No-one with an undetectable viral load, gay or heterosexual, transmits HIV in first two years of PARTNER study

The second large study to look at whether people with HIV become non-infectious if they are on antiretroviral therapy (ART) has found no cases where someone with a viral load under 200 copies/ml transmitted HIV, either by anal or vaginal sex.

Statistical analysis shows that the maximum likely chance of transmission via anal sex from someone on successful HIV treatment was 1% a year for any anal sex and 4% for anal sex with ejaculation where the HIV-negative partner was receptive; but the true likelihood is probably much nearer to zero than this.

When asked what the study tells us about the chance of someone with an undetectable viral load  transmitting HIV, presenter Alison Rodger said: "Our best estimate is it's zero."

The participants

The previous study, HPTN 052, established in 2011 that the efficacy of antiretroviral therapy at reducing HIV transmission from the HIV-positive partner to the HIV-negative one was at least 96% in heterosexual couples, but had too few gay couples in it to establish if the same applied to them (or rather to anal sex).

The PARTNER study was designed to remedy this gap in knowledge. It has so far recruited 1110 couples where the partners have differing HIV status – and nearly 40% of them are gay couples.

In order to be in the study, couples have to be having sex without condoms at least some of the time. The HIV-negative partner cannot be using post-exposure or pre-exposure prophylaxis (PEP or PrEP) and the HIV-positive partner has to be on ART, with the most recent viral load below 200 copies/ml. This is different from HPTN052, which measured the efficacy of the HIV positive partner starting therapy (versus partners who did not).

In total, 767 couples took part in this two-year interim analysis and there were a total of 894 couple-years of follow-up. Among the heterosexual couples, HIV serostatus was split evenly – in half the couples the man had HIV and in the other half, the woman.

Some couples were excluded from this analysis. In most cases, this was because they did not attend follow-up appointments but in 16% of cases it was because the HIV-positive partner developed a viral load above 200 copies/ml, and in 3% of cases because the HIV-negative partner took PEP or PrEP.

There were significant differences between the gay and heterosexual couples. At baseline, gay couples had been having condomless sex for a shorter period on average: 1.5 years versus 2.5 for heterosexual men and 3.5 for women.

During the follow-up period, all the heterosexual HIV-negative partners reported condomless vaginal sex, 72% with ejaculation; 70% of the gay HIV-negative partners reported receptive anal sex, 40% with ejaculation, while 30% reported only being the insertive partner. A significant proportion of the heterosexual couples reported anal sex (to be reported later).

Condomless sex outside the relationship was much more common in the gay men – a third of the HIV-negative partners reported this, versus 3-4% of heterosexuals. No doubt because of this, sexually transmitted infections (STIs) were much more common in the gay couples, with16% of gay men developing an STI (mainly gonorrhoea or syphilis) during the follow-up period versus 5% of the heterosexuals.

At the start of the study, the HIV-positive partner had been on ART for five years in the gay couples and for 7-10 years in the heterosexuals; the proportion reporting an undetectable viral load was 94% in the gay men and 85-86% in the heterosexuals.    

The results

The main news is that in PARTNER so far there have been no transmissions within couples from a partner with an undetectable viral load, in what was estimated as 16,400 occasions of sex in the gay men and 28,000 in the heterosexuals.

Although some of the HIV-negative partners became HIV positive (exactly how many will be revealed in later analyses), genetic testing of the HIV revealed that in all cases the virus came from someone other than the main partner.

Alison Rodger told the conference that if the HIV-positive partners had not been on treatment in this group, 50-100 (median: 86) transmissions would have been expected in the gay couples, and 15 transmissions in heterosexual couples.

No transmissions is not the same as zero chance of transmission. The researchers calculated the 95% confidence intervals for the results seen. What this means is that they calculated the odds of zero transmissions being the ‘true’ figure and what the maximum possible risk of transmission was, given the results seen.

They established that there was a 95% chance that (in a couple whose sexual activity is average for the group studied) the greatest-possible risk of transmission from a partner was 0.45% per year and from anal sex was 1% a year.

At a press conference, PARTNER study principal investigator Dr Jens Lundgren pointed out that this meant that there was a maximum 5% chance that over a ten-year period, one in ten HIV-negative partners in a gay couple who had unprotected anal sex might acquire HIV; equally, though, it was more likely that their chance of acquiring HIV from their partner was nearer to zero, and indeed could be zero.

As the group studied becomes smaller, so confidence intervals become larger and the certainty of a result becomes more ‘fuzzy’. This means that the maximum likely chance of transmission form someone on fully suppressive HIV therapy was 2% a year for vaginal sex with ejaculation, 2.5% for receptive anal sex, and 4% for receptive anal sex with ejaculation. This latter figure implies a more than one-in-three chance of infection if sexual behaviour remains unchanged over ten years, but again this is a 'worst case' scenario and the likelihood is probably lower.

Find out more: Undetectable viral load and HIV transmission

No transmissions occurred despite quite high levels of STIs, especially in the gay couples. When the ‘Swiss Statement’ was released in 2008, it declared that people with an undetectable viral load did not transmit HIV, but made an exception of people with an STI: the PARTNER study may be telling us that STIs (in either the positive or negative partner) don’t increase the likelihood of HIV transmission if the positive partner is on ART and undetectable (though of course they can still be transmitted themselves).

PARTNER is still recruiting gay male couples and, as noted above, its full results will not be out till 2017. Till then we need to be cautious about what it has proved, and, as Jens Lundgren pointed out, it will probably never be possible to show with mathematical certainty that the risk of transmission from someone on successful HIV therapy is absolutely zero. In addition, these results exclude situations where ART failed in the HIV positive partner, though there were relatively few of these cases.