乌鸦传媒

While use of paroxetine (Paxil) and fluoxetine (Prozac) increased the risk of several birth defects, the majority of SSRIs were not associated with significant harm to the fetus during the first trimester, according to the results of a Bayesian analysis.

, of the Centers for Disease Control, and colleagues, found that women taking paroxetine during their first trimester of pregnancy more than doubled their risk of five out of seven serious birth defects, while those taking fluoxetine (Prozac) had a higher risk of two serious birth defects.

However, the authors found no link between citalopram (Celexa), escitalopram (Lexapro), or sertraline (Zoloft) and birth defects in the first trimester, they wrote in .

For the Bayesian analysis, Reefhuis and colleagues combined data from seven previous case-control and cohort studies on birth defects and SSRIs in early pregnancy with data from the CDC's own case-control study, .

The authors note that paroxetine had the most previously reported associations, and this analysis also found similar results. There were five statistically significant birth defects associated with paroxetine:

• Omphalocele: Posterior odds ratio 3.5, 95% credible interval 1.3-8.0
• Anencephaly: Posterior OR 3.2, CI 1.6-6.2
• Gastroschisis: Posterior OR 2.5, CI 1.2-4.8
• Right ventricular outflow tract obstruction (RVOTO) cardiac defects: Posterior OR 2.4, CI 1.4-3.9
• Atrial septal defects: Posterior OR 1.8, CI 1.6-6.2

In this analysis, fluoxetine also had statistically significant association with RVOTO cardiac defects (posterior OR 2.0, 95% CI 1.4-3.1), as well as craniosynostosis (posterior OR 1.9, CI 1.1-3.0).

But that was not the case for the most commonly reported SSRI, sertraline. While prior analyses had found five associations between birth defects and maternal use of the drug, none of these associations were confirmed in the Bayesian analysis. Citalopram had what was described as a "marginal" association with neural tube defects (posterior OR 1.8, 95% CI 1.0-3.0), but no other significant associations were found for citalopram or escitalopram. Sensitivity analyses showed no association between citalopram and neural tube defects, but otherwise yielded similar results.

While not affiliated with the study, , division director for perinatology at Oregon Health and Science University, mentioned some caveats about how the results should be interpreted medically. He pointed out that atrial septal defects will resolve themselves without any treatment and while anencephaly is very serious, it was only found in a small number of cases in the study.

"It's the kind of finding that just by coincidence, if there happened to be one or two more cases than expected in the SSRI exposure group, that can lead to a statistical increase because it's a rare kind of anomaly," he said.

Reefhuis told 乌鸦传媒 via e-mail that due to the conflicting results about birth defects and SSRIs, she wanted to gather all of the available data on the subject. She said the study should be helpful to healthcare providers because it combines knowledge from the literature with one of the largest data sets available to look at these issues, and can help to inform future treatment decisions.

"If healthcare providers are treating women who are already pregnant, it is important to discuss the risk and benefits of antidepressants as well as the safest options available to treat their mental health condition during pregnancy," said Reefhuis.

But Pereira said that neural tube or cardiac defects are determined 4-6 weeks into a woman's pregnancy, often before she knows she is pregnant, so stopping SSRIs in the hopes of preventing birth defects would not impact the fetus. He saw the most relevant use for the study in patients who are new users of SSRIs.

"If they're on paroxetine or fluoxetine already, then I'm not sure that this information would lead me to necessarily change them in the first trimester of pregnancy, but for patients that are going to be new starters, I think it's a reasonable thing to consider," he said.

The NBDPS examined data from 1997 to 2009 and was comprised of 17,293 unexposed cases (with birth defects), 659 cases exposed to any of the five SSRIs, with 9,559 controls (without birth defects), with 298 controls exposed to SSRIs. Patients were excluded if they had diabetes, used teratogenic drugs, had a psychiatric disorder without antidepressant use, or had birth defects unrelated to SSRIs. Sertraline was the most commonly reported SSRI (40%) and tended to favor older mothers. There were no other differences in age distribution.

Limitations to the study include that the study does not address whether the birth defect was caused by the SSRI treatment, maternal disease, or some other factor. The authors also note this was self-reported maternal data based on interviews, and that some numbers of birth defect cases are so small as to potentially result in unstable estimates.

Pereira said the most important takeaway of the study for providers is not to stop SSRI treatment when a patient finds out she is pregnant because the research is far from conclusive.

"None of this demonstrates that there's a cause and effect association and we know that stopping SSRI treatment in patients who are being treated for depression can precipitate major depressive episodes, which can be dangerous for pregnancy and for pregnant women," he said.

Comment