New Data on Birth Defects With Antidepressant Use in Pregnancy

Pauline Anderson

August 18, 2020

Use of venlafaxine (Effexor), a serotonin-norepinephrine reuptake inhibitor (SNRI), is associated with more birth defects than any other antidepressant taken in the first months of pregnancy, new research suggests.

Jennita Reefhuis, PhD

"Our study found that use of the antidepressant venlafaxine during early pregnancy may be linked to several birth defects, including heart defects, defects of the brain and spine, cleft lip and cleft palate, hypospadias, and gastroschisis," senior study author Jennita Reefhuis, PhD, acting director of the CDC's Division of Birth Defects and Infant Disorders in Atlanta, Georgia, told Medscape Medical News.

This finding, though, "needs confirmation given the limited literature on venlafaxine use during pregnancy and risk for birth defects," the researchers write. 

Several selective serotonin reuptake inhibitors (SSRIs), including sertraline (Zoloft), fluoxetine (Prozac), paroxetine (Paxil), and citalopram (Celexa) were each associated with a small number of different birth defects, the study showed.

One SSRI, escitalopram (Lexapro), was associated with the lowest number of birth defects.

In many cases, the risk of birth defects was somewhat reduced after researchers partially accounted for the underlying condition for which the antidepressant was prescribed.

"Fully informed treatment decision-making requires balancing the risks and benefits of proposed interventions against those of untreated depression or anxiety," the researchers note. 

The study was published online August 5 in JAMA Psychiatry.

National Birth Defects Prevention Study

The study provides final data from the US multisite, population-based National Birth Defects Prevention Study (NBDPS). Compared with published interim data, this new information is from a larger sample size and the authors looked at more antidepressants and birth defects.

The NBDPS is among the largest studies worldwide that examines risk factors for birth defects with systematic case verification. This allowed researchers to examine associations with specific defects "more accurately than is possible with administrative data," said the authors.

The analysis included 30,630 cases of birth defects and 11,478 infants born without major birth defects. The mothers, who ranged in age from 12-53 years, gave birth from 1997 to 2011.

Mothers participated in a computer-assisted telephone interview 6 weeks to 24 months after their expected due date. They were asked about the use of antidepressants during the months before conception or during pregnancy.

"This study is unique because researchers asked women what medications they took and for how long," said Reefhuis. "We know that many women stop taking their prescribed medications, particularly antidepressants, once they find out they're pregnant, and you can only find that out by talking to them."

Early Exposure

Early pregnancy exposure was defined as using one or more antidepressants in any dose, duration, or frequency from the month before conception through the third month of pregnancy — the first trimester — when most birth defects develop. Researchers included the month before conception to account for imprecise estimates of conception and medication exposure dates.

Women were considered unexposed if they reported no antidepressant use during the 3 months before conception through the end of their pregnancy.

Researchers also looked at antidepressant exposure outside early pregnancy (during the 2 to 3 months before conception and/or months 4 to 9 of gestation). This was in an effort to account for confounding by indication, which is a major issue for observational studies.

The study therefore "accounted at least partially for confounding by the underlying condition, which, to our knowledge, previous analyses have not commonly done," the authors note.

The researchers examined associations between a range of birth defects and use of individual antidepressants in several classes, including SSRIs (citalopram, escitalopram, fluoxetine, fluvoxamine, paroxetine, sertraline), SNRIs (desvenlafaxine, duloxetine, venlafaxine), and tricyclic and norepinephrine (TCA-NE) reuptake inhibitors (amitriptyline, clomipramine, desipramine, doxepin, imipramine, and nortriptyline).  

They also looked at use of bupropion (Wellbutrin), an atypical antidepressant that acts as a norepinephrine-dopamine reuptake inhibitor and a nicotinic receptor antagonist, and also investigated mirtazapine, nefazodone, and trazodone.

This analytical approach allowed researchers to examine associations for individual antidepressants and specific defects "with more precision" and to explore additional birth defect associations, said the authors. It's unlike that of the many other studies that have assessed the fetal safety of antidepressants, which grouped medications into broad classes that limits clinical usefulness, they noted.

In this new study, early pregnancy antidepressant use was reported by 1562 cases (5.1%) and 467 control mothers (4.1%).

When comparing women exposed to various antidepressants with women who were unexposed, researchers adjusted for race/ethnicity, body mass index, education, smoking, and alcohol use. The authors included covariates in the model if they resulted in an at least 10% change in any point estimate for associations between the exposure and any defects tested.

The analysis showed that different SSRIs were associated with different birth defects. Paroxetine or fluoxetine used in early pregnancy had the highest proportion of elevated adjusted odds ratios (aORs) for specific birth defects.

Elevated associations with fluoxetine were for anomalous pulmonary venous return (APVR) (aOR, 2.56; 95% CI, 1.10 - 5.93]), aortic stenosis, and Dandy-Walker malformation.

Defects with meaningfully elevated associations with paroxetine were APVR and its subset total anomalous pulmonary venous return (TAPVR), pulmonary valve stenosis, anencephaly, and craniorachischisis.

Citalopram and sertraline were the SSRIs with the next highest number of birth defects.

Partial Confounding

When comparing women with early pregnancy exposure to those who took antidepressants only outside early pregnancy, the associations between the SSRIs and birth defects, particularly heart defects, were largely attenuated. This suggests that the elevated associations "may be confounded by the underlying condition," said the authors.

For example, after partially accounting for underlying conditions, the aOR for fluoxetine and APVR was reduced to 1.89 (95% CI, 0.56 - 6.42).

"Although more evidence is needed to be certain, our findings suggest that the increased risk for heart defects among women who reported use of SSRIs might be due to the underlying condition, for example, maternal depression, anxiety, or other mental health concerns, or factors related to the mother's underlying medical conditions and not SSRIs themselves," said Reefhuis.

This is "noteworthy" because most previous research hasn't been able to separate the effect of medication use from that of the underlying health condition, she said.

Non-heart Defects

Associations between SSRIs and noncongenital heart defects (CHD) often persisted after partially accounting for the underlying condition. For example, this was the case for sertraline and diaphragmatic hernia; fluoxetine and esophageal atresia; and citalopram and diaphragmatic hernia.

The finding that a specific SSRI might be the source of increased risk is "novel," said Reefhuis, but she cautioned that "more evidence is needed to be certain of these associations."

There were no elevated aORs between escitalopram and specific birth defects. But Reefhuis noted that this drug was the least commonly used SSRI and researchers were able to assess fewer defects.  

When looking at other antidepressants, the researchers found that use of bupropion had meaningfully elevated aORs for three birth defects: hypoplastic left heart syndrome, intestinal atresia/stenosis, and diaphragmatic hernia.

Venlafaxine had meaningfully elevated associations with heart defects, including left ventricular outflow tract obstruction defects (hypoplastic left heart syndrome and coarctation of the aorta), right ventricular outflow tract obstruction defects (pulmonary valve stenosis), and septal defects (ventricular septal defects and atrial septal defects). Other defects were of the brain and spine, cleft lip and cleft palate, hypospadias, and gastroschisis.

Some of the increased associations with venlafaxine were relatively high (for example, aORs 3.34 [95% CI, 1.69 - 6.60] to 5.26 [95% CI, 1.96 - 14.12]).

"There is not a lot of published information on venlafaxine because it is a relatively new drug, so our findings are new," said Reefhuis.

After partially accounting for underlying conditions, the elevated venlafaxine aORs persisted for most defects.   

The NBDPS interview did not systematically ascertain diagnoses associated with drug treatment, limiting the ability to account directly for mental health conditions in the analyses.

As well, there may be differences in disease severity, relapse risk, or other clinical differences among women who select specific antidepressants, and continue using antidepressants during pregnancy, compared with those who discontinue treatment before pregnancy or initiate treatment after the first trimester. This may have resulted in bias, said the authors.

The study also did not account for the differential half-life of antidepressants, which could have resulted in exposure during periods classified as unexposed. And recall bias could have led to misclassification of exposure.

The researchers stressed that the analysis is "hypothesis generating" and does not determine causality.

Illness or Treatment?

An accompanying editorial by Katherine L. Wisner, MD, Departments of Psychiatry and Obstetrics and Gynecology, Northwestern University School of Medicine, Chicago, Illinois, and colleagues acknowledged the difficulty of distinguishing the adverse effects of psychiatric illnesses from the effects of medications on risks of birth defects.

While the study authors "partially" accounted for the underlying condition, "the reader wonders whether the association of maternal mental illness and symptom severity that guide the decision to use antidepressants during pregnancy has been sufficiently addressed," said the editorial.

In the case of SSRI exposure, the "absence of information on characteristics, such as socioeconomic disadvantage, toxin exposures, and substance use that often accompany poor mental health, renders the results challenging to interpret owing to concern about residual confounding."

Comparing women exposed to antidepressants during the first trimester with those exposed outside this period, the authors adjusted only for maternal education, noted the editorial authors. Information on these mothers' characteristics is "insufficient to demonstrate their comparability" with respect to the psychiatric disorder, comorbid conditions, or concomitant drug exposures (for example, anticonvulsants and/or antimanic agents) that may be associated with increased risk for fetal malformations, they write.

Wisner and colleagues point out that few study subjects took venlafaxine, which is used primarily for SSRI nonresponders. Those who do take this drug are likely to have psychiatric and medical comorbidities, including metabolic syndrome, that affect reproductive outcomes, they said.

The computer-assisted telephone interview during which mothers were asked to recall prenatal psychotropic use raises questions about possible exposure misclassification, the editorial notes. "The recall of women who gave birth to an infant with a congenital malformation is likely to be different than the recall of women who gave birth to a healthy infant."

The editorial writers stressed that many birth defects are extremely rare — with an incidence of 1 in 1000 to 1 in 16,000 births for specific cardiac defects — and that even large increases on the relative scale will correspond to small increases in absolute risk.

As accumulated evidence suggests risks for birth defects associated with antidepressants are "acceptable" compared with those of untreated or undertreated maternal depression, the editorial questions whether it's time "to turn from focusing on identifying risks to discovering potential maternal–fetal benefits of pharmacotherapy."

The research was supported through grants to the Centers for Birth Defects Research and Prevention participating in the National Birth Defects Prevention Study (NBDPS) and/or the Birth Defects Study to Evaluate Pregnancy exposures (BD-STEPS). Reefhuis has disclosed no relevant financial relationships. Wisner is supported by HD085601-Obstetric-Fetal Pharmacology Research Center. Another editorial author, Krista F. Huybrechts, PhD, has been an investigator on grants to Brigham and Women's Hospital from Eli Lilly, the manufacturer of duloxetine.

JAMA Psychiatry. Published online August 5, 2020. Abstract, Editorial

For more Medscape Psychiatry news, join us on Twitter and Facebook


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.