The benefits of breastfeeding are sufficiently important to recommend that breastfeeding should not be discontinued or discouraged unless there is substantial evidence that the drug taken by the mother will be harmful to the infant, and no alternative treatment can be found.

Most drugs are excreted only to a minimal extent in breast milk and in most cases the dosage to which the infant is ultimately exposed is very low and is well below the therapeutic dose level for infants. For this reason, few drugs are totally contraindicated while breastfeeding. However, women who are known to be HIV-infected should be counselled not to breastfeed or to provide their milk for the nutrition of their own or other infants, except in areas where infectious diseases and malnutrition are major causes of infant mortality and where safe alternatives to breastfeeding are unavailable.

In most situations, drugs cross the placenta more efficiently than into breast milk.

When considering prescribing medication (particularly longer-term) during breastfeeding, the following check list may assist in guiding the decision:

The postnatal period may represent an opportunity for the clinician to discuss the modification of maternal drinking and smoking behaviours. Every effort should be made to capitalise on a woman’s motivation to enter treatment during this time.

Educational and behavioural interventions should be the mainstay of treatment. Pharmacological treatments should only be considered after specialist consultation. 

In Australia, approximately half of the women who quit smoking during pregnancy resume smoking in the postnatal period. This is thought to be due to postnatal smoking being perceived as more socially acceptable (as compared to pregnancy), as well as smoking being seen as helpful for losing weight. The powerful influence of having a male partner who smokes has also been emphasised in recent literature.

There is a well-established dose-response relationship between the number of cigarettes smoked and reduced likelihood or duration of breastfeeding.

Smoking has a number of adverse effects upon lactation in terms of decreased production, interference with the let-down reflex and altered taste of breast milk. The infant may be affected through both nicotine in the milk and side-stream smoke. The effects may include colic, irritability, apnoeic episodes and immune system impairment. 

There is debate as to whether or not the adverse effects of nicotine in breast milk are outweighed by the benefits (especially immunological ones) to the infant of being breastfed. Some authorities argue that if a woman cannot cease smoking during the postnatal period she should nevertheless still breastfeed. The resultant boosted immunological competence of the infant has, in some studies, been shown to be protective against the increased risk of respiratory tract infection resulting from passive smoking. Delaying smoking until after having just breastfed the baby, will minimise the nicotine available in the milk by the time of the next breastfeed.

Ongoing exposure to cigarette smoke in the postnatal period is a well-recognised risk factor for sudden infant death syndrome (SIDS).

The use of both nicotine replacement therapy (NRT) and the antismoking drug bupropion during breastfeeding remains controversial. However, the amount of nicotine and cotinine in the milk from patches is similar to that from an equivalent ‘dose’ of cigarettes. Moreover, the milk will not contain other cigarette-derived toxic substances. The additional benefits for a young baby in a nonsmoking household should also be taken into account (eg reduced risk of SIDS). If nonpharmacological measures are unsuccessful, consideration of the use of NRT may be warranted, although many authorities still regard it as contraindicated.

Caffeine is readily transferred into breast milk, and young infants have lower capacity to metabolise it, which raises the possibility of caffeine toxicity. Infants breastfed by mothers ingesting more than 300 mg/day may become jittery, restless and experience sleep difficulties. No long-term adverse effects have been documented. Conflicting findings have emerged regarding the relationship between caffeine and SIDS, and confounding factors (eg smoking) may explain the association found in one study, which was unable to be replicated. For information on the approximate caffeine content in drinks and other preparations, see caffeine use in pregnancy

.

Antiepileptic drugs do appear in breastmilk, but the clinical significance of the amounts excreted has been questioned. However, although breastfeeding is not contraindicated, the infant must be monitored for drowsiness or feeding difficulties. The dose should be kept as low as possible to maintain seizure control.

No adverse reports have been associated with sodium valproate, and this may be the drug of choice in breastfeeding. Hepatotoxicity has been reported in children under 2 years who have been prescribed sodium valproate, but this was as medication to the infant, not by transmission in breast milk.

Occasional reports associate carbamazepine with infant jaundice and hepatic dysfunction.

The American Academy of Paediatrics does not regard these two antiepileptics as contraindicated.

The safety of lamotrigine has yet to be established. Potentially serious skin conditions have been reported in young children treated with this drug.

For acute treatment, paracetamol, ibuprofen, and naproxen can be used. Aspirin should be avoided. Of the triptans, sumatriptan may also be used if migraine headaches are poorly controlled by analgesics alone. The safety of other triptans has not been adequately studied.

For preventive treatment, amitripyline, metoprolol, propranolol, sodium valproate, and verapamil are reasonably safe.

Oral or injectable corticosteroids are generally not contraindicated while breastfeeding. Some caution is required with large doses or prolonged use. However, even at doses of up to 80 mg per day of prednisolone, less than 0.1% is ingested by the infant, which is equivalent to less than 10% of the endogenous cortisol production.

However, particular caution is required in the case of premature infants, or other infants whose ability to metabolise and/or excrete drugs may be impaired.

Box 11.2 can provide a framework for the practitioner to address the decision of whether or not to prescribe a psychotropic drug during breastfeeding. It can also guide discussion with the woman and, preferably, also her partner. 

A categorisation of harmful effects to the infant has been developed utilising 2 classes: short-term toxicity (class 1) and long-term neurodevelopmental effects ('behavioural toxicity') (class 2). The risk of short-term toxicity (class 1) depends on a variety of complex factors, in addition to the drug itself. These include timing of dosage in relation to breastfeeding, differing concentration in foremilk versus hindmilk, the weight of both the mother and the infant, as well as their ability to metabolise and/or excrete the drug and its active metabolites etc. It is this complexity which accounts, in part, for the conflicting findings regarding the extent of transmission of psychotropic drugs into breast milk. Nevertheless, most psychotropic drugs are excreted only to a minimal extent in breast milk, and in most cases the dosage to which the infant is ultimately exposed is very low and well below that which would be expected to have any significant clinical effects. For this reason, there are few drugs which are totally contraindicated while breastfeeding. In most situations, drugs cross the placenta more efficiently than they pass into breast milk. However, particular caution is required in the case of premature infants, or other infants whose ability to metabolise and/or excrete drugs may be impaired.

Some authorities recommend that infants whose mothers are taking psychotropic medications should be monitored for gastrointestinal dysfunction (eg colic, poor feeding), as well as for excessive sedation, irritability and agitation. Given the high incidence of such problems in infants in general, the practitioner should be cautious about attributing these to medication. If severe, a trial of decreased dosage or cessation may be appropriate. For medications given in once-daily dosage, avoidance of breastfeeding for 4 to 6 hours after ingestion may warrant consideration.

Class 2 is largely unexplored, but should not deter prescribing where this is otherwise indicated. The relevance of the unknown risk in class 2 is to impress on the practitioner that no psychotropic can be considered 100% safe in lactation, nonpharmacological interventions always warrant consideration and unnecessarily high dosage and duration of treatment should be avoided. In addition, caution is warranted in prescribing only on the basis of prevention of recurrence of a condition (especially unipolar postnatal depression) which is not currently present. Nevertheless, some authorities do recommend prophylactic prescription of lithium in bipolar patients, especially if there have been previous postpartum relapses.

Important pharmacological determinants for the transfer of drugs into breastmilk include plasma protein binding (PB), oral bioavailability (OB) and adult half-life of the drug (AHL). High protein binding restricts the diffusion into breastmilk (eg olanzapine PB = 93%; sertraline PB = 98%). Oral bioavailability shows what percentage of the drug, from the breastmilk swallowed by the baby, is absorbed. Shorter maternal half-life medications are recommended over those with long half-lives. 

Definitive information is lacking for the atypical antipsychotics (amisulpride, clozapine, olanzapine, quetiapine, risperidone and ziprasidone). Preliminary data suggest that olanzapine and risperidone have low infant doses relative to the maternal weight-adjusted dose and, given the lower adverse effects in the mother, these medications are being increasingly used during lactation. However, additional studies are required to definitively establish safety. Most authorities advise against the use of clozapine in breastfeeding (due to high concentrations in breast milk) unless strongly indicated for maternal well-being. 

The typical antipsychotics, especially at high dose, have occasionally been associated with adverse reactions in breastfed infants (eg urinary retention, dystonic reactions). In lower dosage, ie less than 10 mg haloperidol daily or equivalent, they are usually regarded as safe. 

Most authorities advise against the use of depot preparations if possible. If the infant suffers an adverse reaction to a depot preparation, cessation of breastfeeding will be necessary.

Postnatal depression affects some 10% to 20% of mothers, and there is also a small percentage of women in the population who will require antidepressant treatment throughout their pregnancy and while breastfeeding. Unfortunately, the literature still mainly comprises case reports or very small studies, and the presence of active metabolites are not always taken into account in laboratory studies. However, more methodologically sound studies have recently started to emerge.

In the case of SSRIs, most studies suggest that infant plasma concentrations are either low or not detected, with the exception of fluoxetine where clinically significant concentrations have been measured. Fluoxetine and its metabolites with long half-lives have the potential to accumulate in the breastfed infant, and this medication has been associated with low weight gain and infant gastrointestinal dysfunction in one study. There are now several published studies on sertraline, citalopram, fluvoxamine and paroxetine that show acceptably low relative infant doses, and no class 1 adverse effects in breastfed infants. In recently published studies, venlafaxine had a higher mean relative infant dose than the SSRIs sertraline, citalopram, fluvoxamine and paroxetine. However, the authors considered it to be compatible with breastfeeding. Where infants breastfed by mothers taking SSRIs and venlafaxine have been assessed neurologically or developmentally, no adverse outcomes have been noted. 

Moclobemide has very low transfer into breast milk and is probably safe. Nefazodone, mirtazapine and reboxetine have not been adequately studied. 

Breast milk transfer of TCAs is low, and they are generally considered safe during breastfeeding; however, compared with newer antidepressants they have a less favourable adverse effect profile for the mother, and much higher lethality in overdose. The use of doxepin is controversial because there have been isolated case reports of sedation and respiratory depression in the infant. 

Some authorities recommend use of nortriptyline on the grounds that maternal serum levels can be monitored.

As regards neurodevelopmental toxicity, the only data are for dothiepin (15 infants followed for 4 years). No adverse sequelae were found.

Mianserin and the irreversible nonselective MAOIs have not been adequately studied.

Discontinuation of 'mood stabilisers' in bipolar patients is associated with a significant risk of relapse. The risks versus benefits must be carefully weighed, taking into account the following sections on lithium and antiepileptics. 

High infant serum concentrations have been reported, and renal clearance may be reduced in the young infant. Most authorities advise that lithium be avoided due to adverse reports of infant sedation, dehydration, hypothyroidism and respiratory depression. If used, regular monitoring of infant serum concentrations, thyroid function and electrolytes is recommended, which is an invasive procedure to be avoided wherever possible. Moreover, the interpretation of infant serum lithium concentrations, in terms of safety, may be difficult. The American Academy of Paediatrics states lithium is contraindicated in breastfeeding.

For information on antiepileptics, 

Floppy infant syndrome with symptoms of hypotonia, lethargy and reduced suckling have been reported with the long-acting benzodiazepines (eg diazepam) since they may accumulate in the infant during chronic use. If a benzodiazepine hypnotic is needed during breastfeeding, then temazepam may be the drug of choice as it has a shorter half-life.

Particular caution should be exercised in the first 3 or 4 postnatal days, especially if the infant has physiological jaundice.

There are no data relating to the use of zopiclone or zolpidem in breastfeeding.

Methadone is considered safe to use. Symptoms of withdrawal may occur in the first week in 60% of infants born to mothers on methadone maintenance. Buprenorphine has not been associated with adverse effects on the breastfed infant in a small number of reports.

Topical use of drugs on the breast should be avoided only when substantial absorption is likely. For most drugs topical use on the breast is acceptable provided the nipple and surrounding area is washed before feeding.