HIV infection was related both to women's own risk behaviors and to the perceived risk behaviors of their sexual partners. Underlying societal factors related to heterosexual HIV infection, including deferral of marriage, warrant further study.

HIV infection risk among pregnant women in Lima depends largely on their male partners' risk behaviors. Even monogamous women had very large sexual networks.

We studied the feasibility and acceptability of HIV-1 testing using rapid serological assays in pregnant women presenting in late pregnancy without prior HIV-1 testing. Pregnant women presenting to the emergency room at the Instituto Materno-Perinatal in Lima, Peru, were offered on-site rapid HIV-1 testing. Consenting women were provided HIV counseling and had samples taken for rapid detection of antibodies in the blood and oral secretions. Women testing HIV-1 seropositive by one or both assays were offered standard HIV obstetrical care including antiretrovirals. Between October 2000 and August 2001, 3543 women were tested for HIV-1 antibodies through the rapid testing protocol. Twenty-seven women tested positive by one or both assays. Standard EIA results were negative for two of the 27 women and EIA results were not available for an additional three women. Seventeen of the 19 women (89%) for whom delivery information is available received antiretroviral therapy prior to delivery. Eighty percent of the women treated with nevirapine received the drug at least 1 hr prior to delivery. Study personnel reported problems with performing the rapid assays for less than 1% of the samples and the majority of the women did not express a preference for either the oral fluid or blood sample collection. While HIV-1 testing early in pregnancy is preferable, rapid HIV-1 testing in late pregnancy or at delivery is feasible, acceptable to most women, and allows for timely intervention to reduce maternal-child transmission.