It’s been a year and a half now since TCR last took up the perennially hot topic of SSRI use in pregnancy (TCR Vol 1, No 5, 2003). Our bottom line assessment then was that SSRIs do not cause birth defects, but that third trimester use has been linked to low birth weight and transient but concerning neonatal symptoms such as respiratory distress, jitteriness, and hypoglycemia.
Unfortunately, life has gotten even worse for SSRIs since then.
In June 2004, the FDA asked makers of SSRIs and SNRIs to add a warning about using these antidepressants in the third trimester of pregnancy. Most companies have complied, and you can view the new labeling language by visiting their websites.
The FDA is responding to ever more compelling evidence of what TCR refers to as the “neonatal SSRI syndrome,” hereafter abbreviated as the “NSS.” One particularly large study focused on Paxil (paroxetine). In this study, researchers with the Motherisk program at the University of Toronto recruited 55 pregnant women who were taking Paxil during the third trimester, and telephoned each of them after they delivered, asking them questions about their pregnancy and their infant’s behavior (Arch Pediatr Adolesc Med 2002; 156:1129-1132). These mothers were compared with two matched control groups: 27 women who had taken Paxil during the first and second trimesters only, and 27 pregnant women who had taken no antidepressants at all. The results? Fully 22% of the 55 third trimester-exposed infants experienced complications severe enough to prolong hospitalization (mostly respiratory distress) versus only 6% of infants whose mothers had taken Paxil only in the first two trimesters or not at all. There was also a higher rate of prematurity (20%) in the late exposure group vs. 3.7% of controls.
A perplexing question remains unanswered by this and other similar studies, however. Are the symptoms of the NSS a result of excessive serotonergic stimulation of the neonatal nervous system, or are they caused by SSRI withdrawal? A fascinating study from Finland helps us to answer this question.
Published recently in the Archives of General Psychiatry, the study followed pregnant women who were taking SSRIs (in this case, Prozac/fluoxetine or Celexa/citalopram) and compared them with a non-SSRI control group. But instead of simply interviewing the mothers post-hoc (as was done in the Canadian study above), they developed a serotonergic symptom scale and trained pediatricians to evaluate neonates using this instrument in “real time,” at the moment of delivery. No problems with recall bias here! In addition, blood samples were taken from mothers, from umbilical veins, and from neonates in order to document levels of SSRIs and metabolites over the course of a 2-week post-delivery follow-up. This allowed correlation of symptoms with biochemical parameters. (Arch Gen Psychiatry 2003; 60:720- 726).
The study reported a fourfold higher serotonergic symptom score in the exposed infants than in the controls during the first four days of life. The main symptoms were tremor, restlessness, and rigidity. The good news is that by two weeks of age, the serotonergic symptoms vanished. Furthermore, the decrease in symptoms was associated with a parallel decrease in neonatal Prozac and Celexa serum concentrations, a finding that the authors interpret as being incompatible with a serotonin withdrawal argument. After all, they argue, if the neonatal SSRI syndrome were caused by SSRI withdrawal, you would expect an increase in symptoms as the neonatal concentration went down, and not the rapid improvement that was observed, right?
Well, not necessarily. After all, the time course of adult SSRI withdrawal often mirrors the pattern seen in these neonates, in that the most severe symptoms occur 1-3 days after discontinuation and gradually moderate over the ensuing couple of weeks (J Clin Psychiatry 1997; 58 (Suppl 7):5-10). This means that the time course of symptom disappearance in these infants could, in fact, still be consistent with SSRI withdrawal.
On the other hand, the fact that Prozac was one of the SSRIs that caused NSS symptoms in newborns tends to support the authors’ argument that the symptoms are not a function of withdrawal. Prozac, with its relatively long (2-week) half life, rarely causes withdrawal symptoms in adults, and would figure to be less likely to do so in newborns.
So what are we to conclude from all this?
1. SSRIs do not cause birth defects this piece of wisdom is unaffected by these findings.
2. Infants who have SSRIs in their bloodstream at delivery are more likely than unexposed infants to exhibit symptoms such as tremor, restlessness, rigidity, and respiratory distress, and to be premature.
3. These infants’ symptoms should be completely gone within two weeks. No studies have shown long term sequelae from this neonatal SSRI syndrome, but there is not a lot of controlled data at this point (see TCR May 2003 for a recent review).
4. Whether these symptoms are caused by serotonin stimulation or withdrawal is still unclear, but either way, we speculate that discontinuing SSRIs two weeks before the due date might prevent problems, as there would likely be no medication in the infant’s bloodstream by then.
5. The decision to stop SSRIs is not an easy one, and requires a conversation with your patient about risks vs. benefits. Some argue that maternal depression itself is more harmful than any potential effects of SSRIs on the newborn.
TCR VERDICT: Data accumulates that neonatal SSRI syndrome is real