Perinatal Risk Factors Associated with Autistic Spectrum Disorder

Autism Spectrum Disorder (ASD) is a lifelong neurodevelopmental condition with multifactorial etiology, resulting from complex interactions between genetic susceptibility and environmental exposures. Although numerous studies have identified individual perinatal risk factors for ASD, most have examined these exposures in isolation, limiting understanding of how perinatal complications cluster and jointly influence neurodevelopment. Evidence from Latin America also remains scarce. This study aimed to identify multivariate perinatal risk patterns associated with ASD in a Chilean population, addressing gaps in regional representation and methodological approaches. We conducted a population-based analysis of mothers of children with and without ASD in Chile. A broad set of medical and psychosocial perinatal variables was jointly analyzed using multiple correspondence analysis (MCA) to characterize interrelated risk structures. MCA revealed a clear separation between ASD and non-ASD groups along the first dimension, suggesting that ASD diagnosis is embedded within structured perinatal patterns rather than isolated exposures. MCA-derived, stepwise, and LASSO-penalized logistic regression models were then compared. The most parsimonious model identified maternal vaginal bleeding during pregnancy, prenatal maternal stress or anxiety, and negative pregnancy intention or perception as the strongest factors jointly associated with increased odds of ASD, with a dose-response pattern observed for maternal stress. An unexpected inverse association with neonatal cyanosis may reflect enhanced medical surveillance and warrants cautious interpretation. These findings underscore the importance of integrated perinatal care addressing both obstetric and maternal mental health, and demonstrate the value of multivariate approaches for elucidating complex developmental risk pathways.