Meeting Summary: Optimizing HIV Prevention Research

IOM panel hears from second batch of experts on the pitfalls and promise of conducting HIV prevention research

By Kristen Jill Kresge

The field of HIV prevention research is shifting into a new era. The recently completed trials indicating that male circumcision is about 60% effective at preventing heterosexual HIV transmission in men spurred the World Health Organization to prepare guidelines on implementing this into the comprehensive prevention of HIV/AIDS. Meanwhile, many other HIV prevention strategies, including microbicides, vaccines, the female diaphragm, treatment of herpes simplex virus (HSV-2), and the prophylactic use of antiretrovirals (ARVs), are currently undergoing testing in multiple clinical trials around the world.

But these trials face many potential snags. Among these is a lower-than-expected HIV incidence in many populations, which makes it more difficult to determine the efficacy of these interventions. Higher-than-expected pregnancy rates among women who participate in these prevention studies also impedes data collection and slows the pace of research. These were two of the main subjects addressed when the esteemed Institute of Medicine (IOM), part of the US National Academy of Sciences, convened its first meeting on the methodological challenges in HIV prevention trials in the US this past winter (see Advisory panel considers complexities of HIV prevention trials, IAVI Report 11, 1, 2007).

At the second in a pair of meetings on this subject, held on April 19 in London, the independent panel selected by IOM gathered more information about the challenges and successes in conducting HIV prevention trials involving microbicides, HSV-2 suppressive therapy, the female diaphragm, and male circumcision, while very little, if any, of the discussion centered on AIDS vaccine trials in particular. Some of the overarching concepts discussed at this meeting included measuring adherence to the intervention being tested, selecting the best candidates to advance into clinical trials, and strategies for optimizing trial design.

These meetings, as well as the final report from the IOM that will be released this coming fall, were prompted by a request from the Bill & Melinda Gates Foundation, which is currently funding trials of several new prevention technologies. According to Robin Shattock of St. George's Hospital Medical School in London, funding organizations are increasingly interested in the potential complications of conducting prevention trials and how funds are spent. "Donors are getting much wiser to the costs of these trials and their likelihood of success," he told the panel.


One theme that emerged during the first IOM meeting was the overwhelming need for research groups to have accurate HIV incidence estimates when designing a trial. There are different methods for estimating incidence in a particular community, including the use of previously published data, the BED capture enzyme immunoassay that uses the ratio of anti-HIV immunoglobulin G (IgG) to total IgG to determine early HIV infection, or traditional epidemiological studies that document new infections (see Advisory panel considers complexities of HIV prevention trials, IAVI Report 11, 1, 2007). Obviously, the last of these is the most expensive option, but as Kenneth Mayer, professor of medicine and community health at Brown University in Rhode Island, told the panel: "The most costly and time-consuming methods for determining incidence are the most reliable."

Several examples of how imprecise incidence estimates can negatively affect prevention trials were discussed in the first meeting. During the second session, Gita Ramjee, director of the HIV prevention unit at the South African Medical Research Council, extolled the benefits of conducting preliminary studies to estimate HIV incidence in the communities where a future trial will occur. Before selecting sites to conduct several microbicide trials currently being conducted in South Africa, Ramjee and her colleagues conducted two, 14-month-long feasibility studies at four sites and found that the HIV incidence ranged from 5.0% to 8.5% per year.

"Preparedness studies are extremely valuable," said Ramjee. "You don't want to select a trial site and then find the incidence is too low," she added. One possible fix for the problem is designing clinical trials on a set number of events or end points, in this case HIV infections, rather than on an expected incidence of infection. The idea of end-point-driven trials was mentioned by several of the IOM panelists as a way to eliminate the potential of a trial closing prematurely and inconclusively because it was based upon an inaccurate HIV incidence.

Retaining female volunteers

HIV prevention researchers have observed much higher-than-anticipated pregnancy rates among female volunteers in many trials and since pregnant women must interrupt their use of the experimental intervention due to potential safety concerns, this has become an important issue. These interruptions, which in many cases last nearly the length of the trial, often confound data collection and consequently some trial sponsors are now requiring women to take hormonal contraceptives to be eligible for enrollment. This point was re-emphasized at the second meeting by Helen Rees, executive director of the reproductive health and HIV research unit at the University of the Witwatersrand in South Africa.

Rees proposed that free hormonal contraception should be offered to all women participating in HIV prevention trials, but she did not suggest that it be compulsory. She also raised concern that offering free contraception to women who couldn't otherwise afford it might serve as an unfair inducement for them to participate in the trial.

Rees also noted that pregnancy may actually increase a woman's risk of becoming HIV infected biologically, as well as behaviorally, since pregnant women are even less likely to use condoms. Both of these factors argue for allowing women to continue using the experimental intervention during pregnancy. This is probably most feasible for the PrEP trials and some of the microbicide candidates that incorporate already-licensed ARVs that are currently approved for use by pregnant women. A recommendation on whether or not contraception should be a requirement for participation in HIV prevention research and how trial sponsors should deal with product use during pregnancy will likely come in the IOM's final report.

Determining adherence

The need for researchers to accurately asses a volunteer's adherence to the product/device is another complication shared by most new prevention technologies because they are self-initiated and require consistent and correct use to be effective. Despite extensive volunteer education during trials, adherence is rarely perfect and this can profoundly affect the interpretation of trial results.

Several invited speakers at both meetings addressed different methods researchers use to determine self-reported adherence and sexual behaviors. Coital diaries, where volunteers record their sexual activities as well as use of condoms or the intervention being studied, are one method. Another is coating the applicators used to deliver microbicides with a compound that changes color when it comes in contact with vaginal fluids. Inspection of used applicators can then be tallied with coital diaries.

A further complication related to adherence was encountered in the trials testing the ability of the female diaphragm to prevent HIV infection. Researchers observed an adherence rate of about 70% to both condoms and the diaphragm amongst all volunteers, and although this impressive adherence rate is encouraging, if women who are most likely to use the diaphragm also have the highest adherence to condoms, it could potentially mask the diaphragm's effect. This concern was shared with the IOM panel by Nancy Padian, executive director of the Women's Global Health Imperative at the University of California, San Francisco, who conducted this study.

Andrew Nunn, associate director of the clinical trials unit at the Medical Research Council in the UK, said evidence also suggests that adherence to interventions like microbicides or the diaphragm may decline with time. To avoid this affecting data interpretation, Nunn suggested only following volunteers for one year to determine efficacy of the intervention and then subsequently for longer periods to determine adherence.

New trial design

Nunn also emphasized the importance of smaller Phase IIb trials—which have recently become de rigueur in the AIDS vaccine field—to rank microbicide candidates according to their efficacy before initiating larger and much more costly Phase III trials. He emphasized the importance of collecting efficacy data as soon as possible for the microbicide field in particular, which has been dogged lately by the closure of some Phase III trials that indicated the candidate at best offered no protection, or at worst actually increased a woman's risk of contracting HIV (seeAdvisory panel considers complexities of HIV prevention trials, IAVI Report 11, 1, 2007).

"To be able to show it works at all would be an achievement," said Nunn, who emphasized that preliminary efficacy data will be vital to secure funding for large-scale trials. Shattock shared this concern and said funding agencies are ultimately the ones who make the decision about which candidates are worth moving forward.

Monkey see, monkey do

Shattock also proposed that the microbicide field start using a rational development strategy, including more rigorous preclinical testing using the rhesus macaque model to determine pre-clinical efficacy. "I don't think this is setting the bar too high," he said.

Many of the microbicide candidates that are currently in Phase III efficacy trials have never been tested in non-human primates or against HIV strains that are circulating in Africa. And critically, most have not been evaluated against CCR5-tropic HIV, which is by far the most commonly transmitted virus.

Several of the early-stage microbicides have more specific activity and are based on existing ARVs, including non-nucleoside reverse transcriptase inhibitors like tenofovir and CCR5-inhibitors, and Shattock said these appear to be much more potent. "We can start to look at picking the best drug in a category and taking it into clinical trials," he said.

But Nunn also presented several of the drawbacks to relying too heavily on the limited data that Phase IIb trials provide because it could lead researchers to abandon potentially useful products that only have a modest efficacy, which is all most experts expect from a first-generation microbicide.

Anticipating IOM recommendations

When the final report is issued it is likely to contain recommendations that will be useful for funding organizations, trial sponsors, and the groups conducting clinical trials of new prevention technologies—especially when evaluating new candidates—and is intended to improve trial design and encouraging more efficient use of research funds. It may also help create more consensus within the somewhat disparate field of HIV prevention.

By the time the IOM's report is released, results from the diaphragm trials and perhaps the HSV-2 suppression trials will already be available. If either or both of these interventions show promise, the task—as with male circumcision—will be implementing educational programs and making these prevention technologies available to those who are at highest risk of contracting HIV.

Over the course of the two meetings a few of the IOM panelists kept returning to the notion that different HIV prevention strategies should be tested in combination, rather than in parallel. One panelist remarked that if HIV treatment trials were done similarly to those in the prevention field, "we never would have gotten anywhere." Trials to test combinations of modestly effective, behaviorally-dependent interventions—like microbicides, the diaphragm, and PrEP—together with already-tested methods like male circumcision, may be more complicated to design and more expensive to conduct. But they may also be the best way to determine what is necessary to finally roll back the number of new HIV infections and make headway in curtailing the pandemic's spread until an effective preventive AIDS vaccine is available.