Over the past decade, there has been a steady stream of information and published studies on the safety of antidepressants during pregnancy. Despite small sample sizes, methodological flaws, and other limitations of many published studies, we were grateful for useful information on the reproductive safety of antidepressants, primarily selective serotonin reuptake inhibitors.
Compared with those early years, we now have a vast amount of data and numerous published studies on the subject, providing us with more information regarding the spectrum of outcomes associated with prenatal exposure to SSRIs. Analyses of those data have become more sophisticated, with the appreciation that any assessment of reproductive safety must take into account both the potential risks of exposure to the medication and the risks of untreated psychiatric illness.
Dr. Lee Cohen
In this column, I will not review the extent to which SSRI exposure increases the risk of major congenital malformations, addressed in previous columns. Based on the literature over the last 10-15 years, it is reasonable to conclude that, if the risk of major congenital malformations following first-trimester exposure to SSRIs is increased, the increase is small (N. Engl. J. Med. 2007;356:2732-3).
The effects of SSRI exposure on other outcomes such as preterm delivery, birth weight, and Apgar scores have been less clear. These outcomes are gaining an increased presence in the literature, and interpretation of the data from either small cohort studies or large administrative databases can be extremely confusing for both biostatistician and clinician alike.
An important study, a meta-analysis of 23 studies published online in February, rigorously evaluated pregnancy and delivery outcomes following prenatal exposure to antidepressants (including SSRIs, tricyclics, and monoamine oxidase [MAO] inhibitors) – mostly SSRIs – factoring in the risk of untreated psychiatric illness on these outcomes. There were no significant associations between prenatal exposure to antidepressant medications and the risk of spontaneous abortion, with no significant differences among exposed vs. nonexposed pregnancies (JAMA Psychiatry 2013 Feb. 27:1-8 [doi:10.1001/jamapsychiatry.2013.684]).
There were, however, significant associations between prenatal antidepressant exposure and gestational age, preterm delivery, lower birth weight, and lower Apgar scores. Antidepressant exposure was significantly associated with a shorter gestation, and preterm delivery was more common among women taking antidepressants, whether or not they were compared with all women not exposed to antidepressants or just depressed women who did not take antidepressants during pregnancy.
Similarly, prenatal antidepressant exposure was significantly associated with lower birth weight (a mean difference of 74 g) when compared with birth weights of babies with no prenatal exposure. But when the comparison was limited to depressed mothers with no exposure to antidepressants, the association was no longer significant.
Antidepressant exposure also was significantly associated with lower Apgar scores at 1 and 5 minutes when compared with all the unexposed mothers and to only the mothers who were depressed but did not take antidepressants. (The study was funded by Canadian government grants, and 4 of the 11 authors disclosed having received honoraria, an unrestricted educational grant, and/or research support from companies that included antidepressant manufacturers).
What do these results mean to the clinician trying to translate these scientific findings into clinical practice? For the clinician, the most critical issue to consider is that, even when there were observed differences in outcomes between exposed and nonexposed pregnancies, those differences were small. Gestational age was less than half a week shorter, and the differences in the mean 1- and 5-minute Apgar scores were less than half a point and the scores were considered excellent.
These results are reassuring because the clinical relevance of the cited differences between the two groups is exceedingly small. Perhaps the most critical finding in this published analysis is that even a statistically significant difference in an outcome only has relevance if the difference informs clinical care. This is not an Olympic downhill ski race, where a hundredth of a second can make a critical difference.
Dr. Cohen is the director of the Center for Women’s Mental Health at the Massachusetts General Hospital in Boston, which provides information about pregnancy and mental health at www.womensmentalhealth.org. He has been a consultant to manufacturers of SSRIs.