A study presented today at the 20th International AIDS Conference (AIDS 2014) in Melbourne found positive associations between voluntary medical male circumcision (VMMC) and a reduced incidence of syphilis, not just in men, but in their female partners.
Another study found no evidence of risk compensation among men post-circumcision, while a third used a novel food-voucher scheme as an incentive for getting older men to come forward for circumcision.
Syphilis infections fall in circumcised men and their partners
An analysis of the Partners PrEP study found that the incidence of syphilis during the study was reduced by 42% in men who were circumcised compared with men who were not. It found an even stronger 59% reduction in syphilis incidence in the female partners of circumcised men.
The study was primarily a pre-exposure prophylaxis (PrEP) trial, the results of which have previously been reported, but incidence data on sexually transmitted infections (STIs) were collected, as well as the circumcision status of male partners, allowing a substudy to examine the relationship between circumcision status and syphilis.
The Partners PrEP study enrolled 4758 male/female heterosexual couples of opposite HIV status (62% of the partners living with HIV were women), of which 4716 (i.e. 9432 individuals) were included in this substudy. The average age of participants was 30 for women and 36 for men; most couples were married and had at least one child together. At enrolment, about half the men were circumcised.
There were 221 new syphilis infections observed in trial participants, 99 in women and 122 in men. Circumcised men and female partners of circumcised men had lower syphilis incidence. There was a statistically significant reduction of 42% in syphilis cases in circumcised men (p = 0.017) and in the 1645 men living with HIV, the reduction was even greater (62%, p = 0.013). The 36% reduction seen in HIV-negative men was not statistically significant.
In women, the opposite situation applied. Women who had circumcised partners were 59% less likely to acquire syphilis than women with uncircumcised partners (p = 0.001) and the difference was more pronounced in HIV-negative women, where the reduction in syphilis in those with circumcised partners was 75% (p = 0.014). However, the 48% reduction seen in women living with HIV was also significant (p = 0.04).
The researchers may not have captured all new syphilis infections, because the study only screened people for STIs once a year, and also only re-checked circumcision status once a year. Presenter Jillian Pintye of the University of Washington in Seattle said that the data on circumcision status and other STIs in female partners were conflicting, but that male circumcision did seem to offer some protection against ulcerative STIs such as herpes, and they planned further subanalyses to look at these.
Circumcision and risk compensation: no evidence
One concern with circumcision, as with other prevention interventions, is whether men who were circumcised would change their risk behaviour post-circumcision and thereby partly or totally erase its benefits in preventing HIV.
There has been no evidence of this in the randomised controlled trials (RCTs) of circumcision, but volunteers in RCTs receive higher levels of monitoring and support than men circumcised in the subsequent large roll-out programmes. Researchers from the Population Council of Zambia, a country that aims to circumcise 80% of its eligible male population by 2015, and where one-in-five men have been circumcised, therefore looked at whether there was any evidence of risk compensation by repeatedly sampling randomly selected members of the community – both men and women. So far there have been three rounds that surveyed circumcision status and risk behaviour in the same group of men at two-yearly intervals. Risk factors examined were: the proportion who have had unprotected sex; how many have had two or more partners in a year; how many have had an STI; how many have paid for sex; and the association of alcohol use and sex.
Starting with a circumcision rate of 5%, by the second round of sampling 12% of the men were circumcised and by round three 21%. One problem with this analysis is that this group of men was not randomised to be circumcised but came forward voluntarily. It is possible to control for obvious influences on behaviour like age, ethnicity or income, but there could also be ‘invisible’ confounders. What if, for instance, the men who decided to get circumcised were generally those who were more risk-averse? In that case, without controlling for this, a survey might underestimate any increase in risk behaviour subsequent to circumcision. Equally, the opposite might be the case: the men getting circumcised might be the ones who knew they were at higher risk of HIV already, and in this case the study might overestimate the influence of circumcision on risk behaviour. Three different statistical analyses were used to attempt to factor out these invisible confounders.
These three methods found changes in risk that did reach statistical significance, but the risks varied between method, in some cases circumcised men reduced their risk behaviour, and changes, even if statistically significant, were small. Thus two out of three measures found significant changes in mens' partner numbers post-circumcision, but one measure found men had more partners and the other that they had fewer. Similarly, one method found that more circumcised men paid for sex after circumcision but another found that fewer did. One method found a change post-circumcision in the amount of unprotected sex – but to less of it.
Presenter Erica Soler-Hampesjek said that, overall, there was no evidence of any post-circumcision change in risk behaviour in men in either direction: “You would need to see an unmistakable change documented by all three statistical methods,” she commented. However, she added, the study could not rule out the possibility that men seeking circumcision were more risk-averse.
Several other studies in the same session looked at ways of increasing uptake of circumcision.
Financial incentives for circumcision in older men
Kenyan researcher Kawango Agot, one of the principal investigators of the Kisumu RCT of circumcision, commented that only two countries – Kenya and Ethiopia – had achieved more than 50% of their target of men circumcised and only three others – South Africa, Tanzania and Swaziland – had achieved a cirumcision rate of 20-26%.
A particular challenge, she noted, was older men. Between 2008 and 2011, Kenya had circumcised 57% of its eligible male population but the vast majority of these were under 25: only 14.4% of those aged 25 or over had been circumcised.
The main reason, according to surveys, was financial: men over 25 were more likely to be financially independent family providers, in employment, who could not spare the time and loss of earnings to come and get circumcised.
A study therefore cluster-randomised men in different geographical areas aged 25-49 to receive financial compensation, not in cash, but in the shape of vouchers for food and transport, at three different rates versus a group that received no compensation. The amounts paid were 200, 700 or 1200 Kenyan shillings (KES) which translates to US$2.50, $8.75 and $15.00 respectively. Although the latter two sums amount to two to three days’ wages in Kenya (the average wage in the area was 400KES a day), they are small compared to the cost per circumcision. In particular, Agot said, money is lost if not enough men come forward for circumcision. “The cost per circumcision varies between $30 and $100 per operation,” she commented. “This is because if not enough men come forward, our staff are being paid to sit around and read books.” Even $15 spent on an operation could therefore produce a net saving of anything up to $55 if it resulted in all available slots being taken.
In the event, this is what happened. The total number of men in this study was 1504 or 376 per payment group. Over the course of two months, only 1.6% of the unpaid group and 1.9% of the 200KES group (six and seven individuals respectively) came forward, 6.6% (n = 25) of the 700KES and 9% (n = 34) of the 1200KES group were circumcised, meaning that the latter two groups were respectively 4.3 and 6.2 times more likely than the control group to become circumcised.
Agot said that 9% was actually quite a high proportion of such a group to come forward for circumcision within only two months, and that 25% of the group registered strong interest in the future.
Given that the scheme could be cost-saving, Agot said that, although detailed costings had not been prepared, the Kenyan health ministry was interested in extending the voucher scheme nationwide in the future as part of a time-limited circumcision campaign lasting a year or two among men aged 25 and older.
Pintye J et al. Male circumcision and the incidence of syphilis acquisition among male and female partners of HIV-1 serodiscordant heterosexual African couples: a prospective study. 20th International AIDS Conference, Melbourne, abstract MOPDC0103, 2014.
Hewett PC et al. (presenter Soler-Hampesjek E) Assessing risk compensation post-male circumcision in Zambia’s national programme. 20th International AIDS Conference, Melbourne, abstract MOPDC0105, 2014.
Thirumurthy H et al. (presenter Agot K) The effect of conditional economic compensation on uptake of voluntary medical male circumcision: a randomized controlled trial of a demand creation intervention for male circumcision in Kenya. 20th International AIDS Conference, Melbourne, abstract MOPDC0104, 2014.