Phase 1 results from PARTNER were reported in the Journal of the American Medical Association in July 2016. These included findings for heterosexual and gay couples. Results showed no transmissions when the positive partner’s viral load was under 200 copies/mL. PARTNER2 is continuing with the enrolment and follow-up of solely gay male couples.

There were almost 58,000 episodes of condomless sex in the reported results. This included 22,000 occurrences of anal sex. Results were from almost 900 couples, one third were gay men.

What do these results mean?

As no transmissions occurred between partners, results suggest treatment dramatically reduces transmission risk.

The study will continue to collect further follow-up data in gay men. No study can prove something is 100% safe. Research can only try to quantify levels of risk. If risk is very low, this is generally interpreted as being safe. However, interpretations may vary depending on individual decisions about risk.

Would more transmissions occur if the same sex happened with partners not on medication?

If not on medication, we estimated approximately over 100 infections would have occurred in gay men. Estimations were based on assumptions regarding probability of sexual transmission from two meta-analyses. Boily (2009) estimated transmission per heterosexual sex act in high income countries. Male to female risk was 0.0008 (95% CI: 0.0006-0.0011) and female to male 0.0004 (95% CI:0.0001-0.0014). Baggaley (2010) estimated transmission probability per receptive anal sex act to be 0.014 (95%CI: 0.002-0.025).

Why does the study refer to confidence intervals? What do these mean?

When estimating risk, scientists allow for results possibly occurring by chance. This involves calculating an upper/lower range of possible values called the confidence interval (CI). The 95% CI is the results range potentially observed given possible effects of chance. Generally, in larger studies there is greater confidence in results not being due to chance. In PARTNER, study size is measured by years people are followed, rather than participant numbers. This is why the continued follow-up time in gay men is essential in the study.

Phase one results had an upper 95% confidence limit of 0.3% per year. This was for couples having condomless sex with an undetectable viral load. The upper limit of the 95% CI varies depending on type of sex.

Compared with risks of daily life, this maximum estimate is considered relatively low. However, deciding whether health risks are considered high or low is a personal decision. Deciphering whether benefits outweigh risks is up to the individual.

What do the results tell us about risk for anal sex?

Detailed participant questionnaires enable PARTNER to estimate transmission risk via different sexual routes. PARTNER has the most data on HIV transmission risk during anal sex from an undetectable partner. However, since couples reporting anal sex were low, calculated upper limit of risks for anal sex risk were higher at 0.7% per year.

This does not mean PARTNER found evidence that anal sex is higher risk than vaginal. This is a factor of the amount of data we have. With less data there is more uncertainty as fewer couples reported anal sex. PARTNER 2 plans to increase certainty levels around anal sex risk. This is especially important for receptive anal sex with ejaculation.

Does this mean HIV negative people needn’t use condoms if their partner is 'undetectable'?

Current studies have reported no sexual transmission from someone with a suppressed viral load. However, nobody can exclude it could ever happen. Thus, condomless sex with a partner on stable ART, is extremely low risk. Data from PARTNER further reduces uncertainty about the limited risk of transmission.

The decision on how to rate this level of uncertainty is personal. There are many daily risks which we do consider safe (such as driving a car). Comparatively, stopping using condoms with a stably treated partner can also be considered safe. The results make it easy to see why many couples decide not to use condoms.

However, the study also demonstrates there is HIV risk if having sex without condoms (regardless of sexuality). This is heightened if a sexual partner’s HIV and antiretroviral therapy status is unknown. Condoms will reduce the risk of this. Condoms will also reduce the risk of transmission of other sexually transmitted infections. Therefore, the message is that in such situations condoms use remains recommended.

How does the PARTNER study define ‘undetectable viral load’?

For the study, an undetectable viral load was defined as less than 200 copies/mL. Hospitals use viral load tests with different low-level cut-off values. This can be at 20 or 50 or 200 copies/mL depending on the site.

How long does viral load remain undetectable after a viral load test?

Viral load results give information about viral load at the time blood was taken. If one continues taking treatment on time, viral load is likely to remain undetectable. Approximately 5% of patients stable on treatment see their viral load rebound each year. This is largely thought to be related to adherence difficulties. Viral load rebound in the context of good adherence to meds is unusual. If someone stops treatment/misses medication for several days, viral load will likely become detectable.

For an excellent explanation of the early results please visit: