Bipolar disorder I

BD I is characterized by the occurrence of a manic episode, which may be preceded or followed by major depressive episodes (American Psychiatric Association, 2013). A manic episode is identifiable by feelings of euphoria, “excessively cheerful, high, or feeling on top of the world”; individuals may shift between periods of euphoria, dysphoria and irritability, leading to marked impairment in social and occupational functioning (American Psychiatric Association, 2013). These shifts in mood may occur on a rapid basis and over a brief period of time. During a manic episode, a person with BD may experience inflated self-esteem, a decreased need for sleep, rapid speech, racing thoughts, increased sociability and excessive engagement in plans and activities. Poor insight often accompanies acute mania (Anderson et al., 2012). The depressive episodes are characterized by low mood, sadness, lack of energy, sleep and appetite change, with negative ruminating thoughts.

DSM-5 does not use the term “postpartum psychosis,” but rather makes reference to mood episodes presenting either with or without psychotic features (American Psychiatric Association, 2013). This definition fits most closely with the definition proposed by Kendell and colleagues (1987). Irrespective of the diagnostic status, symptoms of postpartum psychosis closely resemble that of BD in the postpartum period.

There has been significant debate about whether postpartum psychosis is a distinct psychiatric condition or a time- and situation-specific expression of BD, for which the trigger is childbirth. In a review of the relationship between BD and postpartum psychosis, Chaudron and Pies (2003, p. 1286) commented that “postpartum psychosis is not a discrete nosologic entity, but a postpartum presentation of an underlying mood disorder. In many, if not most cases, this underlying disorder appears to be in the bipolar spectrum.” Another study by Munk-Olsen et al. (2012) suggested that “an early postpartum onset of illness has prognostic implications and raises the chance of subsequent conversion to a diagnosis of bipolar disorder.”

The incidence of postpartum psychosis ranges between 1 in 500 to 1 in 1000 and may be more common in primiparous women (American Psychiatric Association, 2013). It is treated as a psychiatric emergency because of its associated high risk of suicide and infanticide. Women with a history of BD have an increased risk for recurrence of psychosis in subsequent pregnancies (25–50% versus 0.1–0.25% in the general population) (Jones & Craddock, 2001). DSM-5 has specified the risk of recurrence with each subsequent delivery as even higher (30–50%) for women with a previous postpartum episode with psychotic features (American Psychiatric Association, 2013). The risk of postpartum episodes with psychotic features is particularly elevated for women with prior postpartum mood episodes, those with a history of depressive or bipolar disorder (especially BD I) and for those with a family history of bipolar disorders (American Psychiatric Association, 2013). Jones and Craddock (2001) reported that 74% of women with puerperal psychosis had a history of postpartum psychosis in first-degree relatives. It is unclear whether the development and/or recurrence of postpartum psychosis is due to biological stressors of genetic predisposition, primiparity, pregnancy/delivery complications, previous episodes (Valdimarsdóttir et al., 2009; Pfuhlmann et al., 2002; Sharma et al., 2004) or psychological stressors associated with the early postpartum period (Pfuhlmann et al., 2002).

The risk of bipolar relapse is highest in the immediate postpartum period, with the onset most often occurring within 72 hours (Sit et al., 2006). Patients can manifest symptoms of a manic, depressed or mixed state (Porter & Gavin, 2010). Risk continues to be elevated for years thereafter. Kendell et al. (1987) reported the relative risk of psychiatric admission 30 days from birth to be 6.0, while for primiparae the risk was 10.9.

Sharma and Mazmanian (2003, p. 102) commented that “sleep loss may be the final common pathway by which various putative risk factors produce psychosis in susceptible women.” Sleep loss has been considered to be a significant risk factor for BD and in precipitating postpartum psychosis (Sharma & Mazmanian, 2003; Ross et al., 2005). It is also the most prominent early symptom of mania and a frequent early symptom of depression (Jackson et al., 2003). The shift towards hypomania or mania may occur as early as within one day of lack of sleep, based on findings in a non-perinatal population (Bauer et al., 2006). Insomnia may be present in 42–100% of women who experience postpartum psychosis (Rohde & Marnerso, 1993; Hunt & Silverstone, 1995). Stabilizing mental health during this period may mitigate the effects of insomnia (Bilszta et al., 2010).

Bipolar disorder II

BD II is no longer thought to be a “milder” condition, because the instability of mood experienced by these individuals is typically accompanied by serious impairment in work and social functioning (American Psychiatric Association, 2013). Hypomania may not always cause impairment initially as it is associated with increased energy and productivity (Anderson et al., 2012). BD II is marked by repeated episodes of major depression alternating with brief periods of hypomania. Individuals with BD II tend to present with depressive symptoms, spend more time in the depressive phase of their illness, have greater chronicity of illness compared to BD I and present with high impulsivity (American Psychiatric Association, 2013). In clinical practice, it is often difficult to differentiate this period of elated mood from the normal period of euphoria following delivery. However, the manifestation of extreme elation with other hypomanic features observed in BD II women occurs between 9 and 20% of women after childbirth (Heron et al., 2009; Sharma et al., 2009). Sharma and Khan (2010) found that 57% of postpartum women with a diagnosis of Major Depressive Disorder (MDD) actually suffered from BD II or Bipolar Disorder Not Otherwise Specified. Misdiagnosis of BD II contributes to symptom exacerbation and prolonged illness trajectory, which can interfere with proper maternal attachment (Sharma et al., 2008). Early intervention and treatment are crucial in ensuring healthy mother-infant interaction, bonding and overall well-being of the family unit.

In a study by Baek and colleagues (2011), BD II was further differentiated from BD I by various characteristics. These included frequent depressive episodes and increased rapid cycling, seasonal variation, psychomotor agitation, suicidal ideation, higher comorbidities and substance use disorder. The prominent mood associated with BD II was one of higher irritability for shorter duration whereas presence of increased manic symptoms occurred for a longer period of time in BD I.

Preconception

Preconception counseling in a stable state is ideal (see Chapter 9); it provides an ideal opportunity to assist women with this decision-making process, inform them about the risks and benefits of their therapeutic options during pregnancy, and plan supports during the perinatal period (see also Chapter 9).

Benefits of pre-pregnancy consultation are multifold. It provides the opportunity to discuss treatment options and switch to a safer medication for pregnancy, decreasing the exposure of the fetus to multiple medications (Viguera et al., 2002). Women need to be informed about the high risk of relapse when discontinuing medication (Cohen & Nonacs, 2005) and this is a time when social support, such as family and friends, need to be involved (Frieder et al., 2008). In a study by Viguera and colleagues (2002), a reported 45% of women had been advised by health professionals not to conceive. Following their pre-pregnancy consultation, 63% of women tried to conceive, but 37% chose not to pursue pregnancy out of fear of adverse effects of medicines on fetal development (56%) and risk of relapse with medication cessation (50%).

Up to 49% of pregnancies in the general US population are unplanned (Finer & Zolna, 2011) and 50% of individuals with planned pregnancies will not have seen a healthcare provider prior to conception (American Academy of Neurology, 1998). By the time they present during the first trimester, the fetus has already been exposed to psychoactive medications (Yonkers et al., 2004; Galbally et al., 2010) (also see Chapter 9).

Genetic counseling can address patients’ fears and educate them about disease inheritance. While the genetics of BD are not yet fully determined, it is clear that the offspring of individuals with this condition have an increased risk of psychiatric illness. The relative risk in individuals with a first-degree relative with BD has been reported to be 7–14 times higher than in those without (Mortensen et al., 2003; Wozniak et al., 2012). A maternal history of BD conferred a 12-fold relative risk, a paternal history a 15-fold greater risk and both parents a 113-fold greater risk than individuals without such a family history (Mortensen et al., 2003). Other studies have also found that offspring of parents with a mental illness had a greater likelihood of an affective disorder (4-fold), any mental disorder (2.7-fold), and BD spectrum disorders (Birmaher et al., 2009; Lapalme et al., 1997). Such risk must, however, be discussed within the context of the relationship between genetic and environmental factors. Monozygotic twin studies show concordance rates greater than those of dizygotic twins, illustrating that while genetic factors are strong, environmental factors do influence the expression of this condition (Smoller & Finn, 2003).

Potential effects of untreated maternal bipolar disorder

As many as 51.2% of patients with BD in the United States might go untreated in a given one-year period (Wang et al., 2005). The literature on the effects of untreated maternal BD on the growing fetus and newborn child is evolving; however, the findings of the impact of untreated maternal depression suggest that this carries direct risks to the fetus and infant, as well as risks secondary to unhealthy maternal behaviors (Bonari et al., 2004).

Untreated maternal depression has been found to be related to negative expression and affect in infants, delays in infant growth, and a three-time increased risk of anxiety and depressive symptoms in later life (Avan et al., 2010; Weissman et al., 2006). Infants may also display more avoidant and disorganized attachment (Korja et al., 2008; Martins & Gaffan, 2000), which has been linked to internalizing and externalizing behavior problems in toddlers (Madigan et al., 2007; Trapolini et al., 2007). Some studies report that boys may be more adversely affected by prenatal and postnatal maternal depression compared to girls (Carter et al., 2001).

Higher cortisol levels in prenatally depressed women may be associated with negative growth development in the infant, such as delayed fetal growth, prematurity and impaired fetal brain development (Field & Diego, 2008; Van den Bergh et al., 2005; Weinstock, 2005). The infants of prenatally depressed women appear to show higher cortisol and lower dopamine and serotonin levels (Field et al., 2004).

In women with BD, acute episodes during pregnancy are of concern due to the risks of substance misuse, neglect of antenatal care, poor judgment, impulsive behavior, poor nutrition, increased risk of committing or being a victim of violence and the risk of self-harm, including suicide in major depression (Bonari et al., 2004; Finnerty, Levin, & Miller, 1996; Viguera et al., 2000). In the postpartum period, maternal recurrences can affect practical infant care and mother-infant attachment (Brockington, 2004; Vemuri & Williams, 2011) (also see Chapter 4). Postpartum psychosis carries the risk of neonaticide and infanticide, particularly in the early months; however, more research is needed to ascertain the exact risk factors for neonaticide and infanticide (Spinelli, 2001; Pearlstein et al., 2009; Flynn et al., 2007, 2013).

Bipolar disorder and pregnancy: pregnancy as a protective factor?

Overall, pregnancy may confer some protection for women (Grof et al., 2000); however, in most cases, this protection may not be enough to discontinue medication. Pregnancy is not likely to be protective for women with a refractory illness, early age of onset, multiple episodes and high relapse rate when unmedicated (Viguera et al., 2000; Viguera et al., 2007).

Psychosocial interventions

An optimal therapeutic outcome is based on combining pharmacotherapy and psychotherapy, along with psychosocial interventions to aid in the recovery process.

Guidelines for BD tend to focus on pharmacotherapy, but also address adjunct and complementary therapies. The guidelines developed by the Canadian Network for Mood and Anxiety Treatments (CANMAT) for the management of patients with BD suggest that “although pharmacotherapy forms the cornerstone of management, utilization of adjunctive psychosocial treatments and incorporation of chronic disease management models involving a healthcare team are required for providing optimal management for patients with bipolar disorder” (Yatham et al., 2005).

Psychoeducation can affect the course of BD in a substantial way. It has been found to increase the time to recurrence of the illness and to reduce relapse rates (Colom et al., 2003; Swartz & Frank, 2001). In addition, psychoeducation has been shown to encourage patients to become actively involved in self-management, sleep regulation, avoidance of substance misuse and promote further collaboration among healthcare professionals (Yatham et al., 2005).

Adjunctive psychosocial therapies should be considered early on in treatment and can include psychoeducation, cognitive-behavioral therapy (CBT), interpersonal (IPT) and social rhythm therapy (IPSRT) and family interventions. According to the CANMAT guidelines, CBT, used as an adjunctive therapy for BD, can improve functioning and adherence, as well as decreased relapses, mood fluctuations, need for medications and hospitalizations (Lam et al., 2000; Lam et al., 2003; Scott, 2003). Individual and/or group CBT may also have a role in patients with BD in improving compliance and increasing function and quality of life (Patelis-Siotis, 2001). IPT did not alter time to relapse, but significantly increased the duration of being euthymic and decreased the duration of depression (Lam et al., 2000; Lam et al., 2003; Scott, 2003). Family interventions have been associated with fewer relapses and hospitalizations and improvements in depressive symptoms and adherence with pharmacotherapy (Rea et al., 2003; Miklowitz et al., 2000; Miklowitz et al., 2003).

Psychosocial therapies for BD may be particularly helpful for the depressive phase of BD I and BD II. The UK National Institute for Health and Clinical Excellence (NICE) clinical management guidelines for antenatal and postnatal mental health also recommend brief psychological treatments (including counseling, CBT and IPT) for mild depressive symptoms in pregnant women with BD and combined medication and psychological treatments for moderate to severe depressive symptoms in this population (Tomson et al., 2007; NICE 2014).

In addition to psychotherapy, social support is a key factor in supporting women with BD during their perinatal period (Yonkers et al., 2004; Viguera, et al., 2002). Limited research has been undertaken on the effect of psychosocial therapies on families and partners of individuals with BD. In a study of patients with a major affective disorder or BD, married for I7 years, those who had psychoeducational marital intervention had improvements in adherence and overall function (but not symptoms), compared to individuals receiving medication only (Clarkin et al., 1998).

Yoga has been shown to be effective in both general and perinatal populations as an adjunct for severe mood disorders, as well as a monotherapy for mild depression (Field, 2011). While no studies to date have evaluated yoga as a treatment for BD, several plausible mechanisms of action through which yoga may exert its therapeutic effect on bipolar patients have been suggested. These include decrease in stress reactivity via regulation of the Hypothalamic-Pituitary-Adrenal axis, reduction of cortisol levels, increase in Gamma-Aminobutyric Acid neuroreceptors, improved sleep efficiency and quality, and decreased rumination (Da Silva, Ravindran, & Ravindran, 2009). Yoga is a component in mindfulness-based cognitive therapy (MBCT), which involves meditation exercises, yoga and a therapeutic focus informed by the CBT framework. Recent studies show that MBCT is effective in bipolar patients as an adjunct to medication for reducing anxiety and depressive symptoms; improving emotional regulation, executive functioning, memory, ability to initiate and complete tasks; and increasing mindfulness (Williams et al., 2008; Ives-Deliperi et al., 2013; Stange et al., 2011; Deckersbach et al., 2012). However, larger and randomized controlled trials are needed to confirm its efficacy, and there is still a dearth in data for the pregnant and postpartum populations.

Pharmacotherapy principles of management

Experts recommend classifying all pregnant women with BD I as “high-risk” pregnancies (Viguera et al., 2011). A comprehensive and integrated approach to prenatal care of these high-risk cases has several elements (ACOG, 2008; Yonkers et al., 2004), including:

In order to ascertain the treatment options for each individual during pregnancy, key patient information requires knowledge of past history of BD, related and unrelated to pregnancy, as well as a family history of puerperal psychosis. Obtaining the medical history is also important to understand response to different agents and time to relapse (Jones & Craddock, 2001; Yonkers et al., 2004). The management of each patient must be decided on a case-by-case basis after evaluating the individual risk/benefit ratio. However, there are some general principles of management. Viguera and colleagues (2002) state that patients who have had one manic episode with rapid, total recovery with subsequent stability may be able to taper medication gradually prior to conception. A recognized drawback of this approach is the risk of relapse while awaiting conception. In an unplanned pregnancy, by the time conception is established, stopping certain medications may not be necessary if the period of greatest risk of fetal exposure has already passed. If medication is discontinued, it should be reinstituted if the patient becomes symptomatic, even during the first trimester. In women with brittle BD, with a history of numerous severe recurrences with potential risk of psychosis or suicidality, maintaining medication preconception and throughout pregnancy is typically recommended (Viguera et al., 2002; Yonkers et al., 2004). In clinical practice, however, medications with high teratogenic risk should be replaced with another mood stabilizer during preconception planning. Ideally, changing one mood stabilizer to another during pregnancy is best avoided to minimize fetal exposure to two sets of medication (see also Chapter 9).

The US Food and Drug Administration (FDA) has approved the use of some psychotropic agents in pregnant women. It is essential that the woman is stable prior to conception as pregnancy can escalate the prior symptomatology of the disorder. Once pregnant, careful monitoring and titrating is recommended. Medication is more difficult to titrate during childbearing as drug pharmacokinetics are affected by the physiological changes of pregnancy and delivery, particularly shifting fluid loads and altered drug clearance rates (Tomson et al., 2013). During parturition and the early postpartum period, the levels of some agents may increase rapidly with abrupt fluid losses; thus, intensive monitoring of maternal hydration status and drug levels is advised.

All psychotropic agents cross the placenta, and can affect fetal organ development/growth in utero and cause neonatal toxicity or withdrawal at birth in the absence of proper hydration (ACOG, 2008). Little is known about the impact of these drugs upon long-term behavior and development. There are specific periods of gestation during which drug exposure increases the risk of particular conditions, for example, exposure in weeks 3–8 can impact cardiac formation, weeks 6–9 can affect palate/lip formation, and up to day 32 can impact neural tube formation and closure (Starr & McMillan, 2011). Exposure in trimesters two and three has risks of low birth weight, preterm delivery, minor malformations and behavioral sequelae (Källén & Reis, 2012). Information on psychoactive drugs in pregnancy generally comes from retrospective cohort studies or case reports; there are few less-biased prospective studies or randomized controlled trials due to practical and ethical reasons. Therefore, prescribing decisions are made with regards to different aspects of perinatal pharmacotherapy (see also Chapters 9 and 11).

Lithium

Fetal and neonatal effects. Lithium is the gold standard for mood stabilization in pregnant women with BD (Viguera et al., 2002; Vieta & Valenti, 2013) as it is not as teratogenic as once perceived (Yatham et al., 2013), with most exposed infants being born without abnormality. However, the risk of Ebstein’s anomaly with lithium exposure continues to be a concern (Galbally et al., 2010). The rate of Ebstein’s anomaly in infants exposed to lithium during gestation is estimated at 1 in 1000 compared to 1 in 20,000 in the general population. Due to this heightened risk, high-resolution ultrasound and fetal echocardiograph are recommended at 16–18 weeks of gestation for women taking lithium in trimester one (Gentile, 2012). Many clinicians prescribe atypical antipsychotics as mood stabilizers if patients do not respond to lithium. The use of novel antipsychotics as a first-line treatment has been recently recommended for mood stabilization by The American Congress of Obstetricians and Gynecologists (ACOG, 2008; Reis & Källén, 2008); however, caution is advised as research is sparse and their efficacy and tolerability profile will be confirmed in the future.

Lithium exposure during delivery has the potential of neonatal toxicity, with hallmarks of “floppy baby syndrome” including neonatal cyanosis and hypotonicity (Galbally, Snellen et al., 2010). Nephrogenic diabetes insipidus and neonatal hypothyroidism have also been reported (Sands & Bichet, 2006; Feingold & Brown, 2010). Fetal thyroid goiter infrequently occurs with lithium exposure during trimesters two and three (American Academy of Pediatrics: Committee on Drugs, 2000) and can complicate delivery. Increased birth weight (by a mean of 80 g, compared to controls) has been reported in neonates with trimester one exposure (Bodén et al., 2012). Long-term neurobehavioral and developmental effects from lithium exposure in utero have been researched in follow-up studies, with no consistent negative sequelae described (Kozma, 2005; Yonkers et al., 2004).

Lithium therapy during pregnancy and delivery. Due to the short half-life of lithium (8–10 h), dosing is recommended 3–4 times/day to maintain optimum steady state levels (Yonkers et al., 2004). Adequate hydration, particularly during concurrent medical illness, is important to avoid lithium toxicity. Maternal polyuria and polydypsia during pregnancy can be aggravated by lithium (Misri & Lusskin, 2004; Nivoli et al., 2011). Lithium levels should be monitored monthly, with increasing frequency as pregnancy progresses and delivery approaches. The required dose typically increases with advancing pregnancy, due to greater drug clearance (Misri & Lusskin, 2004; Gentile, 2012; Yonkers et al., 2004; Burt et al., 2010). There appears to be no consensus regarding lithium use prior to delivery, with some advocating maintenance of the same dose, and others suggesting slow tapering and complete discontinuation. This decision should be made by the treating clinician on an individual basis. Lithium is recommended to be withheld during prolonged labor (1–2 days if necessary) and restarted at preconception levels immediately after delivery (Newport et al., 2005; Bogen et al., 2012). This decision must be balanced against risk of maternal relapse in labor or thereafter. Preserving the patient’s hydration status and intensive monitoring of maternal symptoms and lithium levels is recommended after delivery, because of the risk of toxicity or relapse during childbirth and the immediate postpartum (Gentile, 2012).

One-third of patients with BD do not respond to or tolerate lithium (AAP: Committee on Drugs, 2000) and may receive anticonvulsant therapy. The frequency of congenital anomalies remains high with anticonvulsant exposure (Gentile, 2010a; Gentile, 2010b).

Valproate

Fetal and neonatal effects. Valproate has the highest rate of fetal malformations at 8.7% as a single agent (Vajda et al., 2003). One of the primary concerns of valproate use in pregnancy is the 1–5% risk of neural tube defects (NTDs) (van Dijk et al., 2012). This dose-related risk coincides with the impact of drug exposure on days 17–30 of organogenesis. Polytherapy with carbamazepine can cause a rostral shift in the neural tube defect (Yonkers et al., 2004; van Dijk et al., 2012). Because valproate reduces folate levels (Linnebank et al., 2011), folic acid supplementation is recommended to reduce the risk of NTDs, although recommended amounts vary. Linnebank et al. (2011) recommend 5 mg/day for epileptic women taking anticonvulsants, from preconception until the completion of trimester one at minimum, but the American Academy of Pediatrics (AAP) recommends 4 mg/day before pregnancy and 400 μg after (AAP: Committee on Drugs, 2000). Measuring vitamin B₁₂ levels prior to folate administration is advised to rule out pernicious anemia (Linnebank et al., 2011). Prenatal testing for NTDs includes maternal serum α-fetoprotein, targeted ultrasound, and amniocentesis for α-fetoprotein levels.

DiLiberti et al. (1984) coined the term “fetal valproate syndrome” to describe the pattern of malformation of the “anticonvulsant face.” It is unclear whether intrauterine growth restriction and intellectual disability form part of this syndrome (Yonkers et al., 2004). Neonatal toxicity with heart rate decelerations and symptoms of withdrawal, which include irritability, jitteriness, abnormal tone and feeding problems have been described from valproate administration in pregnant women with epilepsy (Ebbesen et al., 2000). Intrauterine growth restriction, hyperbilirubinemia, skeletal dysplasia and fetal/neonatal distress have also been linked to valproic acid (AAP: Committee on Drugs, 2000). Additionally, the fetus is at an increased risk of spina bifida, atrial septal defect, cleft palate, hypospadias, polydactyly and craniosynostosis (Vajda et al., 2013; Jentink et al., 2010). Studies have found a high recurrence risk (21.9%) of major congenital malformation for women with more than one child (Pennell et al., 2012; Campbell et al., 2013 and see also Chapter 11).

A 2006 study found that valproate clearance increases with concurrent oral contraception use (Galimberti et al., 2006). The FDA has also recently issued a warning against valproate for migraine headaches in pregnant women, noting potential risk of decreased IQ scores in children and increased risk for autism spectrum disorders, childhood autism and reduced cognition (Christensen et al., 2013; Meador et al., 2013). Therefore, the FDA has reclassified valproate from a category D to a category X medication. Due to the high risk of malformations caused by valproate, there is a trend to stop its use in pregnancy and choose atypical antipsychotics instead (Wisner et al., 2011; NICE 2014).

Valproate therapy during pregnancy and delivery. Intensive serum level monitoring is required as amounts can change over the course of pregnancy. The total daily dose is a greater contributing factor to major malformations than peak serum levels. Given the multiple problems associated with valproate use in pregnancy, in clinical practice balancing between adverse effects and maintaining euthymia in the mother is often complex.

Carbamazepine

Fetal and neonatal effects. Like many other psychotropic agents, carbamazepine crosses the placenta and fetal serum levels have been approximated at 44% of maternal levels (Pynnönen & Sillanpää, 1975; Pynnönen et al., 1977). Carbamazepine use in pregnancy carries a risk of NTDs reported as 0.2–1.0% after first trimester exposure, compared to a population rate of 0.03% (Wieck & Gregoire, 2009; Matlow & Koren, 2012). Therefore, as with valproate, folate supplementation is advised. Facial malformation or the “anticonvulsant face” from carbamazepine (as valproate) exposure can also occur, with midface hypoplasia, small nose, anteverted nostrils and a long upper lip (Holmes et al., 2000; Yonkers et al., 2004). A greater rate of major congenital anomalies (2.2–7.9%) and lower birth weight of about 250 g with trimester one carbamazepine exposure has been reported (Nguyen et al., 2009; Diav-Citrin et al., 2001). Cardiovascular malformations, hypospadiasis, craniofacial defects and fingernail hypoplasia have all been described (Nguyen et al., 2009). Carbamazepine exposure also increases the risk of microcephaly (Almgren et al., 2009; Viguera et al., 2002) and thalidomide-like phocomelia (Dursun et al., 2012). There is debate as to whether carbamazepine exposure increases the risk of neurodevelopmental impairment (Cummings et al., 2011; Matlow & Koren, 2012).

Carbamazepine therapy during pregnancy delivery. Clinicians treating pregnant women with carbamazepine must be aware that this agent can result in fetal vitamin K deficiency. This can affect midface development and lower levels of vitamin K dependent clotting factors (Yonkers et al., 2004; Holmes et al., 2005; Lippi & Franchini, 2011; Kaaja et al., 2002), risking neonatal bleeding including intracerebral hemorrhage (AAP: Committee on Drugs, 2000; Kumar et al., 2012). However, there is insufficient evidence to determine whether prenatal vitamin K supplementation reduces neonatal hemorrhagic complications (Harden et al., 2009). Carbamazepine can influence the efficacy of certain anticoagulant drugs (Bauler et al., 2012). The AAP advises maternal treatment with 10–20 mg/day oral vitamin K for the final month of pregnancy (AAP: Committee on Drugs, 2000) and 1 mg intramuscular vitamin K for neonates (Yonkers et al., 2004; Kumar et al., 2012).

It is also important to note that efficacy of the oral contraceptive pill (OCP) can by affected by medications used to treat BD. Carbamazepine is known to induce cytochrome p450, lowering OCP levels (Arnold, 2003).

Lamotrigine

Fetal and neonatal effects. Teratogenic risk with lamotrigine is found at doses exceeding 200 mg/day. The maximum recommended dosage for BD I is 400 mg (Nguyen et al., 2009; Moore & Aggarwal, 2012). The FDA recently issued a warning that women treated with lamotrigine during the first 3 months of pregnancy had an increased risk of cleft lip and cleft palate. While the relative risk of these malformations is high, the absolute risk is low, at 0.89% (Shor et al., 2007). Greater malformation rates have been described with polytherapy, particular with valproic acid (Cunnington et al., 2011). Information on intrauterine growth following lamotrigine exposure in utero has not been described. In a study of development at 1 year, no abnormalities were noted (Mackay et al., 1997).

Experts warn about hepatotoxicity and skin rash, which has occurred in adults receiving lamotrigine (Yonkers et al., 2004; Sedky et al., 2012). A number of studies have shown that lamotrigine levels decrease with concurrent use of combined hormonal contraception methods (Christensen et al., 2007; Herzog et al., 2009; Wegner et al., 2009). Progesterone-only contraceptives, however, have not been found to reduce lamotrigine serum levels (Reimers, Helde, & Brodtkorb, 2005).

Lamotrigine therapy during pregnancy and delivery. Lamotrigine clearance increases during pregnancy and decreases postpartum to preconception levels (Pennell et al., 2008). Changing lamotrigine clearance highlights the importance of monitoring drug levels and dosing over pregnancy and particularly in the immediate postpartum period.

Topiramate

Recent studies show that there is a 4.6% increased risk for major birth defects with topiramate exposure and that it may cause long-term behavioral and cognitive problems (Mølgaard-Nielsen & Hviid, 2011; Rihtman, Parush, & Ornoy, 2012). There is an established connection between topiramate exposure during the first trimester and cleft lip and palate (Margulis et al., 2012; Hernández-Diaz et al., 2012). Consequently, the FDA has issued a warning against its use in pregnancy and has reclassified it as a Pregnancy Category D drug. There is also a high recurrence risk (50%) of major congenital malformations for women with subsequent pregnancies (Campbell et al., 2013). One study found that topiramate crosses the placenta at almost 100%. Additional studies have shown a significant pregnancy-related increase in dose/concentration ratios of topiramate (Ohman et al., 2009). Topiramate may impact upon the contraceptive efficacy of the OCP by reducing ethinyl estradiol levels (Reddy, 2010).

First generation antipsychotics

Fetal and neonatal effects. First generation antipsychotics are recommended as first-line treatment for manic, hypomanic or mixed episodes. Haloperidol is recommended when newer injectible atypical antipsychotics are not available, as it has not been found to increase the risk of congenital anomalies compared to low potency phenothiazines (Diav-Citrin et al., 2005; Owen, 2011; Reis & Källén, 2008). Neonatal toxicity has been described in the context of both withdrawal and extra-pyramidal symptoms with conventional antipsychotics, but it is unclear whether this is due to the disease or to anticholinergic or antihistaminergic properties of these agents. In addition, fetuses exposed to first generation antipsychotics have a slightly higher risk of preterm birth (Lin et al., 2010). Neurobehavioral effects are unclear (Gentile, 2010b).

First generation antipsychotic therapy during pregnancy and delivery. First generation antipsychotics have previously had several applications in pregnant women with BD, particularly to manage acute manic episodes. Their use over lithium or anticonvulsants in pregnancy/first trimester has been advocated because of lower fetal risk. Moreover, these medications were chosen for women who were unmedicated during their pregnancy and became symptomatic (ACOG, 2008). Haloperidol appears both safe and effective for the pregnant woman and her fetus for manic, hypomanic, mixed episodes or mood elevated syndromes (Diav-Citrin et al., 2005; Owen, 2011).

Second generation or novel antipsychotics

Fetal and neonatal effects. Second generation antipsychotics are not associated with any greater risk of fetal abnormality (McCauley-Elsom et al., 2010). However, they may increase the risk of gestational metabolic complications and increased birth weight compared to first generation antipsychotics (Gentile, 2010a). Olanzapine is approved for use in acute mania, but limited data are available for this agent during pregnancy. The manufacturer’s registry collected data prospectively (N = 144) and retrospectively (N = 98) on pregnant women taking olanzapine. No increase in malformation rates were detected (Einarson & Boskovic, 2009). The neonatal toxicity of olanzapine is currently unknown. Quetiapine was approved by the US FDA for treating acute mania but it has now been indicated for other disorders such as bipolar depression, schizophrenia, post-traumatic stress disorder, insomnia and depressive symptoms (Ribolsi, Magni, & Rubino, 2010; Byers et al., 2010; Tadger, et al, 2011; Coe & Hong, 2012). According to the manufacturers, 1.5% of pregnant women taking quetiapine reported congenital anomalies, with the majority of affected women having taken multiple medications during pregnancy. Risperidone is another novel antipsychotic used in BD treatment but data on its use in pregnancy is also limited. The manufacturer has confirmed that rates of major malformations are within rates for the general population, based on 68 cases with a known outcome (Einarson & Boskovic, 2009).

With regard to clozapine, a few case reports found no major malformations. The manufacturer’s registry noted that of 523 babies exposed to clozapine in utero, 22 had unspecified malformations. White cell counts are required in infants of mothers on clozapine, as agranulocytosis can occur in adult patients receiving this medication (Lahdelma, 2012). Case reports have shown conflicting data that ziprasidone may be a possible cause of cleft palate (Ružić et al., 2009). Few studies have evaluated the safety of aripiprazole in pregnancy, but no adverse events have been reported thus far (Einarson & Boskovic, 2009).

Second generation antipsychotic therapy during pregnancy and delivery. Second generation antipsychotics are often prescribed in pregnancy and are preferable to lithium and anticonvulsants. Experts recommend close monitoring of blood pressure, blood sugar levels and weight increases in pregnant women taking olanzapine because data have linked it to weight gain, insulin resistance, gestational diabetes and preeclampsia (Yonkers et al., 2004). It is important to note that the prolactin-sparing qualities of some novel antipsychotics can lead to an increased risk of conception when changing between an older agent and a prolactin-sparing antipsychotic (Green et al., 2013).

Benzodiazepines

Benzodiazepines (BDPs) are not primary agents for the treatment of BD but are implemented to alleviate associated symptoms such as anxiety, agitation and sleeping difficulties (Yonkers et al., 2004). Pregnant women may also be exposed to BDPs when required for medical reasons such as preterm labor.

Fetal and neonatal effects. Teratogenic effects have not been described for lorazepam or clonazepam (which are most often used in BD). One study has found that diazepam may be associated with an increased risk for cleft lip/palate (Kjaer et al., 2007). Many women become unexpectedly pregnant on BDPs. Therefore, it is important to advise women on BDPs regarding the safety of these agents. Case-control studies show a small increased risk of the development of major malformations or oral cleft with BDP use (ACOG, 2008; Enato, Moretti, & Koren, 2011). A level two ultrasound is now recommended when BDPs are used in the first trimester to exclude visible oral cleft (ACOG, 2008), which can be corrected at birth. BDPs should only be used when necessary in the first trimester (Bellantuono et al., 2013). Cases of neonatal toxicity from BDP exposure in third trimester or at delivery have been reported (ACOG, 2008). Withdrawal symptoms (infant irritability, tremor, vomiting and diarrhea, hypertonicity and vigorous sucking) have also been reported from chronic maternal BDP use; however, another report of maternal clonazepam use for panic disorder did not report neonatal toxicity (Hudak et al., 2012; Yonkers et al., 2004). The risk of long-term developmental delay is unclear, as data are limited and existing studies have produced contradictory results (ACOG, 2008).

Benzodiazepine therapy during pregnancy and delivery. In pregnant women, high-potency BDPs may be the best option due to shorter half-life, decreased chance of sedation and reduced accumulation (Yonkers et al., 2004).

Electroconvulsive therapy

Once contraindicated in pregnancy, electroconvulsive therapy (ECT) now plays an important role in the management of acutely psychotic pregnant women with BD. ECT is considered a relatively safe and effective treatment (Nielsen & Damkier, 2012), and should be undertaken by a multidisciplinary team of obstetricians, anesthetists and psychiatrists. A number of studies also provide support for the benefits of ECT in the treatment of postpartum psychosis (Doucet et al., 2011). Recent studies report only transient side effects to postpartum women and no adverse effects for the infants who were breastfed (Babu, Thippeswamy, & Chandra, 2013). A 2009 review found 18 ECT-related maternal complications out of 339 mothers (5.3%), which included vaginal bleeding, uterine contractions and/or preterm labor. Preterm labor was found to be the most common complication (Anderson & Reti, 2009). Although evidence suggests that the use of ECT is safe and effective in women who require it urgently, caution should be exercised to prevent transient fetal effects including bradyarrhythmia. Maternal risks, which are infrequently experienced, can be minimized in the hands of expert obstetricians and anesthetists. Risks to the mother and fetus can also be minimized by adequate oxygenation, right hip elevation, and avoiding both atropine and hyperventilation (Yonkers et al., 2004).

Postpartum ECT risks are considered minimal, but breastfeeding is not recommended for several hours after the ECT because of co-administered pharmacotherapy (Focht & Kellner, 2012). There are important ethical issues when ECT is used in pregnant patients, as obtaining informed consent for a woman and her unborn child can be difficult if a patient’s mental state adversely affects her understanding of the situation and capacity to consent to ECT. Furthermore, the physician is treating two patients and must consider the health of both the mother and the unborn child.

Treatment of postpartum psychosis

The acute treatment of postpartum psychosis typically involves hospital admission and antipsychotic medication. The excess risk of postpartum psychosis in women with BD means that experts recommend administering mood stabilizers with antipsychotics and the careful use of antidepressants because of the risk of rapid cycling (Chaudron & Pies 2003). Cohen and colleagues (1995) found that women with known BD who did not receive a mood stabilizer had 8.6 times greater relative risk of postpartum relapse in the first 3 months after childbirth than those who received pharmacotherapy (mostly lithium). Bergink and colleagues (2012) also found that postpartum pharmacotherapy reduced the rate of psychotic relapse in women with a history of postpartum psychosis, relative to control patients. The use of ECT in patients with acute postpartum psychosis is recommended in situations as previously discussed. A rising amount of data shows ECT as a safe and effective treatment option for postpartum psychosis (Focht & Kellner, 2012).

Breastfeeding and medication

For women with BD, the decision of whether or not to breastfeed is complex. The health benefits of breastfeeding for both mother and child are well-documented and publicly promoted. However, all psychotropic medications pass into the breastmilk and infant serum in varying amounts. Therefore, breastfeeding is decided on a case-by-case basis.

Numerous factors may influence a woman’s decision whether or not to breastfeed. These include previous episodes, a family history of postpartum psychosis, the mother’s desire to breastfeed, risks and benefits to the neonate, and maternal risks from breastfeeding (Chaudron, 2000). Breastfeeding involves sleep disruption throughout the night; generally, sleep deprivation is one of the triggers of inducing a manic episode. Therefore, many women give up nursing in order to maintain mood stability. Providing patients with information about the risks and benefits of breastfeeding and unknown risks such as long-term sequelae is advised (Chaudron & Jefferson, 2000). In women who choose to breastfeed and require mood stabilization, the principles governing drug administration are similar to those used in pregnancy, such as monotherapy where possible, using the lowest therapeutic dose and selecting agents based on the current condition and past responses to medication. (See also Chapter 11.)

Lithium. Postpartum lithium prophylaxis has been reported to decrease the relapse rate in women with BD from about 50% to 10% (Yatham et al., 2005). Lithium should be administered to breastfeeding women with caution, as infant blood concentrations that are 11–56% of therapeutic levels have been reported (Viguera et al., 2007). Studies have found breastmilk lithium concentrations to be about 40% (range 24–72%) and infant levels to be 5–200% of maternal levels (Yatham et al., 2005; Burt & Rasgon, 2004; Yonkers et al., 2004; Ernst & Goldberg, 2002). If women do choose to breastfeed on lithium, lithium concentrations should be monitored closely in both the mother and baby. When it is not possible to measure infant levels, the mother should be advised to observe the baby and report to her healthcare provider when required. In these instances, whether to continue or discontinue breastfeeding will depend on the mother making an informed decision. The long-term impact of extended lithium exposure has not been determined, but limited data indicates no obvious developmental abnormalities (Grandjean & Aubry, 2009).

Anticonvulsants. Both valproic acid and carbamazepine were considered “usually compatible with breastfeeding” by the AAP: Committee on Drugs (2000). In the Reproductive Mental Health Program, our preference is to not use valproate or carbamazepine while breastfeeding due to negative data related to pregnancy.

Valproate. Valproate levels in breastmilk have been reported as 1–10% of maternal serum levels (Tettenborn, 2006; Pennell, 2003), with infant levels varying from undetectable to 40% of maternal levels (Tettenborn, 2006). No adverse effects were reported in several studies involving breastfed infants whose mothers received valproate for BD (Piontek et al., 2000; Wisner & Perel, 1998; Johannessen, Helde, & Brodtkorb, 2005). However, a case of thrombocytopenic purpura, anemia, and reticulocytosis has been described in a breastfed 3-month-old infant whose mother received valproic acid for epilepsy; these symptoms resolved with the cessation of breastfeeding (Stahl et al., 1997). Hepatotoxicity has been reported in children less than 24 months receiving valproate directly (not through breastmilk). In a 2010 study, the IQ of 3-year old children exposed to valproate from breastmilk was measured, and no long-term intelligence deficits were found (Meador et al., 2010). However, due to evidence regarding neurodevelopmental effects of in utero exposure to valproate (Vinten et al., 2005; Eriksson et al., 2005; Gaily et al., 2004; Adab et al., 2001), informed consent should include the possible neurodevelopmental effects from exposure in breastmilk.

Carbamazepine. Relatively high levels of carbamazepine are found in breastmilk, with reports ranging from 32–79% of maternal serum levels (Shimoyama, et al, 2000; Davanzo et al., 2013). Infant serum levels vary from 6 to 65% of maternal concentrations. Notably, in most cases exposure has occurred in utero (Stowe, 2007). Case reports also indicate transient infant hepatotoxicity (Frey et al., 2002).

Lamotrigine. Like other psychoactive agents, lamotrigine crosses over into the breastmilk. A large range of infant levels of lamotrigine from breastmilk exposure is noted due to the variability of studies. Case reports show infant serum levels at approximately 30% of maternal levels (Liporace, et al, 2004; Rubin, et al, 2004). Careful monitoring of infants for rashes is advised, as children receiving lamotrigine are at a heightened risk of life-threatening skin reactions (Guerrini et al., 2012). Another study has shown that exclusively breastfed infants would receive approximately 9% of the maternal weight-adjusted dosage (Page-Sharp et al., 2006). The AAP classifies lamotrigine as an agent of unknown effect, which may be of concern because of potentially therapeutic serum levels in infants (AAP: Committee on Drugs, 2000); therefore, lamotrigine is not recommended while breastfeeding.

Topiramate. In a study of five mother-infant pairs in women with epilepsy taking topiramate (who all concomitantly received either valproic acid or carbamazepine), three of the pairs were breastfed. Topiramate was found to be excreted into the breastmilk between 3–23% of the mother’s weight-adjusted dose. Infant levels were about 10–20% of maternal serum levels and adverse effects were not reported in the infants (Ohman et al., 2002). These levels are confirmed by more recent studies (Froscher & Jurges, 2006). However, more data on this agent is warranted.