Introduction

Psychiatric illness may develop or relapse during pregnancy, affecting approximately 500 000 pregnancies each year.^1,2 A high prevalence of psychiatric illness of 14% was reported during the second trimester of pregnancy with a low percentage (5.5%) receiving treatment.³ Uncertainty of best practices may lead to medication discontinuation without consideration of the risks to mother and child associated with the underlying psychiatric illness (Table 1). On the other hand, medications may be initiated or added based on the psychiatric diagnosis without consideration of pregnancy status. Baseline risk of congenital malformation for all pregnancies range from 3% to 6% with medication exposure thought to account for less than 1% of all birth defects. Environmental issues (eg, maternal conditions, such as poor nutrition or infection) account for around 10% and genetics-related congenital malformations account for 15% to 25% of all birth defects.⁴

TABLE 1 Risk analysis for psychotropic medications in pregnancy^{2,5,7,8,12,16–18,26,27}

View Fullscreen

The British Association for Psychopharmacology (BAP) 2017 consensus guidance provides concise evidence and additional leeway for use of psychiatric medication. As an example of this flexibility, Ebstein’s anomaly associated with lithium exposure is now thought to be more likely associated with maternal psychiatric diagnosis.⁵ Unfortunately, pregnancy is not protective against psychiatric illness, and poor psychiatric health care during pregnancy is strongly predictive of continued postnatal psychiatric illness.⁵ Furthermore, a recent study revealed 0.04% of women aged 20 to 35 years who die by suicide do so during pregnancy.⁶ All pregnant patients should be assessed for risk of suicide and infanticide at each appointment. Suicidality or psychosis should be considered psychiatric emergencies requiring prompt intervention.^1,7

Patient-Centered Care

AK, a 22-year-old, pregnant (21 weeks and 3 days) woman, along with her partner, present for an outpatient psychiatric appointment with the chief complaint: “I need help, my meds are off.” The partner confirms that AK has been demonstrating symptoms of psychosis, making odd statements, and ruminating about them. AK has not been taking medicine for 3 weeks since discovering the pregnancy. During the appointment, AK is calm and cooperative and mood is depressed with flat affect and soft speech. Auditory hallucinations are reported as “hearing 2 to 3 voices, mumbling.” AK is paranoid, reporting “they’re after me,” but denies suicidal or homicidal ideation. Most recent medications included quetiapine 400 mg by mouth at bedtime, citalopram 30 mg by mouth daily, and trazodone 50 mg at bedtime. Past history includes schizophrenia, major depressive disorder (MDD), multiple inpatient hospitalizations, and 2 overdose attempts. Domestic violence and childhood emotional, physical, and sexual trauma are endorsed. Family history includes an uncle who died by suicide, father with alcohol use disorder, and mother with untreated depression. Social substance use was denied, she spent time in foster care, “aging out,” and was previously homeless, but now lives with her boyfriend and father of the baby. Vital signs and laboratory results were unremarkable: urine drug screen (UDS) was negative; urine pregnancy test was positive. Brief Psychiatric Rating Scale score was 42.

Individuals living with mental illness may be able to discontinue medications preconception or during pregnancy for mild illness. In practice, this decision is patient-specific and should be made after careful consideration of all potential risks and benefits.^7,8 Pregnancy is reported as a key reason for discontinuing antidepressant and antipsychotic medication with a goal of avoiding fetal exposure.^9,10 Cited rationale for medication discontinuation includes fear of harm to the baby; pressure from family; symptoms of psychiatric illness, such as paranoid thinking; or poor ability to follow up.^1,7 Untreated psychiatric diagnosis may result in both obstetric (eg, preterm delivery, prolonged labor, fetal distress) and neonatal (eg, decreased developmental scores, congenital malformations, small for gestational age, neonatal intensive care unit admission) difficulties.² Additionally, untreated or inadequately treated maternal psychiatric illness could result in poor adherence with prenatal care; poor maternal self-care and nutrition; increased use of substances, such as alcohol and tobacco; unsupervised use of herbals or prescription medication; poor mother-infant bonding, and poor infant care.²

However, when the risk-versus-benefit analysis is considered for discontinuing psychiatric medications during pregnancy, the risk of untreated illness must be considered alongside the risk associated with medication continuation. Advising medication discontinuation often exchanges the fetal risk of medication exposure for fetal and maternal risks of untreated psychiatric illness.⁵ Often, in research studies that evaluate the risk of psychiatric medications, healthy controls are compared to medicated psychiatric patients with no control for the psychiatric illness itself as a confounder.⁵ Highlighting this effect, a meta-analysis including 23 studies found only 6 attempted to control for depression severity.¹⁰ Antidepressant medications do not seem to confer neonatal risk to the degree previously thought when women are matched based on the severity of depression, lending credence to the theory of depression potentially being a stronger risk factor than the psychiatric medications, and this theory could be extrapolated to other psychiatric diagnoses.^5,7,11

Data regarding the risks and benefits of psychiatric medications in the perinatal period have many weaknesses, especially multiple confounding factors. Therefore, expert opinion recommends avoiding definitive statements about risks and benefits when speaking with patients.⁵ Along with standard education parameters, providers may utilize information in Tables 1 through 4 to inform patients and caregivers about the risks associated with untreated psychiatric illness and the potential risks of medication(s). This information may also be used for a risk-benefit analysis as part of preconception planning with a goal to initiate preferred agents in women of childbearing age in case of pregnancy. Multidisciplinary management is a best practice recommended to facilitate care, and the entire education session should be documented along with patient understanding. For patients who are adamant about discontinuing psychiatric medications, support should be provided as patients are likely to stop drug therapy on their own if not assisted.^8,9 It is important to educate patients that symptomatic relapse may occur if medications are discontinued.¹² Considerations should be given to the disorder being treated, severity of disease, course and most recent episode, number of episodes, response to medication, medication dose, and evidence supporting length of therapy as these factors contribute to the likelihood of symptom relapse occurring.

Treatment Strategies

TM, a 32-year-old, pregnant (34 weeks and 1 day) patient was voluntarily admitted to the inpatient psychiatric unit with suicidal thoughts for “a while.” TM reports, “an attempt about 2 months ago by overdosing on zolpidem but vomiting the pills.” Following that event, psychiatric care was sought, and lithium, which was initially held due to first trimester pregnancy, was restarted when TM was 22 weeks pregnant.^5,7 TM has been experiencing depression off and on for the past 2 years since the birth of her first child with a history of MDD prior to giving birth. Hopelessness, inability to work, loss of interest, low energy and motivation, and feelings of isolation have escalated over the past few months. These features contributed to daily suicidal thoughts, but TM has resisted acting on them because of her other child and current pregnancy. TM reports laying on the couch for hours each day and that caring for her child is difficult. Over the last 3 days, she is hearing music playing with no source for the sound. Current medications include fluoxetine 60 mg by mouth daily (adherent), lithium 300 mg by mouth twice daily (often forgets), zolpidem 5 mg by mouth at bedtime as needed (not using), and a prenatal vitamin (adherent). TM has taken fluoxetine since first being diagnosed with depression and stated it previously worked well. TM states lithium works when taken consistently. Psychiatric history includes depression and postpartum depression with psychosis. Family history is positive for a mother with depression. Socially, all substances are denied, lives with her husband and son, and works as a stay-at-home parent since the birth of her first child. Past medical history is positive for diabetes and hypertension, both at goal without medication. Vital signs, urinalysis, UDS, and electrocardiogram (ECG) were unremarkable; labs were notable for the following: glucose = 98 mg/dL, creatinine clearance = 100 mL/min, A1c 5.2%, and lithium concentration = 0.2 mmol/L. A urine pregnancy test was positive. The multidisciplinary team requested input whether to continue current pharmacotherapy or introduce new medications to achieve therapeutic response. Patient Health Questionaire-9 score was 14.

In mild-to-moderate illness, psychotherapy may often be considered first line, but in moderate-to-severe or recurrent illness, pharmacotherapy should be considered.^7,8,15 Relatively limited data regarding the efficacy of psychiatric medications during pregnancy exists; therefore, in practice, it is usually necessary to extrapolate and generalize efficacy evidence from studies in nonperinatal populations. Data commonly used to estimate the risk associated with medications in pregnancy include animal studies, case reports, case-control studies, prospective cohort studies, historical cohort studies, and voluntary reporting systems. Difficulty with the available data lies in the value of historical reports and noncontrolled studies as birth defects can occur by chance and recall bias may be a concern.

Optimally, shared decision making should occur prior to pregnancy as part of preconception planning.^2,5,7 These discussions early on can help determine pharmacologic therapy. For example, a patient of childbearing age should generally not be prescribed valproic acid (VPA) due to strong associations with teratogenicity.^2,5,7,8,17 If VPA is necessary, oral contraceptives with little potential for nonadherence, such as intrauterine devices, should be considered along with folic acid. Nonpharmacologic treatment options, such as psychotherapy, should be implemented when appropriate. Nonpharmacologic options should be considered for appropriateness to potentially help avoid medication use or assist with reducing medication burden.

The selection of medication to minimize the risk of illness should be based on disease-specific guidelines and efficacy evidence, history of past efficacy with medications, prior exposure during pregnancy, and available reproductive safety information. Medications with fewer metabolites, higher protein binding, and fewer interactions with other medications are generally preferred if possible.^2,5 It is recommended to use the lowest possible dose, but a single medication at a higher dose is preferred over multiple medications to limit overall neonatal exposure.^5,7 If a mother is already prescribed a medication, and the baby has been exposed, efforts should be made to optimize the current medication before utilizing a switch or augmentation strategy.^2,5,7,16 Potential benefits of psychiatric medications, such as treatment response, remission, and prevention of relapse should be part of patient education. Patients should be advised regarding what may occur if treatments are stopped abruptly (eg, potential decompensation and withdrawal) and to seek assistance with discontinuation if desired. It is crucial to conduct a literature review prior to making recommendations for medication use in pregnancy as this information is constantly evolving, and both mother and baby are affected by therapeutic recommendations.

None of the currently available antidepressants is absolutely contraindicated in pregnancy (Table 2). Patient-specific factors must be considered in therapy selection and treatment adjustment during pregnancy. Most guidelines recommend selective serotonin reuptake inhibitors (SSRIs) as first-line pharmacologic treatments for depression because they have the largest evidence for reproductive safety.^2,5,7,18,19

TABLE 2 Evidence for commonly used antidepressant medications in pregnancy^{2,5,7,8,12,16-24,29,30,32-39}

View Fullscreen

Sertraline is currently considered the most preferred agent during pregnancy.^{2,5,7,17–20} Citalopram is considered a preferred agent, and fluoxetine is nonpreferred, but continuation is acceptable.^18,19 Paroxetine has conflicting evidence for use.^2,5,7,16 Of the tricyclic antidepressants (TCAs), amitriptyline, clomipramine, desipramine, imipramine, and nortriptyline have sufficient evidence to support use and are considered to have the fewest adverse effects.^7,19 Most evidence related to adverse effects is associated with clomipramine and includes preterm birth, low birthweight, hypoglycemia, respiratory diagnoses, low Apgar score, jaundice, and cardiac malformations.^5,18 TCA exposure effects are considered to be “not significantly different” when compared with SSRIs or other antidepressants besides monoamine oxidase inhibitors (MAOIs).⁵ MAOIs should be avoided during pregnancy predominantly because of the association with fetal growth restriction and hypertensive crisis.^2,5,7,16,19 Serotonin and norepinephrine reuptake inhibitors, such as duloxetine and venlafaxine, also have limited information and are not considered preferred agents.⁸ Other commonly used agents, such as bupropion, mirtazapine, and trazodone, have limited data. Bupropion is not considered a first-line agent during pregnancy due to possible associations with spontaneous abortion or cardiac malformations though there are conflicting reports highlighting no risk.^5,21–24 Clinically, bupropion may be considered for depressed women who have failed nonpharmacologic smoking cessation attempts as well as nicotine replacement therapy or who have failed other agents and cannot be switched.⁵

In clinical practice, when psychotic symptoms (such as hearing music, muffled talking, or paranoid thoughts) are present, an antipsychotic agent as adjunct therapy may be considered for MDD or postpartum depression, similar to treatment of bipolar disorder.^7,19 Data related to antipsychotic medication use in pregnancy comes from data with patients diagnosed with schizophrenia, bipolar disorder, and depression.^5,25 Reinitiating lithium as adjunctive therapy could be considered for treatment-resistant depression associated with pregnancy though this is more commonly recommended for bipolar disorder or postpartum depression with psychosis. Evidence is lacking regarding use of adjunct agents for MDD in pregnancy, but in clinical practice, selection of adjunct agents is similar for those who are not pregnant (eg, antidepressant with alternative mechanism of action, antipsychotic medications, or lithium to target remaining complaints or symptoms). Another consideration includes increasing the dose of lithium to account for possible pharmacokinetic changes in pregnancy if the patient has been adherent and the medication was reported to have been working previously.^5,7,19,26

Case Outcome

Becausee TM has continual suicidal thoughts, intent, a plan, and means, she is considered high risk and appropriate for inpatient admission to ensure a safe environment.¹⁶ Because the patient had a positive response to fluoxetine in the past, is 34 weeks pregnant, and the baby has already been exposed to fluoxetine, a first-line intervention will be to optimize fluoxetine dosing by increasing to fluoxetine 80 mg by mouth daily.^2,5,7 In the case of poor response to an increased dose of fluoxetine, another antidepressant could be considered. It would be pertinent to speak with TM regarding all past medication trials to determine what previously worked well. Selecting a medication with a history of positive response should be initiated to provide efficacy and limit subsequent drug exposures for the baby. If auditory hallucinations continue with antidepressant treatment, an antipsychotic could be considered. Risks and benefits of lithium were discussed with TM, and rationale was provided for discontinuation though it may be an alternative option (Table 4). Because TM was nonadherent to lithium, exposure to the baby was limited and would not warrant further use and titration at this time. When comparing lithium to initiating another agent with less fetal risk, such as a different antidepressant or adjunctive antipsychotic, the latter may be preferred.^7,15 Recommendations for lithium use during pregnancy generally are to gradually taper until discontinuation prior to conception when possible for mild or infrequent episodes.⁵ For moderate illness or risk of relapse, it is still recommended to taper lithium until discontinuation prior to conception and avoid lithium in the first trimester if possible.^5,7,26 Lithium, along with folic acid, may be restarted in the second trimester if needed, which occurred in this case. For severe illness, lithium may be continued for the duration of pregnancy with clear and complete patient education.^2,5,7 Fetal ECG monitoring and anatomical evaluations are recommended for babies exposed to lithium in the first trimester.^2,7 It is also recommended to discontinue lithium 48 to 72 hours prior to expected delivery or to reduce dose prior to the onset of delivery to assist in preventing possible lithium toxicity in the infant.^2,5,7,26,27 The medication may be restarted at a prepregnancy dose immediately following delivery for relapse prevention if the patient does not wish to breastfeed.^5,7,26,27 Because TM is at 34 weeks and likely to deliver her baby in the next 6 weeks, lithium could be continued or discontinued based on available literature. In this case, the multidisciplinary team decided that the continued use of lithium should be avoided and rationale for use would be reevaluated later. The patient was also instructed not to take zolpidem. A safety plan with clear instructions for what to do during periods of suicidal thoughts was developed with the patient prior to discharge with weekly obstetric follow-up and psychiatric follow-up in 2 weeks. Patients should be educated about all potential medication risks (Tables 2 through 4).^{5,8,12,16–24,26–53}

TABLE 3 Evidence for commonly used antipsychotic medications in pregnancy^{5,8,12,16,18,19,31,40-51,53}

View Fullscreen

TABLE 4 Evidence for commonly used mood stabilizing medications in pregnancy^{5,8,12,16,18,19,26,27,52}

View Fullscreen

Treatment Nonadherence and Long-Acting Injectable Antipsychotics (LAIAs)

SG, a 29-year-old, pregnant (9 weeks and 6 days) patient was brought to the emergency department (ED) by police after being reported missing for days and found wandering around the airport dressed inappropriately for the season. Officers described SG as incoherent, fearful, disheveled, and confused. SG denied suicidal ideation but reported homicidal ideation toward her cousin, endorsed visual hallucinations of a snake and auditory hallucinations of her cousin talking to her boyfriend, being stalked by multiple people, and being unable to sleep for several days. Prior to transfer to the psychiatric unit, SG’s phone was taken away, resulting in agitation; yelling, kicking, and striking at staff; and trying to elope. Her boyfriend, who was present in the ED, stated these outbursts occur frequently following similar, seemingly small requests. Haloperidol 5 mg intramuscular (IM) was given once after verbal de-escalation attempts failed and oral agents were refused.^54,55 SG takes medications off and on, which is consistent with history. Lithium has not been taken in 2 months, aripiprazole was self-stopped almost 3 weeks ago, and her last appointment was missed. Past psychiatric history includes bipolar I, recurrent, severe with psychotic features for 10 years; a mother and grandmother with bipolar disorder, and her father died by suicide in 2018. SG denies substances currently, but reports THC use in the past and currently resides with her boyfriend of 9 years. Past medical history includes hypotension and headache. Vital signs, laboratory tests, and diagnostics (UDS and ECG) were unremarkable except glucose = 199 mg/dL, A1c = 6.6%, and urine pregnancy test positive. Young Mania Rating Scale score was 25.

Limited data is available regarding rapid tranquilization (RT) in pregnancy, but agents with short half-lives are preferred.^54,55 Three primary medications are recommended if nonpharmacologic strategies and oral agents are insufficient. When IM medications are needed, haloperidol, lorazepam, or promethazine are preferred, but none is considered first-line over the others.⁵⁴ These medications may be used alone or in combination with the lowest effective dose used.⁵⁶ Benzodiazepines used near delivery can result in floppy infant syndrome.⁵⁴ Benzodiazepines should not be utilized in pregnancy unless necessary for acute treatment of anxiety or agitation.^54,55

Alternatively, antipsychotics for RT used near delivery may lead to extrapyramidal symptoms for the baby.⁵⁴ Studies report moderately increased risks for several adverse maternal and infant outcomes (Table 3) when pregnant patients exposed to oral antipsychotics are compared to nonpsychiatric, antipsychotic-unexposed pregnancies.^57–59 Conversely, studies report limited associations between antipsychotics and adverse outcomes for the infant or mother when they compared pregnant patients who have a psychiatric diagnosis that are either exposed or unexposed to oral antipsychotics.^25,55,60,61 It is also important to consider that patients prescribed oral antipsychotic medications during pregnancy are more likely to have additional risk factors such as diabetes, hypertension, obesity, alcohol use, tobacco use, or use of other medications when compared with those not taking antipsychotic medications during pregnancy.^25,55,60 When those confounders are controlled, few differences in adverse outcomes are found between medicated and nonmedicated psychiatric illnesses during pregnancy.

To date, most evidence suggests first generation oral antipsychotics (FGAs) and second generation oral antipsychotics (SGAs) are not major teratogens. Available evidence reports no increased risk of stillbirth or spontaneous miscarriage though congenital malformations are reported.^18,61 Overall, there is little difference in fetal risks between FGAs and SGAs or between individual antipsychotics, but data for some individual antipsychotics, such as brexpiprazole, cariprazine, lurasidone, and lumateperone, are limited.^48–53 Oral antipsychotics can be utilized in pregnancy with quetiapine and haloperidol being preferred. Routine use of clozapine should be avoided if possible unless benefits outweigh risks.^18,19,31 In practice, however, patients prescribed clozapine have most likely failed multiple trials other antipsychotic agents and, therefore, must continue with treatment utilizing clozapine. Olanzapine has conflicting evidence for use (Table 3), and aripiprazole and risperidone have limited-to-moderate data (Table 3).

There is higher relapse potential among pregnant patients with severe illness who are medication nonadherent.^62,63 Currently, pregnancy is not a contraindication for using LAIAs.^64–66 Essentially, reported adverse effects from oral antipsychotics, when compared against the LAIA counterparts, are similar. Reinstein et al recommend a similar criterion be applied to determine appropriateness of LAIA, whether pregnant or not, such as medication nonadherence, especially when leading to hospitalizations; severe illness with extended hospitalizations associated with bipolar disorder; schizophrenia and schizoaffective disorder; previous decompensation during pregnancy; or substance use associated with poorly controlled illness.^67,68 LAIAs in pregnancy may be especially beneficial because first pass metabolism is avoided, which may limit fetal exposure to fluctuating plasma concentrations. Pharmacokinetic changes occurring in pregnancy (eg, increased body mass and blood volume) could make an LAIA with more frequent administrations appealing, so dosing can be adjusted as needed or discontinued if adverse effects occur. Alternatively, for patients who have difficulty keeping appointments, LAIAs with longer dosing intervals may be preferable. Patient-specific factors and clinical judgment should be exercised when selecting the LAIA. Women who are taking antipsychotic medications during pregnancy should be monitored closely for gestational diabetes and hypertension.⁶⁷ Data for LAIAs during pregnancy are limited but summarized in Table 5.^{65–70,74–83} In general, the antipsychotic effective in managing symptoms prior to pregnancy should be continued throughout pregnancy.⁶⁷ Considerations for medication selection in pregnancy should be similar to that of nonpregnant patients (eg, affordability, efficacy, and tolerability).

TABLE 5 Long-acting injectable antipsychotics (LAIAs) and evidence to support use^65-70,74-83

View Fullscreen

Conclusion

Whereas data regarding the safety of psychiatric medication during pregnancy is growing, efficacy data is still lacking in this population.²⁷ Preconception planning is paramount; utilization of medications with safety data in pregnancy should be considered for patients who wish to conceive. Special consideration should be given to all women of childbearing age whether they plan to conceive or not because unplanned pregnancy is possible. It is important for providers to encourage their patients to enroll in pregnancy registries for psychiatric medications. Additionally, providers are encouraged to publish evidence, at any level, related to pregnant patients diagnosed with a psychiatric illness whether or not they are receiving medication therapy. Drug safety during pregnancy does not correlate to safety during breastfeeding, and a separate risk-benefit analysis must be performed, but that is outside the scope of this manuscript. It is also important to remember the increased risk of suicidality in this population; therefore, objective monitoring should be conducted at each visit throughout pregnancy. Pregnant patients should be monitored more frequently than their nonpregnant counterparts, and drug concentrations, when appropriate, should be evaluated as there are numerous pharmacokinetic changes taking place during pregnancy. Because evidence is evolving, it is of utmost importance that healthcare providers review published medical literature before making practice decisions or educating patients. This article is intended to provide an overview and should not replace clinical judgment when caring for pregnant patients and their babies.