Haemochromatosis: iron still matters

REVIEW

Haemochromatosis: iron still matters

M. L. BASSETT

Gastroenterology Unit,The Canberra Hospital, Canberra Clinical School, Canberra, Australian Capital Territory andDepartment of Medicine, University of Sydney, Sydney, New South Wales, Australia

INTRODUCTION

Iron overload has been essential to the diagnosis ofhaemochromatosis for a long time. Identification of the major haemochromatosis mutations C282Yand H63D in 1996 in a gene known as HFE hasredefined this genetic disorder. Population studieshave confirmed that haemochromatosis is com-mon: approximately one in 200 of the Australianwhite population is genetically predisposed to thisdisease.1,2 Some of the riddles of haemochromatosishave been resolved, to be replaced by others: (i) whyare some C282Y homozygotes spared the disease,(ii) what is the role of the H63D mutation and (iii) should the diagnosis be made on the basis ofgenetic tests or iron studies? The present article willaddress some of these issues, with emphasis on theAustralian experience, and provide guidelines forassessment of iron storage disorders.

DEFINING HAEMOCHROMATOSIS

A European Association for the Study of the Liver(EASL) Consensus Conference on Haemochro-matosis defined this disorder as primary inheritediron overload, either related or unrelated to HFEmutations (the latter having both juvenile and adultforms) (Table 1).3 The specific genetic abnormalitiesin non-HFE haemochromatosis remain uncertain butare probably not located on chromosome 6: juvenileand adult forms have been described.4,5 Acquired ironoverload and other iron storage disorders not provento have a genetic basis are excluded from the defini-tion of haemochromatosis. This definition thereforeemphasizes the two essential characteristics ofhaemochromatosis: the primary genetic nature of thecondition, and iron overload.

How is iron overload defined? There is a lack ofgeneral agreement on the best method of measuringiron stores, and the cut-off between normal stores andiron overload. This probably indicates that no singletest has completely acceptable utility and reliability.For instance, Barton et al. have defined iron overloadas 4 g or more of iron mobilized by phlebotomy.6 TheEASL Consensus Conference arbitrarily graded iron

Correspondence to: A/Professor M. L. Bassett, Gastroenterology Unit,The Canberra Hospital, PO Box 11,Woden, ACT 2606, Australia.Email: [email protected]

Received 30 August 2000; accepted 8 March 2001.

Internal Medicine Journal 2001; 31: 237–242

Abstract

Our ability to detect those predisposed to haemochro-matosis is greatly enhanced by testing for HFEmutations. Ironically, this diagnostic advance has ledto some confusion regarding the criteria for diagnosisof haemochromatosis, with overreliance on genetictesting instead of investigations for iron overload.Because many people who are homozygous for theC282Y mutation, or compound heterozygous for theC282Y and H63D mutations, either do not express oronly partially express the disease, it is essential toconfirm a diagnosis of haemochromatosis on the basisof increased body iron stores. Liver biopsy remains

the best method of confirming this and has an impor-tant role in the patient with either borderline ironoverload or advanced disease. Persistent elevation ofserum ferritin concentration in the absence of overtliver damage, inflammation or neoplasia, and estima-tion of mobilized body iron by repeated phlebotomy,are reasonable alternatives to liver biopsy. Althoughthe precise definition of iron overload is debated, adiagnosis of haemochromatosis cannot be madewithout demonstrating increased body iron stores.(Intern Med J 2001; 31: 237–242)

Key words: C282Y, diagnosis, haemochromatosis,HFE, iron overload.


Bassett238

overload as minimal (1.5–2 g of mobilizable bodyiron; MBI), modest (2–5 g) and severe (> 5 g).3 Thelatter definition is preferred because it includes milderdegrees of iron overload. While serum ferritin is auseful guide to the level of iron stores, it is frequentlynon-specifically elevated. Thus, serum ferritin shouldnot be relied on as the sole indicator of iron overload:confirmation by either liver biopsy or quantitativephlebotomy is essential.

There is a risk that overreliance on HFE mutations fordiagnosis of haemochromatosis will lead to a ‘loose’definition of the disease, without thorough assessmentof iron stores. C282Y homozygosity, or compoundheterozygosity, without increased body iron stores andstable over a period of time should not be classified ashaemochromatosis. If the definition of haemochro-matosis is broadened to include people who have apredisposing genetic abnormality but no ironoverload, healthy people with no morbidity or risk ofmorbidity will be assigned a diagnosis incorrectly.Thiswill have important implications for employment andinsurance. Furthermore, community resistance toearly detection of haemochromatosis may occur.

HFE MUTATIONS WITHOUT IRONOVERLOAD

The concept of tissue damage because of progressiveiron overload in haemochromatosis has been consid-ered important in the diagnosis.7 As family screeningfor asymptomatic disease became standard practice inthe 1980s, many haemochromatosis patients with

milder degrees of iron overload and no tissue damagewere identified, particularly with the aid of HLAtyping. These patients were usually much youngerthan the proband and had no evidence of hepaticfibrosis or other organ damage.8 However, measure-ment of the hepatic iron concentration (HIC)confirmed iron stores well in excess of that expectedfor age-matched controls, thus providing the basis forthe hepatic iron index, determined by dividing thehepatic concentration in µmol/g dry weight by age.8,9

Although unproven, it was assumed that thesepatients would, if untreated, eventually developcirrhosis and other organ damage, and treatment byregular phlebotomy was advised to prevent suchcomplications.10

A new problem has arisen since the HFE mutationswere described. Case detection by family screening,and population screening studies, have identified asubstantial proportion of C282Y homozygotes whoeither do not express, or only partially express, thedisease. In a population screening study of 3011subjects from Busselton,Western Australia, 16 C282Yhomozygotes were identified aged from 30 to 74years.2 Although only one of 16 had normal trans-ferrin saturation, four had no evidence of increasediron stores based on either serum ferritin or hepaticiron index. Two other subjects had an hepatic ironindex below the conventional cut-off (1.9). Only eighthad symptoms or clinical features of haemochro-matosis, three had hepatic fibrosis and one hadcirrhosis. In another Australian study, 17.3% ofsubjects homozygous for the C282Y mutation did notexpress iron overload to meet current diagnosticcriteria for haemochromatosis.11 Similar findings havebeen reported from other countries in family studiesand population screening studies.12,13–15 For example,McDonnell et al. found iron overload (determined asserum ferritin > 95th percentile and mobilized bodyiron > 99th percentile or hepatic iron index > 1.9) inonly 50% of C282Y homozygotes detected byscreening employees.15 Many C282Y homozygotes,perhaps of the order of 30%, do not develop ironoverload and therefore HFE mutations should not bethe sole diagnostic test for haemochromatosis.

Possible explanations for the phenotypic variationinclude variable penetrance, modifying genes andnon-genetic factors. Non-genetic factors include age,dietary iron content, dietary and pharmacologicalsubstances that modify iron absorption (e.g. ascorbicacid, alcohol), blood loss (physiological and patholo-gical), blood donations and other unknown factors.As yet unidentified genetic factors probably account

Table 1 Classification of iron overload

Primary, genetic or hereditary haemochromatosisHFE-related

C282Y +/+C282Y +/– H63D +/–H63D +/+

Non-HFEJuvenileAdult

Secondary iron overloadIron-loading anaemiasThalassaemia majorHereditary spherocytosis

Chronic liver disease

MiscellaneousNeonatal iron overloadAfrican iron overloadPorphyria cutanea tardaOther iron-loading conditions

239Haemochromatosis: iron still matters

for much of the variation in the degree of ironoverload in haemochromatosis.16–18 Occasionally, theapparent discordance between genotype and pheno-type can be accounted for by methodologicalproblems in performing iron studies and genotyping,emphasizing the importance of repeating these testswhen there is doubt.19

IRON OVERLOAD WITHOUT HFEMUTATIONS

In Australia, the majority of patients withhaemochromatosis are homozygous for the C282Ymutation.11 A small percentage are heterozygous forboth C282Y and H63D (compound heterozygotes)or homozygous for H63D. The latter two groupsusually present with lesser degrees of iron overload,compared to C282Y homozygous patients. Althoughsuch patients appear to have a very low, possibly zero, risk of cirrhosis, they should be treated in thesame way as C282Y homozygotes.14 Non-HFE hae-mochromatosis, uncommon in Australia and othercountries populated by people of Northern Europeanorigin, is much more frequent in other countries such as Italy.20 Increasing migration to Australia from outside the Northern European region is likelyto increase the frequency of non-HFE haemo-chromatosis and secondary iron overload (e.g. thalas-saemia major), relative to HFE-haemochromatosis.

HFE HETEROZYGOTES

HFE C282Y heterozygosity is common: approxi-mately 14% of the Australian Caucasian populationcarry one C282Y mutation and 2% are compoundheterozygotes (one C282Y mutation and one H63Dmutation).2 Compound heterozygotes may developmild to moderate iron overload and should be treatedby phlebotomy if the serum ferritin concentration isincreased. In contrast, a small proportion of C282Ynon-compound heterozygotes may have a slightlyraised transferrin saturation or serum ferritin con-centration, but iron overload in these subjects isextremely rare.2,11,12,21 If there is doubt about thediagnosis in these subjects, iron studies and geneticanalyses should be repeated, a liver biopsy performedand if iron overload confirmed, phlebotomy therapyadvised with quantification of mobilizable iron.

The role of heterozygosity for one or other HFEmutation as a cofactor in liver diseases such as alcohol-related liver disease, non-alcoholic steatohepatitis, andchronic viral hepatitis remains controversial, althoughthere is some evidence in support of this concept.22,23

Patients with these disorders should be assessed forevidence of iron overload.This is frequently impossibleon the basis of serum iron and ferritin studies, whichare non-specifically elevated in the presence of hepaticinflammation and necrosis, and therefore a liver biopsyis usually required to assess iron stores in thesepatients.

AN IMPORTANT BUT DIMINISHINGROLE FOR LIVER BIOPSY

Previously, a firm diagnosis of haemochromatosiswould not have been made without a liver biopsy.Now, many patients question the need for liver biopsyand some decline to undergo this investigation. Liverbiopsy has three main roles in the patient withsuspected iron overload: assessment of the degree ofiron loading particularly in borderline cases, the stageof the disease (precirrhosis or cirrhosis), and concur-rent disease such as alcohol-related liver disease orchronic viral hepatitis (Table 2).1–3

While transferrin saturation and serum ferritin areclinically useful markers of iron overload, proof ofhepatic iron overload, particularly in borderline cases,usually requires liver biopsy with assessment of irongrading by Perls’ stain and measurement of hepaticiron concentration. Histological grading of iron byPerls’ stain is not as accurate as measurement of thehepatic iron concentration. The hepatic iron index ismore reliable than either in distinguishing haemo-chromatosis from other disorders masquerading ashaemochromatosis, such as alcoholic siderosis.8,9 Theraised hepatic iron index in haemochromatosisreflects the life-long increase in iron absorption whichis characteristic of this disease, but is also seen insecondary iron overload states such as thalassaemiamajor. A hepatic iron index of > 1.9 has been the ‘ironstandard’ for diagnosis of haemochromatosis for


Table 2 Indications for liver biopsy in suspectedhaemochromatosis

To resolve doubt about the level of hepatic iron storesBorderline increase in serum ferritinSerum ferritin falsely raised (e.g. because of hepatic

damage)To confirm suspected cirrhosis or advanced fibrosis

Absence of severe fibrosis can be predicted by a combination of serum ferritin < 1000 µg/L, normaltransaminase and absence of hepatomegaly

To detect concurrent disease(e.g. alcohol-related liver disease or chronic viral

hepatitis)

Bassett240

many years but this is now questioned and values inthe range 1.5–1.9 have been reported in C282Yhomozygotes with increased iron stores and evenhepatic fibrosis.2,9,24

Liver biopsy has two other roles in haemochro-matosis. It is useful in staging haemochromatosis asprefibrotic, fibrotic or cirrhotic, and in the assessmentof concurrent liver disease. Assessment of cirrhosis isof considerable importance because hepatocellularcarcinoma is exceedingly rare in patients with hae-mochromatosis who do not have cirrhosis. Mildfibrosis may be reversible but cirrhosis is not. Factorsthat predict absence of cirrhosis or severe fibrosis inhaemochromatosis include a pretreatment serum ferri-tin below 1000 µg/L, normal serum aspartate amino-transferase (AST), and absence of hepatomegaly.25

These variables have good reproducibility for negativeprediction but poor reproducibility for the positiveprediction of severe fibrosis.25 Age of the patient is alsoa factor: cirrhosis is rare under the age of 40.24 Ifadvanced fibrosis or cirrhosis are demonstrated, it is common practice to screen periodically for hepato-cellular carcinoma (serum alpha-fetoprotein andultrasound 6 monthly) but there is no evidence, inAustralia, that this is either reliable or cost-effective.Hopefully, with earlier diagnosis of haemochromatosis,family screening and eventually population-basedscreening, cirrhosis from this disease will become rareand the issue of screening for hepatocellular carci-noma will disappear.

Assessment of cofactors for liver damage in haemo-chromatosis is also important because a proportion ofpatients will have additional disease processes, such asalcohol-related liver disease, viral hepatitis, and non-alcoholic steatohepatitis.22–24 Histological examinationof a liver biopsy and measurement of the iron concen-tration can assess the relative contributions of eachmost reliably. Liver biopsy is recommended if there isevidence of concurrent disease based on either clinicalsuspicion or abnormal liver function tests.

QUANTITATIVE PHLEBOTOMY:IMPROVEMENTS NEEDED

Quantitative phlebotomy is a standard method forassessing body iron stores. Repeated weekly phle-botomy until the serum ferritin reaches the lownormal range allows assessment of the mobilizablebody iron (MBI). It is normally performed bymeasurement of the volume (by weighing) of bloodremoved at weekly phlebotomy, and dividing thevolume by 2 to give the amount of iron removed in

milligrams. A more accurate method is to correct forthe increased haemoglobin concentration that hasbeen described in about two-thirds of patients withuntreated haemochromatosis, and which falls afterexcess iron is removed by phlebotomy therapy.6 Thus,the amount of iron removed can be estimated by thefollowing formula, based on stoichiometry:

Iron removed (mg) = Hb (g/L) × volume (L) × 3.45

Quantitative phlebotomy does not take account ofnon-haemoglobin serum iron (which is very small incomparison), nor iron in other iron-containingmolecules or loosely associated with cells and proteins.Nor is an adjustment made for the increased ironabsorption that occurs in haemochromatosis oncephlebotomy commences as this is difficult to quantify.

While quantitative phlebotomy is only an estimate ofiron removed, it will become an increasingly import-ant technique for confirmation of iron overload inhaemochromatosis as patients decline to undergoliver biopsy in favour of genetic tests. There is a needto establish the normal range for MBI in theAustralian population, which will differ for males andfemales, and to determine more precisely a cut-offvalue that separates normal from increased ironstores. In patients with haemochromatosis phle-botomy is usually performed weekly until the serumferritin is in the low normal range, and then reducedto once each 2–4 months, maintaining the serumferritin concentration under 100 µg/L.

SCREENING FOR HAEMOCHROMATOSIS

There are many opportunities for early detection ofhaemochromatosis. Every patient diagnosed shouldtrigger a mandatory search for affected first-degreerelatives: siblings, parents and offspring. Testing ofsecond-degree relatives (cousins, aunts, uncles, niecesand nephews) is not currently recommended becauseof insufficient evidence supporting this, but may havea role in the future.The per cent saturation of the irontransport protein transferrin (serum transferrin saturation), measured fasting on a morning bloodsample, is the single most sensitive biochemicalscreening test for haemochromatosis. If the trans-ferrin saturation is > 45%, genetic analysis for theC282Y mutation and measurement of serum ferritin(to assess better the level of iron stores) should beperformed. Testing for the H63D mutation is notessential unless C282Y heterozygosity has beendemonstrated, or there is unexplained iron overload.


241Haemochromatosis: iron still matters

In Australia, genetic testing is approved as a MedicalBenefit Schedule Fee item for first-degree relatives,and many families will wish to have this done. Theyshould be informed that C282Y homozygosity doesnot make the diagnosis of haemochromatosis andthat it may affect their insurability even if ironoverload is not demonstrated. In those too young orotherwise unable to give informed consent it isprobably wise to utilize transferrin saturation initiallyand defer genetic testing until the person under-stands the implications. If genetic testing is deferred,transferrin saturation should be performed every fewyears because normal values do not rule out laterdevelopment of iron overload. Genetic testing of the unaffected parent can predict risk of haemo-chromatosis in offspring but could raise issues ofnon-paternity and non-maternity.26

Screening for haemochromatosis should be a routinepart of the assessment of all patients with chronicfatigue, abnormal liver function tests, chronic liverdisease, diabetes mellitus, arthritis, cardiomyopathy,gonadal failure, hepatocellular carcinoma and unex-plained skin pigmentation.

CONCLUSION

Remarkable progress has been made in our under-standing of this common inherited condition in thepast few years.While much of the credit goes to thosewho identified the HFE mutations, it was the lateMarcel Simon from Brittany and coworkers whodiscovered a strong association of haemochromatosiswith HLA antigens.27 Their meticulous clinical andfamily studies predicted a susceptibility locus forhaemochromatosis on the short arm of chromosome6 and prepared the way for the genetic work thatfollowed. A ‘spin-off ’ from the recent genetic dis-coveries has been a better understanding of ironabsorption and the metabolic abnormalities inhaemochromatosis, although knowledge is far fromcomplete. Remarkable as the genetic discoveries havebeen, and notwithstanding their obvious clinicalimpact, we must not lose sight of the essential featureof haemochromatosis: iron overload.

REFERENCES

1 Feder JN, Gnirke A, Thomas W et al. A novel MHC class I-like gene is mutated in patients with hereditaryhemochromatosis. Nat Genet 1996; 13: 399–408.

2 Olynyk JK, Cullen DJ, Aquila S, Rossi E, Summerville L,Powell LW. A population-based study of the clinicalexpression of the hemochromatosis gene. N Engl J Med 1999;341: 718–24.

3 Pietrangelo A, Tygstrup N (eds). European Association forStudy of the Liver. International Consensus Conference onHaemochromatosis, World Congress on Iron Metabolism;1999 May 23–28; Sorrento, Italy.

4 Roetto A, Totaro A, Cazzola M et al. Juvenilehemochromatosis locus maps to chromosome 1q. Am J HumGenet 1999; 64: 1388–93.

5 Pietrangelo A, Montosi G, Totara A et al. Hereditaryhemochromatosis in adults without pathogenic mutations inthe hemochromatosis gene. N Engl J Med 1999; 241: 725–32.

6 Barton JC, Bertoli LF, Rothenberg BE. Peripheral blooderythrocyte parameters in hemochromatosis: evidence forincreased erythrocyte hemoglobin content. J Lab Clin Med2000; 135: 96–104.

7 Grace ND, Powell LW. Iron storage disorders of the liver.Gastroenterology 1974; 67: 1257–83.

8 Bassett ML, Halliday JW, Powell LW. Value of hepatic ironmeasurements in early hemochromatosis and determinationof the critical iron level associated with fibrosis. Hepatology1986; 6: 24–9.

9 Summers KM, Halliday JW, Powell LW. Identification ofhomozygous hemochromatosis subjects by measurement ofhepatic iron index. Hepatology 1990; 12: 20–25.

10 Niederau C, Fischer R, Sonnenberg A, Stremmel W,Trampisch HJ, Strohmeyer G. Survival and causes of death incirrhotic and in noncirrhotic patients with primaryhemochromatosis. N Engl J Med 1985; 313: 1256–62.

11 Crawford DHG, Jazwinska EC, Cullen LM, Powell LW.Expression of HLA-linked haemochromatosis in subjectshomozygous or heterozygous for the C282Y mutation.Gastroenterology 1998; 114: 1003–8.

12 Sham RL, Ou CY, Cappuccio J, Braggins C, Dunnigan K,Phatak PD. Correlation between genotype and phenotype inhereditary hemochromatosis: analysis of 61 cases. Blood CellsMol Dis 1997; 23: 314–20.

13 Brissot P, Moirand R, Jouanolle AM et al. A genotypic study of217 unrelated probands diagnosed as ‘genetic hemochromatosis’on ‘classical’ phenotypic criteria. J Hepatol 1999; 30: 588–93.

14 Adams PC, Kertensz AE, McLaren CE, Barr R, Bamford A,Chakrabarti S. Population screening for hemochromatosis: acomparison of unbound iron-binding capacity, transferrinsaturation, and C282Y genotyping in 5211 voluntary blooddonors. Hepatology 2000; 31: 1160–4.

15 McDonnell SM, Hover A, Gloe D, Ou CY, Cogswell ME,Grummer-Strawn L. Population-based screening forhemochromatosis using phenotypic and DNA testing amongemployees of health maintenance organizations in Springfield.Missouri Am J Med 1999; 107: 30–37.

16 Crawford D, Halliday JW, Summers KM, Bourke MJ, PowellLW. Concordance of iron storage in siblings with genetichaemochromatosis: evidence for a predominantly geneticeffect on iron storage. Hepatology 1993; 17: 833–7.

17 Whitfield JB, Cullen LM, Jazwinska EC et al. Effects of HFEC282Y and H63D polymorphisms and polygenic backgroundon iron stores in a large community sample of twins. Am JHum Genet 2000; 66: 1246–58.

18 Pratiwi R, Fletcher LM, Pyper WR et al. Linkagedisequilibrium analysis in Australian haemochromatosispatients indicates bipartite association with clinicalexpression. J Hepatol 1999; 31: 39–46.

19 Jeffrey GP, Chakrabarti S, Hegele RA, Adams PC.Polymorphism in intron 4 of HFE may cause overestimulationof C282Y homozygote prevalence in haemochromatosis. NatGenet 1999; 22: 325–6.


Bassett242

20 Piperno A, Sampietro M, Pietrangelo A et al. Heterogeneity ofhemochromatosis in Italy. Gastroenterology 1998; 114:996–1002.

21 Simon M, Fauchet R, Hespel JP et al. Idiopathichemochromatosis: a study of biochemical expression in 247heterozygous members of 63 families: evidence for a singlemajor HLA-linked gene. Gastroenterology 1980; 78: 703–8.

22 Stuart KA, Busfield F, Jazwinska EC et al.The C282Ymutation in the haemochromatosis gene (HFE) and hepatitis Cvirus infection are independent cofactors for porphyria cutaneatarda in Australian patients. J Hepatol 1998; 28: 404–9.

23 George DK, Powell LW, Losowsky MS. Thehaemochromatosis gene: a co-factor for chronic liver diseases?J Gastroenterol Hepatol 1999; 14: 745–9.

24 Bacon BR, Olynyk JK, Brunt EM, Britton RS, Wolff RK.HFE genotype in patients with hemochromatosis and otherliver diseases. Ann Intern Med 1999; 130: 953–62.

25 Guyader D, Jacquelinet C, Moirand R et al. Noninvasiveprediction of fibrosis in C282Y homozygoushemochromatosis. Gastroenterology 1998; 115: 929–36.

26 Adams PC. Implications of genotyping of spouses to limitinvestigation of children in genetic hemochromatosis. ClinGenet 1998; 53: 176–8.

27 Simon M, Bourel M, Fauchet R, Genetet B. Association ofHLA-A3 and HLA-B14 antigens with idiopathichaemochromatosis. Gut 1976; 17: 332–4.


Documents

Haemochromatosis: iron still matters