DECREASE OF HIV PREVALENCE OVER TIME AMONG THE MALE POPULATION OF ORANGE FARM, SOUTH AFRICA

Sunday, 29th of July 2012

• DECREASE OF HIV PREVALENCE OVER TIME AMONG THE MALE POPULATION OF ORANGE FARM, SOUTH AFRICA, FOLLOWING A CIRCUMCISION ROLL-OUT

Abstract, also at http://pag.aids2012.org/abstracts.aspx?aid=3903

Presented by Bertran Auvert (France).

B. Auvert¹, D. Taljaard², R. Sitta³, D. Reach², P. Lissouba³, J. Bouscaillou³, B. Singh⁴, D. Shabangu², C. Nhlapo⁵, J. Otchere- Darko², T. Mashigo², R. Taljaard⁶, G. Phatedi², M. Tsepe², M. Chakela², A. Mkhwanazi², P. Ntshangase², G. Peytavin⁷, S. Billy⁵, A. Puren⁴, D. Lewis<sup>4

1</sup>University of Versailles, Versailles, France, ²CHAPS, Johannesburg, South Africa, ³Inserm U1018-CESP, Villejuif, France, ⁴NICD-NHLS, Johannesburg, South Africa, ⁵SFH, Johannesburg, South Africa, ⁶Progressus, Johannesburg, South Africa, ⁷Hopital Bichat - Claude-Bernard, Paris, France

Background: The effect of the roll-out of voluntary medical male circumcision (VMMC) in Southern and Eastern Africa on HIV prevalence is unknown. Three years after the beginning of the Orange Farm (South Africa) VMMC roll-out (ANRS-12126), the project´s impact on HIV prevalence over time was assessed.

Methods: Two cross-sectional surveys were conducted, one in 2007-2008 (n=1971), before the beginning of the roll-out, and one in 2010-2011 (n=3268). The response rates exceeded 80%. Male residents aged 15 to 49 were randomly sampled, interviewed, assessed for circumcision status, counselled and tested for HIV. Blood samples were tested for HIV and for recent HIV infection using the BED assay. Prevalence ratios (PR) were calculated using multivariate (aPR) and propensity score weighted (wPR) Poisson regression models to assess the association of HIV prevalence and recent HIV infections with time and circumcision status. Covariates comprised age group, ethnic group, religion, having children, alcohol consumption, education level, age at first sexual intercourse, marital status and occupation.

Results: In three years, male circumcision prevalence standardized on age increased from 17.0% to 53.9%. In the 2007-2008 and 2010-2011 surveys, the propensity weighted effect of circumcision status on HIV prevalence was wPR=0.37 (95%CI: 0.19−0.58) and wPR=0.48 (95%CI: 0.35−0.65), respectively. HIV prevalence standardized on age decreased from 12.5% to 9.3%, aPR= 0.74 (95%CI: 0.60−0.91), among participants aged 15 to 49, and from 6.2% to 4.2%, aPR= 0.64 (95%CI: 0.49−0.85), among participants aged 15 to 29. Propensity analyses showed that the intervention avoided 536 (95%CI: 135 −1318) HIV infections in 2011 among the 52,000 adult men living in Orange Farm.

Conclusions: This study shows for the first time that the roll-out of VMMC in Africa can, if successfully promoted, lead to a decrease of HIV prevalence over time, detectable three years after its onset.