Introduction

Body iron because of its ability to induce oxidative stress has been considered to be contributing to the pathogenesis of coronary artery disease (CAD). A number of studies carried out to ascertain the relationship between body iron stores and the risk of CAD have yielded conflicting results. A systematic review of the literature on this subject revealed that five studies reported significant association between serum ferritin and coronary heart disease (CHD), while nine studies did not show any such relationship [1]. Most of these studies used serum ferritin as a marker of body iron stores. Determination of serum soluble transferrin receptor (sTfR) concentration along with serum ferritin provided a more reliable tool for estimation of body iron status. In this regard, the concentration ratio of sTfR to ferritin was considered to be “a state-of-the-art measurement of the body iron stores” [2]. This study carried out in Finland showed that sTfR to ferritin ratio was associated with initial acute myocardial infarction (AMI) [2].

CHD is the leading cause of death in the South Asian subcontinent [3]. The major factors contributing to the development of CHD at early age include dyslipidemia, physical inactivity, hyperhomocysteinemia and genetic, environmental and nutritional factors [3].

Few studies have been carried out to investigate the relationship of serum ferritin and risk of CHD in the South Asian region [4], [5]; however, none of them have focused on the entire spectrum of body iron stores employing the ratio of sTfR to ferritin and investigating its relationship with premature CAD. Therefore, the present study was undertaken to find out whether or not iron body stores are associated with the risk of premature AMI in a hospital-based Pakistani population. Since Pakistani urban population has been found to have high levels of blood lead (Pb) which is also associated with hyperhomocysteinemia [6], another objective of the study was to investigate if high blood Pb levels in this population would modulate the relationship (if any) between body iron stores and risk of premature AMI.

Methods

Results

Analysis of the data revealed that concentrations of serum vitamin B12, total cholesterol, LDL-cholesterol, HDL-cholesterol and triglycerides were significantly lower in AMI patients compared to age and gender-matched healthy controls (Table 1). However, serum ferritin levels and blood Pb levels were significantly higher in AMI patients as compared to controls (p value <0.001 and p value = 0.011, respectively). In addition, it was found that the ratio of sTfR to ferritin was significantly lower in AMI patients compared to healthy controls (p value <0.001, Table 1). Moreover, the ratio of sTfR to ferritin had a statistically significant correlation with serum HDL-cholesterol (r = 0.145; p value = 0.003). Table 2 shows serum/plasma concentrations of homocysteine, folate, vitamin B12, cholesterol, LDL-cholesterol, HDL-cholesterol, triglycerides, and blood Pb by quartiles of ratio of sTfR to ferritin. Significant dose-response relationship exists for mean values of serum HDL-cholesterol across quartiles of ratio of sTfR to ferritin (p value = 0.004). We also found some significant results for serum folate, total cholesterol and triglycerides; however, dose-response relationship was not observed for these biomarkers across quartiles of ratio of sTfR to ferritin.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographic characteristics, anthropometric measurements, and concentrations of biomarkers in patients with premature AMI and healthy controls.

https://doi.org/10.1371/journal.pone.0067981.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Serum/plasma concentrations of homocysteine, folate, vitamin B12, cholesterol, LDL-cholesterol, HDL-cholesterol, triglycerides and blood Pb by quartiles of ratio of sTfR to ferritin in Pakistani patients with premature AMI.

https://doi.org/10.1371/journal.pone.0067981.t002

In order to evaluate the risk of premature AMI across the quartiles of ratio of sTfR to ferritin, conditional logistic regression analysis was carried out while adjusting for vitamin B12 and HDL-cholesterol. Compared to quartile 4 (reference) with the highest ratio of sTfR to ferritin (low body iron stores), the risk of premature AMI increases with each low quartile and a clear cut dose-response relationship is observed from highest quartile (low body iron stores) to lowest quartile (high body iron stores) [OR, 3.29(95% CI, 1.54–7.03); p value for trend <0.01; Table 3]. The risk of AMI remained more than 3 fold in individuals in the lowest quartile of sTfR to ferritin when the model was also adjusted for blood Pb [p value for trend = 0.912; Table 3]. Since diabetes mellitus is a well-established risk factor for MI in South Asia [8], the above mentioned logistic regression model was also adjusted for individuals having serum glucose levels greater than 126 mg/dL. Even after this adjustment, the odds (2.47) of having premature AMI remained significant (p value for trend = 0.027; Table 3) with the lowest quartile. Similarly, low HDL-cholesterol (levels less than 35 mg/dL) is another major risk factor of AMI in Pakistani population [9]. When the interactions of low HDL-cholesterol and fasting serum glucose levels greater than 126 mg/dL with body iron status were examined, they were found to be nonsignificant in this regression model (p values for interactions = 0.41 and 0.46, respectively).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Risk of premature AMI in Pakistani adults with quartiles of ratio of sTfR to ferritin.

https://doi.org/10.1371/journal.pone.0067981.t003

Discussion

The relationship between body iron stores and risk of CHD, first proposed by Sullivan in 1981 [10], remained ignored until the results of Kuopio Ischemic Heart Disease Risk study (KIHD), carried out in Finland, became known to the scientific community [11]. In this study, a linear association between serum ferritin concentration and the increased risk of heart attack was shown in men [11], [12]. A strong epidemiological evidence regarding the role of body iron stores in carotid atherogenesis in Austrian men and women provided further support to the Sullivan hypothesis [13]. In a study carried out in Rotterdam, elevated levels of serum ferritin were found to be associated with the increased risk of MI in elderly population [14]. In a case control study involving 145 Indian men, serum ferritin levels were found to be directly correlated with the increased risk of AMI [4]. However, there were several other studies which found no association between body iron stores and risk of CHD [15]. For example, van der A DL et al. and Sun et al. reported lack of association between serum ferritin and CHD in women [16], [17]. Similarly, Sempos et al., Knuiman et al. and Galan et al. found no association between serum levels of ferritin and risk of CHD in both men and women [18]–[20]. It is important to mention that most of the studies examining the relationship of body iron stores and the risk of CHD used serum ferritin as the clinical marker of body iron status [4], [11]–[14], [16]–[23], however, in inflammatory conditions such as AMI, serum ferritin levels have been known to be elevated with concomitant decrease in serum iron [24], therefore, ferritin alone may not be an authentic clinical biomarker for iron status in AMI. On the other hand, sTfR as an index of tissue iron needs would not be influenced by the acute phase response, and perhaps, would provide better assessment of body iron status along with serum ferritin as a marker of tissue iron stores [25]. A ratio of sTfR to ferritin would, therefore, be a more reliable biomarker for body iron status even in an inflammatory condition, and therefore this study investigates the relationship of body iron stores with the risk of premature AMI.

Premature CAD is quite common in Pakistani population. A study from Karachi revealed that 16% of all the patients with AMI admitted to the Aga Khan University Hospital between January 2000–December 2002 were below the age of 45 years and 93% of them were men [26]. Besides genetic predisposition to developing CAD [27], dietary patterns have also been shown to be associated with the risk of AMI in this population [28], [29]. Western dietary pattern involving high consumption of red meat has been found to be associated with increased risk of AMI [28]. Since red meat is one of the major sources of body iron, it is conceivable that high intake of red meat by Pakistani population could be one of the factors contributing to the high body iron status and thus increased risk of AMI.

Pb pollution is a problem in this part of the world [30]. Body iron status can influence Pb absorption [31]. Therefore, we also measured blood Pb levels in cases and controls. AMI patients in this study had significantly high levels of blood Pb compared to controls which is in line with previous reports showing association between blood Pb levels and increased risk of cardiovascular disease [32]. However, when we adjusted our regression model for blood Pb levels, the risk of AMI remained similar to the risk when the adjustment was made for vitamin B12 and HDL-cholesterol. This indicates that body iron status in this population is associated with the risk of AMI independent of blood Pb levels.

According to the INTERHEART Study, the risk of MI among South Asians increases 2.48 folds in presence of diabetes mellitus [8]. In the present study, even adjusting for those individuals having fasting serum glucose levels greater than 126 mg/dL (a cut off level for diagnosis of diabetes mellitus), the risk of premature AMI remained nearly 2.5 fold high in those having highest iron stores of the body compared to those with lowest body iron stores. This also suggests that high body iron could be an independent risk factor for premature CAD.

Ideally, our controls should have been selected from the same communities to which AMI patients belonged. We recruited most of our controls from the personnel of various healthcare centers in Karachi; however, every attempt was made to pick those with socio-economic background similar to that of AMI patients. Both cases and controls belonged to the low income group as 90% of patients and 83% of controls had monthly house-hold income less than US $ 150.

There could be several reasons for increased body iron status in Pakistani AMI patients. Ninety-nine percent of Karachi population is nonvegetarian [33], therefore high consumption of foods rich in iron such as red meat, liver, kidney, heart, shellfish, clams, oyster, eggs could be contributing to high iron stores [34]. Moreover, foods cooked in cast iron pots and regular exposure to tobacco smoke must also be adding to the body iron [35]. Although our questionnaire did have information about consumption of red meat, poultry, fish and eggs, it did not include information about consumption of other food items rich in iron such as, shellfish, clams, oysters, nuts, seeds, or the nature of cooking pots and the extent of an individual’s exposure to tobacco smoke. This could be regarded as one of the limitations of this study. Tuomainen et al. have emphasized to include the information about users and nonusers of aspirin and antioxidants in studies concerning role of body iron stores in CHD because supplementation of these drugs could antagonize the risk-increasing effect of high iron stores [2]. Lack of information about the use of antioxidant vitamins and aspirin in the study population could be considered as another limitation of this study. Mean levels of total cholesterol, triglycerides and LDL-cholesterol were significantly lower in the study patients compared to controls; probably due to anti-lipid medications. Absence of information about the use of these medications is yet another limitation of this study. All of the blood samples from the patients were obtained within 24 hours of hospital admission. Since ferritin levels begin to increase on day 2 following AMI [36], it would be reasonable to assume that increased serum ferritin levels in AMI patients reflect body iron stores without any significant contribution from acute phase response. Measurement of serum levels of acute phase reactants in AMI patients could have been helpful, however, determination of markers of inflammation was not part of the study protocol and this would also remain a limitation of the present study. Despite these limitations, the present study did show an association of body iron stores with the risk of premature AMI in a Pakistani population. While such an association has been shown in adult Finnish men up to the age of 60 years [2], our study shows this relationship in relatively younger people (less than 45 years old) including both men and women. Our results, however, are different from those reported by Sun et al. who have shown lack of association between excessive body iron stores and CHD in female registered nurses in USA aged 30–55 years [17]. Similarly, in a German population serum concentrations of sTfR and ferritin were not found to be associated with CAD [37]. On the basis of these observations, it can be deduced that the relationship of body iron status with CHD could vary from one population to another. More data are, however, needed to further elucidate this relationship, especially in South Asian population which has the highest known rate of CAD [38], [39]. The mechanism by which high iron body status could be contributing to increased risk of AMI merits some discussion. It could be due to increased oxidation of LDL-cholesterol [21], increased oxidative damage [40], the promotion of lipid peroxidation, enhanced reperfusion injury and atherogenic properties [13], increased proliferation of vascular smooth muscle cells [41], or it could be because of the combination of all these along with other risk factors for CHD such as hypertension, inflammation and hyperhomocysteinemia which are highly prevalent in Pakistani population [3], [42].

Acknowledgments

We gratefully acknowledge the help of Ms. Dilshad NoorAli, Center of English Language, Aga Khan University for copyediting the manuscript. We are thankful to Mr. Iqbal Azam, Assistant Professor of Biostatistics, Aga Khan University for his help in study of interactions between body iron status and classical risk factors for AMI.