HFE gene mutations increase the risk of coronary heart disease in women
- First Online:
- Cite this article as:
- Pardo Silva, M.C., Njajou, O.T., Alizadeh, B.Z. et al. Eur J Epidemiol (2010) 25: 643. doi:10.1007/s10654-010-9489-6
- 385 Downloads
The purpose of the present study is to examine HFE gene mutations in relation to newly diagnosed (incident) coronary heart disease (CHD). In a population-based follow-up study of 7,983 individuals aged 55 years and older, we compared the risk of incident CHD between HFE carriers and non-carriers, overall and stratified by sex and smoking status. HFE mutations were significantly associated with an increased risk of incident CHD in women but not in men (hazard ratio [HR] for women = 1.7, 95% confidence interval [CI] 1.2–2.4 versus HR for men = 0.9, 95% CI 0.7–1.2). This increased CHD risk associated with HFE mutations in women was statistically significant in never smokers (HR = 1.8, 95% CI 1.1–2.8) and current smokers (HR = 3.1, 95% CI 1.4–7.1), but not in former smokers (HR = 1.3, 95% CI 0.7–2.4). HFE mutations are associated with increased risk of incident CHD in women.
KeywordsHFE mutationHemochromatosisCoronary heart diseaseSmokingGender
Hereditary hemochromatosis is a genetic disorder characterized by iron overload . Iron overload in patients with hereditary hemochromatosis results in iron depots in the pancreas, liver, joints and heart. In the heart, such depots have been associated with coronary heart disease (CHD) and shortened life expectancy . The most common cause of hereditary hemochromatosis are two common mutations in the HFE gene, C282Y and H63D [3, 4]. Over 80% of the patients with hereditary hemochromatosis are homozygous for the C282Y mutation, about 1% are homozygous for the H63D mutation and about 7% are compound heterozygous . In populations of European origin, an estimated 0.4% of the population is homozygous and 9% is heterozygous carrier of the C282Y mutation and 13% is homozygous and 2% heterozygous for the H63D mutation .
Complications of iron overload may not only occur to homozygous carriers of the risk alleles. Compound heterozygotes and heterozygotes for the C282Y and H63D mutations also have a subtle increases in serum iron, serum ferritin and transferrin saturation [5–8]. These slight changes suggest that heterozygotes are more likely to have a slow accumulation of iron, which may lead to pathology later in life. Iron deposits in the arterial wall trigger the low density lipoprotein cholesterol peroxidation and therefore contributes the formation of atherosclerotic lesions and to the inflammation leading to cardiovascular disease [1, 2]. Some studies also showed that C282Y carriers had an increased risk of myocardial infarction , and coronary heart disease (CHD) . Similarly, smoking has also been associated to cardiovascular disease through increased inflammation, thrombosis, and oxidation of low-density lipoprotein cholesterol . Therefore, to properly disentangle the association between HFE mutations and cardiovascular disease requires taking into account the effect of smoking. Following this, previous work of our group also showed an increased risk of stroke among HFE carriers who smoked . The aim of the present study was to examine the effect of the two common HFE C282Y and H63D mutations on incident CHD accounting for the effect of smoking. Since women are protected early in life from iron-related pathology by menstruation, we examined men and women separately.
The present analysis was performed within the Rotterdam Study, an ongoing population-based study on the determinants of disease and disability in 7,983 subjects aged 55 years and older. Design, rationale and details of the study have been described previously . The present analyses used baseline data (collected between 1990 and 1993) and follow-up morbidity and mortality data (collected until December 2001; mean follow-up of 8.3 years, standard deviation (SD) 2.7 years). The Medical Ethics Committee of the Erasmus MC University Medical Center approved the study protocol and all participants provided written informed consent.
Baseline data collection
During home visits, a trained interviewer obtained information on health status, medical history, medication use and smoking status at baseline. Subsequently, participants were invited to the study center where they underwent an extensive clinical examination in which height, weight, systolic and diastolic blood pressures were measured. Body mass index (BMI) was computed as weight (kg) divided by height squared (m2). Systolic and diastolic blood pressures were measured twice in sitting position, after 5 min rest, using a random-zero sphygmomanometer. The mean of the two measurements was used for the analysis. Hypertension was defined as systolic blood pressure higher than 160 mm Hg, diastolic blood pressure higher than 100 mm Hg or the use of medication indicated to treat high blood pressure (hypertension grades 2 and 3) . Serum glucose, total cholesterol, high-density lipoproteins (HDL) cholesterol and C-reactive protein levels, as a marker of inflammation, were determined using an automated enzymatic procedure [11, 12]. Diabetes was diagnosed based on the use of medication, and/or a random or post-load glucose levels higher than 11.1 mmol/L . Iron and ferritin levels and transferrin saturation were determined in subgroup as previously described .
During follow-up, information on fatal and non-fatal cardiovascular endpoints was obtained from the general practitioners and hospital records. Two research physicians and a cardiologist independently reviewed all information and classified all the events according to the International Classification of Diseases, 10th edition (ICD-10) . Incident CHD was defined as the occurrence of non-fatal myocardial infarction (ICD-10 code I21), revascularization procedure (percutaneous transluminal coronary angioplasty or coronary artery bypass graft) and cardiac death. Cardiac death was defined as death caused by myocardial infarction or other ischemic heart disease (ICD-10 codes I20–I25), sudden cardiac death (ICD-10 code I46), cardiac arrhythmias (ICD-10 code I49), sudden death undefined (ICD-10 code R96) or death from heart failure (ICD-10 code I50).
Genotyping of HFE mutations
C282Y and H63D mutations were genotyped in a random sample of 3,798 individuals as described in details previously . Baseline characteristics did not differ between randomly selected and non-selected participants.
From the 3,798 individuals who had C282Y and H63D genotype available, 3,435 had complete information on cardiovascular risk factors at baseline and were included in the analyses. There were only 10 homozygotes for the C282Y and 80 homozygotes for the H63D mutation among which 2 and 11 cases of incident CHD, respectively. Previous studies on the HFE mutations and iron parameters showed that serum iron, serum ferritin and transferrin saturation were similar for C282Y and H63D heterozygotes [5, 6]. Therefore, we pooled carriers of C282Y or H63D mutations as HFE carriers. Differences in baseline characteristics between the HFE carriers and non-carriers were tested using Chi-squared statistics (for categorical variables) or t tests (for continuous variables). Variables that were not normally distributed, are presented as median with interquartile range and analyzed using the Mann–Whitney U test. CHD incidence rates were calculated as number of events per 1,000 person-years. Risks of incident CHD were quantified as hazard ratios (HRs) from Cox proportional hazards models using age as time scale . Survival time was calculated from age at baseline to age at event. HRs were calculated crude and adjusted for sex, smoking status, hypertension, BMI, total serum cholesterol, HDL cholesterol, diabetes mellitus and CHD at baseline. Further, HRs were additionally adjusted for C reactive protein, to evaluate the role of inflammation in the relationship between HFE mutations and incident CHD. The association between HFE mutations and the risk of incident CHD was investigated overall and in sex and smoking subgroups given that both are associated with CHD and the clinical manifestations of iron overload in our study . Proportionality of all models was tested by Schoenfeld residuals . P-values lower then 0.05 were considered statistically significant.
Baseline characteristics by sex and HFE carrier status
HFE-carriers (n = 656)
Non-carriers (n = 1,174)
HFE-carriers (n = 617)
Non-carriers (n = 988)
Age at entry (years)
Systolic blood pressure (mm Hg)
Diastolic blood pressure (mm Hg)
Body mass index (kg/m2)
Total serum cholesterol (mmol/L)
HDL cholesterol (mmol/L)
C-reactive protein (mg/L)a
Myocardial infarction at baseline
Iron parameters by sex and HFE carrier status
HFE-carriers (n = 37)
Non-carriers (n = 130)
HFE-carriers (n = 35)
Non-carriers (n = 124)
Transferrin saturation (%)
Incidence rates and hazard ratios for incident coronary heart disease by sex and HFE carrier status
Hazard ratio (95% CI)
11.8 (10.2 to 13.7)
13.6 (10.8 to 17.1)
1.3 (1.0 to 1.8)
1.7 (1.2 to 2.4)
10.8 (8.9 to 13.1)
25.3 (22.6 to 28.3)
24.5 (20.4 to 29.4)
0.9 (0.7 to 1.2)
0.9 (0.7 to 1.2)
25.8 (22.4 to 29.8)
Incidence rates and hazard ratios for incident coronary heart disease by sex, smoking status and HFE carrier status
Incidence rate (CHD/1,000 py)
Hazard ratio (95% CI)
We have found that the HFE mutations are associated with higher risk of incident CHD in women. C-reactive protein, a strong predictor of CHD risk was not associated with HFE mutations in women with current or former smoking history. Our study was embedded in a large follow-up study of Caucasian individuals aged 55 and older. This long follow-up allows us to study outcomes that are the result of lifelong exposures and express at older ages like cardiovascular disease associated with iron overload.
A previous report based on a smaller sample of the Rotterdam study (n = 342) showed that HFE genotypes, also those who were heterozygous, were associated with higher levels of transferrin saturation and serum iron levels, albeit that levels in those homozygous for the mutations are exponentially higher . Previous work of our group also showed an increased risk of stroke among HFE carriers who smoked . Altogether, these results are in line with the mechanism linking HFE mutations and smoking to cardio- and cerebrovascular disease through damage to vessel wall and inflammation . The underlying mechanism linking HFE and CHD is thought to be through iron-related oxidative stress and the subsequent damage to vessel wall and inflammation [15, 18]. Such mechanism is also supported by endothelial dysfunction observed in normal subjects after receiving intravenous iron sucrose reported recently , as well as the positive association between markers of oxidative stress and iron therapy . Despite the solid grounds for this association, our data showed that CRP levels were not an intermediate factor in this association. A previous study on the association of chronic inflammation and hemochromatosis phenotype revealed that CRP levels did not change between homozygotes for the C282Y mutations with high or low iron stores . Most likely, blood CRP level is a biomarker for generalized inflammation not specific to the association HFE and CHD. How and through which early intermediate inflammatory pathways HFE is related to inflammation, remains to be determined.
Another issue to be explained is the fact we only found significant evidence for association of HFE and CHD in women and not in men. While there is an ongoing discussion of the validity of subgroup testing by sex in genetic studies , there are reasons to separate men and women for HFE. First, phlebotomy is the key strategy for prevention of hemochromatosis and women are therefore naturally protected due to their menstruation . While this is a common explanation for sex-differences in CHD risk associated with HFE mutations, in our study an increased risk was found in female carriers rather than in male carriers. This may be related to the study design. We studied elderly people and therefore male carriers may have been selected out from the population due to early mortality related to HFE. In women, the gene effect is likely to express clinically later as they are protected by menstruation up to menopause. Second, it also should be noted that the major risks factors for CHD are smoking, body mass index, diabetes and lipid profile as supported by the effect of these co-variates when adjusting HR for CHD in women and men. The extent of CHD risk conferred by these factors is of such magnitude that the genetic predisposition associated with HFE carrier status is rapidly outweighed as seen in Table 4 where it is shown that greater CHD risk is associated with smoking status rather than with HFE mutations. Given the higher prevalence of smoking (current and former) among men compared to women at baseline, it is likely that the modest effect of HFE mutations on the risk of CHD is only apparent among women.
Because of their role in iron overload and their impact on CHD risk, HFE mutations have been of interest in public health [7, 24–29], For long, they have been viewed as a model for genetic testing. A genetic-based strategy for prevention of iron overload and its complications has been investigated after the identification of the two common mutations in the HFE gene . The small effect size, the lack of consistency in the association between HFE mutations and cardiovascular disease and mortality [31–44], and the possible effect modification by environmental factors such as smoking [15, 31–35] implies that the benefit of such screening program varies between different subgroups, imposing further difficulties to the use of genetic testing for population-based screening. The results of our study confirm these previous findings and do not support genetic screening on HFE mutations.
The Rotterdam Study is supported by the Erasmus Medical Center and Erasmus University Rotterdam, the Netherlands Organization for Scientific Research (NWO), the Netherlands Organization for Health Research and Development (ZonMw), the Research Institute for Diseases in the Elderly (RIDE), the Ministry of Education, Culture and Science, the Ministry of Health, Welfare and Sports, the European Commission (DGXII), the Municipality of Rotterdam, and the Centre for Medical Systems Biology (CMSB). This work was supported by an Erasmus University Rotterdam grant to Dr. Omer T. Njajou and by a ZonMW grant to Dr. Behrooz Z. Alizadeh. The contributions of the general practitioners and pharmacists of the Ommoord district to the Rotterdam Study are greatly acknowledged. We are grateful to JM Vergeer, JR Oskamp and WM Luijten for their help in the genotyping.
This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.