A 28-week regimen of antiretroviral prophylactic therapy during breast-feeding reduces the risk of mother-to-child HIV transmission for up to 48 weeks, whether the drugs are given to the mother or the infant.1 In a randomized controlled trial conducted in Malawi, the cumulative risk of HIV transmission at 48 weeks was substantially higher in the control group (7%) than in the infant or mother prophylaxis groups (4% each). All mothers were instructed to wean their child by week 28 (far earlier than is usual in Malawi) to coincide with the cessation of antiretroviral therapy; regardless of intervention group, the rate of adverse events among infants (e.g., diarrhea, growth faltering, death) was higher during weeks 29–48 than during the first 28 weeks (1.1 vs. 0.7 per 100 person-weeks), suggesting that early weaning may increase infant morbidity.

When no safe alternatives to breast-feeding are available, the World Health Organization (WHO) recommends that HIV-infected mothers follow an antiretroviral regimen during breast-feeding; this approach is especially critical in low-resource countries. To determine the relative efficacy of antiretroviral prophylaxis when given to the mother or infant, researchers conducted a trial in 2004–2010 among women at a mother-to-child transmission prevention program in Lilongwe. HIV-positive women aged 14 or older were eligible if they were 30 or fewer weeks pregnant, had CD4 counts of at least 250 cells per μl (for those enrolled before July 2006, at least 200 cells per μl) and had never used antiretroviral drugs. In addition, infants had to weigh at least 2,000 g at birth, and neither mother nor infant could have a disorder that precluded antiretroviral treatment.

Mother-infant pairs were randomly assign- ed to one of three conditions: the maternal-antiretroviral group, the infant-nevirapine group and the control group. At the start of the trial, all mothers and their babies received one oral dose of nevirapine, as well as zidovudine and lamivudine to be taken for seven days. Participants in the control group received no further antiretroviral treatment. Women in the maternal-antiretroviral group, but not their infants, continued taking zidovudine, lamivudine and nevirapine, although after February 2005 other drugs were substituted for nevirapine in this group because of toxicity concerns. Babies in the infant-nevirapine group received escalating doses of nevirapine during the intervention phase.

Mothers were advised to breast-feed exclusively for the first 24 weeks and to wean between weeks 24 and 28. Antiretroviral treatment was stopped when the women reported weaning their baby or after 28 weeks, whichever came first. Participants returned for 10 follow-up visits during the first 28 weeks and four additional visits during the next 20 weeks; breast-feeding practices were assessed at most of these visits, and adverse events were identified at follow-up visits and at any interim sick visits. Infants were tested for HIV at birth, as well as at weeks 2, 12, 28 and 48. Throughout the trial, any mother whose CD4 count fell below 250 cells per μl or who developed an AIDS-defining disease was referred for treatment. The primary outcomes of interest were HIV infection in infants at 48 weeks among all infants and among those uninfected at two weeks; secondary outcomes were infant HIV infection or death among all infants and among those alive and uninfected at two weeks. The Kaplan-Meier method was used to estimate the probability of HIV infection or death by infant age. Researchers used log-rank tests and Gray’s test to identify differences among study groups, and the Cox proportional hazard model to estimate hazard ratios for HIV transmission.

In all, 2,369 mother-infant pairs were randomly assigned. Of these, 1,829 pairs were followed to week 28; by week 48, an additional 20% of the pairs in each group had been lost to follow-up. Social and demographic characteristics were similar across all study groups at the start of the intervention and among those still participating by week 28. The median age of the mothers was about 26 years; most women were married (92–93%), and almost one-third had attended more than eight years of school. The median baseline CD4 count for the women in each group ranged from 429 to 442 cells per μl. During the first 24 weeks, exclusive breast-feeding was reported by 88–98% of mothers; by week 28, this figure had dropped below 10%. The proportion of women who reported no breast-feeding was higher in the intervention groups (96% each) than in the control group (88%) at week 32, but by week 36 the differences between groups were no longer significant.

Ninety-three infants became infected with HIV between weeks 2 and 48; of these, 30% became infected after week 28. The risk of infant seroconversion was higher in the control group (7%) than in either of the antiretroviral groups (4% for each). In comparison with the control group, the risk of HIV transmission during weeks 2–48 was 38% lower for the maternal-antiretroviral group and 48% lower for the infant-nevirapine group. After adjustment for maternal factors, the two antiretroviral regimens remained associated with reduced risk of HIV transmission between weeks 2 and 48 (hazard ratios, 0.3–0.5). The risk of HIV transmission or death was also lower in the two antiretroviral groups than in the control group.

Rates of serious adverse events in infants were higher in weeks 29–48 than during the intervention phase, regardless of group assignment (1.1 vs. 0.7 per 100 person-weeks). Diarrhea, malaria, growth faltering, tuberculosis and death all occurred more frequently after 28 weeks than before.

The researchers acknowledge a few limitations: Some increased morbidity is to be expected after 28 weeks, when infants begin to crawl and are introduced to new foods, and the researchers had fewer opportunities to detect morbidity after 28 weeks because follow-up visits were more infrequent during the trial’s later stages. The investigators note that although both mothers and infants receiving antiretroviral prophylaxis had lower rates of postnatal HIV transmission than did controls, infant morbidity and mortality increased after cessation of breast-feeding and antiretroviral therapy; in keeping with current WHO guidelines, they suggest that “continued breast-feeding with [antiretroviral] prophylaxis given for an extended period could improve infant survival.”—L. Melhado


1. Jamieson DJ et al., Maternal and infant antiretroviral regimens to prevent postnatal HIV-1 transmission: 48-week follow-up of the BAN randomized controlled trial, Lancet, 2012, 379(9835):2449–2458.