Fibrosis in genotype 3 chronic hepatitis C and nonalcoholic fatty liver disease: Role of insulin resistance and hepatic steatosis


  • Potential conflict of interest: Nothing to report.


Hepatic steatosis has been associated with fibrosis, but it is unknown whether the latter is independent of the etiology of fat infiltration. We analyzed the relationship between clinical characteristics, insulin resistance (HOMA-R) and histological parameters in 132 patients with “viral” steatosis caused by genotype 3 chronic hepatitis C (CHC-3) and 132 patients with “metabolic” steatosis caused by nonalcoholic fatty liver disease (NAFLD), matched by age, BMI, and degree of liver fat accumulation. Tests of liver function were comparable in the two study populations. The prevalence of features of insulin resistance was higher in NAFLD, as was HOMA-R (P = .008). Logistic regression analysis confirmed that steatosis was associated with a high viral load and low serum cholesterol in CHC-3, and with high aminotransferase, glucose, ferritin and hypertriglyceridemia in NAFLD. At univariate analysis, advanced fibrosis was associated with steatosis in NAFLD, but not in CHC-3. Other parameters related to fibrosis severity were HOMA-R and a low platelet count in CHC-3, and high aminotransferases, HOMA-R, ferritin and low HDL-cholesterol in NAFLD. On multivariate analysis, only low platelet count (OR = 0.78; 95% CI, 0.67-0.92) and HOMA-R (OR = 2.98; 1.13-7.89) were independent predictors of advanced fibrosis in CHC-3. In NAFLD, severe fibrosis was predicted by fat grading (OR = 3.03; 1.41-6.53), ferritin (OR = 1.13; 1.03-1.25) and HOMA-R (OR = 1.16; 1.02-1.31). In conclusion, insulin resistance is an independent predictor of advanced fibrosis in both NAFLD and CHC-3, but the extent of steatosis contributes to advanced disease only in NAFLD. Virus-induced hepatic steatosis as seen in CHC-3 does not contribute significantly to liver fibrosis. (HEPATOLOGY 2006;44:1648–1655.)

Hepatic steatosis is the hallmark of nonalcoholic fatty liver disease (NAFLD), where liver fat accumulation is caused by increased free fatty acid influx from adipocytes and de novo hepatic triglyceride synthesis,1 related to the underlying insulin resistance.2 Liver steatosis is also common in patients with hepatitis C virus (HCV) infection, where two possible mechanisms of fat accumulation have been suggested: metabolic or viral. In patients with chronic hepatitis C infected with genotypes 1 and 2, steatosis is associated with clinical risk factors for NAFLD (obesity, diabetes, dyslipidemia, hypertension and the metabolic syndrome)3 and with insulin resistance.4 In these individuals the severity of fat accumulation is proportional to the body mass index (BMI) and to the degree of visceral fat, and is not responsive to antiviral therapy.5 On the other hand, the steatosis observed in patients with chronic hepatitis C (CHC) infected by the genotype 3 virus (CHC-3) appears to have a different etiology. The degree of hepatic fat infiltration correlates with the level of HCV replication and protein expression, both in serum and in liver tissue.6 In addition, a sustained response to antiviral therapy is associated with reductions in hepatic triglyceride, while recurrence of infection is generally accompanied by the reaccumulation of steatosis.7 The low apolipoprotein B levels observed in CHC-3 suggest that the underlying mechanism of liver fat accumulation is similar to that observed in congenital hypo-betalipoproteinemia, which is characterized by reduced lipid export from the hepatocyte. Studies in vitro and in experimental animals suggest that “steatogenic” sequences specific to this viral genotype may exist.8

Steatosis per se is considered a major determinant of fibrosis both in NAFLD and in CHC. However, in published studies of CHC cohorts, the presence of different genotypes and host-related factors has often blurred the relative contributions of viral and metabolic factors to steatosis and to liver damage.9 Additionally, while insulin resistance, which is tightly associated with metabolic steatosis, is now emerging as a risk factor for fibrosis in both NAFLD and CHC, it is uncertain whether the fibrosis is a secondary effect of steatosis or a direct consequence of insulin resistance. The interaction between steatosis, insulin resistance, and disease severity is particularly complex in HCV infection because the virus itself might alter intrahepatic insulin signaling.10, 11

Thus, two fundamental questions remain unanswered; first, is hepatic steatosis per se a risk factor for progressive liver damage (namely, necroinflammation and fibrosis) independent of steatosis etiology (viral or metabolic); and second, is insulin resistance the principal determinant of liver fibrosis in steatosis-associated liver disorders?

To resolve this dilemma, we sought to dissect out the relative contributions of (1) virus-induced and metabolic steatosis, and (2) insulin resistance to liver fibrosis. To this end, we compared two large cohorts of patients with “pure” virus-induced steatosis (CHC-3) or “pure” metabolic steatosis (NAFLD), carefully matched for age, BMI and the degree of liver fat accumulation.


NAFLD, nonalcoholic fatty liver disease; HCV, hepatitis C virus; BMI, body mass index; CHC, chronic hepatitis C; ALT, alanine aminotransferase; WHR, waist-to-hip ratio; AST, aspartate aminotransferase; GGT, γ-glutamyltransferase; FBG, fasting blood glucose; HOMA-R, homeostasis model assessment; NASH, nonalcoholic steatohepatitis.

Patients and Methods


From May 1999 to October 2004, 172 consecutive patients with CHC-3 who underwent liver biopsy at Westmead Hospital (University of Sydney) were prospectively enrolled. Chronic hepatitis C was defined by detectable anti-HCV antibodies and serum HCV RNA. HCV genotyping was performed with a second generation reverse hybridization line probe assay (Inno-Lipa HCVII; Innogenetics, Zwijndrecht, Belgium). Patients with concurrent active hepatitis B virus (hepatitis B surface antigen and core antibody positive), autoimmune hepatitis, cholestatic (primary biliary cirrhosis, sclerosing cholangitis) or genetic (hemochromatosis, α1-antitripsin deficiency, Wilson disease) liver disease were excluded. Average current (last 6 months) and previous alcohol intake (g/day) was assessed by interviews extended to family members and general practitioners. The estimated duration of infection was defined as the time elapsed from the presumed date of infection to the date of liver biopsy.

From this initial cohort, 40 subjects were excluded because of fat infiltration <5% at liver biopsy. Pair-matching (by age, BMI, and the degree of steatosis at liver biopsy) was achieved in the remaining 132 individuals with subjects having NAFLD. Patients with NAFLD were recruited at the Gastroenterology Division of the University Hospital of Turin. The diagnosis of NAFLD was based on chronically elevated aminotransferase levels (alanine aminotransferase [ALT] >1.5 times the upper normal values for 6 months or more), negative hepatitis B (hepatitis B surface antigen and core antibody) and C (anti-hepatitis C virus) viral markers, absence of autoimmune hepatitis or celiac disease, and no evidence of genetic, drug-induced, or cholestatic liver disease. Patients with alcohol consumption ≥20 g/day were excluded.

Pair-matching was carried out by selecting from a list of over 300 NAFLD cases, observed in our Unit during the last 5 years, the first case of the same gender, having similar age (± 2 years), BMI (±1 kg/m2), and a similar extent of steatosis. Gender-matching was not possible in 6 cases.

All subjects considered for the study were Caucasian, except for 2 East-Asian subjects in the CHC-3 group and 4 Africans in the NAFLD group. Nineteen patients with CHC-3 had received previous antiviral therapy (interferon monotherapy or combination therapy with interferon and ribavirin) and were either non-responders or had relapsed after treatment. Liver biopsies were performed at least 6 months after the completion of the treatment. The study protocol was approved by the Human Ethics Committee of the Western Sydney Area Health Service and the University Hospital of Turin. Written, informed consent was obtained from all participating subjects.

Anthropometric and Laboratory Evaluations.

All subjects had a complete clinical, anthropometric and laboratory evaluation. BMI was calculated as weight (in kg) divided by height squared (m2). Subjects in the BMI range 25 to 30 kg/m2 and ≥30 kg/m2 were considered overweight and obese, respectively. Waist-to-hip ratio (WHR) was calculated as waist circumference at umbilicus/hip circumference at the maximal circumference over the buttocks. Laboratory investigations included fasting serum levels of albumin, bilirubin, ALT, aspartate aminotransferase (AST), γ-glutamyltransferase (GGT), total and HDL cholesterol, triglycerides, serum iron parameters (iron, transferrin and ferritin), glucose and insulin concentrations.

Plasma glucose levels were measured by the glucose oxidase method (Beckman Instruments, Fullerton CA; interassay CV <4%). Plasma insulin concentrations were assessed by a double-antibody radioimmunoassay (Diagnostic Products Corporation, Los Angeles, CA; interassay CV <13%). Insulin determination in the two centers was standardized. Fasting serum liver function tests and lipid levels were determined by routine laboratory techniques.

Subjects were classified according to their fasting blood glucose (FBG) as follows: normal fasting glucose (FBG <110 mg/dL), impaired fasting glucose (FBG, 110-125 mg/dL) and diabetes mellitus (FBG <125 mg/dL).12

Insulin resistance was calculated on the basis of the fasting glucose and insulin levels, according to the homeostasis model assessment (HOMA-R) method.13 HOMA-R values ≥2.7 were considered to indicate insulin resistance; this cutoff corresponds to the upper quartile of a control population, as previously published.14


In both the CHC-3 and NAFLD groups, steatosis was scored according to the criteria proposed by Brunt et al.15 as mild (<33% of hepatocytes affected), moderate (33%-66%) and severe (>66%). In CHC-3, the degree of necroinflammatory activity and fibrosis were scored according to Scheuer et al.16 Portal or periportal and lobular inflammatory activities were both graded from 0 to 4, fibrosis was staged from 0 to 4 (F0, absent; F1, enlarged fibrotic portal tract; F2 periportal or portal-portal septa but intact architecture; F3, architectural distortion but no obvious cirrhosis; F4, cirrhosis). In NAFLD, the Brunt classification was used to score necroinflammation from 0 to 3 and fibrosis from 0 to 4 (F0, absent; F1, perisinusoidal/ pericellular; F2 periportal; F3, bridging but no obvious cirrhosis; F4, cirrhosis). Nonalcoholic steatohepatitis (NASH) was defined on the basis of the presence of fibrosis (stage 1 or over) or necroinflammation (grade 2 or over).

Statistical Analysis.

Data analyzed using StatView 5.0 (SAS Institute, Inc., Cary, NC.). Unpaired t test (two tailed), χ2 contingency test, Fisher's exact test and linear regression analyses were used whenever appropriate. Non-parametric methods were used for non-normally distributed values (Kruskall-Wallis and Spearman rank correlation). Logistic regression analysis was used to identify factors associated with mild (grade 2 in both groups) and severe (grade 3 in both groups) steatosis, with mild (grade 1-2 in both groups) and severe (grade 3-4 lobular in CHC-3 and grade 3 in NAFLD) necroinflammation and with mild (stage 1-2 in both groups) and severe (stage 3-4 in both groups) fibrosis. All analyses have been adjusted for age, gender and BMI. Independent dichotomous variables were: presence of diabetes, hypertension, dyslipidemia (both hypercholesterolemia and hypertriglyceridemia) and viral load (in HCV cases). Independent parametric variables were biochemical indices of glucose and lipid metabolism (glucose, insulin and HOMA values, total or HDL cholesterol and triglycerides), indices of iron burden (ferritin and transferring saturation), tests of liver function and portal hypertension (platelets, aminotransferase and albumin concentrations). Transformation of HOMA-R values into ln (HOMA-R) did not qualitatively change the results. Hence, for clarity, they are reported without transformation. Results were expressed by odds ratios (95% confidence intervals). All data in the text and in the tables are given as means ±SD, when not otherwise indicated. Values of P less than .05 were considered statistically significant.


Patient Characteristics and Histological Findings.

CHC-3 and NAFLD subjects were carefully matched by age, gender, BMI, WHR, and the degree of hepatic steatosis (Table 1). Obesity was diagnosed in 13% of CHC-3 and 14% of the matched NAFLD subjects. In spite of this, CHC-3 patients had lower insulin, total cholesterol and triglycerides levels, and higher ALT activity, compared with NAFLD patients. Both albumin and bilirubin were slightly higher in NAFLD, while the platelet count was similar in the two groups. As expected, the NAFLD cases had a higher prevalence of metabolic abnormalities. Mean HOMA-R was higher and exceeded the cutoff of 2.7, indicative of insulin resistance, in 57.6% of cases, compared with 35.6% of the CHC-3 cohort (P < .001 Fisher's exact test). Finally, CHC-3 subjects had an higher alcohol intake, but only 7.5% of them exceeded the threshold of 40 g/day.

Table 1. Anthropometric, Clinical, and Laboratory Data in Genotype-3 Chronic Hepatitis C and Matched Patients With NAFLD
 Chronic Hepatitis C (n = 132)NAFLD (n = 132)P
  • NOTE. Data are expressed as mean ± SD, as prevalence (95% confidence interval), or as number of cases (n).

  • *

    Normal glucose tolerance/Impaired glucose tolerance/Diabetes mellitus.

  • Graded as 0–20, 20–40, >40 g/day.

Age (years)40 ± 742 ± 12.117
Male gender (%)7782.178
ALT (IU/L)137 ± 8188 ± 58<.001
Albumin (g/L)42.4 ± 3.343.7 ± 2.2<.001
Platelets (109/L)219 ± 65221 ± 51.799
Bilirubin (mg/dL)0.71 ± 0.330.88 ± 0.40<.001
Transferrin saturation (%)37 ± 1835 ± 12.229
Ferritin (mg/dL)252 ± 235246 ± 259.844
Waist-to-hip ratio0.92 ± 0.080.92 ± 0.07.797
BMI (kg/m2)26.4 ± 4.826.8 ± 4.2.442
Fasting Glucose (mg/dL)98 ± 3098 ± 25.987
Fasting insulin (μU/mL)12.0 ± 8.817.2 ± 13.2<.001
Triglycerides (mg/dL)89 ± 58135 ± 93<.001
Total cholesterol (mg/dL)148 ± 38206 ± 48<.001
HDL cholesterol (mg/dL)49 ± 1551 ± 13.228
HOMA-R3.13 ± 3.194.35 ± 4.17.008
Obesity (%)13 (8–19)14 (8-20).635
Glucose tolerance* (n)121/11/099/12/21<.001
Hypertriglyceridemia (%)10 (6–16)23 (17–31)<.001
Hypertension (%)6 (3–11)27 (20–35)<.001
Alcohol (g/day) (n)102/20/10132/0/0<.001

At liver biopsy, steatosis was grade 1 in 56%, 2 in 32.5% and 3 in 11.5% of both groups. Despite a similar degree of steatosis, fibrosis was significantly more severe in CHC-3 than in NAFLD. A similar trend was observed for necroinflammation (Fig. 1 and Fig. 2). Fibrosis was absent in 10 patients with CHC-3, compared with 50 NAFLD cases. Cirrhosis was present in 16 CHC-3, but only in 7 NAFLD patients. The different fibrosis severity between CHC-3 and NAFLD was maintained when the analysis was carried out after excluding the 30 CHC-3 patients with an average daily alcohol intake in the last 6 months higher than 20 grams and the corresponding 30 matched NAFLD subjects (data not shown).

Figure 1.

Histological data in Genotype-3 Chronic Hepatitis C and matched Non-Alcoholic Fatty Liver Disease patients. Steatosis (left) was scored as mild (5-33% of hepatocytes), moderate (34%-66%) or severe (Grade 3, >66%) according to Brunt classification.40 Necroinflammation (center) was scored as absent, mild (grade 1-2) and severe (grade 3, according to Scheuer16 in CHC-3 and grade 3-4 according to Brunt et al.40 in NAFLD). Similarly, fibrosis was scored as absent, mild (stage 1-2) or severe (stage 3-4), according to Scheuer et al.16 in CHC-3 and according to Brunt et al.40 in NAFLD.

Figure 2.

Distribution of fibrosis according to the severity of steatosis in Genotype-3 Chronic Hepatitis C and in Non-Alcoholic Fatty Liver Disease. Steatosis was scored as mild (5%-33% of hepatocytes), moderate (34%-66%) or severe (Grade 3, >66%) according to Brunt classification.15 Fibrosis was scored according to Scheuer in CHC-316 and according to Brunt et al.15 in NAFLD. No differences were observed in CHC-3, whereas the severity of steatosis was associated with fibrosis stage in NAFLD (P = .013).

Determinants of Steatosis in the CHC-3 and NAFLD Cohorts.

After correction for age, gender and BMI, the univariate analysis confirmed that different risk factors were present for steatosis in the CHC-3 and NAFLD cohorts (Table 2). Thus, in CHC-3, only viral load >850×103 IU/mL and low cholesterol values were associated with mild and severe fat infiltration, in keeping with a virus-induced mechanism, while none of the metabolic parameters had a significant effect (Table 2). Conversely, in NAFLD mild hepatic fat infiltration was associated with increased ALT activity, raised fasting blood glucose, insulin, HOMA-R and ferritin levels. With the notable exception of insulin and insulin resistance, the same biochemical parameters also identified high-grade steatosis, which was also predicted by hypertriglyceridemia.

Table 2. Factors Associated With Steatosis in Chronic Hepatitis C and Matched Patients With NAFLD at Univariate Analysis
 Grade 1-2Grade 3
  1. NOTE. Data (odds ratio and 95% confidence intervals) are adjusted for age, gender, and BMI.

Genotype-3 Chronic Hepatitis C      
Viral Load (>850 × 103 copies/mL)8.462.92–24.50<.0014.301.06–17.40.041
Cholesterol (mg/dL/25)0.740.55–0.99.0410.340.19–0.60.0002
Nonalcoholic Fatty Liver Disease      
Insulin (μU/mL/5)1.181.00–1.40.046   
ALT (mU/mL/10)1.181.05–1.33.0051.381.19–1.61<.001
Ferritin (μg/dL/50)1.181.05–1.31.0041.231.08–1.41.002
Blood glucose (mg/dL/10)1.311.02–1.68.0361.481.09–2.02.012
Hypertriglyceridemia (>150 mg/dL)   5.531.55–19.72.008

Variables Associated With the Severity of Necroinflammation.

In NAFLD, severe steatosis was identified by univariate analysis as the only significant parameter associated with severe necroinflammation (OR = 2.54; CI 1.09-3.84, P = .024), while none of the tested variables predicted the necroinflammatory grade of CHC-3.

Variables Associated With the Severity of Fibrosis.

At univariate analysis, in CHC-3, mild fibrosis was significantly associated with low total cholesterol levels, while severe fibrosis was correlated with insulin resistance (high insulin levels and HOMA-R) and a low platelet count (Table 3). In NAFLD, mild fibrosis was not associated with any of the tested parameters, whereas severe fibrosis was related to increased ALT and AST activity, insulin resistance (high insulin levels and HOMA-R), raised ferritin levels, and decreased HDL-cholesterol. In NAFLD, but not in CHC-3, fat grading was significantly associated with severe fibrosis (Table 3).

Table 3. Univariate Analysis of Factors Associated With Fibrosis in Chronic Hepatitis C and Matched Patients With NAFLD
  1. NOTE. Data (odds ratio and 95% confidence intervals) are adjusted for age, gender, and BMI.

Genotype-3 Chronic Hepatitis C      
Cholesterol (mg/dL/25)0.630.41–0.98.040   
Insulin (μU/mL/5)   5.011.37–18.34.015
HOMA-R   3.191.19–8.53.021
Platelets (109/L/10)   0.780.67–0.90.010
Nonalcoholic Fatty Liver Disease      
ALT (mU/mL/10)   1.121.02–1.23.020
AST (mU/mL/10)   1.261.03–1.55.028
Ferritin (μg/dL/50)   1.211.06–1.38.004
HDL cholesterol (mg/dL/5)   0.740.57–0.94.016
Insulin (μU/mL/5)   1.221.00–1.48.047
HOMA-R   1.181.01–1.36.031
Fat grading   3.371.57–7.21.002

Independent Predictors of Severe Liver Fibrosis.

Next, the variables significantly associated with fibrosis at univariate analysis were separately tested by multivariate analysis for their independent association with severe fibrosis in both patient cohorts (Table 4). In CHC-3, severe fibrosis was independently related to a low platelet count (OR = 0.78; 95% CI, 0.67-0.92; P = .003) and to increased insulin resistance, measured by the HOMA-R (OR = 2.98; 95% CI, 1.13-7.89; P = .028). 43% of CHC-3 patients with F3 (architectural distortion but no obvious cirrhosis) and 75% with F4 (cirrhosis) were in the upper quartile of HOMA-R values. The exclusion of cases with alcohol intake ≥20 g/day did not change significantly the predictive role of low platelets (OR = 0.84; 0.74-0.94) and HOMA-R (OR =3.67; 1.20-11.21). Finally, when the 40 patients without steatosis at liver biopsy were added into the analysis, platelet count (OR = 0.83; 0.77-0.91) and HOMA-R (OR =2.83; 1.14-7.03) were confirmed as the sole predictors of severe fibrosis.

Table 4. Multivariate Analysis for Factors Independently Associated With Severe Fibrosis in Chronic Hepatitis C and Matched Cases of Patients With NAFLD
 OR95% CIP
  1. NOTE. Data (odds ratio and 95% confidence intervals) are adjusted for age, gender, and BMI.

Genotype-3 Chronic Hepatitis C   
Platelets (109/L/10)0.780.67–0.92.003
Nonalcoholic Fatty Liver Disease   
Fat grading3.031.41–6.53.004
Ferritin (μg/dL/50)1.131.03–1.25.013

In NAFLD, severe fibrosis was independently predicted by histological fat grading (OR = 3.03; 95% CI, 1.41-6.53; P = .004), raised ferritin (OR = 1.13; 95% CI, 1.03-1.25; P = .013) and by insulin resistance (HOMA-R: OR = 1.16; 95% CI, 1.02-1.31; P = .021). Eighty-seven percent of NAFLD patients with bridging fibrosis and all NAFLD cases with cirrhosis were in the upper quartile of HOMA-R values. The exclusion of cirrhotic patients (stage 4 in both groups) did not change the association between HOMA-R and fibrosis stage in CHC-3 and in NAFLD.

The severity of fibrosis increased progressively according to the degree of steatosis in NAFLD, but remained almost unchanged in CHC-3 (Fig. 2). By contrast, when the severity of fibrosis was plotted as a function of insulin resistance (quartiles of HOMA), in both groups the relation was statistically significant (Fig. 3).

Figure 3.

Distribution of fibrosis according to the severity of insulin resistance (quartiles of HOMA) in Genotype-3 Chronic Hepatitis C and in Nonalcoholic Fatty Liver Disease. Fibrosis, scored according to Scheuer in CHC-316 and according to Brunt et al.15 in NAFLD, was redefined as absent (stage 0), mild (stage 1), moderate (stage 2) or severe (stage 3) (see text). The severity of insulin resistance was associated with fibrosis stage in both CHC-3 (r = 0.377; P < .001) and NAFLD (r = 0.290; P < .001), but the severity of fibrosis was remarkably higher in CHC-3.


The present study examined the role of steatosis resulting from differing etiologies and of insulin resistance in inducing liver fibrosis. To this end, we examined the relationship between liver fat, insulin resistance and fibrosis in two large cohorts of subjects with a similar degree of steatosis, in whom steatosis was principally of viral (genotype 3 CHC-3) or of metabolic (NAFLD) origin. Our results demonstrate that the association of steatosis with liver fibrosis is dependent on its cause, and highlights the importance of insulin resistance per se as a major determinant of liver fibrosis in liver disorders, irrespective of etiology.

There is evidence that the association between CHC-3 and steatosis is mediated by the “steatogenic” properties of the hepatitis C virus.17, 18 However, in individual patients, host factors, particularly abdominal obesity and type 2 diabetes may contribute to the development of steatosis and liver fibrosis through a mechanism mediated by insulin resistance.19 The coexistence of host and viral factors contributing to steatosis in the same individual has long blurred the analysis of their independent contributions to liver fibrosis in persons infected with HCV. To minimize this possible source of bias, the present study examined a cohort of subjects having a low prevalence of clinical and biochemical features of insulin resistance, and in whom steatosis largely reflects a direct toxic effect of the HCV virus, namely non-obese, non-diabetic subjects infected with genotype 3. In this subset of HCV patients, the viral core protein is ≈3-fold more efficient than the corresponding protein from genotype 1 in inducing lipid accumulation in the liver,18, 20 with intrahepatic levels of core protein and serum HCV-RNA levels correlating with steatosis severity.8, 21 Further, we have previously shown that a sustained viral response to interferon treatment is associated with the disappearance of steatosis.6

In our CHC-3 cohort, low total plasma cholesterol and high viral load were the sole determinants of steatosis, while metabolic, host-related factors were ruled out. This is in keeping with the hypothesis that genotype 3 infection may promote a condition phenotypically similar to genetically-determined hypobetalipoproteinemia, where extensive steatosis results from mutations in the gene encoding MTP.22 In CHC-3 patients, the serum levels of ApoB are decreased and inversely correlate with both the score of steatosis and HCV viral load.23 ApoB levels revert to normal following successful antiviral treatment.24 In liver biopsies of CHC-3 infected patients, liver MTP specific activity is significantly reduced and correlates with reduced serum cholesterol, ApoB, and VLDL.25

We next compared the CHC-3 cohort with a group of NAFLD patients matched by age, gender, BMI and degree of steatosis. Despite the efficacy of matching, the prevalence of clinical features and biochemical indices of insulin resistance were different in the NAFLD cohort; the association of steatosis with metabolic abnormalities as previously reported in these patients was confirmed.26, 27

The different etiological factors for steatosis in the two patient cohorts enabled us to verify the relative impact of liver fat of different etiology on hepatic inflammation and fibrosis. Unexpectedly, steatosis had no association with fibrosis in CHC-3, while it was a major determinant of liver damage in NAFLD, where fibrosis paralleled the severity of fat infiltration. More importantly, in both groups insulin resistance was an independent predictor of severe fibrosis, despite the fact that steatosis correlated with fibrosis severity only in persons with NAFLD. These results were also confirmed after exclusion of patients with cirrhosis (stage 4), thus ruling out the possible effect of a reduced extraction of insulin by the cirrhotic liver. The clear correlation between steatosis and fibrosis observed in our population of NAFLD patients, which has been controversial in previous studies, is probably due to the selection criteria adopted in order to match with the CHC-3 patients. The exclusion of persons with obesity and type 2 diabetes probably unveils the correlation between steatosis and fibrosis, which is often overshadowed by the effect of these two major confounding factors.

The association between steatosis and fibrosis in CHC remains controversial. In some studies, steatosis was associated with more severe fibrosis and with progression of fibrosis, regardless of genotype.28–30 In others, this relationship has been recognized only in some subgroups of patients, i.e., those infected with genotype 14, 31 or 3.32–34 The contribution of steatosis to the onset and progression of fibrosis has been recently challenged in a large cohort of CHC patients.35 Only a few studies, however, have evaluated the impact of insulin resistance on liver damage (i.e., steatosis and fibrosis) and in most reports CHC-3 patients represented only a small proportion of the study population. In a recent study conducted in over 3000 patients with HCV-related chronic hepatitis, the association between steatosis and fibrosis was confirmed in genotype 1, but not in genotype 3–infected subjects.36 However, insulin resistance was not considered in this report. In keeping with the above results, our data from a matched comparison with NAFLD subjects rule out a direct, independent effect of steatosis on fibrosis in CHC-3 and indicate instead that also in this subset of patients fibrosis primarily correlates with insulin resistance.

In CHC-3, insulin resistance and virus-induced liver inflammation appear to be the likely determinants of fibrosis progression. By contrast, in NAFLD, our data suggest that hepatic fibrosis might stem from a complex “triad” of independent risk factors, i.e., steatosis, insulin resistance and systemic inflammation, the latter indicated by the elevated ferritin levels. The mechanisms by which hyperinsulinemia and insulin resistance mediate hepatic fibrogenesis are speculative. Hyperinsulinemia per se induces hepatic stellate cell (HSC) proliferation,37 while high levels of glucose upregulate the expression of pro-fibrogenic cytokines.38 Several lines of evidence further suggest that insulin resistance contributes to a chronic, low-grade systemic inflammatory response. Evidence for this includes the increased serum ferritin levels common in NAFLD and in other metabolic diseases, which has been associated with the extent of liver fibrosis.39

In conclusion, insulin resistance is associated with advanced hepatic fibrosis in metabolic as well as in viral liver disease. In contrast, only “metabolic” steatosis contributes to liver fibrosis. Virus-induced steatosis as seen in CHC-3 does not appear to directly promote necroinflammation and hepatic fibrogenesis. In clinical care, future therapeutic approaches should take into consideration the possible mechanisms for steatosis and properly consider insulin resistance as a specific therapeutic target to prevent progression of viral and non-viral liver disease.


The authors wish to thank the members of the University of Sydney Hepatitis C Pathogenesis Study Group, James Kench (Department of Anatomical Pathology, Westmead Hospital, NSW, Australia), Chris Liddle, Rita Lin and Dev Samarasinghe (Storr Liver Unit and Department of Gastroenterology and Hepatology, Westmead Hospital, NSW, Australia).