Background The pharmacological treatment of depression in nursing women requires information on the magnitude of medication exposure to the infant that may occur through breast milk.
Aims To examine serum concentrations of antidepressants in infants exposed to these medications through breast-feeding.
Method Maternal and infant serum concentrations of sertraline, paroxetine and fluvoxamine were determined with high-performance liquid chromatography (limit of detection=1 ng/ml).
Results No detectable medication was present in any infant exposed to paroxetine (n=16) or fluvoxamine (n=4). Among infants exposed to sertraline (n=30), detectable medication was present in 24% of serum samples. A significant negative correlation was found between infant age and infant serum concentration. Sertraline was significantly more likely to be detected in an infant if the mother's daily dose was 100 mg or higher. No adverse sequelae occurred in any infant.
Conclusions This study shows that paroxetine, fluvoxamine and sertraline produce minimal exposure to infants when taken by nursing mothers.
When new mothers experience depression, they and their clinicians need to know if antidepressant medications are safe to use while breast-feeding. Several recent reports have examined the extent of exposure that occurs to nursing infants whose mothers take antidepressants (e.g. Altshuler et al, 1995; Taddio et al, 1996; Mammen et al, 1997; Stowe et al, 1997; Yoshida et al, 1997; Wisner et al, 1998; Begg et al, 1999; Birnbaum et al, 1999; Kristensen et al, 1999; Ohman et al, 1999; Stowe et al, 2000). However, the research literature has consisted largely of single cases or small studies, and methodological differences have limited the information that can be drawn from them (Yoshida et al, 1999). To expand this literature, we present measurements of maternal and infant serum concentrations in 50 nursing mother—infant pairs in which the mothers took therapeutic doses of sertraline, paroxetine or fluvoxamine.
Fifty nursing mother—infant pairs who presented to UCLA's Pregnancy and Postpartum Mood Disorders Program were included in the study. The women were Caucasian (n=46), Hispanic (n=3) and Asian (n=1), in good physical health, ranging in age from 24 to 41 years old and on standard doses of antidepressant medication, taken once daily for treatment of major depression. Two women additionally were on nortriptyline and another was on alprazolam. None of the infants was on medications of any category. Seventeen women were on antidepressants during pregnancy as well as nursing. Written informed consent was obtained from each mother for collection of serum samples.
Maternal and infant serum samples were obtained at a minimum of 2 weeks following a fixed dose of antidepressant. For women who took the antidepressant during pregnancy, maternal and infant serum samples were obtained a minimum of 2 weeks following delivery. Serum samples were obtained from a total of 50 infants and 48 mothers.
Detection of antidepressant concentrations in serum
The detection of sertraline, desmethylsertraline, paroxetine and fluvoxamine in serum was accomplished via an isocratic high-performance liquid chromatography (HPLC) separation followed by ultraviolet detection at 225 nm. The concentration of each drug (sertraline, desmethylsertraline, paroxetine or fluvoxamine) in the samples was calculated from its peak area ratios by using the slope and intercept of the appropriate calibration curve. The assays had a lower limit of sensitivity of 1 ng/ml, defined by a signal-to-noise ratio of 7 for each drug.
The LIFEREG Procedure using SAS software was used to perform a Tobit analysis on the data. The Tobit model is a regression model for left-censored data, assuming a normal distributed error term. The model parameters are estimated by maximum likelihood. A χ2 test was used to explore whether nursing infants whose mothers took higher daily doses of sertraline (100 mg or more) were more likely to have detectable serum concentrations of medication, as compared with infants of women who took lower doses.
Results of the serum assays for the mother-infant pairs are shown in Table 1. Infant ages ranged from 2 to 60 weeks and weights ranged from 3 to 10 kg at the time of the serum assays.
No detectable medication was present in any infant exposed to paroxetine or fluvoxamine. Detectable medication (parent and/or metabolite) was present in 24% (8/33) of the serum samples obtained from infants exposed to sertraline. Concentrations of sertraline and desmethylsertraline, when present, were 2-8 and 2-12 ng/ml, respectively.
Four mothers (nos 5, 16, 24 and 49) titrated their dose of the antidepressant upwards to help their mood and obtained repeat serum samples on themselves and their infants after being on the higher medication dosage for at least 1 week.
Maternal dosage of sertraline correlated highly with infant serum concentration of desmethylsertraline after controlling for infant age, gestational exposure and breast-feeding exposure (parameter estimate=0.09, d.f.=1, P=0.03). Maternal serum concentrations of sertraline and desmethylsertraline correlated highly with infant serum concentration of desmethylsertraline (parameter estimate=0.20, d.f.=1, P<0.001 and parameter estimate=0.07, d.f.=1, P=0.008, respectively) after controlling for infant age, gestational exposure and breast-feeding exposure. This analysis used all the available maternal and infant serum samples shown in Table 1.
A significant negative correlation was found between infant age and infant serum concentration of desmethylsertraline after controlling for maternal dosage, gestational exposure and breast-feeding exposure (parameter estimate=-1.46, d.f.=1, P=0.002). Among women who breast-fed fully, the likelihood of their infants having a detectable level of medication (sertraline or desmethylsertraline) was significantly higher if their dose was 100 mg or more (χ2=6.81, d.f.=1, P=0.009).
Mothers were questioned about potential adverse sequelae to their infants and did not report any such findings. Specific enquiries were made regarding gastro-intestinal symptoms (e.g. vomiting, watery stool), lethargy, changes in sleep patterns and easy bruising. None of the women in the study was on other medications and the infants were in good health.
This study found that serum concentrations of medication were undetectable in all infants exposed to paroxetine or fluvoxamine and in the majority of infants exposed to sertraline while nursing. When medication was present in the sertraline-exposed infants, it was usually in the form of the metabolite desmethylsertraline. Maternal serum concentrations of sertraline and desmethylsertraline correlated highly with infant serum concentrations of desmethylsertraline. Maternal dosage of sertraline also correlated highly with infant serum concentrations of desmethylsertraline; doses of 100 mg or above were significantly more likely to produce detectable concentrations in the infant. A significant negative correlation emerged between infant age and serum concentration of desmethylsertraline.
This study's findings suggests that paroxetine, fluvoxamine and sertraline are reasonable choices for nursing women requiring treatment for depression. In comparison with fluoxetine, these medications appear to produce less exposure to nursing infants and have not been linked with the adverse events of neonatal irritability, sleep disturbance and poor feeding that have been reported in association with fluoxetine exposure through breast-feeding (Lester et al, 1993; Brent & Wisner, 1998; Chambers et al, 1999; Kristensen et al, 1999). For infants that are healthy and full-term, these findings provide no reason to discourage nursing among women taking paroxetine, fluvoxamine or sertraline at standard therapeutic dosages. The use of additional medications that are commonly taken in the post-partum period (e.g. antihistamines, decongestants, pain medications) should be kept to a minimum until more is known about whether such combinations are safe for the nursing infant (Mitchell, 1999).
Clinical Implications and Limitations
This study found that the use of fluvoxamine, paroxetine and sertraline by nursing women produces minimal medication exposure to the infants.
The presence of low detectable serum concentrations of medication was not associated with adverse effects in the infants.
Maternal serum concentrations and dosage of medication can be employed to estimate infant serum concentrations.
Use of maternal reports rather than paediatric examinations in assessing for potential medication-related adverse events in the infants.
Single serum measurement of medication in the mother—infant pairs.
Lack of evaluation of infants' long-term outcomes associated with their early exposure to antidepressant medications.
This project was supported by grant no. MH0145101 from the National Institute of Mental Health, Bethesda, Maryland. Preparation of this report was supported in part by Smith Kline Beecham and Pfizer.
- Received November 22, 2000.
- Revision received March 15, 2001.
- Accepted March 15, 2001.
- © 2001 Royal College of Psychiatrists