Hyperhomocysteinemia and the MTHFR C677T polymorphism promote steatosis and fibrosis in chronic hepatitis C patients


  • See Editorial on Page 976

  • Potential conflict of interest: Nothing to report.


The factors and mechanisms implicated in the development of hepatitis C virus (HCV)-related steatosis are unknown. Hyperhomocysteinemia causes steatosis, and the methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism induces hyperhomocysteinemia. We investigated the role of these factors in the development of HCV-related steatosis and in the progression of chronic hepatitis C (CHC). One hundred sixteen CHC patients were evaluated for HAI, fibrosis and steatosis grades, body mass index, HCV genotypes, HCV RNA levels, homocysteinemia, and the MTHFR C677T polymorphism. Hyperhomocysteinemia was associated with the TT genotype of MTHFR (r = 0.367; P = .001). Median values of homocysteine in the CC, CT, and TT genotypes of the MTHFR gene were 9.3, 12.2, and 18.6 μmol/L, respectively (P = .006). Steatosis correlated with the MTHFR polymorphism, homocysteinemia, HAI and fibrosis. Steatosis above 20% was significantly associated with fibrosis. Prevalence and high grade (>20%) of steatosis were 41% and 11% in CC, 61% and 49% in CT, and 79% and 64% in TT, respectively (P = .01). Relative risk of developing high levels of steatosis was 20 times higher for TT genotypes than CC genotypes. According to multivariate analysis, steatosis was independently associated with hyperhomocysteinemia (OR = 7.1), HAI (OR = 3.8), liver fibrosis (OR = 4.0), and HCV genotype 3 (OR = 4.6). On univariate analysis, fibrosis was associated with age, steatosis, MTHFR, homocysteinemia and HAI; however, on multivariate analysis, liver fibrosis was independently associated with age (P = .03), HAI (P = .0001), and steatosis (P = .007). In conclusion, a genetic background such as the MTHFR C677T polymorphism responsible for hyperhomocysteinemia plays a role in the development of higher degree of steatosis, which in turn accelerates the progression of liver fibrosis in CHC. (HEPATOLOGY 2005.)

Hepatic steatosis frequently occurs in hepatitis C virus (HCV) infection, more so than in subjects with hepatitis B virus.1 The prevalence of fatty infiltration in the livers of chronic hepatitis C (CHC) has been reported to average around 50%, with a range of 30% to 70%.2–4 For HCV genotype 3 infection, it has been suggested that the virus exerts a direct steatogenic effect, supported by correlation of steatosis with the levels of HCV RNA, and by the reversal of steatosis as the result of antiviral therapy.5–8 Moreover, the degree of steatosis has been related to the extent of hepatic fibrosis in CHC.5, 9–12 The mechanisms involved in HCV-associated steatosis are unknown. Several host (body mass index [BMI], insulin resistance, hypertrigliceridemia) and viral factors may be implicated. Among viral factors, at least two HCV proteins—core and NS5A—have been credited with the ability to alter lipid metabolism in infected cells, thus causing hepatic steatosis in the absence of immune response.13, 14 In the transgenic mouse model, it has also been shown that the HCV core protein induces steatosis.15 Transgenic mice expressing HCV core protein showed a decrease in microsomal triglyceride transfer protein (MTP) activity, which is implicated in the hepatic assembly and secretion of apolipoprotein B containing very low-density lipoprotein.15 On this basis, it has been hypothesized that HCV core protein inhibiting MTP activity alters very low-density lipoprotein assembly and secretion and causes intracytoplasmic accumulation of triglycerides followed by development of steatosis.15 However, several findings do not fully support this hypothesis: (1) HCV core protein does not bind to either apolipoprotein B or MTP; (2) colocalization of MTP with HCV core has not been demonstrated; and (3) no decrease in serum triglycerides or apolipoprotein B has been demonstrated in the HCV core transgenic mice. Therefore, it is possible to hypothesize that other factors may be implicated in the development of HCV-related steatosis.

Patients with cystathionine β-synthase deficiency exhibiting severe hyperhomocysteinemia also develop marked hepatic steatosis.16 Similarly, homozygous cystathionine β-synthase–deficient mice with severe hyperhomocysteinemia develop hepatic steatosis.17 In addition, it has been reported that hyperhomocysteinemia alters intracellular lipid metabolism.18 Thus the data support the view that increased serum levels of homocysteine may be associated with hepatic fat accumulation. Methylenetetrahydrofolate reductase (MTHFR) catalyzes the conversion of 5,10-methylenetetrahydrofolate into 5-methyltetrahydrofolate, while the latter is absolutely required for remethylation of homocysteine to methionine. A common polymorphism for the gene encoding for the MTHFR, the C677T transversion, is associated with elevated plasma homocysteine levels.19 The prevalence of this polymorphism in the general population is estimated to be approximately 12% to 15% for the TT genotype, with considerable variation among different ethnic groups.19 At present, there are no data on the serum levels of homocysteine in CHC patients and its relationship with HCV-associated steatosis.

Accordingly, the aims of the present study were: (1) to evaluate whether HCV-induced liver steatosis is associated with hyperhomocysteinemia; (2) to assess their relationship with the MTHFR C677T polymorphism; and (3) to establish whether such metabolic and genetic alterations are related to the development and severity of steatosis and to the progression of liver fibrosis in CHC patients.


HCV, hepatitis C virus; MTHFR, methylenetetrahydrofolate reductase; CHC, chronic hepatitis C; HAI, histology activity index; BMI, body mass index; MTP, microsomal triglyceride transfer protein; PCR, polymerase chain reaction.

Patients and Methods

Study Design.

The hypothesis of an association between HCV-induced steatosis and plasma levels of homocysteinemia was tested in a prospective study over a 2-year period. At the end of the established period, 116 consecutive CHC patients admitted to our unit were evaluated, and statistical analysis showed an association between steatosis and homocysteine levels. At this time point, we decided to further characterize the population. Therefore, we evaluated the role of a common C677T genetic polymorphism for the MTHFR gene, whose relevant enzyme product plays a crucial role in the remethylation of homocysteine to methionine. Considering that, in the general population, frequency of the MTHFR “TT” genotype is within 12% to 15%, the estimated sample size to be included in the study was of 100 subjects for a power of 0.80, on the basis of an error of α = 0.05 and a supposed difference of μ0a = 0.50. All 116 patients were asked to sign informed consent to perform the genetic analysis that was successfully performed in only 102, due to a lack of consent or technical problems in 2 cases. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the local institutions' human research committee.


Subjects were enrolled if they met the following inclusion criteria: increased aminotransferase levels (> × 1.5 N) for at least 6 months, presence of anti-HCV and HCV RNA in serum, and compatible liver biopsy allowing us to stage HCV disease. Thus patients with clinically decompensated cirrhosis or contraindications to liver biopsy were not included in the study. Also, patients who were hepatitis B surface antigen– and/or HIV-positive and had other potential causes of liver disease, such as excessive alcohol intake (over 30 g/d) or features of autoimmunity (autoimmune hepatitis, primary biliary cirrhosis, sclerosis cholangitis) or metabolic liver disease (hemochromatosis, Wilson's disease), were excluded. Patients with severe cardiac or renal disease, overt diabetes, or active intravenous drug abuse were not included. None of the patients had received previous treatment with interferon or immunosuppressive agents or were taking medication that could cause steatosis (salicylates, nonsteroidal anti-inflammatory drugs, corticosteroids, valproic acid, amiodarone, perhexiline maleate) or modify serum levels of homocysteinemia (folic acid, vitamin B12).

At the time of the liver biopsy, a liver ultrasound scan was performed, epidemiological and demographic data were obtained, and BMI (kg/m2) was calculated for all patients.

Biochemical, Metabolic, Genetic, and Virological Studies.

All serum and plasma samples were collected at the time of the liver biopsy and stored at −80°C. All biochemical tests were performed on the same blood samples. Serum levels of folic acid and vitamin B12 were determined via commercial chemiluminescence assay (Architect; Abbott Laboratories, Abbott Park, IL).

Total homocysteine plasma levels were determined via fluorescence polarization immunoassay (AXYM; Abbott Laboratories). Blood samples were withdrawn in EDTA (1 mg/mL of blood), centrifuged, and stored within 15 minutes from collection to prevent in vitro homocysteine increase, because of its release by red blood cells. Hemolysed samples were not used for homocysteine analysis. Normative reference homocysteine levels were considered to be 12 or less (μmol/L) in males and 10 or less in females.

The presence of the MTHFR-C677T mutation was determined via polymerase chain reaction (PCR) amplification of genomic DNA obtained from patients' leukocytes via phenol/chloroform extraction.20 Genotypization, with respect to C677T polymorphism of the MTHFR gene, was accomplished by examining the occurrence of a HinfI recognition site.21 Briefly, a PCR 198-bp product was obtained via amplification using a Mastercycler gradient thermal cycler (Eppendorf, Hamburg, Germany). The PCR assay mixture contained 50 ng DNA, 0.8 μL dNTP (2.5 mmol/L), 0.5 μL (5 μmol/L) of each primer, and 1.0 μL of reaction buffer in the presence of 2.0 U of AmpliTaq polymerase (Roche Diagnostic SpA, Milan, Italy). PCR conditions were 5 minutes denaturation at 94°C followed by 30 cycles at 94°C for 1 minute, 58°C for 1 minute, and 72°C for 1 minute (primers 5′-CGAAGCAGGGAGCTTTGAGG-3′ and 5′-AGGACGGTGCGGTGAGAGTG-3′). PCR products were then digested with HinfI endonuclease, yielding a major 175-bp fragment in the presence of the C677T transition. Bands were resolved on 3% agarose gel electrophoresis. All determinations were repeated twice in two separate runs.

Anti-HCV antibodies were tested via ELISA (Abbott Laboratories, Chicago, IL) and HCV RNA was tested via PCR assay (Amplicor HCV; Roche Diagnostic SpA). Serum HCV RNA levels were assessed via commercial PCR (Monitor HCV Amplicor; Roche Diagnostic SpA). HCV genotyping was performed using a reverse hybridization line probe assay (INNO LiPA HCV assay, 2nd generation; Innogenetics, Zwijndrecht, Belgium) following the manufacturer's protocol. Hepatitis B markers and HIV antibodies were assayed using commercially available ELISA tests.

Histological Evaluation.

A pathological assessment was made according to Desmet et al.,22 who considered the scores for necroinflammatory changes and architectural alterations separately. The necroinflammatory activity was scored according to the histology activity index (HAI) by Knodell et al.23 The HAI was determined by combining the scores for portal inflammation (0-4), lobular degeneration and necrosis (0-4), and periportal necrosis (0-10). The stage was defined according to the Scheuer24 fibrosis score: 0, absence; 1, fibrous portal expansion; 2, periportal or portoportal fibrosis; 3, bridging fibrosis; 4, cirrhosis. Steatosis was graded as follows: 0, absent; 1, 1% to 10% of hepatocytes affected; 2, 11% to 30% of hepatocytes affected; 3, 31% to 60% of hepatocytes affected; and 4, greater than 60% of hepatocytes affected.

Statistical Analysis.

Absolute frequencies were compared using the χ2 test. Demographic and laboratory data were compared using the Student t test or Kruskal-Wallis test as appropriate. The Mann-Whitney U test was used to compare group differences in the serum HCV RNA levels and in the homocysteinemia levels. Dependence between variables was evaluated using Spearman's rank correlation. ANOVA was used to test for the association between variables. Bonferroni's post hoc test was used to evaluate multiple comparisons. The independent effects of the variables significantly associated with the increased grade of steatosis and progression of fibrosis in univariate analysis was assessed via stepwise multiple logistic regression analysis. The derived odds ration was used to determine the relative risk. Data analysis was performed with the SPSS computer program (SPSS for Windows, Chicago, IL). A 5% level was assumed to denote significance.


General Characteristics of the Study Patients.

The demographic and virological features and the hepatic histological alterations observed are reported in Table 1. Of the 116 patients, 50% had a BMI of 25 or higher; 58% were infected with HCV genotype 1, and 65.5% showed steatosis. Thirty percent of patients drank alcohol (<30 g/d) with an average of 15 g/d.

Table 1. Baseline Characteristics of the 116 CHC Patients Studied
  1. Abbreviations: ALT, alanine aminotransferase; γ-GT, γ-glutamyltransferase.

Demographic data 
 Male/female66/50 (57%/43%)
 Median age, years (range)53 (18–68)
 BMI, mean ± SD (range)26.08 ± 3.82 (17.8–44)
Epidemiological data 
 Transfusion12 (10%)
 Drug users10 (8.6%)
 Unknown94 (81.4%)
Virological parameters 
 Serum HCV RNA, UI × 105, median (range)8.0 (1.0–142.0)
 HCV genotypes 
  1a5 (4.3%)
  1b63 (54.3%)
  2a/c28 (24.1%)
  3a17 (14.7%)
  Mixed3 (2.6%)
Metabolic blood tests 
 Cholesterol, mg/dL (mean ± SD)176 ± 42.1
 Triglycerides, mg/dL (mean ± SD)104 ± 60.1
 Homocysteine, μmol/L, median (range)12.00 (4.78–50.0)
 Folic acid, ng/mL (mean ± SD)3.62 ± 1.1
 Vitamin B12, pg/mL (mean ± SD)592.4 ± 248.2
Hepatic functional blood tests 
 ALT, U/L, median (range)86 (55–529)
 γ-GT, U/L, median (range)46 (6–336)
Liver histology 
 Chronic hepatitis101 (87%)
 Cirrhosis15 (13%)
 HAI score, median (range)5.0 (1–13)
 Fibrosis score (mean ± SD)2.08 ± 1.10
 Number of patients with steatosis76 (65.5%)
 Distribution of grade of steatosis 
  125 (32.9%)
  221 (27.7%)
  312 (15.7%)
  418 (23.7%)

Steatosis, Host and Viral Factors.

There was no statistically significant association between age, sex, BMI, and steatosis (Table 2). Patients infected with HCV genotype 3 showed significantly higher prevalence and levels of steatosis than that observed in patients not infected with genotype 3 (P = .003). Moreover, patients infected with genotype 3 showed lower levels of cholesterol than that observed in genotype 1 and 2 infections (152 ± 36 mg/dL vs. 182 ± 37 mg/dL; P = .013). Overall, there was no correlation between serum HCV RNA levels and the prevalence or grade of steatosis.

Table 2. Demographic, Laboratory, and Histological Characteristics of the 116 Patients According to Grade of Steatosis
 Grade of SteatosisP Value
  • Abbreviations: γ-GT, γ-glutamyltransferase; ALT, alanine aminotransferase.

  • *

    Genotype 3 versus genotypes 1 and 2.

  • Grade 3–4 versus grade 1–2.

  • Grade 3–4 versus grade 0.

Number of patients40 (30.8%)46 (39.6%)30 (25.8%) 
Age in years, median (range)52 (18–68)53 (27–68)52 (29–64) 
BMI (mean ± SD)25.7 ± 4.726.0 ± 2.926.5 ± 4.1 
HCV RNA, UI × 105, median (range)5.4 (1.2–57)5.6 (0.8–142)5.1 (1.0–30) 
HCV genotype    
 1a–1b (n = 68)34%44%22% 
 2a/c (n = 28)50%28.5%21.5% 
 3a (n = 17)18%29%53%.003*
γ-GT, U/L, median (range)35 (14–223)50 (14–336)75 (16–274).002
ALT, U/L, median (range)72 (55–357)91 (55–529)102 (63–326).005
Cholesterol, mg/dL (mean ± SD)174 ± 34178 ± 46186 ± 48 
Triglycerides, mg/dL (mean ± SD)83 ± 31105 ± 46126 ± 96.003
Homocysteinemia, μmol/L (mean ± SD)9.2 ± 3.212.8 ± 6.6321.1 ± 12.0.0001
Folic acid, ng/mL (mean ± SD)3.30 ± 0.93.56 ± 0.93.62 ± 1.4 
Vitamin B12, pg/mL (mean ± SD)567 ± 188548 ± 264648 ± 320 
HAI score, median (range)4.0 (1–11)6.0 (1–13)7.0 (2–14).03
Fibrosis score (mean ± SD)1.67 ± 1.132.09 ± 0.582.60 ± 0.92.002

Steatosis, Histological Features, and Liver Function Tests.

A progressive, significant increase in HAI score was observed in relation to the grades of steatosis (median; no steatosis, 4.0 [range 1–11] vs. steatosis grades 1–2, 6.0 [range 1–13] vs. steatosis grades 3–4, 7.0 [range 2–14]; P = .03 [see Table 2]). A significant correlation between the grades of steatosis, the levels of HAI (r = 0.206; P = .02 [Table 3]), and degree of fibrosis was observed (P = .002 [see Table 2] and r = 0.462, P = .001 [see Table 3]).

Table 3. Multiple Correlations Between the 116 CHC Patients (All Results Are Two-Tailed)

Increased serum levels of alanine aminotransferase and γ-glutamyltransferase were significantly associated with the grade of steatosis (P = .002 and P = .005, respectively [see Table 2]).

According to univariate analysis (see Table 3), levels of liver fibrosis were associated with age, steatosis, and HAI (P = .001). Sex, BMI, genotype, and viremia levels did not correlate with liver fibrosis. Patients with steatosis greater than 20% showed significantly higher levels of both HAI (P = .008) and liver fibrosis (P = .0001) than those observed in subjects with less than 20% steatosis and those without steatosis (Table 4). The levels of HAI and fibrosis in patients with steatosis less than 20% were not significantly different than those observed in subjects without steatosis.

Table 4. Levels of HAI and Fibrosis in CHC Patients in Relationship to Less Than or Greater Than 20% of Liver Steatosis
 No Steatosis (n = 40)Steatosis (n = 76)P Value
<20% (n = 39)≥20% (n = 37)
  • *

    Greater than 20% versus less than 20% and no steatosis.

  • Greater than 20% versus no steatosis (.02 vs. <20%).

HAI median (range)4.0 (1–11)5.0 (1–10)6.0 (1–13).008*
Fibrosis, mean ± SD1.67 ± 1.131.92 ± 1.122.76 ± 1.10.0001

Steatosis, Homocysteine Levels, and Progression of Liver Damage.

Correlations among the grade of steatosis, plasma levels of homocysteine, cholesterol and triglycerides, folic acid, vitamin B12, and histological liver damage were evaluated. As shown in Table 2, the mean values of plasma homocysteine were significantly higher in subjects with grade 3–4 of steatosis than that observed in patients with grade 1–2 of steatosis and those without steatosis. In fact, CHC patients without steatosis had mean levels of homocysteinemia in the normal range (9.2 ± 3.2 μmol/L), whereas an increase in homocysteinemia, although not significant, was observed in grade 1–2 of steatosis (12.8 ± 6.6, P = .096 vs. no steatosis). Instead, a significant increase in grades 3–4 (21.8 ± 12.0 μmol/L, P = .0001 vs. no steatosis and P = .0001 vs. grade 1–2) could be observed. This association was independent of sex, BMI, and age. Serum levels of both folate and vitamin B12 did not correlate with plasma levels of homocysteine or grade of steatosis (see Table 2). As shown in Table 5, 92% of patients with homocysteinemia levels of 14 μmol/L or more had steatosis. At univariate analysis, in addition to steatosis, homocysteinemia was significantly associated with liver fibrosis (P = .031) and MTHFR (P = .001 [see Table 3]). At multivariate analysis, homocysteine plasma levels were associated only with steatosis (odds ratio [OR] = 0.134; 95% confidence interval [CI] 0.061–0.208; P = .0001) and MTHFR (OR = 4.188; 95% CI 1.041–7.335; P = .01).

Table 5. Prevalence of Steatosis in CHC Patients With Plasma Levels of Homocysteine Greater Than or Less Than 14 μmol/L
 HomocysteinemiaP Value
<14 μmol/L≥14 μmol/L
  1. NOTE. Number of patients appears in parentheses.

Patients with steatosis52.5% (41)92% (35) 
Patients without steatosis47.5% (37)8% (3).0001

Serum cholesterol levels were not significantly different for HCV patients with steatosis compared with those without steatosis, whereas triglyceride levels were significantly higher in patients with steatosis (P = .001 [see Table 2]). There was no association between cholesterol or triglycerides and homocysteinemia (data not shown). Homocysteine levels in patients with cirrhosis were not significantly higher than those in patients without cirrhosis (data not shown).

Upon univariate analysis, the grades of steatosis were associated with homocysteine plasma levels (P = .001), liver fibrosis (P = .001), HCV genotype 3 (P = .001), HAI (P = 0.031), γ-glutamyltransferase (P = .008), and tryglicerides (P = .005). There was no association between the grades of steatosis, sex, and serum HCV RNA levels (P = .44). As shown in Table 6, upon multivariate analysis the grades of steatosis (>1) were independently associated with levels of homocysteinemia (OR = 7.1; 95% CI 3.1–16; P = .02), liver fibrosis scores (OR = 4.0; 95% CI 1.7–9.2; P = .02), genotype 3 (OR = 4.6; 95% CI 1.5–14; P = .012), and HAI score (OR = 3.8; 95% CI 1.7– 8.4; P = .05).

Table 6. Multivariate Analysis for Factors Associated With Steatosis Grades in 116 CHC Patients
 OR*95% CIP Value
  • *

    Steatosis grade >1 (>10%) versus homocysteinemia >12 μmol/L, versus HAI >4, versus liver fibrosis >1.

Liver fibrosis4.01.7–9.2.02
Genotype 34.61.5–14.012

C677T Polymorphism of MTHFR Gene, Homocysteinemia, and Steatosis.

The test for the C677T polymorphism of MTHFR was successfully performed in 102 CHC patients. Thirty-eight percent of patients showed the CC genotype, 48% were heterozygous (CT), and 14% were homozygous (TT). In Table 3, the variables significantly associated with the MTHFR polymorphism are reported. In particular, a correlation between the MTHFR polymorphism and the degree of steatosis (r = 0.388; P = .001), liver fibrosis (r = 0.252; P = .017), and homocysteinemia (r = 0.367; P = .001) was observed. The MTHFR polymorphism did not correlate with the HCV genotype and hepatic necroinflammatory activity. Table 7 shows the relationship between the MTHFR polymorphism, homocysteinemia, and steatosis. The median values of homocysteinemia for CC, CT, and TT were 9.33 (range 4.78–46), 12.27 (range 6.71–50), and 18.63 (range 7.70–50), respectively (P = .006 and .0001, TT vs. CT and CC, respectively). Prevalence of patients with above normal levels of homocysteinemia (≥12 μmol/L) was 26%, 55%, and 71% for CC, CT, and TT, respectively (P = .007 for TT and .01 for CT vs. CC). Prevalence of steatosis among the three genomic types was 41%, 61%, and 79% for CC, CT, and TT, respectively (P = .036, TT vs. CC). Median steatosis levels were 0% (range 0%–95%) in patients carrying the CC genotype, 20% (range: 0%–95%) in those carrying the CT genotype, and 30% (range: 0%–80%) in those carrying the TT genotype (P = .006 for TT and .023 for CT vs. CC). As previously demonstrated5 and confirmed in this study, steatosis levels above 20% are of clinical relevance in accelerating liver fibrosis in CHC patients. Therefore, we calculated the relative risk of developing greater than 20% steatosis associated with the MTHFR heterozyous and homozygous state for the T allele. Of the HCV patients carrying CC genotype, only 11% had a high degree of steatosis (>20%), whereas in the CT and TT genotypes the prevalence was 49% and 64%, respectively (P = .0001). As shown in Table 7, HCV patients carrying the CT genotype have an increased risk of developing high levels of steatosis of approximately sixfold greater (OR = 6.4; 95% CI 2.2–24.5; P = .001) than that observed for patients with the CC genotype. Whereas patients carrying the TT genotype for MTHFR have an increased risk of 20-fold greater than that observed in patients with CC (OR = 20; 95% CI 1.6–89.1; P = .0001) and 1.9-fold greater than that observed in patients with the CT genotype (OR = 1.9; 95% CI 0.5–6.5; P = .01).

Table 7. Relationship Between MTHFR Polymorphism, Homocysteinemia, and Steatosis
MTHFR Gene (No. of Patients)Homocysteinemia, μmol/L (Range)Prevalence of Patients With Homocysteinemia >12 μmol/LMedian Steatosis Levels (Range)Prevalence of Patients With SteatosisPrevalence of Patients With Steatosis >20%OR (95% CI) Steatosis >20%
  • *

    P = .01 versus CC.

  • P = .023 versus CC.

  • P = .0001 versus CC.

  • §

    P = .006 versus CT and .0001 versus CC.

  • P = .007.

  • P = .006.

  • #

    P = .036 versus CC.

  • **

    Versus CT.

CC (39)9.33 (4.78–49)26%0% (0–95)41%11%1.0
CT (49)12.27 (6.71–50)55%*20% (0–95)61%49%6.4 (2.2–24.5)
TT (14)18.63 (7.70–50)§71%30% (0–80)79%#64%20.0 (1.6–89.1)1.9 (0.54–6.5)**

Overall Analysis of Variables Associated With Progression of Liver Fibrosis.

All variables that were significantly associated with liver fibrosis upon univariate analysis (age, steatosis, HAI, homocysteinemia, and MTHFR) were included in the multivariate analysis. As shown in Table 8, progression of liver fibrosis was independently associated with age, steatosis, and HAI.

Table 8. Multiple Linear Regression Analysis for Factors Independently Associated With the Progression of Fibrosis in 116 CHC Patients
 Coefficient BSE of Coefficient B95% CIP Value
  1. NOTE. Coefficient = 0.513 (SE = 0.196).



The results of this study demonstrate that, in chronic hepatitis C patients, higher prevalence and levels of steatosis are associated with hyperhomocysteinemia and with the MTHFR C677T polymorphism and confirm that steatosis is an independent cofactor related with an increased HAI and progression of liver fibrosis. Such associations are independent of host factors such as age, sex, BMI, and levels of serum cholesterol and triglycerides, as well as of viral factors (HCV genotypes and levels of HCV RNA). The data demonstrate that a genetic background that induces hyperhomocysteinemia plays an important role in the development and severity of HCV-associated steatosis, which in turn accelerates the progression of HCV-related diseases. The data also demonstrate that for homocysteinemia levels of 14 μmol/L or more, almost all CHC patients (92%) develop steatosis (see Table 5).

Subjects with increased homocysteine plasma levels show increased risk of developing fatty liver,16 the data being consistent with results obtained in animal models of hyperhomocysteinemia.17, 25 Schwahn et al., using MTHFR knockout mice as a model for moderate and severe hyperhomocysteinemia, demonstrated that such mice spontaneously develop marked steatosis and that such steatosis may be prevented by betaine supplementation.25 Hyperhomocysteinemia promotes the development and progression of hepatic steatosis by inducing endoplasmic reticulum stress, which activates the unfolded protein response and the sterol regulatory element-binding proteins, a molecule family that regulates the expression of genes responsible for cholesterol/triglyceride biosynthesis and uptake.18, 26 Homocysteine does not impair lipid export from the liver in vivo.18 In addition, low-density lipoprotein uptake was increased in HepG2 cells treated with homocysteine18 as the result of an upregulation of low-density lipoprotein receptor gene expression, which may be linked to the proposed mechanism of HCV tropism for the hepatocytes via low-density lipoprotein mimicking.27 Thus homocysteine-induced endoplasmic reticulum stress causes fatty liver accumulation by an increased hepatic uptake and biosynthesis of cholesterol and triglycerides without impairing the hepatic export of lipids.18, 26 The hyperhomocysteine-induced steatosis model,18, 26 in contrast to the proposed MTP model that is associated with a reduced lipid synthesis and secretion,15 concurs with HCV-associated steatosis. In fact, the overall analysis of the HCV-infected patients in this study shows that there were no differences in serum levels of cholesterol in patients with and without steatosis, while triglycerides were increased in patients with steatosis (see Table 2). In addition, the hyperhomocysteinemia-induced steatosis model may explain why only a certain number, but not all HCV-infected patients, develop steatosis, and why only a minority of patients (e.g., those with a higher homocysteine levels) accumulate a greater amount of fat in the liver. However, because steatosis behaves differently for the different HCV genotypes (e.g., genotype 3 infection is strongly associated with a high prevalence and severity of steatosis), and considering that genotype 3 patients show lower levels of serum cholesterol,28 it is likely that in genotype 3 infection, as well as in other genotypes, other factors besides hyperhomocysteinemia may contribute to the development and severity of steatosis.

The interrelationship between HCV infection, hyperhomocysteinemia, and steatosis may shed light on the recent report of HCV infection as a risk factor for atherosclerosis in a large series from Japan.29 Hyperhomocysteinemia is a well-known risk factor for atherothrombosis,30 and its association with HCV-induced steatosis might account, at least in part, for the association between HCV infection and atherosclerosis.

Subjects who carry the TT genotype of the MTHFR C677T polymorphism show significant hyperhomocysteinemia and increased cardiovascular risk.21 It is also reported that moderate hyperhomocysteinemia may be present in subjects carrying the CT genotype. Therefore, we studied this polymorphism in the HCV-infected patients included in this study. Our data demonstrate that a genetic background may favor the development and severity of steatosis in CHC patients. In fact, HCV-infected patients carrying the TT genotype of the MTHFR gene showed significantly higher abnormal mean levels of homocysteinemia (21.43 ± 14.26 μmol/L), and higher levels of steatosis than those observed in subjects carrying the CT and the CC genotypes (P = .001). In addition, the prevalence of hyperhomocysteinemia (>12 μmol/L), within genotype groups was 26%, 55%, and 71% for CC, CT, and TT, respectively (P = .007). More pertinent to the aim of this study is the association between the C677T polymorphism of the MTHFR gene and HCV-related steatosis (r = 0.388; P = .001), and that the estimated relative risk of developing higher grades of steatosis (greater than 20%) was six times higher for patients with the CT genotype and 20 times higher for patients with the TT genotype (see Table 6). Such high grades of steatosis are of relevant clinical importance, because they are associated with an accelerating progression of liver fibrosis in chronic hepatitis C5. In this respect, it is important to underscore that HCV patients carrying CC genotype showed a lower prevalence of steatosis and, more importantly, only 11% of these patients had a level of steatosis greater than 20%. In contrast, HCV patients who were heterozygous and homozygous for the MTHFR gene C677T polymorphism showed a high grade of steatosis (>20%) in 49% and 64% of cases, respectively (P = .001).

It is worth noting that the slightly increased levels of homocysteine observed in patients carrying the CT genotype cannot explain the increase in prevalence and grade of steatosis in this subgroup; thus other factors may be involved, such as combined heterozygosity of C677T with A1298C.

Present data indicate that a genetic background that includes hyperhomocysteinemia aggravates HCV-induced steatosis, accelerating the progression of liver disease. The fact that both the MTHFR polymorphism and hyperhomocysteinemia are associated with liver fibrosis only upon univariate analysis, but not upon multivariable analysis, indicates that they may affect liver progression through the development of higher grade of steatosis. Hyperhomocysteinemia may also be induced by liver damage per se, because it has been reported that homocysteine metabolism may be altered in cirrhosis.31, 32 However, the data do not support this hypothesis in this setting of patients, because the levels of homocysteine were not significantly higher in patients with cirrhosis than those observed in CHC with high levels of fibrosis. Furthermore, we cannot exclude an interaction between HCV or HCV components with the homocysteine metabolism, which could be particularly significant in subjects with a favorable genetic background.

This study, in concert with previous reports,5, 9–12 is a further demonstration that hepatic steatosis, age, and HAI are independent factors associated with the progression of liver fibrosis in CHC. The development of HCV-associated steatosis seems to be a multifactorial process in which both host (BMI, insulin resistance, hypertrigliceridemia) and viral factors (HCV genotype) may be involved.33 The present study provides evidence that a genetic background, such as the MTHFR polymorphism through hyperhomocysteinemia-induced derangement of lipid metabolism, may contribute to the development of higher degrees of steatosis, which in turn accelerates the progression of liver fibrosis in chronic HCV infection. Potential mechanisms of this effect may include the increased sensitivity of steatotic livers to oxidative stress, cytokine-mediated injury, and steatosis-related hepatic insulin resistance33.

The results of this study are of potential clinical, prognostic, and therapeutic importance. Hyperhomocysteine and/or the MTHFR polymorphism may be indicators of early metabolic derangements of liver function related to the development of HCV steatosis. They may help us to identify a subgroup of HCV-infected patients who are at a higher risk of developing a severe degree of steatosis. This, in turn, is an important cofactor that triggers the increase of both hepatic necroinflammatory activity and serum levels of liver enzymes (γ-glutamyltransferase, alanine aminotransferase), as well as the acceleration the progression of liver fibrosis in CHC. Moreover, this study could conceivably help us to understand the mechanisms through which HCV interacts with cellular factors and promotes the progression of liver disease. Thus clinical trials aimed at evaluating a metabolic approach to the treatment of CHC (e.g., weight reduction, homocysteine-lowering measures, elimination of alcohol intake) are welcome in the hope of reducing the progression and symptoms of CHC, even when antiviral treatment is unsuccessful.


The authors thank Fiorillo Geltrude and Ciro Carangio for their excellent technical assistance.