While studies like those above indicate that prompt treatment significantly reduces the likelihood of HIV transmission, an Australian study published in the February 4, 2010, advance online edition of AIDS yielded the unexpected finding that the risk of transmission per act of sex between men -- in particular uncircumcised men -- appears not to have fallen since the advent of effective ART in the mid-1990s.

Fengyi Jin and colleagues estimated the per-contact probability of HIV transmission due to unprotected anal intercourse among 1,427 initially HIV negative homosexual men (the researchers' classification) recruited in Sydney between 2001 and 2004. The researchers noted that about 70% of HIV positive men in Australia are on ART and 75% of treated individuals have undetectable viral load.

Follow-up continued until June 2007, at which point a total of 53 men had seroconverted. The estimated per-contact probability of transmission for unprotected receptive anal intercourse (i.e., the "bottom" becoming infected) was 1.43% if ejaculation occurred inside the rectum versus 0.65% if the insertive partner withdrew.

The per-contact probability of transmission for insertive anal sex (i.e., the "top" becoming infected) was 0.11% if the insertive partner was circumcised versus 0.62% if he was uncircumcised.

Taken together, these transmission rates are similar to the 0.82% risk per act of unprotected anal intercourse seen in a U.S. study from the early 1990s. The Australian study, however, sheds more light on specific risk factors.

Unprotected receptive anal sex with ejaculation was about twice as risky as either receptive intercourse with withdrawal or insertive intercourse for uncircumcised men, but ten times as risky as insertive intercourse for circumcised men. While 12 men were infected after having unprotected anal intercourse fewer than ten times, six others did not become infected despite reporting "extremely large numbers" of unprotected sex acts with known HIV positive partners.

Circumcision has been shown to reduce the risk of HIV transmission by as much as 60% among heterosexual men in low-income, high-prevalence settings such as Kenya and South Africa, but most studies looking at men who have sex with men in high-income, low-prevalence countries have not seen a similar protective effect. The Australian team, however, previously reported that circumcision may have some benefit for gay/bisexual tops.

Liz Highleyman (liz@black-rose.com) is a freelance medical writer based in San Francisco.