Low Toenail Chromium Concentration and Increased Risk of Nonfatal
Myocardial Infarction
Eliseo Guallar1,2, F. Javier Jime´nez2,3, Pieter van ’t Veer4, Peter Bode5, Rudolph A. Riemersma6,7, Jorge Go´mez-Aracena8,9, Jeremy D. Kark10,11, Lenore Arab12, Frans J. Kok4, and Jose´ M. Martı´n-Moreno2,13for the EURAMIC–Heavy Metals and Myocardial Infarction Study Group
1
Department of Epidemiology and Welch Center for Prevention, Epidemiology and Clinical Research, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD. 2
National School of Public Health, Institute of Health Carlos III, Madrid, Spain.
3Medical Department, AstraZeneca, Madrid, Spain. 4
Division of Human Nutrition, Graduate School of Food Technology, Agrobiotechnology, Nutrition and Health Sciences, Wageningen University, Wageningen, the Netherlands.
5
Interfaculty Reactor Institute, Delft University of Technology, Delft, the Netherlands.
6
Cardiovascular Research Unit, College of Medicine, University of Edinburgh, Edinburgh, United Kingdom.
7
Department of Medical Physiology, Faculty of Medicine, University of Tromsø, Tromsø, Norway.
8
Department of Preventive Medicine, Faculty of Medicine, University of Ma´laga, Ma´laga, Spain.
9
Nordic School of Public Health, Go¨teborg, Sweden. 10Epidemiology Unit, Hadassah Medical Organization, Jerusalem, Israel.
11Hadassah School of Public Health and Community Medicine, Hebrew University, Jerusalem, Israel. 12
Department of Epidemiology, School of Public Health, University of California, Los Angeles, CA.
13
Department of Preventive Medicine and Public Health, Faculty of Medicine and Odontology, Universidad de Valencia, Valencia, Spain.
Received for publication January 10, 2005; accepted for publication March 16, 2005.
Chromium intake may increase insulin sensitivity, glucose tolerance, and the ratio of high density lipoprotein cholesterol to low density lipoprotein cholesterol. However, the epidemiologic evidence on the association between chromium and cardiovascular disease is very limited. To determine whether low toenail chromium concentrations were associated with risk of nonfatal myocardial infarction, the authors conducted an incident, population-based, case-control study in eight European countries and Israel in 1991–1992. Cases (n¼684) were men with a first diagnosis of myocardial infarction recruited from the coronary units of participating hospitals. Controls (n¼724) were men selected randomly from population registers (five study centers) or through other sources, such as hospitalized patients (three centers), general practitioners’ practices (one center), or relatives or friends of cases (one center). Toenail chromium concentration was assessed by neutron activation analysis. Average toenail chromium concentrations were 1.10lg/g in cases (95% confidence interval: 1.01, 1.18) and 1.30lg/g in controls (95% CI: 1.21, 1.40). Multivariate odds ratios for quintiles 2–5 were 0.82 (95% CI: 0.52, 1.31), 0.68 (95% CI: 0.43, 1.08), 0.60 (95% CI: 0.37, 0.97), and 0.59 (95% CI: 0.37, 0.95). Toenail chromium concentration was inversely associated with the risk of a first myocardial infarction in men. These results add to an increasing body of evidence that points to the importance of chromium for cardiovascular health.
case-control studies; chromium; metals, heavy; myocardial infarction; neutron activation analysis
Abbreviations: CI, confidence interval; EURAMIC, EURopean Multicenter Case-Control Study on Antioxidants, Myocardial Infarction, and Cancer of the Breast.
Chromium is most likely an essential trace element re-quired for normal carbohydrate, lipid, and protein
metabo-lism (1–8). Chromium deficiency results in impaired glucose tolerance, hyperglycemia, and glycosuria that cannot be
Correspondence to Dr. Eliseo Guallar, Welch Center for Prevention, Epidemiology and Clinical Research, 2024 East Monument Street, Room 2-639, Baltimore, MD 21205-2223 (e-mail: [email protected]).
controlled with insulin (9, 10). Although overt chromium deficiency is rare, low chromium intake may be a cause of subclinical insulin resistance and an adverse lipid profile in Western populations (4–6, 9, 11). Use of chromium supple-ments is increasingly popular in some countries, such as the United States (12, 13), although their effects on glucose tolerance, body composition, and lipid parameters are still unclear (7, 8, 14–17).
The association of chromium intake with cardiovascular endpoints is largely unknown. The few early studies available were limited by unreliable analytical methods (18, 19) and/ or the use of serum chromium measurements, which may not adequately reflect long-term chromium intake (1, 3, 19, 20).
To evaluate the hypothesis that long-term exposure to chromium is inversely related to the risk of coronary heart disease, we measured the toenail chromium concentra-tions of participants in the EURAMIC Study (EURopean Multicenter Case-Control Study on Antioxidants, Myocar-dial Infarction, and Cancer of the Breast) (21–23) and as-sessed their association with the risk of a first myocardial infarction.
MATERIALS AND METHODS
Design and participants
The EURAMIC Study was a large international case-control study of antioxidant concentrations and cardiovascu-lar disease risk (21). The target population of the EURAMIC Study was men aged 70 years or younger who were native residents of eight European countries or residents of Israel. Subjects were excluded if they had a previous diagnosis of myocardial infarction, had undergone a recent change in diet, were institutionalized, or had a history of drug or alco-hol abuse or major psychiatric disorders.
Cases were men with a first acute myocardial infarction (International Classification of Diseases, Ninth Revision, code 410), confirmed by characteristic electrocardiogram abnormalities and elevated serum enzyme levels, who had been hospitalized within 24 hours of the onset of symptoms. Cases were recruited from the coronary care units of par-ticipating hospitals.
Controls were men without a history of myocardial in-farction who were recruited from the case’s population catchment area and frequency matched for age (in 5-year intervals). In Finland, Israel, Germany, Scotland, and Switzerland, random sampling from local population regis-ters was used for control selection. In Russia and in the two Spanish study centers, population registries could not be used, because of the lack of complete census data or because of legal restrictions. Therefore, controls were selected from among hospitalized patients with disorders not known to be associated with dietary factors (renal colic, hernia, acute appendicitis or mesenteric adenitis, volvulus or subocclu-sion due to fibrosis, noninfectious prostatism, and rectal or anal disorders other than cancer, hemorrhoids, or chronic infections). When low participation rates from population samples were anticipated, controls were selected by random sampling from the catchment area of the patient’s general practitioner (in the Netherlands) or by inviting apparently
healthy friends and relatives of the patient to participate (in Norway) (21–23).
Cases and controls were recruited concurrently during 1991 and 1992. The overall response rate was 81 percent in cases and 64 percent in controls. Informed consent was obtained from study participants in accordance with the ethical standards of the responsible local committees on human experimentation.
Data collection
Toenail clippings from all 10 toes were collected within 8 weeks of inclusion in the study and were stored in small plastic bags at room temperature (22, 23). A nonfasting venous blood sample was drawn for cholesterol analysis. In cases, blood samples were drawn within 24 hours of hospital admission.
Information on smoking habits, history of hypertension, angina pectoris, and diabetes was collected for all subjects by means of standard questionnaires (21). Socioeconomic status, alcohol intake, and family history of cardiovas-cular diseases were assessed through locally developed questionnaires.
Analysis of biologic samples
Toenail chromium concentrations were measured by in-strumental neutron activation analysis at the Interfaculty Reactor Institute of Delft University of Technology in Delft, the Netherlands (24, 25). The quality assurance system of this laboratory is accredited by the Dutch Council for Accreditation for compliance with European (EN 45001) and international (ISO/IEC Guide 25) standards. Toenail clippings were irradiated for 4 hours in a thermal flux of 5 3 1012/second/cm2. After a decay time of 21 days, the gamma radiation of chromium-51 was measured during 1 hour in a well-type germanium detector. Samples were rotated during irradiation. Irradiation of study samples was conducted from April 1998 through June 1999. The clip-pings of each subject were irradiated and measured together. For each center, samples from cases and controls were analyzed together and randomly distributed across batches. Personnel at the Interfaculty Reactor Institute were blinded with respect to the case-control status of the samples.
Chromium concentrations were reported in lg/g. For each sample, a limit of detection was defined as the concen-tration at which chromium could be detected with 97.5 per-cent certainty. The limit of detection for a sample of average weight (54 mg) was 0.4lg/g. In samples with chromium concentrations below the detection limit (n¼96), the chro-mium concentration was set at one half the reported detec-tion limit.
For quality control, a sample of the certified reference material (BCR-CRM 414, plankton; Institute for Reference Materials and Measurements, Geel, Belgium) was included in each analytical batch (25). The interassay coefficient of variation based on 49 measures of this reference material was 17.3 percent (6.2 percent for log-transformed values).
Serum cholesterol level was determined enzymat-ically using kits obtained from Boehringer-Mannheim
(Boehringer-Mannheim GmbH, Mannheim, Germany). High density lipoprotein cholesterol level was determined after precipitation with dextran sulfate and magnesium chloride. Cholesterol determinations were performed at the National Public Health Institute in Helsinki, Finland.
Statistical methods
Because of the marked right skewness of the distribution of toenail chromium values, logarithmic transformations were used. The distribution of chromium in controls was used to compute cutoff points and medians for quintiles of exposure. Geometric mean chromium concentrations by participant characteristics among controls were compared by linear regression andv2tests.
For multivariate analysis, the association of chromium with the risk of myocardial infarction was estimated with multiple logistic regression. Adjusted odds ratios in quin-tiles 2–5 were calculated using the lowest quintile as the reference category, and trend tests were computed by in-cluding log chromium concentration in the logistic models as a continuous variable. All p values reported are two-tailed. Statistical analyses were performed in S-Plus (26).
RESULTS
The EURAMIC Study recruited 742 men with a first myocardial infarction and 757 controls. Toenail clippings were not available for 58 cases and 33 controls; therefore, we analyzed data from 684 cases and 724 controls. The mean weight of the samples was 54 mg (range, 2–223 mg). Compared with controls, cases had a significantly higher body mass index and a lower level of high density lipoprotein cholesterol. Cases were more likely to be hypertensive, to be diabetic, to smoke, to have angina, and to have a family history of coronary heart disease (table 1). Although we found a lower total cholesterol concentration among cases during hospitalization as compared with controls, this almost certainly reflects the known effect of acute myocardial in-farction. Therefore, total cholesterol was not further consid-ered in case-control comparisons.
Toenail chromium concentrations among controls varied across study centers (table 2). Granada, Spain, had the lowest average chromium concentration (geometric mean¼0.70 lg/g), while Moscow, Russia, and Ma´laga, Spain, had the highest concentrations (1.85 lg/g and 1.87 lg/g, respec-tively)—a 2.7-fold range of variation. The overall geometric mean chromium concentration in controls was 1.30lg/g (95 percent confidence interval (CI): 1.21, 1.40).
Among controls, toenail chromium concentration was inversely related to age (table 2), with an average 9 percent decline (95 percent CI: 2, 16) in chromium with each de-cade of age. Participants with a history of hypertension had lower chromium concentrations than those without such a history (1.11 lg/g vs. 1.35 lg/g; p ¼ 0.05; table 2). Chromium was not associated with history of diabetes, but the small proportion of diabetics among controls lim-ited the power of this analysis. Body mass index, total and high density lipoprotein cholesterol, smoking, alcohol
in-take, and family history of coronary heart disease were not associated with chromium.
The concentration of chromium in myocardial infarction cases was 1.10lg/g (95 percent CI: 1.10, 1.18). After adjust-ment for age, study center, smoking, alcohol drinking, body mass index, high density lipoprotein cholesterol, diabetes, history of hypertension, and family history of coronary heart disease, chromium was 13 percent lower in cases (95 percent CI: 1, 24) (p¼0.04). In risk analyses, there was an inverse trend in the odds ratio for myocardial infarction with increas-ing chromium concentration (table 3). This inverse trend persisted after adjustment for cardiovascular disease risk factors; odds ratios were 0.87, 0.72, 0.54, and 0.65 for quin-tiles 2–5 of chromium, respectively, in comparison with the first quintile (p for trend¼0.04). Adjustment for adipose tissue a-tocopherol, b-carotene, lycopene, linoleic acid, and toenail selenium did not materially modify the risk esti-mates (table 3). Further adjustment for levels of other metals in toenails did not materially alter the results (not shown).
The variability in the association between chromium con-centration and risk of myocardial infarction across centers was consistent with random variation (pfor interaction be-tween center and chromium concentration¼0.12; table 4). In sensitivity analyses, we compared the association be-tween chromium and myocardial infarction at study centers with population controls versus other centers, and at the five centers with the highest response rates among controls ver-sus the other five centers; results were similar (table 4). Finally, we reanalyzed the data while excluding the results from Ma´laga and Moscow, the centers with the highest TABLE 1. Distribution of cardiovascular disease risk factors in myocardial infarction cases and controls, EURAMIC*–Heavy Metals and Myocardial Infarction Project
Risk factor Cases (n¼684)
Controls (n¼724)
p
value
Mean age (years) 54.7 (8.9)y 53.2 (9.3) 0.002 Mean body mass indexz 26.5 (3.9) 25.9 (3.4) 0.004 Mean cholesterol level
(mmol/liter)
Total cholesterol 5.46 (1.11) 5.56 (1.10) 0.11 High density
lipoprotein
cholesterol 0.98 (0.25) 1.09 (0.29) <0.001 History of
hypertension (%) 26.0 17.4 <0.001 Current smoking (%) 61.3 37.5 <0.001 Angina pectoris (%) 14.0 4.5 <0.001 Diabetes mellitus (%) 8.4 3.9 <0.001 Current alcohol use (%) 80.2 82.3 0.31 Family history of
coronary heart
disease (%) 57.6 45.3 <0.001
*EURAMIC, EURopean Multicenter Case-Control Study on Antioxidants, Myocardial Infarction, and Cancer of the Breast.
yNumbers in parentheses, standard deviation.
TABLE 2. Toenail chromium concentrations among controls, by participant characteristics, EURAMIC*–Heavy Metals and Myocardial Infarction Project
Characteristic No. of participants
Geometric mean chromium concentration (lg/g)
95% confidence interval
p
value
Age (years)
<50 254 1.38 1.22, 1.55 0.006y
50–<60 258 1.42 1.26, 1.60
60 212 1.10 0.96, 1.26
Study center <0.001
Helsinki, Finland 62 1.38 1.10, 1.74 Berlin, Germany 97 1.08 0.88, 1.31 Jerusalem, Israel 59 1.01 0.79, 1.29 Zeist, the Netherlands 57 1.04 0.76, 1.41 Sarpsborg, Norway 101 1.31 1.07, 1.60 Moscow, Russia 97 1.85 1.58, 2.17 Edinburgh, United Kingdom 25 1.28 0.90, 1.82 Granada, Spain 52 0.70 0.55, 0.89 Ma´laga, Spain 100 1.87 1.51, 2.33 Zu¨rich, Switzerland 74 1.39 1.13, 1.70
Body mass indexz 0.73y
<25 303 1.38 1.24, 1.54
25–<30 319 1.23 1.10, 1.38
30 93 1.36 1.12, 1.66
Smoking 0.77
Never smoker 181 1.35 1.18, 1.55 Former smoker 270 1.31 1.16, 1.49 Current smoker 271 1.26 1.12, 1.42
Alcohol drinking 0.31
Never drinker 65 1.24 0.94, 1.63 Former drinker 60 1.57 1.20, 2.07 Current drinker 583 1.29 1.19, 1.39
History of hypertension 0.05
No 598 1.35 1.24, 1.46
Yes 126 1.11 0.94, 1.32
Total cholesterol level (mmol/liter)
0.76y
<5.2 226 1.45 1.27, 1.65
5.2–<6.2 202 1.32 1.14, 1.51
6.2 159 1.41 1.21, 1.64
History of diabetes mellitus
0.61
No 692 1.30 1.21, 1.40
Yes 28 1.18 0.77, 1.81
Family history of coronary heart disease
0.30
No 396 1.26 1.14, 1.39
Yes 328 1.36 1.22, 1.51
*EURAMIC, EURopean Multicenter Case-Control Study on Antioxidants, Myocardial Infarc-tion, and Cancer of the Breast.
yBased on a test for trend; all otherpvalues were based on tests for heterogeneity.
concentrations of chromium among controls; the inverse association of chromium with the risk of myocardial infarc-tion persisted (table 4).
DISCUSSION
In this large international case-control study, we found an inverse association between toenail chromium concentra-tion and risk of nonfatal myocardial infarcconcentra-tion. The associ-ation was strong, it was independent of the presence of diabetes, and it persisted after adjustment for cardiovascular disease risk factors and for other nutrients that could be associated with chromium intake, such as alcohol, carot-enoids, or linoleic acid. A similar inverse association be-tween toenail chromium and risk of cardiovascular disease was reported in a nested case-control study (27) and in a cross-sectional study of diabetics (28) in the Health Pro-fessionals Follow-up Study.
Trivalent chromium, the reduced form of the element, is required for insulin action (1–7). A chromium-containing oligopeptide present in insulin-sensitive cells binds to the insulin receptor, markedly increasing the activity of the insulin-stimulated tyrosine kinase (7, 29, 30). Overt chro-mium deficiency has been demonstrated in patients receiving total parenteral nutrition without chromium supplementa-tion (4, 10). It is characterized by hyperglycemia, glycosuria, and weight loss that cannot be controlled with insulin (9, 10, 31); as a consequence, total parenteral nutrition solutions are regularly supplemented with chromium (32). Intraperitoneal
injections of potassium chromate also reversed atheroscle-rotic plaques in rabbits (33, 34).
The main route of exposure to chromium in the general population is dietary intake. Most chromium in the diet is trivalent chromium, and any hexavalent chromium in food or water is reduced to trivalent chromium in the acidic en-vironment of the stomach (1, 3). Foods with high chromium concentrations include whole grain products, green beans, broccoli, and bran cereals (35). The chromium content of meats, poultry, and fish varies widely, since chromium may be introduced during transport, processing, and fortification of foods (35). Foods rich in refined sugars not only are low in chromium but promote chromium loss (36). The use of stainless steel equipment in food processing adds measur-able amounts of chromium (37), especially in acidic foods, salted foods, and foods that are in close contact with stain-less steel when processed. The chromium content of grain products, fruits, and vegetables also varies extensively, pos-sibly for the aforementioned reasons (37) or because of soil properties (38).
Based on the chromium content of well-balanced diets (35), Adequate Intake values for chromium in adults have been established at 35 lg/day in men and 25 lg/day in women (3). Although there are no national survey data on chromium intakes (3), a study of self-selected diets of US adults indicated that the chromium intake of a substantial proportion of subjects may be well below the Adequate Intake (11); similar results have been shown in the United Kingdom, Finland, Canada, and New Zealand (39). Thus, subclinical chromium deficiency may be a contributor to TABLE 3. Odds ratios for myocardial infarction, by quintile of toenail chromium concentration,
EURAMIC*–Heavy Metals and Myocardial Infarction Project
Quintile of chromium concentration pfor trend
First Second Third Fourth Fifth
No. of cases 187 139 143 97 118
No. of controls 145 145 144 145 145 Median chromium level
in controls (lg/g) 0.37 0.73 1.23 2.07 4.66
Model 1 odds ratioy 1z 0.74 0.76 0.54 0.64 0.005 95% CI* 0.54, 1.03 0.55, 1.05 0.38, 0.76 0.46, 0.90
Model 2 odds ratio§ 1z 0.87 0.72 0.54 0.65 0.04 95% CI 0.57, 1.32 0.47, 1.10 0.35, 0.84 0.42, 0.99
Model 3 odds ratio{ 1z 0.82 0.68 0.60 0.59 0.04 95% CI 0.52, 1.31 0.43, 1.08 0.37, 0.97 0.37, 0.95
*EURAMIC, EURopean Multicenter Case-Control Study on Antioxidants, Myocardial Infarction, and Cancer of the Breast; CI, confidence interval.
yAdjusted for age (continuous) and study center (indicator variables).
zReference category.
§ Further adjusted for smoking (indicator variables for current and former smokers), alcohol drinking (indicator variables for current and former drinkers), body mass index (continuous variable), high density lipoprotein cholesterol (continuous variable), diabetes (indicator variable), history of hypertension (indicator variable), and family history of coronary heart disease (indicator variable).
{Further adjusted for toenail selenium adipose tissue levels of a-tocopherol, b-carotene, and lycopene (as continuous log-transformed variables) and major fatty acid peaks (palmitic acid, stearic acid, oleic acid, linoleic acid, arachidonic acid, n-3 fatty acids, andtrans-fatty acids, all as continuous variables).
glucose intolerance, insulin resistance, and cardiovascular disease, particularly in aging populations or populations that have increased chromium requirements because of high-sugar diets (6, 9).
Although there is a biologic basis for a protective effect of chromium on cardiovascular disease, few epidemiologic studies have addressed this hypothesis. This is due in part to the lack of food composition data and to the variability in the chromium content of foods as a function of preparation (1, 3). In addition, the low absorption and bioavailability of chromium offer little confidence in dietary assessment-based estimates of intake and exposure. For all of these reasons, biomarkers of chromium intake offer compelling alternative exposure assessment approaches. However, early studies used unreliable analytical methods for chromium assay (18, 19) and/or serum chromium measurements (19), which do not adequately reflect tissue chromium con-centrations (1, 3).
In our study, we used toenail chromium concentrations to obtain a time-integrated measure reflecting chromium
expo-sure over the past several weeks that has been previously used in epidemiologic studies (28, 40). In a paired compar-ison of toenails collected 6 years apart, the reproducibility of toenail chromium measurements was moderate (40). This suggests that toenail chromium can be used as a biomarker of long-term exposure, although attenuation of the potential associations would be expected to occur because of mea-surement error. As an additional indicator of the validity of toenail chromium measurements, we found in toenails the pattern of decreasing chromium concentrations with age that was previously found in hair, sweat, and serum (41). Both hair and toenail chromium determinations reflect tri-valent chromium, and their biologic similarity makes it likely that chromium concentrations in nails, as in hair, parallel tissue chromium concentrations (40).
Our study presented some additional strengths. We used instrumental neutron activation analysis for chromium de-termination, a highly sensitive technique that is considered the preferred method for determination of chromium con-centrations in biologic materials (4, 24). The international TABLE 4. Odds ratio for myocardial infarction among persons in the 75th percentile of the distribution of
toenail chromium concentrations versus persons in the 25th percentile, EURAMIC*–Heavy Metals and Myocardial Infarction Projecty
No. of cases No. of controls Odds ratio 95% confidence
interval pfor interaction
Overall 684 724 0.81 0.71, 0.93
Center 0.12
Helsinki, Finland 56 62 0.59 0.35, 0.99 Berlin, Germany 75 97 1.17 0.77, 1.76 Jerusalem, Israel 57 59 0.94 0.53, 1.66 Zeist, the Netherlands 64 57 0.68 0.41, 1.12 Sarpsborg, Norway 96 101 0.73 0.50, 1.06 Moscow, Russia 92 97 0.97 0.61, 1.55 Edinburgh, United
Kingdom 39 25 0.50 0.24, 1.05
Granada, Spain 55 52 1.30 0.76, 2.22 Ma´laga, Spain 94 100 0.53 0.37, 0.77 Zu¨rich, Switzerland 56 74 1.32 0.81, 2.15
Type of control group 0.31
Population based 283 317 0.89 0.71, 1.10
Other 401 407 0.76 0.64, 0.91
Response rate among controls
0.67
Five highest study
centers 412 447 0.84 0.70, 0.99
Five lowest study
centers 272 277 0.78 0.62, 0.98
Excluding Ma´laga and
Moscow 498 527 0.85 0.72, 1.00
*EURAMIC, EURopean Multicenter Case-Control Study on Antioxidants, Myocardial Infarction, and Cancer of the Breast.
yBased on logistic regression analyses with a linear term for log-transformed chromium. Results of all analyses were adjusted for age. The 25th and 75th percentiles of toenail chromium concentration in the control group were 0.65lg/g and 2.41lg/g, respectively.
character of the study, including cases and controls from countries with widely different habits, also adds relevance to our findings. Finally, the use of incident cases of myocar-dial infarction makes it unlikely that toenail chromium could have been affected by disease development. For prac-tical purposes, the toenail measurements used in our study represent predisease biomarkers of chromium exposure.
Sample contamination is a major concern when mea-suring chromium in biologic materials (4, 42). In the EURAMIC Study, toenails were clipped using conventional manicure scissors, which could have resulted in the transfer of chromium-containing stainless steel to the toenails. How-ever, using irradiated clippers, Anderson and Morris (43) determined that the transfer of such material to toenails was very small (<100 pg) and that most of the transferred material was removed during washing prior to neutron ac-tivation analysis. Thus, toenail contamination during clip-ping is likely to have been small and nondifferential and is unlikely to account for the observed case-control differences in our study. We further minimized chromium contamina-tion by storing the toenails in plastic bags (42) and by using neutron activation analysis, which does not involve prepa-ratory steps (such as sample homogenization or dissolution) prior to analysis (4, 24).
Although our study provides evidence of an inverse asso-ciation of chromium with the risk of myocardial infarction, some limitations need to be considered. Our analyses, based on single measurements of toenail chromium concentration, were subject to random measurement error that added to the analytical error. Since random errors tend to attenuate risk estimates, we can assume that our analyses underestimated the inverse association of chromium with myocardial infarc-tion. The validity of control selection is also a key issue in case-control studies. Although we aimed at obtaining population-based controls, this approach was only feasible at five of the 10 study centers. However, results were similar for centers with population controls and centers with other types of controls and were also independent of the response rate at each center, reducing the possibility of selection bias. A further limitation of our study was the lack of data on the sources of toenail chromium. Most toenail chromium in the general population is likely to be derived from trivalent chromium intake (1, 3), but we could not separate it from the effect of occupational or environmental inhalation of hexa-valent chromium, which is present in industrial fumes and dust (44) and which may produce different cardiovascular effects. In addition, since we lacked data on dietary intake and on use of multivitamin or mineral supplements, we could not rule out the possibility that the inverse association between chromium and nonfatal myocardial infarction could be explained by dietary determinants of chromium intake or other lifestyle characteristics. Finally, our cases were patients with nonfatal myocardial infarction who sur-vived until hospitalization; chromium may show a different association for fatal events. We note, however, that in the Health Professionals Follow-up Study (27), toenail chro-mium concentration was inversely related to both fatal and nonfatal events.
In summary, our study provides evidence of an inverse association between long-term exposure to chromium and
the risk of myocardial infarction. Despite the scarcity of information on the effects of chromium on cardiovascular disease prevention, chromium supplements are actively used and promoted as improving glucose control, weight loss, exercise capacity, and longevity, particularly in the United States (12, 13), while the use of these supplements is low in Europe and Israel, where this study was conducted. Consid-erably more evidence is needed to substantiate many such claims, as well as to show the long-term safety of chromium supplementation in humans (7, 8, 14–17). Nevertheless, our study indicates that intake of chromium-containing foods may be inversely related to the risk of nonfatal myocardial infarction, and it supports the increasing body of evidence pointing to the importance of chromium for cardiovascular health.
ACKNOWLEDGMENTS
The EURAMIC–Heavy Metals and Myocardial Infarc-tion Project was supported by a BIOMED-2 Concerted Action contract from the European Commission (research contract BMH4-CT98-3565) and was an ancillary project to the EURAMIC Study.
This paper was presented at the American Heart Associ-ation’s 41st Annual Conference on Cardiovascular Disease Epidemiology and Prevention, San Antonio, Texas, March 2, 2001.
Conflict of interest: none declared.
REFERENCES
1. Stoecker BJ. Chromium. In: Shils ME, Olson JA, Shike M, et al, eds. Modern nutrition in health and disease. Baltimore, MD: Williams & Wilkins Company, 1999:277–82.
2. World Health Organization, International Atomic Energy Agency, and Food and Agriculture Organization of the United Nations. Trace elements in human nutrition and health. Geneva, Switzerland: World Health Organization, 1996. 3. Panel on Micronutrients, Food and Nutrition Board, National
Academy of Sciences. Dietary reference intakes for vitamin A, vitamin K, arsenic, boron, chromium, copper, iodine, iron, manganese, molybdenum, nickel, silicon, vanadium, and zinc. Washington, DC: National Academy Press, 2001.
4. Katz SA, Salem H. The biological and environmental chem-istry of chromium. New York, NY: VCH Publishers, 1994. 5. Mertz W. Chromium in human nutrition: a review. J Nutr
1993;123:626–33.
6. Anderson RA. Chromium as an essential nutrient for humans. Regul Toxicol Pharmacol 1997;26:S35–41.
7. Vincent JB. Recent advances in the nutritional biochemistry of trivalent chromium. Proc Nutr Soc 2004;63:41–7.
8. Cefalu WT, Hu FB. Role of chromium in human health and in diabetes. Diabetes Care 2004;27:2741–51.
9. Anderson RA. Chromium, glucose intolerance and diabetes. J Am Coll Nutr 1998;17:548–55.
10. Jeejeebhoy KN. Chromium and parenteral nutrition. J Trace Elem Exp Med 1999;12:85–9.
11. Anderson RA, Kozlovsky AS. Chromium intake, absorption and excretion of subjects consuming self-selected diets. Am J Clin Nutr 1985;41:1177–83.
12. Simon HB. Patient-directed, nonprescription approaches to cardiovascular disease. Arch Intern Med 1994;154:2283–96. 13. Lukaski HC. Chromium as a supplement. Annu Rev Nutr
1999;19:279–302.
14. Vincent JB. The potential value and toxicity of chromium picolinate as a nutritional supplement, weight loss agent and muscle development agent. Sports Med 2003;33:213–30. 15. Nissen SL, Sharp RL. Effect of dietary supplements on lean
mass and strength gains with resistance exercise: a meta-analysis. J Appl Physiol 2003;94:651–9.
16. Pittler MH, Stevinson C, Ernst E. Chromium picolinate for reducing body weight: meta-analysis of randomized trials. Int J Obes Relat Metab Disord 2003;27:522–9.
17. Althuis MD, Jordan NE, Ludington EA, et al. Glucose and insulin responses to dietary chromium supplements: a meta-analysis. Am J Clin Nutr 2002;76:148–55.
18. Schroeder HA, Nason AP, Tipton IH. Chromium deficiency as a factor in atherosclerosis. J Chronic Dis 1970;23:123–42. 19. Newman HA, Leighton RF, Lanese RR, et al. Serum
chro-mium and angiographically determined coronary artery dis-ease. Clin Chem 1978;24:541–4.
20. Simonoff M. Low plasma chromium in patients with coronary artery and heart diseases. Biol Trace Elem Res 1984;6:431–9. 21. Kardinaal AF, Kok FJ, Ringstad J, et al. Antioxidants in adi-pose tissue and risk of myocardial infarction: The EURAMIC Study. Lancet 1993;342:1379–84.
22. Kardinaal AF, Kok FJ, Kohlmeier L, et al. Association be-tween toenail selenium and risk of acute myocardial infarction in European men: The EURAMIC Study. European Antioxi-dant Myocardial Infarction and Breast Cancer. Am J Epide-miol 1997;145:373–9.
23. Guallar E, Sanz-Gallardo MI, van ’t Veer P, et al. Mercury, fish oils, and the risk of myocardial infarction. N Engl J Med 2002;347:1747–54.
24. Bode P, de Goeij JJM. Activation analysis. In: Meyers RA, ed. Encyclopedia of environmental analysis and remediation. New York, NY: John Wiley and Sons, Inc, 1998:68–84.
25. Bode P. Automation and quality assurance in the NAA facil-ities in Delft. J Radioanal Nucl Chem 2000;245:127–32. 26. Data Analysis and Products Division, MathSoft, Inc. S-Plus
2000 user’s guide. Seattle, WA: MathSoft, Inc, 1999. 27. Rimm EB, Guallar E, Giovannucci E, et al. Toenail
chro-mium levels and risk of coronary heart disease among normal and overweight men. (Abstract). In: Final Program and Abstracts of the 42nd Annual Conference on Cardiovascular Disease Epidemiology and Prevention, Honolulu, Hawaii, April 23–26, 2002. Dallas, TX: American Heart Association, 2002:P151.
28. Rajpathak S, Rimm EB, Li T, et al. Lower toenail chromium in men with diabetes and cardiovascular disease compared with healthy men. Diabetes Care 2004;27:2211–16.
29. Davis CM, Vincent JB. Chromium oligopeptide activates insulin receptor tyrosine kinase activity. Biochemistry 1997; 36:4382–5.
30. Vincent JB. Mechanisms of chromium action: low-molecular-weight chromium-binding substance. J Am Coll Nutr 1999; 18:6–12.
31. Morris BW, MacNeil S, Hardisty CA, et al. Chromium ho-meostasis in patients with type II (NIDDM) diabetes. J Trace Elem Med Biol 1999;13:57–61.
32. Department of Foods and Nutrition, American Medical As-sociation. Guidelines for essential trace element preparations for parenteral use: a statement by an expert panel. JAMA 1979;241:2051–4.
33. Abraham AS, Sonnenblick M, Eini M, et al. The effect of chromium on established atherosclerotic plaques in rabbits. Am J Clin Nutr 1980;33:2294–8.
34. Abraham AS, Brooks BA, Eylath U. Chromium and cholesterol-induced atherosclerosis in rabbits. Ann Nutr Metab 1991;35:203–7.
35. Anderson RA, Bryden NA, Polansky MM. Dietary chromium intake: freely chosen diets, institutional diet, and individual foods. Biol Trace Elem Res 1992;32:117–21.
36. Kozlovsky AS, Moser PB, Reiser S, et al. Effects of diets high in simple sugars on urinary chromium losses. Metabolism 1986;35:515–18.
37. Offenbacher EG, Pi-Sunyer FX, Stoecker BJ. Chromium. In: O’Dell BL, Sunde RA, eds. Handbook of nutritionally essen-tial mineral elements. New York, NY: Marcel Dekker, 1997: 389–411.
38. Pfeifer HR, Derron MH, Rey D, et al. Natural trace element input to the soil-sediment-water-plant system: examples of background and contaminated situation in Switzerland, eastern France and northern Italy. In: Markert B, Friese K, eds. Trace elements—their distribution and effects in the environment. Amsterdam, the Netherlands: Elsevier, 2000:33–86.
39. Anderson RA. Chromium. In: Mertz W, ed. Trace elements in human and animal nutrition. Orlando, FL: Academic Press, Inc, 1987:225–44.
40. Garland M, Morris JS, Rosner BA, et al. Toenail trace element levels as biomarkers: reproducibility over a 6-year period. Cancer Epidemiol Biomarkers Prev 1993;2:493–7. 41. Davies S, McLaren HJ, Hunnisett A, et al. Age-related
de-creases in chromium levels in 51,665 hair, sweat, and serum samples from 40,872 patients—implications for the preven-tion of cardiovascular disease and type II diabetes mellitus. Metabolism 1997;46:469–73.
42. Veillon C, Patterson KY. Analytical issues in chromium re-search. J Trace Elem Exp Med 1999;12:99–109.
43. Anderson H, Morris JS. Contamination studies in the use of human nails for dietary studies: the effect of clippers. J Ra-dioanal Nucl Chem 1992;158:251–9.
44. Frantzen K. Chromium. In: Harbison RD, Hardy HL, Hamilton A, eds. Hamilton and Hardy’s industrial toxicology. 5th ed. St. Louis, MO: CV Mosby Company, 1998:51–4.
