Association between Cesarean section and neurodevelopmental disorders in a Japanese birth cohort: the Japan Environment and Children’s Study

Cesarean section (CS) is a lifesaving and important mode of delivery for neonates and mothers. Its use has increased around the world in recent years [1]. In Japan, CS deliveries as a percentage of total births have doubled since the 1980s and recently exceeded 20% of all deliveries [2]. There are many reports regarding the negative effects of CS on several health outcomes such as obesity, allergy, and asthma [3‐5]. The mechanisms underlying the link between CS and future health and mental disorders are believed to be early-term birth [6], altered microbiota [3], decreased serum oxytocin level [7], and use of general anesthesia during CS [8].

The long-term effects of a CS birth on child neurodevelopment are of increasing interest [9‐13]. The estimated worldwide prevalence of autism spectrum disorder (ASD) is 0.62% [14], and its incidence has increased 20-fold since the 1980s [15]. A recent study in Japan also reported an increasing trend in ASD diagnoses, from 2.23% (2009) to 3.26% (2014) [16]. There are conflicting results regarding the association between ASD and CS. In the UK, no association was found between CS and ASD [9]. However, in a meta-analysis including 13 studies, CS birth was linked to a greater risk of future ASD (odds ratio (OR) 1.23) than vaginal birth [11]. Although ASD is a highly inheritable disorder with male preponderance, relevant environmental factors may be sex-specific [9]; for example, the indications for CS and the anesthesia method used during CS were found to have some effects on ASD morbidity depending on the infants’ sex [8, 17, 18]. Some environmental factors in the perinatal period may also be involved in ASD morbidity. It is interesting to note that several birth cohort studies have reported increased risks of motor delay and intellectual disability associated with CS, but not the risk of ASD [9, 12, 13, 19]. Zhang et al. investigated the association between CS and neurodevelopmental prognosis in a population exceeding 1 million [10], and their findings suggested that CS was associated with a moderately increased risk of neurodevelopmental disorders in children; however, this risk was mostly explained by familial factors. However, the rate of CS was relatively low (12.4% of all births) among the many children studied and the risks associated with CS stratified by sex were not reported.

The inconsistent results reported about the effects of mode of delivery on subsequent neurodevelopmental problems, including motor delay, intellectual disability, and ASD [9, 10, 12, 19] have mainly been reported for Western populations and the effects in Asian populations are not known. Moreover, there have been very few studies of sex effects on neurodevelopmental disorders according to mode of delivery. Accordingly, in this study, we investigated possible associations between mode of delivery and neurodevelopmental disorders at 3 years of age, using data from the Japan Environment and Children’s Study (JECS), a large nationwide birth cohort study. The associations were also analyzed separately by sex to evaluate the sex-specific relationship between CS and neurodevelopmental disorders.

In this large cohort study of 65,701 mother–infant pairs, we found that children who were born by CS had a higher ASD risk at 3 years of age than those born by vaginal delivery. Moreover, females born by CS had higher risks of motor delay and ASD than females born by vaginal delivery. However, males born by CS showed no significantly higher risk of neurodevelopmental disorders.

After adjustment for potential confounding factors, the risk of ASD was 38% higher in children born by CS than in those born by vaginal delivery in this study. In the literature, there are conflicting reports regarding the association between CS and ASD [7, 10, 11, 17]. Previous studies and a meta-analysis indicated that CS increased the risk of ASD [7, 11, 17], which is line with the findings of the present study. However, the large cohort study by Zhang et al. (2021) in Sweden showed no association between CS and ASD [10]. That study investigated the diagnosis of ASD at an average age of 17 years, whereas we investigated this at 3 years of age. This large difference in the time of diagnosis may be a reason for this conflicting finding. Another reason could be the rate of CS and/or the ratio of emergent CS to elective CS. The rate of CS was 12.4% of all births in Zhang et al.’s study [10], while Japan’s CS rate is higher, at more than 20% [2]. The indication of CS might be different between Sweden and Japan. Although we did not investigate whether CS was emergent or elective in present study, emergent CS occurred in more than half of all CS cases in Zhang et al.’s study [10]. Common indications for emergent CS are related to fetal hypoxia, such as fetal distress, placental abruption, and dystocia, whereas those for elective CS are previous CS, cephalopelvic disproportion, and breech presentation, which have few hypoxic effects on the fetus. General anesthesia during CS was found to increase the risk of ASD compared with local anesthesia [8]. Many cases of emergent CS involve general anesthesia, whereas elective CS tends to involve local anesthesia. Moreover, elective CS tends to be conducted earlier, such as at 37–38 gestational weeks of age, when the neonate’s temperature and blood sugar levels are unstable and brain development has been interrupted [25]. Therefore, the ratio of emergent to elective CS would have some impact on the prognosis of infants.

In terms of neurodevelopmental outcomes, there was no difference between males delivered by CS and males delivered vaginally. However, females delivered by CS had higher morbidities of motor delay and ASD than females delivered vaginally (Table 3). Several studies have examined the sex-specific effects of CS on early childhood neurodevelopmental outcomes. For example, emergent CS was found to be associated with an increased rate of ASD in males, whereas elective CS was associated with an increased rate in females [17]. The risk of ASD in females born by CS was twice as high as that for females born by vaginal delivery [18]. Grace et al. reported that younger maternal age, smoking during early pregnancy, and stress during late pregnancy were associated with motor delay in girls [19]. They suggested that boys and girls were differently affected by antenatal and perinatal risk factors, due to sex-specific developmental pathways. The existence of sex differences in CS birth and subsequent health problems has been reported, with increased incidence of acute lymphoblastic leukemia and hepatoblastoma evident in females [26] and increased incidence of respiratory tract infections evident in males [27]. Females are thought to be more sensitive to the long-term effects of CS than males [17]. Sex differences in CS birth-related health problems have not been generalized, however, and the underlying mechanisms have not been elucidated. The sex-specific propensity may be at least partly due to sex chromosomal gene dosage and sex hormone levels. Experimental studies may be needed to clarify the factors associated with sex-specific incidence of ASD.

In this study, the risk of motor delay was 88% higher in females born by CS than in females born by vaginal delivery (Table 3). Females with ASD tend to have additional complications such as sleep disorder, developmental disorder, and emotional problems [28]. ASD in childhood is often complicated by gross and fine motor delay [29‐31]. Because females born by CS were more likely to develop ASD in the present study, this result may contribute to the higher rate of motor delay in females born by CS. Previous studies have reported associations of antenatal and perinatal risk factors, such as maternal pre-eclampsia, CS, and low income, with motor delay in childhood, and these perinatal risk factors may have a lasting effect on fetal neurological systems and postnatal motor development [12, 19]. On the other hand, children born by elective CS also showed motor delay at 9 months of age, although this delay had disappeared at 3 years of age [12]. Moreover, although CS birth was found to be associated with less white matter development in widespread regions of the brain and with less functional connectivity, these effects disappeared with age [32]. These reports indicate that children need to be followed up for longer periods to investigate the relationship between CS birth and neurodevelopmental effects.

The strength of our study is that we used data from a very large sample of mother–child pairs from all over Japan, including both rural and urban locations, so our results are likely to be representative of the Japanese general population. Moreover, we conducted the analysis after adjusting for many potential confounders. Nonetheless, our work had several limitations. First, we used mothers’ answers to JECS questions about whether their child had a neurodevelopmental disorder and it was therefore not clear how motor delay, intellectual disability, and ASD were diagnosed. However, every child in this study had received medical examinations several times, so we consider these diagnoses correct. Second is the timing of ASD diagnosis. Mild ASD cannot be detected at 3 years of age. Although the mean age of diagnosis for children with ASD is 4–6 years, clinical signs are usually present by 3 years of age [23, 33, 34]. In Japan, every healthy child has a medical examination at 3 years old to detect motor delay, language delay, and intellectual disability, and ASD is often diagnosed based on clinical signs such as delayed language. Therefore, we performed this investigation at 3 years. Third, no significant relationship was found between mode of delivery and intellectual disability in this study. However, two articles have reported that children born by CS showed reduced cognitive ability at school age [13, 35]. In our study, children were evaluated at 3 years old, which may be too young to be assessed for cognitive ability. We therefore plan to re-evaluate the participants at school age. Finally, we did not include familial factors, which matters because ASD tends to run in families. Although we did not include familial factors in our analysis, we did include many maternal psychological disorders such as depression, anxiety disorder, dysautonomia, and schizophrenia as cofounders, which may support the reliability of our data.

This study provides evidence of a significant association between mode of delivery and ASD in early childhood and contributes to the growing body of research attempting to identify the healthcare needs of females with autism and individuals with motor delay. Further research is required to determine the mechanisms underlying this association between CS birth and childhood neurodevelopmental delay and to identify the long-term impact on mental health.