Previously, we reported that maintenance peginterferon in the HALT-C Trial was not effective in preventing clinical or histological outcomes.4
It has been suggested that the limited duration of follow-up in these reports may not have been sufficient to observe a clinical benefit of peginterferon as chemoprevention for HCC;8
in this vein, retrospective studies have suggested that a benefit of interferon in reducing the occurrence of HCC may not be observed for 5 years or more.9
Supporting this notion is a recent meta-analysis demonstrating that patients with hepatitis C who received interferon therapy had a significantly lower incidence of HCC but significant heterogeneity in the results limited confidence in this conclusion, and the duration of follow-up was an independent contributor to the heterogeneity.3
The process of hepatocarcinogenesis is well recognized to take place over many years, and, even after HCC has developed, many months to years may elapse before the tumor can be detected clinically. The HALT-C Trial investigators acknowledged the possibility that the initially planned follow-up period of 4 years from randomization (6 months after completion of randomized treatment) may have been too short to detect a benefit.10
Therefore, the protocol was amended to extend the duration of follow-up.
With continued observation of this large cohort of patients with advanced hepatitis C for up to 8.7 (median 6.1) years, we found that the number of patients with HCC increased to 88 from 53 (including 5 prevalent cases diagnosed in the first year) in an earlier report.5
Continued follow-up of the HALT-C Trial cohort for up to 5 years after the end of therapy did not show a statistically significant benefit of low dose peginterferon in reducing the incidence of HCC in patients with advanced hepatitis C who did not achieve SVR. In sub-group analysis, however, patients with cirrhosis at baseline assigned to treatment had lower risk of HCC.
Because of the unexpected difference among patients with cirrhosis between our previous and the current findings, we performed additional analyses to verify the consistency of the results. A beneficial effect of peginterferon was observed even when the analysis was restricted to patients with definite HCC (i.e., after exclusion of 20 patients with presumed HCC as defined by the protocol). In addition, a dose-response effect of peginterferon was apparent, as suggested by three findings. First, among patients randomized to receive peginterferon, multivariate analysis showed that those who stopped treatment before 2 years did not have a reduction in the risk of HCC after adjusting for laboratory markers of advanced liver disease. Second, among patients receiving peginterferon, those experiencing a ≥2-point reduction in HAI had an approximately 3-fold reduction in HCC compared to those who did not have a reduction in HAI on follow-up liver biopsy. Third, in a regression analysis of factors associated with the development of HCC, including treatment assignment, peginterferon therapy was associated with an approximate 50% reduction in the incidence of HCC over a median of 6.1 year observation period among patients with cirrhosis at baseline even when adjusted for other risk factors including age, race, smoking history, and laboratory tests associated with advanced liver disease. Nevertheless, while analysis by the two pre-specified strata of histological severity (bridging fibrosis and cirrhosis) was warranted, why the effect of peginterferon was confined to patients with cirrhosis is unclear, and the possibility that this is a chance finding cannot be excluded.
A small randomized controlled trial by Nishiguchi et al11
was the first to suggest that interferon may reduce the risk of HCC in patients with HCV-related cirrhosis. This report was followed by many other reports, mostly from Japan and Europe. The majority of these studies were retrospective or cohort studies comparing the incidence of HCC in patients who had received interferon for 6–12 months with historical or concurrent patients who had not received treatment. The median duration of follow-up ranged from 2 to 15 years. Most studies showed a benefit of treatment, although the effect was predominantly seen in patients who achieved SVR. In addition, three independent meta-analyses suggested that interferon therapy decreased the incidence of HCC in patients with chronic hepatitis C3, 12, 13
In the most recent of these meta-analyses, the relative risk (RR) of HCC for treated vs. untreated patients was 0.43 (95% CI 0.33–0.56, P<0.00001) and for patients with SVR vs. without SVR was 0.35 (95% CI 0.26–0.46, P<0.0001); however, only four of the studies included were randomized controlled trials.
Few studies have compared the incidence of HCC in interferon nonresponders and untreated patients. In two large retrospective studies, the incidence of HCC in these two groups was similar. HCC developed in 14% (20/148) of nonresponders and 12% (19/144) of untreated patients in the study by Imai et al14
and in 11% (39/342) of nonresponders and 10% (54/562) of untreated patients in the study by Yu et al.15
As was true for the 3.5-year randomized phase of the HALT-C Trial, two other studies of long-term, low-dose peginterferon (COPILOT and EPIC3) in patients with advanced fibrosis or cirrhosis showed no benefit of treatment on overall clinical outcomes or HCC.16, 17
To date, however, only preliminary data on outcomes up to 4 years have been reported in these two studies.
The mechanisms by which peginterferon might prevent HCC are not clear. Certainly, achieving an SVR is associated with a reduction in HCC. Indeed, patients in HALT-C trial who achieved SVR after 48 weeks of combination therapy had lower risk of HCC compared to lead-in phase nonresponders, HR 0.19, 95% CI 0.04–0.80.18
Patients who had >2 log decline in HCV RNA after 1.5 years of low-dose peginterferon also had a lower incidence of HCC compared to those with <2 log decline but the difference was not significant and only 15.4% of the patients receiving low-dose peginterferon had >2 log decrease in HCV RNA after 18 months of treatment. Interferon has been shown to have antitumor effects in some malignancies and may have other beneficial effects including antiproliferative and antiangiogenic effects.19, 20
Chronic inflammation has been postulated to play a role in hepatitis B virus-(HBV) and HCV-related HCC. The observation that reduction in HCC incidence in the HALT-C Trial cohort was greater among the treated patients experiencing a reduction in HAI suggests that interferon may have acted through this pathway; however, absence of benefit among control patients with a reduction in hepatic inflammation indicates that interferon may exert its effects through other mechanisms and that a reduction in HAI on follow-up biopsies in interferon-treated patients may be a surrogate of other interferon effects.
The reasons for the apparent delay in the chemopreventive effect of peginterferon are also not known, but we speculate that the delay is related to the slow growth rate of HCC, estimated to have a median doubling time of 117 days.21
In fact, the Gompertzian model would predict that an HCC may have been present for several years before reaching a detectable size.22
Therefore, the HCCs diagnosed in the first 3 to 5 years may well have been present at enrollment into the HALT-C Trial. If, indeed, peginterferon prevented HCC, several years would have been required to detect this effect. In this study, the divergence in HCC incidence between the treated and control patients with cirrhosis at baseline was not apparent until after year 4. Unfortunately, with the termination of this study, we could not determine whether the protective effect disappears or increases with longer follow-up.
Several other interesting observations emerged during this study. Severity of liver disease and portal hypertension, as reflected by more abnormal biochemical tests, lower platelet counts, and presence of esophageal varices were associated with an increased risk of HCC. On the other hand, several studies have shown that patients with increased BMI have a higher rate of HCC;23, 24
while patients in the HALT-C Trial with higher BMI had a significantly lower incidence of HCC. An association between obesity and HCC in other studies may be related to the inclusion of patients with cirrhosis secondary to non-alcoholic fatty liver disease or the inclusion of patients with earlier stage liver disease in whom obesity and steatosis may have accelerated fibrosis progression thereby increasing the risk of HCC. The HALT-C Trial excluded patients with severe steatohepatitis and all patients had advanced fibrosis or cirrhosis. Alcohol and smoking have been reported to be associated with HCC. In the HALT-C Trial, although the average life-time alcohol consumption prior to entry into the HALT-C Trial was high, neither duration of regular alcohol consumption nor life-time alcohol consumption was associated with an increased risk of HCC. Smoking is a risk factor for many cancers, but its association with HCC is not as well described. Several prior reports suggested an association between smoking and histological severity of liver disease,25
and a few reports have appeared of an association between smoking and HCC.24, 26
In this study, smoking was a significant risk factor for HCC, even after adjustment for other risk factors. These data highlight the importance of counseling patients with liver disease on smoking cessation. Esophageal varices and more recently increase in hepatic venous pressure gradient had been shown to be predictors of HCC.27, 28
We found that both presence of varices at baseline and incident varices were associated with increased risk of HCC underscoring the need to evaluate for portal hypertension in patients with advanced liver disease.
In an earlier analysis of the incidence of HCC in the HALT-C Trial cohort, we found that 17% of patients with HCC had no evidence of cirrhosis on serial histology.5
In the current analysis, 8% of HCC developed in patients who had no evidence of cirrhosis. The lower rate of HCC in noncirrhotic patients in the current analysis is related to the subsequent finding of cirrhosis on explant livers in some of the earlier patients and a higher probability that patients diagnosed with HCC during the last two years of the HALT-C study had progressed from bridging fibrosis to cirrhosis. These data confirmed that HCC can occur in non-cirrhotic patients but the rate is low and the cost-effectiveness of HCC surveillance in these patients is unclear.
In conclusion, extended follow-up of the HALT-C Trial cohort showed a modest benefit of long-term peginterferon therapy in reducing the incidence of HCC in patients with hepatitis C and cirrhosis but not in those with advanced precirrhotic fibrosis, an effect that took several years after completion of therapy to become apparent. Although the results were corroborated in multiple sub-analyses, the clinical implications are not clear. Given the marginal beneficial effect and its restriction to only part of the HALT-C Trial cohort, that no overall long-term benefit on mortality was observed in this cohort,29
and given the side effects of peginterferon, necessitating dose reduction or discontinuation in a substantial number of patients, the utility of maintenance peginterferon therapy to prevent HCC in patients with HCV-related cirrhosis is doubtful.