The use of psychotropic medications during lactation has not been investigated in a controlled and systematic fashion. The literature is laden with case reports and small case series containing numerous confounds that render the establishment of definitive treatment guidelines tenuous. The increasing number of women who plan to breast-feed and the high rate of psychiatric illness during the postpartum period underscore the need to develop such guidelines. A MEDLINE search was conducted for key words either in the titles or abstracts of publications citing the use of psychotropic medications in lactating women and describing the pharmacokinetics of medication excretion into breast milk. The publications identified span over three decades. The largest single study by one group of investigators examined 12 mother-infant pairs. The majority of studies report their results as a ratio of the breast milk concentration to the maternal serum concentration (milk/plasma [M/P]) ratio. Estimations that use the M/P ratio of the infant daily dose range from 0.1% to 6.2% of the maternal dose. Few studies attempt to account for the complex variations in the maternal, breast milk, and infant physiologic environments. The major confounds of the studies reviewed include (1) failure to document portion of breast milk assayed (foremilk versus hindmilk), (2) limited metabolite assay, (3) limited assay sensitivity (1-25 ng/mL), not of research quality, (4) concomitant maternal and/or infant medications, and (5) medication exposure during pregnancy. Despite these confounds, there are remarkably few reports of adverse effects on nursing infants exposed to psychotropic medications in breast milk. The limited data confirm that psychotropic medications are excreted into breast milk and that the infant is exposed to these medications. The ideal breast milk study that accounts for the confounds identified has not been completed. The complex matrix of breast milk and the changing infant metabolic capacity will require a more detailed analysis with assays of improved sensitivity. Despite the limited reports of adverse effects on nursing infants, the limitations of the available literature and minimal sample sizes make it premature to recommend specific medications from a given class. There is inadequate data on nursing infant exposure to multiple medications to support changing medication to a different agent in an otherwise stable patient. An individualized risk/benefit assessment with the empirical goal of minimizing infant exposure while maintaining maternal emotional health is the ideal approach.