The role of interferon in the prevention of hepatocellular carcinoma remains controversial.
The role of interferon in the prevention of hepatocellular carcinoma remains controversial.
In this meta-analysis we evaluated the hepatocellular carcinoma incidence in interferon-treated and -untreated patients with hepatitis C virus-related cirrhosis.
Eleven studies with 2178 patients were found to fulfil our inclusion criteria. The pooled odds ratio (OR) and 95% confidence intervals (CI) were calculated from the raw study data.
Hepatocellular carcinoma development was significantly more frequent in untreated (21.5%) than in interferon-treated patients (8.2%; OR: 3.0, 95% CI: 2.3–3.9). In the five studies reporting hepatocellular carcinoma incidence in patients with and without sustained response to interferon, hepatocellular carcinoma was detected at a much higher rate in patients without (9%) than with a sustained response (0.9%; OR: 3.7, 95% CI: 1.7–7.8). Moreover, hepatocellular carcinoma developed significantly more frequently in the untreated patients than in the non-sustained responders (OR: 2.7, 95% CI: 1.9–3.9). The benefit from interferon on hepatocellular carcinoma incidence was not influenced by the study type (prospective or retrospective), the follow-up duration, or the study origin.
Interferon therapy significantly reduces the hepatocellular carcinoma risk in patients with hepatitis C virus cirrhosis. Hepatocellular carcinoma development becomes almost negligible among sustained responders, but a reduction in hepatocellular carcinoma incidence is also achieved even in the non-sustained responders.
Hepatocellular carcinoma is a major cause of death in patients with chronic hepatitis C virus infection.1 It usually develops in patients with cirrhosis and very rarely in those with chronic hepatitis C without established cirrhosis.1, 2 The annual incidence of hepatocellular carcinoma in cirrhotics varies significantly among different geographical regions, ranging from 3% to 10%.3–5 Currently, hepatitis C virus is considered to promote the development of hepatocellular carcinoma mainly through cirrhosis, although the possibility of a direct carcinogenic effect via interaction with cellular genes that regulate cell growth and differentiation also exists.6, 7
The prognosis of untreated hepatocellular carcinoma is extremely poor, while therapeutic interventions are generally ineffective in advanced stages.6 Screening programmes have been applied trying to detect hepatocellular carcinoma in early stages. However, common screening programmes with determinations of serum alpha-fetoprotein (aFP) levels and ultrasonography every 4–6 months are rather unsatisfactory, because hepatocellular carcinoma is detected in the early stages in less than 50% of cases.8 Thus, development of primary prevention strategies is imperative.
Treatment with alpha interferon, a drug with anti-viral, immunomodulatory and anti-tumoral activities, may slow down the disease progression and improve liver histology in patients with chronic hepatitis C, particularly in those with a sustained virological response. However, sustained response is achieved in only a minority of patients (15–25%) and is particularly low (< 12%) in those with established cirrhosis, who frequently experience interferon-related side-effects.9, 10 Interferon therapy has been reported to reduce the incidence of hepatocellular carcinoma in hepatitis C virus cirrhosis, but such an effect has not been confirmed by others.11–14 Although most of the studies have shown a reduction in the hepatocellular carcinoma incidence in the treated patients, the effect of interferon is still considered to be uncertain.15 Therefore, we performed a meta-analysis of all published studies including interferon-treated and untreated patients with hepatitis C virus cirrhosis in order to review the subject in depth and to objectively evaluate whether interferon therapy reduces the incidence of hepatocellular carcinoma.
The source of the reviewed studies was the MEDLINE database from January 1990 to January 2000. Keywords used in our search were ‘hepatocellular carcinoma’, ‘interferon’ and ‘hepatitis C’. We also checked the reference lists of all available review articles and original studies to identify other studies not found in the primary search.
Studies included in our meta-analysis fulfilled the following criteria: (i) they were published as a full article in English; (ii) the study population included both interferon-treated and untreated patients with hepatitis C virus cirrhosis; and (iii) data on the incidence of hepatocellular carcinoma in interferon-treated and untreated patients were given.
The selected end-point was the development of hepatocellular carcinoma.
Data were extracted independently from each study (by G.V.P. and V.C.P.) using a predefined review form and disagreement was resolved by consensus.
The pooled odds ratio (OR) and 95% confidence intervals (CI) were calculated from the raw study data using the Peto method (fixed effect model) and the Der Simonian and Laird method (random effect model). We used a statistical evaluation of heterogeneity by the χ2-test to assess whether the variation in treatment effect within trials of the same group was greater than might be expected. We considered heterogeneity to be present if P < 0.05. In the absence of statistically significant heterogeneity, OR and 95% CI by the fixed effect model are only given in the results. When statistical significance was present, we calculated the ‘number needed to treat’ and the ‘publication bias assessment’.
We also performed separate sensitivity analyses for the incidence of hepatocellular carcinoma according to the following parameters. First, we considered that the response to interferon therapy might influence the results of our analysis. Hence, we performed a separate sensitivity analysis according to the response to interferon treatment (three strata: sustained response, no sustained response, or no treatment). Second, we questioned whether the study design could influence the results of the meta-analysis. Hence, we performed a sensitivity analysis according to the study design (two strata: prospective or retrospective). Third, since the incidence of hepatocellular carcinoma may increase with time and therefore the duration of follow-up might influence the results, we performed a separate analysis according to the duration of follow-up (two strata: < 48 months or 48 months). The cut-off point of 48 months was chosen because it represented the median value of the follow-up times in interferon-treated patients in the studies of our meta-analysis. Fourth, because there have been suggestions that the epidemiology of hepatocellular carcinoma and even the effect of interferon treatment on hepatocellular carcinoma incidence may differ between Japanese and non-Japanese patients, we performed a separate analysis for studies coming for Japan and those coming from other countries.
There were 212 relevant articles in the literature searches. In total, 11 studies involving 2178 patients with hepatitis C virus cirrhosis met our inclusion criteria;11–14, 16–22 1223 patients were treated with interferon and 955 were untreated. Agreement between reviewers for the selection of relevant articles was 100%. Four studies had a prospective design;11, 16, 18, 20 six a retrospective one;12, 13, 17, 19, 21, 22 only one of them was randomized controlled trial.11 There was one further randomized controlled trial in which the efficacy of interferon in hepatitis C virus-related cirrhosis was evaluated, but it had no prospective design for extended follow-up and 20% of control patients were treated with interferon later on; it was therefore considered as a retrospective trial (Table 1).14
The patient numbers and the inclusion criteria in each study are presented in Table 1. Patient demographic characteristics, severity of cirrhosis, and follow-up times are presented in Table 2. Mean or median duration of follow-up was less than 3 years in only one study;18 between 3 and 4 years in four studies;13, 14, 20, 22 and more than 4 years in six studies.11, 12, 16, 17, 19, 21 Screening for hepatocellular carcinoma was performed by both ultrasonography and aFP determinations every 3–12 months in nine studies;11–13, 16–20, 22 and by only ultrasonography every 3–6 months in two.14, 21 Hepatocellular carcinoma diagnosis was confirmed by histology in five studies;12, 16, 18–20 by hepatic angiography with or without histology in four;11, 13, 21, 22 and by histology and/or elevated aFP associated with a liver mass in two (Table 3).14, 17 Interferon was given in various dosages from 3 to 6 MIU thrice weekly for 3–12 months (Table 3). The treatment response was defined only by biochemical criteria in eight studies;12–14, 16–19, 22 and by both biochemical and virological criteria in three (Table 3).11, 20, 21
Hepatocellular carcinoma developed significantly more frequently in untreated (205 out of 955 or 21.5%) than in interferon-treated patients (100 out of 1223 or 8.2%; OR: 3.0, 95% CI: 2.3–3.9, P < 0.001; Table 4, Figure 1). There was no evidence of statistical heterogeneity (P=0.57). Publication bias assessment showed that 171 null or negative studies would be needed to render the result of this analysis not significant, and the number needed to treat to prevent hepatocellular carcinoma in one patient was nine (95% CI: 7–11). The incidences of hepatocellular carcinoma in treated and untreated patients in individual studies are presented in Table 4.
Data about hepatocellular carcinoma incidence in relation to the response to interferon were reported in five studies (Table 5).11, 17, 18, 21, 22 Hepatocellular carcinoma was diagnosed in 9% (51 out of 568) of non-sustained responders and in only 0.9% (one out of 115) of sustained responders (OR: 3.7, 95% CI: 1.7–7.8, P=0.001; Table 5, Figure 2). There was no evidence of statistical heterogeneity (P=1.00). In addition, hepatocellular carcinoma developed significantly more frequently in untreated patients (22%) than in non-sustained responders (9%) (OR: 2.7, 95% CI: 1.9–3.9, P < 0.001; Table 5, Figure 3). There was no evidence of statistical heterogeneity for this analysis (P=0.62).
The effect of interferon on hepatocellular carcinoma was not associated with the type of study. In particular, hepatocellular carcinoma incidence was significantly higher in the untreated than in interferon-treated patients in both the prospective (19.5% vs. 4%, OR: 4.6, 95% CI: 2.7–7.8, P < 0.001) studies;11, 16, 18, 20 and the retrospective studies (22% vs. 10%, OR: 2.7, 95% CI: 2.0–3.5, P < 0.001).12–14, 17, 19, 21, 22 The effect of interferon on hepatocellular carcinoma development was also not found to be associated with the duration of study follow-up. In particular, the incidence of hepatocellular carcinoma was significantly higher in untreated than in interferon-treated patients both in studies with a follow-up of less than 48 months (22.4% vs. 7.4%, OR: 3.0, 95% CI: 1.8–4.9, P < 0.001);13, 14, 18, 20, 21 and studies with a mean follow-up of more than or equal to 48 months (21.1% vs. 8.5%, OR: 3.0, 95% CI: 2.2–4.0, P < 0.001).11, 12, 16, 17, 19, 22 Finally, the origin of the study was not found to influence the effect of interferon on hepatocellular carcinoma development. Hepatocellular carcinoma developed significantly more frequently in untreated than in interferon-treated patients in both Japanese studies;11, 13, 21, 22 and non-Japanese studies (Table 6).12, 14, 16–20
Anti-viral therapy with interferon plus ribavirin or interferon alone, in cases with a contra-indication to ribavirin, is currently recommended for chronic hepatitis C patients with moderate or severe necroinflammation and/or fibrosis.23 However, the necessity of such a treatment is considered to be doubtful for patients with established cirrhosis because of both the relatively low response rate and the frequent and potentially severe side-effects.23
Our meta-analysis shows that interferon treatment reduces the development of hepatocellular carcinoma, because the incidence of hepatocellular carcinoma was found to be significantly lower in interferon-treated (8.2%) than in untreated patients with hepatitis C virus cirrhosis (21.5%). The results of a meta-analysis certainly depend on the quality of the individual studies. From the 11 analysed studies, only one was indeed randomized, while the remaining 10 were prospective or retrospective, mostly observational cohort studies. The results of the only randomized trial by Nishiguchi and colleagues, which was also the first published study on the subject, were similar to the findings of our meta-analysis.11 There were, however, considerable differences between the studies in the interferon regimens as well as in the methods of screening for and of diagnosis of hepatocellular carcinoma (Table 3). On the other hand, both interferon-treated and untreated patients all had compensated hepatitis C virus-related cirrhosis (Child class A in the vast majority of cases), no evidence of current hepatitis B virus (HBV) infection, and no history of heavy alcohol drinking. Moreover, there were no significant differences in sex, age, and Child–Pugh class between treated and untreated patients in all studies with available data (Table 2). However, it is widely accepted that a lack of randomization can introduce bias, because treated groups may differ from untreated ones in both known and unknown baseline variables. In addition, data on several factors which may determine the severity of liver disease or the risk of hepatocellular carcinoma (e.g. platelet count, spleen size or other markers of portal hypertension, duration of infection, hepatitis C virus genotypes, viral load, and prevalence of past HBV infection), were not available for interferon-treated and untreated patients in most of the studies.4, 6, 24, 25 Despite all these disadvantages, the difference in hepatocellular carcinoma incidence between treated and untreated patients cannot be easily explained only by unrecognized differences in baseline characteristics between the two groups. Furthermore, the severity of cirrhosis, a factor that cannot be easily estimated in cirrhotics and may differ between treated and untreated patients, has not been clearly shown to be associated with the risk of hepatocellular carcinoma.24, 25 Although there is a need for confirmation of our findings in at least one large, randomized, controlled, long-term trial including patients with early hepatitis C virus cirrhosis, such a trial may be quite difficult to conduct in the future, because most patients are unwilling to remain untreated, particularly in the current era of combination therapy.
The mechanism of hepatocarcinogenesis in hepatitis C virus infection remains unclear.6, 24 Hepatitis C virus is an RNA virus that cannot be integrated in the host genome, such as HBV. Although a direct carcinogenic effect cannot be excluded, hepatitis C virus is generally considered to promote the development of hepatocellular carcinoma.7 This is mainly indirectly through chronic inflammation and cirrhosis, and the subsequent DNA damage resulting from increased hepatocyte necrosis and regeneration.6, 24 If the last hypothesis is valid, any therapy sufficient to suppress viral replication and hepatocyte necroinflammation may inhibit or delay the progression of liver disease and the development of hepatocellular carcinoma. It should be noted that interferon therapy does not seem to have a strong beneficial effect in hepatocellular carcinoma prevention in patients with HBV-related cirrhosis, in whom integration of the host genome is believed to be mainly responsible for liver carcinogenesis.6, 15 Several studies have suggested that interferon treatment may prevent the development of hepatocellular carcinoma even in non-cirrhotic patients with chronic hepatitis C.26, 27 Although the benefit of any therapy in hepatocellular carcinoma prevention may be difficult to disclose in such patients with very low hepatocellular carcinoma risk, these data further support the role of interferon in hepatocellular carcinoma prevention in patients with hepatitis C virus-related liver disease.
The response to interferon therapy was found to influence the rate of hepatocellular carcinoma development in the five studies with available data.11, 17, 18, 21, 22 Hepatocellular carcinoma development becomes almost negligible among sustained responders (< 1%) and certainly significantly less frequent than in non-sustained responders (9%). The difference between sustained and non-sustained responders is not influenced by the lack of randomization between treated and untreated patients. Based on these findings, one can hypothesize that the combination therapy of interferon plus ribavirin, which achieves better sustained response rates than interferon monotherapy in patients with chronic hepatitis C, and probably even in cirrhotics, may further reduce the development of hepatocellular carcinoma in this setting.28–30
We also found that hepatocellular carcinoma incidence was significantly lower in non-sustained responders than in untreated patients (9% vs. 22%). If this is the case, then interferon probably has an effect in hepatocellular carcinoma prevention independent of its direct anti-viral activity, which supports the suggestions that there is a beneficial effect of interferon even in the non-responder patients.31 Interferon is a cytokine with anti-viral activities but immunoregulatory and anti-tumoral effects as well. It results in inhibition of cell growth and proliferation via reduction of expression of growth factors, change of action of transcriptional factors, and induction of natural killer cells and cytotoxic lymphocytes activity. Although it has been approved for the treatment of several neoplasmatic conditions, mainly haematological malignancies, the administration of interferon is probably ineffective in the treatment of hepatocellular carcinoma.32 However, if interferon can reduce the hepatocellular carcinoma risk even in non-responder cirrhotics, its use must be recommended for all patients with compensated hepatitis C virus cirrhosis regardless of its actual efficacy.
The type of study, prospective or retrospective, did not have any influence on the effect of interferon therapy on hepatocellular carcinoma incidence. Moreover, the benefit from interferon was greater in the prospective than retrospective studies (O.R: 4.6 vs. 2.7); this finding potentiates the overall result of our analysis. The duration of follow-up was also not found to influence the incidence of hepatocellular carcinoma in interferon-treated and untreated patients. Hepatocellular carcinoma development was significantly less frequent in treated than untreated cirrhotic patients irrespective of duration of follow-up (below or above 48 months). This result also strengthens the validity of a beneficial effect of treatment on the development of hepatocellular carcinoma. The incidence of hepatocellular carcinoma was generally higher in Japanese than non-Japanese studies, but the beneficial effect of interferon therapy was not influenced by the origin of the study. In particular, hepatocellular carcinoma incidence was reported to be 14% in interferon-treated, 34% in untreated Japanese patients, and 6% in interferon-treated, 17% in untreated patients in non-Japanese studies. The universal reduction of hepatocellular carcinoma incidence after interferon therapy, irrespective of the hepatocellular carcinoma risk in the untreated population, further strengthens the beneficial effect of interferon.
The most serious complications of chronic hepatitis C are hepatic failure and hepatocellular carcinoma. Cirrhosis develops in 20–30% of patients with chronic hepatitis C and usually remains silent for many years.1 Complications of portal hypertension and hepatic failure occur late and related mortality is 2–5% per year.3 The annual incidence of hepatocellular carcinoma ranges from 3% to 10%, but hepatocellular carcinoma is not the most frequent cause of death in patients with hepatitis C virus cirrhosis.3–5 Therefore, the effect of interferon or other type of therapy on the overall survival must be always evaluated. This is particularly important, because, at least in one study, relatively fewer treated than untreated patients developed hepatocellular carcinoma, but relatively more treated patients died.14 Unfortunately, there are very few data about the effect of interferon on patients’ survival and no clear conclusions can be drawn.3, 33
In summary, our meta-analysis supports that interferon therapy reduces the overall risk of hepatocellular carcinoma development in patients with hepatitis C virus-related cirrhosis. Sustained responders are the patients who most benefit from treatment, but a reduction in hepatocellular carcinoma incidence is achieved even in non-sustained responders. Based on these findings, the presence of compensated cirrhosis should not be considered a reason for excluding patients with hepatitis C virus-related liver disease from interferon therapy. However, more studies are needed to evaluate the effect of interferon therapy on the actual survival of patients with chronic hepatitis C.