Tricyclic antidepressants (TCA) and serotonin/noradrenaline reuptake inhibitors (SNRIs) are less well studied than SSRIs, but may be used in women who do not respond to SSRIs. Most data exists for amitriptyline, clomipramine, imipramine, nortriptyline, and venlafaxine (Schaefer et al. 2015). Data do not indicate any increased risk of major malformations, but most studies have been too small or have other methodological limitations to draw definite conclusions. An increased risk of poor neonatal adaptation has been reported after use of venlafaxine during pregnancy, as for other serotonergic antidepressants.
15.2.2 Antipsychotics
Antipsychotics are prescribed to 0.2–0.3 % of all pregnant women (Petersen et al. 2014), mainly for bipolar disorder and schizophrenia. First-generation antipsychotics are known to affect the menstrual cycle and may also reduce fertility secondary to the group’s prolactin-increasing effects. Second-generation antipsychotics, on the other hand, have metabolic side effects that may increase the risk of excessive maternal weight gain during pregnancy, gestational diabetes, and preeclampsia. None of the published studies indicate an increased risk of teratogenic effects, though less data are available for the newest second-generation antipsychotics (Galbally et al. 2014).
Concern has been regarding the use of antipsychotics in the third trimester and transient dyskinesias in the newborn. These findings, however, have been mostly related to the use of first-generation antipsychotics. Most guidelines nevertheless recommend continuing antipsychotic treatment until delivery to avoid the risk of relapse and to monitor the newborn for possible neuromuscular adverse effects and feeding problems. General recommendations for the use of antipsychotics in pregnancy are summarized in (Table 15.1).
15.2.3 Mood Stabilizers
Lithium has traditionally been considered a teratogenic drug due to an association with cardiac malformations, in particular Ebstein’s anomaly. This risk, however, may have been overestimated, and today it is generally considered that the absolute risk of Ebstein’s anomaly after first-trimester exposure to lithium is low (0.5–1:1000, as compared to 1: 20,000 in the general population). Lithium should be reserved for cases where drugs with a more favorable fetal risk profile are unsuitable. Women using lithium in pregnancy could be offered an early ultrasound examination to exclude possible fetal cardiac defects (Schaefer et al. 2015). The newborn should be monitored for signs of neuromuscular toxicity first week postpartum.
Carbamazepine, valproic acid, and lamotrigine are antiepileptic drugs frequently used in patients with bipolar disorders. Carbamazepine and valproic acid should be avoided in pregnant women with psychiatric illness due to an increased risk of malformations and adverse neurodevelopmental effect in the offspring. Both carbamazepine and valproic acid interfere with folate metabolism, which may be one of the mechanisms involved. First-trimester exposure to carbamazepine has been associated with a twofold to fivefold increased risk of neural tube defects. An increased risk of other malformations, e.g., cleft palate, cardiac malformations, and hypospadias, has also been reported. A meta-analysis including more than 60,000 women with epilepsy using carbamazepine as monotherapy reported an overall malformation rate of 5.7 % (95 % CI 3.7–7.7 %) (Meador et al. 2008), which is significantly higher than the expected baseline rate of 3–4 %. There is plausible evidence of a dose-response effect with increased risk with doses above 1000 mg/day (Tomson et al. 2011).
Valproic acid is one of the most teratogenic antiepileptic drugs and should not be prescribed to women in reproductive age. The meta-analysis mentioned above found a malformation rate of 17.6 % (95 % CI 5.3–30.0 %) among women on valproic acid monotherapy (Meador et al. 2008). The most common malformations include neural tube defects, cardiac defects, and defects of the extremities and urogenital tract, with a dose-dependent increased risk at least for spina bifida and hypospadia (Vajda and Eadie 2005). It has also been suggested that a fetal valproic acid syndrome, including specific anomalies of the face and fingers, may exist (Kozma 2001). Finally, exposure to valproic acid is linked to impaired cognitive function with lower IQ scores, especially after exposure to doses above 800–1000 mg/day (Bromley et al. 2014).
A principal limitation of the literature on the safety of carbamazepine and valproic acid in pregnancy is that most studies have been conducted among women with epilepsy where the underlying disease in itself is associated with an increased fetal risk. Moreover, the doses are generally higher than those used in psychiatric disorders. However, as the risks related to use of these drugs are clearly significant, it is not justifiable to be less restrictive in the recommendations in bipolar disorder than in epilepsy, although the risk possibly is lower.
Knowledge on the safety of lamotrigine in pregnancy stems from thousands of pregnancies showing no increased risk of malformations or long-term adverse effects in prenatally exposed children. Lamotrigine could therefore be considered the antiepileptic drug of choice for pregnant women with bipolar disorders (Schaefer et al. 2015). General recommendations for the use of mood stabilizers in pregnancy are summarized in (Table 15.1).
15.2.4 Anxiolytics and Hypnotics
Among benzodiazepines and benzodiazepine-like hypnotics (zopiclone, zolpidem, and zaleplon), most data on safety during pregnancy exists for diazepam and zolpidem. One meta-analysis and a recent literature review concluded that there is little evidence for major malformations after use of benzodiazepines during pregnancy (Dolovich et al. 1998; Bellantuono et al. 2013). A study including more than 500 early pregnancy exposures to zolpidem found no overall increased risk of major malformations (Wikner and Källen 2011). There were four cases of non-atresial intestinal malformations (0.8 cases expected), but the authors suggest that this finding could be spurious due to multiple testing (Wikner and Källen 2011).
Prolonged use of benzodiazepines, particularly in high doses and close to delivery, has been associated with neonatal pharmacological effects and withdrawal symptoms, including poor suckling, floppy infant syndrome, and respiratory distress. Moreover, after prolonged use of benzodiazepines throughout pregnancy, concern has been raised about unfavorable long-term neurodevelopmental effects. However, although some data indicate that a small number of children may have a delayed development during the first year or so, they have generally developed normally by 4 years of age (McElhatton 1994).
Several important limitations exist for studies on the safety of benzodiazepines in pregnancy. Most studies did not have information about duration of therapy or indication for use, they did not differentiate between the various benzodiazepines, and in most studies there was a high degree of co-medication with other psychotropic drugs. Moreover, recall bias has been a major problem in the studies employing a case-control design (Dolovich et al. 1998). Women using benzodiazepines also drink more alcohol, are more often smokers, and have lower socioeconomic status than women not using benzodiazepines in pregnancy, factors that clearly could confound the results.
According to current knowledge, benzodiazepines and benzodiazepine-like hypnotics could be prescribed for short-term use in the lowest possible doses during pregnancy. Special precautions should be taken close to delivery. General recommendations for the use of anxiolytics and hypnotics in pregnancy are summarized in (Table 15.1).
15.3 Drug Safety During Breastfeeding
Human milk represents the ideal source of nutrients for small infants and provides superior immunological and antioxidant protection to milk substitutes (Newton 2004). As the infant should not unnecessarily be denied the benefits of breast milk, women are strongly encouraged to breastfeed whenever possible (World Health Organization 2003; American Academy of Pediatrics 2012). The obvious dilemma when treating a breastfeeding mother with a psychotropic drug is weighing the potential risk to the infant due to drug exposure through the milk against the disadvantage of not receiving breast milk. Another alternative, stopping or not commencing maternal drug treatment, might be even more harmful, taking into account the risk for the mother and thereby indirectly also for the infant if the mother is not receiving adequate treatment for her mental disorder (Cornisha et al. 2005).
Specific questions to be answered when deciding how to handle drug treatment in a woman with a mental disorder postpartum include: What are the risks for the mother and the infant if the maternal disease is not adequately treated? How strong is the mother’s desire to breastfeed? What are the disadvantages for the infant of not receiving breast milk? What are the risks for the infant of being exposed to the medication through breast milk? Is there any evidence to suggest that some specific drugs within a therapeutic group are more favorable than others to use related to infant risk? If necessary, could any practical strategies be used to reduce drug exposure to the infant? And finally, if there is a (often small or even just theoretical) risk of adverse effects in the infant due to drug exposure and breastfeeding nevertheless is allowed, should the infant be monitored in any way?
In order to provide precise answers to these questions, knowledge about to which extent the various psychotropic drugs are excreted into breast milk and about the infant’s age-related capacity to metabolize these drugs is a prerequisite. Moreover, knowledge about the theoretical infant dose and plasma concentrations of the drugs that could be expected in the infant and whether any adverse effects have been reported is also required. Unfortunately, available information about milk or infant plasma concentrations and effects of psychotropic drugs in breastfed infants is almost exclusively based upon case reports and small case series rather than upon prospective studies with unexposed control groups. For many drugs, information is extremely sparse.
15.3.1 Pharmacokinetic Considerations
Passage of psychotropic drugs between maternal plasma and breast milk is based upon principles of passive diffusion and will therefore follow a gradient from high to low concentration of free (unbound) drug. The timing of breastfeeding in relation to the time of maternal ingestion of the drug will thus be an important determinant for the concentration of the drug in the milk. When the concentration reaches its peak level in the maternal plasma, usually within a couple of hours after intake, it will, after a short delay, also reach its highest level in breast milk. Thereafter, as the maternal plasma drug concentration gradually declines, the concentration in breast milk will decrease until the mother ingests the next dose. For drugs with a short elimination half-life, such as zolpidem (t1/2 = 2–3 h), the risk of adverse effects in the infant can thus be reduced by breastfeeding at times when the drug concentration in breast milk is at its lowest level. This can be achieved – as obvious for zolpidem but also valid for non-hypnotic drugs with short elimination half-lives – by taking the daily dose in the evening and avoiding breastfeeding during the night. However, the vast majority of psychotropic drugs have long or very long elimination half-lives with quite stable concentrations in breast milk. For such drugs, avoiding breastfeeding during the peak concentration period will only reduce the infant’s drug intake to a small extent. Simulation of data from a study from our group of the excretion of paroxetine (Öhman et al. 1999) showed that total infant exposure would be reduced only by about 20–30 % by avoiding the peak phase in milk.
As psychotropic drugs are lipophilic, concentrations in breast milk increase and decrease in parallel with the milk triglyceride content. In two previous studies from our group, we have shown that the concentrations of paroxetine and aripiprazole were considerably higher in hindmilk than in foremilk, related to the higher triglyceride content in hindmilk (Öhman et al. 1999; Nordeng et al. 2014a). However, as also the nutritional value of milk is linked to its triglyceride levels, efforts to avoid additional drug exposure by discarding milk with high triglyceride levels, such as hindmilk, cannot be recommended.
Because newborns, and in particular premature infants, have an immature liver and kidney function, they eliminate drugs at a lower rate than older children and adults (for an overview related to psychotropic drugs, see Spigset and Hägg 1998). This applies both to enzymes belonging to the cytochrome P-450 (CYP) system (by which most psychotropic drugs are metabolized) and to the family of glucuronidating (UGT) enzymes (by which, e.g., oxazepam and lamotrigine are metabolized). For example, the elimination half-life of diazepam (which is mostly metabolized by CYP2C19) is approximately 80 h in premature infants, about 30 h in full-term newborns, and 10–20 h in infants after the newborn period. During the first months of life, hepatic function gradually matures, and after about 3–4 months of age, the metabolic capacity reaches adult levels. In contrast, renal function is not fully developed until the infant is at least 6 months old. Therefore, drugs with a high degree of renal elimination, such as lithium, are of particular concern related to accumulation in young infants.
In addition to the pharmacokinetic aspects discussed above, premature or seriously ill infants will often have a lower tolerance to the pharmacological action of drugs than healthy children, including possible unfavorable effects.
15.3.2 Calculation of Infant Dose
If the drug concentration in breast milk is known, the infant’s theoretical dose can be estimated by multiplying this value with the volume of milk that the infant ingests. In such calculations, milk intake in an infant who is fully breastfed is standardized to 150 ml/kg body weight per day. To assess the risk of adverse effects, the estimated dose can then be related, e.g., to the recommended pediatric dosage for that drug for individuals of the same age. For example, for lamotrigine a fully breastfed infant is exposed to amounts corresponding to between 25 % and 50 % of the therapeutic pediatric dosage for infants, indicating that there might be a risk of pharmacological effects in the infant (Nordmo et al. 2009; Hale 2014).
It is even more common to calculate the infant’s weight-adjusted relative dose, i.e., the dose the infant ingests per kilogram bodyweight in relation to the maternal dose per kilogram bodyweight. Infant exposure is regarded as minimal when the relative dose is below 2 %, small when the relative dose is 2–5 %, moderate when the relative dose is 5–10 %, and high when the relative dose is above 10 %. With relative doses above 10 %, it is generally considered that a risk of pharmacological effects in the infant does exist (Hale 2014; Ito 2000). With lower relative doses than 10 %, breastfeeding is in principle assumed to be safe, although infrequent cases of possible untoward infant affection have been reported also for drugs with relative doses below (but close to) 10 %, such as citalopram and aripiprazole (Hale 2014; Berle and Spigset 2011). On the other hand, breastfeeding is not necessarily always contraindicated during maternal treatment with drugs for which the relative doses are above 10 %, such as lithium and lamotrigine (see later).
15.3.3 Evidence of Adverse Effects
Most data on adverse effects in breastfed infants are derived from case reports (Rubin et al. 2004). Such reports are clearly of interest for drugs for which no or very little previous data exist, but it is often complicated or even impossible to determine whether there is a causal connection between reported infant symptoms and drug exposure.
A literature review of 183 reports on the use of psychotropic drugs during breastfeeding (Fortinguerra et al. 2009) found that infant adverse effects had been published for all groups of psychotropic drugs (but not for all individual drugs!). Notably, another review found that in about 80 % of the cases with adverse effects after maternal drug use in general, the infant was younger than 2 months of age (Anderson et al. 2003). This is accordance with the expected gradual maturation of hepatic and renal function during the first months of life and clearly illustrates that infant age is a critical factor to take into account when assessing the individual infant risk.
Very few prospective and systematic studies have compared the occurrence of symptoms possibly related to psychotropic drug exposure in breastfed infants of mothers taking and not taking the medication under study. One of a few such studies has been performed by our group (Berle et al. 2004). In that study, excretion of selective serotonin reuptake inhibitors (SSRIs) or venlafaxine to breast milk was investigated in 25 mothers with 26 exposed infants. Ten common symptoms of SSRI exposure (regurgitation/vomiting, irritability, tremor, suckling or feeding problems, decreased or increased sleep, yawning, etc.) were rated by the mothers and compared to a control group of 68 breastfed infants of the same age where the mother did not use any medication. There were no differences between the groups neither regarding any of the specific symptoms nor regarding the total symptom score (5.9 in the antidepressant group vs. 7.6 in the control group on a scale ranging from 0 to 30).
15.3.3.1 Antidepressants
Depression is a common illness among women in the postpartum period with an incidence of 10–15 % (O’Hara and Schwain 1996). In addition to the potentially harmful effects the depression may have on the mother, challenges exist related to caretaking of the newborn infant. Chronic maternal depression in the first year postpartum is associated with delayed psychomotor development in the child at 15 months and may also affect the cognitive and emotional development (Cornisha et al. 2005; Poobalan et al. 2007).
Key data on infant exposure for antidepressants are presented in Table 15.2. Regarding the choice of specific drug, it is sometimes recommended that sertraline should be preferred over other SSRIs due to the low infant exposure to that drug (Gentile 2007). It has also been recommended that when possible, fluoxetine and citalopram should be avoided or used with caution due to the higher infant plasma levels (Table 15.2) and the possible risk of adverse effects such as irritability, sleep disturbances, colic, and poor suckling.
Table 15.2
Infant doses and plasma concentrations of newer antidepressants after exposure via breast milk
Drug | Relative infant dosea | Absolute infant plasma concentration | Relative infant plasma concentrationb |
|---|---|---|---|
Selective serotonin reuptake inhibitors | |||
Citalopram | 3–10 % | Negligiblec | Up to 10 %d |
Escitalopram | 3–6 % | <5 ng/ml | <4 % |
Fluoxetinee | <12 % | Up to 100 ng/mlf | Up to 80 %f |
Fluvoxamine | <2 % | <LODg | – |
Paroxetine | 0.5–3 % | <LODg | – |
Sertraline | 0.5–3 % | <LODg | – |
Other newer antidepressants | |||
Venlafaxineh | 6–9 % | Up to 40 ng/ml | Up to 30 % |
Duloxetine | <1 % | <LODg | – |
Reboxetine | 1–3 % | <5 ng/ml | <2 % |
Bupropioni | 2 % | <LODg | – |
Mirtazapine | 0.5–3 % | 0.2 ng/ml | <1 % |
Nevertheless, irrespective of which SSRI if the mother has been treated with during pregnancy, we suggest that the same drug could also be used in the postpartum period (Berle and Spigset 2011; Berle et al. 2004). Based upon the current literature including a prospective study from our group (Berle et al. 2004), we recommend that when antidepressant treatment is indicated in the postpartum period, the woman should generally not be advised to discontinue breastfeeding. Some reviews and guidelines recommend infant monitoring, particularly if the infant is sick, is premature, or has a low body weight (Weissman et al. 2004; Lanza di Scalea and Wisner 2009). However, given the very low risk of any untoward effects, we consider there is no general need for routine follow-up examinations.
The numbers of exposed infants vary significantly between drugs, with about 100 cases for fluoxetine, paroxetine, sertraline, and citalopram, but less than 25 for the other newer antidepressants. Some degree of uncertainty inevitably exists for the drugs with the lowest numbers of exposed infants, even when no adverse effects have been reported. On this basis, drugs for which little data exist, such as fluvoxamine, venlafaxine, duloxetine, reboxetine, bupropion, and mirtazapine, should not be viewed as first-line therapies, but could be considered in special cases.
15.3.3.2 Antipsychotics
The risk of a psychotic reaction is higher in the postpartum period than anytime during a woman’s life. In a study of female patients with bipolar disorder, 36 % had had their first psychotic episode in the postpartum period (Hunt and Silverstone 1995). In another study of women who previously had been hospitalized with psychiatric symptoms, the risk of postpartum psychosis was increased about 100-fold (Nager et al. 2008). Thus, treatment with antipsychotics related to breastfeeding is not an uncommon issue.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
