Serum Concentrations of Antidepressants and Benzodiazepines in Nursing Infants: A Case Series

DISCUSSION

Anecdotal reports of infant adverse effects are described in cases in which children have breastfed while their mothers used psychotropic medications. These reports exist across classes of compounds including benzodiazepines, tricyclic antidepressants, and older and newer SSRIs.^{42 ,43} However, systematic investigation of the safety of breastfeeding while using psychotropics is lacking.

Although all psychotropic medications are secreted into breastmilk,⁴⁴ infant absorption of these medications varies. Characteristics of a medication that determine the extent of its transport into breastmilk include molecular weight, protein binding, lipid solubility, ionization, pH, and half-life.¹⁰ Factors that affect the extent of infant exposure to medication taken by a lactating mother include individual medication characteristics, variability of maternal and infant metabolism, composition of breast milk, and timing of feedings.¹⁰ Wisner and Perel⁴⁵ describe the absence of medication in serum of 5 newborns whose mothers breastfed while taking nortriptyline, although the hydroxy metabolite of nortriptyline was noted in two of the samples (assay sensitivity <1–5 ng/mL). Wisner and colleagues¹⁹ also describe the absence of clomipramine and its metabolites in the serum of 4 infants whose mothers breastfed while taking this medication. Other investigators⁴⁶ have reported trace concentrations of SSRIs in the serum of 12 infants whose mothers used these agents while breastfeeding (assay sensitivity <1 ng/mL). One additional report²⁹ of a mother treated with sertraline while breastfeeding failed to detect drug levels in a newborn (assay sensitivity <0.5).

In the current series, medication or its metabolite was detected in 9 of 35 infants. Of these infants' mothers, 7 had been treated with fluoxetine during pregnancy and the postpartum period. Adverse effects were not reported in any of these infants. Although there have been reports of perinatal toxicity-associated fluoxetine use during pregnancy⁴⁷ and lactation,⁴⁸ no adverse effects were noted in any of the infants described in the current series. Other data supporting the absence of perinatal toxicity associated with fluoxetine use during pregnancy have been described more recently. Inconsistent findings regarding the risk for perinatal toxicity associated with use of this drug during pregnancy and/or lactation may derive from the still relatively small numbers of exposures to this drug and to differential sensitivity to drug across infants. Use of fluoxetine during pregnancy was common in our sample given the abundant reproductive safety data available for this compound compared with other SSRIs. Fluoxetine has an elimination half-life of 4 to 6 days after chronic administration, and its active metabolite (norfluoxetine) has an elimination half-life of up to 16 days after chronic administration. These long elimination half-lives suggests that the potential exists for medication to be detected in newborns with histories of antenatal drug exposure. This might be observed for several weeks after delivery and depends on individual mother and infant characteristics, as well as maternal dosage and duration of treatment before delivery. Therefore, presence of fluoxetine in newborn serum during the acute postpartum period after maternal use of the medication during the third trimester of pregnancy is not surprising given the extended half-life of fluoxetine and norfluoxetine. However, the absence of higher infant plasma concentrations of antidepressant during retesting (infants 8, 12, and 13) is of interest. The absence of infant accumulation of fluoxetine/norfluoxetine in these particular infants suggests that postnatal exposure via breastmilk did not increase plasma concentrations of the drug or its metabolite. Because fluoxetine has a longer time to drug elimination after discontinuation of this compound compared with other SSRIs with shorter duration of activity that might clear infant plasma more quickly, treatment with SSRIs, at this time, is best guided by the amount of information available regarding reproductive safety and risk for perinatal toxicity,⁴⁹ which is extensive for fluoxetine.⁴⁹

Clonazepam has a relatively long half-life and was detected in an infant (infant 17) whose mother used the medication during the third trimester of pregnancy and in whom there also was evidence of immature infant hepatic function (hyperbilirubinemia). Both prematurity and medical conditions associated with immature hepatic metabolism are assumed to increase likelihood of medication detection in infant serum when the mother has breastfed while using psychotropic medication.¹⁶ Clonazepam was not detected in the 9 other infants whose mothers used clonazepam during pregnancy and in the postpartum period. Neither fluoxetine nor clonazepam was detected in the sera of infants whose mothers were treated with these medications exclusively during the puerperium.

The current report is limited by lack of information regarding feeding schedule, as well as regarding time and duration of maternal dose. Varying infant age precluded meaningful calculation of intensity of infant exposure.⁴⁶ Data regarding the relationship between time of medication ingestion and the time of feedings also were not available. Infant serum concentrations of medication were obtained at various community and hospital-based laboratories with assays that differed in sensitivity, and the assay protocol was not standardized. Some infants were exposed to medication in utero, whereas others were not. Follow-up serum concentrations for infants who had detectable serum concentrations of medication were not always obtained. In addition, maternal serum levels of medications were not obtained to confirm compliance with prescribed regimens. Another limitation of this study was that maternal interviews regarding infant adverse effects were not standardized. Nonetheless, the current report describes the single largest series of cases in which mothers breastfed while using psychotropics and in which concentrations of medication in infant serum are described, in particular, the SSRI fluoxetine (n = 12 infants) and the benzodiazepine clonazepam (n = 11 infants). No readily noticeable infant adverse effects were reported by mothers across the sample. In addition, there was no evidence of infant accumulation of longer half-life medications such as fluoxetine or its metabolite norfluoxetine after retesting infants who were noted initially to have detectable drug (or metabolite) concentrations in serum during index screening. These data seem to support the low incidence of infant toxicity and adverse effects associated with antidepressant and benzodiazepine use during breastfeeding, particularly for fluoxetine and clonazepam which were used most commonly in our sample.

Decisions regarding whether women should breastfeed while using psychotropics must be made on a case by case basis. Properties of each medication as well as the renal and hepatic function of both infant and mother must be considered. Some investigators have suggested that immature infant hepatic function should be a relative contraindication to breastfeeding while taking psychotropic medication.¹⁶Given the limited accumulated data regarding serum concentrations of psychotropic medications noted in children whose mothers are treated with these agents for postpartum illness, no single agent seems to be particularly safer than another with respect to risk for infant toxicity. Therefore, decisions regarding choice of medication for women who suffer from postpartum illness should be guided by the likelihood that its use will result in improvement of maternal mood and functioning. Factors that may predict response to antidepressants include past or family history of response to a given medication. Lithium carbonate, which has been reported consistently to have relatively high concentrations in breastmilk, is an exception.⁵⁰ Its use during lactation is avoided frequently.

Women who are treated with psychotropic medications and who breastfeed should be counseled regarding the benefits of this activity and the small known risk of severe infant adverse effects as well as the unknown, but possible, impact on brain development. Women who choose to breastfeed while taking psychotropic medication should work collaboratively with a psychiatrist and a pediatrician. Infant serum should be assayed for the presence of medication after ∼2 to 4 weeks. For women who have been treated with longer acting psychotropics such as fluoxetine during pregnancy, testing may be more informative at 6 to 8 weeks postpartum. This extension of the time of testing minimizes the likelihood of detecting residual medication (or its metabolite) derived from maternal use of the agent during pregnancy.

Most assays available from commercial and hospital-based laboratories have sensitivities ranging from 2 to 50 ng/mL depending on the medication being tested and the laboratory technique used. Pending greater routine availability of assays with heightened sensitivity compared with those used by most commercial and hospital-based laboratories, clinicians may use currently available assays foremost as a measure of potential toxicity. Although the results of these assays may not confirm an absolute presence or absence of medication in the infant, they may be helpful in ruling out situations in which there is idiosyncratic significant secretion of medication into breastmilk or an inability of the infant to metabolize psychotropic medication.

Whether women choose to stop breastfeeding when laboratories with extremely sensitive assays (ie, 1–2 ng/mL) detect small amounts of medication is a decision to be made collaboratively between the patient and the physician. The impact of untreated maternal depression on infants is well described and must be factored critically into any decision regarding the treatment of postpartum mood disorder. However, the possibility that chronic exposure to even very low doses or trace amounts of psychotropic medications during the infant period might affect longer term neurobehavioral development must also be considered and must be balanced with the wishes and needs of mothers and families. The latter issue is an important one. Currently, data are not available to support the conclusion that neonatal exposure to trace amounts of any particular medication is associated with different effect on the outcome variable of ultimate interest, namely long-term neurobehavioral function. Even if it could be demonstrated that maternal use of a particular SSRI was associated consistently with lower infant plasma levels of the medication, how the variable effects of these agents at the receptor translate into measurable differences in behavior is a subject about which it is only possible to speculate at this time. Controlled, prospective neurobehavioral evaluation of infants of mothers who choose to breastfeed while taking psychotropic medications and for whom extent of exposure is well documented is needed and ultimately may help to refine our understanding of the risks and benefits of psychotropic medication use in postpartum women who wish to breastfeed.