Infant Feeding for Mothers Living with HIV
I. Definition of the Prevention Area
HIV can be transmitted from a mother to her child during pregnancy or delivery. When a mother is HIV infected, her baby can also be infected through breastfeeding. Several interventions reduce the risk of transmission in the first two phases. Until recently, there was uncertainty regarding appropriate and healthy options for feeding infants of mothers living with HIV.
The 2010 World Health Organization (WHO) Guidelines on HIV and infant feeding have changed this landscape dramatically. The benefits of breastfeeding can now be achieved by mothers living with HIV, with a very low risk of transmitting HIV to the infant when breastfeeding is combined with maternal or infant antiretroviral (ARV) interventions. This avoids the risks associated with formula feeding and attains the ultimate goal of "HIV-free survival" for more infants.
II. Epidemiological Justification for the Prevention Area
Without interventions to prevent mother-to-child transmission, an estimated one-third of infants born to mothers living with HIV will become infected. Breastfeeding is responsible for 30 to 60 percent of the infections among children with HIV.
Mothers with high viral loads or low CD4 cell counts are most likely to transmit the virus during breastfeeding. Antiretroviral treatment before pregnancy or during early pregnancy and continuing postnatally can cut the risk of mother-to-child transmission by at least 75 percent. Other risk factors include mixed feeding (giving formula or other substances in addition to breast milk) in the first 6 months, various breast pathologies (such as mastitis, abscesses, or nipple lesions), and poor maternal nutrition.
III. Core Programmatic Components
The major programmatic options include promoting breastfeeding or formula feeding. Unfortunately, in regions with poor sanitation, unclean water used to mix formula has caused serious illness and deaths among formula-fed infants, who are more likely than breastfed infants to die from diarrhea, malnutrition, and serious infections. Alternatives such as heat treating expressed breast milk or providing breast milk from HIV-uninfected wet nurses are not practical in most resource-limited settings.
The 2010 WHO Guidelines advise national authorities to advocate a single infant feeding practice based on local conditions (socioeconomic and cultural contexts; availability and quality of health services; local epidemiology including HIV; the primary causes of infant and child mortality; and maternal and child undernutrition). Breastfeeding with antiretroviral (ARV) treatment/prophylaxis or avoidance of all breastfeeding are optimal choices. Most governments are expected to opt for breastfeeding with ARV treatment/prophylaxis. In resource-limited countries that have opted for breastfeeding as the primary option, exclusive breastfeeding for the first 6 months should be promoted even if ARV treatment or prophylaxis are not yet accessible; breastfeeding is not conditional on ARV treatment or prophylaxis.
Major messages in the 2010 WHO Guidelines for countries that opt for breastfeeding are:
- "Mothers known to be HIV-infected (and whose infants are HIV uninfected or of unknown HIV status)should exclusively breastfeed their infants for the first six months of life, introducing appropriate complementary foods thereafter, and continue breastfeeding for the first 12 months of life."
- "If infants and young children are known to be HIV-infected, mothers are strongly encouraged to exclusively breastfeed for the first six months of life and continue breastfeeding as per the recommendations for the general population, that is, up to two years or beyond."
WHO also promotes counseling and support for infant feeding and informed choice on the part of the mother. Support should not be contingent on whether a mother's choice aligns with national recommendations.
The absence of comprehensive and widely accessible postnatal care (including infant feeding) has been a major gap in PMTCT programs overall. These new WHO recommendations urgently need to be implemented in order to achieve the ultimate goal of "HIV-free survival" of infants.
IV. Current Status of Implementation Experience
Cultural barriers, financial and manpower constraints, and other obstacles made it difficult to implement maternal, newborn, and child health and PMTCT programs recommended by WHO in 2006. In many cases, counseling and support for women were inadequate. The stigma associated with HIV compromised the ability of women to breastfeed, and the cost and complexity of using formula created further barriers to implementation.
Successful implementation of the 2010 WHO Guidelines will require several steps: national guidelines and policies related to HIV and infant feeding will need to be changed; training curricula altered; new training provided for health workers; and efforts to secure community support reinforced.
Finally, efforts to widely promote breastfeeding for all mothers--which have been compromised in recent years due to fears about HIV transmission--will need to be reinvigorated.
This comprehensive review article summarizes key studies that examine infant feeding practices and antiretroviral treatment and prophylaxis to reduce mother-to-child transmission of HIV. The studies are grouped by category in three tables: 1) maternal treatment with highly active antiretroviral therapy (HAART) [Note: although the authors refer to HAART, the table includes women given highly active antiretroviral drugs solely as prophylaxis to prevent mother-to-child transmission]; 2) infant HIV prophylaxis; and 3) upcoming trials of maternal antiretroviral therapy or infant HIV prophylaxis. The authors conclude that well-designed studies show that infant prophylaxis with antiretroviral drugs during breastfeeding can reduce HIV transmission to as low as 1.2 percent at 6 weeks to 6 months. A combination of maternal HAART (or antiretroviral drugs for prophylaxis) and infant prophylaxis for the duration of breastfeeding, say the authors, could "effectively eliminate" mother-to-child transmission of HIV where formula is not acceptable.
Prevention of HIV transmission is not enough, according to the authors of this commentary, who say that measures to improve maternal health and child survival must also be addressed. The authors review key pitfalls of some prevention programs, from earlier attempts to encourage all women living with HIV to abstain from breastfeeding (which led to a sixfold increase in child mortality in Uganda), to the tendency to conduct prevention activities such as voluntary testing and counseling in urban areas. Because some prophylactic antiretroviral regimens are administered during, prior to, or just after a child's birth, many children of mothers with HIV are left unprotected during breastfeeding--a period when 40 percent of children with HIV will acquire the virus.
This clinical trial conducted in Malawi evaluated HIV-free survival rates of children whose mothers were categorized into one of three groups: group A, highly active antiretroviral therapy (HAART)-ineligible (defined as a CD4 count of 250 or greater); group B, HAART-eligible but untreated women; and group C, HAART-eligible and treated women. At 24 months, HIV-free survival of infants was 82 percent in group A; 68 percent in group B; and 81 percent in group C. When survival rates were adjusted for various infant prophylaxis regimens, the risk of death among the children of treated mothers and treatment-ineligible mothers were each approximately half that of the children of treatment-eligible but untreated women.
This study, conducted in Burkina Faso and Kenya, randomized 824 pregnant women living with HIV and with CD4 counts of 200 to 500 to receive either triple antiretroviral (ARV) prophylaxis or short-ARV prophylaxis. Over three-quarters of infants in both groups were breastfed. Infants in both groups were treated with single-dose nevirapine. At 12 months, the cumulative HIV infection rates among infants born to women receiving triple-ARV and short-ARV were 5.6 percent and 9.3 percent, respectively; however, there was no statistically significant difference in overall mortality.
This study randomized 2,637 mother-infant pairs, in which the mothers had CD4 counts higher than 250, to receive either maternal triple drug prophylaxis (referred to by the authors as highly active antiretroviral therapy), infant nevirapine, or no additional prophylaxis (all pairs were treated with single-dose nevirapine and one week of two-drug prophylaxis immediately after birth). Treatment extended up to 28 weeks of breastfeeding, and mothers breastfed exclusively for 24 weeks, followed by rapid weaning. Preliminary data show that at one week, 4.9 percent of infants in the two treatment groups were infected, reflecting transmission during pregnancy. At 28 weeks, 6.4 percent of infants in the control arm (those with no additional treatment) developed HIV; 3.0 percent of infants born to mothers treated with triple drug prophylaxis became HIV-positive; and 1.8 percent of infants treated with nevirapine became infected. Revised results will be published in an upcoming issue of the New England Journal of Medicine.
Five hundred and sixty women living with HIV with CD4 counts of 200 or higher were randomized to receive one of two triple drug regimens (referred to by the authors as highly active antiretroviral therapy): Arm A received abacavir/zidovudine/lamivudine and Arm B received lopinavir/ritonavir/Combivir, which were prescribed at 26 to 34 weeks of pregnancy through 6 months following birth. A control group of 170 women with CD4 counts less than 200 were treated with nevirapine/Combivir. Seven of 10 mothers breastfed for five months or longer. Neither infant mortality nor HIV transmission differed significantly among the groups. Infant mortality at six months did not differ by groups and was very low in all arms.
This systematic review of clinical trials of prevention of mother-to-child transmission of HIV during breastfeeding found that up to 42 percent of transmission occurs during breastfeeding. Mothers who had a high viral load; were young; did not breastfeed exclusively; or who had breast lesions were more likely to transmit HIV to their infants. In low-resource areas with poor sanitation, infants were more likely to die of diarrheal diseases and pneumonia when not exclusively breastfed--completely offsetting deaths from HIV at 2 years. The authors suggest that for mothers who initiate breastfeeding, a) breastfeeding should be exclusive, and b) extended prophylaxis should be given to the infant (nevirapine alone or nevirapine with zidovudine).
An outbreak of diarrhea led to an abrupt increase in acute malnutrition and deaths among children younger than 5 years old in Botswana. The outbreak occurred during a period of heavy rains when sewage may have flooded the water supply. The study authors examined risk factors for diarrhea, including the HIV status of mother and child; contamination of the water supply; presence of flush toilets; and breastfeeding. Approximately one-third of mothers were living with HIV and, as a result of local prevention counseling against breastfeeding, almost none of these mothers breastfed, whereas almost half of uninfected mothers did breastfeed. HIV status of the infant and failure to breastfeed were correlated with the risk of diarrhea.
In an attempt to balance the benefits of breastfeeding with reduced risk of HIV transmission by reducing exposure time, researchers randomized mother-infant pairs to early, abrupt cessation of breastfeeding at four months or to a control group that breastfed as long as desired. Children who were infected with HIV at four months were somewhat less likely to survive to 24 months if breastfeeding was stopped early. However, there was no statistically significant difference in overall child survival at 24 months in the two groups: approximately one-quarter of the children in the intervention arm and the control arm had died by 24 months. The authors conclude that the failure of early breastfeeding cessation and its associated costs did not justify such an approach.
Mothers living with HIV who chose to breastfeed were compared to those who chose to formula feed their infants. Eligible mothers received antiretroviral treatment and those who were not eligible received antiretroviral drugs to prevent mother-to-child transmission of HIV. Only one-quarter of mothers practiced exclusive breastfeeding at one month. Although formula-fed infants were less likely to become HIV-infected at one month, they were six times as likely to die by 12 months compared to breast-fed infants. Formula feeding was particularly risky for HIV-positive infants; all 3 formula-fed HIV-positive infants died by 12 months, compared to none of the 12 breast-fed HIV-positive infants.
Formula prepared by mothers in Durban, South Africa, had high levels of fecal contamination. The mothers, who participated in a program to prevent mother-to-child transmission of HIV, all said that they boiled the water before preparing formula. Some of the mothers stored the boiled water for later use, however, which significantly increased the risk of contamination. Other factors associated with contamination included storing leftover formula for later feeds and the use of unboiled water to rinse previously sterilized bottles. Although mothers correctly diluted powdered formula for younger infants, they tended to overdilute formula for older children. Some mothers put extra scoops of powder in the feed, mistakenly thinking the baby "needs it," a practice that puts the child at risk of potentially fatal problems from too much salt and sugar in the bloodstream.
This prospective, observational study examined the hypothesis that exclusive breastfeeding would reduce mother-to-child HIV transmission. Some study mothers received either single-dose nevirapine for prophylaxis or antiretroviral treatment. Infants who were exclusively breastfed were at least half as likely to acquire HIV at four months of age as infants of mothers who practiced non-exclusive breastfeeding. Eighty-four percent of the 958 mothers in the study practiced exclusive breastfeeding at four months. The researchers review factors hypothesized to increase HIV transmission during mixed feeding, and add a new hypothesis of their own.
Mothers in this study were encouraged to either exclusively breastfeed or to exclusively use replacement feeds if they chose not to breastfeed. Infants who received mixed feedings were, by three months, nearly 11 times as likely to become HIV infected as infants receiving exclusive breastfeeding. Infants who received replacement feeds for three months were more than twice as likely to die as infants who were exclusively breastfed. The authors state that 210,000 to 270,000 new HIV infections acquired by breastfeeding annually could be averted if mothers are encouraged to stop exclusive breastfeeding at six months (around 4 percent transmission) rather than the "more usual 18-24 months of mixed feeding" (16 percent transmission).
Mothers interviewed about their infant-feeding practices were selected from a cohort that had achieved some success in either exclusive breastfeeding or exclusive formula feeding. Despite the relative success of this cohort, 80 percent of breastfeeding mothers introduced other liquids to the infant's feeds within the first month. Factors that led to mixed feedings included conflicting health care messages (posters promoting breastfeeding and formula feeding were found in the same clinic); pressure from mothers, mothers-in-law, and grandmothers to add other liquids; and secrecy regarding the diagnosis of HIV, which led many women to reject recommended feeding practices out of fear that it would make their diagnosis obvious if they chose to exclusively formula feed or to stop breastfeeding early.
This small study examines transmission of HIV from mother to child during the postnatal (after birth and during breastfeeding) period. Breastfed infants who received zidovudine prophylaxis for six months were compared to formula-fed infants given one month of infant zidovudine. Three-quarters of mothers practiced mixed feeding. HIV infection at seven months was somewhat higher (9 percent) in the breastfed group than in the formula-fed group (5.6 percent); however, mortality at the same time was higher in the formula-fed group (9.3 percent) compared to the breastfed group (4.9 percent). HIV-free survival was similar between the two groups by 18 months.
This large study assessed the risk of mother-to-child transmission (MTCT) of HIV among mothers who practiced exclusive breastfeeding, predominant breastfeeding, or mixed feeding. Mixed breastfeeding was associated with the highest risk and exclusive breastfeeding with the lowest risk of MTCT. Mixed breastfeeding infants were 2.6 times as likely to become HIV positive at 18 months, and predominantly breastfed infants were 1.6 times as likely as exclusively breastfed infants to become HIV positive. Only 7.6 percent of infants were exclusively breastfed for at least three months. By six months, 93 percent of infants received early mixed feedings. More than two-thirds of the breastfeeding transmission occurred after six months. Maternal CD4 <200 was significantly associated with postnatal transmission.
The authors previously reported that the risk of mother-to-child transmission of HIV from breastfeeding is 16 percent. Among infants with HIV, up to 44 percent acquired the infection from breastfeeding. In this study, mothers living with HIV were randomized to either breastfeed or formula-feed their infants. Infant mortality at two years was similar in the two groups (approximately 1 of every 4 or 5 infants died by their second birthday). However, as HIV-free survival was higher in the formula-fed group, the authors concluded that "formula feeding can be a safe alternative to breastfeeding." The authors acknowledge that their study represents a "best-case scenario" as the mothers all had access to potable water, a reliable source of formula, and ready access to medical care.
This chapter describes the evidence on three strategies of infant feeding (exclusive breastfeeding, early weaning, and treatment of expressed breast milk) to reduce the likelihood of HIV transmission from an HIV-positive mother to her infant. The evidence for the reduced transmission of HIV to infants when mothers exclusively breastfeed is strong. The theories given on why mixed feeding may increase HIV transmission include: 1) mixed feeding damages an infant's intestinal mucosa which then facilitates HIV infection, and 2) mixed feeding may cause the mother's breasts to have subclinical mastitis that increase the breast milk's viral loads. However, further studies on these two theories have disproven them, and research is still necessary to further investigate them. Early weaning was thought to be a strategy in reducing HIV transmission to infants since it reduced their exposure to the virus; however, it was found that early weaning could cause increased morbidity and mortality to HIV-positive infants and had no effect among HIV-negative infants. The World Health Organization (WHO) no longer recommends this strategy. Treating breast milk with microbicides to inactivate HIV-1 was successfully reported in one study but subsequent research could not replicate the findings. Another study attempted to inactivate HIV by leaving breast milk at room temperature for six hours, but this was found to be ineffective as well. Direct and indirect heat treatment of breast milk are two prevention of mother-to-child transmission strategies. However, direct heat treatment often leads to loss of important nutrients in the milk. There are two types of indirect heat treatments--"Pretoria pasteurization" and "flash heat"--and both had positive results, but more research is needed to conclusively recommend these methods. WHO updated its guidelines on HIV-1 and infant feeding in 2010 based on the current evidence.
This article describes a follow-up study of the Breastfeeding, Antiretrovirals, and Nutrition (BAN) clinical trial, which assessed the effectiveness of 28 weeks of use of maternal or infant antiretroviral (ARV) prophylaxis on postnatal HIV infection at 48 weeks. The study also examined the effects on breastfeeding cessation and on maternal and infant deaths, as well as serious adverse events. The original BAN randomized clinical trial provided mothers and infants with antiretroviral therapy (maternal triple ARVs, infant nevirapine, or ARVs for both before and during labor but not after the birth) to prevent postnatal HIV-1 transmission. According to the authors, the 48-week follow-up study has demonstrated that both infant and maternal prophylaxis effectively reduces postnatal HIV-1 transmission, and that the protective benefits continue after breastfeeding ends. They report that once mothers wean their infants from breastfeeding, there is an increase in HIV transmission, infant morbidity, and mortality. They particularly point out that weaning at six months or less may increase infant morbidity. Due to these results, the authors recommend breastfeeding with prophylaxis for more than 28 weeks. They also remind readers that the World Health Organization now recommends that HIV-positive mothers or uninfected infants receive antiretroviral prophylaxis for 12 months of breastfeeding.
These World Health Organization guidelines were updated in 2010 in response to significant new programmatic experience and evidence regarding HIV and infant feeding. The 2010 recommendations recognize the important impact of ARVs during the breastfeeding period, and recommend that national authorities in each country decide which infant feeding practice, i.e. breastfeeding with an ARV intervention to reduce transmission or avoidance of all breastfeeding, should be promoted and supported by their Maternal and Child Health services. The document includes the principles, recommendations, and summary of the evidence. Nine key principles focus on overall HIV-free survival; integrated versus vertical infant feeding interventions; setting national strategies for infant feeding; information and counseling for mothers; and promoting breastfeeding among the general population. Seven recommendations are provided on breastfeeding and young child feeding practices.
This report summarizes new findings, conclusions, and recommendations from the World Health Organization HIV and Infant Feeding Technical Consultation held in October 2006. Newer evidence has reversed many of these conclusions.
The authors performed a rapid assessment of HIV and infant feeding practices in Malawi, Kenya, and Zambia, and found serious problems, such as "lack of space and time for proper counseling"; lack of support for nutrition as a preventive measure; few linkages with community-based infant feeding programs; weak monitoring and evaluation; and lack of resources devoted to this activity. They found a "strong bias against breastfeeding" in all three countries, and they cite a United Nations assessment that found that despite training in PMTCT, Kenyan health care workers had "uniformly poor" knowledge of PMTCT. The authors emphasize the need to focus on overall child survival, not just reduction of the risk of HIV transmission.
HIV-negative and HIV-positive women in KwaZulu-Natal, South Africa, who received frequent counseling and support to practice exclusive breastfeeding (EBF) were twice as likely to practice EBF as those mothers who did not adhere to regular counseling visits. However, large numbers of mothers did not adhere to the full series of visits. Adherence to EBF at six months was 45 percent of HIV-negative women and 40 percent of women living with HIV. The authors conclude that EBF can be successfully promoted through the provision of a home support program with well-trained lay counselors.
HIV-exposed infants in Kenya experienced high rates of diarrhea during weaning, according to the Kisumu Breastfeeding Study. Researchers tried to reverse this situation by providing safe water storage vessels, hygiene education, and bleach for household water treatment. The interventions successfully reduced diarrhea before and after weaning, but not during the weaning period itself. Provision of safe water was not found to be protective against weaning-associated diarrhea among infants who were weaned early.
The study explored the feasibility of whether HIV-positive mothers who were breastfeeding could flash-heat (FH) their breast milk for their child to reduce the transmission of HIV. Mothers were visited weekly by a health worker and also had monthly clinic visits. Mothers who were eligible or currently taking antiretrovirals were excluded from the study to better determine the intervention's effectiveness on the infants. Baseline information on the mothers' demographics, and infant feeding, growth, and morbidity were collected. Mothers also kept daily journals on infant feeding habits, and health workers measured the peak milk temperatures and collected pre-heated and post-heated milk samples for bacterial cultures. Seventy-two mothers were eligible for the intervention since their infants were HIV-negative, and 51.4 percent (37 mothers) chose to FH their milk once their child turned 6 months. The median duration of FH was 9.6 weeks (range: 1 day-15.6 weeks). Women who expressed more often also expressed a larger amount of breast milk. It was found that mothers were successful in cleaning the utensils, heating the milk, and feeding their infants with the FH milk. About 42 percent of milk samples were found to be contaminated with bacterial growth. Mothers reported little stigma associated with expressing and FH but did report hiding their behaviors about 77 percent of the time, mostly from neighbors and infrequently from friends and family. The study was successful in demonstrating that HIV-positive mothers were both interested and able to express and FH their breast milk to decrease the likelihood that their infants would become infected with HIV.
This generic training package for a five-day prevention of mother-to-child transmission (PMTCT) course was updated in 2008 and serves as a comprehensive, evidence-based tool that can be updated and implemented in country or regional resource-constrained settings. It consists of nine modules, each of which contains a participant manual, slides, and trainer manual. It is targeted to staff such as nurses, midwives, physicians, social workers, outreach workers, counselors, and program managers working in PMTCT programs or health care settings that provide PMTCT services. Guidance is given for closing the course, offering an optional field visit, and answering frequently asked questions. A glossary is also available. The document may be freely reproduced.
These tools were created to help health workers better support HIV-infected mothers. The tools include counseling cards, a reference guide, and an orientation guide for trainers. The counseling cards and reference guide are also available in French and Spanish. These tools do not reflect the November 2009 WHO guidelines.
Rapid Advice: Use of Antiretroviral Drugs for Treating Pregnant Women and Preventing HIV Infection in Infants.
These recently-updated, evidence-based recommendations on PMTCT serve as a reference for countries to adopt and adapt according to conditions found at the national level (resource availability and limitations, etc.). They are intended as an aid to simplify and standardize previous recommendations by making guidelines for initiating antiretroviral treatment (ART) for pregnant women the same as for non-pregnant women and by taking into account programmatic considerations that affect implementation. They provide guidance for policy makers and program managers responsible for PMTCT programs and are a resource for health workers involved in prevention, care, and treatment of pregnant women and their infants. These recommendations focus on two key areas: a) when to start ART and which regimen to use for pregnant women who are treatment-eligible (and their infants); and b) when to start and which antiretroviral prophylaxis regimen to give non treatment-eligible women (and their infants). Recommendations are given for breastfeeding and non-breastfeeding infants. These recommendations form part of a larger guideline which is expected to be published and disseminated in March-April 2010.
View Report (PDF, 428 KB)
View Report Summary (PDF, 128 KB)
State of the World's Children
The tables found in the State of the World's Children report contain data for all countries as well as regional summaries on breastfeeding patterns.
View Report (PDF, 420 KB)