Pharmacological treatment of bipolar disorder in pregnancy: An update on safety considerations

Abstract

Pregnancy in women with bipolar disorder (BD) can be considered a high-risk pregnancy in view of several clinical and pharmacotherapeutic considerations. Pharmacological treatment during pregnancy requires a careful weighing of psychotropic drug exposure against the risk of BD relapse. An untreated bipolar illness can negatively affect the health of mother as well as unborn child in the event of a relapse. Availability of well balanced, latest information on safety of prophylactic drugs for BD is crucial for making informed decisions. The review provides an evidence-based update (2015–2021) on the drug safety considerations involved in providing care for women with BD who are either pregnant or planning to conceive in near future. Literature review based on systematic reviews, meta-analyses, and data available from studies based on large-scale cohorts and birth registries has been synthesized and presented along with clinically relevant recommendations.

Introduction

Bipolar disorder (BD) is a severe mood disorder, with an episodic course with varying intervals of remission lasting months or years. Globally, BD has a lifetime prevalence of 1% in representative community samples, with rates close to 4%–5% after inclusion of bipolar spectrum and other subthreshold conditions.[1] Moreover, the typical age of onset for BD is in the early twenties and course is characterized by multiple, recurrent episodes. This leads to substantial overlap with periods of pregnancy and childbirths in affected women.[2] The reproductive milestones increase the risk of relapses because of medication discontinuation, hormonal factors, or other well-known triggers of BD (e.g. sleep deprivation).[2,3,4] The continuation of medication prophylaxis for BD in pregnant women is associated with several clinical and ethical considerations. Management of BD in pregnant women requires careful weighing of risks to the unborn fetus against the risk of relapse of BD. Thus, the provision of readily available, well-balanced, and evidence-based information is crucial in making right informed decisions.

Few prior reviews[5,6,7] had focused on the safety of psychotropic agents used in the management of BD among pregnant women. The reviews by Epstein et al.[5] and Scrandis[6] included studies available till December 2013 and January 2017, respectively. In light of several newer large-observational studies and meta-analyses in the past 5 years, there is a need to provide an updated review of available literature. The review by Clark[7] is relatively recent but had not focused on antidepressants and anti-anxiety medications, which are frequently co-prescribed in acute care of patients with BD during bipolar depression. It will be useful to focus on all medications which are potentially used in both acute and chronic care of pregnant women with BD.

With this background, this review aims to provide an evidence-based update on the treatment considerations in care for pregnant women or those who are planning to conceive in near future.

Search strategy

A systematic search based on PubMed/Medline database was conducted for time period between January 2015 and July 2021. The search strategy employed consisted of MeSH terms and Boolean operators as described in Box 1. We restricted PubMed search to include systematic reviews, meta-analyses, and comparative and observational studies. Studies conducted on animals and whose full text was not available in English language were excluded.

Box 1.

Search strategy

PubMed search 1: (“psychotropic drugs”[MeSH Terms] AND “congenital abnormalities”[MeSH Terms]) AND ((meta-analysis [Filter] OR observational study[Filter] OR comparative study[Filter] OR systematic review[Filter]) AND (2015:2021[pdat]))

PubMed search 2: (“psychotropic drugs”[MeSH Terms] AND “pregnancy”[MeSH Terms]) AND ((meta-analysis[Filter] OR observational study [Filter] OR comparative study [Filter] OR systematic review[Filter]) AND (2015:2021[pdat]))

Supplementary search

Cross-references of select papers

Google Scholar search for additional articles not indexed in PubMed/Medline

A total of 158 articles were obtained on PubMed search, of which 49 (17 meta-analyses, 18 systematic reviews, and 14 observational studies) were included for further review after manually screening their title/abstract to assess their relevance. Further, the cross-references of select articles were searched for suitable articles of importance. In addition, Google Scholar search was carried out to further identify recent studies related to theme which were not indexed in PubMed.

A literature review based on systematically retrieved systematic reviews, meta-analyses, and findings from large-scale cohorts and birth registries has been synthesized (2015–2021), and presented in a clinically relevant narrative. Studies prior to 2015 have been referred to, only if they were deemed to be key papers and/or carry important findings which continue to be relevant to the field till date.

Risk of Relapse of Bipolar Disorder in Pregnancy

Pregnancy was found to be a strong predictor of drug discontinuation in women due to excessive concerns for fetal safety.[8] The recurrence rate was 71% without any maintenance therapy and 24% with maintenance therapy, with 66% relative risk reduction with treatment.[9] The recurrence risk was greater following an abrupt or rapid (1–14 days) discontinuation compared to slower taper.[10] Pregnancy is not protective for relapse of BD and postpartum risk increases further.[11] However, individual case considerations such as frequency and severity of episodes must guide the decision.

The following options can be considered for women with BD, who decide to conceive with continuation of prophylaxis:

• Continuation, preferably monotherapy at minimum effective dose

• Discontinuation or drug-free (with psychological interventions, if indicated)

• Limited discontinuation (for the first 12 weeks of organogenesis only)

• Substitution (gradual cross-taper with a safer agent) prior to conception.

Ideally, such discussions must take place prior to conception. In real life, however, pregnancies are often unplanned.[12] Therefore, contraceptive counseling should be offered to all women in follow-up for BD,[13,14] and reproductive safety of an agent must be considered while planning prophylaxis in reproductive age group women.

Safety and Choice of Pharmacotherapy for Bipolar Disorder in Pregnancy

The safety of pharmacotherapeutic agents used to treat BD needs to be considered in terms of associated (a) teratogenicity, (b) poor neonatal adaptation, and (c) long-term neurobehavioral sequelae.

One important point in the interpretation of drug safety data is that the data must be viewed in context of baseline rates of congenital malformations (2%–3.5%) in general population studies.[15,16] Further, patients with BD may be at slightly higher risk of adverse pregnancy outcomes not only due to medication exposure but also potentially due to lifestyle factors, obesity, or frequently associated comorbidities (acting as confounders in studies).[15,16,17]

Table 1 summarizes the safety concerns associated with drugs used in management of BD in pregnant women.

Table 1.

Reproductive safety profile of psychotropic drugs used in bipolar disorder

Medication	Potential fetal/newborn risk summary	Clinical recommendation
Mood stabilizers
Lithium	Small risk of Ebstein’s anomaly (0.1%–0.05% live births) with first-trimester exposure No evidence of behavioral or cognitive problems in children exposed to lithium till 5 years of age	Fetal echocardiography Frequent serum lithium monitor in view of fluid shifts over pregnancy
Valproic acid Divalproate	Increased rate of congenital malformations (approximate 10%), including neural tube defects Linked to neurodevelopmental disorders in children with in-utero exposure	Use in pregnancy for BD banned in several countries Must be avoided in pregnancy High-dose folic acid supplementation (4 mg/day) advised prior to conception
Carbamazepine	Congenital malformations: Around 3%–6% Risk of bleeding diathesis in neonates due to Vitamin K deficiency	If possible, use in pregnancy should be avoided High-dose folic acid supplementation (4 mg/day) advised prior to conception Oral Vitamin K supplementation for newborn
Oxcarbazepine	Congenital malformations: Around 3% Risks appear to be low, though more data is required Possible link to autism with in-utero exposure but need more studies
Lamotrigine	Risk of major congenital malformations in most studies (2%–3%) is comparable to baseline population rates No significant neurobehavioral sequelae	Preferred agent to use, if clinically indicated (e.g., bipolar depression or depression predominant course) Dose may be increased due to enhanced clearance rates across pregnancy stages
FGAs
Haloperidol Trifluoperazine	Increased risk of preterm labor and reduced birth weight; but not always clinically significant No increased risk of congenital malformations or neurodevelopmental delays	May watch for extrapyramidal and other adverse effects Appear safe to use in pregnancy, though FGA use is often restricted to acute care rather than prophylaxis in BD
SGAs
Olanzapine Quetiapine Aripiprazole Risperidone	Risk of congenital malformations for most SGAs: Around 3.5%, not substantially different from general population rates (2%–3.5%) Risperidone is associated with possible small increase in risk, but needs replication Increased risk of maternal weight gain, gestational diabetes mellitus, and large-for-gestational-age baby	Relatively safe to use in pregnancy, especially olanzapine, quetiapine, and aripiprazole Periodic monitor for maternal glucose and weight with olanzapine/quetiapine Ultrasound for fetal size in late pregnancy
Clozapine Ziprasidone Lurasidone	Limited/scarce quality data on safety in pregnancy	Weekly monitoring for agranulocytosis for infants who were exposed to clozapine in utero till six months after birth
SSRIs
Fluoxetine Escitalopram Sertraline Paroxetine Fluvoxamine	Small increased risk of spontaneous abortion, preterm labor, and lower weight at birth Small increased absolute risk (2/1000 births) of congenital cardiac defect with first-trimester paroxetine exposure Conflicting evidence for persistent pulmonary hypertension of newborn with third-trimester exposure Mild poor neonatal adaptation syndrome in about one-third of cases with exposure during late pregnancy	Largely safe to use in pregnancy, if indicated Paroxetine use may be avoided
Antidepressant drugs (other than SSRIs)
Bupropion Mirtazapine Venlafaxine Duloxetine	Less data on their safety as compared to SSRIs Reported risks of congenital malformation appear similar to SSRIs in limited available data
Benzodiazepines
Clonazepam Lorazepam Alprazolam Midazolam diazepam Chlordiazepoxide	Older evidence pointed to association with oral cleft defects, which was not found in recent meta-analytic evidence Third-trimester exposure associated with neonatal toxicity and withdrawal symptoms Case reports of floppy infant syndrome	Regular or high-dose use should be avoided as far as possible Tapering of dose should be considered before delivery Use in acute care, no role in prophylaxis

FGAs=First-generation antipsychotics, SGAs=Second-generation antipsychotics, SSRIs=Selective serotonin reuptake inhibitors, BD=Bipolar disorder

Mood Stabilizers

Lithium

A recent meta-analysis found that lithium was associated with an increased odds of congenital anomaly (29 studies with inclusion of 13 in analysis; n = 23,300; odds ratio [OR] = 1.81, 95% confidence interval [CI] = 1.35–2.41; number needed to harm = 33; 95% CI of 22–77; prevalence: 4.1%) and cardiac malformations (exposure during the first trimester).[18] In another meta-analytic study of 727 pregnancies with lithium exposure and 21,397 pregnancies in controls, first-trimester lithium exposure was associated with some risk of congenital malformations (OR: 1.62, 95% CI of 1.12–2.33; pooled prevalence of 7·4% vs. 4·3% in controls). However, the elevated risk of cardiac malformations was not found.[19]

In another recent cohort involving 1.3 million pregnancies, cardiac abnormalities were found in 2.4% of lithium-exposed pregnancies.[20] The overall risk of right ventricular outflow tract obstruction defects (inclusive of Ebstein's anomaly) was 0.60 per 100 live births among lithium-exposed infants (compared to 0.18 per 100 among lithium-unexposed infants). The risk appeared to be dose related. The adjusted risk ratios for cardiac malformations ranged from 1.11 for daily dose of ≤600 mg/day to 3.22 for >900 mg/day.[21]

Older studies had indicated an increased risk of fetal tricuspid valve defect (Ebstein's anomaly) with antenatal lithium exposure. The more recent studies have reported the rates to be 1in 1000–2000 lithium-exposed pregnancies. While it still represents a 10- to 20-fold increased risk over baseline, the absolute rates reported are low (0.1%) in the exposed pregnancies.[5,21]

Poor neonatal adaptation syndrome has been reported with lithium. Floppy baby syndrome, hypotonia, respiratory distress, lethargy, neonatal goiter, diabetes insipidus, etc., have also been reported.[5,22,23] Furthermore, another recent large retrospective cohort study (n = 443 pregnancies among women with BD-I on lithium therapy) reported the risk of miscarriage with lithium even after adjusting for possible underlying maternal condition causing multiple miscarriages and maternal age at conception (adjusted OR: 2.94, 95% CI of 1.39–6.22).[24]

Clinical points

• Lithium has relatively low teratogenic risk and can be considered for specific indications (e.g. presence of lifetime suicidality, lithium responsive patients, etc.), though most atypical antipsychotic agents and lamotrigine have even better safety profile than lithium

• A small increased risk of Ebstein's anomaly (1–2/1000) can be screened in level II ultrasound and fetal echocardiography

• Frequent therapeutic drug monitoring and the lowest effective dosage should be maintained during pregnancy

• Women should be well hydrated, especially around delivery to curtail the chances of lithium toxicity

• Lithium may be withheld 24–48 h before labor to reduce the risk of neonatal toxicity and/or poor neonatal adaptation syndrome

• Neonatal electrocardiogram and monitoring for any signs of toxicity.

Divalproate/valproic acid

Prenatal exposure to valproic acid or its preparations, especially in the first trimester, is associated with a markedly high rate of multiple congenital malformations (9%–11%) and several-fold higher risk for neural tube defects.[25,26] A constellation of specific congenital anomalies in neonates known as fetal valproate syndrome has also been reported in the available literature. A prospective cohort from the EURAP registry (1999–2016) reported a congenital malformation rate of 10.3% with valproate, with increased rates for all doses of valproate exposure.[25] The risks increase with higher dose and combination therapy. The Cochrane database review also reported a greater risk of malformation (10.93%, 95% CI: 8.91–13.13) among any anti-epileptic monotherapy.[26] The characteristic fetal malformations included spina bifida, atrial septal defect, hypospadias, and cleft palate, among others.

Increased risk has also been reported for poor neonatal adaptation syndrome.[27] Valproate has been significantly associated with adverse neurodevelopmental outcomes including increased risk of developing autism spectrum disorder.[28] Neurobehavioral sequelae, including intellectual impairments, have also been reported.

Contraception must be discussed with all women on valproate therapy to prevent unplanned pregnancies. European countries have essentially banned the use of valproate in pregnancy. Valproate must be avoided in women of childbearing potential in view of the above-stated risks and availability of medications with better safety profile.

Clinical points

• Higher teratogenicity; avoided in favor of safer agents; banned in some countries

• Prior to conception, women with BD preferably to be shifted to another safer prophylactic agent

• Folate supplementation should be started prior to conception.

Carbamazepine

Prenatal exposure, more so in the first trimester, is associated with an increased risk of congenital malformations ranging between 3% and 6%,[25,29,30] which increases further at doses above 1000 mg/day.[31] A recent meta-analysis has reported more congenital malformations with carbamazepine exposure compared to controls (OR: 1.37; 95% CI of 1.10–1.71).[32] There is a risk of spina bifida with first-trimester carbamazepine exposure, though some studies have conflicting finding.[26] It has been variably associated with other craniofacial and skeletal abnormalities, small-for-gestational-age babies, and developmental delays. Carbamazepine exposure during pregnancy might also cause Vitamin K deficiency in neonates, leading to an increased risk of bleeding diathesis in neonates.

Clinical points

• Higher teratogenicity; better avoided in favor of safer available agents

• Oral Vitamin K (20 mg/day) supplementation in the last month of pregnancy in women taking carbamazepine

• Neonate may be given parenteral Vitamin K (1 mg IM), shortly after birth.

Lamotrigine

Multiple pregnancy registries indicate a low risk of congenital malformations (2%–3%) with lamotrigine (at doses <300 mg per day), which is comparable to the baseline rate of congenital malformations in general population.[25,26,30,31,32] The risk of malformations was 1.9% among lamotrigine-exposed infants; however, risks increased to 9.1% for lamotrigine plus sodium valproate (OR: 5.0; 95% CI of 1.5–14.0), and 2.9% for lamotrigine combined with any other anti-epileptic drug (OR: 1.5; 95% CI of 0.7–3.0).[33]

The findings from a pregnancy registry of antiepileptic drugs reported a higher risk of oral cleft defects with an OR of 10.4 (95% CI of 4.3–24.9),[33] but more recent case–control study (3.9 million births from 19 different population registries) failed to show any significant association with oral cleft defects on lamotrigine monotherapy.[34] No marked risk of adverse fetal outcomes in the last trimester was observed in lamotrigine-exposed pregnancies.[35] The favorable safety profile of lamotrigine has been further supported by recent large observational studies and meta-analyses.[36,37] Available literature so far does not reveal any significantly increased risk of poor neonatal adaptation syndrome or neurobehavioral sequelae till 12 months of age with lamotrigine.

Clinical points

• This is among relatively safer mood-stabilizing agent for pregnant females

• Dose increase may be needed to maintain therapeutic levels as its metabolism is significantly affected by changes in estrogen levels

• Keep a watch for lamotrigine-induced rash in baby.

Oxcarbazepine

The EURAP registry study reported malformations in 3% of oxcarbazepine-exposed pregnancies.[25] Relatively limited information is available on teratogenic risks associated with oxcarbazepine use in pregnant women as compared to carbamazepine.[26] The risks appear to be low, but more substantial data are required for a conclusive statement. There may be some increased risk of developing autism in children with antenatal oxcarbazepine exposure (OR: 13.51, 95% CI of 1.28–221.40).[28]

Clinical point

• Risk appears low, but more data are needed.

Atypical antipsychotic agents

Available evidence on second-generation antipsychotics (SGAs) suggests that these are relatively safe in pregnancy, with a small association with adverse neonatal and obstetrical outcomes (heart defects, preterm labor, decreased birth weight, etc.) which does not necessarily imply causation.[38,39,40] A systematic review of 12 studies (1782 cases and 1,322,749 controls) reported a OR of 2.03 for SGA-exposed pregnancies (95% CI: 1.41–2.93).[41] Further, SGAs do not appear to differ from first-generation agents such as haloperidol, which have more substantial reproductive safety literature.[21,42]

The cumulative data for first-trimester-exposed pregnancies reported malformation rates of 3.5%–3.6% with olanzapine, quetiapine, and aripiprazole. However, a slightly increased rate of 5.1% with risperidone needs a further assessment.[42] There is not much pregnancy-related safety data available for lurasidone.

A recent systematic review concluded that rates of gestational diabetes mellitus are higher in the SGA exposed group (2.6%–22%), compared to the unmedicated pregnant women (0.95%–10.7%), but confounders such as the underlying maternal conditions (e.g. obesity) were not controlled.[43]

Poor neonatal adaptation syndrome has been reported. Neurodevelopmental delays were reported at 6 months of age, which were no longer significant by 12 months of age.[44] There is a need for generating more high-quality evidence regarding the long-term safety of SGAs.

Key points

• Most SGAs, such as olanzapine, quetiapine, or aripiprazole, appear to be relatively safe

• Risperidone appears to have slightly increased risk but needs more literature

• Newly marketed drugs such as lurasidone remain underrepresented in safety studies

• Pregnant women on olanzapine and quetiapine should be screened for gestational diabetes mellitus.

Antidepressants

The antidepressants are used to treat depressive episode or significant refractory mood or anxiety symptoms in patients with BD under adequate cover of mood stabilizing agent (s).

Selective serotonin reuptake inhibitors

A large cohort including 949,504 pregnant women and their live-born infants suggests no significant risk of fetal cardiac malformations with selective serotonin reuptake inhibitor (SSRI) use.[45] However, several recent meta-analyses have indicated a small but significant increase in such risk. A recent systematic review (from January 2010 to April 2020) of 15 meta-analyses also suggested an association between the use of SSRIs and congenital anomalies.[46] Of SSRIs, paroxetine in particular has been linked to an increased risk of cardiovascular anomalies. A small risk of persistent pulmonary hypertension of newborn and postpartum hemorrhage was reported with third-trimester pregnancy exposure.[47,48] Overall, the risk does not appear to be substantial clinically and might not justify withdrawing antidepressant, as untreated perinatal depression might also cause adverse maternal and fetal outcomes.[49] Antenatal exposure to antidepressants could lead to higher likelihood of preterm delivery, but not low-birth weight babies.[50]

Poor neonatal adaptation syndrome is observed but is often mild and self-limiting in majority of cases. There is limited and conflicting evidence of association with autism spectrum disorder.[51] Further, there were no significant or consistently reported long-term neurodevelopmental problems reported in studies assessing emotional or behavioral problems, and hyperactivity-inattention symptoms in children born to mothers taking antidepressants during pregnancy.[52,53]

Antidepressant agents (other than selective serotonin reuptake inhibitors)

A systemic review examining the safety of mirtazapine use in pregnancy, including 31 studies (390 neonates with exposure to mirtazapine), reported no significant increase in risk of major congenital malformations with it.[54] There might be an increased risk of spontaneous abortions, respiratory difficulties, and hypoglycemia in neonates with maternal mirtazapine use. However, lack of adequate safety data to draw definitive conclusions was acknowledged. A single prospective comparative cohort study, involving 281 venlafaxine-exposed pregnancies matched to antidepressant-unexposed (n = 1405) and SSRI-exposed (n = 843) comparator groups, concluded no significantly increased risk of spontaneous abortions or any other adverse pregnancy outcomes with venlafaxine.[55] Similarly, very limited data are available for bupropion (congenital malformations rate of 2.3%), warranting more studies.[56] Studies on duloxetine are also rather limited. A recent cohort study found it to be safer for overall congenital malformations, though with a marginal rise in occurrence of cardiac malformations and postpartum hemorrhage.[57]

Key points

• Try and manage milder depressive symptoms with nonpharmacological measures, and prefer to use SSRI if clinically indicated

• Small teratogenic risk is associated with SSRIs (especially paroxetine)

• Limited safety data for non-SSRI antidepressants.

Benzodiazepines

They are frequently used as adjunctive agents in acute treatment of BD. Early evidence, including a meta-analysis by Enato et al.,[58] had suggested an association with oral cleft in the benzodiazepine (BZD)-exposed group (0.7%), compared to the nonexposed/control group (0.06%). Another study from the Swedish Medical Birth Register (1996–2011)[59] reported no congenital risk with antenatal use of BZDs except for alprazolam (OR = 1.97, 95% CI of 1.38–2.02).[59] However, this study finding was based on a relatively small first-trimester alprazolam exposure group of 444 infants (31/444).

More recently, the meta-analysis by Grigoriadis et al.[60] included eight prospective cohort studies, and showed no significant congenital anomaly with in utero exposure to BZD (OR = 1.13, 95% CI of 0.99–1.30). However, combination of BZD and antidepressants (three studies) posed a significant risk (OR = 1.40, 95% CI of 1.09–1.80).[60] In spite of the encouraging recent evidence in terms of congenital malformations, clinicians should still use them with caution during pregnancy. They might increase the risk of spontaneous abortion, preterm birth, and/or other adverse neonatal outcomes.[61]

Key points

• Most BZDs were not associated with congenital anomaly risk

• BZDs should be avoided in pregnancy (especially in later gestational months) due to potential adverse neonatal outcomes

• Must be considered on PRN basis (as required), only if the benefits outweigh risks.

General Considerations

Pharmacokinetic considerations merit attention and dose adjustments as pregnancy progresses. Pregnancy is associated with physiological changes such as weight gain, plasma volume expansion, and increased renal clearance, which can reduce serum drug concentrations of medications. Wesseloo et al.[62] in a retrospective cohort of 113 pregnancies found an increase in lithium elimination clearance during the first two trimesters, and a decline during the third trimester and postpartum. The activity of CYP450 isoenzymes might vary during pregnancy, but the precise characterization is complicated due to their genetic polymorphisms and multiple metabolic pathways.

Recent literature and clinical practice guidelines for BD[2,12,13] provide a set of treatment considerations in clinical care of pregnant women with BD as below

• Shared decision-making, risk–benefit analysis, and case-specific considerations to be kept in mind

• Choose medications with better safety profile as per the latest evidence

• Prefer minimal effective dosages

• Consider dose adjustments as per drug pharmacokinetics and physiological changes

• Minimize the total number of drug exposures for the fetus

• More frequent monitoring of drug levels in pregnancy

• Monitor closely for mood symptoms at each follow-up

• Consider pregnancy in BD as a high-risk pregnancy and offer collaborative care.

Conclusion

Women of childbearing age must be advised to plan ahead for conception. A shared, decision-making must be done, based on the latest evidence on drug safety and risk–benefit analysis on an individual basis. Certain agents such as valproate/divalproic acid are contraindicated in pregnant women, while lamotrigine and most atypical antipsychotic drugs might be chosen if indicated. Lithium has a small but relatively low teratogenic risk and might be preferred in subset of lithium-responsive patients whose episodes are severe or suicidal. Pregnancy in women with BD must be monitored closely to enhance the reproductive as well as mental health outcomes in women.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.