In this study we have used a mathematical projections model and HIV-1 seroprevalence data from samples of pregnant women to estimate infant HIV-1-infection prevalence and HIV-1-attributable mortality among children. This paper presents the results of samples from six developing country capitals and Martinique. The estimated impact of HIV-1 infection on overall infant mortality is low in regions where maternal HIV-1 seroprevalence is low and overall infant mortality is high, but substantial in locations where either maternal HIV-1 seroprevalence is high, or where maternal HIV-1 seroprevalence is moderate and overall infant mortality is low. The estimates for child (ages 1-4 years) and under-5 mortality suggest that the impact of HIV-1 infection on overall mortality in children may exceed the impact on infant mortality. The recent gains made in ensuring child survival are likely to be increasingly reversed in regions where HIV-1 infection is being transmitted in a substantial proportion of pregnancies and births.
PIP: A mathematical projections model was applied to human immunodeficiency virus (HIV)-1 data from samples of pregnant women in 6 capital cities in sub-Saharan Africa and the Caribbean as well as Martinique in order to estimate HIV-1 prevalence and HIV-attributable child mortality. Aside from Martinique, which was selected as a control, the 6 capital city samples were all from developing countries with high overall levels of infant and child mortality and low gross national products. The infant HIV-1 infection prevalence was arrived at by multiplying maternal seroprevalence rates in the capital city sample by current estimates of the vertical transmission rate (currently set at 24-39%). This methodology resulted in the following estimated percentages of infant HIV-1 infection prevalence: Martinique, 0.0.4%; Mozambique, 0.0-0.3%; Ivory Coast, 0.2-2%; Zaire, 1.8-3.5%; Haiti, 1.7-3.8%; Zambia, 2.5-5.0%; and Uganda, 4.5-11.5%. These prevalences were then multiplied by the probability of an HIV-infected liveborn dying at under 1 year and under 5 years of age. The estimates of under-5 years HIV-1-attributable mortality were as follows: Martinique, 1-3/1000 live births; MOZAMBIQUE, 0-2/1000; Ivory Coast, 1-17/1000; Zaire, 7-28/1000; Haiti, 7-30/1000; Zambia, 10-40/1000; and Uganda, 18-89/1000. In developing countries with low HIV-1 seroprevalence, children are at greatest risk of death in the 1st year of life. However, the estimates obtained in this study suggest that the impact of HIV-1 infection on child mortality may exceed the impact on infant mortality in regions with high rates of HIV-1 infection and undermine any gains that have been made in the area of child survival.