Maternal-newborn human immunodeficiency virus infection in Harlem

To determine the prevalence of human immunodeficiency virus type 1 (HIV-1) infection and its association with illicit drug use for mothers being delivered of infants at an inner-city municipal hospital.

We anonymously tested the umbilical cord blood for HIV-1 antibody of 98.1% (2971/3028) of singleton infants with birth weight greater than 500 g born during 1989 and linked the results to a maternal-infant database from which all identifying information had been removed.

Overall, HIV-1 seroprevalence was 3.3% (99/2971). Among HIV-1-seropositive mothers, 79% (78/99) gave no history of ever using injected drugs. Seropositivity for HIV-1 was independently associated with history of maternal cocaine use during pregnancy (odds ratio, 3.55; 95% confidence interval, 2.18, 5.78), history of ever using injected drugs (odds ratio, 6.02; 95% confidence interval, 3.14, 11.6), positive serologic test result for syphilis during pregnancy (odds ratio, 3.37; 95% confidence interval, 1.94, 5.88), and increasing maternal age per year (odds ratio, 1.04; 95% confidence interval, 1.00, 1.09). Voluntary testing programs failed to identify 71% (70/99) of all HIV-1-infected women. Infants placed into foster care were eight times more likely to be HIV-1 seropositive than those discharged to their mothers.

Most HIV-1-infected mothers seem to have acquired the infection via heterosexual transmission rather than via injected drug use. Associations of maternal HIV-1 infection with cocaine use, syphilis, and increasing age probably operate through behaviors that increase maternal risk of exposure to an HIV-1-infected sexual partner or, in the case of syphilis, also through biologic factors that may predispose to HIV-1 transmission. The failure of voluntary testing to identify most HIV-1-infected mothers provides a strong rationale for routine HIV-1 testing during pregnancy and in the newborn period.