HIV Transmission Probabilities among MSM
Lydia N. Drumright, UC San Diego
The probability of HIV transmission per coital act has been estimated among heterosexuals based on transmission between monogamous, HIV-discordant partners in a number of different populations. However, few studies have examined estimates of the transmission probability per anal sexual contact among men who have sex with men (MSM) and estimates were not based on monogamous or phylogenetically linked partner pairs in which transmission between the partners could be established. Transmission probabilities of HIV may differ by type of sexual activity, therefore it cannot be assumed that the transmission probabilities for vaginal and anal intercourse are equivalent. Additionally, it is important to explore the contribution of cofactors that may increase the likelihood of transmission such as, stage of infection or viral load, coinfection with other sexually transmitted infections (STIs), and behavioral factors, in increasing the probability of transmission among MSM. We propose to estimate the probability of HIV transmission per episode of unprotected anal intercourse (UAI) among pairs of MSM in which either one individual is infected or both partners are infected, and viral phylogenetic analysis supports epidemiological linkage between the partners. We will compare estimates of transmission probability by stage of infection, by viral load, and by the presence of herpes simplex type 2 (HSV-2) infection. Biological and behavioral data will be collected on newly recruited, recently-HIV infected MSM participating in a well established, NIH-funded study of acute and early HIV infection and their their sexual partners. Among HIV-infected individuals we will determine viral load and stage of infection. All participants will receive testing for HIV-1, syphilis, gonorrhea, chlamydia, and herpes simplex virus type-2 (HSV-2) and will be asked to complete a computer assisted survey instrument (CASI) regarding number and type of sexual acts with a given partner, partnership dynamics, and illicit substance use. Transmission probabilities will be examined by determining if HIV transmission has occurred within a partnership and will be stratified by stage of infection, viral load, and STI infection within each member of the partnership. All data on number of sexual acts within partnerships will be determined by participant response to CASI. Discrete time Cox models will be used to provide estimates of transmission probability per UAI episode. Additionally we will explore factors that predict the number of partners, and number of acts of unprotected anal intercourse, including use of methamphetamine and other substances and patterns of partnering. Data will be collected by participant response to CASI. Zero- truncated negative binomial and Cox proportional hazard models will be used to analyze these data. Three different outcomes will be addressed including number of sexual partners reported with the time period of infection (based on estimated date of infection); number of sexual acts reported with a given partner (including sexual positioning and use of condoms); and duration of partnership. The predictors for these models will include use of methamphetamine or other substances, concurrency, and partner type. These estimates of HIV transmission probabilities among MSM will result in a greater understanding HIV transmission in California and will be helpful in planning vaccine and microbicide intervention studies.