Background: An understanding of the risk and timing of mother-to-child transmission of HIV-1 in the postnatal period is important for the development of public-health strategies. We aimed to estimate the rate and timing of late postnatal transmission of HIV-1.
Methods: We did an international multicentre pooled analysis of individual data from prospective cohort studies of children followed-up from birth born to HIV-1-infected mothers. We enrolled all uninfected children confirmed by HIV-1-DNA PCR, HIV-1 serology, or both. Late postnatal transmission was taken to have occurred if a child later became infected. We calculated duration of follow-up for non-infected children from the time of negative diagnosis to the date of the last laboratory follow-up, or for infected children to the mid-point between the date of last negative and first positive results. We stratified the analysis for breastfeeding.
Findings: Less than 5% of the 2807 children in four studies from industrialised countries (USA, Switzerland, France, and Europe) were breastfed and no HIV-1 infection was diagnosed. By contrast, late postnatal transmission occurred in 49 (5%) of 902 children in four cohorts from developing countries, in which breastfeeding was the norm (Rwanda [Butare and Kigali], Ivory Coast, Kenya), with an overall estimated risk of 3.2 per 100 child-years of breastfeeding follow-up (95% CI 3.1-3.8), with similar estimates in individual studies (p=0.10). Exact information on timing of infection and duration of breastfeeding was available for 20 of the 49 children with late postnatal transmission. We took transmission to have occurred midway between last negative and first positive HIV-1 tests. If breastfeeding had stopped at age 4 months transmission would have occurred in no infants, and in three if it had stopped at 6 months.
Interpretation: Risk of late postnatal transmission is consistently shown to be substantial for breastfed children born to HIV-1-positive mothers. This risk should be balanced against the effect of early weaning on infant mortality and morbidity and maternal fertility.