Dynamic partnerships and HIV transmissions by stage

Effectiveness of some control programs of human immunodeficiency virus (HIV) transmission depends on what proportion of new infections are attributable to a particular stage of HIV infection. Most model analyses for the transmission of HIV by stage have neglected real-world details such as sexual partnerships, risk fluctuation and sexual role segregation. To examine the effects of those real-world details on the transmission of HIV by stage, we constructed various models of HIV transmission using both individual-based and deterministic compartmental approaches.;Transmissions through long-term sexual partnerships generate local network structure in which infected individuals are connected to fewer susceptible partners compared with the population average. The increasing depletion of susceptible partners around infected individuals monotonically decreases basic reproductive ratio, R0 , and endemic prevalence of HIV infection with increasing partnership duration. The role of primary HIV infection (PHI), i.e., fractional contribution to R0 of PHI or the fraction of transmissions from PHI at endemic phase, has a U-shaped relationship with partnership duration. It drops in shorter partnerships, but rises in longer partnerships. This pattern is determined by the difference in relative depletion of susceptible partners by stage of infection. As the risk of transmission is made increasingly different by type of sex act while keeping the total population risk unchanged, endemic prevalence and the role of PHI become smaller. The decreased role of PHI is only observed when partnerships are long lasting. If individuals fluctuate between high- and low-risk phases, susceptible individuals are replenished from low- to high-risk phase and infection is spread from high- to low-risk phase. This increases endemic prevalence in the overall population. Risk fluctuation also causes individuals with PHI to be more likely to be in high-risk phase, which increases the role of PHI.;Realistic details like sexual partnerships, sexual role segregation and risk fluctuation can strongly influence the transmission of HIV and do so differentially by stage of HIV infection. Model analyses intended to evaluate control program options or assess the role of a particular stage of infection need to take these details into account.