Autism Risk in Children Born to Women with Perinatal Psychiatric Diagnoses

Background: Psychiatric disorders affect up to one-quarter of the U.S. population and are particularly common in women of reproductive age.  When women with mental health issues become pregnant, their conditions are often unrecognized and untreated because of inadequate screening and persistent stigma that deters spontaneous disclosure of symptoms and previous diagnoses.  When a psychiatric disorder is identified, concerns arise regarding its impact on the pregnancy and developing fetus.  Maternal mental health disorders have been associated both individually and collectively with autism spectrum disorder in multiple contexts, but most investigations have been performed in restricted populations or have targeted specific maternal conditions.

Objectives: To determine the extent to which maternal psychiatric diagnoses reported during the birth hospitalization were associated with offspring autism risk in a very large, socio-demographically diverse 18-year statewide birth cohort.

Methods:  This retrospective cohort study linked California hospital records for 8,951,763 singleton births occurring 1/1/91-12/31/08 from the Office of Statewide Health Planning and Development with neurodevelopmental diagnostic and treatment service records provided by the Department of Developmental Services (DDS). Maternal psychiatric diagnoses - schizophrenia, bipolar disorder, major depressive disorder, dissociative and factitious disorders, somatoform disorder, dysthymic disorder, and depressive disorder-NOS - were culled from delivery discharge diagnoses via ICD-9 codes.  DDS diagnosis of autism was the outcome in 42,423 children. Prospective data collection insured that maternal exposures were recorded prior to the child’s autism diagnosis. Logistic regression models were created to explore the relationship between maternal psychiatric disorders and autism, adjusting for maternal age, education, race, and country of birth.

Results:  California children with autism were similar to those in other investigations, with inflated risk for males, older parents, higher parental educational attainment, and lower risk for Hispanics and mothers born in Mexico. Reported prevalence of psychiatric disorders in pregnancy was far lower than expected, likely reflecting under-recognition of these conditions and under-reporting by clinicians for all but the most severe of presentations.  In adjusted analyses, mothers diagnosed with each individual psychiatric condition - mood and anxiety disorders as well as schizophrenia - were 1.4-2.2 times more likely to have a child who developed autism (Figure).  Mothers diagnosed with any mood, anxiety, or schizophrenic disorder during pregnancy were 80% more likely to have a child who developed autism compared with women without mental health conditions (RR= 1.79; 95% CI 1.66, 1.92).

Conclusions:  Women diagnosed with a psychiatric disorder during pregnancy were nearly twice as likely as women without such diagnoses to bear a child ultimately diagnosed with autism.  Mechanisms underlying these associations are likely to represent a complex web involving genetic predisposition, alterations in overlapping neurodevelopmental pathways, maternal stress, variations in mother-child interactive behaviors, and / or fetal exposure to certain psychiatric medications. These findings highlight the need for routine prenatal screening for and documentation of psychiatric conditions to improve outcomes for both mother and child.  Further, we recommend enhanced neurobehavioral screening for children born to mothers with a history of psychiatric conditions in order to detect early signs of autism and optimize the timeliness of interventions.