Unfortunately, the bottom line for most of these important questions is that we really don’t know as much as we probably should.
Just when we’ve read a convincing finding from a reputable journal that establishes a link between prenatal SSRI use and an untoward outcome, we see it disputed the next month. Why is this always happening, and why can’t we really know anything with certainty? Much of the confusion can be attributed to research methods and the obvious difficulty of using randomized, controlled trials to control for potential confounding factors. While statistical techniques have become increasingly sophisticated in addressing these confounding factors, they remain imperfect. For example, one of the most difficult challenges that remains is separating any effects of a medication from any effects caused by the condition it was designed to treat. Comparing women with the same underlying condition, some of whom are treated with a medication and some of whom are not, is a step forward, but there may be important reasons that one group decides to seek treatment and the other doesn’t. One clever research design that was employed to look at congenital anomalies in the offspring of women taking SSRIs accounted for siblings of these children who were born when their mother was not taking an SSRI. This study demonstrated that these women were more likely to have children with congenital malformations even when they weren’t taking the SSRIs.1 Other factors that render this literature difficult to interpret include small sample sizes when looking at specific SSRIs (many studies cluster them all), dose effects, timing (which trimester), duration of treatment, and method of recording compliance.
The potential link between SSRIs and autism has received a fair amount of attention lately, especially after a very well-designed study in 2016 suggested a significantly increased risk.10 However, as with many of the findings, this study was quickly disputed by other high-quality, well-powered research that found no increased risk after controlling for maternal illness.11,12
ADHD generally has not been found to be related to maternal SSRI use, although one study did find a link between ADHD and tricyclic antidepressants.12,13
In terms of other neurodevelopmental outcomes, there have been many negative studies examining IQ, nonverbal communication, as well as speech and motor skills.14,15,16 However, as with so many other outcomes, some other studies contradict these negative results. According to a recent, large cohort study, there may be some concern regarding SSRI exposure prenatally and an increase in speech disorders by age 14 years, as well as lower language competence at age 3 years.17,18 Likewise, mild motor abnormalities have been observed, with maternal depression severity as an independent but contributing factor.19
Several studies demonstrate a connection between prenatal SSRI exposure and childhood internalizing symptoms, such as depression and anxiety, independent of maternal depression.12,20 These findings must be balanced with our knowledge of the serious mental health conditions in offspring that are associated with untreated maternal illness, including both internalizing and externalizing disorders.21,22
How does one come to any firm conclusions to guide a primary care clinician’s practice and recommendations? Hopefully, the evidence will become clearer over time as we adopt more sophisticated designs and accumulate observations. A larger number of observations would allow us to decrease heterogeneity by studying subgroups according to type of SSRI and duration of exposure. Enhanced understanding of the role of genetic factors also may shed some light on individual variation as the serotonin transporter gene has been suggested as a potential moderator of sensitivity.23
Dr. Guth is an assistant professor in the department of psychiatry at the University of Vermont Medical Center and the University of Vermont Robert Larner College of Medicine, both in Burlington. She works with children and adolescents as well as women in the perinatal period. She has no relevant financial disclosures.