Background: There are limited data on characteristics of acceptors of safe male circumcision (SMC), risk compensation and effectiveness of SMC service programs. We compared baseline characteristics of SMC acceptors and non-acceptors, determined the effectiveness of SMC and compared trends in sexual behaviors of the two groups using data on men aged 15-49 years enrolled in the Rakai community cohort study in Uganda.
Methods: 1192 non-Muslim HIV-negative SMC acceptors were compared to a stratified random sample of 2384 uncircumcised men. Baseline behaviors and trends over time were compared using multivariable modified Poisson with generalized estimating equations. HIV incidence rates between the groups were compared using the incidence rate ratio (IRR) from a multivariable Poisson regression model.
Results: Overall SMC acceptors were younger, less likely to be currently or previously married, and had higher education attainment. Among sexually active men, baseline sexual behaviors were comparable between the groups. However SMC acceptors had 26 percent higher prevalence of genital ulcers compared to non-acceptors (p=0.025). After circumcision, the rate of increase in prevalence of sexual activity was 2.6 percent higher among SMC acceptors (p<0.001) compared to non-acceptors. On stratification by age, the difference was 3.2 percent higher, p=0.08 among youths (15-24) but no difference was seen above 24 years. The prevalence of sexual activity with women in higher risk occupations (bar attendants, alcohol brewers, restaurant workers, itinerant traders, fisher folk, housemaids), increased by 10.2 percent per year among SMC acceptors (p=0.007) but no change occurred among uncircumcised men. Trends in other sexual behaviors were similar between the groups. HIV incidence among SMC acceptors was 0.61/100 person years and 1.11/ 100 person years among non-acceptors (adj. IRR=0.50, p=0.05, 95 percent CI=0.25-1.01). Conclusions: The higher prevalence of genital ulcers among sexually active SMC acceptors suggests that higher risk sexually active men self-selected to receive SMC. The suggestion of faster increase in sexual activity among circumcised youths and the increase in partnerships with higher risk women suggest possible behavioral disinhibition and need to be investigated in other settings. Though these behaviors, did not attenuate the effectiveness of SMC, there is need to add avoidance of high risk partners to the current SMC messaging.