The impact of HIV-1 infection on mortality in children under 5 years of age in sub-Saharan Africa

OBJECTIVE: To estimate the effects of the HIV-1 epidemic on mortality in children under 5 years of age in urban and rural populations in eastern and central, and southern Africa. METHODS: A lifetable method that allows for the effects of competing causes (i.e., mortality due to HIV and other causes) was used to estimate mortality. Our calculations used published and unpublished data on HIV-1 infection in African adults and children (incidence and prevalence, vertical transmissions, transmission by blood transfusion and natural history), and typical baseline fertility and child mortality data. The results were applied to model rural and urban populations to explore the effects of parameters such as mortality in HIV-1-infected children, fertility in infected mothers and overall population growth. RESULTS: We estimate that child mortality will rise substantially because of the prevalence of HIV-1 in urban areas. There will be little difference in the absolute levels of increase in mortality between areas in eastern and central, and southern Africa with similar levels of HIV infection; however, in relative terms the effect will be more noticeable in southern Africa because of the lower baseline mortality. Towns with severe epidemics (30% adult seroprevalence) might experience a rise in child mortality of one-third in eastern and central Africa and three-quarters in southern Africa. This will cancel or reverse existing advantages in urban over rural levels of child mortality and this effect will be more pronounced in southern Africa. The exact impact of HIV-1 will vary according to mortality among HIV-1-infected children and to fertility among infected women. However, changes in age structure and population growth have relatively little impact on mortality. CONCLUSIONS: There are likely to be substantial increases in child mortality in sub-Saharan Africa as a result of HIV-1 infection. The main determinant of childhood infection is the scale of the epidemic among adults. Increases in mortality will depend on local adult seroprevalence but are hard to predict precisely because of possible variation in death rates among HIV-1-infected children. In rural areas with low seroprevalence other diseases will remain the main cause of mortality. However, in urban areas families and health services will have to face considerably increased demands from ill and dying children.