Hemochromatosis, iron toxicity and disease


  • L. R. Zacharski

    1. From the Department of Medicine, Dartmouth Hitchcock Medical Center, 1 Medical Center Drive, Lebanon, New Hampshire 03756 and the Research Service (151), VA Hospital, White River Junction, Vermont 05009, USA
    Search for more papers by this author

Leo R. Zacharski, MD, Research Service (151), VA Hospital, White River Junction, VT 05009, USA.
(fax: 802-296-6308; e-mail: leo.r.zacharski@dartmouth.edu).

In this issue of the Journal, Ellervik et al. reported a significant association between both the hereditary hemochromatosis (HFe) genotype (C282Y/C282Y) and extreme elevation of the transferrin saturation (>80%) either individually or combined, and essential hypertension defined as being on antihypertensive medication [1]. This association was discovered upon analysis of data from three populations observed for cardiovascular disease (CVD). They noted that both hypertension and deposition of iron in the myocardium with HFe impair cardiac function with increased thickness of the ventricular wall and dilation cardiomyopathy. Their hypothesis was that hypertension and HFe might, therefore, co-exist. They cited prior evidence for modulation of vascular tone by oxidative stress [2], a high prevalence of increased ferritin levels in males with essential hypertension [3] and regression of increased arterial wall thickness characteristic of presymptomatic HFe with iron depletion [4] in support of their hypothesis. Limitation of statistical associations to older men was considered due to increased penetrance of HFe in males whose iron burden rises earlier in life and to a greater extent with ageing compared to females whose iron levels are lower because of menstrual blood loss [5]. Lack of association with HFe heterozygosity was presumed to be due to lower body iron burden. An association with antihypertensive therapy but not single blood pressure readings was explained by the fact that subjects were under medical care that included diagnosis and treatment of hypertension. An advantage of the experimental design used in this study was that no assumptions were made on effects of treatment – of either hypertension or iron burden – on basic pathophysiologic relationships. A limitation of this study was the lack of data on ferritin levels that mark body iron levels and, indirectly, oxidative stress [6].

Previous studies have also shown increased levels of markers of oxidative stress [7, 8] and possible benefits of antioxidant nutrients [9] in hypertension. Tzoulaki et al. found correlations between blood pressure measurements and dietary heme iron – but not total dietary or nonheme iron content [10]. However, the possibility that iron-driven oxidative stress may contribute to the pathogenesis of hypertension is not well recognized and the term ‘hypertension’ does not appear in the index to the comprehensive treatise on HFe edited by Barton and Edwards [11].

These intriguing results invite exploration of the mechanisms linking CVD in general and hypertension in particular to iron status. Exaggerated oxidative stress resulting from unphysiologic iron levels is associated with laboratory and clinical features of both CVD and HFe [12, 13]. Under-representation of HFe in atherosclerosis, the development of which is dependent on lipid oxidation and macrophage foam cell formation, represents an apparent contradiction that has been explained in a recent review by Sullivan [14]. Levels of the key iron regulatory peptide, hepcidin, increase with inflammatory cytokine signalling or rising body iron burden. Hepcidin binds to the cell membrane iron exporter, ferroportin, leading to its internalization and degradation resulting in reduced iron transport across the intestinal endothelium and retention of iron within macrophages. Iron excess in wild-type individuals is associated with increased hepcidin expression that reduces intestinal iron transport, and promotes macrophage iron retention and foam cell formation necessary for atherogenesis. Paradoxically, HFe is characterized by loss of hepcidin expression in spite of massive iron accumulation – the opposite of the hepcidin response in wild-type (ambient) iron excess. Loss of feedback inhibition of iron absorption by hepcidin in HFe results in unrelenting intestinal iron absorption and massive iron accumulation whilst macrophages retain iron export function and inability to form foam cells. These details are relevant because the association between atherosclerosis and hypertension (discussed in their references 16, 37 and 38) but not HFe may be unnecessarily confusing. Whilst their pathogenesis may differ in detail, the common denominator between atherosclerosis, HFe and hypertension may be elevated body iron burden.

Studies in alternative populations should be informative. In the study cited [3], the mean ferritin level in normotensive controls was 146 ng mL−1, whilst the mean ferritin in patients with hypertension was 228 ng mL−1 (= 0.05). Eight hypertensive patients but no normotensive controls had ferritin levels in excess of 350 ng mL−1. Individuals homozygous for the C282Y mutation were excluded from analysis for unexplained reasons. However, 22.2% of hypertensive patients with relatively high ferritin levels (mean 541 ng mL−1) were compound heterozygotes for the C282Y and H63D HFe mutations compared to none in either hypertensive patients with normal ferritin levels (mean 145 ng mL−1) or normotensive controls (mean 146 ng mL−1) (< 0.01).

These data challenge interpretive skills. Perhaps iron excess, whether ambient (i.e., as a result of nutritional excess alone) or mutational (as a result of HFe), contributes to hypertension in general and the presence of the mutation signifies a sub-population with accelerated iron accumulation. Perhaps relatively low levels of iron are present in some patients with hypertension because of correspondingly low threshold sensitivity to iron-catalysed oxidative stress. Perhaps the mutation itself contributes to hypertension in some patients apart from effects on iron accumulation.

Precedent for such alternative interpretations exists in the literature that explains analogous variability in effects of iron on cancer risk. Syrjakoski et al. [15] reviewed evidence linking elevated body iron burden and the HFe genotype to risk of female breast cancer. In contrast, whilst the HFe genotype was not over represented in a cohort of male breast cancer patients, the risk of HFe genotype was increased twofold in the subset of male breast cancer cases that were also carriers of the common BRCA2 mutation 9346(-2) A-to-G compared to male breast cancer cases without the BRCA2 mutation. These investigators suggested that the HFe genotype might be a male breast cancer risk modifier gene amongst BRCA2 mutation carriers. Relatively, low levels of body iron may also contribute to female breast cancer in a subset of patients simultaneously deficient in compensatory protective antioxidant mechanisms [16]. The HFe genotype may also influence response of malignancy to chemotherapy [17].

It seems that the appropriate starting point for defining effects of iron burden on hypertension might be to test the hypothesis that iron reduction may ameliorate hypertension regardless of genotype or ferritin level. This approach was used to advantage in a prospective randomized trial of calibrated iron reduction in previously cancer-free patients with advanced peripheral arterial disease [13]. For practical reasons, entry criteria excluded individuals with ferritin levels >400 ng mL−1, which likely excluded most patients with HFe (patients were not genotyped). However, there was no floor entry ferritin requirement. Patients randomized to iron reduction were subjected to calibrated phlebotomy to reduce ferritin levels from baseline (mean ferritin 122 ng mL−1) to approximately 25 ng mL−1. This level was maintained by 6-monthly blood removal for the 6-year study duration.

Trends towards improved outcomes by age quartile were statistically significant for death plus nonfatal myocardial infarction and stroke (P for interaction = 0.004). A steep age-related benefit from iron reduction was observed in the youngest age quartile for both all cause mortality (= 0.02) and death plus nonfatal myocardial infarction and stroke (< 0.001). Age analysed as a continuous variable was significantly associated with improved outcome for both endpints (= 0.04 and 0.001, respectively) [13]. Iron reduction also resulted in a 37% lower risk of new cancer (= 0.025), reduced cancer risk on Kaplan–Meier (= 0.036) and Cumulative Incidence analysis (= 0.019), and reduced cancer-specific (= 0.003) and all-cause mortality in patients diagnosed with cancer (= 0.009) [18]. Ferritin levels below about 70 ng mL−1 seemed to be protective. Iron reduction is known to cause remission of arrhythmias in patients with HFe. However, remission of arrhythmia with iron reduction has recently been described in two patients with lesser degrees of ferritin elevation without HFe [19].

In summary, several factors identified in previous studies are potentially capable of contributing to the variable manifestations of iron-associated (ferrotoxic) disease. These factors, summarized in Fig. 1, provide guidance for future protocol design. Justification exists for extending findings of Ellervik and associates by means of controlled clinical trials that stand to clarify complex pathophysiology and, if positive, impact morbidity, mortality and cost of care. Ellervik and coauthors may have underestimated the potential public health significance of their findings [20].

Figure 1.

Factors that may contribute to the variable manifestations of iron-associated (ferrotoxic) disease.

Conflict of interest