Response to antiviral therapy for chronic hepatitis C and risk of hepatocellular carcinoma occurrence in Japan: a systematic review and meta-analysis of observational studies

In Japan, hepatocellular carcinoma (HCC) is a leading cause of cancer mortality and hepatitis C virus infection is a major cause of HCC. We conducted a systematic review and meta-analysis of published studies evaluating patient response to antiviral therapy for chronic hepatitis C on the risk of HCC occurrence in Japan. Articles were searched using terms determined a priori through PubMed, screened by title and abstract, and selected by full-text assessment according to criteria determined a priori, including HCC occurrence in response to interferon (IFN)-based or IFN-free therapy, Japanese study, and 2 or more years of follow-up. We excluded studies on HCC recurrence. We calculated the pooled estimate of the crude incidence rate ratio with data from the selected studies using the person-years method with Poisson regression model and pooled estimate of the hazard ratio adjusted for potential confounders reported by the studies using a random effects model. A total of 26 studies were identified, all of which examined only IFN-based therapy as a result of the selection process. The pooled estimate (95% confidence interval [CI]) of 25 studies was 0.37 (0.33–0.43) for sustained virologic response (SVR) and 1.70 (1.61–1.80) for non-SVR for the HCC incidence rate per 100 person-years, and 0.22 (0.19–0.26) for the incidence rate ratio (SVR vs. non-SVR). The pooled estimate of the hazard ratio (95% CI) of HCC incidence adjusted for potential confounders of 8 studies was 0.25 (0.19–0.34). SVR to interferon therapy for chronic hepatitis C reduces the risk of HCC occurrence.


Methods
Study selection. Studies were identified through PubMed as of April 5, 2022, using the following search terms determined a priori: (interferon or daclatasvir or asunaprevir or sofosbuvir or ledipasvir or ombitasvir or paritaprevir or ritonavir or elbasvir or grazoprevir or beclabuvir or glecaprevir or pibrentasvir or velpatasvir or direct acting antiviral) AND (hepatitis c) AND (hepatocellular carcinoma or liver cancer or liver neoplasm) AND (Japan or Japanese). Two investigators (YY and KT) independently screened all article titles and abstracts, and conducted full-text assessment to determine eligibility, according to criteria determined a priori. Inclusion criteria were (1) studies on Japanese populations, (2) examination of HCC incidence related to SVR and non-SVR and/or non-IFN treatment as chronic hepatitis C therapy, and (3) follow-up duration of 2 or more years. The exclusion criteria were (1) examination of HCC recurrence, (2) examination of continuing therapy for chronic hepatitis C, (3) studies including participants with liver transplantation, (4) studies including participants with co-infection by human immunodeficiency virus, and (5) studies with less than 20 total participants or with less than 10 participants with SVR or non-SVR. For overlapping study populations identified by study period and institute, studies that included a more comprehensive population and more complete data were selected. Inconsistencies in study selection between the reviewers were solved by discussion.
Data extraction and assessment of risk of bias. Data were extracted by YY using a spreadsheet developed a priori that included HCC cases with SVR/non-SVR, patients with SVR/non-SVR, follow-up period, hazard ratio (HR) adjusted for covariates, age, sex, advanced fibrosis, HCV genotype, and methods for HCC diagnosis. The data were checked by KT. Risk of bias in the selected studies was assessed by YY using the Newcastle Ottawa Scale for cohort studies, which consists of three items (selection, compatibility, and outcome) with eight sub-items 30 , and checked by KT. A "good" quality score required 3 or 4 stars in selection, 1 or 2 stars in comparability, and 2 or 3 stars in outcome. A "fair" quality score required 2 stars in selection, 1 or 2 stars in comparability, and 2 or 3 stars in outcome. A "poor" quality score reflected 0 or 1 star in selection, or 0 stars in comparability, or 0 or 1 star in outcome.
Data synthesis and analysis. The crude incidence rate and incidence rate ratio for each study were calculated using the person-years method assuming a Poisson distribution of the observed number of HCC cases with SVR/non-SVR during the follow-up period. Pooled estimates of both measures were derived using a Poisson regression model with a random intercept and a random coefficient for IFN response (SVR vs. non-SVR) according to each study. A pooled estimate of the HR adjusted for potential confounders reported in the studies was calculated using a random effects model with restricted maximum likelihood. Subgroup analysis and sensi- www.nature.com/scientificreports/ tivity analysis were conducted on studies that did or did not include/report data on patients with liver cirrhosis, HBsAg, and Peg-IFN use. Heterogeneity among studies was assessed using Q statistics and I 2 statistics. Publication bias was tested using funnel plots and Egger's test. All analyses were performed with Stata 17.0 (Stata Corp LLC, College Station, TX). This systematic review was not registered.

Results
Study selection. A total of 932 articles were identified through PubMed using search terms determined a priori on antiviral therapy for chronic hepatitis C (Fig. 1). Of these, 119 articles were selected by screening titles and abstracts. Full-text assessment according to the inclusion and exclusion criteria led to the exclusion of 93 studies due to unavailable data on HCC (n = 30), studies on HCC recurrence (n = 5), follow-up period less than 2 years (n = 5), population less than 20 (n = 4), duplicate population (n = 40), and incomplete data, including the number of HCC occurrences among those with SVR/non-SVR, patients with SVR/non-SVR, and followup period, or adjusted hazard ratio (n = 9). As a result, 26 studies on only IFN-based therapy, but not IFN-free therapy, were selected.
Study characteristics and risk of bias assessment. The characteristics of individual studies are summarized in Table 1.
Year of publication, sample size, average follow-up duration and average age ranged from 1997 to 2017, 118 to 4302, 2.4 years to 11.8 years, and 47.2 years to 60 years, respectively. SVR was defined as undetectable HCV RNA at 24 weeks after the end of treatment in all studies, and the SVR rate ranged from 19.6% to 55.3%. In our risk of bias assessment using the Newcastle-Ottawa Scale for individual studies (Table 2), the exposure and outcome were reported based on medical records in most studies as they were hospital-based studies. The mean follow-up period was 5 or more years in 11 studies. The definition of "follow-up period" was inconsistent or unreported. The number lost to follow-up was reported in 6 of 26 studies. The overall study quality was regarded as "good" in 11 studies and "poor" in 15 studies. All eight studies that reported adjusted HRs had a "good" quality score. The magnitude of the association was consistent among studies ( Table 3, Fig. 2). Publication bias cannot be ruled out due to asymmetry in the funnel plot ( Pooled estimates of the crude incidence rate ratio. The crude incidence rate ratio was calculated using data from 25 studies (Table 3). HCC was not observed in patients with SVR in 7 studies. The pooled esti-   (Fig. 2). In subgroup analysis, pooled estimates of the adjusted HR did not differ by the inclusion or exclusion of patients with Peg-IFN use (Fig. 4). In sensitivity analysis, the pooled estimate of the adjusted HR of studies that did and did not include Likewise, pooled estimates did not change in subgroup analyses on study design, study period, follow-up period, sample size, and age (Supplementary Table 1 and 2).

Discussion
We conducted a systematic review of studies that examined the association between response to antiviral therapy for chronic hepatitis C and risk of HCC occurrence in a Japanese population. Based on pooled estimates of both the crude incidence rate ratio calculated using data from individual studies and the HR adjusted for potential confounders reported by the studies, we concluded that SVR to IFN-based therapy for chronic hepatitis C reduces the risk of HCC occurrence. In subgroup analysis and sensitivity analysis related to HBsAg-positivity, liver cirrhosis, and Peg-IFN use, pooled estimates of each effect size did not change.
A previous meta-analysis of observational studies in Asia, including Japan, Europe and North America reported an effect size of similar magnitude to that reported in the present study, irrespective of whether patients had advanced liver fibrosis: SVR to IFN-based therapy reduced the risk of HCC occurrence overall (pooled HR = 0.24, 95% CI: 0.18-0.31) and in advanced liver fibrosis (pooled HR = 0.23, 95% CI: 0.16-0.35) 21 .
Eradication of HCV by IFN is thought to reduce HCC occurrence by improving hepatic inflammation, regression of hepatic fibrosis, and the antitumor effects of IFN including tumoricidal, antiproliferative, or immunomodulatory effects [31][32][33] . These actions are considered unique to IFN compared to IFN-free therapy. Although achievement of SVR to antiviral therapy reduces HCC occurrence, HCC risk can remain during follow-up in patients with SVR. Advanced fibrosis, older age, alcohol intake, and diabetes mellitus have been suggested to increase the risk of HCC 34 . Strategies to improve lifestyle factors along with surveillance for HCC occurrence in patients with SVR are still needed.
Due to the selection process, this study did not include articles on IFN-free therapy. Although the effect of SVR to IFN-free therapy on HCC occurrence was controversial in earlier studies due to the older age and more advanced fibrosis among patients receiving IFN-free therapy 35,36 , later studies suggested that SVR to IFN-free therapy did in fact reduce the risk of HCC 37 . In Japan, Kobayashi et al. reported a cumulative HCC incidence (3-/5-year) of 1.30/3.03% for IFN-free and 1.02/2.19% for IFN-based therapy during the follow-up period (median, 4.0 years and 7.3 years, respectively), with a log-rank test indicating no significant differences in either group 27 . Similarly, Nagata et al. demonstrated that the 3-year cumulative incidence of HCC occurrence in patients with SVR did not differ by therapy in propensity score-matched analysis (3.3% for IFN-based, 1.4% for IFNfree therapy; P = 0.49, log-rank test) 28 38 .
Given that the effectiveness of IFN-based therapy is linked to host and viral factors, one strength of this study was that we focused our analysis specifically on the Japanese population. Among host factors such as age, sex, race, and liver fibrosis, a comparison of patient response to Peg-IFN therapy combined with ribavirin in those infected with HCV genotype 1 identified interleukin-28B polymorphisms as a contributing factor [22][23][24] . The SVR rate was 14% in patients with the unfavorable allele TG/GG at rs809917, and 50% in those with the favorable allele TT 39 . These polymorphisms are associated with spontaneous clearance of HCV, which is observed in around 30% of those with acute infection with HCV in Japan. The minor allele frequency of rs809917 or rs12979860 is race specific, and the frequency of the favorable allele is higher in Asians compared with Caucasians and African-Americans [40][41][42] . In terms of viral factors, HCV genotype is related to response to antiviral therapy and the distribution of genotypes differs by region and route of transmission. Genotype 1 and poor response to IFN therapy were found in an estimated 65% of patients with chronic hepatitis C in 2015 in Japan 26 . Interestingly, Table 2. Risk of bias assessment using the Newcastle-Ottawa Scale for 26 studies. a Single (○) or multicenter (•) hospital-based study. b Drawn from the same community as the exposed cohort. c Based on medical records. d Demonstrated. e Adjusted for fibrosis. f Adjusted for age and other factors. g Based on medical records. h Mean follow-up duration > 3 years (○) and > 5 years (•). i Reported number lost to follow-up.

Article
Selection Comparability Outcome

Author (year)
Representativeness of the exposed cohort a Selection of the non-exposed cohort b  www.nature.com/scientificreports/ Table 3. Pooled estimates of the crude incidence rate (per 100 person-years) and incidence rate ratio of HCC in patients treated with antiviral therapy (SVR vs. non-SVR) calculated using data from 25 studies. HCC hepatocellular carcinoma, SVR sustained virologic response, CI confidence interval. www.nature.com/scientificreports/ evidence suggests that the distribution of genotypes has been altered with recent changes to the route of transmission, such as through intravenous drug abuse and tattoos, leading the prevalence of genotype 2b to be higher in patients born after 1970 regardless their place of birth in Japan 43 . The reported SVR rate to Peg-IFN therapy combined with ribavirin for genotype 2 is 80% 20 . Our findings on Japanese patients in this study suggest that eradication of HCV with antiviral therapy could be reducing HCC occurrence. Given the difficulty of comparing SVR and non-SVR to IFN-free therapy due to the high effectiveness of IFN-free therapy, further studies comparing SVR to IFN-free therapy and SVR to IFN are needed to evaluate the effects of SVR to IFN-free therapy on HCC occurrence. Several limitations of this study should be noted. First, as most studies did not report the number of patients lost to follow-up, we could not rule out some risk of selection bias. Second, the definition of follow-up period was inconsistent or unreported among the selected studies; thus, HCC incidence may have been incorrectly evaluated in this study. Third, adjustment for confounders may have been insufficient: adjustment for alcohol consumption, smoking status, diabetes mellitus, and obesity was not necessarily conducted in each study. Furthermore, the effect size of HCC occurrence in patients with non-SVR to IFN therapy could have been overestimated, since older age and higher fibrosis stage are unfavorable factors for both response to IFN-based therapy and risk of HCC occurrence. Fourth, HCC cases with SVR were not observed and adjusted HRs were not available in earlier studies, possible because the effectiveness of antiviral therapy and technology used to diagnose HCC  www.nature.com/scientificreports/ improved over the study period. Fifth, we used Pubmed only to search for reports during the selection process due to availability. Finally, we could not contact the study authors to clarify any unclear information because the studies were conducted a long time ago. Future studies should examine the effects of SVR to IFN-free therapy on HCC occurrence by comparing SVR to IFN-free therapy and SVR to IFN. The findings of the present study suggest that eradication of HCV by IFN therapy may reduce HCC occurrence in Japanese patients. However, among the articles selected based on the eligible criteria in this study, only two articles reported outcomes other than HCC by response to antiviral therapy: unchanged incidence of malignancy other than HCC by response to IFN 44 and decreased progression to LC in patients without LC before IFN therapy by SVR to IFN 45 . Outcomes other than HCC including overallmortality, hepatic decompensation, and liver-related mortality by response to antiviral therapy are also important issues and warrant further study.
In conclusion, our systematic review on the association between response to antiviral therapy for chronic hepatitis C and HCC occurrence in a Japanese population suggests that eradication of HCV using antiviral therapy for chronic hepatitis C reduces HCC occurrence.

Data availability
The datasets generated during and/or analyzed during the current study are not publicly available due to the agreement of the Research Group for the Development and Evaluation of Cancer Prevention Strategies in Japan but are available from the corresponding author on reasonable request.  . Forrest plot of the pooled estimate of the hazard ratio of HCC incidence adjusted for potential covariates in patients treated with antiviral therapy (SVR vs. non-SVR) by studies that did and did not include Peg-IFN use. HCC hepatocellular carcinoma, SVR sustained virologic response, Peg-IFN pegylated interferon, HR hazard ratio, CI confidence interval.