Condom use in French PrEP trial: half only used PrEP, a quarter used condoms and PrEP, one in six used neither

An analysis of condom use in the placebo-controlled phase of the French IPERGAY trial of intermittent pre-exposure prophylaxis (PrEP), presented at the Conference on Retroviruses and Opportunistic Infections (CROI 2016), found that just over half of the participants had high levels of PrEP use but rarely used condoms, and about a quarter were “belt-and-braces” users who had high levels of both PrEP and condom use.

However, this left about one in six trial participants who had low levels of use of both PrEP and condoms. While their condom use did not change, in a minority of this group their PrEP use declined significantly during the study.

In this randomised phase of the study, participants did not know if they were taking PrEP or placebo. A second presentation of results from the open-label phase of the study, when all participants knew they were on PrEP, found that the use of condoms where the participant was the receptive partner declined slightly but significantly, with a relative fall from previous condom use rates of about 15 to 20%.

About IPERGAY

The IPERGAY trial of intermittent PrEP stopped its randomised phase in October 2014 and at the subsequent CROI in February 2015 it was announced that there were 86% fewer HIV infections in the group of participants allocated to PrEP than to placebo. Its full results were published in December 2015.

Details can be found in these linked reports but, in brief, the innovative regimen that IPERGAY offered participants was that they take two tenofovir/emtricitabine (Truvada) pills in the 24 hours before they anticipated having sex, and if they did have sex, then to continue daily PrEP if they continued sex, and to take one pill on both of the two days after sex if they stopped having it.

The presentation on PrEP and condom use confirms that PrEP use as measured by pharmacy refills and returned bottles was generally high. Adherence, defined as the proportion of cases of anal sex covered by at least the double dose beforehand, two doses afterwards, or both, was also high.

PrEP and condom use in the placebo-controlled trial

The researchers, further analysing the data, discovered that participants fell into four fairly distinct patterns of PrEP use.

Nearly 40% of trial participants were consistent users who used PrEP 95 to 100% of the time they had sex. In fact after the third month of the trial, usage was 100% in this group.

Slightly over 30% of users were high-level but inconsistent users. Their coverage of sex with PrEP varied between 70 and 90% over the course of the trial but did not significantly increase or decline over time. While one reason for this variation could have been poor adherence, another could be that they changed their PrEP status according to the known or perceived HIV or viral load status of partners.

Another group, forming 16% of participants, only used PrEP to cover sex occasionally. During the first 16 months of the trial, their PrEP use fell from about 20% to zero, rallying slightly in the last eight months, though numbers of participants were small by this time and these changes were not statistically significant.

In the smallest of the four groups, forming 13% of participants, PrEP use declined significantly during the trial. These participants started off with 90% use on average in the first two months but usage had fallen to 50% by month ten. It rallied slightly at their one-year visit but had fallen to pretty much zero by month 16.

As for condom use, participants fell into two groups. A majority – 70% – were low-level users whose use of condoms during anal sex varied between 10 and 25% during the trial. The remaining 30% were higher-level users, though on average their condom adherence varied between 80 and 45%, so this is a relative term. Their adherence had a tendency to fall during the trial, from 78% at month one to 45% at month 18, though it increased to 71% at month 24.

Combining the figures for PrEP and condom use, the researchers found that:

• 54% of trial participants were consistent or high-level users of PrEP but did not often use condoms
• 23.5% had high levels of both PrEP and condom use
• 6.5% had high levels of condom use but used PrEP only rarely
• 16% rarely used either PrEP or condoms.

It is of course this latter group who are of greatest concern. If compared to the most-protected 23.5%, this group included older participants (50% more likely to belong to this group for every ten years older). They were also twice as likely to not have had college education, twice as likely to say they were dissatisfied with their sex lives and – worryingly – were nearly three times more likely to have had sex with partners they did not know. They were also somewhat more likely (p = 0.08) to be the active partners in anal sex.

Qualitative studies from IPERGAY will elucidate further why individuals chose particular PrEP and condom use strategies.

Data from the open-label trial

Another presentation gave data from the open-label phase of the trial when, from November 2014, all participants were given the option to use Truvada PrEP.

Among the 400 participants initially enrolled in IPERGAY, 336 were eligible to join the open-label study (the remainder having dropped out of the study for various reasons) and all but three joined. Another 29 new participants were enrolled, making 362. All but one transgender woman were gay men.

The data shown at this conference were for the ten months up to the end of September 2015 and the cumulative follow-up time was 248 participant-years. Study retention was good with only 13 participants discontinuing follow-up (3.6%).

Their average age was 35, and 84% were of French or European origin. At the start of the open-label phase, 10% had had no anal intercourse in the previous month and the average number of partners in the previous two months was seven.

During the open-label phase, one person became HIV positive. Like the two people allocated to PrEP who acquired HIV during the randomised phase, he had stopped using PrEP. He was in the randomised trial for eight months and was diagnosed one month and ten days into the open-label study. By self-report he had not used PrEP since starting the open-label study and had had no drug detectable in in his blood. His HIV had no drug resistance mutations.

Putting together HIV incidence in the randomised and open-label phases, HIV incidence was 0.4% a year in people allocated to PrEP. This compared with 0.91% in people allocated to PrEP in the randomised phase alone and to 6.6% in people allocated to placebo; it implies 94% fewer HIV infections in those given PrEP versus those given placebo.

Participants used an average of 18 pills a month according to pharmacy returns but this is probably an overestimate because, after the results from the randomised study were announced, participants became reluctant to return their PrEP bottles in case PrEP became unavailable after the trial (in fact, fully reimbursable PrEP became available in France from the beginning of this year.)

During the open-label study a third of participants were diagnosed with a new sexually transmitted infection. Of these 38% were diagnosed with gonorrhoea, 42% with chlamydia and 21% with syphilis. Three individuals (1%) were diagnosed with hepatitis C.

There was no significant change between the randomised phase and the open-label phase in the median number of occasions of sex or sexual partners. But there was a significant decrease in condom use for receptive anal intercourse Condom use as the receptive partner varied between 40 and 25% during the randomised study. During the open-label study, it varied between 20 and 30%.

Safety was good with a low rate of serious adverse events (4%). One participant discontinued PrEP because of a decrease in creatinine clearance, but more as a precaution as high actual relative loss in kidney function was slight: his creatinine clearance was 81 millilitres per minute at the start of the open-label study and 76 at discontinuation.

Drug-related gastrointestinal adverse events were fairly common. Diarrhoea was reported by 10% of participants, nausea by 11%, abdominal pain by 8% and other gastrointestinal symptoms. The only life-threatening adverse event was a stroke, in a participant who already had a cerebral aneurysm (blood vessel dilation). This was not regarded as drug-related.

Further data will be forthcoming from IPERGAY and the French health ministry has made the collection of behavioural and safety data a requirement for joining the rollout programme.