Safety of selective serotonin reuptake inhibitors in pregnancy

Selective serotonin reuptake inhibitors (SSRIs) are recommended by the National Institute for Clinical Excellence as first-line drugs for the treatment of depression, but there is emerging evidence that they might not be entirely safe in pregnancy. We reviewed the literature in this area.

Some data indicate an association between first-trimester SSRI exposure, particularly with paroxetine, and an increased risk of some major congenital malformations. Stronger evidence supports an association with small reductions in gestational age and neonatal withdrawal symptoms.


Risks and benefits of using SSRIs during pregnancy should be discussed with the patient, and a joint decision made between prescriber and patient regarding treatment. Limited data suggest that other SSRIs are safer than paroxetine in pregnancy.

Risks and benefits of using SSRIs during pregnancy should be discussed with the patient, and a joint decision made between prescriber and patient regarding treatment. Limited data suggest that other SSRIs are safer than paroxetine in pregnancy.

Depression is common in women of child-bearing age with an annual prevalence of approximately 10% (World Health Organization, 2001). Women who develop depression during pregnancy may be less likely to care for themselves, placing their own health and that of their unborn child at risk. For those who have an established diagnosis of depression and who are taking antidepressant drugs when they conceive, the risk of relapse on stopping their antidepressant is high. Cohen et al (2006) found that 44 of 65 women (68%) who discontinued their antidepressant relapsed compared with 21 of 82 women (26%) who continued their treatment. Women with a history of depression, particularly post-partum, are at a high risk of post-partum relapse. For those with bipolar disorder this may be up to 50% (Cohen et al, 1995). These figures suggest that most psychiatrists who work with adults of working age are likely to be faced with making decisions about prescribing antidepressants for women who are pregnant or who wish to conceive.

The National Institute for Clinical Excellence (NICE) recommends selective serotonin reuptake inhibitors (SSRIs) as first-line treatment for moderate to severe depression (National Institute for Clinical Excellence, 2004). There have been a number of recent publications reporting on the safety of SSRIs in pregnancy and these studies are reviewed here.

Assessing the effects of drugs on pregnancy outcome
It is obviously unethical to conduct randomised controlled trials in pregnant women as this would involve the intentional exposure of the foetus to a potential teratogen. Other methods commonly used to detect congenital malformations associated with drugs include pre-licensing studies in laboratory animals, and post-licensing observational studies in humans. Extrapolation of animal data to humans is complex and studies of this type will not be considered further here. Human data can originate from prospective or retrospective studies, all of which have limitations. Examples of prospective studies include cohort studies such as those conducted through teratology information services; women who contact these services may not be representative of pregnant women as a whole. Examples of retrospective studies are cohort and case-control studies; these types of study usually utilise administrative data from healthcare databases that have been set up for an entirely different purpose. Assumptions are made, such as every woman who is prescribed an antidepressant drug will actually take it. People cannot be directly questioned or physically examined.

The rate of congenital malformations in control groups identified from healthcare databases can vary widely depending on the demographic and clinical characteristics of the women selected as controls and the size of the group. It is important that the control group are matched for factors other than drug treatment that may affect the outcome that is being investigated. For example, if low birth weight is the outcome being measured, women should be matched for smoking status.

Risks associated with SSRIs

Prospective cohort studies
Einerson & Einerson (2005) conducted a meta-analysis of studies from teratology information services. Women who contacted these services to ask for advice about the use of antidepressants in pregnancy were followed-up prospectively and pregnancy outcomes were compared with women who contacted the same information services requesting advice about drugs known not to be teratogenic. The overall malformation rate in the 830 women exposed to SSRIs and related antidepressants in the first trimester of pregnancy was 2.7% compared with 2.6% in women who were taking drugs that are known not to be teratogenic; no increase was seen in the malformation rate overall or in any specific malformation. This analysis was powered to detect specific malformations that occurred at a frequency of up to 1 in 240 exposures (i.e. were relatively common). The major limitations of this analysis were that:

  * only live births were included; no information is available on the number of spontaneous abortions, elective abortions or still births
  * women who contact teratology information services are likely to be more educated and from a higher social class than most women who are treated with drugs during pregnancy.

Since this meta-analysis was published, there has been a further prospective study from a teratology information service. Diav-Citrin et al (2005) followed up 330 paroxetine- and 230 fluoxetine-exposed pregnancies and compared outcomes with 1141 women who sought advice about exposure to drugs that are known not to be teratogenic. There was a higher rate of major abnormalities overall (RR=1.92, 95% CI 1.01-3.65) and of cardiovascular abnormalities in particular (RR=3.46, 95% CI 1.06-11.24) in infants exposed to paroxetine in the first trimester compared with controls. These calculations are based on 13 paroxetine-exposed infants with major congenital malformations and 5 with cardiac malformations, so should be treated with caution. Diav-Citrin et al (2005) also report that perinatal complications were more common in the SSRI-exposed group than the control group. This study is available in abstract form only and no further details are available.

In a further cohort study, Levinson-Castiel et al (2006) reported that almost a third of neonates exposed to SSRIs in the third trimester (n=60) experienced symptoms that were attributed to SSRI withdrawal, such as sleep disturbance, tremor, hypertonicity, tachypnoea and gastrointestinal disturbance. None of the 60 unexposed neonates, matched for birth weight, gestation, Apgar score and mode of delivery, experienced these symptoms. Zeskind & Stephens (2005) examined 17 infants exposed to SSRIs in utero and 17 non-exposed infants matched for maternal age, social class and smoking. Those infants who had been exposed to SSRIs showed disruption in a range of neurobehavioural outcomes such as tremulousness and sleep. The authors comment that it is unknown if these changes are enduring, and if they are, what the long-term consequences might be.

Allison Donnelly, Clinical Pharmacist
Oxleas NHS Foundation Trust

Carol Paton, Chief Pharmacist
Oxleas NHS Foundation Trust, Pinewood House, Pinewood Place, Dartford, Kent DA2 7WG, email: .(JavaScript must be enabled to view this email address)

Declaration of interest
Psychiatric Bulletin (2007) 31: 183-186. doi: 10.1192/pb.bp.106.012898

Full article

Provided by ArmMed Media