Postnatal depression and puerperal psychosis
Section 4: Prescribing issues in pregnancy and lactation

4.1 Introduction

Clinicians are cautious about prescribing drugs during pregnancy or when a mother is breast feeding due to the possible risks to the fetus and infant. In early pregnancy the risk of teratogenesis is the main concern. The main risks associated with psychotropic drugs in later pregnancy are neonatal toxicity or withdrawal syndrome following delivery and the possibility of a long term impact on the infant's neurodevelopment.119 During breast feeding many drugs taken by the mother are excreted in the milk and ingested by the infant, with consequent concerns about their impact on the infant with regard to both short term toxicity and longer term neurodevelopment.120 The level of concern about prescribing during pregnancy and lactation is reflected in the Terms of Marketing Authorisation, with most psychotropic drugs not being licensed for use in pregnancy and lactation. This underlines the need to give very careful consideration to the risks and benefits of prescribing psychotropic medication at this time.

Decisions may have to be made about commencing psychotropic drugs during pregnancy if a woman becomes ill or, more commonly, stopping medication if a woman finds that she is pregnant. As many pregnancies are unplanned, some are exposed inadvertently to psychotropic medication. Advice may then be sought on the need for further investigation and possible therapeutic termination of pregnancy. Studies based on systematic case registers such as the Swedish Medical Birth Registry121 are contributing to more substantial evidence for such decisions, particularly in relation to teratogenicity.

Much of the evidence base for risk associated with drug treatments is based on case reports and case series. As new antidepressants are introduced there will continue to be a time-lag in evidence becoming available on which decisions can be made. Evidence is scant for newer agents and for long term developmental risk. It is reasonable to assume that the fetus and newborn are as, if not more, susceptible to the same side effects as adults. The evidence identified demonstrates a relatively low risk with most psychotropic agents, however no treatment is risk-free. Pregnant women and their families have the right to expect that treatments prescribed are clearly indicated and are associated with lowest known risk.


4.2 Drug treatment in the first trimester


Evidence indicates no increased risk of major malformation in the newborn or spontaneous abortion following exposure to antidepressants in early pregnancy.119,121,122,123,124,125,126 This evidence applies to a variety of tricyclic antidepressants (TCAs) and selective serotonin re-uptake inhibitors (SSRIs);119,121,123,124 and more specifically, fluoxetine,122,125 citalopram,121 fluvoxamine,126 paroxetine,126 and sertraline.126 Evidence level 2++

Adequate evidence is not available to make recommendations on the use of other antidepressant medications.




Lithium is regularly used on a maintenance basis in the prevention of relapse of bipolar affective disorder. Such a relapse is more likely to occur following childbirth and when lithium is withdrawn.127 Evidence level 2+

Current practice for women with bipolar disorder who plan to or become pregnant ranges from the discontinuation of lithium treatment accompanied by close monitoring or prescription of antipsychotic medication, through to maintenance of lithium throughout pregnancy in cases where the risk of relapse is significant. Evidence is not available to allow comparison of these strategies, but recent studies have allowed a review of the risks associated with lithium treatment.

Early studies of lithium in pregnancy suggested that the risk of major fetal malformations, in particular Ebstein's anomaly, was increased by exposure to lithium in early pregnancy.128 Recent evidence, based on prospective studies, suggests that the risk to the fetus of lithium exposure may have been over-estimated and the risk to the mother and child of lithium withdrawal may have been under-estimated.127 Evidence level 2+

Other mood stabilisers

The antiepileptic drugs (AEDs) carbamazepine, valproate and, more recently, lamotrigine are also used as mood stabilisers. The evidence from studies of women with epilepsy suggests that exposure to AEDs in early pregnancy increases the risk of congenital malformations and this effect is related to the use of antiepileptic drugs, not the epilepsy.129 The relative risk is higher with valproate than carbamazepine and, in particular, with doses of valproate over 1000 mg per day. Evidence level 2+

Several AEDs, including carbamazepine, are folate antagonists. Folic acid supplements are recommended for women on AEDs from preconception to the end of the first trimester.130 (See the SIGN guideline Epilepsy in Adults due for publication in late 2002). Evidence level 2+

There is as yet no evidence available on the risks of lamotrigine in early pregnancy.


Evidence suggests that exposure to benzodiazepines in early pregnancy increases the risk of major malformations and oral cleft in the fetus.131 Evidence level 1+

4.3 Drug treatment beyond the first trimester


The evidence base on the risks of perinatal problems and impaired infant development following exposure to psychotropic medication in pregnancy is limited to single case reports covering a wide variety of types of psychotropic medication, and a limited number of cohort and case-control studies. Perinatal toxic syndromes and withdrawal syndromes following delivery by mothers who received psychotropic medication during pregnancy have been identified, and there are understandable concerns regarding the long term effects of such medication on the neurological development of the infant.119 In particular, newborn infants of women treated with lithium in later pregnancy face potential risks of neonatal toxicity, thyroid and renal dysfunction.132

One cohort study found no evidence of impairment of neurological development at two years after delivery following exposure to antidepressants.133 Evidence level 2+

A case control study found poor neonatal function and withdrawal syndromes in a significant proportion of infants exposed to antipsychotic medication in the last three months of pregnancy.119, 134 Evidence level 2++

4.4 Drug treatment and lactation

Women who develop mental illness following childbirth and need psychotropic medication are likely to be discouraged from breast feeding because of the risks to the infant. However, the risk of breast feeding in this situation may be over-estimated and the advantages under-estimated.135 The process of excretion of psychotropic medication is complex, with variation in milk/maternal plasma ratios for different drugs and between foremilk and hindmilk. The level of metabolic maturity of the infant will also influence any effect of drugs taken by the mother. The evidence base is limited due to the small number of breast feeding women who have been exposed to any specific drug and the lack of any systematic approach to monitoring and registering information about the use of psychotropic medication in breast feeding women.136


The use of TCAs is diminishing with the introduction of the SSRIs and other novel antidepressants. The evidence base consists mainly of case studies, which confirm that TCAs are excreted in breast milk in higher concentrations in hindmilk than foremilk and the milk/maternal plasma ratio exceeds one. The limited evidence available suggests no short term toxic effects for the infant except in the use of doxepin. 120,134,135,137,138,139,140,141 No long term developmental effects for the infant have been demonstrated. Evidence level 2+


SSRIs are being used increasingly in clinical practice. The evidence identified on the use of SSRIs in breast feeding is based on individual case reports and a small case series.

Cumulative evidence on the use of sertraline142,143,144,145,146 indicates no significant adverse effects on the breast-fed infant. In one cohort study, 24% of serum samples from infants of nursing mothers taking sertraline had detectable levels of sertraline, particularly where the mother's daily dose was 100 mg or more.147 No adverse clinical effects have been reported in breast-fed infants of mothers taking paroxetine.147,148,149,150 Infant serum levels were undetectable in all studies.147,148,149,150 No adverse neurodevelopmental effects were noted in a study of fluoxetine and breast feeding.151 However, fluoxetine and its metabolite nor-fluoxetine have been detected in the serum of a proportion of breast-fed infants whose mothers were taking fluoxetine152 and nor-fluoxetine has a particularly long metabolic half life.153 Paroxetine has the lowest milk/plasma ratio of these three drugs.146,148,149 Evidence level 2+

There is very limited evidence on citalopram but the milk/maternal plasma ratio is relatively high, as is the calculated infant dose.154,155,156 There is little evidence on fluvoxamine but it has a low milk/maternal plasma ratio and the calculated dose to the infant is low.157,158 Evidence level 3


Evidence indicates that lithium is excreted in breast milk at a level of approximately 40% of maternal serum level. Lithium toxicity has been described in a breast-fed infant and lithium is known to impair thyroid and renal function in adults.159, 160 Evidence level 3

Sodium valproate is excreted in breast milk in low levels and infant serum levels are between one and two per cent of the maternal serum level. No adverse clinical effects have been noted in breast-fed children when mothers are taking sodium valproate.159, 161 Evidence level 2-

The evidence on carbamazepine suggests that it is excreted into breast milk in significant quantities and infant carbamazepine levels in serum range from 6% to 65% of maternal serum levels. Evidence level 3


Benzodiazepines are excreted in breast milk with a low milk/plasma ratio.120, 162 Evidence level 4


Note: this recommendation does not cover drug dependence, where breast feeding may be beneficial if the infant has been exposed to benzodiazepines in utero.

All forms of antipsychotic medication are excreted in breast milk but as yet there is no evidence to suggest that breast-fed infants are at risk of toxicity or impaired development.119,120,137,162,163 There is little evidence on the new atypical antipsychotic drugs.120, 164

The evidence on other antidepressant drugs including moclobemide,165,166 venlafaxine,167 and nefazodone168 in breast feeding is very limited, but they are all excreted in breast milk. Evidence level 3

[Contents][Back][Top] [Next]