The main objective of this one year multidisciplinary research is to estimate the frequency of seropositive men having sex with men (MSM) under antiretroviral therapy (ART) having an undetectable blood viral load (VL) and a detectable VL in semen.

Design: Participants'inclusion (n=150) will be done in 6 HIV hospital departments participating to the research project. To warrant feasibility and diversity of participants needed, the centers will be located in Paris and nearby suburbs. The study design is based on one blood and semen sample taken at day 0 and at day 30. Blood and semen samples will be taken in the enrolment centers. Biological and virological analyses will be performed by the laboratory of microbiology at Necker hospital, on blood and semen samples. Pharmacological analyses are planned in a subsequent study. Socio-behavioral data will be collected through a self-administered questionnaire at day 0 and day 30.

Schedule: Patients' enrolment, collection of biological samples and questionnaires will last 10 months (end of first quarter 2012). Biological and virological analyses will be performed until the end of the first semester 2012. Quantitative and socio-behavioral data will be analyzed during the third quarter 2012. Results will be released at the beginning of year 2013.

The secondary objective is to characterize the biological and behavioral components that might influence VL discordance between blood and semen.

Background: In most cases the semen VL reflects the blood VL. However some seropositive people with an undetectable blood VL have a detectable semen VL and thus might be at risk of HIV transmission in case of unprotected sexual behaviors. Most of the previously published studies were made on rather small numbers of HIV+ heterosexual men, were cross-sectional and did not address factors that might be associated with discordant VL between blood and semen. Few studies have been performed among seropositive MSM who are one of the most at-risk populations for HIV.

Issue: The working hypothesis is that virus present into reservoirs and risky sexual behaviors might influence discordance in VL between blood and semen in MSM. Our hypothesis is that this rate might be higher among MSM than among the general population due to the existence of biological and specific factors that could lead to HIV local replication. The main reasons to address this question in MSM are i) no or very few data available in this population; ii) the possibility to measure in this population the impact of life style determinants on semen VL (high number of sexual partners; risky sexual behaviors, multi-drugs consumption…); and iii) a strong expectation from this community to have new prevention strategies and clear information about the HIV transmission risk.

Outcomes: Quantitative data analysis will allow us to evaluate the proportion of people having a detectable semen VL, its variation according to time. Socio-behavioral data analysis will allow us to describe the sexual behaviors that might influence discordance of the VL between blood and semen in HIV+ MSM. These results will provide information on the HIV residual risk of transmission in treated MSM and to better target people who might be at risk for HIV transmission in case of unprotected sexual behaviors. Describing the socio-behavioral characteristics of these people will help to define adaptated interventions in order to improve HIV prevention in MSM

Inclusion Criteria:

• Age ≥ 18 and
• Able to give written consent
• Covered by French Social Security
• be a men and have sex with men
• be treated with the same ARV treatment for at least 3 months
• blood plasma have a CV undetectable (<50 copies / ml) for at least 6 months
• accept the constraints imposed by the study, that is to say one or more masturbation in hospital

Exclusion Criteria:

• Under protection(saving) of justice