Baffling Failure of Vaginal Gel Laced With Anti-HIV Drug

A study in sub-Saharan Africa of a method to help prevent heterosexual transmission of HIV to women has been stopped midcourse as early data have shown that the microbicide, a vaginal gel laced with an antiviral drug, failed to reduce the rate of new infections. A previous trial had shown that the gel was marginally effective, so researchers suspect that this complete failure occurred because too many participants failed to use the microbicide as instructed.

The researchers running the study, known as the Vaginal and Oral Interventions to Control the Epidemic (VOICE), announced on 25 November that a review of data by an independent board that monitors the trial found that 6% of approximately 1000 women who were given the drug-laced gel and asked to use it daily became infected by HIV. A similarly sized control group that gave women an inert placebo gel had an incidence of 6.1%. "I'm very pleased that the answer was so clear," says the University of Pittsburgh's Sharon Hillier, who heads the Microbicide Trials Network that is conducting VOICE. "It didn't work a little bit. It just flat out didn't work."

VOICE began in 2009 and enrolled 5029 HIV-negative women in South Africa, Uganda, and Zimbabwe to test three different strategies of so-called pre-exposure prophylaxis (PrEP). In addition to the vaginal gel laced with the drug tenofovir, VOICE also set out to evaluate daily use of oral tenofovir as well as a second drug, Truvada, that combines the drug with another antiretroviral. In September, VOICE stopped the oral tenofovir arm of the study because it, too, did not work. Some 2000 women continue in the study to evaluate oral Truvada versus a placebo.

The failure of the tenofovir-laced vaginal gel surprised and disappointed many in the field. "I don't think anyone suspected that the tenofovir gel would not work," says immunologist John Moore, who works on microbicides at Weill Cornell Medical College in New York City. "One has to assume it had something to do with adherence," he says, referring to the willingness of trial participants to follow instructions and actually use the product as instructed. "No matter how effective it is, if it stays in the tube, it's not going to work," Moore says.

The same product did work in a large South African study reported in July 2010. In that trial, dubbed CAPRISA 004, women were instructed to use the tenofovir-laced gel before and after sex rather than every day. CAPRISA 004 found that the product reduced the risk of infection by 39%, and when researchers did a subset analysis of the "high adherers," who reported using the gel more than 80% of the time that they had sex, the efficacy went up to 54%. Similarly, studies of oral PrEP have found much higher efficacy in highly adherent participants.

Hillier stresses that because VOICE is ongoing, the monitoring group offered only the most general information about the failure of the tenofovir gel, and that the researchers will have to wait for more data to understand why it failed. She, too, suspects adherence played a critical factor. "We've always worried that taking anything every day was going to be hard slog," Hillier says. "It's hard to do something every day for something you don't have yet."

Connie Celum, an epidemiologist at the University of Washington, Seattle, who ran a large oral PrEP study with Truvada that had 73% efficacy, says it will be important to compare VOICE and CAPRISA 004 for differences in sexual behaviors, sexually transmitted infections, cervicovaginal inflammation, and contraceptive use. Adherence can also differ by population. Celum's study focused on long-term couples in which one partner had a known infection and thus was highly motivated to use PrEP—adherence was 97%. In VOICE, the women were predominantly in their 20s and not necessarily even in a long-term relationship. "While we wait to understand the VOICE data, I think we recognize the importance of trials in different populations and that we often cannot predict trial outcomes," Celum says.

Other studies that will soon begin test whether less user-dependent delivery methods—such as a vaginal ring that dispenses an antiretroviral for a month at a time—will lead to more effective microbicides.

Follow News from Science

A 3D plot from a model of the Ebola risk faced at different West African regions over time.
dancing shoes