Vaginal HIV Prevention Gel Fails in Major Study
Disappointing new study results have dealt a serious blow to hopes for a vaginal gel to help women protect themselves from HIV infection. In research presented at CROI 2015 in Seattle, Washington, a gel containing 1% tenofovir (Viread) failed to show any greater efficacy than a placebo gel. Poor adherence appeared to be the primary reason for the gel's failure.
Key findings from the study, known as FACTS 001, are:
- A total of 61 HIV infections occurred in the tenofovir arm versus 60 in the placebo gel arm; both arms had an infection rate of 4.0 per 100 person-years.
- Participants in both arms only used the gel before 50 to 60% of sexual acts.
- Only 13% of participants used the gel before sex at least 80% of the time, the threshold for tenofovir gel efficacy.
- Among women who reported recently having sex and who had detectable tenofovir levels in vaginal tissue, the tenofovir gel conferred 52% protection against HIV infection.
FACTS 001 is a Phase 3 study of young (mean age 23), sexually active, HIV-negative women recruited from locations in several South African provinces. The acronym stands for Follow-on African Consortium for Tenofovir Studies, an alliance of researchers and organizations that sought to build off of initial successes in vaginal tenofovir gel research.
It was less than five years ago that scientists and activists were literally brought to tears over the early success of the CAPRISA 004 trial, which provided the first-ever sign that a female-controlled biological intervention could potentially protect women from HIV infection.
CAPRISA 004 offered convincing proof of concept that a vaginal gel could be effective. One of the key questions, however, was whether the gel could realistically be used frequently enough to achieve protective drug levels. The FACTS 001 study results, presented at CROI 2015 by Helen Rees on Feb. 24, strongly suggest the answer is no.
This Phase 3 study was large: It enrolled 2,059 women, each of whom was randomized to receive either 1% tenofovir gel or a placebo gel. Participants were instructed to use the gel pericoitally: one gel applicator within 12 hours of sex and another within 12 hours after sex (but no more than two in any 24-hour period). Rees noted this regimen was identical to that used in CAPRISA 004.
These were motivated volunteers; the study protocol involved monthly follow-up visits to study sites, and when the trial concluded after reaching its endpoint (118 total HIV infections, reached 27 months after the first participant was enrolled), more than 90% of active participants showed up for their final scheduled visit. In total, 3,036 person-years of data were obtained.
In light of disappointing results from the 2011 VOICE study, the FACTS 001 study designers built in additional efforts aimed at maximizing adherence among participants. For instance, it offered a range of participant-centered counseling, education efforts and motivational messaging. "We also had a whole strategy that looked at a collective persuasion to try and use the gel, which included FACTS clubs, films and barometers, all for collective discussion," Rees said.
However, despite participants' commitment to participation in the study and the extent to which they were saturated with adherence interventions, actual rates of gel utilization were low. Participants were asked to bring back used gel applicators for each monthly follow-up visit, and were interviewed about their sexual activity. Calculations based on that information yielded the sobering numbers at the top of this article.
In addition, researchers conducted a prespecified case-cohort substudy of participants who were using the tenofovir gel. The substudy population included 56 women who seroconverted while on the tenofovir gel arm plus 158 who did not; quarterly cervicovaginal lavage samples (1,075 total) were collected to gauge tenofovir levels.
Just 22% of the women had detectable levels of tenofovir in all of their quarterly samples, suggesting high adherence; 65% of the women had detectable tenofovir in at least one sample, suggesting periodic adherence; and 13% had no detectable tenofovir in any sample.
Among women who had detectable tenofovir in a sample and who had reported sex in the previous 10 days, the adjusted hazard ratio for HIV infection was 0.48, which translated to a 52% protective effect (P=.04).
"What's important about this is that the similar result has been shown from other studies that have also looked at HIV incidence and tenofovir levels in high-adherent participants," Rees said.
As disappointing as the FACTS 001 results are, Rees felt that they also yielded very clear directions for future research into female-controlled HIV interventions. "We urgently need a range of prevention options for young women which may be easier to integrate into their lives," she said. "We have extensive social science data which shows that women's lives, really, are one of biggest barriers to them being able to integrate this gel use ... despite, as you can see, a real willingness to try and do it."
FACTS 001 was also designed to gauge whether the tenofovir-containing vaginal gel could curb incidence of herpes simplex virus-2 in the participants; Rees said that the researchers were still waiting for the results from that part of the trial.
Myles Helfand is the editorial director of TheBody.com and TheBodyPRO.com.
Follow Myles on Twitter: @MylesatTheBody.