in patients with HBV genotype C, and associated with more advanced disease such as liver cirrhosis and hepatocellular carcinoma [35].

Hepatitis C is also a major public health problem. There are more than 170 million people infected worldwide [36]. Approximately 80% of HCV infected patients develop chronic hepatitis C. About 20% of these patients will develop severe chronic hepatitis C and cirrhosis, which becomes detectable in the second and third decade after infection. The natural history of chronic hepatitis C infection is characterized by a predominantly asymptomatic course and a variable clinical outcome. For these reasons it is difficult to define the rate of progression to cirrhosis and HCC. The risk of cirrhosis in chronic hepatitis C is less than 10% in women infected at a young age and >30% in men infected after the age of 40 over a 20 year period [37, 38]. Five prospective studies from Europe and the US have shown that during the first 10–15 years after initial infection, liver cancer is a rare occurrence [39‐43]. In patients with hepatitis C, there is an increased risk of HCC coinciding with the establishment of cirrhosis with yearly incidence between 3–8% [6, 7, 44‐47]. In Japan, the mean interval between infection and development of HCC is 30 years [48]. A study from the US shows a long time lag (mean 28 years, range 8–42) between transfusion-associated hepatitis and development of HCC [49].

There is conflicting information on the relationship between HCV genotype and progression to HCC in longitudinal studies. It is suggested by some authors that genotype 1b (most prevalent in Europe and Japan) is associated with a higher incidence of HCC than infection with other genotypes [50, 51]. In other studies, however, this was not observed [52, 53].

Both HBV and HCV are transmitted parenterally and coinfection is not uncommon in intravenous drug users and in countries with a high prevalence of HBV [54]. Coinfection of HBV and HCV seems to result in more severe liver disease than either infection alone [55]. The risk of developing HCC in subjects with both infections has been investigated in a meta-analysis of 32 epidemiological studies between 1993 and 1997 [56]. The odds ratio for development of HCC in HBsAg positive, anti-HCV/HCV RNA negative subjects was 20.4; in HBsAg negative, anti-HCV/HCV RNA positive subjects 23.6; and subjects positive for both markers 135. These data suggest a more than additive but less than multiplicative effect of HBV and HCV coinfection on the relative risk for HCC. The viruses may act through common as well as different pathways in the carcinogenic process.

It has been reported that HBV DNA is still present after seroconversion of HBsAg in patients with hepatitis B. Several reports indicate that prior HBV infection, characterized by presence of anti-HBc, affects the development of HCC in patients infected with HCV [57‐59]. Given these data, in patients with chronic HCV infection, serologic markers of past HBV infection should be checked, not just HBsAg. Other authors, however, were not able to document any adverse event of occult HBV infection on the clinicopathologic course of chronic HCV infection [60].

In case of coinfection with HBV (whether active or past), a more aggressive surveillance to detect early HCC could be suggested [61]. However, to date screening and surveillance programs have not demonstrated a significant survival benefit.

In view of the role of HBV as cofactor in the development of HCV related cirrhosis and HCC, vaccination of patients with chronic hepatitis C against HBV has been advocated with the presumption of avoiding additional liver injury [62, 63].

Verme et al [11] suggested that HBsAg positive patients with HDV superinfection develop cirrhosis and HCC at an earlier stage (mean age 48 year) than HBsAg carriers without HDV infection (mean age 62 years).

Coinfection of HBV and HCV with HIV is common because these diseases share the same routes of transmission. Recently a series of HCC in HIV-HCV coinfected patients was published, indicating an unusually rapid development of HCC in these patients [64]. This is not surprising, as chronic hepatitis C is more aggressive in HIV positive subjects, leading to cirrhosis and end-stage liver disease in a shorter period of time [65].

An Egyptian study showed that Schistosoma infection increased the risk of HCC, only in the presence of HCV, whereas isolated S. mansoni infection does not [66].

Reports suggest that HBV and ethanol act synergistically to promote HCC [67, 68]. Habitual heavy drinking was reported to be a significant risk factor for HCC in patients with HCV-related liver cirrhosis by multiple logistic regression analysis [57]. A recent study showed synergism between alcohol drinking and HBV or HCV infection, with approximately a twofold increase in the odds ratio for each hepatitis virus infection for drinkers' > 60 g/d, with a more than additive but less than multiplicative risk [69]. Although two case-control studies did not show a relationship of alcohol consumption with the occurrence of HCC [70, 71], another case-control study found a positive interaction between HBsAg positivity and HCV RNA positivity and heavy alcohol intake in the development of HCC [72]. Furthermore, Hassan et al. [73] showed synergistic interaction (more than additive) between heavy alcohol consumption ≥ 80 ml/d and chronic HBV or HCV infection (odds ratio 53.9) and insulin or non-insulin dependent diabetes mellitus (odds ratio 9.9).

A significant proportion of HBV-related HCCs arise in an otherwise normal liver, implicating that the virus can also be directly oncogenic [124].

It has been demonstrated that HBV integrates into the DNA of the host cells. This integration may dysregulate the control mechanisms on the cell cycle by chromosomal abnormalities, production of viral proteins or alteration of human genes and proto-oncogenes. It is, however, controversial whether viral integration plays an important role in the process leading to development of HCC. The hepadnaviral integration process appears to involve recombination mechanisms that do not preserve the viral genome sequence. Thus it is impossible for the viral integrant to function as a template for subsequent virus replication. Several studies suggest that DNA integration sites are at random and that integration occurs at random times during the course of a chronic viral infection [130, 131]. HBV integration can be present in chronically infected liver tissue without evidence of HCC [132]. Non-neoplastic hepatocytes may have a similar pattern of rearrangement of viral sequences following integration into human DNA.

HBV DNA may induce malignant transformation in another way.

Mammalian hepadnaviruses contain a gene (the HBX gene), of which the protein (HBX protein) can trans-activate several cellular promotors and upregulate their expression of different cellular and viral genes [137, 138]. Integrated HBX, even when truncated, frequently encodes functionally active trans-activator proteins [139]. This protein has been shown to transform mouse fetal hepatocytes into a full malignant phenotype [140]. There are studies in transgenic mice with the HBX gene that developed multifocal areas of altered hepatocytes, adenomas and HCCs [110].

In contrast to mammalian hepadnaviruses associated with HCC, avian hepadnaviruses such as the duck hepatitis virus or heron hepatitis virus, lack the HBX gene and are not associated with HCC [123].

A gene that may be affected by the HBX gene is the p53 tumor suppression gene. This gene has been shown to play an important role in hepatocarcinogenesis. It is considered to negatively regulate the cell cycle. The HBX protein has been shown to complex p53 protein and to inhibit its function [141, 142]. In a transgenic mouse model it was shown that HCC development correlates with p53 binding to HBX [143].

It has been proposed that HBV acts as an insertion mutagen by integrating into the host genome and activating the cellular proto-oncogenes c-myc, ras and c-fos [144].

The preS2/S gene is integrated in most HCCs associated with HBV. When 3'-truncated it generates a truncated protein that is oncogenic by trans-activating proto-oncogenes c-myc and c-fos [145].

Growth factors and their receptors function as positive or negative modulators of cell proliferation and differentiation. Insulin-like growth factor-II and transforming growth factor-β expression correlate with HBX protein expression in animal models [146, 147], suggesting trans-activation of these growth factors facilitating tumor formation.

PreS deletion mutants accelerate the storage of large envelope proteins in hepatocyte cytoplasm, which could induce cytotoxic effects toward the development of end-stage liver disease [148]. The accumulation of large envelope protein can activate cellular promoters by inducing endoplasmic reticulum stress [149]. Furthermore, pre-S1 sequences can stimulate the transcription of transforming growth factor α (TGFα). Coexpression of TGFα and HBsAg could accelerate hepatocellular carcinogenesis by stimulation of hepatocyte proliferation [150].

Allelic loss of chromosome 4q is one of the most frequent genetic aberrations found in HCC. It was found to be associated with HBV-related hepatocarcinogenesis, probably by inactivation of a putative tumor suppressor gene included in it [151].

The most effective tool to prevent HCC is avoidance of the risk factors such as viral infection by HBV or HCV. Any action diminishing the potential transmission of contaminated blood products (uncontrolled blood transfusion, needle sharing, invasive procedures without proper health standards) will decrease the likelihood of viral spread.

The major advance has come from the availability of an effective vaccine that protects against HBV.

In 1969, Taiwan was an hyperendemic area of HBV infection with a high rate of HBsAg positivity, 19% of the population being infected before the fourth decade of life. In 1976, HBsAg prevalence was > 80% in HCC in Taiwan [167]. In 1984 a program to control cirrhosis and HCC began. All neonates born to HBsAg positive mothers were given hepatitis B vaccine in order to counter perinatal infection. In 1986 all neonates were included in the program. As a consequence, there was a decrease in HBsAg positivity in six-year-olds from 10.6% in 1983–1984 to 0.8% in 1993–1994. There was a parallel decline in incidence of childhood HCC (6–14 years old), in the cohort born between 1980 and 1984. The incidence of liver cancer in children between 6 and 14 years old decreased to zero for children born in 1986 and 1987 [168]. The decline of HCC in children after universal vaccination can be considered as an early indicator of the effectiveness of vaccination in reducing the rate of HCC. Since the incidence of HCC in Taiwan peaks in the sixth decade of life, it may take 40 years or longer to see an overall decrease in the rate of HCC as a result of the vaccination program. Vaccination against HBV should become a health priority together with the promotion of adequate health standards.

Unfortunately, there is no vaccine against HCV. Up to now, the only effective method to prevent its transmission is the avoidance of contamination with infective blood products.

In patients with chronic hepatitis B, clearance of the HBeAg after treatment with interferon-alfa is associated with improved clinical outcome in terms of survival and development of complications of cirrhosis [180]. Another study confirmed these results and showed a reduction of incidence of HCC in the responders [181]. As most of these patients were non-cirrhotics at entry of the study, the prophylactic effect of interferon on development of HCC can be explained by prevention of cirrhosis development. In Chinese patients with chronic hepatitis B infection, however, interferon-alfa was of no long-term benefit in inducing HBeAg conversion, or in the prevention of HCC and other cirrhosis-related complications [182].

Seven studies investigated the possible effect of interferon treatment on development of HCC in patients with already established cirrhosis [183‐189] (Table 4). A meta-analysis was performed on these studies [190]. Interferon seemingly decreased the rate of HCC in all trials, while a significant difference was observed in 2 studies [183, 186]. Virologic response was strongly associated with reduced risk for HCC in the studies of Oon [183] and Mazzella [184], suggesting that arrest of viral replication is a critical factor. Subgroup analysis in relation to ethnic origin of patients (European, Oriental) showed no preventive effect of interferon on the development of HCC in the European patients [190].

It should be noted that the studies are very heterogeneous and that none of them were randomized controlled trials, so that the results should be interpreted with caution.

A recent study showed a significant reduction of the risk of HCC in patients with chronic hepatitis B and advanced fibrosis or cirrhosis, treated with lamivudine for a maximum of five years, compared to placebo [191].

Three studies assessed whether interferon treatment prevents the development of HCC in noncirrhotic patients with chronic hepatitis C [192‐194] comprising 3,798 noncirrhotic patients treated with interferon-alfa monotherapy. Pooled together, the incidence of HCC was 60/2,532 (2.37%) in sustained virological responders and 76/1,266 (5.29%) in nonresponders. In a study of 291 noncirrhotic patients with chronic hepatitis C who were nonresponders to interferon therapy and followed for 6–117 months after therapy, the incidence of HCC was significantly lower in patients who received > 500 MU of interferon. Patients with a transient response (i.e. relapse after end of treatment) had a significant lower rate of HCC development (4/166 = 2.4%) than nonresponders (12/125 = 9.6%) [195].

This anti-oncogenic benefit can presumably be explained by an arrest or slowing down of the cirrhogenic process.

The findings of 13 studies of interferon treatment and development of HCC in HCV-infected patients with compensated cirrhosis are summarized in Table 5[45, 184‐186, 194, 196‐204]. Only 3 studies were randomized [199, 201, 202, 204], the remainders were observational cohort studies. Statistical combination of data is not possible because of different definitions of response (biochemical, virological), different dose schedules for interferon and different duration of follow-up. All studies showed a lower risk for development of HCC in the interferon-treated patients, suggesting that interferon may prevent HCC in compensated cirrhosis caused by hepatitis C. The overall result was largely influenced by three Japanese studies [194, 198, 201, 202], which had the highest incidence of HCC in untreated patients (5–6% per year). This may be explained by intensiveness of the screening programs, but also by genetic, environmental or viral factors. Four European studies failed to document a significant reduction in risk of developing HCC [45, 196, 199]. In the studies of Fattovich et al [196] and Bruno et al [45], interferon-alfa treatment showed a decrease in incidence of HCC in univariate analysis. However, this was not present in multivariate analysis. In the study of Fattovich [196], a very low natural incidence of HCC was observed, rendering difficult to show a significant decrease. The prospective randomized controlled trial of Valla et al [199] also failed to show a significant effect of interferon treatment on the development of HCC. However, the number of patients in this study was limited and the follow-up relatively short. Also a recently published randomized controlled study from Italy comprising 51 interferon-treated and 71 untreated patients with compensated hepatitis C-cirrhosis, failed to demonstrate any reduced risk in development of HCC after a mean follow-up of 96.5 months [204].

In most studies, virological and/or biochemical response are associated with a lower risk of development of HCC, which is less clear in nonresponders. In the study of Imai et al. [198], patients with sustained biochemical response after interferon therapy were at low risk for development of HCC (risk ratio versus controls 0.06; 0.95 in nonresponders). Also in the study of Mazzella [184], a statistically significant effect of interferon treatment was demonstrated when biochemical responders were compared with controls but not when compared with nonresponders. In the study of Benvegnù et al [188], the beneficial effect of interferon treatment on development of HCC was independent of the type of response. In the study of Yoshida et al [194] the risk for HCC was reduced especially among patients with sustained virological but also merely biochemical response that tested positive for HCV RNA. Okanoue et al [200] studied 1,148 patients with chronic hepatitis C treated with interferon-alfa, 40 of them having cirrhosis (fibrosis stage F4). They were followed for 1–7 years after therapy. The cumulative incidence of HCC was significantly decreased in sustained biochemical responders, compared to nonresponders and transient responders, in patients with stage F2 fibrosis, but not in the more advanced stages F3 and F4. In the study of Testino et al [204] HCC did also develop in sustained biochemical responders. Tanaka et al [205], however, demonstrated in 55 patients with HCV-cirrhosis that long-term administration of interferon prevented HCC in those with biochemical and virological response, whereas HCC only appeared in nonresponders.

The mechanisms by which an interferon treatment might reduce the risk of HCC development in cirrhosis caused by HCV independent of virological response remains speculative. Maintenance of serum transaminases at low levels may protect against the development of HCC as hepatocyte necrosis, cell damage and increase in hepatocyte replication result in increased DNA damage, influencing hepatocarcinogenesis. Other possible mechanisms for prevention of HCC are the direct and indirect effects of interferon. It is, however, perplexing that only 6 or 12 months of therapy can produce this benefit without virological response. Because of potential biases in the published trials it is premature to advocate the use of interferon as established therapy in HCV infected patients with cirrhosis to prevent HCC. Prospective randomized controlled trials should reproduce the findings in large numbers of patients before a definitive conclusion on the long term effects of interferon in HCV cirrhosis can be established.

It must also be realized that a sustained virological response to interferon-alfa monotherapy can be obtained only in 0–8% of patients with cirrhosis [206‐208]. New treatments are now available for chronic hepatitis C, which are more performant in difficultly to treat cases as patients with cirrhosis. The combination of interferon-alfa and ribavirin results in a sustained virologic response in up to 25% of cirrhotics due to hepatitis C [207]. A sustained virologic response of 32% was reported after peginterferon-alfa-2a monotherapy [207] and of 43% after combination of peginterferon-alfa2a and ribavirin [173]. It should be investigated in prospective trials, taking into account the sustained virological and biochemical responses if these more performant treatment regimens will also influence favorably the incidence of HCC, as no data on the long-term effects of these treatments are available up to now.