Intended for healthcare professionals


Safety of antipsychotic drugs for pregnant and breastfeeding women with non-affective psychosis

BMJ 2004; 329 doi: (Published 21 October 2004) Cite this as: BMJ 2004;329:933

This article has a correction. Please see:

  1. Louise Howard (l.howard{at}, senior lecturer,
  2. Roger Webb (r.webb{at}, research fellow,
  3. Kathryn Abel (kathryn.m.abel{at}, senior lecturer
  1. Health Services Research Department, Institute of Psychiatry, London SE5 8AF
  2. Centre for Women's Mental Health Research, University of Manchester, Manchester M13 9PL

    Multicentre cohort studies are needed as randomised trials are impractical

    Around 2% of women develop a non-affective psychotic disorder, and more than half of them have children.1 2 Women with nonaffective psychoses are less fertile than controls, partly because of hyperprolactinaemia secondary to antipsychotic drugs.3 4 The use of atypical drugs such as clozapine and olanzapine, which do not have this effect, may increase fertility rates. Weighing risks against benefits in treating pregnant and breastfeeding women with antipsychotics requires assessment of clinical effectiveness versus risk of toxicity to mother, fetus, newborn, and the developing child. Withholding antipsychotic treatment may expose mother and fetus to more harm than benefit as, in addition to behavioural disturbance which may put both at risk, physiological changes associated with psychosis could affect fetoplacental integrity and development of the central nervous system. Subsequently, poor clinical outcomes for mothers may result in poorer interaction between mother and infant.5 But the impact of antipsychotic medication on the fetus is also unclear as no randomised controlled trials have evaluated the use of antipsychotics in pregnancy.6

    The only large controlled studies of antipsychotics in pregnancy have been conducted on women with hyperemesis gravidarum. Much lower doses of medication (generally drugs such as phenothiazines) are used for these patients than for patients with schizophrenia, and residual confounding is commonplace. Conclusions are therefore limited. A metaanalysis reported that exposure to low potency antipsychotics during the first trimester was associated with a small additional risk of congenital anomalies.7 Antipsychotics can also produce toxic effects in newborn infants medicated in the womb, including respiratory depression and neonatal behavioural abnormalities such as extrapyramidal movements and difficulty with oral feeding, although these effects usually resolve within days. The few small studies examining longer term neurobehavioural effects in children exposed to antipsychotics in the womb have reported no behavioural or intellectual abnormalities.

    Even less is known about any adverse effects of atypical antipsychotics. On comparing case registry data (n = 37) with historical controls the rates of adverse effects with olanzapine during pregnancy were not found to be higher.8 However, data from the National Teratology Information Service indicate an increased incidence of malformations of 10% (compared with an expected incidence of 2-3%).9 However, as only three cases were seen, no clear relation can be established.9 One survey of pregnant women taking clozapine reported a higher than expected rate of malformations (five cases in 61 pregnancies).10 Reduced variability of the fetal heart rate before and during labour has also been noted in one woman taking clozapine.11 Case reports and data from the National Teratology Information Service do not indicate a problem with quetiapine and risperidone during pregnancy. No data are available on newer drugs such as aripiprazole. An increased risk of neural tube defects has also been proposed as many atypical antipsychotics are associated with weight gain and subsequent folate deficiency.12 This association has not been investigated empirically.

    Thus very little evidence exists regarding the use of antipsychotics during pregnancy. When a woman with schizophrenia presents pregnant, wanting to come off medication, many psychiatrists will be concerned about the high risk of relapse during pregnancy if medication is discontinued. Most will suggest that ideally she should be maintained on the lowest possible dose of antipsychotic during and after pregnancy, with fetal screening throughout pregnancy. The woman is likely to be concerned about the risk of medication to the fetus. But neither the psychiatrist nor the pregnant woman will be able to find data on the size of the effect of these risks.

    Pharmacological treatment after delivery depends on the mother's symptoms and whether she is breast feeding. Antipsychotic drugs are excreted in breast milk, but to date breast fed infants have not shown signs of toxicity or impaired development in most case reports of antipsychotics, although manufacturers advise avoidance of these drugs during breast feeding. Recent recommendations from the World Health Organization and the UK government, stressing the desirability of breast feeding exclusively for six months, make it imperative that women know whether they can do this safely while on medication.

    Little is known about the effect of newer antipsychotics; preliminary experience shows no adverse effects attributable to olanzapine, but long term data are required.10 13 Dev and Krupp reported on four infants breast feeding from mothers taking clozapine-one developed excess sleepiness and another agranulocytosis, which resolved with discontinuation of breastfeeding.10 Mothers on clozapine should thus not breast feed. The evidence for prescribing antipsychotic medication while breast feeding is therefore based on even fewer reports. Although mothers are advised to continue the same medication given during pregnancy when breast feeding, to avoid drugs with a long half life, and to time feeds to coincide with trough concentrations of drugs in breast milk, this is based on common sense and not on any high level evidence.

    The lack of evidence is partly attributable to ethical constraints restricting randomisation of pregnant and breastfeeding women. In addition, drug companies do not usually sanction the evaluation of antipsychotics in pregnant or breastfeeding women in randomised controlled trials. This raises ethical and clinical dilemmas regarding appropriate management of an already vulnerable group of women.

    Many women are likely to be taking atypical antipsychotic drugs at conception. This provides an opportunity to investigate short and long term adverse effects of antipsychotic drugs on pregnant women and their offspring, including neurobehavioural outcomes in these children. Although randomised controlled trials may be impractical, coordinated multicentre prospective cohort studies could provide good evidence. Obstetricians, psychiatrists, and general practitioners need robust evidence on which to base prescribing of antipsychotic drugs for pregnant and breastfeeding women, and women must be provided with dependable information to support them in decisions about medication during pregnancy and after the birth.


    • Competing interests None declared.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    13. 13.
    View Abstract