Prenatal and Perinatal Factors Associated with Pediatric Autism Spectrum Disorder in East China


 Background

Autism spectrum disorder (ASD) is a complex neurodevelopmental condition characterized by persistent challenges in social communication and interaction and restricted/repetitive patterns of behavior, interests or activities. An increasing number of studies have revealed that environmental exposure is a potential risk factor for ASD. The relationship between prenatal and perinatal risk factors and ASD has rarely been studied in large samples in China. Therefore, in this study, we compared children with ASD with typical developing (TD) children to assess the ASD-associated prenatal and perinatal risk factors and provide effective information for ASD prevention.
Methods

A case-control study of 709 children with autism spectrum disorder (ASD) and 709 gender-matched children with typical development was conducted to investigate the prenatal and perinatal risk factors of children with ASD compared with children with typical development ( TD). Through a self-developed general information questionnaire, the basic information (name, age, gender), prenatal factors (parents’ age at the child’s birth, parents’ education levels, use of assisted reproductive technology, history of miscarriage, gestational diabetes mellitus, gestational hypertension), and perinatal factors (delivery mode, full-term birth, parity, birth weight) of the children in the two groups were examined.
Results

The prenatal and perinatal factors of the groups were submitted to univariate analysis, the parent’s age at childbirth, education level, history of miscarriage, use of ART, pregnancy-induced hypertension, and GDM differed significantly between the two groups (P<0.05), and that among perinatal factors, infant parity and maturity also differed significantly between the two groups (P<0.05). These statistically significant factors were included in a binary logistic regression model. The results showed that the prenatal factors of young maternal age at the child’s birth (≤24 years vs 25-29 years, OR=2.408,95%CI:1.335~4.345), old paternal age at the childbirth(≥45years vs≤24years, OR=4.744, 95%CI:1.281~17.570 ), pregnancy induced hypertension (OR=6.178, 95%CI:2.311~16.517) and GDM(OR=0.220,95%CI:0.149~0.324), the perinatal factors of preterm birth(OR=4.434, 95%CI:2.872~6.846) and non-firstborn child(OR=1.387, 95%CI:1.029~1.869) are likely risk factors for ASD.
Conclusion

We show that some prenatal and perinatal factors are associated with a high prevalence of ASD in children.

The relationship between prenatal and perinatal risk factors and ASD has rarely been studied in large samples in China (25)(26)(27). Therefore, in this study, we compared children with ASD with normally developing children to assess the ASD-associated prenatal and perinatal risk factors and provide effective information for ASD prevention.

Patient Group
A total of 709 patients who received clinical treatment at the Children's Mental Health Research Center of Nanjing Brain Hospital from July 2020 to November 2020, including 562 males and 147 females with an average age of (2.965±0.925) years, were investigated in this study. Frist of all, they were evaluated as positive for ASD according to the Chinese version of the Modi ed Checklist for Autism in Toddlers (M-CHAT), scores of ABC≥67(The ABC is a well-established parent report checklist used to screen for and diagnose autism).
And then attained the childhood autism diagnostic critical value on the Childhood Autism Rating Scale (CARS) scores≥30, diagnosed by two psychiatrist, met the diagnostic criteria for ASD in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) , If the two psychiatrist don't agree on a diagnosis, they're excluded.
Children with neurological diseases and serious physical, genetic or metabolic diseases, such as Rett syndrome and Fragile X syndrome, were excluded.

Control Group
A total of 709 healthy children, including 562 males and 147 females with an average age of (2.989±0.336) years, who were born at Nanjing Maternity and Child Health Hospital and underwent a follow-up physical examination were included in the control group. These children were assessed for the Bayley Scales of Infant and Toddler Development (Bayley-III) at the age of one year, con rmed to have typical intelligence and development, showed no physical or mental illness based on a physician's examination and evaluations, and were evaluated as negative for ASD according to the Chinese version of the Modi ed Checklist for Autism in Toddlers (M-CHAT).

Questionnaire and Information Collection
To collect the information of the patient group, we rst developed a questionnaire regarding possible risk factors for the onset of ASD based on a survey of the relevant literature (14,15,28,29). The questionnaire included (1) the children's basic information (name, gender, and age); (2) prenatal factors [parents' age at childbirth and education levels; maternal history of gestational hypertension, gestational diabetes mellitus (GDM), miscarriage and use of assisted reproductive technology (ART), etc.]; (3) perinatal factors (delivery mode, infant maturity, parity, birth weight, etc.). The questionnaire was distributed to the parents, and completed questionnaires were collected by the staff of the evaluation o ce of the Children's Mental Health Research Center of Nanjing Brain Hospital, who also ensured the completeness of the questionnaires.
The information for the control group was collected through online questionnaires or o ine interviews conducted by the staff of Nanjing Maternity and Child Health Hospital, who also ensured the completeness of the information.

Statistical Analysis
Some binary variables (e.g., history of pregnancy-induced hypertension, GDM, miscarriage, use of ART, diagnosed with ASD) were divided into "yes" and "no". The binary variable parity was divided into rstborn and non-rstborn. The binary variable infant maturity was divided into full-term birth 37-42weeks and preterm birth 28-37weeks . The binary variable delivery was divided into vaginal delivery and caesarean delivery.
The graded variables of parents' education levels was divided in three as junior high school or below , high school and college or above. The graded variables of paternal age at childbirth was divided in four as ≤24 years , 25-35 years , 36-44 years and ≥45 years. The graded variables of maternal age at childbirth was divided in three as ≤24 years, 25-29 years and ≥30 years. The graded variable birth weight was divided in three as low birth weight 2500g , normal birth weight (2500g-4000g) and high birth weight( 4000g). The difference in parental age at childbirth as continuous variable.
The data were analyzed using the SPSS 23.0 software. In the univariate analysis, count data were expressed as frequency and percentage, and the intergroup difference was analyzed with the chi-square test, measurement data were expressed as the mean ± standard deviation, and the intergroup difference was analyzed with a t-test. The risk factors that had statistical signi cance (P<0.05) in the univariate analysis were selected using the enter method and were included in the unconditional multivariate logistic regression model, which was subjected to the two-sided test at a level of α=0.05.

Univariate Analysis of the Two Groups According to Prenatal and Perinatal Factors
The results of the questionnaire survey showed that among the prenatal factors, the parent's age at childbirth, education level, history of miscarriage, use of ART, pregnancy-induced hypertension, and GDM differed signi cantly between the two groups (P<0.05), while the absolute value of the difference in parental age at childbirth did not(P>0.05), and that among perinatal factors, infant parity and maturity Among the perinatal factors, non-rstborn status signi cantly increased the risk of ASD in the offspring (OR=1.387, 95% CI 1.029-1.869, P<0.05), and preterm birth also signi cantly increased the risk of ASD in the offspring (OR=4.434, 95% CI 2.872-6.846, P<0.001) ( Table 2).

Multiple Regression analysis of prenatal and perinatal factors on severity of ASD
We also collected the CARS scores of 709 children with ASD, use a t-test, univariate analysis of variance, spearman correlation analysis to discuss whether there is a difference in the CARS scores among the investigated prenatal and perinatal factors. The risk factors that had statistical signi cance (P<0.05) in the univariate analysis were substituted into the multiple linear regression model, which was subjected to the two-sided test at a level of α = 0.05.
The results showed that, in the univariate analysis, the score of CARS is signi cantly different in paternal education level(F=1.158, junior high school or below vs college or above, p=0.004), maternal education level(F=0.905, junior high school or below vs college or above, p=0.035) and parity(non-rstborn vs rstborn, t=-2.216, p=0.027), multiple Linear regression analysis showed that the prenatal and perinatal factors investigated by this study did not affect the severity of symptoms in children with ASD. The results of multiple linear regression are shown in the Table3. In this study, we found that compared with fathers who were ≤24 years at childbirth, a paternal age of ≥45 years at childbirth signi cantly increased the risk of ASD in offspring. Our study was consistent in previous studies. Some studies showed that paternal age at childbirth is independently correlated with the incidence of ASD in offspring (30)(31)(32)(33), perhaps because de novo mutations (34,35) and epigenetic changes (36-38) that are affected by paternal age increase the risk of ASD. Other studies found that after adjusting for other perinatal factors and he parents' psychiatric history, the association between the risk of ASD and the age of mother or father is not statistically signi cant (39), or that after adjusting for the year of birth and socioeconomic status (SES), maternal age of ≤35years, only advanced paternal age at childbirth is related to the risk of ASD in offspring (32,40).
In the future, we will increase the sample size so to expand the range of paternal ages at childbirth. Furthermore, the confounding factors related to age at childbirth, such as socioeconomic factors and personality traits, will also be examined to understand their effects on the risk of ASD in offspring.

Maternal Age at Childbirth
Previous studies showed that advanced maternal childbearing age increases the risk of ASD in offspring (16, [41][42][43]. Accumulations of mutations, an increased incidence of complications, increased exposure to drugs or pollution (14), uniparental disomy (44), endocrine disruptors (45), and omega-3 fatty acid supply (46) may be the causes of the increased risk of ASD in the offspring of older pregnant women.
We found that younger mothers (≤24 years old vs 25-29 years old) had an increased risk of ASD in their offspring, which is consistent with the nding of a previous study. This may be because young age is related to education, employment, and other SES indicators. Studies have shown that low SES may be a risk factor for ASD and is causally related to a young age at childbirth. This complex interaction between SES and maternal age may be accompanied by risks of other adverse consequences, including maternal infections and poor health surveillance (47), and is likely to be associated with adverse pregnancy outcomes (48, 49). In addition, without adjusting for SES, the maternal childbearing age at childbirth shows a U-shaped relationship with illness in the offspring (50) or with ASD with intellectual disability in offspring (40). In other SES-adjusted studies, a monotonic association with maternal age is reported (30). Previous studies suggest that a very young or old age at childbirth will increase the risk of ASD in offspring. Compared with those studies, we need to increase our sample size and include samples from more regions while thoroughly recording household socioeconomic factors, such as parental occupation, education level, and income, so that the impact of age and other confounding factors on ASD in offspring can be better understood.

Parental Education Levels
In this study, we found that as the education levels of parents increases, the risk of ASD in the offspring is reduced, likely because low education levels in parents affect their SES. It was reported that the low measured prenatal SES is correlated with an increased risk of ASD in offspring (51). In addition, children with ASD from families with lower SES show increased emotional and behavioral problems (52) and a greater risk of intellectual disability (53,54). Moreover, the SES-based differentiation of diagnostic services may also cause families in poorer communities to have a delayed diagnosis, which delays intervention and increases the severity of ASD symptoms in children from these communities (54)(55)(56)(57)(58).
Studies have also shown that language delay is related to the paternal education level (59) and that the use of positive parenting styles is related to the maternal education level (60). The higher the education levels of the parents, the more knowledgeable they are about the disease; as a result, they will nd various professional resources to help them understand their child's symptomatic behavior and then adopt a more rational way to treat the child, while the individuals with low SES often have insu cient resources for these purposes (61).
In the initial data collection stage of this study, we found that the patient group was dominated by those from outside the city who were making special trips to the clinic and that the parents of these patients were mostly migrant workers with low education levels, while the subjects in the normal control group were mostly children seeking outpatient consultations, physical examinations, or assistance with other medical conditions, and these families were evenly distributed throughout the city's household registration. This may bias our results; therefore, in future studies, we will increase the sample size and include more socioeconomic factors (e.g., occupation and family income) for a comprehensive consideration.

High Blood Pressure
In this study, we found that hypertensive disorder of pregnancy (HDP) increases the risk of ASD in the offspring, which is consistent with the results of previous studies. It has been reported that HDP is associated with an increased risk of mental disorders in offspring (62)  This may be because that HDP not only complicates 5-10% of pregnancies (69), which affects maternal health and pregnancy outcomes, but it also affects fetal brain development due to an unfavorable fetal environment, such as poor placental vascularization, which makes the infant more susceptible to neurobehavioral disorders. In addition, due to inadequate uterine-placental perfusion and hypoxia, HDP may be related to reduced oxygen and nutrient supply to the fetus (70,71), which can damage the fetus's

GDM
In our study, we nd that GDM increases the risk of ASD in the offspring, which is consistent with some previous reports (79)(80)(81)(82)(83)(84)(85). Studies have suggested that GDM's impact on neurodevelopment may be mediated by oxidative stress (86), pro-in ammatory cytokines (87), and epigenetics (88). There were also studies that suggest that GDM causes a variety of adverse obstetric outcomes (e.g., preeclampsia, fetal macrosomia, perinatal mortality, caesarean section, and premature delivery) rather than simply increasing the risk of neurodevelopmental disorders (89-91). In addition, studies have shown that a higher body mass index or pregestational diabetes mellitus instead of GDM increases the risk of ASD in offspring (84, 92), or that only pregestational diabetes mellitus and GDM in obese mothers increases the risk of ASD in offspring (93), suggesting that the exposure to accompanying in ammation, lipotoxicity, metabolic stress, and hyperglycemia may have stronger neurological impact than hyperglycemia alone (94)(95)(96)(97). Some studies indicate that GDM at different stages of pregnancy can lead to different neurodevelopmental conditions (83, 98).

ART
In this study, we found that ART is signi cantly associated with ASD and may be a protective factor against the disease, which is inconsistent with the results of some previous studies (99,100). But the results are consistent with some previous literatures (101,102). This may be because women with assisted conception have better pre-pregnancy preparation and reduce behaviors that negatively affect their health through close contact with the healthcare system before and during pregnancy (103). The health status of pregnant women is related to the intellectual disability of their offspring, and therefore, it may be bene cial for pregnant women to have access to advice on promoting good growth conditions for the fetus before and during pregnancy. Some studies indicate that folic acid (FA) intake in early pregnancy and during pregnancy plays an important role in the etiology and severity of ASD; thus, taking FA may be a protective factor (104). Another possibility is that certain elements in the maternal diet (fatty acids, vitamin D, and iron) may play a protective role by combating some of the core ASD symptoms (105,106). For example, in the late stage of pregnancy, higher iron intake is associated with a reduced risk of ASD since iron contributes to neurotransmitter production, myelination, and immune function. In addition, vitamin D supplementation during pregnancy can reduce the risk of ASD in offspring (107, 108).
However, some studies have shown that ART can increase the risk of ASD in offspring (109,110). This may be due to several procedures used in ART, such as hormone stimulation, egg retrieval, in vitro fertilization, intracytoplasmic sperm injection, gamete micromanipulation, and exposure to culture media, which may cause environmental stress in gametes and early embryos, which in turn may be associated with an increased risk of birth defects (109). The inconsistency of our ndings with those of other studies indicates that further prospective, large-scale, and high-quality research is needed. The limitation leading to this inconsistency is derived from the heterogeneity of the studies, especially in terms of study design (cohort and case-control), ART data recruitment strategies (registration, medical records, parent interviews), and the evaluation of confounding factors.

Preterm Birth
In this study, we found that preterm birth increased the risk of ASD. This is consistent with previous research reports that abnormal gestational age (premature delivery, late delivery) can increase the risk of ASD (39,82,111).
This nding may re ect damage to neurodevelopment that occurred during the prenatal period. Notably, intrauterine growth restriction (IUGR) derived from placental insu ciency is a condition associated with chronic hypoxia, metabolic disorders, and acidosis (112,113). The neonatal consequences of intrauterine acidosis include hypoxic-ischemic encephalopathy, intraventricular hemorrhage, and periventricular white matter softening (114). Therefore, IUGR is associated with poor neurodevelopmental results (115), which can lead to behavioral and cognitive di culties, cognitive de cits (116-118), and ASD manifestations in children. In addition to placental dysfunction, IUGR can be caused by intrauterine infection or genetic abnormalities, both of which are related to neurodevelopmental abnormalities (119,120).
The increased risk of ASD in children with post-term birth may re ect abnormalities in maternal metabolism, such as obesity, in ammatory changes associated with abnormal maternal metabolism, delivery complications, birth injuries, or postpartum metabolic disorders. For example, circulating in ammatory cytokines and acute phase proteins are systemically increased in obese women (121). In ammatory mediators have been shown to be able to pass through the placenta and alter neurodevelopment in animal models (122). It has been speculated that in addition to having a direct impact on fetal neurodevelopment, exposure to intrauterine in ammation can trigger lifelong immune dysfunction in the fetal immune system, which may lead to the underlying abnormal neurodevelopment and behavioral abnormalities observed in ASD (123).

Parity
In this study, we found that non-rstborn children had a higher risk of ASD, which is not completely consistent with the ndings of previous studies. Previous studies have shown that in families with no children with ASD, a pregnancy interval of less than 24 months is related to an increased chance of a second or third child suffering from ASD, and a pregnancy interval of less than 12 months presents the highest risk; this indicates that with shorter pregnancy intervals, children born later have a continuously increasing risk of having ASD (124). This may be due to (1) the association of shorter pregnancy intervals with adverse pregnancy outcomes (including low birth weight and premature delivery) (125), (2) depletion of maternal nutrients, especially FA (FA is required for DNA synthesis and cell division during pregnancy. Without supplementation, FA levels in serum and red blood cells begin to decline in the second trimester, and the FA level in red blood cells continue to decline for at least 12 months after delivery (126, 127)), and (3) other possible mechanisms, such as maternal iron and polyunsaturated fatty acid levels (128) or stress (129). Conversely, some studies have shown that rstborn children have a higher risk of having ASD, likely because families with a child with ASD may delay further pregnancies, a phenomenon called reproductive stoppage (130)(131)(132)(133)(134). Moreover, the ASD phenotype is highly heterogeneous. In some patients, symptoms persist beyond early childhood, while in others, these symptoms only appear in a milder form in late childhood. Therefore, even for a family with a child with ASD, the severity of the child's ASD phenotype may affect the family's decision to have another child (135). The inconsistency of our results with those of other studies suggests that the birth interval of between children should be considered when studying the impact of parity on ASD and that it is important to review the neurodevelopment of children in families with multiple children to perform a classi cation analysis of the family's ASD exposure. In addition, we need to pay attention to the psychological stress that parents are likely to experience when raising a child with ASD. With the current changes in fertility policy in China, these investigations can provide references for families planning to have a third child.
This study has certain limitations. First, the sample size was small compared with that of the same type of studies conducted in other countries, and consequently, we were unable to include some possible risk factors or negative results in this study. Additionally, the ages of the children in the patient group and those in the control group were not matched, and because this study was retrospective, the results may be subject to recall bias. However, there was no signi cant difference in gender between the two groups of children, which eliminates the impact of gender on the results. This study collected the largest reported sample of participants from East China; therefore, the results can be used as a reference for clinical and scienti c research.
In short, in this study, we found that the prenatal factors of a maternal age at childbirth of ≤24 years and pregnancy-induced hypertension and the perinatal factors of non-rstborn children and preterm birth may be risk factors for ASD. Improving the living environment during pregnancy, reducing complications during pregnancy, and avoiding exposure of the mother to adverse environments during the prenatal and perinatal periods can be effective entry points for the prevention and treatment of ASD. Moreover, we also found that factors that may be related to the SES of parents may affect the incidence of ASD in offspring. Against the background of China's current three-child policy, it is very important to pay attention to the factors that affect the health of offspring. However, due to the limited sample size and representativeness of this study, its conclusions may not fully represent all prenatal and perinatal risk factors for ASD. It is necessary to conduct more studies in the future to understand the adverse factors of ASD so that more information on the prevention and treatment of pediatric ASD can be provided to families.

Limitations
The present study has limitations. The number of cases in this study is still relatively small compared to similar risk factor studies abroad, leading to some possible risk factors can't be included in this study, or negative results. Conclusions may not be representative of all prenatal and perinatal risk factors for ASD. Future research should further expand the sample size, in-depth study of environmental factors on the impact of disease, to provide more effective and comprehensive prevention and treatment information for children with ASD.

Conclusion
We show that some prenatal and perinatal factors are associated with a high prevalence of ASD in children. Improving the living environment during pregnancy, reducing complications during pregnancy, and avoiding exposure of the mother before and during the perinatal period can be an effective approach to the prevention and treatment of ASD. In this study, we also found that the socioeconomic status of parents may be related to

Declarations
All methods were carried out in accordance with relevant guidelines and regulations.
Ethics approval and consent to participate All parents of children involved in the questionnaire signed an informed consent form. All experimental protocols were approved by the Nanjing Brain Hospital Ethics Committee (Approval certi cate number: 2017-KY098-01). Ethical approval for the study was granted by the China Clinical Trial Registration Center (Name of the Ethic Committee: Nanjing Brain Hospital Ethics Committee), ChiCTR-OPC-1701 1995.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used or analysed during the current study available from the corresponding author (Xiaoyan Ke) on reasonable request.

Competing interests
The authors declare that they have no competing interests ND and ZYZ and LH support the part of data collection YL support the part of data resources XYK designed the study and oversight the manuscript