Abstract
Background:
Breastfeeding-related adverse drug reactions (ADRs) are thought to be uncommon as reported cases are globally low. The nonspecific nature of these reactions and a lack of awareness and difficulty in identification of ADRs by mothers and clinicians may result in these ADRs being underreported.
Aims:
This study hypothesized that breastfeeding-related infant ADRs are more frequent than reported. As a first-hand account of breastfeeding mothers, this study aimed to evaluate the impact of the perceived ADRs on the continuation of breastfeeding and maternal treatment.
Methods:
Women currently breastfeeding or having breastfed in the last 12 months were invited to complete an online survey. The survey comprised 42 questions in 5 sections to obtain data from breastfeeding mothers, including their use of medicines during lactation, perceptions of infant adverse reactions attributable to maternal medication use and its potential impact on breastfeeding.
Results:
This online survey was completed by 339 women, 42% of whom reported taking at least one medication during breastfeeding. ADRs were reported in 23 infants where a possible or probable causal relationship indicated by a Naranjo score of 1–8 was established in 16 (11.3%). Antibiotics (n = 12) and opioids (n = 2), including tramadol and oxycodone were identified as the most common adverse reaction-causing drugs. The average age of infants at the time of the perceived ADR was 25.6 days (95% confidence interval; 4–85 days; median age 17.5 days).
Conclusion:
Suspected ADR reporting in this study was significantly greater than those reported to the regulatory body, the Australian Therapeutics Goods Administration, which shows that common breastfeeding-related infant ADRs are underreported.
Introduction
Breastfeeding women frequently take medicines for common postpartum conditions, such as cesarean wound pain, urinary incontinence, constipation, hemorrhoids, minor illnesses, backache, infections, mastitis, stimulation of lactation, depression, and other neuropsychiatric illnesses.1,2 The reported extent of medicine use in postpartum women varies significantly ranging from 34% to 100%, which is largely attributed to a lack of uniformity in medicine use reporting systems, and differences in study designs.2–4 Some studies suggest that women are more likely to take medicines when breastfeeding compared with when pregnant. 4 While most medicines are considered safe and compatible with breastfeeding, adverse drug reactions (ADRs) do occur in breastfed babies that are thought to result from exposure through breast milk. Globally, breastfeeding-related ADRs are infrequently reported and are thought to be uncommon.5–7
This underreporting could be attributed to the nonspecific nature of infant ADRs (e.g., diarrhea or unsettled behavior) that could be easily mistaken for other illnesses. Furthermore, inadequate training of health care professionals to even consider establishing a causal relationship between a maternal medication use and an ADR in a breastfed infant may be additional contributing factors.8–14
While it is possible that infant ADRs are infrequent, it is more likely that this method of reporting may affect apparent frequency of infant ADRs resulting in global underreporting. Due to a lack of baseline data, it is difficult to quantify the extent of the problem and its economic impacts. Investigation of infant ADRs can place an unnecessary burden on the health system when they are carried out to rule out more serious illnesses. Mothers choosing to forgo treatment for their medical condition for the fear of inadvertently exposing their breastfed infant to a medication can have a detrimental impact on the health and wellbeing of both mother and infant. Undetermined long-term effects on neurobehavioral development remain the biggest risks of inadvertent infant exposure through breast milk, which can potentially have great economic consequences.
The perceived ethical and financial risks associated with testing of drugs in pregnant and breastfeeding mothers is a disincentive for pharmaceutical companies to conduct drug safety tests in this population. Consequently, these data are unavailable for most drugs, and pharmaceutical companies often recommend their products not be used in this population by default.15,16 For instance, a scan of the electronic Therapeutic Guidelines (eTG), which is considered a leading source of independent, evidence-based, practical treatment advice to assist practitioners with decision making at the point of care, shows that for more than half (54%) of the drugs listed in the eTG, the recommendation is to exercise caution or avoid use during lactation due to lack of data. 17 Therefore, clinicians must prescribe medicines “off-label” for lactating women without the usual guidance or evidence base used in prescribing for other populations. 18
Often clinicians are forced to resort to small observational studies or case studies, to obtain their lactation safety data. While case reports are valuable, they often lack robust methodology and study design, which when added to interindividual differences makes it even more difficult to provide high-quality evidence-based information to women. For some drugs, safety information is limited to data from animal studies, which may not correlate well with human physiology. 19
Several studies showed that breast milk-related ADRs are more frequent in the first 2 months of life when an infant's metabolic and excretory organs are underdeveloped but can occur at any time during breastfeeding.5,20 In Australia, the reporting of ADRs, including those that may result from passage of the drug through breast milk is a voluntary process, which relies on health care professionals and members of the public to report to the Therapeutics Goods Administration (TGA), the authority responsible for regulating medicines. To gain a better understanding of the Australian data on breastfeeding-related ADR reporting, a report of ADRs in breastfeeding was obtained from the TGA under the Freedom of Information Act using MedDRA terminology “exposure during breastfeeding” as the key search term for the requested reports. During 2003–2016, there were a total of 65 cases that met the requested search criteria that is, “exposure to breastfeeding.” Of these 65 reports, it was found that 5 reports were duplicated (Refer to Supplementary Appendix SA1 for details).
Analysis of case reports provided by the TGA found that the data recorded by the TGA lacked consistency, uniformity, and vital information. The infant age at the time of reporting, a crucial factor, was not stated in most of the reports and it was difficult to confidently know the reporter/source in most cases. According to the studied reports, medicine use by mother was discontinued or breastfeeding was ceased in many cases. Currently, no long-term follow-up of these cases is conducted (as confirmed with TGA) to enable assessment of long-term effects of the ADRs on the infant. Drugs acting on the central nervous system, such as antidepressants and antipsychotics, and antimicrobial agents were implicated in most of the cases. The incomplete nature of the TGA data reduces the usefulness of this report to ascertain the impact of the ADRs on the breastfed infant. What is clear is that regardless of whether a country has mandatory ADR reporting like France or voluntary reporting like Australia, it appears that overall, the numbers of reported breastfeeding-related ADRs are low.
To gain a better understanding of the occurrence (proven or perceived) of infant ADRs, this study was undertaken seeking a first-hand account of women who were breastfeeding. It sought to investigate the reporting rate and nature of these infant ADRs and document an account of the impact of such ADRs on their breastfeeding experience.
Survey Design and Methodology
Participants were recruited by advertising on the Australian Breastfeeding Association (ABA) website, through word of mouth and through poster displays at the Ngala Association office in Perth and child health centers in Western Australia seeking participation by women currently breastfeeding or having breastfed in the last 12 months. The ABA and Ngala were selected as these organizations are frequented by breastfeeding women. The ABA is a national organization that promotes breastfeeding and provides many useful breastfeeding-related resources through their website. Ngala is a West Australian organization that delivers services such as parenting support, breastfeeding support, and childcare centers for Western Australian families. Therefore, the ABA and Ngala were deemed as suitable targeted advertising platforms for this survey.
West Australian metropolitan and regional Child Health centers were also utilized for this purpose, as these centers provide care and support to infants in early life, including postbirth appointments, baby weighing facilities, breastfeeding and general infant advice, and childhood immunizations starting at 2 months of age. The use of maternity hospital as a recruitment platform was considered. However, it was not pursued due to ethical challenges.
A priori sample size estimation was based on an average of 300,000 births per year in Australia. 21 With a margin of error of ±5% (confidence interval [95% CI]), the sample size needed for an adequately powered study was determined to be 385 participants. This study was conducted using a self-administered online questionnaire validated through several pilot studies. The anonymous survey was presented on a Microsoft Azure supported website, fully secured with a username and password. Participants were required to be 18 years or older, breastfeeding, or had breastfed in the past 12 months. Participants were asked to provide an email address to ensure the integrity of data by preventing duplications.
The online survey comprised 42 questions in six sections. Each section was designed to obtain an in-depth overview of the mother's health and perceptions regarding breastfeeding.
Section 1: General details (6 questions)
Section 2: Mother's general health and wellbeing (4 questions)
Section 3: Medicine use during breastfeeding (10 questions)
Section 4: Perceived reactions to food products (4 questions)
Section 5: ADR probability (14 questions)
Section 6: Impact on breastfeeding (4 questions)
Section 5 of the survey based on the Naranjo Algorithm aimed to establish a causal relationship between an identified untoward clinical event and a suspected medicine (drug) by employing critical causation variables.22,23 A Naranjo total score of 1–4 indicates a possible causal relationship between the suspected drug and the ADR, a total score of 5–8 indicates a probable causal relationship and a total score greater than 9 indicates a definite causal relationship between the drug and the ADR. While the commonly used Naranjo algorithm is a reliable way of establishing causal relationships between an ADR and a suspected drug, it is aimed for use by health care professionals.23,24 To make the questions more understandable and applicable to the survey participants, questions were rephrased, simplified and the order rearranged. Refer to Supplementary Appendix SA2 for the modified Naranjo questions.
Results
Demographic and general details of participants
This online survey was completed by a total of 376 participants over a 7-month period. Of these, 37 surveys were deemed ineligible for inclusion due to duplication and/or incompleteness. A total of 339 responses (95% CI; ±5.3% margin of error) were included in the analysis of this study. The majority (83.8%) of the 339 participants (95% CI) were from the state of Western Australia. Participants who were breastfeeding at the time of the survey formed 78.2% of the respondents (Table 1). The average breastfeeding duration was 45 weeks (range 0.86–312 weeks; standard deviation ±42 weeks). The education level of the respondents was generally high with 62% having at least an undergraduate degree and over a quarter of total respondents having a postgraduate degree.
Demographic Data of Survey Participants
TAFE, Technical and Further Education Colleges.
Mother's general health and medicine use during pregnancy and lactation
Approximately one-third of the survey respondents (29.2%; n = 99) indicated that they had a medical condition but only a small proportion of these participants actually disclosed the nature of their condition as shown in Table 1. The use of prescribed and over-the-counter medicines (including complementary medicines) was high among breastfeeding mothers with a total of 142 (42%) women taking one or more medicines at some stage while breastfeeding with an average of 1.42 medicines/supplements per person as shown in Table 2. Approximately, one third of the participants (n = 108) stated having started a new medicine after the birth of their baby, which they were not exposed to or had taken during pregnancy. A small proportion of respondents (16%) reported taking a new short-term medicine after the birth of their baby (Table 3). These medicines included antibiotics and analgesics, including nonsteroidal anti-inflammatory drugs and opioids.
Long-Term Medications (Prescribed and Over-the-Counter) Used by Breastfeeding Mothers
The drug classes are bold and the drugs under those classes are presented in regular font. In literature, breastfeeding ADRs are often classified in drug classes. In order to not deviate from this, we chose to include drug classes as well as the actual drugs taken by breastfeeding women.
NSAIDs, nonsteroidal anti-inflammatory drugs.
Short-Term Medications (Prescribed and Over-the-Counter) Used by Breastfeeding Mothers
The drug classes are bold and the drugs under those classes are presented in regular font. In literature, breastfeeding ADRs are often classified in drug classes. In order to not deviate from this, we chose to include drug classes as well as the actual drugs taken by breastfeeding women.
NSAIDs, nonsteroidal anti-inflammatory drugs.
Of the 339 participants, 134 (39.5%) were concerned about the transfer into their breast milk of both long and short-term medicines. A great majority of these women (n = 130) sought professional advice regarding their concerns. It is not known whether the professional advice sought, or the information obtained by these mothers affected their choice of treatment. Forty-eight participants (37%) of those who sought professional advice stated the source of their advice. The internet (n = 10) was reported to be the most used resource to obtain this information, followed by hospital (n = 7), pharmacist (n = 7), and a specialist doctor (n = 7)
Establishing causal relationship between the offending drug and the appearance of the perceived ADR using Naranjo algorithm
Twenty-three participants (16.2%) indicated that they noticed or suspected an ADR in their breastfed infant after they took a medicine. A possible or probable causal relationship with maternal medicine use could be established in only 16 (11.3%) cases, 13 had a Naranjo total score of between 1 and 4 (possible causal relationship), while the remaining 3 had a total Naranjo score of between 5 and 8 (probable causal relationship). No ADRs were identified to have been definitely caused by maternal medicines (Naranjo score >9). Antibiotics (n = 12) and opioids, including tramadol and oxycodone (n = 2) were identified as the most common causing drugs. The average age of infants at the time of the perceived ADR was 25.6 days (95% CI; 4–85 days; median 17.5 days), which aligns with evidence that infant ADR mostly occur under 2 months of age. 5 Ten participants reported the ADR to their clinician.
Impact of ADR on continuation of breastfeeding and maternal treatment
Only one participant reported cessation of breastfeeding due to the perceived ADR (Table 4). Five participants taking antibiotics stopped taking the suspected offending drugs. It is unclear as to whether they replaced it with a different medicine or stopped after the course was completed. It is unusual for patients to discontinue antibiotics when there is unresolved infection. Ten participants (62.5%) reported the perceived ADR to a health care professional. However, the mothers did not know if the clinician then reported this ADR to the TGA.
Medicines Suspected to Have Possibly or Probably Caused Infant Adverse Drug Reaction in the Study Population (Naranjo Score > 0)
Juice Plus® contains a blend of 20 fruits and vegetables in the form of juice powder, available in both capsule and chewable forms.
ADR, adverse drug reactions.
Study limitations
While this study was intended as national survey, most of the responses were from the state of Western Australia (WA). Participant recruitment in WA was multichannel through child health clinics, Ngala parenting centers, and the ABA website, whereas participant recruitment for the rest of the country was through the ABA website only. The account of reported adverse reactions were based on mothers' perception and not a medical assessment by a health care professional. Therefore, it is possible that mothers who thought their infant had an ADR were more likely to take the survey. To mitigate the risk of recall bias, participation was limited to women who had ceased breastfeeding no >12 months before the survey. Despite this, almost one fifth of the survey participants had stopped breastfeeding at the time of the survey possibly resulting in some recall bias.
In this survey, the participants were allowed to provide the names of their medications in free text. This was done to ensure information provided by the participants was accurate acknowledging that not all participants were expected to be familiar with both the generic and brand names of their medications. This resulted in a significant number of responses where the participants had stated the class of their medication for example, antibiotic but not the name of medication, limiting the depth of this information.
Another limitation of this study was the reported education level of the respondents, which was generally high with 62% having at least an undergraduate degree and over a quarter of total respondents having a postgraduate degree. This is indicative of a high level of health literacy suggesting that the participants may have been more likely to report infant ADR due to greater awareness and higher literacy.
Discussion
Through a first-hand account of breastfeeding mother in evaluating occurrence of infant ADRs, this study showed that medication use is widespread in breastfeeding women with 42% using one or more medicines while breastfeeding. Concern of inadvertently exposing their infant to medications was prevalent with greater than one-third of the survey respondents indicating that they sought professional advice regarding the lactation safety of their medication. While most drugs used during lactation are considered safe, some women reported using medicines such as antipsychotics (olanzapine and quetiapine), benzodiazepines (temazepam), opioid analgesics (codeine and tramadol), anti-inflammatories (mesalamine), oral corticosteroids, propylthiouracil, pregabalin, and Garcinia cambogia, where lactation safety data and studies evaluating long-term effects on breastfed infants is limited.
Overall, only a small number of women reported noticing an ADR in their breastfed infant. Out of these, a possible or probable causal relationship with maternal medicine use could be established in only 11.3% cases, which is similar to reports by Ito et al who found minor, nonserious ADRs in breastfed infants to be around 11.2%. 25 Antibiotics and opioids, including tramadol and oxycodone were identified as the most common adverse reaction-causing drugs.
The average age of infants at the time of the perceived ADR was 25.6 days (range: 4–85 days; median age: 17.5 days), which confirmed previous reports that breast milk-related ADRs are most common in the first 2 months of life. 5 Ten participants reported the ADR to their clinicians, but it is not known whether these ADRs were then relayed to the TGA. In this study population, breastfeeding-related ADRs appear to be low with only 23 ADR reports from 339 respondents over a 7-month period. However, when compared with the national reporting of breastfeeding-related ADR to the TGA (60 ADRs over a 13-year period from 2003 to 2016), it is significantly higher. However, the ADRs reported in this study were mostly nonserious reactions.
With the average new births of 35,000 per year in Western Australia, this study was completed by an estimated 1% of women who gave birth in that year. 26 Furthermore, due to the known decline in the number of women who continue breastfeeding over time and this study, including participants at various stages of breastfeeding (including those who ceased breastfeeding in the preceding 12 months), these data lack a true denominator. Hence, the prevalence of breastfeeding-related ADRs is likely an underestimation. Despite the limitations, this study shows that common and nonserious breastfeeding-related ADRs are not reported to the TGA. While it can be argued that some of the ADRs reported in this study are common side effects of these medications and do not warrant reporting to the TGA, in the absence of large clinical studies, and prevalent off-label prescribing of medicines to lactating women, increased pharmacovigilance through reporting is the best way of identifying safety concerns in breastfed children who are exposed to medicines.
Despite antibiotics being considered largely compatible with breastfeeding, the findings of this study identified antibiotics to be implicated in majority of perceived infant ADRs. These findings are supported by studies published by Soussan et al, and Ito et al, which also show anti-infectives to be causing breastfed infant ADRs.5,25 Antibiotics are largely considered “compatible” with breastfeeding but transient side-effects such as diarrhea, rashes, and vomiting are commonly seen in breastfed infants whose mothers use these drugs. 27
Recent studies have shown that antibiotic exposure in the early years of life has been associated with adverse neurodevelopmental and neurocognitive outcomes later in life.28–30 These associations were only made possible with long-term follow-up of those children. Therefore, a drug being compatible with breastfeeding does not equate to infant safety in the long term and is further reason for implementation of more thorough pharmacovigilance systems. This also calls into question the “blanket rule” approach taken by current references and guidelines where medicine safety in breastfeeding is classified into broad categories of compatibility namely “compatible and noncompatible” or “safe, caution, and avoid” as per the eTG.
This survey highlights that “breastfeeding-compatible” drugs, such as antibiotics and opioids are commonly used by breastfeeding women and they account for a significant proportion of reported ADRs, which seem to be mild in nature. However, due to a lack of long-term follow-up and appropriate pharmacovigilance, their long-term effects will not be fully known. Studies have shown that breastfeeding women are likely to be noncompliant with anti-infective treatments due to safety concerns and have been shown to discontinue breastfeeding or not take their antibiotic as prescribed. 31 Noncompliance to prescribed medicines in pregnancy and breastfeeding due to safety concerns for the baby are commonly reported.32–34 In this study, one participant reported taking her prescribed antibiotic in four divided doses instead of two to minimize its transfer in her breast milk. Making changes such as this to prevent infant ADR can have unfavorable effects in the mother, including potential treatment failure and antibiotic resistance.
Conclusion
This study demonstrates that nonserious ADRs in breastfed infants is more common than reported. These findings are supported by previous such studies putting infant ADRs at just over 11%. ADR reporting to the TGA was found to be generally very low with an average of 13 reports annually from 2003 to 2016. Study participants who reported the perceived ADR to their health care professional did not know if the clinician had then informed the TGA. It is likely that these ADRs were potentially not reported as most were expected side-effects of antibiotics for example, gastrointestinal upset. Considering the emerging evidence linking antibiotics (a drug class deemed largely compatible with breastfeeding) exposure in early life to long-term negative neurobehavioral outcomes, it is imperative that any drugs used during lactation be subject to stricter pharmacovigilance and any ADR, perceived or real should be reported.
Due to the current ethical challenges surrounding controlled trials, enhanced pharmacovigilance through better reporting is the second-best way of ensuring lactation medication safety data are improved. Additionally, it is important to acknowledge that labeling of drugs as “compatible” or otherwise with breastfeeding does not assure absolute safety in breastfed infants as it does not evaluate long-term safety.
Footnotes
Acknowledgments
The authors wish to acknowledge the Australian Breastfeeding Association (ABA) for allowing the survey to be available in their website and the Ngala Association office in Perth, child health centers in Western Australia for allowing poster to be displayed to recruit participants.
Authors' Contributions
H.A.: Conceptualization (equal); methodology (equal); survey design (lead); formal analysis (lead); writing—original draft (lead); and writing—review and editing (equal). L.B.G.T.: Conceptualization (equal); methodology (supporting); survey design (co-lead); formal analysis (supporting); writing—original draft (supporting); writing—review and editing (equal); and supervision of project (cosupervisor). A.C.: Conceptualization (equal); methodology (supporting); formal analysis (supporting); writing—original draft (supporting); writing—review and editing (equal); and supervision of project (primary supervisor).
Human Ethics
This project had ethics approval from Curtin University Human Research Ethics Committee (HR2012/110). All participants' written consent have been archived in secured research drive within Curtin University.
Disclosure Statement
No competing financial interests exist.
Funding Information
There is no external funding for the project. This project is part of a PhD study conducted in Curtin University.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
