Potential conflict of interest: Nothing to report.
Previous studies examining the relationship between HFE mutations and severity of nonalcoholic steatohepatitis (NASH) have been limited by small sample size or ascertainment bias. The aim of this study was to examine the relationship between HFE mutations and histological severity in a large North American multicenter cohort with NASH. Data from 126 NASH patients were collected from 6 North American centers. Liver biopsy and genotyping for the C282Y and H63D HFE mutations were performed in all subjects. Serum transferrin–iron saturation and ferritin levels as well as hepatic iron content were recorded whenever available. Univariate and multivariate logistic regression analyses were performed to identify factors associated with advanced hepatic fibrosis. The prevalence of heterozygous C282Y and H63D HFE mutations was 14.3% and 21.4%, respectively, in the overall cohort. Among Caucasians, C282Y heterozygotes were more likely to have bridging fibrosis or cirrhosis (44% versus 21% [P = 0.05]) and stainable hepatic iron (50% versus 16% [P = 0.011]) compared with patients with other genotypes. Diabetes mellitus was the only independent predictor of advanced hepatic fibrosis (OR 4.37, 95% CI 1.41-13.54 [P = 0.010]) using multiple logistic regression analysis adjusting for age, sex, ethnicity, body mass index, and HFE genotype status. Conclusion: The HFE C282Y heterozygous mutation is associated with advanced fibrosis among Caucasians with NASH. Additional studies are warranted to examine the possible mechanisms for this relationship. (HEPATOLOGY 2007.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Nonalcoholic Fatty Liver Disease (NAFLD) encompasses the spectrum of liver disorders ranging from “simple steatosis” to steatosis with hepatic necroinflammation, liver cell ballooning with or without Mallory's hyaline with varying degrees of fibrosis (nonalcoholic steatohepatitis [NASH]) in the absence of significant alcohol consumption (<20 g/day in women, <30 g/day in men).1 There is an increasing prevalence of NAFLD in the United States and worldwide,2–4 and this liver disorder is now considered to be the most common cause of elevated serum aminotransferase levels identified through routine blood testing.5 NASH, a more serious form of NAFLD, may ultimately lead to cirrhosis, hepatocellular carcinoma, and end-stage liver disease. A commonly accepted model for the pathogenesis of NASH is the so-called “two-hit” hypothesis, wherein the first “hit” leads to accumulation of hepatic free fatty acids resulting in a histological picture of macrovesicular steatosis.6, 7 A subsequent “hit” associated with oxidative stress may result in liver injury, which over time may lead to eventual development of hepatic fibrosis and possible progression to cirrhosis.
Hereditary hemochromatosis is associated with deposition of iron in many parenchymal tissues and may lead to cirrhosis and hepatocellular carcinoma, cardiomyopathy, diabetes mellitus, arthropathy, and skin pigment changes.8 A homozygous mutation in the hemochromatosis (or HFE) gene is present in the majority of Caucasians with the phenotype of hemochromatosis type 1.9 Individuals heterozygous for HFE mutations may have mildly increased iron stores but are generally not believed to be at risk for complications from iron overload unless another cause for iron overload coexists.
However, several studies have suggested that heterozygosity for HFE mutations may exacerbate the severity of other chronic liver diseases, presumably via increased iron accumulation.10, 11 We have previously shown that HFE mutations are associated with increased hepatic iron loading and more rapid progression of liver disease among patients with chronic hepatitis C.11
It has been suggested that excess iron deposition may contribute to oxidative stress within the liver via the production of reactive oxygen species.6 It is therefore plausible that HFE mutations may also contribute to oxidative stress via hepatic iron loading among patients with NASH, serving as the second hit leading to more severe disease. Previous studies examining whether HFE mutations contribute to the severity or progression of liver disease among patients with NASH have found conflicting results.12–22 Furthermore, many such studies were likely limited by inadequate sample size and/or possible referral or ascertainment bias.23, 24 Thus the relationship between HFE gene mutations, iron status, and severity of NASH remains unclear. The goals of the present study were to (1) test the hypothesis that HFE mutations are independently associated with advanced hepatic fibrosis among patients with NASH and (2) examine whether NASH patients with HFE mutations have increased hepatic iron content.
Patients were enrolled prospectively and retrospectively into the study from the following 6 North American centers: the University of Washington Medical Center, the Mayo Clinic, the University of Toronto, Indiana University, The University of Texas Medical Branch, and the University of Calgary. The study was approved by the human subjects institutional review board at each participating institution. Inclusion criteria were: ≥18 years of age; minimal alcohol use (<20 g/day in women, <30 g/day in men) as reported by the patient, the patient's family, and the patient's referring physician; and appropriate exclusion of other liver diseases, including viral hepatitis, autoimmune hepatitis, drug-induced liver disease, primary biliary cirrhosis, Wilson's disease, and α1-antitrypsin deficiency. All patients met the criteria for NASH as defined by Brunt et al.25 (i.e., macrovesicular steatosis and lobular inflammation along with associated features such as hepatocyte ballooning degeneration with or without fibrosis). Exclusion criteria were as follows: (1) a known history of phenotypic hemochromatosis, defined as C282Y homozygosity, and either hepatic iron index >1.9 or 3 to 4+ hepatic iron staining; (2) use of medications known to cause steatosis (e.g., steroids, high-dose estrogens, and amiodarone) in the year prior to enrollment; and (3) a history of jejunoileal bypass surgery.
Clinical data collected included body mass index and presence of diabetes mellitus. Serum ALT, AST, cholesterol, triglycerides, glucose, iron, ferritin, and transferrin–iron saturation were recorded whenever available as follows: serum iron indices, including serum iron (n = 88), transferrin–iron saturation (n = 74), and ferritin (n = 81); serum aminotransferase levels (n = 99); and fasting triglycerides and total cholesterol (n = 87). In all cases, values closest to the time of liver biopsy were recorded. Genotyping for the HFE H63D or C282Y mutations were performed using previously described methods.11
Liver biopsies were performed and processed according to the standard protocols at each participating center. All biopsies were read by experienced hepatopathologists as part of baseline pretreatment biopsy evaluations to enter various treatment studies for NASH or for routine staging. Criteria for the diagnosis of NASH were based on the method of Brunt et al.25 after routine staining with H&E and Masson's trichrome for assessment of fibrosis. Perls' Prussian Blue stain was used to determine iron staining grade (0 to 4+) according to previously described methods.26 The data on fibrosis and iron staining score were captured from pathology records by study personnel at each site. Biopsies were classified as showing “advanced” fibrosis if bridging fibrosis or cirrhosis was described or if the stage was given as 3 or 4. Fibrosis was classified as “none or mild” if the report described fibrosis as absent, mild, or moderate or if fibrosis stage was reported as 0-2. Hepatic iron concentration was measured from paraffin-embedded liver tissue using previously described methods.27 Adequate tissue for analysis (0.4 mg) could be obtained in the majority of subjects (93/126 [74%]).28 Hepatic iron concentration was expressed in μg/g dry weight, and hepatic iron index was calculated (μmol/age).
Descriptive statistics of continuous variables were calculated and expressed as means ± SD. Descriptive statistics of categorical data were calculated and expressed as a proportion. Several variables were nonnormally distributed; continuous variables were compared between groups using a t test for parametric data or the Mann-Whitney U test for nonparametric data. Categorical variables were compared between groups using the Mann-Whitney U test, χ2 test, or Fisher exact test, where appropriate. Univariate associations between variables of interest were calculated and expressed as P values. Multivariate analysis using logistic regression was used to investigate variables leading to the presence of advanced fibrosis. All statistical procedures were performed using Stata software version 9.0 (Stata, College Station, TX).
Demographic and Laboratory Characteristics.
The clinical and laboratory features of the 126 NASH patients are shown according to ethnicity in Table 1. The average age of the subjects was 47.9 years. There were more men (52.4%) than women and most patients were of Caucasian descent (77.7%). The study also included 9.5% Asians, 9.5% Hispanics, and 1.6% Native Americans. Ethnicity was unknown in 2 subjects.
Table 1. NASH Patient Characteristics According to Ethnicity
Abbreviations: BMI, body mass index; HIC, hepatic iron content; NA, not available; TIBC, total iron binding capacity.
P < 0.01 compared with Caucasian (Mann-Whitney U test, Student's t test, or χ2 test).
P < 0.05 compared with Caucasian (Mann-Whitney U test, Student's t test, or χ2 test).
Asians had a significantly lower mean body mass index (29.0 ± 3.6 versus 32.7 ± 5.3 [P = 0.016]) and lower total iron binding capacity (315.6 ± 34.4 versus 356.8 ± 59.6 [P = 0.04]) compared with Caucasians. Asians also had significantly higher mean serum ALT levels (129.1 ± 57.1 versus 90.2 ± 63.9 [P = 0.037]). Serum ferritin levels tended to be higher in Asians compared with other ethnic groups, although the difference was not statistically significant. Caucasians had significantly lower serum ALT levels (P = 0.008) compared with other ethnic groups.
Prevelance of HFE Mutations and Association With Serum and Hepatic Iron Studies.
The prevalence of heterozygous C282Y (C282Y+/−), heterozygous H63D (H63D+/−), and C282Y/H63D compound heterozygous HFE mutations overall was 14.3%, 21.4% and 4.8%, respectively. The prevalence of these mutations among Caucasians (n = 98) was 16.3%, 23.4%, and 5.1% respectively, which is not significantly different than the prevalence of these HFE mutations reported among Caucasians in the general population in 2 recent large North American studies.29, 30
Mean levels of serum iron indices are shown according to HFE genotype in Table 2. Compared with wild-type (WT) individuals, C282Y/H63D compound heterozygotes had significantly higher serum iron (126.0 ± 37.0 versus 87.4 ± 34.1 [P = 0.029]), ferritin (657.3 ± 254.6 versus 321.7 ± 285.8 [P = 0.005]), and transferrin–iron saturation (42.8 ± 7.05 versus 25.5 ± 11.2 [P = 0.004]). There was no significant difference in serum iron studies between C282Y or H63D heterozygotes compared with WT patients.
Table 2. NASH Patient Characteristics According to HFE Genotype
Any HFE Mutation
Abbreviations: BMI, body mass index; HIC, hepatic iron content; TIBC, total iron binding capacity; WT, wild-type.
P < 0.05 compared with WT+/+ (Mann-Whitney U test, Student's t test, or χ2 test).
P < 0.01 compared with WT+/+ (Mann-Whitney U test, Student's t test, or χ2 test).
Hepatic iron concentration (available in 55/75 WT patients and 38/51 patients with HFE mutations) was higher in C282Y heterozygotes and C282Y/H63D compound heterozygotes but not H63D heterozygotes compared with WT individuals, although these differences did not reach statistical significance (Table 2). Hepatic iron staining score was available in 58 of 75 WT patients and 39 of 51 patients with HFE mutations. C282Y/H63D compound heterozygotes had a significantly greater proportion of hepatic iron staining compared with WT patients (P = 0.033) (Table 2). Patients with HFE mutations were more likely to have stainable hepatic iron compared with WT patients (28% versus 19%, respectively) (Table 2) and C282Y heterozygotes were even more likely to have any stainable iron in the liver compared with WT patients (43% versus 19%, respectively) (Table 2). Among Caucasians, stainable hepatic iron was significantly more common among C282Y heterozygotes compared with WT patients (50% versus 15% [χ2 = 6.37, P = 0.012]). However, interpretation of iron staining data was limited by the lack of detailed information on pattern and distribution of iron staining in the liver.
Factors Associated With Fibrosis.
The proportion of subjects with either no or mild hepatic fibrosis (defined as stage 0-2) compared with subjects with advanced hepatic fibrosis (stage 3-4) according to the HFE genotype is shown in Fig. 1. Advanced hepatic fibrosis was significantly more common among Caucasian patients carrying the C282Y+/− mutation (P = 0.05) compared with WT Caucasian patients (Fig. 1B). Univariate logistic regression analysis was used to investigate possible risk factors associated with advanced fibrosis. A history of diabetes mellitus was the only variable predictive of advanced fibrosis (OR 3.90, 95% CI 1.55-9.84 [P = 0.004]). There was a trend toward significance for the association of the heterozygous C282Y mutation (C282Y±) with advanced hepatic fibrosis (OR 2.35, 95% CI 0.82-6.75 [P = 0.112]), which was even stronger among Caucasians alone (n = 98) (OR 2.97, 95% CI 0.97-9.14 [P = 0.057]). When multiple logistic regression modeling was employed adjusting for age, sex, ethnicity, body mass index, and HFE genotype status, diabetes mellitus again remained the only independent predictor of advanced hepatic fibrosis (OR 4.37, 95% CI 1.41-13.54 [P = 0.010]).
Investigation of Possible Ascertainment Bias.
To address the issue of possible ascertainment bias affecting the prevalence or significance of HFE mutations in this cohort, we reviewed the reasons for referral in each case. The reason for referral to a hepatologist could be determined in 72 of the NASH subjects, 19 of whom were referred for 1 or more of the following conditions: elevated iron studies, a family history of hemochromatosis, and known HFE mutations. The C282Y+/− and H63D+/− mutation frequency among the remaining Caucasian patients with NASH (n = 40) was 17.5% and 32.5%, respectively, after excluding these 19 patients, which was not significantly different from the overall cohort. In this subset, 37% were WT patients (1 of 7 had advanced fibrosis), 32% were C282Y+/− (1 of 6 had advanced fibrosis), 21% were C282Y/H63D compound heterozygotes (0 of 4 had advanced fibrosis), and 11% were H63D heterozygotes (0 of 2 had advanced fibrosis). In the remaining subset of 40 Caucasians, the association of the C282Y+/− mutation with advanced hepatic fibrosis remained significant (χ2 = 4.17, P = 0.041); therefore, we think ascertainment bias is unlikely to have confounded our results.
Our data suggest that the presence of the C282Y+/− mutation is a risk factor for development of advanced hepatic fibrosis (bridging fibrosis or cirrhosis) among Caucasian patients with NASH. The prevalence of the C282Y+/−, H63D+/−, and C282Y/H63D HFE mutations overall was 14.3%, 21.4%, and 4.8%, respectively. These mutations were more common among Caucasians (n = 98); 16.3%, 23.4%, and 5.1%, respectively. However, the prevalence of HFE mutations was not significantly different than that reported among Caucasians in the general population in 2 recent large North American studies (together comprising more than 100,000 subjects from 6 geographically diverse centers),29, 30 even after exclusion of 19 subjects that were referred for evaluation of possible hemochromatosis. Therefore, our study suggests that the prevalence of HFE mutations among Caucasian patients with NASH is not higher than expected in the general population.
Several previous reports with smaller sample size have found either an increased prevalence of HFE mutations in NASH patients and/or an association between HFE mutations and severity of NASH12–15 (Table 3). In addition, Valenti et al.16 found a higher prevalence of the C282Y mutation among 134 Italian patients with NAFLD (based on ultrasonographic findings of an echogenic liver) compared with normal control subjects; however, the prevalence of C282Y mutations was not increased in a smaller subset of 42 patients with biopsy-proven NASH. Bonkovsky et al.14 studied 57 patients with NASH, including 36 patients who underwent genotyping for HFE mutations; the prevalence of HFE mutations among patients with NASH in that smaller study was similar to the findings in our study. These authors also found that HFE mutations were associated with advanced hepatic fibrosis; Bonkovsky et al. did not find a significant difference in hepatic iron content (HIC) between patients with and without HFE mutations, or a significant relationship between HIC, hepatic iron stain, and degree of fibrosis.14 By contrast, George et al.15 determined that HIC or hepatic iron staining but not carriage of the C282Y mutation was significantly associated with severe fibrosis. However, because the C282Y mutation was significantly associated with both HIC and hepatic iron staining, the authors postulated that the C282Y mutation is indirectly associated with an increased risk of fibrosis in NASH via increased hepatic iron stores. Four studies have reported that patients with NASH who carry HFE mutations have higher hepatic iron stores compared with those WT for HFE.12, 13, 15, 16 Several other studies have not found an increased prevalence of HFE mutations among NASH patients or an association between presence of HFE mutations and disease severity.17–21 Three of these studies were conducted in countries where HFE mutations are infrequent,31 namely, Brazil,19, 20 Japan,21 and India.22 Two additional studies in different Italian NAFLD cohorts also did not find an association between presence of HFE mutations and fibrosis stage.17, 18
Table 3. Sample Sizes and HFE Genotype Frequencies in Previous Studies
Abbreviations: NR, not reported; Pwt, frequency of stage 3-4 fibrosis patients among WT+/+ cases; PC282Y, frequency of stage 3-4 fibrosis patients among C282Y+/− cases.
One possible explanation for differences in findings from previous studies is the lack of sufficient power to address the role of HFE mutations in the disease severity of NASH because of the lack of sufficient numbers of patients with both advanced fibrosis and HFE mutations. Our study included the largest sample of C282Y heterozygotes with advanced fibrosis reported to date (Table 3). Furthermore, our study is the first to attempt to account for potential ascertainment bias and one of the few to examine whether HFE mutations are independently associated with advanced fibrosis.
In parallel with our observation that heterozygous C282Y mutations were associated with advanced fibrosis in Caucasians, we also found that stainable hepatic iron was more common among C282Y+/− Caucasian patients compared with WT patients suggesting that hepatic iron loading may be a factor in the development of advanced fibrosis among Caucasian patients with NASH who carry the C282Y mutation.
However, it should be emphasized that determination of hepatic iron may be inaccurate in the severely steatotic liver because iron granules may be displaced to the side of the cell and be fewer in number; in addition, histopathological assessment of iron staining may or may not include sinusoidal iron deposition. It is also possible that sampling and biological variability in iron deposition may be greater in steatotic liver tissue, leading to inaccurate measurements in either hepatic iron staining or biochemical determination. Another source of variation in HIC is related to the sample processing of the formalin-fixed paraffin-embedded tissue we used to measure HIC. When the paraffin-embedded tissue is de-waxed with xylene, fat in the steatotic liver tissue may also be eliminated. This may result in reduced tissue dry weight of the sample leading to spuriously increased HIC in patients with severe steatosis. This could explain the disparity between our HIC and hepatic staining results if a greater proportion of WT subjects compared with C282Y+/− had severe steatosis leading to erroneously increased HIC.
These factors may explain why we did not find a significant relationship between biochemical hepatic iron concentration in the NASH patients with either C282Y+/− or H63D+/− mutations compared with WT patients, although mean hepatic iron concentration was higher in C282Y heterozygotes and C282Y/H63D compound heterozygotes compared with WT individuals. Additionally, our sample size may have lacked sufficient power to distinguish relatively small differences in hepatic iron content. Our study was powered to detect differences in HFE mutation frequencies between patients with and without advanced fibrosis, but not to detect differences in HIC between groups.
Other possible limitations of our study deserve mention. These include the largely retrospective study design and lack of complete data on all subjects. We did not have a central pathology review by a single pathologist because of the retrospective nature of the study and the lack of availability of biopsy slides from all subjects. We also did not collect information on grade of inflammation, ballooning, or steatosis or distribution of hepatic iron staining, because the primary a priori end point in this study was fibrosis in relation to HFE mutation status.
In conclusion, there does not appear to be an increased prevalence of HFE mutations among Caucasian patients with NASH. However, presence of C282Y+/− mutations was associated with advanced hepatic fibrosis among Caucasian patients with NASH. We speculate that the mechanism is related to increased oxidative stress in the liver due to increased iron deposition. Additional studies are warranted to further explore the role of iron and HFE mutations in the pathogenesis of NASH.
We would like to acknowledge the following pathologists for their contributions: Rodger Haggitt (deceased), Shari Taylor, Mary Bronner, Matthew Yeh (University of Washington Medical Center); Lawrence Burgart (Mayo Clinic College of Medicine); Ian Wanless (University of Toronto); Oscar Cummings (Indiana University); and Abida Haque (University of Texas).