Methods
Drugs considered for the present review have been selected as the most frequently used among those categorized as major antiepileptic (Code N03) by the Anatomical Therapeutic Chemical Classification System (ATC) [
3] together with the anxiolytic diazepam (N05BA01). Before assessing the lactation risk of each AED, we have collected information on their main pharmacokinetic parameters: plasma protein binding, half-life, milk-to-plasma ratio, oral bioavailability (see Table
1 for definitions). We decided to present data on AEDS pharmacokinetic parameters as they represent the theoretical basis on which the lactation risk assessment should be developed. As an example, we can assume that if a drug has a short half-life (<3 hours), its level in the maternal plasma will be declining when the infant feeds again, considering a 3-hour interval. Moreover, if the drug is highly protein bound, it cannot enter the milk compartment easily. Milk/plasma ratio has the primary use of quantifying the extent of drug transfer into the milk; nevertheless, its use in assessing a lactation risk is limited as the amount of drug transfer into milk is mainly determined by the maternal plasma level. Medications for the breastfeeding mother should have a low oral bioavailability, as the result of either a poor gut absorption, or the liver sequestration prior to entering the plasma compartment. Table
2 summarises each AED main pharmacokinetic characteristics.
Table 1
Definitions and clinical relevance of the pharmacokinetic parameters used to assess the lactation risk following maternal intake of medications
Half-life or “T ½”
| The half-life of a substance is the time it takes for its plasma concentration to halve. If the half-life is long (>12-24 hrs), drugs may accumulate in maternal plasma and the time of drug transfer from plasma to breast milk is longer. |
Maternal plasma protein binding (PB)
| This parameter is expressed in percentage. The higher the percentage of the drug bound to the maternal plasma proteins, the less the drug passes into breast milk. An ideal drug to be taken during breastfeeding should have a plasma protein binding > 80%. |
Milk-to-plasma ratio (M/P)
| It denotes the ratio of the drug concentration in the mother’s milk (M) divided by its concentration in the mother’s plasma (P). It is an indicator of drug transfer into breast milk. A M/P ratio greater than 1.0 suggests that the drug may be present in breast milk in high concentrations. |
Oral bioavailability
| It describes the fraction of one orally administered dose of a drug that reaches the systemic circulation. It is expressed as a percentage of the administered dose. When the intestinal absorption of a drug is impaired, the risk of adverse effects may be lower. |
Table 2
Main pharmacokinetic characteristics of antiepileptic drugs
Carbamazepine | 74 | 100 | 18 – 54 | 0.69 |
Clonazepam | 50 – 86 | 100 | 18 – 50 | 0.33 |
Diazepam | 99 | 100 | 43 | 0.2 – 2.7 |
Ethosuximide | NA | 100 | 30 – 60 | 0.94 |
Gabapentin | < 3 | 50 – 60 | 5 – 7 | 0.7 – 1.3 |
Lamotrigine | 55 | 98 | 29 | |
Levetiracetam | < 10 | 100 | 6 – 8 | |
Oxcarbazepine | 40 | 100 | 9 (oxcarbazepine metabolyte) | 0.5 |
Phenobarbital | 51 | 80 – 100 | 20 – 133 | 0.4 – 0.6 |
(45–500 in newborns) [ 41] |
Phenytoin | 89 | 70 – 100 | 6 – 24 | 0.18 – 0.45 |
(20–160 in preterm infants) [ 48] |
Pregabalin | NA | 90 | 6 | NA |
Primidone | 25 | 90 | 5 – 18 | 0.72 |
Tiagabine | 96 | 90 | 7 – 9 | NA |
Topiramate | 15 | 75 | 18 – 24 | |
Valproate | 94 | 100 | 14 | 0.42 |
Vigabatrin | NA | 50 | 7 | < 1 |
Zonisamide | 40 | NA | 63 | 0.93 |
Beyond presenting pharmacokinetic parameters on AEDs, we chose to collect more relevant clinical parameters such as the theoretical infant dose (TID), the therapeutic dose in the neonatal period and the relative infant dose (RID). All these parameters were reported in a table (Table
3) together with the assessment of the lactation risk (see below), in order to have a clinical overview of each AED.
Table 3
Clinical relevant parameters for each antiepileptic drug and assessment of their lactation risk
Carbamazepine | 0.7 | 10 - 20 | 3.8 - 5.9 | L2 | ▪ CBZ levels are relatively high in breast milk | Safe |
▪ Breastfed infants have serum levels that are usually below the therapeutic range. |
▪ Side effects were rarely reported as sedation, decreased suckling, withdrawal reactions and 3 cases of liver dysfunction. |
▪ Infant should be monitored for jaundice, drowsiness, adequate weight gain, and developmental milestones especially in premature infants, exclusively breastfed and in combination with other antipsychotics. |
Clonazepam | 0.002 | 0.1 - 0.2 | 2.8 | L3 | ▪ Monitor growth, sedation, developmental milestones, especially in preterm neonates, exclusively breastfed infants and if mother is receiving psychotropic drugs. | Contraindicated |
▪ Monitoring of serum concentration in breastfed infant, if excessive sedation occurs. |
Diazepam | 0.05 | IV dose available: | 7.1 | L3 | ▪ Accumulates in maternal milk and serum of breastfed infant. Other agents are preferred, especially while nursing a newborn or preterm infant. | Contraindicated |
0.1- 0.3 |
Oral dose |
▪ Single dose does not require delaying feeding. |
|
Ethosuximide | 11.5 | 15 – 40 | 31.4-73.5 | L4 | ▪ Monitor infant for drowsiness, adequate weight gain and psychomotor development. | Contraindicated |
▪ Measurement of an infant serum level might help rule out toxicity, if there is a concern. |
Gabapentin | 1.7 | Only paediatric dose available: | 1.3 - 6.6 | L2 | ▪ Monitor the infant for drowsiness, adequate weight gain, and developmental milestones, especially in younger, exclusively breastfed infants and when using combinations of anticonvulsant or psychotropic drugs. | Moderately safe |
10-15 |
Lamotrigine | 0.7 | 1-6 with valproate, 5-15 with enzyme inducing AEDs | 9.2 | L3 | ▪ It is not necessary to discontinue breastfeeding, but any adverse effects such as apnoea, rash, drowsiness, decreased sucking are to be monitored and serum levels are to be measured. | Moderately safe |
▪ Monitoring of the platelet count may also be advisable. |
Levetiracetam | 3.9 | | 3.4 - 7.8 | L3 | ▪ Monitor infant for the appearance of sleepiness, increase appropriate weight, normal psychomotor development. | Moderately safe |
Oxcarbazepine | NA | 27.7 – 50 | 1.5-1.7 | L3 | ▪ Monitor the infant for drowsiness and decreased feeding, and developmental milestones especially in the first 2 months of life. | Moderately safe |
|
Phenobarbital | 0.4 | 3-4 | 24 | L3 | ▪ The presence of phenobarbital in breast milk may mitigate possible neonatal abstinence. | Safe |
▪ Monitor the breastfed infant for the possible onset of drowsiness, adequate weight gain and developmental milestones, especially in younger, exclusively breastfed infants and antiepileptic polytherapy. |
▪ Measurement of the infant’s serum drug concentration might help rule out toxicity. |
Phenytoin | 0.4 | 5-8 | 0.6-7.7 | L2 | ▪ The proportion ingested by infants is small and generally brings about no problems except in rare cases of idiosyncratic reactions. | Safe |
Pregabalin | NA | | NA | L3 | ▪ Compatible with breastfeeding. | Moderately Safe |
▪ An alternate drug may be preferred, especially while nursing a newborn or preterm infant. |
Primidone | 0.9 | 12-20 | 8.4-8.6 | L3 | ▪ The presence of phenobarbital in breast milk may mitigate possible neonatal abstinence. | |
▪ Monitor the breastfed infant for the possible onset of drowsiness, adequate weight gain and developmental milestones, especially in younger, exclusively breastfed infants and antiepileptic polytherapy. |
Safe |
▪ Measurement of the infant’s serum drug concentration might help rule out toxicity. |
Tiagabine | NA | <12 years: limited data available. | NA | L3 | ▪ Monitor the infant for the onset of drowsiness, for adequate weight gain and for developmental milestones especially in younger, exclusively breastfed infants and when using combinations of anticonvulsant or psychotropic drugs. | Moderately safe |
▪ Other drugs should be preferred especially while nursing a newborn or preterm infant. |
Topiramate | 0.3 | 1 - 6 | 24.5 | L3 | ▪ Monitor the infant for the onset of diarrhea, drowsiness, increase appropriate weight and psychomotor development. | Moderately safe |
|
Valproate | 0.7 | Limited data available in the neonatal period. | 1.4-1.7 | L3 | ▪ Breastfed infants are at risk for hepatotoxicity. | Safe |
▪ Monitor the infant for unusual bleeding (a case of thrombocytopenia has been reported). |
Vigabatrin | 0.1 | 25-50 | 1.5 – 2.7 | L3 | ▪ Until more data are available, vigabatrin should only be used with careful monitoring during breastfeeding. | Moderately safe |
S-enantiomer
|
Zonisamide | 1.9 | 5 – 8 | 28.9 -36.8 | L4 | ▪ Monitor infant for drowsiness, adequate weight gain and psychomotor development. | Contraindicated |
| | | | | ▪ Measurement of an infant serum level might help rule out toxicity if there is a concern. | |
TID is the maximum estimated amount of ingested drug with breast milk; in other words, it is the estimation in milligrams per kilogram per day of the theoretical infant dose (TID). We calculated it by using the formula by Atkinson (TID = daily breast milk intake (150 ml/kg/day) × maximum breast milk concentration of medication). [
4] As AEDs may be used in the neonatal period as therapy, we presented their therapeutic dose, expressed in mg/kg/day in order to have a direct comparison with the TID. A therapeutic dose which is higher than the TID reassures of the drug safety during breastfeeding. RID represents one of the most useful parameters for assessing the lactation risk. RID is calculated by dividing the infant dose via milk in “mg/kg/day” by maternal dose in “mg/kg/day”, assuming a 70 kg weight. [
5] Many authors agree that anything less than 10% of the maternal dose is considered probably safe [
5].
To review the lactation risk of AEDs, we consulted the following two most accredited English sources: Medications and Mother’s Milk 2012, a Manual of Lactational Pharmacology [
1], and the 2013 Lactmed database (in TOXNET) [
6] of the National Library of Medicine. In his textbook, updated every two years, Hale collects data on many current medications and their use during breastfeeding. After evaluating information on pharmacokinetics and what is currently published in the scientific literature for each drug, including its reported side effects, he makes a personal recommendation, using a 5 categories of lactation risk: L1: safe drugs at the highest level, L2: safe drugs, L3: moderately safe drugs, L4: drugs possibly dangerous, L5: contraindicated drugs. LactMed database is part of the National Library of Medicine’s Toxicology Data Network (TOXNET), and it includes information on the levels of drugs and other chemicals to which breastfeeding mothers may be exposed in breast milk and infant blood, together with the possible adverse effects in the nursing infant. All data published in LactMed are derived from the scientific literature and fully referenced.
To complete our synopsis, we have also performed a non-systematic Medline search of the literature with the keywords “antiepileptic drugs” AND “breastfeeding”, retrieving studies from 2004 to April 9
th 2013 , including the most relevant and up-dated studies on lactation risk. Data from the Committee on Drugs of the American Academy of Pediatrics (AAP) [
7], and the Goodman & Gilman’s Textbook of Pharmacology [
8] were also included in our review.
For each AED we organized the relevant data into small summaries.
As the result of our review we classified AEDs in 3 categories: safe, moderately safe and contraindicated during breastfeeding. The moderately safe category has a less documented safety profile due to a short clinical experience and lack of studies. However, the moderately safe AEDs can be used with caution. The lowest dose of the drug should be chosen and the nursing infant should be monitored and, when possible, his/her plasma level should be checked.
Discussion
The use of AEDs is often unavoidable for most mothers with epilepsy. Moreover, the post partum period is a vulnerable time for women with epilepsy, owing to the changes in the drug metabolism and possibly to sleep deprivation. Both these conditions may trigger an epileptic crisis [
65].
As the newborn might be indirectly exposed to AEDs via breast milk, pharmacologically treated epilepsy has been and sometimes is still considered a contraindication to breastfeeding, irrespective of which AED is taken by the mother [
66]. On the contrary, health professionals have currently become more aware of the need to weigh the short and long-term risks for the nursing infant against the well demonstrated nutritional, immunological, developmental, economic and ecological benefits of breastfeeding [
1,
2,
67]. When evaluating the safety profile of an AED during lactation, the physician should collect and process data which are often dispersed in the scientific literature and difficult to retrieve [
66]. We have therefore completed the present review in order to facilitate the consultancy on the use of AEDs during breastfeeding.
The comparison of advices from the main and most recent accredited literature sources has highlighted the existence of a certain degree of inconsistency [
66], which is probably the consequence of the combination of limited scientific evidence and of a certain degree of arbitrariness. This heterogeneity of advice is not specific of AEDs. We have previously documented a metavariability in the assessment of the lactation risk of other classes of drugs, such as beta-blockers [
68], corticosteroids [
69] and antidepressants [
70].
Evaluating the lactation risk of any drug is complex and should take into account several aspects. In our review, we have shown that neither pharmacokinetic information nor more clinical parameters (such as TID and RID) are alone good predictors of the lactation risk. As an example, a drug such as phenobarbital, which is poorly bound to the mother plasma proteins and has a relatively long half-life and an high RID, can still be considered compatible during breastfeeding [
1].
When evaluating the lactation risk of a drug, we should always check whether or not toxicity in breastfed infants is reported. AEDs are expected to determine in the nursing infant a series of symptoms related to their pharmacological effects on the central nervous system, such as altered sleep patterns, abnormal tone, poor feeding and possibly poor growth secondary to inadequate breast sucking. However, the clinical relevance of these attributed side-effects vary widely. As an example, the lactation risk of infant apnea after maternal use of clonazepam [
11] deserves a higher level of caution than the risk of infant sedation after maternal use of CBZ.
Another variable to take into account when considering the infant tolerance to AEDs, is his/her ability to metabolize the drug. This is much lower when the maternal treatment begins before or during pregnancy (as the exposure of the baby through breast milk adds to the one through the placenta), or in case of prematurity and during the first 2 months of life [
71]. Maternal polytherapy is also expected to increase the lactation risk, as the result of the possible synergistic interaction of different medications [
39].
We might also speculate that the alleged long-term negative effects on child psychomotor development resulting from the passage of small amounts of the drug through breast milk, though up to now not well-documented, could be well-balanced by the benefits that breastfeeding confers to the cognitive, social and relational development. This hypothesis seems to be supported by the results from a study conducted in the US on 199 children exposed in uterus to AEDs (phenytoin, carbamazepine, valproate and lamotrigine) and followed in the first 3 years of life. In this cohort, there was no significant difference regarding cognitive scores between breastfed and not breastfed children [
24].
Breastfeeding mothers should be provided sound and clear information on the lactation risk of the prescribed medications. Nevertheless, breastfeeding mothers taking AEDs happen to receive inconsistent and sometimes conflicting advices on whether or not to breastfeed from different clinicians (general physician, neurologist, pediatrician, obstetrician, etc.). Unluckily, many such given advices are not evidence-based, but simply emphasize not circumstantial assessments of the lactation risk.
We believe that mothers treated with AEDs should be encouraged to start breastfeeding immediately after childbirth, even if a definite and complete evidence-based information has yet not been collected. In fact, breast milk should not be considered as a “presumed guilty”, because, after a complete scrutiny, only very few AEDs resulted contraindicated during breastfeeding [
24,
72,
73].
The present study is not without limitations. First of all, we did not conduct a systematic review of the published literature, but we offered a synopsis of the two most updated and authoritative clinical sources on the lactation risk (namely: Hale and Toxnet). Secondly, the existing reports on the lactation risks of medications in general [
73], and particularly of AEDs [
44] , are usually anecdotal or based on case series and therefore of poor methodological quality. However, we should remember that any recorded side-effect, while biologically and pharmacologically plausible in the nursing infant, is rarely attributable with certainty to a particular drug taken by her/his mother [
74].
Conclusions
According to the present review, we grouped AEDs into 3 main classes of lactation risk:
1.
AEDs safe to use during lactation. Carbamazepine, valproic acid and phenytoin are compatible with breastfeeding. Phenobarbital and primidone are also compatible, although particular attention should be paid on infant monitoring, possibly measuring the infant plasma drug level.
2.
AEDs with a less documented safety profile during lactation (moderately safe). Gabapentin, lamotrigine, levetiracetam, oxcarbazepine, pregabalin and vigabatrin may be used during lactation, but their lowest dose should be prescribed. The nursing infant should be carefully monitored and, if required, the infant drug plasma-level tested. When lamotrigine is taken by the mother, it is advisable to check the presence of thrombocytopenia in the nursing infant. Although based on scarce available data, tiagabine and topiramate can be considered compatible with breastfeeding.
3.
AEDs contraindicated during lactation. Ethosuximide should be used only when there are no alternatives and after informed consent of the mother. Lastly, zonisamide and the continuous use of diazepam or clonazepam should be avoided.
Even if an AED can be safely taken by the breastfeeding mother, it is a good clinical practice to call the mother attention on the behavior, sleep, feeding patterns and growth of the nursing infant, especially in the first 2 months of life [
75]. When possible, we might advice to monitor the infant plasma levels, especially for the moderately safe AEDs. However, we are aware that there is no clear agreement on when to test them. We believe that after 4–8 weeks is a reasonable time window, as, if the child is fully breastfed, breast milk would have reached its highest production and the consequent intake by the nursing infant would be at the maximum.
Competing interests
This study was performed without any funding or grants. All the authors declare that they have no competing interest and do not have any financial relationships with any biotechnology and/or pharmaceutical manufacturers.
Authors’ contributions
RD conception and design. LM, SDB, RD analysis and interpretation of the data. LM, SDB RD drafting of the article. RD critical revision of the article for important intellectual content. All authors listed here have seen and approved the final version of the report. JB administrative, technical, or logistic support. MC, EZ collection and assembly of data.