Pregnant women infected with HIV-1 were enrolled in a prospective mother-to-infant transmission study from 1992 through 1994 in Bangkok. In participating hospitals, voluntary HIV testing was routinely offered at the beginning of antenatal care and again in the middle of the third trimester of pregnancy. Women who seroconverted to HIV during pregnancy were compared with women who had tested positive on their first antenatal test. Maternal HIV RNA levels were determined during pregnancy, at delivery, and postpartum using RNA polymerase chain reaction (PCR), and infection status in infants was determined by DNA PCR. No infants were breast-fed, but prophylactic antiretroviral therapy was not yet used in Thailand to prevent transmission from mother to infant. Among enrolled women, 16 who seroconverted during pregnancy and 279 who were HIV-1-seropositive at their first antenatal test gave birth. Median plasma RNA levels at delivery were similar for the two groups (17,505 and 20,845 copies/ml, respectively; p =.8). Two (13.3%) of 15 infants born to women who seroconverted and 66 (24.8%) of 266 infants born to previously HIV-seropositive women were infected with HIV (p =.5). There was no increased risk for mother-to-infant HIV transmission and no significant difference in viral load at delivery between HIV-infected women who seroconverted to HIV during pregnancy and those who were HIV-seropositive when first tested.