STATS ARTICLES 2007
HIV and Breastfeeding
Is it better for HIV mothers to breastfeed or not breastfeed – and does it depend on where they live?
A new study published in The Lancet has found that when a HIV-positive mother exclusively breastfeeds her child, the health benefits are much greater than if she were to partially breastfeed (that is, use both breast milk and other sources of nutrition).
Unfortunately, the study’s results were misunderstood by the media. Reuters Africa reported that
“Researchers in South Africa, writing in the Lancet medical journal, tracked 1,372 HIV-infected women and found a 4 percent risk of postnatal transmission of the human immunodeficiency virus to babies fed only breast milk for six months after birth.”
And Associated Content wrote that
“Research shows that exclusive breast feeding can significantly reduce the risk of HIV being transmitted from mother to child in infants under age six months.”
This is all patently false. The rate of HIV infection among breastfed babies is almost 20 percent. The four percent figure is among babies who were uninfected after six weeks. And the reduced risk attributed to breast feeding is compared to babies who also have formula, not compared to those who are exclusively formula fed.
The main issue with this study is that it is not comparing breastfeeding and not breastfeeding; the risks of breastfeeding are significant, and well-documented. The problem is that so is the risk of bad water (use to mix formula) on an undeveloped gut. What is not clear is whether an underdeveloped gut that is also exposed to water, formula, or solids is also more vulnerable to HIV transmission through breast milk.
Among those who are fed both formula and breast milk, the study is unclear about the percentage who were uninfected after six weeks that became infected by six months. While “twice as high” is mentioned several times, this figure is cited out of context – it does not refer to those babies who are HIV-negative at six weeks. One tell-tale paragraph about babies who tested negative at six weeks reads “In Cox regression analysis with exclusive breastfeeding as reference, the hazard ratio (HR) for breastmilk plus other food or fluids was 1·56 (0·66–3·69, p=0·308).” Translation: for babies who are HIV-negative at six weeks, there is no increased rate of HIV infection among those who have mixed feeding or are given solids, compared to those who have breast milk alone.
So where does this “twice as high” figure come from? The article is not entirely clear on this, though it seems that this may be a measure of the difference of HIV infection rate among all babies from birth. Unfortunately, we don’t know if more of the babies in this “mixed-feeding” group were already HIV positive at six weeks, perhaps due to gestational or during-birth infection. We also don’t know if they were more likely to develop HIV before six weeks (when almost all mothers were exclusively breastfeeding) because the CD-4 cell counts were lower.
The Lancet study was observational, not controlled. That is to say, the women were not told whether to supplement their breast milk with solids or formula or not: they were simply provided with the formula should they choose to use it. While in theory, those who decided to breastfeed exclusively and those who decided to do “mixed feeding” were similar groups of women, both sets of womenwere HIV-positive South African women with young children. But HIV infection does not affect all people equally at all times before the infection turns lethal. If the women who are sicker are also less likely to breastfeed alone, the whole data could be skewed.
How would the data be skewed if sicker women introduced formula into the diet of their babies? It could be that they have higher levels of HIV virus in their bodies and in their breast milk, which is then passed to the babies. It could also be that they are themselves immuno-compromised, which in turn diminishes the positive immunological effects of the breast milk that their babies do get. Yet another possibility is that there are other, secondary infections that these women suffer from – which, when passed to their babies, make the babies more vulnerable to the HIV exposure through breast milk. In all these cases, there would be no greater positive effect for their babies were these women to nurse exclusively.
A Technical Problem in the Study
One of the most worrisome aspects of the study is the how the study accounted for CD-4 cell counts. CD-4 cells are a type of T-cells that are suggestive of how much the HIV virus has progressed. The lower the CD-4 cell count, the sicker the mother.
It is clear that CD-4 cell counts were significantly correlated with mothers’ decisions to nurse, and whether to nurse exclusively or not. Perhaps some of the sicker mothers were simply unable to nurse, or their families took on a larger role of caring for the infants than the babies of healthy mothers. Regardless of the reason, of the women who nursed exclusively, about eight percent had CD-4 counts below 200 cells per micro-liter of blood. In contrast, over 12 percent of those who started with exclusive nursing and then switched to mixed feeding had CD-4 counts below 200 cells per micro-liter.
The authors of the study do not seem to control for this difference, which could have had significant impact on transmission rates as well. Since those mothers with lower cell counts were more likely to pass HIV on to their babies as well as more likely to introduce formula and/or solids to their babies diet, we may inadvertently blame the foods instead of the low cell count for the increased HIV transmission.
The difference between United States and the third world
One of the main claims of this paper is that survival rates were higher with exclusive breastfeeding, as compared to exclusive formula feeding, in this population of mothers with a high rate of HIV infection.
This study should not be used to evaluate the risks or benefits of breastfeeding in developed countries.
The two main factors leading to infant deaths in this study were illness resulting from poor water quality, and infection by HIV. Those who are given formula are exposed to the water (with an immature gut), while those who are given breast milk are exposed to the HIV virus. In making these comparisons, the study is trying to assess which of the two evils is lesser.
Both of these are different calculations for people living in North America and Europe. First of all, water quality is high and babies do not typically die of waterborne illness and associated diarrhea and dehydration. Over 70 percent of American babies are formula fed, with no indication that this has an impact on death rates (we noted last year that the American Academy of Pediatrics’ claim that using formula increases death rates was based on injury rates, not formula use).
The HIV calculation is also different in developed countries. Those mothers with access to drugs that significantly lower their CD4 cell count (a type of T-cell) are much less likely to pass HIV infection on to their infants. In this study on South African women, those with cell counts of less than 200 cells per micro-liter of blood were much more likely to pass HIV infection on to their children than those with counts of higher than 500 cells per micro-liter.
In developed countries, women who are able to keep their CD4 cell count low face a reduced risk for transmitting HIV infection to their babies through breastfeeding. But they can eliminate risk entirely by exclusively using formula.
Death rates versus HIV-free rates
The six-month survival rate was similar for both the babies who were exclusively breastfed and those who were exclusively formula fed, at just under 88 and 85 percent, respectively. Yet little mention is given of the fact that, among those living who were breastfed, a significant portion has HIV due to the breastfeeding – and little access to modern, expensive drug therapy. The elephant in the room is that survival rates and HIV-free survival rates are not equivalent from a public health point of view.
A look at the HIV-free survival rates suggests that formula fed babies and breastfed babies do about equally well – both have a 75 percent HIV-free survival rate at six months of age. But unfortunately the study does not distinguish between babies who were born HIV-positive, and those who acquire it (through nursing). If babies who are born with HIV are more likely to be given formula, the data are significantly skewed. Indeed, there was a highly significant correlation (p=.003) between replacement feeding and a very low CD-4 cell count of fewer than 200 cells per micro-liter of blood.
What can we take from this?
We can conclude very little from this study about what role breastfeeding should play in American life, except that HIV-positive mothers should continue to avoid breastfeeding to minimize the possibility of transmitting infection to their babies.
For South Africans, the same data could be reworked to control for the obvious factors such as low CD-4 cell counts to see if the results still hold. If so, women who breastfeed should be encouraged not to introduce any other nutrients or water until their babies are six months old. If not, these same women should be encouraged to give their babies formula and to avoid breastfeeding entirely.
View the Technorati Link Cosmos for this entry