HIV impact on mother and child mortality in rural Tanzania

HIV affects child mortality directly, due to mother to child transmission, and indirectly, because maternal illness and death has negative consequences for child health. A longitudinal community-based study in Kisesa ward, Mwanza region is used to show the contribution of HIV infection to child mortality. HIV status data of 4,273 mothers, from three rounds of serological testing (1994, 96 and 99) are linked to survival information for 6,049 children born between 1994 and 2001. Impacts of maternal survival and HIV status on child mortality are assessed using hazard analysis with time varying co-variates. 584 child deaths were recorded during 10,002 person-years of observation. Infant mortality among children of HIV positive mothers was 158 per 1,000 compared to 74 for children of HIV negative mothers. By age 5 child mortality risks were 265 and 135 respectively. 51 deaths were observed among women who gave birth, 14 of these were among the 149 mothers known to be HIV positive at parturition. Infant mortality among children whose mothers died was 257 compared to 87 amongst children of surviving mothers. Mortality risks for children whose mother died were as high in the year preceding the mother"s death as in the first year of orphanhood. Statistical analysis showed that the effect of maternal death was independent of maternal HIV status, though numbers were too small to study interactions in children over age 3. After allowing for the effects of age, sex, twinning, birth interval, maternal education and residence, the child death hazard ratio for maternal HIV infection was 2.2 (1.6 - 3.1), the hazard ratio associated with maternal death was 4.6 (2.6 - 8.1). In a population in which HIV prevalence among pregnant women is 4.3%, the fraction of infant mortality attributable to maternal HIV infection is 8.1%. (Author abstract)