Many new mothers who need antidepressant or mood-stabilising drug treatment may wish to breastfeed their infants, but are hesitant to do so because of possible harmful effects of the medication on the infant. This article reviews current data on drug excretion into breast milk and the effects on the breast-fed child, and provides recommendations for the use of the different psychotropic drugs in lactating women. Relevant literature was identified through systematic searches of MEDLINE, EMBASE and the Science Citation Index Expanded (ISI) from 1966 to February 2005. The present knowledge is based on the accumulation of case studies. No randomised controlled trials in breast-fed infants have been performed and there is a lack of long-term follow-up studies. Use of SSRIs and TCAs (except doxepin) is compatible with breastfeeding. However, if treatment with an SSRI is started in the postpartum period, fluoxetine and citalopram may not be drugs of first choice. With regard to other antidepressants, such as venlafaxine, trazodone, mirtazapine, reboxetine, moclobemide and other MAOIs, very little knowledge exists. Breastfeeding should be avoided while using lithium. Carbamazepine and sodium valproate (valproic acid) are generally better tolerated by the breast-fed infant than lithium. Data on lamotrigine are still sparse. Knowledge is also scarce on the novel antipsychotics and thus recommendations in lactating women cannot be made for these agents. It is unwise to expose infants unnecessarily to drugs that may have severe adverse effects. As such, clozapine should probably be avoided because of the risk of agranulocytosis. Our knowledge of the impact of drug exposure through breast milk is still limited. Infant drug exposure is, however, generally higher during pregnancy through placental passage than through breast milk. Despite the low dosage transferred to the infant through breast milk, premature infants and infants with neonatal diseases or inherited disturbances in metabolism may be vulnerable to such exposure.