Paper # 36
Longer-term Effects of Male Circumcision on HIV Incidence and Risk Behaviors during Post-trial Surveillance in Rakai, Uganda
Xiangrong Kong*1, G Kigozi2, V Ssempija2, D Serwadda2,3, F Nalugoda2, F Makumbi2,3, T Lutalo2, S Watya2,3, M Wawer1,2, and R Gray1,2
1Johns Hopkins Univ Bloomberg Sch of Publ Hlth, Baltimore, MD, US; 2Rakai Hlth Sci Prgm, Entebbe, Uganda; and 3Makerere Univ, Kampala, Uganda
Background: Three trials demonstrate the efficacy of
male circumcision for HIV prevention over 2 years, but longer-term
effectiveness and impact on risk behaviors are unknown.
Methods: A randomized trial of male circumcision enrolled
4996 HIV– men aged 15 to 49 in Rakai, Uganda. Following trial
closure for efficacy, we offered male circumcision to control participants and
maintained post-trial surveillance of intervention and control arm men up to
2.28 years. HIV incidence per 100 person-years was assessed by an as-treated
analysis. Sociodemographic and sexual risk characteristics of controls opting
for or declining male circumcision, and within-individual changes in behaviors
(∆) were assessed during the first post-trial follow-up using Wilcoxon
rank test or linear models for repeated categorical responses.
Results: Of control arm participants, 80.4% accepted
male circumcision. During ~2.28 years of post-trial surveillance, overall HIV
incidence was 0.54/100 person-years in circumcised men and 1.66/100 person-years
in uncircumcised men (effectiveness 67%, 95%CI 41 to 82%). Post-trial HIV
incidence was 0.55/100 person-years in circumcised men. In analyses restricted
to control arm participants during post-trial surveillance, HIV incidence was
0.53/100 person-years in circumcised controls and 1.65/100 person-years in
uncircumcised controls (effectiveness 68%, 95%CI 35 to 85%). Controls accepting
or declining circumcision were comparable in age, education, marital status,
number of sex partners, and condom and alcohol use with sex. During the
post-trial follow-up, both groups reported no change in number of non-marital
sex partners, decreased alcohol use with sex (∆ –9.3%, p
<0.001), decreased consistent (∆ –4.3%, p <0.001), and any
condom use(∆ –6.2%, p <0.001). However, the risk behavior
changes were not statistically different between circumcised (n = 1321) and
uncircumcised (n = 372) men.
Conclusions: The effectiveness of male circumcision during
a post-trial observational study was comparable to the efficacy of circumcision
for HIV prevention during a randomized trial. Post-trial male circumcision acceptance
was high among controls, with no evidence of self-selection. Condom use
declined in controls both opting for and declining male circumcision, however,
the changes were similar between groups and there was no evidence of risk
compensation associated with circumcision.
|
