of their infection, the opportunity for the early diagnosis and treatment of their infected infants frequently may be missed. Similarly, interventions to prevent postpartum transmission, such as avoidance of breast-feeding, are not available to women who do not know that they are infected with HIV. Such interventions may also not be recommended in regions of the world where risks of HIV transmission are overshadowed by risks for other adverse health outcomes (e.g., diarrheal disease) among infants who are not breast-fed.

In perinatal transmission, a variety of factors, usually associated with latter-stage disease, including the presence of maternal p24 antigenemia and low maternal CD4⁺ lymphocyte counts at the time of conception, correlate with the likelihood of infection of a neonate. Additional risk factors in perinatal transmission include high maternal CD8⁺ T-lymphocyte counts, placental membrane inflammation, and maternal fever. It is likely that many HIV infections in infants are acquired at birth through contact with contaminated blood or secretions. Among twins born to HIV-infected mothers, a higher risk of HIV infection is seen in the firstborn, even for twins delivered by cesarean section, suggesting that factors related to the delivery process affect the risk of infection.

The fact that both the virus donor and the recipient are known in the case of perinatal HIV infection provides potential opportunities for interventions to decrease the risk of viral transmission. Recent studies have shown that antiviral treatment of an HIV-infected mother with zidovudine (AZT) can significantly decrease the likelihood of HIV infection in her offspring (Connor et al., 1994). The availability of an effective intervention to decrease perinatal HIV transmission has increased interest in screening pregnant women for the presence of HIV infection so that perinatal HIV infection may be limited.

In summary, much is known about the various modes of HIV transmission. However, it is not possible to provide accurate estimates of their relative efficiency. It is difficult to identify accurate denominators for the numbers of individuals and encounters in which exposure has occurred; moreover, transmission of the virus depends on factors other than the mode of exposure. The dose of virus transferred (inoculum size), the frequency of exposure, differences in host susceptibility, variations in infectiousness of an infected person over time, and the differences in virulence among HIV isolates, as well as the presence of factors such as particular sexual practices and the presence of sexually transmitted diseases, may all influence the likelihood of transmission.