Steatohepatitis/Metabolic Liver Disease
HFE Cys282Tyr homozygotes with serum ferritin concentrations below 1000 μg/L are at low risk of hemochromatosis†‡
This study was funded by National Institute of Diabetes, Digestive and Kidney Diseases (USA) (1 RO1 DK061885-01 A2) and the National Health and Medical Research Council (Australia) (grants no. 251668, 209057). Cohort recruitment was funded by VicHealth and The Cancer Council Victoria. The National Health and Medical Research Council also provided funding for the following authors: L.C.G. and K.J.A. (Career Development Awards), M.B.D. and J.K.O. (Practitioner Fellowships), G.J.A. (Senior Research Fellowship) and J.L.H. (Australia Fellowship).
Potential conflict of interest: Nothing to report.
Hemochromatosis gene (HFE)-associated hereditary hemochromatosis (HH) is a genetic predisposition to iron overload and subsequent signs and symptoms of disease that potentially affects approximately 80,000 persons in Australia and almost 1 million persons in the United States. Most clinical cases are homozygous for the Cys282Tyr (C282Y) mutation in the HFE gene, with serum ferritin (SF) concentration >1000 μg/L as the strongest predictor of cirrhosis. The optimal treatment regimen for those with SF concentrations above the normal range but <1000 μg/L is unknown. We assessed HFE mutations in a prospective cohort of 31,192 participants of northern European descent, aged 40-69 years. An HFE-stratified random sample of 1438 participants including all C282Y homozygotes with iron studies 12 years apart were examined by physicians blinded to participants' HFE genotype. All previously undiagnosed C282Y homozygotes (35 male, 67 female) and all HFE wild-types (131 male, 160 female) with baseline and follow-up SF concentrations <1000 μg/L were assessed for HH-associated signs and symptoms including abnormal second/third metacarpophalangeal joints (MCP2/3), raised liver enzymes, hepatomegaly, and self-reported liver disease, fatigue, diabetes mellitus, and use of arthritis medication. The prevalence of HH-associated signs and symptoms was similar for C282Y homozygotes and HFE wild-types for both normal and moderately elevated SF concentrations. The maximum prevalence difference between HFE genotype groups with moderately elevated SF was 11% (MCP2/3, 95% confidence interval = −6%, 29%; P = 0.22) and for normal SF was 6% (arthritis medicine use, 95% confidence interval = −3%, 16%; P = 0.11). Conclusion: Previously undiagnosed C282Y homozygotes with SF concentrations that remain below 1000 μg/L are at low risk of developing HH-associated signs and symptoms at an age when disease would be expected to have developed. These observations have implications for the management of C282Y homozygotes. HEPATOLOGY 2010
Hereditary hemochromatosis (HH) refers to symptoms and signs of disease that result from an inherited predisposition to iron overload. Iron overload is preventable, but can lead to significant health problems, including arthritis, hepatic cirrhosis, hepatocellular carcinoma, fatigue, and diabetes mellitus, if it is left untreated.1 More than 80% of patients presenting with symptomatic iron overload2, 3 are homozygous for the 845G→A mutation in the hemochromatosis (HFE) gene, which leads to the Cys282Tyr (C282Y) substitution in the HFE protein.4 The prevalence of C282Y homozygotes is at least 1 in 200 for people of northern European descent.5, 6 The majority of C282Y homozygotes have elevated iron indices7, 8 but the serum ferritin (SF) concentration threshold at which there is an increased risk of developing HH-associated signs and symptoms other than cirrhosis is not known.
We have recently shown that at least 28% of male C282Y homozygotes develop iron overload–related disease (as defined by both the presence of documented iron overload9 and one of the following five objective HH features: hepatocellular carcinoma, cirrhosis/fibrosis, physician-diagnosed symptomatic HH, elevated liver enzymes, or evidence of HH-associated arthritis),7 with onset in the majority by age 55 years. Other studies have shown that individuals with SF concentrations >1000 μg/L are at significantly increased risk of cirrhosis.10, 11 Assessment of HH-associated signs and symptoms for C282Y homozygotes has largely been limited to clinical case series where the sample sizes were greater for those with both SF concentrations >1000 μg/L and symptoms compared with those with only moderately elevated SF concentrations (i.e., above the upper limit of the normal range but below 1000 μg/L).10–12 Several studies have reported prevalence estimates for C282Y homozygotes who were identified through cascade screening of relatives of a hemochromatosis-affected proband.13, 14 The relatedness of individuals, however, could lead to within-family correlation between both iron indices and the risk of disease, which has the potential to bias prevalence estimates of HH-associated signs and symptoms for C282Y homozygotes.
Several population-based studies have demonstrated that the majority (60%-80%) of untreated C282Y homozygotes develop SF concentrations that are elevated but below the threshold of 1000 μg/L.8 Assuming a C282Y homozygosity prevalence of 0.44%8 and a white population of 223,965,009,15 we estimate that in the United States alone there are almost 700,000 C282Y homozygotes who will develop SF concentrations that are elevated but below 1000 μg/L8 and almost 55,000 of these individuals in Australia. Given the greater prevalence of HFE mutations in the northern European population, the corresponding figure for the United Kingdom is likely to exceed 200,000. However, there is currently no population-based evidence from any country for the risk of developing HH-associated signs and symptoms for those individuals with moderately elevated SF. Such data would have implications for both clinical practice and population-based genetic screening for HH.16
We used an HFE genotype–stratified random sample of participants in a cohort study prospectively sampled and followed over a 12-year time period to assess the prevalence of HH-associated signs and symptoms for C282Y homozygotes with SF concentrations <1000 μg/L and to compare this with the corresponding prevalence for controls with neither the C282Y nor His63Asp (H63D) mutation using data collected when both participants and physicians were blinded to HFE genotype. Our findings on the prevalence of HH-associated signs and symptoms for C282Y-H63D compound heterozygotes and the other HFE genotype groups have been published elsewhere.7, 17
Patients and Methods
The present study, known as HealthIron, is a substudy of the Melbourne Collaborative Cohort Study (MCCS).18 Between 1990 and 1994, 41,514 people (24,469 of whom were women) with a target age range of 40-69 years were enrolled in the MCCS. Participants were recruited via the electoral roll (voting is compulsory in Australia), advertisements, and community announcements in local media. The majority of participants gave a blood sample at baseline, which was aliquoted as blood spots on Guthrie cards and stored at room temperature. In addition, 1 mL samples of buffy coats and plasma were stored in liquid nitrogen.
For the HealthIron study, the DNA samples from a subsample of participants were extracted from Guthrie cards (n = 23,484) using Chelex reagent or from frozen buffy coats (CorProtocol 14102; Corbett, Sydney, Australia) (n = 7708) and genotyped for the nucleotide changes that correspond to the amino acid substitutions C282Y and H63D in the HFE protein, using TaqMan (Applied Biosystems, Carlsbad, CA) real-time polymerase chain reaction (PCR) probes as previously described.7 Only samples from participants actively participating in the cohort who reported being born in Australia, the United Kingdom, Ireland, or New Zealand were processed. Participants born in southern Europe (Italy, Greece, or Malta) were excluded due to the lower prevalence of the HFE C282Y mutation in populations from that region.
A comprehensive active follow-up of MCCS participants began in 2003 and was completed in June 2007. Letters of invitation to participate in the HealthIron study were sent to a sample of 1438 participants that included all C282Y homozygotes identified in the MCCS (n = 203) and a stratified random sample of approximately equal numbers of participants from each of the other five HFE genotype groups. All participants gave written, informed consent to participate in both MCCS and the HealthIron study. Both study protocols were approved by the Human Research Ethics Committee of the Cancer Council of Victoria. Participants attending a study center completed a computer-assisted personal interview (that included questions on medical history, blood donation history, and venesection), provided a cheekbrush DNA sample for confirmatory HFE genotyping using real-time PCR assay with TaqMan probes (Applied Biosystems), and underwent a clinical examination of the abdomen and metacarpophalangeal (MCP) joints by study physicians blinded to HFE genotype. Blood samples were collected for measurement of iron indices, alanine aminotransferase (ALT) and aspartate aminotransferase (AST) concentrations using Roche automated assays (Roche Diagnostics, Indianapolis, IN) and were paired for analysis with stored baseline plasma samples for each participant. Blood samples were usually collected in the morning at both baseline and follow-up, and participants were requested to fast.
Definitions and Exclusion Criteria.
We defined sex-specific and menopause-specific SF upper limit of normal thresholds to be >300 μg/L for men and postmenopausal women and >200 μg/L for premenopausal women. We categorized participants according to their baseline SF concentration. Those below the threshold at baseline were defined as “normal SF” and those above the threshold but below 1000 μg/L at baseline were defined as “moderately elevated SF.”
We investigated the prevalence of eight outcomes (with examining physicians blinded to genotype) known or suspected from previous research to be associated with primary iron overload. These included: abnormality (bony spur, effusion, or tenderness) of the second and third MCP joints on either hand (MCP2/3), raised AST (>45 IU/L) concentration or raised ALT (>40 IU/L) concentration, a liver span of 13 cm or more (hepatomegaly), self-reported liver disease, self-reported fatigue, self-reported fatigue using the Modified Fatigue Impact Scale (MFIS),19 self-reported diabetes mellitus, and self-reported use of arthritis medication. The MFIS, a shortened version of the Fatigue Impact Scale,20 is a measure of self-reported fatigue based on 21 questions in three domains (physical, cognitive, and psychosocial) and scored on a scale of 0-84 (where a higher score indicates a greater impairment of daily activities due to fatigue). With the exception of diabetes mellitus and use of arthritis medication, which were recorded at baseline, all outcomes were measured at follow-up.
We excluded from the analyses those participants who had baseline SF concentrations >1000 μg/L, who had been diagnosed and treated for HH prior to baseline, or who were missing baseline SF concentrations and therefore could not be categorized. We also excluded participants with follow-up SF concentrations >1000 μg/L because current evidence suggests treatment should be recommended due to the high risk of irreversible cirrhosis.
Participants with neither the C282Y nor H63D mutation (referred to as “HFE wild-types”) were the control group for comparison with C282Y homozygotes. Participants from all other HFE genotype groups except C282Y homozygotes were excluded.
The prevalence of HH-associated signs and symptoms, stratified by sex, HFE genotype (C282Y homozygote or HFE wild-type), and normal/moderately elevated SF, was estimated as the observed proportion. Confidence intervals (CIs) for prevalence differences and P values for two-sample comparison of proportions were generated by assuming the normal approximation to the underlying binomial distribution to quantify sampling variability. For statistical analyses of SF concentrations, the values were (natural) log-transformed. SF concentrations were summarized using the geometric mean and were compared between groups by using the SF ratio, which is calculated by exponentiating the difference of the mean log SF values.21 Values for transferrin saturation and the MFIS were summarized using the sample mean and compared between groups using the two-sample t test.
One hundred sixty-one C282Y homozygotes (75 male and 86 female) and 336 HFE wild-types (153 male and 183 female) completed at least one of the following components of the HealthIron study: the HealthIron follow-up questionnaire, attendance at a follow-up clinic, or provision of a blood sample at either baseline or follow-up. Thirty-one C282Y homozygotes (27 male and 4 female) and one male HFE wild-type were excluded due to having baseline SF concentrations >1000 μg/L or being diagnosed and treated for HH prior to baseline. Data on these participants have been published previously.7 We further excluded 21 homozygotes (10 male and 11 female) and 38 HFE wild-types (17 male and 21 female) who were missing baseline SF concentration and five C282Y homozygotes (two male and three female) and one male HFE wild-type who had SF concentration >1000 μg/L at follow-up. After applying these exclusion criteria, 102 C282Y homozygotes (35 male and 67 female) and 291 HFE wild-types (131 male and 160 female) remained. Although data from those participants with missing baseline SF concentrations or SF concentrations >1000 μg/L at follow-up are included in Table 2 for completeness, they were removed for all comparative analyses of the prevalence of HH-associated signs and symptoms.
Not all participants contributed data for each outcome, explaining the variation in denominators for the calculation of prevalence statistics. The majority of participants completed the HealthIron follow-up questionnaire (143/161 [88%] C282Y homozygotes and 320/336 [95%] HFE wild-types) or provided a blood sample at follow-up (134/161 [83%] C282Y homozygotes and 309/336 [92%] HFE wild-types). A lower proportion attended the follow-up clinics (109/161 [68%] C282Y homozygotes and 260/336 [77%] HFE wild-types).
Participant Demographics and Health-Related Characteristics.
Summary statistics for age, body mass index, alcohol consumption, and blood donation at baseline are displayed in Table 1.
Table 1. Participant Demographics at Baseline as Per Inclusion Criteria for Prevalence of HH-Associated Signs and Symptoms Analysis*
|Age (years), mean (SD)||55.6 (9.2)||54.4 (8.3)||53.7 (9.2)||53.5 (9.1)|
|BMI (kg/m2), mean (SD)†||24.9 (2.9)||25.0 (3.9)||26.9 (4.2)||26.4 (4.8)|
|Alcohol (g/day), mean (SD)||12.6 (11.1)||6.1 (8.1)||21.1 (30.0)||7.0 (11.8)|
|Blood donation at baseline, number (%)|
| Never||18 (51%)||39 (58%)||60 (46%)||87 (54%)|
| Former||9 (26%)||18 (27%)||44 (33%)||42 (26%)|
| Current||8 (23%)||10 (15%)||27 (21%)||31 (20%)|
Table 2 presents sample sizes and the prevalence for five HH-associated signs and symptoms, stratified by HFE genotype, sex, baseline SF, and follow-up SF, including data from participants with missing baseline SF concentration. Although no formal analysis of the prevalence of HH-associated signs and symptoms for these participants was undertaken, expression in this group was low for both C282Y homozygotes and HFE wild-types, and it is unlikely that their exclusion would alter the conclusions of our analyses.
Table 2. Prevalence of HH-Associated Signs and Symptoms Categorized by HFE Genotype, Baseline Serum Ferritin (SF) Concentration, and Sex
|Treated before baseline†‡||4||2|| || || || || || || || || || || || || |
| || || || || || || || || || || || || || || || |
|SF>1000 μg/L||23||2|| || || || || || || || || || || || || |
| || || || || || || || || || || || || || || || |
|Moderately elevated||26||36||SF>1000 μg/L||2||2||0/2||0/0||0/2||1/2||1/2||1/2||0/2||0/0||0/2||0/2|
| || || ||SF<1000 μg/L||12||22||3/8||2/17||2/12||0/22||0/11||3/21||0/8||0/17||0/12||2/21|
| || || ||missing SF||4||4||0/0||0/0||0/0||0/0||0/1||0/2||0/0||0/0||0/1||0/2|
| || || ||treated§||||8||8||2/5||1/7||2/7||0/8||3/8||1/8||0/5||0/7||1/7||0/8|
| || || ||total||26||36||5/15||3/24||4/21||1/32||4/22||5/33||0/15||0/24||1/22||2/33|
| || || || || || || || || || || || || || || || |
| || || ||SF<1000 μg/L||6||24||1/4||3/20||0/6||3/24||0/6||5/22||1/4||1/20||0/6||3/22|
| || || ||missing SF||3||6||0/0||0/0||0/0||0/0||0/0||0/1||0/0||0/0||0/0||0/1|
| || || ||treated§||2||3||0/2||0/3||0/2||0/3||0/2||1/3||0/2||0/3||0/2||0/3|
| || || ||total||11||34||1/6||3/24||0/8||4/28||0/8||6/27||1/6||1/24||0/8||3/27|
| || || || || || || || || || || || || || || || |
|Missing SF¶||10||11||SF>1000 μg/L||4||1||2/4||0/1||2/4||0/1||1/4||0/1||0/3||0/1||1/4||0/1|
| || || ||SF<1000 μg/L||4||5||0/3||0/5||0/4||0/5||0/4||0/5||0/2||0/5||0/4||0/5|
| || || ||treated§||2||5||0/2||0/3||0/2||0/5||0/2||0/5||0/2||0/3||0/2||0/4|
| || || ||total||10||11||2/9||0/9||2/10||0/11||1/10||2/11||0/7||0/9||1/10||0/10|
|SF>1000 μg/L||1||0|| || || || || || || || || || || || || |
| || || || || || || || || || || || || || || || |
|Moderately elevated||32||4||SF>1000 μg/L||0||0|| || || || || || || || || || |
| || || ||SF<1000 μg/L||30||4||3/25||0/4||4/29||0/4||4/30||0/4||3/24||0/4||2/30||0/4|
| || || ||missing SF||2||0||0/0||0/0||0/0||0/0||0/1||0/0||0/0||0/0||0/1||0/0|
| || || ||total||32||4||3/25||0/4||4/29||0/4||4/31||0/4||3/24||0/4||2/31||0/4|
| || || || || || || || || || || || || || || || |
| || || ||SF<1000 μg/L||91||144||14/79||12/121||14/91||3/144||8/89||28/142||2/78||2/115||2/88||10/142|
| || || ||missing SF||8||12||0/0||0/1||0/1||0/0||0/5||2/8||0/0||0/1||0/5||0/7|
| || || ||total||100||156||14/80||12/122||14/93||3/144||9/95||30/150||2/79||2/116||2/94||10/149|
| || || || || || || || || || || || || || || || |
|Missing SF¶||17||21||SF>1000 μg/L||0||0|| || || || || || || || || || |
| || || ||SF<1000 μg/L||17||21||1/11||4/18||1/17||2/21||2/15||3/20||0/10||1/15||0/14||0/20|
| || || ||total||17||21||1/11||4/18||1/17||2/21||2/15||3/20||0/10||1/15||0/14||0/20|
Table 3 displays the prevalence of HH-associated signs and symptoms and summary measures of iron indices for participants with SF concentrations <1000 μg/L at baseline, stratified by sex and HFE genotype. Despite significantly higher mean SF and transferrin saturation in C282Y homozygotes compared with HFE wild-type controls, the prevalence of HH-associated signs and symptoms was similar in these two groups for both sexes with the exception of male C282Y homozygotes for whom the prevalence of abnormal MCP2/3 was increased compared with male HFE wild-types (32% versus 16%; prevalence difference = 16%; 95% CI = −7%, 37%; P = 0.12).
Table 3. Prevalence of HH-Associated Signs and Symptoms in C282Y Homozygotes and HFE Wild-Types with Serum Ferritin (SF) Concentration <1000 μg/L at Baseline, by Sex
|Geometric mean SF (95% CI)||315.7 (203.9, 488.9)||149.6 (127.7, 175.2)||2.1‡ (1.4, 3.1)||<0.001||141.0 (93.8, 212.0)||58.6 (49.8, 68.8)||2.4‡ (1.7, 3.5)||<0.001|
|Mean TS (95% CI)||63 (55.3, 70.9)||29.6 (27.9, 31.4)||33.4 (28.3, 38.7)||<0.001||51.1 (45.5, 56.6)||24.6 (23.3, 25.9)||26.5 (22.4, 30.5)||<0.001|
|MCP 2/3 prevalence (%)||6/19 (32%)||17/104 (16%)||16 (–7, 37)||0.117||6/47 (13%)||12/126 (10%)||3 (–8, 14)||0.534|
|Raised AST or ALT prevalence (%)||4/27 (15%)||18/121 (15%)||0 (–15, 15)||0.994||3/57 (5%)||3/148 (2%)||3 (–3, 9)||0.218|
|Hepato-megaly prevalence (%)||1/19 (5%)||5/102 (5%)||0 (–11, 11)||0.947||1/47 (2%)||2/120 (2%)||0 (–4, 5)||0.840|
|Liver disease prevalence (%)||1/28 (4%)||4/124 (3%)||0 (–7, 8)||0.926||5/57 (9%)||10/153 (7%)||2 (–6, 11)||0.576|
|Fatigue: self-report prevalence (%)||3/28 (11%)||12/125 (10%)||1 (–11, 14)||0.858||10/57 (18%)||30/154 (19%)||–2 (–14, 10)||0.750|
|Fatigue: MFIS mean (SD)||19.5 (13.4)||20.8 (15.2)||–1.3 (–7.7, 5.0)||0.685||25.5 (17.0)||26.6 (16.1)||–1.1 (–6.2, 4.0)||0.672|
|Diabetes prevalence (%)||1/35 (3%)||3/131 (2%)||1 (–6, 7)||0.846||0/67 (0%)||1/160 (1%)||–1 (–2, 1)||0.517|
|Arthritis medicine prevalence (%)||2/35 (6%)||2/131 (2%)||4 (–4, 12)||0.151||6/67 (9%)||12/160 (8%)||1 (–7, 9)||0.711|
Table 4 displays the prevalence of HH-associated signs and symptoms in C282Y homozygotes compared with HFE wild-types, stratified by baseline SF. There was little difference in the prevalence of HH-associated signs and symptoms for C282Y homozygotes compared with HFE wild-types, or for C282Y homozygotes with moderately elevated SF compared with those with normal SF. The two exceptions were abnormal MCP2/3, which occurred more frequently for C282Y homozygotes with moderately elevated SF than for HFE wild-types with moderately elevated SF (prevalence difference = 11%; 95% CI = −6%, 29%; P = 0.22) and hepatomegaly, which was less common for C282Y homozygotes than HFE wild-types (prevalence difference = −11%; 95% CI = −22%, 0%; P = 0.04). Similar results for MCP2/3 and hepatomegaly were observed when comparing C282Y homozygotes with moderately elevated SF to those homozygotes with normal SF.
Table 4. Prevalence of HH-Associated Signs and Symptoms in C282Y Homozygotes and HFE Wild-Types, by SF Concentration at Baseline
|Male (n)||24||32|| || ||11||99|| || || || |
|Female (n)||34||4|| || ||33||156|| || || || |
|Total (n)||58||36|| || ||44||255|| || || || |
|Signs and symptoms|| || || || || || || || || || |
|MCP 2/3 prevalence (%)||8/37 (22%)||3/29 (10%)||11 (–6, 29)||0.22||4/29 (14%)||26/201 (13%)||1 (–13, 14)||0.90||8 (–10, 26)||0.41|
|Raised AST or ALT prevalence (%)||4/49 (8%)||4/33 (12%)||–4 (–17, 10)||0.55||3/35 (9%)||17/236 (7%)||1 (–8, 11)||0.77||–1 (12, 12)||0.95|
|Hepatomegaly prevalence (%)||0/37 (0%)||3/28 (11%)||–11 (–22, 0)||0.04||2/29 (7%)||4/194 (2%)||5 (–5, 14)||0.13||–7 (–16, 2)||0.11|
|Liver disease prevalence (%)||3/51 (6%)||2/35 (6%)||0 (–10, 10)||0.97||3/34 (9%)||12/242 (5%)||4 (–6, 14)||0.35||–3 (–14, 9)||0.60|
|Fatigue: self-report prevalence (%)||7/51 (14%)||4/35 (11%)||2 (–12, 16)||0.75||6/34 (18%)||38/244 (16%)||2 (–12, 16)||0.76||–4 (–20, 12)||0.62|
|Fatigue: MFIS mean (SD)||22.6 (15.8)||23.1 (12.8)||–0.5 (–7.2, 6.2)||0.88||25.0 (16.6)||24.2 (16.3)||0.8 (–5.1, 6.7)||0.79||–2.4 (–9.7, 4.8)||0.51|
|Diabetes prevalence (%)||1/28 (2%)||1/36 (3%)||–1 (–7, 5)||0.71||0/44 (0%)||3/255 (1%)||–1 (–2, 0)||0.47||2 (–2, 5)||0.38|
|Arthritis medicine prevalence (%)||3/58 (5%)||1/36 (3%)||2 (–5, 10)||0.58||5/44 (11%)||13/255 (5%)||6 (–3, 16)||0.11||–6 (–17, 5)||0.25|
We conducted a sensitivity analysis, excluding participants with body mass index >30 kg/m2 or high alcohol consumption (>60 g/day for men and >40 g/day for women) (classified according to the Australian National Health and Medical Research Council guidelines)22 from the calculation of prevalence statistics for raised AST or ALT, hepatomegaly, and self-reported liver disease. This allowed us to assess the sensitivity of the results to these additional exclusion criteria, which are based on known risk factors for elevated liver enzymes and liver disease.
Exclusion of participants with heavy alcohol consumption and/or obesity changed the prevalence of raised AST or AST, hepatomegaly, and self-reported liver disease by less than 3% for each sex-specific and SF concentration-specific HFE genotype group.
We found little evidence that C282Y homozygotes with SF concentrations below 1000 μg/L at either baseline or follow-up 12 years later were at increased risk of HH-associated signs and symptoms compared with HFE wild-types, despite having, on average, significantly greater SF at baseline. Furthermore, C282Y homozygotes with moderately elevated SF concentrations were not at increased risk of HH-associated signs and symptoms compared with those C282Y homozygotes with normal SF concentrations at baseline, after an average of 12 years follow-up. Although we observed a higher prevalence of arthritis for male C282Y homozygotes compared with male HFE wild-types, the association remained when patients were stratified by SF concentration rather than sex, which suggests arthritis might occur independently of iron overload for C282Y homozygotes. This hypothesis is supported by the clinical observation that arthritis has often been present in patients for an extended period prior to diagnosis of HH,3, 23 and reports that it does not respond well to venesection treatment.3 However, the suggestion that the lack of treatment is causally related to the development of arthritis requires further scrutiny.
Our study has several strengths. It is the largest sample of C282Y homozygotes followed prospectively over a long period.24, 25 Data were collected with both physicians and participants blinded to HFE genotype, limiting recall bias. Data on modifying factors such as heavy alcohol consumption and obesity were also recorded prospectively, although, due to their low prevalence, we were unable to assess the extent to which they contributed to the prevalence of HH-associated signs and symptoms for C282Y homozygotes.
One limitation of our study is that the majority of participants were recruited after 45 years of age; therefore, our findings do not necessarily apply to younger C282Y homozygotes. However, previous population studies of hemochromatosis where the average age of participants was much younger have not found a high prevalence of disease.16 Moreover, the prevalence of C282Y homozygosity observed in our sample was larger than established estimates of this prevalence from large cross-sectional studies,2 a scenario that is unlikely if an appreciable fraction of eligible C282Y homozygotes declined to participate due to ill health. Data on the use of magnetic resonance imaging scanning or liver biopsies to quantify liver iron content were not collected systematically, and therefore we are unable to exclude the presence of cirrhosis or fibrosis. However, in a consecutive clinical series of 672 C282Y homozygotes, cirrhosis was not detected in any patient with SF < 1000 μg/L.10
Treated C282Y homozygotes were included in this study for completeness. We cannot infer that they were more or less likely to have HH-associated signs and symptoms. Some were ascertained through presentation with symptoms (and therefore more likely to have HH-associated signs and symptoms), but further data on the reasons for diagnosis are not available. Others were ascertained through cascade or other opportunistic screening and were asymptomatic. We note that one previous study that excluded treated C282Y homozygotes from the analysis concluded that most C282Y homozygotes do not develop iron overload–related disease.26 This approach is likely to have underestimated the prevalence of HH-associated signs and symptoms.27
The association between iron indices and the risk of HH-associated signs and symptoms has also been examined among community-recruited participants in the Hemochromatosis and Iron Overload Screening (HEIRS) study, which is the largest cross-sectional population-based study of iron indices in C282Y homozygotes to date. HEIRS assessed the prevalence of HH-associated signs and symptoms after participants were informed of both their iron and HFE genotype status, and the examining physicians were not blinded to genotype.8, 28 The HEIRS authors found that the prevalence of chronic fatigue and MCP2/3 was greater for C282Y homozygotes either previously diagnosed or newly diagnosed with any elevated SF, compared with HFE genotype controls. However, they did not stratify based on SF concentrations <1000 μg/L, as in the present study, and there were no longitudinal data on iron studies, so the results are not directly comparable with those presented here.
Our results raise the question as to whether C282Y homozygotes with SF concentrations <1000 μg/L should be managed aggressively or simply monitored to prevent SF rising over the critical threshold of 1000 μg/L. Further evaluation of the clinical benefits of therapeutic venesection should be undertaken to definitively confirm our suggestion that careful observation is a viable alternative to venesection therapy of such subjects. Ideally, a randomized controlled trial of phlebotomy versus a “wait and watch” approach for C282Y homozygotes with SF < 1000 μg/L would be mounted, although the follow-up period required for such a study to produce definitive results may be prohibitively long. If such a trial demonstrated that phlebotomy therapy was not superior, then the “wait and watch” approach would save many thousands of C282Y homozygotes worldwide from unnecessary venesection.
Dr. Sue Forrest from the Australian Genome Research Facility, Melbourne, supervised HFE genotyping of the cohort samples. Andrea A. Tesoriero with Dr. Melissa C. Southey (both at The University of Melbourne) supervised DNA extraction. Ashley Fletcher provided assistance with study coordination and sample retrieval. This study was made possible by the contribution of many people, including the original investigators and the diligent team who recruited the participants and who continue working on follow-up. We would like to express our gratitude to the many thousands of Melbourne residents who continue to participate in the study.