Over 35 million people worldwide are known to be infected with HIV, according to the World Health Organization. With such a heavy global burden, research into vaccines against HIV are underway in order to prevent its further spread. For such research to move from the laboratory bench into the field, lengthy clinical trials must be conducted to assess their safety, immunogenicity and efficacy. In an effort to accelerate the clinical development of HIV vaccine strategies, Rodolphe Thiébaut and Laura Richert from the Bordeaux School of Public Health, France, and colleagues, explore early stage trial designs for four HIV vaccine strategies and propose a randomised multi-arm phase I/II design. Here Thiébaut and Richert explain the key elements of their optimised trial design, as published in a recent methodology study in Trials, and discuss whether this could be applied to later phase trials and vaccine development strategies for other diseases.
What got you interested in HIV vaccine studies, and in particular methods to accelerate their clinical development?
As part of the biostatics core of the French Vaccine Research Institute (VRI) we are directly involved in HIV clinical trial designs and thus confronted with applied methodological questions in our daily research activities. Given the many unknowns in HIV vaccine development, in particular the absence of validated surrogate markers, early clinical trials are part of an iterative ‘discovery’ science, in which candidate strategies go back and forth between preclinical and clinical studies in order to learn the most from the data. Many prophylactic candidate vaccine strategies are currently in early development, and the most promising among them should be identified as rapidly as possible in order to use available resources efficiently in the overall development plan. Acceleration of the early clinical evaluation of these strategies is thus one key aspect for moving HIV vaccine research forward.
Adaptive design methods in clinical research have become more popular in recent years, especially in cancer. Do you think this is also the way forward for randomised controlled trials in HIV vaccine development?
We are completely inspired by the evolution in designs in cancer research and we think that adaptive designs should also be relevant for prophylactic HIV vaccine development. However, this latter field is very different from cancer research and requires specific trial designs.
Can you briefly outline the key aspects of your optimised study design for accelerating early stage clinical development of HIV vaccines?
Our proposed trial design combines phase I and phase II evaluations into one single trial, and therefore allows for a gain in the development timelines of heterologous prime-boost vaccine strategies. Continuous safety monitoring with Bayesian methods is implemented for the phase I evaluation of a new candidate vaccine, with the aim to stop the administrations of this vaccine before the boost phase, should it not be safe enough. This approach is justified if there is prior knowledge making it very likely that the new vaccine is indeed safe. In our portfolio this is the case, since the new vaccine consists of a vector that has been extensively studied before, and the novelty only concerns the HIV antigen insert used. Moreover, our design is a randomised multi-arm trial allowing for an unbiased evaluation of several vaccine strategies in parallel in the phase II part of the trial.
How do you hope your proposed methodology will impact the clinical research community?
We hope that our approach is useful for clinical researchers working on the development of complex vaccine strategies, which not only concern the HIV field, but for instance also malaria or tuberculosis vaccine research. In cases where our proposed design is not directly adoptable in a given research context, we hope that our publication gives an impetus to researchers to search for an appropriate ‘optimised’ design and to thoroughly evaluate its features before implementation.
Do you think some aspects of the early stage design put forth in your study could be extended to phase III/IV trials?
Continuous safety monitoring methods, which rely on sequential statistical methods, could be relevant for any stage of clinical research, in which there is a particular concern for the safety of an intervention. This is mostly the case in early stage development, where uncertainty about the properties of the evaluated intervention are usually higher than in later stage development, but in some circumstances it could also be warranted in phase III or IV trials.
The other aspect of our trial is the multi-arm design, which is a method also used in later stage development.
Do you think this study design could be applied to other vaccine development trials, including those requiring dose escalation?
In trials requiring dose escalation, a randomised design from the beginning is difficult to conceptualise since the phase II dose is unknown at trial start. Moreover, in our opinion, thorough methodological considerations are required as to whether dose escalation is suitable in vaccine development and if so, which method should be used. Indeed, most dose escalation designs have been developed in the context of therapeutic interventions, and their appropriateness for prophylactic vaccine research needs to be evaluated, given that a) prophylactic candidate vaccines are generally very well tolerated so it is plausible to not observe any major toxicity event in the considered dose range; b) if toxicities are observed, it is not clear whether the classical assumption of increasing toxicity with increasing dose also holds for vaccines; and c) and not only toxicity but also immunogenicity may need to be taken into account in order to define the dose range for subsequent development. This should be subject to methodological research before considering integration of dose escalation into phase II HIV vaccine designs.
Do you think an effective prophylactic vaccine strategy against HIV will emerge in the near future?
Many promising results have been obtained in preclinical studies of prophylactic HIV vaccine strategies in recent years, and there is an increasing understanding of the immune mechanisms likely to be required for an efficacious vaccine strategy. It is thus possible that proof of concept of an efficacious strategy in humans is possible in the future. However, even with optimised or adaptive clinical trial designs, the full development of a HIV vaccine strategy is lengthy due to the fact that large sample sizes and long follow-up times are required to demonstrate efficacy with regards to a HIV acquisition endpoint in phase IIB/III. Developing a strategy that not only provides proof of concept but also has the desired properties allowing for large-scale roll-out on an operational level is likely to require more time.
What’s next for your research?
Our current work focuses on the definition of immunogenicity endpoints in early stage HIV vaccine trials in the absence of validated surrogate immunogenicity markers. Our approach includes modeling the dynamics of the different types of immune responses to HIV vaccines over time in order to better understand their temporal patterns and interrelationships, thus allowing us to better define which immunogenicity markers should be assessed and at what time points in order to get the most information from early stage vaccine trials. Modeling HIV vaccine response could help in performing in silico trials.