Key points

Introduction

While rates of breastfeeding early postpartum are greater than 74.6%, the US Centers for Disease Control and Prevention’s 2012 breastfeeding report card found breastfeeding at 6 months was 47.2%, and breastfeeding at 12 months was 25.5%. Although there are many social factors that lead to this high rate of discontinuation, the use of medications must be considered. The average number of different medications (excluding iron, minerals, folic acid, vitamins) taken per mother in a small American study was 4 throughout lactation (0.9 medications per month).

With so many women using drug therapy during lactation, pediatricians and obstetricians are faced with the challenge of determining which medications are suitable for breastfeeding mothers. Although there is more literature available about the transfer of medications into breast milk, this is often not communicated to students and clinicians; therefore, many women are advised to stop breastfeeding or avoid drug therapy based on information obtained from product monographs.

Even without specific medication data from human studies, a good understanding of the kinetic principles and mechanisms of medication entry into breast milk can help a clinician make an informed decision that often allows the mother to continue breastfeeding while treating her medical condition. This article discusses the most important concepts needed to understand how medications enter breast milk to aid in clinical decisions and highlights suitable medications for breastfeeding mothers.

Introduction

While rates of breastfeeding early postpartum are greater than 74.6%, the US Centers for Disease Control and Prevention’s 2012 breastfeeding report card found breastfeeding at 6 months was 47.2%, and breastfeeding at 12 months was 25.5%. Although there are many social factors that lead to this high rate of discontinuation, the use of medications must be considered. The average number of different medications (excluding iron, minerals, folic acid, vitamins) taken per mother in a small American study was 4 throughout lactation (0.9 medications per month).

With so many women using drug therapy during lactation, pediatricians and obstetricians are faced with the challenge of determining which medications are suitable for breastfeeding mothers. Although there is more literature available about the transfer of medications into breast milk, this is often not communicated to students and clinicians; therefore, many women are advised to stop breastfeeding or avoid drug therapy based on information obtained from product monographs.

Even without specific medication data from human studies, a good understanding of the kinetic principles and mechanisms of medication entry into breast milk can help a clinician make an informed decision that often allows the mother to continue breastfeeding while treating her medical condition. This article discusses the most important concepts needed to understand how medications enter breast milk to aid in clinical decisions and highlights suitable medications for breastfeeding mothers.

Key concepts of medication entry into breast milk

Although all medications enter milk to some degree, clinically relevant levels are seldom attained. Most drugs simply transfer in and out of the milk compartment by passive diffusion from a region of high concentration to a region of low concentration. Some active transport systems exist for immunoglobulins, electrolytes, and particularly iodine, but facilitated transport systems are limited. Fewer than 10 drugs are known to be selectively transported into human milk.

Medications that enter breast milk often have certain physicochemical characteristics. They are generally low in molecular weight (less than 500 Da), they often attain higher maternal plasma levels, they are generally poorly bound to plasma proteins, and they have a higher pK _a (pH at which a drug is equally ionic and nonionic; polar or ionic medications are less likely to leave the breast milk compartment). Human milk has a low pH (7–7.2) which causes some medications with a higher pK _a (>7.2) to become ionized and trapped in milk.

In addition, clinicians also need to consider the oral bioavailability of the drug in the infant’s gastrointestinal tract. Many drugs are simply not absorbed in the gastrointestinal tract of infants. The stage of lactation is important. More medication can enter breast milk in the colostral phase, but only minimal doses are transferred to the infant during this phase because of the limited volume of colostrum. With mature milk, there is a larger volume but less medication enters breast milk because of tight cell to cell junctions.

Calculating infant exposure

Perhaps the most useful tool in clinical practice is to calculate the actual dose received by the infant. To do this, the actual concentration of medication in the milk and the volume of milk transferred must be known. Although not always available, data on milk levels are available for many drugs. More recent studies now calculate the average area under the curve (AUC) value for the medication (C _ave ). This methodology accurately estimates the average daily level of the drug in milk, and hence the average intake by the infant. Although the actual volume of milk ingested is highly variable and depends on the age of the infant and the extent to which the infant is exclusively breastfed, many clinicians use a value of 150 mL/kg/d to estimate the amount of milk ingested by the infant. The most useful and accurate measure of exposure is to calculate the relative infant dose (RID) as shown in Fig. 1 .

Calculating the RID.

The RID is generally expressed as a percentage of the mother’s dose, and it provides a standardized method of relating the infant’s dose in milk to the maternal dose. Bennett was the first to recommend that an RID of more than 10% should be the theoretical level of concern for most medications. However, the 10% level of concern is relative and each situation should be evaluated individually according to the overall toxicity of the medication.

Unique infant factors

To evaluate the risk of the medication, infants should be categorized as low, moderate, or high risk. Infants at low risk are generally older (6–18 months), receive lower volumes of breast milk, and are able to metabolize and handle drugs more efficiently. Mothers in the terminal stage of lactation (>1 year) often produce relatively lower quantities of milk. Thus, the absolute clinical dose transferred is often low.

Infants at moderate risk are term infants who are aged between 2 weeks and 6 months. Those at higher risk are premature, newborns, or infants with acute or chronic medical conditions that may be affected by the medication or may impair the clearance of medications in the infant (eg, renal dysfunction).

Psychiatric conditions

Recent data from 17 American states indicate that postpartum depression affects 12% to 20% of women. Fortunately for practitioners, there is increasing information available about the use of antidepressants during lactation that support the treatment of the condition while breastfeeding ( Table 1 ). The selective serotonin reuptake inhibitors (SSRIs) are presently the mainstay of antidepressant therapy in women who are breastfeeding. Table 1 provides the RID for common SSRIs. Clinical studies in breastfeeding patients consuming sertraline, fluvoxamine, and paroxetine clearly suggest that the transfer of these medications into human milk is low and uptake by the infant is even lower. Thus far, no or minimal untoward effects have been reported following the use of these 3 agents in breastfeeding mothers. Sertraline is overwhelmingly favored because more than 50 mother-infant pairs have been evaluated in numerous studies, and milk and infant plasma levels are low to undetectable.

Fluoxetine has also been studied in more than 50 mother-infant pairs, and transfers into human milk at concentrations as high as 9% of the maternal dose. Because of its long half-life and active metabolite, clinically relevant plasma levels in infants have been reported, but without major complications. Long-term studies, however, are not yet available. Because of a higher RID, fluoxetine is perhaps less preferred unless lower doses are used during pregnancy and early postpartum. However, in reality, the incidences of untoward effects are probably remote, and mothers who cannot tolerate other SSRIs should be maintained on the product that works best for them. In essence, if a product works, it is not advisable to change breastfeeding mothers to another product. Although almost all tricyclic antidepressants produce low RIDs and are well tolerated by the infant, they are seldom used due to intolerable anticholinergic side effects in the mother.

Benzodiazepines are often used with antidepressants to help with anxiety or can be used as sleep aids for short periods (when non-drug measures have failed). Kelly and colleagues conducted a study to determine adverse event rates in infants exposed to benzodiazepines via breast milk. The 3 most commonly used benzodiazepines in this study were lorazepam (52%), clonazepam (18%), and midazolam (15%). Of 124 women taking benzodiazepines only, 1.6% (2 of 124) of their infants (2–24 months old) had depression of the central nervous system. There was no correlation between sedation and the benzodiazepine dose or duration of breastfeeding. Of the 2 mothers who reported sedation, 1 was taking alprazolam occasionally and the other was taking 2 benzodiazepines (clonazepam and flurazepam) chronically. If these agents are used, choose a product with a short half-life and use the lowest effective dose for the shortest duration to minimize exposure.

Other medications used for sedation include first-generation antihistamines; the most commonly used medications are diphenhydramine, dimenhydrinate, and doxylamine. Dimenhydrinate’s active ingredient is diphenhydramine and doxylamine has a similar structure to diphenhydramine; therefore, the RID for diphenhydramine (0.7%–1.5%) is often extrapolated to all 3 of these drugs and it is believed that none of them readily enter breast milk.

There are 2 medications that are not part of the benzodiazepine family but are indicated for insomnia: zopiclone and zolpidem. A study of 12 mothers who took zopiclone 7.5 mg orally, 2 to 6 days postpartum for sleep, found that the RID was about 1.4% of the maternal dose and no adverse effects were reported in the infants. Monitoring the infant for sedation and the ability to breastfeed is recommended with all sedating medications.

Atypical antipsychotics are another class of medications that are being used more commonly for many disorders such as psychosis, bipolar disorder, depression, and so forth. Three of the most commonly used atypical antipsychotics are risperidone (RID 4.3%), quetiapine (RID 0.09%), and olanzapine (1.6%). These medications have low RIDs and are believed to be more suitable during breastfeeding than the older antipsychotics (phenothiazines), which have been associated with drowsiness and lethargy.

Methylphenidate (RID 0.2%) and dextroamphetamine (RID 5.7%) have relatively low penetration into breast milk and no history of causing adverse effects in breastfed infants. However, the infants of all breastfeeding mothers on stimulants require monitoring for irritability, weight loss, or poor weight gain.

Clinicians are often faced with the decision of whether or not antiepileptic or mood stabilizer medications are suitable while breastfeeding. If a mother is stable on a drug, vigilant monitoring of the infant can be done to evaluate safety, such as monitoring drug levels or signs of sedation. Table 2 provides the RIDs of most antiepileptic and mood stabilizer drugs.

Valproic acid levels in milk are low with an RID of approximately 1.4% to 1.7%. In a study of 16 patients receiving 300 to 2400 mg/d, valproic acid concentrations in milk ranged from 0.4 to 3.9 mg/L (mean = 1.9 mg/L). Although it is generally agreed that the amount of valproic acid transferring to the infant via milk is low, given the high risk of teratology of valproic acid, this drug should probably be avoided in women in the early postpartum period, and certainly in women at high risk of pregnancy.

Lamotrigine has been studied in at least 26 breastfeeding mothers. Levels in milk seem significant with RIDs ranging from 9.2% to 18.3%. The use of lamotrigine in breastfeeding mothers produces significant plasma levels in some breastfed infants, although they are apparently not high enough to produce side effects. Some investigators suggest it is advisable to monitor a symptomatic infant’s plasma levels to ensure safety. Premature infants should be closely monitored for apnea, sedation, and weakness. The maternal use of lamotrigine is probably compatible with breastfeeding of premature and full-term breastfeeding infants as long as the infant is closely observed for untoward symptoms.

The transfer of topiramate into human milk is significant. In a study of 3 women who received topiramate (150–200 mg/d) at 3 weeks postpartum, the RID ranged from 3% to 23% of the maternal daily dose. Plasma levels were detectable in 2 of the 3 infants and were low, 10% to 20% of the maternal plasma level. At 4 weeks, the milk/plasma ratio dropped to 0.69 and infant plasma levels were less than 0.9 μM and 2.1 μM, respectively. The breast milk and plasma levels were drawn 10 to 15 hours after the last dose of topiramate, which gives an underestimate of the amount found in breast milk. Because the plasma levels found in breastfeeding infants were significantly less than in maternal plasma, the risk of topiramate in breastfeeding mothers is probably acceptable. Close observation including monitoring plasma levels in symptomatic infants is advised.

Recreational drugs

Alcohol readily transfers into human milk, with an average milk/plasma ratio of about 1.0. Yet the clinical dose of alcohol in human milk is not necessarily high. In a well-controlled study of 12 women who ingested 0.3 g/kg of ethanol, the mean maximum concentration of ethanol in milk was only 320 mg/L. In an interesting study of the effect of alcohol on milk ingestion by infants, the rate of milk consumption by infants during the 4 hours immediately after exposure to alcohol (0.3 g/kg) in 12 mothers was significantly less. Reduction of letdown is apparently dose dependent and requires alcohol consumption of 1.5 to 1.9 g/kg. Other studies have suggested psychomotor delay in infants of moderate drinkers (2+ drinks daily). These reports also suggest that alcohol suppresses milk production significantly, which is secondary to alcohol suppression of oxytocin release. Metabolism of alcohol in adults is approximately 28 g in 3 hours, so mothers who ingest alcohol in moderate amounts may return to breastfeeding after waiting for approximately 2 hours for each drink consumed. Thus, mothers should avoid consuming alcohol or avoid breastfeeding during and for at least 2 hours per drink after consuming alcohol. Chronic or heavy consumers of alcohol should not breastfeed.

Studies on the use of cannabis in pregnant women seem to be somewhat inconsistent. Commonly called marijuana, the active component tetrahydrocannabinol (delta-9-THC) is rapidly distributed to the brain and adipose tissue. It is stored in fat tissues for long periods (weeks to months). Small to moderate secretion into breast milk has been documented. In 1 mother who consumed marijuana 7 to 8 times daily, milk levels of THC were reported to be 340 μg/L.

In 1 mother who consumed marijuana once daily, milk levels were reported to be 105 μg/L. Analysis of breast milk in a chronic heavy user revealed an 8-fold accumulation in breast milk compared with plasma, although the dose received is apparently insufficient to produce significant side effects in the infant. Studies have shown significant absorption and metabolism in infants although long-term sequelae are conflicting. In 1 study of 27 women who smoked marijuana routinely during breastfeeding, no differences were noted in outcomes on growth, mental, and motor development. In another study, maternal use of marijuana was shown to be associated with a slight decrease in the motor development in infants at 1 year of age, especially when used during the first month of lactation. The data from this study were conflicted by the use of marijuana during the first trimester of pregnancy. Maternal use of marijuana during pregnancy and lactation had no detectable effect on the mental development of the infant at 1 year of age.

There are few documented hazards reported following the limited use of marijuana while breastfeeding. Recent data suggest significant changes in the endocannabinoid system after fetal exposure to marijuana. This system has a major role in the development of the central nervous system (CNS) and is involved in mood, cognition, and reward and goal-directed behavior. Both animal and human data strongly suggest that marijuana exposure in pregnancy, and potentially lactation, may lead to neurobehavioral complications. Until further data can confirm these studies, use of this drug should be strongly discouraged.

Cigarette smoking not only exposes breastfed infants to nicotine and its metabolite cotinine but it also exposes the infant to toxic xenobiotics in the cigarette and environmental cigarette smoke. A study by Ilett and colleagues assessed the difference in nicotine and cotinine exposure from smoking cigarettes or using nicotine patches in breastfeeding mothers. This study enrolled 15 women who smoked an average of 17 cigarettes a day into an 11-week smoking cessation program using nicotine patches (21 mg/d in weeks 1 to 6, 14 mg/d in weeks 7 and 8, 7 mg/d patch in weeks 9 and 10, weaning around week 11). This study found that the absolute infant dose (in nicotine equivalents) decreased by about 70% by the time the mother was using the 7-mg patch compared with the dose generated by smoking. In addition, the breast milk concentrations of nicotine and cotinine also decreased by 50% and 66%. The average nicotine equivalents for infants exposed via breast milk was 25.2 μg/kg/d for smoking, 23 μg/kg/d for the 21-mg patch, 15.8 μg/kg/d for the 14-mg patch, and 7.5 μg/kg/d for the 7-mg patch. Therefore, as the mother progresses through the patch strengths, the transfer of nicotine equivalents to the infant via breast milk is significantly decreased and the exposure to other toxins from cigarettes is eliminated.

It is not recommended that women smoke near their infants, in the home or before breastfeeding; therefore, should a mother who smokes wish to breastfeed, it would be suitable to recommend nicotine replacement therapy to help her quit while continuing to breastfeed.

Caffeine is a naturally occurring CNS stimulant present in many foods and drinks. The half-life in adults is 4.9 hours, but the half-life in neonates is as high as 97.5 hours. The half-life decreases with age to 14 hours at 3 to 5 months and 2.6 hours at 6 months and older. The average cup of coffee contains 100 to 150 mg of caffeine depending on preparation and country of origin.

Peak levels of caffeine are found in breast milk 60 to 120 minutes after ingestion. In a study of 5 patients after ingestion of 150 mg of caffeine, peak concentrations of caffeine in serum ranged from 2.39 to 4.05 μg/mL and peak concentrations in milk ranged from 1.4 to 2.41 mg/L, with a milk/serum ratio of 0.52. The average concentration of caffeine in milk at 30, 60, and 120 minutes after ingestion was 1.58, 1.49, and 0.926 mg/L, respectively. Another study included 7 breastfeeding mothers who consumed 750 mg of caffeine per day for 5 days, and were 11 to 22 days postpartum. The average concentration of caffeine in the milk was 4.3 mg/L, and the mean concentration of caffeine in the serum of the infants on day 5 was 1.4 μg/mL.

The occasional use of coffee or tea is not contraindicated, but persistent chronic use of caffeine may lead to high plasma levels of caffeine in the infant, particularly during the neonatal period.

Pain/analgesia

Analgesics are one of the most commonly used medications while breastfeeding. Options for pain control include acetaminophen, nonsteroidal antiinflammatory drugs (NSAIDs), and opioids. Most NSAIDs are used to reduce pain and inflammation and are generally a suitable choice in breastfeeding women. Ibuprofen, acetaminophen, and naproxen are probably the most commonly used analgesics in North America. Their RID in milk ranges from 0.65% for ibuprofen, 8.81% for acetaminophen, to 3.3% for naproxen.

Opioids are often used for acute pain after cesarean delivery or for other procedures in breastfeeding mothers. Morphine is generally the preferred opioid used in breastfeeding mothers because of its poor oral bioavailability (26%) in the infant and low RID of 9.1%. Hydrocodone is a suitable alternative, its active metabolite is hydromorphone, and its RID ranges from 0.2% to 9% (average 2.4%). There have been 2 reports of adverse events with infants exposed to hydrocodone via breast milk. In the first case, both mother and infant were sedated after the mother took 2 hydrocodone 10 mg/acetaminophen 650 mg tablets every 4 hours for mastitis. Once the dose was reduced to 1 tablet every 3 hours, the sedation resolved. The second infant required intubation after exposure to a combination of opioids his mother had taken for a migraine (hydrocodone and methadone).

The use of codeine has started to decline since the death of an infant whose mother was taking codeine while breastfeeding in 2005. Both codeine and oxycodone are less favorable opioids because their metabolism is unpredictable (CYP 2D6 enzyme) producing active metabolites and data showing CNS depression in infants. In a cohort of mothers using oxycodone, codeine, and acetaminophen for pain during lactation, infant sedation was reported in 20.1%, 16.7%, and 0.5% for each drug, respectively. All opioids should be used with caution in breastfeeding mothers, using low doses and short courses, avoiding combinations with other opioids, monitoring the mother and child continuously for sedation/side effects, and constantly reevaluating the need for the opioid. Although methadone is not given as a pain medication and a full discussion of methadone is beyond the scope of this review, maternal use of methadone is not a contraindication to breastfeeding.

When possible, treating the cause of the underlying pain and using acetaminophen/NSAIDs are recommended.

Hypertension

Several medications are used to treat hypertension, including diuretics, β-adrenergic blockers, calcium channel blockers, angiotensin-converting enzyme inhibitors (ACEIs), and angiotensin receptor blockers (ARBs). Many of these medications are suitable during breastfeeding; however, some in the β-blocker family are known to cause problems for breastfed infants. The β-blockers of choice are metoprolol (RID 1.4%) and propranolol (RID 0.3%); neither medication has been associated with any adverse events in infants. In addition, labetalol has not been associated with any adverse effects in infants and has a low RID of 0.6%. Although rare, atenolol and acebutolol have both been associated with adverse effects in infants, such as cyanosis, tachypnea, bradycardia, hypotension, and low body temperature, and are not preferred agents. There is presently no information about the transfer of carvedilol or bisoprolol into breast milk. In summary, monitoring the infant for hypotension, bradycardia, and lethargy is suggested when using β-blockers during lactation.

The most common calcium channel blockers are amlodipine, felodipine, nifedipine, verapamil, and diltiazem. Studies on nifedipine suggest a low RID of 2.3%, 1 hour after a 30-mg dose. In a patient who took verapamil 80 mg 3 times a day, the average steady-state milk concentration of verapamil was 25.8 μg/L; no drug was found in the infant’s plasma and the RID was estimated to be 0.15%. There is 1 report of a patient who received diltiazem 60 mg 4 times a day; in this case, the RID was low (0.9%). There have been no reports of adverse events in breastfed infants exposed to nifedipine, verapamil, or diltiazem.

ACEIs are not only used for hypertension; they have numerous other indications such as heart failure, myocardial infarction, diabetes, kidney disease, and so forth. The 2 ACEIs with the most breastfeeding data are captopril and enalapril. In a study of 12 women who took captopril 100 mg 3 times a day, breast milk levels were about 4.7 μg/L 4 hours after the dose, the estimated RID was 0.002%, and no adverse effects were found. In a study in which 5 mothers were given a single 20-mg dose of enalapril, the average maximum milk concentrations of enalapril and its active metabolite enalaprilat were 1.74 μg/L and 1.72 μg/L, respectively; the RID was estimated to be about 0.175%. There are many ACEIs on the market, but captopril and enalapril are preferred until there are sufficient data available to confirm their safety in breastfed infants.

There are no data available on the use of ARBs in breastfeeding mothers. Until this information becomes available, ACEIs should be used instead of ARBs (candesartan, irbesartan, losartan, and so forth).

Diuretics are often used to help lower blood pressure and decrease edema. There are no published data on the amount of furosemide that enters breast milk. There is 1 case report of a woman who received hydrochlorothiazide 50 mg daily. On day 28, the mean milk concentration of hydrochlorothiazide was 80 ng/mL, resulting in a total infant daily dose of 0.05 mg of hydrochlorothiazide. Plasma levels of hydrochlorothiazide in this infant were undetectable and no adverse events were reported.

Despite suggestions in the past that diuretics may suppress milk production, no further details of this study or other studies have been published that confirm this controversy; at this time, there is no substantial evidence to suggest that diuretics reduce milk volume and or that diuretics are contraindicated in breastfeeding.