Abstract
The aim of the present study was to develop recommendations for antenatal care and monitoring for women with bipolar disorder and schizophrenia who are on lithium carbonate, antipsychotic or anti-epileptic medication during pregnancy. A literature search and review of original research, published reviews and guidelines was undertaken for mood stabilizers and antipsychotics in pregnancy and for the management of bipolar disorder and schizophrenia in pregnancy. This information was summarized, condensed and then reviewed by representatives of psychiatry, pharmacy, paediatrics and obstetrics to produce an information booklet and subsequently monitoring recommendations and tables. A model of antenatal monitoring and care for women with schizophrenia, bipolar disorder and related disorders who are maintained on psychotropic medication was developed. This included an online and published booklet for clinicians summarizing psychotropic medication in pregnancy, and lactation and monitoring tables that could be part of patient case files. These were to assist in reminding and educating staff on the need for additional monitoring and assessment above standard antenatal care for women on mood stabilizers and antipsychotic medication. Women with bipolar disorder and schizophrenia have an increased risk of complications in pregnancy from their illness and from the medications they are prescribed. A summary of the risks and a model of suggested additional monitoring during pregnancy have been developed in consultation across a number of clinical disciplines.
The antenatal management of women with bipolar disorder and schizophrenia is a challenge for both obstetric and mental health services because both the illness and the medications used to treat pose significant risk for pregnancy. Both illnesses are associated with increased risk of a number of pregnancy complications such as antepartum haemorrhage, prematurity, low birthweight, and intrauterine growth retardation [1–4]. In addition the medications that are used to treat these illnesses (antipsychotic and mood stabilizers medications) are associated with complications such as teratogenicity (e.g. cardiac malformations, neural tube defects, facial deformity), pregnancy complications (e.g. gestational diabetes, polyhydramnios, fet al macrosomia), neonatal complications (e.g. neonatal toxicity and withdrawal effects, neonatal hypothyroidism, hypotonia) and longer term adverse neurodevelopmental effects [5–8]. Given the significant potential pregnancy and neonatal risks associated with exposure to these medications, there is the need for careful and close monitoring in an appropriate multidisciplinary antenatal setting to ensure best outcomes.
In response to these concerns, many women are advised to cease or reduce their dose of medications, but this significantly increases their risk of relapse. For instance, those women with bipolar disorder who discontinue a mood stabilizer are more than twice as likely to relapse in pregnancy (85.5% vs 37%) and are ill for fivefold longer during their pregnancy than those who continue their medication [9]. Even among women continuing their medication in pregnancy, the pregnancy-related changes in pharmacokinetics may render pre-pregnancy doses of medication subtherapeutic due to changes in bodyweight, drug absorption, metabolism and excretion. For instance, plasma levels of medications such as lithium carbonate, which are renally excreted and fluid volume dependent, reduce significantly during the third trimester due to the increased volume of distribution. To use a strategy of reducing the dose further significantly increases the risk of relapse, with previously effective medications rendered subtherapeutic.
Despite the documented increased risks there has not been a strong development of specific expertise in obstetrics around the antenatal management of women who suffer from major mental illness and the medications they are prescribed. Therefore, guidelines both in Australia and overseas have tended to come from mental health clinicians rather than obstetrics. It is imperative, however, that guidelines that involve recommendations about antenatal obstetric care come from joint initiatives between obstetrics and mental health professionals for this to be practical and capable of implementation.
In a new initiative at the Mercy Hospital for Women, the Division of Obstetrics and Maternity Services has established a multi-disciplinary clinic for women with bipolar disorder and schizophrenia that offers coordinated care and access to a specialist perinatal psychiatrist, consultant obstetrician, a neonatal paediatrician and specialist physicians as well as liaison with preexisting community services involved in these women's care. It is hoped that this will also allow the training of obstetric and psychiatric registrars with specific expertise in this important area of health care.
As part of the establishment of this clinic the available research and guidelines were reviewed and recommendations were developed for optimal antenatal care. Summarized information was made available for clinicians regarding psychotropic medication in pregnancy and lactation. A set of guidelines and tables for improved monitoring in pregnancy for the specific conditions and medications were also developed to further streamline the management of women with bipolar disorder, schizophrenia and related disorders in pregnancy.
Method
A review of published research, review articles and guidelines developed for the psychiatric and pharmacological management of women with bipolar disorder and schizophrenia was undertaken. This included a computerized search of the electronic databases Medline, EMBASE and PsychInfo for articles of original research that were relevant to clinical assessment of risk and care in pregnancy for schizophrenia, bipolar, mood stabilizers (sodium valproate, carbamazepine, lamotrigine, lithium carbonate) and antipsychotic medications (typical and atypical). Searches were also undertaken of relevant guidelines, reference books and bibliographies of identified articles. The key findings and recommendations were then condensed into a summary for general practitioners (GPs) and obstetricians and published as a booklet and online [10]. The specific monitoring recommendations related to lithium carbonate, antipsychotic and mood stabilizer medications in pregnancy were summarized into three tables that would be suitable for use in a patient's clinical file (Tables 1–3). The development of both the booklet and tables involved a process of modification and discussion by psychiatric, obstetric, neonatalogy and pharmacy staff as relevant to each section. The three pregnancy monitoring tables for use in patient clinical files were for anti-epileptic drugs (AEDs), lithium carbonate and antipsychotics.
Mercy Hospital for Women lithium carbonate: pregnancy monitoring
FBE, full blood examination; FT3, Free Thyroxine T3; FT4, Free Thyroxine T4; LFT, liver function tests; NT, nuchal translucency;TFT, thyroid function tests; TSH, thyroid-stimulating hormone; U&E, urea and electrolytes.
Mercy Hospital for Women anti-epileptic mood stabilizers: pregnancy monitoring
FBE, full blood examination; FT3, Free Thyroxine T3; FT4, Free Thyroxine T4; LFT, liver function tests; NT, nuchal translucency; TFT, thyroid function tests; TSH, thyroid-stimulating hormone; U&E, urea and electrolytes.
Mercy Hospital for Women antipsychotics: pregnancy monitoring
BMI, body mass index; BP, blood pressure; ECG, electrocardiography; FBE, full blood examination; GTT, Glucose Tolerance Test; LFT, liver function tests; NT, nuchal translucency; TFT, thyroid function tests; U&E, urea and electrolytes
Results
The key findings and recommendations developed within the clinic around antenatal care are outlined within the six areas of considerations for management in pregnancy. These are preconception, teratogenic risk, pregnancy, delivery, neonatal and postnatal care.
Preconception
Where appropriate there should be consideration of pre-pregnancy modification of treatment regimens to minimize teratogenic risk and to ensure optimal pre-pregnancy physical and mental health. If treatment regimens are to be altered a discussion with women around adequate contraception is important given the high rate of unplanned pregnancies. Carbamazepine and sodium valproate are known to interfere with folate metabolism, therefore to minimize teratogenic risk, 5 mg of folic acid is recommended 1 month prior to and during pregnancy. The pregnancy and delivery risks for women with bipolar disorder and schizophrenia include those due to the illness itself, such as the risk of poor self-care and inadequate nutrition. Therefore, consideration should be made of an extensive organic screen prior to conception or early in pregnancy with tests such as full blood examination, urea and electrolytes, liver function tests, thyroid function tests, vitamin B12, folate, iron studies and vitamin D. The last could be considered a contentious inclusion, but a recent Australian study has found higher levels of vitamin D deficiency within patients admitted for psychiatric illness than controls, and female patients were found to have lower vitamin D levels than male patients [11]. In recent studies there has also been found an association between maternal vitamin D deficiency and abnormal brain development, and also a possible increased risk of subsequent mental illness in offspring [12–14]. For women on atypical antipsychotics there should be a recent evaluation of glucose tolerance. Where there is associated obesity an assessment for poor nutrition and folate deficiency, both of which can increase risk of pregnancy complications and neural tube defects, is recommended.
Consideration should be given to where women attend for their antenatal care and delivery to ensure, where possible, it is at a hospital that has facilities that are able to manage the pregnancy and neonatal risks associated with these conditions and medications.
Teratogenic risk
There is a well-recognized increase in teratogenic risk associated with the use of sodium valproate, carbamazepine, lamotrigine and lithium [15]. There is a paucity of data regarding the teratogenic risk of antipsychotics, but the limited data available would suggest that the risk is not greatly increased above background [16,17]. The data regarding the malformation risk for mood stabilizers such as sodium valproate, carbamazepine and lamotrigine, are mostly drawn from AED registries [8,18]. A recent review of prospective registry data suggests a malformation rate of 2.7% and 2.9% for lamotrigine and carbamazepine, respectively [19]. As a single agent, sodium valproate has been consistently shown to have the highest rate of congenital malformations (8.7%), with several investigators reporting a dose–response relationship with a greater risk when doses exceed 1 g [18–21]. The majority of these structural abnormalities are central nervous system anomalies, particularly neural tube defects, as well as cardiac, facial and urogenital malformations. In addition to the increased risk of structural abnormalities, there has been more recent interest in the potential for these AEDs to negatively impact on infant neurocognitive development [22–24]. Again, valproate and polytherapy appear to be associated with the highest risk, but more data are awaited [8,24–26]. It needs to be remembered that other confounders may play a role in both structural malformation and neurocognitive development in women receiving these medications as AEDs. Most of the research to date on the AEDs has been in those women with epilepsy rather than bipolar disorder. There may be an independent effect of the seizures themselves as well as a genetic predisposition; these considerations may not apply to the same extent in women receiving AEDs for mood disorders.
Lithium is associated with an increased risk of cardiovascular malformation, specifically, Ebstein's anomaly (0.05–0.1%), although the relative increase in risk is lower than previously thought [6,27–30]. Beyond the first trimester, lithium exposure is also associated with an increased risk of diabetes insipidus, polyhydramnios, thyroid dysfunction and floppy baby syndrome [31].
There is limited information on the safety of atypical antipsychotics in pregnancy. Although there does not appear to be a demonstrable increase in the risk of teratogenicity over background rate, further data are awaited [32,33]. A recent review identified only seven studies on pregnancy outcomes and antipsychotic exposure in women with psychiatric disorders although none showed a clear association with any specific malformation [17]. This review also only identified one study which examined long term neurodevelopmental outcomes following exposure to typical antipsychotics in utero and none for atypical antipsychotics [17]. The one study identified found no differences in outcomes for 4 year old children compared to controls [17].
Pregnancy care
In women receiving sodium valproate, lamotrigine or carbamazepine, an 11–13 week ultrasound examination should be offered. Acrania (the precursor of anencephaly) should be recognized at this gestation, and the nuchal translucency measurement is a useful screening test for cardiac and other structural malformations [34]. Among women receiving lithium, a preliminary assessment of the fetal heart may be performed at 16 weeks’ gestation, but even if this is normal a detailed echocardiography should still be performed with the anatomical survey at the mid-trimester (18–22 weeks) scan. Among women with bipolar disorder or schizophrenia, the referral for the mid trimester morphology scan should include detail on medication use so that a targeted examination, with particular attention to the neural axis, heart and face, can be performed. Where there are adequate resources for expert mid-trimester morphological assessment using ultrasound, the value of screening for neural tube defects with maternal serum a-feto-protein is limited. While serum screening will identify nearly all pregnancies with anencephaly, the detection rate for neural tube defects overall is 65%, rising to 85% if gestational age is confirmed on ultra-sound. By contrast, detection rates of open neural tube defects with high-level ultrasound range from 97% to 100% [35].
Therapeutic drug monitoring is necessary for women receiving medications such as lithium, carbamazepine, sodium valproate and lamotrigine. Changes in gastric emptying, increased volume of distribution, decreased gastrointestinal motility, decreased drug-binding capacity and increased hepatic metabolism during pregnancy may alter the therapeutic dose required. Lithium levels in particular require careful monitoring during the third trimester, to ensure that they remain in the therapeutic range. This should include monthly serum lithium estimations throughout pregnancy, increasing to weekly after 36 weeks gestation. Serum levels are also recommended for women taking sodium valproate, carbamazepine and lamotrigine due to the decrease in serum concentration of these drugs during pregnancy. Although it may be sufficient to perform serum levels of sodium valproate and carbamazepine each trimester among women with stable disease and dosage without side-effects, monthly checks of lamotrigine levels are recommended because highly variable and more clinically significant fluctuations in drug concentrations have been observed [36].
Other issues to consider in pregnancy include the need for increased fetal growth surveillance given that schizophrenia, bipolar disorder and all three groups of medications can potentially alter fetal growth during the second half of pregnancy. Antipsychotic medications have been associated with both low-birthweight babies and large-for-date babies, the latter mainly with exposure to the atypical antipsychotics [37]. Given the potential diabetogenic effects of atypical antipsychotics, it is suggested that in patients receiving antipsychotics, a glucose tolerance test, rather than glucose challenge test, should be performed at 28 weeks gestation, and a glucose tolerance test early in pregnancy (14–16 weeks’ gestation).
Administration of supplemental vitamin K to mothers receiving enzyme inducers, such as carbamazepine and sodium valproate, is no longer routine, but Konakion must be administered to all newborns to minimize the risk of neonatal bleeding [38].
Delivery planning
Women on lithium in pregnancy need to have developed a delivery plan that is communicated to obstetric staff [39]. This should include ceasing lithium 24–48 h prior to delivery or at delivery, a lithium level when admitted for delivery, ensuring adequate hydration during labour and delivery, including consideration of i.v. hydration if necessary. When possible, administration of potential nephrotoxins should be minimized or avoided, including amnionoglycosides and non-steroidal anti-inflammatory drugs. Where renal throughput is compromised (such as in pre-eclampsia or haemorrhage), the dose should also be lowered or withheld. Angiotensin-converting enzyme inhibitors are not used in pregnancy due to their potential teratogenesis, but calcium channel blockers are routinely used in preterm labour. Other medications should be given as usual during the peripartum period, although dose adjustments after delivery are likely.
In view of the risks of neonatal complications, women should deliver where there are adequate facilities for neonatal resuscitation. There should be liaison with paediatrics staff to ensure assessment and monitoring of the neonate after delivery for complications associated with in utero exposure. If there are concerns about the possibility of fetal toxicity then a cord blood lithium level should be estimated at delivery. The mother also requires clear information around recommendations not to breast-feed while taking lithium, but pharmacological suppression is contraindicated due to the small increased risk of psychotic relapse. Adequate communication to ensure appropriate postnatal staff support is of paramount importance. As part of the delivery plan there should also be included information to ensure that lithium is restarted at the pre-pregnancy dose once the patient is medically stable.
Neonatal management
Lithium has been associated with neonatal complications including floppy baby syndrome, cardiac dysfunction, diabetes insipidus, hypothyroidism, low muscle tone, lethargy, hepatic abnormalities, and respiratory difficulties. Sodium valproate has been associated with a withdrawal syndrome, hepatic toxicity and hypoglycaemia. Carbamazepine and lamotrigine have both been associated with neonatal hepatic toxicity. Therefore, adequate assessment and monitoring through paediatrics should be instigated as a matter of course. It is not clear what, if any, are the potential neonatal complications with atypical antipsychotics. Typical antipsychotics, however, have been associated with transient extra-pyramidal side-effects in the neonate, and babies should always be monitored for sedation and withdrawal symptoms.
All psychotropic medications are excreted in breast milk, but the relative infant doses of sodium valproate (0.68%) and carbamazepine (4.3%) are low, and these medications are regarded as compatible with breast-feeding [40]. Lamotrigine has a relatively high excretion with a relative infant dose of 22.7% [40], and the product information for all these drugs cautions that they should be used during breast-feeding if the benefits outweigh the risks. As such, these decisions need to be made on a case-by-case basis. Breast-feeding is not recommended with lithium because breast milk contains 30–50% of maternal serum level and, combined with immature neonatal renal function, leads to a high risk of toxicity [40], but a recent study of 10 infants, however, did find low lithium levels while being breast-fed [41].
Safety data regarding atypical antipsychotics are extremely limited [42]. Data to date suggest that breast milk levels are low, with relative infant doses generally <3% of maternal dose. The infant should be observed for sedation because this may also result in poor feeding and low weight gain. For clozapine the adverse pharmacological and toxicological effects that occur in adults may also occur in the fetus/neonate, including a risk of agranulocytosis in the fetus/neonate, and therefore if breast-feeding is considered the infant must be monitored accordingly. Breast milk excretion for clozapine is low, with a relative infant dose of 1.2% [40]. Given the very limited evidence available for all the antipsychotic medications, decisions around breast-feeding should be made on a case-by-case basis.
Postpartum care: hospital and beyond
The postpartum period comes with a high risk of maternal relapse in addition to the medication-related risks for both the neonate and the mother [9,31,43,44]. Any monitoring or management plan should include regular assessments and ongoing monitoring of the woman's mental state through this high-risk period, with a clear management plan, should she relapse antenatally or during the postpartum period. Adequate psychosocial support in addition to pharmacological treatment is important for the mother both during pregnancy and the postpartum period. It is recommended that the woman and, if appropriate, partner/ family supports, where possible, meet with the maternal child health nurse prior to delivery, consider extending paternity leave, and the avoidance of potential stressors during pregnancy and the early postpartum period, such as home renovations, house move, excessive visitors and so on. The importance of adequate postnatal social supports cannot be overemphasized.
Given the high risk of relapse and the potential for exacerbation with sleep deprivation, it is important for these women to avoid sleep deprivation during the initial postpartum period. Therefore clinicians should have a low threshold to use hypnotics, such as temazepam, to assist maternal sleep and consider while mother and baby are in patients nursery care of the baby overnight to minimise potential sleep deprivation. After delivery a low-stimulus environment may also be of benefit, and this includes the limitation of the number of visitors and being placed in a single room when possible. There may also be benefit in allowing a woman a longer stay in hospital after delivery to ensure that there is adequate support and monitoring in the early postpartum period. A clear and comprehensive management plan for monitoring for relapse, pharmacological treatment and psychosocial support for the postpartum period should be in place prior to discharge, and communicated to all those involved.
Discharge planning should also include the mother–infant relationship and, ideally, support for the partner or family involved in supporting the mother–infant dyad. Mothers, particularly with schizophrenia, have been found to have more compromised interactions with their infants, characterized by less sensitivity towards the infant and more intrusiveness [45–47]. This is more pronounced in those with active symptoms compared to those who are well. Sensitivity of parenting is a predictor of attachment in infants and this has been related to a range of child and adult mental health outcomes [48]. There are few specific models or services that focus on mother–infant interventions for women with illnesses such as bipolar disorder or schizophrenia [47,49]. A recent review identified no studies that examined interventions specifically for mothers with psychosis and their infants, but that paper did discuss sensitivity-focused behavioural techniques that may be useful for these mothers and infants [49]. Consideration of services that support mothers and infants in early parenting may therefore be useful. In addition the mother's main supports: partner and/or family, are an important part of a comprehensive discharge plan for both the mother and infant.
Monitoring procedures
The information on suggested antenatal monitoring for pregnancy and neonatal effects of antipsychotic, anti-epileptics mood stabilizers and lithium carbonate developed are given in Tables 1–3.
Summarized information on the specific risks associated with these identified medications that required additional antenatal monitoring and assessment was made available to clinicians in a published booklet [10] and online (http://www.mercy.com.au/files/NRR6CEQQCO/Psychotropic%20drugs%20%20pregnancy%202nd%20Edn.pdf).
Discussion
Mood stabilizers and antipsychotics are established and effective treatments that assist in the prevention of relapse in bipolar disorder and schizophrenia, and may also be used in the treatment and prevention of puerperal psychosis [50,51]. The recent improvement in the range of effective treatments both for acute episode and relapse prevention has revolutionized the management and lives of those suffering from these illnesses. For these women this means the capacity to form stable partnerships and the consideration of children. Like all women in our community, however, there is also a risk of unplanned pregnancy. It is estimated that >50% of pregnancies are unplanned and many women do not realize that they are pregnant until the end of the first trimester or later [52]. A study of New Zealand mental health professionals, however, found that they discussed family planning with only 17% of their female patients [53]. Given the possibility of an unplanned pregnancy, the consideration and discussion of pregnancy risks of medications prescribed in women of child-bearing age should become standard practice.
Optimal antenatal management of women with schizophrenia and bipolar disorder involves the formation of a close liaison between the treating psychiatrist/mental health services, obstetricians, GPs, maternal child health nurses and neonatal paediatricians. A management plan should be formulated that addresses all contingencies in unison with the patient and is recorded in a prominent place in the case notes. This will assist in ensuring that mothers with a significant mental illness are monitored carefully throughout both pregnancy and the early postpartum period. These suggestions and recommendations are intended to assist clinicians in providing optimal care to women and their families with bipolar disorder, schizophrenia and also related disorders such as schizoaffective disorder and puerperal psychosis.
Footnotes
Acknowledgements
The authors thank Elizabeth McCarthy, Alison Fung, Paul Drinkwater, Gordon Spalding, Courtney Munro, Gillian Opie, Andrew Watkins, Klara Szego, Anne Buist, Adaobi Udechuku.