HIV and Infant Feeding: Knowledge, Gaps and Challenges for the Future

1. 1 HIV and Infant Feeding: Knowledge, Gaps and Challenges for the Future Jay Ross & Ellen G. Piwoz Academy for Educational Development This presentation was given by Jay Ross (Policy Adviser, the LINKAGES Project, Academy for Educational Development) at the Colloquium for the Asia Pacific Region on Infant feeding and HIV in New Delhi on November 28, 2003. It is adapted from one originally given by Ellen Piwoz at the WABA/UNICEF Colloquium on HIV and infant feeding, held in Arusha,Tanzania on September 19, 2002. This presentation was given by Jay Ross (Policy Adviser, the LINKAGES Project, Academy for Educational Development) at the Colloquium for the Asia Pacific Region on Infant feeding and HIV in New Delhi on November 28, 2003. It is adapted from one originally given by Ellen Piwoz at the WABA/UNICEF Colloquium on HIV and infant feeding, held in Arusha,Tanzania on September 19, 2002.

2. 2 Outline of the Presentation Overview Review of evidence on risk factors HIV and infant feeding risk analysis Research planned and underway Challenges for the future The presentation covers these topics: 1) Introduction: Basics facts about MTCT of HIV and timing of HIV transmission to set the context, 2) An overview of the risk factors for postnatal or BF-related HIV transmission. Because this presentation is billed as a �technical update,� I will emphasize new information. 3) The results of our efforts to quantify the risks of different infant feeding strategies using a simulation model 4) A brief description of the many current and planned field trials of interventions to prevent mother-to-child transmission through breastfeeding; and 5) Recommendations and a discussion of challenges for the futureThe presentation covers these topics: 1) Introduction: Basics facts about MTCT of HIV and timing of HIV transmission to set the context, 2) An overview of the risk factors for postnatal or BF-related HIV transmission. Because this presentation is billed as a �technical update,� I will emphasize new information. 3) The results of our efforts to quantify the risks of different infant feeding strategies using a simulation model 4) A brief description of the many current and planned field trials of interventions to prevent mother-to-child transmission through breastfeeding; and 5) Recommendations and a discussion of challenges for the future

3. 3 Introduction: The Context First an overview of the context� First an overview of the context�

4. 4 Timing of Mother-to-Child Transmission HIV transmission from mothers to infants occurs during pregnancy, at the time of labor and delivery, and postnatally through breastfeeding. This slide summarizes the timing of HIV transmission during these periods. In the absence of any interventions to prevent or reduce transmission, about 5-10 percent of HIV infected mothers pass the virus to their infants during pregnancy; between 10-20 percent during labor and delivery; and another 10-20 percent postnatally through breastfeeding to 24 months. One reason for the lack of precision here is that there is no reliable method of determining whether an infant is infected or not until about 6 weeks of age so the precise timing cannot be determined. Labor and delivery is the single time point of greatest risk with as much infection occurring within 24 hours as occurs postnatally within 24 months of breastfeeding. Most ARV prophylaxis regimens aim to reduce HIV transmission during this time. HIV transmission from mothers to infants occurs during pregnancy, at the time of labor and delivery, and postnatally through breastfeeding. This slide summarizes the timing of HIV transmission during these periods. In the absence of any interventions to prevent or reduce transmission, about 5-10 percent of HIV infected mothers pass the virus to their infants during pregnancy; between 10-20 percent during labor and delivery; and another 10-20 percent postnatally through breastfeeding to 24 months. One reason for the lack of precision here is that there is no reliable method of determining whether an infant is infected or not until about 6 weeks of age so the precise timing cannot be determined. Labor and delivery is the single time point of greatest risk with as much infection occurring within 24 hours as occurs postnatally within 24 months of breastfeeding. Most ARV prophylaxis regimens aim to reduce HIV transmission during this time.

5. 5 MTCT in 100 HIV+ Mothers by Timing of Transmission This slide illustrates this data on HIV transmission in a slightly different way. Imagine 100 HIV+ women. Using the midpoints of the ranges described in the previous slide, you would expect 7 of their infants to be infected with HIV during pregnancy, another 15 during labor and delivery; and another 15 over the course of about 2 years of breastfeeding. 63 infants would not become infected with HIV, even if breastfed and without any intervention in place to prevent transmission. This slide illustrates this data on HIV transmission in a slightly different way. Imagine 100 HIV+ women. Using the midpoints of the ranges described in the previous slide, you would expect 7 of their infants to be infected with HIV during pregnancy, another 15 during labor and delivery; and another 15 over the course of about 2 years of breastfeeding. 63 infants would not become infected with HIV, even if breastfed and without any intervention in place to prevent transmission.

6. Major Causes of Death among Children around the World When considering the issue of HIV and infant feeding it is important to also bear in mind that undernutrition, a consequence of suboptimal infant feeding, is an underlying cause of approximately 60% of deaths in children less than 5 years of age worldwide. Sources: For cause-specific mortality: EIP/WHO. For deaths associated with malnutrition: Caulfield LE, Black RE. Malnutrition and the global burden of disease: underweight and cause-specific mortality. When considering the issue of HIV and infant feeding it is important to also bear in mind that undernutrition, a consequence of suboptimal infant feeding, is an underlying cause of approximately 60% of deaths in children less than 5 years of age worldwide. Sources: For cause-specific mortality: EIP/WHO. For deaths associated with malnutrition: Caulfield LE, Black RE. Malnutrition and the global burden of disease: underweight and cause-specific mortality.

7. 7 Proportion of all < 5 yrs deaths that could be prevented with infant feeding interventions Other recently published estimates indicate that optimal breastfeeding would avert about 13% of the 10 million deaths that occur each year among children under 5. In contrast, the same source suggests that if all HIV infected mothers use perinatal nevirapine and avoid breastfeeding, about 2% of under five deaths would be averted. Note that this estimate does not account for the increased mortality that would occur due to artificial feeding among these children. SO although breastfeeding is a risk factor for transmission of HIV, it is, at the same time, one of the most important and effective child survival interventions available. The challenge is how to balance the risks and benefits when formulating public health policies related to HIV and infant feeding -- a dilemma that is being confronted at every level: by public health policy makers, health care managers, health workers and mothers themselves. Other recently published estimates indicate that optimal breastfeeding would avert about 13% of the 10 million deaths that occur each year among children under 5. In contrast, the same source suggests that if all HIV infected mothers use perinatal nevirapine and avoid breastfeeding, about 2% of under five deaths would be averted. Note that this estimate does not account for the increased mortality that would occur due to artificial feeding among these children. SO although breastfeeding is a risk factor for transmission of HIV, it is, at the same time, one of the most important and effective child survival interventions available. The challenge is how to balance the risks and benefits when formulating public health policies related to HIV and infant feeding -- a dilemma that is being confronted at every level: by public health policy makers, health care managers, health workers and mothers themselves.

8. 8 Overview of HIV Transmission during Breastfeeding For the remainder of the presentation, we will focus on postnatal HIV transmission during breastfeeding.For the remainder of the presentation, we will focus on postnatal HIV transmission during breastfeeding.

9. 9 Risk Factors For Postnatal Transmission Mother Immune/health status Plasma viral load Breast milk virus Breast inflammation (mastitis, abscess, nipple lesions) New HIV infection Viral Characteristics Infant Breastfeeding duration Non-exclusive BF Age (first months) Lesions in mouth, intestine Prematurity Infant immune response This slide summarizes the known risk factors for postnatal HIV transmission. The slide is divided into risk factors for mothers and those for infants. Most (though not all) of these risk factors will be discussed in this presentation. Throughout the talk I/we will highlight new information. This slide summarizes the known risk factors for postnatal HIV transmission. The slide is divided into risk factors for mothers and those for infants. Most (though not all) of these risk factors will be discussed in this presentation. Throughout the talk I/we will highlight new information.

10. 10 How does HIV transmission during breastfeeding occur? -1 Exact mechanisms unknown HIV virus in blood passes to breast milk cell-associated, cell-free virus observed virus shed intermittently (undetectable ~ 25-35%) levels vary between breasts in samples taken at same time (Willumsen et al, 2001) Virus may also come directly from infected cells in mammary gland produced locally in mammary macrophages, lymphocytes, epithelial cells (Becquart et al, 2002) First, how does HIV transmission during breastfeeding occur? The answer is that the exact mechanisms are not yet known. HIV appears to pass from the mother�s circulation into her breastmilk. Cell-free and cell-associated virus have been found in milk samples. The virus appears to be shed intermittently over time. In studies with a limit of detection of 200 viral copies/ml (somewhat high), about 25-35 percent of samples have no detectable virus present (it is either not there or virus is present in levels below the limit of detection). Studies also show that the amount of virus in breast milk varies between breasts in samples taken at the same time. A recent study suggests that infected cells may also enter the milk from the mammary gland. First, how does HIV transmission during breastfeeding occur? The answer is that the exact mechanisms are not yet known. HIV appears to pass from the mother�s circulation into her breastmilk. Cell-free and cell-associated virus have been found in milk samples. The virus appears to be shed intermittently over time. In studies with a limit of detection of 200 viral copies/ml (somewhat high), about 25-35 percent of samples have no detectable virus present (it is either not there or virus is present in levels below the limit of detection). Studies also show that the amount of virus in breast milk varies between breasts in samples taken at the same time. A recent study suggests that infected cells may also enter the milk from the mammary gland.

11. 11 How does HIV transmission during breastfeeding occur? -2 Infant consumes HIV HIV enters/infects infant through permeable mucosal surfaces, lymphoid tissues, and/or lesions in mouth, intestines Although BF infant may consume >500,000 virons, >25,000 infected cells per day, majority do NOT become HIV infected (Lewis et al, 2001) Why? Immune factors in BM, saliva play a role (Miller et al, 2002; Sabbaj et al, 2002; Farquhar et al, 2002; Van der Perre et al, 1999; 1993; 1988) Infants consume HIV in breast milk. The virus enters or infects the baby through permeable mucosa, lymphoid tissues, or through lesions in the gastro-intestinal tract (mouth to intestine). The most remarkable aspect of HIV transmission during breastfeeding is that although an infant exposed to HIV may consume a half million virons and 25,000 infected cells per day, the majority do not become infected. Immune factors in breast milk and infant saliva are believed to play a role in preventing transmission. Infants consume HIV in breast milk. The virus enters or infects the baby through permeable mucosa, lymphoid tissues, or through lesions in the gastro-intestinal tract (mouth to intestine). The most remarkable aspect of HIV transmission during breastfeeding is that although an infant exposed to HIV may consume a half million virons and 25,000 infected cells per day, the majority do not become infected. Immune factors in breast milk and infant saliva are believed to play a role in preventing transmission.

12. 12 Risk factors for postnatal transmission: Maternal immune status Maternal immune status is one of the most important predictors of HIV transmission risk at all time points. CD4 T-cells are a class of lymphocytes, a part of the immune system that is infected and destroyed by the virus so a low CD4 count is both an indicator and a functional consequence of disease progression. Low CD4 T-lymphocyte counts have been associated with a greater risk of postnatal HIV transmission in studies in Malawi, Kenya, South Africa, Uganda, and, as shown in this slide, in West Africa (Leroy et al 2003). This study is a pooled analysis of data from 641 HIV+ mothers in Abidjan (Cote d�Ivoire) and Bobo Dioulasso (Burkina Faso), who participated in an ARV trial. Breastfeeding initiation was 98 percent and median duration of breastfeeding was > 13 months in both sites. This slide shows the cumulative HIV transmission from 6 weeks to 24 months by maternal CD4 at baseline (during pregnancy) for the infants in the treatment group (findings are similar for the placebo group but these data are not shown). The transmission due to breastfeeding was 11 times higher in women with CD4 counts < 500, than in women with CD4 counts > 500. These data underscore the importance of maternal health for preventing HIV transmission during breastfeeding. Keeping mothers healthy is critical for infant health and survival. Maternal immune status is one of the most important predictors of HIV transmission risk at all time points. CD4 T-cells are a class of lymphocytes, a part of the immune system that is infected and destroyed by the virus so a low CD4 count is both an indicator and a functional consequence of disease progression. Low CD4 T-lymphocyte counts have been associated with a greater risk of postnatal HIV transmission in studies in Malawi, Kenya, South Africa, Uganda, and, as shown in this slide, in West Africa (Leroy et al 2003). This study is a pooled analysis of data from 641 HIV+ mothers in Abidjan (Cote d�Ivoire) and Bobo Dioulasso (Burkina Faso), who participated in an ARV trial. Breastfeeding initiation was 98 percent and median duration of breastfeeding was > 13 months in both sites. This slide shows the cumulative HIV transmission from 6 weeks to 24 months by maternal CD4 at baseline (during pregnancy) for the infants in the treatment group (findings are similar for the placebo group but these data are not shown). The transmission due to breastfeeding was 11 times higher in women with CD4 counts < 500, than in women with CD4 counts > 500. These data underscore the importance of maternal health for preventing HIV transmission during breastfeeding. Keeping mothers healthy is critical for infant health and survival.

13. 13 Risk factors for postnatal transmission: Maternal immune status Again looking at CD4 counts, this slide shows data from meta-analysis of data from 9 sites in Africa (as reported at the 10th Conference on Retroviruses and Opportunistic Infections in Boston in February 2003). After adjusting for study site and infant sex, mothers with baseline (in pregnancy) CD4 counts < 200 were 8 times more likely to infect their infants through breastfeeding compared with women with baseline CD4 levels of 500 or greater. Women with intermediate CD4 counts (between 200-499) had an intermediate risk of transmission (3.7). Again looking at CD4 counts, this slide shows data from meta-analysis of data from 9 sites in Africa (as reported at the 10th Conference on Retroviruses and Opportunistic Infections in Boston in February 2003). After adjusting for study site and infant sex, mothers with baseline (in pregnancy) CD4 counts < 200 were 8 times more likely to infect their infants through breastfeeding compared with women with baseline CD4 levels of 500 or greater. Women with intermediate CD4 counts (between 200-499) had an intermediate risk of transmission (3.7).

14. 14 Risk factor: Maternal viral load Viral load is an important predictor of intra-partum MTCT (Leroy et al, 2001; Semba et al, 1999) Plasma viral load is also a risk factor during breastfeeding Mothers newly infected during lactation ? temporary �pulse� in circulating virus ? 29% transmission risk (Dunn et al, 1992) Plasma viral load associated with increased risk of PN transmission in studies in: Kenya (John et al, 2001; Richardson et al, 2003) Tanzania (Fawzi et al, 2002) West Africa (Leroy et al, 2003) The level of virus in the mother�s blood is also an important predictor of infant HIV infection at birth and at 6 weeks of age, as observed in several studies. Viral load in maternal blood is high in the first weeks after new infection until the body begins to manufacture antibodies that suppress the virus, reducing it to low levels until the immune system begins to break down as the disease progresses, usually years later. The initial peak in viremia following infection is the probable explanation for the increased risk of transmission among infants of newly infected mothers, as demonstrated in an early meta-analysis by Dunn et al (1992). Increased risk of postnatal transmission associated with maternal viral load is also suggested from an analysis of data from the randomized trial of breastfeeding and formula feeding in Nairobi. This study found that maternal plasma viral RNA levels higher than the median (>43k copies/ml) were associated with a 2.6-fold (95% CI 1-6.4) greater risk of HIV infection after two months (John et al, 2001). Likewise, in Tanzania, mothers with viral loads >=50K copies/mL were 5 times more likely to pass on HIV during breastfeeding than women with lower levels of virus in their blood. In West Africa, transmission risk increased by a factor of 2.65 for each 10-fold (1 log unit) increase in viral load, even after adjusting for CD4 count and age of the infant. [Viral load is usually expressed in log10 units.] The level of virus in the mother�s blood is also an important predictor of infant HIV infection at birth and at 6 weeks of age, as observed in several studies. Viral load in maternal blood is high in the first weeks after new infection until the body begins to manufacture antibodies that suppress the virus, reducing it to low levels until the immune system begins to break down as the disease progresses, usually years later. The initial peak in viremia following infection is the probable explanation for the increased risk of transmission among infants of newly infected mothers, as demonstrated in an early meta-analysis by Dunn et al (1992). Increased risk of postnatal transmission associated with maternal viral load is also suggested from an analysis of data from the randomized trial of breastfeeding and formula feeding in Nairobi. This study found that maternal plasma viral RNA levels higher than the median (>43k copies/ml) were associated with a 2.6-fold (95% CI 1-6.4) greater risk of HIV infection after two months (John et al, 2001). Likewise, in Tanzania, mothers with viral loads >=50K copies/mL were 5 times more likely to pass on HIV during breastfeeding than women with lower levels of virus in their blood. In West Africa, transmission risk increased by a factor of 2.65 for each 10-fold (1 log unit) increase in viral load, even after adjusting for CD4 count and age of the infant. [Viral load is usually expressed in log10 units.]

15. 15 Risk Factor: Maternal Viral Load In the Nairobi formula feeding trial, breastfeeding transmission was 4 times greater among mothers with viral loads greater than the median (43,120 copies/mL) than among mothers with lower viral loads. In the Nairobi formula feeding trial, breastfeeding transmission was 4 times greater among mothers with viral loads greater than the median (43,120 copies/mL) than among mothers with lower viral loads.

16. 16 Risk factor: Maternal viral load These hazard ratios for log10 plasma viral load and CD4 count < 500/mm3 are from a multivariate statistical analysis that includes both, so these effects are independent. Note that the treatment effect of perinatal ZDV was not significant (maternal perinatal therapy with ZDV at delivery and for one week after delivery was not effective in reducing transmission after 1 month). These hazard ratios for log10 plasma viral load and CD4 count < 500/mm3 are from a multivariate statistical analysis that includes both, so these effects are independent. Note that the treatment effect of perinatal ZDV was not significant (maternal perinatal therapy with ZDV at delivery and for one week after delivery was not effective in reducing transmission after 1 month).

17. 17 Risk Factor: Breast Pathology Prevalence of breast pathologies in HIV+ women in Africa Mastitis (clinical or sub-clinical): Clinical exam: 7-11% (Embree, 2000; John et al, 2001) Na+/K > 1.0: 11-12% at 6, 14 wk (Willumsen et al, 2000) Na+ > 12 mmol/L: 16.4% at 6 wk (Semba et al, 1999) Nipple lesions: Clinical exam: 11-13% (Embree, 2000; John et al, 2001) Clinical exam: 10% (Ekpini et al, 1997) Hospitalized infants: 11% (Kambarami et al, 1997) Breast abscesses: Clinical exam: 12% (John et al, 2001) Clinical exam: 3% (Ekpini et al, 1997) Breast inflammation is another important risk factor for HIV transmission during breastfeeding. According to available data, 11-13 percent of HIV+ women experience one or more breast pathologies during breastfeeding. The conditions are usually more common during the first weeks of lactation. The prevalence of these conditions does not appear to be greater than among HIV-uninfected women (e.g., data from Semba -15.6 percent; Filteau - 12-13 percent). [Note: Breast milk sodium or sodium/potassium ratios are easy to measure and have been reported for lactating women with and without HIV over different stages of lactation. Sodium is a plasma constituent that passes into breast milk when tight junctions between mammary epithelial cells become permeable. This occurs, and milk sodium levels are high, during lactogenesis (when cellular pathways have not yet closed), during breast inflammation (leukocytes are also present) and systemic infection, and when breastmilk production decreases, especially below 200 ml/d (during weaning).] Breast inflammation is another important risk factor for HIV transmission during breastfeeding. According to available data, 11-13 percent of HIV+ women experience one or more breast pathologies during breastfeeding. The conditions are usually more common during the first weeks of lactation. The prevalence of these conditions does not appear to be greater than among HIV-uninfected women (e.g., data from Semba -15.6 percent; Filteau - 12-13 percent). [Note: Breast milk sodium or sodium/potassium ratios are easy to measure and have been reported for lactating women with and without HIV over different stages of lactation. Sodium is a plasma constituent that passes into breast milk when tight junctions between mammary epithelial cells become permeable. This occurs, and milk sodium levels are high, during lactogenesis (when cellular pathways have not yet closed), during breast inflammation (leukocytes are also present) and systemic infection, and when breastmilk production decreases, especially below 200 ml/d (during weaning).]

18. 18 Risk factor: Breast Pathology Breast inflammation & mastitis ? increased risk of postnatal transmission (Embree et al; John et al; Semba et al) Nipple lesions, breast abscesses ? increased transmission (Fawzi et al, 2002; Embree et al, 2000; Ekpini et al, 1997) Sub-clinical mastitis ? higher viral load in BM (Willumsen et al, 2000; Semba et al, 1999, Hoffman, 2003) This slide summarizes the literature on the increased risk of transmission due to various breast conditions. Breast inflammation, mastitis, nipple lesions and breast abscesses are significantly related to transmission. I haven�t included all the risk ratios and confidence intervals but the graphic summarizes our calculations from 2 studies that provide both a relative risk estimate and a prevalence of the condition. These indicate that the fraction of transmission that can be attributed to breast inflammation in Malawi and Kenya are 18 and 20%, respectively. If this is a causal association and all of the excess transmission occurs during breastfeeding, this suggests that as much as half of breastfeeding transmission may be due to breast conditions that are preventable and treatable. Finally, we know that sub-clinical mastitis is associated with a higher level of virus in breastmilk so this points to a mechanism. This slide summarizes the literature on the increased risk of transmission due to various breast conditions. Breast inflammation, mastitis, nipple lesions and breast abscesses are significantly related to transmission. I haven�t included all the risk ratios and confidence intervals but the graphic summarizes our calculations from 2 studies that provide both a relative risk estimate and a prevalence of the condition. These indicate that the fraction of transmission that can be attributed to breast inflammation in Malawi and Kenya are 18 and 20%, respectively. If this is a causal association and all of the excess transmission occurs during breastfeeding, this suggests that as much as half of breastfeeding transmission may be due to breast conditions that are preventable and treatable. Finally, we know that sub-clinical mastitis is associated with a higher level of virus in breastmilk so this points to a mechanism.

19. 19 Association between breast inflammation and breast milk virus These data are from a recent report of a small pilot study conducted in Malawi (n=33 mothers sampled on two occasions). A comparison of whole milk samples between 4 mothers with clinical signs of breast inflammation and the rest showed significant differences in both viral load and sensitivity. Women with breast inflammation were 2.5 times more likely to have detectable levels of virus in their breastmilk and had breastmilk viral loads about 3 logs (1000 times) higher. [These results are not adjusted for plasma viral load, or CD4 count, which is also associated with breastmilk viral load in this study. However, confounding is unlikely since uninfected mothers have similar rates of breast problems, suggesting that there is no association between maternal infection itself and breast problems.] These data are from a recent report of a small pilot study conducted in Malawi (n=33 mothers sampled on two occasions). A comparison of whole milk samples between 4 mothers with clinical signs of breast inflammation and the rest showed significant differences in both viral load and sensitivity. Women with breast inflammation were 2.5 times more likely to have detectable levels of virus in their breastmilk and had breastmilk viral loads about 3 logs (1000 times) higher. [These results are not adjusted for plasma viral load, or CD4 count, which is also associated with breastmilk viral load in this study. However, confounding is unlikely since uninfected mothers have similar rates of breast problems, suggesting that there is no association between maternal infection itself and breast problems.]

20. 20 Impact of lactation counseling on sub-clinical mastitis: Bangladesh But can anything be done to prevent breast problems? The positive impact of lactation counseling on breast health/sub-clinical mastitis is illustrated in this slide from a study in Bangladesh (not among HIV-infected mothers) where women were randomized to receive lactation management support at delivery versus no special lactation support. The study showed that a single counseling session on good breastfeeding techniques at the onset of lactation resulted in significantly less breast inflammation among new mothers. Mild inflammation was twice as prevalent and the prevalence of severe inflammation was 3 times greater in women who did not receive this counseling. This suggests that early breastfeeding counseling can reduce the risk of sub-clinical and clinical mastitis - important risk factors for HIV transmission during breastfeeding. But can anything be done to prevent breast problems? The positive impact of lactation counseling on breast health/sub-clinical mastitis is illustrated in this slide from a study in Bangladesh (not among HIV-infected mothers) where women were randomized to receive lactation management support at delivery versus no special lactation support. The study showed that a single counseling session on good breastfeeding techniques at the onset of lactation resulted in significantly less breast inflammation among new mothers. Mild inflammation was twice as prevalent and the prevalence of severe inflammation was 3 times greater in women who did not receive this counseling. This suggests that early breastfeeding counseling can reduce the risk of sub-clinical and clinical mastitis - important risk factors for HIV transmission during breastfeeding.

21. 21 Risk factor: First months Higher in the first months of life (Nduati et al, 2000; John et al, 2001) Why? Maybe: higher prevalence of mastitis, breastfeeding problems infant gut more immature, permeable greater exposure (higher concentration of cells) Some studies indicate that the risk of transmission may be higher in the first months of life. There are several possible explanations for this - one is the higher prevalence of mastitis and other breastfeeding problems; the infant gut is more immature and thus possibly also more vulnerable and permeable; and it could be a greater risk because more breast milk is consumed (higher exposure) in young versus older infants.Some studies indicate that the risk of transmission may be higher in the first months of life. There are several possible explanations for this - one is the higher prevalence of mastitis and other breastfeeding problems; the infant gut is more immature and thus possibly also more vulnerable and permeable; and it could be a greater risk because more breast milk is consumed (higher exposure) in young versus older infants.

22. 22 Risk Factor: First month In this slide we compare the rates of HIV seroconversion from birth to 4, 6, or 8 weeks in breastfeeding versus non-breastfeeding infants in several studies. These rates are expressed per month and are compared here with estimates from the Read et al meta-analysis estimate of the rate of late postpartum transmission (after 6 weeks). These comparisons are not perfect. In Kenya infants were randomized to breastfeeding or not, while in the other studies feeding patterns were self-selected. In SAINT infants were exposed to ARV prophylaxis, whereas in the other studies they were not. For the SAINT trial data we used the least effective perinatal regimen (ZDV/3CT) to minimize the effect of ARVs and for the PETRA trial the estimate is based on the difference at 6 weeks between BF and non-BF infants in the placebo group. Sources: a) Kenya: Nduati R, John G, Mbori-Ngacha D, et al. JAMA 2000; 283: 1167-74. b) Durban: Coutsoudis A, Pillay K, Kuhn L, Spooner E, Tsai W-Y, Coovadia HM. AIDS 2001, 15:379-387. c) SAINT: D Moodley - presentation at the Global Strategies Conference, September 2001. updated in Moodley et al. JID 2003 187: 725-35. d) Petra: Petra Study Team. Lancet 2002; 359: 1178-86 e) Read: see next slideIn this slide we compare the rates of HIV seroconversion from birth to 4, 6, or 8 weeks in breastfeeding versus non-breastfeeding infants in several studies. These rates are expressed per month and are compared here with estimates from the Read et al meta-analysis estimate of the rate of late postpartum transmission (after 6 weeks). These comparisons are not perfect. In Kenya infants were randomized to breastfeeding or not, while in the other studies feeding patterns were self-selected. In SAINT infants were exposed to ARV prophylaxis, whereas in the other studies they were not. For the SAINT trial data we used the least effective perinatal regimen (ZDV/3CT) to minimize the effect of ARVs and for the PETRA trial the estimate is based on the difference at 6 weeks between BF and non-BF infants in the placebo group. Sources: a) Kenya: Nduati R, John G, Mbori-Ngacha D, et al. JAMA 2000; 283: 1167-74. b) Durban: Coutsoudis A, Pillay K, Kuhn L, Spooner E, Tsai W-Y, Coovadia HM. AIDS 2001, 15:379-387. c) SAINT: D Moodley - presentation at the Global Strategies Conference, September 2001. updated in Moodley et al. JID 2003 187: 725-35. d) Petra: Petra Study Team. Lancet 2002; 359: 1178-86 e) Read: see next slide

23. 23 Risk Factor: Duration of breastfeeding BHITS meta-analysis Duration of breastfeeding is another risk factor for postnatal transmission. The risk of HIV transmission persists for as long as breastfeeding is practiced. This slide shows preliminary data from a meta-analysis (presented at the Barcelona AIDS Conference in July 2002). The meta-analysis uses data from 4343 children born to HIV+ mothers from 9 studies all over Africa. 999 children were infected with HIV. For nearly half (45.1 percent), the timing of transmission could not be ascertained. For the remaining infants, approximately 4 percent became infected with HIV between 4 weeks and 6 months; another 5 percent between 6 months and 12 months; and another 7 percent from 12-18 months of infancy. This analysis adjusts for actual breastfeeding practices (once infants stop breastfeeding they are excluded). The monthly risk from this preliminary meta-analysis is about 0.9% per month (as shown in the previous slide). It is important to note that the amount of transmission in the first month of infant life due to breastfeeding is not known and is difficult to ascertain/differentiate from infection that occurs during late pregnancy and labor/delivery. But it is likely higher than later on (as shown in the previous slide). Duration of breastfeeding is another risk factor for postnatal transmission. The risk of HIV transmission persists for as long as breastfeeding is practiced. This slide shows preliminary data from a meta-analysis (presented at the Barcelona AIDS Conference in July 2002). The meta-analysis uses data from 4343 children born to HIV+ mothers from 9 studies all over Africa. 999 children were infected with HIV. For nearly half (45.1 percent), the timing of transmission could not be ascertained. For the remaining infants, approximately 4 percent became infected with HIV between 4 weeks and 6 months; another 5 percent between 6 months and 12 months; and another 7 percent from 12-18 months of infancy. This analysis adjusts for actual breastfeeding practices (once infants stop breastfeeding they are excluded). The monthly risk from this preliminary meta-analysis is about 0.9% per month (as shown in the previous slide). It is important to note that the amount of transmission in the first month of infant life due to breastfeeding is not known and is difficult to ascertain/differentiate from infection that occurs during late pregnancy and labor/delivery. But it is likely higher than later on (as shown in the previous slide).

24. 24 Risk Factor:Duration of Breastfeeding Nairobi, Kenya: Randomized trial of formula vs breastfeeding Statistical model developed (n=358 infants with 75 infections, 52 possibly through breastmilk) Overall probability of HIV transmission per day of BF = 0.00028/day (=0.85% per month) Risk continues as long as breastfeeding continues Median duration of breastfeeding: 17 months Using data from the Nairobi randomized trial of formula vs breastfeeding, a statistical model was developed to account for actual breastfeeding practices and loss to follow-up. Overall the probability of HIV transmission was 0.00028/day (equivalent to 10% per year). [No statistical difference between risk <4 months vs. > 4 months was observed. (explained by the authors as due to the high early rates being �diluted� by 4 months).] The point here is that the risk of transmission here continued as long as breastfeeding continued, which here was a median of 17 months. Using data from the Nairobi randomized trial of formula vs breastfeeding, a statistical model was developed to account for actual breastfeeding practices and loss to follow-up. Overall the probability of HIV transmission was 0.00028/day (equivalent to 10% per year). [No statistical difference between risk <4 months vs. > 4 months was observed. (explained by the authors as due to the high early rates being �diluted� by 4 months).] The point here is that the risk of transmission here continued as long as breastfeeding continued, which here was a median of 17 months.

25. 25 Proportion of postnatal HIV transmission occurring after 6 months in selected studies (excludes first 4-6 weeks) Although the first month of life appears to convey a relatively greater risk of HIV transmission, the contribution of late postnatal transmission to total breastfeeding (postnatal) transmission is significant (the meta-analysis estimate for transmission after 6 months=12%). This slide shows the proportion of infant postnatal infections occurring AFTER 6 months in several studies. The analysis excludes the first 4-6 weeks since it is difficult to ascertain whether these infections were due to breastfeeding or intra-partum transmission. Data suggest that continued breastfeeding after 6 months is responsible for a very significant portion of BF-related transmission (excluding the first 4-6 weeks). If you assume that there is a 3-4 percent risk of BF transmission in the first 4-6 weeks, the contribution of late (> 6 months) postnatal transmission is about 50% in all studies except Kenya (Nduati et al). In conclusion, early breastfeeding cessation (after 6 months) is likely to prevent a significant proportion of BF-related transmission and increase HIV-free survival if it can be done safely - with nutritionally adequate and appropriate replacement foods. Although the first month of life appears to convey a relatively greater risk of HIV transmission, the contribution of late postnatal transmission to total breastfeeding (postnatal) transmission is significant (the meta-analysis estimate for transmission after 6 months=12%). This slide shows the proportion of infant postnatal infections occurring AFTER 6 months in several studies. The analysis excludes the first 4-6 weeks since it is difficult to ascertain whether these infections were due to breastfeeding or intra-partum transmission. Data suggest that continued breastfeeding after 6 months is responsible for a very significant portion of BF-related transmission (excluding the first 4-6 weeks). If you assume that there is a 3-4 percent risk of BF transmission in the first 4-6 weeks, the contribution of late (> 6 months) postnatal transmission is about 50% in all studies except Kenya (Nduati et al). In conclusion, early breastfeeding cessation (after 6 months) is likely to prevent a significant proportion of BF-related transmission and increase HIV-free survival if it can be done safely - with nutritionally adequate and appropriate replacement foods.

26. 26 Risk Factor: Early Mixed breastfeeding The type of breastfeeding practiced (exclusive versus partial) may also be an important risk factor for HIV transmission. This slide shows the well-known data from Anna Coutsoudis and colleagues� study in Durban SA. It compares the cumulative rates of HIV transmission in infants exclusively or partially breastfed (to 3 months). The rates of infection were similar at birth, but partially breastfed infants had higher rates of postnatal transmission at 3, 6, and up to 15 months of age when compared with babies who had been exclusively breastfed. These data remain very controversial since they are observational data and have not been replicated. There are several studies in progress to try to replicate these findings (in South Africa, Zambia, Zimbabwe, and elsewhere). The type of breastfeeding practiced (exclusive versus partial) may also be an important risk factor for HIV transmission. This slide shows the well-known data from Anna Coutsoudis and colleagues� study in Durban SA. It compares the cumulative rates of HIV transmission in infants exclusively or partially breastfed (to 3 months). The rates of infection were similar at birth, but partially breastfed infants had higher rates of postnatal transmission at 3, 6, and up to 15 months of age when compared with babies who had been exclusively breastfed. These data remain very controversial since they are observational data and have not been replicated. There are several studies in progress to try to replicate these findings (in South Africa, Zambia, Zimbabwe, and elsewhere).

27. 27 Feeding mode and Morbidity of children born to Women with HIV From the same Durban study we just looked at but from a more recent publication, this shows the relationship between mode of feeding and (instead of transmission) child illness and hospitalization. There were no significant differences between the illness experiences of children mixed vs. exclusively breastfed. However comparing �never� with �ever� breastfed children, the never breastfed children had a higher incidence of illness and hospitalization in the first 2 months of life. Both of these differences were statistically significant. Thus even in the context of a research trial the health effects of not breastfeeding are very evident. From the same Durban study we just looked at but from a more recent publication, this shows the relationship between mode of feeding and (instead of transmission) child illness and hospitalization. There were no significant differences between the illness experiences of children mixed vs. exclusively breastfed. However comparing �never� with �ever� breastfed children, the never breastfed children had a higher incidence of illness and hospitalization in the first 2 months of life. Both of these differences were statistically significant. Thus even in the context of a research trial the health effects of not breastfeeding are very evident.

28. 28 Bacterial Contamination and Improper Preparation of Commercial Infant Formula in a PMTCT Program in Durban, South Africa Characteristics of mothers (n=94) 54% completed high school or greater 66% had indoor piped water 70% flush toilet Contamination of milk samples 64% E Coli 26% Enterococci Over dilution of milk samples 22% for infants <= 12 months 78% for children > 12 months These data come from a study of infant replacement feeds in a PMTCT program, also in South Africa. Despite the relatively high educational attainment, and good conditions of hygiene and sanitation, more than half of all milk samples were contaminated with diarrheal pathogens and a significant proportion of the samples were inappropriately diluted, exposing infants and young children to malnutrition. This illustrates the reality that safe preparation of infant formula and bottle feeding is extremely challenging, even under relatively good conditions. These data come from a study of infant replacement feeds in a PMTCT program, also in South Africa. Despite the relatively high educational attainment, and good conditions of hygiene and sanitation, more than half of all milk samples were contaminated with diarrheal pathogens and a significant proportion of the samples were inappropriately diluted, exposing infants and young children to malnutrition. This illustrates the reality that safe preparation of infant formula and bottle feeding is extremely challenging, even under relatively good conditions.

29. 29 Higher Rates of Hospitalization for Non-Breastfed Infants of HIV+ Mothers in a PMTCT Program in Pune, India And another illustration� As part of a multi-center trial of perinatal ARVs (ZDV short course), investigators in Pune, India conducted an observational study of the health effects of replacement feeding, by comparing hospitalization rates between breastfed infants and infants who were replacement fed, mostly with diluted animal milk. Whereas none of the 62 breastfed infants died or were hospitalized, among the 86 non-breastfed infants there were 27 hospitalizations and 4 deaths. Hospitalization rates were particularly high in the early postpartum period about 2.5 times greater in the 1st month than during the 2nd and 3rd month. And another illustration� As part of a multi-center trial of perinatal ARVs (ZDV short course), investigators in Pune, India conducted an observational study of the health effects of replacement feeding, by comparing hospitalization rates between breastfed infants and infants who were replacement fed, mostly with diluted animal milk. Whereas none of the 62 breastfed infants died or were hospitalized, among the 86 non-breastfed infants there were 27 hospitalizations and 4 deaths. Hospitalization rates were particularly high in the early postpartum period about 2.5 times greater in the 1st month than during the 2nd and 3rd month.

30. 30 Breastfeeding Saves Lives There is a wealth of literature on the survival benefits of breastfeeding but until recently, age-specific estimates covering all of infancy were not available. To address this gap, WHO published these age specific relative risks of death due to artificial feeding a couple of years ago, and explicitly suggested that they be used to help decide when replacement feeding by an infected mother would be safer than continuing to risk HIV transmission through breastfeeding. There is a wealth of literature on the survival benefits of breastfeeding but until recently, age-specific estimates covering all of infancy were not available. To address this gap, WHO published these age specific relative risks of death due to artificial feeding a couple of years ago, and explicitly suggested that they be used to help decide when replacement feeding by an infected mother would be safer than continuing to risk HIV transmission through breastfeeding.

31. 31 Quantifying theBalance of Risks Miriam Labbok and I took up this challenge and developed a mathematical risk simulation model to examine the balance of risks � between HIV transmission through breastfeeding and on the other hand the additional risk of mortality from not breastfeeding. Miriam Labbok and I took up this challenge and developed a mathematical risk simulation model to examine the balance of risks � between HIV transmission through breastfeeding and on the other hand the additional risk of mortality from not breastfeeding.

32. 32 Timing the Introduction of Replacement Feeding This is the theoretical framework for our age-specific risk assessment model which we have used to answer policy relevant questions like At what age, under �typical� conditions, is it safer to use replacement feeds? At what level of infant mortality is replacement feeding from birth safer than beginning with breastfeeding and switching later on. How do different infant feeding strategies compare in terms of HIV-free survival? This is the theoretical framework for our age-specific risk assessment model which we have used to answer policy relevant questions like At what age, under �typical� conditions, is it safer to use replacement feeds? At what level of infant mortality is replacement feeding from birth safer than beginning with breastfeeding and switching later on. How do different infant feeding strategies compare in terms of HIV-free survival?

33. 33 HIV and infant feeding risk model 4 feeding strategies compared: We simulated 4 different infant feeding strategies: B24: No postnatal intervention: all mothers breastfeed as usual for 24 months B0: HIV-positive mothers do not breastfeed and commercial infant formula is provided. B6: HIV-positive mothers breastfeed for 6 months then switch to replacement feeding. SB6: As in the previous strategy HIV-positive mothers breastfeed for 6 months but an intervention is in place to reduce breastfeeding transmission by 50%. This �safer breastfeeding� intervention would involve exclusive breastfeeding, prevention and treatment of breast problems, and use of antiretroviral drugs by the infant and/or mother.We simulated 4 different infant feeding strategies: B24: No postnatal intervention: all mothers breastfeed as usual for 24 months B0: HIV-positive mothers do not breastfeed and commercial infant formula is provided. B6: HIV-positive mothers breastfeed for 6 months then switch to replacement feeding. SB6: As in the previous strategy HIV-positive mothers breastfeed for 6 months but an intervention is in place to reduce breastfeeding transmission by 50%. This �safer breastfeeding� intervention would involve exclusive breastfeeding, prevention and treatment of breast problems, and use of antiretroviral drugs by the infant and/or mother.

34. 34 Cumulative HIV-free Survival Among Infants of HIV-infected Mothers IMR=91 (average for sub-Saharan Africa) This slide shows the results of simulation of these 4 scenarios for a �typical� sub-Sarahan African country with an IMR of about 91/1000 live births. As shown in the slide, a program to support six months of safer BF (SB6) - the red line - gives the best outcome overall because it results in the highest HIV-free survival. Providing ARV prophylaxis but not addressing infant feeding issues for HIV+ women (continued BF to 2 years, B24) produces the worst outcome - the most infant HIV infections and deaths (blue line). At six months the worst outcome occurs when there is no breastfeeding (B0) (green line). Although this strategy avoids breastfeeding transmission, this is outweighed by the additional deaths, mainly from infectious diseases, due to not breastfeeding.This slide shows the results of simulation of these 4 scenarios for a �typical� sub-Sarahan African country with an IMR of about 91/1000 live births. As shown in the slide, a program to support six months of safer BF (SB6) - the red line - gives the best outcome overall because it results in the highest HIV-free survival. Providing ARV prophylaxis but not addressing infant feeding issues for HIV+ women (continued BF to 2 years, B24) produces the worst outcome - the most infant HIV infections and deaths (blue line). At six months the worst outcome occurs when there is no breastfeeding (B0) (green line). Although this strategy avoids breastfeeding transmission, this is outweighed by the additional deaths, mainly from infectious diseases, due to not breastfeeding.

35. 35 Cumulative HIV-free Survival Among Infants of HIV-infected MothersINDIA: IMR=66 (SRS, 2001) In India, where the IMR is lower (66 per 1000 live births) the additional risk of death due to artificial feeding is therefore also lower and the difference between not breastfeeding and breastfeeding (with no intervention to make breastfeeding safer) is negligible. However, the safest strategy is again �safer breastfeeding.� These analyses are intended only to help guide programs and policies at the population level and are not intended for use in individual counseling, where each mother�s specific situation should be taken into account, including social and economic factors. In India, where the IMR is lower (66 per 1000 live births) the additional risk of death due to artificial feeding is therefore also lower and the difference between not breastfeeding and breastfeeding (with no intervention to make breastfeeding safer) is negligible. However, the safest strategy is again �safer breastfeeding.� These analyses are intended only to help guide programs and policies at the population level and are not intended for use in individual counseling, where each mother�s specific situation should be taken into account, including social and economic factors.

36. 36 Intervention Research on Postnatal Transmission

37. 37 Interventions to Prevent Postnatal Transmission Mother Avoid breastfeeding Wet nursing Early cessation Maternal ARV Preventing and treating breast conditions Exclusive Breastfeeding Heat treatment Nutrition? Infant Post-exposure ARV prophylaxis Preventing and treating lesions in mouth, intestine Immunization ? Given all that we know about risk factors for postnatal transmission, this slide summarizes the implied preventive interventions, again distinguishing between those for mothers and those for infants. Trials of regimens to prevent postnatal transmission are in progress. One randomised trial on the use of formula has been reported. Given all that we know about risk factors for postnatal transmission, this slide summarizes the implied preventive interventions, again distinguishing between those for mothers and those for infants. Trials of regimens to prevent postnatal transmission are in progress. One randomised trial on the use of formula has been reported.

38. 38 Overview of Trials Planned or Underway 32 intervention field trials on MTCT reviewed 18 related to some aspect of infant feeding: Post-exposure prophylaxis (PEP) during BF (8) PEP + immunization (1) Early cessation (3) Vitamin A + EBF (1) EBF (1) Highly Active Antiretroviral Therapy (HAART) (3) Chloroquin (1) 14 on perinatal transmission: ARVs (12) Immunization (2) This is the current list of field trials planned, underway or in analysis [reference: Ghent PMTCT trials inventory at http://www.isped.u-bordeaux2.fr/ISPED/RECHERCHE/GHENT/WWW_BASE/US-GHENT-Existing-Trial.asp ]. Unlike the situation until very recently � when virtually all the research on PMTCT was directed at ARV therapies to prevent perinatal transmission � most studies are now directly related to some aspect of infant feeding. However, there is still an alarming absence in many of the perinatal ARV trials of any appreciation of the importance of infant feeding. It would add greatly to our understanding of the risks of transmission, and would take little extra effort, if, in the many studies on the effects of therapies to prevent perinatal transmission, infant feeding practices were carefully monitored to complement the data on transmission and infant health. This is the current list of field trials planned, underway or in analysis [reference: Ghent PMTCT trials inventory at http://www.isped.u-bordeaux2.fr/ISPED/RECHERCHE/GHENT/WWW_BASE/US-GHENT-Existing-Trial.asp ]. Unlike the situation until very recently � when virtually all the research on PMTCT was directed at ARV therapies to prevent perinatal transmission � most studies are now directly related to some aspect of infant feeding. However, there is still an alarming absence in many of the perinatal ARV trials of any appreciation of the importance of infant feeding. It would add greatly to our understanding of the risks of transmission, and would take little extra effort, if, in the many studies on the effects of therapies to prevent perinatal transmission, infant feeding practices were carefully monitored to complement the data on transmission and infant health.

39. 39 SIMBA Study � Uganda and Rwanda -1 Design: 405 HIV+ women received ZDV+DDI (from 36 wks+ 1 wk pp) All counseled to EBF and wean from 3-6 months 397 infants randomized to daily NVP or 3TC from 1 wk to 1 month after BF stopped Findings: Overall rate HIV transmission (KM): 8% 6% in-utero; 1% from birth-4 wks; 1% 4 wks-6 mo These data are from the only postnatal ARV trial to be publicly presented to date, the Simba trial in Uganda and Rwanda, presented at the International AIDS Society meeting in Paris in July 2003. 405 HIV+ women received a combination of ARVs (ZDV+DDI) from 36 weeks of pregnancy through 1 week after delivery. All were counseled to breastfeed exclusively and to wean from 3-6 months. 397 infants were randomized to receive either nevirapine or 3TC daily from 1 wk to 1 month after BF stopped. The overall rate of HIV transmission was only 8% with 6% occurring in-utero, 1% from birth to 4 wks and 1% from 4 wks to 6 mo. Much was made of this latter statistic, which was compared with the 15% of infants infected through breastfeeding in the absence of ANY preventive intervention. [ZDV=zidovudine; DDI=didanosine; 3TC=lamivudine]These data are from the only postnatal ARV trial to be publicly presented to date, the Simba trial in Uganda and Rwanda, presented at the International AIDS Society meeting in Paris in July 2003. 405 HIV+ women received a combination of ARVs (ZDV+DDI) from 36 weeks of pregnancy through 1 week after delivery. All were counseled to breastfeed exclusively and to wean from 3-6 months. 397 infants were randomized to receive either nevirapine or 3TC daily from 1 wk to 1 month after BF stopped. The overall rate of HIV transmission was only 8% with 6% occurring in-utero, 1% from birth to 4 wks and 1% from 4 wks to 6 mo. Much was made of this latter statistic, which was compared with the 15% of infants infected through breastfeeding in the absence of ANY preventive intervention. [ZDV=zidovudine; DDI=didanosine; 3TC=lamivudine]

40. 40 SIMBA Study � Uganda and Rwanda -2 Conclusions/Concerns: Infant prophylaxis a promising strategy for reducing PN transmission Concerns: Widely cited comparison with 15% transmission rate in other studies is misleading lack of control arm asymptomatic population so risk of transmission was low to begin with What was the intervention? Dual treatment of mother, infant prophylaxis, EBF (88%), early cessation (3.3 months) Although infant prophylaxis is a promising strategy for reducing PN transmission, comparing the 1% transmission from 4 weeks to 6 months with the 15% transmission rate in other studies is misleading. The lack of a control arm means that we do not know how much transmission would have otherwise occurred during this period. It is likely much less than 15% because this was an asymptomatic population so risk of transmission would have been low anyway, and the intervention included dual treatment of mother, infant prophylaxis, exclusive breastfeeding (88%), and early cessation (3.3 months). Therefore the effects cannot all be attributed to infant prophylaxis. Although infant prophylaxis is a promising strategy for reducing PN transmission, comparing the 1% transmission from 4 weeks to 6 months with the 15% transmission rate in other studies is misleading. The lack of a control arm means that we do not know how much transmission would have otherwise occurred during this period. It is likely much less than 15% because this was an asymptomatic population so risk of transmission would have been low anyway, and the intervention included dual treatment of mother, infant prophylaxis, exclusive breastfeeding (88%), and early cessation (3.3 months). Therefore the effects cannot all be attributed to infant prophylaxis.

41. 41 Challenges for the Future It is important to integrate all of the information provided in this presentation in order to make concrete recommendations for individuals and populations - and to consider the future challenges. This review suggests the following: It is important to integrate all of the information provided in this presentation in order to make concrete recommendations for individuals and populations - and to consider the future challenges. This review suggests the following:

42. 42 1.Focus on maternal health & nutrition Keeping HIV+ mothers well may be among the most important things we can do to prevent P/N transmission BF transmission was ~2% between 6 w-24 months in women with CD4 >500 (Leroy et al, 2003) Nutrition depletion, weight loss during BF may increase risk of maternal mortality, especially in immune compromised mothers (Nduati et al, 2001) Keeping mothers alive will improve child�s chances for survival (Nduati et al, 2001; Nakiyingi et al, 2003) We need to focus our thinking about PMTCT more on maternal health and nutrition. Keeping HIV+ mothers healthy may be among the most important things we can do to prevent postnatal HIV transmission. Recall that breastfeeding transmission was only about 2 percent between 6 weeks and 24 months in the West Africa study in women who had baseline CD4 counts over 500. Furthermore, nutrition depletion and weight loss may increase the risk of maternal death during the postnatal period (for breastfeeding and non-breastfeeding mothers alike). The study in Kenya also found that keeping mothers alive improved their children�s chances of survival. How to do this through ARV treatment, improved maternal care, and other health and nutrition interventions, is being studied. We need to focus our thinking about PMTCT more on maternal health and nutrition. Keeping HIV+ mothers healthy may be among the most important things we can do to prevent postnatal HIV transmission. Recall that breastfeeding transmission was only about 2 percent between 6 weeks and 24 months in the West Africa study in women who had baseline CD4 counts over 500. Furthermore, nutrition depletion and weight loss may increase the risk of maternal death during the postnatal period (for breastfeeding and non-breastfeeding mothers alike). The study in Kenya also found that keeping mothers alive improved their children�s chances of survival. How to do this through ARV treatment, improved maternal care, and other health and nutrition interventions, is being studied.

43. 43 2. Expanding Use of ARVs Lower prices, wider variety of available regimens, easier logistics, trial results spurring demand (including HAART PMTCT trials) ? expanding postnatal use and availability of ARVs increasing concerns about long term effects on infant (toxicity and resistance) � studies looking at this now Legitimate demand for a single standard of care regardless of socioeconomic conditions ? currently HAART for mother, perinatal ARV therapy, and replacement feeding from birth (with all necessary support) ARV treatment for mothers in resource limited settings will expand due to: lower prices, wider variety of available regimens, easier logistics, trial results spurring supply and demand. Part of the challenge will involve responding to: concerns about long term effects on infant (toxicity and resistance) � studies looking at this now demands for a single standard of care regardless of socioeconomic conditions, including HAART for the mother (where appropriate), perinatal ARV therapy, and replacement feeding from birth (with all the necessary support) ARV treatment for mothers in resource limited settings will expand due to: lower prices, wider variety of available regimens, easier logistics, trial results spurring supply and demand. Part of the challenge will involve responding to: concerns about long term effects on infant (toxicity and resistance) � studies looking at this now demands for a single standard of care regardless of socioeconomic conditions, including HAART for the mother (where appropriate), perinatal ARV therapy, and replacement feeding from birth (with all the necessary support)

44. 44 3. Strengthen approaches for making breastfeeding safer for ALL women Provide adequate lactation counseling and support, involving families/communities increase adherence to exclusive breastfeeding promote good breastfeeding techniques prevent cracked nipples, maintain breast health Immediate treatment for mastitis, other systemic infections that could affect viral load in BM could prevent a sizeable fraction of BF transmission may be most important in early month(s) Safe sex/condom use for prevention In the meantime, there is an opportunity and an obligation here to make breastfeeding safer for ALL women. Many practices considered best practice for uninfected mothers also reduce the risk of HIV transmission through breastfeeding. Community-wide promotion of these practices would 1) bring benefits to uninfected mothers and their infants, 2) help reduce transmission among the many women who do not know their HIV status 3) reduce the stigma associated with these practices for infected mothers who do know their status. This could be done through several means - improving family and community support for mothers� health and nutrition and optimal breastfeeding practices. Such support will increase adherence to exclusive breastfeeding; promote good breastfeeding techniques and breast health; and prevent cracked nipples and other risky conditions. The support for safer breastfeeding should also include immediate treatment for mastitis and other systemic infections that affect viral load in breast milk. This is particularly important for HIV-positive mothers. We believe that this could prevent a sizeable fraction of postnatal transmission. It may be most important in the early months when these conditions are more common and transmission risk is also highest. Finally, safe sex practices should be promoted among HIV-negative women to PREVENT HIV infection during the postnatal breastfeeding period. In the meantime, there is an opportunity and an obligation here to make breastfeeding safer for ALL women. Many practices considered best practice for uninfected mothers also reduce the risk of HIV transmission through breastfeeding. Community-wide promotion of these practices would 1) bring benefits to uninfected mothers and their infants, 2) help reduce transmission among the many women who do not know their HIV status 3) reduce the stigma associated with these practices for infected mothers who do know their status. This could be done through several means - improving family and community support for mothers� health and nutrition and optimal breastfeeding practices. Such support will increase adherence to exclusive breastfeeding; promote good breastfeeding techniques and breast health; and prevent cracked nipples and other risky conditions. The support for safer breastfeeding should also include immediate treatment for mastitis and other systemic infections that affect viral load in breast milk. This is particularly important for HIV-positive mothers. We believe that this could prevent a sizeable fraction of postnatal transmission. It may be most important in the early months when these conditions are more common and transmission risk is also highest. Finally, safe sex practices should be promoted among HIV-negative women to PREVENT HIV infection during the postnatal breastfeeding period.

45. 45 4. Make breastfeeding safer for HIV+ women Assist families with decisions about early breastfeeding cessation assess health status of mother and infant prepare for the process so that the transition is safe (cup-feeding, safe preparation/hygiene, stigma) heat treat breast milk if weaning is gradual could prevent � to � of BF transmission Provide adequate infant nutrition after breastfeeding ends appropriate breast milk substitutes and/or multi-nutrient supplements should be provided to prevent malnutrition In addition to making early BF safer for all mothers, we also have to help HIV+ mothers and their families to make appropriate decisions about early cessation. This decision about when to stop breastfeeding should be based on an assessment of the health status of both the mother and the infant. Parents need to prepare for early cessation to ensure that the transition is safe for them and their baby. This includes teaching infants to feed from a cup, teaching parents how to prepare replacement milks and foods safely and hygienically, and helping parents to address any pressure or stigma in the community. Heat treating breast milk may help in the transition process, and where it is culturally acceptable, it should be taught. As we have noted in this presentation, early breastfeeding cessation will prevent a sizeable fraction of postnatal transmission - half or more if cessation is around 6 months. However, in order to prevent malnutrition and ensure adequate child growth, health, and development, early cessation must be accompanied by adequate nutrition for the infant. Since breast milk is an essential part of the infant diet throughout the first year (and beyond), programs that encourage early cessation should provide nutritious breast milk substitutes and/or multi-nutrient supplements. The use of �ready to eat foods� is being studied for this purpose. In addition to making early BF safer for all mothers, we also have to help HIV+ mothers and their families to make appropriate decisions about early cessation. This decision about when to stop breastfeeding should be based on an assessment of the health status of both the mother and the infant. Parents need to prepare for early cessation to ensure that the transition is safe for them and their baby. This includes teaching infants to feed from a cup, teaching parents how to prepare replacement milks and foods safely and hygienically, and helping parents to address any pressure or stigma in the community. Heat treating breast milk may help in the transition process, and where it is culturally acceptable, it should be taught. As we have noted in this presentation, early breastfeeding cessation will prevent a sizeable fraction of postnatal transmission - half or more if cessation is around 6 months. However, in order to prevent malnutrition and ensure adequate child growth, health, and development, early cessation must be accompanied by adequate nutrition for the infant. Since breast milk is an essential part of the infant diet throughout the first year (and beyond), programs that encourage early cessation should provide nutritious breast milk substitutes and/or multi-nutrient supplements. The use of �ready to eat foods� is being studied for this purpose.

46. 46 5. Make replacement feeding safer for HIV+ women Provide safe water & environmental conditions rural and urban areas may vary Family support, community understanding Postnatal follow-up and enhanced care essential child health interventions Screen mothers, target use to those most at risk Take measures to prevent unnecessary use of RF need to strengthen efforts to support optimal infant feeding for all In settings where replacement feeding is being provided or practiced, there must also be efforts to make it safer for HIV+ women and to prevent spillover in the general (non-HIV) population. Safe RF requires not only the provision of adequate supplies of breast milk substitutes, but also safe water and environmental conditions. Family support and community understanding are also needed. As mentioned before, postnatal follow-up and enhanced care are required for all mothers to ensure that they can safely feed their children � and this is especially true among infants who are not breastfed/no longer breastfeeding. This care should include all the essential child health and survival interventions (immunizations, growth promotion, nutrition management, vitamin A supplementation, use of bednets, etc). In settings where replacement feeding is being provided or practiced, there must also be efforts to make it safer for HIV+ women and to prevent spillover in the general (non-HIV) population. Safe RF requires not only the provision of adequate supplies of breast milk substitutes, but also safe water and environmental conditions. Family support and community understanding are also needed. As mentioned before, postnatal follow-up and enhanced care are required for all mothers to ensure that they can safely feed their children � and this is especially true among infants who are not breastfed/no longer breastfeeding. This care should include all the essential child health and survival interventions (immunizations, growth promotion, nutrition management, vitamin A supplementation, use of bednets, etc).

47. 47 Thank You!