In Malawi, HIV-positive women who breast-feed their infants are no more likely to become ill or die than their counterparts who do not breast-feed, according to an analysis of longitudinal data from HIV-infected mothers and their newborns.1 This finding was not affected by women's frequency or pattern of breast-feeding. In addition, the women's breast-fed infants were about 60% less likely to die in the first two years of life than their non–breast- fed counterparts, whether or not they were infected with HIV.

The data came from a pair of clinical trials conducted in Malawi in 2000–2003 that tested antiretroviral therapy for preventing mother-to-infant transmission of HIV. Blood samples were collected from HIV-positive mothers at the time of delivery to measure their HIV load and their hemoglobin level, and from their newborns to test for HIV infection. Social, demographic, medical and reproductive information was recorded at delivery. At each of 10 visits over the next two years, mothers were examined and were asked if they were breast-feeding; those who were breast-feeding their infants were asked how frequently they did so and whether they were giving their infants only breast milk (classified as exclusive breast- feeding) or breast milk plus other liquids or solids (classified as mixed breast-feeding). Multivariate analyses assessed associations of breast-feeding with maternal health and survival, and with infant survival, controlling for maternal age, initial maternal viral load and hemoglobin level, and body mass index (weight for height) at the follow-up visits.

A total of 2,000 women and their singleton infants were enrolled in the trials. On average, the women were about 25 years old and had had three live births. Eleven percent had not attended school, 63% had attended primary school and 26% had a higher level of education.

Slightly more than 2% of mothers died in the two years after delivery. The cumulative probability of death was 18 per 1,000 at one year and 32 per 1,000 at two years. Maternal deaths were most commonly due to tuberculosis, pneumonia, malaria and diarrhea; the cause was unknown in about one-fifth of cases.

About 16% of infants died in the two years after delivery. The cumulative probability of death was 132 per 1,000 at one year and 195 per 1,000 at two years. Infant and child deaths were most commonly due to respiratory infections, gastroenteritis and septicemia; the cause was unknown in about one-seventh of cases. The estimated proportion of infants and children surviving and not infected with HIV was 80% at one year and 73% at two years.

On average, women breast-fed their infants for 15 months overall, exclusively breast-fed for 2.4 months and practiced mixed breast-feeding for 11.7 months. In a comparison of measures of maternal health between women who did and did not breast-feed in the first year (to assess the possibility that health itself influenced this practice), the two groups of women did not differ with respect to initial HIV viral load or hemoglobin level, or body mass index at visits.

Women who breast-fed did not have a significantly different risk of death than their counterparts who did not breast-feed. In addition, the risk did not vary between women who breast-fed five or more times in a 24-hour period and those who did so less frequently, or between women who exclusively breast-fed and those who did not breast-feed. Women who practiced mixed breast-feeding had a lower risk of death than did those who did not breast-feed (hazard ratio, 0.3). In terms of other factors, the likelihood of maternal death was positively associated with initial viral load (3.8–3.9), and negatively associated with initial hemoglobin level (0.8) and body mass index (0.9).

Breast-feeding in general, its frequency and its pattern were not associated with increased risks of illness among the women, as assessed with three measures—hospitalization and use of medicines, the presence of HIV symptoms and the need for assistance with daily activities. In fact, the likelihood of hospitalization and use of medicines was lower among women who exclusively breast-fed than among those who did not breast-feed (odds ratio, 0.8), and the likelihood of needing help with daily activities was lower among women who breast-fed in general (0.7) and those who practiced mixed breast-feeding (0.7) than among women who did not breast-feed. All three measures of illness were positively associated with initial viral load (1.2–1.6), and negatively associated with initial hemoglobin level (0.9–1.0) and with body mass index (0.9). Also, women younger than 25 years of age were less likely to experience illness than their older counterparts (0.6–0.7).

Infants and children who were breast-fed had a lower risk of death than their non–breast-fed counterparts (hazard ratio, 0.4). Both mixed and exclusive breast-feeding were protective when compared with no breast-feeding (0.5 and 0.4, respectively). In addition, infants' and children's risk of death was positively associated with their mother's initial viral load (2.6), but was negatively associated with maternal body mass index (0.9).

The association between breast-feeding and lower mortality remained when infants and children were stratified by their HIV status at 6–8 weeks of age. Specifically, compared with the risk of death among infants who were not breast-fed, the risk was lower among HIV- negative and HIV-positive infants alike who were breast-fed, regardless of pattern (hazard ratios, 0.3 and 0.4, respectively), those who received both breast milk and supplemental foods (0.4 and 0.4) and those who were exclusively breast-fed (0.1 and 0.4).

The researchers conclude that breast-feeding by HIV-positive mothers does not appear to hasten the progression of their illness or their death; moreover, this practice can be life-saving for their children, although it also poses a risk of infection. These findings, they assert, support recommendations adopted by several countries for breast-feeding when breast milk substitutes are not available, despite maternal HIV infection. They note that AIDS, as measured by viral load, remains the main risk factor for death among mothers and children alike. "Therefore, providing antiretroviral treatment to mothers (and their children) should be a major priority in order to save lives," they contend.—S. London


1. Taha TE et al., The impact of breastfeeding on the health of HIV-positive mothers and their children in sub-Saharan Africa, Bulletin of the World Health Organization, 2006, 84(7):546–554.