DSpace at VNU: Occurrence of Perchlorate and Thiocyanate in Human Serum From E-Waste Recycling and Reference Sites in Vietnam: Association With Thyroid Hormone and Iodide Levels

DSpace at VNU: Occurrence of Perchlorate and Thiocyanate in Human Serum From E-Waste Recycling and Reference Sites in Vi...

Occurrence of Perchlorate and Thiocyanate in Human Serum

From E-Waste Recycling and Reference Sites in Vietnam:

Association With Thyroid Hormone and Iodide Levels

Akifumi Eguchi•Tatsuya Kunisue•

Qian Wu• Pham Thi Kim Trang•Pham Hung Viet•

Kurunthachalam Kannan•Shinsuke Tanabe

Received: 23 October 2013 / Accepted: 18 March 2014 / Published online: 10 April 2014

Ó Springer Science+Business Media New York 2014

Abstract Perchlorate (ClO4-) and thiocyanate (SCN-)

interfere with iodide (I-) uptake by the sodium/iodide

sym-porter, and thereby these anions may affect the production of

thyroid hormones (THs) in the thyroid gland Although

human exposure to perchlorate and thiocyanate has been

studied in the United States and Europe, few investigations

have been performed in Asian countries In this study, we

determined concentrations of perchlorate, thiocyanate, and

iodide in 131 serum samples collected from 2 locations in

Northern Vietnam, Bui Dau (BD; electrical and electronic

waste [e-waste] recycling site) and Doung Quang (DQ; rural

site) and examined the association between serum levels of

these anions with levels of THs The median concentrations

of perchlorate, thiocyanate, and iodide detected in the serum

of Vietnamese subjects were 0.104, 2020, and 3.11 ng mL-1,

respectively Perchlorate levels were significantly greater in

serum of the BD population (median 0.116 ng mL-1) than

those in the DQ population (median 0.086 ng mL-1), which

indicated greater exposure from e-waste recycling operations

by the former Serum concentrations of thiocyanate were not significantly different between the BD and DQ populations, but increased levels of this anion were observed among smokers Iodide was a significant positive predictor of serum levels of FT3and TT3and a significant negative predictor of thyroid-stimulating hormone in males When the association between serum levels of perchlorate or thiocyanate and THs was assessed using a stepwise multiple linear regression model, no significant correlations were found In addition to greater concentrations of perchlorate detected in the e-waste recycling population, however, given that lower concentra-tions of iodide were observed in the serum of Vietnamese females, detailed risk assessments on TH homeostasis for females inhabiting e-waste recycling sites, especially for pregnant women and their neonates, are required

Perchlorate (ClO4-) is an anionic compound, and its salts are used as oxidizing agents in rocket propellants, explo-sives, and fireworks and as dopant materials in the pro-duction of polyvinyl chloride (PVC) (Interstate Technology and Regulatory Council 2005) Perchlorate also occurs naturally in some fertilizers (Urbansky et al.2001), and is presumably generated in the atmosphere (Rao et al.2010)

It has been reported in the United States that such sources and high-hydrophilic property of perchlorate led to the widespread presence of this anion in the aquatic environ-ment and in drinking water (Blount et al 2010) In addi-tion, perchlorate has been detected in some food items and vegetables (Sanchez et al 2005) Therefore, studies per-formed in the United States have shown that perchlorate is found in various human bodily fluids such as urine (Blount

et al.2006), breast milk (Kirk et al.2007), saliva (Oldi and Kannan2009), and blood (Oldi and Kannan2009; Blount

A Eguchi
T Kunisue
S Tanabe

Center for Marine Environmental Studies, Ehime University,

Bunkyo-cho 2-5, Matsuyama 790-8577, Japan

T Kunisue ( &)

Faculty of Agriculture, Tottori University, 4-101

Koyama-minami, Tottori 680-8553, Japan

e-mail: kunisue@muses.tottori-u.ac.jp

Q Wu
K Kannan

New York State Department of Health and Department of

Environmental Health Sciences, School of Public Health,

Wadsworth Center, State University of New York at Albany,

P.O Box 509, Albany, NY 12201-0509, USA

P T K Trang
P H Viet

Center for Environmental Technology and Sustainable

Development, Hanoi University of Science, 334 Nguyen Trai,

Hanoi, Vietnam

DOI 10.1007/s00244-014-0021-y

et al.2009) Nevertheless, few studies on human exposure

to perchlorate are available in Asian countries where

per-chlorate salts are produced and used

It is known that thiocyanate (SCN-) is formed by way

of the detoxification process of hydrogen cyanide

con-tained in cigarette smoke (Tuncel et al.1994) This anionic

compound is also a metabolite of cyanogenic glucosides

present in plant foods such as cabbage, broccoli, and

mustard (VanEtten et al 1969; Foss and Lund-Larsen

1986; Bertelsen and Hegedus1994) Thus, it is thought that

intake of cigarette smoke and plant foods is a main

expo-sure source of thiocyanate for humans In Europe and the

United States, the detection of thiocyanate in human urine

and serum has been reported (Foss and Lund-Larsen1986;

Blount et al.2006)

Iodide (I-) is essential for the production of thyroid

hormones (THs) (Bianco et al 2002) Both perchlorate

and thiocyanate can competitively inhibit iodide uptake by

the thyroid gland (TG) by way of the sodium/iodide (Na?/

I-) symporter (NIS) consequently decreasing the synthesis

of tri-iodothyronine (T3) and thyroxine (T4) (Tonacchera

et al 2004; Dohan et al 2007) The United States

National Academy of Sciences and the United States

Environmental Protection Agency (USEPA) have adopted

an oral reference dose (RfD) of 0.7 mg perchlorate/kg

body weight (bw)/d based on a study of inhibition of

iodide uptake by perchlorate into the TG (Greer et al

2002; Zewdie et al 2010) Exposure to perchlorate was

associated with increased serum TSH levels in adult

women with urinary iodine levels \100 lg L-1 (Blount

et al.2006) It was also shown that concentrations of T3

and T4 decreased significantly in TG and serum of rats

coadministered with perchlorate and an iodide-deficient

diet (Kunisue et al 2010, 2011a) Although no RfD for

thiocyanate is available, an in vivo study has shown that a

single dose of thiocyanate to pregnant mice led to

decreased concentrations of free T3 and T4 and to

increased concentrations of TSH in plasma of their pups,

which was associated with decreased iodine levels in the

pup TG (Ghorbel et al.2008) Thus, the adverse effects of

these two anions on TH homeostasis can be exacerbated

under iodide-deficient conditions

In Vietnam, it was reported that urinary concentrations

of iodine in school-aged children were in the range of

20–49 lg L-1, \100 lg L-1, indicating insufficient iodide

intake by this population (de Benoist et al 2008;

Zim-mermann et al 2008) Considering the above-mentioned

information, it is likely that Vietnamese people may be at

high risk of TH disruption by perchlorate and thiocyanate;

however, human exposure to these anions has been never

investigated in this country In addition, primitive methods

of recycling of electrical and electronic wastes (e-wastes)

in Vietnam have raised considerable concern because

hazardous substances, such as heavy metals and persistent organohalogen compounds (Silicon Valley Toxics Coali-tion and Basel AcCoali-tion Network2002), can be released into the environment Our research group recently suggested that Vietnamese workers in an e-waste site are occupa-tionally exposed to polychlorinated biphenyls and poly-brominated diphenyl ethers during recycling activities (Tue

et al.2010) Given that perchlorate is contained in PVC and lithium-ion batteries as a dopant material, it is probable that residents and workers in such e-waste recycling sites are exposed to this anion

The present study was aimed at determining the serum concentrations of perchlorate and thiocyanate in residents

at an e-waste recycling site and a rural reference site in Northern Vietnam We examined the relationship between serum concentrations of THs and perchlorate/thiocyanate,

to assess the effects of these anions on TH homeostasis in the Vietnamese population

Materials and Methods

Chemicals and Devices

Ammonium perchlorate ([99.9 %) and methylamine (40 weight % solution in water) were purchased from Sigma-Aldrich (St Louis, Missouri, USA) Potassium iodide and thiocyanate solutions ([99.5 %) were from AccuStandard Inc (New Haven, Connecticut, USA) Isotopically labelled sodium perchlorate (Cl18O4-, [90 %) and potassium thiocyanate (S13CN-, [95 %) were purchased from Cam-bridge Isotope Laboratories (Andover, Massachusetts, USA) Vivaspin 2 centrifugal-filtration devices (CFDs) were obtained from Sartorius Stedim Biotech (Goettingen, Germany)

Samples

Human serum samples (n = 131) were collected from 50 males and 81 females age 10–64 years in Bui Dau (BD, e-waste recycling site, n = 83) and Duong Quang (DQ, rural site, n = 48), Vietnam, during January 2010 to Jan-uary 2011 (Fig.1) These 131 donors were informed beforehand about the purpose of the study at local gov-ernment health stations where volunteers registered their consent to participate, and they consented to participation

in our study All of the participants were randomly selected without arbitrary criteria

Informed consent was obtained from all 131 donors, and this study was approved by the Ethical Committee of Ehime University, Japan Age, body mass index (BMI), living site, and health conditions (smoking habits, dietary habits, and

pregnancy status) were recorded by interview with the

par-ticipants, and these data are listed in Table1 The interview

was performed by volunteers from the Center for

Environ-mental Technology and Sustainable Development, Hanoi

University of Science, according to a standardized

ques-tionnaire Whole blood samples were collected by a certified

physician, and serum samples were obtained by way of

centrifugation after heparin treatment, frozen in liquid

nitrogen, transported to Japan, and stored at -25°C or

-80°C in the Environmental Specimen Bank (es-Bank) of

Ehime University (Tanabe2006) until analysis

Analysis of Perchlorate, Thiocyanate, and Iodide

Perchlorate, thiocyanate, and iodide were analyzed according to the procedures reported previously (Oldi and Kannan2009; Zhang et al.2010) Each serum sample was thawed at room temperature, and an aliquot of 0.5 mL was transferred to a Vivaspin 2 CFD Two hundred microliters

of an internal standard mixture containing 0.1 ng (100 lL [1 ng mL-1]) of Cl18O4- and 10 ng (100 lL [100 ng mL-1]) of S13CN-and 300 lL of Milli-Q water were added The diluted sample was vortexed to

Fig 1 Map of Vietnam

showing serum sampling

locations

Table 1 Cohort characteristics

and demographics of the

Vietnamese populations from

e-waste recycling and rural sites

a No females were smokers

Characteristics and demographics

All participants (n = 131)

E-waste recycling workers (n = 83)

Residents in reference site (n = 48)

Fish consumption Freshwater fish

Marine fish

Meat consumption

Vegetable consumption

incorporate the internal standard into the sample matrix.

The Vivaspin 2 CFD was then centrifuged for 30 min at

4,0009g The filtrate was transferred to a sample vial for

the following instrumental analysis

Instrumental Analysis

The instrumental analysis was performed using an Agilent

1100 Series high-performance liquid chromatograph

(Ag-ilent Technologies, Santa Clara, California, USA) coupled

with a Micromass Quattro LC tandem mass spectrometer

(Waters, Milford, Massachusetts, USA) One hundred

microliters of the filtrate were injected using a Gilson 215

liquid handler (Gilson, Middleton, Wisconsin, USA) and a

Gilson 819 injection module equipped with a 100-ll

injection loop Separation of perchlorate, iodide, and

thiocyanate in the sample was performed using an IonPac

AS-21 column (guard column 50 mm 9 2 mm, regular

column 250 mm 9 2 mm; Dionex, Sunnyvale, California,

USA) An isocratic mobile phase of 200 mM aqueous

methylamine was used at a flow rate of 0.3 mL/min

Electrospray negative ionization was employed in the

multiple reaction monitoring mode with the following mass

transitions to identify and quantify perchlorate, Cl18O4-,

thiocyanate, S13CN-, and iodide: 99 (35ClO4-) [ 83

(35ClO3-), 101 (37ClO4-) [ 85 (37ClO3-), 107

(35Cl18O4-) [ 89 (35Cl18O3-), 58 (SCN-) [ 58 (SCN-),

59 (S13CN-) [ 59 (S13CN-), and 127 (127I-) [ 127

(127I-) Relative responses of the native standard to the

isotope-labeled internal standard and the ratios of35Cl to

37

Cl (for perchlorate) were used for the confirmation of

target analytes The ratios (35Cl:37Cl) were considered

acceptable at 3.12 (± 25 %) Data acquisition and

calcu-lation were accomplished using Micromass MassLynx 3.5

(Waters)

Quality Assurance and Quality Control

Recoveries of internal standards Cl18O4-and S13CN-spiked

into serum samples (n = 131) were in the range of 60–110 %

and 88–135 %, respectively A 10-point calibration standard

(in Milli-Q water) comprising concentrations ranging from

0.01 to 50 ng mL-1for perchlorate, 2–2,000 ng mL-1for

thiocyanate, and 0.02–100 ng mL-1for iodide were injected

with each batch of 30 samples The regression coefficients

for calibration curves were[0.999 for all target analytes A

laboratory reagent blank and an instrument blank were run

with each batch of samples Blanks contained trace levels of

iodide, which were subtracted from concentrations of iodide

detected in samples The limits of quantitation (LOQs) for

perchlorate, thiocyanate, and iodide in serum samples were

0.05, 0.5, and 0.5 ng mL-1, respectively

Perchlorate-Equivalent Concentration

An in vitro study using Chinese hamster ovary cells express-ing human NIS reported that the relative potency of ClO4-to inhibit125I-uptake by the NIS was 15 times greater than that

of SCN-on a molar basis (Tonacchera et al.2004) Based on the potency factor, we calculated perchlorate-equivalent concentrations (PECs) in serum sample, using the following formula: PEC (lmol L-1) = [ClO4-] (lmol L-1) ? [SCN-] (lmol L-1)/15

TH Analysis

Concentrations of THs in serum were analyzed by electro-chemiluminescence immunoassay method as described in our previous study (Kunisue et al.2011b) TSH, total T3(TT3), total T4 (TT4), free T3(FT3), and free T4(FT4) were measured using Elecsys kits (Roche Diagnostics, Mannheim, New York, USA) and Modular Analytics E170 systems (Hitachi, Tokyo, Japan) The expected reference intervals in euthyroid humans are within 0.270–4.2 lIU mL-1 for TSH, 0.80 –2.0 ng mL-1 for TT3, 45–117 ng mL-1 for TT4, 2.0 –4.4 pg mL-1for FT3, and 9.7–17 pg mL-1for FT4(Roche Diagnostics GmbH2008)

Statistical Analysis

Statistical analyses were performed using R program Version 2.15.1 with its graphical user interface EZR (Saitama Medical Center, Jichi Medical University, The R Foundation for Statistical Computing, version 2.13.0) (Kanda2013) This interface is a modified version of the R commander (version 1.6-3) and is designed to add fre-quently used biostatistical functions Steel–Dwass test (Kruskal–Wallis post hoc test) was used to assess differ-ences in serum concentrations of anions and THs between the residents of e-waste recycling and reference sites or among females, nonsmoking males, and smoking males Spearman’s rank correlation coefficients were calculated to assess the strength of relationships between serum con-centrations of each anion and age or BMI of the donors Associations between serum levels of THs and perchlorate, thiocyanate, or iodide were assessed using a stepwise multiple linear regression model The normality of distri-bution for each parameter was assessed by Shapiro–Wilk test Although the data for TT4 and FT4 were normally distributed in serum, the concentrations of TSH, TT3, FT3, perchlorate, thiocyanate and iodide were not normally distributed; hence, these data were log-transformed for the regression analyses Characteristics, such as age, residen-tial area, BMI, pregnancy status, and dietary habits (meat, freshwater, and marine fish consumption), which can

influence TH levels, were used as explanatory variables A

p-value of \0.05 denoted significance

For any model, the parameters that optimize the

approximation of the likelihood can be found numerically

Following this, the optimized likelihoods from different

models can be compared through Akaike’s information

criterion (AIC) as follows: AIC = 2k-2ln L In this

model, k is the number of parameters, and L is the

maxi-mized likelihood The model with the smallest AIC was

selected, thus providing a tradeoff between model

com-plexity (preferring models with fewer parameters) and the

maximized likelihood of the model

Results

Perchlorate

Perchlorate was detected in 129 of 131 serum samples with

the concentrations ranging from 0.050 to 1.25 ng mL-1

(median 0.104 ng mL-1), whereas the levels in two donors

were lower than the LOQ (Table2) Serum concentrations

of perchlorate in BD (e-waste recycling site) residents

(median 0.116 ng mL-1) were significantly greater

(p \ 0.05) than those in DQ (rural site) residents (median

0.086 ng mL-1) No significant relationships were found

between serum concentrations of perchlorate and smoking,

dietary habits, sex, age, or BMI (Table2, Fig.2)

Thiocyanate

Thiocyanate was found in all 131 serum samples analyzed

(median 2,020 ng mL-1) (Table2) Significantly greater

concentrations of thiocyanate were found in males (median

2,850 ng mL-1) than in females (median 1,760 ng mL-1)

It is of note among both BD and DQ residents, male

smokers had significantly greater thiocyanate levels than

did male nonsmokers (all females were nonsmokers)

(Fig.3) No significant associations were found between

thiocyanate levels and location of sampling, dietary habits,

or BMI, but an age-dependent increase in serum

concen-trations of thiocyanate was observed (Fig.2)

Iodide

Iodide was also found in all 131 serum samples analyzed in

this study (median 3.11 ng mL-1) (Table 2) Significantly

greater (p \ 0.05) concentrations of iodide were found in

the serum of males (median 3.38 ng mL-1) and those who

consumed less fish (median 3.27 ng mL-1) compared with

those of females (median: 2.91 ng mL-1) and frequent fish

eaters (median 2.49 ng mL-1) Other demographic

char-acteristics, such as smoking, meat consumption, age, and

BMI, were not significantly associated with serum iodide levels

PEC

The PECs calculated from serum concentrations of per-chlorate and thiocyanate in the Vietnamese populations were in the range of 0.365–23.1 lmol L-1 (median 2.28 lmol L-1) (Table2), and thiocyanate levels accoun-ted for [99 % of the PEC values Therefore, the associa-tions between serum PECs and demographic characteristics

of donors were identical with the trends observed for thiocyanate levels described previously, i.e., greater PECs were found in males and smokers than in females and nonsmokers

Association of Serum TH Levels With Perchlorate, Thiocyanate, and Iodide Concentrations

Concentrations of THs measured in Vietnamese sera are listed in Table3 Serum concentrations of FT3and TT3in

BD residents (e-waste site) were significantly lower (p \ 0.05) than those in DQ residents (Table3) To examine the relationship between TH levels and concen-trations of perchlorate, thiocyanate, and iodide in sera, we used a stepwise multiple linear regression model Because serum concentrations of thiocyanate and iodide differed significantly between males and females, as described previously, subsequent analyses were separated by sex The multilinear regression analyses showed that iodide was a significant positive predictor of FT3 (p \ 0.01) and TT3 (p \ 0.01) and a significant negative predictor of TSH (p \ 0.01) in males (Table4) In contrast, no significant associations between serum concentrations of THs and perchlorate, thiocyanate, or PEC were found for either males or females

Discussion

Perchlorate

To our knowledge, this is the first study to determine serum levels of perchlorate in Vietnamese populations Interest-ingly, serum concentrations of perchlorate in BD residents were significantly greater than those in DQ residents (Table2), and the measured concentrations were compa-rable with those reported for Albany and New York City, United States (Oldi and Kannan2009); however, the levels

in BD residents were relatively lower than those reported for New Jersey and Lansing, United States (Oldi and Kannan 2009; Blount et al 2009) (Fig.4) These results indicate the presence of perchlorate sources in the e-waste

Table 2 Serum concentrations

of perchlorate, thiocyanate, and

iodide and PEC and selected

demographic characteristics

Concentrations and characteristics

Perchlorate (ng mL-1)

Thiocyanate (ng mL-1)

Iodide (ng mL-1)

PEC (lmol L-1) Total (n = 131)

Location

DQ (rural) (n = 48)

BD (e-waste (n = 83)

Sex Female (n = 81)

Male (n = 50)

Habit of eating marine fish

0 times/week (n = 119)

1–3 times/week (n = 12)

Habit of eating freshwater fish 1–3 times/week (n = 58)

recycling site and the specific exposure of BD residents to

perchlorate Given that perchlorate is used in PVC and

lithium-ion batteries as a dopant material (Interstate

Technology and Regulatory Council 2005), BD residents

might be exposed to relatively high levels of this anion

during e-waste recycling operations

Based on the serum concentrations of perchlorate

detected in this study, we estimated the perchlorate

expo-sure dose for the Vietnamese populations, using the

fol-lowing equation, developed by Gibbs (2006), as follows:

Log E = [log(C/99.5) -1.052]/0.9193,

where E is the estimated exposure dose (mg/kg bw/d), C

is the serum perchlorate level (lg L-1), and 99.5 is the

molar mass of perchlorate The median exposure doses of

perchlorate estimated for DQ and BD residents were

0.033 lg/kg/d and 0.048 lg/kg/d, respectively, and the

values were 1 order of magnitude lower than the USEPA

RfD of 0.7 lg/kg/d (Zewdie et al 2010) However, the

estimated exposure dose (0.81 lg/kg/d) for an e-waste

recycling worker, who had the highest serum concentration

of perchlorate, was comparable with the USEPA RfD,

indicating that some individuals might be at risk from

perchlorate exposure Further studies are needed to identify

hot spots in the e-waste recycling site and exposure sources

of perchlorate for BD residents in Vietnam

Thiocyanate

Greater serum concentrations of thiocyanate in males than

in females were found (p \ 0.05) Especially at both the

BD and DQ sites, male smokers had greater levels of this anion than did male nonsmokers and females (all females were nonsmokers) (Fig.3), suggesting that smoking is a major exposure source of thiocyanate In previous Nor-wegian (Foss and Lund-Larsen 1986) and Danish (Laur-berg et al 2004) studies, greater serum concentrations of thiocyanate in smokers than in nonsmokers have been reported (Fig.5) In fact, it is known that thiocyanate is formed by way of the detoxification process of the hydrogen cyanide contained in cigarette smoke (Tuncel

et al.1994) In this study, a significant positive correlation between serum thiocyanate levels and donor age was observed (Fig.2) However, this correlation no longer became significant after exclusion of data for smokers (p = 0.24), whereas serum thiocyanate levels in smokers were positively correlated with age (p \ 0.05) (data not shown) Thus, it is probable that the age-dependent increase in serum concentrations of thiocyanate observed

in this study (Fig 2) is attributed to the smoking habit

As for nonsmokers, serum concentrations of thiocyanate were comparable with the data reported from Norway (Foss

Table 2 continued

Qu Quartile, Max Maximum,

Min Minimum

a Lower than detection limit

b Significantly greater than

rural site

c Significantly greater than

females

d Significantly greater than

marine fish eater

e Significantly greater than

freshwater fish eater

*p \ 0.05

**p \ 0.01

Concentrations and characteristics

Perchlorate (ng mL-1)

Thiocyanate (ng mL-1)

Iodide (ng mL-1)

PEC (lmol L-1)

[4 times/week (n = 73)

Habit of eating meat 1–3 times/week (n = 24)

[6 times/week (n = 107)

and Lund-Larsen 1986) and Denmark (Laurberg et al.

2004) and were greater than those reported from the United

States (Blount et al.2009) (Fig.5) Given that thiocyanate

is also a metabolite of cyanogenic glucosides present in

plant foods such as cabbage, broccoli, and mustard

(Va-nEtten et al 1969, 1976; Foss and Lund-Larsen 1986;

Bertelsen and Hegedus 1994), the differences in serum

concentrations of this anion observed between Vietnamese

and American nonsmokers may be related to the frequency

of vegetable ingestion

Iodide

Iodide concentrations found in Vietnamese serum samples

ranged from 1.26 to 13.1 ng mL-1 (median 3.11 ng mL-1)

(Table2) Normal range of serum concentrations of iodine in healthy subjects were defined to range from 50 to

100 ng mL-1, and approximately 5 % (2.5–5 ng mL-1) of this form in the serum are thought to be inorganic iodine (i.e.,

95 % as organic iodine derivatives) (Wagner et al.1961; Fisher

et al.1965; Nagataki et al.1967; Sternthal et al.1980) Iodide concentrations in 71.7 % of serum samples (n = 94) analyzed

in this study were[2.5 ng mL-1, but the levels in 28.2 % samples (n = 37) were less than the reference value The World Health Organization (World Health Organization, United Nations International Children’s Emergency Fund, International Council for Control of Iodine Deficiency Disor-ders2001) has reported that the optimal values for a popula-tion’s urinary iodine levels is in the range of 100 and

200 ng mL-1 In Vietnam, however, it was recently reported

0 5000 10000 15000 20000

r = 0.23

p = 0.008

2 4 6 8 10 12

r = 0.09

p = 0.32

0.0 0.2 0.4 0.6 0.8 1.0 1.2

r = 0.17

p = 0.055

0 5000 10000 15000

20000

r = 0.13

p = 0.13

2 4 6 8 10 12

r = 0.01

p = 0.99

0.0 0.2 0.4 0.6 0.8 1.0 1.2

r = 0.08

p = 0.24

-1 )

-1 )

-1 )

I M B )

s r y ( e g A

Fig 2 Relationships between

serum anion levels and age or

BMI of Vietnamese subjects

that urinary concentrations of iodine in school-age children

were in the range of 20–49 ng mL-1(de Benoist et al.2008;

Zimmermann et al.2008) These observations indicate that

Vietnamese people consume a relatively low amount of iodide

In the present study, serum concentrations of iodide in

females were significantly lower than those in males

(Table2) In fact, serum iodide levels in 35.8 % of females

(n = 28) were \2.5 ng mL-1 It has been reported that

females generally consume less (iodized) salt than do

males (Brown et al.2009), and hence the sex difference in

iodide levels observed in this study may be attributed to the

consumption of iodized salt, although no such information

on salt intake is available for Vietnam Considering that serum iodide levels in the Vietnamese population were not associated with meat consumption and were greater in those who ate less fish (Table2), the intake of iodide-enriched food items, e.g., seaweed (Tokudome et al.2004) other than meat and fish may be less among Vietnamese females

PEC

PECs calculated from serum concentrations of perchlorate and thiocyanate among Vietnamese subjects ranged from 0.365 to 23.1 lmol L-1, and thiocyanate accounted for [99 % of the PEC values PECs observed in 84 % of the subjects (n = 110) exceeded the IC50 (1.22 lmol L-1) values reported for inhibition of iodide uptake in an in vitro study using Chinese hamster ovary cells expressing human NIS (Tonacchera et al.2004) Recently, in the calculation

of PECs, Bruce et al (2013) adopted a relative potency of 17.6 for thiocyanate When we recalculated the PECs using this value (17.6), those in 79 % of Vietnamese subjects (n = 103) were greater than the IC50(1.22 lmol L-1) These results imply that the PEC in Vietnamese popu-lations can inhibit iodide uptake into the TG An epide-miological survey suggested that serum PEC values 6.7 lmol L-1were associated with adverse effects on TH homeostasis in humans under iodine-deficient conditions (Gibbs 2006) Such differences in the threshold values estimated between in vitro and in vivo studies are mainly

0

5000

10000

15000

a

a

a a

b

b

-1 )

Fig 3 Concentrations of thiocyanate in the serum from Vietnamese

females (nonsmokers), male nonsmokers, and male smokers

Differ-ent letters (a and b) on the top of each column represDiffer-ent the statistical

differences by Steel–Dwass test (p \ 0.05)

Table 3 Concentrations of

thyroid hormones in

Vietnamese serum analyzed in

this study

Qu Quartile, Max Maximum,

Min Minimum

a Significantly lower than rural

site

*p \ 0.05

**p \ 0.01

Cohort TT3(ng mL-1) TT4(ng mL-1) FT3(pg mL-1) FT4(ng mL-1) TSH (lIU mL-1) Total (n = 131)

Location Rural (n = 48)

E-waste (n = 83)

due to the differences in kinetics of perchlorate and

thio-cyanate Approximately 12 % of the Vietnamese subjects

(n = 16) analyzed in this study had the PECs greater than

the threshold value of 6.7 lmol L-1 reported by Gibbs

(2006)

Although nitrate was not analyzed in this study, it was

reported in the National Health and Nutrition Examination

Survey subjects (Suh et al 2013; Bruce et al 2013) that

this anion might have affected thyroid NIS status Suh et al

(2013) showed that urinary nitrate was a significantly

negative predictor of serum free T4, and Bruce et al (2013)

reported that nitrate accounted for [75 % of PEC values calculated from urinary perchlorate, thiocyanate, and nitrate levels Thus, the effect of exposure of Vietnamese populations to perchlorate, thiocyanate, and nitrate on TH homeostasis is a concern and should be the subject of further investigation

Associations of Serum TH Levels With Perchlorate, Thiocyanate, and Iodide Concentrations

Associations between serum levels of THs and perchlorate and related anions were assessed using a stepwise multiple linear regression model No significant associations between serum concentrations of THs and PECs were found for either males or females (Table4) The multiple regression analyses for the concentrations of perchlorate and thiocyanate individually yielded similar results (p [ 0.05) These observations indicated that serum con-centrations of perchlorate and thiocyanate in most of the Vietnamese donors are lower than the levels that would affect TH homeostasis As described earlier, PECs in 87.8 % of the Vietnamese serum samples (n = 115) were lower than the threshold value of 6.7 lmol L-1 (Gibbs 2006) In addition, an epidemiological study reported sig-nificant upregulation of TSH and downregulation of TT4in the serum of a population exposed to thiocyanate (mean serum concentration of thiocyanate 13,000 ng mL-1)

Table 4 Association coefficients (b) between serum concentrations of anions and THs in the Vietnamese populations by single and multiple linear regression models

Unadjusted ba Adjusted bb Unadjusted b Adjusted b Unadjusted b Adjusted b Unadjusted b Adjusted b Female (n = 81)

Male (n = 50)

Blank cell Data omitted by stepwise procedure

a Unadjusted coefficients were calculated by single regression analysis

b Coefficients were adjusted by age, BMI, pregnancy status (yes/no), living site (e-waste/reference), and habit of eating freshwater fish (1–3 times/week vs 4 times/week), marine fish (0 times/week vs 1–3 times/week), and meat (1–3 times/week vs [4 times/week)

c Values of TT3, FT3,TSH, perchlorate, thiocyanate, iodide, and PEC were log-transformed

* p \0.05

** p \0.01

Fig 4 Comparison of serum concentrations of perchlorate in

Viet-namese subjects analyzed in this study with previous United States

data 1 = Oldi and Kannan (2009); 2 = Blount et al (2009)