Circumcision reduces men’s HIV risk for up to seven years, according to follow-up data from a randomized trial conducted in Rakai, Uganda.1 In analyses that combined data from the two-year trial with nearly five years of posttrial surveillance data, the incidence of HIV among circumcised men—including those who had the procedure after the trial ended—was 0.5 per 100 person-years, compared with 1.9 per 100 among those who were uncircumcised. This corresponds to a circumcision effectiveness of 74%. In a survival analysis that controlled for men’s background characteristics and sexual behaviors, the adjusted hazard ratio of HIV acquisition among circumcised men was 0.3, and the circumcision effectiveness was 73%.

Although circumcision has been shown to be effective in HIV prevention among men, its long-term effectiveness has not been established. When the Rakai randomized trial was halted in 2006, after analysis demonstrated the efficacy of circumcision in preventing HIV acquisition in a sample of 4,996 men aged 15–49, all trial participants were enrolled in a posttrial surveillance study, and circumcision was offered to uncircumcised participants. Of the 1,602 HIV-negative men who were not circumcised during the trial but attended at least one posttrial follow-up visit, 79% had been circumcised by December 2010 (the cutoff for the present analysis). At each posttrial visit, men were tested for HIV, examined for circumcision status and asked about their sexual behaviors in the preceding year (number of partners, whether they had had a nonmarital partner, consistency of condom use and alcohol use before sex). The posttrial analysis used data from visits over 4.8 years; incidence rate ratios were estimated using Poisson regression, and Kaplan-Meier survival analysis and Cox regression were used to calculate hazard ratios of HIV acquisition. These analyses controlled for men’s social and demographic characteristics (age, marital status and education) at the end of the initial trial interval, allowing the researchers to assess whether self-selection bias occurred.

During the initial two-year trial period, HIV incidence was 0.5 per 100 person-years among circumcised men and 1.1 per 100 among those who were uncircumcised. In the posttrial period, HIV incidence was 0.5 and 1.9 per 100 person-years, respectively, among circumcised and uncircumcised men. The corresponding effectiveness levels for the two periods—59% during the trial and 74% during follow-up—were not significantly different from each other. When background characteristics and sexual behavior were controlled for in the survival analysis, the adjusted hazard ratio of HIV acquisition during the posttrial period was 0.3, and the effectiveness of circumcision was 73%.

In an analysis of participants assigned to the original study’s control arm, HIV incidence during the posttrial period was 0.5 per 100 person-years among circumcised men and 1.7 per 100 among the uncircumcised. These rates translate to a circumcision effectiveness of 68%. In the survival analysis, the adjusted hazard ratio of HIV acquisition during the posttrial period was 0.3, while the circumcision effectiveness was 67%. The background characteristics and sexual behavior of participants who were circumcised during the posttrial period did not differ from those of men who remained uncircumcised, at either the last trial visit or the first posttrial visit, suggesting that self-selection bias was minimal.

The researchers note that the elevated HIV incidence among uncircumcised men in the posttrial period may be attributed in part to some unmeasured high-risk behavior or to reduced exposure to preventive education. They caution against generalizing their findings to other populations or programs because of several study limitations: Participants were self-selected; they received voluntary counseling and treatment at enrollment and after seroconversion, as well as health education at each visit; the timing of HIV infection within a follow-up interval was unknown; and behavioral differences between circumcised and uncircumcised men may emerge only after longer time intervals. Nonetheless, the researchers note that during the posttrial period, "the effectiveness of male circumcision for HIV prevention … was comparable to or higher than" that observed during the initial randomized trial, which "suggests that male circumcision confers long-term protection from HIV infection in men."—J. Thomas

REFERENCE

1. Gray R et al., The effectiveness of male circumcision for HIV prevention and effects on risk behaviors in a posttrial follow-up study, AIDS, 2012, 26(5):609–615.