Association between the Concentrations of Metallic Elements in Maternal Blood during Pregnancy and Prevalence of Abdominal Congenital Malformations: The Japan Environment and Children’s Study

Abdominal congenital malformations are responsible for early mortality, inadequate nutrient intake, and infant biological dysfunction. Exposure to metallic elements in utero is reported to be toxic and negatively impacts ontogeny. However, no prior study has sufficiently evaluated the effects of exposure to metallic elements in utero on abdominal congenital malformations. The aim of the present study was to evaluate associations between metallic elements detected in maternal blood during pregnancy and congenital abdominal malformations. Data from participants in the Japan Environment and Children’s Study was used in the present study, and contained information on singleton and live birth infants without congenital abnormalities (control: n = 89,134) and abdominal malformations (case: n = 139). Heavy metals such as mercury (Hg), lead (Pb), cadmium (Cd), and trace elements of manganese (Mn) and selenium (Se) were detected in maternal serum samples during mid- and late-gestation. Infant congenital abnormalities were identified from delivery records at birth or one month after birth by medical doctors. In a multivariate analysis adjusted to account for potential confounders, quartiles of heavy metals and trace elements present in maternal blood were not statistically correlated to the prevalence of abdominal congenital malformations at birth. This study is the first to reveal the absence of significant associations between exposure levels to maternal heavy metals and trace elements in utero and the prevalence of abdominal congenital malformations in a large cohort of the Japanese population. Further studies are necessary to investigate the impact of exposure to heavy metals and trace elements via maternal blood in offspring after birth.


Introduction
Congenital anomalies are responsible for 11% of neonatal mortality worldwide [1]. In Japan, congenital anomalies, including congenital malformations, morphological anoma-

Selected Participants in Present Study
The flowchart in Figure 1 illustrates when the various types of data were collected from participants. Concentrations of ME in maternal blood from 96,696 participants in their mid-and late-trimesters were analyzed. Of these, 94,661 sets of participant data contained the following: self-administered questionnaires, birth records without missing data on maternal age, maternal smoking and drinking habits, paternal smoking habits, birth year of infant, and sex of infant. We excluded infants with chromosomal abnormalities, congenital malformations without abdominal anomalies, miscarriages, and those that were stillbirths or from multiple births, but included singleton and live birth infants without congenital abnormalities (control: n = 89,134) and abdominal malformations (n = 139).
were stillbirths or from multiple births, but included singleton and live birth infants without congenital abnormalities (control: n = 89,134) and abdominal malformations (n = 139).

Statistical Analysis
The differences in the concentrations of ME detected in maternal blood, parental characteristics, and infant characteristics between abdominal malformed infants and nonabdominal malformed infants were analyzed using Pearson's chi-square tests, Student's t-tests, Spearman's rank correlation test, Mann-Whitney U-test, and Kruskal-Wallis test (Tables 1-4).

Statistical Analysis
The differences in the concentrations of ME detected in maternal blood, parental characteristics, and infant characteristics between abdominal malformed infants and nonabdominal malformed infants were analyzed using Pearson's chi-square tests, Student's t-tests, Spearman's rank correlation test, Mann-Whitney U-test, and Kruskal-Wallis test (Tables 1-4).  ICD-10 code: Transcriptions of both the ICD-10 code and congenital anomaly names in medical records were collected during the data collection phase.  The participants were divided into quartiles according to distributed concentrations of each of the five ME. In a multivariable logistic regression analysis, with abdominal malformation as the dependent variable, the odds ratio (OR) and 95% confidence interval (CI) of abdominal malformation for the second, third, and fourth quartiles of the concentrations of ME present in maternal blood, compared with the lowest quartile as reference, were calculated. To evaluate the dose-response relationship, a p-trend was calculated for the quartile groups of categorized concentrations of ME in maternal blood, as an ordinal variable ( Figure 2, Tables 5-7). Outcomes of all abdominal malformations and each of the seven types of anomalies, including omphalocele, gastroschisis, esophageal atresia with or without fistula, duodenal atresia, intestinal atresia, anorectal atresia, and diaphragmic hernia were used in a multivariable logistic regression analysis, adjusted for potential confounders. These confounders comprised factors from the present study or those previously reported to be significantly associated with maternal ME and abdominal congenital malformations, such as maternal age at delivery (categorical) and maternal smoking habit (never, quit, smoking during pregnancy), drinking habit (never, quit, drinking during pregnancy), paternal smoking habit (never, quit, smoking during pregnancy), birth year of infants, and sex of infants.
Statistical significance was set at p < 0.05. Statistical analyses were performed using SPSS software for Windows (version 21.0J; IBM, Armonk, NY, USA).

Parental Characteristics
For 89,273 participants in the present study, the median concentrations of maternal ME, including Pb, Cd, Hg, Se, and Mn, were 5.84 ng/g, 0.661 ng/g, 3.64 ng/g, 168.0 ng/g, and 15.3 ng/g, respectively. There were no significant differences in the concentrations of maternal ME between infants with and without malformation (Table 1). Abdominal congenital malformations were observed in 139 infants, included in the current study by transcription of medical records at birth or one month after birth ( Table 2). Parental and infant characteristics are shown in Tables 3 and 4, respectively. The prevalence of abdominal congenital malformations significantly differed according to maternal smoking habit, type of delivery, sex of infants, gestational age, and birth weight, but not maternal age, drinking habit, educational duration, and household income (Tables 3 and 4). Concentrations of maternal ME were significantly different between maternal, paternal, and infant characteristics.
Logistic regression analysis revealed no statistically significant (p < 0.05) associations between maternal ME and any abdominal congenital malformations ( Figure 2 and Table 5). In relation to specific abdominal congenital malformations, compared with the lowest quartile (p-trend = 0.089), the third quartile of Hg, adjusted for ORs of diaphragmatic hernia, was significantly increased (OR (95% CI): 3.22 (1.05, 9.90)) ( Tables 6 and 7). A significant inverse trend was observed for diaphragmatic hernia throughout the quartiles of Mn (p-trend = 0.010). The highest (OR (95% CI): 0.15 (0.03, 0.68)) of Mn, adjusted for ORs of diaphragmatic hernia, were significantly decreased, compared with the lowest quartiles (Tables 6 and 7). A significant inverse trend was observed for omphalocele throughout the quartiles of Pb (p-trend = 0.033). The second (OR (95% CI): 9.96 (1.27, 77.88)) and third quartiles (OR (95% CI): 10.06 (1.28, 78.77)) of Se, adjusted for ORs of omphalocele, were significantly increased, compared with the lowest quartiles (Tables 6 and 7). We found no significant associations between maternal ME and specific abdominal congenital malformations, including gastroschisis, esophageal atresia with or without fistula, duodenal atresia/stenosis, intestinal atresia/stenosis, and anorectal atresia/stenosis (Tables 6 and 7). scription of medical records at birth or one month after birth ( Table 2). Parental and infant characteristics are shown in Tables 3 and 4, respectively. The prevalence of abdominal congenital malformations significantly differed according to maternal smoking habit, type of delivery, sex of infants, gestational age, and birth weight, but not maternal age, drinking habit, educational duration, and household income (Tables 3 and 4). Concentrations of maternal ME were significantly different between maternal, paternal, and infant characteristics.

Discussion
We found no statistically significant associations between maternal ME and any abdominal congenital malformations. However, significant positive or inverse associations were observed among maternal Pb, Mn, Hg, and Se and specific abdominal congenital malformations, including diaphragmatic hernia and omphalocele, with a wide range in 95% CI, due to the small sample size. The results indicated that presence of maternal ME was not a definite risk factor for congenital malformation in the offspring. This is the first study to show that exposure to ME via maternal blood is unlikely to cause fetal abdominal congenital malformations.
Direct comparison of results between previous studies and the present study is difficult, because there are no previous studies assessing the effects of ME in maternal blood. In North Carolina, a human study was conducted to evaluate the association between metal concentrations in private well water and birth defect prevalence, including heat defects, cleft palate/lip, hypospadias, pyloric stenosis, and gastroschisis. Individual exposure was designated as the average metal concentrations in the census tract encompassing the geocoded maternal residence. No association was observed between abdominal congenital malformation and Mn and Cd, excluding the inverse association between Cd and prevalence of pyloric stenosis (RR: 0.4 95% CI: 0.3-0.7) [15]. In another study, based on the classification of exposure to emissions from municipal solid waste incinerators, data from exposed (n = 194) and unexposed (n = 2678) settlement populations was used to evaluate the relative risks of congenital malformations. The rate of congenital anomalies, including diaphragmatic hernia and gastroschisis, was not significantly higher in exposed than unexposed communities [16]. In a recent study, metal concentrations in maternal blood from 140 pregnant women living in industrial areas were measured; however, Cd was detected in urine samples from only two pregnant women who birthed infants without congenital anomalies (detection rate: 1.4%, median: 5.3 µg/L), which means that the effect on congenital anomalies could not be effectively statistically evaluated [17]. While the three afore-mentioned studies described inadequate and weak effects of ME, our present study provides more tangible results contending that maternal exposure to ME via maternal blood (at levels detected in this study) are unlikely to cause fetal abdominal congenital malformation.
Adjusted ORs for infant diaphragmatic hernia and omphalocele increased the upper Hg and Se. The potential biological mechanisms are that Hg can disturb cell differentiation and growth, and Se can affect fetal growth, because the normal range of Se is very narrow. Moreover, previously reported exposure concentrations of Hg among JECS participants estimated that almost 30% of the participants had blood Hg levels within the range where detrimental health effects cannot be excluded with adequate certainty [14]. On the other hand, we cannot exclude the possibility that this result was due to chance, because there is no evidence of the toxicity of Hg and Se on the specific biological mechanisms of diaphragmatic hernia and omphalocele, gastroschisis, esophageal atresia with or without fistula, duodenal atresia/stenosis, intestinal atresia/stenosis, or anorectal atresia/stenosis. Conversely, the ORs of infantile omphalocele and diaphragmatic hernia decreased in the upper Pb and Mn. A previous JECS study reported a positive correlation between maternal Pb and abstinence from alcohol during pregnancy. Another study among Japanese women reported that the major contribution rates to the total estimated uptake of Mn were Japanese tea (33.9%), rice (23.7%), and vegetables (13.6%) [18]. Therefore, our study suggests that Pb and Mn concentrations in maternal blood were at levels that did not negatively impact fetal development, and that high-quality nutritional intake may safeguard against anomalous fetal development.
In a previous JECS study on congenital anomalies, which defined an index scaling of occurrence per 10,000 pregnancies, prevalence for diaphragmatic hernia was 1.9, omphalocele was 2.6, gastroschisis was 1.5, esophageal atresia with or without fistula was 1.5, and duodenal atresia/stenosis was 1.1, intestinal atresia/stenosis was 0.8, and anorectal atresia/stenosis was 2.9 [13]. Applying the same index scaling to the present study targeting infants who were singleton and live births, the prevalence of abdominal congenital malformations was lower in all participants [13]. In the present study, malformed abdominal infants had lower gestational ages and birth weights than non-malformed infants (Table 4), which is consistent with reports from previous human studies [19]. Congenital anomalies begin to develop in the early onset of fetal development and involve risk factors common to structural abnormalities and fetal growth retardation, such as prenatal complications, chromosomal anomalies, placental insufficiency, and polyhydramnios [19]. Maternal smoking is a known risk factor for birth defects [20]. However, in the present study, mothers with malformed abdominal infants had a lower frequency of smoking during pregnancy and a higher frequency of quitting smoking before pregnancy than those without abdominal malformed infants. The results suggest that maternal smoking during pregnancy cannot be responsible for abdominal malformed infants, and mothers may behave in a protective manner according to maternal awareness of self-risk of birth defects such as older childbearing age. Previously, associations between maternal alcohol consumption and malformed abdominal infants were weak [21]. In the present study, maternal alcohol consumption during pregnancy was not associated with malformed abdominal infants, a finding consistent with previous studies that reported that maternal alcohol consumption during pregnancy has a weak effect on the risk of cryptorchidism [22] and congenital heart defects [23].

Strengths and Limitations
The strengths of this study are that basic data on early gestations in participants have been prospectively collected using a national cohort study, which had a large sample size and included a sizable geographic proportion of Japan. ME in maternal blood were detected at a comprehensive detection rate. Cases of abdominal malformation were identified from hospital records by the obstetrician, which minimalized miscalculation bias. Our study did have some limitations. Firstly, we have no data on the history of abdominal congenital malformation in participating parents. Unobserved potential confounders may also be associated with maternal ME and abdominal malformations. Further studies are needed to evaluate the effect of exposure to maternal ME in utero on offspring health outcomes, including growth and development.

Conclusions
To our knowledge, this study is the first to reveal no significant associations between exposure levels in utero of maternal ME such as Hg, Pb, Cd, Mn, and Se, and the prevalence of abdominal congenital malformations in a large cohort of the Japanese population. However, there is a possibility that maternal ME may indeed have adverse effects on fetal development and subsequent health in children. Further studies with follow-up of children are necessary to investigate the effects of maternally present heavy metals on offspring after birth.