Over the last several years, the reproductive safety of selective serotonin reuptake inhibitors (SSRIs) has been the focus of continual interest and concern, with data from several studies informing the question of relative risks of fetal exposure to these medicines. Two recent analyses from large birth defects surveillance programs suggest that the risk for major malformations associated with first-trimester exposure to SSRIs is extremely small (N. Engl. J. Med. 2007;356:2684-92, 2675-83).
On the basis of the results of other studies that are particularly consistent, there now appears to be a consensus that late-trimester exposure to SSRIs is associated with an increased risk for symptoms broadly described as “poor neonatal adaptation” including: tremulousness, jitteriness, and myoclonus. These symptoms tend to be transient, and their clinical relevance appears to be limited—because infants born with these symptoms typically do not require any particular clinical intervention.
More recently, an increased risk of a far more serious condition, persistent pulmonary hypertension of the newborn (PPHN), has been associated with exposure to SSRIs late in pregnancy. The first study to describe this association, derived from a case-control study using data from a birth defects database, reported that in utero SSRI exposure after 20 weeks' gestation was associated with an increased risk of PPHN (about sixfold higher), with an absolute risk approaching 1% (N. Engl. J. Med. 2006;354:579-87). When it was published, the study elicited great concern among providers and patients; the Food and Drug Administration also suggested inclusion of the finding in the labels of SSRIs.
Factors driving the risk for PPHN have been studied extensively and were the focus of a case-control study of the same birth defects database published last year that found a strong association between PPHN and several common prenatal and perinatal factors, including cesarean delivery, high maternal BMI, being African American or Asian, and being large for gestational age (Pediatrics 2007;120:e272-e82). Though extensive discussion was not given in that article about the relative contributions of these common factors to risk for PPHN, compared with SSRI exposure late in pregnancy, the data suggest that large BMI and cesarean delivery, for example, drive PPHN to a far greater extent that SSRI exposure.
The association between SSRI exposure during pregnancy and PPHN was again addressed recently in a case-control study using data from the Swedish Medical Birth Register of over 800,000 infants born between 1997 and 2005; 506 infants had a discharge diagnosis of PPHN. Across a wide variety of SSRIs to which women had been exposed, the risk of PPHN associated with maternal use of an SSRI among babies born after 34 weeks' gestation was increased by 2.4, after adjusting for other risk factors (Pharmacoepidemiol. Drug Saf. 2008;17:801-6). This finding was among the women who reported using an SSRI when interviewed early in pregnancy, before the end of the first trimester—a marginally statistically significant increase. These results are consistent with the results of the 2006 study, but suggest a more attenuated risk and might explain why physicians with patients treated with an SSRI during pregnancy rarely see this condition.
The results may be considered somewhat reassuring because they suggest that even if there is an increased risk, the increase is small. Considering the prevalence of SSRI use during pregnancy—as high as 6%–8%, according to several recent studies—PPHN presumably would be seen more often if the risk were as high as 1%. (On the other hand, it is common to hear about neonatal adaptation symptoms in babies who have been exposed in late pregnancy to SSRIs, where estimates of the prevalence of such symptoms are 25%–30%.)
One of the obvious advantages of the Swedish study was that the information about SSRI use was obtained prospectively and did not rely on retrospective recall, as was the case in the 2006 study. But no study is perfect and, in the Swedish study, the investigators were not able to distinguish between early and late exposure. Late exposure was extrapolated by assuming that women on the drug early in pregnancy were also on the drug late in pregnancy. That is not a trivial limitation, because of evidence indicating that is not uncommon for women who are on an antidepressant in early pregnancy to discontinue treatment after either documentation of pregnancy or later in pregnancy.
The results of new studies sometimes either refute current associations or suggest greater or lesser associations. Ultimately, providers and their patients make collaborative decisions, and treatment will be tailored to match the clinical situation using the best available data possible.