Associations between maternal exposure to particulate matter ( ≤ 10 microns) and congenital anomalies in Liaoning Province, China: A case-control study (2010–2015)

Background: A growing number of reports suggest that maternal exposure to ambient air pollution is a potential risk factor for congenital anomalies (CAs). However, most studies have focused on specic CAs, especially congenital heart defects, while relatively few have investigated the effects of exposure to particulate matter of ≤ 10 microns (PM10) on CAs overall, but with inconsistent results. This study aims to investigate the associations between maternal exposure to PM10 and the risks of CAs in offspring in Liaoning Province. Methods: Cases of CAs (n = 31,407) and controls (n = 7,958) were selected from the Maternal and Child Health Certicate Registry of Liaoning Province from 2010 to 2015. PM10 concentrations were obtained from the Environment Protection Bureau in the same region. A multivariable logistic regression model combined with variables was used to analyze the association between maternal PM10 exposure and the risks of CAs. Results: Exposure to higher levels of PM10 signicantly increased the risks of CAs. In adjusted model I, the odds ratio (OR) of moderate and severe exposure at 3 months before pregnancy was 1.25 (95% condence interval [CI] = 1.11–1.40) and 1.40 (95% CI = 1.25–1.57), respectively. In the rst trimester, the OR of moderate and severe exposure was 1.27 (95% CI = 1.12–1.44) and 1.71 (95% CI = 1.16–1.93), respectively. Conclusions: Maternal exposure to PM10 was signicantly associated with increased risks of CAs in Liaoning Province from 2010 to 2015. Pre-pregnancy (3 months) and the rst trimester were identied as potential windows of susceptibility.


Background
There is growing evidence that air pollution has adverse effects on newborns. According to the World Health Organization, "congenital anomalies (CAs) can be de ned as structural or functional anomalies that occur during intrauterine life, and can be identi ed prenatally, at birth, or sometimes may only be detected later in infancy" [1]. The incidence of CAs is about one in 33 births worldwide, accounting for an estimated 303,000 deaths within 4 weeks after birth annually [1,2]. However, in surviving infants, CAs may be accompanied by long-term disabilities, which pose serious implications to the well-being of individuals, families, and society in general [3]. Previous studies have shown that non-genetic risk factors, including maternal characteristics [4], maternal infection [5], and maternal nutritional status [6], may be associated with increased risks of CAs.
Ambient particulate matter (PM), considered as a diverse class of air pollution, has attracted much attention from the public sector and the scienti c community. The Global Burden of Disease, Injuries, and Risk Factors Study 2017 attributed a total of 2.94 million deaths to PM pollution globally [7]. Moreover, several studies have reported associations between in utero exposure to PM with aerodynamic diameters of ≤ 10 µm (PM10) and increased risks of certain types of CAs, such as cardiovascular malformations [8,9,10], orofacial defects [11], neural tube defects [12] and abdominal wall defects [13]. However, evidence regarding the effect of environmental pollutants on CAs is limited, as various studies have failed to con rm an association or have demonstrated an inverse correlation between maternal exposure to PM10 and multiple CAs [14,15,16,17]. Since there exist inconsistencies and uncertainties regarding the impact of PM10 on the risks of CAs, further studies are required in order to arrive at rm conclusions.
Liaoning Province is an important heavy industry base in northeast China that has experienced enormous economic development over the past 10 years, as well as severe air pollution. In addition, the incidence of birth defects among newborns in Liaoning Province is about 2% [18]. Therefore, the aim of this population-based case-control study was to investigate possible relationships between the risks of CAs in offspring and prenatal exposure to PM10, and to identify the most sensitive time of exposure to PM10 before and during pregnancy.

Study design and population
Liaoning Province, located in northeast China (area, 148,600 km 2 ; population, 43.517 million), is a hub of heavy industry with ambient air pollution that is mainly caused by soil particles, biomass combustion, and tra c emissions. Maternal information and delivery outcomes of pregnant women between January 1, 2010 and December 31, 2015 were obtained from the Maternal and Child Health Certi cate Registry of Liaoning Province. A total of 39,365 neonates were selected for inclusion in the present case-control study. Demographic and obstetrical characteristics included the season of conception, sex of the infant, gestational age, birth weight, maternal age, gravidity, parity, and educational level of the mother.

De nition of outcomes (CAs)
According to the "Surveillance of Congenital Anomalies" toolkit of the Centers for Disease Control and Prevention, a CA is an anomaly of body structure or function that exists at birth and has a prenatal origin.
Major structural anomalies associated with CAs, which are important causes of death, morbidity, and disability, include ano-/microphthalmia, ano-/microtia, cleft lip/palate, congenital heart defects craniosynostosis, diaphragmatic hernia, Down syndrome, esophageal atresia, gastroschisis, hypospadias, microcephaly, neural tube defects, omphalocele formation, and reduction defects of the upper/lower limbs. The study protocol was approved by the Institutional Review Board of Liaoning Women and Children's Health Hospital and conducted in compliance with local and national regulations.

Exposure assessment
Exposure assessment data were collected from 77 air quality monitoring stations across 14 cities in Liaoning Province (Shenyang, Liaoyang, Dalian, Yingkou, Anshan, Benxi, Fushun, Dandong, Panjin, Jinzhou, Fuxin, Tieling, Huludao, and Chaoyang). The monthly average ambient PM10 concentrations were calculated from data obtained from the Environmental Protection Bureau of each city. In consideration of the metabolic process of air pollutants in the body, PM10 levels were determined from 3 months preconception throughout the rst trimester.

Descriptive statistics
The basic characteristics of the pregnant women and 39365 newborns in this study, including 31,407 (79.80%) cases of CAs, delivered from January 1, 2010 to December 31, 2015 are shown in Table 1. In the control group, most newborns were conceived in the summer and the least in the winter. The majority of newborns were delivered at or beyond gestational week 37. Most of the pregnant women in both the control and CA groups were aged 20-34 years. The highest proportion of women were in their rst pregnancy (gravidity 1) and the rst delivery (parity 0) for most in the control group. Among all newborns, there were more males than female (21,786 [53.3%] vs. 17,579 [44.7%], respectively) and most birth weights ranged from 2500 to 4000 g. In the case group, the highest educational level of most of the mothers was junior high school. The distribution of ambient PM10 concentrations (μg/m 3 ) in the case and control groups during different gestation periods is shown in Table 2. Table 3 shows the relationship between PM10 exposure and the risk of CAs at 3 months before pregnancy and during the rst trimester using the crude model and adjusted model I, respectively. The PM10 exposure level was arti cially divided into three levels (mild [reference value], moderate, and severe) in order to equalize the sample sizes. The results revealed differences between the crude model and adjusted model I (odds ratio [OR] of the crude model and adjusted model I in the rst trimester for severe exposure: 1.02 [95% con dence interval (CI) = 0.96-1.09] and 1.71 [95% CI = 1.16-1.93], respectively). In other words, the cofounders in uenced the relationship between ambient PM10 exposure and CAs through various underlying mechanisms. For adjusted model I, the ORs of moderate and severe exposure at 3 months before pregnancy were 1.25 (95% CI = 1.11-1.40) and 1.40 (95% CI = 1.25-1.57), respectively. In the rst trimester, the ORs of moderate and severe exposure were 1.27 (95% CI = 1.12-1.44) and 1.71 (95% CI = 1.16-1.93), respectively.

Discussion
In China, there exists a long-standing registry for all types of births, including fetal death and selective termination of pregnancy. The timing of exposure within the critical window of fetal development and an extended exposure window at 3 months before conception was explored in order to better identify underlying factors. In the present study, from 2006 to 2015, exposure to higher levels of PM10 at both 3 months preconception and early pregnancy signi cantly increased the risk of CAs among offspring in Liaoning Province. The study results also revealed that the most crucial time windows for susceptibility were 3 months before pregnancy and the rst trimester of gestation. In addition, this association did not appear to be substantially in uenced by a shorter time scale of 1 month.
This study provides evidence of a robust association between maternal exposure to ambient PM10 and the risks of CAs in offspring. A recent systematic review reported evidence of an association between ambient pollutants (including PM10) and CAs, although the results were limited and inconsistent [19]. Previous reviews of the effects of air pollution on birth outcomes demonstrated a clear association between PM10 exposure and risks of CAs [20,21]. Nonetheless, it is worth noting that most studies conducted over the past decade have shown an association between prenatal PM10 exposure and the risk of some speci c defects, especially cardiovascular malformations [8][9][10][11][12][13], but not CAs overall. 95% CI = 1.019-1.049) and the highest risks were observed in the second month of pregnancy (RR = 1.031, 95% CI = 1.020-1.042) [24]. More positive associations between PM10 exposure and overall CAs have been observed in Korea and Italy, but these associations were not statistically signi cant [25,26].
Moreover, studies that included the full spectrum of CAs found little evidence of such associations [15,13,17]. A possible explanation for these inconsistent ndings could be due to differences in sample sizes, air pollutant levels, exposure assignments, and confounders [22,19,17]. Notably, the cohort of the present study was a population exposed to relatively high mean PM10 levels of 90.264-93.626 µg/m 3 , while all of the above-mentioned studies, which were conducted in areas with relatively low mean PM10 levels of ≤ 40 µg/m 3 , reported irrelevant or negative associations. Given the limited number of studies, these ndings should be interpreted with caution.
A great deal of evidence suggests that oxidative stress [27], pulmonary and placental in ammation [28], blood coagulation [29], endothelial function [30], and hemodynamic responses [31] are involved in the mechanisms underlying the effects of air pollution. In addition, maternal exposure to PM10 may affect embryos and fetuses by in uencing transplacental transport of oxygen and nutrients [32]. Recent studies have suggested that prenatal PM10 exposure during the last trimester of pregnancy may cause mitochondrial dysfunction and shorten the lengths of telomeres in the cells of newborns [33,34,35].
Despite this evidence, the effects of air pollutants remain uncertain.
There were several limitations to this study that should be addressed. First, assessment of the dosage of prenatal exposure may have been inaccurate because the method used to measure PM10 concentrations may have affected the accuracy of exposure estimation and led to misclassi cation. In the present study, the monthly average of prenatal PM10 exposure concentrations was calculated from daily readings of all air monitoring stations throughout Liaoning Province. However, the municipal exposure level estimates are generally lower than those based on personal assessment of exposure [36]. Second, the time that people spend indoors was not taken into consideration, which may explain the difference between personal and municipal exposure. However, this error was assumed to be irrelevant with respect to the differentiation between cases and controls. Such misclassi cation would lead to underestimation of the effect estimates [14]. Third, although any random migration would reduce the accuracy of prenatal exposure assessment, the median distance of migration in this study tended to be short and, therefore, was not likely to result in misclassi cation [37]. Fourth, some of the potential risk factors for CAs, such as smoking, alcohol consumption, and folic acid use, were not included in multivariate analyses. Although it is unlikely that these factors were associated with exposure to PM10, they were partially accounted for by adjusting for the educational level of the mother [22]. Fifth, the study design only tested associations for a single pollutant, as the co-effects of other air pollutants, such as carbon monoxide, nitrogen dioxide, and ozone, which may be associated with CAs, were not considered in the analyses [38]. Hence, future studies should include mixtures of pollutants.
In general, the results of the present study demonstrated a strong association between maternal exposure to environmental PM10 and the risk of CAs in offspring. However, additional studies are needed to elucidate the underlying mechanisms in order to further explore the impact of maternal PM10 exposure on CAs.

Conclusions
In Liaoning Province, maternal exposure to PM10 was signi cantly associated with increased risks of CAs in the period of 2010 to 2015. These results indicate potential windows of susceptibility before and during the rst trimester.

Abbreviations
CAs congenital anomalies PM particulate matter PM10 particulate matter of ≤ 10 microns CI con dence interval OR odds ratio Declarations Ethics approval and consent to participate The study protocol was approved by the Institutional Review Board of LiaoningWomen and Children's Health Hospital and conducted in compliance with local and national regulations.

Consent for publication
Not Applicable.

Availability of data and material
The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Competing interests
The authors declare that they have no known competing nancial interests or personal relationships that could have appeared to in uence the work reported in this paper "Associations between maternal exposure to particulate matter (≤10 microns) and congenital anomalies in Liaoning Province, China: A case-control study (2010-2015)".

Funding
This study was supported by the Liaoning Providence Science and Technology Project (grant no. 2015225025 to Yan-Hong Huang) and the Shenyang Science and Technology Project (grant no. F15-139-9-09 to Yan-Hong Huang). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Authors' contributions HY conceptualized and designed the study. HS analyzed the data and wrote the paper. CY, LS and CZ were responsible for acquisition of data. LJ and JC performed the statistical measurement and analyzed the data. LL and ZC evaluated the manuscript. All authors have read and approved the manuscript.  Table 3 The associations between maternal exposure to ambient PM 10 during different study periods and CAs