|Home | About | Journals | Submit | Contact Us | Français|
There are limited data comparing hepatic phenotype among hemochromatosis patients with different HFE genotypes. The goal of this study was to compare hepatic histopathologic features and hepatic iron concentration (HIC) among patients with phenotypic hemochromatosis and different HFE genotypes.
We studied 182 U.S. patients with phenotypic hemochromatosis. Degree of hepatic fibrosis, pattern of iron deposition, presence of steatosis or necroinflammation, and HIC were compared among different HFE genotypes.
C282Y/H63D compound heterozygotes and patients with HFE genotypes other than C282Y/C282Y were more likely to have stainable Kupffer cell iron (31.1% vs 9.5%; p=0.02), portal or lobular inflammation (28.9% vs 15.6%; p=0.03) and steatosis (33.3% vs 10.2%; p<0.01) on liver biopsy than C282Y homozygotes. Mean log10 HIC (p<0.05) and log10 ferritin (p<0.05) were higher among C282Y homozygotes than in patients with other HFE genotypes. In a logistic regression analysis using age, gender, HFE genotype, log10 ferritin, and log10 HIC as independent variables, log10 SF (p = 0.0008), male sex (p = 0.0086), and log10 HIC (p=0.047), but not HFE genotype (p = 0.0554) were independently associated with presence or absence of advanced hepatic fibrosis.
C282Y/H63D compound heterozygotes and other non-C282Y homozygotes who express the hepatic hemochromatosis phenotype frequently have evidence of steatosis or chronic hepatitis and lower body iron stores than C282Y homozygotes. These data suggest that presence of concomitant liver disease may explain expression of the hemochromatosis phenotype among non-C282Y homozygotes. Increased age, HIC and ferritin are associated with advanced hepatic fibrosis, regardless of HFE genotype.
Hereditary hemochromatosis is characterized by increased dietary iron absorption that leads to excessive iron accumulation in the liver and other tissues in some patients, ultimately resulting in organ dysfunction. In 1996, homozygosity for a mutation in the HFE gene located on chromosome 6 was found in the majority of patients with the phenotype of hereditary hemochromatosis (1). Two HFE mutations were initially identified, a substitution of cysteine for tyrosine at amino acid 282 (the Cys282Tyr or C282Y) and the His63Asp or H63D mutation at amino acid 63 (2,3). A third mutation was subsequently identified at amino acid 65 and results from the substitution of serine for cysteine (S65C); this mutation is less common than C282Y or H63D, and is infrequently associated with iron overload (4).
Subsequent studies showed that among Caucasian patients with hemochromatosis phenotypes, 80–90% are C282Y homozygotes and 4–7% are C282Y/H63D compound heterozygotes. A smaller proportion (3–10%) have other HFE genotypes (i.e., C282Y/wild type, H63D/wild type, H63D/H63D or wild type/wild type) (5). By contrast, population-based screening studies have found very low rates of phenotypic expression among C282Y/H63D compound heterozygotes (6). In contrast, among patients with chronic hepatitis C, alcoholic liver disease, or nonalcoholic fatty liver disease, heterozygosity for C282Y or H63D, or C282Y/H63D compound heterozygosity was associated with increased iron stores and more advanced liver disease than was observed in HFE wild type patients, suggesting that mild to moderate iron loading due to heterozygosity for common HFE mutations may exacerbate other liver diseases. Furthermore, iron loading in the liver occurs in some patients with steatohepatitis and viral hepatitis (7,8).
It is therefore possible that the presence of other undiagnosed liver diseases may lead to hepatic iron overload among patients who are heterozygous or wild type for HFE mutations. In support of this hypothesis, Brunt et al. showed in that the histological pattern of iron deposition varied among patients with different HFE genotypes; furthermore, only 42 of 72 (58%) subjects with a hemochromatosis pattern of hepatic iron overload were homozygous for C282Y (9). This group also found that HIC was higher among C282Y homozygotes than in C282Y/H63D compound heterozygotes (9,10).
The goals of the current study were the following: 1) to compare differences in the pattern of iron deposition across different HFE genotypes among a large multicenter cohort of patients with the hepatic hemochromatosis phenotype; and 2) to examine whether C282Y/H63D compound heterozygotes and those with other HFE genotypes are more likely to have steatosis or other features of chronic liver injury than C282Y homozygotes. A secondary aim was to examine whether there are differences in hepatic iron concentration (HIC) and other variables associated with advanced hepatic fibrosis among patients with different HFE genotypes.
Histopathology records were reviewed from 182 patients with phenotypic HH collected from six tertiary centers in the United States. The centers (University of Washington Medical Center, Seattle, Washington; Mayo Clinic, Rochester, Minnesota; William Beaumont Hospital, Royal Oak, Michigan; Southern Iron Disorders Center, Birmingham, Alabama; Rochester General Hospital, Rochester, New York; and University of Vermont Medical Center, Burlington, Vermont) represented different geographical regions in the United States: the Northeast, Midwest, Southeast, and Northwest. We have previously described the relationship between serum ferritin level and advanced fibrosis in this cohort (11).
The following clinical and laboratory data were available at the time of diagnosis on all patients who were considered for the study: age, sex, pretreatment serum ferritin levels, and HFE mutation status. In addition, all patients had undergone liver biopsy and the pathology reports were available for review. All patients had increased stainable iron (≥ 3+) by using Perls’ Prussian blue stain with a periportal to pericentral gradient.
Inclusion criteria were 1) hepatic phenotype associated with hemochromatosis (on the basis of hepatic iron concentration of ≥ 4,000 µg/g dry weight or hepatic iron index of ≥ 1.9, calculated as hepatic iron concentration [µmol/g] divided by age [years]), or >4 g of iron removed by quantitative phlebotomy; 2) no other known causes of chronic liver disease; 3) absence of secondary cause for iron overload; 4) available pretreatment serum ferritin value; 5) alcohol intake less than 20 g/d; and 6) no risk factors for or histologic evidence of nonalcoholic steatohepatitis (body mass index < 33 kg/m2, serum triglyceride level < 5.65 mmol/L [<500 mg/dL], and no type 2 diabetes mellitus). Histologic data were available for all 182 patients included in this report; hepatic iron concentration (HIC) and serum ferritin measurements were available in 172 (11).
The institutional review board at each participating institution approved the study. Two investigators who were blinded to the laboratory and clinical data reviewed the reports of pathologists' interpretations of liver biopsy specimens. The following information was recorded: degree of stainable iron (0 to 4+); presence or absence of cirrhosis (bridging fibrosis was classified as cirrhosis) on the basis of the description by the attending pathologist; and evidence of other liver diseases, such as fatty change, steatohepatitis, or chronic hepatitis. Hepatic iron concentration was measured by atomic absorption spectrophotometry or colorimetry.
Statistical analyses were performed using Stata program (Stata Corp., College Station, TX), Excel 2000® (Microsoft Corp., Redmond, WA), and GB-Stat® (v. 10.0, 2003, Dynamic Microsystems, Inc., Silver Spring, MD). Initial evaluations revealed that serum ferritin and HIC data in the 172 subjects were not normally distributed. Therefore, these data were normalized for analysis using log10 transformation; some results were reported as antilog values. Descriptive data are displayed as enumerations, percentages, or mean ± 1 S.D. (or 95% confidence intervals, as appropriate). Frequency values were compared using chi-square analysis or Fisher exact test, as indicated. Mean values were compared using student t-test; a Bonferroni correction was applied for multiple comparisons. Multiple regression analysis was performed using age, sex (male or female), HFE genotype category (C282Y/C282Y, C282Y/H63D, or other), log10 SF, and presence of cirrhosis as independent variables, and log10 hepatic iron concentration (log10 HIC) as the dependent variable. In a logistic regression, we used age, sex (male or female), HFE genotype category (C282Y/C282Y, C282Y/H63D, or other), log10 SF, and log10 HIC as independent variables, and presence (or absence) of cirrhosis as the dependent variable. Values of p <0.05 were defined as significant.
Characteristics of the 182 subjects are displayed in Table 1. There were 147 C282Y homozygotes (82.5%), 15 C282Y/H63D compound heterozygotes (7.6%), and 20 patients (10.5%) with other HFE genotypes (C282Y/wt, H63D/wt, H63D/H63D, or wt/wt). Approximately three-quarters of the subjects were men (Table 1).
In univariate analyses, mean log10 ferritin levels were greater in C282Y homozygotes than in C282Y/H63D compound heterozygotes (p = 0.0097) or in subjects with other HFE genotypes (p = 0.0242). In multiple comparisons across the three HFE genotype groups, the mean log10 ferritin was highest in C282Y homozygotes after Bonferroni correction (p <0.05) (Table 1).
In univariate analyses, the mean log10 HIC level in C282Y homozygotes did not differ significantly from that of C282Y/H63D compound heterozygotes (p = 0.1707). The mean log10 HIC level in C282Y homozygotes was greater than that of subjects with other HFE genotypes (p = 0.0167) (Table 1). The mean log10 HIC level in C282Y/H63D compound heterozygotes did not differ significantly from that of subjects with other HFE genotypes (p=0.7169). In multiple comparisons across the three HFE genotype groups, the mean log10 HIC was significantly higher in C282Y homozygotes than that of persons with other HFE genotypes after Bonferroni correction (p <0.05). As we have previously reported, cirrhosis was detected in 22% of C282Y homozygotes, but this proportion did not differ significantly from that observed in subjects with other HFE genotypes (Table 1).
C282Y/H63D compound heterozygotes were more likely to have to have Kupffer cell iron staining (26.7 vs 9.5%; p=0.04), portal or lobular inflammation (40% vs 15.6%; p=0.03), and steatosis (46.7% vs 10.2% p<0.01) than C282Y homozygotes. Patients with other HFE genotypes also had a higher prevalence of Kupffer cell iron staining (50% vs 9.5%; p<0.01), inflammation (35% vs 15.6%; p=0.03), and steatosis (40% vs 10.2%; p<0.01) than C282Y homozygotes. There were no significant differences between C282Y/H63D compound heterozygotes and patients with other HFE genotypes in prevalences of Kupffer cell staining (26.7% vs 50%, respectively; p=0.16), inflammation (40% vs 35%, respectively; p=0.76), or steatosis (46.7% vs 40%, respectively; p=0.69) (Figure 1).
In univariate analyses, Kupffer cell iron staining, portal or lobular inflammation, and steatosis were significantly more prevalent in C282Y/H63D compound heterozygotes or in subjects with other HFE genotypes than in C282Y homozygotes (Table 1). In 3 × 2 Chi-square tables, greater proportions of C282Y/H63D compound heterozygotes or subjects with other HFE genotypes had Kupffer cell iron staining, portal or lobular inflammation, and steatosis than C282Y homozygotes (p <0.0001, p = 0.0152, and p <0.0001, respectively).
Among subjects with available HIC, mean HIC was significantly higher among C282Y homozygotes than C282Y/H63D compound heterozygotes (13,563 ± 10,400 µg/g and 7,781 ± 4,954 µg/g, respectively; p = 0.002). Mean HIC in C282Y homozygotes was significantly higher than in C282Y/H63D compound heterozygotes both among patients with cirrhosis (23,719 ± 14,991 µg/g vs 18,300 ± 98 mcg/g, , respectively; p=0.05) and those without cirrhosis (10,700 ± 6,282 µg/g vs 5,869 ± 1,817 µg/g, respectively; p<0.01) (Figure 2).
Multiple regression analysis was performed using age, sex (male or female), HFE genotype category (C282Y/C282Y, C282Y/H63D, or other), log10 SF, and presence of cirrhosis as independent variables, and log10 HIC as the dependent variable. The significant independent variables were presence of cirrhosis (p = 0.0001), log10 serum ferritin (p = 0.0002), and HFE genotype (favoring C282Y homozygosity) (p = 0.0045). In a logistic regression, we used age, sex (male or female), HFE genotype category (C282Y/C282Y, C282Y/H63D, or other), log10 SF, and log10 HIC as independent variables, and presence (or absence) of cirrhosis as the dependent variable. The significant independent variables were log10 SF (p = 0.0008), male sex (p = 0.0086), and log10 HIC (0.0470). HFE genotype was not a significant independent variable (p = 0.0554).
In this multicenter U.S. study, the hepatic histopathologic features among C282Y homozygotes with phenotypically defined hemochromatosis were distinct from those of C282Y/H63D compound heterozygotes, and patients with other HFE genotypes. C282Y homozygotes were less likely to have Kupffer cell iron staining, portal or lobular inflammation, or steatosis than patients with other HFE mutation patterns. These data suggest that the homozygous C282Y genotype alone is sufficient for expression of the hepatic hemochromatosis phenotype, possibly as a consequence of significantly greater body iron stores than are observed in persons with other common HFE genotypes. In contrast, expression of the hepatic hemochromatosis phenotype among non-C282Y homozygotes may require a secondary process (a “second hit”) to promote hepatic iron accumulation such as cryptogenic hepatitis or fatty liver disease. An alternative explanation is that deleterious mutations in other iron-related genes such ferroportin, transferrin-receptor 2, hepcidin, or hemojuvelin may contribute to iron accumulation in non-C282Y homozygotes and may result in chronic liver injury. However, we believe the latter explanation to be less likely, given the lack of previous data showing chronic hepatitis or steatosis in patients with hepatic iron overload due to mutations in other iron-related genes and the rarity of these mutations.
Several recent studies have attempted to define the rate of phenotypic expression among patients with HFE mutations in different populations, ranging from otherwise healthy subjects identified by screening to patients with liver disease and iron overload. A previous study estimated that the odds-ratio of iron overload was >4000 for White C282Y homozygotes, 32 for C282Y/H63D compound heterozygotes, 4.1 for C282Y/wild type heterozygotes, and 1.6 for H63D/wild type heterozygotes (12). Furthermore, in screening studies, only a minority of C282Y/H63D compound heterozygotes have evidence of iron overload based on serum ferritin levels (13). These data suggest that expression of the hepatic phenotype of hemochromatosis among non-C282Y homozygotes could be explained by a “second hit” leading to hepatic iron accumulation such as chronic hepatitis or fatty liver disease that leads to necroinflammation in the liver.
The findings of the current study confirm and extend previous findings reported by Bacon and colleagues in a smaller cohort from a single U.S. center; these authors also found that C282Y homozygotes had significantly higher HIC compared to patients with other HFE genotypes (14). A large study from Australia examined HIC among C282Y homozygotes and patients with other genotypes. Patients were classified based on HFE genotype and hepatic histological features were recorded. Compared to C282Y homozygotes, C282Y/H63D compound heterozygotes were significantly more likely to have hepatic steatosis, lobular inflammation, and steatosis. These authors found that HIC was significantly lower among C282Y/H63D compound heterozygotes than C282Y homozygotes (15). In contrast to our study, HIC was also lower among C282Y/wild type and H63D/wild type heterozygotes and HFE wild type patients than C282Y homozygotes; however, this difference between the study by Walsh and the current report is likely explained by the larger sample size in that study. Similar to Walsh and colleagues, we found that HIC was significantly higher among C282Y homozygotes than C282Y/H63D compound heterozygotes in patients with or without cirrhosis.
We and others have previously demonstrated that serum ferritin level>1,000 ng/ml is associated with the presence of advanced hepatic fibrosis among C282Y homozygotes with phenotypic hemochromatosis (16). In the current study increased age, serum ferritin and HIC were independently associated with advanced hepatic fibrosis. This was observed in both C282Y homozygotes and patients with other HFE genotypes, highlighting the central role of hepatic iron deposition in promoting hepatic fibrogenesis among patients with hemochromatosis as previously suggested by Adams (17).
Limitations of the current study deserve mention. There was a relatively small number of patients in this study who were C82Y/wild type or H63D/wild type or wild-type for HFE. For this reason, we did not analyze these patients separately from C282Y homozygotes and C282Y/H63D compound heterozygotes, and interpret any differences between these subsets and the other HFE genotype groups with caution. We excluded chronic viral hepatitis using serologic assays for hepatitis B and C and alcohol-related liver disease based on clinical history; in addition, we attempted to exclude patients with obvious risk factors for non-alcoholic steatohepatitis. Nonetheless, it is possible that our data collection may have not identified patients with other causes of chronic hepatitis and may have included patients who under-reported their alcohol consumption or individuals with nonalcoholic fatty liver disease. Because our study was retrospective, we were not able to retrieve liver histopathologic specimens for a centralized pathology review, and we did not have stored DNA samples for HFE S65C, transferrin-receptor 2, hepcidin or hemojuvelin mutation analysis. Regardless, the present study provides clinical confirmation of the hypothesis that a co-factor is needed to lead to phenotypic expression of hemochromatosis among non-C282Y homozygotes. Future studies will need to further examine whether these co-factors are genetic or environmental.
In conclusion, we have shown that in a large, multicenter U.S. cohort of patients with the hepatic phenotype of hemochromatosis, HIC is significantly higher among C282Y homozygotes than in C282Y/H63D compound heterozygotes, although the threshold level associated with cirrhosis is similar in the two groups. In addition, patients who did not have the C282Y homozygous genotype were more likely to have Kupffer cell iron, hepatic necroinflammation, and steatosis. These data suggests that expression of the hepatic hemochromatosis genotype among non-C282Y homozygotes is associated with a presence of a “second hit” that contributes to liver disease leading to hepatic iron overload. These findings extend understanding of the interaction between genetic and environmental factors in phenotypic expression of hemochromatosis among patients without C282Y homozygosity, and highlight the central role of hepatic iron accumulation in the development of hepatic fibrosis in primary iron overload disorders.
Supported in part by NIH Grant DK-02957