Eight years after it started, the final results of a large French study that offered a choice of daily or event-driven oral PrEP (using tenofovir disoproxil/emtricitabine) were presented at the Conference on Retroviruses and Opportunistic Infections (CROI 2026)Â in Denver, US yesterday.
The Prévenir Study enrolled over 3200 people, most of them gay or bisexual cisgender men, at 26 sites in the Paris region. The results from its first three years were published in 2022.
Key findings of the completed study included the following:
- With remarkable consistency, at any time during the study, half of the participants were taking daily PrEP and half event-driven (‘on demand’).
- This masked considerable dynamism in the choice of PrEP made by individuals. During the study, 59% of participants who had started on daily PrEP changed at least once to on-demand PrEP, and 52% changed at least once the other way.
- Fourteen participants acquired HIV, equating to a very low overall annual HIV incidence of 0.11%. HIV incidence was twice as high in participants who were primarily on-demand users, but because there were so few HIV acquisitions, this difference was not statistically significant.
- One of Prevenir’s targets was to contribute towards a general reduction in HIV cases in the Paris region of 15%, compared to pre-study levels. This was not achieved, and the population-level incidence only fell by 2% – during a period when the overall number of PrEP users in the region increased from a few hundred to over 27,000.
- This, however, concealed a big difference between the French-born majority, where HIV cases fell by 33%, and people born abroad, where cases increased by 75%.
More details
Nearly 99% of the 3209 people in Prévenir were gay or bisexual men, and 85% of participants were White. Their average age was 36. There were 27 heterosexual women and men in the study and 14 transgender women. Almost one in six participants (17.5%) were born outside France, many of them in Africa. Nearly 60% had already used PrEP when they entered the study, and 47% chose to use on-demand PrEP at study entry. This overall proportion scarcely changed throughout, with 49% using it at the study’s close.
But as reported above, changing from daily to event-driven or vice-versa was the rule rather than the exception for individual participants. Because of this, the study population was divided into three groups: people who spent more than three-quarters of the time taking daily PrEP (38%), people who spent more than three-quarters of the time taking on-demand PrEP (38%) and people who switched more often from one method to the other and back (23%).
Professor Jean-Michel Molina said that people tended to switch to on-demand during periods when they were having less sex and those who started as on-demand users were somewhat older than average.
Over the study’s eight years, 66% of on-demand users eventually left the study, 60.5% of daily users, and 56% of switchers, or about 15% a year on average. Only 0.3% of discontinuations per year were due to PrEP-related side effects.
There were 14 HIV acquisitions altogether, with an overall annual incidence of 0.11%. One participant did not give consent for other details to be recorded. Among the other 13, there were seven HIV acquisitions among consistent on-demand users, and three among daily users and three among switchers, equating respectively to annual HIV incidences of 0.15%, 0.06% and 0.09%, but these differences were not statistically significant.
Nine of the 13 acquired HIV while not taking PrEP, while drug levels in the other four showed evidence of suboptimal adherence.
There were a lot of gonorrhoea and chlamydia infections in participants, with an annual incidence of 33% and 26%, respectively, but a relatively low syphilis incidence of 1%. One of the surprises of the study was a high hepatitis C incidence of 0.45% a year (61 infections altogether) which was attributable largely to injection drug use rather than sexual transmission.
The only drug-related adverse events that reached significance were gastrointestinal, which were experienced by 4.5% of on-demand users compared with half that percentage in daily users. This is not surprising as other studies have found symptoms more likely in people starting or restarting PrEP. Daily PrEP users had a higher rate of raised liver enzymes than on-demand users, but not of severely raised levels. Significantly poor kidney function (eGFR below 50) was rare in all groups, with only 11 instances across the whole study, for an annual incidence of less than 0.15%.
The number of sexual partners people had decreased somewhat across the eight years, by 23% in daily users and 37% in switchers, but the proportion who had condomless receptive sex last time they had sex increased from about 40% to about 50% over eight years, with the largest increase (about 50%) in switchers.
If there was one message coming from the eight years of Prévenir, Professor Molina said, it was that both daily and on-demand oral PrEP were safe and effective in the study population. However, there is clearly a need to implement PrEP in France in gay and bisexual men and transgender women born abroad and other populations at risk of HIV acquisition.
References
Molina J-M et al. Final Results of the ANRS Prévenir Study with Daily or On Demand Oral PrEP with TDF/FTC in Paris, France. Conference on Retroviruses and Opportunistic Infections, Denver, abstract 127, 2026.
