Potential conflict of interest: Dr. Poynard owns stock in Biopredictive. He also consults for and is on the speakers' bureau of Merck. Dr. Pol consults for, advises, and is on the speakers' bureau of Boehringer Ingleheim, Abbott, and GlaxoSmithKline. He consults for, advises, is on the speakers' bureau of, and received grants from Roche, Gilead, Bristol-Myers Squibb, and MSD.
Members of the Swiss Hepatitis C Cohort Study Group: Francesco Negro (Geneva, Chairman), Antoine Hadengue (Geneva, Chairman of Scientific Committee), Laurent Kaiser, Laura Rubbia-Brandt (Geneva); Darius Moradpour, Cristina Cellerai (Lausanne); Martin Rickenbach (Lausanne Data Center); Andreas Cerny, Gladys Martinetti (Lugano); Jean-François Dufour, Meri Gorgievski, Virginie Masserey Spicher (Berne); Markus Heim, Hans Hirsch (Basel); Beat Müllhaupt, Beat Helbling, Stephan Regenass (Zurich); Raffaele Malinverni (Neuchatel); David Semela, Guenter Dollenmaier (St Gallen); Gieri Cathomas (Liestal).
Genetic polymorphisms near IL28B are associated with spontaneous and treatment-induced clearance of hepatitis C virus (HCV), two processes that require the appropriate activation of the host immune responses. Intrahepatic inflammation is believed to mirror such activation, but its relationship with IL28B polymorphisms has yet to be fully appreciated. We analyzed the association of IL28B polymorphisms with histological and follow-up features in 2335 chronically HCV-infected Caucasian patients. Assessable phenotypes before any antiviral treatment included necroinflammatory activity (n = 1,098), fibrosis (n = 1,527), fibrosis progression rate (n = 1,312), and hepatocellular carcinoma development (n = 1,915). Associations of alleles with the phenotypes were evaluated by univariate analysis and multivariate logistic regression, accounting for all relevant covariates. The rare G allele at IL28B marker rs8099917—previously shown to be at risk of treatment failure—was associated with lower activity (P = 0.04), lower fibrosis (P = 0.02) with a trend toward lower fibrosis progression rate (P = 0.06). When stratified according to HCV genotype, most significant associations were observed in patients infected with non-1 genotypes (P = 0.003 for activity, P = 0.001 for fibrosis, and P = 0.02 for fibrosis progression rate), where the odds ratio of having necroinflammation or rapid fibrosis progression for patients with IL28B genotypes TG or GG versus TT were 0.48 (95% confidence intervals 0.30-0.78) and 0.56 (0.35-0.92), respectively. IL28B polymorphisms were not predictive of the development of hepatocellular carcinoma. Conclusion: In chronic hepatitis C, IL28B variants associated with poor response to interferon therapy may predict slower fibrosis progression, especially in patients infected with non-1 HCV genotypes. (HEPATOLOGY 2012)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Hepatitis C virus (HCV) is a major human pathogen responsible for chronic hepatitis that may progress toward cirrhosis and hepatocellular carcinoma (HCC).1 Therapy with peginterferon-alpha and ribavirin is successful in only about half of genotype 1-infected patients, is costly, and burdened with numerous side effects, adding to the HCV-related direct and indirect costs.2 Thus, treatment should be preferentially administered to patients more likely to benefit from it in the long term, i.e., those presenting with features predictive of liver disease progression.3 Baseline and on-treatment factors associated with sustained response to current therapies have been identified and are used to tailor regimens in order to spare drug exposure.4
Recently, genetic polymorphisms near the interleukin-28B (IL28B) gene were reported to be strongly associated with spontaneous5, 6 and treatment-induced clearance of HCV,6-9 although the functional link between IL28B polymorphisms and HCV clearance remains elusive. Nonetheless, the association is meaningful, because IL28B encodes for interferon-λ3 (IFN-λ3), a type III IFN together with IFN-λ1 (encoded by IL29) and IFN-λ2 (encoded by IL28A). Type III IFNs exhibit in vitro10, 11 and in vivo12 antiviral activity against HCV. Although type III IFNs may contribute to host defenses by activating a classical antiviral state through mechanisms similar to, but independent of, type I IFNs,13 most of their antiviral properties depend on the proper stimulation of the host immune system.14 IL28B is capable of establishing a robust T-cell adaptive immune response.15, 16 This may be relevant because a proper activation of the CD8+ response has been shown to predict rapid and sustained virological response to therapy.17 As a consequence, the IL28B polymorphisms associated with viral persistence and poor responsiveness to therapy of HCV infection may be the hallmark of an impaired/inappropriate activation of the adaptive immune response. Because the histological counterpart of this response is believed to be the intrahepatic mononuclear infiltrate, it is intuitive to investigate the association (if any) between IL28B polymorphisms and the presence/degree of inflammatory infiltrate in the liver of chronic hepatitis C patients. Historically, there is evidence linking liver inflammation (often indirectly measured as serum alanine aminotransferase [ALT] levels) and response to therapy,18 although the association is less striking than observed in chronic hepatitis B19 and overshadowed by other, more robust predictors.18 Thus, we analyzed the association of IL28B polymorphisms with the intensity of the necroinflammatory infiltrate in a large population of HCV-infected Caucasian patients enrolled in two large and well-characterized cohorts. Because the intrahepatic grade of necroinflammatory activity is the strongest predictor of fibrosis, we also assessed whether IL28B polymorphisms may be associated with the fibrosis stage and/or, whenever assessable, the fibrosis progression rate and the development of HCC.
ALT, alanine aminotransferase; BMI, body mass index; IFN, interferon; IL28B, interleukin-28B; FPR, fibrosis progression rate; HCC, hepatocellular carcinoma; HCV, hepatitis C virus; SCCS, Swiss Hepatitis C Cohort Study; SVR, sustained virologic response; SNP, single nucleotide polymorphism.
Patients and Methods
Patients were included from the Swiss Hepatitis C Cohort Study (SCCS) and a French Cohort of chronically HCV infected patients. The features and inclusion criteria of the two cohorts have been reported in detail before.20, 21 Patients with chronic HCV infection who provided written consent for genetic analyses were included in the genetic study.
Single nucleotide polymorphisms (SNPs) near IL28B (rs8099917 and rs12979860) were extracted from a genome-wide association (GWA) study-generated dataset6 or genotyped with a TaqMan SNP genotyping assay (Applied Biosystems, Foster City, CA), using the ABI 7500 Fast real time thermocycler. TaqMan probes and primers were designed and synthesized using Applied Biosystems software as described.22 Automated allele calling was performed using SDS software from Applied Biosystems. Positive and negative controls were used in each genotyping assay.
Patients with at least one liver biopsy with fibrosis or activity scoring performed prior to any antiviral treatment were included in the biopsy study. Patients with concomitant liver diseases, including hepatitis B, autoimmune hepatitis, α-1 antitrypsin deficiency, Wilson's disease, or hemochromatosis were excluded. Significant alcohol consumption was defined as >40 g/day over a period of ≥5 years. Liver biopsy specimens were collected from all patients, formalin-fixed, and paraffin-embedded for histological evaluation and analyzed by experienced pathologists. Necroinflammation score was dichotomized as score 0 (absence of necroinflammation or minimal changes) or 1 (presence of mild/moderate/severe necroinflammation).23 Fibrosis stage was scored according to METAVIR24 and dichotomized into two groups (Metavir score F0-1 and Metavir score F2-4). Steatosis was graded as absent or minimal when affecting less than 5% of hepatocytes and present otherwise.25, 26
Statistical analyses were performed using Stata (v. 11.1, StataCorp, College Station, TX). Patients were stratified into two groups of stage-constant fibrosis progression rate (FPR), calculated as the ratio of the fibrosis stage to the estimated duration of infection in years (fibrosis units per year).24, 27 The median rate (0.074 fibrosis units/year) was used as a cutoff. Factors associated with severe FPR (i.e., higher than the median) were analyzed by univariate and multivariate regression. Age, duration of infection, and body mass index (BMI) were treated as continuous variables. Covariates in multiple logistic regression models were selected by backward selection, by using P < 0.2 as a cutoff. For IL28B SNPs, comparisons were made using a dominant model for the rare allele unless otherwise indicated.
The HCC study included only patients from the SCCS, as HCC was an exclusion criterion for the patients of the French cohort. HCC was diagnosed based on the European Association for the Study of the Liver criteria.28 The association between IL28B polymorphisms and HCC was first performed with a case control design by using a logistic regression model. Because this approach does not take into account the time at risk to develop HCC, we calculated the cumulative incidence of HCC for the different IL28B polymorphisms, using the putative infection date as a starting point, with censoring at death or lost follow-up.
The study included 2,335 patients (1,915 from the SCCS and 420 from the French cohort): 1,527 were assessable for the liver histology study and 1,915 for the HCC study (Table 1). Most patients were males (58%) and infected with HCV genotype 1 (57%) (Table 2).
Table 1. Patient Selection
Swiss Hepatitis C Cohort Study (SCCS)
Total number of patients (genetic consent and DNA)
Liver histology study (genotype 1-4 and biopsy before any treatment)
Fibrosis stage and assessable date of infection to estimate the fibrosis progression rate (FPR)
Assessable date of infection (Cox model)
Table 2. Patient Characteristics
Liver Histology Study
Liver Histology Study Patients by Viral Genotype
Numbers are the proportion of patients with the indicated characteristics. HCC, hepatocellular carcinoma.
Date of infection was missing in 215 out of 1,527 liver histology study patients and 323 out of 1915 HCC study patients. IQR, interquartile range. Duration is provided from estimated infection date to the date of biopsy.
For the liver histology study, only patients with genotype 1-4 were selected. HCV genotype was missing (or other than genotype 1-4) in 33 of 1,915 HCC study patients.
HCV RNA was missing in 427 of 1527 liver histology study patients and 28 out of 1,915 HCC study patients.
For the liver histology study, the closest measurement before liver biopsy is reported (data missing in 420 patients); for HCC study and all patient data, the maximal ALT value per patient at any time is considered.
Alcohol consumption data are provided before liver biopsy for the liver histology study patients (data missing in 17 patients) and during any time of follow up for the HCC study patients (data missing in 33 patients).
Diabetes data were missing in 440 liver histology patients.
BMI data are the closest BMI to liver biopsy for liver histology study patients (data missing in 440 patients), and the maximal BMI at any time of follow up for HCC study and all data patients.
HIV serology was missing in 186 liver histology study patients and 337 HCC study patients.
Liver histological activity was missing in 429 patients.
Steatosis data were missing in 556 patients.
rs8099917 data were missing in 9 liver histology study patients and 24 HCC study patients. The P value for HWE was 0.11 (liver histology study).
rs12979860 data were missing in 38 liver histology study patients and 45 HCC study patients. The P value for HWE was 0.58 (liver histology study).
In the liver histology study, necroinflammatory activity was present in 59% of patients and significant fibrosis (≥F2) in 51%. An estimated date of infection was assessable in 1,312 of 1,527 patients, allowing for FPR calculation: the median and mean FPRs were 0.074 and 0.118 fibrosis units per year, respectively, with the fastest progression observed in HCV genotype 3-infected patients (median FPR = 0.087). The demographic and histological characteristics differed in the two cohorts (Supporting Table S1). The French cohort included a lower percentage of males (44% versus 62%), fewer drug users (33% versus 41%), and fewer heavy drinkers (13% versus 19%), but more patients infected with HCV genotype 1 (70% versus 56%). Despite the older age at infection (median 24 versus 19 years), the duration of infection was longer in French patients (median 22 versus 19 years). Overall, these characteristics may explain why the French cohort had a lower percentage of patients with severe fibrosis (≥F2) compared with the SCCS (38% versus 56%).
In the HCC study, 1,593 out of the 1,915 SCCS patients (83.2%) had an estimated date of infection and were assessable for a cumulative incidence analysis. HCC was diagnosed in 62 patients (3%), among whom 43 had an assessable date of infection. The minor allele frequencies (MAFs) of rs8099917(T/G SNP) and rs12979860(C/T) were 0.26 (allele G) and 0.40 (allele T) among the whole study population, respectively, similar to frequencies observed in other Caucasian populations with chronic hepatitis C. Consistent with previous data, the R2 and D' values for linkage disequilibrium between these two SNPs were 0.43 and 0.97, respectively.
Necroinflammatory activity tended to be less pronounced in patients carrying the minor alleles of IL28B SNPs than in patients carrying the major allele, i.e., considering a dominant model for the corresponding rare allele. For rs8099917 the proportion of patients with an activity score of 1 was 0.56 in GT/GG individuals as compared with 0.60 in TT patients (P = 0.12; Table S2). However, important differences were noted when patients were stratified according viral genotypes (Fig. 1A). In patients infected with non-1 genotypes, there was a significantly lower (P = 0.003) proportion of high activity scores in GT/GG (0.53) than in TT subjects (0.67). In genotype-1 infected patients, the association was present only when using the recessive mode of inheritance (P = 0.05). Similar results were observed in multivariate analyses (Table 3), with a significant effect of rs8099917 in the whole sample (odds ratio [OR] for having necroinflammatory activity in GT/GG versus TT subjects = 0.719, 95% confidence interval [CI] 0.528-0.979, P = 0.04). This protective effect was again stronger in patients infected with non-1 genotypes with an OR at 0.482 (95% CI 0.300-0.776, P = 0.003). Similar results, although less significant, were observed for rs12979860: here, the proportion of patients with HCV non-1 genotypes presenting with necroinflammatory activity was 0.58 in CT/TT subjects as compared to 0.67 in CC subjects (P = 0.05; Table S2).
Table 3. Factors Associated with Presence of Liver Necroinflammation
Odd ratios are for the comparison of patients with no or minimal necroinflammation vs. those with any degree of inflammation. All variables associated with the endpoint (P < 0.2) in univariate analyses (age, sex, infection duration, viral load, HCV risk groups, alcohol use, BMI and diabetes) were entered in multivariate models. The final model was obtained by backward selection (P < 0.2), with sex, age and viral genotypes (when appropriate) forced into the model. P values are provided for the dominant model, P values for other models are indicated by a footnote, whenever appropriate.
OR=0.410, 95% CI 0.166-1.016, P = 0.05 when using the recessive model
For rs8099917 the proportion of patients with fibrosis ≥F2 was 0.48 in GT/GG and 0.54 in TT patients (P = 0.03; Table S2). Accordingly, the proportions of rapid progressors (FPR ≥0.077) were 0.46 and 0.54 in GT/GG and TT patients, respectively (P = 0.004; Table S2). Important differences were noted when patients were stratified according to viral genotypes (Fig. 1C,D). The associations between rs8099917 and both fibrosis and FPR were stronger in patients infected with non-1 genotypes, e.g., the proportions of those patients with fibrosis ≥F2 were 0.48 and 0.61 in GT/GG and TT patients, respectively (P = 0.001; Table S2), and the proportions of rapid progressors were 0.52 and 0.62, respectively (P = 0.02). In multivariate analyses, the associations were also highly significant (Tables 4, 5). For GT/GG versus TT patients infected with non-1 genotypes, the OR of developing a fibrosis ≥F2 was 0.431 (95% CI 0.265-0.703, P = 0.001), and the OR of being a rapid progressor was 0.564 (95% CI 0.348-0.916, P = 0.02). As previously noted for necroinflammation similar results, although less significant, were observed for rs12979860, with, as an example, the proportion of fibrosis ≥F2 in patients infected with non-1 genotypes being 0.52 in CT/TT subjects as compared with 0.62 in CC subjects (P = 0.02; Table S2).
Table 4. Factors Associated with Significant Fibrosis
Odd ratios are for the comparison of patients with no or portal fibrosis (stages F0-F1) vs. those with stages F2 to F4. All variables associated with the endpoint (P < 0.2) in univariate analyses (age, sex, infection duration, viral load, HCV risk groups, alcohol use, BMI and diabetes) were entered in multivariate models. The final model was obtained by backward selection (P < 0.2), with sex, age and viral genotypes (when appropriate) forced into the model. P values are provided for the dominant model, P values for other models are indicated by a footnote, whenever appropriate.
Age at infection (years, cont.)
Duration of infection (years, cont.)
Log HCV RNA copies/mm3 (cont.)
Reported risks for HCV infection
Inv. procedure, needle stick
Table 5. Factors Associated with Rapid Fibrosis Progression Rate
Odd ratios are for the comparison of patients with slow progression rate (FPR < median) versus rapid progression (FPR > median). All variables associated with the endpoint (P < 0.2) in univariate analyses (age, sex, viral load, HCV risk groups, alcohol use, BMI and diabetes) were entered in multivariate models. The final model was obtained by backward selection (P < 0.2), with sex, age and viral genotypes (when appropriate) forced into the model. P values are provided for the dominant model, P values for other models are indicated by a footnote, whenever appropriate.
OR = 0.467, 95% CI 0.174-1.252, P = 0.13 when using the recessive model.
Because patient demographic and histological characteristics differed in the two cohorts, we performed stratified analyses to determine whether the effects of IL28B SNPs on fibrosis and its progression were comparable (Fig. S1). Despite different baseline proportions of patients with fibrosis ≥F2, rs8099917 influenced these proportions in both cohorts, although the level of significance after stratification was not reached in the French cohort due to the smaller sample size (Fig. S1A). The SNPs effect differed according to viral genotypes, and this was highly consistent in both cohorts. IL28B rs8099917 did not affect the proportion of fibrosis ≥F2 among genotype 1-infected patients (0.39 in both rs8099917 GT/GG and TT carriers in the French cohort, P = 1.0; 0.53 among GT/GG carriers versus 0.52 among TT carriers in the SCCS, P = 0.8, Fig. S1B). However, rs8099917 influenced these proportions among patients infected with genotypes 3 (0.22 among rs8099917 GT/GG carriers versus 0.37 among TT carriers in the French cohort, P = 0.16, 0.52 versus 0.69 in the SCCS, P = 0.006; Fig. S1C). Similarly, the proportion of rapid fibrosis progressors differed according to IL28B rs8099917 in both cohorts, although significance was not reached in the French cohort (proportion of rapid progressors 0.51 among GT/GG carriers versus 0.42 among TT carriers in the French cohort, P = 0.08; 0.56 versus 0.49 in the SCCS, P = 0.003; Fig. S1D). However, the effect differences in patients infected by HCV genotype 1 (Fig. S1E) versus genotype 3 (Fig. S1F) were less striking than those observed for the fibrosis stage. Similar but less significant results were found for rs12979860 (Fig. S2).
Elevated ALT levels tended to be less frequent among patients carrying the minor alleles of rs8099917 and rs12979860, but none of these differences were significant (Table S2), even after stratification by viral genotypes (Fig. 1B). Finally, no association was detected between IL28B SNPs and the development of HCC (Fig. S1). Among patients with assessable response to treatment, a fibrosis stage ≥F2 was associated with a reduced sustained virologic response (SVR) (OR = 0.553, 95% CI 0.351-0.872, P = 0.01), but necroinflammatory activity (P = 0.7) and FPR (P = 0.7) were not.
Using well-characterized chronic hepatitis C patients from two large cohorts, we showed that IL28B polymorphisms linked to a poor virological response to therapy are protective against liver necroinflammation and fibrosis progression, especially in patients with HCV genotypes other than 1. Previous observations on the role of IL28B polymorphisms with regard to necroinflammatory activity, fibrosis stage, transaminases, or gamma-glutamyl transpeptidase levels were reported from Japan,29 and, in abstract form, from the IDEAL trial performed in the U.S.30, 31 However, these studies were largely29 or exclusively30, 31 limited to patients infected with HCV genotype 1. A more recent work on a limited series of patients failed to show any association between IL28B genotype and FPR.32 In the present study we analyzed the association of IL28B polymorphisms with necroinflammatory activity, ALT, fibrosis stage, and FPR in two large, well-pedigreed series of patients infected with the four most frequent HCV genotypes. Our data show a clear association between the poor treatment-response associated alleles of IL28B and low fibrosis stages as well as slow FPR in patients infected with HCV non-1 genotypes, but not in those infected with genotype 1, in agreement with the IDEAL and the Milan studies.32 A weak association between the poor treatment response allele of rs8099917 and low fibrosis stages was observed in the Japanese study, but the discrepancy may be explained by the contribution of genotype 2-infected patients, who were analyzed together with those infected with genotype 1.29 Most previous studies did not analyze the role of these polymorphisms with FPR, which is a more robust measure of the risk to progress to cirrhosis. Altogether, our data show a significant association between poor treatment-response associated alleles of IL28B and slow FPR in genotype non-1 infected patients.
Our data also show an association between the poor treatment-response associated allele of IL28B and decreased necroinflammatory activity among patients with non-1 genotypes. In genotype 1 patients, a similar association was detected, but only when using the recessive mode of inheritance. This is consistent with the Japanese study, in which the association with rs8099917 was significant when using the recessive mode of inheritance.29 The IDEAL study showed an association between the poor treatment response allele of another marker (i.e., rs12979860) and lower necroinflammatory activity in genotype-1 infected patients, when using the dominant mode of inheritance.30 Differences in the significance of the association may be explained by the use of different study designs (treatment-oriented clinical trial versus cohort study) and sample sizes. Differences in the significance level of rs8099917 and rs12979860 with regard to viral clearance were also reported among studies. Both polymorphisms may be in linkage disequilibrium with another or several other polymorphism(s) with a functional effect. Until such polymorphism(s) has/have been identified, it is difficult to speculate why rs8099917 or rs12979860 may provide stronger associations with any of the above phenotypes. These differences probably result from differences in the statistical power (due to different allele frequencies) and/or the degree of linkage disequilibrium that each SNP has with the functional polymorphism(s). Notwithstanding these apparent discrepancies, these studies establish a link between the poor treatment-response alleles of IL28B and lower necroinflammatory activity. Regarding FPR, our study suggests that this association is stronger in patients infected with HCV non-1 genotypes than in those infected with genotype 1. These data add to the factors influencing FPR in chronic hepatitis C, and may be useful to implement targeted therapeutic interventions in patients at risk of rapid liver disease progression.
Elevated ALT levels tended to be less frequent among patients carrying the minor alleles of IL28B, but the association did not reach significance. In contrast, data from the IDEAL study showed a significant association between these alleles and lower ALT levels in genotype 1-infected patients.30 Necroinflammatory activity and/or elevated ALT levels have sometimes been associated with a good treatment response,33-35 including in the IDEAL study.36 However, necroinflammatory activity is not universally considered a favorable predictor of virological response to therapy of chronic hepatitis C,18 although there is evidence that a strong HCV-specific CD8+ response predicts both a fast viral decline during therapy and SVR.17 In our study, neither necroinflammatory activity nor ALT levels predicted SVR. This hints at a potential heterogeneity of the inflammatory infiltrate and underscores the need for more detailed immunophenotypic analyses using markers specific for the major immune cell subsets, such as CD8, CD4, NK, and especially Treg. It is likely that different profiles of immune cell subpopulations may be better predictors of response outcome than merely the grade of the infiltrate taken as a whole, as suggested by ex vivo analyses.33, 34 Furthermore, ALT levels may be influenced by genetic and metabolic factors and thus they may not necessarily mirror the degree of the immune response.
Paradoxically, our data show that the minor alleles of IL28B (i.e., rs8099917 G and rs12979860 T) can at the same time be unfavorable to the host, by reducing the chances of viral clearance, and favorable, by reducing the degree of liver inflammation and the rate of fibrosis progression in case of viral persistence. Studies showed that the minor alleles of IL28B were associated with reduced expression level of IL28B in peripheral blood mononuclear cells.9 IL28B induces strong adaptive immunity, blunting the Treg responses and stimulating CD8+ cytotoxic T-cell-mediated killing15 and increasing granzyme B expression and perforin release.16 The inflammatory infiltrate of chronic hepatitis C patients is mostly represented by CD8+ T cells,37-42 which are supposed to play a major role in viral containment,39, 43 and are also associated with the severity of the inflammatory infiltrate.42, 43 Thus, IL28B alleles leading to increased IL28B expression may partially revert the inhibition brought about on the HCV-specific CD8+ infiltrate by Tregs. Conversely, IL28B polymorphisms incapable of achieving spontaneous viral resolution would characterize an effector T-cell response that, even in the presence of a dysfunctional and/or exhausted virus-specific response, would be associated with persistent liver damage. This is only one hypothesis, because the effector functions of activated T cells are multifaceted, and may even include cytoprotective effects mediated by IL-22.44 Thus, the definitive immunopathogenetic interpretation of our results can only rely on a thorough phenotypic and/or functional analysis of the T-cell infiltrate.
Our data did not show an association between IL28B polymorphisms and the occurrence of HCC among chronically HCV-infected patients, but the number of patients with HCC was likely insufficient to detect a significant effect, especially as the majority of patients with HCC were infected with HCV genotype 1. Other investigators found that the poor treatment response rs12979860 T allele was associated with HCC in a heterogeneous group of 412 patients with endstage liver cirrhosis due to mixed viral and nonviral etiologies.45 However, the study design did not allow for a specific analysis of the role of IL28B SNPs on the risk of developing HCC among HCV-infected patients. Main limitations included the selection of endstage cirrhotic patients and the lack of analyses stratified according the viral etiologic agent, as the poor treatment response associated rs12979860 T allele was overrepresented among HCV infected patients with endstage cirrhosis compared with noncirrhosis patients, but not in those with cirrhosis due to HBV or other etiology. A Japanese study presented in abstract form showed that the poor treatment response rs8099917 G allele was associated with a higher risk to develop HCC in chronic hepatitis C.46 Large, well-designed prospective studies are warranted to ascertain the role of IL28B polymorphisms in liver carcinogenesis, also in view of the reported antiproliferative effects of IFN-λ in several tumor cell lines.47-51
In conclusion, IL28B variants associated with a poor virological response to therapy are also associated with decreased necroinflammation in the liver and predict slower liver fibrosis progression, especially in patients infected with HCV genotypes other than 1. Our results indicate that IL28B plays a role in influencing the intensity, and very likely the phenotype, of the intrahepatic inflammatory infiltrate, which warrants further immunophenotypical analyses. Independently of its association with the grade of activity, IL28B polymorphisms are an additional host factor with prognostic value of liver disease progression.