Mortality in HIV-infected and uninfected children of HIV-infected and uninfected mothers in rural Uganda.

OBJECTIVE: To estimate 2-year mortality rates in HIV-1-infected and uninfected infants born to HIV and HIV mothers.
METHODS: Data are from a prospective study in rural Rakai District, Uganda. Infant HIV status (determined by polymerase chain reaction) was evaluated at 1 to 6 weeks postpartum and during breast-feeding, and maternal HIV viral load and CD4 levels were measured at the postpartum visit. Multivariate Cox proportional hazards models and Kaplan-Meier survival analysis were used to assess survival of infants by maternal and infant HIV status and by quartiles of viral load. Log-rank tests were used to test the equality of survival functions.
RESULTS: Of the 4604 pregnant women, 16.9% were HIV, and the proportion of children infected was 20.9%. Median survival of HIV-infected infants was 23 months. Two-year child mortality rates were 128 of 1000 children born to HIV mothers, 165.5 of 1000 uninfected children born to HIV mothers, and 540.1 of 1000 HIV-infected children (P < 0.0001). Compared with children of HIV mothers, the hazard of child mortality was 2.04 (P < 0.001) if the mother was HIV and 3.78 (P < 0.001) if the infant was also infected. In the adjusted model, the highest quartiles of log10 HIV viral load in infants and mothers were associated with significantly increased hazard of child mortality (hazard ratio [HR] = 8.54 and HR = 2.50, respectively). Maternal CD4 counts <200 cells/mL were also significant predictors of child mortality (HR = 2.61). A total of 67.6% of HIV-infected children with viral loads above the median died by the age of 2 years and are in need of early antiretroviral therapy (ART).
CONCLUSIONS: More than half of HIV-infected infants died at less than 2 years of age. Therefore, ART may need to be initiated earlier in HIV-infected African children.