Potential conflict of interest: Nothing to report.
Although steatosis is a common histological feature in chronic hepatitis C (CHC), nonalcoholic steatohepatitis (NASH) has not yet been clearly characterized in this context. The aim of this prospective study was to investigate the characteristics of patients with NASH and CHC. Biopsies were categorized as CHC alone (178 patients [57%]), CHC+steatosis (94 patients [34%]), or CHC+NASH (24 patients [9%]). Patients with CHC+NASH had significantly higher AST and triglyceride levels and lower high-density lipoprotein (HDL) cholesterol or total cholesterol than patients with CHC+steatosis. They also showed more steatosis and higher METAVIR fibrosis stage than patients with CHC+steatosis. Genotype 3 was more frequent in patients with CHC+NASH than in patients with CHC+steatosis or CHC alone. Patients with genotype 3 and CHC+NASH were similar to those with CHC+steatosis or with CHC alone according to triglyceride or the homeostasis model for assessment of insulin resistance (HOMA-IR), whereas in patients with genotype 1, HOMA-IR and triglyceride increased progressively from CHC alone to CHC+steatosis to CHC+NASH. In multivariate analysis, triglyceride and HDL cholesterol were predictors of NASH in patients with genotype 1, whereas in patients with genotype 3, AST was the only predictor. Conclusion: Patients with CHC+NASH differ significantly from those with CHC+steatosis and CHC alone in terms of biological and metabolic parameters and more advanced histopathological lesions. NASH is more common in genotype 3 and is not associated with metabolic dysfunctions in this subgroup, suggesting that NASH may complicate steatosis in CHC irrespective of etiology of steatosis. (HEPATOLOGY 2007.)
Nonalcoholic fatty liver disease (NAFLD) encompasses a spectrum of lesions including simple steatosis, nonalcoholic steatohepatitis (NASH), fibrosis, and cirrhosis in patients who deny alcohol intake.1 Only recently has this condition been recognized as a chronic liver disease affecting a substantial proportion of the population.2 Main risk factors associated with NAFLD include obesity, dyslipidemia, type 2 diabetes mellitus, and other conditions characterized by insulin resistance.2, 3 Recent evidence suggests that similar lesions might also be observed in association with other chronic liver diseases, including chronic hepatitis C (CHC), another common and clinically significant condition.4–6 There are now strong arguments suggesting that the association of these 2 common liver diseases is much more frequent than predicted by chance alone and that the coexistence of the 2 may be significant in disease progression.
Steatosis is a very common lesion in CHC, with a prevalence of 40% to 86%.7 It is now acknowledged that the pathogenesis of steatosis in CHC might differ according to the virus genotype. Strong clinical and experimental evidence suggests that steatosis in patients infected with genotype 3 is partly related to a direct cytopathic effect, whereas in genotype 1, steatosis is mainly related to an associated metabolic syndrome and the ensuing insulin resistance.8–10 Because NASH in patients without CHC has been described in the context of the metabolic syndrome and is strongly associated with steatosis, it is not surprising to observe NASH in patients with CHC. Indeed, NASH has also been reported in this context, though the prevalence of the lesion varies.4–6
In the absence of reliable clinical and biochemical surrogate markers, diagnosis of NASH relies purely on histopathological bases.1, 10, 11 Within the context of NAFLD, only patients with biopsy-proven NASH have convincingly been shown to progress to cirrhosis; therefore, it is of major importance to clearly delineate the histopathological criteria of NASH in CHC. However, the histopathological criteria required for diagnosis of NASH are ambiguous, because it is a progressive disease associating several different features and different definitions and grading have been proposed.11, 12 Most of these scoring systems include various associations of steatosis, lobular inflammation, clarification/ballooning of liver cells, and sinusoidal fibrosis. Within the context of CHC, lesions related to the background chronic inflammatory disease increase the complexity of NASH identification, because some of these elementary features might overlap (steatosis, lobular inflammation). Therefore, diagnosis of NASH in patients with CHC requires careful histopathological evaluation of selected criteria, the relevance of which should be validated by clinico-biological correlates. More specifically, it remains to be determined whether histopathological criteria for the definition of NASH in patients with CHC are relevant to characterize patients that differ from those having CHC and steatosis alone. Finally, and based on the difference in pathogenesis of steatosis according to HCV genotype, it is also relevant to investigate whether viral genotype may also influence NASH in CHC.
In this prospective study of a large group of patients with CHC, we identify those patients who also have NASH and investigate their clinico-biological, metabolic, histological, and virological characteristics compared with those with CHC plus steatosis and CHC alone. Finally, we identified predictive factors of NASH in patients with CHC.
BMI, body mass index; CHC, chronic hepatitis C; HDL, high-density lipoprotein; HOMA-IR, homeostasis model for assessment of insulin resistance; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis.
Patients and Methods
A total of 278 consecutive naïve patients with CHC who were admitted to our hospital between September 2005 and October 2006 for liver biopsy were included in the study. Liver biopsy was performed for staging and grading of chronic hepatitis prior to any antiviral treatment. All patients gave their informed consent prior to liver biopsy.
Inclusion criteria were positive antibodies against HCV (AxSYM Anti HCV, Abbott), detectable HCV RNA (TMA, Bayer's Versant HCV RNA Qualitative Assay), and an adequate liver biopsy. Exclusion criteria were positive hepatitis B surface antigen or autoimmune hepatitis, HIV coinfection, insulin requiring diabetes mellitus, decompensated liver cirrhosis (Child-Pugh grade B or C), hemochromatosis, excessive alcohol intake, α1-antitrypsin deficiency, and Wilson's disease.
A questionnaire was completed prospectively for all patients. Risk factors for fibrosis and steatosis were assessed. In addition, all parameters of the metabolic syndrome as defined by the National Cholesterol Education Program classification, were collected. The questionnaire included details on: sex, age, age at infection, source of infection, alcohol consumption (grams per day), weight (kilograms), height (meters), waist circumference (centimeters), and arterial blood pressure. Body mass index (BMI) was calculated as weight divided by height squared (kg/m2). Overweight was defined as a BMI >25 kg/m2; obesity was defined as a BMI >30 kg/m2. Central obesity was defined as waist circumference >102 cm and >88 cm in males and females, respectively. The metabolic syndrome was diagnosed according to the National Cholesterol Education Program definition.13 Alcohol intake was assessed via patient interview. Patients were considered excessive alcohol consumers if their daily intake was ≥30 g (males) or ≥20 g (females); such patients were excluded from the study. The estimated duration of infection was defined as the time that elapsed from the presumed date of infection (date of transfusion, first year of intravenous drug use) to the date of liver biopsy. After an overnight fast of 12 hours, venous blood flow was drawn to determine serum levels of ALT, AST, γ-glutamyltransferase, ferritin, glucose, triglyceride, high-density lipoprotein (HDL) cholesterol, insulin, and C-peptide. Serum insulin and C-peptide were determined via electrochemiluminescence immunoassay (Elecsys 2010, Roche Diagnostics).
Virological tests included HCV genotype (sequencing) and quantitative serum HCV RNA level (Bayer's VERSANT HCV RNA 3.0 Assay) for all patients.
Each biopsy was individually coded and processed for histopathological interpretation by a single experienced liver pathologist (P. B.) blinded to patient identity and clinical and biological data. A minimum of 6 portal tracts per biopsy specimen for noncirrhotic patients was deemed adequate for study analysis. All biopsies were processed routinely, formalin-fixed, serially sectioned, and stained with hematoxylin-eosin, Masson's trichrome, and Sirius red for fibrosis evaluation and Perls staining for iron. A detailed descriptive scoring sheet was employed, including the grading and staging of chronic hepatitis according to METAVIR,15 semiquantitative evaluation of steatosis (0, <5%; grade 1, 5%-30%; grade 2, 30%-60%; grade 3, >60%), predominant zonal distribution of fat (pericentral, periportal, azonal, diffuse), clarification/ballooning of liver cells (0, absent; 1, mild; 2, marked), perisinusoidal fibrosis (0, absent; 1, mild; 2, marked), Mallory hyaline (0, absent; 1, present) and presence of glycogenated nuclei (0, absent; 1, present) as well as other more less common histological features as described by Kleiner et al.11 Finally, patients were classified as CHC alone, CHC with steatosis but without NASH (CHC+steatosis), or CHC with NASH (CHC+NASH). The criterion for steatosis was grade 1 or higher. NASH was defined according to a score made by addition of clarification/ballooning and perisinusoidal fibrosis semiquantitation. A score of >3 was considered diagnostic for NASH. Neither steatosis nor lobular inflammation was required for the diagnosis of NASH, because both could be related to CHC alone.
Quantitative variables were expressed as the mean ± SE. Univariate analysis was performed using the Fisher exact test and χ2 test for frequencies and the Student t test and Mann-Whitney rank-sum test for means for quantitative variables. Multiple logistic regression was employed to model the presence of NASH in CHC as a function of laboratory results and demographic factors. Multivariate analysis was performed separately in all patients and in the group of patients with genotype 1 and those with genotype 3. A P value of 0.05 or less was considered significant (NSCC, Kaysville, UT).
Of the 278 consecutive patients, 158 were male and 120 were female. The mean age was 47.3 years (range, 19–75). One hundred forty-four patients (52%) were infected with HCV genotype 1, 24 (9%) were infected with genotype 2, 50 (18%) were infected with genotype 3, 57 (20%) were infected with genotype 4, and 3 (1%) were infected with genotype 5.
Patients were divided into 3 groups based on histopathological criteria: 160 cases (57%) had CHC alone, 94 (34%) had CHC with steatosis but without NASH (CHC+steatosis), and 24 (9%) had CHC with NASH (CHC+NASH).
The 24 patients with CHC+NASH had a mean age of 48 years (range, 31–73); there were 16 males and 8 females. The mean BMI was 25.7 + 0.7 kg/m2 (range, 20–32); 8 of the 24 patients (33%) were overweight, including 4 (17%) who were considered obese. The metabolic syndrome was present in 3 patients (12%); diabetes was present in none. Typical histopathological features of NASH in CHC are shown in Fig. 1.
The main clinical data according to histopathological groups are shown in Table 1.
Table 1. Baseline Clinical and Demographic Features
Patients with CHC+NASH had significantly higher serum AST levels than those with CHC+steatosis (P < 0.05). The same trend was observed for ALT and γ-glutamyltransferase, but the values did not reach significance. Total cholesterol, triglycerides, and HDL cholesterol were also significantly different when patients with CHC+NASH were compared with those with CHC+steatosis. The data are shown in detail in Table 2.
Table 2. Biological Parameters in Patients with CHC
CHC, All Patients (n = 278)
CHC, No Steatosis, No NASH (n = 160)
CHC+Steatosis (n = 94)
CHC+NASH (n = 24)
P Value (CHC+Steatosis vs. CHC+NASH)
Abbreviations: GGT, γ-glutamyltransferase; NS, not significant.
When considering patients according to virus genotype, a difference was observed for insulin resistance and lipid metabolism. Patients with genotype 1 showed significant difference in the homeostasis model for assessment of insulin resistance (HOMA-IR) and triglycerides according to histopathological groups, whereas those with genotype 3 did not display any difference (Fig. 2). In patients with genotype 1, insulin resistance, as assessed via the HOMA-IR index, increased progressively from CHC alone to CHC with steatosis (P < 0.01), reaching the highest value in patients with CHC+NASH (P < 0.01 for CHC+NASH versus CHC+steatosis). No significant change for HOMA-IR was observed in the subgroup of patients infected with genotype 3 according to histopathological parameters. The same difference was also observed for triglycerides according to genotype and histopathological groups. By contrast, total cholesterol was constantly lower in genotype 3 than genotype 1 when similar histopathological groups were compared. The difference was significant only in the CHC+NASH group (CHC+NASH: 135 + 18 in patients with genotype 3 versus 183 + 11 in CHC+NASH with genotype 1; P < .05).
Duration of infection was known for 133 patients. The median duration was 23 years and was not significantly different when the 3 groups were compared.
Serum HCV RNA level and genotype according to histological groups are shown in Table 3. There was no difference in viral load in the 3 histopathological groups. Among the 24 patients with CHC+NASH, 8 (33%) were infected with genotype 3, 10 (42%) were infected with genotype 1, and 6 (25%) were infected with genotype 4. Genotype 3 was more common in patients with CHC+NASH (33%; P < 0.01) or CHC+steatosis (23%; P < 0.05) than in patients with CHC alone (13%).
Table 3. Virological Characteristics of Patients with CHC According to Histopathological Group
CHC, All Patients (n = 278)
CHC Alone (n = 160)
CHC+Steatosis (n = 94)
CHC+NASH (n = 24)
P Value (CHC+Steatosis vs. CHC+NASH)
Abbreviation: NS, not significant.
Years of infection, median (range)
Log HCV RNA
6.3 ± 0.04
6.3 ± 0.06
6.3 ± 0.08
6.2 ± 0.2
No. with genotype 1
No. with genotype 3
No. with genotype 4
No. with others
The mean length of liver biopsy was 15.1 mm and was not significantly different between the 3 histopathological groups. Twenty-four patients had CHC+NASH, 21 of whom had a score (clarification + perisinusoidal fibrosis) of 3 and 3 of whom had a score of 4. Although we did not include steatosis as a criterion for NASH, all patients with NASH also had steatosis.
The major histopathological patterns according to the 3 main histological subgroups are shown in Table 4.
Table 4. Pathological Characteristics in Patients with CHC and According to Histopathological Groups
Regarding fibrosis, 108 patients (39%) were stage F0 or F1, 101 (36%) were stage F2, 40 (15%) were stage F3, and 29 (10%) were stage F4. Sinusoidal fibrosis was graded as absent, mild, or significant in 195 (70%), 65 (23%), and 18 (6%) patients, respectively. Steatosis was absent or minimal (<5%) in 161 patients (58%), mild (5%–30%) in 62 patients (22%), moderate (30%–60%) in 33 patients (8%), and severe (>60%) in 22 patients (8%). Grading of activity was A0, A1, A2, and A3 in 67 (24%), 190 (68%), 19 (7%), and 2 cases (1%), respectively. Clarification and ballooning were absent in 211 patients (76%), mild in 51 patients (18%), and significant in 16 patients (6%).
The amount of steatosis was higher in patients with CHC+NASH than in those with CHC+steatosis without NASH (58% vs. 29%, P < 0.05). Sixty-nine patients (25%) displayed an advanced stage of fibrosis (F3–F4), a prevalence that was not significantly different in the group of patients with known duration of disease (30 of 133 [23%]) and those who disease duration was unknown (39 of 145 [27%]; P value not significant). Patients with CHC+NASH had a more advanced METAVIR stage of fibrosis (54% with stage F3 or F4) than patients with CHC+steatosis (32%; P < 0.05) or patients with neither steatosis nor NASH (16%; P < 0.01). Grade of activity was not significantly different between the 3 groups.
In the group of CHC+NASH, neither significant steatosis or advanced stage of fibrosis were influenced by virus genotype. Seven of 10 patients (70%) with genotype 1 and 6 of 8 (75%) with genotype 3 had significant steatosis (ns) and 5 of 10 patients (50%) with genotype 1 and 4 of 8 (50%) patients with genotype 3 had advanced stage of fibrosis.
To identify factors that might be associated with the presence of NASH in patients with CHC, we performed a multiple logistic regression analysis in which NASH was the dependent variable. Parameters that were significant in univariate analysis were introduced in the model. Results are shown for all patients but also according to virus genotype in Table 5. Triglyceride, cholesterol, and AST were predictors for NASH in patients with CHC when all genotypes were studied together. Additionally, in patients with genotype 1, HDL cholesterol and triglyceride significantly predicted NASH, whereas in genotype 3, only AST predicted NASH.
Table 5. Multivariate Analysis for Predicting NASH in Patients with CHC
OR (95% CI)
NAFLD covers a spectrum of histopathological lesions that includes steatosis and steatohepatitis in patients who deny alcohol consumption. Although NASH is often associated with the metabolic syndrome, some elements suggest that similar lesions can be observed within the context of different chronic liver diseases, including CHC.4–7, 15, 16 Steatosis is a frequent histopathological finding in CHC and may be present in as many as 51% of patients with CHC in a meta-analysis of individual data.18 The prevalence and characteristics of patients with NASH in CHC are less clear, because most studies are retrospective using different criteria for the definition of NASH, with some patterns overlapping those observed in CHC.4, 5, 15 Recent studies suggest that patients with CHC may display typical NASH lesions in the absence of alcohol intake in as many as 18% of patients with CHC.16 Another study suggested that the perisinusoidal pattern of fibrosis similar to that of NAFLD might be observed in some patients with CHC.18
To assess the prevalence of NASH unambiguously, we carefully selected morphologic criteria for NASH identification in a CHC background. Several histopathological definitions of NASH have been proposed, most of which are not applicable within the context of CHC, because some histopathological features (eg, lobular inflammation or steatosis) may be related to NASH or CHC only.19, 20 For the purpose of the present study, we selected 2 criteria that seem to be characteristic of NASH but are not present in CHC: perisinusoidal fibrosis and clarification/ballooning of liver cells. Each of these lesions was semiquantified and, based on the threshold that we have defined, only those cases displaying both lesions with at least 1 present at a significant level were included in this group. This strict definition and this threshold excluded inclusion of intermediate cases, which are the main causes of interobserver variation and sampling errors.21 Clarification/ballooning of liver cells is a well-accepted criterion for NASH. Matteoni et al.22 clearly showed that this lesion is associated with a poor prognosis in patients with NASH. We also included perisinusoidal fibrosis, a characteristic pattern of NASH. Therefore, and with our definition, the group of NASH in chronic CHC included 24 (9%) clear-cut cases in which NASH was superimposed on CHC in a large cohort of consecutive patients with CHC who were admitted for liver biopsy, mainly prior to antiviral treatment.
Thus, we sought to investigate whether identification of CHC with NASH, as described with these selected criteria, is clinically relevant and whether this definition enables characterization of a distinctive clinico-biological profile different from that with CHC alone or CHC with steatosis. If such a distinction exists, it would lend support to NASH characterization in patients with CHC. For this purpose, we correlated clinico-biological data and metabolic and virological status according to the histopathological grouping. First, we confirmed that the patients with CHC+steatosis differ strongly from those with CHC alone regarding metabolic disturbances, biochemical changes, and higher stage of fibrosis. Furthermore, it was clearly evident that patients with CHC and NASH share characteristics with patients with CHC and steatosis, but there are some differences. Patients with NASH have significantly higher levels of AST but not ALT. Interestingly, this group with NASH did not display higher grade of necroinflammation than the other 2 groups, suggesting that the increase in aminotransferases in patients with NASH is not related to CHC but rather superimposed NASH. Interestingly, some of these laboratory parameters distinguishing NASH from simple steatosis in patients with CHC were also reported in patients with NAFLD in the absence of CHC.23 Patients with CHC and NASH also differed from those with CHC and steatosis in terms of lipid metabolism, because they displayed significantly higher triglyceride and lower HDL cholesterol and total cholesterol. HOMA-IR also increased progressively from CHC alone to CHC+steatosis, reaching the highest levels in CHC with NASH. These results strengthen the histopathological characterization of a group of patients with NASH as defined in this study, because they displayed a more severe metabolic syndrome than those with CHC+steatosis or CHC alone.
At the histopathological level, we showed that patients with CHC and NASH displayed a higher grade of steatosis. Because NASH was not selected or graded according to steatosis in our study, this supports the hypothesis that development of NASH may be promoted by major steatosis. This result is concordant with a recent study suggesting that perisinusoidal pattern of fibrosis in some CHC patients is associated with higher BMI or grade of steatosis.18
We also showed that patients with CHC and NASH have higher METAVIR stage of fibrosis than those with CHC+steatosis or CHC alone. This relationship warrants further discussion. Ong et al.15 reported that lesions of NASH in CHC correlated with severity of fibrosis. Although the duration of disease was known in a limited number of patients only, and because the median duration of disease did not differ according to whether NASH or steatosis was present in association with CHC, our results suggest that NASH should promote a more rapid fibrosis progression in these patients, as already shown within the context of metabolic syndrome.24 However, our study is cross-sectional, and this result needs to be confirmed with longitudinal follow-up of these patients.
It is interesting to note that, as in the group of CHC+steatosis patients, patients with CHC+NASH are more often infected with HCV genotype 3 than patients with CHC alone. Patients with genotype 3 more often have steatosis—and frequently at an higher grade—supporting the notion that NASH in CHC is a complication of severe steatosis independent of virus genotype.
Furthermore, and when breaking down patient data according to genotype, different trends for triglyceride and HOMA-IR were observed according to histopathological groups. Although these parameters did not change in patients with genotype 3 according to the different histological groups, a progressive increase was observed in patients with genotype 1, with CHC+NASH having more metabolic disturbances than CHC+steatosis. Interestingly, cholesterol level was always lower in patients with genotype 3 than in those with genotype 1 when similar histological groups were compared. This result is in concordance with previous reports suggesting that HCV genotype 3 lowers serum cholesterol, a property not shared by other genotypes.25, 26 It is also interesting to observe that predictors for NASH in multivariate analysis strikingly differed according to genotype. In patients with genotype 1, metabolic factors were the main predictors of NASH, whereas in patients with genotype 3, AST appeared to be the only blood predictive marker of NASH. Taken together, these results demonstrate that in patients with genotype 3, NASH might occur in the absence of any metabolic dysfunction.
The higher grade of steatosis in patients with NASH suggests that the 2-hit theory holds true in CHC.27 In this concept, steatosis—the first hit—is a prerequisite for the development of fibrosis that occurs after a second insult, where production of profibrotic compounds through oxidative stress enables and fuels the fibrogenesis mechanism. Interestingly, we and others have already shown that lipid peroxidation by-product can be demonstrated in certain biopsies of patients with CHC.28, 29 Furthermore, the higher prevalence of genotype 3 in patients with CHC+NASH—which mirrors the increase in patients with CHC+steatosis alone in the same subgroup—also supports the notion that NASH might complicate severe steatosis independently of the mechanism leading to steatosis.
In conclusion, this study supports the notion that a histopathological definition of NASH that includes perisinusoidal fibrosis and liver cell ballooning in CHC is relevant for characterizing a group of patients with more severe clinicopathological lesions than those with chronic CHC and steatosis alone. NASH, like steatosis, is more common in patients with genotype 3. The absence of an association with metabolic disturbances in these patients suggests that the physiopathology of NASH might occur as a complication of steatosis, whatever its etiolopathology, and is a determinant for faster progression of fibrosis.