Fig. 2

a A forest plot of the association between cesarean section (CS) and Autism Spectrum Disorder (ASD) in total study samples; b Adjusted odds ratios and two-sided 95% confidence intervals of ASD following CS compared with vaginal delivery in different anesthesia regimen subgroups (regional anesthesia [RA] vs. general anesthesia [GA]); c Adjusted odds ratios and two-sided 95% confidence intervals of ASD following CS compared with vaginal delivery in different surgery indication subgroups (emergency CS, voluntary CS and indicated CS). Gray dashed horizontal lines separate the results of different stratification analyses. Adjusted odds ratio for CS was estimated from logistic regression after adjusting for confounding factors (e.g., parity number, past miscarriage history, subject’s sex and age, and past miscarriage history*CS). Adjusted odds ratio for different surgery indication was adjusted for confounding factors (parity number, past miscarriage history, children’s sex and age). Adjusted odds ratio for anesthesia regimen was adjusted for confounding factors (parity number, past miscarriage history, infant’s weight and children’s sex and age)