As we said in the previous section, only 3% of couples in the HPTN 052 study were
male/male, which is not nearly enough to calculate a transmission risk or
reduction in risk for MSM. Some other studies about this group are relevant
though.
A prospective study examining the effect of antiretroviral
treatment on HIV transmission in gay men was first presented in February 2009
to the 16th Conference on Retroviruses and Opportunistic Infections in Montreal,1 with further information
presented in April 2009 to the 15th Annual British HIV Association Conference
in Liverpool.2 This study, involving 1144
gay men attending an HIV-treatment centre in Brighton
between 2000 and 2006, is the first to examine the risk of transmission in a
cohort of gay men and other men who have sex with men (MSM).
The investigators used clinical and epidemiological
information to identify the factors involved in new HIV infections in gay men
and, by performing phylogenetic analysis on the HIV from 859 individuals, 41
'likely transmitters' were identified, 29 (70%) of whom had never taken HIV
treatment and nine of whom (22%) had interrupted their treatment at the time of
transmission. The study found an association between a higher viral load and a
greater risk of HIV transmission, with each tenfold increment in viral load increasing the
risk of HIV transmission by 68%.
Taking HIV treatment was associated with a 96% reduction in the
risk of HIV transmission. Of the three transmissions on treatment seen during
3556 person-years of follow-up, one is thought to have originated in an
individual with an undetectable viral load. No further details are available.
This is not the first recorded case of HIV transmission
during sex between men where the infected partner has an undetectable viral
load, however. In a case report from Germany, published in August 2008,
and confirmed by phylogenetic analysis, a gay man who had maintained an
undetectable viral load on treatment since 2000 apparently infected his partner
between 2002 and 2004 after reporting unprotected anal intercourse on a number of
occasions.3 Neither partner reported
a sexually transmitted infection and both reported that their relationship was
monogamous.
A study from San
Francisco has found that in cases where HIV is transmitted between gay men, there is
a strong correlation between the viral load in the infecting person and the viral
load in the person infected.4
In this study, gay men with acute HIV infection were matched
phylogenetically with infecting partners, as in the Brighton
study. Twenty-four recently infected men were matched with 23 source partners,
one of whom had infected two people. The average viral load in the recently infected
men was 86,332 copies/ml (4.94 logs) and in the source partners was 23,951 copies/ml
(4.38 logs). For every one log (tenfold)
rise in the infecting partner’s viral load, the mean viral load of the partners
infected increased by 0.43 logs (a 2.7-fold increase).
(To explain more clearly, this would imply that if the average
viral load in the infecting men had been 100,000 copies/ml, the average viral
load in partners infected would be 128,000 copies/ml, but if the average viral
load in source partners had been 10,000 copies/ml, the viral load in infected
partners would be 47,500 copies/ml. It is important to emphasise that the viral
load in source and infected partners in this study was being measured at
different time points in the evolution of their infection – in acute infection
in the infected partners and in relatively recent, but chronic, infection in
the source partners).
Two or three of the 23 source individuals in this study had
been on ARVs at the time of infection, but none had an undetectable viral load
when it was measured (minimum 6776 copies/ml3).
In short, treatment seems to reduce the chance of infection
between men (and, by the same token, between heterosexuals who have unprotected
anal intercourse) to the same degree as it does for vaginal intercourse. Given,
however, that the best estimate we have for the per-contact risk of HIV
infection via anal intercourse is that the risk of infection is between five
and ten times greater than that seen in vaginal intercourse,5 we can speculate that it
would take a more profoundly suppressed viral load to render anal intercourse
negligibly unsafe than it would to have the same effect with vaginal
intercourse.