It is now estimated that autism spectrum disorders (ASD) affect about 1% to 2% of the population. Data derived from twin and sibling studies indicate that ASD is highly heritable. While genetic factors clearly play an important role, other studies point to an interplay between environmental and genetic factors in the etiology of this disorder. Over the years, various environmental exposures have been implicated, including vaccinations, mercury, and insecticides; however, it has been difficult to generate solid evidence linking specific environmental factors with increased risk for autism.
A recent article published first online in the Archives of General Psychiatry  explores the association between autism and maternal SSRI use during pregnancy. In this population-based case-control study, the investigators evaluated the medical records of 298 children with autism spectrum disorder (82.9% male) and 1507 healthy controls from the Childhood Autism Perinatal Study. All were born between January 1995 and June 1999 at a Kaiser Permanente facility in Northern California. Medical records were used to identify children with autism, Asperger syndrome, and pervasive developmental disorder.
This study found a greater risk of autism spectrum disorders among children of SSRI users compared to non-users during pregnancy: 6.7% of the children with ASD (n = 20) versus 3.3% of the healthy controls (n = 50) had prenatal exposure to at least one antidepressant. The authors attempted to control for various potential confounding factors, including maternal history of psychiatric illness. Maternal symptoms and burden of illness during pregnancy were not assessed.
The risk for ASD was significantly increased for children with exposure to any SSRI during the year before delivery (adjusted odds ratio [AOR]=2.2); the risk was highest when the exposure occurred during the first trimester (AOR=3.8). Among those children with mothers who had a history of mental health treatment but did not take SSRIs during pregnancy, there was no increase in risk for ASD.
The study generates questions for further study, but does not necessarily support a cause-and-effect relationship between prenatal SSRI exposure and autism. In an interview for Medscape , Dr. Lisa Croen lead author of the article and director of the Autism Research Program at Kaiser Permanente Northern California (KPNC) stated, “It is possible that the association we found between maternal SSRI use and autism is in fact explained by the underlying condition for which the woman took the medication.”
Because children with ASD are more likely to have a family history psychiatric illness, this study attempted to distinguish between the effects of medication exposure versus the effects of the underlying disease that led to treatment. While previous history of mood and anxiety disorders is factored into the analyses, acute illness during pregnancy is not factored in. Multiple studies suggest that untreated depression may lead to physiologic changes which negatively affect fetal development and birth outcomes; thus, this omission greatly diminishes the value of the article to assess the impact of SSRIs in pregnancy on the risk of autism in children.
Current use of SSRIs may be a better indication of acute illness during pregnancy than history of previous diagnoses (as many women are undertreated during pregnancy due to concerns about fetal exposure, hence leading to a dual exposure to both medication and untreated maternal illness). Since the authors state that “the combined effect of maternal serotonin transporter genotype and prenatal stress may contribute to autistic-like behaviors in offspring,” it is disappointing that their study did not control for prenatal maternal stress or burden of illness, nor did it control for postpartum depression or other disorders. Therefore, it is not possible to separate out genetic predisposition and symptoms experienced by mothers during pregnancy – suspected to contribute to risk of autism- from medication exposure in this study.
At this time, we need further research to definitively determine whether maternal anxiety or depression or medication use may increase the risk of autism spectrum disorders, and whether other unknown factors associated with maternal mental health conditions or medicine use modulate this risk. Until risk factors are clearer, women with serious mood and anxiety disorders that do not find relief from non-medication treatments such as psychotherapy should discuss the risks and benefits of medication treatment with their doctors, and antidepressants are still considered first-line treatment for moderate to severe illness. Women with more mild illness, as before this report was published, should be encouraged to try psychotherapy first, and perhaps other non-medication strategies.
Marlene Freeman, MD
Ruta Nonacs, MD PhD
Croen L, Grether J, Yoshida C, Odouli R, Hendrick V. Antidepressant Use During Pregnancy and Childhood Autism Spectrum Disorders  Arch Gen Psychiatry [epub ahead of print].
Commentary in Medscape