Underlying factors of the association between cesarean section and autism spectrum disorder

Background: Autism spectrum disorder (ASD) is a lifelong neurodevelopmental condition characterized by impairment in social communication, and restricted, repetitive patterns of behavior, interests, or activities. A wide variety of prenatal and perinatal factors have proposed as risk factors of ASD in the last couple of years. Among these, the association of Cesarean section with ASD is particularly interesting given the continuous increase in the implementation of this procedure worldwide. Thus, understanding the underlying factors of this association is a public health priority.
Objectives: The goal of this study was to explore the contribution of various prenatal, perinatal and neonatal variables to the association between CS and ASD.
Method: We studied a wide range of prenatal, perinatal and neonatal characteristics in 347 children diagnosed with ASD, 117 children diagnosed with other forms of developmental delay (DD), and 2226 matched controls (matched by sex, birth date, and ethnicity at a 1:5 case-control ratio). Both cases and controls were ascertained from all single-live-born children at the Soroka University Medical Center (SUMC) between the years 2009 - 2016. Diagnosis of ASD and DD was determined according to DSM-V criteria by a child psychiatrist or child neurologist. Prenatal, perinatal and neonatal variables were obtained for both cases and controls from the electronic database of the obstetrics and gynecology department (OGD) of SUMC. The associations between each of these variables and ASD or DD were examined using appropriate univariate analyses. Significantly associated variables were further included in conditional logistic regression models that tested the adjusted effect of CS on the risk of either ASD or DD.

Results: Delivery by CS was significantly associated with ASD but not with DD (P-values = 0.019 and 0.540 respectively). Additional variables that were associated with ASD included general anesthesia (GA), parity number, and amniocentesis (P-value < 0.05). Stratification of CS deliveries to those, which were conducted with and without GA, revealed that only CS + GA elevate the risk of ASD (OR = 1.618, 95% CI: 1.176-2.226). Further stratification of the CS + GA deliveries to emergency and elective surgeries revealed a slightly higher risk associated with emergency surgeries (OR = 1.97 vs. 1.56 respectively), however this difference was not statistically significant (Breslow-Day test of homogeneity P-value = 0.174). Finally, repeating these analyses in subgroups of children with ASD according to their DSM-V severity levels, indicated that exposure to CS + GA affects only the most severely diagnosed children with ASD (OR = 2.52; 95% CI = 1.488 - 4.275).
Conclusions: Our findings indicate that the reported association between CS and ASD is restricted only to CS performed with GA. Nevertheless, our subgroup analysis suggests that exposure to GA during CS is unlikely causing to ASD, but rather contributes to the manifestation of additional symptoms that worsen the clinical condition of children with ASD.