This article is a part of the Interventions and policy approaches to promote equity in breastfeeding collection, guest-edited by Rafael Pérez-Escamilla, PhD and Mireya Vilar-Compte, PhD

Introduction

In 2016, South Africa recorded 67.3% of infants initiating breastfeeding within 1 h of birth, and only 31.6% being exclusively breastfed, with a mean exclusive breastfeeding duration of 2.9 months [1], the lowest rates on the African continent [2]. South Africa, an upper-middle income country in Sub-Saharan Africa has a population of about 58 million [3, 4]. While South Africa is rich in natural resources, its social indices reflect structural vulnerabilities and inequities, such as healthcare gaps and the uneven impact of HIV/AIDS [5]. South Africa’s maternal mortality ratio is 119/100,000 live births [6], and child mortality rates have been steadily declining with infant and under-5 mortality rates of 28 and 35 per 1000 live births compared to Sub-Saharan Africa’s rates of 52 and 76 respectively [7, 8].

Breastfeeding is vital for an infant’s development and survival as it reduces morbidity and mortality from diarrhea, pneumonia and malnutrition, particularly in infants under-5 years [9,10,11]. Breastfeeding also reduces the risk of childhood obesity and fosters cognitive development [10, 12,13,14]. Furthermore, it confers health benefits to mothers including reduced risk of cancers (breast, ovarian), hypertension and diabetes [10, 15, 16].

The World Health Organization (WHO) and the United Nations Children’s Fund (UNICEF) recommend exclusive breastfeeding (EBF) of infants from birth to 6 months, followed by the introduction of nutritious and safe complementary foods with continued breastfeeding for at least 2 years [9]. The United Nations (UN) policy brief on The UN Decade of Action on Nutrition includes five targets to eradicate all types of malnutrition, as well as six targets to improve maternal, child health and nutrition by 2025 [17]. Target 5 on breastfeeding, aims to increase exclusive EBF rates to at least 50% by 2025 [17]. Hence, it is of concern that globally just 44% of infants are breastfed soon after birth, and 40% of those less than 6 months old are exclusively breastfed [9, 18]. In 1985, HIV infection through human milk or breastfeeding was first recognized in the United States [19, 20] and the Centers for Disease Control and Prevention issued guidelines for HIV-positive women to adopt replacement feeding [21].

In terms of HIV-positive women, WHO guidelines on infant feeding have drastically evolved over time. Initially in 2001, the WHO advised women not to breastfeed when affordable and safe human milk substitutes were available due to evidence supporting the mother to child transmission (MTCT) of HIV through human milk or breastfeeding [22]. By 2006, WHO advised that early cessation of breastfeeding before 6 months was no longer required for HIV-positive women [23]. In 2010, they advised women on anti-retroviral treatment (ART) to exclusively breastfeed for 6 months and continue breastfeeding for 1 year [10, 24]. Finally in 2016, the timeline for breastfeeding was extended to at least 24 months, and mixed feeding was no longer considered a risk factor for MTCT as long as ART was available [9].

In 2011, South Africa changed its infant feeding policy to EBF for all women regardless of HIV status [25]. This shift in policy conveyed through the Tshwane declaration was aligned with WHOs 2010 guidelines on HIV and infant feeding that realigned breastfeeding guidelines of HIV-positive women with HIV-negative women as long as ARTs are available [25, 26]. The declaration also called for not distributing free infant formula for HIV-positive women (except for medical reasons) at public health facilities, and for these facilities to become “baby-friendly” by 2015 [18, 25].

The UNICEF/WHO Baby Friendly Hospital Initiative (BFHI) launched in 1991 was designed to strengthen the capacity of maternity facilities to protect, promote, and support successful breastfeeding [27]. The Tshwane declaration also called for the government to establish legislation to enforce the WHO International Code on Marketing of Breast Milk substitutes (The Code) [25]. The Code includes specific guidelines to protect, promote and support breastfeeding by regulating the advertising and sales of breastmilk substitutes (BMS), bottles and teats [28, 29].

While The Code was approved since 1981 by the World Health Assembly [29] violations continue to be widespread globally due to weak Code monitoring and enforcement mechanisms across countries, especially in low-to-middle income countries (LMIC) [30,31,32]. As a result, infant formula sales have increased exponentially in LMICs, including in South Africa resulting in higher rates of morbidity and mortality [30, 31, 33, 34].

South Africa also has one of the highest rates of HIV in the world with an adult prevalence of 19.2% [4]. Breastfeeding outcomes are suboptimal in South Africa [11, 33, 35]. Some have attributed it, at least in part, to the relatively easy access that women and families have had to infant formula as a result of infant formula distribution programs designed to prevent MTCT of HIV [36]. Apparently, this policy may have had an undesirable spill-over effect on HIV-negative women as well. Hence, South African academics and advocates have called for enforcing The Code within the framework of the most recent WHO guidelines on infant feeding in the context of HIV [9, 33]. Enforcing The Code is paramount in the context of WHOs most recent guidelines on infant feeding [9] to improve EBF. Thus, the aims of this scoping review were to: (a) describe EBF practices in South Africa, (b) determine how EBF has been affected by the HIV infant feeding policies followed since 2006, and (c) assess if the renewed interest in The Code has had any impact on breastfeeding practices in South Africa.

Methods and analyses

We applied the PRISMA Extension (PRISMA-ScR) and Joanna Briggs Institute guidelines for scoping reviews [37, 38]. The review protocol was registered in Open Science Framework (OSF) (https://osf.io/sxcfv/). A highly experienced medical librarian (KN) conducted a peer-reviewed comprehensive search of multiple databases.

Information sources and methods

The databases searched in this project and their platforms were: MEDLINE All (Ovid), PsycINFO (Ovid), Embase (Ovid), Global Health (Ovid), Web of Science Core Collection (as licensed by Yale University, Core Collection included Science Citation Index Expanded (SCI-EXPANDED), Social Sciences Citation Index (SSCI), Arts & Humanities Citation Index (A&HCI), Conference Proceedings Citation Index- Science (CPCI-S), Conference Proceedings Citation Index- Social Science & Humanities (CPCI-SSH), Book Citation Index– Science (BKCI-S), Book Citation Index– Social Sciences & Humanities (BKCI-SSH), Emerging Sources Citation Index (ESCI) --2015-present, Current Chemical Reactions (CCR-EXPANDED), and Index Chemicus (IC), Dissertations and Theses Global (ProQuest), Africa-Wide Information (Ebsco), CENTRAL (Cochrane Library), CINAHL (Ebsco), and Africa Index Medicus (Global Index Medicus). All titles/abstracts and texts were screened in Covidence except for the South African National ETD Portal (via netd.ac.za). This portal was searched online, but because records’ export was not possible, the potentially relevant records were screened by an author (DV) online and only records requiring full text screening were added to Covidence. Each database was searched individually, using a combination of keywords and (if available) controlled vocabulary. No study registries were searched. No citation chaining was done.

Records published between 2006 and the dates of the searches in 2020 were retrieved. This date limit was used because this was the period of major policy shifts in WHOs infant feeding guidelines in the context of HIV. No limits were imposed with respect to study design and languages. Conference papers were not retrieved. Papers with animal subject indexing were excluded in some databases, but only if they did not also have human subject indexing. This search strategy was developed independently and did not use any published or validated filters. The MEDLINE search strategy was peer reviewed by an independent medical librarian using the Peer Review of Electronic Search Strategies Guideline [39].

Search strategies

The search strategies for each database are available in full. Search terms for Ovid Medline are presented in Table 1. While the strategies used appropriate syntax, indexes, and controlled vocabulary for the various databases, each one includes queries for the setting (South Africa) and queries for the broad topic (infant feeding). The strategies do not require an explicit reference to the WHO HIV infant feeding guidelines or The Code.

Table 1 Search terms for Ovid MEDLINE
Full size table

Inclusion and exclusion study criteria

For studies to be included, they needed to be peer reviewed, grey literature technical reports or theses and dissertations. We considered quantitative, mixed-methods and qualitative study designs. Studies had to be conducted in South Africa.

Types of participants

Eligible participants were women in South Africa who delivered a healthy live birth (birth to 24 months) with known infant feeding practices.

Interventions

We focused on two policy-level interventions, namely, the World Health Organization (WHO) Updates on HIV and Infant Feeding Guidelines (2016) [9], and The International Code of Marketing of Breastmilk Substitutes (The Code) [28, 29]. The updated 2016 (most recent) infant feeding guidelines recommend lifelong antiretroviral therapy for anyone diagnosed with HIV, including those who are pregnant and breastfeeding. It also provides guidance on appropriate infant feeding practices for mothers living with HIV [9]. Women who are HIV-positive and receiving antiretrovirals are advised to breastfeed following the same breastfeeding recommendations for HIV-negative women. The objective of The Code initially endorsed in 1981 by the World Health Assembly of WHO was to ensure safe and adequate nutrition for infants by protecting and promoting breastfeeding. It specifically sought to regulate the marketing of products such as breastmilk substitutes including formula, other types of milk, beverages and equipment (bottles, teats) [28, 29].

Outcomes

The infant feeding/breastfeeding outcomes considered were: Breastfeeding initiation within 1 h of birth; Exclusive breastfeeding up to 6 months; Any breastfeeding up to 12 months of age; and Continued breastfeeding from 12 to 23 months.

Screening

All references retrieved were first de-duplicated by KN, then uploaded into Covidence [40]. Two independent reviewers (DV and MVC) initially screened a sample of 20 references (titles/abstracts) to ensure consistency and measure inter-rater literature screening reliability.

Data collection and analyses

Data extraction

Data were independently extracted and entered on a standard Microsoft Excel spreadsheet form by two authors (DV and MVC). Studies which did not meet inclusion criteria and outcomes were excluded. Any disagreement was resolved in consultation with the senior author (RPE). Data extraction fields included author names, journal name and year of publication, study design, location of study, sample characteristics, number of participants, and outcomes. These outcomes included: Breastfeeding initiation within 1 h of birth; Exclusive breastfeeding up to 6 months; Any breastfeeding up to 12 months of age; and Continued breastfeeding from 12 to 23 months.

Assessment of methodological quality of the included studies

We did not evaluate the methodological quality and risk of bias of the included studies since this is not required for a scoping review [37, 38].

Analyses and reporting

Our findings are reported based on the PRISMA Extension and Joanna Briggs Institute guidelines for scoping reviews. We analyzed the data extracted from included studies based on the HIV status of participants and their infant feeding practices and outcomes wherever provided.

Results

A total of 5431 citations were retrieved from eleven databases. Duplicates were removed in EndNote and by Covidence. Records from all the databases except National EDT Portal underwent title-abstract screening in Covidence; records from National EDT Portal underwent title-abstract screening on the National EDT Portal website. 35 National EDT Portal Records were identified as deserving full text screening and were added to Covidence at that stage. 1588 unique records were processed in Covidence and 179 records met the criteria for full-text screening. The final number of included records was 83 (Fig. 1 PRISMA flow diagram).

Fig. 1
figure 1

PRISMA flow diagram)

Full size image

Characteristics of included studies

Of the 83 studies included, nine were qualitative, 11 RCTs, and 63 observational. All nine South Africa provinces were represented; the highest number of studies were conducted in Kwa-Zulu Natal (34) and the lowest number in Northern Cape (2). Other studies were in Western Cape (25), Gauteng (15), Eastern Cape (14), Limpopo (13), North West (7), Mpumalanga (6) and Free State (5). Studies were conducted primarily in rural (31) and urban (25) settings; other settings included semi-rural (3), semi-urban (6) and peri-urban (22).

While the articles were published between 2006 and 2020, we interpreted findings in the context of the time period when the study was conducted. The number of articles were about evenly divided between the two periods of major changes in WHOs breastfeeding guidelines for HIV-positive women; prior to 2011 (n = 40 articles) and from 2011 onward (n = 43 articles).

Infant feeding practices in South Africa (2006–2010)

The studies conducted during 2006–2010, i.e. before the release of the 2010 guidelines recommended HIV-positive women on ARVs to exclusively breastfeed for the first 6 months after birth (Table 2) revealed that HIV-positive women with 0–6 months infants mainly practiced formula and mixed feeding [41,42,43,44]. EBF rates in this age group ranged from 13 to 77% [44, 45].

Table 2 Included studies 2006–2010
Full size table

During this period, it was common for HIV-positive women who initiated breastfeeding to stop doing so prior to 6 months after birth (1–3 months) [46,47,48,49]. Low rates of breastfeeding were observed for a variety of reasons including free formula distribution, not wanting to transmit the virus to the infant, and work/school obligations [43, 50,51,52,53]. However, for HIV-positive women who practiced exclusive formula feeding (EFF) initially, lack of access to infant formula was one of the reasons for mixed feeding [54, 55].

While breastfeeding initiation rates were high among HIV-negative women, ranging from 52.7 to 97% [48, 49, 56,57,58,59], EBF rates declined as liquids and solids were introduced before the infants were 6 months old [60,61,62,63]. Other reasons for infant formula use included non-disclosure of HIV status, family pressure, and cultural practices [41, 54, 55, 57].

Infant feeding practices in South Africa (2011–2020)

The post-2010 period (Table 3) after South Africa endorsed the 2010 WHO infant feeding guidelines for HIV-positive women recommending EBF for 6 months irrespective of HIV status resulted in a wide range of EBF rates among HIV-positive (26–99%) and HIV-negative (12–92%) women [64,65,66,67,68,69]. However, EBF duration was brief (1-3 months) [66, 70, 71]. While initiation of breastfeeding was high, there were low rates of EBF among HIV-negative women with introduction of liquids and other foods prior to 3 months [72,73,74,75]. In some instances, weaning occurred as early as 2 months [76]. HIV-positive women expressed fear of HIV transmission to their infants, as well as school and work commitments as reasons for discontinuation of EBF [68, 77,78,79,80].

Table 3 Included studies 2011–2020
Full size table

RCTs testing interventions to improve breastfeeding practices and outcomes

Two major multi-country RCTs conducted during 2005–2008 among HIV-positive women and infants ≤6 months old were the Kesho Bora trial [81, 82] and the PROMISE-EBF trial [83].

In the Kesho Bora trial [81, 82], conducted in Burkina Faso, Kenya and South Africa (Durban, Somkhele), HIV-positive pregnant women were randomized into two groups who received either triple ART during pregnancy through the breastfeeding period (intervention), or short course prophylactic therapy until delivery (control). All HIV-exposed infants received single-dose nevirapine at birth. Women were also counselled to either breastfeed with cessation by 6 months or formula feed from birth (free infant formula was provided for 6 months). The main objectives were to determine rates of HIV transmission, infant survival at 6 weeks and 12 months and adverse events. Bork and colleagues [81], found breastfeeding initiation to be lower in Durban (57.1%) than rural Somkhele (80.9%) among 751 HIV-exposed infants. Overall, they found that non-breastfed infants (0–6 months) had higher morbidity risks than those breastfed, with increased risk for serious infections (e.g. severe diarrhea) between 0 and 2.9 months [81].

The PROMISE-EBF trial was a behavior change intervention to promote EBF using peer counsellors in Burkina Faso, South Africa (Paarl, Rietvlei and Umlazi), and Uganda [47, 83,84,85,86]. This counselling intervention (N = 2579 mother-infant pairs) included one antenatal breastfeeding peer counselling visit and four postnatal peer visits. The two main outcomes of interest were prevalence of EBF and diarrhea at ages 12 and 24 weeks. Overall, Tylleskar and colleagues [83] found that EBF prevalence (all countries) in the intervention groups at 12 weeks was double that of the control groups. However, South Africa had exceedingly low EBF rates at baseline (10%), compared with 79% in Burkina Faso and 82% in Uganda [83]. There were no significant differences in the prevalence of diarrhea (all countries) between the two groups at either 12 or 24 weeks of age [83]. Finally, the authors [83] found that while the peer counselling intervention was effective in increasing EBF rates in Uganda and Burkina Faso, it was not effective in improving breastfeeding rates in South Africa.

Violations of the international code of marketing of breastmilk substitutes

South Africa took another bold step in 2012 to enforce The Code via legislative action (i.e., regulation R991 of 2012), which sought to regulate the sales, advertising, marketing, information and education of foods promoted for infants and young children [87, 88]. The specific objective of regulation R991 was to protect and support breastfeeding by creating an environment free from the relentless marketing strategies of BMS manufacturers, and prevent conflicts of interest among the healthcare staff or other child care providers [87, 88]. Unfortunately, this legislation has had limited positive effects as violations of the Code are still prevalent, highlighted in the three articles discussed below.

Of the three included studies which focused on Code violations, two were Master of Public Health theses [89, 90] and the other was an expert commentary from academics [33] (Table 3). Muravha [89] investigated Code violations among health workers in four Provinces and 40 health facilities and found that four health workers accepted free gifts (pens, calendars, posters) from a BMS company, despite being aware of the R991 regulations. Health workers also received education materials (leaflets, booklets) and equipment (South African water bags for adult usage) which were branded with the manufacturer’s name, but not the name of a specific product marketed by the BMS company. Radebe [90] examined media (radio, television, print) infractions and identified 30 marketing violations from 117 baby products (formula, bottles and teats) advertised in magazines targeting primarily pregnant women or families. The author reported that these numbers are likely to be underestimated, since the analysis did not include all media sources.

Finally, Lake and colleagues’ commentary [33] documented anecdotally the marketing strategies of the BMS industry and their reach to health workers and other stakeholders through sponsorships of conferences and other scientific meetings, misleading information on infant formula on company websites, health promotion materials and support for staff salaries. The authors called for improved leadership efforts in enforcing the Code and strengthening breastfeeding interventions.

Discussion

During the early years of the HIV pandemic, recommendations for infant feeding were guided by scientific evidence indicating that when safe feeding with breast milk substitutes was universally available, as was the case in the United States and other high income countries, it was appropriate to recommend for HIV-positive mothers to not breastfeed at all [21].

By contrast, as part of their infant feeding decisions HIV-positive mothers in low income countries needed to consider the risk of their children dying if they were not breastfed as a result of having access to safe replacement feeding alternatives. Since then, HIV infant feeding guidelines have evolved over time. WHO guidelines in this area switched from exclusive breastfeeding with abrupt weaning from the breast before 6 months to current guidelines recommending EBF for 6 months followed by breastfeeding continuation for at least 12 months if the mothers have ART access [9, 22,23,24].

The initial guidelines implemented in South Africa in 2008 [36] provided HIV-positive mothers with widespread access to cost-free infant formula at public health facilities. Unfortunately, these polices, albeit well-intentioned may have inadvertently negatively impacted the EBF behaviors among HIV-negative mothers and led to subsequent increases in infant morbidity and mortality in South Africa and similar settings [91]. Our review indeed suggests as reported in other studies that there may have been a spill-over from an infant feeding policy driven by HIV-positive women to HIV-negative mothers explaining why EBF rates are still low in South Africa [44, 92]. Unfortunately, there is no evidence that EBF rates significantly increased after the Tshwane declaration.

One of the challenges highlighted in this review is that HIV usually coexists with poverty creating a syndemic-like effect in the lives of women, as breastfeeding was disincentivized in the early WHO guidelines, generating inequities in access to infant feeding choices. In addition, as the South African HIV infant feeding guidelines evolved and pushed for phasing out free formula distribution [25, 36], some key challenges remain: (i) changing breastfeeding behaviors, social norms and medical practices have been slower than desired; and (ii) marketing of BMS and Code violations have persisted. These factors have disproportionately affected HIV-positive mothers mainly located in rural and peri-urban areas of South Africa where not breastfeeding was associated with increased risk for serious infections (chronic diarrhea, lower respiratory tract infections) and death [45, 81, 84]. Further research is needed to address how poverty affects women’s ability to successfully implement the national infant feeding guidelines for HIV-positive and HIV-negative women.

One of the strengths of this review is that it highlights diverse studies conducted in South Africa on breastfeeding in the era of HIV among both HIV-infected and uninfected women, and that EBF rates are subpar in both groups as a result of the premature introduction of liquids and solids. However, there is a dearth of policy-responsive implementation research to inform how the more recent HIV infant feeding national guidelines can be successfully implemented in South Africa, a country where this should be possible because there is now almost universal access to ART among HIV-positive women [93,94,95,96], and a call for strong protection, promotion and support for breastfeeding including those who are HIV-positive [33, 97]. As it has been found in other countries [98,99,100,101], our findings suggest that strong Code enforcement combined with increased investments in breastfeeding protection, promotion and support programs in South Africa are needed for the country to make progress towards meeting the United Nations target to increase breastfeeding rates to at least 50% by 2025 [17].

Because the infant formula distribution program was coordinated by the health sector, the findings from our review are highly consistent with previous studies showing that health workers play an important and influential role in counseling or advising mothers on infant feeding practices [77, 102,103,104,105,106].

Therefore, it is key that the WHO Code is properly enforced to allow for the environment conditions surrounding mothers in South Africa to become stronger enablers of optimal breastfeeding practices [106]. Unfortunately, in our review we did not find evidence that this has already started to happen yet in South Africa.

Although penalties exist for first and subsequent violations of regulation R991, with the first penalty being a fine and at most 6 months imprisonment [107], to date no individual or organization has been prosecuted for Code violations, a strong indicator of lack of enforcement or commitment to change the status quo. Thus, routine monitoring of this key legislation is required as encouraged by other academics and advocates [33, 97, 108]. As such, enforcing regulation R991, particularly the outreach to health workers by BMS manufacturers and their associates is imperative.

Conclusion

The free distribution of infant formula combined with the BMS industry’s marketing practices that violate the WHO Code have played a role in suboptimal breastfeeding practices among both HIV-positive and negative women in South Africa [33, 109, 110].

This scoping review integrated evidence on infant feeding practices, especially EBF rates among HIV-positive and HIV-negative women in South Africa in the context of rapidly evolving HIV infant feeding guidelines from the WHO. Although highly effective ARTs has made breastfeeding for HIV-positive women safe, and South Africa has widespread access to ART [93, 95, 96], it is discouraging that women continue to cite fear of HIV transmission to their infants as a reason for either not breastfeeding or doing so for short periods of time. This finding calls for improved access to high quality breastfeeding counselling, education and awareness campaigns from the local health care facility to the national level.

Although monitoring and enforcement of the Code remain nonexistent, there has been progress in strengthening legislation. The most recent Status Report (2020) from WHO, UNICEF and IBFAN found 70% (136/194) of WHO Member States had new Code legislation [111]. However, only 31 countries regulated milk products for infants up to at least 36 months. Moreover, just 58% (79/136) ban the promotion of BMS at health facilities [111].

Even though enforcing the most recent WHO guidelines and the WHO Code are necessary to improve breastfeeding outcomes in South Africa, they are not sufficient because as our review shows, there are additional barriers that impact breastfeeding outcomes, including lack of social support among women returning to work or school after the birth of their children.

Mixed-methods research, including in-depth interviews with key informants representing different government sectors and civil society is needed to prioritize actions and strategies to make this happen in South Africa. This effort should be followed by implementation research and policy instruments [112, 113] that can guide South Africa in its efforts to scale up the protection, promotion, and support of breastfeeding programs at the national level in the context of the HIV pandemic.