Abstract

BACKGROUND:

Defining the effect of antiretroviral regimens on breast milk HIV type-1 (HIV-1) levels is useful to inform the rational design of strategies to decrease perinatal HIV-1 transmission.

METHODS:

Pregnant HIV-1 seropositive women (CD4+ T-cell count >250 and <500 cells/mm3) electing to breastfeed in Nairobi, Kenya were randomized to highly active antiretroviral therapy (HAART; zidovudine [ZDV], lamivudine and nevirapine [NVP]) during pregnancy and 6 months post-partum or to short-course ZDV plus single-dose NVP (ZDV/NVP). Breast milk samples were collected two to three times per week in the first month post-partum.

RESULTS:

Between November 2003 and April 2006, 444 breast milk samples were collected from 58 randomized women during the first month after delivery. Between 3 and 14 days post-partum, women in the HAART and ZDV/NVP arms had a similar prevalence of undetectable breast milk HIV-1 RNA. From 15 to 28 days post-partum, women in the HAART arm had significantly lower levels of breast milk HIV-1 RNA than women randomized to ZDV/NVP (1.7 log10 copies/ml [limit of detection] versus >2.10 log10 copies/ml, P<0.001). In contrast to breast milk HIV-1 RNA, suppression of plasma HIV-1 RNA during the neonatal period was consistently several log10 greater in the HAART arm compared with the ZDV/NVP arm.

CONCLUSIONS:

HAART resulted in lower breast milk HIV-1 RNA than ZDV/NVP; however, ZDV/NVP yielded comparable breast milk HIV-1 RNA levels in the first 2 weeks post-partum. Breast milk HIV-1 RNA remained suppressed in the ZDV/NVP arm despite increased plasma HIV-1 levels, which might reflect local drug effects or compartmentalization.