The role of transferrin saturation as a screening test for hereditary haemochromatosis in an Irish population seeking medical care

Transcription

1 The role of transferrin saturation as a screening test for hereditary haemochromatosis in an Irish population seeking medical care R O Hara 1, N Cavanagh 1, M Cassidy 1 and M Cullina 2 Original Article Abstract Addresses 1 Biochemistry Laboratory Portiuncula Hospital Ballinasloe Co. Galway, Ireland 2 Immunology Department University College Hospital Galway, Ireland Correspondence Ms R O Hara Biochemistry.Portiuncula@whb.ie Background Genetic studies have indicated a high prevalence of mutations for hereditary haemochromatosis (HH) in Irish populations. Given the high cost of genetic screening and the ethical implications thereof, we assessed the role of transferrin saturation (TS) as a primary screening test for HH in an Irish population seeking medical care. Methods TS and ferritin were measured on 330 consecutive blood specimens received in the laboratory for routine screening. Patients with TS 445% were genetically screened for the C282Y and H63D mutations. Results Twenty-six patients had TS values 445%. Of these, specimens were available for genetic screening on 20 patients. Three previously undiagnosed patients were found to be C282Y homozygotes and one a compound heterozygote (C282Y/ H63D). The prevalence of C282Y homozygotes was 0 93%, which is similar to the reported prevalence found in Irish populations by genetic screening. Conclusion Given the number of positive findings in this study, we conclude that, in the absence of a national programme, TS could be used as part of a healthscreening panel in the Irish setting. Patients expressing iron overload would be detected before organ damage occurred, leading to less severe clinical disease and better patient prognosis. Introduction Hereditary haemochromatosis (HH) is an autosomal recessive disorder of iron metabolism. 1 It is one of the most common hereditary metabolic diseases in Caucasians, particularly those of Northern European descent, a ecting one in every individuals. 2 It is characterized by an increase in intestinal absorption of iron, resulting in deposition of iron in the parenchymal cells of the liver, heart, pancreas and other organs, 3 which, if untreated, can lead to subsequent complications, such as hepatic cirrhosis, arthritis, cardiomyopathy and diabetes. 4 In 1996, Feder et al. 5 discovered two mutations in the gene associated with HH (now called the HFE gene). 6 These mutations are the missense mutation C282Y, which is found in % of cases of HH, and the H63D mutation whose role in the disease is not so signi cant. 5,7 A recent study of Irish neonates found1% to be homozygous for the C282Y mutation, with 4% compound heterozygotes (C282Y/H63D). 8 An allele frequency of11% for C282Y was reported, which is one of the highest worldwide. 8 A ected individuals who are diagnosed and treated at an early stage, before the onset of diabetes or cirrhosis, are less symptomatic and have an improved life-expectancy. 9,10 Treatment rst described in 1952 has changed little, and consists of a programme of therapeutic phlebotomy to remove the excess iron. 11 HH would therefore seem to be an ideal candidate for population screening. However, there is still debate as to whether screening is bene cial, 12 which populations would bene t and which type of screening is most e ective. 13,14 Population screening has not yet been recommended in Europe, and a consensus panel has found that, if population screening were introduced, it would be in the absence of good evidence of a favourable bene t. 15 However, the College of American Pathologists has recommended that all adults over 20 years of age should be screened for iron overload, 2003 The Association of Clinical Biochemists 169

2 170 O Hara et al. initially by a transferrin saturation (TS) test, followed by con rmatory tests if the TS is elevated. 1 There have been a number of Irish studies on HH, mainly to determine the prevalence in the Irish population. 8,16,17 These have all used genetic screening and, given the high prevalence found, some would advocate population screening in the future. However, genetic screening is expensive 18 and is usually only performed in larger centres where the expertise and facilities are available. Before the discovery of the HFE gene, the TS test was considered the single most reliable screening test for haemochromatosis 19,20 and is still the most sensitive phenotypic test. 21 It is inexpensive, easy to perform and is available in most routine clinical chemistry laboratories. Because there is no screening programme for HH in Ireland, and given the high prevalence of HH and the high cost of genetic screening, phenotypic screening using TS may be a better alternative. The aim of the present study is to assess the role of TS as a screening test for HH in an Irish population seeking medical care. Subjects and methods Three hundred and thirty consecutive fasting blood specimens, received in the laboratory for routine screening tests over a 2-month period, were included in the study. A total of 24% of the specimens were from inpatients (mainly elective surgery cases), 27% from outpatients and 49% were patients of general practitioners. Iron, unbound iron-binding capacity (UIBC), alanine aminotransferase (ALT), g-glutamyltransferase (GGT) and C-reactive protein (CRP) were measured for each specimen using a Roche Integra 700 (Roche Diagnostics Ltd, Lewes, UK). Iron was measured using the Roche FerroZine method without deproteinization; the between-batch percentage coe cient of variation (%CV) was 1Í1 at 19Í7 mmol/l. The UIBC was also measured using the Roche FerroZine method (between-batch %CV of 3Í2 at 25Í7 mmol/ L). Total iron-binding capacity (TIBC) was calculated from iron and UIBC results (Iron+UIBC=TIBC). TS was calculated using the formula [(Iron/TIBC)6100]. Ferritin was measured by a sandwich immunoassay using chemiluminescence technology on an ACS 180 Plus analyser (Bayer plc, Newbury, UK). The betweenbatch %CV for ferritin was 4Í3 at 137 mg/l. Liver function tests (LFTs) were used to assess whether liver damage had occurred when iron overload was detected. CRP results were used to rule out raised ferritin due to infection rather than iron overload. Although these tests were carried out on all patients included in this study, this would be unnecessary in general screening. Instead, cascade testing could be used ^ i.e. when a patient is found to have abnormal TS results, LFTs, ferritin and CRP could be assayed. Data, including age, sex and diabetic status, were recorded for each patient to whom a unique study number was assigned. A total of 22% of the fasting specimens received were from diabetics. Because diabetes is a known complication of HH, a patient s diabetic status was noted to see if a higher prevalence of HH was present in the diabetic groups. Blood samples from patients with a raised TS were genetically screened for the C282Y and H63D mutations in the HFE gene when suitable samples were available. The method used to detect the HFE mutations was a multiplex polymerase chain reaction, with appropriate uorescence-labelled oligonucleotide hybridization probes, on a LightCycler (Roche Diagnostics Ltd) as described by Mangasser- Stephan et al. 22 Melting curve analysis of the ampli cation products was then used for genotyping, with each genotype having a unique curve. 22 Transferrin saturation Reported cut-o s for TS from di erent studies have ranged from 45% to 70%. 14,23,24 McLaren et al. used a cut-o of 45% for TS, giving a sensitivity of 98% and a speci city of 100%. 25 To establish a valid cut-o for TS in our study, a search of the laboratory database was made for previous haemochromatosis genetic screen requests. Genetic results for 85 patients (59 men, mean age 50Í6 years; 26 women, mean age 51Í0 years) were found. TS values for all of these patients were calculated retrospectively, using the results of iron studies. These results were used, together with the results of the genetic tests, to measure the sensitivity and speci city of the TS test in screening for HH (these are presented in Table 1). A TS cut-o value of 45% gave a test sensitivity of 100%, with a speci city of 72Í7%. The speci city was increased by raising the cut-o value, or by using a combination of TS values and ferritin, but this resulted in a loss of sensitivity. HH is a serious, but treatable, disease, and it was felt that 100% sensitivity is preferable; 26 therefore, a cut-o of 45% was adopted for this study. Table 1. Sensitivity and speci city of the transferrin saturation assay in screening for hereditary haemochromatosis Cut-off point Sensitivity (%) Speci city (%) TS 45% TS 60% TS 45% and ferritin 300 mg/l TS 60% and ferritin 300 mg/l The results were obtained from previously screened patients in the laboratory database. TS ˆ transferrin saturation.

3 Screening for hereditary haemochromatosis 171 Table 2. Summary of study results Men Ethical considerations This study was approved by the Ethics Committee of Portiuncula Hospital. Specimens for genetic screens were sent to an outside laboratory. Study numbers only were used to identify these patients, so their genetic screens were performed anonymously. Results from the genetic screen were only recorded by study numbers and were not recorded in the hospital database. Statistical analysis Statistical analyses were performed using the Minitab software package. Means, standard deviations (SD) and ranges were calculated using descriptive statistics. TS values were compared between groups using the Student s t-test. Women Diabetic Non-diabetic All Diabetic Non-diabetic All Number of patients Age (years) Range TS (%) Range Ferritin (mg/l) Range TS 445% TS 460% Ferritin 4300 mg/l Number of C282Y homozygotes TS=transferrin saturation. Results A summary of the study results is given in Table 2. A total of 330 patients were included, 149 men, mean age 56Í5 years (SD ˆ 14Í8), and 181 women, mean age 61Í8 years (SD ˆ 14Í3). The distribution of TS is shown in Fig. 1. The mean values of TS for men (33Í0%, SD ˆ 13Í3%) was signi cantly di erent from women (26Í8% SD ˆ 13Í3%, P50 05). There was no signi cant di erence (P40Í05) between female diabetics and non-diabetics and between male diabetics and non-diabetics. The mean ferritin concentration for all men studied was 184 (SD ˆ 141 mg/l, range 7Í0-914) and mean ferritin for women studied was 122 (SD ˆ mg/l, range 5 0^1345). Transferrin saturation values ranged from 2^76% for men and 3^94% for women. A total of 26 patients had TS values greater than 45%; these were sent for genetic screening if an EDTA specimen was available. An EDTA specimen was not available for ve patients; thus, 21specimens were sent for genetic screening. DNAwas extracted successfully Figure 1. Distribution of transferrin saturation in (a) women and (b) men. from 20 specimens, one being unsuitable. Genetic results were available for 20 patients who had a TS value greater than 45%. None of these patients had an abnormal CRP value. The pro le for each of these patients is shown in Table 3. Overall, there were three C282Y homozygotes and one compound (C282Y/H63D) heterozygote. The prevalence of C282Y homozygotes in this population was thus 3 in 324 (0Í93%); the prevalence of compound heterozygotes was 1 in 324 (0Í3%). The C282Y homozygotes were two female patients and one male patient. All these patients had TS values greater than 70% and raised ferritin values. Neither of the women had abnormal LFTs, but the man had mildly

4 172 O Hara et al. Table 3. Biochemical and genetic results for patients with raised transferrin saturation Study number Sex Age (years) Diabetic status* TS (%) ALT (IU/L) GGT (IU/L) Iron (mmol/l) TIBC (mmol/l) UIBC (mmol/l) Ferritin (mg/l) CRP (mg/l) Genetic test** 299 M 47 N Normal 140 M 40 Y Normal 294 M 41 N H63D+ 62 M 53 Y H63D++ 36 M 54 N H63D++ 39 M 52 Y H63D F 77 N Normal 285 M - Y C282Y+ 57 M 67 Y Normal 150 F 68 N Normal 100 M 29 Y H63D+ 96 M 23 Y Normal 224 M 55 N Normal 264 M 23 N C282Y+ 132 F 44 N C282Y+ 142 F 44 Y C282Y+ 86 M 56 N C282Y++ 11 M 66 N C282Y+/ H63D+ 186 F 74 N C282Y F 77 N C282Y++ *Y ˆ diabetic, N ˆ non-diabetic. TIBC ˆ total iron-binding capacity. **+ ˆ heterozygote; ++ ˆ homozygote. TS ˆ transferrin saturation; ALT ˆ alanine aminotransferase; GGT ˆ g-glutamyltransferase, CRP ˆ C-reactive protein. raised ALT and GGT levels. The compound heterozygote was a man who had abnormal LFTs. Discussion The prevalence of C282Y homozygotes in the study population was found to be 0Í93%. This is higher than the published prevalence of C282Y homozygotes in Northern Europe (0Í3-0Í5%) 2 and in the UK (0Í7%), 27 but is similar to that found in a recent genetic study of 800 Irish neonates (1% C282Y homozygotes). 8 Other studies of Irish populations have estimated a homozygote frequency of 1Í96% 16 and 1%, 28 but these numbers were based on estimates calculated from the prevalence of C282Y heterozygotes. Although a study 17 of 187 Irish blood donors showed an apparent C282Y homozygote prevalence of 2Í1%, this was based on a relatively small population compared with the neonatal group mentioned above. The prevalence of 0Í93% found in this study is similar to the 1% prevalence found in the neonatal study, indicating a high penetrance of the gene in the study population. It would have been interesting to further assess the homozygotes and heterozygotes for clinical signs and symptoms of HH, but this was not possible because the design of the study was anonymous. Some homozygotes may possibly have been missed, due to false negatives, if any of the non-expressing C282Y homozygotes had TS values less than 45%. However, McLaren et al. predicted that a fasting TS value of 45% or higher would identify 98% of individuals who are C282Y homozygotes but no normal subjects. 25 In contrast to McLaren s ndings, there were normal subjects with TS values greater than 45% among the previously screened patients in our database as well as in the patients included in this study. This concurs with the ndings of Olynyk et al., who also found normal individuals with TS values greater than 45%. 29 Their ndings of 0Í5% of normal individuals with a TS value greater than 45% is lower than the 2Í1% found in the present study. Altogether, 8Í2% of patients from our database who were genetically normal had an increased TS, but this may be because these were not fasting TS results. It is recommended that, if a patient has a raised TS value, a repeat test should be carried out on a fasting sample. 30 Overall, in the present study, there were 16 patients with a raised TS who were neither C282Y homozygotes nor compound heterozygotes. If the TS cut-o value is increased to 60%, the number of patients in this category is reduced to three. Interestingly, these three patients are all C282Y heterozygotes. It has been reported that C282Y heterozygotes tend to have higher TS values than the normal population, but lower than homozygotes. 25,31,32 Previous studies have also found increased TS levels associated with alcohol consumption. 19,29 However, information on the alcohol intake of the patients in this study was not

5 Screening for hereditary haemochromatosis 173 available. As Ireland has a high prevalence of C282Y heterozygotes, this may also contribute to a higher false positive rate. It has been recommended that appropriate local reference values for screening for iron overload be established, as signi cant regional variations occur due to environmental and genetic factors. 19 There were no homozygotes with TS values between 45% and 60%, so increasing the cut-o to 60% should not markedly decrease the sensitivity of the test although it would increase the speci city. Although 45% was adopted as the cut-o value for TS in this study, in the light of the results, a higher level may be appropriate for screening purposes. Patients with TS values between 45% and 60% could be rescreened at a later date. Alternatively, ferritin could be used in conjunction with TS, as this would increase the sensitivity (see Table 1). As TS values are higher in men than in women, it might also be appropriate to use a lower cut-o point for women when selecting patients for genotyping, although all homozygotes in this study had TS 460%. Even if the TS test misses some C282Y homozygotes, it is not as important to nd them all as it is to nd those expressing iron overload. This is one of the advantages of using initial phenotypic screening tests and then con rming the results by genotyping. There was only one (0Í3%) compound heterozygote (C282Y/H63D) with iron overload found in this study, whereas a study using genetic screening found a prevalence of 4% in the Irish population. 8 This concurs with the view that there is a low penetrance associated with this genotype. 33 Of the known diabetics included in this study, none were found to be C282Y homozygotes. This nding would support the ndings of Ferandez-Real et al., who reported that the frequency of the C282Y mutation is the same in patients with type 2 diabetes as in control subjects. 34 A recent study in the UK measured TS only on specimens with raised ALT, and found a much lower prevalence of HH than expected. 27 However, two of the genetically positive patients in the present study had normal ALT while the male homozygote had a raised ALT.These results are similar to those of George et al., 35 who found that only 48% of their patients had raised ALT; therefore, pre-selecting specimens with raised ALT will reduce the usefulness of the screening test. Family screening is recommended, particularly in siblings, when patients are found to be homozygous. 36 Patients who are C282Y homozygotes, or compound heterozygotes and have a raised TS but a normal ferritin, should be monitored annually with iron studies. 29 Treatment is commenced when the ferritin concentration rises above the reference range. 29 As shown in this and other studies, 8,16,28 Ireland has one of the highest gures for C282Y homozygosity worldwide. In the absence of population screening in Ireland, opportunistic screening would be useful. As pointed out by George et al., 37 individuals (such as diabetics) who are already having blood tests are more likely to present themselves for a haemochromatosis screening test than an asymptomatic individual from the general population. A TS value is readily obtained from the same blood specimen using the same instruments as those used for lipid screening and analysed at very little extra cost. Although there are those who would advocate genetic screening, this has a number of disadvantages, as pointed out earlier. It is more expensive: per test compared with 1Í61 for TS and 3Í40 for ferritin as used in this study. There are also ethical implications inherent in genetic testing, particularly as the patient may not go on to develop the disease. Genetic screening will only detect patients with HFE-associated haemochromatosis, whereas phenotypic screening detects all forms of iron overload. An additional advantage of phenotypic screening is that it not only detects iron overload but will also detect patients with anaemia. 38 A TS value less than 15% is indicative of iron de ciency and these patients should be further evaluated. 38,39 In the present study, 15 patients (three men and 12 women) had TS values 515% and ferritin 520 mg/l, indicating that 2Í0% of the men and 6Í6% of the women had iron de ciency. There is no large published study of the phenotypic expression of the C282Y gene in Ireland; such a study would be useful in determining the necessity for population screening. Given the results of the present study, it seems that TS could play a major role as a screening test for HH in Ireland. A study by Bacon and Sadiq 3 has shown that TS, if included in general biochemistry screening panels, would detect patients expressing iron overload before symptoms occurred. Patients thus identi ed by screening have been shown to have less severe clinical, biochemical and pathological ndings than patients who have in the past been identi ed by symptoms and signs of the disease. 3,18 Opportunistic screening would therefore greatly improve the prognosis of a ected individuals. Patients identi ed by a raised TS value while LFTS and ferritin concentrations are still normal are unlikely to have organ damage. 40 Ferritin concentrations can then be monitored and treatment commence before any serious damage occurs. 40 Acknowledgements This paper forms part of a dissertation submitted for the award of MSc in Biomedical Sciences to the School of Biomedical Sciences, University of Coleraine, Coleraine, UK.

6 174 O Hara et al. References 1 Press RD. Hereditary haemochromatosis: impact of molecular and iron-based testing on the diagnosis, treatment, and prevention of a common, chronic disease. Arch Pathol Lab Med 1999; 123: Lyon E, Frank EL. Hereditary haemochromatosis since discovery of the HFE gene. Clin Chem 2001; 47: Bacon BR, Sadiq SA. Hereditary haemochromatosis: presentation and diagnosis in the 1990s. Am J Gastroenterol 1997; 92: Stehney MA. Genetics in practice: 1. Hereditary haemochromatosis. Hosp Pract 2000; 35: Feder JN, Gnirke A, Thomas W, Tsuchihashi Z, Ruddy DA, Basava A, et al. A novel MCH class 1-like gene is mutated in patients with hereditary haemochromatosis. Nat Genet 1996; 13: Brandhagen DJ, Fairbanks VF, Batts KP, Thibodeau SN. Update on hereditary haemochromatosis and the HFE gene. Mayo Clin Proc 1999; 74: Barton JC, Acton RT. Population screening for haemochromatosis: has the time nally come? Curr Gastroenterol Rep 2000; 2: Byrnes V, Ryan E, Barrett S, Kenny P, Mayne P, Crowe J. Genetic haemochromatosis, a Celtic disease; is it now time for population screening? Genet Test 2001; 5: Haddow JE, Ledue TB. Preventing manifestations of hereditary haemochromatosis through population-based screening. J Med Screen 1994; 1: Crawford DH, Leggett BA, Powell LW. Haemochromatosis. Baillieres Clin Gastroenterol 1998; 12: Barton JC, McDonnell SM, Adams PC, Brissot P, Powell LW, Edwards CQ, et al. Management of haemochromatosis. Ann Intern Med 1998; 129 (Suppl): Beutler E, Felitti VJ, Koziol JA, Ho NJ, Gelbart T. Penetrance of 845G*A (C282Y) HFE hereditary haemochromatosis mutation in the USA. Lancet 2002; 359: Allen K, Williamson R. Should we genetically test everyone for haemochromatosis? J Med Ethics 1999; 25: Adams PC, Kertesz AE, McLaren CE, Barr R, Bamford A, Chakrabarti S. Population screening for haemochromatosis: a comparison of unbound iron-binding capacity, transferrin saturation, and C282Y genotyping in 5211 voluntary blood donors. Hepatology 2000; 31: Consensus Panel EASL. International Consensus Conference on Haemochromatosis Part III. Jury Document. J Hepatol 2000; 33: Ryan E, O Keane C, Crowe J. Haemochromatosis in Ireland and HFE. Blood Cells Mol Dis 1998; 24: Ryan F, Vaughan J. Haemochromatosis mutation analysis in a normal Irish population. Br J Biomed Sci 2000; 57: Edwards CQ, Griffen LM, Ajioka RS, Kushner JP. Screening for hemochromatosis: phenotype versus genotype. Semin Hematol 1998; 35: Porto G, Vincente C, Fraga J, da Silva BM, de Sousa M, Haemochromatosis Clinical and Research Group. Importance of establishing appropriate local reference values for the screening of haemochromatosis: a study of three different control populations and 136 haemochromatosis family members. J Lab Clin Med 1992; 119: Edwards CQ, Griffen LM, Goldgar D, Drummond C, Skolnick MH, Kushner JP. Prevalence of haemochromatosis among 11,065 presumably healthy blood donors. N Engl J Med 1988; 318: Brissot P, Guyader D, Loreal O, Laine F, Guillygomarch A, Moirand R, et al. Clinical aspects of hemochromatosis. Transfus Sci 2000; 23: Mangasser-Stephan K, Tag C, Reiser A, Gressner AM. Rapid genotyping of haemochromatosis gene mutations on the lightcycler with uorescent hybridization probes. Clin Chem 1999; 45: McCullen MA, Crawford DH, Dimeski G, Tate J, Hickman PE. Why there is discordance in reported decision thresholds for transferrin saturation when screening for hereditary haemochromatosis Hepatology 2000; 32: Olsson KS, Eriksson K, Ritter B, Heedman PA. Screening for iron overload using transferrin saturation. Acta Med Scand 1984; 215: McLaren CE, McLachlan GJ, Halliday JW, Webb SI, Leggett BA, Jazwinska EC, et al. Distribution of transferrin saturation in an Australian population: relevance to the early diagnosis of haemochromatosis. Gastroenterology 1998; 114: Galen RS, Gambino SR, eds. Beyond Normality: The Predictive Value and Ef ciency of Medical Diagnoses. New York: Wiley, Bhavnani M, Lloyd D, Bhattacharyya A, Marples J, Elton P, Worwood M. Screening for genetic haemochromatosis in blood samples with raised alanine aminotransferase. Gut 2000; 46: Merryweather-Clarke AT, Pointon JJ, Shearman JD, Robson KHJ. Global prevalence of putative haemochromatosis mutations. J Med Genet 1997; 34: Olynyk JK, Cullen DJ, Aquilia S, Rossi E, Summerville L, Powell LW. A population-based study of the clinical expression of the haemochromatosis gene. N Engl J Med 1999; 341: Dooley JS, Walker AP. Genetic haemochromatosis: detection, management, and population screening. Genet Test 2000; 4: Crawford DHG, Jazwinska EC, Cullen LM, Powell LW. Expression of HLA-linked haemochromatosis in subjects homozygous or heterozygous for the C282Y mutation. Gastroenterology 1998; 114: Bulaj ZJ, Griffen LM, Jorde LB, Edwards CQ, Kushner JP. Clinical and biochemical abnormalities in people heterozygous for haemochromatosis. N Engl J Med 1996; 335: Adams PC. Population screening for haemochromatosis. Gut 2000; 46: Fernandez-Real JM, Vendrell J, Baiget M, Gimferrer E, Ricart W. C282Y and H63D mutations of the haemochromatosis candidate gene in type 2 diabetes. Diabetes Care 1996; 22: George DK, Fletcher LM, Monk CD, Walker NI, Do K, Powell LW. Is hepatic iron the cause of raised transaminases in haemochromatosis. Hepatology 1998; 28: 421A 36 Worwood M. Pathogenesis and management of haemochromatosis. Br J Haematol 1999; 105 (Suppl 1): George DK, Evans RM, Crofton RW, Gunn IR. Testing for haemochromatosis in the diabetic clinic. Ann Clin Biochem 1995; 32: McDonnell SM, Phatak PD, Felitti V, Hover A, McLaren GD. Screening for hemochromatosis in primary care settings. Ann Intern Med 1998; 129: Phatak PD, Sham RL, Cappuccio JD. Screening for haemochromatosis: additional considerations {correspondence}. Gastroenterology 1996; 110: Powell LW, George DK, McDonnell SM, Kowdley KV. Diagnosis of haemochromatosis. Ann Intern Med 1998; 129 (Suppl): Accepted for publication 8 October 2002