On average, each time a monogamous, heterosexual couple in which one partner is HIV-positive has intercourse, the probability that the virus will be transmitted to the uninfected partner is 0.11%, according to an analysis of data from rural Rakai, Uganda.1 The probability rises significantly as the infected partner's viral load (the amount of virus in the blood) increases, and it is elevated if the HIV-positive partner has genital ulcers. Although some researchers have hypothesized that viral subtype is a factor in HIV infectivity, no such relationship was evident in this population.

The data on which the analysis was based came from an AIDS prevention study conducted in 1994-1998, for which researchers gathered information from more than 15,000 men and women aged 15-59. At follow-up visits every 10 months, participants were asked to provide blood and urine samples, and women were asked for a self-collected vaginal swab; all samples were tested for HIV and a variety of other sexually transmitted diseases (STDs). In addition, participants completed interviews in which they were asked about their background characteristics and sexual behavior, including their number of partners, use of condoms and usual frequency of intercourse with each partner.

At the end of the study, the investigators identified couples in which one partner had been HIV-positive and the other HIV-negative at enrollment. To assess the probability of infection associated with various factors, they used data from the 174 couples in which both partners reported that they were monogamous and that the uninfected partner was monogamous throughout follow-up.

In 77 couples, the woman had been infected with HIV at enrollment, and in 97, the male had been the infected partner. The HIV-positive men and women had a median age of 29 and a median viral load of 12,476 copies per mL. By the end of the study, 27% of infected women and 18% of infected men had transmitted the virus to their partners. Participants reported having intercourse an average of 8.9 times per month; a high level of agreement between partners' reports supports the reliability of this information.

Using Poisson regression, the investigators estimated rate ratios of HIV transmission, controlling for the HIV-infected partner's sex, age, viral load, and STD symptoms and diagnoses. These calculations revealed no significant difference between men and women in the risk of transmission. While rate ratios also did not vary significantly by age, results of chi-square testing showed a significant trend toward a lower risk of transmission as an infected partner's age increased.

The risk of transmission climbed sharply and steadily as viral load increased: Compared with men and women whose viral load was less than 1,700 copies per mL, those with a viral load of 1,700-12,499 copies per mL were 16 times as likely to transmit the virus (rate ratio, 16.1); the rate ratio rose to 27.7 for individuals with a viral load of more than 38,500 copies per mL. An infected individual with genital ulcer disease was at increased risk of transmitting the virus (2.6), but no other STD-related factors were associated with the risk.

Overall, the probability of HIV transmission was 0.11% per act of intercourse. The probability was higher for individuals younger than 30 (0.13-0.17%) than for those aged 30 or older (0.06-0.09%), and it rose as viral load increased (from 0.01% at the lowest level to 0.23% at the highest). The age pattern was the same regardless of viral load, and the viral load pattern was the same regardless of age. Similarly, the probability of transmission was higher among those with genital ulcers (0.41%) than among those without (0.11%), and this pattern held across levels of viral load. Infected women appeared to be more likely than infected men to transmit the virus (0.13% vs. 0.09%); the researchers note that while this difference was not statistically significant, it is consistent with incidence data from Rakai and with findings on transmission in other developing countries. No difference in the probability of transmission was detected between the two virus subtypes that are responsible for the HIV epidemic in Uganda.

While previous studies have examined the probability of HIV transmission per act of intercourse in the United States, Europe and Thailand, the researchers note that theirs was the first to do so in Sub-Saharan Africa, and their overall finding was similar to those from the earlier work. Given this similarity and the fact that different strains of HIV are prevalent in these various settings, the investigators conclude that the rapid spread of the virus in parts of Africa is unlikely to be attributable to characteristics of a particular subtype that facilitate transmission. Commenting on the association they found between the probability of transmission and viral load (which is consistent with results of other research indicating increased transmission among people with compromised immune systems), they suggest that interventions aimed at lowering viral load could reduce transmission.--D. Hollander


1. Gray RH et al., Probability of HIV-1 transmission per coital act in monogamous, heterosexual, HIV-1-discordant couples in Rakai, Uganda, Lancet, 2001, 357(9263): 1149-1153.