Viswanathan M, Middelton JC, Stuebe A, et al. Maternal, fetal, and child outcomes of mental health treatments in women: A systematic review of perinatal pharmacological interventions [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2021 Apr. Report No.: 21-EHC001. AHRQ Comparative Effectiveness Reviews. PMID: 33950611. doi: 10.23970/AHRQEPCCER236.
National Collaborating Centre for Mental Health (UK). Antenatal and Postnatal Mental Health: Clinical Management and Service Guidance: Updated edition; 2020. NICE guideline (CG192). https://www.nice.org.uk/guidance/cg192
McAllister-Williams RH, Baldwin DS, Cantwell R, et al; endorsed by the British Association for Psychopharmacology. British Association for Psychopharmacology consensus guidance on the use of psychotropic medication preconception, in pregnancy and postpartum 2017. J Psychopharmacol. 2017 May;31(5):519-552. doi: 10.1177/0269881117699361. Epub 2017 Apr 25. PMID: 28440103. https://pubmed.ncbi.nlm.nih.gov/28440103
Perinatal depressive and anxiety disorders are of great clinical and public-health concern. When untreated, these problems are associated with poor maternal health outcomes, obstetric complications, relationship problems with partners, impaired maternal-infant attachment, less breastfeeding, and a range of developmental problems in the offspring (see Perinatal Depressive and Anxiety Disorders for details on these negative outcomes).
Prior to initiating treatment, clinicians must discuss with pregnant persons, mothers, and birthing parents the potential benefits and risks of treatments for perinatal depression and anxiety, along with the risks associated with untreated illness. While psychologic therapies are thought to pose minimal risks to these individuals and their offspring, there is a growing body of research describing the relationship between perinatal use of psychotropic medications and electroconvulsive therapy (ECT) and offspring outcomes.
This chapter describes the existing evidence on the use of psychotropic medications and ECT during pregnancy and lactation.
Risks of treatment during PregnancyTop
Before reviewing extant research examining the effects of psychotropic medications, such as antidepressants, on maternal and neonatal outcomes, it is important to note that there are no randomized controlled trials (RCTs) comparing the use of psychotropic medications with psychologic therapies or placebo interventions among pregnant individuals with depressive and/or anxiety disorders. Therefore, it is not possible to definitively conclude whether gestational use of psychotropic medications (eg, selective serotonin reuptake inhibitors [SSRIs]) is causally related to the risks described below or whether these associations are confounded by indication (ie, due to active depression/anxiety). Moreover, studies examining the risks associated with antenatal use of psychotropic medications are limited by the fact that many other variables that confound associations between medication exposure and offspring adverse effects (eg, substance misuse, poor prenatal care, maternal health problems) are also not adjusted for.
Furthermore, as many of the observational studies examining these links fail to recruit a comparator group of pregnant persons with depression/anxiety, it is difficult to disentangle whether the poor obstetric or neonatal outcomes seen in these observational studies are related to the use of psychotropic medications or are due to active mental health problems. As a result, in the absence of RCTs, the magnitude and nature of these risks are not completely understood.
In discussion with patients about the risks and benefits of the use of psychotropic medications during pregnancy, it is critical for clinicians to be aware of the fact that a ≥2-fold increase in relative risk (ie, relative risk ≥2) is traditionally accepted as clinically meaningful or significant in perinatal psychiatric practice, but it is the absolute increase that matters most. Therefore, clinicians must be aware that although some psychotropic medications may be associated with a statistically significant increase in the risk of poor outcomes, they may not be clinically significant at the individual level.
In addition to describing the multiple adverse outcomes that are commonly associated with untreated antenatal depression/anxiety (see Perinatal Depressive and Anxiety Disorders), clinicians must also consider these methodologic limitations when engaging in a conversation with pregnant mothers about the potential risks of using psychotropic medications during pregnancy.
1. Risks of using antidepressants during pregnancy: Gestational use of an SSRI during the first trimester has been linked to a small increase in the risk of any type of congenital malformation in the neonate (risk ~1.1), including cardiovascular malformations (risk ~1.2-1.3). However, risks can differ between individual SSRIs. For example, paroxetine exposure during the first trimester may pose a small risk of developing any type of congenital malformation in the neonate (risk ~1.2), but the risk of any type of cardiovascular malformation appears to be stronger (risk ~1.3-1.6). In particular, first-trimester use of paroxetine may clinically elevate the risk of atrial septal defects (risk ~2.4) and obstruct the right ventricular outflow tract (risk ~2.2-2.3). First-trimester exposure to fluoxetine has also been linked to a small increase in the risk of any type of congenital (risk ~1.2) or cardiovascular (risk ~1.3-1.4) malformation in the fetus. First-trimester use of sertraline may be associated with a small risk of inducing any type of cardiovascular malformation (risk ~1.3-1.4), as is citalopram (any congenital [risk ~1.2] or cardiovascular [risk ~1.2] malformation). Despite these increased risks, it is critical to note that they may not be clinically significant at the level of the individual, and a number of these complications resolve spontaneously and do not pose significant functional impairment to the offspring. Some have suggested that increased vigilance (eg, more frequent and detailed ultrasound imaging) account for these increased risks, while others, that it may be active depression/anxiety (or the increased severity that leads to the need for antidepressants) that may account for these risks.
Research suggests that first-trimester use of any serotonin-norepinephrine reuptake inhibitor (SNRI; particularly venlafaxine; risk ~1.7) or bupropion (ie, a type of norepinephrine-dopamine reuptake inhibitor; risk ~1.2) may be associated with an increase in the risk of congenital heart malformations in the fetus. Insufficient evidence exists as to whether mirtazapine or tricyclic antidepressants are associated with congenital malformations, with the exception of clomipramine. Indeed, population-level data have suggested that clomipramine exposure during early gestation may clinically raise the risk of developing any type of cardiovascular malformation in the neonate (risk ~1.9-2.0), including atrial and ventricular septal defects (risk ~2.2).
In terms of obstetric outcomes, there may be a modest link between gestational antidepressant use and spontaneous abortion (risk ~1.5), but this may not be present for all agents (eg, it may be higher for paroxetine and venlafaxine). Also, antidepressant use during pregnancy may be associated with an increased risk of small-for-gestational-age (SGA) infants (risk ~1.5), low birth weight (risk ~1.4-1.7), preterm birth (risk ~1.2-1.8), low Apgar scores (risk ~1.7-2.2), and admission to specialized neonatal care (risk ~1.7). Despite these risks, the magnitude of these findings on an individual level is generally small among mothers exposed to antidepressants during pregnancy. For example, 2 meta-analyses have suggested that infants born to individuals receiving antidepressants during pregnancy have only 4 days shorter gestational age, weigh 70 g to 75 g less at birth, and score 0.4 to 0.5 points lower on the Apgar scale compared with infants born to mothers not receiving antidepressants during pregnancy. When examining these associations only among pregnant mothers with depression, one meta-analysis demonstrated that antidepressant use during pregnancy was not linked to preterm birth, low birth weight, SGA infants, or low Apgar scores. However, that meta-analysis did show that SSRI use during gestation may independently increase the risk of preterm birth (risk ~1.6) regardless of the presence of antenatal depression.
At delivery, fetuses exposed to SSRIs in the third trimester are at an increased risk of poor neonatal adaptation syndrome, which is marked by symptoms such as convulsions, jitteriness, tremors, irritability, hypoglycemia, feeding problems, temperature dysregulation, and excessive crying. Occurring in 10% to 15% (but in up to 30% of infants, depending on the agent), these symptoms are time limited (typically resolving in 2-14 days) and are not associated with an increased risk of mortality or long-term neurodevelopmental problems (and resolve with supportive care). This risk may be highest with paroxetine, venlafaxine, and fluoxetine.
Antidepressant exposure (SSRI or SNRIs) throughout pregnancy may also be associated with an increase in the risk of persistent pulmonary hypertension of the newborn (PPHN) (risk ~1.5-1.8). When administered during late gestation, the risk of PPHN may be higher. Despite these findings, the absolute risk of PPHN among infants exposed to antidepressants in utero ranges between 0.6 and 3.0 per 1000 live births, which is comparable with the risk found in the general population: 2/1000 live births. Furthermore, the number needed to harm has been reported to be between 1000 and 1615, indicating that 1000 to 1615 additional pregnant individuals need to be treated with an antidepressant to produce one additional case of PPHN in an infant. Among the individual SSRIs, gestational use of sertraline appears to have the lowest risk of inducing PPHN followed by escitalopram, paroxetine, citalopram, and fluoxetine.
Data exploring the long-term effects of antidepressant exposure during gestation and cognitive, emotional, and behavioral problems among the offspring suggest that there is minimal to no risk.
Despite multiple meta-analyses of observational studies suggesting a link between antenatal antidepressant use and autism spectrum disorder (ASD) in the offspring, clinicians must be aware of methodologic limitations of the studies that are included in these meta-analyses. Many of these studies did not adequately or even partially control for variables that could confound this association, such as maternal age, maternal ethnicity, multiparity, preconception/preexisting depression, and concurrent gestational diabetes. It has been shown that controlling for such confounders strongly suggests that there is no link between gestational antidepressant use and ASD risk in the offspring. In addition, none of the studies found in these meta-analyses used a comparator group composed of cases with concurrent gestational depression. This is critical in order to disentangle whether the risk of ASD in the offspring is related to fetal antidepressant exposure or active maternal depression during pregnancy. After all, maternal depressive disorders have been independently linked to increasing the risk of ASD in the offspring. Furthermore, only studies with a control group derived from general population samples found a significant association between antenatal antidepressant use and ASD risk in the offspring. When the control group was pregnant mothers with a past history of a mental illness, this association no longer remained significant. At present, this body of evidence does not suggest that gestational antidepressant use is causally related to ASD in offspring.
Lastly, the same methodologic limitations described above for gestational antidepressant use and ASD risk in the offspring apply to work examining the link between fetal antidepressant exposure and the risk of attention deficit hyperactivity disorder (ADHD) in the offspring. One meta-analysis of observational studies demonstrated that this link was only significant when the comparator group was derived from the general population. When the comparator group was composed of pregnant mothers with a past history of affective disorders, this relationship no longer remained significant.
2. Risks of using benzodiazepines and benzodiazepine receptor agonists during pregnancy: In terms of teratogenicity, current data suggest that the use of benzodiazepines or benzodiazepine receptor agonists (BZRAs) during pregnancy may not increase the risk of any type of congenital or cardiovascular malformation in the fetus. However, the combination of using benzodiazepines with antidepressants during gestation may modestly increase the risk of congenital malformations in the neonate (risk ~1.4).
As for adverse obstetrical outcomes, antenatal benzodiazepine use may be associated with an increased risk of SGA (risk ~1.4), spontaneous abortion (risk ~1.4), preterm birth (risk ~2.0), low birth weight (risk ~2.2), low Apgar scores (risk ~2.2), induced abortion (risk ~2.0), and neonatal intensive care unit (NICU) admission (risk ~2.5). Furthermore, benzodiazepine exposure during pregnancy may also be associated with an increased risk of caesarean delivery and respiratory support for infants. The use of benzodiazepine receptor agonists and z-drugs (BZRAs; eg, zopiclone, zolpidem) in pregnancy is also linked to poor delivery outcomes, including preterm birth (risk ~1.5), low birth weight (risk ~1.5), and SGA (risk ~1.3).
A major complication of using benzodiazepines during the third trimester and close to delivery is that it can cause withdrawal symptoms (eg, irritability, restlessness, and tremors) and signs of intoxication (eg, lethargy, respiratory distress, and hypothermia) in the neonate. It is likely that these negative effects contribute to the aforementioned risks of low Apgar scores, NICU admission, and need for respiratory support. Moreover, gestational use of benzodiazepines has also been associated with floppy infant syndrome (ie, signs of hypotonia). Finally, some studies have suggested that fetal exposure to benzodiazepines may be associated with delays in the development of the infant during the postpartum period. However, these results need to be taken with caution, as many of these studies did not control for pertinent confounding variables (eg, active maternal mental illness) in their study design or statistical analyses.
3. Risks of using antipsychotics during pregnancy: Clinicians should be aware that second-generation antipsychotics can induce weight gain and modestly increase the risk of metabolic complications in pregnancy, such as gestational diabetes (risk ~1.2-1.3). In particular, a couple of population-based studies suggest that gestational use of either quetiapine or olanzapine appears to carry the greatest risk of gestational diabetes among all second-generation antipsychotics. In addition, the use of second-generation antipsychotics may also cause extrapyramidal symptoms and sedation in the pregnant mother as well.
Meta-analyses of observational studies examining the link between overall gestational use of second-generation antipsychotics with major congenital malformations have demonstrated mixed results, with most of these meta-analyses suggesting that there may not be an association. Indeed, quetiapine and aripiprazole so far have not appeared to carry a risk of congenital malformations in the neonate. However, recent population-level data have suggested that in-utero exposure to olanzapine or risperidone may increase the risk of congenital malformations in the fetus even after controlling for multiple confounders. Despite these findings for olanzapine and risperidone, clinicians need to be aware that these results have yet to be replicated by other population-based data sets and require replication before they guide practice.
While fetal exposure to second-generation antipsychotics may modestly increase the risk of preterm birth (risk ~1.6-1.9), the difference in gestational age between exposed infants and unexposed infants has been reported to be statistically insignificant. However, one population-based cohort study has suggested that olanzapine and quetiapine may be linked to a modest increase in the risk of the neonate being born large for gestational age, even after adjusting for potential confounding variables.
Clozapine can be used during pregnancy, but significant caution should be exercised given that it can cross the placenta and possibly cause agranulocytosis in the fetus. Furthermore, gestational use of clozapine has also been associated with floppy infant syndrome (ie, signs of hypotonia) at delivery, seizures, and poor adjustment in the infant. Lastly, clozapine increases the risk of gestational diabetes in the mother as well as in the infant being born large for their gestational age.
Like the studies examining risks associated with psychotropic medications during gestation, those investigating adverse events from ECT use during pregnancy are small and methodologically limited. Indeed, most of these consist of case-study designs with no control participants, small sample sizes, different ECT procedures, and varying periods of follow-up.
In addition to the general adverse effects that are commonly found in ECT treatment (eg, headaches, confusion, nausea), one systematic review of case studies suggests that ECT use during the second or third trimester may lead to premature uterine contractions and preterm labor. Additionally, pregnant mothers undergoing ECT may frequently report vaginal bleeding as well as abdominal and pelvic pain. ECT use during pregnancy may also be linked with bradycardia and other cardiac arrhythmias in the fetus. Moreover, multiple reviews examining case studies have noted that 3% to 8% of patients undergoing ECT during pregnancy report abortions, stillbirths, or neonatal deaths.
Risks of treatment during BreastfeedingTop
Breastfeeding is an individual decision made by mothers and should be supported by physicians, as the use of most psychotropic medications while nursing is not an absolute contraindication to breastfeeding. However, monitoring by mothers and clinicians is required when breastfeeding is undertaken while taking psychotropic medications. Concerns about breastfeeding while taking medications often regard short-term adverse reactions and longer-term neurodevelopmental effects.
Studies examining the effect of using psychotropic medications (eg, antidepressants) on infant outcomes among breastfeeding mothers are greatly limited by the fact that most of them are based on case reports, which often lack the statistical power to detect associations. Furthermore, many of these studies do not include a comparator group, examine a wide range of medication dosages, use different follow-up periods, and do not control for confounding variables. Therefore, it is difficult to conclude whether the use of psychotropic medications such as SSRIs are causally related to any of the poor outcomes described below.
Along with describing the multiple poor maternal and offspring outcomes that can develop during untreated postnatal depression/anxiety (see Perinatal Depressive and Anxiety Disorders), clinicians must also consider these methodologic shortcomings in any conversation surrounding the potential risks of using any psychotropic medication while breastfeeding the infant.
1. Risks of using antidepressants during breastfeeding: Exposure to antidepressants in breastfed infants is 5- to 10-fold lower than exposure in utero. Serum levels in infants born preterm or those with liver and/or kidney impairment can be higher and consultation with a pediatrician can help guide decisions in these infants. It is generally believed that relative infant doses (RIDs) of medications <10% are generally safe (RID estimates infant drug exposure through breast milk using a known milk concentration and comparing it with a weight-adjusted therapeutic dose); all of the SSRIs and SNRIs tested to date appear to produce exposures below this threshold. It is also important to note that brexanolone demonstrates a low RID of 1% to 2%. Of the SSRIs, sertraline, fluvoxamine, and paroxetine have the lowest RID and milk-to-plasma (M/P) ratios. Just 1 to 2 minor infant reactions have been noted in case studies of >200 infants breastfed by mothers treated with sertraline and paroxetine. Citalopram had higher rates of infant reactions (4%-5%), but these are reversible and generally limited to short-term increases in irritability, restlessness, somnolence, or insomnia. Given its relatively low RID, nortriptyline can be a useful choice if mothers prefer or require treatment with a tricyclic antidepressant. Doxepin should be avoided during breastfeeding, as adverse events have been reported in 2 infants, such as pallor, impaired breathing, vomiting, and weight loss. Postnatal use of bupropion is also possible although 2 reports of seizures in breastfed infants exist. Fluoxetine use is lower in the postpartum period given the lengthy half-life of its active metabolite and its exposure has been linked to increases in infant restlessness, somnolence, and colic.
Unfortunately, next to no data exist on the use of monoamine oxidase inhibitors (MAOIs) during lactation. There is also a paucity of data on the long-term neurodevelopmental outcomes associated with breastfeeding on antidepressants. However, the limited existing evidence fails to support an increased risk of adverse long-term neurodevelopmental effects. This should be considered given the known risks associated with untreated maternal depression/anxiety during the postpartum period (see Perinatal Depressive and Anxiety Disorders).
2. Risks of using benzodiazepines during breastfeeding: Low levels of benzodiazepines can be found in the breast milk of mothers who are taking these medications during the postpartum period. In one study adverse events were reported in <2% of exposed infants. The most common adverse event seen among exposed infants is sedation, followed by lethargy, poor weight gain, apnea, and irritability.
3. Risks of using antipsychotics during breastfeeding: Data on most second-generation antipsychotics indicate a RID <10% and an M/P ratio <1, suggesting that they are generally safe for short-term use while breastfeeding infants. However, the use of olanzapine postnatally has led to reports of adverse events in exposed infants. The most frequent adverse event seen is sedation/lethargy, followed by irritability, tremors, insomnia, diarrhea, and poor suckling. Amisulpride should be avoided in general given its RID ~11% and M/P ratio ranging from 11 to 20, suggesting that it may not be safe for use during breastfeeding. Also, clozapine exhibits a high M/P ratio of 2.8 and there was one case of agranulocytosis in the exposed infant. Other case studies of breastfeeding patients taking clozapine have also reported sedation and irregular vocal development in their exposed infants.
Studies investigating the link between postpartum ECT use and negative outcomes among breastfeeding mothers are also limited, with most studies using case-study designs, no control groups, varying ECT procedures, different follow-up periods, and no control for confounders.
Common adverse effects that can occur in the mother include transient memory loss, confusion, anterograde amnesia, and prolonged seizures. Among postpartum mothers with venous thromboembolism, it is recommended that ECT should be avoided in the first 6 weeks of the postnatal period, given that the procedure can displace thrombi. Although it is suggested that anesthetic drugs may pose minimal risks to breastfed infants, clinicians should consult anesthetists to determine whether ECT is a viable option for breastfeeding mothers.
Despite the presence of some research suggesting modest links between perinatal use of psychotropic medications and ECT with poor outcomes in parents and their offspring, clinicians must be cognizant of the multiple methodologic limitations present in studies using observational designs. Indeed, most of the research examining the negative outcomes of perinatal use of psychotropic medications is subjected to confounding by indication (ie, outcomes may be due to active perinatal depression/anxiety rather than the use of psychotropic medications), they often lack a comparator group of perinatal mothers with active depression/anxiety, and either did not adjust or inadequately controlled for potential confounders of these associations (eg, perinatal substance misuse). Therefore, it is difficult to definitively conclude whether psychotropic medications are causally linked to the poor offspring outcomes that are reported in these observational studies.
The work investigating perinatal ECT use also presents with several methodologic shortcomings, including most of the research using very small samples (which are likely statistically underpowered), case-study designs, no comparator groups, varying ECT procedures, different follow-up periods, and no control of confounders. Hence, more methodologically rigorous research is required to understand the effects of perinatal ECT use on adverse maternal and offspring outcomes.
Overall, clinicians need to be aware of these methodologic limitations and that a majority of these reported risks are small in magnitude and do not exceed the threshold to be considered clinically significant on an individual level (ie, relative risk >2). Such information must be discussed alongside the adverse outcomes associated with untreated perinatal depression/anxiety in any conversation that involves the potential risks of using psychotropic medication and/or ECT in the perinatal period.