To Prescribe or Not to Prescribe: That is the Question
The last time TCPR covered the topic of psychotropics in pregnancy was less than two years ago, in February of 2009. Since then, however, some new treatment guidelines have been published, and they appear to be useful to clinicians —a nice change from most guidelines that are so all-inclusive that they might as well be called textbooks.
These new guidelines were issued jointly by our APA and by the American College of Obstetricians and Gynecologists (ACOG), and were published in both a psychiatric journal (Yonkers et al, General Hospital Psychiatry 2009;31(5):403–413) and an obstetric journal (Yonkers et al, Obstetrics and Gynecology 2009;114(3): 703–713). In addition to providing a nice summary of the neonatal risks of antidepressant exposure, the paper offers clear, concise, and easy to follow algorithms for the management of depression in three scenarios doctors are likely to encounter. (See the sidebar on page 2 for a summary of these recommendations.)
The guidelines suggest an individualized approach. Psychotherapy might be your first line treatment for depressed women who present to your office either already pregnant or who are planning to become pregnant and want to avoid medications. Cognitive behavioral therapy (CBT) and interpersonal therapy (IPT) have both been shown to be effective in these circumstances (Spinelli M, Endicott J, Am J Psychiatry 2003;160(3):555–562). The guidelines advise you to use antidepressants for women with moderate to severe symptoms (ie, a patient with suicidal or psychotic symptoms), women with a history of severe or recurrent depression, women with significant psychiatric comorbidities (ie, panic disorder, eating disorder, or substance use disorder), or women who did not respond previously to psychotherapy.
Since these guidelines were released, a number of articles have been published expanding the evidence base further. Here’s an update.
The risk of cardiac malformations from SSRIs
In 2005, the FDA required GlaxoSmithKline to change Paxil’s (paroxetine) pregnancy risk category from C (ie, not enough studies to feel confident) to risk category D (ie, there is positive evidence of fetal risk, so try to avoid if at all possible). Most psychiatrists have studiously avoided prescribing Paxil to pregnant women since then. But as Adrienne Einarson discusses in this month’s interview, Paxil’s black sheep status is rather controversial.
Recently, two meta-analyses investigating the effects of exposure to SSRIs in early pregnancy (with a focus on paroxetine) have been published (Wurst et al, Birth Defects Research 2010; 88:159–170, and Gentile S et al, J Clin Psychiatry 2009;70(3):414–422). In the Wurst paper, which included a review of twenty reports, first trimester paroxetine use was associated with an increased prevalence of both cardiac defects specifically (prevalence odds ratio, 1.46; confidence interval, 1.17 to 1.82), and all congenital defects (prevalence odds ratio, 1.24; confidence interval 1.08 to 1.43). The Gentile meta-analysis, on the other hand, appeared to give paroxetine a clean bill of health. While both papers included some of the same studies, the Wurst meta-analysis included only published studies, whereas the Gentile meta-analysis included data from unpublished and not yet published studies to minimize publication bias. This may explain the two papers’ different conclusions.
In a retrospective cohort study of more than 490,000 children born in Denmark, researchers mined a huge teratogenic database called Eurocat (European Surveillance of Congenital Anomalies). (For an explanation of some of this methodology jargon, see this month’s interview.) After analyzing prescription records and reported birth defects, they found an apparent increased prevalence of cardiac septal defects in children whose mothers were prescribed an SSRI (particularly sertraline (Zoloft) and citalopram (Celexa)) in early pregnancy. Children born to women who had taken an SSRI in pregnancy had a rate of cardiac defects of 0.9%—nearly twice that of children born to women who had not taken an SSRI in pregnancy, 0.5%. Interestingly, the highest prevalence of cardiac defects (2.1%) was found for children of women who had filled prescriptions for more than one type of SSRI (Pedersen LH et al, BMJ 2009;339:b3569).
However, these results have to be taken with a grain of salt because of the inherent methodological shortcomings of retrospective studies. For example, the study relied on mothers’ memories of which drugs they took, and people may be more likely to remember taking an SSRI if their baby was born with a heart defect. This so-called “recall bias” could artificially inflate the dangers of antidepressants.
The risks beyond congenital malformations: other adverse outcomes of antidepressant use
Antidepressant exposure in pregnancy may lead to adverse outcomes other than birth defects, such as pre-term birth, small for gestational age (SGA) birth, or birth complications. A recent well-designed controlled prospective investigation gives us some insight. A total of 238 pregnant women were enrolled in the study, and they fell into five categories: 1) Continuous SSRI during most or all of the pregnancy (N=48); 2) Partial SSRI at some point during pregnancy, but at least one full trimester without exposure (N=23); 3) No SSRI, but continuous depression (N=14); 4) No SSRI, partial depression (N=22) (these were women who were depressed during some portion of their pregnancy, but who took no SSRIs); 5) No SSRI, no depression (N=131). All women were followed with maternal assessments at 20, 30, and 36 weeks of gestation, and after birth, with neonatal outcomes assessed by a blinded review of delivery records.
The results were mostly reassuring. Neither SSRIs nor depression increased the risk for minor physical anomalies or reduced maternal or infant weight. The major finding of the study, however, was that infants exposed to either SSRIs or depression continuously during pregnancy had preterm birth rates exceeding 20%, compared to 6% for nonexposed, nondepressed women (Wisner et al, Am J Psychiatry 2009;166(5):557–566).
Contrary to these results, however, Toh and colleagues found that SSRIs taken late in pregnancy were associated with a significant risk of small for gestational age (SGA) babies, and women receiving non-SSRI antidepressants were more likely to deliver both premature and SGA babies (Toh et al, J Clin Psychopharmacol 2009;29(6):555–560). This was a retrospective cohort study and is therefore prone to the same methodological deficiencies previously noted.
As you can see, the research findings throughout the field of “perinatal psychiatry” are frustratingly contradictory. It is clear, however, that antidepressants are unlikely to cause significant birth defects, which is good news. But whether they cause other problems such as preterm births or miscarriages is a matter of continuing debate. The problem is that untreated depression alone can cause a variety of poor pregnancy outcomes, and it has been difficult to control for this effect in studies of antidepressant exposure.
Beyond therapy: Non-prescription options for the treatment of depression in pregnancy
Potential alternatives to medications or therapy for depression in pregnancy include acupuncture, exercise, phototherapy, natural remedies, and transcranial magnetic stimulation. (For a review of these, see Tjoa C et al, Women’s Health 2010;6(4):565–576). As an example, researchers at Stanford University enrolled 150 depressed pregnant women into a controlled trial of acupuncture. Patients were randomly assigned to receive either acupuncture specific for depression, sham acupuncture, or massage. Women receiving the specific acupuncture regimen had a higher antidepressant response rate (63%) than the combined control groups (44%) (Manber et al, Obstet Gynecol 2010;115(3):511–20).
TCPR VERDICT: Antidepressants, in general, probably do not cause birth defects, though there is still debate regarding paroxetine’s effects. Antidepressants may cause problems with premature birth, low birth weight, and miscarriage, but the data are too contradictory and incomplete to know how significant these risks are.