A Dizzying Second Twist in Trial of Anti-HIV Drugs as Preventives

Yet another large study that aims to prevent the spread of HIV by giving uninfected women antiretroviral (ARV) pills has had to redesign the trial because of startling—and negative—interim results.

The Microbicide Trials Network (MTN), which is funded by the U.S. National Institutes of Health, today announced that it decided to stop one arm of a study involving more than 5000 women in South Africa, Zimbabwe, and Uganda. The decision followed an interim review of the ongoing trial by an independent monitoring board, which found that the drug tenofovir when used as pre-exposure prophylaxis (PrEP) had less effect in protecting women than anticipated. Although the board did not offer any specifics on how many women became infected on the drug versus placebo, they said continuing with the tenofovir arm was "futile" as it would not yield meaningful results. They didn't give numbers because parts of the trial are still going on, testing other prevention measures.

Sharon Hillier, who heads the MTN, says her team is "incredibly disappointed" and that she personally is humbled and somewhat flummoxed. "I've got to quit guessing how studies will turn out," Hillier says. "It breaks your heart."

The study, Vaginal and Oral Interventions to Control the Epidemic (VOICE), began in September 2009 and is not scheduled to end until about 1 year from now. The MTN designed VOICE to compare three different PrEP strategies: tenofovir pills, tenofovir in a vaginal gel, and Truvada pills (a combination of tenofovir and a second antiretroviral, emtricitabine). The tenofovir gel and Truvada pills arms of the study are still under way. "HIV prevention research is a constant reminder of how much we don't know," says Mitchell Warren, who heads the AIDS Vaccine Advocacy Coalition, which closely follows PrEP studies.

The new results particularly baffled people who follow this promising prevention strategy because there were mixed but encouraging findings in two similar studies reported earlier this year. In April, researchers stopped a study called FEM-PrEP that evaluated Truvada pills in nearly 2000 uninfected young women in South Africa, Kenya, and Tanzania after an interim analysis revealed that continuing was futile. But then in July, an early look at infections in a study that evaluated taking either tenofovir or Truvada pills as PrEP found that both worked well in women. A key difference in the second trial, called Partners PrEP, is that it involved more than 4700 couples in Kenya and Uganda in which one partner tested positive at the outset, while both VOICE and FEM-PrEP primarily enrolled young, single women.

Timothy Mastro, who helped lead the truncated FEM-PrEP for FHI 360 in Durham, North Carolina, says teasing out why the same drugs would fail in one population and work in another will require analyzing two main factors: biology and behavior. Studies have shown that small amounts of antiretroviral drugs taken orally reach the vaginal mucosa. That protection may have been overwhelmed in VOICE and FEM-PrEP if male partners had higher levels of HIV than those in Partners PrEP. Or it could be that women in Partners PrEP had more motivation to "adhere" to study protocols and take pills each day as instructed, given that they knew for certain—unlike women in the other two trials—that they were having sex with an HIV-infected man. The infected men in Partners PrEP may have also encouraged their uninfected partners to adhere. "It's really important for our two groups to compare the data and study populations," Mastro says.

VOICE's Hillier says she's looking forward to Mastro's group completing their detailed analyses of the factors behind FEM-PrEP's disappointing results. "There's not going to be a simple answer about who should get PrEP, and it's very clear that different populations can get different results," she says.

Hillier emphasizes that these conflicting findings underscore that much remains to be learned about PrEP, which also worked well in gay men in yet another large study recently completed. That analysis needs to be done before public health officials roll out recommendations for its use. "The data are trying to tell us something very important," Hillier says. "People thought ARVs were going to be magic, and you could sprinkle them out there and people would use them all time and they'd prevent all infections. These studies are teaching us loud and clear that how and when they'll work raises very nuanced questions. And these studies are teaching us things we didn't want to know."

VOICE hopes to have results for the other two arms of the study by the end of next year.