Haemochromatosis

Authors


Correspondence to: Dr K. V. Kowdley, Associate Professor of Medicine, University of Washington, Box 356174, Seattle, WA 98195, USA. E-mail: kkowdley@u.washington.edu

Summary

Hereditary haemochromatosis is the prototype disease for primary iron overload. The disorder is very common, especially amongst subjects of Northern European extraction. It is characterized by an autosomal recessive mode of inheritance, and most cases are homozygous for the C282Y mutation in the HFE gene. Haemochromatosis is now recognized to be a complex genetic disease with probable significant environmental and genetic modifying factors. The early diagnosis of individuals at risk for the development of haemochromatosis is important, because survival and morbidity are improved if phlebotomy therapy is instituted before the development of cirrhosis. The cost-effectiveness and utility of large-scale screening for haemochromatosis have been questioned given that many individuals with the homozygous C282Y mutation do not have iron overload or end-organ damage. However, the use of phenotypic tests, such as serum transferrin-iron saturation, for initial screening avoids the problem of the identification of non-expressing homozygotes. Liver biopsy remains important in management to determine the presence or absence of cirrhosis, particularly amongst patients with serum ferritin levels greater than 1000 ng/mL or elevated liver enzymes. Those with non-HFE haemochromatosis who cannot be identified on genotypic testing should have a liver biopsy to establish diagnosis. Patients with end-stage liver disease may develop liver failure or primary liver cancer, and liver transplantation may be required. Liver transplantation for haemochromatosis is associated with a poorer outcome compared with other indications because of infections and cardiac complications.

Introduction

Iron is indispensable for basic cellular functions. However, this metal is also a catalyst for chemical reactions associated with the production of reactive oxygen species, which may lead to oxidative stress and cellular damage. Thus, iron levels within cells must be precisely regulated to promote essential functions and provide an appropriate abundance to maintain adequate stores and yet minimize the risk of potential toxicity. Body iron homeostasis is regulated primarily by duodenal and upper small intestinal absorption and is responsive to body iron stores. Hence, iron absorption is increased during iron deficiency and down-regulated when iron stores are replete.1 There is no effective method for the elimination of excess body iron, and iron overload from iatrogenic or idiopathic pathological causes can lead to multiple systemic complications.

Hereditary haemochromatosis, the prototype disorder of iron overload due to misregulated iron homeostasis in humans, is caused by an inappropriate increase in iron absorption in the duodenum and upper small intestine. Most cases of hereditary haemochromatosis amongst subjects of Northern European descent are associated with a homozygous mutation in HFE, the so-called haemochromatosis gene. Loss of the functional HFE protein product is associated with inappropriately increased iron absorption, resulting in the deposition of iron in parenchymal organs, notably the liver, pancreas, heart, joints, skin and pituitary gland, and consequent end-organ damage. HFE protein is similar to major histocompatibility complex class 1-type proteins, a group of proteins that bind to β2-microglobulin.

Iron overload in hereditary haemochromatosis may result in severe systemic disease, such as cirrhosis, restrictive cardiomyopathy, diabetes mellitus, arthropathy, skin hyperpigmentation and gonadal failure. End-stage liver disease and portal hypertension may necessitate liver transplantation. Hepatocellular carcinoma or primary liver cancer may develop once cirrhosis is established. The risk of primary liver cancer is increased up to 200-fold in males 55 years or older with cirrhosis due to hereditary haemochromatosis.2

Hereditary haemochromatosis associated with a homozygous mutation of the HFE gene is the most common cause of primary iron overload; other causes include juvenile haemochromatosis and other rare disorders (Table 1). Secondary iron overload is used to describe iron overload resulting from additional mechanisms other than continuous inappropriately increased gastrointestinal iron absorption with subsequent excessive iron deposition in body tissues. The most common causes of secondary iron overload include disorders of ineffective erythropoiesis, such as sideroblastic anaemia and thalassaemia major. Iron overload in some of these disorders may be a result of a combination of excess iron absorption and red cell transfusions as well as iatrogenic iron therapy. Excess alcohol consumption and/or chronic hepatitis C may also cause secondary iron overload, particularly in the setting of end-stage liver disease.3, 4 Parenteral iron overload is the result of excess iron from transfusions and/or parenteral iron administration. African iron overload may result from a combination of genetic and environmental factors. Secondary iron overload and parenteral iron overload need to be excluded when diagnosing hereditary haemochromatosis.

Table 1.  Haemochromatosis variants
TypeCommentsLocation
  1. wt, wild type.

1. HFE Chromosome 6
 C282Y/C282Y> 80% cases world-wide 
 C282Y/H63D11% present phenotypically 
 H63D/H63DOften has a potentiator, e.g. hepatitis C infection, β-thalassaemia trait 
 C282Y/S65CMild 
 H63D/S65CMild 
 S65C/S65CMild 
 C282Y/wtVery rare 
 H63D/wtVery rare 
 S65C/wtVery rare 
 wt/wtVery rare 
2. HFE2 (juvenile haemochromatosis)Presents before age 30 yearsChromosome 1
Very severe 
Equal sex ratio 
3. HFE3Mutation in TfR2 geneChromosome 7
Tyr250Stop mutation 
Other mutations identified 
4. HFE4Autosomal dominantChromosome 2
Mutation in ferroportin 1 gene (SLC11A3) 
5. Other Unknown
 African iron overload
 Iron overload in Solomon Islanders

Genetics

Hereditary haemochromatosis is characterized by an autosomal recessive pattern of inheritance. It is the most common Mendelian inherited disorder of Northern Europeans, with a prevalence of 1 : 200 to 1 : 500.5, 6 An even higher prevalence of 1 : 100 is likely in Ireland.7–10 The identification of the HFE gene in 1996 was a major breakthrough in the understanding of hereditary haemochromatosis. The HFE gene encodes for a novel 343-amino acid major histocompatibility complex class 1 molecule. Initially, two missense mutations were identified in the HFE gene, which is located on the short arm of chromosome 6.11 A single mutation of G to A at nucleotide 845 results in the substitution of tyrosine for cysteine at amino acid 282. This is known as the Cys282Tyr or C282Y mutation. A second mutation of C to G at nucleotide 187 results in a substitution of aspartate for histidine at amino acid 63. This His63Asp or H63D mutation does not, except in very rare instances, occur on the same allele as C282Y.12, 13 A third mutation resulting in a serine to cysteine substitution at amino acid 65 (Ser65Cys or S65C) has recently been suggested to be associated with a mild form of haemochromatosis.14 There are at least 38 other allelic variants of the HFE gene.15–17 Most do not appear to be clinically significant at this time.

Amongst subjects of Northern European descent, 80–100% of those with clinical features of hereditary haemochromatosis are homozygous for the C282Y mutation.5, 7, 8 Amongst subjects of European (including Northern and Southern Europe) descent, homozygosity of C282Y is associated with hereditary haemochromatosis in 50–100% of patients, with lower rates amongst Mediterranean and Southern European populations.15, 18 Only a minority of compound heterozygotes (C282Y/H63D) develop clinical symptoms of haemochromatosis (11% of cases).5, 19, 20 The homozygous H63D mutation is not as penetrant for phenotypic hereditary haemochromatosis as the C282Y mutation, but there are rare reported cases of haemochromatosis with this genotype.21, 22 Goochee et al., studying a population of Northern European extraction, found that homozygosity of H63D did not result in clinically significant iron overload.23 However, the H63D mutation may lead to iron overload in the setting of another risk factor, such as β-thalassaemia trait24 or hepatitis C.25 Whilst the C282Y mutation is largely confined to subjects of European extraction, the H63D mutation is much more widespread.26 In Saguenay-Lac-Saint-Jean, a geographically isolated region 200 km north-east of Quebec City in Canada, an unusual distribution of hereditary haemochromatosis genotypes in phenotypically affected individuals has been reported. The genotypes include H63D homozygotes, C282Y, H63D and S65C heterozygotes and a number of patients with no HFE mutations. The researchers presumed a founder effect or the presence of a large number of families with the as yet unidentified gene for juvenile haemochromatosis.27

Hereditary haemochromatosis not associated with the characteristic mutations in the HFE gene is much less common than HFE-related haemochromatosis and is termed non-HFE-related haemochromatosis. As non-HFE mutations are identified, a numbering system is being produced.28 Juvenile haemochromatosis (HFE2) is a rare disorder reported in different ethnic groups. It is the most common of the non-HFE-related haemochromatoses and occurs equally in both sexes. HFE2 is more severe than HFE-related haemochromatosis and clinical symptoms present before 30 years of age because of cardiac disease and hypogonadism. The HFE2 locus maps to chromosome 1q, but the gene remains to be identified.29

The molecular basis of another form of non-HFE-related haemochromatosis (HFE3) has been recently identified. The HFE3 locus maps to chromosome 7q22 and the gene, transferrin receptor 2 (TfR2), encodes a possible transmembrane protein with moderate homology to transferrin receptor.28, 30 Multiple mutations have been identified.31 It is now possible to test for the Tyr250Stop mutation in the TfR2 gene mutation by collecting DNA from buccal swabs.32 A fourth form of non-HFE-related haemochromatosis (HFE4), which is autosomal dominant, has been mapped to chromosome 2. It is associated with a mutation in the ferroportin 1 (SLC11A3) gene.28, 33, 34 Other forms of haemochromatosis, such as African iron overload disorder, are known to have a genetic basis but are yet to be more clearly defined.35, 36

Pathophysiology

The identification of the HFE gene has greatly increased the understanding of the mechanism of normal iron metabolism and the putative defect in hereditary haemochromatosis. The most important pathophysiological step in body iron loading appears to be the inappropriately high intestinal iron absorption. Knowledge of the effects of mutations on iron absorption is central to the understanding of the pathological basis of hereditary haemochromatosis.37

Dietary iron is absorbed either as haem or ionic iron by the enterocytes in the duodenum. The ‘set point’ for iron absorption is determined in crypt cells, which migrate to the villi after a period of 2–3 days. This ‘set point’ is dependent on body iron requirements at the time enterocytes are formed and is deranged in patients with hereditary haemochromatosis, such that absorption remains inappropriately high despite increasing body stores. In normal subjects, HFE is highly expressed in the deep crypts of the small intestine.

Dietary haem is absorbed by an as yet unidentified transporter, and iron is subsequently released from haem within the enterocyte. Dietary free iron is first reduced from a ferric to a ferrous state by ferric reductase which is expressed on the luminal surface of the duodenum.38 The resultant ferrous iron is taken up by the apical transporter, divalent metal transporter 1.39 Iron may be stored within the mucosal absorptive cell as ferritin and lost when the senescent enterocyte is sloughed or transferred across the basement membrane to the plasma. Transfer to the plasma occurs via ferroportin 1.40 This process requires the oxidation of iron to the ferric state in which another protein, hepahaestin, plays a role.41 Patients with hereditary haemochromatosis demonstrate increased ferric reductase activity,42 increased divalent metal transporter 1 expression,43 decreased enterocyte ferritin stores44 and increased ferroportin 1 expression.38, 45 Up-regulation of ferric reductase, divalent metal transporter 1 and ferroportin 1, as displayed by DBA/2 HFE–/– mice,46 and decreased enterocyte ferritin stores, as predicted by stably transfected Chinese hamster ovary cells,47 could explain the significant increase in iron absorption seen in patients with hereditary haemochromatosis. The role of other newly described proteins involved in iron transport, such as TfR2 and hepcidin, is unknown.

HFE is a transmembrane protein with both intracellular and extracellular domains (Figure 1). HFE has a binding site for β2-microglobulin. Both HFE and β2-microglobulin are needed for normal iron regulation. This is shown by β2-microglobulin knockout –/– mice as well as HFE knockout –/– mice having a haemochromatosis phenotype.48 The C282Y mutation in the HFE gene results in the disruption of a critical disulphide bridge, with subsequent conformational changes to the protein.11 However, the C282Y mutation is not a null allele, as C282Y –/– mice have milder hepatic iron loading than HFE–/– mice.49 In cell culture, the mutant C282Y protein is trapped in the endoplasmic reticulum and middle Golgi compartments, does not undergo late Golgi processing, and is more quickly degraded.50 This reduced cell surface expression of the C282Y mutation is also present in patients with hereditary haemochromatosis.51

Figure 1.

Proposed structure of HFE. This transmembrane protein is similar to other major histocompatibility complex class 1-like proteins, and is localized on the basolateral aspect of the enterocyte; the locations of the C282Y and H63D mutations are shown. Reproduced with permission from the Journal of Hepatology 2000;32 (supplement 1), with permission from Elsevier Science.

HFE forms a complex with TfR, the receptor by which cells acquire iron-loaded transferrin. Uncomplexed HFE protein is rapidly degraded. HFE and transferrin compete for overlapping binding sites on TfR.28 Interaction between HFE and TfR is probably necessary in the normal control of iron absorption, as compound mutant mice hemizygous for TfR and HFE–/– accumulate more liver iron than HFE–/– mice alone.49 The HFE–TfR complex may exert effects on cellular iron homeostasis by influencing the level of transferrin-mediated cellular iron uptake in crypt cells. One proposed hypothesis for the mechanism of increased iron absorption in hereditary haemochromatosis is that the crypt cell in hereditary haemochromatosis ‘senses’ a state of iron deficiency because of decreased intracellular iron content as a result of impaired HFE–TfR-dependent iron uptake. Consequently, divalent metal transporter 1, the mucosal iron transporter, is up-regulated, resulting in the hyperabsorption of iron once these cells have migrated to the villi. However, the over-expression of HFE in HeLa cells leads to a decrease in transferrin-mediated iron uptake. This would suggest an inhibitory role for normal HFE in iron uptake.52–55 These somewhat conflicting data on normal and mutant HFE function have led to the speculation that the altered activity of other iron-related proteins, such as ferroportin 1, may contribute to the hyperabsorption of iron.56 It is likely that new research findings will lead to the further evolution of our understanding of iron metabolism in hereditary haemochromatosis.

Diagnosis of hereditary haemochromatosis

Phenotypic evaluation

Prior to the identification of the HFE gene, prevalence studies of hereditary haemochromatosis were based primarily on phenotypic markers of the disease, such as serum ferritin and transferrin-iron saturation. An over- or under-estimation of the prevalence of hereditary haemochromatosis may have occurred using these methods.

Serum ferritin measurement has several advantages: it is automated, relatively inexpensive and highly sensitive for iron overload. The disadvantages include the fact that serum ferritin values increase with age, are dependent on gender and may be elevated in many conditions other than iron overload due to hereditary haemochromatosis, such as cancer, inflammatory disorders, cytolysis and dysmetabolic hepatosiderosis (insulin-resistant-associated hepatic iron overload disorder). Elevated serum ferritin is also common in non-alcoholic steatohepatitis, which can be confused clinically with hereditary haemochromatosis, especially amongst those carrying the C282Y mutation.

Measurements of transferrin-iron saturation may vary greatly if performed using different techniques or equipment. The gold standard for the measurement of transferrin-iron saturation requires colorimetric measurement of iron and immunological measurement of transferrin.57, 58 Transferrin-iron saturation may then be calculated using the molecular weight of transferrin. Most epidemiological studies have used a calculated total iron binding capacity saturation as a surrogate marker of transferrin-iron saturation. Manual radial immunodiffusion methods used in this calculation are not as precise as automated nephelometric methods, and a significant bias may occur.58 This may lead to different ‘cut-off’ levels of transferrin-iron saturation when screening for hereditary haemochromatosis. A threshold value of 45% for transferrin-iron saturation to warrant additional testing for hereditary haemochromatosis has been proposed using the ‘gold standard’ based on good sensitivity and acceptable cost.59

Transferrin-iron saturation is an excellent phenotypic screening test for iron overload, and the reagent costs are acceptable. It is influenced by disease, age and hormones and may also increase due to the recent ingestion of iron or vitamins. It is also subject to a diurnal variation. Edwards et al. showed that this diurnal variation was lower amongst those with hereditary haemochromatosis.60 Although increased transferrin-iron saturation is a sensitive marker for hereditary haemochromatosis, it is not specific. Increased transferrin-iron saturation can also be found with excessive alcohol consumption, decreased transferrin synthesis due to liver dysfunction, hepatic cytolysis and other disorders of iron overload with or without blood transfusions. Transferrin is also a negative acute phase reactant and may be decreased in acute inflammation.

Genotypic (mutational) expression

It is important to emphasize that homozygosity for C282Y does not confer a diagnosis of haemochromatosis, but rather indicates a proclivity towards iron overload. Non-expression is common, particularly in women.19 However, women can have full phenotypic expression of the disease, including cirrhosis.61 Factors that influence expression include age, sex, diet, alcohol intake, hepatitis C infection, blood donation and abnormally decreased absorption or loss of iron. Ingestion of excessive alcohol has a major modifying influenceon the expression of hereditary haemochromatosis. Fletcher et al. found that subjects with hereditary haemochromatosis who drink more than 60 g of alcohol per day are approximately nine times more likely to develop cirrhosis than those who drink less than 60 g of alcohol per day.62

Phenotypic expression of hereditary haemochromatosis appears to be different in different ethnic populations. A recent study in a Sicilian population found a correlation between carriage of the C282Y allele and a significant increase in longevity in women.63 Conversely, a decreased life expectancy was found in female C282Y heterozygotes in a Danish study.64 Whilst appearing to be contradictory, these findings mirror previous work on the haemochromatosis-associated disorder porphyria cutanea tarda. Roberts et al. found a positive association of C282Y with porphyria cutanea tarda in patients in the UK.65 This work was subsequently confirmed in Australian patients.66 By contrast, porphyria cutanea tarda was associated with the H63D mutation in Italian patients, and the frequency of the C282Y mutation was not significantly different from controls.67 These findings point to a possible difference in penetrance or expression of the genes responsible for hereditary haemochromatosis between Northern and Southern Europeans, as well as differing pathogenic mechanisms for porphyria cutanea tarda in these different populations.

A large Australian study in subjects of Northern European ancestry has shown that, in addition to HFE genotype, there may be highly significant effects of as yet unidentified genes on iron stores.68 The possible presence of a modifying gene(s) or another mutation affecting hereditary haemochromatosis patients, located in the D6S105 region of the ancestral haplotype, has been demonstrated.69 In addition, a polymorphism in haptoglobin has been shown to be over-represented in C282Y homozygous hereditary haemochromatosis patients.70

In a recent study of volunteers at a health appraisal clinic in San Diego, Beutler et al. suggested that the rate of clinical manifestations or penetrance of haemochromatosis may be less than 1%.71 A prevalence of 1 in 200 for C282Y homozygosity in San Diego would give a clinical presentation for hereditary haemochromatosis of approximately 1 in 20 000 persons if these estimates were correct. These figures are similar to the data reported in 1955, when Finch and Finch estimated the frequency of hereditary haemochromatosis in the USA as 1 in 20 000 based on hospital admissions which noted the classical features of cirrhosis, diabetes, pigmentation and arthralgia.72 Subsequent autopsy studies found this figure to be a gross under-estimation, and the frequency was revised to 1–2 per 1000.73, 74 A prevalence of 1 in 500 or 1 in 1000 for organ damage due to haemochromatosis would be consistent with the 50% penetrance found by Olynyk et al.75 There were no liver biopsies performed in the study of Beutler et al. and the study did not allow for selection bias. Furthermore, a number of previously diagnosed haemochromatotics were excluded from the analysis.

A major confounding factor in the study by Beutler et al. could be the population investigated, as San Diego has nearly a 500-year or 20-generation association with Spain. There are probably different modifying genes in Northern vs. Southern European populations, as discussed above. In a preliminary study, Beutler et al. found that the frequency of C282Y was 12% and H63D was 30%.76 Similarly high H63D frequencies have been reported by Merryweather-Clarke et al. in Basques6 and by Sanchez et al. in Spanish haemochromatosis patients.77 The only Northern European population to have a similarly high carrier frequency of H63D of 32% has been reported in Norwich, UK.78 In a study of health maintenance organization employees in the USA, McDonnell et al. in Missouri found an H63D allelic frequency of 0.16.79 This allelic frequency is closer to that usually noted in populations of Northern European descent. The allelic frequency in the study of Sanchez et al. was 0.22.77 McDonnell et al. also found a prevalence of 8 per 1000 of hereditary haemochromatosis genotypes consisting of C282Y homozygotes, compound heterozygotes, homozygous normals and one H63D homozygote who had evidence of hereditary spherocytosis. This study also found a prevalence of 4 per 1000 of iron overload in 1653 employees. Eighty-three per cent of the employees were women. These are very different findings from the studies of Beutler et al.71 The San Diego white population may therefore represent an ethnically diverse population with a greater admixture of Northern and Southern European genes.

In summary, the probable presence of multiple factors influencing the expression of hereditary haemochromatosis shows it to be a complex genetic disease with significant environmental and genetic modifiers.

Combined diagnostic approach

Screening strategies for hereditary haemochromatosis may utilize phenotypic tests alone, may begin with phenotypic tests followed by genotypic confirmatory tests, or may begin with genotypic tests followed by phenotypic tests.80, 81 The great variability in the expression and prevalence of hereditary haemochromatosis, as well as the availability of genetic testing, will command different strategies for different populations. The identification of a genotype associated with hereditary haemochromatosis is not in itself sufficient for the diagnosis of hereditary haemochromatosis. Rather, the presence of the genetic abnormality identifies individuals with an increased susceptibility to development of the phenotype.81 Testing for the C282Y mutation has also been shown to be a cost-effective method of screening relatives of patients with hereditary haemochromatosis.82 Diagnostic screening should initially target high-risk groups, such as those with suspicious organ involvement, a family history of hereditary haemochromatosis and those with a biochemical or radiological abnormality suggestive of iron overload.83 There are societal concerns about discrimination by insurers and employers if widespread genetic screening for hereditary haemochromatosis is implemented in healthy populations. Until these are resolved, it is unlikely to be adopted as a public policy.

In the near term, some of these controversies can be resolved by focusing effort on the education of physicians and the public, and by targeted screening for hereditary haemochromatosis in high-risk populations. Selection of the appropriate test for each population being studied is important to improve the diagnostic yield. Whilst C282Y genotyping may be the preferred screening strategy for hereditary haemochromatosis in Ireland,10 it may not be appropriate in Greece where the C282Y allele has been found in only 50% of haemochromatosis patients.18 The optimal strategy for screening in the USA, given the significant variation in ethnicity, is nowhere near as clear. The American Association for the Study of Liver Diseases has endorsed a low-cost phenotypic approach for screening of the general population in its Practice Guidelines.83 It recommends the use of a measure of transferrin-iron saturation which is expressed as the ratio of serum iron to a calculated total iron binding capacity consisting of iron plus unsaturated iron binding capacity. The advantage of this less expensive method is that it allows the clinician to judge the significance of a raised transferrin-iron saturation by noting those that might be spuriously elevated by a low total iron binding capacity. Decreased cost makes this recommendation particularly attractive for large-scale population screening. A suggested algorithm adapted from the above guidelines for the management of hereditary haemochromatosis is shown in Figure 2.

Figure 2.

Suggested algorithm for the screening and management of iron overload: 2002. Abnl, abnormal; HCC, hepatocellular cancer; LFT, liver function test.

Liver biopsy

Prior to the availability of HFE gene testing, liver biopsy, with the assessment of hepatic iron by staining and the biochemical measurement of the hepatic iron concentration, was generally considered to be the gold standard for the diagnosis of haemochromatosis. The most important prognostic factor at the time of diagnosis of hereditary haemochromatosis is the presence or absence of hepatic fibrosis or cirrhosis. Those without hepatic fibrosis may be expected to have a normal life expectancy with phlebotomy therapy.84

The liver biopsies of patients with hereditary haemochromatosis characteristically show grade 2–4+ hepatocellular iron stores on staining, with the greatest density of iron staining in periportal areas.85 Kupffer cell iron staining is usually not present unless total iron storage is severe, with grade 4+ iron stores and fibrosis. Liver biopsies of patients with secondary iron overload do not show the characteristic periportal deposition in hepatocytes and may demonstrate increased iron in reticulo-endothelial cells. HFE4 patients may demonstrate increased iron staining in both hepatocytes and reticulo-endothelial cells.33, 34

The current recommendations in Australia for the role of liver biopsy in HFE-related haemochromatosis, where most hereditary haemochromatosis patients are of Anglo-Celtic descent, are to reserve liver biopsy for patients who are homozygous or compound heterozygous for the C282Y mutation and have at least one of the following criteria:86 serum ferritin > 1000 µg/L; abnormal serum aminotransferase levels; hepatomegaly; or age greater than 45 years.

The American Association for the Study of Liver Diseases Practice Guidelines recommend the consideration of liver biopsy in all homozygotes with clinical evidence of liver disease, serum ferritin > 1000 ng/mL and, particularly, in those over 40 years of age with other risk factors for liver disease. They also recommend that liver biopsy should be considered in compound (C282Y/H63D) or simple C282Y heterozygotes with elevated transferrin-iron saturation, particularly in those with abnormal liver enzyme levels or clinical evidence of liver disease.83 The use of either of these criteria will probably detect the vast majority of patients at risk for cirrhosis who require more intensive follow-up, and will avoid liver biopsy in younger patients in whom biopsy would not alter management. Compound heterozygotes (C282Y/H63D) with iron overload often have other contributing factors, such as excessive alcohol consumption, non-alcoholic steatohepatitis, chronic hepatitis C or porphyria cutanea tarda.87 Rarely, C282Y heterozygotes may develop hereditary haemochromatosis in the presence of other mutations, such as IVS3 + IG?T.88 If the patient presents with a phenotypic picture of hereditary haemochromatosis and is wild type for C282Y, other causes of iron overload and non-HFE forms of hereditary haemochromatosis should be considered. Liver biopsy is essential in these patients for diagnosis and management.

When performing a liver biopsy, a biopsy core of at least 2.5–3.0 cm in total length should be obtained. Of this, a 0.5–1.0 cm piece of the core tissue should be removed and placed in a dry tube or in 10% formalin (not in saline which may leach out iron). The remainder of the fixed tissue can be processed for routine histopathological evaluation and a Perls' stain for iron. If tissue is not separated and saved before fixation and embedding, it can be removed from the paraffin block and used for quantitative iron measurement.83 A liver biopsy performed in this manner will permit the confirmation of iron overload, the determination of the pattern and degree of hepatic iron overload, the calculation of the hepatic iron concentration and hepatic iron index and the detection of associated lesions, such as steatosis or hepatitis.89 The hepatic iron index is obtained by dividing the hepatic iron concentration (in µmol/g) by the age (in years). A hepatic iron index of 1.9 µmol/g/years is highly suggestive of hereditary haemochromatosis. In the diverse US population, a ‘threshold’ hepatic iron concentration of > 71 µmol/g dry weight can almost always distinguish patients with phenotypic hereditary haemochromatosis from patients with other liver diseases, and is a useful adjunct to the hepatic iron index in the diagnosis of hereditary haemochromatosis.90 The assessment of the severity of liver disease is important for the prognosis and risk of primary hepatocellular cancer. Patients noted to have cirrhosis should undergo regular screening for hepatocellular cancer.83

Liver biopsy is particularly useful in identifying the presence or absence of cirrhosis amongst patients with hereditary haemochromatosis who have a history of significant alcohol use. Non-invasive methods of detecting cirrhosis in hereditary haemochromatosis patients, such as elevated serum type IV collagen, also show promise but need to be evaluated further.91 Other methods of diagnosis of hereditary haemochromatosis besides liver biopsy have limited utility. Severe iron overload can be detected by computed tomography scanning, but milder degrees of iron overload may not be detected.92 Standard magnetic resonance imaging has a limited role in the diagnosis and assessment of hereditary haemochromatosis at present as it lacks sensitivity for the lower ranges of hepatic iron concentrations.93, 94 Only semi-quantitative estimates in patients with high tissue iron concentrations are possible in routine practice.95 When performing magnetic resonance imaging for liver iron quantification, a complete hepatic examination is preferable to a simple signal measurement in patients with cirrhosis, as it can also screen for primary liver cancer.96

Treatment

The goal of treatment in hereditary haemochromatosis is to remove excess body iron stores, usually via phlebotomy. This is usually performed by weekly or twice-weekly (if tolerated) venesections of 500 cm3 until the serum ferritin is < 50 ng/mL; a secondary but less important goal is to reduce the transferrin-iron saturation to < 50%.83 Induction phlebotomy using this regimen, compared to less frequent phlebotomy, is also a more reliable method to estimate the quantity of mobilizable iron stores. After the serum ferritin has been lowered to < 50 ng/mL, maintenance therapy may be needed three to four times per year. When iron overload and anaemia are present concomitantly, chelation with deferoxamine may be required. Vitamin C supplementation should be avoided because vitamin C can enhance iron toxicity.83

Liver transplantation

Orthotopic liver transplantation has been performed for end-stage liver disease due to hereditary haemochromatosis and is the only effective treatment for patents with decompensated cirrhosis and otherwise resectable hepatocellular carcinoma. It is difficult to diagnose hereditary haemochromatosis using the transferrin-iron saturation and hepatic iron index in end-stage liver disease as they lack specificity for the disease.3 Genotyping for the HFE mutation C282Y is currently the only reliable means to identify hereditary haemochromatosis in end-stage liver disease. However, only 10–15% of patients with hepatic iron overload in the setting of end-stage liver disease are C282Y homozygotes. Therefore, additional research is needed to evaluate other forms of hereditary haemochromatosis in these patients.

There is a high prevalence of hepatocellular cancer in end-stage liver disease due to hereditary haemochromatosis. The first study on the prevalence of hepatocellular cancer in patients with hereditary haemochromatosis undergoing orthotopic liver transplantation showed a high prevalence with the majority being unsuspected.97 This multicentre study, which was the first with detailed observations on outcome, also showed the need to diagnose hereditary haemochromatosis prior to orthotopic liver transplantation, as hereditary haemochromatosis patients were shown to have a poorer average survival after transplantation. This poorer outcome could not be explained solely by the presence of concomitant hepatocellular cancer. Subsequent studies confirmed the decreased survival in patients undergoing orthotopic liver transplantation for hereditary haemochromatosis. Increased infections (especially fungal) in the first year after transplantation and later cardiac complications were the most common causes of post-transplantation deaths in hereditary haemochromatosis patients.98 Powell showed that the long-term survival of patients with hereditary haemochromatosis who had their iron load adequately removed prior to transplantation was not significantly different from that of other transplanted diseases.99 Decreased post-transplantation survival in patients with iron overload appears to be independent of HFE gene status, suggesting that, regardless of cause, iron overload may be detrimental in patients undergoing orthotopic liver transplantation.100 Recent interim results from the National Hemochromatosis Transplant Registry for survival after orthotopic liver transplantation amongst patients with hepatic iron overload concluded that C282Y homozygotes or C282Y/H63D compound heterozygotes had a decreased 1-year and later survival rate after transplantation compared to controls with hepatic iron overload and other genotypes. This survival difference was present even after controlling for age, United Network for Organ Sharing status or year of transplant. Patients with a hepatic iron index of 1.9 showed a trend towards increased risk of death post-transplantation compared to those with an index of less than 1.9. These data suggest that increased body iron stores may adversely affect outcome after orthotopic liver transplantation in a dose-dependent manner.101 Alternatively, Stuart et al., in an Australian study of 282 consecutive adult patients with cirrhosis, reported that increasing hepatic iron concentration in cirrhosis is a surrogate marker of the severity of the underlying disease, and that poorer post-transplant survival amongst non-hereditary haemochromatosis patients with hepatic iron overload may be due to more severe liver disease.102

The American Association for the Study of Liver Diseases Practice Guidelines for Hereditary Haemochromatosis recommend that liver transplantation is indicated for carefully selected patients with hereditary haemochromatosis. However, because the results have been disappointing, they also recommend that more research is needed to determine the optimum use of transplantation in hereditary haemochromatosis patients. The ongoing National Institutes of Health-sponsored 15 centre study by the National Hemochromatosis Transplant Registry into the outcome after liver transplantation for iron overload may provide answers to some of these questions, and may help to improve the outcome of orthotopic liver transplantation in hereditary haemochromatosis patients.

Conclusions

The understanding of haemochromatosis and iron metabolism has made great strides since the identification of the HFE gene in 1996. Other mutant genes have also been identified and the earlier diagnosis of haemochromatosis is now possible, thus allowing reduced morbidity and mortality. However, this discovery has also raised questions about screening, as not all C282Y homozygotes develop haemochromatosis. Liver biopsy is no longer essential to establish a diagnosis of haemochromatosis, but remains an important diagnostic and prognostic tool.

Acknowledgement

This study was supported in part by K24DK02957 (KVK).

Ancillary