Duty of care

Who in this setting is the “expert”? The volunteer could claim the expertise to define the limits of acceptability of a trial on ethical grounds, given that it is he who is consciously and knowledgeably taking the risk. The volunteer could argue that it is his assessment of the risk-benefit equation that must prevail. But should the triallist take a “paternalistic” role because of his obligation as the person who must take responsibility for trial conduct, including its ethics? What are the limits of beneficence in agreeing to the self-determination of a willing and well-informed volunteer? Whose perception of risk takes precedence? Whose “conflict of interest” should prevail? The clinical investigator surely still has a duty of care to ensure that trial participation does not create hazard for the patient, as well as enabling the best research opportunities to be taken.

Precipitateness is another issue in this unusual setting. The volunteers have put themselves forward in a way that could be construed as putting pressure on vaccine triallists to undertake such a study. That may indeed have been, in part, their intent. The urgency is real enough. Yet that urgency has been present for well over a decade. Pressure to take some bold steps, even short cuts, is certainly fuelled by the lack of progress with viral protein vaccines. Is there also a sense of global solidarity because of the increasing disparity between what can be offered as treatment to those in the industrialised, and to those in the developing, world? These arguments would appear to justify proceeding apace on the grounds of the ethics of emergency.

The balance in favour of attempting the live attenuated approach may have been tipped by the impressive effectiveness of combination anti-retroviral therapy. This could have been seen as providing something of a safety net for the vaccinees, if the vaccine strains proved to be, or reverted to, pathogenic strains. Is the urgency to proceed based on the need and on a perception of the available data, rather than its realistic appraisal? Can we genuinely identify non-pathogenic strains? Is there a sufficiency of data now to justify the view that immune protection can indeed be provided, or are we still too unclear about the nature, correlates and induction of such protection? Is a false sense of security being provided by the partial data that we have to date?

Yet one could argue that there will always be uncertainties, always more data to be gathered, but that ultimately the only way forward is to try something that we know from other settings to be the most likely way of achieving an effective vaccine. While some data can be obtained from animal experiments including those on primates, the reality is that HIV is a uniquely human pathogen. Therefore it is only in the setting of clinical trials that we are going to find out the vital answers. So why not just try it, if there are willing guinea pigs? The argument is an appealing offer to break the gridlock of incomplete and unsatisfactory knowledge derived from only partially relevant extrapolations. After all, Jenner didn't wait for the laboratory data to stack up before applying the original vaccine against smallpox. (However, Jenner did have a raft of human experience, almost literally “in the field”, to draw upon!)

If we do look at the available and relevant data, can we draw any conclusions? The Sidney cohort of transfusion recipients is a key group that can offer insights.³ In this group of transfusion recipients, and the relevant donor, the strain infecting them has a deletion in the nef gene. Their clinical course seemed to be non-progressive over a long period of observation. Primate studies had also indicated that deletions in the nef gene could render the virus less pathogenic in a model system.⁴ These two strands of evidence seemed to provide encouragement.

Timely reminder

However, the clinical data from Sidney are necessarily time-limited, and more recent studies on the cohort³ have indeed started to suggest that there is evidence of disease progression, or immune decline, after longer periods of observation. This is a timely reminder of the importance of a sufficiently long follow-up period before judgments are made about a disease that may, in the “wild”, take 10 years to cause disease in 50% of subjects. The possibility remains that the strain has reduced pathogenicity and hence a longer time to progression, and that this can lead to the discovery of an even less pathogenic strain for vaccine use. However, data from primate models suggest that the pathogenicity of nef deletion strains varies.⁵ Other data show that the capacity for recombination, as well as mutation, in HIV and related viruses is very real. Thus it seems dangerous, on theoretical grounds, to rely on an assumption the the vaccine strain will remain non-pathogenic.

As more data emerge therefore, the validity of the concept of an identifiable stable attenuated strain of HIV, rendering it suitable for vaccine use, is seen as premature. The data also underpin the need for very long term follow-up to validate any observational experience of wild-type variants before conclusions are drawn. This also highlights just how long the follow-up period of any trial would need to be before it could be judged safe. A vaccine safety trial (phase I) of the necessary ten to fifteen years' duration seems neither realistic, nor likely to be an acceptable pathway for those who would wish to stress the urgency. Yet how could one justify extrapolating from initial preliminary results of an apparently safe initial phase in order to proceed to large-scale phase II and III trials of efficacy, let alone safety? The Sidney cohort offers a brutal reminder of the dangers of reading the results too soon.

The balance of these various strands of argument is difficult to resolve and different views can be readily justified. There is no question that the gauntlet thrown down to the scientific community by the volunteers for live attenuated vaccines has sharpened the debate on vaccine development. Indeed, it has lent other approaches some added urgency. The volunteers have also challenged ethicists to re-examine some of their basic concepts in a very different setting, and have enhanced our insights into the conduct of clinical trials.

Perhaps an almost overwhelming argument against attempting such trials with the present state of knowledge, is that there is a real risk that a truly pathogenic strain could turn out to have been used, or could emerge over time, in the vaccinees. If that were to happen, even in one recipient, it would almost certainly mean that the live attenuated approach could never be tried again. So to proceed with this adventurous approach, even with willing volunteers, would probably jeopardise the possible validity of the concept and a potentially valuable approach to HIV vaccine development for all time. That heavy burden of responsibility suggests that the challenge should not be taken up at this time. Rather, the validity of the concept should be recognised and developed further in preclinical settings. The critical risk of the premature application of this approach, even in willing and altruistic volunteers, is not just the risk to the individuals, which they knowingly accept, but the significant threat to a whole strand of future investigation and application. The challenge has, however, facilitated an acceptance of the legitimacy of the concept of a live attenuated vaccine approach, which in itself is an achievement.

Author's note

This paper is based on a talk given at a conference organised by the Royal Free Centre for HIV medicine at the Royal College of Physicians, London in October 1998.