Correlation between Congenital Heart Defects and Maternal Copper and Zinc Concentrations Hui Hu1,2, Zhen Liu1,3, Jun Li4, Shengli Li5, Xiaoxian Tian6, Yuan Lin7, Xinlin Chen8, Jiaxiang Yang9, Ying Deng1,3, Nana Li1,3, Yanping Wang1, Ping Yuan2, Xiaohong Li*1,3, and Jun Zhu*1
Background: The aim of this study was to investigate the correlation between maternal concentrations of copper and zinc and the risk of having an infant with a congenital heart defect (CHD). Methods: A multi-center hospital-based case-control study was conducted in China. A total of 212 cases and 212 controls were recruited from pregnant women who received prenatal examinations in four tertiary hospitals accredited to perform prenatal diagnosis in the cities of Shenzhen, Zhenzhou, Fuzhou and Wuhan between February 2010 and November 2011. Correlation between CHDs and maternal copper and zinc concentrations was estimated by a 1:1 conditional logistic regression. Also the interaction between copper and zinc was analyzed. Results: Compared with the controls, mothers with hair copper concentrations of 17.77 lg/g or more were more likely to have a child with a CHD than those with a lower concentration. The adjusted odds ratio was 5.70 (95% confidence interval, 2.58–12.61) for CHDs and 6.32 (95% confidence interval, 2.11–18.92) for conotruncal defects. Zinc concentrations were not significantly different in the case and control groups. The results suggest that
1 National Center for Birth Defect Monitoring, West China Second University Hospital, Sichuan University, Chengdu, Sichuan, China 2 West China School of Public Health, Sichuan University, Chengdu, Sichuan, China 3 Laboratory of Molecular Epidemiology for Birth Defects, West China Second University Hospital, Sichuan University, Chengdu, Sichuan, China 4 Department of Ultrasound, Xijing Hospital, Fourth Military Medical University, Xi’an, Shanxi, China 5 Department of Ultrasound, Shenzhen Maternity and Child Healthcare Hospital, Shenzhen, Guangdong, China 6 Department of Ultrasound, Maternal and Child Healthcare Hospital of Guangxi Zhuang Autonomous Region, Nanning, Guangxi, China 7 Department of Obstetrics & Gynecology, Fujian Provincial Maternal and Child Healthcare Hospital, Fuzhou, Fujian, China 8 Department of Ultrasound, Hubei Provincial Maternal and Child Healthcare Hospital, Wuhan, Hubei, China 9 Department of Ultrasound, Sichuan Provincial Maternal and Child Healthcare Hospital, Chengdu, Sichuan, China
The study was supported by the National Basic Research Program of China (Research Grant 2010CB529502) and the National “Twelfth Five-Year” Plan for Science & Technology Support (2014BAI06B01) as well as by the Natural Science Foundation (No. 81273086), and the Program for Changjiang Scholars and Innovative Research Team in University (ID: IRT0935). Drs. Hui Hu, Zhen Liu, and Jun Li are joint first authors. *Correspondence to: Jun Zhu and Xiaohong Li, National Center for Birth Defect Monitoring, West China Second University Hospital, Sichuan University, Sec. 3, No. 17, South RenMin Road, Chengdu, Sichuan, China. Code: 610041. Email: [email protected]
(Jun Zhu), [email protected]
(Xiaohong Li), These authors contributed equally to this study. Published online 0 Month 2012 in Wiley Online Library (wileyonlinelibrary.com). Doi: 10.1002/bdra.23284
C 2014 Wiley Periodicals, Inc. V
mothers whose zinc content was 104.60 lg/g or less did not have a significantly higher risk of having a child with a CHD. No interaction between maternal copper and zinc concentrations was observed in the multiplicative or additive model. Conclusion: Women with excessive copper concentrations have a significantly increased risk of having offspring with a CHD. A low maternal zinc status might have a correlation with CHDs, and an interaction between copper and zinc might exists, but an epidemiological study with a larger sample size is needed to confirm this finding. Birth Defects Research (Part A) 00:000–000, 2014. C 2014 Wiley Periodicals, Inc. V
Key words: Pregnancy; Congenital Heart Defect; zinc and copper concentrations; hair analysis
Introduction Congenital heart defects (CHDs) are among the most prevalent abnormalities presented at birth and remaining the main cause of mortality resulting from birth defects (Rosamond et al., 2007). The average total prevalence of Chinese CHDs was 40.95 per 10,000 births in 2011 (P.R. China, 2012). The risk factors of CHDs are various and include maternal diabetes, obesity and deficiency in nutrients such as zinc (Duffy et al., 2004; Morgan et al., 2008; Mills et al., 2010). Animal studies have demonstrated that suboptimal maternal micronutrient status can lead to a significantly elevated risk of adverse pregnancy outcomes, including heart defects (Keen et al., 2003). Copper ions serve as an important catalytic cofactor in the redox chemistry of proteins that carry out fundamental biological functions, such as: Cu/Zn SOD (Cu/Zn superoxide dismutase), cytochrome c oxidase, and ceruloplasmin (Shim and Harris, 2003). Copper deficiency induces heart teratogenesis in rats (Beckers-Trapp et al., 2006), and copper excess leads to an increase in the mortality of the embryos and larvae of Pagrus major (Cao et al., 2010). Zinc is an essential trace element for humans and is involved in the synthesis of many proteins, prokaryotic and eukaryotic nucleic acids and lipids (Falchuk and Montorzi, 2001). Dietary zinc deficiency during gestation has been recognized to induce teratogenic
effects, leading to heart malformations or orofacial clefts in humans and animals (Lopez et al., 2008; Hozyasz et al., 2009). In addition, an antagonistic interaction between copper and zinc has been demonstrated in animal experiments, and to affect human health (Sundaresan et al., 1996). Zinc-induced copper deficiency seems to be the cause of neutropenia and anaemia (Kroft et al., 2005). So far, only a small number of epidemiological studies have been conducted to examine the correlation between trace elements and CHDs. Their conclusions cannot be extrapolated because there are numerous factors affecting CHDs and reference concentrations of trace elements are lacking. Some studies have used blood as a biomarker of copper and zinc, but the results may only reflect people’s short-term exposure. However, hair concentrations have been confirmed to be suitable to reflect the chronic exposure of copper and zinc in the body (Klevay et al., 2004), especially in pregnant women during the first trimester, which may represent the exposure of the fetus. Therefore, hair copper and zinc concentrations may be the optimal biomarker. The method used to detect the hair concentrations of copper and zinc, namely ICP-MS, is presently considered to be state of the art in metal analysis because of its high sensitivity and its ability to detect multiple metals simultaneously. In 2009, we conducted a project designed to be a hospital-based case-control study based on the National Birth Defects Surveillance System in China. It aimed to explore the effects of some environmental and genetic risk factors on the occurrence of CHDs, including geneenvironment interactions. A questionnaire-based interview and biological samples were collected to obtain information on the subjects. This article aims to demonstrate that maternal copper deficiency or excess during pregnancy raises the risk of having a child with a CHD. Similarly maternal zinc deficiency raises the risk of having a child with a CHD. Finally, preliminarily data were obtained on the interactive effect of copper and zinc on heart development.
Materials and Methods
HEART DEFECTS AND MATERNAL COPPER AND ZINC LEVELS
cians, obstetricians and pathologists from each hospital; The diagnosis established on stillbirths or terminated pregnancies relied on autopsy reports. In total, 1032 subjects were recruited for the study (more details about the recruitment procedure have been discussed in our previous publication [Li et al., 2013]). Exclusion criteria are as follows: the pregnant woman’s gestational age was smaller than fourteen weeks; the child had a CHD associated with a chromosome disorder or a single-gene disorder; the diagnosis was unclear; or it was a multiple birth. To comply with the purpose of this article, several additional exclusion criteria were established a priori for the 424 subjects: the hair of the pregnant woman was color-treated or over-processed; the pregnant woman had a mental disorder; the CHD was associated with extracardiac malformations. The matching factor was the gestational age. When a fetus was recruited as a case, a fetus whose gestational age was no more than 2 weeks different from that of the case was recruited as a control. After matching and exclusion, we came up with 212 cases and 212 controls. CLASSIFICATION OF CHDS
Each child recruited in the study had at least one type of cardiac defect and an absence of any non-CHD defect. In line with both anatomic defects and clinical classification, the cases of CHDs were divided into six major categories: (1) Septal defect, including atrial septal defects (ASDs) and ventricular septal defect (VSDs), atrioventricular septal defects (AVSDs), etc.; (2) Conotruncal defect, including transposition of the great arteries, tetralogy of Fallot, double-outlet right ventricle, etc; (Pope et al., 2007). (3) Left ventricular outflow tract obstruction, including aortic valve stenosis, hypoplastic left heart syndrome, etc.; (4) Right ventricular outflow tract obstruction, including pulmonary valve stenosis, pulmonary atresia, tricuspid atresia, and Ebstein anomaly; (5) Anomalous pulmonary venous return, including total and partial anomalous pulmonary venous return; (6) Other heart defects (Malik et al., 2008).
TRACE ELEMENT ANALYSIS IN HAIR
In this study, subjects were recruited from pregnant women who received prenatal examination in four tertiary hospitals accredited to perform prenatal diagnosis located in Shenzhen, Zhenzhou, Fuzhou and Wuhan between February 2010 and November 2011. Pregnant women whose fetuses had been diagnosed with a CHD by echocardiography were recruited in the case group, while those whose fetuses had not been diagnosed with any abnormality (from the same hospital) were recruited in the control group once a case was ascertained. Case fetuses were live births, stillbirths or terminated pregnancies, and control fetuses were live births. The diagnosis of each live birth child was confirmed within 3 months after birth by ultrasonographers, paediatri-
We decided to use hair concentration as a biomarker because it is replicable, easy to get and reflects a longterm exposure period. Hair strands that were approximately 3 to 5 cm long, weighting approximately 1 g were cut at the scalp from the occipital area of pregnant woman by trained specimen collectors using stainless steel scissors. Each 100 mg of hair sample was microwave-digested (CEM) at 180 C following the manufacturer’s procedure, and 5 ml of HNO3 was added to the microwave-digested tubes. Then, the tubes were heated to near-dryness on a heating plate and subsequently diluted with 2% HNO3 to a volume of 2 ml. The diluted samples were stored at 4 C for analysis.
BIRTH DEFECTS RESEARCH (PART A) 00:00–00 (2014)
ICP-MS was performed using an Agilent 7500cx ICP/ MS system (Agilent Technologies, Wilmington, DE) equipped with a G3160B I-AS integrated autosampler. Sample introduction was performed with a micromist nebulizer combined with a Scott-type double pass spray chamber (Agilent Technologies, Wilmington, DE). The internal standard was a multi-element standard solution (SPEX CertiPrep, Metuchen, NJ). QUESTIONNAIRE
Using a detailed questionnaire, face-to-face interviews with the subjects were performed to collect information. The questionnaires obtained information on each participant and her husband regarding their demographic characteristics, pregnancy history, lifestyle, medication use, family history, folic acid supplementation, etc. Every individual who was willing to participate in the study signed an informed consent form. This study was approved by the Medical Ethics Committee of Sichuan University (No.2010004). STATISTICAL ANALYSIS
The distributions of copper and zinc concentrations are presented as medians (interquartile range, IQR) because of their non-normal distributions. Differences between the case and control groups in maternal age, copper and zinc concentrations, and other maternal characteristics were tested using the paired Student’s t test, the Wilcoxon signed-ranks test or the v2 test according to the distribution observed. The concentrations of copper and zinc and confounding variables were used in a 1:1 conditional logistic regression. In view of the large variations in the reference values for trace elements among the different ethnic groups and residences of subjects, the different maternal age groups and the unknown normal range of zinc and copper concentrations in maternal hair, we used the 5th to 95th percentile of concentrations in the control group’s hair tissue (n 5 212) as reference ranges in our study. The other two portions represented lower or higher concentration, respectively. Crude odds ratio (OR), adjusted OR, and 95% confidence interval (CI) were calculated to estimate the risk of having a child with a CHD. Confounding variables were the mother’s age (brought into the logistic regression in the form of a continuous variable), maternal residence (urban, suburban or rural), taking folic acid (yes or no), and whether the mother had previous pregnancies (yes or no). The interaction between copper and zinc concentrations was analyzed in a multiplicative model by logistic regression and in additive model by a four-by-two table and an Excel spreadsheet (Zou, 2008). To test the hypothesis of an additive interaction, subjects’ copper and zinc concentrations were divided into two binary variables: unexposed (2) for women whose copper or zinc concen-
tration was inside the reference range, exposed (1) for women whose copper or zinc concentration was outside the reference range (Qiu et al., 2008). We defined OR00 as a reference category when case and control mothers were unexposed to either element. OR11 was calculated when mothers’ copper and zinc concentrations were both exposed, OR10 and OR01 were calculated when one of the two elements was exposed. The additive interaction had three measures: relative excess risk due to interaction (RERI), attributable proportion due to interaction (AP) and the synergy index (S). RERI was the value of the interaction of the two factors. The bigger the absolute value of RERI was, the stronger the interaction would be. The following formula was used to calculate RERI: RERI5OR11 2ðOR10 1OR01 21Þ
If there’s no additive interaction between the two factors, the confidence intervals of RERI and AP include 0, and that of S includes 1. A conditional logistic regression model was used to calculate the covariance matrix for the regression coefficients for each of the three exposure categories, then the covariance matrix was used as input in an Excel spreadsheet to carry out ORs, RERI, AP, S and their 95% confidence intervals (Andersson et al., 2005). Data were statistically analyzed using SPSS v.20.0 (SPSS Inc., Chicago: IL) and Microsoft Excel 2010 (Microsoft Corporation, Redmond, WA). The significance level was 0.05.
Results From February 2010 through October 2011, we got 1032 subjects in this study. Of these participants, 27 case and 14 control women’s gestational age were smaller than fourteen weeks, 152 case children were diagnosed with a chromosome disorder, a single-gene disorder or an extracardiac malformations, 31 case children had an unclear diagnosis, 50 controls were not able to participate in the follow-up, 12 case children were issued from multiple births. In addition, 59 case and 34 control mothers’ hair were color-treated or over-processed. 6 case and 4 control mothers were suffering from gestational diabetes mellitus. These participants with their matched cases or controls were excluded from our analysis. Ultimately, 212 case and 212 control fetuses remained in the study. The most common heart defects in our study were conotruncal defects (89, 41.98%); the second-most common were septal defects (53, 25.00%), followed by right ventricular outflow tract obstructions (50, 23.58%), left ventricular outflow tract obstructions (49, 23.11%), other heart defects (19, 8.96%), and anomalous pulmonary venous return (6, 2.83%). The characteristics of the subjects are shown in Table 1. The cases and controls were significantly different with respect to
HEART DEFECTS AND MATERNAL COPPER AND ZINC LEVELS
TABLE 1. Characteristics of Case and Control Groups
Cases (n5212) n(%)
Controls (n5212) n(%)