Cover Image
close this bookNutrition and HIV/AIDS United Nations Administrative Committee on Coordination, Sub-Committee on Nutrition. Nutrition Policy Paper #20. Report of the 28th Session Symposium Held 3-4 April 2001, Nairobi, Kenya (UNAIDS - UNSSCN, 2001, 85 p.)
View the document(introduction...)
View the documentForeword and Acknowledgements
View the documentNutrition and HIV/AIDS
View the documentOverview of the 29th Session Symposium - Nutrition and HIV/AIDS
View the documentKeynote Address
View the documentAchievements of the AIDS Support Organization (TASO) in Uganda
View the documentHIV/AIDS and Development: Unsolved Challenges for Africa
View the documentHIV/AIDS, Food and Nutrition Security: Impacts and Actions*
View the documentNutrition and the Care Package
View the documentPanel Discussion on the Implications of HIV/AIDS for Nutrition Programmes
View the documentDr. Abraham Horwitz Memorial Lecture - Infant Feeding Options for Mothers with HIV: Using women’s Insights to Guide Polices
View the documentAnnex 1 - The facts about nutrition and HIV/AIDS
View the documentAnnex 2 - Effect of breastfeeding on mortality among HIV-infected women
View the documentList of abbreviations
View the documentNutrition Policy Papers Series

Panel Discussion on the Implications of HIV/AIDS for Nutrition Programmes

Dr. Elizabeth Marum
Centre for Disease Control and Prevention, Kenya

Dr Piot talked about food and the lack of food being the number one problem mentioned by people living with HIV/AIDS in Malawi. Professor Saasa raised the issue of “ill conceived home-based care programmes”. In Malawi we did an assessment of home-based care programmes throughout the country. Our findings relate to the issue of hunger among people living with HIV/AIDS, and the problem of home-based care programmes, which are not well designed. We interviewed 50 patients or their care givers in their homes and 47 home-based care volunteers serving these patients. We also had focus group discussions. We looked at a number of issues including food and nutrition. Sixty-eight percent of the patients were women and only one in four female patients was married and living with her husband. The rest were being cared for primarily by female care givers, similar to what has been observed by TASO in Uganda. The average age of the male patients was 46; the women were considerably younger, with an average age of 32. We found that the needs and requests they had were very simple. They needed food, pain-killing medications, soap, bed clothes and help with household tasks. The sad story is that these basic needs were not provided by the home-based care programmes. There was an acute unmet need for food supplementation for these patients and their children.

Our concerns about sustainability and the complexity of the problem should not prevent us from making sure that people living with HIV/AIDS in Africa have enough to eat

Eighty-six percent of the patients reported needing food but only one person was receiving food on a regular basis. It is important to note that these were not people living with AIDS who we found in the community. These were patients who were enrolled in HIV/AIDS home-based care programmes and yet only one was receiving food on a regular basis. Another 13 received food sometimes, but not often enough. The food they were receiving unfortunately came directly from the volunteers themselves, from their own food supplies. The home-based care programmes, some of which were reasonably well funded by external donors, did not supply any food supplements. We heard repeatedly from the volunteers that they could not continue doing this, that it was putting their own families under constraint. These data highlight that food supplementation for people living with HIV/AIDS is an acute problem and one which home-based care programmes do not adequately address. Colleagues in Kenya report a similar situation. In Uganda the availability of food supplementation was a motivation to be enrolled in home-based care. It was also motivation to get tested.

We also found that the treatment of even simple opportunistic infections was very inadequate. For example, 43% of the adult patients reported recent symptoms of oral thrush but not one was receiving correct treatment. Two of these patients had received some form of traditional treatment, and one had received pain medications. However, none had received the relatively simple and inexpensive drug which treats oral thrush. Untreated oral thrush can contribute to dehydration and malnutrition because eating becomes difficult and painful.

Another major issue is that of compensation for volunteers working in these programmes. In almost every home-based care programme and community HIV/AIDS programme that I know of, there is an excessive dependence on volunteer labour. In Malawi, Uganda and Kenya there are people who are motivated to help their communities but they have their own problems at home. To always assume that these counsellors can do this work on a volunteer basis is really not very helpful. Why don’t we use food aid as a way to compensate people doing volunteer HIV/AIDS work? The model of “food for work” programmes in India and Bangladesh could be adapted to “food for HIV/AIDS work” for people in rural communities who are giving their time. The HIV/AIDS community needs to pay more attention to nutrition issues. Many programmes do not factor in nutrition adequately. Just as there are now gender assessments, there should also be nutrition impact assessments. The gender community have been very effective in making sure that gender issues receive attention in all kinds of development programmes, including HIV/AIDS programmes. The nutrition community needs to speak up and to engage the HIV/AIDS community, to ensure that nutrition is taken fully into account.

Another important problem is the issue of testing and the lack of voluntary counselling and testing. New technologies can help us in this regard. The latest rapid test kits are very simple; we do not need machines any more. The new test kits are also reliable and accurate. They can be done with just a finger prick. What this means is that voluntary counselling and testing can move away from labs, away from hospitals, out of health centres, to the communities. These test kits were introduced for same-hour results in Malawi. There was a four-fold increase in people coming in for voluntary counselling and testing. Early diagnosis of HIV infection means that nutritional interventions for people living with HIV/AIDS can be offered early.

The final issue I would like to discuss is sustainability. I recognize, without being an expert in this area, that there are complex relationships between food aid and food security. Giving out food to people living with HIV/AIDS may affect local food production. However, I feel very strongly that no person living with HIV/AIDS should also be living with hunger. Our concerns about sustainability and the complexity of the problem should not prevent us from making sure that people living with HIV/AIDS in Africa have enough to eat.

Dr Ruth Nduati
University of Nairobi

HIV/AIDS is expanding rapidly in Africa, and contributing to poverty, malnutrition and endemic illness. In adolescent girls HIV/AIDS and malnutrition co-exist. About two thirds of people who are infected actually get infected when they are adolescents1. It is not unusual for a teenager to get pregnant and become HIV-infected the first time she has sex. These girls are the mothers of the undernourished underfives; some 30% of underfive children in this region are stunted. These children will grow up to be malnourished mothers who give birth to low birthweight babies, and the cycle continues. In the Sub-Saharan region 20-30% of pregnant women are HIV-infected, and 25-35% of these will transmit the virus to their infants, either during pregnancy, at delivery or through breastfeeding. We are able to reduce mother-to-child transmission (MTCT) through the use of anti-retroviral drugs. In the developed world we talk about eradicating MTCT. In the developing world we have had some significant advances with these drugs, however, gains have been modest because the infants are still exposed to HIV when they breast feed. In Africa, reduced transmission has not translated into improved survival of children of HIV-infected women. In addition to preventing viral transmission, much more needs to be done to improve the well-being of young infants in Africa.

I want to focus my discussion on HIV, nutrition and women. First, what effect does maternal malnutrition have on the outcome of pregnancy? Second, what is the impact of increased nutritional demands of pregnancy and lactation on disease progression in mothers and what is the impact of nutritional supplementation on infant and maternal outcomes? We know that HIV-infected individuals have increased resting energy expenditure (REE) of approximately 10% above that of non-infected persons2. This is equivalent to about 150 kcals per day. If there is co-infection, such as tuberculosis or diarrhoea, there is a 34% higher resting energy expenditure compared to an uninfected individual who does not have any co-morbidity. There is also a correlation between REE and viral load; as viral load increases so does REE3.

Most research has focused on infant outcomes and not on impacts on the mother

We know that death in patients with HIV, or in other chronic wasting diseases, can be predicted by the amount of weight lost. Death usually occurs when one third of lean body mass is lost. There are several studies in the developed world, mainly in men, that have looked at weight loss in HIV infected people4,5. In addition to the elevated REE, anorexia and depression reduce energy intake in HIV-infected patients while some drugs suppress appetite. The HIV-infected person has an abnormal metabolic response to her disease. Under normal circumstances if energy intake is low, REE falls to compensate for the decrease. However, with HIV infection it is not possible to reduce REE. The body conserves energy by reducing physical activity. An HIV-infected woman will reduce her physical activity by curtailing cooking, gardening and other household tasks. This will further aggravate her disease because she does not have access to the meals she would have cooked for herself and her family.

Lactation is energy demanding. Healthy women produce around 600 to 700 ml of milk in a 24-hour period. This requires about 600 kcals per day in additional food. Healthy women meet these demands by eating more, by breaking down the fat they stored during pregnancy, and possibly by reducing physical activity. Malnourished women are able to make adequate milk for their babies. However, if food intake is poor and there are no fat reserves, maternal muscle will be broken down to meet the demands of the baby. Women who are malnourished and who get enough to eat during lactation actually gain weight when they lactate. The co-existence of lactation and HIV is a very significant metabolic challenge. The requirements of increased resting energy expenditure, plus energy needed to support exclusive breastfeeding, are equivalent to one whole extra meal per day, providing at least 750 kcals. HIV-infected women who have experienced child death have special needs. These women tend to have shorter birth intervals because they try to replace the child who has died. Frequent reproductive cycles may contribute to nutritional depletion.

I want to highlight data from recent studies that bring home this interaction of nutrition and HIV in women starting with maternal nutrition and infant outcome. Malnourished women have increased rates of premature deliveries, low birthweight and intrauterine growth retardation6. In 1998, a meta-analysis was carried out on 31 published cohort studies comparing HIV-infected women to uninfected women7. HIV-infected women had a higher risk of low birth-weight (LBW), higher risk of preterm delivery, higher risk of intrauterine growth retardation, and also increased risk of perinatal mortality. In an analysis of developed versus developing countries these differences were much more marked7. This interaction between poor nutrition and non-fatal outcomes has been confirmed by the work of Dr Fawzi and others in Tanzania. In a randomized clinical study of micronutrients supplementation in pregnant HIV infected women, these authors showed that use of vitamins reduced preterm births by 39%, low birthweight by 44%, intrauterine growth retardation by 43% compared to women who received iron/folate and prophylaxis for malaria.8 These findings are exciting and have important implications for programmes.

Many studies, particularly in the developed world have looked at whether pregnancy accelerates HIV progression9,10. Most have found that it does not. However, a recent meta-analysis demonstrated that there was a non-significant increase in maternal deaths among HIV-infected women and a progression of their HIV disease. Also, they were more likely to have an AIDS-defining illness11. This was more prominent in women from the developing world. A community-based study in Malawi looked at maternal mortality and found that among HIV-infected women the rate was similar to uninfected women (that is during pregnancy and the first 42 days after). However, in the first year after delivery the risk of death in HIV-infected women was tenfold more than in uninfected women12.

Most research has focused on infant outcomes and not on impacts on the mother. In Sub-Saharhan Africa, breastfeeding is the choice for many HIV-infected women. Our randomized clinical trial of breastfeeding and formula feeding among infants of HIV-infected women carried out between 1992-98 showed that the rate for breastmilk transmission was 16.2%13. The unique design of this study allowed us to look at the impact of lactation on maternal health. In our study, breastfeeding and formula feeding mothers were comparable both for health and social and demographic parameters. The median duration of breastfeeding was 17 months; 95% were breastfeeding at three months, 89% at six months and 80% at 12 months. The median age of introduction of complementary feeding was 3.8 months. Overall mortality among breastfeeding women was ten percent, while among formula feeders it was three percent. This was a threefold increased risk of dying among breastfeeding women. Sixty-nine percent of maternal deaths in the breastfeeding group were attributable to breastfeeding. Median weight loss in the first five to nine months was one kg, while the group using formula did not lose weight. The women in this study were all relatively malnourished; they lost very little because they had little to lose. Normally a woman will lose about 0.6 kg or less per month in the first six months of lactation14,15. The average weight loss of the women who died in this study was at least 4.2 kg13. We concluded that these deaths were attributable to nutritional depletion aggravated by the metabolic demands of lactation.

We must target women to improve birth outcomes and nutritional status of children. We can improve lactation performance and improve survival of mothers so there are fewer orphans. We have to strengthen health services. This is not the time to reinvent the wheel. We need to take what we already know and apply it wisely and aggressively. We have many nutritionally valuable foods in this region that are under-exploited and need to be promoted. We should distribute iron/folate and B vitamins during pregnancy, prevent malaria through chemoprophylaxis and provide anti-helminths to reduce intestinal parasites. An unpublished study in this region found that only 30% of pregnant mothers who turn up in government health facilities actually receive anti-helminths. Many obstetricians do not know that we must give iron and folate to pregnant women because there is not enough in the diet to meet nutritional requirements.

Dr Phillip Mwalari
National AIDS Control Council, Kenya

Africa accounts for 70% of the world’s HIV cases but 80% of deaths due to AIDS. This indicates that there is another factor in Africa which makes those infected not survive as long. In Kenya, before the 1999 census, the projected population was 30 million but the census showed we had 28 million. We looked at trends related to HIV/AIDS and showed that HIV/AIDS contributed to the drop to 28 million. Approximately 75% of infected people in Kenya live in the rural areas. According to the census about 80-85% of the population live in rural areas - so in fact the prevalence of HIV/AIDS might be getting close to what we see in the urban areas. We need to move very fast to have control programmes in the rural areas. When HIV/AIDS landed in Africa it found a very fertile environment to grow and spread - poverty, poor nutrition and an immuno-compromised population. Socio-cultural norms have enabled HIV to spread as well. Poverty and overcrowding promote opportunistic infections that accelerate death. AIDS also found poor health-seeking behaviours in the population.

...we hope to shift as much as 60% of the resources to the community

In 1999, Kenya realized that the magnitude of the problem was such that we needed to take action. Hence, HIV/AIDS was declared a national disaster and this led to the formation of the National HIV/AIDS Council. This Council has set up a network that takes the message down to the grassroots level. At the provincial level the Provincial HIV/AIDS Coordinating Committee has HIV/AIDS coordinating units in the public (e.g., ministries) and private sectors. At the community level we have Constituency HIV/AIDS Coordinating Committees to promote community mobilization. We realize that there is a lot of common ground between the spread of HIV/AIDS and malnutrition. Knowing that poor nutrition will make HIV/AIDS worse and vice versa, there is a need for us to integrate HIV education into our programmes. Within this approach we hope to shift as much as 60% of the resources to the community, leaving 40% or less for other aspects of fighting AIDS. A strong multi-sectoral approach and an opportunity for nutrition promotion in this country involves all the ministries and takes programmes to the grassroots.

There is a need for applied research on the nutrient composition of traditional food crops as well as operations research on the best ways to promote consumption of these crops. These crops are not widely consumed anymore, although some of them are very nutritious.


Question from the audience Preliminary data have been published recently from a study carried out in South Africa on mortality among HIV-infected mothers16. There was no difference between breastfeeding and non-breastfeeding mothers. How does this study compare with the Kenyan one?

Question from the audience Regarding the quick test for HIV in mothers and children, in Tanzania, we find that if we, as counsellors, spend a few minutes or even an hour to counsel these mothers before the quick test, it is more beneficial than simply doing the test with no counselling. The problem is how to scale up. Scaling up will require much larger counselling facilities to accommodate everyone. Every mother should be counselled on how to manage and how to care for herself. If mothers are not counselled properly and the child dies she will certainly come back pregnant the next year. This test may be very cheap, but it is not available free even in Tanzania. Only at the blood banks and in areas where projects are operational are quick tests available. Finally, we should not rush to formula feeding. In Tanzania only 30% of our people have access to clean water. Similar rates are found elsewhere in Africa. As long as we cannot improve the water supply we cannot improve hygiene.

Comment from the audience So far we have discussed only women. It is important to involve men. Men must be more fully involved so that they can participate in the care of the mother and child.

Question from the audience Solutions to the HIV/AIDS pandemic are not just sectoral or country specific. When it comes to drugs, why do we not strategize as a continent? Is there a continental initiative at the level of the Organization of African Unity? The pharmaceutical companies are very powerful and to negotiate effectively with them one needs a collective position.

Comment from the audience This was the first panel that spoke about the mother and infant together. This issue of the mother and HIV and breastfeeding has to be looked at, as well, with the infant and HIV and breastfeeding. Dr. Nduati’s work is important in this respect because it begins to look at the dyad. Finding “cheaper formula” is not the issue. With “cheaper formula”, the birth interval will be shorter and the needs of the new baby must be met. Even if there is enough money to buy any kind of formula, the mother can be fed much cheaper than the baby. In the USA, it would cost an estimated 50 cents a day to give the mother enough extra energy for breastfeeding. It would be less in other countries. It is important to think about the mother’s survival and the baby’s survival together, not separately.

Question from the audience Regarding supplementation, nutrient requirements of people living with HIV/AIDS are very high and it may be difficult to meet these requirements through food alone. There are many companies in Uganda that are marketing products they claim meet these high needs. The companies are making a lot of money from the sale of these food supplements. Should we, as nutritionists, encourage these companies? Should we encourage people living with HIV/AIDS to purchase these supplements?

Question from the audience Firstly, is the rapid test kit reliable and affordable? Is it available? Secondly, I have a comment on the use of the term “prevention of mother to child transmission”, or PMTCT. So long as we continue to use the term prevention of mother-to-child transmission of HIV/AIDS, we lay the entire responsibility and the blame for the transmission at the mother’s doorstep. Can we not make a recommendation to be a little more progressive in the language we use? It is not a semantic issue; it is a matter of principle. Can we not move to using prevention of parent-to-child transmission where the father’s role is recognized?

Question from the audience I am a community nutritionist working with people living with HIV/AIDS. I am concerned about sustainability of our nutrition efforts for children. We already have child-headed families. Firstly, how sustainable are our programmes targeting this very deprived group? Secondly, I am concerned about economic empowerment at the community level. How are we approaching economic empowerment for these communities? Are we promoting locally available resources? We hear about food supplements and medications. However, we do not hear about traditional knowledge. Some of the opportunistic infections can be treated locally at the community level. Should we not promote local solutions? In our programme, we find that UNIMIX (a porridge meant for children) is used by some mothers because they cannot take solid food.

Ruth Nduati The South African study did not find an increase in mortality, whereas the Kenyan study did. The two studiesa are different. Our study is a randomized clinical trial whereas the other is a cohort study. The duration of breast-feeding in the two studies is very different. In the South African study the medium duration of breastfeeding was six months. In our study it was 17 months. I think our study is the first to report this and we need to replicate it. It is also the first to report increased mortality among lactating women. We need funding now for a randomized clinical trial of supplementation. We will probably try energy plus micronutrients to see whether we can mitigate the deterioration of mothers’ health during lactation.

a Editor’s Note: on 7 June, 2001 the World Health Organization issued a statement on the effect of breastfeeding on mortality among HIV-infected women. This statement is reproduced in Annex 2.

Regarding formula, there are women who will choose to formula feed when they are HIV-infected as a way of protecting their infant, especially mothers who have lost other children to HIV infection. I believe they have a right to make that choice and we should be able to support them. The majority of mothers will breastfeed and we need to support them in their efforts to exclusively breastfeed. We are not doing this currently.

The question of supplements for HIV-infected women or for pregnant women is very important. Adequate energy can come from food alone, but some micronutrients, iron in particular, will be needed as a supplement. This has a beneficial impact on the infant’s iron status as well. Zinc is also a micronutrient needed, most probably, as a supplement. In general, where there is little animal protein in the diet there is a need for micronutrient supplementation. The private sector is already involved. Widespread fortification of foods that are commonly consumed by the whole population will help to improve the overall nutritional status of the population.

There is the biological definition of prevention of mother-to-child transmission and the social definition. When we talk of prevention of mother-to-child transmission in the biological sense, we are speaking specifically of the time points of pregnancy, delivery and breastfeeding when transmission takes place. Fathers are indeed significant regarding transmission in the social sense. However, for the purposes of this discussion we should adhere to science and use terminologies that will help us communicate exactly what we are talking about.

Regarding orphans and child-headed households, we have not begun to scratch the surface in terms of developing programmes that will support their needs. Programmes are needed to address their nutritional well-being, but also their social development.

Elizabeth Marum Regarding HIV testing, we should not do wholesale testing without counselling. Pretest and post-test counselling is the essential component of voluntary counselling and testing. A test result, in and of itself, does not help people change their behaviour. However, counsellors do sometimes discourage people from getting tested. We need to encourage an environment in which knowledge of sero-positive status is one of the first and most important steps in determining what behaviour and choice should be in decisions on pregnancy and family planning. We can take advantage of this new technology along with the counselling. There is no shortage of counsellors, in my opinion, in many countries in Africa. Hundreds of nurses in Kenya, Uganda and Malawi have been trained as counsellors. They can use the new test kits, simply and easily, to inform people who come for testing and counselling. The rapid kits are exceedingly reliable, in part because they are read on the same day, eliminating lab and recording error. Costs range from US$1.20 to $1.80 per test.

Phillip Mwalari On the question of sustainability of programmes, we have Constituency HIV/AIDS Control Committees at the community level. They have been in existence for awhile and have shown managerial skills and capacity. These groups will be supported so they can continue their important work. This will be more sustainable than pushing activities and interventions onto the communities without their involvement. Of course, we are also trying to mobilize local resources within the communities.

The question concerning traditional healers is relevant. We believe everybody who shows interest in HIV/AIDS control should be invited to come on board. We have created a forum for traditional healers to come forward and work towards a common goal. If we truly want a decentralized approach to HIV/AIDS we need to bring traditional healers on board. On the question of nutritional supplements promoted by the commercial sector, some of these products are good. However, prices are too high for most people. We look to the professional nutrition community to monitor the composition and overall quality of these products.

Minister Dlamini Regarding child-headed families and sustainability, when people come to assist they sometimes think we need to build orphanages or somehow bring the children to homes where they can be cared for. When a parent is lost, it is a lot for the child to deal with. However, when the child is removed from the family environment, siblings, school friends and neighbours are also lost. So in fact the orphans are traumatized by parental loss, sibling loss and loss of friends. Some efforts to help them have actually damaged further an already extremely difficult situation. This is what we have come up with in Swaziland.

Swaziland is a small country where we have traditional chiefs. In other countries, you would call this local government. These chiefs know the families in their community. The communities, or villages, can be clustered geographically so that 10 to 15 homes form a cluster. The children in a cluster know one another. When a mother falls ill, the children can tell the neighbour within the cluster. The neighbours can assist and give support. When the mother dies, the children will not lose their home, even if the father is gone. They stay right there. We are building this into the legislation, but the state should assist these people to look after the children and keep them in school. This means, also, that our educational system should change. We were trained, by the British system, to look for jobs and to be employed. We should review our educational system and train our children to create employment and to be creative. Then, even as youth they can have little projects and earn a living and still stay in school. The children are supported by a network. This network rests on our traditional values because in the past we always supported each other.

As far as sustainability, people living with HIV/AIDS want to do something. They may not be able to garden, but they can become involved locally in income-earning projects. We teach them and they teach each other. Marketing can be facilitated by government. Above all, it is crucial to sit and learn from people living with HIV/AIDS and ask them what they would like to do.


1. Williams B, Gouws E, Wilkinson D, et al (2001) Estimating HIV incidence rates from age prevalence data in epidemic situations. Statistics in Medicine 20:2003-2016.

2. Melchior JC, Raguin G, Boulier A, et al (1993) Resting energy expenditure in human immunodeficiency virus-infected patients: comparison between patients with and without secondary infection. American Journal of Clinical Nutrition 57:614-619.

3. Mulligan K, Tai VW, Schambelan M, et al (1997) Energy expenditure in human immunodeficiency virus infection. New England Journal of Medicine 336:70-71.

4. Mulligan K, Tai VW, Schambelan M (1997) Cross-sectional and longitudinal evaluation of body composition in men with HIV infection Journal of Acquired Immune Deficiency Syndromes and Human Retrovirology 15:43-48.

5. Macallan DC (1999) Wasting and HIV infection and AIDS. Journal of Nutrition 129: 238S-242S.

6. Gonzalez-Cossio T and Delgado H (1991) Functional consequences of maternal malnutrition. In: Selected Vitamins, Minerals and the Functional Consequences of Maternal Malnutrition Simopoulos AP (ed) World Review of Nutrition and Dietetics. Karger: Basel.

7. Brocklehurst P and French R (1998) The association between maternal HIV infection and perinatal outcome: a systematic review of the literature and meta-analysis. British Journal of Obstetric Gynaecology 105:836-848.

8. Fawzi WW, Msamanga GI, Spiegelman D, et al (1998) Randomised trial of effects of vitamin supplements on pregnancy outcomes and T cell counts in HIV-1-infected women in Tanzania. Lancet 351:1477-1482.

9. Saada M, Le Chenadec J, Berrebi A, et al (2000) Pregnancy and progression to AIDS: results of the French prospective cohorts. SEROGEST and SE-ROCO Study Groups. AIDS 14:2355-2360.

10. Brettle RP, Raab GM, Ross A, et al (1995) HIV infection in women: immunological markers and the influence of pregnancy. AIDS 9:1177-1184

11. French R and Brocklehurst P (1998) The effect of pregnancy on survival in women infected with HIV: a systematic review of the literature and meta-analysis. British Journal of Obstetric Gynaecology 105:827-835.

12. McDermott JM, Slutsker L, Steketee RW, et al (1996) Prospective assessment of mortality among a cohort of pregnant women in rural Malawi. American Journal of Tropical Medicine and Hygiene 55(1 Suppl):66-70

13. Nduati R, Richardson BA, Grace J, et al (2001) Effect of breastfeeding on mortality among HIV-1 infected women: A ransomised trial. Lancet 357:1651-1655.

14. van Raaij JM, Schonk CM, Vermaat-Miedema SH, et al (1991) Energy cost of lactation, and energy balances of well-nourished Dutch lactating women: reappraisal of the extra energy requirements of lactation. American Journal of Clinical Nutrition 53:612-619.

15. Haiek LN, Kramer MS, Ciampi A, et al (2001) Postpartum weight loss and infant feeding. Journal of the American Board of Family Practice 14:85-94.

16. Coutsoudis A, Coovadia H, Pillay K, et al (2001) Are HIV-infected women who breastfeed at increased risk of mortality? AIDS 15:653-655.