New kind of AIDS vaccine, new kind of trial?

At the recent Retrovirus conference in Boston, Steve Self of the University of Washington in Seattle - part of the US NIH HIV Vaccine Trials Network - explored the implications for clinical trial design of vaccines expected not to prevent infection but to change the course of illness and reduce infectivity.

Several vaccines which may fall into this category are now in clinical trials. These include the DNA-MVA subtype A vaccines now in Phase I and II trials sponsored by IAVI in England, Kenya and Uganda; the Merck DNA-adenovirus vaccine in trials in the USA; another DNA-adenovirus combination being developed by the US National Institutes of Health Vaccine Research Center, and a DNA-MVA vaccine developed by Harriet Robinson’s group at Emory University in the USA, which has just begun a phase I trial in the USA. (Initials are decoded at the end of this article!)

Self asked how a Phase III trial could best be designed, to find out if a vaccine of this kind would have an impact, not only on individuals’ disease risks but on a community’s ability to live with an HIV epidemic.

The main outcome considered in the recently concluded AIDSVAX trial (results awaited), namely reduction of individual risk of becoming infected with HIV, may not be relevant to such vaccines. While AIDSVAX trial volunteers who become infected are being followed up to look for delayed progression, animal studies suggest that vaccines that generate cellular immune responses are more effective in doing this than the antibody-generating AIDSVAX products from VaxGen. As previously reported by the same group at the International Conference on AIDS in Barcelona, there are special problems in assessing delayed progression in people treated with a vaccine that sometimes, but not always, prevents infection (Hudgens).

An initial trial might still randomise people individually, in a similar design to the AIDSVAX trials. In a population with a 2% risk of HIV infection every year, it should be possible to get answers from a trial that recruited 5,000 people over 12 months and followed them for another four years. This would show if the vaccine did, in fact, have an impact on HIV infection rates – and would be large enough to tell the difference between a vaccine that was 30% effective and one that was 60% effective. One of the differences, he said, between vaccines that prevent infection and those which may influence the later course of disease, is that the former can be boosted to extend their effect while the latter probably cannot be boosted, once the person is infected with HIV. (Some might question this, especially if the person is treated with ARVs.)

With no protection against infection, this trial design would also make it possible to compare two groups of people with HIV for an average follow-up of around 18 months. This might be long enough to look at some markers of progression (viral load and falling CD4 counts) although this could be obscured by treatment. As follow-up increased, he thought the likelihood of treatment obscuring any effect of the vaccine would become greater. (It might also be reduced, if treatment is driven by CD4 counts, since better control of the virus should lead to deferred treatment.)

The greatest public health benefit would be likely to come from using a vaccine that reduced onward transmission of the virus. Where much of the transmission is driven by newly infected people, passing on the virus while their viral load is at its peak, this could be a major benefit. But how could it best be detected?

Self’s conclusion is that future trials should randomise whole communities to two different interventions, differing only in the presence or absence of the vaccine. Both arms of the study would promote voluntary counselling and testing and provide clinical care and ARV treatment, but only one arm would offer the trial vaccine (to people who test HIV negative). The treatment programme would be essential, to secure high rates of testing and participation among those at risk of HIV, and he argued that it would not compromise the assessment of onward transmission.

The greatest benefit from a vaccine would be if it reduced viral load in vaccinees in the period immediately after infection and before antibodies were produced. If so, this should be reflected relatively rapidly in lower rates of new HIV diagnoses over the course of the study, in communities that have received the vaccine compared to those which have not.

This proposed trial design sets up a number of challenges. Results could vary, as for different STI control strategies, depending on the maturity of the epidemic. The size of the communities compared, the actual level of vaccine coverage achieved, and the extent to which community membership remains stable over the period of the study, may all be issues. So, too, is the question of how and from whom consent should be sought for such a study.

Nonetheless, this proposal would eliminate the conflict sometimes perceived between treatment and care, including ARV access, and prevention, including vaccines. It could also strengthen the case for providing HIV/AIDS treatment and care in smaller and more rural communities, where such trials would be most likely to give clear results.