The purpose of this study is to see if antibiotic drugs given to treat an infection of the uterus during pregnancy can reduce the chances of HIV being passed from an HIV-positive mother to her baby.

A link between bacterial disease of the vagina, premature birth, infection of the uterus during pregnancy, and the passing of HIV from a mother to her baby has been found. Early treatment of these problems may reduce the risk of passing HIV from an HIV-positive mother to her baby.

[Note: As of 02/21/03, enrollment into this study was halted because preliminary data showed that the study antibiotics were not effective in preventing mother-to-child HIV transmission.]

Obstetric risk factors for HIV maternal-child transmission (MCT) include preterm birth, prolonged rupture of the membranes, and chorioamnionitis. Many preterm births are associated with and likely caused by chorioamnionitis. The relationship between bacterial vaginosis, preterm birth, histologic chorioamnionitis, and perinatal transmission of HIV has been consistently demonstrated. Perinatal HIV transmission is more common in preterm infants, and there is now evidence that subclinical chorioamnionitis is a substantial risk factor for MCT. For this study, the primary hypothesis is that early and appropriate treatment of subclinical chorioamnionitis prior to the onset of spontaneous preterm labor, and/or antibiotic treatment during labor, to prevent premature rupture of membrane-associated-chorioamnionitis, will reduce the risk of perinatal HIV transmission.

[Note: As of 02/21/03, enrollment into this study was halted because preliminary data showed that the study antibiotics were not effective in preventing mother-to-child HIV transmission.]

At 20 to 24 weeks, women who are randomized to receive antibiotics receive metronidazole and erythromycin for 7 days. Women randomized to the control group receive identically appearing placebos. With the onset of contractions and/or premature rupture of membranes, study participants will initiate a second oral course of antibiotics consisting of metronidazole and ampicillin or placebo every 4 hours, continuing after delivery until the course is completed. All HIV-infected women and their neonates will be offered the HIVNET 012 nevirapine (NVP) regimen. If the mother accepts the NVP for herself and her baby, she will be given 1 dose of NVP to be taken at onset of labor, and her baby will receive 1 dose of NVP at 72 hours post-birth or discharge, whichever occurs earlier. If the mother refuses NVP or is uninfected, she will receive a matched placebo at the 26- to 30-week visit to preserve participant confidentiality. This study takes place in Blantyre and Lilongwe, Malawi, in Lusaka, Zambia, and in Dar es Salaam, Tanzania.

Inclusion Criteria

• HIV positive.
• 20 to 24 weeks pregnant.
• Willing to take the planned antibiotic treatment.
• Planning to deliver at 1 of the study sites.
• Willing to come back for follow-up visits for 1 year after the baby is born.

Exclusion Criteria

• Have taken antibiotics, except for syphilis or gonorrhea, within the last 2 weeks.
• Are allergic to penicillin, ampicillin, erythromycin, or metronidazole.
• Have major illnesses, such as diabetes, severe kidney or heart disease, or active tuberculosis, which might affect the pregnancy.
• Are having major problems with the pregnancy, such as placenta previa, ruptured membranes, or multiple pregnancy.
• Have a central nervous system disease, such as seizures.
• Are taking anticoagulant drugs.