A Common Clinical Conundrum: Antidepressant Treatment of Depression in Pregnant Women

Abstract

Objective:

Depression during pregnancy is associated with adverse maternal, pregnancy, and infant outcomes. Treatment during pregnancy requires a balanced discussion of the risks of both drug exposure and untreated depression. An updated review of the epidemiology, outcomes, and management of maternal depression is presented.

Findings:

Adverse outcomes are associated with both maternal depression and antidepressants. Research gaps include data on the longitudinal developmental trajectory of offspring exposed to antidepressants compared to depression, with assessment of in utero symptom exposure and environmental exposures. Additionally, neonatal syndrome associated with antidepressant use during pregnancy has no consensus definition or mechanistic explanation.

Conclusion:

With sophisticated large-scale epidemiologic studies, there has been progress in distinguishing the impact of depression processes from medication used for treatment. Optimal treatment of perinatal depression includes close symptom monitoring and medication adjustments to maintain symptom remission. This evolving field requires frequent consultation with reproductive data sources included in this article.

1. Introduction

Depression affects 12–15% of women during pregnancy^1,2. The cumulative prevalence of depression increases to a striking 21.9% the first year after delivery³. Psychiatric disorders and suicide are the second leading cause of death in the postpartum period⁴ and contribute to approximately a quarter of maternal deaths in the first year after birth^5,6. Major Depressive Disorder (MDD) during pregnancy has been associated with multiple risks that contribute to infant morbidity and mortality, including preterm birth, low birth weight, gestational diabetes and hypertensive disorders^7–9. Exposure to depression in utero and postpartum increases the risk of offspring developmental effects, such as insecure attachment, childhood psychopathology and other conditions^10,11.

Representatives from the American Psychiatric Association and the American College of Obstetricians and Gynecologists (ACOG)¹² reviewed the literature on the risks associated with both depression and antidepressant treatment published through 2008. The authors concluded that methodological advances were needed to differentiate the reproductive effects of medications from the impact of the underlying psychiatric illness and its psychosocial sequelae. Progress has been made in addressing this challenging issue. We provide an updated review from 2008 to the present, discuss outcomes associated with both depression and antidepressant use during childbearing and describe an approach to optimize treatment.

2. Diagnosing Depressive Disorders in Perinatal Women

The United States Preventive Services Task Force (USPSTF) recommended that health care professionals screen for depression in all adults, including pregnant and postpartum women^13–15. The Edinburgh Postnatal Depression Scale (EPDS)¹⁶ and Patient Health Questionnaire (PHQ-9)¹⁷ are validated and frequently used brief self-report screening measures. The EPDS is specific to the perinatal period while the PHQ-9 has broader use in screening general medical populations. The recommendation from the Council on Patient Safety in Women’s Health Care was to screen at the initial pregnancy appointment, at least once during pregnancy, and again in the postpartum period^9,18, which is consistent with ACOG guidelines¹⁴. This strategy identifies chronically depressed women who become pregnant, women whose depression begins during pregnancy, and women with onset in the postpartum period. These recommendations for multiple screening times were supported by a hospital-based epidemiologic study of newly delivered women diagnosed with depression at 4–6 weeks postpartum. In this cohort, 27% reported episode onset before pregnancy (chronic disease), 33% experienced the onset during the 9-month pregnancy period, and 40% had onset within the first 6 weeks after birth¹⁹.

3. Maternal Depression and Adverse Reproductive Outcomes

3.1. Miscarriage

The risk of pregnancy loss after gestational week 5 is 21.3%²⁰. Factors associated with depression, such as severe psychosocial stress and resource disadvantage, contribute to the risk for early pregnancy loss^21,22. Pre-existing mental health disorders are significant contributing factors²³. In a retrospective cohort analysis, patients with previous diagnoses of affective disorders (OR=1.6) were at a significantly higher risk for fetal mortality in pregnancy compared to patients without a previous mental health diagnosis.

3.2. Preterm Birth and Growth Effects

Using a national register database²⁴, investigators found that mothers with depression were at significantly higher risk for preterm birth compared to mothers without depression. Similar results were reported in a meta-analytic study²⁵, which defined the relative risk for preterm birth (PTB) in women with MDD as 1.39 [95% CI, 1.19–1.61]. In the same meta-analytic study, the relative risk of having an infant with low birth weight (LBW) for women with antenatal depression was 1.10 (95% CI, 1.01–1.21). MDD was associated with PTB among women of lower socioeconomic status. Although birth outcomes were linked to maternal depression, post-birth growth rates for weight, length, or head circumference during the first year after birth were comparable in infants exposed to depression in utero and infants born to non-depressed women².

3.3. Neonatal Effects

Infants born to mothers with depression were at higher risk for admission to a neonatal intensive care unit (NICU) and had longer hospital stays (7 or more days) compared to infants born to mothers without depression²⁴. In a study comparing the neurobehavior of newborns exposed during pregnancy to MDD, SSRIs or no exposure²⁶, newborns exposed to MDD had lower attention scores than the other exposure groups. Brain imaging revealed an association between antenatal maternal depressive symptom exposure and infant cortical thinning, which is associated with depression in adults²⁷. Additionally, infants born to mothers with depressive symptoms during pregnancy had greater functional connectivity of the amygdala with the left temporal cortex and insula as well as with the left and right anterior cingulate, medial orbitofrontal, and ventromedial prefrontal cortices. These observations are similar to the connectivity patterns observed in MDD in adults and adolescents²⁸. Other associations with maternal mood symptoms during pregnancy include gray matter volume reductions in the prefrontal cortex, the medial temporal lobe, the lateral temporal cortex, and the postcentral gyrus as well as the cerebellum extending to the middle occipital gyrus and the fusiform gyrus²⁹.

3.4. Long-term Effects on Offspring

Fetal exposure to maternal depression was associated with difficult temperament during infancy^30,31 and a two-fold increase in childhood behavioral problems and psychiatric disorders, as well as depression in adolescence and adulthood^32–34. These findings are independent of postnatal maternal mood symptoms. Infants exposed to maternal MDD in utero have elevated concentrations of cortisol (a biomarker for depression^35,36) compared to infants of mothers who were not depressed^30,31. This biochemical change continues through adolescence, which places the offspring at higher risk for developing emotional and behavioral problems³⁷. Studies are limited in relation to perinatal paternal MDD; however, researchers reported that children with depressed fathers had a higher risk of emotional and behavioral problems in young childhood through adolescence, as well as an increased risk of experiencing depressive symptoms³⁸.

The number of studies of fetal exposures and their impact on adult health have grown exponentially from the developmental origins of health and disease (DOHaD) hypothesis. The premise is that the quality of the intrauterine environment shapes the offspring’s risk for disease throughout the lifespan³⁹. Many of the DOHaD studies focus on the lasting effects of LBW, which we described above as an adverse outcome associated with maternal depression. LBW has been linked to several psychiatric disorders later in life, including Attention Deficit Hyperactivity Disorder, negative impact on intelligence and educational achievement, and depression³⁹.

A limitation to studying child and adolescent outcomes as they relate to in utero exposures is the lengthy time for continuing influence by postnatal environmental events. Longitudinal studies beginning proximal to birth with long-term follow-up can address these limitations; however, such studies are costly and difficult to conduct.

4. The Impact of Antidepressants on Birth Outcomes

In a study published in 2011, Mitchell et al reported that 82.3% of American women took at least one medication (prescription or over the counter, not including vitamins)⁴⁰. Nearly half of pregnant women were prescribed a medication during the first trimester. Approximately 8% of pregnant women are exposed to an antidepressant drug⁴⁰ with 2.8% of women continuing treatment throughout pregnancy⁴¹. Women who discontinued antidepressants proximal to pregnancy were more likely to experience a depression recurrence than women who continue antidepressants during pregnancy (68% vs 26%), and symptoms usually emerged in the first trimester⁴². Clinicians and patients must balance the risks of untreated psychiatric illness with the potential adverse fetal and infant effects of exposure to pharmacotherapy. Wisner et al.⁴³ provided a model for the risk-benefit decision-making process.

4.1. Miscarriage

An elevated risk of miscarriage was reported in antidepressant exposed compared to non-exposed pregnant women (hazard ratio (HR) =1.08; 95% CI=1.04–1.13)⁴⁴. However, factors associated with the underlying depression, such as maternal health habits and lifestyle, were confounding variables that increased risk estimates. Miscarriage rates in women with antidepressant use were within the expected range of miscarriages within the general population^45–49.

4.2. Structural malformations

Congenital malformations occur in 3–5% of births^50–52. Studies that assess the risk of congenital malformations related to drug exposure focus on exposures during organogenesis. Consistent with a prominent theme in this literature, adjustment for variables associated with both the underlying disease and malformations has been a challenge. Huybrechts et al.⁵³ utilized a robust study design with data from the Medicaid Database to define the association between antidepressant use and the risk for cardiac malformations. This study included 949,504 pregnant women with 6.8% prescribed an antidepressant during the first trimester. Women exposed to antidepressants were compared to those who were unexposed and the rate of cardiac defects was reported. The antidepressants included paroxetine, sertraline, fluoxetine, tricyclic antidepressants as a group, serotonin norepinephrine reuptake inhibitors (SNRI) as a group, monoamine oxidase inhibitors and bupropion. The initial unadjusted analysis showed an elevated odds ratio (OR 1.25; 95% CI= 1.15–1.36) which suggested an association between cardiac malformations and first trimester exposure to antidepressants. However, after restricting the analysis to only women with a diagnosis of MDD, the OR in treated compared to untreated women was reduced (adjusted odds ratio (aOR) 1.12; 95% CI= 1.01–1.25). Further adjustment with propensity score stratification reduced the OR to non-significance (aOR 1.02; 95% CI=0.09–1.15). Propensity score matching statistically accounts for variables that affect both the exposure (antidepressant use) and the outcome (congenital cardiac malformation). Although women taking an antidepressant are at greater risk for having a child with a congenital cardiac defect, the risk is attributable to factors other than antidepressant medications.

4.3. Growth Effects and Preterm Birth

Investigators who conducted a retrospective study of women treated with antidepressants compared to untreated pregnant women with MDD reported that both MDD and antidepressant use were associated with preterm birth, NICU admission, and severe neonatal morbidity/mortality⁵⁴. The risk was higher for infants exposed to antidepressants, but other factors may have mediated results, such as greater severity of symptoms and illness course, which are associated with antidepressant use. In an early prospective study of women who were treated with SSRIs compared to untreated women with MDD, both groups had higher rates of preterm birth (23% and 21% respectively), than women without depression or exposure to antidepressants (6%)⁵⁵. Investigators in a recent Scandinavian study reported that SSRI use during pregnancy was associated with a lower rate of late preterm (32 to 37 weeks gestational age) (OR=0.84, 95% CI; 0.74–0.96) and very preterm birth (28 to 32 weeks gestational age) (OR=0.52, 95% CI; 0.37–0.74) compared to un-medicated women with psychiatric disorders²⁴. The authors suggested that the protective effect they observed could be due to reduction of depressive symptoms with pharmacotherapy or to an increased risk of PTB in mothers with untreated depression. Remission of the mood and somatic symptoms of depression may result in a maternal physiologic milieu more conductive to carrying a pregnancy to term; however, depressive illness severity data were not available in this epidemiologic study.

The risk of LBW in infants exposed and unexposed to antidepressants in utero is similar⁸. This suggests that LBW may be attributable to sequelae of the underlying depressive disease, rather than to antidepressant drug exposure. The growth of in utero SSRI-exposed infants the first year after birth was similar to that of unexposed offspring².

4.4. Neonatal outcomes

Poor neonatal adaptation, often termed neonatal adaptation syndrome (NAS), has been described in infants exposed to antidepressant medications in utero. This syndrome consists of a group of newborn signs including restlessness, rigidity, tremor, tachypnea, hypoglycemia, temperature instability, and irritability^56–59. NAS has been reported in up to 30% of infants exposed to SSRI during the third trimester^55,56,60 and is more frequently observed in infants exposed to venlafaxine, paroxetine, and fluoxetine⁶¹. The pharmacologic characteristics of individual drugs have rarely been considered in the expression of NAS, and the evolution of signs and the underlying mechanisms differs by drug. NAS may be a manifestation of increased serotonergic stimulation at birth, withdrawal (rapid drug decline after birth) or neurobehavioral teratologic effects on fetal brain function. Rare cases of severe symptoms (seizures, dehydration, excessive weight loss, hyperpyrexia, and intubation) were reported in 0.3% of neonates in newborns reported to the FDA Adverse Events Reporting System⁵⁶. Neither a consensus definition nor standardized measurement tool has been developed for SSRI-associated NAS, which contributes to the wide range in reported incidence and outcomes. The use of benzodiazepines concomitantly with antidepressant medications increases the likelihood and duration of NAS⁶². The use of concomitant medications in the evolution of NAS also remains understudied. Supportive management in the hospital is sufficient for most newborns with NAS⁵⁶.

SSRI antidepressant use in the final trimester also has been associated with persistent pulmonary hypertension of the newborn (PPHN). Researchers utilizing the Medicaid Database with nearly 3.79 million pregnant women compared outcomes of women taking an SSRI, another class of antidepressant, or no antidepressant in the last 90 days of pregnancy⁶³. After adjusting for confounders, restricting the sample to women with depression, and performing propensity score matching, the OR for PPHN was nonsignificant (OR=1.10; 95% CI, 0.94–1.29) as well as for non-SSRI antidepressants (OR=1.02; 95% CI, 0.77–1.35). The adjusted odds ratio for a specific subtype of primary PPHN (not secondary to aspiration or lung hypoplasia), was 1.28 (95% CI, 1.01–1.64) and for non-SSRI antidepressants was 1.14 (95% CI, 0.74–1.74). The authors concluded that the absolute risk of PPHN is small and that any contribution for SSRI medications was more modest than previously suggested.

4.5. Long-term Effects on Offspring

In a study of infants assessed with the Bayley Scale of Infant Development at 12, 26, 52, and 78 weeks of age, SSRI exposed infants were compared to infants with exposure to maternal depression and no exposure⁶⁴. Mental development scores did not differ between groups across the assessment 78-week period. Psychomotor development was less favorable in infants exposed in utero to antidepressants at 26 and 52 weeks of age, although the scores remained within the normal range. At 78 weeks, no significant difference was observed. Several investigators have reported this motor delay in early infancy in antidepressant-exposed offspring^65–67.

Researchers analyzed speech and language in children up to 14 years of age who were exposed in utero to SSRIs, un-medicated depression, or no exposure⁶⁸. Children born to mothers who filled prescriptions for SSRI at least twice during pregnancy were at a significantly increased risk for speech/language disorders compared to children in the other groups (HR, 1.37; 95% CI, 1.11–1.70; P = .004). The results remained statistically significant after controlling for maternal severity of depression.

Full-scale IQ of sibling pairs in children ages 3–7 years old were compared⁶⁹. One sibling was exposed to antidepressant in utero and the other was not exposed. No difference in childhood behavior symptoms was observed; however, the severity of maternal depression during pregnancy significantly predicted both internalizing and externalizing behavior problems in all offspring⁶⁹.

Studies on the association of autism spectrum disorder and exposure to SSRIs in utero yielded heterogeneous results. Large-scale studies using health registry data that allowed for adjustment of confounding factors, including maternal psychiatric disorders, other diseases, and familial risk factors, did not demonstrate an increased risk of autism attributable to in utero exposure to antidepressant medications^70–72. These findings were supported by a recent systematic umbrella review of 45 meta-analyses of the association of antidepressant use with adverse health outcomes⁷³.

5. Considerations for Antidepressant Treatment During Pregnancy

5.1. Discontinuation of Antidepressants prior to or during Pregnancy

Pregnant women who discontinued their antidepressant proximal to conception were nearly three times more likely to suffer recurrence of depressive symptoms during pregnancy than women who continued their medication⁴². In contrast, investigators who conducted a prospective study⁷⁴ found that the risk for recurrence of depression was similar in women who stopped antidepressants compared to those who maintained treatment during pregnancy. The discrepant outcomes can be attributed to the differences in study populations. The first study⁴² focused on women in an academic perinatal psychiatry practice while the second⁷⁴ included women from obstetrical practices and hospital-based prenatal clinics. The women in the first study also had less severe symptoms and onset later in life⁴². Both investigators observed that women with a recent episode of depression or who had four or more prior episodes of depression were at higher risk for recurrence of depression in pregnancy (hazard ratio (HR) 3.6; 95% CI, 1.9–7.0; P<.001).

Women who choose to discontinue their antidepressants during pregnancy require close monitoring for symptom recurrence. A plan defining the woman’s preferences treating a recurrence before symptoms return is good clinical practice. Alternative treatment options for women who discontinue medication are briefly discussed below.

5.2. Dosing During Pregnancy

Some women who maintain antidepressants during the perinatal period have a recurrence of depressive symptoms. One explanation is inadequate dosing due to dramatic changes in pharmacokinetics (PK) across pregnancy, which result in progressively lower plasma drug concentrations of most antidepressants^75–78. Small case series studies show that the concentrations of the SSRI fluoxetine⁷⁹, sertraline⁷⁵, citalopram and escitalopram⁷⁵ decrease throughout pregnancy. Doses may need to be adjusted for pregnant women taking SSRIs⁸⁰. Monitoring symptoms frequently, especially in the second and third trimesters, and dose increases with early symptoms of relapse is a strategy to avoid symptom recurrence.

To optimize antidepressant treatment during pregnancy, establishing the trajectory and magnitude of changes in plasma concentrations would be preferable to relying on reactive dose adjustment with symptom emergence. Genotyping may be a tool to predict cytochrome P450 (CYP) activity and plasma concentration across pregnancy. Ververs et al.⁸⁰ published evidence of this concept by studying the metabolism of paroxetine, which is metabolized solely by CYP2D6. They evaluated the CYP2D6 genotypes of 74 pregnant women and separated them into categories of poor, intermediate, extensive, and ultra-rapid metabolizers. The extensive and ultra-rapid metabolizers had decreasing concentrations of paroxetine throughout pregnancy with an increase in depressive symptoms. Intermediate and poor metabolizers had increasing concentrations throughout pregnancy. Determining a patient’s genotype may provide guidance for prescribing during pregnancy.

6. Other treatment Options during the Perinatal Period

Evidence-based psychotherapies have established efficacy for perinatal women with depression. Interpersonal psychotherapy is effective for depression during pregnancy⁸¹. An adapted, brief interpersonal psychotherapy resulted in decreased depressive symptom scores in low-income women⁸². Behavior activation (BA) is also an effective psychotherapeutic treatment for perinatal depression. Participants receiving BA therapy had significantly lower depressive symptoms and a higher remission rate (56.3% vs. 30.3%, p = .003) compared to those receiving treatment as usual⁸³.

Additional interventions include electroconvulsive therapy (ECT), transcranial magnetic stimulation (TMS), and bright light therapy. ECT is used for severe depression and/or mania that compromise maternal physical function and health⁸⁴. TMS is not associated with significant adverse effects when used to treat depression in pregnancy^85,86 and has been associated with a decrease in depressive symptoms⁸⁷. Protocol-based morning bright light therapy effectively treated pregnant women with depression in a randomized controlled trial^88,89.

7. Future Directions and Conclusion

Most adverse events associated with antidepressant exposure in utero are also associated with exposure to maternal depression. Separating the impact of exposures creates a challenge that is inherent in observational studies when randomization is not feasible. Depression is often a confounding variable in evaluations of adverse events in antidepressant-exposed infants, as women taking psychotropics are frequently not in full symptom remission. Severity of depressive symptoms may mediate adverse events associated with both SSRI exposure and MDD exposure during pregnancy. An investigation to optimize medication management in mothers with depression throughout pregnancy is underway⁹⁰. The goal is to provide the optimal drug doses across the changing milieu of pregnancy to maximally reduce disease burden and maintain remission while minimizing adverse effects.

Research on neonatal and long-term adverse events remains a largely unmet need. Reports of the frequency of NAS among infants with in utero exposure to SSRIs vary dramatically due to the lack of a consensus definition. The characteristics of SSRI-associated neonatal signs differ from withdrawal associated with other drugs, such as opioids. The mechanism(s) behind NAS must be clarified and data suggest that increased serotonergic stimulation, rather than withdrawal, is the likely cause of newborn signs⁹¹. Current NAS research focuses on the first few weeks after birth. Minimal information is available on later developmental outcomes in newborns with NAS.

To advance the treatment of MDD during pregnancy, researchers must frame treatment questions with a focus on optimal outcomes and assess whether the long-term benefits outweigh potential short-term risks. This requires prioritizing an individualized approach to treatment, monitoring symptoms to measure efficacy, and adjusting dose or treatment type when necessary to maximize efficacy. When the decision to treat a pregnant woman with an antidepressant is made, clinicians are compelled to provide the highest quality pharmacologic care. The justification for treating a pregnant woman is that the overall negative impact of the disease is greater than that of the pharmacotherapy, which reduces disease expression and thereby its impact on pregnancy, infant, childhood and family outcomes. Once the decision to continue pharmacotherapy during pregnancy has been made, optimal pharmacological treatment must be provided, with dose adjustments during pregnancy and postpartum and consideration for breastfeeding.