Abstract
Background
IL28B polymorphisms are associated with spontaneous clearance of HCV infection and response to therapy. Whether IL28B genotype affects fibrosis progression or clinical outcome is unclear.
Aims
To study the relationship between IL28B genotype and both histological and clinical outcomes in patients with chronic hepatitis C (CHC).
Methods
Hepatic fibrosis was scored using the Ishak (0-6) scale; progression was defined as a 2-point increase in Ishak score between biopsies. Multiple logistic and Cox regressions were used to identify variables associated with fibrosis progression.
Results
1483 patients were included in a baseline cross-sectional analysis, from which 276 were eligible for a paired biopsy analysis (median time between biopsies-4 years), and 400 for a clinical outcome analysis. At baseline biopsy, patients with IL28B CC genotype had significantly higher portal inflammation (2.4 vs. 2.2) and ALT levels (133 vs. 105 U/L), p <0.05 for all. In the paired biopsy analysis, there was no difference in the frequency of fibrosis progression between patients with IL28B CC and non-CC genotypes (17% vs. 23%). In logistic regression, only higher baseline alkaline phosphatase, lower platelets and greater hepatic steatosis were associated with fibrosis progression. Patients with IL28B CC were twice as likely to develop adverse clinical outcomes compared to non-CC (32% vs. 16%), P= 0.007.
Conclusions
IL28B CC genotype was associated with greater hepatic necroinflammation, higher ALT, and worse clinical outcomes in CHC patients. This suggests that IL28B CC is associated with a state of enhanced immunity that on one hand can promote viral clearance, but alternately, can increase necroinflammation and hepatic decompensation without enhancing fibrosis progression.
Keywords: Hepatitis C virus, IL28B, Fibrosis progression
Introduction
Chronic hepatitis C (CHC) is a global health problem and can lead to cirrhosis, end-stage liver disease and hepatocellular carcinoma (HCC).1, 2 It is the most common cause of death from liver disease and indication for adult liver transplantation in the United States.3 However, not all subjects with CHC will develop these serious sequelae; indeed, a majority of individuals will die with their disease rather than from their disease. Although several host, viral and environmental factors have been linked with outcome of CHC,4, 5 they do not completely explain the variable outcome of the disease. Recently, genome-wide association studies have identified several single nucleotide polymorphisms (SNPs), within and in the vicinity of three genes that encode IFN-λ.6-10 The CC genotype of rs12979860 was strongly associated with resolution of HCV infection following treatment with peginterferon and ribavirin and was independent of race, with similar sustained virological response (SVR) rates among individuals of both European and African ancestry.9 Moreover, rates of spontaneous and treatment-associated clearance of HCV infection for patients with the CC genotype were approximately double those for the TT genotype.6, 9 These studies underscore the importance of the IL28B gene in the outcome of acute HCV infection and response to peginterferon-based therapy. However, the role of IL28B in the natural history of chronic HCV infection is not well understood.
A recent study suggested that the T allele of IL28B rs12979860 was more prevalent among patients with HCV-related cirrhosis compared to patients with mild CHC and that carriage of the T allele was associated with an increased risk of developing HCC.11 Further, more severe necroinflammatory activity and fibrosis were found to be associated with homozygosity for the major T allele of another IL28B SNP (rs809917) in patients with CHC from Japan.12 In contrast, Maribita et al13 showed no relationship between IL28B genotype and severity of liver fibrosis. Moreover, none of the previous studies have examined the relationship between the IL28B genotype and disease outcome as assessed by fibrosis progression using serial liver biopsies and hard clinical outcomes. Therefore, the primary aims of the current study were to investigate whether the previously identified IL28B SNP rs12979860 (CC, CT or TT genotype) was associated with histological progression on serial liver biopsies in a large cohort of patients with CHC and to assess if there was any association of IL28B SNP rs12979860 with clinical outcomes.
Methods
Patients
Adult patients (age 18 or above) were analyzed from two cohorts: 1) patients participating in a long-term natural history study of CHC conducted at the Clinical Center of the National Institutes of Health (NIH)14 including patients who were referred for evaluation and possible therapy who elected not to undergo treatment14 (NIH Cohort) and; 2) the Hepatitis C Long-Term Treatment Against Cirrhosis (HALT-C) Trial (HALT-C Cohort).15 Patients in the NIH cohort underwent either a protocol liver biopsy or a recommended standard of care biopsy approximately every 5 years and were never treated before or between biopsies. The design of the HALT-C Trial has been described previously.15 Briefly, patients with CHC who had failed to achieve an SVR after treatment with interferon with or without ribavirin and who had advanced fibrosis on liver biopsy (Ishak fibrosis score >3), with no history of hepatic decompensation or hepatocellular carcinoma (HCC) were treated with peginterferon alfa-2a and ribavirin for 6 months (the lead-in phase of the trial). Patients who remained viremic during the lead-in phase of treatment (lead-in patients), those who experienced virological breakthrough or relapse after initial response (breakthrough/relapser patients) and those who were non-responders to peginterferon and ribavirin outside of the HALT-C trial (express patients) were randomized to maintenance therapy (peginterferon alfa-2a 90 μg weekly) or to remain as untreated controls for the next 3.5 years. Liver biopsy was performed within 12 months prior to entry into the trial and then at 2 and 4 years following enrollment. Hepatic necroinflammation was scored using the histology activity index (HAI) scale (0-18) and hepatic fibrosis using the Ishak scoring system (0-6).16 In the HALT-C trial hepatic steatosis was graded as 0 (<1%), 1 (1%-5%), 2 (5%-33%), 3 (33%-67%), and 4 (>67%) according to the percentage of hepatocytes with fat. In the NIH cohort, hepatic steatosis was graded on a 6-point scale as none, <5%, 5% to 25%, 26% to 50%, 51% to 75%, and 76% to 100% based on the proportion of hepatocytes with fat. For this analysis, the scoring scales were merged to none (<1% / 0%), trace (1-5% / trace), mild (6-33% / trace to 25%) and moderate to severe (all other categories).
All biopsies from the NIH cohort were read and scored by one, expert hepatopathologist; biopsies from the HALT-C cohort were read and scored by a panel of ten hepatopathologists, one of whom included the NIH hepatopathologist. Alcohol use was categorized as none, social drinking ≤ 3 drinks per day and heavy alcohol consumption as > 3 drinks per day. Patients were excluded if they had another cause for liver disease or if essential clinical or laboratory data were missing. None of the patients included in this analysis experienced an SVR.
Rationale for Combining the NIH and HALT-C cohorts
The primary purpose for combining the two cohorts was to ensure representation of all stages of fibrosis in the analysis. We acknowledge the fact that the two cohorts are inherently different: the NIH cohort was untreated and had less severe fibrosis compared to the HALT-C cohort, while the HALT-C had all failed treatment previously and had advanced fibrosis. To control for this difference, all analyses were performed with the cohorts combined as well as separately to determine if the results were similar, supplementary Table S1.
Primary outcome analyses
Three primary outcome analyses were conducted
1) Cross-sectional analysis
IL28B genotypes were compared with liver biopsy grading and staging using the initial liver biopsy from the NIH cohort (n=246) and the entry biopsy for the HALT-C cohort (n=1237).
2) Longitudinal analysis
To determine whether IL28B genotype was associated with fibrosis progression, we correlated the change in fibrosis score between biopsies with IL28B genotype. For this analysis we included only patients with paired liver biopsies a minimum of 1 year apart (and no more than 10 years apart) and absence of cirrhosis on the baseline biopsy since these patients could not progress. Patients from the HALT-C cohort who were randomized to receive low-dose peginterferon alfa-2a between liver biopsies were also excluded from this analysis as were breakthrough/relapsers because they received therapy beyond the lead-in phase of the trial (24 weeks of peginterferon alfa-2a and ribavirin). Based on these parameters, 168 patients from the NIH cohort were excluded: 161 because they did not have two biopsies within the required time period and 7 with cirrhosis at entry, Figure 1. One thousand and thirty-nine patients from the HALT-C cohort were excluded: 306 who were not randomized, 134 randomized breakthrough/relapsers, 397 randomized to long-term peginterferon alfa-2a, 66 with no follow-up liver biopsy and 136 with cirrhosis on entry biopsy, Figure 1. Thus, data from a total of 276 patients were included, 78 from the NIH cohort and 198 from the HALT-C cohort. The 78 patients in the NIH cohort were never treated and 30 express patients in the HALT-C cohort received no treatment between biopsies. The remaining HALT-C patients were treated for 24 weeks after the first biopsy. The median duration between biopsies was 4 years, (range 1.7 to 9.9 years). To limit the issue of biopsy sampling error, fibrosis progression was defined as a 2-point increase in Ishak score between the paired biopsies. Patients were divided into two groups: 1) patients with no fibrosis progression, defined as difference in the Ishak score of <2 between the biopsies 2) patients with fibrosis progression, defined as 2 or greater increase in the Ishak score between biopsies.
Figure 1.
Derivation of the 3 analyses Cohorts: for the baseline analysis: There were 1382 subjects enrolled in the HALT-C trial and 309 followed at NIH. 145 subjects from the HALT-C cohort were excluded either for missing IL28B results n=143 or missing histology data n=2 and 63 subjects from the NIH cohort were excluded either for missing IL28B results n=58 or missing histology data n=5, leaving 1483 for analysis.*Clinical outcome cohort: Only subjects randomized to the control arm of the HALT-C trial because data on the clinical outcomes were prospectively collected over 3.85 years. This was not done for the NIH cohort. **Paired biopsy cohort: For this analysis we included only patients with paired liver biopsies a minimum of 1 year apart (and no more than 10 years apart) and absence of cirrhosis on the baseline biopsy.
3) Clinical outcomes analysis
For this analysis, only subjects from the control arm of HALT-C cohort, n=400, were included because data on the clinical outcomes were prospectively collected over 3.85 years and adjudicated by a panel of 3 principal investigators using stringent criteria to confirm that a clinical event had indeed occurred. A clinical outcome was defined as one of the following: death, development of ascites, spontaneous bacterial peritonitis, variceal hemorrhage, hepatic encephalopathy, hepatocellular carcinoma and increase in Child-Pugh-Turcotte score by 2 or more points on two consecutive clinic visits 12 weeks apart.
Both studies were approved by the Institutional Review Board of the NIDDK, NIH and both cohorts signed a separate consent form for genetic testing.
IL28B SNP analysis
Genotyping of the rs12979860 SNP was performed on all patients from the HALT-C and NIH cohorts with available DNA samples and who provided genetic consent as previously described,17 and supplementary material.
Statistical analysis
Baseline clinical characteristics and laboratory values of these patients and their relationships to fibrosis were examined. Variables analyzed included demographic factors including age, sex, race, and ethnicity, anthropometric indices (body mass index, BMI), duration of infection, presence of diabetes and alcohol consumption. The following laboratory and histological tests were included: serum alanine aminotransferase (ALT), aspartate aminotransferase (AST) levels, alkaline phosphatase, total bilirubin, albumin, prothrombin time, platelet count, ferritin and hepatic steatosis. Baseline variables were compared using chi-square, t-test, or analysis of variance. Logistic regression was used to calculate odds ratios for the relationship between fibrosis progression and IL28B (CC versus CT or TT). Analyses of the combined cohorts included a variable indicating cohort (NIH or HALT-C). Other predictors of fibrosis progression were evaluated and those significant after backward selection were also included in the model. Change in fibrosis, HAI, and ALT were analyzed using an analysis of variance controlling for baseline levels. Clinical outcome rates were estimated using Kaplan-Meier estimates and significance was tested using the log-rank test and Cox proportional hazards regression. Analyses were conducted by cohort and with both cohorts combined. Data are presented as percent or mean and standard deviation (SD) unless otherwise noted. SAS® (Statistical Analysis Software, Cary, NC) version 9.2 was used for statistical analyses.
Results
Patient demographics and Baseline Characteristics
A total of 309 patients were followed in NIH natural history studies and 1382 patients were enrolled into the HALT-C trial. Sixty-three patients were excluded from the NIH cohort because of missing IL28B data in 58 and missing histology data in 5 patients. One-hundred and forty-five patients were excluded from the HALT-C trial because of missing IL28B data in 143 and missing histology data in 2 patients, Figure 1. Thus, a total of 1483 patients were available for the cross-sectional analysis, of whom 246 were from the NIH cohort and 1,237 from the HALT-C cohort.
The baseline demographic, laboratory and histologic data are shown in Table 1. Overall, the mean age was 49 years, 69% were male, 75% were white and the mean duration of infection was 28 years. In total, 21% of subjects had diabetes, 19% were teetotalers and 25% had heavy alcohol consumption. The mean body mass index (BMI) was 30 and 86% were infected with HCV genotype 1. At baseline, the mean serum ALT was 112 U/L, mean serum AST was 87 U/L and mean serum albumin, total bilirubin and platelet count were within normal limits (Table 1). The distribution of IL28B genotypes was as follows: 25% of patients had CC genotype, 53% CT, and 22% TT. The mean Ishak fibrosis and HAI scores were 3.8 and 7.7 respectively; one-third of patients had cirrhosis at baseline, hepatic steatosis was absent or mild in 60% and 9% had moderate-severe steatosis. No subject had concomitant non-alcoholic steatohepatitis.
Table 1.
Baseline Characteristics of the NIH and HALT C studies
| Variable | NIH | HALT C | Overall | p value** NIH vs. HALT- C | |
|---|---|---|---|---|---|
| N Patients | 246 | 1237 | 1483 | NA | |
| Age* | 45.4, 9.0 | 49.9, 7.1 | 49.2, 7.7 | <0.0001 | |
| Male, % | 57.6 | 71.8 | 69.4 | <0.0001 | |
| Race | <0.0001 | ||||
| White, % | 76.6 | 74.3 | 74.7 | ||
| African American, % | 16.4 | 15.4 | 15.5 | ||
| Hispanic, % | 0.0 | 8.3 | 6.9 | ||
| Other, % | 7.0 | 2.1 | 2.9 | ||
| Duration of infection, years* | 18, 9 | 28, 8 | 27.9 | <0.0001 | |
| Diabetes, % | 1.1 | 23.8 | 20.9 | <0.0001 | |
| Alcohol History | <0.0001 | ||||
| None, % | 39.2 | 16.5 | 18.6 | ||
| Social, % | 29.6 | 59.6 | 56.8 | ||
| Heavy, % | 31.2 | 23.9 | 24.6 | ||
| BMI* | 27.7, 5.4 | 29.8, 5.5 | 29.5, 5.5 | <0.0001 | |
| IL28B | 36.2 | 22.7 | 25.0 | <0.0001 | |
| CC, % | 43.1 | 54.5 | 52.6 | ||
| CT, % | 20.7 | 22.8 | 22.4 | ||
| TT, % | |||||
| HCV genotype 1, % | 68.3 | 88.8 | 85.5 | <0.0001 | |
| ALT* | 118, 111 | 111, 79 | 112, 84 | 0.34 | |
| AST* | 77, 75 | 88, 61 | 87, 63 | 0.045 | |
| Albumin* | 4.2, 0.34 | 3.9, 0.40 | 3.9, 0.40 | <0.0001 | |
| Alkaline phosphatase* | 79, 25 | 99, 44.0 | 96, 43 | <0.0001 | |
| Platelets* | 208, 61 | 167, 65 | 173, 66 | <0.0001 | |
| Total bilirubin* | 0.74, 0.33 | 0.79, 0.42 | 0.78, 0.41 | 0.086 | |
| Serum ferritin* | 169, 199 | 354, 391 | 338, 382 | <0.0001 | |
| Log serum ferritin* | 2.0, 0.41 | 2.3, 0.46 | 2.3, 0.46 | <0.0001 | |
| Fibrosis | <0.0001 | ||||
| 0-2, % | 56.9 | 8.3 | 16.3 | ||
| 3-4, % | 33.3 | 53.1 | 49.8 | ||
| 5-6, % | 9.8 | 38.6 | 33.9 | ||
| Fibrosis* | 2.3, 1.7 | 4.1, 1.3 | 3.8, 1.5 | <0.0001 | |
| HAI* | 8.6, 2.8 | 7.6, 2.1 | 7.7, 2.3 | <0.0001 | |
| Lobular inflammation* | 3.3, 1.0 | 2.6, 0.9 | 2.7, 1.0 | <0.0001 | |
| Portal inflammation* | 2.2, 1.2 | 2.2, 0.8 | 2.2, 0.8 | 0.76 | |
| Periportal inflammation* | 3.0, 1.6 | 2.8, 1.0 | 2.8, 1.1 | 0.029 | |
| Steatosis*** | <0.0001 | ||||
| None | 40.2 | 18.4 | 22.0 | ||
| Trace | 15.9 | 42.0 | 37.6 | ||
| Mild | 36.6 | 30.4 | 31.4 | ||
| Moderate to severe | 7.3 | 9.3 | 9.0 | ||
| Steatosis* | 1.1, 1.0 | 1.3, 0.9 | 1.3, 0.9 | 0.005 | |
Mean, SD
P values from analysis of variance or chi-square test (not adjusting for study)
First grouping was used by HALT-C, grouping in () was used by NIH
Patients in the HALT C trial were more likely to be older (50 versus 45 years), male (72% versus 58%), of Hispanic race (8.3% versus 0%), have longer duration of infection, 29 versus 19 years), have a higher BMI (30 versus 28), a higher prevalence of diabetes (24% versus 1%), more likely to consume alcohol, and be infected with HCV genotype 1 (89% versus 68%) compared to the NIH cohort. Serum AST (but not ALT), alkaline phosphatase, and ferritin were higher and serum albumin and platelet counts were lower in the HALT-C cohort as compared to the NIH cohort. The HALT-C cohort had a lower frequency patients with IL28 genotype CC compared to the NIH cohort, 22% versus 36%, (p<.0001). The NIH patients had a higher mean HAI score (8.6 vs. 7.6; P <0.0001), higher lobular and periportal inflammatory scores and a greater proportion of subjects with no hepatic steatosis compared to the HALT-C cohort but the mean fibrosis score was significantly higher in the HALT-C cohort compared to the NIH cohort (4.1 vs. 2.3, P<0.0001). Fifty-seven percent of the NIH cohort had mild disease, Ishak fibrosis scores ranging between 0-2, compared to only 8% of HALT-C patients. In contrast, 39% of the HALT-C cohort had cirrhosis (Ishak 5-6) compared to 10% of the NIH cohort.
Cross-sectional Analysis
Association of Baseline Clinical and Laboratory Factors with IL28B genotype
White patients had a higher frequency of IL28b genotypes CC and CT compared to TT, 79% and 79% vs. 61%, respectively, p<.0001, Table 2. Conversely, African-American patients had a higher frequency of IL28B genotype TT compared to CC and CT, 31% vs. 6% and 14%, respectively (p<0.001). Interestingly, patients with the IL28B CC genotype were less likely to be infected with HCV genotype 1 than the CT and TT genotypes (76% vs. 87% vs. 93%, respectively, P<0.0001), table 2, figure 2. Serum ALT, a surrogate marker of inflammation, was higher in patients with the CC genotype compared to patients with non-CC genotype.
Table 2.
IL28B genotype with Clinical, Laboratory and Histological Characteristics n=1483 (NIH and HALT-C Cohorts Combined)
| Variable | CC | CT | TT | p value** 3 groups | p value** CC vs. not CC |
|---|---|---|---|---|---|
| N Patients | 370 | 780 | 333 | NA | NA |
| Age* | 49.5, 7.4 | 48.9, 7.6 | 49.5, 8.2 | 0.28 | 0.33 |
| Male, % | 72.2% | 70.5% | 64.0% | 0.041 | 0.19 |
| Race | <0.0001 | <0.0001 | |||
| White, % | 78.9 | 78.7 | 60.7 | ||
| African American, % | 6.0 | 13.6 | 30.6 | ||
| Hispanic, % | 7.0 | 6.4 | 7.8 | ||
| Other, % | 8.1 | 1.3 | 0.9 | ||
| Duration of infection, years* | 27.1, 9.2 | 27.0, 8.4 | 26.5, 8.9 | 0.71 | 0.68 |
| Diabetes, % | 17.3 | 22.8 | 20.8 | 0.11 | 0.057 |
| Alcohol History | 0.53 | 0.59 | |||
| None, % | 17.3 | 17.9 | 21.6 | ||
| Social, % | 56.3 | 57.5 | 55.7 | ||
| Heavy, % | 26.4 | 24.7 | 22.6 | ||
| BMI* | 29.3, 5.6 | 29.8, 5.5 | 29.3, 5.4 | 0.28 | 0.31 |
| HCV Genotype | <0.0001 | <0.0001 | |||
| 1, % (n=1263) | 76.4 | 86.9 | 92.5 | ||
| 2, % (n=97) | 12.0 | 5.7 | 2.7 | ||
| 3, % (n=91) | 10.1 | 5.5 | 3.3 | ||
| Other, % (n=26) | 1.6 | 1.9 | 1.5 | ||
| ALT* | 133, 113 | 107, 73 | 100, 66 | <0.0001 | <0.0001 |
| AST* | 93, 83 | 84, 548 | 86, 56 | 0.068 | 0.065 |
| Albumin* | 4, 0.4 | 3.9, 0.4 | 3.9, 0.4 | 0.0045 | 0.017 |
| Alkaline phosphatase* | 94, 44 | 97, 43 | 98, 41 | 0.43 | 0.25 |
| Platelets* | 171, 65 | 174, 66 | 172, 67 | 0.82 | 0.67 |
| Total bilirubin* | 0.7, 0.4 | 0.8, 0.4 | 0.8, 0.5 | 0.18 | 0.38 |
| Serum ferritin* | 257, 270 | 354, 409 | 384, 404 | <0.0001 | <0.0001 |
| Log serum ferritin* | 2.2, 0.4 | 2.3, 0.5 | 2.4, 0.5 | <0.0001 | <0.0001 |
| Fibrosis | 0.076 | 0.046 | |||
| 0-2, % | 20.3 | 15.8 | 13.2 | ||
| 3-4, % | 48.9 | 48.7 | 53.5 | ||
| 5-6, % | 30.8 | 35.5 | 33.3 | ||
| Fibrosis* | 3.6, 1.6 | 3.8, 1.5 | 3.8, 1.5 | 0.070 | 0.021 |
| HAI* | 8.0, 2.5 | 7.7, 2.2 | 7.6, 2.2 | 0.015 | 0.0069 |
| Lobular inflammation* | 2.7, 1.0 | 2.7, 0.9 | 2.8, 0.9 | 0.43 | 0.84 |
| Portal inflammation* | 2.4, 0.9 | 2.2, 0.8 | 2.1, 0.8 | <0.0001 | <0.0001 |
| Periportal inflammation* | 2.9, 1.2 | 2.8, 1.1 | 2.7, 1.0 | 0.060 | 0.025 |
| Steatosis*** | <0.0001 | <0.0001 | |||
| None | 34.3 | 18.1 | 17.4 | ||
| Trace | 34.1 | 39.4 | 37.5 | ||
| Mild | 24.9 | 32.1 | 37.2 | ||
| Moderate to severe | 6.8 | 10.5 | 7.8 | ||
| Steatosis* | 1.0, 0.9 | 1.4, 0.9 | 1.4, 0.9 | <0.0001 | <0.0001 |
Mean, SD
P values from analysis of variance or chi-square test (not adjusting for study)
First grouping was used by HALT-C, grouping in () was used by NIH
Figure 2.
Frequency of IL28B genotypes by HCV viral genotype (Cross-sectional analysis): 1483 patients were included in the cross-sectional analysis.
Association of IL28B with Histological Severity
On cross-sectional analysis, subjects with IL28B CC genotype had lower Ishak fibrosis scores compared to those with IL28B CT and TT genotypes combined, 3.6 vs. 3.8, p value = 0.021 but this was not significant when all three groups were compared, 3.60 vs. 3.80 vs. 3.81, p value= 0.07, table 2 and supplementary figure 1, or when the cohorts were analyzed separately, supplementary table S1. Subjects with IL28B CC genotype had higher total HAI scores as well as peri-portal and portal necroinflammation compared to non-CC genotypes, table 2. Subjects with IL28B CC genotype had milder hepatic steatosis compared to CT and TT, 1.0 vs., 1.4 and 1.4 respectively, p=<.0001 and were more likely to have no hepatic steatosis. Similar results were obtained when the two cohorts were analyzed separately, supplementary Table S1.
Longitudinal Analysis
This analysis examined the association between IL28B genotype with fibrosis progression based upon paired biopsies from the untreated NIH cohort (n=78) and the control arm of HALT-C trial, (n=198). Baseline characteristics of the two cohorts were different in several aspects (supplemental tables S2 and S3). Compared to the HALT-C cohort, subjects in the NIH cohort were younger, more likely to be female, white, have a shorter duration of infection, less likely to have diabetes or consume alcohol and have laboratory and histological features consistent with the presence of milder liver disease, supplemental table S2. The IL28B CC genotype was twice as frequent in the NIH cohort compared to the HALT-C cohort, 26% vs. 13%, respectively, p value= 0.02. Overall, the distribution of IL28B genotypes was 17% CC, 54% CT and 30% TT.
Association of IL28B Genotype with Histological Progression
Overall, a two-point increase in Ishak fibrosis score was observed in 60/276 (22%), table 3a. Progression of fibrosis occurred more frequently in the HALT-C cohort compared to the NIH cohort, p=0.0037. There was no difference in the frequency of fibrosis progression between patients with IL28B genotype CC (17%) and non-CC (23%), both in unadjusted analysis and after adjusting for baseline platelets, alkaline phosphatase and hepatic steatosis, p value =0.51. The mean change in Ishak fibrosis scores was 0.41 among patients with IL28B CC genotype and 0.51 among those with IL28B CT or TT genotype, p =0.95, table 4, supplementary figure 2. Results were similar when the cohorts were analyzed separately, HALT-C (0.46 versus 0.58, p = 0.70) and NIH cohorts (0.35 versus 0.33, p = 0.60), Table 4.
Table 4.
Change in mean fibrosis, HAI Scores and Serum ALT Level between Biopsies by IL28B Genotype and Cohort
| Variable* | P value Study** | P value IL28B** | P value interaction** | |
|---|---|---|---|---|
| Ishak fibrosis score | ||||
| ALL | ||||
| CC (n=46) | 0.41, 1.26 | 0.08 | 0.95 | 0.55 |
| CT/TT (n=230) | 0.51, 1.25 | |||
| NIH | ||||
| CC (n=20) | 0.35, 1.04 | NA | 0.60 | NA |
| CT/TT (n=58) | 0.33, 0.87 | |||
| HALT-C | ||||
| CC (n=26) | 0.46, 1.42 | NA | 0.70 | NA |
| CT/TT (n=172) | 0.58, 1.35 | |||
| HAI score | ||||
| ALL | ||||
| CC (n=46) | -0.13, 2.91 | 0.0069 | 0.53 | 0.060 |
| CT/TT (n=230) | 0.49, 2.74 | |||
| NIH | ||||
| CC (n=20) | 0.10, 2.99 | NA | 0.40 | NA |
| CT/TT (n=58) | 0.71, 3.18 | |||
| HALT-C | ||||
| CC (n=26) | -0.31, 2.90 | NA | 0.30 | NA |
| CT/TT (n=172) | 0.42, 2.58 | |||
| ALT U/L | ||||
| ALL | ||||
| CC (n=46) | -51.8, 105.1 | 0.79 | 0.19 | 0.37 |
| CT/TT (n=230) | 2.7, 75.9 | |||
| NIH | ||||
| CC (n=20) | -84.4, 125.8 | NA | 0.35 | NA |
| CT/TT (n=58) | 11.8, 114.4 | |||
| HALT-C | ||||
| CC (n=26) | -33.6, 89.2 | NA | 0.64 | NA |
| CT/TT (n=172) | -0.1, 59.8 | |||
Mean, SD
P values are from an analysis of variance with factors for study, IL28B (CC vs. CT & TT), study and IL28B interaction, and fibrosis, HAI, or ALT at baseline.
Association of IL28B Genotype with Change in HAI Score and Serum ALT Level
We also explored the relationship between change in HAI scores and serum ALT level between liver biopsies with IL28B genotype. At baseline, patients with IL28B genotype CC had higher total HAI scores as well as portal and periportal inflammatory scores. Serum ALT and AST levels were higher and log serum ferritin lower at baseline in patients with IL28B genotype CC compared to those with genotypes CT/TT, supplementary table 2. HAI scores and serum ALT levels improved between biopsies in patients with IL28B CC genotype (mean change -0.13 and -52 U/L, respectively) compared to IL28B non-CC genotype (mean change 0.49 and 3 U/L, respectively) but these differences were not significant, table 4, Supplementary figures 3 and 4.
Factors associated with fibrosis progression
In a logistic regression model to identify factors associated with a two-point increase in Ishak fibrosis score, low platelet count, elevated alkaline phosphatase and more severe hepatic steatosis at baseline liver biopsy were the best predictors of fibrosis progression. IL28B genotype non-CC versus CC was not significantly associated with fibrosis progression and addition of IL28B genotype to the model did not improve the fit.
Clinical Outcomes Analysis
To determine whether IL28B genotype was associated with clinical outcome, we restricted the analysis to the untreated HALT-C cohort who were prospectively observed every 3 months for 3.85 years for the development of clinical outcomes (death n=7, ascites n=7, spontaneous bacterial peritonitis n=1, variceal hemorrhage n=3, hepatic encephalopathy n=6, hepatocellular carcinoma n=11 and increase in Child-Turcotte-Pugh score by ≥2 points at 2 consecutive study visits n=37, total events n=72). There was no histological requirement for this analysis, thus, 400 subjects who were randomized to the control arm of HALT-C were analyzed and included 50 IL28B genotype CC and 350 IL28B genotype CT/TT subjects. Interestingly, untreated subjects with IL28 CC genotype were twice as likely to develop an adverse clinical outcome compared to subjects with IL28B non-CC genotype by life table analysis, 32% versus 16% respectively, p value= 0.003, table 3b and figure 3. This difference became apparent within the first 6 months of randomization and continued to increase over the period of follow-up. This finding remained significant after adjusting for presence of diabetes at baseline, cirrhosis, albumin and bilirubin levels, p value= 0.004. The result was also independent of fibrosis stage and observed in those with baseline cirrhosis and bridging fibrosis on baseline biopsy.
Figure 3.
Life table analysis of clinical outcomes by IL28B genotypes: Clinical outcomes were defined as death, ascites, spontaneous bacterial peritonitis, variceal hemorrhage, hepatic encephalopathy, hepatocellular carcinoma and increase in Child-Turcotte-Pugh score by ≥2 points at 2 consecutive study visits) in 400 untreated HALT-C subjects prospectively observed for 3.8 years.
Discussion
Several recent studies that have examined the association between IL28B genotype and disease severity (hepatic fibrosis and necro-inflammation) in patients with CHC have yielded contradictory results.11-13, 18-20 Some studies have shown an association between IL28B rs12979860 genotype CC (or rs 809917 TT) with more advanced fibrosis or cirrhosis,12, 20 and some have shown an association of the minor, IL28B rs12979860 genotype TT (or rs 809917 GG) with more advanced fibrosis or cirrhosis,11, 18, 19 while other studies have reported no association of IL28B genotype with fibrosis.13 There may be multiple reasons for the discrepant results including small sample size in some studies, different patient cohorts, sample bias, study design-cross-sectional versus longitudinal and imprecise estimates of fibrosis progression. Two studies determined a fibrosis progression rate by calculating the ratio of fibrosis stage to the estimated duration of infection.13, 20 This method assumes that fibrosis progression occurs linearly over time. Using this approach, one study showed no association between IL28B genotype and fibrosis progression13 while the other suggested that the IL28B rs 809917 GG genotype was associated with a slower rate of fibrosis progression, particularly in patients with non-1 HCV genotype.20
In this analysis, when fibrosis progression was assessed using a stringent definition of a 2-point worsening in Ishak score between paired biopsies, we found no association between IL28B genotype with fibrosis progression after controlling for baseline platelet count, alkaline phosphatase and hepatic steatosis. This result strongly suggests no association between IL28B genotype and fibrosis progression.
A significant finding of this analysis was the observation that HALT-C subjects with IL28B genotype CC, who received no treatment beyond the 24 week lead-in period, had twice the rate of adverse outcomes when compared to subjects with IL28B genotype non-CC. This finding was present even after controlling for baseline factors associated with a poor outcome. We speculate that prior non-responders with the IL28B CC genotype may have had a worse outcome than non-responders with IL28B non-CC genotypes due to a more vigorous immune response that was insufficient to result in viral clearance, but sufficient to cause greater liver cell injury as evidenced by greater hepatic necroinflammation and serum ALT levels. Two other studies have noted an association of greater necroinflammation and higher serum ALT levels in patients with IL28B genotype CC but neither examined its relationship to clinical outcomes.20, 21 It is also possible that other recently identified IL28B variants that are in linkage disequilibrium with IL28B CC genotype may account for the differences.22, 23
An apparent paradox in this study was that the higher indices of inflammation observed in subjects with IL28B genotype CC were associated with more severe clinical outcomes, but not with fibrosis progression. We offer several possible explanations. First the duration between the paired biopsies (median 4 years) may not adequately capture the rate of fibrosis progression with a small sample size and slow fibrosis progression, resulting in type II error. Second fibrosis progression, though important, might not be the only cause of adverse clinical outcomes in patients with chronic hepatitis C. Indeed, natural history studies have reported that a majority of patients with cirrhosis maintain stable liver disease without clinical decompensation for many years, suggesting that other factors likely contribute to the development of clinical events.24-26 Third, production of selective pro-inflammatory, antiviral cytokines by cells of the innate immune response in patients with IL28B genotype CC, such as interferon gamma could result in more hepatic inflammation and higher rates of viral clearance during early infection, but conversely precipitate clinical decompensation in subjects with late stage disease even in the absence of fibrosis progression.27 Fourth, it is possible that differences in baseline characteristics between the two cohorts may account for some of the associations found with IL28B. However, analyzing the cohorts separately yielded similar findings (Supplementary table S1). Finally, clinical outcomes are much better endpoints of disease progression than fibrosis staging, and therefore should be the more biologically relevant measure of phenotype-genotype correlation.
IL28B genotype CC was associated with a lower frequency of HCV genotype 1 infection and hepatic steatosis compared to IL28B genotypes CT or TT. These appear to be consistent findings among studies that have examined these relationships.11, 28-30 The reason for the association with HCV genotype 1 is not clear but the findings suggests that either HCV non-1 genotypes preferentially infect subjects with IL28B genotype CC or that HCV genotype 1 infections may be more effectively cleared by those with IL28B CC genotype compared to non-1 HCV genotype. It is possible that prior therapy of the HALT-C cohort may have introduced bias into the analysis by enrichment of the HALT-C cohort (predominantly HCV genotype 1) with IL28B non-CC genotypes. However, when we examined the NIH cohort, all of whom were untreated, the same association with IL28B genotype and HCV genotype was found.
This study has several strengths that are worth highlighting. Patients were derived from two pedigreed cohorts that encompassed a wide spectrum of disease severity. Liver biopsies were evaluated by a panel of expert hepatopathologists in the majority of cases and a robust definition of fibrosis progression was used, a 2 point increase in fibrosis score, that should limit the effect of sampling error on liver biopsy. No previous study has used paired biopsies, which we believe is the best approach to explore the relationship between IL28B and disease progression. Subjects in the HALT-C trial were prospectively followed every three months for the development of clinical decompensation and all events were adjudicated by a panel of three rotating investigators before being entered into the trial database. A potential limitation was that all the subjects in the HALT-C cohort had received a prior course of therapy that may have influenced liver histology such that patients with IL28B genotype CC with milder disease may have responded to therapy thereby enriching the cohort with subjects with more severe activity and fibrosis. However, since response to interferon-based therapy is independent of baseline ALT level31 (a surrogate marker of HAI score) this would likely be true for other IL28B genotypes with milder activity and fibrosis. Moreover, the last course of interferon was completed a median of 15 months prior to the baseline biopsy in HALT-C and a minimum of 3.5 years prior to the follow-up biopsy. In addition, therapy was shown to have no influence on fibrosis progression in the HALT-C trial15 and subjects with IL28B genotype CC from the untreated NIH cohort had greater inflammation than subjects with IL28B non-CC genotypes, supplementary table S1.
In summary, we demonstrate that IL28B genotype was not associated with fibrosis progression in patients with CHC. However, the IL28B CC genotype was associated with greater hepatic necroinflammation, higher serum ALT levels and a higher rate of clinical outcomes. This suggests that IL28B genotype CC may be associated with a state of enhanced antiviral immune response that promotes viral clearance and inflammation, but not fibrosis progression. This further suggests that there are mechanisms for fibrogenesis that are independent of inflammation.
Supplementary Material
Table 3.
Fibrosis Progression (2 point increase in Ishak score) and Clinical Outcomes by IL28B genotype and Cohort
| Table 3a | ||||
|---|---|---|---|---|
| n | Fibrosis progression n (%) | Odds Ratio (95% CI) | P Value | |
| All | ||||
| CC | 46 | 8 (17%) | Unadjusted: 0.85(0.37-1.981) Adjusted1: 0.72 (0.28-1.89) |
0.71 0.51 |
| CT/TT | 230 | 52 (23%) | ||
| NIH | ||||
| CC | 20 | 2 (10%) | Unadjusted: 0.96 (0.18-5.21) Adjusted2: 0.38 (0.04-3.70) |
0.97 0.40 |
| CT/TT | 58 | 6 (10%) | ||
| HALT-C | ||||
| CC | 26 | 6 (23%) | Unadjusted: 0.82 (0.31-2.17) Adjusted2: 0.89 (0.30-2.62) |
0.69 0.83 |
| CT/TT | 172 | 46 (27%) | ||
| Table 3b | ||||
|---|---|---|---|---|
| n | Clinical Outcomes n (Cumulative %) | Hazard Ratio (95% CI) | P Value | |
| HALT-C | ||||
| CC | 34 | 16 (32%) | Unadjusted: 2.33 (1.34-4.06) Adjusted3: 2.30 (1.31-4.03) |
0.0029 0.0037 |
| CT/TT | 294 | 56 (16%) | ||
Odds ratios and p values adjusted for cohort and baseline platelets, AP, and steatosis using logistic regression
Odds ratios and p values adjusted for baseline platelets, AP, and steatosis using logistic regression
Hazard ratio and p value adjusted for baseline diabetes, cirrhosis, albumin, and total bilirubin using Cox regression
Acknowledgements
We wish to acknowledge the contributions of Dr. Richard Chen who passed away during revision of this work. We also wish to acknowledge the HALT-C investigators without whom this work could not have been done.
Financial support: This research was supported by the Intramural Research Program of the NIDDK, NIH.
Abbreviations
- CHC
chronic hepatitis C
- SVR
sustained virological response
- HCC
hepatocellular carcinoma
- HALT-C
Hepatitis C Long-Term Treatment Against Cirrhosis
- HAI
histological activity index
- BMI
body mass index
- ALT
alanine aminotransferase
- AST
aspartate aminotransferase
Footnotes
Conflict of interest: No conflicts of interest exist
References
- 1.Liang TJ, Rehermann B, Seeff LB, Hoofnagle JH. Pathogenesis, natural history, treatment, and prevention of hepatitis C. Annals of internal medicine. 2000;132:296–305. doi: 10.7326/0003-4819-132-4-200002150-00008. [DOI] [PubMed] [Google Scholar]
- 2.El-Serag HB. Epidemiology of viral hepatitis and hepatocellular carcinoma. Gastroenterology. 2012;142:1264–1273. e1. doi: 10.1053/j.gastro.2011.12.061. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Alter MJ. Epidemiology of hepatitis C. Hepatology. 1997;26:62S–65S. doi: 10.1002/hep.510260711. [DOI] [PubMed] [Google Scholar]
- 4.Pawlotsky JM. Has genetics eradicated the good old predictors of hepatitis C treatment response? Clinics and research in hepatology and gastroenterology. 2011;35:157–8. doi: 10.1016/j.clinre.2011.03.006. [DOI] [PubMed] [Google Scholar]
- 5.Chung RT, Poordad FF, Hassanein T, Zhou X, Lentz E, Prabhakar A, Di Bisceglie AM. Association of host pharmacodynamic effects with virologic response to pegylated interferon alfa-2a/ribavirin in chronic hepatitis C. Hepatology. 2010;52:1906–14. doi: 10.1002/hep.23947. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ge D, Fellay J, Thompson AJ, Simon JS, Shianna KV, Urban TJ, Heinzen EL, et al. Genetic variation in IL28B predicts hepatitis C treatment-induced viral clearance. Nature. 2009;461:399–401. doi: 10.1038/nature08309. [DOI] [PubMed] [Google Scholar]
- 7.Tanaka Y, Nishida N, Sugiyama M, Kurosaki M, Matsuura K, Sakamoto N, Nakagawa M, et al. Genome-wide association of IL28B with response to pegylated interferon-alpha and ribavirin therapy for chronic hepatitis C. Nature genetics. 2009;41:1105–9. doi: 10.1038/ng.449. [DOI] [PubMed] [Google Scholar]
- 8.Suppiah V, Moldovan M, Ahlenstiel G, Berg T, Weltman M, Abate ML, Bassendine M, et al. IL28B is associated with response to chronic hepatitis C interferon-alpha and ribavirin therapy. Nature genetics. 2009;41:1100–4. doi: 10.1038/ng.447. [DOI] [PubMed] [Google Scholar]
- 9.Thomas DL, Thio CL, Martin MP, Qi Y, Ge D, O'Huigin C, Kidd J, et al. Genetic variation in IL28B and spontaneous clearance of hepatitis C virus. Nature. 2009;461:798–801. doi: 10.1038/nature08463. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Rauch A, Kutalik Z, Descombes P, Cai T, Di Iulio J, Mueller T, Bochud M, et al. Genetic variation in IL28B is associated with chronic hepatitis C and treatment failure: a genome-wide association study. Gastroenterology. 2010;138:1338–45. 1345, e1–7. doi: 10.1053/j.gastro.2009.12.056. [DOI] [PubMed] [Google Scholar]
- 11.Fabris C, Falleti E, Cussigh A, Bitetto D, Fontanini E, Bignulin S, Cmet S, et al. IL-28B rs12979860 C/T allele distribution in patients with liver cirrhosis: role in the course of chronic viral hepatitis and the development of HCC. Journal of hepatology. 2011;54:716–22. doi: 10.1016/j.jhep.2010.07.019. [DOI] [PubMed] [Google Scholar]
- 12.Abe H, Ochi H, Maekawa T, Hayes CN, Tsuge M, Miki D, Mitsui F, et al. Common variation of IL28 affects gamma-GTP levels and inflammation of the liver in chronically infected hepatitis C virus patients. Journal of hepatology. 2010;53:439–43. doi: 10.1016/j.jhep.2010.03.022. [DOI] [PubMed] [Google Scholar]
- 13.Marabita F, Aghemo A, De Nicola S, Rumi MG, Cheroni C, Scavelli R, Crimi M,S, et al. Genetic variation in the interleukin-28B gene is not associated with fibrosis progression in patients with chronic hepatitis C and known date of infection. Hepatology. 2011;54:1127–34. doi: 10.1002/hep.24503. [DOI] [PubMed] [Google Scholar]
- 14.Ghany MG, Kleiner DE, Alter H, Doo E, Khokar F, Promrat K, Herion D, et al. Progression of fibrosis in chronic hepatitis C. Gastroenterology. 2003;124:97–104. doi: 10.1053/gast.2003.50018. [DOI] [PubMed] [Google Scholar]
- 15.Di Bisceglie AM, Shiffman ML, Everson GT, Lindsay KL, Everhart JE, Wright EC, Lee WM, et al. Prolonged therapy of advanced chronic hepatitis C with low-dose peginterferon. The New England journal of medicine. 2008;359:2429–41. doi: 10.1056/NEJMoa0707615. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Ishak K, Baptista A, Bianchi L, Callea F, De Groote J, Gudat F, Denk H, et al. Histological grading and staging of chronic hepatitis. Journal of hepatology. 1995;22:696–9. doi: 10.1016/0168-8278(95)80226-6. [DOI] [PubMed] [Google Scholar]
- 17.Germer S, Holland MJ, Higuchi R. High-throughput SNP allele-frequency determination in pooled DNA samples by kinetic PCR. Genome research. 2000;10:258–66. doi: 10.1101/gr.10.2.258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Di Marco V, Bronte F, Calvaruso V, Capra M, Borsellino Z, Maggio A, Renda MC, et al. IL28B polymorphisms influence stage of fibrosis and spontaneous or interferon-induced viral clearance in thalassemia patients with hepatitis C virus infection. Haematologica. 2012;97:679–86. doi: 10.3324/haematol.2011.050351. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Falleti E, Bitetto D, Fabris C, Cussigh A, Fornasiere E, Cmet S, Fumolo E, et al. Role of interleukin 28B rs12979860 C/T polymorphism on the histological outcome of chronic hepatitis C: relationship with gender and viral genotype. Journal of clinical immunology. 2011;31:891–9. doi: 10.1007/s10875-011-9547-1. [DOI] [PubMed] [Google Scholar]
- 20.Bochud PY, Bibert S, Kutalik Z, Patin E, Guergnon J, Nalpas B, Goossens N, et al. IL28B alleles associated with poor hepatitis C virus (HCV) clearance protect against inflammation and fibrosis in patients infected with non-1 HCV genotypes. Hepatology. 2012;55:384–94. doi: 10.1002/hep.24678. [DOI] [PubMed] [Google Scholar]
- 21.Thompson AJ CP, Zhu M, Zhu Q, Ge D, et al. Genome Wide-Association Study Identifies IL28B Polymorphism to be Associated with Baseline ALT and Hepatic Necro-Inflammatory Activity in Chronic Hepatitis C Patients Enrolled in the IDEAL Study. Hepatology. 2010;52:1220A. [Google Scholar]
- 22.Prokunina-Olsson L, Muchmore B, Tang W, Pfeiffer RM, Park H, Dickensheets H, Hergott D, et al. A variant upstream of IFNL3 (IL28B) creating a new interferon gene IFNL4 is associated with impaired clearance of hepatitis C virus. Nature genetics. 2013;45:164–71. doi: 10.1038/ng.2521. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Urban TJ, Thompson AJ, Bradrick SS, Fellay J, Schuppan D, Cronin KD, Hong L, et al. IL28B genotype is associated with differential expression of intrahepatic interferon-stimulated genes in patients with chronic hepatitis C. Hepatology. 2010;52:1888–96. doi: 10.1002/hep.23912. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Fattovich G, Giustina G, Degos F, Tremolada F, Diodati G, Almasio P, Nevens F, et al. Morbidity and mortality in compensated cirrhosis type C: a retrospective follow-up study of 384 patients. Gastroenterology. 1997;112:463–72. doi: 10.1053/gast.1997.v112.pm9024300. [DOI] [PubMed] [Google Scholar]
- 25.Serfaty L, Aumaitre H, Chazouilleres O, Bonnand AM, Rosmorduc O, Poupon RE, Poupon R. Determinants of outcome of compensated hepatitis C virus-related cirrhosis. Hepatology. 1998;27:1435–40. doi: 10.1002/hep.510270535. [DOI] [PubMed] [Google Scholar]
- 26.Hu KQ, Tong MJ. The long-term outcomes of patients with compensated hepatitis C virus-related cirrhosis and history of parenteral exposure in the United States. Hepatology. 1999;29:1311–6. doi: 10.1002/hep.510290424. [DOI] [PubMed] [Google Scholar]
- 27.Le Bon A, Tough DF. Links between innate and adaptive immunity via type I interferon. Current opinion in immunology. 2002;14:432–6. doi: 10.1016/s0952-7915(02)00354-0. [DOI] [PubMed] [Google Scholar]
- 28.Montes-Cano MA, Garcia-Lozano JR, Abad-Molina C, Romero-Gomez M, Barroso N, Aguilar-Reina J, Nunez-Roldan A, et al. Interleukin-28B genetic variants and hepatitis virus infection by different viral genotypes. Hepatology. 2010;52:33–7. doi: 10.1002/hep.23624. [DOI] [PubMed] [Google Scholar]
- 29.McCarthy JJ, Li JH, Thompson A, Suchindran S, Lao XQ, Patel K, Tillmann HL, et al. Replicated association between an IL28B gene variant and a sustained response to pegylated interferon and ribavirin. Gastroenterology. 2010;138:2307–14. doi: 10.1053/j.gastro.2010.02.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Tillmann HL, Patel K, Muir AJ, Guy CD, Li JH, Lao XQ, Thompson A, et al. Beneficial IL28B genotype associated with lower frequency of hepatic steatosis in patients with chronic hepatitis C. Journal of hepatology. 2011;55:1195–200. doi: 10.1016/j.jhep.2011.03.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Zeuzem S, Diago M, Gane E, Reddy KR, Pockros P, Prati D, Shiffman M, et al. Peginterferon alfa-2a (40 kilodaltons) and ribavirin in patients with chronic hepatitis C and normal aminotransferase levels. Gastroenterology. 2004;127:1724–32. doi: 10.1053/j.gastro.2004.09.050. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.