Environment International 106 (2017) 127–134

Contents lists available at ScienceDirect

Environment International
journal homepage: www.elsevier.com/locate/envint

Associations between toxic and essential trace elements in maternal blood MARK
and fetal congenital heart defects
Yanqiu Oua,1, Michael S. Bloomb,1, Zhiqiang Niea,1, Fengzhen Hanc, Jinzhuang Maia, Jimei Chend,
Shao Linb, Xiaoqing Liua,⁎,2, Jian Zhuangd,⁎,2
a
Department of Epidemiology, Guangdong Cardiovascular Institute, Guangdong Provincial Key Laboratory of South China Structural Heart Disease, Guangdong General
Hospital, Guangdong Academy of Medical Sciences, Guangzhou, China
b
Departments of Environmental Health Sciences and Epidemiology and Biostatistics, University at Albany, State University of New York, Rensselaer, NY, USA
c
Department of Obstetrics, Guangdong General Hospital, Guangdong Academy of Medical Sciences, Guangzhou, China
d
Department of Cardiac Surgery, Guangdong Cardiovascular Institute, Guangdong Provincial Key Laboratory of South China Structural Heart Disease, Guangdong General
Hospital, Guangdong Academy of Medical Sciences, Guangzhou, China

A R T I C L E I N F O A B S T R A C T

Keywords: Prenatal exposure to toxic trace elements, including heavy metals, is an important public health concern. Few
Heart anomalies studies have assessed if individual and multiple trace elements simultaneously affect cardiac development. The
Teratogen current study evaluated the association between maternal blood lead (Pb), cadmium (Cd), chromium (Cr),
Trace elements copper (Cu), mercury (Hg), and selenium (Se) levels and congenital heart defects (CHDs) in offspring. This
Risk factor
hospital-based case-control study included 112 case and 107 control infants. Maternal peripheral blood draw
Environment
was made during gestational weeks 17–40 and used to determine trace element levels by inductively coupled
plasma mass spectrometry. Multivariable logistic regression was used to assess associations and interactions
between individual and multiple trace elements and fetal CHDs, adjusted for maternal age, parity, education,
newborn gender, migrant, folic acid or multivitamin intake, cigarette smoking, maternal prepregnancy body
mass index, and time of sample collection. Control participants had medians of 2.61 μg/dL Pb, 1.76 μg/L Cd,
3.57 μg/L Cr, 896.56 μg/L Cu, 4.17 μg/L Hg, and 186.47 μg/L Se in blood. In a model including all measured
trace elements and adjusted for confounders, high levels of maternal Pb (OR = 12.09, 95% CI: 2.81, 51.97) and
Se (OR = 0.25, 95% CI: 0.08, 0.77) were harmful and protective predictors of CHDs, respectively, with positive
and negative interactions suggested for Cd with Pb and Se with Pb, respectively. Similar associations were
detected for subgroups of CHDs, including conotruncal defects, septal defects, and right ventricle outflow tract
obstruction. Our results suggest that even under the current standard for protecting human health (10 μg/dL), Pb
exposure poses an important health threat. These data can be used for developing interventions and identifying
high-risk pregnancies.

1. Introduction of epidemiological literature has shown that a multitude of noninher-

Congenital heart defects (CHDs), structural abnormalities of the
cardiovascular system characterized by anatomical departures of the
heart and great vessels from normal, are the most common birth defects
in newborns, affecting 6 to 12 per 1000 live births worldwide (Donofrio
et al., 2014). CHDs remain the leading cause of morbidity, mortality,
and disability in neonates and children, associated with lifelong
physical and mental comorbidities, and involving substantial financial
expenses (Donofrio et al., 2014). Over the past decade, a growing body
ited risk factors may increase the risk of CHDs, including maternal
illnesses like pregestational diabetes, maternal medication exposures,
environmental exposures, and paternal exposures, while periconcep-
tional multivitamin or folic acid intake may reduce the risk (Donofrio
et al., 2014). Environmental pollutants, including toxic trace elements
such as heavy metals, may play a contributing role (Di Renzo et al.,
2015).
Heavy metal pollution in China and trace element deficiencies
constitute an important public health concern. Lead (Pb) exposure is

⁎
Corresponding authors at: Guangdong Cardiovascular Institute, Guangdong Provincial Key Laboratory of South China Structural Heart Disease, Guangdong General Hospital,
Guangdong Academy of Medical Sciences, 96 Dongchuan Road, Guangzhou 510080, China.
E-mail addresses: drxqliu@163.com (X. Liu), drzhuangjian5413@163.com (J. Zhuang).
1
Yanqiu Ou, Michael S. Bloom and Zhiqiang Nie contributed equally to this paper.
2
Jian Zhuang and Xiaoqing Liu contributed equally to this paper.

http://dx.doi.org/10.1016/j.envint.2017.05.017
Received 18 December 2016; Received in revised form 19 May 2017; Accepted 19 May 2017
Available online 20 June 2017
0160-4120/ © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
Y. Ou et al.
widespread. The public may be exposed to Pb through contaminated
food, water, house dust, and through industrial activities such as metal
recycling and the battery industry (Barbosa et al., 2005). In Guangdong,
daily exposure also exists through consumption of cultured fish, rice
paddy soil, and atmospheric pollution (Liang et al., 2016; Yang et al.,
2004; Zeng et al., 2016). Maternal exposure to Pb during pregnancy has
been associated with a higher probability for congenital abnormalities
among offspring (Liu et al., 2015). Exposure to other trace elements
were also reported to increase the likelihood for congenital abnormal-
ities, such as cadmium (Cd) and chromium (Cr), that are associated
with modern industrial processes, and are absorbed in significant
quantities from cigarette smoke (Thompson and Bannigan, 2008; Gil
et al., 2011). Exposure of experimental animals to oral or parenteral Cd
causes a wide range of embryo abnormalities, conditional on the stage
of exposure and the administered dose (Thompson and Bannigan,
2008). Selenium (Se) is an essential element, required for anti-oxidant
enzyme activities, and deficiency can occur with inadequate dietary
intake. Detectable traces of Se in drinking water were associated with a
lower frequency of any CHD than was observed among children
exposed to drinking water without detectable Se levels (Zierler et al.,
1988). Other findings further indicated that maternal Se deficiency
during pregnancy might contribute to neural tube defects (NTDs) in
offspring (Cengiz et al., 2004).
Like Cd, chromium (Cr) exposure mostly results from modern
industrial process and cigarette smoking (Gil et al., 2011). Human data
showed a suggestive but not necessarily causal association between higher
parental Cr exposure and increased CHDs, even at levels approaching but
not exceeding the 0.05 mg/L action level for protecting human health
(Goldberg et al., 1990). Copper (Cu) is an essential element required in
trace concentrations for normal physiology. Yet, the relation between
maternal plasma Cu concentrations and birth defects in offspring is
unclear as few data are available to address the risk. A small case-control
study reported higher serum Cu among the mothers of elective termina-
tions with NTDs, relative to matched live births without a defect (Cengiz
et al., 2004). Seafood consumption is the most important source of
mercury (Hg) exposure among non-occupationally exposed populations
(Wells et al., 2016). A suggestive association was found between higher
placental Hg and higher NTD risk in offspring from a Chinese population
(Jin et al., 2013).
The aforementioned elements cross the placenta and are thus
transferred from mother to a developing fetus (Caserta et al., 2013).
Possible biological mechanisms underlying their teratogenic effects
include oxidative damage to the DNA and modification of epigenetic
patterns (Hansen, 2006; Valinluck et al., 2004; Pilsner et al., 2009;
Ercal et al., 2001). Moreover, complex antagonistic interactions be-
tween Se and toxic trace elements have been reported in previous
studies. For example, Se may exert an antagonistic effect on Pb induced
expression of inflammatory factors and heat shock proteins genes
(Zheng et al., 2016), the cataractogenic effects of Hg may be offset by
Se (Lemire et al., 2010), and Se has been identified as a potential
countermeasure against Cd inducted toxicity (Zwolak and Zaporowska,
2012).
To date, biomarker based observational studies of gestational trace
element exposures and human CHDs remain sparse. There are even
fewer data available to assess the simultaneous effects of multiple trace
elements, and their interactions with respect to CHDs. Given the
widespread nature of the exposures and tantalizing results from the
few studies published to date, additional investigation is necessary. The
current study was designed to address the existing data gap with an
investigation of the individual and combined effects of Pb, Cr, Cd, Cu,
Hg, and Se, on the occurrence of CHDs among mother-infant pairs in
southeastern China.
128
Environment International 106 (2017) 127–134
2. Materials and methods
2.1. Study design and subjects
This is a hospital-based case-control study conducted from
December 2012 to September 2013 at one provincial general hospital
in China. Pregnant women receiving routine prenatal care in the
hospital obstetric department mostly reside in the Pearl River Delta
Region of Guangdong Province. Both pregnant women treated by the
hospital's obstetrics department as well as mothers treated at other
hospitals in the Pearl River Delta Region are referred to the Department
of Fetal Echocardiography in our institute if a CHD is suspected
following an obstetric ultrasound examination. In 2001, the cardiovas-
cular institute reported 92% sensitivity and 95% specificity for
echocardiogram detection of fetal CHDs (Pan et al., 2001). Consecutive
singleton fetuses diagnosed with defined structural cardiac defects were
recruited to the study at the time of the prenatal diagnosis, between 17
and 40 weeks. Among 1379 fetuses examined with fetal echocardio-
graphy, 170 of them were diagnosed with CHDs after review by two
physicians specialty-trained in echocardiography. We excluded fetuses
with: 1) only mitral or tricuspid valvular lesions or enlarged atrial
septal defects (n = 6); 2) chromosomal malformations or syndromes, or
CHDs co-existing with ex-cardiac birth defects (n = 7); 3) self-reported
family history of CHDs (n = 4); 4) diabetes (n = 8); 5) multiple
gestations (n = 8); and 6) prenatal CHD diagnosis not confirmed after
delivery (n = 3). Of 134 eligible case 112 (84%) agreed to participate
in our study. Each CHD case was matched by gestational age
( ± 3 months) and maternal age ( ± 5 years) to a control mother
without diabetes or a family history of CHDs, who was undergoing
routine prenatal consultation for a normal singleton fetus randomly
selected from the obstetrics clinic, on the same day. Of 119 eligible
controls, 107 agreed to participate in our study; a participation rate of
90%, which was not differ from that of cases.
All live births with a prenatal CHD diagnosis received B-mode
echocardiographic examination to confirm the diagnosis after delivery.
Live births without a prenatal CHD diagnosis were routinely evaluated
before discharge to determine whether further investigation for a
potential CHD was needed. At least two senior pediatric cardiologists
confirmed each CHD case diagnosis after delivery, using computed
tomography, cardiac catheterization, surgery, or autopsy (i.e., for cases
of stillbirth and elective termination) as clinically necessary. CHD cases
were coded based on the International Classification of Diseases version
10 (ICD-10) with codes Q20-Q28.
2.2. Ethical approval
All procedures performed in studies involving human participants
were in accordance with the ethical standards of the Guangdong
General Hospital Human subjects committee (No. 2011120H) and with
the 1964 Helsinki declaration and its later amendments or comparable
ethical standards. Oral informed consent was obtained from the
mothers of each fetus prior to study enrollment.
2.3. Classification of CHD case subtypes
Cases were classified into five cardiac phenotypes after modification
based on prior reports (Leirgul et al., 2014), including: (i) conotruncal
defects (d-transposition of the great arteries [TGA], tetralogy of Fallot
[ToF], double outlet right ventricle [DORV], conoventricular ventricle
septum defect [VSD], truncus arteriosus, and interrupted aortic arch
[IAA] type B or type C); (ii) left ventricle outflow tract obstructions
[LVOTO] (coarctation of aorta [CoA], aortic valve stenosis [vAS], and
hypoplastic left heart syndrome [HLHS]); (iii) right ventricle outflow
tract obstructions [RVOTO] (hypoplastic right heart syndrome [HRHS],
tricuspid atresia, Ebstein anomaly, pulmonary atresia [PA] or valvular
PA, and pulmonary valve stenosis [vPS] with or without septal defects);
Y. Ou et al.
(iv) septal defects (VSD, atrial septum defect [ASD], and atrioventri-
cular septal defect [AVSD]); and (v) ‘other’ defects were classified into a
single group due to small sample sizes, including heterotaxia, pulmon-
ary venous return (APVR), other complex heart defects (e.g., single
ventricle, congenital corrected transposition of the great arteries), and
other specific heart defects (e.g., isolated valvular malformations not
classified as LVOTO or RVOTO and venous malformations).
2.4. Information collection
Information about maternal demographic factors, pregnancy his-
tory, family history of CHDs, perinatal diseases (i.e. diabetes), medica-
tion or supplement use, and fetal diagnosis was captured from the
clinical record. Case and control mothers completed a face-to-face
interviewer administered structured, standardized questionnaire to
collect information on maternal education, exposure to trace elements
or metals in the workplace, migrant status, life behavior and environ-
mental factors, prepregnancy body weight and height, and also
provided a blood sample. Two investigators double-entered and
examined the data for errors and accuracy prior to statistical analysis.
2.5. Blood samples collection, preparation, and analysis
Blood samples were collected from mothers with gestational ages
from 17 to 40 weeks. Maternal peripheral venous blood was collected
using trace-element-free syringes and transferred into a heparin coated
anticoagulant tube mixed by reversing the tube. All samples were
immediately frozen at − 80 °C and stored in the hospital central
laboratory for 3–15 months before shipment to KingMed Diagnostics
Company (Guangzhou, China) for trace elements analysis. Prior to
analysis, all samples were thawed, mixed thoroughly by vortex, and
subsequently diluted with 0.1% Triton X-100 and 0.1% HNO3 at a ratio
of 1:19. The diluted samples were stored for 5 min at 4 °C for analysis.
Whole blood Pb, Cr, Cd, Cu, total Hg, and Se concentrations were
determined by inductively coupled plasma mass spectrometry (ICP-MS)
using the Agilent 7700 × ICP-MS system (Agilent Technologies, Santa
Clara, CA, USA). Specimens were analyzed using external calibration
and an internal standardization procedure with monitoring for isotopes:
206
Pb, 207Pb, 208Pb, 114Cd, 111Cd, 52Cr, 65Cu, 202Hg, 77Se, and 82Se.
Seronorm™ Trace Elements Whole Blood L-2, a certified reference
material, was used to ensure quality control. Limits of detection
(LOD), in whole blood, were 0.39 μg/dL (Pb), 4.8 μg/L (Cr), 3.7 μg/L
(Cd), 26.4 μg/L (Cu), 2.8 μg/L (Hg), and 73.1 μg/L (Se). No trace
elements were measured below the LOD.
2.6. Statistical analysis
Variables were first characterized by their measures of center and
variability. For categorical factors, percentage was used to describe the
distributions and Chi-square or Fisher exact tests were used to compare
differences between groups as appropriate. Skewed continuous vari-
ables were presented as median (interquartile range [IQR]
[25th ~ 75th percentiles]), and compared by the Wilcoxon-Mann-
Whitney U test. The distribution of all of the trace elements were also
shown in graphs by cases and controls. We also assessed the inter-
correlations among measured trace elements.
Because no information was available a priori to indicate a possible
dose threshold, trace element concentrations were divided into low,
middle and high levels according to tertiles of control distributions
prior to analysis (Supplementary materials, Table S1). The associations
between individual maternal trace elements as predictors and CHDs as
outcomes, including total CHDs and CHD subtypes, were assessed as
unadjusted and confounder-adjusted odds ratios (aOR) and their 95%
confidence intervals (95% CI) using separate logistic regression models.
We incorporated factors associated with exposure to trace elements and
the risk for CHDs, into a directed acyclic graph (DAG) (Supplementary
129
Environment International 106 (2017) 127–134
materials, Fig. S1) based on the literature (Wells et al., 2016; Aoki et al.,
2016; Hinwood et al., 2013; Liu et al., 2013). We retained maternal age,
parity, maternal education, newborn gender, migrant status, folic acid
or multivitamin intake, cigarette smoking, maternal prepregnancy body
mass index (maternal BMI), and gestational week at blood sample
collection as confounders in the regression models. Each was operatio-
nalized as a categorical variable except for maternal BMI and gesta-
tional week at sample collection, which were left as continuous. We
adjusted all models for the same set of confounders to allow for
interpretation of results in a uniform context. To examine associations
with a mixed exposure, we also constructed multi-element models,
simultaneously including all measured maternal trace elements as
predictors. In sensitivity analyses, we repeated all models using
continuous trace elements as predictors to assess the robustness of
our findings, and we also limited the analysis to a subsample of cases
and the controls residing in the Pearl River Delta Region to assess the
impact of selection bias.
We used three approaches to explore modifying effects among trace
elements and to assess interactions: 1) use of categorical variables; 2)
interaction on the multiplicative scale; and 3) interaction on the
additive scale. For the categorical approach, levels of Pb, Cd and Se
were re-categorized as ‘low’ (1st and 2nd tertiles) and ‘high’ (3rd
tertile). We analyzed effects as low Pb with high Se, low Pb with low Se,
high Pb with high Se, and high Pb with low Se. Likewise, we assessed
analogous effects between Pb and Cd, and Cd and Se. Pairwise
interactions were modeled on the multiplicative scale by introducing
cross-product terms into regression models and assessing significance of
the coefficients using the Wald statistic. For the additive scale, we
assessed interaction contrast ratios (ICR). The computations for ICR
was: ICR = OR11-OR10-OR01 + 1, where OR11 denotes co-exposure
(e.g. high Pb and low Se, high Pb and high Cd, or high Cd and low
Se), OR10 or OR01 denotes the presence of one exposure and the absence
of the other (Knol et al., 2011). An ICR above 1 indicates a super-
additive effect, whereas an ICR below 1 indicates a sub-additive effect.
Confidence intervals (95%) for ICRs were computed by the multivariate
delta method as ICR ± 1.96 √ var(ICR) (Nie et al., 2010).
Participants missing data for education (n = 6) were excluded from
multivariable models, leaving n = 213 analyzed. Data were analyzed
using SPSS (v.23.0; SPSS Inc., IBM, Chicago, IL USA) and SAS (v.9.3;
SAS Institute, Inc., Raleigh, NC USA). We used PASS v.13 (NCSS, LLC
Kaysville, UT USA) to identify, a priori, n = 94 cases needed to achieve
80% power for detecting different proportions of Pb exposure by case-
control status at α = 0.05 in single-pollutant models. Statistical tests
were two sided with a significance level of P < 0.05.
3. Results
3.1. Basic characteristics of the study participants
Cases were categorized (not mutually exclusive) as: n = 42 con-
otruncal defects; n = 62 septal defects; n = 16 LVOTOs; n = 32
RVOTOs; and n = 11 ‘other’ defects. Maternal basic characteristics
are presented in Table 1 compared by case-control status. Maternal
parity, education level, migrant status, folic acid or multivitamin
intake, and maternal smoking were significantly different between
the case and control groups. No mothers self-reported daily or occupa-
tional-contact with trace elements or metals.
3.2. Trace elements in maternal blood samples
Median (IQR) Pb [3.81 (3.13–4.84) vs. 2.61 (2.02–3.40) μg/dL,
P < 0.001] and Cd [2.25 (1.72–3.24) vs. 1.76 (1.35–2.55) μg/L,
P = 0.001] levels in cases were significantly higher than in controls,
respectively, while median (IQR) Cu [838.38 (778.90–936.98) vs.
896.56 (817.62–979.44) μg/L, P = 0.02], Hg [3.63 (2.73–4.82) vs.
4.17 (3.17–5.58) μg/L, P = 0.03], and Se [Se = 172.90
Y. Ou et al. Environment International 106 (2017) 127–134

Table 1 3.3. Maternal blood trace elements level and fetal CHDs
Descriptive characteristics of the study sample.
Associations between the trace elements and CHDs were first
Variable Case (n = 112) Control P-valueb
n (%) or median (n = 107) investigated using single-element regression models. The unadjusted
(IQR)a n (%) or median effect estimates are shown in the Supplementary materials (Table S3).
(IQR)a Multivariable models were adjusted for maternal age, parity, education,
newborn gender, migrant status, folic acid or multivitamin intake,
Maternal age (years)
≤ 25 31 (27.7) 20 (18.7) 0.47 cigarette smoking, maternal pre-pregnancy BMI, and gestational week
26–30 50 (44.6) 52 (48.6) at the time of sample collection. As shown in Table 3, compared with
31–35 26 (23.2) 29 (27.1) the low level of Pb, mothers in the middle and high levels had higher
≥ 36 5 (4.5) 6 (5.6) odds for total CHDs, with ORs of 6.95 (95% CI: 1.69, 28.62) and 14.46
Parity (95% CI: 3.59, 58.17), respectively. On the other hand, mothers with
Multiparous 42 (37.5) 15 (14.0) < 0.001 high level Se bore the lowest odds for total CHDs [OR = 0.35, 95% CI:
Primiparous 70 (62.5) 92 (86.0)
0.15, 0.81] vs. the low level. The middle level of Cr was also protective
Newborn gender for total CHDs [OR = 0.22, 95% CI: 0.08, 0.57], yet the high level was
Male 48 (42.9) 58 (54.2) 0.24
not associated. The association of the high level of Pb and Se was also
Female 54 (48.2) 41 (38.3)
Unknown 10 (8.9) 8 (7.5) present in subtypes of conotruncal defects, septal defects, and RVOTO.
However, the association of the middle level of Cr was not statistically
Educationc
≤ 12 years 65 (58.0) 18 (17.8) < 0.001
significant in CHD subtypes. For Cu, protective effects were statistically
> 12 years 47 (42.0) 83 (82.2) significant only for LVOTO defects (Supplementary materials, Table
S4).
Migrantd
Yes 23 (20.5) 11 (10.3) 0.04 To further elaborate on the associations detected between trace
No 89 (79.5) 96 (89.7) elements and CHDs, multi-element models were employed to simulta-
Folic acid or multivitamin intake neously incorporate all measured trace elements as predictors of CHDs,
No 14 (12.5) 5 (4.7) 0.04 adjusted for confounders. As shown in Table 4, with the multi-element
Yes 98 (87.5) 102 (95.3) models, associations for the high levels of Pb and Se maintained
Maternal occupation as manual worker statistical significance for total CHDs (Pb: OR = 12.09, 95% CI:2.81,
Yes 10 (8.9) 4 (3.7) 0.12 51.97; Se: OR = 0.25, 95% CI: 0.08, 0.77), conotruncal defect, septal
No 102 (91.1) 103 (96.3) defect, and RVOTO defects. Associations for the middle level of Hg was
Maternal smoking also statistical significance with RVOTO defects (OR = 5.20, 95% CI:
Yes 15 (13.4) 1 (0.9) < 0.001 1.03, 26.18), but not in total CHDs or other subtypes. Other elements
No 97 (86.6) 106 (99.1)
including Cd, Cr and Cu did not show any statistically significant
Living in a newly renovated home association. In a sensitivity analysis using trace elements as continuous
Yes 8 (7.1) 6 (5.6) 0.64
predictors (Supplementary materials, Table S5), higher blood Pb was
No 104 (92.9) 101 (94.4)
Maternal prepregnancy BMI 19.8 (18.4–21.3) 20.3 (19.0–20.7) 0.29
also associated with a higher odds for total CHDs, conotruncal defect,
(kg/m2) septal defect, and RVOTO defects. However, protective effects were
Gestational week at sample 27.0 (25.0–29.0) 28.0 (24.0–28.0) 0.26 significant only for RVOTO defects with higher continuous Se expo-
collection (weeks) sures. We also conducted a sensitivity analysis limited to cases (n = 96)
a
and controls (n = 93) residing in the nearby Pearl River Delta Region,
For categorical variables, percentage was used to describe the distribution; for
abnormal distributed continuous variables, median (IQR) was used to describe the
with similar results as for the overall analysis (Table S6).
distribution.
b
P-value was calculated with Chi-square or Fisher exact tests. 3.4. Interaction between Pb and Cd, and Se
c
n = 6 missing observations.
d
Migrant means the people were not permeant residents in the local region, and did
We tested pairwise interactions among maternal blood Pb, Cd, and
not have local medical insurance.
Se to assess for synergic or antagonistic effects on the occurrence of
CHDs. Maternal blood Pb, Cd, and Se levels were first dichotomized as
Table 2
Median (interquartile range) for blood trace element levels measured in cases and ‘low’ (the 1st and 2nd tertiles) and ‘high’ (the 3rd tertile), and then
controls (n =219 )a. cross-product terms entered into logistic regression models, adjusted for
confounders (Table 5). There was no significant interaction on the
Trace elements Case median (IQR) Control median (IQR) P-value multiplicative or additive scales for Pb and Se, Pb and Cd, or for Cd and
Pb (μg/dL) 3.81 (3.13–4.84) 2.61 (2.02–3.40) < 0.001 Se. The only significant results came from the use of categorical
Cd (μg/L) 2.25 (1.72–3.24) 1.76 (1.35–2.55) 0.001 variables. The positive associations for Pb or Cd and total CHDs was
Cr (μg/L) 3.63 (2.09–4.10) 3.57 (3.27–3.99) 0.16 stronger among low Se mothers than among high Se mother. Likewise,
Cu (μg/L) 838.38 (778.90–936.98) 896.56 (817.62–979.44) 0.02 having high Cd in addition to high Pb appeared to elicit a stronger
Hg (μg/L) 3.63 (2.73–4.82) 4.17 (3.17–5.58) 0.03
Se (μg/L) 172.90 (153.87–192.23) 186.47 (172.45–207.34) < 0.001
association on fetal cardiac development than high Pb alone, although
ICR confidence intervals included the null (Table 5).
Pb, lead; Cd, cadmium; Cr, chromium; Cu, copper; Hg, mercury; Se, selenium.
a
Wilcoxon-Mann-Whitney U test was used to compare the difference between groups. 4. Discussion

(153.87–192.23) vs. 186.47 (172.45–207.34) μg/L, P < 0.001] levels
were significantly lower in cases than in controls, respectively (Table 2
and Supplementary materials, Fig. S2). As shown Table S2 in the
Supplementary materials, we detected modest to moderate correlations
between Pb and Cd (r = 0.29, P < 0.001), Pb and Cu (r = - 0.14,
P = 0.04), Cd and Cu (r = − 0.23, P < 0.001), Cr and Cu (− 0.27,
P < 0.001), and Hg and Se (r = 0.57, P < 0.001).
To the best of our knowledge, this is the first study to identify
associations between CHDs in offspring and maternal exposure to
multiple trace elements during pregnancy as measured in blood.
Specifically, we observed that maternal blood Pb and Se were harmful
and protective predictors of total CHDs respectively. The findings were
consistent for CHD subtypes, including conotruncal defects, RVOTO,
and septal defects. Finally, our results suggest that mothers with low Se

130
Y. Ou et al. Environment International 106 (2017) 127–134

Table 3
Adjusted ORs (95% CI) for associations between maternal blood trace element levels and total congenital heart defects using multivariable single-element logistic regression modelsa.

Trace element Low level Middle level High level

n cases/controls n cases/controls OR (95%CI) n cases/controls OR (95%CI)

Pb 2/36 24/35 6.95 (1.69, 28.62) 86/36 14.46 (3.59, 58.17)

Cd 21/36 30/36 1.23 (0.54, 2.82) 61/35 1.49 (0.65, 3.40)
Cr 50/37 13/35 0.22 (0.08, 0.57) 49/35 1.07 (0.50, 2.30)
Cu 55/36 28/35 0.48 (0.21, 1.10) 29/36 0.63 (0.29, 1.39)
Hg 48/36 40/36 0.82 (0.38, 1.77) 24/35 0.52 (0.22, 1.22)
Se 63/36 29/36 0.58 (0.27, 1.26) 20/35 0.35 (0.15, 0.81)

Pb, lead; Cd, cadmium; Cr, chromium; Cu, copper; Hg, mercury; Se, selenium; OR: odds ratio; CI, confidence interval.
a
Low level serves as the reference group. Models were adjusted for maternal age (≤ 25; 26–30; 31–35 or ≥ 36 years old), parity (multiparous/primiparous), education (≤ 12 years/
> 12 years), newborn gender (male/female), migrant (yes/no), folic acid or multivitamin intake (yes/no), cigarette smoking (yes/no), maternal pre-pregnancy body mass index (kg/m2),
and time of sample collection (gestational week).

Table 4
Adjusted ORs (95% CI) for associations between maternal blood trace element levels and congenital heart defects using multivariable multi-element logistic regression modelsa.

Total defects (n = 112) Conotruncal defects (n = 42) Septal defects (n = 62) RVOTO (n = 32)

Middle level High level Middle level High level Middle level High level Middle level High level

Pb 5.89 (1.38, 25.20) 12.09 (2.81, 51.97) 12.89 (1.02, 162.94) 21.86 (1.70, 280.77) 4.18 (0.70, 24.90) 15.28 (2.63, 88.84) 1.10 (0.15, 8.28) 7.06 (1.06, 47.27)
Cd 1.28 (0.51, 3.24) 1.26 (0.48, 3.31) 0.56 (0.14, 2.14) 1.09 (0.29, 4.03) 2.33 (0.73, 7.49) 1.77 (0.54, 5.82) 0.53 (0.11, 2.49) 0.97 (0.21, 4.50)
Cr 0.24 (0.08, 1.69) 0.84 (0.36, 1.96) 0.44 (0.10, 1.89) 1.34 (0.40, 4.50) 0.19 (0.05, 1.74) 0.86 (0.31, 2.39) 0.30 (0.05, 1.67) 0.66 (0.15, 2.80)
Cu 0.57 (0.22, 1.47) 0.77 (0.31, 1.89) 0.50 (0.12, 2.15) 0.58 (0.16, 2.17) 0.81 (0.25, 2.63) 1.14 (0.38, 3.46) 0.74 (0.16, 3.38) 1.06 (0.21, 5.50)
Hg 1.36 (0.54, 3.48) 1.42 (0.45, 4.45) 1.62 (0.49, 5.37) 0.84 (0.17, 4.25) 1.15 (0.37, 3.53) 1.35 (0.37, 4.89) 5.20 (1.03, 26.18) 1.99 (0.26, 15.29)
Se 0.55 (0.22, 1.38) 0.25 (0.08, 0.77) 0.53 (0.15, 1.86) 0.13 (0.03, 0.68) 0.34 (0.11, 1.07) 0.20 (0.06, 0.71) 0.27 (0.06, 1.27) 0.10 (0.01, 0.72)

Pb, lead; Cr, chromium; Cd, cadmium; Cu, copper; Hg, mercury; Se, selenium; RVOTO, right ventricle outflow tract obstructions; OR: odds ratio.
The results for left ventricle outflow tract obstructions (LVOTO) was not shown because of limited space. There was no statistical significance for LVOTO.
a
Low level serves as the reference group. Models were adjusted for maternal age (≤ 25; 26–30; 31–35 or ≥ 36 years old), parity (multiparous/primiparous), education (≤ 12 years/
> 12 years), newborn gender (male/female), migrant (yes/no), folic acid or multivitamin intake (yes/no), cigarette smoking (yes/no), maternal prepregnancy body mass index (kg/m2),
and time of sample collection (gestational week).

or high Cd in addition to high Pb had inflated odds for CHDs relative to
mothers with elevated blood Pb only. These data add to the growing
concern for developmental toxicity with maternal Pb exposures at levels
lower than the Chinese reference level for child health equal to 10 μg/
dL (Li et al., 2015).
Trace element levels measured in our study population were
generally higher than reported by previous studies of pregnant women.
The median value of blood Pb among controls (2.61 μg/dL) far
exceeded that reported for a representative sample of pregnant U.S.
women in 2003–2004 (0.60 μg/dL); in fact, our lowest exposure
category (< 2.15 μg/dL) corresponded approximately to the 95th%
tile in that study (1.8 μg/dL) (Woodruff et al., 2011). Our mean Pb level
(2.90 μg/dL) was closer to that reported for 178 Spanish women with
occupational exposure to heavy metals (2.77 μg/dL) (Gil et al., 2011).
The median levels of blood Cd (1.76 μg/L) and Hg (4.17 μg/L) among
controls in our study also exceeded those reported for pregnant
Canadian women (0.20 μg/L for Cd and 0.56 μg/L for Hg) (Arbuckle
et al., 2016). For blood Se, the median level (186.47 μg/L) in our
control participants was higher than that reported for 230 primiparous
women (1.31 μmol/L, equal to 103.44 μg/L) participating in a British
preeclampsia prevention trial (Rayman et al., 2015).
4.1. Associations between maternal blood Pb and CHDs
Experimental neural crest cell ablation induces conotruncal anoma-
lies in mice (Porras and Brown, 2008) and high Pb concentrations
altered rat neural crest cell proliferation and differentiation in vitro
(Huang and Schneider, 2004), so it is plausible that high Pb levels could
affect conotruncal anomalies. Previous studies also described associa-
tions between maternal or parental lead exposure and heart defects in
offspring, including defects in APVR (Jackson et al., 2004), PA (Correa-
Villasenor et al., 1993), and single ventricle (Steinberger et al., 2002).
However, these prior investigations were limited by self-reported or
occupational exposure assessments. A widely validated biomarker that
integrates all sources of exposure and reflects body burden over
extended intervals in the context of ongoing exposures (Hu et al.,
2007), is likely to offer a stronger design for assessing associations with
congenital abnormalities. To our knowledge, only two previous studies
evaluated the effect of lead on birth defects using maternal exposure
biomarkers. In Fallujah, Iraq, adult tooth enamel Pb was 1.4 times
higher in the parents of children with birth defects than in the parents
of controls (Al-Sabbak et al., 2012). A hospital-based Chinese study of
n = 316 cases and n = 348 controls reported a 2–4 times higher odds
for carrying a CHD baby (including total CHDs, septal defects,
contruncal defect, RVOTO, and LVOTO) in association with higher
maternal hair Pb (Liu et al., 2015). In the current study, we detected
similar effects using maternal blood Pb as a biomarker of exposure,
although our sample size was smaller and effect estimates imprecise. In
contrast to the prior Chinese study, our results do not indicate an
association between maternal Pb exposure and LVOTO, possibly due to
different CHD subtypes considered, as well as differences in the
covariates included in regression models. Still, our overall data are
consistent with the results of previous studies using biomarkers of
exposure to assess associations between maternal Pb and fetal CHDs.
Lead readily crosses the placenta, transferred from the mother to the
developing fetus. Yet, the biological mechanisms underlying Pb′s
teratogenic effects are poorly elucidated. Excessive maternal Pb can
reduce total antioxidant capacity, which is regarded as a common
mechanism of teratogenesis (Hansen, 2006), and specifically for CHDs
(Hobbs et al., 2005). Oxidative damage to DNA could also impact
methylation patterns, possibly leading to inherited changes in gene
expression (Valinluck et al., 2004); in a prior study, prenatal Pb
exposure was inversely associated with genomic DNA methylation in
cord blood (Pilsner et al., 2009). In addition, blood Pb influenced
homocysteine level, which is also associated with reduced DNA
methylation (Pilsner et al., 2009). While it is tempting to speculate

131
Y. Ou et al. Environment International 106 (2017) 127–134

Table 5 4.4. Interactions between trace elements

Interaction effects among blood Pb, Cd, and Se on the odds for total congenital heart
defects using multivariable logistic regression (n = 219)a.
To our knowledge, there have been no previous studies of trace
Exposure OR (95% CI) P-value element interactions on CHDs. Antagonism between Se, Pb (Zheng
et al., 2016), and Cd (Zwolak and Zaporowska, 2012) has been well
Pb/Se category documented, and therefore Se status may have modified associations
Low Pb and high Se Reference –
between Pb and CHDs in our study. In addition, it has been proposed
Low Pb and low Se 2.03 (0.60, 6.89) 0.26
High Pb and high Se 3.47 (0.88, 13.65) 0.07 that co-exposure to Pb and Cd might exert a synergistic detrimental
High Pb and low Se 9.63 (2.83, 32.74) < 0.001 effect on reproductive capabilities (Gollenberg et al., 2010). Pb and Cd
Cross-Product term for interaction 1.37 (0.29, 6.46) 0.69 may act through a similar oxidative-stress related biologic mechanism
ICR for interaction on the additive scale 5.13 (− 2.32, 12.57) 0.18 (Ercal et al., 2001). Although not statistically significant, our results
Pb/Cd category suggested that co-exposure to higher Pb and Cd, or higher Pb and lower
Low Pb and low Cd Reference – Se, or higher Cd and lower Se, potentiated the odds for CHDs, although
Low Pb and high Cd 1.61 (0.57, 4.54) 0.37
possibly due to random chance. However, this null result may have
High Pb and low Cd 4.32 (1.86, 10.04) < 0.001
High Pb and high Cd 5.21 (2.25, 12.07) < 0.001 been due to insufficient power because of the relatively small sample
Cross-Product term for interaction 0.75 (0.19, 2.89) 0.67 size. Still, given the paucity of research on the joint effects of toxic
ICR for interaction on the additive scale 0.28 (− 4.10, 4.65) 0.90 elements for fetal cardiac development, further studies are warranted to
Cd/Se category more definitively interpret these findings.
Low Cd and high Se Reference –
Low Cd and low Se 1.55 (0.62, 3.88) 0.35 4.5. Strengths and limitations
High Cd and high Se 0.98 (0.28, 3.49) 0.98
High Cd and low Se 3.47 (1.29, 9.30) 0.01
Cross-Product term for interaction 2.29 (0.53, 9.88) 0.27 Our study offers several strengths, which lend validity to our results.
ICR for interaction on the additive scale 1.94 (− 0.50, 4.38) 0.12 We used a biomarker to capture maternal exposures during pregnancy,
and clinical records to capture demographic, pregnancy, and birth data,
Low Pb: 1st and 2nd tertiles of blood Pb level; high Pb: 3rd tertile of blood Pb level. which are not vulnerable to recall bias associated with differential
Low Se: 1st and 2nd tertiles of blood Se level; high Se: 3rd tertile of blood Se level.
reporting between case and control mothers. For each case, we matched
Low Cd: 1st and 2nd tertiles of blood Cd level; high Cd: 3rd tertile of blood Cd level.
OR: Odds ratio; CI, confidence interval; ICR: interaction contrast ratio.
controls on the same day as the diagnosis and from the same obstetrics
a
Adjusted for maternal age (≤25; 26–30; 31–35 or ≥ 36 years old), parity (multi- clinic to ensure representativeness of the sampling frame. We also
parous/primiparous), education (≤ 12 years/ > 12 years), newborn gender (male/fe- assessed for associations with additional trace elements measured in
male), migrant (yes/no), folic acid or multivitamin intake (yes/no), cigarette smoking maternal blood specimens, and adjusted for a panel of potentially
(yes/no), maternal prepregnancy body mass index (kg/m2), and time of sample collection
confounding variables.
(gestational week).
However, the results of this study are also limited by several
important factors. Firstly, our higher-tier hospital may experience
on a role for epigenetic changes in mediating Pb-CHD associations, to
differential referral patterns for residents of Guangdong Province living
our knowledge, no data have been reported to describe associations
outside of the Pearl River Delta, in which cases are referred from a large
between changing epigenetic patterns and CHDs. Additional research is
area, whereas controls mostly reside nearby; this might lead to a
warranted to assess whether prenatal Pb exposure influences cardiac
selection bias. However, we found similar results in a sensitivity
gene-specific DNA methylation profiles, and to characterize the impact
analysis limited to study participants (both cases and controls) residing
of different methylation profiles on cardiac development.
in the nearby Pearl River Delta area and so we believe any impact to
have been minimal. Secondly, our sample size was relatively small. In
4.2. Associations between Se and CHDs the subgroup analysis for Pb and conotruncal defects and LVOTO
defects, no cases were available for low level exposure, potentially
There is a paucity of evidence showing the contribution of Se to limiting statistical power to detect modest associations and interactions,
CHDs. However, several studies underscore the importance of Se to and producing the extremely wide confidence intervals that we
NTDs. In a prospective cohort study, serum Se was significantly lower in observed. However, we did find significant differences and consistent
mothers of infants with congenital malformations compared to the effect trends for Pb and Se, even with the limited sample size. Thirdly,
mothers of infants without the malformation (Hammouda et al., 2013). blood was collected from participants in mid- and late-pregnancy, after
Another study indicated that maternal Se deficiency during pregnancy the critical period for cardiac development during the 3rd-8th week of
might contribute to NTDs (Cengiz et al., 2004). Compelling evidence gestation (Sadler, 2004), which may have misclassified exposure for
suggests that although toxic at high doses, Se is essential at trace some participants. However, Pb measurements reflect not only recent
concentrations for antioxidant enzyme activities and normal fetal exposure, but also provide an indication for historic exposures as a
development (Martin et al., 2004). Changes in epigenetic patterning result of Pb mobilization from bone (Gulson et al., 2003). In subjects
were also associated with Se exposure in human studies and with without excessive Pb exposure, 45–75% of the blood Pb may reflect
animal models (Pilsner et al., 2011; Davis et al., 2000). bone stores (Barbosa et al., 2005). Arbuckle recently reported decreases
in blood Pb without changes in blood Cd over pregnancy (Arbuckle
et al., 2016). Mistry reported relatively stable Se levels across preg-
4.3. Associations between other trace elements and CHDs nancy, unless women are depleted (Mistry et al., 2012). So, exposure
misclassification due to mistimed specimen collection is likely to be
Although some association were found between Cr, Cu, or Hg and non-differential with respect to CHDs and bias effects estimates towards
total or certain subtypes of CHDs in the single-element or multi-element the null hypothesis. Furthermore, we matched on gestational age and so
multivariable models, these seemed to be chance findings given mobilized bone Pb is unlikely to have contributed differentially
inconsistencies throughout the analysis. Null associations for these between cases and control. Still, any exposure misclassification due to
elements with CHDs are consistent with several previous studies, which mistimed blood specimen collection is likely to have been similar
reported no associations between these trace elements and CHDs in between CHD cases and controls, which may have undermined our
animals or humans (Hansen, 2006; Ercal et al., 2001; Al-Sabbak et al., ability to detect modest associations for other trace elements with null
2012). results. Finally, we did not capture data for additional potential

132
Y. Ou et al.
confounders, such as diet, and so we are unable to assess the effect on
associations between trace elements and CHDs.
5. Conclusions
Environmental Pb contamination remains a public health concern in
China. Blood Pb levels in our study were mostly lower than the Chinese
reference level for child health equal to 10 μg/dL (Li et al., 2015). Still,
we found strong effects for CHDs with higher maternal blood Pb at
these levels of exposure; an effect that may have been enhanced by
concurrent Cd exposure and mitigated by concurrent Se exposure. Our
study results should prove useful for developing interventions to reduce
exposures among women of childbearing age residing in Pb-polluted
and Se-deficient areas of Guangdong Province, and to help to identify
high risk pregnancies for referral to fetal echocardiography and
intensive prenatal monitoring.
Funding
This work was supported by Guangdong Science and Technology
Department, China [grant number 2014A050503048]; the National
Natural Science Foundation of China [grant number U1401255].
Conflict of interest
The authors declare that they have no competing interests.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.envint.2017.05.017.
References
Al-Sabbak, M., Sadik Ali, S., Savabi, O., Savabi, G., Dastgiri, S., Savabieasfahani, M.,
2012. Metal contamination and the epidemic of congenital birth defects in Iraqi
cities. Bull. Environ. Contam. Toxicol. 89 (5), 937–944. http://dx.doi.org/10.1007/
s00128-012-0817-2.
Aoki, Y., Brody, D.J., Flegal, K.M., Fakhouri, T.H., Axelrad, D.A., Parker, J.D., 2016.
Blood lead and other metal biomarkers as risk factors for cardiovascular disease
mortality. Medicine (Baltimore) 95 (1), e2223. http://dx.doi.org/10.1097/MD.
0000000000002223.
Arbuckle, T.E., Liang, C.L., Morisset, A.S., Fisher, M., Weiler, H., Cirtiu, C.M., et al., 2016.
Maternal and fetal exposure to cadmium, lead, manganesium and mercury: the
MIREC study. Chemosphere 163, 270–282. http://dx.doi.org/10.1016/j.
chemosphere.2016.08.023.
Barbosa Jr., Fernando, Tanus-Santos, José Eduardo, Gerlach, Raquel Fernanda, Parsons,
Patrick J., 2005. A critical review of biomarkers used for monitoring human exposure
to lead: advantages, limitations, and future needs. Environ. Health Perspect. 113
(12), 1669–1674.
Caserta, D., Graziano, A., Lo Monte, G., Bordi, G., Moscarini, M., 2013. Heavy metals and
placental fetal-maternal barrier: a mini-review on the major concerns. Eur. Rev. Med.
Pharmacol. Sci. 17, 2198–2206.
Cengiz, B., Soylemez, F., Ozturk, E., Cavdar, A.O., 2004. Serum zinc, selenium, copper,
and lead levels in women with second-trimester induced abortion resulting from
neural tube defects: a preliminary study. Biol. Trace Elem. Res. 97 (3), 225–235.
http://dx.doi.org/10.1385/BTER:97:3:225.
Correa-Villasenor, A., Ferencz, C., Loffredo, C., Magee, C., 1993. Paternal exposures and
cardiovascular malformations. The Baltimore-Washington infant study group. J.
Expo. Anal. Environ. Epidemiol. 3 (Suppl. 1), 173–185.
Davis, C.D., Uthus, E.O., Finley, J.W., 2000. Dietary selenium and arsenic affect DNA
methylation in vitro in Caco-2 cells and in vivo in rat liver and colon. J. Nutr. 130
(12), 2903–2909.
Di Renzo, G.C., Conry, J.A., Blake, J., DeFrancesco, M.S., DeNicola, N., Martin Jr., J.N.,
et al., 2015. International Federation of Gynecology and Obstetrics opinion on
reproductive health impacts of exposure to toxic environmental chemicals. Int. J.
Gynaecol. Obstet. 131 (3), 219–225. http://dx.doi.org/10.1016/j.ijgo.2015.09.002.
Donofrio, M.T., Moon-Grady, A.J., Hornberger, L.K., Copel, J.A., Sklansky, M.S.,
Abuhamad, A., et al., 2014. Diagnosis and treatment of fetal cardiac disease: a
scientific statement from the American Heart Association. Circulation 129 (21),
2183–2242. http://dx.doi.org/10.1161/01.cir.0000437597.44550.5d.
Ercal, N., Gurer-Orhan, H., Aykin-Burns, N., 2001. Toxic metals and oxidative stress part
I: mechanisms involved in metal-induced oxidative damage. Curr. Top. Med. Chem. 1
(6), 529–539.
Gil, F., Hernandez, A.F., Marquez, C., Femia, P., Olmedo, P., Lopez-Guarnido, O., et al.,
133
Environment International 106 (2017) 127–134
2011. Biomonitorization of cadmium, chromium, manganese, nickel and lead in
whole blood, urine, axillary hair and saliva in an occupationally exposed population.
Sci. Total Environ. 409 (6), 1172–1180. http://dx.doi.org/10.1016/j.scitotenv.2010.
11.033.
Goldberg, S.J., Lebowitz, M.D., Graver, E.J., Hicks, S., 1990. An association of human
congenital cardiac malformations and drinking water contaminants. J. Am. Coll.
Cardiol. 16 (1), 155–164.
Gollenberg, A.L., Hediger, M.L., Lee, P.A., Himes, J.H., Louis, G.M., 2010. Association
between lead and cadmium and reproductive hormones in peripubertal U.S. girls.
Environ. Health Perspect. 118 (12), 1782–1787. http://dx.doi.org/10.1289/ehp.
1001943.
Gulson, B.L., Mizon, K.J., Korsch, M.J., Palmer, J.M., Donnelly, J.B., 2003. Mobilization
of lead from human bone tissue during pregnancy and lactation–a summary of long-
term research. Sci. Total Environ. 303 (1–2), 79–104.
Hammouda, S.A., Abd Al-Halim, O.A., Mohamadin, A.M., 2013. Serum levels of some
micronutrients and congenital malformations: a prospective cohort study in healthy
saudi-arabian first-trimester pregnant women. Int. J. Vitam. Nutr. Res. 83 (6),
346–354. http://dx.doi.org/10.1024/0300-9831/a000176.
Hansen, J.M., 2006. Oxidative stress as a mechanism of teratogenesis. Birth Defects Res.
C. Embryo Today 78 (4), 293–307. http://dx.doi.org/10.1002/bdrc.20085.
Hinwood, A.L., Callan, A.C., Ramalingam, M., Boyce, M., Heyworth, J., McCafferty, P.,
et al., 2013. Cadmium, lead and mercury exposure in non smoking pregnant women.
Environ. Res. 126, 118–124. http://dx.doi.org/10.1016/j.envres.2013.07.005.
Hobbs, C.A., Cleves, M.A., Zhao, W., Melnyk, S., James, S.J., 2005. Congenital heart
defects and maternal biomarkers of oxidative stress. Am. J. Clin. Nutr. 82 (3),
598–604.
Hu, H., Shih, R., Rothenberg, S., Schwartz, B.S., 2007. The epidemiology of lead toxicity
in adults: measuring dose and consideration of other methodologic issues. Environ.
Health Perspect. 115 (3), 455–462.
Huang, F., Schneider, J.S., 2004. Effects of lead exposure on proliferation and
differentiation of neural stem cells derived from different regions of embryonic rat
brain. Neurotoxicology 25, 1001–1012.
Jackson, L.W., Correa-Villasenor, A., Lees, P.S., Dominici, F., Stewart, P.A., Breysse, P.N.,
et al., 2004. Parental lead exposure and total anomalous pulmonary venous return.
Birth Defects Res. A Clin. Mol. Teratol. 70 (4), 185–193. http://dx.doi.org/10.1002/
bdra.20014.
Jin, L., Zhang, L., Li, Z., Liu, J.M., Ye, R., Ren, A., 2013. Placental concentrations of
mercury, lead, cadmium, and arsenic and the risk of neural tube defects in a Chinese
population. Reprod. Toxicol. 35, 25–31. http://dx.doi.org/10.1016/j.reprotox.2012.
10.015.
Knol, M.J., VanderWeele, T.J., Groenwold, R.H., Klungel, O.H., Rovers, M.M., Grobbee,
D.E., 2011. Estimating measures of interaction on an additive scale for preventive
exposures. Eur. J. Epidemiol. 26, 433–438. http://dx.doi.org/10.1007/s10654-011-
9554-9.
Leirgul, E., Fomina, T., Brodwall, K., Greve, G., Holmstrom, H., Vollset, S.E., et al., 2014.
Birth prevalence of congenital heart defects in Norway 1994–2009 — a nationwide
study. Am. Heart J. 168 (6), 956–964. http://dx.doi.org/10.1016/j.ahj.2014.07.030.
Lemire, M., Fillion, M., Frenette, B., Mayer, A., Philibert, A., Passos, C.J., et al., 2010.
Selenium and mercury in the Brazilian Amazon: opposing influences on age-related
cataracts. Environ. Health Perspect. 118 (11), 1584–1589. http://dx.doi.org/10.
1289/ehp.0901284.
Li, M.M., Cao, J., Gao, Z.Y., Shen, X.M., Yan, C.H., 2015. The trend of lead poisoning rate
in Chinese population aged 0–18 years old: a meta-analysis. BMC Public Health 15,
756. http://dx.doi.org/10.1186/s12889-015-2103-9.
Liang, P., Wu, S.C., Zhang, J., Cao, Y., Yu, S., Wong, M.H., 2016. The effects of
mariculture on heavy metal distribution in sediments and cultured fish around the
Pearl River Delta region, south China. Chemosphere 148, 171–177. http://dx.doi.
org/10.1016/j.chemosphere.2015.10.110.
Liu, Z., Li, X., Li, N., Li, S., Deng, K., Lin, Y., et al., 2013. Association between maternal
exposure to housing renovation and offspring with congenital heart disease: a multi-
hospital case-control study. Environ. Health 12, 25. http://dx.doi.org/10.1186/1476-
069X-12-25.
Liu, Z., Yu, Y., Li, X., Wu, A., Mu, M., Li, N., et al., 2015. Maternal lead exposure and risk
of congenital heart defects occurrence in offspring. Reprod. Toxicol. 51, 1–6. http://
dx.doi.org/10.1016/j.reprotox.2014.11.002.
Martin, I., Gibert, M.J., Pintos, C., Noguera, A., Besalduch, A., Obrador, A., 2004.
Oxidative stress in mothers who have conceived fetus with neural tube defects: the
role of aminothiols and selenium. Clin. Nutr. 23 (4), 507–514. http://dx.doi.org/10.
1016/j.clnu.2003.09.010.
Mistry, H., Broughton Pipkin, F., Redman, C.W.G., Poston, L., 2012. Selenium in
reproductive health. Am. J. Obstet. Gynecol. 206, 21–30. http://dx.doi.org/10.1016/
j.ajog.2011.07.034.
Nie, L., Chu, H.T., Li, F., Cole, S.R., 2010. Relative excess risk due to interaction:
resampling-based confidence intervals. Epidemiology 21 (4), 552–556. http://dx.doi.
org/10.1097/EDE.0b013e3181e09b0b.
Pan, W., Wu, G.P., Li, Y.F., Qian, M.Y., Wang, H.S., Li, H., et al., 2001. The experience of
diagnosing congenital heart disease by fetal echocardiography on 900 fetuses. South
China J. Cardiovasc. Dis. 7 (3), 180–182 (Article in Chinese).
Pilsner, J.R., Hu, H., Ettinger, A., Sanchez, B.N., Wright, R.O., Cantonwine, D., et al.,
2009. Influence of prenatal lead exposure on genomic methylation of cord blood
DNA. Environ. Health Perspect. 117 (9), 1466–1471. http://dx.doi.org/10.1289/ehp.
0800497.
Pilsner, J.R., Hall, M.N., Liu, X., Ahsan, H., Ilievski, V., Slavkovich, V., et al., 2011.
Associations of plasma selenium with arsenic and genomic methylation of leukocyte
DNA in Bangladesh. Environ. Health Perspect. 119 (1), 113–118. http://dx.doi.org/
10.1289/ehp.1001937.
Y. Ou et al.
Porras, D., Brown, C.B., 2008. Temporal-spatial ablation of neural crest in the mouse
results in cardiovascular defects. Dev. Dyn. 237, 153–162.
Rayman, M.P., Bath, S.C., Westaway, J., Williams, P., Mao, J.Y., Vanderlelie, J.J., et al.,
2015. Selenium status in UK pregnant women and its relationship with hypertensive
conditions of pregnancy. Br. J. Nutr. 113, 249–258. http://dx.doi.org/10.1017/
S000711451400364X.
Sadler, T.W., 2004. Langman's Medical Embryology, 9th ed. Lippincott
Williams & Wilkins, Baltimore.
Steinberger, E.K., Ferencz, C., Loffredo, C.A., 2002. Infants with single ventricle: a
population-based epidemiological study. Teratology 65 (3), 106–115. http://dx.doi.
org/10.1002/tera.10017.
Thompson, J., Bannigan, J., 2008. Cadmium: toxic effects on the reproductive system and
the embryo. Reprod. Toxicol. 25 (3), 304–315. http://dx.doi.org/10.1016/j.reprotox.
2008.02.001.
Valinluck, V., Tsai, H.H., Rogstad, D.K., Burdzy, A., Bird, A., Sowers, L.C., 2004. Oxidative
damage to methyl-CpG sequences inhibits the binding of the methyl-CpG binding
domain (MBD) of methyl-CpG binding protein 2 (MeCP2). Nucleic Acids Res. 32 (14),
4100–4108. http://dx.doi.org/10.1093/nar/gkh739.
Wells, E.M., Herbstman, J.B., Lin, Y.H., Jarrett, J., Verdon, C.P., Ward, C., et al., 2016.
Cord blood methylmercury and fetal growth outcomes in Baltimore newborns:
potential confounding and effect modification by omega-3 fatty acids, selenium, and
134
Environment International 106 (2017) 127–134
sex. Environ. Health Perspect. 124 (3), 373–379. http://dx.doi.org/10.1289/ehp.
1408596.
Woodruff, T.J., Zota, A.R., Schwartz, J.M., 2011. Environmental chemicals in pregnant
women in the United States: NHANES 2003–2004. Environ. Health Perspect. 119 (6),
878–885. http://dx.doi.org/10.1289/ehp.1002727.
Yang, Q.W., Shu, W.S., Qiu, J.W., Wang, H.B., Lan, C.Y., 2004. Lead in paddy soils and
rice plants and its potential health risk around Lechang lead/zinc mine, Guangdong,
China. Environ. Int. 30 (7), 883–889. http://dx.doi.org/10.1016/j.envint.2004.02.
002.
Zeng, X., Xu, X., Zheng, X., Reponen, T., Chen, A., Huo, X., 2016. Heavy metals in PM2.5
and in blood, and children's respiratory symptoms and asthma from an e-waste
recycling area. Environ. Pollut. 210, 346–353. http://dx.doi.org/10.1016/j.envpol.
2016.01.025.
Zheng, S., Song, H., Gao, H., Liu, C., Zhang, Z., Fu, J., 2016. The antagonistic effect of
selenium on lead-induced inflammatory factors and heat shock protein mRNA level in
chicken cartilage tissue. Biol. Trace Elem. Res. 173 (1), 177–184. http://dx.doi.org/
10.1007/s12011-016-0630-6.
Zierler, S., Theodore, M., Cohen, A., Rothman, K.J., 1988. Chemical quality of maternal
drinking water and congenital heart disease. Int. J. Epidemiol. 17 (3), 589–594.
Zwolak, I., Zaporowska, H., 2012. Selenium interactions and toxicity: a review. Cell Biol.
Toxicol. 28 (1), 31–46. http://dx.doi.org/10.1007/s10565-011-9203-9.