Background: Relatively little information exists regarding either the risk or timing of breastfeeding transmission, or potential risk factors for such transmission. We analyzed individual patient (pt) data from randomized, placebo-controlled clinical trials to 1) estimate the contribution of late postnatal transmission (LPT) of HIV through breastfeeding to the overall risk of mother-to-child transmission of HIV; 2) to characterize the timing of breastfeeding transmission; and 3) to identify determinants of LPT.

Methods: Eligible trials were conducted among populations in which breastfeeding was common; included at least 2 scheduled HIV diagnostic tests by 3 months (mos) of age and, if follow-up continued, 2 additional tests between 3 and 12 mos, and provided regular assessment of infant feeding modality. Uniform definitions of HIV infection status of children were applied across trials. Data for children born before January 2000 were analyzed.

Results: Nine (9) of 10 eligible trials agreed to participate. Of 5,327 deliveries with available data, there were 4,343 children who were breastfed singletons with HIV diagnostic testing. Of these, 999 (23%) were definitively infected, of whom 545 could be categorized according to the timing of infection (314 children [58%] with infections before 4 weeks of age and 231 children [42%] with LPT). LPT occurred throughout the period of breastfeeding, and the estimated hazard function for time to LPT was roughly constant. The risk of LPT was lower for girls (relative risk [RR] for girls vs boys = 0.7, 95% confidence interval [CI] 0.5-0.9), and higher for children whose mothers had lower CD4⁺ lymphocyte counts (cells/microliter) (CD4⁺) measured around the time of delivery (CD4⁺ < 200: RR = 7.7 [95% CI: 4.7–12.5]; CD4⁺ = 200–499: RR = 3.7 [95% CI: 2.4–5.5]; CD4⁺ = 500+ [reference]). The gender difference was not explained by differences in duration of BF.

Conclusions: LPT represents a significant proportion of overall HIV transmission and occurs throughout breastfeeding. LPT risk varies according to maternal CD4⁺, with lower CD4⁺ associated with higher risk, reflecting more advanced maternal HIV disease and, possibly, higher viral loads in maternal blood and breast milk. LPT risk also varies by the child’s gender, with boys experiencing a higher risk of LPT. The association of gender with LPT of HIV should prompt further research into the potential underlying mechanism(s), both biological and cultural.