Circumcision may protect both men and their partners from syphilis, according to an analysis of data from an HIV prevention trial conducted in Uganda and Kenya.1 During the study’s three-year follow-up period, the risk of a new syphilis infection was lower among circumcised men than among their uncircumcised peers (hazard ratio, 0.6), and lower among their partners than among the partners of uncircumcised men (0.4). In analyses that stratified participants by HIV status, circumcision was associated with reductions in syphilis risk among HIV-infected men (0.4), HIV-infected women (0.5) and women without HIV (0.3), though not among men without HIV.
Data from at least three clinical trials have shown that circumcision reduces the risk of HIV acquisition among heterosexual men, and observational studies have suggested that it may also reduce transmission of HPV, herpes and other STIs. At least some of these benefits may extend to women as well. Evidence concerning syphilis transmission is less conclusive, however; although observational studies suggest that circumcision reduces the risk of syphilis, two clinical trials have failed to show protective effects, and no data are available on women.
In the current study, Pintye and colleagues analyzed syphilis data from a prospective clinical trial (the Partners PrEP Study) designed to examine whether pre-exposure therapy reduced the risk of HIV transmission between HIV-serodiscordant partners. All participants received HIV prevention services, including referrals for circumcision, and those randomized to the treatment group received antiretroviral drugs. Participants were recruited from 2008 to 2010 and followed for up to three years; the placebo group was discontinued in 2011 when an interim analysis revealed convincing evidence that outcomes were superior in the treatment group. Throughout follow-up, uninfected partners had monthly counseling visits and infected partners had quarterly visits; testing for syphilis was done annually or at interim visits if clinically indicated.
At enrollment, participants provided information on their demographic characteristics, medical history and sexual behavior in the past 30 days; circumcision status was determined via physical exam at baseline and annually thereafter. The relationship between circumcision and syphilis incidence was assessed using Andersen-Gill survival models that treated circumcision as a time-dependent event (i.e., researchers were able to account for circumcisions that occurred during follow-up). Analyses were conducted separately for men and women by HIV status to examine whether the relationship between circumcision status and syphilis incidence differed by HIV status. All statistical models adjusted for age at enrollment, whether the participant had had unprotected sex in the past 30 days and serum level of HIV RNA in the infected partner; other demographic and behavioral variables were not included in models because they did not have a meaningful impact on odds ratios.
The analytic sample consisted of 4,716 couples; the HIV-infected partner was male in 38% of couples and female in 62%. Median age at enrollment was 36 for men and 30 for women; nearly all couples were married. On average, couples had had sex four times in the past month; one-fourth had had unprotected sex during that time. Men were more likely than women to report having had an outside partner in the past month (14% vs. 1%), though women were more likely to have a curable STI (10% vs. 5%).
Nearly half (46%) of men were circumcised when they entered the study. Compared with their uncircumcised peers, circumcised men had slightly higher levels of education and had more children, and in the past month they were less likely to have had unprotected sex with their partner or to have had sex with an outside partner. They were also less likely to have laboratory-confirmed syphilis (2% vs. 6%) at study enrollment; their partners had a reduced prevalence of infection as well (2% vs. 5%). In multivariate analyses, the odds of syphilis infection at enrollment were reduced among both men and their partners, regardless of HIV status, if the man was circumcised (odds ratios, 0.3–0.5).
During follow-up, 221 new cases of syphilis emerged—122 among men and 99 among women. In multivariate analyses, circumcision was associated with a 42% reduction in the risk of syphilis infection among men (hazard ratio, 0.6). The risk of infection was reduced to an even greater extent among men with HIV (0.4); among men without HIV, the risk reduction fell short of statistical significance. Among women, having a circumcised partner was associated with reductions in the risk of syphilis incidence for the full sample (0.4), those with HIV (0.5) and women without HIV (0.3).
Findings were generally similar (though not necessarily statistically significant) in sensitivity analyses that excluded possible cases of treatment failure (i.e., individuals who tested positive after a previous syphilis infection was detected and treated); that excluded cases in which only one partner had a positive syphilis test; or that were restricted to men who were circumcised during follow-up.
Limitations of the study, according to the authors, include that syphilis testing was done annually (unless men had symptoms that warranted interim testing); hence, cases that were treated by outside providers (including cases that were indirectly treated when men received antibiotics for unrelated conditions) may have gone undetected. Moreover, the circumcision status of any outside partners who transmitted the infection to study participants was not known, potentially leading to misclassification. Nonetheless, the authors note that the study is the first to find that circumcision reduces the risk of syphilis infection in men with HIV or among women of any HIV status. "If confirmed," the authors note, the results "suggest that male circumcision could significantly reduce syphilis incidence and related sequelae in both men and their female partners," which would be a particular public health benefit in settings where HIV is highly prevalent.—P. Doskoch
REFERENCE
1. Pintye et al., Association between male circumcision and incidence of syphilis in men and women, Lancet Global Health, 2014, 2(11):e664–e671, doi: 10.1016/S2214-109X(14)70315-8, accessed Jan. 15, 2015.