By Kristen Jill Kresge

The prestigious US Institute of Medicine (IOM), an independent advisory group on public health policy, convened a series of meetings last year on the methodological challenges of conducting non-vaccine HIV prevention trials. The final report based on these proceedings, as well as site visits by IOM committee members to clinical trial sites in Uganda and South Africa, was just issued in February (www.nap.edu/catalog/12056.html). 

These meetings and the final report were commissioned by the Bill & Melinda Gates Foundation. The foundation requested that the IOM committee focus in particular on research involving microbicides and pre-exposure prophylaxis (PrEP; see VAX May 2006 Spotlight article, Treatment as prevention), and provide recommendations on how future trials could be conducted in a way that could increase the likelihood of success and enable donors to optimally invest their limited financial resources.

At the public meetings, committee members and leading researchers in the field discussed several of the most-pressing issues surrounding the design and conduct of large-scale HIV prevention trials (see Advisory Panel considers complexities of HIV prevention trials, IAVI Report, January-February 2007 and Optimizing HIV prevention research, IAVI Report, March-April 2007).

The final report outlines the recent spate of late-stage clinical trials in the HIV prevention field that have failed to provide any benefit in reducing the risk of HIV infection (see Spotlight, this issue), leading the authors to conclude that, “A near-perfect biomedical intervention for preventing HIV infection is unlikely to be available in the near future.”

The importance of accurately estimating HIV incidence is among the main issues highlighted in the report. This became a concern when multiple prevention trials were stopped early because the HIV incidence observed during the trial was lower than initial estimates on which the trial was based (see VAX July 2007 Primer on Understanding HIV Incidence). The IOM committee recommends that all late-stage trials be designed based on incidence estimates collected through traditional cohort follow-up studies of HIV-uninfected individuals in the communities where the trial will occur. The authors also suggest that this estimate should be corroborated by at least one other source.

High pregnancy rates during HIV prevention trials, and the impact on retention of female volunteers, was another critical issue that was discussed at the committee meetings and is addressed in the report (see Primer, this issue). Female volunteers are typically not allowed to receive the experimental intervention during pregnancy because of potential safety risks to the fetus. But their exclusion from the trial can confound the results. On this issue, the authors suggest that researchers should try to determine the safety of the intervention in pregnant women to determine circumstances where women could potentially continue to participate in HIV prevention trials while pregnant.

The report also outlines several other ways that trials can be designed to more efficiently determine the influence of behavior and adherence on the final results.

What are some of the considerations regarding recruitment and retention of women in AIDS vaccine clinical trials?

To determine the safety and efficacy of an AIDS vaccine candidate, it must be tested in populations that are most affected by the disease. This requires conducting AIDS vaccine clinical trials in developing countries in which there are the highest HIV infection rates.

It is also imperative that a vaccine candidate be tested in individuals and populations that would eventually benefit the most from a preventive AIDS vaccine. This includes groups that are at high risk of HIV infection, either through sexual contact or through blood-to-blood transmission, which occurs in injection-drug users. In many countries, women are at increasingly high risk of HIV infection. According to the latest report from the Joint United Nations Programme on HIV/AIDS (UNAIDS), released in November 2007, 68% of the world’s HIV-infected individuals live in sub-Saharan Africa and the majority of them are women. It is therefore critical that AIDS vaccine candidates are evaluated in HIV-uninfected female volunteers.

Reaching the target

Specific targets are often set for the number of women that will be enrolled in an AIDS vaccine clinical trial. If the percentage of women participating is too low, researchers may be unable to draw conclusions about the safety or efficacy of the vaccine candidate in women.

During an efficacy or preliminary efficacy trial, such as a Phase IIb test-of-concept trial, it is also important that the women are at risk of HIV infection (see VAX July 2007 Primer on Understanding HIV Incidence). This issue was highlighted in the recently-conducted STEP trial, testing Merck’s AIDS vaccine candidate (see VAX October-November 2007 Spotlight article, A STEP back?). The majority of volunteers enrolled at sites in North and South America, Australia, and the Caribbean were men who have sex with men. One-third of the participants were women, but during the course of the trial only a single HIV infection occurred in a female volunteer. As a result, all of the women were excluded from the final data analysis. For the Phambili trial, a companion study to the STEP trial that was conducted in South Africa, investigators planned to enroll mostly women, but this study was stopped early by the trial’s data safety monitoring board based on the results of the STEP trial.

Recruiting women

Recruiting women for AIDS vaccine trials can be challenging. In some places it is difficult for women to participate because they are the sole caregivers for their families and are therefore unable to make regular trial site visits. To make it easier for women to participate, some clinical trial sites offer supervised child-care services and encourage women to bring their children along on clinic visits.

In other situations women are hesitant to participate without the permission of their husbands or male partners. One strategy used to encourage participation in this case is to offer couples voluntary counseling and testing for HIV (seeVAX October 2005 Primer on Understanding Couples Voluntary Counseling and Testing). At many clinical trial sites where couples cohorts are established, researchers have been able to recruit higher numbers of HIV-uninfected female volunteers for AIDS vaccine trials.

Pregnancy and participation

Women may also be unwilling to participate in a trial if they wish to become pregnant. Pregnant women are not allowed to enroll in AIDS vaccine clinical trials because of safety concerns regarding the effect of the product on the woman or the fetus. If a woman becomes pregnant during the course of an AIDS vaccine trial, she is not allowed to receive further vaccinations. Women who become pregnant during an AIDS vaccine trial, as well as their babies, are usually followed beyond the end of the trial to monitor any potential adverse effects of the vaccine. During microbicide or pre-exposure prophylaxis trials—where antiretrovirals are administered to women to try to prevent HIV infection—women must discontinue use of the product for the duration of their pregnancy.

In all HIV prevention trials, women are counseled to use some form of contraception to prevent pregnancy. Some trials require that women use hormonal contraception, either oral or injectable, in addition to a barrier method such as condoms to prevent pregnancy. But this is somewhat controversial—some studies have suggested that hormonal contraceptives can increase a woman’s risk of HIV infection. However, this association has not been proven. Whether or not hormonal contraception is required, female volunteers are usually offered it free of charge. These services, however, are not always provided at the clinical trial sites. Instead women are given a referral to a clinic in the area that provides hormonal contraception.

Despite efforts to provide access to contraceptives, pregnancy rates during some HIV prevention trials have been quite high. All women are tested for HIV infection before enrollment and researchers speculate that women who find out they are not infected may choose that time to become pregnant. In a microbicide trial conducted in Nigeria, 7% of women who were screened for participation in the trial were already pregnant, and during the trial 30% of the participants became pregnant. In a trial testing pre-exposure prophylaxis, the total pregnancy rate at all sites in Cameroon, Nigeria, and Ghana was 56% during the trial. If such a high percentage of women are excluded from the trial for an extended period of time, the trial can lose its statistical power. This limits the ability of investigators to interpret the data and draw conclusions about the safety and efficacy of the intervention being tested.