The challenges of medically complex breastfed children and their families: A systematic review

Abstract Exclusive breastfeeding for the first 6 months and then alongside solid food for the first 2 years and beyond is the gold standard in young child nutrition. There is an abundance of literature relating to the preventative nature of breastmilk and breastfeeding against many infectious diseases and chronic conditions. However, despite medically complex infants and children being a group that could benefit most from continued breastfeeding, breastfeeding duration and exclusivity are lower among more complex paediatric populations. The reasons for this are not well known, and there is a paucity of data relating to supporting infants who have acute or chronic illness, disability or congenital anomaly to breastfeed. This systematic review aimed to understand the challenges of breast/chestfeeding the medically complex child and to establish the gaps in healthcare provision that act as barriers to optimal infant and young child feeding. The search was limited to studies published in English, focused on breastfed sick infants in hospital, with no date limits as there is no previous systematic review. Of 786 papers retrieved, 11 studies were included for review, and seven themes identified. Themes included practical and psychological challenges of continuing to breastfeed in a hospital setting, complications of the condition making breastfeeding difficult, lack of specialist breastfeeding support from hospital staff and a lack of availability of specialist equipment to support complex breastfeeding. The findings affirm the lack of consistent high‐quality care for lactation support in paediatric settings and reinforce the need for further focused research in this area.


| INTRODUCTION
Breastfeeding is the biologically normal way to feed infants and children up to the age of 2 years and beyond (World Health Organization (WHO), 2018). Not breastfeeding is associated with a greater risk of adverse health issues for infants including infectious disease, obesity and diabetes and a rise in reproductive cancers, heart disease and diabetes for women (Victora et al., 2016). Not meeting breastfeeding goals is also associated with higher rates of maternal depression and grief (Borra et al., 2015;Brown, 2016;Fahlquist, 2016). Breastfeeding is perhaps even more important for medically complex children, due to the immunologic protection it confers; for energy consumption when solid foods may be difficult to digest; and for comfort through illnesses and procedures (Edwards & Spatz, 2010;Spatz, 2012;Thomas, 2020). Breastfeeding is also known to be effective in managing procedural pain, though it has so far mainly been studied in the context of vaccination (Harrison et al., 2016). Breastfeeding is a way of parenting and is often used as a tool to calm and soothe a young child; it cannot be distilled down to caloric intake alone (Brown, 2018;Brown et al., 2018).
However, evidence from studies around the world suggests that breastfeeding duration and exclusivity is lower and more difficult to achieve among medically complex children for a variety of reasons.
For example, one UK literature review found that children with Down syndrome have lower rates of breastfeeding and those who managed to breastfeed only could because they had more support (Sooben, 2012). This was echoed in a Puerto Rican study (Colón et al., 2009). A recent case report compared volumes ingested by an infant with Down syndrome with an infant without Down syndrome and found that low intra-oral pressure, large tongue and less effective suckling were clinically significant (Coentro et al., 2020). Meanwhile, Torowicz et al. (2015) studied infants with a congenital heart defect, noting that the high-stress environment makes establishing a milk supply more challenging. Finally, Rivera et al. (2008) explored the complexities of breastfeeding infants with spina bifida and concluded infant instability after surgery was not the biggest barrier-rather, it was the clinical environment, lack of medical staff knowledge and hospital routines.
It is important to understand how and why breastfeeding is more challenging for parents of medically complex children in order to target services and support to enable them to meet their feeding goals.
Understanding the challenges could also lead to more specific training for healthcare staff so that they are able to support families more effectively and skilfully. Currently, there is little formal guidance for breastfeeding medically complex children. Although the policies of the Baby Friendly Hospital Initiative (BFHI) and the Baby Friendly Initiative (BFI) UK promote, protect and support breastfeeding, their policies do not cover paediatrics where medically complex children will be cared for. Whereas children who are diagnosed antenatally with a congenital anomaly will be cared for in the neonatal unit, where staff are likely to have been trained in how to support breastfeeding and maintain milk supply, children cared for in the cardiac intensive care unit (CICU), paediatric intensive care unit (PICU), emergency department or general medical or surgical paediatric ward may be cared for by staff with very little breastfeeding training. Furthermore, infants who start off in the neonatal unit may be transferred to one of these paediatric settings, meaning that their experience of breastfeeding support may change.
Despite the known difficulties of feeding medically complex children (Coates & Riordan, 1992), there is no systematic review of their breastfeeding experience within the paediatric setting. A synthesis of studies exploring the experiences of parents breastfeeding their medically complex children may illuminate areas for prioritization of training and support. The purpose of this systematic review is therefore: 1. To establish the existing body of knowledge around the challenges and needs of parents breastfeeding their sick infants or children in the paediatric setting.
2. To identify gaps in healthcare provision that act as barriers to maintaining breastmilk supply and facilitating breastfeeding in the medically complex paediatric population.

| METHODOLOGY
To address the research questions, the search strategy (Table 1) and eligibility criteria (Table 2) were designed in line with the PICOS criteria (population, intervention, comparator, outcomes, study design or setting). This is a modification of the PICO criteria, which omits

Key messages
• There are limited data relating to breastfeeding medically complex children. Existing literature focuses on specific infant conditions. Support to facilitate breastfeeding is generally considered suboptimal.
• Breastfeeding medically complex infants is more challenging. There are practical and psychological parent challenges, as well as infant-related difficulties. There are also shortcomings with lactation support within paediatrics, a lack of training for healthcare professionals and challenges with provision of and education in using specialized infant feeding equipment.
• Parents of medically complex breastfed children have unique needs, which need to be addressed with research, policy and training in order to optimize outcomes for this population.

| Eligibility criteria
Both published and unpublished studies using any methodology were eligible if they met the inclusion criteria (see Table 2). Literature was included from anywhere in the world, as there may be examples of good practice, as well as higher breastfeeding rates in resource-poor as well as resource-rich countries (Victora et al., 2016). No date limits were set as there is no previous systematic review, although where located studies were dated, their results were treated with caution.
All studies whose focused population was breastfed children with acute or chronic illness, disability or congenital anomaly were considered. An acute illness is experienced by a child who is usually healthy After the initial exclusion of articles that did not meet the inclusion criteria, there remained 127 article abstracts to read. Reasons for exclusion are noted in Figure 1. All articles that could not be conclusively accepted or rejected after reading the abstract were kept. After applying the inclusion criteria to the full texts of the remaining papers, there remained 11 articles for review (see Figure 1). A narrative synthesis and thematic analysis were then conducted with the eligible studies.

| RESULTS
A Overall, there were eight qualitative and three mixed-methods studies representing a total sample size of 599 (range: n = 5-194).
All the studies explored the impact on breastfeeding of illness, disability or congenital anomaly. There was a small clustering of studies in 1998, and of 11 included studies, six were conducted in the United States or Canada. There was just one very small 3.1 | Narrative synthesis

| Study quality
The studies all had clearly defined aims and recruitment strategies. All of them explored the impact of various medical conditions on breastfeeding outcomes. The studies addressed and commented on a range of potential confounding factors, including socio-economic status (SES), degree of infant illness or disability and infant age. Only four studies commented on prenatal intention to breastfeed (Lambert & Watters, 1998;Madhoun et al., 2019;Moe et al., 1998;Rendón-Macías et al., 2002), which could be significant, as parental motivation is known to be a factor in breastfeeding duration and exclusivity (Claesson et al., 2019).
Only two studies commented on whether the hospital facility was baby friendly accredited ( (see Table 5).

| Study themes
Despite the variable study quality, all included studies contributed to the development of themes because there is so little literature available and all the experiences of the study participants are arguably valid and meaningful. Themes were developed by reading and rereading the papers to become familiarized with their purpose, methods and results.
A theoretical thematic analysis was chosen, because the themes were not necessarily linked to the data being collected in the study in question, although they were explicit within the meaning of the papers. Rather, the data collected were analysed in relation to this research question (Braun & Clarke, 2006).
Seven themes emerged from the literature: parental factors, both practical and psychological; infant factors, relating to both infant illness and instability, as well as how a chronic condition affected their ability to effectively breastfeed; the availability of specialist lactation support; support and information from healthcare professionals; and specialized equipment and resources needed.
Eight studies mentioned at least six out of the seven themes and the most prevalent theme related to the inadequacy of healthcare professional support-this was identified as a barrier in every study (see Table 6).

| The practical impact of infant hospitalization on the parent
Infant hospitalization impacts parents in a very practical way. Eight of the 11 studies found various practical problems, ranging from issues of logistics to practical breastfeeding problems. For example, the way in which the infant was fed breastmilk often had to change due to separation or necessary adaptations due to the illness/disorder. Many mothers had to start expressing milk or work to maintain supply when under normal circumstances they may have been able to directly feed (Lambert & Watters, 1998). This can add in a layer of complication in finding the time to express, store and deliver the milk.
Expressing milk rather than directly breastfeeding is also not necessarily straightforward. It is associated with a higher risk of blocked ducts, mastitis, engorgement (Kvist, 2010)

| The impact of infant hospitalization on the parent: Psychological
Nine of the studies explored data relating to the parental psychological aspects of breastfeeding their hospitalized child. These could be both negative and positive. For example, from a negative perspective, Duhn and Burke (1998) and Lewis and Kritzinger (2004) found that exhaustion, overwhelm, disappointment, frustration, inadequacy, loneliness and sadness were common. Likewise, Madhoun et al. (2019) described high levels of anxiety and depression in their sample, with over half the mothers suffering with anxiety and one third struggling with postnatal depression.

T A B L E 5 Summary of strengths and limitations of studies
However, Lambert and Watters (1998) found some positive aspects relating to breastfeeding a child with a chronic condition.
Some of the perceived benefits of breastfeeding included a greater sense of calm, decreased stress, an opportunity for relaxation and an increased sense of self-efficacy.

| The impact of infant acute critical illness or instability affecting infant ability to breastfeed
Eight of the studies commented on infant instability or severity of illness as a factor affecting their ability to feed easily. Of these studies, one was a study of children with acute bronchiolitis in otherwise healthy children, and the others all described experiences of children with congenital conditions.
Several studies exploring congenital heart defect found infant critical illness or instability to be a barrier to breastfeeding. Duhn and Burke (1998) found that infants with heart defects were likely to struggle to feed and be too fatigued to feed, and surgery and ventilation were cited as factors affecting an infant's ability to breastfeed. Barbas and Kelleher (2004) found that infants were more likely to be breastfed postoperatively if they had attempted breastfeeding before their surgery. Meanwhile, Lambert and Watters (1998) and Rendón found that infants with Down syndrome were more likely to have problems with breastfeeding if they also had a cardiac defect, low tone and poor suck.
The only study to explore an acute illness was Heilbronner et al. (2017). It appeared that admission to hospital for acute bronchiolitis was associated with reduced exclusivity or cessation of breastfeeding, but the reasons for this were multifactorial and mostly unrelated to illness severity.
3.5 | The impact of infant chronic condition on their ability to effectively breastfeed Most of the studies exploring the challenges of breastfeeding infants with cardiac defects had many findings in common. Increasing and measuring milk consumption was a common theme. Barbas  Kelleher (2004) focused mainly on the need for more calories, whereas Lambert and Watters (1998) and Duhn and Burke (1998) found that many mothers needed to provide small frequent feeds to Some mothers turn to other sources for information and support instead. For example, Lewis and Kritzinger (2004) found peer support from another mother who had breastfed a child with Down syndrome particularly helpful, although this was not a standardized service.
Meanwhile, Madhoun et al. (2019) and Moe et al. (1998) found that mothers accessed online support groups and organized breastfeeding support groups were also identified.
When high-quality support was provided, it had a positive impact.
For example, Barbas and Kelleher (2004) studied 68 infants with congenital heart disease (CHD) in the United States, where 6 years previously they had established a designated lactation support programme led by a full-time IBCLC. The IBCLC also had extensive paediatric nursing experience and went on to establish a programme of education for all staff. This intervention led to an increase in breastfeeding rates from 14% to 47% in that time. In this study, mothers cited the IBCLCs as very supportive.

| The support, training and attitudes of healthcare professionals
In general, most of the studies highlighted inadequate support. Many of the healthcare staff were acknowledged as caring, but most parents did not get the breastfeeding support they needed. Some staff were perceived as ambivalent about the importance of breastmilk (Barbas & Kelleher, 2004) Lambert and Watters (1998) reported that women rated paediatric staff knowledge as poor.
A lack of training and skills was identified by a number of studies as central to the lack of support (Lambert & Watters, 1998;Madhoun et al., 2019). Moe et al. (1998) found that parents perceived physicians to be theoretically supportive of breastfeeding, but without adequate training to be able to provide support for breastfeeding challenges. Rendón-Macías et al. (2002) found that advice to supplement or stop breastfeeding by a medical professional was prevalent.
In the Ryan et al. (2013) study, mothers highlighted various gaps in knowledge. Banta-Wright et al. (2015) found that parents had to be creative, finding their own breastfeeding support or utilizing peer and family support.
3.8 | The necessity and availability of specialized equipment or resources bottles, nasogastric tube (NGT), gastrostomy tube (GT) and syringes but did not in general find adequate support to help them with their breastfeeding journeys. Lambert and Watters (1998) identified lack of privacy, lack of access to pumps and inconsistent advice as barriers. Heilbronner et al. (2017) noted that many parents complained of breast pump shortage and that pumping was difficult, creating a significant barrier to expressing.
In the Madhoun et al. (2019) study, there is specific mention of six different specialty feeding bottles, as well as NGT and GT. Some of the mothers were disappointed that the hospital staff did not know how to help the parents use the equipment. Finally, Ryan et al. (2013) found that mothers were reliant on practical aids such as a specialized bra that enabled hands-free pumping. Three mothers who used NGT felt that they were at times overused instead of attempting breastfeeding.  et al., 2018). The papers in this study had limited numbers of low-SES families, which may be significant as these problems will have the greatest negative impact on vulnerable and low-income families (Beck et al., 2017;Thomson et al., 2016).

| DISCUSSION
Most parents in this study cited some level of psychological distress surrounding their child's admission to hospital, although some specifically described the positive benefits that breastfeeding brought to the experience. The parents cited exhaustion, stress, anxiety and depression frequently. The psychological aspects were not all negative however, with many of the mothers describing breastfeeding as something that made them part of the solution, and one mother stated that she felt breastfeeding helped to re-establish trust with her toddler after their surgery. Essentially, breastfeeding was hard work, but the parents were motivated to continue despite the challenges.
The psychological challenges relating to breastfeeding may on the one hand negatively impact a parent's confidence and experience of feeding and caring for their child, but breastfeeding also provides an opportunity to empower parents to feel included in their child's care.
Supporting parents to be able to overcome a challenge rather than feel defeated by it may lead to a greater sense of self-efficacy.
Throughout many of the papers, there was a sense that the parents managed to persevere with breastfeeding in spite of their experience within the paediatric setting, rather than because of it.
There were also infant-related feeding challenges, distinctly different from the practical challenges of maintaining healthy lactation in the parent. Not all congenital conditions affect a child's immediate physiological stability, such as cleft palate or Down syndrome. However, even when a child is initially stable, their condition can change, or corrective surgery can make them more unstable.
Some conditions necessitate specific breastfeeding adaptations due to the infant's condition, and not their medical instability. These may be related to positioning for breastfeeding, fat or calorie content, specialized techniques or frequency of feeding. Effective breastfeeding involves both the infant and parent. The infant has to be able to use their tongue, lips, jaw and cheeks to stabilize the breast in their intra-oral palate, create negative pressure and be able to safely suckle and swallow while also coordinating breathing (Genna, 2013).
However, for ongoing successful lactation, milk must be removed from the breast/chest according to the infant's individual metabolic and caloric need. The infant will need to be positioned sustainably for breastfeeding in a way that supports a safe suck-swallow-breathe sequence.
Some children are born with conditions that require breastfeeding modification. For example, infants with chylothorax cannot receive breastmilk unless it has been separated in a centrifuge to remove the fat (Davis & Spatz, 2019), and infants with phenylketonuria (PKU) cannot breastfeed exclusively because although breastmilk contains less phenylalanine than formula, these infants usually need specialized Phe-free formula to a greater or lesser extent depending on their Phe levels-which must be monitored closely. Conversely, infants with hypotonia may not only tire easily but are also more difficult to hold and position, and they may not be able to effectively create a seal at the breast. Supporting a mother-baby dyad in these specialist cases is more difficult and requires specialist knowledge compared with supporting healthy breastfeeding infants.
In many clinical settings, such as maternity or neonatal units, specialist lactation support is a clearly defined sub-specialty. This type of support involves more than simple breastfeeding management in uncomplicated situations and requires the ability to be able to assess and treat complications, at a level far higher than standard breastfeeding training. Globally, the IBCLC credential is the recognized leading qualification in breastfeeding support, and IBCLCs have the most comprehensive and robust skill sets (Chetwynd et al., 2019).
However, the number of IBCLCs globally varies, as does the scope of practice. In the United States, IBCLCs are often part of the wider healthcare team, serving neonatal and obstetric departments (Haase et al., 2019). Conversely, in other countries, such as the United Kingdom, IBCLCs usually only work in the hospital setting if they are also a health professional. Although their additional skills enable them to effectively carry out their role, the credential itself is often incidental, and not formally part of the person specification. Other staff may not always have specialist breastfeeding knowledge and skills, meaning parental experience can differ depending on who they encounter (Holaday et al., 1999;Dykes, 2006;McLaughlin et al., 2011). Additionally, lactation support is often limited to maternity and neonatal care units, meaning it is often not routinely present on paediatric units.
Alongside specialist services, we know that breastfeeding is best facilitated when all health professionals looking after a mother recognize its value and have the skills to support her or signpost for more specialist support if needed (McFadden et al., 2017;Thomas, 2020).
However, although UNICEF Baby Friendly standards support and protect breastfeeding on the neonatal and maternity wards, these do not currently extend into paediatrics (Carney & Bruce, 2011). Therefore, there are no standardized, mandatory training programmes for paediatric nurses, physicians and allied health professionals such as dieticians, speech and language therapists, physiotherapists and occupational therapists-all of whom are likely to work with medically complex children. The World Breastfeeding Trends Initiative (WBTi, 2020) identified many risk factors among the training curriculums of health professionals, noting that there are many gaps (Gupta et al., 2019).
The support, training and attitudes of health professionals are considered as a separate theme, as some units and hospitals in the review had designated lactation support that was considered alongside medical treatment. Other units and hospitals had no such identified service, and therefore, any lactation support was provided by the medical team-who may or may not have the required skills and training to offer support.
When direct, exclusive, responsive breastfeeding is not possible, extra feeding equipment will be needed, both for parents and for infants. For example, parents will need to maintain their milk supply with a breast pump (Marasco, 2008) -either a hospital-grade double electric pump, single electric pump or manual breast pump, together with hand expressing and breast massage (Geddes et al., 2013;Morton et al., 2009;Morton et al., 2012;Morton, 2014;Witt et al., 2016;Strauch et al., 2019). Different approaches work best for individual mothers (Meier et al., 2016). Specialized bottles and teats (e.g. squeeze bottles and one-way valves), cups, spoons, syringes, NGT, nipple shields, palatal prostheses, at-breast supplementers and GT may also be needed (Rosenberg et al., 2008;Boyce et al., 2019;Rudra et al., 2016). A thin silicone nipple shield may increase the effectiveness of milk transfer for infants unable to achieve good intraoral pressure at the breast (Meier et al., 2017). Parents will likely need further education around using these products and maximizing milk supply.
Very few of the studies specifically studied the use, education or availability of equipment. There are many aspects of using specialized equipment that are missing, such as the possibility of expressing milk at the infant's bedside, how to optimize milk production in difficult circumstances and utilize specialist equipment and specific techniques for positioning infants with low tone, fatigue or orofacial anomalies.

| Limitations of this review
A major limitation of this review is that it was conducted by a single reviewer. This was unavoidable as it forms part of a PhD. The process was made more rigorous by a second reviewer checking the criteria used and being involved in the development of the review. This systematic review is also small, so all co-authors became familiar with the studies analysed.
No study explored the impact of illness in a general sense on breastfeeding. There is a paucity of research related to infant acute illness and serious conditions that do not specifically affect the head, mouth, palate or face.
The available studies have all explored the relationship between illness/disability and breastfeeding outcome in a disease-specific way, without drawing out more general themes. Becasuse all infant and child conditions will affect breastfeeding differently, with so few conditions studied it is hard to know whether some aspects of infant feeding difficulty have not yet been identified. The data are therefore not necessarily generalizable.
Additionally, most paediatric wards admit children with a range of diseases and illnesses. It is perhaps more user-friendly for health pro- The studies included also tended to have limited racial diversity and SES among the included participants. This may be representative of the ongoing higher prevalence of breastfeeding in high-SES groups and among predominantly White, married, heterosexual, women with higher levels of education (Bartick et al., 2017). Again, this limits the generalizability of the findings to the wider population.
In addition, the studies came from different parts of the world where healthcare systems are disparate. Half of the studies were conducted in the United States or Canada. This is potentially problematic in terms of exploring healthcare-based lactation support, as the provi-

| CONCLUSIONS
There is much we know about breastfeeding in terms of risk reduction of various illnesses (Victora et al., 2016), yet we know far less about what it is like to breastfeed a medically complex child. Although breastfeeding reduces the risk of many conditions, it does not eliminate risk. In the important work of continuing to promote of breastfeeding in general, we must not forget the children who are unwell despite having been breastfed.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.