Blood donation, body iron status and carotid intima-media thickness

Transcription

1 Atherosclerosis 196 (2008) Blood donation, body iron status and carotid intima-media thickness Mariëlle F. Engberink a,, Johanna M. Geleijnse a, Jane Durga a,b, Dorine W. Swinkels c, Wim L.A.M. de Kort d, Evert G. Schouten a, Petra Verhoef a,b a Wageningen University, Division of Human Nutrition, Wageningen, The Netherlands b Wageningen Centre for Food Sciences, 1 Wageningen, The Netherlands c Department of Clinical Chemistry, Radboud University, Nijmegen Medical Centre Nijmegen, The Netherlands d Sanquin Blood Bank, Southeast Region Nijmegen, The Netherlands Received 8 September 2006; received in revised form 25 January 2007; accepted 29 January 2007 Available online 6 March 2007 Abstract Iron could promote free radical formation, which may lead to injury of the arterial wall and atherosclerosis. Blood donation may reduce cardiovascular risk by lowering body iron status. We collected data on blood donation history and intima-media thickness of the common carotid artery (CIMT) in 819 subjects (50 70 years), who were recruited from municipal and blood bank registries in The Netherlands. Serum iron parameters were assessed, including non-transferrin bound iron (NTBI) that has recently been found in conditions of iron overload. Serum ferritin was lower in current donors (n = 443; 44 g/l) than in ex-donors (n = 120; 114 g/l) and never-donors (n = 256; 124 g/l, P for trend <0.001). For NTBI, values were 2.33, 2.54, and 2.51 mol/l, respectively (P < 0.05). CIMT was slightly reduced in frequent donors (i.e., 49 times during life or 2 times per year), although not statistically significant. CIMT was not significantly related to NTBI. Frequent blood donation, resulting in lowered body iron, might give some protection against accelerated atherosclerosis Elsevier Ireland Ltd. All rights reserved. Keywords: Blood donation; Body iron status; Non-transferrin bound iron; Atherosclerosis; Carotid intima-media thickness 1. Introduction In 1981, Sullivan proposed that menstrual loss of iron protects women against ischemic heart disease, and that higher levels of stored iron in men and postmenopausal women may explain their higher incidence of heart disease [1]. In line with this iron hypothesis, a number of epidemiological studies have shown a positive relationship between body iron and cardiovascular disease (CVD) [2 4]. Blood donation, an effective way to lower body iron, has been linked to a reduced risk of CVD [5 8]. Other studies of CVD, however, failed to show an association with iron status or blood donation [9 13]. Corresponding author at: P.O. Box 8129, 6700 EV Wageningen, The Netherlands. Tel.: ; fax: address: marielle.engberink@wur.nl (M.F. Engberink). 1 The Wageningen Centre for Food Sciences is an alliance of major Dutch food industries, Maastricht University, TNO Nutrition and Food Research in Zeist, and Wageningen University and Research Centre, with financial support by the Dutch Government. Iron status has also been associated with markers of vascular function and atherosclerosis. Recently, Zheng et al. reported that volunteer blood donors with a high donation frequency had significantly greater flow-mediated dilation than low frequency donors [14]. In a study by Kiechl et al., iron stores were related to progression of carotid atherosclerosis [15], although this could not be confirmed by others [16,17]. Discrepant findings may partly be explained by the choice of iron parameters. Serum ferritin can readily be assessed and is often used in epidemiological studies [18]. It has been shown that one blood donation per year lowers serum ferritin by almost 50% and more frequent donations are associated with further decreases [19]. Lee and Jacobs, however, postulated that common markers of stored body iron, such as ferritin, transferrin saturation, iron, or total iron binding capacity, are inappropriate for examining harmful effects related to iron overload [20]. Free iron, as reflected by reactive non-transferrin bound iron (NTBI), is more likely to be involved in the atherosclerotic process [20]. It has been found /$ see front matter 2007 Elsevier Ireland Ltd. All rights reserved. doi: /j.atherosclerosis

2 M.F. Engberink et al. / Atherosclerosis 196 (2008) that free iron not only exists in conditions of iron overload (such as hemochromatosis), but that NTBI is also present in serum with incomplete transferrin saturation [21,22]. NTBI could promote atherosclerosis by catalyzing the formation of highly reactive oxygen species, thereby enhancing lipid peroxidation [23]. Recently, Van der A et al. examined the relation between tertiles of NTBI and risk of coronary heart disease and acute myocardial infarction in a populationbased sample of 11,471 postmenopausal women, but found no excess risk within the highest NTBI tertile compared with the lowest [24]. The effect of blood donation on NTBI has not yet been established. We collected data on blood donation history and serum iron parameters, including NTBI, in healthy subjects aged years. Both blood donation and iron parameters were studied in relation to intima-media thickness of the common carotid artery (CIMT), a marker of atherosclerosis and a surrogate marker of vascular disease in the general population. 2. Subjects and methods 2.1. Study population We used baseline data of 819 men and postmenopausal women who participated in the folic acid and carotid intimamedia thickness (FACIT) study. The FACIT study is a 3-year, randomized trial of folic acid supplementation and atherosclerosis [25]. Subjects were recruited from a blood bank registry (Sanquin Bloodbank) and municipal registries in the eastern part of The Netherlands. Inclusion criteria were age years; postmenopausal state; homocysteine levels mol/l; no renal, thyroid or intestinal disorders; and no use of B-vitamin supplements, lipid-lowering drugs or hormone replacement therapy. The Medical Ethics Committee of Wageningen University approved the present study and all subjects gave written informed consent Data collection Data on blood donation history in general, self-reported medical history, including family history of premature vascular disease (onset <60 years in first degree family), smoking habits and other lifestyle factors were obtained by questionnaires that were reviewed by a trained research assistant. More detailed information on blood donation history, like type of donation (whole blood versus plasma), first and last occasion of donation (years), total number of donations, and reason for switching from whole blood to plasma donation were obtained by an additional questionnaire. Height and weight were measured with light indoor clothing without shoes, and body mass index (kilogram per meter squared) was calculated. Blood pressure was measured using an automated device (Dinamap Compact Pro 100, General Electric). The average of eight blood pressure measurements was used for the analysis. Double data entry was performed Laboratory measurements Venipuncture was performed after an overnight fast at least 6 weeks after the last blood donation. Blood was immediately processed and aliquots were stored at 80 C until determination. Total serum iron and total iron binding capacity were measured on a Hitachi 747 analyzer (Roche Diagnostics). Transferrin saturation is expressed as the ratio ( 100%) of total serum iron concentration and total iron binding capacity. Serum ferritin levels were determined on the Immulite 1 of DPC (Diagnostic Product Corporation) using a two-site immunometric assay (references values: men g/l, and postmenopausal women g/l, respectively). Non-transferrin bound iron was measured by a fluoresencebased one step chelation method [26]. C-reactive protein (hcrp) was determined with ELISA using polyclonal antibodies (Dako, Glastrup, Denmark). Serum lipids were determined using Hitachi 747 (Roche Diagnostics). Intra-assay and inter-assay variation coefficients were all below 15% Carotid ultrasound High-resolution B-mode ultrasonography was performed using a 7.5 MHz linear-array transducer (ATL Ultramark IX). In a dark quiet room, the subject was in supine position with his head tilted 45 in the direction opposite of the carotid being measured. Longitudinal images of the distal common carotid arteries were obtained at four predefined angles of 30 steps ( on the right side and on the left side). Images were frozen on the top of the R-wave of the electrocardiogram and recorded on videotape. The mean and mean of the maximum distance of the near and far wall of the distal 10 mm to the bifurcation of the right and left common carotid arteries was determined using an automated edge detection program. Ultrasound examination was performed in duplicate within three weeks; the average was used for analyses. The same sonographer performed both examinations in 91% of the subjects. A single reader interpreted all images. The mean difference (±S.D.) between sonographers was 0.09 mm (±0.10 mm) for maximum carotid intimamedia thickness and 0.05 mm (±0.04 mm) for mean carotid intima-media thickness. The intraclass correlation coefficient for maximum carotid intima-media thickness was 0.90 for sonographer 1 and 0.88 for sonographer 2. The intraclass correlation coefficient for mean carotid intima-media thickness was 0.96 for both sonographers. Both sonographers and the reader were blinded to participants study information Statistical analysis Data analysis was performed using SPSS 11.0 for Windows. Mean and maximum CIMT were examined in categories of iron parameters, donor status, lifetime number of blood donations, and donation frequency using analysis of covariance (GLM). Because there was a non-linear relationship between CIMT and iron parameters and women

3 858 M.F. Engberink et al. / Atherosclerosis 196 (2008) have lower body iron levels than men, sex-specific quartiles of iron parameters were used in the analysis. In multivariate analyses, model 1 included age and sex as a covariates, model 2 included age, sex, body mass index (BMI), C-reactive protein (CRP), smoking status (never, past, current), and pack-years of smoking, whereas model 3 also included LDL and HDL cholesterol. Statistical significance was defined as a two-sided P-value < P-values for linear trend over the categories of iron parameters and lifetime blood donation were obtained by entering median values within quartiles into the linear regression models. The present study had a power of 81% for detecting a 0.03 mm difference in CIMT between current and never-donors assuming a standard deviation of 0.13, with α = 0.05 (two-sided). 3. Results Of 819 subjects, 256 (31%) had never donated blood, 120 (15%) were ex-donors and 443 (54%) were current donors. Among donors, 90% exclusively donated whole blood and 10% donated plasma, mostly after a long history of whole blood donation. Overall, mean CIMT was 0.83 mm (range: ), and maximum CIMT was 1.02 mm (range: ). Serum ferritin concentration, which showed a skewed distribution, had a median value of 67 g/l (range: g/l). Mean NTBI was 2.42 mol/l (range: mol/l). Characteristics of donor status groups were essentially similar (Table 1), except for a higher percentage of selfreported CVD and diabetes mellitus among ex-donors, which may have been a reason for discontinuing blood donation. Serum iron parameters differed according to blood donation status (Table 1). As expected, ferritin, transferrin saturation and NTBI were all reduced in current donors whereas total iron binding capacity was increased. In ex-donors, only serum ferritin was reduced compared to never-donors. Table 2 presents mean CIMT and maximum CIMT by donor status. CIMT was increased in ex-donors compared to the other groups, but this difference was no longer statistically significant after adjustment for age and sex. Further adjustment for confounders and the exclusion of 116 subjects with self-reported CVD or diabetes mellitus did not change the results (Table 2). In addition, no significant linear trend was observed between the lifetime number of blood donations and mean or maximum CIMT (Table 3). However, CIMT appeared to be reduced in subjects who had donated blood very frequently ( 49 times). A more detailed analysis was Table 1 Characteristics of a general population of Dutch men and women (n = 819) by blood donation status Never-donors (n = 256) Ex-donors (n = 120) Current donors (n = 443) Age (years) 61.3 (5.7) 62.1 (5.6) 59.1 (5.4) Sex (% men) Body mass index (kg/m 2 ) 26.5 (3.9) 26.5 (3.5) 26.6 (3.4) Blood donation history a Total number 0 24 (6; 49) 50 (30; 70) Number of years 0 16 (4; 32) 32 (18; 38) Smoking % Current % Past Blood pressure (mmhg) Systolic 133 (16) 133 (19) 133 (15) Diastolic 76 (8) 77 (9) 78 (8) Self-reported CVD (%) Self-reported diabetes mellitus (%) Alcohol intake (g/day) a 11 (3; 23) 13 (5; 22) 13 (5; 24) Serum lipids (mmol/l) Total cholesterol 6.20 (1.29) 5.73 (1.12) 5.63 (0.93) HDL cholesterol 1.30 (0.42) 1.15 (0.33) 1.21 (0.32) LDL cholesterol 4.32 (1.12) 3.97 (0.99) 3.85 (0.84) Triglycerides a 1.15 (0.87; 1.60) 1.29 (0.89; 1.67) 1.14 (0.81; 1.60) Serum CRP (mg/dl) a 1.31 (0.72; 2.69) 1.30 (0.62; 2.40) 0.96 (0.53; 2.12) Serum iron parameters Total iron ( mol/l) 18.2 (5.4) 18.8 (5.7) 18.5 (7.0) Ferritin ( g/l) a 124 (78; 190) 114 (62; 163) 44 (22; 70) Transferrin saturation (%) 33 (10) 34 (12) 30 (12) Total iron binding capacity ( mol/l) 56 (7) 57 (8) 63 (8) NTBI ( mol/l) 2.51 (0.80) 2.54 (0.89) 2.33 (0.94) Data are presented as unadjusted means (S.D.). a Median value (interquartile range) is given because of skewed distribution.

4 M.F. Engberink et al. / Atherosclerosis 196 (2008) Table 2 Carotid intima-media thickness (CIMT) by blood donation status in a general population of Dutch men and women Never-donors (n = 256) Ex-donors (n = 120) Current donors (n = 443) P a Mean CIMT (mm) Crude 0.83 (0.008) 0.85 (0.011) 0.82 (0.006) Model 1 b 0.83 (0.007) 0.83 (0.011) 0.83 (0.005) 0.99 Model 2 c 0.83 (0.007) 0.82 (0.011) 0.82 (0.005) 0.96 Model 2, in subjects free of CVD and diabetes mellitus d 0.82 (0.007) 0.81 (0.012) 0.82 (0.005) 0.93 Model 3 e 0.83 (0.007) 0.82 (0.011) 0.82 (0.005) 0.95 Maximum CIMT (mm) Crude 1.03 (0.011) 1.06 (0.016) 1.01 (0.008) Model 1 b 1.03 (0.010) 1.03 (0.014) 1.02 (0.008) 0.61 Model 2 c 1.03 (0.010) 1.03 (0.014) 1.02 (0.007) 0.61 Model 2, in subjects free of CVD and diabetes mellitus d 1.01 (0.010) 1.01 (0.016) 1.01 (0.007) 0.95 Model 3 e 1.03 (0.010) 1.02 (0.014) 1.02 (0.007) 0.72 Data are presented as means (S.E.). P-value obtained by ANCOVA. Model 1: adjusted for age (continues) and sex. c Model 2: adjusted for age (continues), sex, BMI (continues), CRP (continues), smoking status (never/past/current) and pack-years of smoking (continues). d n = 703. e Model 3: adjusted for age (continues), sex, BMI (continues), CRP (continues), smoking status (never/past/current), pack-years of smoking (continues), LDL cholesterol (continues), and HDL cholesterol (continues). performed among current donors, comparing CIMT in highfrequent donors ( 2 donations per year) with low-frequent donors (<2 donations per year). Although CIMT was lower in high-frequent donors than in low-frequent donors, this was not statistically significant (Table 4). Since plasma donors usually donate more frequently than whole blood donors and are overrepresented in the group with high-frequency donors (21% versus 19%), we performed a post hoc analysis excluding plasma donors. Mean CIMT in whole blood donors, adjusted for age, sex, BMI, CRP, smoking status and pack-years of smoking, was 0.82 mm in high-frequent donors compared with 0.83 mm in low-frequent donors (p = 0.32; data not in table). Values for maximum CIMT were 1.02 mm and 1.00 mm (p = 0.28; data not in table). Table 5 presents mean CIMT in sex-specific quartiles of serum iron parameters. No significant trends over the quartiles were observed, except for age-adjusted CIMT which was slightly increased in the upper quartile of serum ferritin (0.84 mm, compared to 0.82 mm in lower quartiles; P = 0.05). Further adjustment for confounders (model 2) made the relationship with serum ferritin non-significant. Age-adjusted maximum CIMT (data not presented in table) was 1.01, 1.01, 1.02 and 1.04 mm in consecutive quartiles of serum ferritin (P = 0.04), but this trend disappeared after further adjustment for confounders. Maximum CIMT was not related to other serum iron parameters. 4. Discussion In the present study, in a general population of men and postmenopausal women, blood donors had a significantly Table 3 Association of lifetime blood donation with carotid intima-media thickness (CIMT) in a general population of Dutch men and women Lifetime number of blood donations a P trend 0(n = 256) 1 24 (n = 114) (n = 116) (n = 113) >68 (n = 112) Mean CIMT (mm) Model 1 b 0.83 (0.007) 0.83 (0.011) 0.83 (0.011) 0.81 (0.011) 0.81 (0.011) 0.48 Model 2 c 0.83 (0.007) 0.83 (0.011) 0.83 (0.011) 0.81 (0.011) 0.81 (0.011) 0.43 Model 3 d 0.83 (0.007) 0.83 (0.011) 0.83 (0.011) 0.81 (0.011) 0.81 (0.011) 0.43 Maximum CIMT (mm) Model (0.010) 1.03 (0.015) 1.03 (0.014) 1.00 (0.015) 1.01 (0.015) 0.33 Model (0.010) 1.03 (0.015) 1.03 (0.014) 1.00 (0.014) 1.00 (0.015) 0.31 Model (0.010) 1.03 (0.014) 1.03 (0.014) 1.00 (0.014) 1.00 (0.015) 0.33 Data are presented as means (S.E.). a Based on quintiles: first quintile are the never-donors; subsequent quintiles are based on number of subjects with median lifetime number of blood donations 12, 38, 56, and 80, respectively. b Model 1: adjusted for age (continues) and sex. c Model 2: adjusted for age (continues), sex, BMI (continues), CRP (continues), smoking status (never/past/current) and pack-years of smoking (continues). d Model 3: adjusted for age (continues), sex, BMI (continues), CRP (continues), smoking status (never/past/current), pack-years of smoking (continues), LDL cholesterol (continues), and HDL (continues).

5 860 M.F. Engberink et al. / Atherosclerosis 196 (2008) Table 4 Association of blood donation frequency a with carotid intima-media thickness (CIMT) in current Dutch blood donors Low frequent donors (n = 93) High frequent donors (n = 272) Mean CIMT (mm) Crude 0.82 (0.013) 0.81 (0.007) 0.47 Model 1 b 0.83 (0.012) 0.81 (0.007) 0.36 Model 2 c 0.83 (0.011) 0.81 (0.005) 0.30 Model 3 d 0.83 (0.012) 0.81 (0.007) 0.34 Maximum CIMT (mm) Crude 1.01 (0.016) 1.00 (0.009) 0.47 Model 1 b 1.02 (0.015) 1.00 (0.008) 0.26 Model 2 c 1.02 (0.015) 1.00 (0.008) 0.23 Model 3 d 1.02 (0.015) 1.00 (0.008) 0.29 Data are presented as means (S.E.); P-value obtained by ANCOVA. a Based on donation frequency among current donors: low-frequent donors are defined as donors who donate on average <2 times per year; highfrequency donors are defined as those who donate on average 2 times per year. Median serum ferritin among low-frequent donors: 54 (32; 79); median serum ferritin among high-frequent donors: 39 (20; 65). b Model 1: adjusted for age (continues) and sex. c Model 2: adjusted for age (continues), sex, BMI (continues), CRP (continues), smoking status (never/past/current) and pack-years of smoking (continues). d Model 3: adjusted for age (continues), sex, BMI (continues), CRP (continues), smoking status (never/past/current), pack-years of smoking (continues), LDL cholesterol (continues), and HDL (continues). lower iron status than ex-donors and never-donors. However, body iron status was not clearly related to CIMT, a marker of atherosclerosis and vascular disease, even though CIMT was slightly elevated in the highest quartile of serum ferritin. Blood donation was not consistently associated with CIMT, P but we cannot exclude the possibility of less atherosclerosis in frequent blood donors. We studied an asymptomatic outcome (CIMT) and it is therefore unlikely that use of medication or intentional changes in dietary and lifestyle factors have affected iron status. Donors, however, must satisfy rigorous screening criteria and if CVD is detected they are excluded from blood donation. In the present study we showed that ex-donors had a significantly higher prevalence of CVD and diabetes mellitus compared to current and never-donors. When we performed additional analyses excluding 116 subjects with a history of CVD or diabetes, however, similar results were obtained for iron status and blood donation in relation to CIMT. In The Netherlands, the number of whole blood donations is limited to three per year for women and four per year for men, whereas the number of plasma donations is unrestricted. Generally, the total number of donations will thus be strongly associated with age, sex, and plasma donation, which may diminish the effect. Adjustment for age and sex (Tables 3 and 4) indeed led to a more, but not statistically significant, reduced CIMT in donors who donate more frequently. Excluding plasma donors from the analysis, however, increased CIMT in frequent donors. In our study that included both donors and non-donors, a wide range in body iron could be studied. Various iron parameters were assessed, including serum iron, ferritin, total iron binding capacity, and (calculated) transferrin saturation. Blood donation had a clear effect on these iron parameters. Ferritin is the most frequently used clinical measure, and a good indicator of stored body iron [18]. Transferrin is responsible for the transport of iron from sites of absorption to sites Table 5 Association of serum iron parameters with carotid intima-media thicknes (CIMT) in a general population of Dutch men and women Mean CIMT (mm) in sex-specific quartiles of iron parameters P trend Q1 Q2 Q3 Q4 Total iron ( mol/l) Model 1 a Model 2 b Ferritin ( g/l) Model 1 a Model 2 b Transferrin saturation (%) Model 1 a Model 2 b Total iron binding capacity ( mol/l) Model 1 a Model 2 b NTBI ( mol/l) Model 1 a Model 2 b Data are presented as means; S.E. = 0.01 for all values. a Model 1: adjusted for age (continues). b Model 2: adjusted for age (continues), BMI (continues), CRP (continues), smoking status (never/past/current) and pack-years of smoking (continues). Sex was adjusted for by means of sex-specific quartiles.

6 M.F. Engberink et al. / Atherosclerosis 196 (2008) of use [27]. Binding to ferritin or transferrin prevent iron from catalytic reactions that produce free oxygen radicals. These reactive forms of oxygen may play a role in the development of CVD by oxidizing LDL, which could damage the vascular endothelium. Bound iron is considered to be less harmful than free iron. However, a proportion of iron is unable to bind and circulates as NTBI. This unbound iron may be a better predictor of atherosclerosis than overall body iron status. NTBI was initially detected in iron-overloaded patients with over 100% transferrin saturation, but in more recent studies it was also identified in sera with incomplete saturation [21,22]. We measured NTBI by a fluorescence-based one step chelation method [22,26,28]. Alternatively, the redoxactive component of NTBI could be assessed, [29 31] which may be even more clinically relevant. However, these methods are more appropriate to measure NTBI when serum iron exceeds 40 mol/l or transferrin saturation exceeds 70%, which is not the case in our healthy population. Transferrin saturation was strongly correlated with NTBI in our study (r = 0.97, P < 0.001), as was also shown by Jacobs et al. [32]. In the present study, no association was found between NTBI and atherosclerosis. Although our results are in line with the findings of a previous study by Van der A et al, who found no relation between NTBI and risk of coronary heart disease [24], the added value of NTBI in studies of CVD needs further investigation. Until now, the atherogenic pathway is believed to be the most plausible mechanism by which body iron levels might affect the risk of CVD. The present study in individuals with body iron levels within the normal range, however, does not provide strong evidence for this hypothesis. However, we cannot completely rule out the possibility that lifelong blood donation has a small beneficial effect on CIMT. A number of epidemiological studies, however, did show a positive association between body iron and risk of CVD. Possibly, mechanisms other than atherosclerosis could be responsible for the harmful effect of iron, such as direct oxidative damage to the myocardium [33], mechanisms that involve the thrombotic pathway [34], or decreased insulin sensitivity [33]. In conclusion, the present epidemiological study in a general Dutch population does not provide strong evidence for the hypothesis that blood donation or specific parameters of body iron status, including NTBI, influence the atherosclerotic process. However, we cannot exclude the possibility that blood donation, at least frequent, might give some protection against accelerated atherosclerosis. Acknowledgements We thank Rudy Meijer and Annemarie Rijswijk of the Vascular Imaging Center, University Medical Center Utrecht for CIMT image acquisition and reading. Furthermore, we thank Dr. Berry van Tits, Department of General Internal Medicine, Radboud University Nijmegen Medical Center, Nijmegen, The Netherlands, for the measurement of the NTBI. This study was supported by grants from The Netherlands Organisation for Health Research and Development (ZonMw, grant number ), Wageningen University and Wageningen Centre for Food Sciences (WCFS), and Sanquin Blood Bank, The Netherlands (grant number ). Contribution: W.L.A.M. and E.G.S. participated in designing the research; M.F.E., J.D. and P.V. participated in performing the research; M.F.E. and J.M.G. analyzed the data; all authors participated in interpretation of the data; M.F.E. drafted the manuscript; all authors critically evaluated the manuscript. References [1] Sullivan JL. Iron and the sex difference in heart disease risk. Lancet 1981;1: [2] Morrison HI, Semenciw RM, Mao Y, Wigle DT. Serum iron and risk of fatal acute myocardial infarction. Epidemiology 1994;5: [3] Salonen JT, Nyyssonen K, Korpela H, Tuomilehto J, Seppanen R, Salonen R. High stored iron levels are associated with excess risk of myocardial infarction in eastern Finnish men. Circulation 1992;86: [4] Tuomainen TP, Punnonen K, Nyyssonen K, Salonen JT. Association between body iron stores and the risk of acute myocardial infarction in men. Circulation 1998;97: [5] Meyers DG, Jensen KC, Menitove JE. A historical cohort study of the effect of lowering body iron through blood donation on incident cardiac events. Transfusion 2002;42: [6] Meyers DG, Strickland D, Maloley PA, Seburg JK, Wilson JE, McManus BF. Possible association of a reduction in cardiovascular events with blood donation. Heart 1997;78: [7] Salonen JT, Tuomainen TP, Salonen R, Lakka TA, Nyyssonen K. Donation of blood is associated with reduced risk of myocardial infarction. The Kuopio Ischaemic Heart Disease Risk Factor Study. Am J Epidemiol 1998;148: [8] Tuomainen TP, Salonen R, Nyyssonen K, Salonen JT. Cohort study of relation between donating blood and risk of myocardial infarction in 2682 men in eastern Finland. BMJ 1997;314: [9] Baer DM, Tekawa IS, Hurley LB. Iron stores are not associated with acute myocardial infarction. Circulation 1994;89: [10] Danesh J, Appleby P. Coronary heart disease and iron status: metaanalyses of prospective studies. Circulation 1999;99: [11] Liao Y, Cooper RS, McGee DL. Iron status and coronary heart disease: negative findings from the NHANES I epidemiologic follow-up study. Am J Epidemiol 1994;139: [12] Sempos CT, Looker AC, Gillum RE, McGee DL, Vuong CV, Johnson CL. Serum ferritin and death from all causes and cardiovascular disease: the NHANES II Mortality Study. National Health and Nutrition Examination Study. Ann Epidemiol 2000;10: [13] Sempos CT, Looker AC, Gillum RF, Makuc DM. Body iron stores and the risk of coronary heart disease. N Engl J Med 1994;330: [14] Zheng H, Cable R, Spencer B, Votto N, Katz SD. Iron stores and vascular function in voluntary blood donors. Arterioscler Thromb Vasc Biol 2005;25: [15] Kiechl S, Willeit J, Egger G, Poewe W, Oberhollenzer F. Body iron stores and the risk of carotid atherosclerosis: prospective results from the Bruneck study. Circulation 1997;96: [16] Moore M, Folsom AR, Barnes RW, Eckfeldt JH. No association between serum ferritin and asymptomatic carotid atherosclerosis. The Atherosclerosis Risk in Communities (ARIC) Study. Am J Epidemiol 1995;141:

7 862 M.F. Engberink et al. / Atherosclerosis 196 (2008) [17] Rauramaa R, Vaisanen S, Mercuri M, Rankinen T, Penttila I, Bond MG. Association of risk factors and body iron status to carotid atherosclerosis in middle-aged eastern Finnish men. Eur Heart J 1994;15: [18] Cook JD, Lipschitz DA, Miles LE, Finch CA. Serum ferritin as a measure of iron stores in normal subjects. Am J Clin Nutr 1974;27: [19] Finch CA, Cook JD, Labbe RF, Culala M. Effect of blood donation on iron stores as evaluated by serum ferritin. Blood 1977;50: [20] Lee DH, Jacobs Jr DR. Serum markers of stored body iron are not appropriate markers of health effects of iron: a focus on serum ferritin. Med Hypotheses 2004;62: [21] de Valk B, Addicks MA, Gosriwatana I, Lu S, Hider RC, Marx JJ. Non-transferrin-bound iron is present in serum of hereditary haemochromatosis heterozygotes. Eur J Clin Invest 2000;30: [22] Gosriwatana I, Loreal O, Lu S, Brissot P, Porter J, Hider RC. Quantification of non-transferrin-bound iron in the presence of unsaturated transferrin. Anal Biochem 1999;273: [23] de Valk B, Marx JJ. Iron, atherosclerosis, and ischemic heart disease. Arch Intern Med 1999;159: [24] van der AD, Marx JJ, Grobbee DE, et al. Non-transferrin-bound iron and risk of coronary heart disease in postmenopausal women. Circulation 2006;113: [25] Durga J, Bots ML, Schouten EG, Kok FJ, Verhoef P. Low concentrations of folate, not hyperhomocysteinemia, are associated with carotid intima-media thickness. Atherosclerosis 2005;179: [26] Breuer W, Cabantchik ZI. A fluorescence-based one-step assay for serum non-transferrin-bound iron. Anal Biochem 2001;299: [27] Ponka P, Beaumont C, Richardson DR. Function and regulation of transferrin and ferritin. Semin Hematol 1998;35: [28] Loreal O, Gosriwatana I, Guyader D, Porter J, Brissot P, Hider RC. Determination of non-transferrin-bound iron in genetic hemochromatosis using a new HPLC-based method. J Hepatol 2000;32: [29] Esposito BP, Breuer W, Sirankapracha P, Pootrakul P, Hershko C, Cabantchik ZI. Labile plasma iron in iron overload: redox activity and susceptibility to chelation. Blood 2003;102: [30] Gutteridge JM, Rowley DA, Griffiths E, Halliwell B. Low-molecularweight iron complexes and oxygen radical reactions in idiopathic haemochromatosis. Clin Sci (Lond) 1985;68: [31] von Bonsdorff L, Lindeberg E, Sahlstedt L, Lehto J, Parkkinen J. Bleomycin-detectable iron assay for non-transferrin-bound iron in hematologic malignancies. Clin Chem 2002;48: [32] Jacobs EM, Hendriks JC, van Tits BL, et al. Results of an international round robin for the quantification of serum non-transferrin-bound iron: need for defining standardization and a clinically relevant isoform. Anal Biochem 2005;341: [33] Parkes JG, Hussain RA, Olivieri NF, Templeton DM. Effects of iron loading on uptake, speciation, and chelation of iron in cultured myocardial cells. J Lab Clin Med 1993;122: [34] Day SM, Duquaine D, Mundada LV, et al. Chronic iron administration increases vascular oxidative stress and accelerates arterial thrombosis. Circulation 2003;107: