Olanzapine in pregnancy and breastfeeding: a review of data from global safety surveillance

Women with psychiatric conditions may become pregnant, and motherhood is common in such women: in one sample, 63% of women with psychotic disorders were mothers [1]. Women with psychotic disorders who are pregnant or breastfeeding are often treated with antipsychotics [2]. Due to ethical constraints restricting inclusion of pregnant and breastfeeding women in clinical trials, there is a paucity of data available on the use of antipsychotic drugs in this population. The current literature on antipsychotic use during pregnancy and breastfeeding stems from case reports and large, uncontrolled reports of prospective and retrospective data [3, 4], making it difficult to draw conclusions concerning the safety of these medications for the mother and child. Adverse outcomes during pregnancy, delivery, post-natal care as well as birth defects, and perinatal complications have been reported in patients treated with antipsychotics [4].

Women with schizophrenia and bipolar disorder have a greater likelihood of complications, including placental abnormalities; hemorrhaging; fetal distress; congenital anomalies, such as cardiovascular defects; and neonatal complications [5]. Additionally, cessation of antipsychotic treatment for women with psychotic disorders may increase the risk of relapse, which in turn could lead to poor pre- and post-natal care as well as obstetric-related adverse events [4]. Therefore, clinicians and women must carefully weigh the benefits and the risks of remaining on or terminating antipsychotic treatment.

Olanzapine has been shown to be effective for treating the symptoms of schizophrenia and bipolar disorder in adults and schizophrenia and acute manic or mixed episodes associated with bipolar disorder in adolescents [6–10]. There have been reports about the safety of olanzapine in the fetus and infant, with some cases reporting normal births without complications and others reporting adverse outcomes, such as differences in birth weight [11] and neural development [2]. In preclinical trials, prenatal treatment with olanzapine did not disrupt spatial memory and short-term memory in rats, whereas other antipsychotics did [12].

In the absence of adequate and well-controlled clinical trials, spontaneous post-marketing adverse event data provides information on the safety of treatment with antipsychotic medications during pregnancy and breastfeeding. Spontaneous post-marketing data typically involve a much greater number of exposures to a much broader population of patients compared with data from clinical trials. This is particularly true in this population (women who are pregnant or breastfeeding), since pregnant women are excluded from clinical trials. However, this should be considered within the limitations associated with spontaneously reported data, which may sometimes under represent the true incidence of events, as not all patients or clinicians report pregnancies and their outcomes to Eli Lilly and Company. Here we present prospectively collected data from Eli Lilly and Company’s safety database including spontaneous reports, clinical trial cases, and post-marketing observational study reports, regarding the use of olanzapine during pregnancy and/or breastfeeding.

Through 31 December 2010, it is estimated that more than 33 million patients were treated with olanzapine. Based on the analysis of prospectively reported pregnancy where there was an outcome available, 66% of the outcomes of women treated with olanzapine at any time during pregnancy were normal. This rate of normal birth outcomes is comparable to that of the general population, which ranges from 61% to 64% [19]. Premature births (9.8%) and spontaneous abortion (9.3%) rates in this dataset were comparable to those of the general population (12.8% [20] and 10% to 20% [14, 15], respectively).

Prospective reports are subject to fewer reporting biases compared with retrospective reports [26]. Women whose children have major birth defects (abnormality that can affect the structure or function of an organ) are more likely to report the outcome, compared with women who have healthy babies [27]. Prospectively collected information provides more details about a case and helps to provide a risk estimate [28], whereas retrospective reports may be susceptible to systematic recall bias and underestimation of exposure to maternal psychiatric illness and non-psychotropic agents [29], and cannot be used in the calculation of outcome rates [30]. In addition, prospective cohorts are able to include specific variables, and the data can be collected with more reliability and accuracy [31], thus providing an opportunity to obtain follow-up information, and a more accurate ascertainment of exposure during pregnancy and in the perinatal period. Therefore, in this population, prospective reports were used to calculate frequency of pregnancy outcomes for comparison to the general population.

The risk of unfavorable outcomes in neonates and infants whose mothers were treated with olanzapine during pregnancy in this dataset did not appear to differ from that of the general population. This information is complementary to the findings of a previous prospective study of pregnancy outcomes in patients treated with atypical antipsychotics. The defects in the infant of the olanzapine-treated patient included the midline defects of cleft lip, encephalocele, and aqueductal stenosis. Yet all these events occurred in one child of an olanzapine-treated mother, compared with none in the other treatment groups [2]. Given the well-known limitations of the data under review, the findings need to be interpreted with caution.

Other congenital defects, including meningocele/ankyloblepharon [32], hip dysplasia [33], acheiria [34], and atrioventricular canal defect/unilateral clubfoot [35], have been reported in infants exposed to olanzapine in utero.

In a postmarketing safety database assessment of 68 prospective pregnancies, spontaneously reported with a known outcome in women who received risperidone [3], 37 pregnancies reported a normal outcome. Organ malformations were reported in 3.8% of the reports and spontaneous abortions in 16.9% of the reports. The denominator used was subtracting the number of induced abortions (predominantly undertaken for nonmedical reasons) from the total 68 reports. In utero exposure to risperidone does not appear to increase the risk of spontaneous abortions, structural malformations, and fetal teratogenic risk above that of the general population.

In a prospective, controlled cohort study of 215 pregnancies in women exposed to haloperidol or penfluridol [36], rates of major abnormalities were compared to a control group of pregnancies (N=631) exposed to non-teratogenic agents reported to a European counseling center. Compared to the controlled group, the women treated with antipsychotics were older; and a significantly higher proportion smoked 5 or more cigarettes a day. No differences were noted in the number of previous miscarriages, number of pregnancies, history of elective terminations, or gestational age at first contact. No differences were observed between the antipsychotic-exposed group and the control group in the rates of major malformations (3.4% vs. 3.8% respectively) or in the rates of major malformations in live births with first trimester exposure (3.1% vs. 3.8% respectively). There were statistically significant differences observed between the antipsychotic-exposed group and the control group in the rates of delivery (81.9% vs. 90.3% respectively), elective termination of pregnancy (8.8% vs. 3.8% respectively), preterm birth (13.9% vs. 6.9% respectively), Cesarean section (25.5% vs. 16.3% respectively), and in the median gestational age [interquartile range] at delivery (40 [38-40] vs. 40 [39-41] weeks respectively), median birthweight [interquartile range] (3155 [2800–3500] vs. 3370 [3030–3700] g respectively), and median birthweight [interquartile range] of full-term infants (3250 [3000–3590] vs. 3415 [3140–3750] g respectively).

In this dataset, 66% of births were classified as normal births; premature births occurred in 9.8% of pregnancies. This is not consistent with the findings of Newham et al. [37], who found higher rates of premature births in infants whose mothers were treated with typical antipsychotics compared with a non-medicated reference group, although there was no significant difference between the atypical antipsychotic and the reference groups. This analysis, as ours, did not control for concomitant mediations or other potential confounders, which may have potentially affected gestational age.

Although this analysis did not examine birth weight, previous studies have found higher rates of low birth weight (31%) and neonatal intensive care unit admission (31%) reported in neonates whose mothers were treated with olanzapine during pregnancy compared with neonates of mothers who were treated with other atypical antipsychotics [11], although the differences were not statistically significant. Another study found that infants born to mothers who were treated with olanzapine or clozapine during pregnancy had significantly higher birth weights compared with infants of mothers who were treated with typical antipsychotics [37].

One study examining placental passage (defined as the ratio of umbilical cord to maternal plasma concentration) of antipsychotics found that olanzapine concentrations in the placenta were 72% of the mother’s, which was significantly greater compared with concentrations of haloperidol, quetiapine, and risperidone [11]. However, another study reported the rate of placental passage of olanzapine to be much lower (17%) [38]. Potential fetal exposure to olanzapine should be taken into consideration when weighing the benefits and risks of remaining on olanzapine therapy.

Infant exposure to olanzapine via breast milk of olanzapine-treated mothers was lower than exposure in utero. Adverse events were reported in 15.6% of infants who were exposed to olanzapine through breast milk, and the most common events included somnolence, irritability, tremor, and insomnia. This differs from several studies of olanzapine in which either no adverse events were noted [39–43], or the types of adverse events (e.g., respiratory difficulties and hypotonia) were different [44]. The safety profile of olanzapine is based on exposures to the drug in adults and adolescents, the safety profile of exposure through uterus, during early and late developmental stages or via breast milk is likely to differ. Gilad et al. [44] found that the rate of adverse events in olanzapine-exposed breastfed infants were not statistically significantly different from infants not exposed to olanzapine.

The information reviewed in the current paper was voluntarily reported by (or on behalf of) the patient (spontaneous adverse event reports). There are several, well-known limitations to spontaneously reported data. Spontaneous reporting of adverse events may be highly variable [45]: clinicians may be less likely to report adverse events that are not serious or have been seen in the general population; this may skew the data toward more serious adverse events. Similarly, the information collected may lack details in particular key information, such as concurrent treatments, relevant medical history, and long-term follow-up of infant developmental outcomes. The data reported here included all reported cases regardless of concomitant medications, relevant medical co-morbidities, and potential causal relationship to olanzapine exposure. Since this is not a clinical trial, data are not collected like they typically are in a clinical trial or in a prospective observational study where there is a defined protocol. The follow-up rate cannot be calculated from spontaneously reported data as there is no pre-defined duration of follow-up for individual patient reports. Moreover, due to the nature of spontaneous reporting, and given the process of soliciting and obtaining follow-up information, a follow-up rate often does not provide an informative assessment of the completeness and accuracy of the follow-up. Although follow-up information may have been received, often the most relevant follow-up is the outcome of pregnancy/breastfeeding, which may not always be available. In addition, it is not possible to confirm the exact trimester in which the patients were exposed to olanzapine from spontaneously reported data. In summary, spontaneous data alone are not adequate to make definitive conclusions regarding the potential risk of adverse events following exposure to olanzapine.

Based on reports from the Adverse Event Reporting System (AERS), the Food and Drug Administration’s (FDA) computerized information database designed to monitor for new adverse events and support post-marketing safety surveillance for approved drugs, various major regulatory agencies recently updated the prescribing information label language of all antipsychotics regarding use during pregnancy, informing of the potential risk of abnormal muscle movements (extrapyramidal signs) and/or withdrawal symptoms in neonates whose mothers were treated with antipsychotics during the third trimester of pregnancy [46]. It is important to assess the risks and benefits of treating women who are pregnant or breastfeeding with antipsychotics, and weigh these against possible risks of anomalies and developmental problems to the fetus and infant.

There are no controlled studies for the use of olanzapine therapy in pregnant women or in women who are breastfeeding. Given the well-known limitations of data under review, the findings need to be interpreted with caution. The relative risks and benefits of olanzapine treatment during pregnancy and/or breastfeeding should be carefully weighed by the clinician and the patient on a case-by-case basis. Because of limited experience in humans, olanzapine should be used in pregnancy only when potential benefit justifies potential risk to the fetus [25]. Olanzapine should only be considered during breastfeeding when the potential benefit justifies the potential risk to the infant. Women should be advised to notify their clinician if they become pregnant or intend to become pregnant during treatment with olanzapine. Presently, data are not sufficient to make definitive conclusions regarding the safety of olanzapine therapy during pregnancy and/or breastfeeding. However, acknowledging the limitations of the existing data, our review found that the frequency of fetal outcomes in prospectively identified pregnancies exposed to olanzapine did not differ from rates of outcomes reported in the general population. In the absence of data from clinical trials, the current analysis of post-marketing data attempts to provide greater information about the safety of olanzapine exposure during pregnancy and/or breastfeeding.