Detection of human immunodeficiency virus type 1 (HIV-1) in breast milk by culture and polymerase chain reaction does not necessarily mean that breastfeeding is a route of transmission, although evidence from several case-reports points in that direction. We undertook a systematic review of published studies meeting criteria that allowed determination of quantitative risk of transmission via breastfeeding. Based on four studies in which mothers acquired HIV-1 postnatally, the estimated risk of transmission is 29% (95% Cl 16-42%). Analysis of five studies showed that when the mother was infected prenatally, the additional risk of transmission through breastfeeding, over and above transmission in utero or during delivery, is 14% (95% Cl 7-22%). Where there are safe alternatives to breastfeeding, universal named testing of pregnant women would provide an opportunity to advise more infected women not to breastfeed and might thereby reduce the number of vertically infected children. Since breastfeeding protects against infant deaths from infectious diseases, breastfeeding is still recommended where infectious diseases are a common cause of death in childhood, despite the additional risk of HIV transmission.
PIP: Epidemiologists reviewed 9 published studies of possible transmission of HIV-1 via breast milk to determine estimates of quantitative risk. They identified 4 studies in which the infants were breast feeding when the mother became infected: 2 retrospective studies in Australia and Kinshasa. 19 seronegative mothers at delivery in Lusaka seroconverted by 12 months after delivery. 3 infants also tested positive for HIV-1. In Kigali, 5 HIV-1 infected children out of 9 mothers who seroconverted after 3 months had an earlier negative polymerase chain reaction. The rates of mother to child transmission for Australia was 3/11 and 1/3 for Kinshasa. The overall risk of HIV-1 transmission via breast milk was 29%. The other 5 studies were birth cohort studies which included infants of mothers infected with HIV-1 at delivery: Europe; Miami, Florida; France; Switzerland; Kinshasa, Zaire; and Australia. The epidemiologists compared the risk of HIV-1 transmission in breast-feeding infants with that of bottle-fed infants. The summary estimate of the additional risk of transmission via breast feeding was 14% with a wide confidence interval of 7-22%. This indicated that 14% more infants were infected with HIV-1 through breast milk over and above the proportion infected in utero or during delivery. The estimates of additional risk ranged from 5% in Miami to 33% in Australia. None of the 10 bottle-fed infants in Kinshasa became infected with HIV-1. Only 1 study had done a multivariate analysis so it was not possible to determine confounding factors. In countries where safe alternatives to breast milk exist, HIV-1 positive mothers should not breast feed their infants. On the other hand, in countries where infectious diseases cause considerable infant death, HIV-1 positive mothers should consider breast feeding.