Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 May 1.
Published in final edited form as: Environ Res. 2016 Mar 2;147:356–364. doi: 10.1016/j.envres.2016.02.032

Metal mixtures in urban and rural populations in the US: The Multi-Ethnic Study of Atherosclerosis and the Strong Heart Study☆

Yuanjie Pang a,*, Roger D Peng b, Miranda R Jones a, Kevin A Francesconi c, Walter Goessler c, Barbara V Howard d,e, Jason G Umans d,e, Lyle G Best f, Eliseo Guallar a,g,h, Wendy S Post a,g,h, Joel D Kaufman i, Dhananjay Vaidya h, Ana Navas-Acien a,g,j
PMCID: PMC4827253  NIHMSID: NIHMS767794  PMID: 26945432

Abstract

Background

Natural and anthropogenic sources of metal exposure differ for urban and rural residents. We searched to identify patterns of metal mixtures which could suggest common environmental sources and/or metabolic pathways of different urinary metals, and compared metal-mixtures in two population-based studies from urban/sub-urban and rural/town areas in the US: the Multi-Ethnic Study of Atherosclerosis (MESA) and the Strong Heart Study (SHS).

Methods

We studied a random sample of 308 White, Black, Chinese-American, and Hispanic participants in MESA (2000–2002) and 277 American Indian participants in SHS (1998–2003). We used principal component analysis (PCA), cluster analysis (CA), and linear discriminant analysis (LDA) to evaluate nine urinary metals (antimony [Sb], arsenic [As], cadmium [Cd], lead [Pb], molybdenum [Mo], selenium [Se], tungsten [W], uranium [U] and zinc [Zn]). For arsenic, we used the sum of inorganic and methylated species (∑As).

Results

All nine urinary metals were higher in SHS compared to MESA participants. PCA and CA revealed the same patterns in SHS, suggesting 4 distinct principal components (PC) or clusters (∑As-U-W, Pb-Sb, Cd-Zn, Mo-Se). In MESA, CA showed 2 large clusters (∑As-Mo-Sb-U-W, Cd-Pb-Se-Zn), while PCA showed 4 PCs (Sb-U-W, Pb-Se-Zn, Cd-Mo, ∑As). LDA indicated that ∑As, U, W, and Zn were the most discriminant variables distinguishing MESA and SHS participants.

Conclusions

In SHS, the ∑As-U-W cluster and PC might reflect groundwater contamination in rural areas, and the Cd-Zn cluster and PC could reflect common sources from meat products or metabolic interactions. Among the metals assayed, ∑As, U, W and Zn differed the most between MESA and SHS, possibly reflecting disproportionate exposure from drinking water and perhaps food in rural Native communities compared to urban communities around the US.

Keywords: Metals, Urine, Biomarker, Statistical methods, Exposure sources

1. Introduction

Exposure to metals is widespread in the environment. Experimental and epidemiologic evidence support a role for low-to-moderate chronic exposure to certain toxic metals in the development of cardiovascular disease, kidney disease, morbid neurocognitive outcomes and some cancers (Hu, 2000; Navas-Acien et al., 2005, 2009a). Urinary biomarkers are commonly used to assess metal exposure and internal dose as they integrate multiple exposure sources including air, water and food (Aitio et al., 2007). Metals in urine might be related to each other due to common environmental sources or to similarities in metabolism. Multivariate analysis, including principal component analysis (PCA) and cluster analysis (CA), is widely used in environmental research to identify metal sources in air, soil and water (Lee et al., 2006; Loska and Wiechuła, 2003; Yongming et al., 2006) and to describe underlying patterns of metal biomarkers (Basu et al. 2011; Nowak, 1998; Wang et al., 2009). By reducing the initial dimension of the variables (Everitt et al. 2011; Hotelling, 1933), these methods can facilitate interpretation and identification of common sources and metabolic pathways for urinary metals.

Few studies have evaluated common sources of metal exposures in general populations, as most studies on metal-mixtures have focused on occupationally-exposed populations or populations living in contaminated areas (Basu et al. 2011; Nowak, 1998; Wang et al., 2009). Urban or rural residency might be an important source of variation in metal exposures as natural and anthropogenic sources could differ. While it is often assumed that urban areas are more contaminated than rural areas due to the high number of potential sources (Davis et al., 2009; Diamond and Hodge, 2007), some rural communities can sometimes be affected by important contamination (Carpenter, 2014; Hoover et al., 2012). Compared with urban areas, groundwater sources contaminated with naturally occurring metals are more commonly used for drinking water in rural and sub-urban areas [U.S. Environmental Protection Agency (USEPA, 2015)]. Sociocultural factors could also influence differences in metal exposure across different communities and ethnic/racial groups.

Our study population was drawn from two separate cohorts, American Indian participants in the Strong Heart Study (SHS) residing in rural areas and towns of Arizona, Oklahoma, and North/South Dakota, and White, Black, Hispanic, and Chinese-American participants in the Multi-Ethnic Study of Atherosclerosis (MESA) residing in urban and sub-urban areas of Baltimore, MD; Chicago, IL; Los Angeles, CA; New York, NY; St. Paul, MN; and Winston-Salem, NC. Both studies are funded by the National Heart, Lung, and Blood Institute. The communities and ethnic groups included in the study were selected with their main goal of evaluating cardiovascular disease and its risk factors in diverse populations around the United States.

Our objective was to characterize metal-mixtures in urine and identify patterns of metal mixtures which could suggest common environmental sources and/or metabolic pathways of different urinary metals in MESA and SHS. In addition to PCA and CA, we used linear discriminant analysis (LDA) to determine which metal (s) differed the most between MESA and SHS, as well as between different US regions and race/ethnic groups. To evaluate the consistency of the metal patterns across different communities, we compared the principal component (PC) score levels in each study area. We specifically hypothesized that arsenic, uranium and tungsten would cluster together due to common exposure from contaminated groundwater in the Southwestern and Midwestern States (McMahon et al., 2015; Salinas and Ingram, 2005). Understanding patterns of metal-mixtures in US communities could help to identify sources of metal exposures and to guide future assessment of the health implications of metal-mixtures.

2. Methods

2.1. Study population

MESA is a population-based cohort study evaluating cardiovascular disease and its risk factors in participants aged 45–84 years who were free of cardiovascular disease at baseline (2000–2002) in 6 urban and sub-urban communities in the United States (Bild et al., 2002). We recently measured baseline urinary metal concentrations in an overall sample of 310 participants from the 6 study sites (90 White, 75 Black, 75 Hispanic, and 70 Chinese American participants). These 310 participants were selected using random stratification by site and race group with a predetermined distribution of participants per race and site to ensure sufficient numbers for stratified analyses. The selected sample size was also based on funding available. We excluded 2 participants with abnormal levels of tungsten in urine (37.5 and 230.0 times higher than the 90th percentile), leaving a total of 308 participants for this analysis.

The SHS is a population-based cohort study of cardiovascular disease and its risk factors in 13 rural American-Indian communities (reservations and small towns) from Arizona, Oklahoma, and North/South Dakota that started in 1989–1991 (North et al., 2003). The names of the tribes are not provided following the desire of the communities. In 1998–2003, relatives of the original SHS participants were recruited into a family study that included 96 extended families (Arizona, 33; Oklahoma, 36; and North/South Dakota, 27) totaling 3665 participants from all three centers ranging in age from 14 to 93 years. Urinary metals were measured in 2456 of these participants as part of an ancillary study to evaluate gene-environment interactions for diabetes and the metabolic syndrome. Among them, we randomly sampled three individuals from each family within the same age range as MESA participants. Urinary metals were measured in 95 of 96 families. One family had only one individual within the MESA participant age range and six additional families had only two participants, resulting in a total of 277 participants for this analysis. The rationale for selecting up to three family members per family was to obtain a balanced sample size between SHS and MESA. Sensitivity analyses were conducted to confirm similar results by selecting one single family member per family.

The MESA study protocols were approved by each field center's institutional review board. The Strong Heart Study protocol was approved by the Institutional and Indian Health Service Review Boards and the participating American Indian communities. All the participants provided oral and written informed consent.

2.2. Urinary metals

In both MESA and SHS, spot urine specimens were collected in the morning at the time of the clinical examinations. Urinary antimony (Sb), arsenic (As), cadmium (Cd), lead (Pb), molybdenum (Mo), selenium (Se), tungsten (W), uranium (U) and zinc (Zn) were measured using inductively coupled plasma mass spectrometry (ICPMS) at the Trace Element Laboratory of University of Graz, Austria following the same protocol (Scheer et al., 2012). The nine metals were selected a priori following a previous study done in the original cohort study of SHS (Navas-Acien et al., 2009b). In addition, these metals were selected because data from NHANES have shown that most of them were relevant in the general US population (Navas-Acien et al., 2005, 2009b). Only nine metals were selected due to the challenges in achieving high quality exposure assessment and the need to incorporate appropriate standards and quality control methods even when using ICPMS (Scheer et al., 2012). For arsenic, we also measured inorganic arsenic, monomethylarsonate, dimethylarsinate and arsenobetaine and other arsenic cations using high performance liquid chromatography coupled to ICPMS, using the sum of inorganic and methylated species (∑As) in all analyses. In MESA, we accounted for arsenobetaine to remove the impact of relatively high levels of organic arsenic species from seafood imputing non-seafood urinary concentrations of inorganic and methylated arsenic species by regressing their original concentrations by arsenobetaine and extracting the model residuals (Jones et al., unpublished results; Navas-Acien et al., 2011). In SHS, seafood intake is rare and there was no need to account for organic arsenic species from seafood (Navas-Acien et al., 2009b). The limits of detection (LOD) were 0.1 μg/L for all arsenic species, 0.015 μg/L for Cd, 0.1 μg/L for Mo, 0.08 μg/L for Pb, 0.006 μg/L for Sb, 2 μg/L for Se, 0.008 μg/L for U, 0.005 μg/L for W and 10 μg/L for Zn. The percentages of participants with concentrations below the LOD are summarized in Table S1 (Supplementary Table S1). For samples below the LOD, we replaced their values by the LOD divided by the square root of two. Using Multiple Imputation (MI) program in Stata to estimate U and W levels below the LOD yielded similar results (data not shown). For all metals, we accounted for urine dilution by standardizing their concentrations using specific gravity according to urinary metal concentration*(mean urinary specific gravity-1)/(urinary specific gravity-1) (Suwazono et al., 2005).

2.3. Other variables

The interviews, physical examinations and collection of biospecimens were conducted in MESA and SHS by trained and certified staff using similar procedures (Bild et al., 2002; North et al., 2003). Sociodemographic (age, sex, race/ethnicity) and lifestyle information was collected using standardized questionnaires. Body mass index was calculated from measured weight (kg) divided by measured height (m2). Estimated glomerular filtration rate (eGFR) was calculated from recalibrated creatinine, age and sex using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula (Inker et al., 2011; Levey et al., 2009).

2.4. Statistical analysis

Descriptive analyses were conducted for each study separately (MESA and SHS). Essential metals (Mo, Se, and Zn) were normally distributed and analyzed untransformed. Toxic metals (∑As, Cd, Pb, Sb, U, and W) were right skewed and log-transformed for all analyses. Histograms showed that normalization was approximately achieved (Supplementary Figs. S1 and S2). Pearson's correlation coefficient was calculated to examine the bivariate correlations between each pair of metals.

PCA can reduce data complexity with minimum loss of original information by extracting latent factors (principal components, PCs) from the observed variables (Hotelling, 1933). The coefficients defining these linear combinations, called factor loadings, are the correlation coefficients of each variable with that component. Hierarchical CA can identify relatively homogeneous groups of variables combining them into agglomerative clusters until only one cluster is left (Everitt et al., 2011). We ran PCA and CA within each study (main analysis) as well as in both studies combined (secondary analysis), then compared the findings of PCA and CA to assess common patterns between the two approaches. Concentrations of the nine metals evaluated varied by differing orders of magnitude. We therefore normalized each variable to unit variance and zero mean before conducting PCA and CA. The PCA with varimax normalized rotation was applied to maximize the variances of the factor loadings across variables for each factor (Kaiser, 1958). We retained all principal factors with eigenvalues ≥1.0 (Kaiser, 1958). For CA, dendrograms were constructed to assess the cohesiveness of the clusters formed. We used Ward's method with Euclidean distances as the criterion for forming clusters. Scores for each of the retained PCs were compared across communities in MESA and SHS, respectively.

LDA discriminates groups by seeking the linear combination of the variables that provides maximal separation between the groups compared. Maximal separation is determined from an eigen analysis of W−1B, where B is the between-group sum-of-squares and cross-products matrix, and W is the within-group cross-products matrix (Fisher, 1936). The method extracts N-1 discriminant functions, N being the number of groups to differentiate. In a two-group situation, the discriminant function has the following mathematical form:

D=d1Z1+d2Z2+…+dnZn

where D is the score on discriminant function, d is the weighting coefficient of a discriminating variable, Z is the standardized value of a discriminating variable used in the analysis, and underscore n is the number of variables, in our case the nine metals. The magnitude of the weighting coefficients indicates the relative importance of that variable to group difference when controlling for the other variables. We conducted three discriminant analyses with the following grouping variables: study cohort (MESA and SHS), US regions (East (Maryland, New York and North Carolina), Midwest (Illinois, Minnesota, North/South Dakota and Oklahoma), and West (Arizona and California)), and race/ethnicity (White, Black, Hispanic, Chinese American and American Indian). When regions and race/ethnicity were used as grouping variables, we considered differences between two groups at a time for generating discriminant functions.

We ran several sensitivity analyses. First, we conducted PCA and CA in MESA adjusting for rice intake, with similar results (Supplementary Table S2 and Fig. S3). The aim was to control for the influence of rice intake on urinary ∑As and potentially better discriminate other sources of arsenic such as drinking water. Second, we repeated PCA and CA restricting to individuals with eGFR ≥60 ml/min/1.73 m2 and to non-current smokers, in separate analyses, with similar results (Supplementary Table S3, Table S4, Fig. S4, and Figure S5). Third, to control for shared environmental exposure within a family, we repeated PCA and CA in SHS with a random sample size of one and two individuals in each family instead of three, separately, also with similar findings (Supplementary Table S5 and Fig. S6). All statistical analyses and graphical displays were performed using R software (version 2.14.2; R Core Team 2014).

3. Results

3.1. Metal levels in urine

MESA participants were older, more likely to be men, never smokers, current alcohol drinkers, and had lower body mass index than SHS participants (Supplementary Table S6). SHS participants had higher urinary concentrations of all nine metals than MESA participants (Table 1, Fig. 1). In unadjusted analyses, participants in the Midwest and West regions had higher urinary ∑As, U, W and Zn than those in the East regions. Participants in the Midwest regions had higher urinary Cd than those in the East regions, and participants in the West regions had higher urinary Mo than those in the East regions. Compared with non-smokers, current smokers had higher levels of ∑As, Cd, Se, W and Zn. In MESA, compared with Whites, Blacks and Chinese Americans had higher levels of Cd and Zn. Mo was higher in Chinese Americans compared with Whites. ∑As levels were higher in Chinese Americans compared to Whites, and Pb levels were higher in Whites compared to other race/ethnic groups.

Table 1.

Geometric means of metal levels in urine (μg/L) by participant characteristicsa.

Cohort
 Region
 Smoking status
 Race/ethnicity
MESA SHS East Midwest West Non-smoker Smoker White Black Hispanic Chinese American American Indian
∑As 3.42 8.25** 3.03 5.28** 7.03** 4.87 6.23 2.80 2.80 3.71 4.95** 8.25**
Cd 0.61 0.87** 0.60 0.79* 0.70 0.63 1.14 0.54 0.72* 0.49 0.77* 0.87**
Mo 39.25 45.15** 36.23 41.26 47.47** 42.10 41.68 36.23 33.78 40.04 49.90* 45.15
Pb 1.62 2.08* 1.65 1.90 1.82 1.72 2.19 2.59 1.72 1.12** 1.23** 2.08
Sb 0.10 0.15** 0.10 0.11 0.10 0.10 0.12 0.11 0.08 0.10 0.11 0.15
Se 49.40 58.56** 46.99 57.97** 50.91 52.46 57.40* 51.42 46.53 48.42 51.94 58.56
U 0.01 0.05** 0.01 0.02** 0.04** 0.02 0.03 0.01 0.01 0.01 0.02** 0.05**
W 0.04 0.19** 0.03 0.09** 0.18** 0.08 0.12** 0.04 0.03 0.03 0.07** 0.19**
Zn 314.19 906.87** 320.54 632.70** 512.86** 468.72 720.54** 284.29 383.75* 284.29 327.01* 906.87**
Open in a new tab

Abbreviations: ∑As: the sum of inorganic and methylated arsenic species (after correction for arsenobetaine levels in MESA because of frequent seafood intake).

a

Metal levels are standardized by specific gravity to account for urine dilution. P-values are estimated based on one-way ANOVA and Bonferroni correction for multiple comparisons. Reference groups are participants from MESA, East US region, non-smokers and white.

*

p<0.05

**

p<0.01.

Fig. 1.

Fig. 1

Open in a new tab

Box plots of urinary metals (μg/L) standardized by specific gravity in MESA and SHS. As refers to the sum of inorganic and methylated arsenic species (after correction for arsenobetaine levels in MESA because of frequent seafood intake). Horizontal lines within boxes indicate medians; boxes, interquartile ranges; error bars, values within 1.5 times the interquartile range. We excluded outlying data points beyond 1.5 times the interquartile range. P-values are estimated based on one-way ANOVA, with MESA as the reference group.

We observed moderate positive correlations between Zn-Sb, Sb-W, Mo-W, U-W in MESA and between Mo-Se, Pb-Sb, ∑As-U, ∑As-W, and U-W in SHS (Supplementary Table S7).

3.2. Principal component and cluster analyses

Four PCs explained 65.9% and 67.3% of the total variance in MESA and SHS (Table 2), respectively. In MESA, the 4 PC (% variance explained) were characterized by Sb-U-W (20.3%), Pb-Se-Zn (17.9%), and Cd-Mo (14.4%) with an inverse correlation of Cd and Mo, and ∑As (13.3%). In SHS, the four PCs (% variance explained) were characterized by ∑As-U-W (18.5%), Pb-Sb (17.8%), Cd-Zn (15.8%), and Mo-Se (15.2%). PC score levels by communities in MESA and SHS were summarized in box plots (Supplementary Figs. S7 and S8). The distributions of the PC levels were similar across communities both in MESA and SHS, except for higher levels of PC-1 (∑As-U-W) in SHS for Arizona, and higher levels of PC-1 (Sb-U-W) in MESA for Los Angeles. When combining MESA and SHS in PCA, four PCs explained 68.5% of the total variance, including ∑As-U-W (25.0%), Mo-Se-Zn (16.4%), Cd-Mo (13.6%), and Pb-Sb (13.5%).

Table 2.

Standardized rotated factor loadings in principal component analysis in MESA and SHSa

MESA SHS
Component 1 2 3 4 1 2 3 4
∑As 0.01 −0.05 0.03 0.81 0.67 −0.24 0.11 0.06
Cd 0.07 0.07 0.69 0.30 0.04 0.07 0.68 −0.21
Mo 0.07 0.03 −0.67 0.27 0.09 0.09 −0.17 0.75
Pb −0.15 0.61 0.11 −0.07 −0.03 0.65 −0.03 −0.01
Sb 0.52 0.29 0.01 −0.15 −0.06 0.65 0.11 0.08
Se −0.08 0.42 −0.14 0.38 −0.14 0.11 0.33 0.56
U 0.58 −0.26 0.14 0.10 0.51 0.22 0.02 −0.18
W 0.56 0.02 −0.15 0.02 0.50 0.16 −0.14 0.13
Zn 0.21 0.55 0.01 −0.06 0.05 −0.02 0.59 0.16
Eigenvalue 1.84 1.61 1.30 1.19 1.66 1.60 1.43 1.37
Total variance
(%)		 20.27 17.94 14.41 13.25 18.50 17.77 15.84 15.18
Cumulative
(%)		 20.27 38.20 52.62 65.87 18.50 36.27 52.11 67.29
Open in a new tab
a

For adjustment of urine dilution and interpretation of ∑As, see Table 1. Factor loadings are bolded if > 0.40.

In MESA we found two large clusters: ∑As-Mo-Sb-U-W and Cd-Pb-Se-Zn (Fig. 2). In SHS, we found four clusters: Mo-Se, Cd-Zn, Pb-Sb, and ∑As-U-W (Fig. 2). When combining MESA and SHS in CA, we found three clusters: ∑As-U-W, Cd-Se-Zn, and Mo-Pb-Sb. When using the full population in SHS (n=2458), the PC/cluster of ∑As-W-U was consistent, but we failed to observe the PC/cluster of Cd-Zn.

Fig. 2.

Fig. 2

Open in a new tab

Dendrograms of metals in urine in MESA and SHS. All metals were standardized by specific gravity. As refers to the sum of inorganic and methylated arsenic species (after correction for arsenobetaine levels in MESA because of frequent seafood intake).

The results were similar in analyses stratified by sex for both studies. In analysis stratified by age, the results were similar for <60 and ≥60 years in MESA, and in <60 years in the SHS, but showed different patterns for SHS, particularly in CA, although the number of SHS participants ≥60 years was small (n=60) (data not shown).

3.3. Linear discriminant analyses

The weighting coefficients showed that the discriminant function contrasting MESA and SHS was negatively weighted most by ∑As, U, W, and Zn (Table 3). This indicated that ∑As, U, W, and Zn differed the most comparing MESA and SHS participants. Fig. S9 displays good group separation based on nine urinary metals (Supplementary Fig. S9).

Table 3.

Weighting coefficients in linear discriminant analysisa.

Cohort Region Race/ethnicity
MESA:SHS East:Midwest East:West Midwest:West Black:White Hispanic:White Chinese American:White American Indian:White
∑As −0.42 −0.37 −0.30 −0.32 0.02 −0.65 −0.67 −0.45
Cd 0.05 0.00 0.02 0.18 0.32 0.22 −0.21 −0.06
Mo 0.16 0.10 0.08 −0.11 −0.01 0.09 −0.25 0.04
Pb −0.04 −0.05 −0.00 0.02 −0.65 0.86 0.47 0.24
Sb 0.19 0.25 0.38 0.24 −0.42 −0.25 0.37 0.11
Se −0.17 −0.20 −0.05 0.32 −0.35 0.18 0.18 −0.10
U −0.41 −0.30 −0.36 −0.41 0.14 0.08 −0.32 −0.39
W −0.38 −0.49 −0.75 −0.58 −0.27 0.46 −0.27 −0.40
Zn −0.49 −0.30 −0.04 0.47 0.69 −0.36 −0.22 −0.37
Open in a new tab
a

For adjustment of urine dilution and interpretation of ∑As, see Table 1. Weighting coefficients are boldedif > 0.30.

Comparing East and Midwest US regions, the discriminant function was negatively weighted most by ∑As, U, W, and Zn. Comparing East and West regions, the discriminant function was positively weighted by Sb and negatively weighted by ∑As, U, and W. Comparing Midwest and West regions, the discriminant function was positively weighted by Se and Zn, and negatively weighted by ∑As, U, and W. In a plot of discriminant functions, participants in the East appeared to separate more from participants in the other two regions (Supplementary Fig. S10).

Comparing Black and White, the discriminant function was positively weighted by Cd and Zn, and negatively weighted by Pb, Sb, and Se. Comparing Hispanic and White participants, the discriminant function was positively weighted by Pb and W, and negatively weighted by ∑As and Zn. Comparing Chinese American and White participants, the discriminant function was positively weighted by Pb and Sb, and negatively weighted by ∑As and U. Comparing American Indian and White participants, the discriminant function was negatively weighted most by ∑As, U, W, and Zn.

4. Discussion

We used PCA, CA and LDA to identify potential common environmental sources and/or metabolic pathways using metal-mixtures in urine from participants from two well-established cohort studies that cover six urban/sub-urban and three rural/ town areas across the US, MESA and SHS. In SHS, PCA and CA provided consistent results. The ∑As-U-W cluster and PC in SHS and in MESA and SHS combined might reflect groundwater contamination in rural and sub-urban areas. The score level for this PC was higher for Arizona. The Cd-Zn cluster and PC in SHS could reflect common sources, for instance from organ meat consumption (Fretts et al., 2012), or interactions in metabolic pathways for those metals (Bridges and Zalups, 2005). In MESA, Black and Chinese American participants also had higher Zn levels, maybe also because of higher intake of organ meats and shellfish. SHS participants had a higher burden of all nine urinary metals compared to MESA. Moreover, urinary ∑As, U, W, and Zn concentrations differed the most among nine urinary metals in participants from MESA and SHS in LDA, reflecting disproportionate exposure from drinking water (∑As, U, W) and maybe diet (Zn) in rural Native communities compared to urban communities around the US.

Our study is important in that understanding metal-mixtures in general populations is a needed first step before evaluating the implications of those mixtures on health effects. The PCs and clusters identified in our study might provide useful insights into evaluating health effects associated with metal-mixtures. For instance, while arsenic, tungsten and uranium are highly toxic metals, most studies in the US have evaluated health effects of each metal individually instead of in combination. Given their possible source from groundwater and potential interactions among them, our findings might warrant future studies linking metal co-exposures to health outcomes in those two populations.

4.1. Arsenic, uranium and tungsten

Elevated levels of ∑As-U-W in groundwater are found in areas with high naturally occurring levels of these metals in rocks and soil [Agency for Toxic Substances and Disease Registry (ATSDR, 2005a); USEPA, 2006; U.S. Geological Survey (USGS, 2000a)]. Drinking water from groundwater sources might be a common route of exposure for these metals in some communities. Before 2006, arsenic levels were above the arsenic standard (10 μg/L) in several public drinking water systems in Arizona and North/South Dakota and some small public drinking water systems remain affected today. In private wells, As levels probably often exceeded 10 μg/L in Arizona, Oklahoma, and North/South Dakota (USGS, 2000b). The average U concentrations in drinking water were reported to reach 2.5 pCi/L in Arizona, 2.5 pCi/L in Oklahoma, and 2.7 pCi/L in California (USEPA, 2006). Although tungsten levels in drinking water are generally unknown, releases to groundwater typically occur in regions where natural formations of tungsten minerals are prevalent, including Arizona, California, and North/ South Dakota (USGS, 2014). In MESA, a cluster of ∑As-U-W could also be related to groundwater as evidenced by the first cluster in CA, which included those three metals. In PCA, there was also a PC for U and W, with a higher score for Los Angeles.

Among our nine metals, ∑As, U, and W had high discriminant ability to distinguish participants from MESA and SHS in LDA as well as participants from the West and Midwest regions compared to the East. Evidence from SHS has shown that arsenic is a risk factor for the development of cardiovascular disease and some cancers, and possibly for diabetes and chronic kidney disease (García-Esquinas et al., 2013; Kuo et al., 2015; Moon et al., 2013; Zheng et al., 2013), consistent with findings in other populations (Moon et al., 2012; Zheng et al., 2014). Uranium compounds are associated with chronic kidney disease, although few epidemio-logic studies have evaluated other health outcomes (ATSDR, 2013). Tungsten is thrombogenic and proinflammatory (ATSDR, 2005a) and has been linked to cardiovascular disease (Agarwal et al., 2011) and peripheral arterial disease (Navas-Acien et al., 2005). The markedly higher urinary ∑As, U, and W levels in SHS than MESA participants represent an additional call for action to prevent metal exposure, especially inorganic arsenic, in drinking water in rural communities in the US, including Native communities. In recent years, efforts have been made to ensure that community water systems are compliant with the As standard of 10 μg/L. Levels between 5 and 10 μg/L and even between 1 and 5 μg/L, are likely to occur disproportionately in small community water systems (Frost et al., 2003; USGS, 2000a). Private wells, moreover, are not required to comply with these arsenic standards.

4.2. Cadmium and zinc

Smoking is the major source of Cd in most populations (ATSDR, 2012). Among non-smokers, leaf and root vegetables, organ meats and shellfish are major sources of Cd exposure (ATSDR, 2012). In SHS, urinary Cd levels were relatively high and differed little by smoking status, suggesting another major source of cadmium in this population (Tellez-Plaza et al., 2013). Cd and Zn levels are higher in whole grains, shellfish and meat products, including organ meats (ATSDR, 2005b, 2012). High intake of meat products in SHS could represent a source of Cd and Zn and explain relatively high urinary Cd levels in SHS participants. In MESA, Zn levels were higher in Blacks and Chinese Americans compared to Whites, race/ethnic groups that are characterized by relatively higher consumption of shellfish and meat products, including organ meats. In MESA, Chinese Americans had higher intake of shellfish (81.4% with more than once per month) compared to Whites (52.2%). Blacks and Chinese Americans had higher intake of organ meats (27.1% and 26.7% with more than once per month) compared to Whites (13.3%). Cd and Zn may also be interrelated due to interactions of Cd and Zn in metabolism. Cd is able to replace or mimic Zn in the early steps of transport (e.g divalent metal transporter-1 and luminal Zn transporter-1), but then it is incapable of mediating the subsequent vital functions of Zn (Bridges and Zalups, 2005). Ample dietary intakes of Zn increase the induction of metallothionein (MT). The increased availability of MT could also increase the accumulation of Cd in the kidney (Jacquillet et al., 2006; Fatima et al., 2010).

Urinary Zn levels also distinguished participants from MESA and SHS in LDA. For American Indians living in rural areas or reservations, dietary choices are influenced by the limited selection of foods available at local stores or through the US Department of Agriculture commodity foods assistance program, such as processed meats (Fretts et al., 2012). Meat products are generally high in Zn and might contribute to higher urinary Zn concentrations in SHS compared to MESA participants.

4.3. Half-lives of urinary metals

Urinary As, U and W are reliable biomarkers of metal exposure (ATSDR, 2005a, 2013; Fowler et al., 2007). Urinary As is excreted triphasically, with half-lives of approximately 2, 9, and 38 days (Fowler et al., 2007). W is excreted through the kidneys with half-lives of 5 days (70%) and 100 days (30%) (ATSDR, 2005a). U half-life in the kidney has been reported as 1–6 days (ATSDR, 2013). Urinary Cd and Zn are also reliable biomarkers of long-term exposure to these metals, with half-lives in kidney of 10–30 years and 280 days, respectively (Nordberg et al., 2007). Urinary Mo reflects a dietary intake and has half-life between 4 min and 30 h (Werner et al., 2000). Se is excreted from urine through three phases, with half-times of approximately 1 day, 8–20 days, and 65–116 days (Nordberg et al., 2007). The half-life of renal excretion of Sb has been reported to be approximately 4 days in humans (Stemmer, 1976; Chulay et al., 1988; Kentner et al., 1995). The interpretation of lead in urine is unclear and urinary lead is considered as an inadequate biomarker of lead exposure (Nordberg et al., 2007). Thus, although some differences exist in the biomarkers evaluated, many of them have relatively long half-lives. Also, with chronic exposure, even biomarkers with relatively short half-life such as As could reflect long-term exposure (Navas-Acien et al., 2009b).

4.4. Limitations and strengths

Our study has several limitations. First, for some metals, such as Pb, Se and Zn, urinary concentrations are considered less reliable biomarkers of exposure and internal dose (ATSDR, 2005b, 2007; Bleys et al., 2009). However, since the major route of excretion for these metals is via urine, urine has been suggested as a useful marker of status of these metals (ATSDR, 2003; Lowe et al., 2009). Second, although both studies were selected by the National Heart, Lung and Blood Institute as typical communities to evaluate cardiovascular disease in diverse populations, they might not be representative of environmental metal exposures in urban and rural communities. For instance, metal analyses in the SHS were conducted after confirming moderate arsenic exposure in a pilot study (Navas-Acien et al., 2009b). Third, as in most epidemiologic studies we used spot urine samples, requiring adjustment for urine dilution. There is scientific debate about whether it is better to adjust for urine dilution using urinary specific gravity or creatinine. In our study, we adjusted for urinary specific gravity because urinary creatinine is also a marker of creatinine production, and thus it is associated with age, sex, and muscle mass (Sauvé et al., 2015). Specific gravity corrections could introduce less variability than urinary creatinine corrections. In a sensitivity analysis with adjustment of urinary creatinine, we found similar results in both MESA and SHS.

Strengths of this study include the inclusion of urban and rural populations in the US, the wide geographical and race/ethnic coverage, the rigorous laboratory methods with extensive quality control, the reliability of urine as biomarkers of exposure for most of the metals studied, and the simultaneous use of three multivariate statistical approaches to disentangle differences in metal exposures across two studies conducted in urban and rural communities. This study also benefited from the similar time periods and methods of collection of urine samples for metal analysis in both MESA and SHS, and the use of the same laboratory and analytical procedures.

5. Conclusions

In conclusion, the present study showed marked differences in the distribution and correlations of selected urinary metals in urban/sub-urban and rural/town populations, as represented by participants of MESA and SHS. On average, metal levels were significantly higher in urine in SHS participants than in MESA participants. Groundwater contaminated with ∑As-U-W or with U-W could explain the PCA and CA findings in SHS and MESA. Diets rich in meat products, including organ meats, could explain the Cd-Zn cluster and PC in SHS as well as higher Zn levels in some race/ethnic groups compared to Whites. The separation of the two studies in LDA with highest weight of ∑As, U, W and Zn suggests the impact of groundwater contamination and dietary differences across these populations. Additional research is needed in metal-mixtures to confirm the potential sources and metabolic pathways and to evaluate the potential health impacts of metal co-exposures. From a public health perspective, the marked difference in As, U, and W exposure between MESA and SHS highlights the importance of preventing exposure in rural communities affected by these metals in drinking water.

Supplementary Material

supplement

Abbreviations

∑As

the sum of inorganic and methylated species

As

arsenic

CA

cluster analysis

Cd

cadmium

CKD-EPI

the Chronic Kidney Disease Epidemiology Collaboration

eGFR

estimated glomerular filtration rate

ICPMS

inductively coupled plasma mass spectrometry

LDA

linear discriminant analysis

LOD

limits of detection

MESA

the Multi-Ethnic Study of Atherosclerosis

Mo

molybdenum

Pb

lead

PC

principal component

PCA

principal component analysis

Sb

antimony

Se

selenium

SHS

the Strong Heart Study

Zn

zinc

U

uranium

W

tungsten

Footnotes

Competing financial interests

None.

Appendix A. Supporting information

Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.envres.2016.02.032.

☆

MESA was supported by contracts N01-HC-95159 through N01-HC-95167, N01-HC-95169, and R01-HL077612 from the National Heart, Lung, and Blood Institute (NHLBI). SHS was supported by grants HL41642, HL41652, HL41654 and HL65521 from NHLBI. Metal analyses in SHS and MESA were supported by R01HL090863 from NHLBI and by R01ES021367 from the National Institute of Environmental Health Sciences. M. R. Jones was supported by a National Cancer Institute (NCI) National Research Service Award (T32CA009314). The MESA study protocols were approved by each field center's institutional review board. The Strong Heart Study protocol was approved by the Institutional and Indian Health Service Review Boards and the participating American Indian communities.

References

  1. Agarwal S, Zaman T, Tuzcu EM, Kapadia SR. Heavy metals and cardiovascular disease: results from the National Health and Nutrition Examination Survey (NHANES) 1999–2006. Angiology. 2011;62(5):422–429. doi: 10.1177/0003319710395562. [DOI] [PubMed] [Google Scholar]
  2. Aitio A, Bernard A, Fowler B, Nordberg G. Biological monitoring and biomarkers. In: Nordberg GF, Fowler BF, Nordberg M, Friberg L, editors. Handbook on the Toxicology of Metals. 3rd ed Elsevier; Amsterdam: 2007. pp. 65–78. [Google Scholar]
  3. Agency for Toxic Substances and Disease Registry (ATSDR) Toxicological Profile for Selenium. U.S. Department of Health and Human Services, Public Health Service; Atlanta, GA: 2003. [PubMed] [Google Scholar]
  4. ATSDR (Agency for Toxic Substances and Disease Registry) [accessed 01.05.15];Toxicological profile for Tungsten. 2005a Available: ⟨ http://www.atsdr.cdc.gov/toxprofiles/tp186.pdf⟩. [PubMed]
  5. ATSDR (Agency for Toxic Substances and Disease Registry) [accessed 01.05.15];Toxicological profile for Zinc. 2005b Available: ⟨ http://www.atsdr.cdc.gov/toxprofiles/tp60.pdf⟩. [PubMed]
  6. ATSDR (Agency for Toxic Substances and Disease Registry) [accessed 01.05.15];Toxicological profile for Lead. 2007 Available: ⟨ http://www.atsdr.cdc.gov/toxprofiles/tp13.pdf⟩.
  7. ATSDR (Agency for Toxic Substances and Disease Registry) [accessed 01.05.15];Toxicological profile for Cadmium. 2012 Available: ⟨ http://www.atsdr.cdc.gov/toxprofiles/tp5.pdf⟩. [PubMed]
  8. ATSDR (Agency for Toxic Substances and Disease Registry) [accessed 01.05.5];Toxicological profile for Uranium. 2013 Available: ⟨ http://www.atsdr.cdc.gov/toxprofiles/tp150.pdf⟩. [PubMed]
  9. Basu N, Nam DH, Kwansaa-Ansah E, Renne EP, Nriagu JO. Multiple metals exposure in a small-scale artisanal gold mining community. Environ. Res. 2011;111(3):463–467. doi: 10.1016/j.envres.2011.02.006. [DOI] [PubMed] [Google Scholar]
  10. Bild DE, Bluemke DA, Burke GL, Detrano R, Diez Roux AV, Folsom AR, et al. Multi-ethnic study of atherosclerosis: objectives and design. Am. J. Epidemiol. 2002;156(9):871–881. doi: 10.1093/aje/kwf113. [DOI] [PubMed] [Google Scholar]
  11. Bleys J, Navas-Acien A, Laclaustra M, Pastor-Barriuso R, Menke A, Ordovas J, et al. Serum selenium and peripheral arterial disease: results from the national health and nutrition examination survey, 2003-2004. Am. J. Epidemiol. 2009;169(8):996–1003. doi: 10.1093/aje/kwn414. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Bridges CC, Zalups RK. Molecular and Ionic mimicry and the transport of toxic metals. Toxicol. Appl. Pharmacol. 2005;204(3):274–308. doi: 10.1016/j.taap.2004.09.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Carpenter DO. Environmental exposure in indigenous communities: an international perspective. Rev. Environ. Health. 2014;29:3–4. doi: 10.1515/reveh-2014-0004. [DOI] [PubMed] [Google Scholar]
  14. Chulay J, Fleckenstein L, Smith DH. Pharmacokinetics of antimony during treatment of visceral leishmaniasis with sodium stibogluconate or meglumine antimoniate. Trans. R. Soc. Trop. Med. Hyg. 1988;82:69–72. [PubMed] [Google Scholar]
  15. Davis HT, Aelion CM, McDermott S, Lawson AB. Identifying natural and anthropogenic sources of metals in urban and rural soils using GIS-based data, pca, and spatial interpolation. Environ. Pollut. 2009;157:2378–2385. doi: 10.1016/j.envpol.2009.03.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Diamond ML, Hodge E. Urban contaminant dynamics: from source to effect. Environ. Sci. Technol. 2007;41:3796–3805. doi: 10.1021/es072542n. [DOI] [PubMed] [Google Scholar]
  17. Everitt BS, Landau S, Leese M, Stahl D. Hierarchical clustering. In: Everitt BS, Landau S, Leese M, Stahl D, editors. Cluster Analysis. 5th ed Wiley; Chichester, UK: 2011. pp. 71–98. [Google Scholar]
  18. Fatima H, Nouha A, Baati T, Imed M, Abdelhamid K. Cadmium retention increase: a probable key mechanism of the protective effect of zinc on cadmium-induced toxicity in the kidney. Toxicol. Lett. 2010;196:104–109. doi: 10.1016/j.toxlet.2010.04.006. [DOI] [PubMed] [Google Scholar]
  19. Fisher RA. The use of multiple measurements in taxonomic problems. Ann. Eugen. 1936;7:179–188. [Google Scholar]
  20. Fowler BA, Selene CH, Chou J, Jones RL, Chen CJ. Arsenic. In: Nordberg GF, Fowler BF, Nordberg M, Friberg L, editors. Handbook on the Toxicology of Metals. 3rd ed Elsevier; Amsterdam: 2007. pp. 367–406. [Google Scholar]
  21. Fretts AM, Howard BV, McKnight B, Duncan GE, Beresford SA, Mete M, et al. Associations of processed meat and unprocessed red meat intake with incident diabetes: the Strong Heart Family Study. Am. J. Clin. Nutr. 2012;95(3):752–758. doi: 10.3945/ajcn.111.029942. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Frost FJ, Muller T, Petersen HV, Thomson B, Tollestrup K. Identifying US populations for the study of health effects related to drinking water arsenic. J. Expo. Anal. Environ. Epidemiol. 2003;13(3):231–239. doi: 10.1038/sj.jea.7500275. [DOI] [PubMed] [Google Scholar]
  23. García-Esquinas E, Pollán M, Umans JG, Francesconi KA, Goessler W, Guallar E, et al. Arsenic exposure and cancer mortality in a US-based prospective cohort: the Strong Heart Study. Cancer Epidemiol. Biomark. Prev. 2013;22(11):1944–1953. doi: 10.1158/1055-9965.EPI-13-0234-T. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Hoover E, Cook K, Plain R, Sanchez K, Waghiyi V, Miller P, et al. Indigenous peoples of north America: environmental exposures and reproductive justice. Environ. Health Perspect. 2012;120:1645–1649. doi: 10.1289/ehp.1205422. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Hotelling H. Analysis of a complex of statistical variables into principal components. J. Educ. Psychol. 1933;24:417–441. [Google Scholar]
  26. Hu H. Exposure to metals. Prim. Care. 2000;27(4):983–996. doi: 10.1016/s0095-4543(05)70185-8. [DOI] [PubMed] [Google Scholar]
  27. Inker LA, Eckfeldt J, Levey AS, Leiendecker-Foster C, Rynders G, Manzi J, et al. Expressing the CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) cystatin C equations for estimating GFR with standardized serum cystatin C values. Am. J. Kidney Dis. 2011;58(4):682–684. doi: 10.1053/j.ajkd.2011.05.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Jacquillet G, Barbier O, Cougnon M, Tauc M, Namorado MC, Martin D, et al. Zinc protects renal function during cadmium intoxication in the rat. Am. J. Physiol. Ren. Physiol. 2006;290:127–137. doi: 10.1152/ajprenal.00366.2004. [DOI] [PubMed] [Google Scholar]
  29. Jones MR, Tellez-Plaza M, Vaidya D, Grau M, Francesconi KA, Goessler W, et al. Development of a biomarker to estimate inorganic arsenic exposure in populations with frequent seafood intake: evidence from the Multi-Ethnic Study of Atherosclerosis (MESA) and the National Health and Nutrition Examination Survey (NHANES) unpublished results.
  30. Kaiser HF. The varimax criterion for analytic rotation in factor analysis. Psychometrika. 1958;23:187–200. [Google Scholar]
  31. Kentner M, Leinemann M, Schaller KH, Weltle D, Lehnert G. External and internal antimony exposure in starter battery production. Int. Arch. Occup. Environ. Health. 1995;67:119–123. doi: 10.1007/BF00572235. [DOI] [PubMed] [Google Scholar]
  32. Kuo CC, Howard BV, Umans JG, Gribble MO, Best LG, Francesconi KA, et al. Arsenic exposure, arsenic metabolism, and incident diabetes in the Strong Heart Study. Diabetes Care. 2015;38(4):620–627. doi: 10.2337/dc14-1641. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lee CS, Li X, Shi W, Cheung SC, Thornton I. Metal contamination in urban, suburban, and country park soils of Hong Kong: a study based on GIS and multivariate statistics. Sci. Total Environ. 2006;356(1–3):45–61. doi: 10.1016/j.scitotenv.2005.03.024. [DOI] [PubMed] [Google Scholar]
  34. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AR, 3rd, Feldman HI, et al. A new equation to estimate glomerular filtration rate. Ann. Intern Med. 2009;150(9):604–612. doi: 10.7326/0003-4819-150-9-200905050-00006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Loska K, Wiechuła D. Application of principal component analysis for the estimation of source of heavy metal contamination in surface sediments from the Rybnik reservoir. Chemosphere. 2003;51(8):723–733. doi: 10.1016/S0045-6535(03)00187-5. [DOI] [PubMed] [Google Scholar]
  36. Lowe NM, Fekete K, Decsi T. Methods of assessment of zinc status in humans: a systematic review. Am. J. Clin. Nutr. 2009;89:2040S–2051S. doi: 10.3945/ajcn.2009.27230G. [DOI] [PubMed] [Google Scholar]
  37. McMahon PB, Bohlke JK, Dahm KG, Parkhurst DL, Anning DW, Stanton JS. Chemical considerations for an updated national assessment of brackish groundwater resources. Ground Water. 2015. [DOI] [PubMed]
  38. Moon K, Guallar E, Navas-Acien A. Arsenic exposure and cardiovascular disease: an updated systematic review. Curr. Atheroscler. Rep. 2012;14(6):542–555. doi: 10.1007/s11883-012-0280-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Moon KA, Guallar E, Umans JG, Devereux RB, Best LG, Francesconi KA, et al. Association between exposure to low to moderate arsenic levels and incident cardiovascular disease: a prospective cohort study. Ann. Intern. Med. 2013;159(10):649–659. doi: 10.7326/0003-4819-159-10-201311190-00719. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Navas-Acien A, Francesconi KA, Silbergeld EK, Guallar E. Seafood intake and urine concentrations of total arsenic, dimethylarsinate and arsenobetaine in the US population. Environ. Res. 2011;111(1):110–118. doi: 10.1016/j.envres.2010.10.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Navas-Acien A, Silbergeld EK, Sharrett R, Calderon-Aranda E, Selvin E, Guallar E. Metals in urine and peripheral arterial disease. Environ. Health Perspect. 2005;113(2):164–169. doi: 10.1289/ehp.7329. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Navas-Acien A, Tellez-Plaza M, Guallar E, Muntner P, Silbergeld E, Jaar B, et al. Blood cadmium and lead and chronic kidney disease in US adults: a joint analysis. Am. J. Epidemiol. 2009a;170(9):1156–1164. doi: 10.1093/aje/kwp248. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Navas-Acien A, Umans JG, Howard BV, Goessler W, Francesconi KA, Crainiceanu CM, et al. Urine arsenic concentrations and species excretion patterns in American Indian communities over a 10-year period: the strong heart study. Environ. Health Perspect. 2009b;117(9):1428–1433. doi: 10.1289/ehp.0800509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Nordberg GF, Nogawa K, Nordberg M, Friberg L. Cadmium. In: Nordberg GF, Fowler BF, Nordberg M, Friberg L, editors. Handbook on the Toxicology of Metals. 3rd ed Elsevier; Amsterdam: 2007. pp. 445–486. [Google Scholar]
  45. North KE, Howard BV, Welty TK, Best LG, Lee ET, Yeh JL, et al. Genetic and environmental contributions to cardiovascular disease risk in American Indians the strong heart family study. Am. J. Epidemiol. 2003;157(4):303–314. doi: 10.1093/aje/kwf208. [DOI] [PubMed] [Google Scholar]
  46. Nowak B. Contents and relationship of elements in human hair for a nonindustrialised population in Poland. Sci. Total Environ. 1998;209(1):59–68. doi: 10.1016/s0048-9697(97)00298-2. [DOI] [PubMed] [Google Scholar]
  47. Salinas S, Ingram JC. Uranium and arsenic characterization of well water and sediment in black falls, AZ. Abstr. Pap. Am Chem. Soc. 2005;229:U834. U834. [Google Scholar]
  48. Sauvé JF, Lévesque M, Huard M, Drolet D, Lavoué J, Tardif R, et al. Creatinine and specific gravity normalization in biological monitoring of occupational exposures. J. Occup. Environ. Hyg. 2015;12(2):123–129. doi: 10.1080/15459624.2014.955179. [DOI] [PubMed] [Google Scholar]
  49. Scheer J, Findenig S, Goessler W, Francesconi KA, Howard B, Umans JG, et al. Arsenic species and selected metals in human urine: validation of HPLC/ICPMS and ICPMS procedures for a long-term population-based epidemiological study. Anal. Methods. 2012;4(2):406–413. doi: 10.1039/C2AY05638K. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Stemmer KL. Pharmacology and toxicology of heavy metals: antimony. Pharm. Ther. A. 1976;1:157–160. [Google Scholar]
  51. Suwazono Y, Akesson A, Alfven T, Jarup L, Vahter M. Creatinine versus specific gravity-adjusted urinary cadmium concentrations. Biomarkers. 2005;10:117–126. doi: 10.1080/13547500500159001. [DOI] [PubMed] [Google Scholar]
  52. Tellez-Plaza M, Guallar E, Howard BV, Umans JG, Francesconi KA, Goessler W, et al. Cadmium exposure and incident cardiovascular disease. Epidemiology. 2013;24(3):421–429. doi: 10.1097/EDE.0b013e31828b0631. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. USEPA (U.S. Environmental Protection Agency) [accessed 01.05.15];Environmental radiation data. Report 128 October–December 2006. 2006 Available: ⟨ http://nepis.epa.gov/Exe/ZyPDF.cgi/P100H75Q.PDF?Dockey=P100H75Q.PDF⟩.
  54. USEPA (U.S. Environmental Protection Agency) [accessed 01.05.5];Private drinking water wells. 2015 Available: ⟨ http://water.epa.gov/drink/info/well/index.cfm⟩.
  55. USGS (U.S. Geological Survey) U.S. Geological Survey; [accessed 01.05.15]. 2000a. Arsenic in ground-water resources in the United States. Available: ⟨ http://pubs.usgs.gov/fs/2000/fs063-00/pdf/fs063-00.pdf⟩. [Google Scholar]
  56. USGS (U.S. Geological Survey) [accessed 01.05.15];A retrospective analysis on the ccurrence of arsenic in ground-water resources of the United States and limitations in drinking-water-supply characterizations. 2000b Available: ⟨ http://pubs.usgs.gov/wri/wri994279/pdf/wri994279.pdf⟩.
  57. USGS (U.S. Geological Survey) [accessed 01.05.15];Geochemical and mineralogical maps for soils of the conterminous United States. 2014 Available: ⟨ http://pubs.usgs.gov/of/2014/1082/pdf/ofr2014-1082.pdf⟩.
  58. Wang T, Fu J, Wang Y, Liao C, Tao Y, Jiang G. Use of scalp hair as indicator of human exposure to heavy metals in An electronic waste recycling area. Environ. Pollut. 2009;157(8-9):2445–2451. doi: 10.1016/j.envpol.2009.03.010. [DOI] [PubMed] [Google Scholar]
  59. Werner E, Roth P, Heinrichs U, Giussani A, Cantone MC, Zilker TH, et al. Internal biokinetic behaviour of molybdenum in humans studied with stable isotopes as tracers. Isot. Environ. Health Stud. 2000;36:123–132. doi: 10.1080/10256010008032938. [DOI] [PubMed] [Google Scholar]
  60. Yongming H, Peixuan D, Junji C, Posmentier ES. Multivariate analysis of heavy metal contamination in urban dusts of Xi'sn, central China. Sci. Total Environ. 2006;355(1–3):176–186. doi: 10.1016/j.scitotenv.2005.02.026. [DOI] [PubMed] [Google Scholar]
  61. Zheng L, Kuo CC, Fadrowski J, Agnew J, Weaver VM, Navas-Acien A. Arsenic and chronic kidney disease: a systematic review. Curr. Environ. Health Rep. 2014;1(3):192–207. doi: 10.1007/s40572-014-0024-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Zheng LY, Umans JG, Tellez-Plaza M, Yeh F, Francesconi KA, Goessler W, et al. Urine arsenic and prevalent albuminuria: evidence from a population-based study. Am. J. Kidney Dis. 2013;61(3):385–394. doi: 10.1053/j.ajkd.2012.09.011. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplement
NIHMS767794-supplement-supplement.doc (752KB, doc)

RESOURCES