HTB South

No HIV transmissions between gay couples when viral load is undetectable: preliminary results from “Opposites Attract” study in Australia, Thailand and Brazil

CROI logoSimon Collins, HIV i-Base

Preliminary results from a prospective observational study in serodifferent gay couples reported no linked HIV transmissions when the HIV positive partner had an undetectable viral load, even when not using condoms. [1]

This was from an interim prespecified analysis, but the findings are important for adding to those from the European PARTNER study that were presented at CROI last year. [2]

By December 2014, the Opposites Attract study had enrolled 234 couples in Australia (n=135), Bangkok (n=52) and Rio de Janeiro (n=47) who were already not using condoms. As with the PARTNER study, phylogenetic analysis is used to determine whether any new infections are linked to the HIV positive partner.

Only limited results were included in the poster on baseline characteristics of participants, but mean age was 36 (no SD given), and couples had been together for <12 months (39%), 1-5 years (33%) and >5 years (28%).

This analysis contained results from 150 couple years of follow up (CYFU) with only 91/150 from when condoms were not used. 152 couples contributed to follow-up 43% of which (n=65) were open relationships and 88/150 who had condomless sex. Unlike the PARTNER study, not all HIV positive people were on treatment: only 84.2% were on treatment at baseline and overall viral load was undetectable (<200 copies/mL) in 82.9%. Also at baseline, STI prevalence (details not given) was 11% and 6% in the positive and negative partners respectively.

No linked HIV transmissions were reported from 5905 times in 88 couples when condoms were not used (based on reports from the negative partner), with upper limit of 95% confidence intervals of the annual HIV incidence rates ranging from 2.46 to 6.46 per 100 couple years of follow up (CYFU). By comparison this upper limit was 184.31 when viral load was detectable at >200 copies/mL. See Table 1.

Table 1: HIV incidence in interim analysis of Opposites Attract
Type of condomless (CL) anal sex New linked trans-missions Couple years of follow up (CYFU) No. of acts Incidence rate (95% CI)
Overall 0 150 5905 0 (0 to 2.46)
Any CL 0 91 5905 0 (0 to 4.06)
Insertive 0 78 3569 0 (0 to 4.74)
Receptive 0 57 2337 0 (0 to 6.46)
Any when VL <200 c/mL 0 89 5656 0 (0 to 4.16)
Any when VL >200 c/mL 0 2 237 0 (0 to 184.31)


This is good news, even though it is from a very short follow-up time.

Based on the limited data so far, the Opposites Attract study is reporting that although there were no linked transmissions, there is 95% confidence that the real risk is that 0 to 6.5/100 men could become HIV positive over a year. This is from being the receptive partner and based on having sex without condoms an average of 67 times a year. However, using a one-side confidence interval, there is also a 5% (1 in 20) chance that the real risk is above 6.5/100 men per year.

By comparison, the results from the PARTNER study, reported a much smaller upper risk of 2.7/100 men based on receptive sex (with ejaculation, slightly lower without) over the same time period. The lower estimate is because the PARTNER study contains a much larger number of years of observation of men having condomless sex. Because PARTNER uses a two-sided confidence interval there is a 2.5% (1 in 40) chance that the real rate could be higher than the upper confidence limit of 1.7/100 men per year.

This detail is important for explaining why the headline result of finding zero linked transmissions means very different things in each study.

Continued follow up is important in both studies for defining a similar level of confidence in the level of risk for gay men as the PARTNER study already reported for heterosexual couples. Of interest, the Opposites Attract study includes a sub study looking at viral load in semen. [3]

In Europe, this follow-up continues in the PARTNER 2 study, which is currently looking to enrol an additional 530 gay couples from 2014-2017. Support from doctors and researchers to help this enrolment is important and new patient materials, including leaflets and posters, are now available. [4]


  1. Grulich AE et al. HIV transmission in male serodiscordant couples in Australia, Thailand and Brazil. 2015 Conference on Retroviruses and Opportunistic Infections (CROI 2015), 23-26 February 2015, Seattle. Late breaker poster 1019 LB.
  2. Rodger A et al. HIV transmission risk through condomless sex if HIV+ partner on suppressive ART: PARTNER Study. 21st CROI, 3-6 March 2014, Boston. Oral late breaker abstract 153LB.
    See report in HTB March 2014 for details of exposure risks and confidence intervals.
  3. Opposites Attract study website.
  4. PARTNER 2 study website.