Background:  The semen of an HIV-infected man is the most common mode of HIV transmission. HAART often results in an undetectable blood HIV RNA viral load, and it has been suggested that there may be no risk of sexual transmission in this context. However, the influence of effective HAART on HIV levels in semen is unclear and requires further study.         

Methods:  We performed a prospective, longitudinal study of semen and blood HIV RNA levels after HAART initiation, and a cross sectional study in men on long-term effective HAART. Effective HAART was defined as the achievement of an undetectable blood viral load. Anti-retroviral drug concentrations were measured using high-performance liquid chromatography (HPLC) and HIV genotype and sequence analysis were performed using real-time polymerase chain reaction (RT-PCR).

Results:  In 25 participants followed prospectively after HAART initiation, the blood viral load was consistently undetectable by week 16. Isolated semen HIV RNA shedding was detected in 12 of 25 (48%) participants despite effective HAART, and at a high level (>5000 copies/mL) in 4 of 25 (16%). Isolated semen shedding was detected at 19 of 116 (14%) visits with an undetectable blood viral load, and was associated with the pre-therapy semen HIV viral load. Semen isolates did not contain drug resistance mutations, and were infectious in vitro. Isolated semen HIV shedding was not associated with antiretroviral drug concentrations in semen plasma. In 13 participants with much more prolonged suppression of the blood viral load (median; 126 months), isolated semen HIV shedding was detected in 4 of 13 (31%) participants.

Conclusions:  Although effective HAART often eliminated HIV RNA from the semen, isolated HIV semen shedding was common, even after extremely prolonged suppression of blood viral load. Public health messages and policy must be tailored carefully to reflect this reality.