The risk of HIV infection through breastfeeding appears to be greatest in the first few months after birth, and is lower among infants who are fed breast milk exclusively than among those who receive breast milk with supplements. In a recent South African study, infants who were exclusively breastfed were significantly less likely to become infected in the first three months than were those who also received supplemental foods and liquids, but had a risk similar to that among infants who received only infant formula.¹ A second study, conducted in Malawi, found that the risk of HIV transmission fell from 0.7% per month among infants aged 1-5 months to 0.3% per month among those aged 12-17 months.² Moreover, mothers aged 25 or older and those with four or more births were less likely to transmit the virus through breastfeeding than were other women.

The South African Study

The data for analysis came from 549 mother-infant pairs who participated in a clinical trial investigating the effect of vitamin A taken during pregnancy on vertical transmission. HIV-infected pregnant women were recruited for the intervention between July 1995 and April 1998 at two prenatal clinics in Durban, South Africa.

After receiving counseling during their prenatal visits, women chose to feed their infant breast milk alone, breast milk supplemented with solids or liquids, or formula alone. The infants born to these women were tested for HIV on their first day of life; 6% of the 549 babies tested positive at that time. They were tested again at one and six weeks and at three months.

The cumulative probability of a positive HIV test by three months of age was assessed among the three infant-feeding subgroups--156 babies who were never breastfed, 103 who were still being breastfed exclusively at three months and 288 who had been given other foods besides breast milk--using Kaplan-Meier life-table methods. (Data on the type of breastfeeding were missing for two babies.)

Of the 549 babies born to HIV-infected mothers, an estimated 19% of those who were never breastfed were HIV positive by three months of age, compared with 21% of ever-breastfed infants, a nonsignificant difference. However, the estimated proportion infected by that age was significantly lower among babies who were breastfed exclusively than among those who received both breast milk and other foods (15% vs. 24%). When the 32 infants who tested positive on their first day of life were excluded from the analysis, the proportion testing positive at three months was significantly lower among babies who were breastfed exclusively (8%) than among those who received mixed feeding (20%), but did not differ significantly from the proportion among those who were never breastfed (13%).

There were no significant differences in the estimated proportions testing positive at three months according to four maternal background characteristics--maternal education, maternal employment, electricity in the home and source of water. However, the risk of infection was elevated among babies born to women who had a low CD4 count at enrollment (adjusted hazard ratio of 2.1), those who tested positive for syphilis (1.8) and those who delivered before term (2.0). When the effects of maternal immune status, syphilis test results and preterm delivery were accounted for, infants who had been exclusively breastfed for three months were significantly less likely to be infected than were those who had received mixed feeding (0.5).

The researchers caution that their results are limited by their inability to validate self-reported feeding choices and by possible differences in unmeasurable variables that may influence infant feeding choices. Nevertheless, they say, the large differences in the risk of infection by three months between babies who were exclusively breastfed and those who received mixed feedings probably "reflect genuine differences in postnatal transmission due to feeding practices." According to the researchers, contaminants or allergens in supplemental foods could undermine breast milk's benefits to the immune system; contaminants might injure membranes lining the gastrointestinal tract, thus facilitating infection with HIV.

The Malawian Study

The study, conducted in a tertiary care hospital in Blantyre, Malawi, between 1994 and 1997, examined the HIV status of 672 infants born to infected women who had made at least two follow-up visits and were still breastfeeding at the time of the second visit. Only babies who were uninfected at the first visit--when they were about six weeks of age--were eligible for the study. After the second follow-up, at three months postpartum, the mother-infant pairs returned at three-month intervals until the infants were two years old or until they were weaned. Subsequent tests for HIV infection were performed at six-month intervals.

By the end of the 24-month follow-up, 7% of the 672 infants had become infected; no infant seroconverted after being weaned. The HIV-infection rates in the first two years of life declined significantly over time, falling from 0.7% per month in months 1-5 to 0.6% per month in months 6-11, and to only 0.3% when the babies were 12-17 months old. The cumulative risk of infection with HIV-1 among babies breastfed for more than one month was 4% after five months of life, 7% after 11 months, 9% after 17 months and 10% after 23 months.

According to the univariate risk-factor analysis, parity was the only one of six maternal factors examined (age, progression to symptomatic disease or death, mode of delivery, parity, infant birthweight and breast problems such as cracked or bleeding nipples) that significantly affected the risk of HIV infection. Women who had already had at least four children had a significantly lower risk of transmitting the virus than did women who had had 1-3 children (risk ratio of 0.39). A stepwise regression analysis found a stronger protective effect against infection for high parity (risk ratio of 0.23), as well as a significantly lower risk of transmission among women aged 25 or older (0.44).

The investigators acknowledge that their study is limited by the lack of data on the women's immunologic and virologic status, which both influence the rate of transmission. They also note that they could not assess risk before the baby completed one month of life, so their data, by definition, underestimate the overall risk of transmission through breastfeeding.

The researchers suggest that the peak in infection rates within the first few months of breastfeeding might be explained by the concentration of HIV-infected cells in colostrum and early breast milk, and by the newborn's especially immature immune system. They also hypothesize that the unexpected higher risk of transmission among lower parity women and younger women might reflect these women's relative inexperience with breastfeeding and their greater likelihood of subclinical mastitis or cracked nipples. As no other obvious groups of women emerged as being at elevated risk of transmission, the researchers suggest that efforts to instruct young and inexperienced mothers about breastfeeding might be useful.

Because of the difficulty in balancing the risk of HIV transmission with the benefits of breastfeeding, the researchers conclude that "recommendations may be most usefully made at the level of the individual mother," especially because women in the developing world have such limited access to safe alternatives to breast milk.--L. Remez

REFERENCES

1. Coutsoudis A et al., Influence of infant-feeding patterns on early mother-to-child transmission of HIV-1 in Durban, South Africa: a prospective cohort study, Lancet, 1999, 354(9177):471-476.

2. Miotti PG et al., HIV transmission through breastfeeding: a study in Malawi, Journal of the American Medical Association, 1999, 282(8):744-749.