“Ms. A” was a 38-year-old married woman with a history of psychosis who sought consultation about the safety of antipsychotic use during pregnancy and lactation. Ms. A had done well taking quetiapine for 5 years. A few years earlier, she had undergone an abortion after being told that her medication was not compatible with pregnancy. Subsequently, she and her husband sought consultation at a university hospital.
Ms. A had a history of psychiatric symptoms dating back to her teens. In high school, Ms. A remembered having depression, anorexia, and suicidal ideation. A music teacher at the time suggested to her parents that they seek psychiatric care. Her parents ignored the advice.
In college, Ms. A became bulimic, and her depression worsened. She attempted suicide three times in her early 20s. At age 24, her weight fell to 90 lb, and she was severely depressed and developed delusions about communicating with God. She was initially diagnosed with bipolar disorder and treated with lithium and typical antipsychotics, including haloperidol and thiothixene. The addition of fluoxetine at one time caused such severe agitation that she again attempted suicide. She was diagnosed with schizophrenia after her final suicide attempt, which occurred when Ms. A thought she was the devil and that the world would be cured of evil if she died.
In 1995, Ms. A moved and began seeing a psychiatrist who switched her to atypical antipsychotics. Ms. A did better while taking olanzapine but had significant residual symptoms, including depression, prominent negative thinking, psychomotor retardation, and paranoia. She was later switched to quetiapine, during which time she began to feel “normal.”
In 2000, Ms. A became pregnant while taking quetiapine but was told to have an abortion because of exposure of the fetus to medication. She and her husband were not yet married and decided for both reasons to not bear the child. Subsequently, several years after they married, Ms. A was referred by her psychiatrist to a consultant more familiar with the use of medications during pregnancy.
Ms. A’s medical history was unremarkable. She was an otherwise healthy Asian American woman with regular menstrual cycles. She was a nonsmoker who did not use alcohol or illicit drugs. She had used oral contraceptives for contraception.
The initial psychiatric evaluation consisted of a detailed psychiatric and medical history followed by a discussion of the risks and benefits of medication use during pregnancy and the postpartum. The first consultant reviewed three possible courses of treatment: 1) discontinuation of medications before pregnancy, 2) use of medications until Ms. A had a positive pregnancy test and subsequent withdrawal of medication, and 3) use of antipsychotics throughout the pregnancy. Given Ms. A’s severe history of psychosis and suicide attempts, she and her husband agreed with the consultant that medications should be continued throughout pregnancy and possibly during lactation.
The consultant then reviewed the published literature on antipsychotics that existed at that time. Olanzapine had the most cases documented in the literature, and the consultant used this as a rationale for suggesting that Ms. A switch to olanzapine. Ms. A was titrated off of quetiapine and began taking olanzapine.
Within 3 months, Ms. A had gained 20 lb. She was depressed, lethargic, anhedonic, and had severe memory and concentration difficulties. She slept at least 14 hours each day and found it increasingly difficult to work. Her outlook was extremely negative, and she ruminated about many kinds of potential adversities—e.g., the death of her husband. She had recently developed passive suicidal ideation and, as a result, her treating psychiatrist asked Ms. A to have a second consultation with a different psychiatrist who was also knowledgeable about the use of medications during pregnancy.
The second psychiatrist met with Ms. A and her husband. Ms. A was overweight and appeared to have psychomotor retardation and to be lethargic. Her thinking was concrete, and her self-esteem was diminished. She had paranoid ideation and worsening interpersonal relationships at work. Her husband felt that Ms. A’s clinical condition had declined considerably since the switch to olanzapine.
The second consultant suggested that Ms. A’s mental health was essential to establishing and maintaining a pregnancy as well as to her functioning as a mother. Given her poor response to olanzapine and her previous excellent one to quetiapine, this consultant suggested that Ms. A go back to taking quetiapine. The consultant pointed out that the original decision to take olanzapine was based on a small number of cases reported by a manufacturer’s registry. At the time of the second consultation, there were two case reports of normal pregnancies while the mothers were taking quetiapine. The consultant said that in light of limited information about the risks of each medication, the decision needed to be based on the mother’s psychiatric well-being.
The couple was troubled by this. Ms. A said that although she was delighted by the prospect of taking quetiapine again, she wanted to determine a treatment course based solely on the potential effects of the medication on her fetus. Her husband also struggled with this decision. They took the published articles the consultant had copied for them and decided to return to discuss what to do further in several weeks.
By their second meeting, the couple had decided that Ms. A would restart quetiapine. They said they realized that Ms. A’s mental health had been a secondary consideration at the first consultation, and they felt that too much emphasis had been placed on the very small number of olanzapine cases. Ms. A commented that “if my mind chatter is so negative, I wonder if my chemistry is affected by that.” She feared the effects of her anxiety and any attendant additional stress on a fetus.
Ms. A decided to switch back to quetiapine and did so over the following month, increasing up to her normal dose of 300 mg/day. Within 8 weeks, Ms. A was socializing and functioning better at work. She had the energy to exercise daily and had lost 5 lb. After taking quetiapine for 3 months, her work and relationships had improved to the degree that she was being considered for a more substantial role in her department. She had lost all of the weight gained while taking olanzapine. She joked that she was so much better cognitively that she was able to worry about her future. She summarized the changes by saying, “My life is my own now. What I choose to do with it is my responsibility, and I have the wherewithal now to plan for it. I feel lucky.”
Ms. A’s case illustrates the complexity of issues faced by a woman with schizophrenia who wants to conceive. After aborting a fetus mainly because she feared its exposure to an antipsychotic, Ms. A sought consultation to review the risks and benefits to both her and her offspring from pregnancy and from medication exposure. This type of preconception evaluation allows couples to formulate a plan well before pregnancy, thus minimizing exposure of the fetus to unnecessary medications as well as preventing possible recurrence of psychiatric illness that may ensue from medication changes. It also allows for adjunctive interventions, such as psychotherapy or behavioral programs that could bolster psychiatric well-being.
Significant gender differences exist in the course, characteristics, and treatment response of schizophrenia. Although schizophrenia has classically been thought to have equal incidence in men and women, the risk may be lower in women (1). The peak onset of schizophrenia is later in women, occurring mainly in the mid to late 20s, with a second, much smaller peak between 45 and 49 years of age. One hypothesis for the differences in the age of onset is the protective antidopaminergic effect of estrogen on psychosis, leading to later onset in women in the earlier years and increased onset in the perimenopausal and menopausal years. It is also possible that these different ages of onset represent different subtypes of the disease (1, 2).
These differences in age of onset have implications for the course of illness. In contrast to men, women are more likely to have finished school, started working, and be married with children by the time of diagnosis (3). Women show greater functionality in all domains in the first 15 years of illness; this trend appears to reverse by the third decade of illness, perhaps related to the decrease in estrogen levels with menopause. Women also tend to demonstrate milder symptoms than men, fewer and shorter hospitalizations, greater posthospitalization employment, more intimate relationships, fewer legal problems, and a lower risk for suicide (4).
Women with schizophrenia are more likely than men to have positive family histories of schizophrenia and affective disorders (2). The diagnosis is often missed initially because of women’s “atypical” form of the disease that is characterized by more mood and fewer negative symptoms and faster resolution of both, with lower doses of antipsychotics (3, 4). Lower doses may be necessary because of several factors, including women’s higher percentage of body fat in which to store lipophilic antipsychotics, the protective effects of estrogen, and higher blood levels of antipsychotics. Women, however, experience more side effects, particularly akathisia, tardive dyskinesia, extrapyramidal symptoms, blood dyscrasias, and gastrointestinal and dermatological conditions.
Schizophrenia and Pregnancy
Although it appears that women with schizophrenia may have lower fertility rates than women in the general population, the majority of women with psychotic disorders do have children (5). They have a higher risk of unplanned and unwanted pregnancies and are more likely to be unmarried and to have limited social support (6).
The majority of women with schizophrenia experience loss of custody, with their children most often going into intermittent foster care (7). Such inconsistency in their upbringing can confer additional risk to the offspring, who already have increased genetic and in utero vulnerability to the disease.
Although the course of schizophrenia during pregnancy is not well defined, these pregnancies should be considered high risk (6). Women with schizophrenia tend to receive less prenatal care, have poorer nutrition, and use more tobacco, alcohol, and illicit drugs compared to women without schizophrenia. The offspring of women with schizophrenia appear to face an increased risk of lower APGAR scores, low birth weight, intrauterine growth retardation, preterm delivery, stillbirth, malformation, and infant death (6, 8). Of note, many of the studies examining the effects of schizophrenia on pregnancy did not control for maternal medication use. Studies evaluating the effects of medication on the fetus generally did not control for the impact of maternal illness.
Psychosis itself appears to be particularly harmful to the fetus: Nilsson et al. (9) found a doubling of adverse pregnancy outcomes in women who experienced a psychotic episode during pregnancy. These adverse outcomes, including stillbirth, death, prematurity, and small size for gestational age, were found despite control for other variables, including single motherhood, smoking, parity, age, education, and pregnancy-induced hypertensive disorders.
Additionally, a woman with psychosis is often agitated and anxious. Prenatal stress, particularly in the third trimester, is associated with a long-term impact on the hypothalamic-pituitary-adrenal axis in offspring (10) and may increase the risk of congenital malformations (11) and low birth weight (12). Further, research suggests that prenatal anxiety can negatively affect behavior, emotionality, and cognition in the growing child (13). Thus, although fetal medication exposure is more obviously identified as a risk to pregnancy outcome, maternal mental illness and its risks to the fetus must also be clearly evaluated when one considers treatment during pregnancy.
Given the risks of psychosis during pregnancy, the risks of fetal medication exposure must be weighed against the risks of no treatment; for example, stopping medication often leads to relapse. The classic triad of medication exposure risks includes the following: 1) congenital malformations associated with medication use during the first trimester, the major time of organogenesis; 2) perinatal complications reflecting medication exposure or withdrawal late in the third trimester; and 3) behavioral teratogenesis, which describes disturbances in behavior and cognition observed in a developing child exposed to medication in utero.
Increasingly, atypical antipsychotics are being prescribed as first-line agents for new-onset schizophrenia, and many women with existing schizophrenia are switching to them as well. Therefore, the group of reproductively aged women taking atypical antipsychotics may well be expanding. But the issue of whether to use typical versus atypical medication during pregnancy is a difficult one. There is more data on low-potency typical antipsychotics, which carry a slightly higher risk of congenital malformations. There is less safety data on high-potency agents, and they may require additional potentially teratogenic medications to manage side effects. On the other hand, data on atypicals during pregnancy are quite limited. Furthermore, these agents are associated with medical conditions that can themselves pose risks to both the mother and the fetus.
Literature Review of Antipsychotics
Much of the existing data on the use of typical antipsychotics during pregnancy comes from studies of the treatment of Hyperemesis gravidarum, which requires lower doses than those used for schizophrenia. A meta-analysis by Altshuler et al. (8) included 2,591 babies exposed in the first trimester to typical antipsychotics. That study found a small but statistically significant increase in the relative risk of congenital malformations from a background of 2.0% to 2.4% in offspring exposed during the first trimester to low-potency antipsychotics. No pattern of malformations emerged.
Data on high-potency antipsychotics are more limited. There have been three reports of limb malformations in offspring with first-trimester haloperidol exposure and one such report after first-trimester exposure to penfluridol, producing some concern of a possible association between high-potency antipsychotics (primarily haloperidol) and limb deformities (14, 15). Other studies did not confirm this association. A recent prospective controlled study involving 215 pregnant women exposed to either haloperidol or penfluridol had twice the risk of prematurity and had elevated rates of low birth weight among full-term infants (14).
A perinatal syndrome associated with typical antipsychotics includes symptoms of respiratory depression, difficulty feeding, floppy infant syndrome, hypertonicity, sluggish primitive reflexes, extrapyramidal symptoms, tremor, abnormal movements, irritability, and agitation (8, 15, 16). These symptoms are fairly rare and generally resolve within days (8).
Data on behavioral teratogenesis are sparse. Kris found normal social, emotional, and cognitive development in a cohort of 52 children born to mothers who took chlorpromazine during pregnancy (17). Stika et al. (18) also found no statistical differences in behavior or cognition in 66 9-and 10-year-olds with in utero exposure, only after the 20th week of gestation, to chlorpromazine or chlorprothixene.
To date, to our knowledge, there have been no blinded or randomized studies examining birth outcomes in women taking atypical antipsychotics, and these are unlikely given ethical considerations. There is one prospective study, and the remainder of the data come from case reports and two manufacturers’ data collections. Both case reports and manufacturers’ data may show a reporting bias that could overrepresent the rate of adverse outcomes. See data supplement 1 (available with the online version of the article at http://ajp.psychiatryonline.org) for information about these studies.
Of note, the prospective study done by McKenna et al. (19) grouped together subjects exposed to olanzapine, risperidone, quetiapine, and clozapine without any breakdown of results by medication, except for noting one malformation in a baby exposed to olanzapine. The exposed group showed no statistical differences in the rates of miscarriage, stillbirth, prematurity, congenital malformations, and perinatal syndromes in relation to healthy comparison subjects. There was an increased rate of babies with low birth weight and, in the mothers, greater body mass index and more elective abortions.
We found 129 cases reported in the literature of women taking olanzapine during their pregnancies, including 118 prospective (including case reports) and 11 retrospective cases. Sixty of these come from the study by McKenna et al (19). There was no recurrent pattern found in the four reported malformations. Two cases documented normal development; one reported delayed motor development at 7 months that resolved by 11 months.
We found 61 prospectively identified cases of fetal exposure to risperidone, including 49 from the study by McKenna et al. (19). There were no reported congenital malformations and two reports of normal development up to 1 year postpartum.
There were 39 prospectively identified cases of fetal exposure to quetiapine, including 36 in the study by McKenna et al. (19). We found no reports of congenital malformations. One case demonstrated normal development at 6 months.
Ziprasidone and Aripiprazole
To our knowledge, there are no reports of fetal exposure to ziprasidone or aripiprazole.
Nineteen prospective reports document fetal exposure to clozapine, including 13 case reports and six cases in the prospective study. There are two reports of perinatal/neonatal seizures and one child born prematurely with several anomalies and delayed development at 7 months. Normal development was reported in seven subjects evaluated up to 5 years of age. Gestational diabetes and/or pregnancy-induced hypertension complicated five pregnancies.
Retrospective data included 102 pregnancies with 59 deliveries resulting in 61 births. Data for 22 pregnancies were unavailable. Five had congenital malformations, and five had perinatal difficulties that were unspecified. There were no data on developmental outcomes.
Additional concerns include the known risks of agranulocytosis and orthostatic hypotension associated with clozapine. Although agranulocytosis has not been reported in a fetus, this possible risk warrants WBC monitoring in newborns, especially if the mothers are nursing. Additionally, maternal orthostatic hypotension could decrease placental blood flow, with potentially damaging effects to the fetus.
Medical Sequelae of Atypical Antipsychotics
Major side effects posed by atypical antipsychotics include weight gain, diabetes, sedation, and hypertension. These all confer significant risks for pregnancy. The rate of obesity has increased significantly in recent years. It is associated with high rates of obstetrical complications, including gestational diabetes mellitus, preeclampsia, and cesarean delivery (20). The rise in obesity may well account for increased perinatal adversity rates (21), as well as increased rates of gestational diabetes mellitus. For the mother, gestational diabetes is a strong risk factor for diabetes mellitus in a later nonpregnant state. Immediate risks to the fetus include macrosomia, hypoglycemia, shoulder dystocia, and associated birth injuries, such as fractures and nerve palsies.
Boney et al. (22) found that a neonate who is both large for gestational age and exposed to maternal obesity—with or without gestational diabetes mellitis—is at increased risk for metabolic syndrome during its childhood. Previously, abnormal birth weight (large or small for gestational age) and in utero exposure to maternal diabetes were found to be significant risk factors for the onset of type II diabetes mellitus in Pima Indian children. These findings suggest that fetal hyperinsulinemia may develop as a result of maternal hyperglycemia in an obese mother, potentially causing both the short-term problem of being either large or small for gestational age and the more insidious risk of glucose intolerance later in the offspring’s childhood.
Other studies have found that prepregnancy body mass index is also a major risk factor for stillbirth and neonatal death. Kristensen (23), in a prospective study, found a doubling of the risk of these outcomes in children born to obese mothers compared with those born to normal weight women. Of interest, adjustment for maternal cigarette smoking, caffeine and alcohol intake, maternal age, parity, education, and cohabitation with partner did not change the effect, nor did the exclusion of women with hypertension or diabetes. Diabetes was 10-fold more common in obese women compared with those of normal weight (5.5% versus 0.4%, respectively), whereas the proportion of women with hypertension rose from 3.7% in women of normal weight to 16.6% in those with obesity.
Maternal diabetes mellitus before pregnancy is an established risk factor for congenital malformations including CNS defects. Anderson et al. (24) found a substantially increased risk of anencephaly, spina bifida, or isolated hydrocephaly in the offspring of obese mothers (body mass index ≥30.0 kg/m2). In an earlier study, Moore et al. (25) found that the offspring of pregnant women with both obesity and diabetes were three times as likely to have a major, nonchromosomal congenital defect, suggesting that there may be a synergistic teratological effect between these disorders.
Another complication of rising body mass index is hypertension. Ricart et al. (26) prospectively collected information on body mass index and glucose tolerance on more than 9,000 women. He found that prepregnancy body mass index surpassed abnormal blood glucose tolerance as a risk factor for macrosomia, cesarean section, pregnancy-induced hypertension, and small-for-gestational-age babies. In fact, the upper quartile of prepregnancy body mass index (>26.1 kg/m2) accounted for 50% of the pregnancy-induced hypertension compared with only 9% caused by abnormal blood glucose levels. Hypertension is itself a major risk factor for small size for gestational age, preterm delivery, and resultant long-term developmental issues. In a prospective study of nearly 2,000 women with hypertension during pregnancy, Ray et al. (27) found that the greatest risk for obstetrical adversity was in mothers who developed preeclampsia, with or without underlying chronic hypertension. After adding control for maternal obesity and prior history of preterm delivery, Ray et al. found that preeclampsia doubled the risk of prematurity, prolonged ventilation, and perinatal death.
Risk for Postpartum Psychosis and Impact on Mother and Child
In the postpartum period, the risk for relapse of schizophrenia is worst during the first 3 months after parturition (28). McNeil (29) found a rate of postpartum relapse of nearly 24% in patients with schizophrenia, and their relapse occurred somewhat later than in women with predominately affective illness. Women with schizoaffective disorder are particularly susceptible (30).
Within the first few days after delivery, levels of estrogen and progesterone fall from as much as hundreds of times normal to their lowest levels during the follicular phase. These rapid shifts have been implicated in postpartum depression and psychosis by other authors. The loss of estrogen’s antidopaminergic activity may increase the schizophrenic mother’s risk for decompensation.
A mother’s worsening mental state is clearly dangerous for her neonate. Thought disorganization, attention to psychotic symptoms, blunted emotional awareness, anergy, and unresponsiveness can cause disruptions in mother-infant interactions (31) and may well have a lasting impact on a child’s cognitive and behavioral development (32). Furthermore, loss of a responsive, empathic mother may harm the child’s ability to attach or to develop object constancy. This is worsened when a mother must be hospitalized.
The worst outcome of a mother’s psychotic decompensation is infanticide. Although most postpartum psychoses are due to affective illness, infants with schizophrenic mothers still may be at great risk if the infant becomes part of the mother’s delusions. A woman with schizophrenia is more likely to kill her infant when her psychosis worsens as a result of medication discontinuation or stressors associated with the postpartum (33).
Treatment of Schizophrenia Postpartum
Fourteen case reports have shown that nurslings receive a small fraction of the maternal dose of olanzapine, risperidone, and quetiapine. Nurslings generally did well with mothers taking olanzapine, although sedation has been reported (34). We found three documented infant exposures to risperidone through lactation with no report of adversity (35–37). To our knowledge, there is only one case report of breast-feeding with quetiapine with normal findings in an infant at 4.5 months (38). Breast-feeding should not be encouraged in mothers taking clozapine because agranulocytosis and excess sedation have been reported in nurslings (39), and seizures are potential side effects. We found no reports of babies exposed to ziprasidone or aripiprazole.
Pregnancies of women with schizophrenia often are unplanned, suffer from complications, and result in loss of custody of the child. Therefore, psychiatrists should initiate and document discussions about contraception and pregnancy early and throughout treatment. Women should be informed of the risks of fetal medication exposure, the risks of psychosis during pregnancy, and the importance of prenatal care for themselves and their fetuses. A prepregnancy consultation is one way to optimize pregnancy outcome through education and establishment of a treatment plan.
A prepregnancy consultation should ideally involve the patient’s partner, other family or support, and other providers, such as the obstetrician. Her psychiatrist should review the patient’s history, including the nature of her previous episodes and current level of functioning and support. This history informs the risk-benefit analysis for medication use. Because of the high risk of relapse and the effects of psychosis on both mother and fetus, many psychiatrists suggest continuation of medications during pregnancy.
Many women with schizophrenia are using atypical antipsychotics because of subjective improvement and greater tolerability that obviates the need for additional medication exposure. Associated risks, such as obesity, diabetes, and hypertension, pose health risks for the mother and child and may lead to adverse pregnancy outcomes. Therefore, it is advisable to coordinate with the patient’s obstetrician to closely monitor weight gain, blood glucose, and blood pressure. Furthermore, the patient may benefit from meeting with a nutritionist and obtaining other supports to optimize her nutrition and weight.
An individual’s past response to treatment informs the choice of medication. Thus, in Ms. A’s case, quetiapine was ultimately chosen over olanzapine based on both her more complete therapeutic response to quetiapine and her history of decompensation and weight gain with olanzapine.
There may be cases in which one decides not to medicate, at least during the first trimester, in a woman in sustained remission from schizophrenia. However, one must also consider the woman’s past level of psychosis and risk for relapse off of medication. Indeed, medication maintenance may ultimately limit fetal exposure both to illness and to higher medication doses if maintenance prevents acute relapse.
Because there is risk for a perinatal syndrome, newborns should be monitored before discharge in case they do develop this complication. However, these syndromes are fairly rare and generally without long-term sequelae. One approach to treatment has been to taper down or off of medication before delivery to minimize the risk of a perinatal syndrome. Such benefit must be weighed against the patient’s risk of relapse during the postpartum—a particularly vulnerable time—with potentially adverse consequences to the mother’s ability to parent her newborn.
Data on behavioral teratogenicity from antipsychotic exposure in utero are limited, although they are greatest for typical antipsychotics. In all cases, optimizing the mother’s health and ability to parent should be considered crucial for the health of the developing child.
If medication was tapered before delivery, it should be restarted as soon as the baby is born. Postpartum relapse can impair attachment and lead to neglect, custody loss, and cognitive and behavioral sequelae in the offspring and, in the worst cases, infanticide. Furthermore, doses may need to be increased if there is worsening during the postpartum with the loss of estrogen. If this is necessary, it further complicates the issue of nursing while taking atypical antipsychotics, about which there is only limited information.
Antipsychotics, like all psychotropics, pass into breast milk. While breast-feeding does confer benefits to the infant, these are likely outweighed by potential risks. We recommend abstaining from breast-feeding until further data are available. However, this decision must consider the individual patient.
Supportive care is important in the postpartum to keep the mother healthy and able to care for her baby. New mothers with schizophrenia need to be assessed frequently for any change in their psychiatric well-being. Because they have regular appointments with obstetricians and pediatricians during the neonatal period, coordination of care among all providers offers a more complex system for monitoring a new mother.
Additional social supports may be particularly helpful in preventing relapse. Because insomnia may increase the risk of relapse, new mothers may benefit from someone else feeding the baby at night with either formula or expressed breast milk. All new mothers can benefit from parenting classes, mother-infant programs, and lactation consultation. When available, groups for new mothers with mental illnesses can offer instruction and support, minimizing a woman’s sense of stigma and guilt. Clearly, there is a need for better resources to aid these women, as well as for additional research that can further inform the risk-benefit analysis.