Clinical Science

Research article

A common polymorphism in the ABCB11 gene is associated with advanced fibrosis in hepatitis C but not in non-alcoholic fatty liver disease

Rika Iwata, Katharina Baur, Bruno Stieger, Joachim C. Mertens, Ann K. Daly, Pascal Frei, Julia Braun, Athanasios Vergopoulos, Felix Stickel, Karim Sabrane, Ina V. Martin, Johannes Schmitt, Oliver Goetze, Chris P. Day, Beat Müllhaupt, Andreas Geier,


Chronic HCV (hepatitis C virus)-associated cirrhosis represents a major indication for liver transplantation. Bile acids contribute to hepatic stellate cell activation as a key event in fibrogenesis. The aim of the present study was to investigate the role of bile acids and polymorphisms in bile acid level-regulating genes on fibrosis progression. A total of 206 subjects with chronic HCV infection were included for ABCB11 (ATP-binding cassette, subfamily B, member II) 1331T>C and NR1H4 (nuclear receptor) −1G>T genotyping, 178 of which were analysed for fibrosis stage. Exclusion criteria were HBV (hepatitis B virus) or HIV coinfection, alcohol >40 g/day and morbid obesity. A total of 358 patients with NAFLD (non-alcoholic fatty liver disease) were genotyped for comparison with a non-viral liver disease. Caucasian individuals (n = 110), undergoing liver resection for focal hepatic metastasis, served as controls. The ABCB11 1331C allele was significantly overrepresented in HCV patients compared with controls {allelic frequency 62.9%; OR (odds ratio), 1.41 [95% CI (confidence interval), 1.012–1.965]}. Median plasma bile acid levels were not significantly increased in the CC compared with TT genotype [7.2 (1–110) μmol/l compared with 3.5 (1–61) μmol/l; values are medians (range). A significant association between the presence of cirrhosis and ABCB11 genotype (CC compared with CT or TT, P=0.047) was observed in the χ2 test and independent of other risk factors of age, gender, body mass index and disease duration in multivariate analysis (P = 0.010). No such association could be observed in fatty liver patients with regard to advanced fibrosis (F≥2). The common ABCB11 1331CC genotype, which is present in 40% of HCV patients and renders the carrier susceptible to increased bile acid levels, is associated with cirrhosis.

  • bile acid transport
  • bile salt export pump (BSEP)
  • cirrhosis
  • farnesoid X receptor (FXR)
  • non-alcoholic fatty liver disease (NAFLD)
  • single nucleotide polymorphism
  • viral hepatitis


With more than 170 million people infected worldwide, chronic HCV (hepatitis C virus) infection represents a leading cause of end-stage liver disease [1]. Mortality associated with chronic hepatitis C results mainly from the progression to cirrhosis with subsequent complications such as chronic liver failure and the evolution of hepatocellular carcinoma. In the United States and Europe decompensated cirrhosis due to chronic hepatitis C infection represents the main indication for liver transplantation [2]. However, the natural history of liver fibrosis progression in patients with chronic HCV infection is quite heterogeneous [3].

Factors such as age, gender, alcohol use, age-at-infection and co-infection with HIV influence the progression to cirrhosis, but cannot accurately predict the individual risk of developing cirrhosis in patients with chronic hepatitis C [4,5]. In addition to these well-characterized cofactors, genetic factors that modulate fibrogenesis are considered as an explanation for the heterogeneous prognosis of chronic hepatitis C. Epidemiological studies have identified SNPs (single nucleotide polymorphisms) in a number of candidate genes that may influence the progression of liver fibrosis in humans [6]. However, some of the studies have yielded contradictory results, and only very few genetic markers involved in fibrosis progression in hepatitis C patients have been reproduced in independent sample sets [6,7]. Recently, a whole-genome-scan approach in patients with chronic hepatitis C infection led to the identification of a seven-gene signature and the calculation of a cirrhosis risk score, which proved to be a better predictor of cirrhosis risk than established clinical factors [8]. However, the identities of these SNPs are not disclosed to the scientific community, so far, thus limiting the clinical applicability to date. Therefore, conclusive studies on candidate genes related to fibrosis progression in well-defined cohorts of patients with HCV infection such as the SCCS (Swiss Hepatitis C Cohort Study) [9] are still warranted today. Ideally, a pathophysiological relevance of these investigated polymorphisms of interest should be documented to strengthen a causal link to disease progression.

Recent observations indicate that hepatocellular accumulation of bile acids in several forms of cholestatic liver injury in either pediatric patients or animal models of obstructive cholestasis contributes to an up-regulation of hepatocyte-derived MCP-1 (monocyte chemotaxis protein-1), which in turn results in hepatic stellate cell recruitment and therefore may represent an early event in liver fibrogenesis [10].

Hepatocellular bile acid levels are determined by the function of the BSEP [bile salt export pump; ABCB11 (ATP-binding cassette, subfamily B, member 11)], an FXR (farnesoid X receptor; NR1H4)-target gene product located at the canalicular membrane of hepatocytes [11]. Functional inhibition of BSEP leads to a reduced bile salt secretion and a decreased bile flow, a condition defined as cholestasis. Mutations in canalicular transporter genes and, in particular, the ABCB11 gene, induce functional defects that either lead to hereditary forms of cholestasis or render the patients more susceptible to acquired cholestasis [11]. BSEP expression levels vary significantly among otherwise healthy individuals, and low BSEP protein expression levels were recently found to be associated with the non-synonymous 1331C allele (444A) of ABCB11 [12]. As an example for the functional consequences, carriers of the 1331T>C polymorphism have an increased susceptibility for intrahepatic cholestasis of pregnancy and contraceptive-induced cholestasis, both of which are clinically characterized by accumulation of bile acids [13]. Functional heterozygous variants in the NR1H4 gene have been shown to affect hepatic FXR-dependent target gene expression [14]. Among other variants, the most frequent polymorphism NR1H4 −1G>T has been recently reported to predispose to intrahepatic cholestasis of pregnancy, thereby, demonstrating the central role of FXR in regulating bile acid homoeostasis [15].

On the basis of these data, the aim of the present study was to investigate a potential role of polymorphisms in bile acid level-regulating genes on fibrosis progression in the SCCS.


Patients and control cohorts

All patients with chronic hepatitis C infection were recruited from the SCCS [9]. Overall, 206 patients with available blood samples and PBMC (peripheral blood mononuclear cell) pellets within 12 months before therapy were included for the assessment of ABCB11 genotype distribution and its correlation with bile acid levels. Further analysis of patients according to their fibrosis stage was performed in 178 patients with available liver biopsy in the context of antiviral therapy (Table 1). All patients were 18 years and over and tested HCV-antibody-positive by immunoblot. Chronic HCV infection was confirmed by detectable HCV RNA in these patients. Exclusion criteria were co-infection with HBV (hepatitis B virus) or HIV, alcohol consumption >40 g/day and morbid obesity [BMI (body mass index) >40 kg/m2]. ALT (alanine aminotransferase) and bilirubin levels were available from standard autoanalyser determinations from the SCCS databank.

View this table:
Table 1 Patient characteristics and clinical parameters of study subjects

Ages are given as medians (range) and ALT levels are given as means±S.D. Clinical characteristics for liver histology and virology only apply for the subset of patients with available liver biopsy.

A total of 110 individuals of Caucasian origin undergoing liver resection for focal hepatic metastasis from extrahepatic malignancies with histologically normal surrounding liver tissue were used as the control group [12]. Cholestatic disease was excluded in all patients. ALT levels were normal in the majority of subjects and mildly elevated in 16%.

As another group of patients with chronic liver disease of non-viral origin, patients with NAFLD (non-alcoholic fatty liver disease) of Caucasian origin (n = 358) with biopsy-proven disease at different stages were recruited from Newcastle Hospitals NHS Foundation Trust, Newcastle upon Tyne, U.K. Clinical characteristics of most cases have been described in detail previously [16], but 52 additional cases recruited recently were also included in the current study. The severity of steatosis, necroinflammatory grade and stage of fibrosis were scored according to modified Brunt criteria [17]. In line with previous data, we considered the presence of stage 2 fibrosis as progressive disease, since stage 1 fibrosis is minor and reflects predominantly pericellular fibrosis. A summary of all of the clinical data is provided in Table 1.

The studies were approved by the local ethics committees following the guidelines of the Declaration of Helsinki. All subjects gave written informed consent.

DNA isolation and genotyping

DNA from the SCCS cohort and healthy controls was isolated from whole blood/PBMC using the QIAamp DNA Minikit (Qiagen) [12]. In these cohorts, genotyping of the SNP ABCB11 1331T>C (V444A, rs2287622) was performed with the TaqMan SNP genotyping assay C_16182459_10, genotyping of the SNP NR1H4 −1G>T with a custom-made TaqMan SNP genotyping assay SNP–FXR3–FXR3 [18] (Applied Biosystems). Allelic discrimination was processed using a TaqMan 7900 HT Fast (Applied Biosystems) sequence detector according to the manufacturer's instructions.

For the NAFLD cohort, DNA was prepared from peripheral blood lymphocytes as described previously [19]. All PCR–RFLP (restriction-fragment-length polymorphism) assays were validated by DNA sequencing of PCR products of known genotype with controls of known genotype and ‘no DNA’ controls included in all runs. For ABCB11 1331T>C, the primers were 5′-CACACAGACACCGAGTATCAACACA-3′ and 5′-CAGGACAGTCTCAATGTATGCTACACCT-3′. A PCR product of 333 bp was obtained, and 5 μl was digested with HaeIII (1 unit) for 16 h. Samples positive for the C allele were digested to give bands of 206 and 127 bp, whereas the T allele was not digested. For FXR −1G>T, the primers were 5′-GCATTCCCACAGTCACAAAC-3′ and 5′-TGAGGAAATGCCTAGATGATGA-3′. This product (5 μl) was digested with FokI at 37 °C for 16 h. Wild-type (G) alleles were digested to bands of 146 and 104 bp, whereas mutant (T) alleles were not digested. All genotyping data were analysed with the SDS 2.3 software.

Bile acid quantification

Plasma bile acid levels were determined from 100 μl plasma diluted in 100 μl of saline using direct spectrophotometry of total bile acids (Trinity Biotech) with an upper reference limit of 8 μmol/l. To correct for fluctuations of plasma bile acid levels between fasting and postprandially, a mean value of sequential samples drawn at random time points was calculated whenever available (n = 79).

Statistical analysis

Statistical analysis was performed using SPSS 15.0. Allelic and genotype frequencies for ABCB11 and NR1H4 in HCV patients compared with healthy controls were first analysed using the χ2 test. Differences between ABCB11 genotypes or fibrosis stage and bile acid levels were tested for significance with the Mann–Whitney U test or Kruskal–Wallis test as appropriate. Differences in ABCB11 genotypes and allelic frequencies between cirrhotic compared with non-cirrhotic patients were calculated using the χ2 test (and Fisher exact test as appropriate). Multivariate analysis was performed using logistic regression. ORs (odds ratios) are given with 95% CIs (confidence intervals). P≤0.05 was considered statistically significant; Bonferroni correction was used if appropriate. Power calculation was performed using PS: Power and Sample Size Calculation Version 3.0.2 [20].

Values are as means±S.D., if not otherwise indicated.


Patient characteristics

A total of 206 patients with chronic hepatitis C infection and 110 individuals undergoing partial hepatectomy for focal liver disease were included in this study. Demographic data and clinical parameters of both groups are summarized in Table 1.

In the hepatitis C patient cohort, a liver biopsy was available for 178 patients, and bile acid levels could be determined for 184 patients. No major differences between males and females were observed for age, HCV genotype distribution and ALT level in the total HCV-infected patient collective. As expected, there was a male preponderance in the hepatitis C patient group. Gender as an established risk factor for fibrosis has been reproduced in our cohort [mean Metavir fibrosis score F 2.00±1.15 (male) compared with 1.82±1.42 (female); P=0.045]. The control group was largely balanced. Liver tissue from partial hepatectomy was available from all individuals and presurgery serum liver parameters from 92 individuals.

Genotype distribution of ABCB11 SNP 1331T>C in hepatitis C patients and healthy controls

SNP genotyping of ABCB11 1331T>C was performed in 206 HCV patients and compared with 110 Caucasian individuals regarding allele frequency (Table 2). The frequency of the C allele (associated with bile acid retention [13]) was slightly but significantly overrepresented in the patient cohort (62.9%) compared with the control group (54.5%; P=0.042). The OR for the presence of C compared with T alleles in HCV patients compared with healthy controls was 1.41 (95% CI, 1.01–1.96) in logistic regression analysis. Similarly, the 1331CC genotype was encountered in 40.3% of HCV patients and in 26.4% of controls (P=0.014). The OR for CC compared with any other genotype was 1.89 (95% CI, 1.14–3.13) in logistic regression analysis.

View this table:
Table 2 Genotype distribution of ABCB11 SNP 1331T>C in hepatitis C patients and healthy controls from the same genetic background with NAFLD patients given separately owing to their different genetic background

P values were calculated using a χ2 test.

The minor allele frequency of the NR1H4 −1G>T polymorphism in both groups (n = 318) amounts only to 2.0%. Nine out of 206 hepatitis C patients and four out of 112 control individuals had a heterozygous NR1H4 −1GT genotype [4.4% compared with 3.6% respectively; P=0.738; OR, 1.61 (95% CI, 0.43–6.25)].

Bile acid levels in hepatitis C patients according to different ABCB11 1331 genotypes and fibrosis stages

To determine whether different ABCB11 1331 genotypes are accompanied by a relevant difference in bile acid levels in HCV patients (n = 184) as described previously for patients with intrahepatic cholestasis of pregnancy [13], plasma bile acid levels were quantified (Figure 1). For patients carrying the homozygous 1331CC genotype, median (range) bile acid levels were 7 (1–110) μmol/l (mean, 13.7±18.8 μmol/l) compared with 4 (1–61) μmol/l (mean, 10.1±13.7 μmol/l) for the 1331TT genotype and 6 (1–60) μmol/l (mean, 10.1±12.1 μmol/l) in heterozygous patients (1331TC genotype) respectively. However, these differences are not statistically significant (P=0.208 for 1331CC compared with TT).

Figure 1 Bile acid levels in hepatitis C patients harbouring different ABCB11 1331 genotypes

Plasma bile acid levels in hepatitis C patients for different ABCB11 1331 genotypes (total, 184 patients; TT, n = 28; TC, n = 86; and CC, n = 70) are given in box plots. Bile acid levels were compared using the Kruskal–Wallis test (P=0.387).

Furthermore, we investigated bile acid levels according to fibrosis stages in those 163 patients with available liver biopsy (Figure 2). Fibrosis stages were categorized in three groups defined as F0 (‘no fibrosis’), F1–3 (‘fibrosis’) and F4 (‘cirrhosis’). Median bile acid levels of cirrhotic patients (n = 25) [23 (1–110) μmol/l] were significantly elevated compared with 5 (1–61) μmol/l in fibrotic patients (n = 115) and 6 (1–18) μmol/l in non-fibrotic patients (n = 23). This difference in plasma bile acid levels between patients with and without cirrhosis is statistically highly significant (P=0.0001).

Figure 2 Bile acid levels in hepatitis C patients according to fibrosis stage

Plasma bile acid levels in hepatitis C patients are given in box plots according to the fibrosis stage. A total of 163 patients were included; n = 23 for no fibrosis, n = 115 for fibrosis and n = 25 for cirrhosis. Bile acid levels of two different stages were compared using the Mann–Whitney U test (P=0.0001 for no fibrosis compared with cirrhosis). After Bonferroni correction P=0.017 was considered as significant.

Fibrosis stage in hepatitis C patients according to ABCB11 1331 genotype and alleles

To determine whether ABCB11 and NR1H4 genotypes, predisposing to a cholestatic phenotype [13,15,2123], are associated with the development of cirrhosis as they were with increased bile acid levels, 163 patients with available liver biopsy were categorized (‘cirrhosis’/F4 compared with ‘no cirrhosis’ defined as F0–3) and genotyped. The distribution of ABCB11 and NR1H4 genotypes across the different stages of fibrosis is shown in Table 3. Owing to the small number of patients in the individual subgroups, no statistically significant differences in the presence of different ABCB11 and NR1H4 genotypes according to fibrosis stage could be identified (Figure 2).

View this table:
Table 3 Distribution of the ABCB11 and NR1H4 genotypes across the different stages of fibrosis in hepatitis C patients

Categorization allowed to identify a cirrhosis rate of 23.6% among patients with the ABCB11 1331CC genotype compared with a mean of 12.3% in all other genotypes (1331TC 12.8% and 1331TT 10.7%, respectively). This association of the ‘cholestatic’ ABCB11 1331CC genotype compared with both of the others (CC compared with TC and TT) is significant (P=0.047) (Figure 3). The OR for the presence of cirrhosis (compared with no cirrhosis) for patients carrying an ABCB11 1331CC genotype is 2.21 (95% CI, 0.998–4.898) in the logistic regression analysis. Regarding the HCV genotype 1 and 2/3 subgroups, there were no significant differences in the prevalence of cirrhosis according to different ABCB11 1331 genotypes. In line with the findings, the presence of the 1331C allele is overrepresented in the cirrhotic patients compared with the non-cirrhotic patients, with a clear trend close to statistical significance (P=0.054). Of note, the genetic association of the ABCB11 1331CC genotype with cirrhosis is highly significant and even more prominent when controlling for the known risk factors of age, gender, BMI, HCV genotype and disease duration [P=0.010; OR, 5.102 (95% CI, 1.465–17.776)] (Table 4). Power analysis indicates an 89.3% power to detect a significant difference with the given number of patients.

Figure 3 Presence of cirrhosis in hepatitis C patients according to ABCB11 1331 genotype distribution and alleles

The frequency of cirrhosis in hepatitis C patients according to ABCB11 1331 genotype distribution and alleles is given. (A) 10.7% (three out of 28 patients) of the patients carrying the 1331TT genotype had cirrhosis, 11.5% (9/78 patients) of patients with 1331TC and 22.2% (16/72 patients) of patients with 1331CC genotype. (B) 11.9% (16 out of 134 alleles) of the T allele and 19.8% (44/222 alleles) of the C allele were detected in hepatitis C patients with cirrhosis. The difference between the 1331CC and any other genotypes (1331CC compared with TT and TC; P=0.047), respectively, the two alleles (C compared with T; P=0.054), was calculated using χ2 test.

View this table:
Table 4 Univariate and multivariate analyses for the presence of cirrhosis in hepatitis C patients

Owing to the low minor allelic frequency of the NR1H4 −1G>T polymorphism in HCV-infected patients (minor allele frequency 2.0%, six heterozygous −1GT in a total of 179 patients; see Table 2), no association between the distribution of the genotypes in relation to fibrosis stage can be observed.

Genetic association of the ABCB11 1331 genotype to fibrosis stage in NAFLD as a distinct chronic liver disease of non-viral origin

To confirm a potential genetic association of these polymorphisms with the presence of cirrhosis in a second independent cohort and to address the question whether the observed bile acid effects on the presence of cirrhosis are HCV-specific or generally apply for chronic liver disease, a cohort of patients with NAFLD as a non-viral disease was analysed accordingly. ABCB11 genotypes for these patients are summarized in Table 2 with data relating both ABCB11 and NR1H4 genotype to fibrosis stage shown in Table 5. As in the previous study on the KLF6 genotype in NAFLD [16], we compared the distribution of ABCB11 genotypes in cases with F0–1 with F≥2. The genotype distributions between low and high fibrosis scores were not significantly different (P=0.470), [52% (25 out of 48 patients) of the patients with the 1331TT genotype, 42% (75 out of 179 patients) of patients with the 1331TC genotype and 43% (56 out of 130 patients) of patients with the 1331CC genotype had advanced fibrosis]. In addition, when genotype distributions for F4 (cirrhosis) cases were compared with all others as a more direct comparison with the hepatitis C analysis, no significant difference was seen (P=0.230). As for the hepatitis C and control groups, the NR1H4 polymorphism was too rare for further analysis.

View this table:
Table 5 Distribution of the ABCB11 and NR1H4 genotypes across the different stages of fibrosis in NAFLD patients


Hepatocellular accumulation of bile acids contributes to an up-regulation of inflammatory cytokines and chemokines such as TNFα (tumour necrosis factor α) and MCP-1, which in turn results in hepatic stellate cell activation as a key event in hepatic fibrogenesis [10,24]. The common ABCB11 1331T>C (V444A) polymorphism in the BSEP gene, which is encountered in about half of the Caucasian population, has been linked to increased serum bile acid levels in patients with cholestasis of pregnancy [13]. The main finding in the present study is an association of this ABCB11 1331T>C polymorphism with progression to liver cirrhosis in a viral liver disease such as chronic hepatitis C with a more than 2-fold carrier risk. This association cannot be observed in NAFLD as a non-viral liver disease. Furthermore, an increased frequency of the 1331C allele in patients with hepatitis C compared with healthy controls also points to a potential causal role for chronicity of the infection.

In previous studies, the ABCB11 1331T>C polymorphism has been associated with increased susceptibility for cholestatic diseases, including ICP (intrahepatic cholestasis of pregnancy) and drug-induced cholestasis [12,13,22,25]. A recent comprehensive analysis of ABCB11 missense mutations and SNPs showed that aberrant pre-mRNA splicing and protein processing/function contribute significantly to BSEP deficiency [23]. Particularly, the ABCB11 1331T>C polymorphism has been associated with reduced expression levels of the mature BSEP protein [12,13,23,25]. Although elevated bile acid levels have been observed in ICP patients [13], we observed only a weak trend towards increased mean bile acid levels in chronic hepatitis C patients carrying the 1331CC genotype (Figure 1). The finding that this genotype is present in 26% of control individuals may explain the fact that the mean bile acid level of the 1331CC group is still within the ‘normal range’ of the general population. Increased bile acid levels were also observed in patients with HCV-induced cirrhosis compared with non-cirrhotic patients (Figure 2), although this may largely be a consequence of cirrhosis. Accordingly, a significant association of the ‘cholestatic’ ABCB11 1331CC genotype compared with both others (CC compared with TC and TT) has been observed (P=0.047). Most importantly, the genetic association of the ABCB11 1331CC genotype with cirrhosis was independent of known risk factors such as age, gender, BMI, HCV genotype and disease duration and even more prominent in multivariate analysis (P=0.010). Combined with the fact that the OR for the presence of cirrhosis is 2.2 in those patients with the 1331CC genotype compared with 1331TC and TT genotypes (Figure 3), these results suggest an association of the ‘cholestatic’ 1331CC genotype with progression of fibrogenesis during chronic hepatitis C infection. However, these data need to be confirmed in a separate cohort of patients. Of note, no genetic association of this polymorphism with the presence of advanced fibrosis and cirrhosis has been observed in a second independent cohort with NAFLD as a non-viral disease. This leads to the speculation that the observed bile acid effects on the presence of cirrhosis could be HCV-specific and may not generally apply for chronic liver disease of other non-viral causes. The absent correlation of HCV viral load and bile acid levels in another study focused on IFN (interferon) therapy response [26] does not support the concept of an activated viral replication as cause of accelerated fibrosis. A potential alternative mechanistic explanation could be the modulation of endogenous inflammatory signalling pathways by bile acids, but a study in 27 liver biopsy samples of patients with chronic HCV infection did not detect significant differences for hepatic IFN-γ, IL (interleukin)-10 and IL-12 mRNA expression for the ‘cholestatic’ ABCB11 1331CC genotype (A. Geier and F. Strickel, unpublished work).

As a member of the nuclear receptor family of transcription factors, FXR is the principal transactivator of ABCB11 [11]. For the most frequent NR1H4 polymorphism −1G>T, we found only nine heterozygote HCV patients in our collective, which is concordant with a minor allele frequency in Caucasians of approx. 2% [15]. Unexpectedly, none of these individuals had cirrhosis, and the mean bile acid levels in −1G>T heterozygous patients were within the normal range. Owing to the limited number of HCV patients for this supplementary genetic analysis, no definite conclusions can be drawn from this part of the results.

For interpretation of our data, it is important to note that all patients were recruited from the SCCS national databank in a retrospective fashion [9]. Several studies demonstrated the variability of histological scoring based on specimen size, location and quality of the liver biopsy and on the experience of the pathologist [2729]. This variability is mostly limited to single stages and has been overcome by categorization of all patients in only two groups, namely cirrhosis compared with no cirrhosis, to minimize the effect of this sampling error. As another inevitable variability caused by the heterogeneity of sampling, plasma samples for bile acid quantification were acquired at different times of the day. To correct for the diurnal variation of bile acid synthesis with two major peaks during daytime [30], we calculated a mean value of sequential samples drawn at random time points whenever available (79 out of 184).

The association of ‘cholestatic’ ABCB11 polymorphisms with cirrhosis development in HCV patients suggests that patients with elevated bile acid levels are an attractive target for therapeutic intervention. Several studies have been carried out in HCV patients either as a short-term treatment with the choleretic UDCA (ursodeoxycholic acid) alone or in combination with antiviral IFN treatment [31]. Changes in liver histology regarding portal fibrosis following UDCA administration are not evident from short-term observation up to 1 year [32,33]. Long-term effects of UDCA therapy in chronic hepatitis C patients are yet to be elucidated. Nevertheless, one retrospective study showed a 50% reduction in the incidence of hepatocellular carcinoma within 5 years from the onset of HCV-related early cirrhosis by UDCA treatment [34].

Interestingly, our data document a small but significant difference in the allelic frequency of the ABCB11 1331C allele in patients with chronic hepatitis C infection compared with healthy subjects (Table 2). This finding may be a hint towards a potential causal role of increased bile acids for a susceptibility to develop chronicity during the acute phase of HCV infection. Although there are no data from the literature to date, this speculation would be in line with an established negative effect of bile acids on IFN signalling in cell culture studies [35].

In summary, we show that the common ABCB11 1331CC genotype, which is present in 40% of HCV patients, is associated with progression to cirrhosis in HCV. Our hypothesis-driven association study establishes this polymorphism as a new and frequent risk factor for progression to cirrhosis in Caucasian patients with chronic hepatitis C and warrants future studies with bile acid-lowering intervention to prevent disease progression in this subset of patients.


This work was supported by the Swiss National Science Foundation (SNF) [grant number 310000–122310/1 (to A.G.)], the Foundation for Research at the Medical Faculty, University of Zurich (to A.G.), and Roche Pharmaceuticals Switzerland (to A.G. and B.M.).


Rika Iwata, Katharina Baur and Ann Daly performed the genotyping and statistical analysis. Bruno Stieger was involved in study design and analysis. Joachim Mertens, Pascal Frei and Felix Stickel recruited the patients. Oliver Goetze and Julia Braun acted as the statistical advisors. Athanasios Vergopoulos performed the bile acid quantification. Karim Sabrane, Ina Martin and Johannes Schmitt supervised the genotyping and contributed to data analysis. Chris Day and Beat Müllhaupt were clinical supervisors and were involved in the study design. Andreas Geier was the clinical supervisor, designed the study and wrote the manuscript.


We thank Joba Arikkat for her excellent technical assistance, Dr Yvonne Meier (formerly University of Zurich) for technological support as well as Dr Franziska Schöni-Affolter (University of Lausanne, CHUV, SCCS Cohort) for her epidemiological advice and logistic support.

Abbreviations: ABCB11, ATP-binding cassette, subfamily B, member 11; ALT, alanine aminotransferase; BMI, body mass index; BSEP, bile salt export pump, CI, confidence interval; FXR, farnesoid X receptor, HCV, hepatitis C virus; ICP, intrahepatic cholestasis of pregnancy; IFN, interferon; IL, interleukin; MCP-1, monocyte chemotaxis protein-1; NAFLD, non-alcoholic fatty liver disease; NR, nuclear receptor; SCCS, Swiss Hepatitis C Cohort Study; SNP, single nucleotide polymorphism; PBMC, peripheral blood mononuclear cell; UDCA, ursodeoxycholic acid


View Abstract