Anna Coutsoudis, Prof, PhD, BSc Hons
Expected Key Learning Outcomes
The safety of breastfeeding for HIV positive mothers
Recommendations for antiretroviral (ARV) treatment
Considerations for decision makers
Since the first report in 1985 that Human Immunodeficiency Virus (HIV) could be transmitted via breast milk there has been considerable debate, controversy and confusion around breastfeeding by HIV-infected women. The initial response from the US Centers for Disease Control and Prevention recommended that HIV-infected women should not breastfeed but instead make use of replacement milk feeds . This caused problems for poorer countries, which were unlikely to be able to follow this guidance without putting their infants at considerable risk of mortality. The World Health Organization (WHO) and other agencies then faced the dilemma of dual guidelines, one for rich countries and one for poor. WHO finally advocated a shorter duration of breastfeeding (3–4 months) as the safest option, or the complete avoidance of breastfeeding if it could be done safely .
Considerable efforts were leveraged to provide “safe” replacement feeding to HIV-exposed infants, and several countries and international agencies distributed free formula milks to HIV-infected mothers to reduce HIV transmission through breastfeeding. Several calls were made to review these policies ,  but these generally went unheeded with governments and non-governmental organisations choosing to concentrate on minimising the risks of environmental contamination during the preparation of formula milk feeds. They ignored the fact that no matter how “hygienic” the formula milk feed, excluding breast milk from the infant’s diet also excluded vital immune components that confer the protection  needed to reduce the risk of infectious disease mortality.
During this time, a surprising observation was documented in a study in South Africa. This study reported that mothers who had exclusively breastfed their infants for at least 3 months had a significantly lower risk of transmitting HIV to their infants during the breastfeeding period compared to mothers who had not exclusively breastfed .
These findings were soon replicated in several other African studies, and the information was incorporated into the revised WHO guidelines in 2006 . These revisions recommended 6 months of exclusive breastfeeding for women in whom replacement feeding was not considered to be acceptable, feasible, affordable, sustainable or safe. In recognition that risk associated with formula feeding was variable and dependent on many factors, including access to clean water and health services and other maternal and social factors, WHO recommended that replacement feeding should be considered only when it was deemed to be “acceptable, feasible, affordable, sustainable or safe”.
As these guidelines were put into place, many programmes were reporting that mothers and health care workers were interpreting the guidelines to recommend 6 months exclusive breastfeeding only. This resulted in early cessation of all breastfeeding when the period of exclusive breastfeeding ended. A randomised controlled trial in Zambia reported that early cessation of breastfeeding resulted in increased mortality of all infants but especially of those infants who were HIV infected . Country reports and several studies also began to detail increasing diarrhoeal and pneumonia deaths by excluding breastfeeding to prevent HIV transmission , , .
With the reminder of the important role that breastfeeding plays in child survival, researchers turned their attention to ways of preserving breastfeeding while making it safer by reducing HIV-transmission risk. It had been clearly established that the risk of transmission by breastfeeding is closely associated with the mothers CD4 cell count and viral load. Thus, it followed that antiretroviral drugs (ARVs), which are effective in reducing HIV viral load, could play a role in reducing HIV transmission during breastfeeding (as for in utero and intra partum transmission). Studies showed that when mothers were provided with a combination of ARV drugs or when children were provided with single or dual drug prophylaxis, the risk of transmission was reduced to < 2%. Several studies in the years 2000–2009  showed that such drug use resulted in significant reductions in HIV transmission during breastfeeding.
Following these studies, further revision of the guidelines became necessary. In 2010, WHO recommended ARVs to prevent postnatal transmission of HIV through breastfeeding , . In addition, the WHO recommended a public health approach that member states should principally promote and support a feeding practice for all HIV-infected women who accessed public health facility care. Previous guidelines had been dependent on health care workers as they encouraged an individualised approach of counselling HIV-infected women on feeding options dependent on each woman’s individual household and social circumstances.
The new guidelines called for a change from this individualised approach to a public health approach, where countries were encouraged to consider the socio-economic and cultural contexts of their populations and then choose to support either breastfeeding or formula feeding in their public health services. It was recommended that women should breastfeed for about 12 months and that those women with CD4 cell count ≤ 350 should be prioritised for ARV treatment. For women with CD4 cell count > 350, WHO recommended that either infants received Nevirapine (NVP) prophylaxis for the duration of breastfeeding (known as Option A) or mothers received ARV prophylaxis with three drugs (known as Option B).
Again, new research findings necessitated reformulating guidelines. Some studies were extrapolated to suggest that if HIV-infected pregnant or lactating women were started on ARV treatment at CD4 cell counts < 350, it would improve their health and make them less likely to transmit the virus to their sexual partners. In view of the potential added benefit to maternal health, WHO recommended that countries adopt an Option B approach instead of Option A .
When the United Nations International Children’s Fund (UNICEF) published the Clinton Health Access Initiative Business Case , there was strong encouragement that countries extend their programmes and adopt the newly coined programme, Option B + . This option recommended that all HIV infected women started ARVs early in pregnancy and stayed on them for life; thus, women with higher CD4 cell counts were no longer expected to discontinue ARVs after cessation of breastfeeding. Consequently in 2013, when WHO published Consolidated Guidelines on the use of ARV drugs for treating and preventing HIV infection they included the Option B + recommendation on the use of ARVs for prevention of post-natal transmission . In 2016, the WHO Guidelines were updated again and of note was new guidance recommending that the period of breastfeeding should increase from 12 to 24 months .
There is now considerable evidence that transmission of HIV during breastfeeding over 6–12 months is negligible, particularly when mothers begin ARV prophylaxis during pregnancy from 14 weeks, are adherent to their ARVs during pregnancy, intra-partum and breastfeeding, and have an undetectable viral load. This has been confirmed in a recent study that reported a transmission rate of 0.28% through 12 months of breastfeeding .
Although ARV interventions seem to be a guaranteed method for reducing HIV transmission, two major problems to the prevention of transmission by breastfeeding are particularly common in the developing world:
Late presentation for antenatal care, resulting in mothers commencing ARV late in pregnancy and therefore entering the delivery and breastfeeding periods with inadequately suppressed viral load, which is a major risk factor for transmission.
Inadequate adherence to ARVs during the breastfeeding period, with the infant probably most vulnerable during the early breastfeeding period. Inadequate adherence to ARVs is a major problem in the developing world. There are many reasons for this: stigma, lack of social support, food insecurity, lack of understanding of the dangers of non-adherence, and health system failures resulting in depleted drug stocks.
As discussed, non-exclusive breastfeeding is a risk factor for HIV transmission. This resulted in calls to encourage exclusive breastfeeding in the first 6 months of birth. Similarly, high HIV viral load was identified as a major risk factor. Programmes were therefore implemented to provide ARVs to mothers and/or infants to lower their viral load, including that in breast milk, and thus reduce transmission.
The extent to which exclusive breastfeeding and ARVs are independent variables of HIV transmission is not clear. Currently, there is no clear answer as to whether the risk of non-exclusive breastfeeding is eliminated by ARV use. However, it is clear that even with ARVs it is important to promote and encourage exclusive breastfeeding because of its independent effects on preventing morbidity and late-onset metabolic disorders.
Before the use of ARVs to prevent HIV transmission during breastfeeding, mothers with mastitis and other breast pathology were reported to be at increased risk of transmitting HIV to their infants. Women with any breast pathology were encouraged not to feed from the affected breast but to express and discard breast milk from the affected side while continuing to feed from the unaffected side. As with non-exclusive breastfeeding, there is insufficient information to clarify whether the risks of breast pathology are eliminated by ARVs; as such, the current recommendation is to use the old guidelines until new data become available.
Waitt et al. conducted a systematic review and meta-analysis of 24 studies investigating the concentrations of ARVs in breast milk . Relatively low ARV penetration into breast milk was found when compared to levels in maternal plasma, although there was considerable variability in methodologies of extraction. Overall, studies appeared to indicate that nucleoside reverse transcriptase inhibitors (NRTIs) have higher and more variable breast milk penetration than non-NRTIs or protease inhibitors (PIs). Accumulation of PIs in breast milk was found to be minimal. Transfer of these drugs to the infant is also variable; lamivudine (3TC) and NVP appear to have higher transfer rates (5–10%) while Efavirenz had a much lower rate (2–3%).
A concern is that transfer of low levels of individual drugs to the breastfed infant would potentially confer high rates of drug resistance should the infant later become infected. However, given the negligible risk of infection, this may not be an over-riding issue. Other risks associated with transfer of ARVs to the infant are minimal, and usually include increased risk of anaemia and neutropenia.
Questions remain unanswered about which combination of ARVs to use, whether providing ARVs to mother or infant give better outcomes overall and in terms of pregnancy, infant and maternal side effects. Over the last few years there have been reports of associations between anaemia and Zidovudine (ZDV), neutropenia and NVP, and cardiac toxicity and adrenal dysfunction and Kaletra. A systematic review suggested that extended ARV prophylaxis does not have a negative impact on the growth and incidence of non-HIV infections in HIV-exposed infants , although more data are needed.
The recently published PROMISE study has provided valuable information to answer the question as to whether ARV prophylaxis should be given to mother or infant. The study with approximately 2400 breastfeeding mother-infant pairs, randomized the pairs either to mother receiving ARV prophylaxis or infant receiving prophylaxis with nevirapine for the duration of breastfeeding (median time 16 months). The results showed that both strategies were safe and resulted in similarly low breastfeeding transmission rates (0.57 and 0.58%).
Recent high-quality research has shown that HIV infected and uninfected mothers have similar concentrations of serum and breast milk immunoglobulins and cytokines . However, additional information on the impact of ARVs on the immunological and nutritional components of breast milk is required.
Immune therapy and vaccination are currently being investigated as possible strategies to reduce HIV transmission during breastfeeding . Data are needed on whether the vaccine should be administered to mothers or infants. A variety of trials has been conducted and several vaccines have been tested, which show the strategies to be well-tolerated (efficacy data are still awaited). Preliminary studies using human monoclonal antibodies (mAb) are promising. Currently, studies are investigating use of VRC01, a human mAb targeting the CD4 binding site of HIV-1 gp120, which has broad neutralising activity against viruses .
As detailed above, using ARVs in mother and/or infant is the mainstay of making breastfeeding safer for infants born to HIV-infected mothers. However, there are two other important strategies, exclusive breastfeeding and pasteurisation of breast milk.
It is well established that exclusive breastfeeding compared to non-exclusive breastfeeding reduces HIV transmission to infants. Smith and Kuhn, 2000, discussed the physiological mechanisms to explain this robust finding . Of importance is that exclusive breastfeeding has been found to be associated with a significantly lower risk for breast pathology , which would account for its reduced risk for HIV transmission.
Exclusive breastfeeding is also known to influence the establishment of the microbiota of primate infants, which in turn influences the development of the immune system. A recent study reported that exclusively breastfed rhesus macaques developed robust populations of memory T cells and T helper-17 cells within the memory pool, unlike their formula-fed counterparts, which would explain the differences in protection against infection .
Early work from South Africa showed that heating a single bottle of breast milk in a pot of water to 62.5 °C for 30 minutes was effective in destroying HIV. This methodology was based on the Holder pasteurisation method, which is used widely to pasteurise breast milk in human milk banks. In 2000, Chantry and colleagues simulated a different method known as flash pasteurisation or high temperature, short time heating, using temperatures of 72 °C for 15 seconds to destroy HIV in breast milk .
Using this methodology, Chantry’s group was able to demonstrate that such “flash-heating” was capable of inactivating spiked cell-free HIV-1, as detected by reverse transcriptase activity . Additionally, flash-heating was shown to inactivate HIV in naturally-infected breast milk samples collected from HIV infected mothers in South Africa . Importantly, the researchers showed that flash-heat methodology inactivated cell-associated HIV as well as cell-free HIV . Pilot data suggested limited negative impact of flash-heating on vitamins and proteins . Work on naturally infected breast milk samples from women in South Africa showed similar limited negative impact .
Flash-heating is capable of eliminating pathogenic and non-pathogenic bacteria and an 8-hour storage period outside the refrigerator does not result in a significant increase of bacteria . Furthermore, the heat treatment did not diminish the bacteriostatic activity of breast milk . Importantly, the majority of breast milk’s immunoglobulin activity survives the heating, suggesting flash-heated breast milk is immunologically superior to breast milk substitutes. Additional studies showed that flash heating resulted in a decrease of 20% in total immumoglobulin (Ig) A and 33% of total IgG. Similar decreases were seen in anti-HIV-1 gp120 IgG, anti-pneumococcal polysaccharide and antipoliovirus IgA. Although the latter was most affected, 66% of the unheated antigen-binding ability was still retained. In contrast, the binding capacity of IgA and IgG to influenza increased after heating . This flash-heating method has been shown to be feasible for use in neonatal intensive care units in developing countries  and in older infants where ARVs are not available .
As much as policy makers strive to have one policy for all, two different scenarios have evolved for infant feeding by mothers with HIV, divided mainly by the degree of country development.
Historically, mothers with HIV in the developed world were discouraged/prohibited from breastfeeding their infants and there are reports of mothers in the UK and US facing prosecution for such action. Following clinical evidence of reduced HIV transmission when a mother was receiving ARV therapy, the British HIV Association (BHIVA) and the Children’s HIV Association published a revised position paper in March 2011 . The default recommendation still remains as formula feeding but, additionally, it now allows after careful consideration a virally-suppressed woman on effective ARVs to choose to exclusively breastfeed for the first 6 months, provided that she is fully adherent to her ARV therapy and remains virally suppressed.
In 2013, the American Academy of Pediatrics (AAP) applied a similar rationale and reversed its previous stance against breastfeeding . The revised AAP guidelines recommend formula feeding by HIV-infected mothers but, similar to the UK, they encourage clinicians to support mothers if they express any interest in breastfeeding.
As discussed earlier, in developing countries, replacement feeding resulted in unacceptable rates of infant and young child morbidity, mortality, and malnutrition , . Even before ARV use in developing countries with high infant mortality rates, modelling showed that breastfeeding carried an improved infant HIV-free survival relative to formula feeding. It was therefore no surprise that with the introduction of ARV prophylaxis, breastfeeding together with ARV therapy resulted in a significantly greater infant HIV-free survival than replacement feeding.
The importance of a different default feeding option in developing countries is highlighted by the renewed emphasis on breastfeeding as a strategy to improve household food security and protection from infectious disease. Re-positioning of good breastfeeding practices is key to attaining the second goal of the United Nations  Sustainable Development Goals (i.e., to end hunger and improve nutrition and by 2020 to end all forms of malnutrition with special attention to stunting and wasting in children under 5 years of age).
For the few infants who escape the protection afforded by ARVs and are infected with HIV, there is considerable scientific evidence confirming that HIV-infected infants who receive breastfeeding for 2 years or more have a better health outcome than those who receive replacement feeding. As such, WHO recommends that HIV-infected infants are breastfed.
It is vital that HIV-infected infants are identified as early as possible. Point-of-care, same-day results for an HIV diagnosis in infants is becoming widely available, allowing HIV-infected infants to be identified soon after delivery. This is to be encouraged to ensure that mothers do not discontinue breastfeeding without knowing the HIV status of their infants.
Scientific evidence on the risk factors associated with HIV transmission during breastfeeding and the efficacy of ARV drugs given either to the mother or infant to reduce this risk, has given mothers the confidence to breastfeed. This is vital considering the strong maternal drive to feed their young, and to nurture the mother-child bond. Since lack of confidence is known to influence a mother’s ability and resolve to breastfeed, health care workers play an important role in continuing to supply mothers with up-to-date information to guide and encourage them. Support is needed to encourage mothers to exclusively breastfeed during the first 6 months and to continue for 24 months or longer depending on the setting.
A key strategy to promote breastfeeding is encouragement of the UNICEF Baby Friendly Hospital Initiative in all baby delivery facilities . It is also imperative to engage communities and household members as supporters of breastfeeding. Each community is unique, so policy makers/ health care workers need to work with communities to try to elucidate the constraints and facilitators for supporting breastfeeding mothers .
Policy makers also need to engage multi-sectors and advocate for laws to protect breastfeeding mothers, allowing them sufficient maternity leave to enable 6-months exclusive breastfeeding followed by a place and time at work to express milk to ensure a continued milk supply. Other important policies are those for HIV-infected women (and others) who have problems such as difficult deliveries, which prevent them from producing sufficient breast milk for their infants.
Breastmilk is especially important for low birth weight and/or preterm infants. Strategies need to be in place to ensure that vulnerable infants can be provided with donor breast milk, which imparts gut protection. There is evidence that low birth weight infants who receive formula milk instead of breast milk are at increased risk of necrotising enterocolitis, which increases their risk of HIV infection. Additionally, recent evidence of exclusive breastfeeding in older infants, as measured using the deuterium dilution method, suggests that higher percentages of breast milk intake are associated with lower levels of gut inflammation. Those infants whose intake was 100% breast milk had the lowest levels of inflammatory markers as measured in faecal samples .
Donor human milk banks supplying donor milk are quite common in Brazil, US, Canada, Europe, and Australia but their use in the developing world is severely limited. This is an important strategy for protecting infants, especially HIV-exposed infants during vulnerable periods. One developing country (South Africa) has made progress by introducing simple technology to enable human milk banks to be set up in neonatal intensive care units. Thus, vulnerable infants who cannot access their own mother’s milk are able to access donor milk in the early vulnerable period to afford gut protection , . Internationally, there has been renewed commitment to promote breastfeeding, particularly using human milk banks as a strategy for child survival. A framework for countries to set up human milk banks was recently developed at an International Milk Bank Technical Advisory Group meeting convened by PATH .
Additional to the importance of mothers maintaining optimum breastfeeding practice, equally important is that mothers adhere to their ARV regimens in spite of side effects. Many of these side effects may not be sufficiently serious to warrant a clinical decision to discontinue or change an ARV regimen. It is therefore vital that policy makers consider strategies for facilitating ARV adherence, especially among disadvantaged and disempowered mothers who may not have sufficient support to motivate drug adherence. Ways of using community structures, faith-based institutions and other “safe” places to distribute ARVs, and providing counselling need to be established. Policy makers in developing countries, e.g., in Africa, also need to consider the increasing burden of food insecurity and the problems experienced by mothers taking ARVs when hungry .
WHO, UNICEF, and countries continue to look for simple algorithms to improve access to ARVs. Seen to hold great promise, the introduction of Option B + has not been easy to implement because one drug is not always possible and because severe side effects from the combination pill may be attributed to one or more of its three drugs. Lack of adherence to ARVs, which exposes the infant to an increased risk of infection by transmission, also poses a problem. Health care workers may be under the misconception that all mothers issued ARVs are virally suppressed. It is therefore important for countries to consider monitoring viral loads rather than CD4 cell counts. In cases where mothers are not adherent or started drugs late in pregnancy, additional prophylactic cover is recommended for infants during the breastfeeding period. Instead of providing only 6 weeks of prophylaxis with NVP, infants should receive 12 weeks of dual prophylaxis (NVP and ZDV). Where mothers are not able or refuse to adhere to prophylaxis because of side effects, the option of providing infants with prophylaxis for 12 months or for the duration of breastfeeding should be considered. A recent study reported that 12 months prophylaxis with either Kaletra or 3TC to infants resulted in the very low transmission rate of 1.4– 1.5% at 12 months .
This welcome change in infant feeding guidelines will now result in many HIV-exposed infants receiving the benefits of breastfeeding for longer periods with a negligible risk of HIV infection. This change has now brought into focus another policy, which is in need of re-examination viz. the daily cotrimoxazole prophylaxis policy for HIV-exposed uninfected infants. This policy was initially implemented by WHO  almost 15 years ago based on evidence of its effectiveness in protecting HIV infected infants. It was believed that given the risk of HIV infection through breastfeeding, infants were at considerable risk of HIV infection and therefore would benefit from the known benefits in the case of their becoming infected. However, with the new infant feeding guidelines, the risk of HIV infection is negligible and the risks of daily antibiotic administration may likely outweigh the small purported benefits especially with prolonged breastfeeding . A recently published study from South Africa has confirmed that breastfeeding, HIV exposed, uninfected infants derive no health benefits from cotrimoxazole prophylaxis and the investigators have consequently called for a discontinuation of this policy.
Finally, the foundation to prevent transmission during breastfeeding must surely be primary prevention, as re-iterated in the first two of the four United Nations’ recommendations  for elimination of mother-to-child transmission of HIV (i.e., primary prevention of HIV infections in young women and prevention of unplanned pregnancies in HIV infected women).
In mothers that adhere to the recommended antiretroviral prophylaxis, the transmission of HIV through to 12 months of breastfeeding is negligible
In countries that have opted to promote and support breastfeeding with ART, HIV-infected mothers should be encouraged to be adherent to ART; practice exclusive breastfeeding for 6 months and then add complementary feeding while continuing to breastfeed until 24 months or beyond
Review of policy frameworks around HIV and breastfeeding needs to be done on a global basis in order to optimise infant feeding practices in regions with high HIV infection rate
[Editorial Note: Reference  is cited in section 19.3.1 to match printed version]