Opposites Attract study: HIV-positive/negative couples used multiple strategies to prevent HIV transmission

Two years ago, the Opposites Attract study found no transmissions between 343 regular gay couples of differing HIV status where the HIV-positive partner had a viral load under 200 copies/ml, the negative partner was not taking pre-exposure prophylaxis (PrEP), and they did not use a condom. In doing so, alongside the PARTNER study, it provided the evidence that people with HIV who had an undetectable viral load were not infectious or, to quote the slogan that has become a global brand, undetectable = untransmittable (U=U).

A new analysis has now delved deeper into the actual sex and prevention behaviour of those who took part in Opposites Attract and found that couples were by no means just relying on the positive partner’s undetectable viral load to avoid the negative partner becoming infected. Some used different strategies altogether, while others combined strategies in a ‘belt-and-braces’ fashion.

This new study looks at couple's sexual prevention behaviours at all time points in the study. The previous analysis, of whether there was any risk of transmission by the HIV-positive partner if he had an undetectable viral load, excluded all sex acts that were covered by condoms or PrEP. In fact, a high proportion of couples continued to use condoms all or some of the time and a third of the negative partners took PrEP during at least part of the study.



A sexual risk reduction strategy for gay men having anal sex without a condom. Different sexual positions are adopted according to HIV status: the HIV negative man takes the insertive role (‘top’) and the HIV-positive man the receptive role (‘bottom’).


Undetectable = Untransmittable (U=U)

U=U stands for Undetectable = Untransmittable. It means that when a person living with HIV is on regular treatment that lowers the amount of virus in their body to undetectable levels, there is zero risk of passing on HIV to their partners. The low level of virus is described as an undetectable viral load. 

exclusion criteria

Defines who cannot take part in a research study. Eligibility criteria may include disease type and stage, other medical conditions, previous treatment history, age, and gender. For example, many trials exclude women who are pregnant, to avoid any possible danger to a baby, or people who are taking a drug that might interact with the treatment being studied.


Receptive anal intercourse refers to the act of being penetrated during anal intercourse. The receptive partner is the ‘bottom’.

post-exposure prophylaxis (PEP)

A month-long course of antiretroviral medicines taken after exposure or possible exposure to HIV, to reduce the risk of acquiring HIV.

There was also high usage of ‘seropositioning’ (the HIV-negative partner taking the insertive role and the positive partner the receptive role), which also reduces the risk of HIV, but because some risk still remains, times when couples practised seropositioning were not excluded.

Opposites Attract started in Australia but later extended to sites in Brazil and Thailand, and there were interesting differences between the HIV prevention choices in the different countries. In Thailand in particular, a high proportion of the HIV-positive partners had not started antiretroviral therapy (ART) when they entered the study, though nearly all had by its end. Viral load tests are conducted less often there, and this may explain why Thai participants were more likely to use condoms or PrEP, alone or in combination, as their main protective behaviour.

The new analysis provides an interesting snapshot of how gay men’s sexual risk and HIV prevention behaviour was changing during a period of transition when universal ART on diagnosis was becoming established and PrEP was becoming more available, but the awareness of them both was perhaps spreading more slowly.

The couples

This new analysis includes all 343 couples from the original study: 153 in Australia, 93 in Brazil and 97 in Thailand. The study started in 2012 in Australia and in 2014 in Brazil and Thailand, before closing at the end of 2016. The Brazilian and Thai participants therefore contributed fewer couple-years of data to the analysis.

The average age of participants was 35 years and 10 months, with almost no difference between the HIV-positive and negative partners. The Australians tended to be older (around 40 years) than the Thais (around 30 years). The HIV-positive partners were somewhat less likely to be university-educated: 4% less likely in Thailand and Brazil, and 8% less likely in Australia.

Just over 40% of couples had been in their sexual relationship for less than a year, about a third from one to five years, and just under a quarter more than five years. Brazilian and Thai couples were more likely to have been together for less than a year (54% and 46% respectively) and less likely to have been together for over five years (15% and 17.5%).

As already mentioned, not all of the HIV-positive participants were on ART when they entered the study. Eighty-five per cent of Australians and 80% of Brazilians were, but only 50.5% of Thais. All but six participants were on ART by the end of the study. Not surprisingly then, whereas 85% of Australians and 81% of Brazilians had viral loads below 200 copies/ml throughout the study, only 55% of Thais did.

Prevention strategies

The differences in the availability of ART were reflected in participants’ prevention strategies. For instance, during the study, 11% of Australians, 31% of Brazilians, and 45% of Thais never had condomless anal sex, meaning that up to a quarter of all participants continued relying on condoms as their sole prevention method, or one of them, and 47% of all acts of anal intercourse in the study involved using a condom.

Multiple prevention methods were used: one of the surprises of the study was the high proportion of people who used what was defined as “daily PrEP” at least part of the time. A third of the HIV-negative participants recorded at least some PrEP use, and usage was higher in Brazil and Thailand (40% and 37%) than Australia (27%).

Some participants may have obtained PrEP via the Princess PrEP programme in Thailand and the demo projects in Australian states such as EPIC-NSW, which both started in early 2016, but PrEP was not available in Brazil officially until 2017. Even allowing for confusion between PrEP and PEP (post-exposure prophylaxis), there seems to have been significant online purchasing of PrEP or other informal sourcing among participants.

In more detail

The researchers took all acts of anal intercourse recorded during the study and classed each as to whether it offered protection from HIV via one or more of four effective methods:

  • The perceived viral undetectability of the HIV-positive partner (U=U): 77% of all acts were covered by this.
  • Seropositioning which offers a 90% reduction in the chance of HIV transmission when used consistently: 62% of all acts were covered by this.
  • Condom use by either partner: 47% of all acts were covered by condoms.
  • PrEP use by the negative partner: 24% of all acts were covered by PrEP.

As can be seen, these figures are not exclusive and most couples used more than one way to protect themselves. For U=U, the proportion of sex acts covered only by this and no other method was just 12%. With condoms, seropositioning and PrEP, only 5%, 3% and 1% of sex acts respectively were covered by this one method alone.

The most popular strategy (assuming it was a deliberate strategy) was to add in seropositioning. The ‘top five’ of mixed strategies were:

  • U=U and seropositioning: 23% of all acts
  • Condoms, U=U and seropositioning: 12% of all acts
  • Condoms and U=U: 9% of all acts
  • Condoms and seropositioning: 9% of all acts
  • U=U, PrEP and seropositioning: 7% of all acts.

Where the HIV-positive partner was perceived as undetectable, condoms were also used (in any combination) 38% of the time, and PrEP was also used 10% of the time.

Just 4% of anal sex acts were covered by all four strategies but conversely only 1.6% were not covered by any strategy – i.e. the positive partner was detectable, neither condoms nor PrEP were used, and the negative partner was receptive.

There were strong differences between countries. Thais and Brazilians were very much more likely to use condoms than Australians: 85% of all anal sex involved using a condom in Thai participants and 64% in Brazilians, versus 26% of Australians. Thais were also more likely than others to use condoms and PrEP (21% of all sex acts involved both) and while the absolute number using PrEP was smaller than for other methods, two-thirds of Thais who did use PrEP used it as their sole method of protection.

Lumping together U=U and PrEP, 69% of Australians used one or both of these as their sole protection versus 35% of Brazilians and only 5% of Thais. Only 5% of participants used strategic positioning as their sole prevention method but it was used by 10% of Thais.

Were people correct about their partner’s viral load?

The data begs the question: were the partners correct about the HIV-positive partner’s viral load? The answer is: in Australia and Brazil, most of the time (88% and 78% respectively). In Thailand, possibly because viral load testing is less common, partners were more likely to perceive their partner’s viral load as detectable when in fact it was not (61%).

If the positive partner did have a viral load over 200 copies/ml, the negative partner was generally correct in either knowing this, or saying he did not know his viral load.

Reassuringly, there were very few cases – just 15 in all – where the negative partner thought their partner’s viral load was undetectable but it wasn’t: just 1% in Australia, 1.2% in Brazil and no cases at all in Thailand. Nonetheless, in all these 15 cases, the HIV-positive partner had a viral load of over 1000 copies/ml and therefore was potentially infectious.


Bavinton B et al. Strategies used by gay male serodiscordant couples to reduce the risk of HIV transmission from anal intercourse in three countries. Journal of the International AIDS Society 22e25277, 2019. (Read full text here).