Infants born to HIV-infected women in a clinical trial in Kenya were almost twice as likely to contract the virus by age two if they were given breast milk as if they were fed formula, and the difference in risk was entirely attributable to breastfeeding. More than half of those who were infected through breast milk contracted the virus within the first six weeks of their lives. Infants who were nursed and those who were fed only formula had similar mortality rates, but the rate of infection-free survival was significantly higher among formula-fed infants than among those who were given breast milk.1

Study participants were recruited from among women attending four prenatal clinics in Nairobi between November 1992 and July 1998 who tested positive for HIV and had access to clean water. Participants received counseling about mother-to-infant HIV transmission, the risks and benefits of formula feeding and breastfeeding, and the nature of the study; they also agreed to adhere to a randomly assigned feeding method.

Randomization took place at about 32 weeks' gestation; at that time, participants in the formula-feeding group received free dried formula and instruction on how to prepare it. Counseling on infant feeding practices remained available throughout the study period. Women and their infants were followed up monthly for the first year and quarterly for the second year. At each follow-up visit, the woman provided an interim history, and both she and her baby had a physical examination. Breast milk and the infant's blood were tested at regular intervals.

The analyses are based on 401 women and their infants (excluding second-born twins), of whom 197 were assigned to the breastfeeding group and 204 to the formula-feeding group. In both groups, the women's median age was 23, and three-quarters were married; socioeconomic status was low. At enrollment, the two groups had similar HIV-related laboratory test results; their pregnancy, labor, delivery and neonatal characteristics also were the same. HIV status was available for 83% of infants at the end of the study, and mortality data for 93%.

Women in the formula-feeding group were classified as having complied with the feeding method if they gave their infants only formula; those in the breastfeeding group were considered to have complied if they nursed at all. The level of compliance was significantly lower in the formula-feeding group (70%) than in the breastfeeding group (96%). The median duration of breastfeeding was 17 months.

In all, 92 infants acquired HIV by the age of two--61 who were breastfed and 31 who were formula-fed. The cumulative probability of infection was 37% among breastfed babies and 21% among those given formula. Because the two groups of women were comparable at entry to the study, the investigators explain, this difference in infection rates should be attributable to the feeding method. Furthermore, the 16-point difference in rates suggests that 44% of infections in the breastfeeding group were transmitted through breast milk. However, the researchers also point out that since a considerable proportion of women in the formula-feeding group did not comply with the method (that is, they gave their baby breast milk at some time), the results underestimate the role of breast milk in transmitting HIV.

By comparing the two groups' cumulative probabilities of infection at various points throughout the study period, the investigators discovered that most breastfeeding-related HIV transmission occurred shortly after birth. While the probability of infection was not significantly different between the two groups at birth, it was 10 percentage points higher among breastfed than among formula-fed babies at six weeks of age; this difference represents 63% of the cumulative 16-point difference at two years. Three-quarters of the cumulative difference was accounted for by six months of age and 87% by one year.

Roughly one in five infants in each group died during the study period; mortality rates did not differ at any point over the two years. By contrast, the proportion of babies who were alive and free of infection at age two was significantly higher in the formula-fed group (70%) than in the breastfed group (58%).

While the findings suggest that the exclusive use of formula could substantially reduce the rate of mother-to-infant HIV transmission, the researchers note that "risks associated with formula are community specific....In developing country communities in which clean water and formula-feeding knowledge are limited, the balance of risks and benefits could be shifted." Furthermore, they observe that most HIV-infected women in Sub-Saharan Africa cannot afford infant formula. "The current priority," they conclude, "is to...make interventions to prevent infant HIV-1 infection widely available." --D. Hollander


1. Nduati R et al., Effect of breastfeeding and formula feeding on transmission of HIV-1: a randomized clinical trial, Journal of the American Medical Association, 2000, 283(9):1167-1174.