A pilot study of several environmental endocrine disrupting chemicals in children with autism spectrum disorder in south China

Autism spectrum disorders (ASD) is a group of heterogeneous neurodevelopmental disorders. Evidence has implied that environmental pollutants are important factors related to ASD. In this study, several environmental endocrine-disrupting chemicals, including parabens, benzophenone-type ultraviolet filters, hydroxyl polycyclic aromatic hydrocarbons, triclosan and tetrabromobisphenol A were analyzed in blood plasma in ASD children (n = 34) and the control children (n = 28). The results showed that parabens were the most concentrated chemicals (2.18 ng/mL, median value), followed by hydroxyl polycyclic aromatic hydrocarbons (0.73 ng/mL), benzophenone-type ultraviolet filters (0.14 ng/mL), triclosan (0.13 ng/mL) and tetrabromobisphenol A (0.03 ng/mL). ASD children accumulated significantly lower 2-hydroxy-4-methoxybenzophenone, 2,4-dihydroxybenzophenone, 4-hydroxybenzophenone and triclosan but higher 2-hydroxyphenanthrene and tetrabromobisphenol A than the control children (0.02/0.09 ng/mL of 2-hydroxy-4-methoxybenzophenone, p < 0.05; 0.04/0.07 ng/mL of 2,4-dihydroxybenzophenone, p < 0.05; 0.03/0.04 ng/mL of 4-hydroxybenzophenone, p < 0.05; 0.13/1.22 ng/mL of triclosan, p < 0.01; 0.03 ng/mL/not detected of 2-hydroxyphenanthrene, p < 0.05; 0.03/0.004 ng/mL of tetrabromobisphenol A, p < 0.05). Gender differences in certain environmental endocrine-disrupting chemicals were evident, and the differences were more inclined toward boys. Positive associations between 2-hydroxy-4-methoxybenzophenone and triclosan, and tetrabromobisphenol A and 2-hydroxyphenanthrene were found in ASD boys. Binary logistic regression analysis showed that the adjusted odds ratio value of 2-hydroxyphenanthrene in ASD boys was 11.0 (1.45–84.0, p < 0.05). This is the first pilot study on multiple environmental endocrine-disrupting chemicals in children with ASD in China.


Introduction
Autism spectrum disorders (ASD) are a group of heterogeneous neurodevelopmental disorders characterized by impairment in communications and social interactions accompanied by repetitive and restrictive behaviors (APA, 2013;Morales-Suárez-Varela et al., 2017). Symptoms of ASD usually occur in the early developmental period (Morales-Suárez-Varela et al., 2017). Infants and children are highly susceptible to ASD because they are more sensitive to neurological disorders than adults (Liew et al., 2015). The prevalence of ASD has increased dramatically in the past few decades (Baxter et al., 2015;Wing & Potter, 2010). According to the Centers for Disease Control and Prevention of the United States, the average incidence of ASD in North America, Europe and Asia has reached 1-2% (Japan, Hong Kong, Taiwan, and South Korea) by 2016. Since the early 1980s, the prevalence of ASD in children in China has increased significantly (Kim et al., 2011;Lai et al., 2014;Tammimies et al., 2015). Between 2000 and2004, the proportion of children diagnosed with ASD was estimated to be 0.85%, but by 2011 it had reached 1.64%, doubling in a decade (Sun et al., 2013).
Environmental endocrine-disrupting chemicals (EDCs) are a class of chemicals that are highly suspected to be involved in autistic behaviors in children (Cock et al., 2012;Moosa et al., 2017;Salvador et al., 2018). They may alter the endogenous axis, interfere with steroid-dependent neurodevelopment, and thus alert the risk of ASD (Barkoski et al., 2019;Braun, 2012;Landrigan et al., 2012;Tran & Miyake, 2017;Ye et al., 2017). A Korean study found that mothers of autistic children were exposed to higher levels of PBDEs, PCBs, BPA and dioxins than mothers of normally developing children (Sun & Allison, 2010). Gestational and prenatal exposure to OCPs, dicofol, endosulfan and organophosphate insecticide chlorpyrifos may also increase the risk of autistic behaviors in offspring (Miodovnik et al., 2011;Roberts et al., 2007). Perfluorooctanoic acid (PFOA), PCB-178, PBDE-28 and PBDE-85 were reported to be associated with autistic behaviors in 4-to 5-year-old children . Several cohort studies and case-control studies also found elevated serum BPA levels in children with ASD Kardas et al., 2016;Miodovnik et al., 2011;Stein et al., 2015;Testa et al., 2012). The impact of EDCs on the pathogenesis of ASD remains unclear. However, studies have shown that exposure to certain EDCs has a strong impact on reciprocal social, repetitive, and stereotypic behaviors in children, and to varying degrees revealed a link between EDCs and ASD. In this study, we determined five types of environmental EDCs, including parabens, benzophenonetype ultraviolet (BP-type UV) filters, hydroxyl polycyclic aromatic hydrocarbons (OH-PAHs), triclosan (TCS), and tetrabromobisphenol A (TBBPA) in plasma in ASD and non-ASD children in South China. Although the target compounds analyzed have different chemical properties and activities in humans, they have a certain influence on biological neurodevelopmental and neurobehavioral changes or promotion of the sequential absorption of some non-absorbed toxic hydrophilic substances (Zeliger, 2013). Parabens, BPtype UV filters and TCS are widely exposed EDCs in the Chinese population. A large number of studies have documented the occurrence of these EDCs in Chinese children, and children are primarily exposed to parabens, BP-type UV filters and TCS via ingestion with foodstuffs and pharmaceuticals and dermal absorption with personal care products (Chisvert et al., 2012;El Hussein et al., 2007;Kunisue et al., 2010;Liao et al., 2013;Lu et al., 2017;Wang et al., 2012;Witorsch, 2014). Polycyclic aromatic hydrocarbon (PAHs) and TBBPA are wellrecognized neurotoxic endocrine disruptors. Sheng et al. reported that benzo(a)pyrene (B(a)P) exposure could directly decrease the activity of gene promoters, down-regulate the levels of RNA and protein, and eventually lead to dysregulation of their expression patterns (Sheng et al., 2010). TBBPA was reported to be associated with hyperactivity and cognitive function in some neuropsychiatric disorders (Zieminska et al., 2016). The purposes of this study were to figure out the pollution characteristics of these EDCs in children with ASD and explore the potential relationship between these chemicals and ASD.

Study Participants
In 2016, a total of 34 children with ASD (22 boys and 12 girls) were randomly recruited at Shenzhen Children's Hospital, South China. All the children in the case group were newly diagnosed with ASD and had no medical treatment until blood collection. For comparison, 28 healthy children (21 boys, 7 girls) from a public kindergarten in Shenzhen were recruited as control subjects. Blood samples were collected into vacuum glass tubes to prevent contamination. The collected blood from all participants was processed by centrifugation at 3000× g (g is gravitational acceleration) for 4 min, and plasma was stored at −40 °C in pre-cleaned containers until analysis. At the time of sample collecting, each participant was subjected to a questionnaire survey. The questionnaire included detailed information on demography, general lifestyle, and parents' working situation (summarized in Table 1). This study was examined and approved by the ethics committee of the First Affiliated Hospital of Jinan University, China.
Enzymatic deconjugation and a liquid-liquid extraction method were applied for the extraction procedure. Briefly, after thawing at room temperature, 0.3 mL of plasma was transferred into a 12 mL glass tube and spiked with 20 ng of the standard internal mixture. The plasma was equilibrated for 2 h. Subsequently, 0.3 mL of 1.0 M ammonium acetate buffer containing 300 units of β-glucuronidase from Helix pomatia and 0.3 mL of water were added. The sample was placed in incubation for 12 h at 37 °C. After deconjugation, 3.0 mL of MTBE was added, and ultrasonic extraction was applied for 20 min. The extraction was then shaken vigorously for 40 min and centrifuged for 10 min. The organic fraction was collected into another glass tube. The extraction procedure was repeated two times. Organic fractions were combined and added 1.0 mL of water for removing watersoluble substances. The mixed solution was centrifuged at 4000× g for 10 min after 30 s vortex. The organic layer was collected and evaporated to near-dryness under a gentle stream of high-purity N 2 . Finally, 0.3 mL of methanol was added.
All target compounds were quantified by highperformance liquid chromatography equipped with a tandem mass spectrometry system (HPLC-MS-MS; Shimadzu LC-30A LC system plus an AB Sciex 5500 triple quadrupole MS). The instrument analysis methods of all types of target compounds are improved by referring to previous studies (Gao et al., 2015;Ma et al., 2013;Xue et al., 2015). Detailed information about the parameters involved in instrumental analysis is shown in the Supplementary Material. All processes of sample preparation and instrument analysis were completed at Guangdong Key Laboratory of Environmental Pollution and Health, Jinan University.
Quality Assurance/Quality Control (QA/QC) A series of necessary quality control measurements were conducted in the process of sample collection, storage, preparation and analysis to eliminate possible contaminants. First, blood samples were collected using pre-checked glass vacuum collective tubes, and plasma samples were stored in pre-checked sterile tubes at −40 °C before further analysis. None target EDCs were detected in both glass or sterile tubes after liquidliquid extraction with MTBE. Second, a special control experiment was performed to make sure that no target EDCs migrated from tubes during storage. In general, most EDCs in human fluids such as plasma and serum exist in both free and conjugated forms. However, studies demonstrated that some EDCs (e.g., parabens, phenols and TCS) metabolize rapidly in humans and exist mostly as a form of conjugated (Ye et al., 2006;Ye et al., 2009). Thus, the presence of a free form of EDCs in the blood may indicate contamination from other sources (Yamamoto et al., 2016). In order to ensure that the plasma were not contaminated during collection and transportation, we randomly checked eight plasma samples (four from ASD children and four from the control children), and found that there was no free form of target EDCs in the plasma, which indicated that the samples analyzed in this study were not contaminated by extraneous sources. Third, all glassware used in the experiment was baked at 400 °C for 6 h and pre-washed with methanol before use. The detailed results of QA/QC are shown in Table 2. In each sample batch, two method blanks, two blanked spiked samples, and two matrix-spiked samples were carried out. The recoveries of internal standards in samples were 36-85%. Trace concentrations of several EDCs (from 0.01 ng/mL for PrP, BuP, HepP and TCS to 1.01 ng/mL for MeP) were found in procedural blanks. Trace concentrations found in procedural blanks were subtracted from the reported concentrations.

Statistical Analysis
Data analysis SPSS Software (Version 22) was applied to perform the statistical analysis. Non-parametric tests (Kruskal-Wallis H Test and Mann-Whitney U Test) were used to compare concentration differences of all target compounds between ASD and control groups, boys and girls, as well as children with different demographic characteristics due to abnormal distribution of the data. Concentrations below the limit of Table 2 The results of quality assurance/quality control system a LOD means the limit of detection; b LOQ means the limit of quantification; c Average and standard deviation. quantification (LOQ) were assigned values equal to half the value of LOQ. Spearman's rank correlation coefficients were used for the analysis of the relationship between two sets of data because of the limited sample size and the non-normal distribution of all data. Binary logistic regression was applied to investigate the relationship between specific chemicals and children with ASD. Statistical significance was set at p < 0.05.

Subject Characteristics
The characteristics of ASD and the control children are summarized in Table 1. For all participants, demographic information, including age, height, and body weight, as well as parental smoking, drinking, and engaging in poisonous or harmful work, were investigated. There was no significant difference in height and body weight between the two groups. The average age of children with ASD was slightly lower than that of the control group. However, there was no significant correlations between concentrations of each target compound and age, as well as height and body weight (p > 0.05). The smoking rate of fathers of ASD children (47%) was slightly higher than that of the control group (21%).

Occurrence of EDCs in ASD Children
Concentrations and detection frequency of parabens, BP-type UV filters, OH-PAHs, TCS and TBBPA in the plasma of ASD and the control children are listed in Table S2. Parabens were the dominant EDCs in ASD and the control children, with a median concentration of 2.18 and 2.47 ng/mL, accounting for 55.9% and 37.7% of total EDCs (Fig. S1). Other chemical groups followed the sequence of OH-PAHs (0.73 and 4.35 ng/mL for ASD and the control children) > BP-type UV filters (0.14 and 0.28 ng/mL) ~ TCS (0.13 and 1.22 ng/mL) > TBBPA (0.02 and 0.004 ng/mL), accounting for 29.0% and 35.5%, 5.05% and 4.0%, 8.49% and 22.5%, and 1.59% and 0.34% of the total EDCs, respectively. Compared with the control group, ASD children accumulated significantly higher TBBPA (p < 0.05) but lower BP-type UV filters, OH-PAHs and TCS (p < 0.05) (Fig. 1).
Parabens Parabens were widely found in the plasma of ASD children, with a detection frequency of 88-100% for MeP, EtP, PrP, BuP and BzP and 62% for HepP. The median concentrations of MeP, EtP, PrP, BuP, BzP and HepP in ASD children were 1.09, 0.32, 0.04, 0.003, 0.004 and 0.001 ng/mL, respectively. MeP and EtP were the major components, which accounted for more than 90% of total parabens (Fig. 2). The detection frequencies and concentrations of parabens in the control children were comparable to those of ASD children (Table S2). MeP and EtP accounted for more 88% of total parabens. No significant difference was observed between the two groups (p > 0.05). Parabens are commonly used as preservatives in a variety of daily necessities. To date, parabens in urine and blood of healthy children have been widely reported ( Kawaguchi et al. reported that rats exposed to iso-BuP in their early life showed impaired social behavior (Kawaguchi et al., 2010). However, this relationship was not observed between paraben compounds and children with ASD in this study. Significant differences in parabens were also not observed between ASD and the control children.

BP-Type UV Filters
BP-1, BP-8 and 4-HBP were detected in more than 90% of ASD and the control children, while BP-3 and BP-2 were detected in 53% and 38% (ASD group) and 93% and 32% (the control group), respectively. The median concentrations of BP-3, BP-1, BP-8 and 4-HBP in ASD and the control children were 0.02 and 0.09 ng/mL, 0.05 and 0.07 ng/mL, 0.02 and 0.02 ng/mL, and 0.03 and 0.04 ng/mL, respectively. BP-1 was the most abundant component, accounting for 39% and 34% of total BPtype UV filters, respectively, followed by 4-HBP (21% and 19%), BP-8 (20% and 9%), BP-3 (15% and 31%) and BP-2 (5% and 7%). BP-type UV filters are commonly used as sunscreen agents in a variety of cosmetic products (Chisvert et al., 2012;Kunisue et al., 2010). Due to their extensive usage, BP-type UV filters are widely detected in human (Gao et al., 2015). Significant differences in BP-3, BP-1 and 4-HBP were observed between ASD and the control children. ASD children had significantly lower BP-3 (0.02 ng/mL),  BP-1 (0.05 ng/mL) and 4-HBP (0.03 ng/mL) than the controls (0.09 ng/mL, 0.07 ng/mL and 0.04 ng/mL) (p < 0.05) (Fig. 3). It is worth noting that the differences between ASD and the control children occurred mainly in boys. The concentrations of BP-3, BP-1 and 4-HBP in ASD boys were significantly lower than those in the control boys (p < 0.05), but no significant differences were found between ASD and the control girls (p > 0.05). This indicated that there were gender differences in BP-3, BP-1 and 4-HBP between ASD and the control group, that is, unlike ASD girls, ASD boys showed significant differences in some EDCs levels than the control boys.
TCS TCS was detected in 97% and 100% of ASD and the control children. The median concentration of TCS in ASD children was 0.13 ng/mL, which was much lower than that in the controls (1.22 ng/mL) (p < 0.01). Similar to BP-type UV filters, the gender difference in TCS between ASD and the controls was also observed in boys (p < 0.05). The effects of TCS on ASD in children remain controversial. Etzel et al. reported that prenatal exposure to TCS was not associated with behavior and cognitive ability in children after birth (Etzel et al., 2018). However, Hao et al. speculated that TCS might induce autistic behaviors in rats by down-regulating cellular RA signaling, such as by altering the microbial composition of mother rats and their offspring (Hao et al., 2019). Jackson-Browne et al. also reported that urinary TCS concentrations were associated with significantly lower children's cognitive test scores in children (Jackson-Browne et al., 2018). The finding of this study cannot provide definitive evidence for the impact of TCS on ASD in children, but there is some evidence for a gender difference in TCS between ASD and the controls.
OH-PAHs 2-OH-Nap, 2-OH-Fluo and 2-OH-Phen were detected in 100% of ASD children, followed by 3-OH-Phen (91%) and 1-OH-Nap (79%). 2-OH-Nap and 2-OH-Fluo were major OH-PAHs in ASD children, with median concentrations of 0.33 and 0.29 ng/ mL. A previous study reported that specific PAHs, such as B(a)P, have a direct negative effect on the regulatory development of autism risk gene expression, which is associated with negative behavioral learning and memory outcomes (Sheng et al., 2010). The transcription and developmental expression patterns of the autism risk gene are highly sensitive to B(a) P. Compared with the control group, a relatively high detection rate and concentrations of 2-OH-Phen were found in ASD children. 2-OH-Phen was detected in 100% of ASD children, while it was only detected in 25% of the controls. Although the direct effect of 2-OH-Phen on children with ASD has not been demonstrated, the association between 2-OH-Phen and children with ASD should be seriously taken into account. After all, in a wild-type mouse model study, PAH-exposed Cpr lox/lox offspring showed a dosedependent increase in the production of PAH metabolites (Sheng et al., 2010).
TBBPA TBBPA was detected in 79% of ASD children, slightly higher than that in the controls (64%). ASD children had higher TBBPA (0.02 ng/mL, median) than the control group (0.004 ng/ mL) (p < 0.05), and the difference occurred only in boys. Evidence that TBBPA induces ASD is scarce. But it has been shown that TBBPA could increase levels of the presynaptic protein SNAP-25, which plays an important role in intracellular vesicular transport and exocytosis and is associated with hyperactivity and cognitive function in some neuropsychiatric disorders (Zieminska et al., 2016). The effects of EDCs on ASD in children are complex. Studies on the relationship between gestational EDC exposure and the behavior of ASD children revealed that some EDCs are associated with certain autistic behaviors, while others are either weakly or negatively associated with autistic behaviors . This is consistent with the results of this study, that is, children with ASD accumulated significantly higher 2-OH-Phen and TBBPA, but they exhibited lower BP-3, BP-1, 4-HBP and TCS. This may be due to several reasons. First, ASD might have some influence on the processing or metabolism of EDCs in children. Previous studies have shown that children with ASD exhibited a lower capacity for glucuronidation in the metabolic pathway of certain EDCs (diethylhexyl phthalate and bisphenol A). Lower capacity for glucuronidation could remarkably influence the ability to detoxify and eliminate toxic substances of the organism, and finally lead to higher levels of toxic substances in ASD children than in healthy children Stein et al., 2015). Second, the negative association between EDCs and ASD may be interpreted by the specific hormonal effects of EDCs. Although the specific hormonal effects of parabens, BP-type UV filters, OH-PAHs, TCS and TBBPA on ASD have not been reported, the target compounds in this study all have estrogenic activity. A study of perfluoroalkyl substances (PFAS) showed that amniotic fluid PFAS was inversely related to the risk of ASD and speculated this is related to the weaker estrogen and antiandrogenic effects of PFAS (Long et al., 2019). Similarly, it has been reported that higher levels of perfluorooctane sulfonate (PFOS) in the mother's blood are associated with reduced autistic behavior and that children with higher levels of prenatal PFOS have the better cognitive ability . A protective link was also found between PFAS exposure and cognitive ability in older adults (Power et al., 2012). Third, the sources of exposure to EDCs may be another factor to be considered. Human can be exposed to EDCs through ingestion, inhalation and dermal contact pathways. In this study, exposure source as one of the influencing factors is mainly because it is a factor that can not be completely ignored. Despite we tried best efforts to investigate the relevant information for each child, the contamination of EDCs in their daily living environment and the diet are still difficult to control. Therefore, the differences in EDCs between ASD and the control children may be partly attributable to the different sources of exposure.

Gender Difference in EDCs
Gender differences of BP-3, BP-1 and 4-HBP, TCS and TBBPA were evident between ASD and the control children, and the differences were more inclined to boys. This is in good agreement with those reports of a high prevalence of ASD in boys (CDC, 2012;NINDS, 2016;Nor et al., 2014). Although the direct evidence for this assumption is limited, studies have shown that some EDC exposure may affect children's behavioral outcomes in a gender-specific way (Eskenazi et al., 2015;Lin et al., 2017;Perera et al., 2012). This is confirmed by the high prevalence of ASD in boys (CDC, 2012;NINDS, 2016;Nor et al., 2014). It is hypothesized that there might be a hormonal etiology of ASD (Baron-Cohen et al., 2005). Maternal exposure to EDCs has been reported to cause different behavioral outcomes for offspring, and males are generally more sensitive to individual or mixtures of chemicals (Sobolewski et al., 2014). Furthermore, the biochemical, physiological and neural mechanisms of gender differences in response to environmental factors may also provide clues to the different susceptibility of males and females to ASD.

The Relationship between EDCs and ASD
Correlations between the concentrations of different EDCs are shown in Table S3. Significant positive correlations between 2-OH-Nap and total OH-PAHs were observed in both ASD and the control children, which resulted from the fact that 2-OH-Nap had the highest concentration among all OH-PAH compounds. Correlations between BP-3 and total BP-type UV filters were also observed in ASD and the control children, and this may be explained by the fact that BP-3 could be metabolized into other BP-type UV filters, such as BP-1, BP-2 and BP-8. A similar correlation was also found in our previous study on the urinary BP-type UV filters in young Chinese adults (Gao et al., 2015). For individual compounds, significant correlations between BP-3 and TCS, and 2-OH-Phen and TBBPA were observed in ASD boys (Fig. 4). However, these correlations were not observed in ASD girls as well as the control boys or girls.
Regression analysis was used to explore the possible relationship between specific EDCs and children with ASD. Firstly, binary logistic regression analysis was applied. BP-3, 4-HBP, TCS, 2-OH-Nap, 2-OH-Phen and TBBPA were regarded as independent variables. Concentrations of BP-3, 4-HBP, TCS, 2-OH-Nap, 2-OH-Phen and TBBPA were converted into categorical variables depending on the median concentration. In order to analyze the potential confounding factors, the covariates were selected by the following criteria: altered the β-value more than 10% for analytes, associated with the dependent variable (p < 0.01), and reported to a confounder (Vander-Weele, 2019). Finally, age, height and body weight were introduced as covariates to adjust results. The odds ratio (OR) refers to the ratio of exposure to non-exposure in the case group divided by the ratio of exposure to non-exposure in the control group. In a case-control study, OR values are indicators that reflect the degree of association between exposure risk factors and the disease. (Barkoski et al., 2019;VanderWeele, 2019) For the boy group, the OR values of 2-OH-Phen and TBBPA were 19.1 (95% CI: 4.12-88.9, p < 0.01) and 4.29 (1.20-15.4, p < 0.05) and their adjusted OR (aOR) values (aOR means correlations adjusted by age, height and body weight) were 11.0 (1.45-84.0, p < 0.05) and 1.78 (0.35-9.16, p > 0.05) (Table 3). Due to the limited sample size, binary logistic regression analysis was not applied to girls. Multi-factor regression analysis was conducted to identify the impact of multiple compounds on ASD. All compounds were introduced as independent variables. For the boy group, 2-OH-Phen and TBBPA were finally introduced to the equation as variables.
The OR values of 2-OH-Phen and TBBPA were 28.0 (2.92-268, p < 0.01) and 0.03 (0.03-0.33, p > 0.05).  This is consistent with the results of binary logistic regression analysis. Although relatively high OR values of 2-OH-Phen were obtained in ASD boys, no matter whether in single-factor regression analysis or multi-factor regression analysis, we still cannot conclude that 2-OH-Phen is a risk factor of ASD in children due to sampling size limitation. Further studies are needed to clarify the association between 2-OH-Phen and ASD.

Conclusions
This is a pilot study on multiple EDCs in children with ASD in South China. It firstly provides basic exposure data of some EDCs in ASD children in South China. Concentrations of several chemicals were significantly different between ASD and the control children. Gender differences in certain EDCs were evident, and the differences were more inclined toward boys. However, there were still some limitations. First, although all characteristics of concentration distributions in different groups were very clear, our study is still a pilot one, as the sample size is small, especially for each subgroup involved in the comparison. Limited sample size may lead to deviation in comparisons between genders, case and control groups, as well as between children with different demographic characteristics. Therefore, a large sample size study is needed in the future to further confirm our conclusions. Second, most of the target EDCs have short lives after entering the human body, so the effect of those chemicals on children was based on the hypothesis that children are continuously exposed to those chemicals for a long time. Third, in binary logistic regression analysis, though the results only suggested 2-OH-Phen may be associated with ASD, it needs to be emphasized that other compounds which were introduced to the regression models should not be overlooked due to their significant concentration differences in ASD children. Finally, we focused only on the association between individual chemical groups and ASD. The combined effects of mixtures of chemical groups on ASD have not been discussed due to the lack of relevant toxicological data. But in fact, these pollutants do not exist independently in humans but have synergistic or antagonistic effects in living organisms. This study cannot directly draw conclusions about these EDCs as risk factors for ASD, but it could serve as an early warning that EDCs such as 2-OH-Phen and TBBPA might be related to ASD in children, which is potentially harmful to children's health. Further studies are needed to clarify the association between EDCs exposure and ASD in Children.
Data availability Datasets used in this study will be available upon request.

Declarations
All authors have read, understood, and have complied as applicable with the statement on "Ethical responsibilities of Authors" as found in the Instructions for Authors". Yan-Yan Qin is the major corresponding author.

Conflicts of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The authors declare that they have no competing interests.

Institutional review board
The methodology applied in the present research was conducted in compliance with protocols approved by the Ethics Committee of Shenzhen Children's Hospital (20140176).
Informed consent All authors of this paper consent to participate.