More than 170 million people are estimated to be infected with hepatitis C virus (HCV) worldwide, and its prevalence in Korea is approximately 1.29% in people 40 years of age and older.^1,2 Most cases of HCV infection progress to chronic hepatitis C (CHC), which leads to cirrhosis in 15% to 56% of patients with CHC. Once cirrhosis develops in these patients, hepatocellular carcinoma and death occur, with incidence rates of 1.4% to 4.9% and 2% to 4% per year, respectively.³ Globally, one-fourth of cirrhosis or hepatocellular carcinoma cases are related to HCV infections.⁴

Previous studies have demonstrated that sustained virological response (SVR) with interferon-based antiviral treatment could improve the prognosis of patients with CHC by reducing hepatic decompensation, hepatocellular carcinoma, and liver-related death.^5?7 However, SVR rates in patients with advanced fibrosis or cirrhosis have been unsatisfactory, with the result being 33% to 51% for patients with HCV genotypes 1/4 and 57% to 61% for patients with HCV genotypes 2/3.⁸ Fortunately, however, several combination trials of new oral and direct-acting antiviral agents have reported SVR rates exceeding 90% even in patients with advanced fibrosis who failed to respond to interferon-based treatments, regardless of HCV genotypes.^9,10 However, effective emerging therapies remain out of reach for most patients in the near future due to their high costs, leaving peginterferon plus ribavirin as a standard therapy for CHC in Korea for the time being.

Although some Korean studies have reported relatively higher SVR rates in patients with HCV genotype 1 and non-1 genotypes compared to Western studies,^11,12 the reasons for these improved responses have not been clearly delineated, and the response rate in cirrhotic patients remains uncertain. Some Western studies demonstrated that successful antiviral treatment in patients with CHC halted progression of hepatic fibrosis, but the risk of hepatocellular carcinoma remained even after achieving SVR.^8,13 However, this evidence is insufficient to establish concrete surveillance strategies against liver cancer for patients with CHC who achieve SVR with antiviral treatment, since there are few reports regarding long-term posttreatment outcomes in patients with CHC according to liver fibrosis stage before treatment.

Therefore, we have extensively analyzed retrospective clinical practice data regarding treatment response to standard peginterferon and ribavirin combination therapy and long-term posttreatment outcomes in relation to hepatic fibrosis stage in Korean patients with histologically confirmed chronic hepatitis or clinically evident cirrhosis due to HCV infection.

1. Patients

A total of 376 patients with CHC received peginterferon and ribavirin combination therapy at Korea University Anam Hospitals from February 2003 to February 2014. After excluding 25 patients who were still under treatment at analysis, 17 without pretreatment serum HCV RNA data, and 30 patients lost to follow-up before SVR assessment, 304 patients were included in the final analysis. Eight patients had a past history of curatively treated hepatocellular carcinoma. This study conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the institutional review board for human investigations of Korea University Anam Hospital (ED15233-09-2015). Informed written consent was waived because of the retrospective nature of this study.

2. Diagnostic methods

HCV genotyping, HCV RNA quantification, and analysis of IL28B genotypes (rs12979860) were performed using the techniques described in Supplementary Table 1. The histological progression of hepatic fibrosis in liver biopsy specimens was divided into five stages using the Batts-Ludwig system: F0 (no fibrosis), F1 (portal fibrosis), F2 (periportal fibrosis), F3 (septal fibrosis), and F4 (cirrhosis).¹⁴ Cirrhosis was evident in 62 patients based on imaging and clinical findings. Hepatocellular carcinoma was diagnosed by pathology or based on clinical diagnostic criteria from the European Association for the Study of the Liver (EASL).¹⁵ Clinically relevant portal hypertension (CRPH) was defined by the presence of esophagogastric varices or thrombocytopenia (<10⁵/mm³) with splenomegaly. Decompensated cirrhosis was defined by the presence of ascites, variceal bleeding, hepatic encephalopathy, and Child-Pugh score ≥7. Well-compensated cirrhosis was defined as the state of compensated cirrhosis corresponding to Child-Pugh score 5.

3. Treatment

We administered 180 μg peginterferon α-2a (Pegasys; Roche Pharma AG, Reinach, Switzerland) weekly regardless of body weight or 1.5 μg/kg peginterferon α-2b (Pegintron; Merck & Co., Kenilworth, NJ, USA) weekly, in combination with 1,000 or 1,200 mg/day ribavirin (body weight ≤75 kg or >75 kg, respectively) for 48 weeks to patients with HCV genotype 1, and 800 mg/day ribavirin regardless of body weight for 24 weeks to patients with non-1 HCV genotypes. Peginterferon α-2a was used in 43.1% (131/304) and peginterferon α-2b in 57% (173/304) of patients. The type of peginterferon to be used was just determined by the judgement of attending physician.

4. Outcomes

Virological responses were defined by EASL guidelines for CHC.¹⁶ All virological responses were determined by intention-to-treat analysis. All patients were evaluated for other virological responses except rapid virological response (RVR). Long-term outcomes were evaluated by the occurrence of hepatic decompensation (defined by ascites, variceal hemorrhage, hepatic encephalopathy, or Child-Pugh score >6), hepatocellular carcinoma, all-cause death, and liver-related death.

5. Statistical methods

Student t-test was used to compare the mean of continuous variables. Chi-square and Fisher exact tests were applied to compare the frequencies of categorical variables. Univariate and multivariate logistic regression analyses were used to identify factors associated with SVR. Univariate and multivariate Cox regression analyses were performed to identify factors associated with long-term outcomes. In both analyses, variables with p<0.1 in univariate analysis were included in the stepwise multivariate analysis. Cumulative incidence rates and curves for each event were plotted using the Kaplan-Meier method and compared using the log-rank test. All statistical analyses were performed using SPSS version 20.0 (SPSS Inc., Chicago, IL, USA) and p-values <0.05 were considered statistically significant in all analyses.

1. Baseline characteristics

A total of 304 patients were followed up for 4 to 118 months (median, 54 months); their baseline characteristics according to HCV genotype and cirrhosis are summarized in Table 1. There were 152 patients each with HCV 1 (six genotype 1a, 141 genotype 1b, and five unclassified) and non-1 genotypes (149 genotype 2 and three genotype 3). Liver cirrhosis was found in 92 patients, of whom 49 and 43 belonged to HCV genotypes 1 and non-1, respectively; 52 (57%) had clinically relevant portal hypertension; 18 (20%) and one (1%) had Child-Pugh score of 6 and 7, respectively; four (4%) showed signs of decompensated cirrhosis, and there were no genotype differences associated with these characteristics.

The incidence of the IL28B (rs12979860) CC genotype was 88% as a whole, and did not vary according to HCV genotypes or the presence of cirrhosis. Baseline serum HCV RNA level was significantly higher in patients with HCV genotype 1 than that in those with non-1 genotypes (all p<0.001) among both cirrhotic and noncirrhotic patients; among patients with identical HCV genotypes, those with cirrhosis showed significantly lower or a tendency towards lower serum HCV RNA levels at baseline than those without cirrhosis (p=0.038 for genotype 1 and p=0.073 for genotype non-1).

Peginterferon α-2a was more often administered to patients with HCV genotype 1 (66%) and peginterferon α-2b in patients with non-1 genotypes (80%) (both p<0.001), but no preference was observed between patients with and without cirrhosis (Table 1).

2. Adherence to treatment

Treatment was prematurely discontinued in 43% (34/103) and 33% (21/49) of HCV genotype 1 patients with and without cirrhosis, respectively, compared to 16% (7/43) and 7% (8/109) of among patients with identical HCV genotypes, those with cirrhosis showed significantly lower or a tendency towards non-1 genotype patients with and without cirrhosis. Drug discontinuation occurred more frequently in patients with HCV genotype 1 than patients with non-1 genotypes among patients with and without cirrhosis (all p<0.01). The proportion of patients who received drugs of more than 75% of scheduled dosage was lower in patients with cirrhosis than those without cirrhosis for both HCV genotype 1 (38% vs 55%, p=0.056) and non-1 (44% vs 73%, p<0.001). The most common cause of discontinuation was severe adverse drug effects (54%, 38/70) (Table 1, Supplementary Table 2).

3. Virological response

Achievement of virological response to standard peginterferon and ribavirin combination therapy is shown in Table 1. Among patients with HCV genotype 1, SVR was achieved in 36.7% (18/49) and 67% (69/103) of patients with and without cirrhosis (p<0.001), respectively. In comparison, among patients with non-1 HCV genotypes, SVR rates were 86.0% (37/43) in cirrhotic patients and 86.2% (94/109) in noncirrhotic patients. The SVR achievement rate did not differ significantly between the F0?2 and F3 groups containing patients with HCV genotype 1; similarly, the rate did not differ significantly among the F0?2, F3, and F4 (cirrhosis) groups consisting of patients with HCV non-1 genotypes (Fig. 1A). In addition, among patients with cirrhosis, the SVR rate did not differ significantly between those with and without portal hypertension, regardless of HCV genotype (Fig. 1B).

4. Factors associated with SVR

Multivariate logistic regressions analyses confirmed that well-known baseline factors were independently associated with SVR, including serum HCV RNA levels and other variables such as the presence of cirrhosis in patients with HCV genotype 1 and age in patients with HCV non-1 genotypes (Table 2). Along with these baseline predictors, SVR was independently and positively correlated with RVR achievement and treatment adherence, including actual drug dosage in patients with HCV genotype 1 and actual duration of drug exposure in patients with non-1 genotypes (Table 2). In fact, the SVR rate in genotype 1 patients with and without cirrhosis who received >75% of the planned dosage reached 58% and 88%, respectively, which further increased to 83.3% and 93% with the additional presence of RVR. Among patients with non-1 genotypes, SVR rates exceeded 90% in both cirrhosis and noncirrhosis patients with drug exposures for >75% of expected duration, and reached 97.6% in those under 65 years of age with serum HCV RNA levels ≤5 log₁₀ IU/mL (Supplementary Tables 3 and Supplementary Tables 4).

Furthermore, among patients with HCV genotype 1, the IL-28B CC genotype was associated with significantly higher SVR rates after adjusting for baseline predictors (p=0.019 vs TC/TT genotypes) (Fig. 1C), while peginterferon α-2b was linked with a significantly lower SVR rate (p=0.007 vs peginterferon α-2a) (Table 2, Fig. 1D).

5. All-cause and liver-related deaths

Among the 304 patients, there were 19 deaths during the follow-up period (median, 54 months) after treatment initiation, including 12 liver-related deaths. Detailed causes of death were as follows: hepatocellular carcinoma in five patients, liver failure in four, spontaneous bacterial peritonitis in one, variceal bleed in one, cholangiocarcinoma in one, pneumonia in two, chronic kidney disease in one, myocardial infarction in one, pancreatic cancer in one, lung cancer in one, suicide in one.

When expressed as events per 10⁴ person-years, incidence rates of death were remarkably higher in patients with cirrhosis (F4) than that in those without (353 vs 48 for all-cause death, 277 vs 10 for liver-related death, both p<0.001 using Cox regression analysis) (Table 3). Additionally, the rates of all-cause death and liver-related death were significantly lower in patients with SVR than that in those without (all-cause death: 87 vs 257; hazard ratio [HR], 0.34; 95% confidence interval [CI], 0.14 to 0.83; p=0.018; liver-related death: 48 vs 180 per 10⁴ person-years; HR, 0.27; 95% CI, 0.09 to 0.86; p=0.026). When we considered the degree of hepatic fibrosis, the impact of SVR on mortality was only obvious in a subgroup of patients with advanced fibrosis (F3+F4) who maintained well-compensated liver functions (Child-Pugh score, 5) at baseline (p=0.006 and 0.016 for all-cause and liver-related death, respectively) (Table 3). Multivariate Cox regression analysis demonstrated an interaction between SVR and well-compensated cirrhosis for all-cause (p=0.063) and liver-related deaths (p=0.04), suggesting that a substantial reduction in liver-related death due to SVR would be expected only in patients with well-compensated cirrhosis at baseline (Table 4). In fact, the difference in cumulative incidence rates of liver-related death between the SVR and non-SVR groups was statistically significant only in patients with well-compensated cirrhosis (p=0.034 by log-rank test), and not in cirrhotic groups including patients with Child-Pugh scores >5 (Fig. 2A and B, Supplementary Table 5).

6. Development of hepatic decompensation

Among the 300 patients without hepatic decompensation at baseline, 26 developed hepatic decompensation during the follow-up period. The incidence rate was significantly higher in patients with cirrhosis than in those without (674 vs 19 per 10⁴ person-years, p<0.001), but was significantly lower in patients who attained SVR compared to those who did not (79 vs 491 per 10⁴ person-years; HR, 0.16; 95% CI, 0.07 to 0.36; p<0.001) (Table 3). The effect of SVR on hepatic decompensation in relation to hepatic fibrosis stage was only significant in patients with cirrhosis (F4) (p<0.012) and advanced fibrosis (F3+F4) (p<0.001) (Table 3). Multivariate analysis showed independent relationships between the occurrence of hepatic decompensation and achievement of SVR (p=0.002) or the presence of cirrhosis (p=0.001) (Table 4). The cumulative incidence rates of hepatic decompensation consistently increased with progression of fibrosis and with lack of SVR, as shown in Fig. 2C and Supplementary Table 5 (p=0.027 between F0?2 and F3 groups, p=0.020 between F3 group vs F4 group with SVR, and p=0.008 between F4 groups with and without SVR).

7. Development of hepatocellular carcinoma

Among the 296 patients without past history of hepatocellular carcinoma at baseline, 21 developed hepatocellular carcinoma during the follow-up period. The incidence rate of hepatocellular carcinoma was remarkably higher in cirrhotic patients than that in noncirrhotic patients (503 vs 39 per 10⁴ person-years, p<0.001), and significantly lower in patients who attained SVR compared to those who did not (80 and 360 per 10⁴ person-years; HR, 0.22; 95% CI, 0.09 to 0.54; p<0.001) (Table 3). This SVR effect against hepatocellular carcinogenesis was only evident (p=0.005) in patients with advanced fibrosis (F3+F4) compared to other stages of hepatic fibrosis (Table 3). Multivariate analysis corroborated the independent impact of cirrhosis (p<0.001) and SVR (p=0.032) on hepatocellular carcinogenesis (Table 4). Reflecting these findings, the cumulative incidence rates of hepatocellular carcinoma increased significantly with the progression of fibrosis and lack of SVR, as shown in Fig. 2D and Supplementary Table 5 (p<0.001 between F0?2 and F3 groups, p=0.048 between F4 groups with and without SVR by log-rank test).

In the present study, standard peginterferon and ribavirin combination therapy resulted in fairly high SVR rates in CHC patients with cirrhosis, including 36.7% among those with HCV genotype 1 and 86.0% for non-1 genotypes. These results were better than those reported in Western studies (13% to 33% for genotype 1/4 and 31% to 57% for genotype 2/3)^8,17?19 and in small Korean studies (21% to 24% for genotype 1 and 33% to 53% in non-1 genotypes).^20,21 In addition, SVR rates in the current study reached 67% among patients without cirrhosis with HCV genotype 1 and 86% in those with non-1 genotypes, also better results compared to large-scale Western studies (57% to 60% for genotype 1 and 76% for non-1 genotype groups),^8,22 but similar to other Korean data (74% for HCV genotype 1 and 83% in genotype non-1).¹¹

The improved treatment responses in Korean patients with CHC could be partly attributed to a higher frequency of the specific IL28B genotype (rs12979860 CC type and rs8099917 TT type) in Koreans, which is closely linked to favorable responses to interferon treatment among patients with HCV genotype 1. The frequency of IL28B rs12979860 CC genotype in the present study was 88%, markedly higher than the 39% prevalence in European Americans, 16% in African Americans, and 35% in Hispanic Americans reported in a previous study.²³ Indeed, in the present study, SVR achievement among patients with HCV genotype 1 was significantly higher in those with IL28B CC genotype than in those with non-CC genotypes (Fig. 1C).

Another factor accounting for better SVR rates in the present study, especially in patients with cirrhosis, might be higher treatment adherence. Previous studies on patients with cirrhosis showed that the rate of premature drug cessation was as high as 67% and 33% in patients with and without HCV genotype 1, respectively.¹⁹ In comparison, the premature withdrawal rate in this study was 43% among cirrhotic patients with HCV genotype 1 and only 16% among those with non-1 HCV genotypes. This higher adherence to treatment in our cirrhotic patients, especially in those with non-1 HCV genotype, must have contributed to a higher SVR rate despite the presence of cirrhosis. Concordant with a previous study,²⁴ treatment response in patients with HCV genotype 1 improved with drug exposures >75% of the total target dose in the present study, with the SVR rate reaching 58% even in cirrhotic patients; it further increased to 83.3% in the presence of RVR. Therefore, a combination therapy of peginterferon and ribavirin might be cost-effective in some subgroups. Unfortunately, however, the proportion of patients who received >75% of the total target dose was only 38% in cirrhotic patients with HCV genotype 1, as compared to 55% in noncirrhosis patients with the same genotype (p=0.056), probably resulting in a far lower SVR rate in cirrhotic patients (37% vs 67%).

Furthermore, a relatively lower level of viremia observed in non-1 HCV genotype cirrhotic patients in the present study, as compared to those reported in previous studies (4.90 vs 6.04 to 6.11 log HCV RNA IU/mL),^20,21 probably had a positive influence on SVR rate, for viral load is a well-known predictive factor of SVR as confirmed in the present study.

Some of our study results are also noteworthy in view of factors affecting the effect of peginterferon and ribavirin on CHC. First, there has been controversy about the therapeutic effects of different peginterferon types.^25?27 In the present study, peginterferon α-2a exhibited a significantly higher SVR rate in patients with HCV genotype 1, particularly those without cirrhosis, compared to peginterferon α-2b. Secondly, some studies reported very low SVR rates of 13% to 14% in compensated cirrhosis patients with HCV genotype 1 and portal hypertension.^17,18 However, in the present study, the SVR rate in these patients was as high as 40%, and did not differ significantly from the SVR rate in those without portal hypertension. This might be owing to their relatively high adherence to treatment, not inferior to that in patients without portal hypertension as shown in Supplementary Table 3. Thirdly, Bruno et al. demonstrated a decrease in SVR rate with progression of hepatic fibrosis in patients without cirrhosis.⁸ In the present study, however, no significant difference in the SVR rate was observed between those with and without bridging fibrosis (F3 vs F0?2), irrespective of HCV genotypes. Yet, the results of present study on the therapeutic efficacy according to the type of peginterferon and patients’ characteristics should be interpreted with caution as our study has unavoidable limitations arising from its nature of retrospective study design.

Sporadic reports have demonstrated that SVR achievement reduces the occurrence of liver-related death, hepatic decompensation, and hepatocellular carcinoma in patients with CHC.^5,6 A recent meta-analysis reported that SVR achievement reduced liver-related death (0.23-fold), risk of hepatic decompensation (0.21-fold), and risk of hepatocellular carcinoma (0.21-fold) in patients with CHC in all stages of fibrosis.⁷ A large-scale observational study conducted in the United States reported that SVR achievement reduced all-cause mortality 0.51- to 0.70-fold.²⁸ In the present study, we also confirmed that SVR achievement reduced the risk of overall death 0.34-fold; liver-related death 0.27-fold; risk of hepatic decompensation 0.16-fold; and risk of hepatocellular carcinoma 0.22-fold in patients with CHC in all stages of fibrosis.

However, it should be noted that pretreatment fibrosis stages in the present study greatly affected patient prognosis after successful treatment. Hepatic decompensation or liver-related death did not occur during the follow-up period in patients without cirrhosis who achieved SVR. In contrast, the risk of hepatic decompensation, while reduced, still existed in patients with cirrhosis even after achieving SVR. Particularly, the risk of liver-related death was reduced only in patients with cirrhosis with Child-Pugh scores of 5. This suggests that SVR needs to be achieved before liver function deteriorates beyond a Child-Pugh score of 5 in order to improve the prognosis of patients with HCV-related cirrhosis. In addition, it is noteworthy that hepatocellular carcinoma did not occur in patients without advanced fibrosis (F0?2) once they achieved a SVR, but did develop at a high rate in patients with advanced fibrosis (F3+F4) even after they achieved a SVR. This indicates that, irrespective of SVR achievement, patients with advanced hepatic fibrosis due to CHC should undergo continued surveillance for hepatocellular carcinoma.

In summary, we were able to achieve a high SVR rate in Korean patients with CHC, including those with cirrhosis, using a standard peginterferon and ribavirin combination therapy, and showed that SVR achievement could reduce the incidence of liver-related death, hepatic decompensation, and hepatocellular carcinoma. However, the risk of liver-related complications still exists in patients with advanced fibrosis even after achieving SVR. Therefore, all patients with HCV-related advanced hepatic fibrosis should receive close follow-up irrespective of successful antiviral treatment.

No potential conflict of interest relevant to this article was reported.