The ability of iron to cycle reversibly between its ferrous and ferric oxidation states is essential for the biological functions of iron but may contribute to vascular injury through the generation of powerful oxidant species. We examined the association between chemical forms of iron that can participate in redox cycling, often referred to as “catalytic” or “labile” iron, and cardiovascular disease (CVD). In our cross-sectional study of 496 participants, 85 had CVD. Serum catalytic iron was measured using the bleomycin-detectable iron assay that detects biologically active iron. The odds of existing CVD for subjects in the upper third of catalytic iron were 10 times that of subjects with lower catalytic iron in unadjusted analyses. The association was decreased by 1/2 by age adjustment, but little additional attenuation occurred after adjusting for age, Framingham Risk Score, estimated glomerular filtration rate, hypertension status, high-density lipoprotein cholesterol, and systolic blood pressure, with the association remaining strong and significant (odds ratio 3.8, 95% confidence interval 1.4 to 10.1). In conclusion, we provide preliminary evidence for a strong detrimental association between high serum catalytic iron and CVD even after adjusting for several co-morbid conditions; however, broader prospective studies are needed to confirm these findings, which would support therapeutic trials to assess the beneficial effects of iron chelators on CVD.
Cardiomyopathy is known to occur in several iron overload states ; however, a role for iron in atherosclerotic cardiovascular disease (CVD) in the absence of iron overload is less clear. Although several in vitro and animal studies have supported a role for iron in atherosclerosis, human observational studies have provided inconsistent results. Several factors, including the fact that total body iron is not reliably related to the level of biologically active iron, may have contributed to these inconsistencies. In the present study we evaluated groups of patients with several long-term conditions to assess the association between serum catalytic iron and CVD after controlling for various co-morbidities.
Methods
This cross-sectional study was approved by the institutional ethics committees at the participating centers, and written informed consent was obtained from all study participants. The 568 subjects who agreed to participate included 349 participants from a survey of healthy governmental workers, 147 patients with chronic kidney disease (CKD) who attended Muljibhai Patel Urological Hospital, and 72 patients with angiographically established stable coronary artery disease who attended cardiology clinics at Bhaila Amin General Hospital. To adjust for the Framingham 10-year coronary heart disease (CHD) risk score, subjects <20 years of age and those ≥80 years of age were excluded. Given the potential impact of hemodialysis on laboratory values, we excluded 58 patients on hemodialysis, leaving 496 participants for analyses.
All subjects underwent a detailed clinical evaluation from July 2007 through June 2008 at the Muljibhai Patel Urological Hospital, where risk factor profiles and medical histories including medical diagnoses and tobacco use were carefully recorded. Body mass index was calculated as weight (kilograms) divided by height (meters) squared. With the participant in a seated position ≥3 blood pressure measurements were taken 5 to 10 minutes apart and the average blood pressure was used in these analyses. Hypertension was defined as blood pressure >140/90 mm Hg, self-reported diagnosis of hypertension, or use of antihypertensive medications. Diabetes was diagnosed based on American Diabetes Association criteria or current use of oral hypoglycemic agents or insulin. Stable coronary artery disease was defined as patients with angiographically established coronary disease excluding those with effort angina, unstable angina, and acute myocardial infarction in the previous 3 months. Electrocardiography was performed in all participants to evaluate for changes that would suggest ischemic heart disease. In addition to stable obstructive coronary artery disease, patients from any site with previous angina, chronic stable angina, or previous myocardial infarction were included as subjects with CVD. CKD was defined from National Kidney Foundation KDOQI guidelines (2006) based on estimated glomerular filtration rate (eGFR), proteinuria, and structural damage. For men eGFR was calculated by the Cockcroft-Gault equation: eGFR = (140 – age) × weight (kilograms)/72 × serum creatinine (milligrams per deciliter); for women the value was multiplied by 0.85. Framingham Risk Score was defined based on the National Cholesterol Education Program definition of 10-year CHD risk, where any tobacco use was substituted for smoking.
Participants had blood drawn after fasting for ≥12 hours. Measurements of blood glucose, lipid profiles, serum ferritin, and high-sensitivity C-reactive protein were performed on a fully automated biochemistry analyzer. Serum creatinine was measured by the Jaffe kinetic reaction with alkaline picric acid using a kit prepared by Erba Diagnostics Mannheim (Mumbai, India).
Catalytic iron was measured from serum using the bleomycin-detectable iron assay as described previously. Briefly, the assay uses bleomycin, which binds to and degrades DNA in the presence of labile iron, forming a product that reacts with thiobarbituric acid to form a chromogen. To avoid external iron contamination, reactions were carried out in disposable polypropylene tubes and all soluble reagents except bleomycin were treated overnight with Chelex (Bio-Rad Laboratories India Pvt. Ltd., Mumbai, India) (300 mg for a 10-mL solution). Intra-assay coefficients of variation for serum bleomycin-detectable iron (catalytic iron) for low (mean 0.04 μmol/L), medium (mean 0.57 μmol/L), and high (mean 3.15 μmol/L) levels of catalytic iron were 8.1%, 8.8%, and 4.0%, respectively, whereas interassay coefficients of variation were 13.0%, 14.9%, and 10.7%, respectively.
Serum catalytic iron was log-transformed or categorized into thirds using SAS (SAS Institute, Cary, North Carolina) quantiles because of markedly skewed distribution. Subjects in the lowest 2/3 of catalytic iron formed the reference group because of the very low CVD prevalence in the lowest third. Boxplots depicted the distribution of serum catalytic iron on the log scale for the overall sample and for participants with any and those without all the co-morbid conditions in the present study (diabetes mellitus, CKD, CVD, and hypertension). Differences in clinical characteristics for participants in the upper 1/3 versus the lower 2/3 of catalytic iron were tested with chi-square tests (categorical variables) and with Wilcoxon rank-sum tests (continuous variables). Logistic regression models tested associations between serum catalytic iron and CVD. To control for important clinical characteristics, multivariable models were developed using purposeful selection with a statistical significance criterion of a p value ≤0.05 (Wald test for covariates and deviance test for interactions). Covariates were tested for confounding and were retained if they produced an ≥15% change in the odds ratio (OR) for level of catalytic iron. Because of marked correlation with major disease categories and risk factors, center effects could not be tested in logistic models. Possible interactions between serum catalytic iron and major clinical characteristics identified initially by Breslow-Day tests of homogeneity were evaluated by the deviance test in the fully adjusted logistic model. SAS 9.2 was used in all analyses.
Results
This middle-aged population of mostly men had several co-morbid conditions ( Table 1 ). Prevalences of CVD for the highest to lowest thirds of catalytic iron were 39%, 8.6%, and 0.9%. Therefore, to have sufficient events in the reference group, we combined subjects within the middle and lower thirds of catalytic iron. Several significant detrimental associations were found for subjects in the upper third of serum catalytic iron (range 0.14 to 16.7 μmol/L) compared to lower levels (range 0.01 to 0.13 μmol/L): they were older and had higher systolic blood pressure, abnormal lipid levels (lower high-density lipoprotein, higher low-density lipoprotein, and higher triglycerides), higher high-sensitivity C-reactive protein, higher serum creatinine, higher ferritin, lower eGFR, and higher prevalences of co-morbidities (CVD, CKD, diabetes mellitus, hypertension, and obesity; Table 1 ). No significant differences in gender or prevalence of tobacco use were found for level of catalytic iron. However, the limited numbers of women could have prevented identification of gender differences.
