By Jonathan Pugh and Dominic Wilkinson.
The emergence of the Omicron variant has prompted a great deal of uncertainty. One significant area of uncertainty is the the extent to which the variant can escape the protection afforded by current vaccines. One early South African pre-print suggests that Omicron has more extensive Pfizer vaccine escape than the Beta variant. However, there is also data to suggest that existing vaccine booster doses can neutralise the Omicron variant. At this early stage, all this data has to be interpreted cautiously.
In any case, in light of the emergence of this variant, various developers have announced that they are pursuing Omicron-specific variant vaccines. Even if it turns out that it will not be necessary to roll-out variant-specific vaccines for Omicron, it is quite possible that future variants of the coronavirus will necessitate the roll-out of variant-specific vaccines.
Of course, the development of variant-specific vaccines is a significant scientific challenge. However, it is also one that gives rise to important ethical questions for regulators: what sort of evidence base should be deemed necessary for establishing sufficient safety and efficacy for the regulatory approval of these modified vaccines? In March 2021, the MHRA announced that variant-specific vaccines will not require brand new approval or “lengthy” clinical studies. In this post, we consider the ethical justification for this sort of strategy.
Precedents for Regulatory Approval of COVID Variant Vaccines
Naturally, the strength of the evidence that different trials can generate depends in large part on their size and design. In many cases though, trials that are able to provide more robust forms of evidence will often take a longer time to complete. This is a problem in the context of a pandemic, particularly if we were to face a more contagious and/or virulent variant capable of significant vaccine escape; in such circumstances, time is lives.
Accordingly, there is an ethical trade-off here between speedy regulatory approval (to minimise the loss of life to the new variant) and scientific certainty in the safety and efficacy of approved vaccines. We have faced versions of this ethical trade-off before in this pandemic, both with the initial approval of the COVID-19 vaccines, and the more recent approval of booster doses. Before considering the approach regulators that might take to the approval of variant-specific vaccines, it is instructive to consider the approach they took to these prior decisions.
The Pfizer and Astra Zeneca vaccines were authorised by the MHRA in December 2020, following data from ongoing blinded, randomised, placebo-controlled phase 3 trials with large numbers of participants. 43,448 participants received either a control or active intervention in the Pfizer trial; participant screening and randomisation began on 27th July 2020 and ended on 14th November 2020. The Astra Zeneca interim analysis of safety and efficacy pooled data from four trials of different phases (COV001; COV002; COV003;COV005) 23’848 participants were enrolled overall, and 11’636 participants were included in the interim primary efficacy analysis. The phase 2/3 studies that contributed to this interim analysis began in May and June 2020.
The rapid development and testing of these vaccines has been well-documented. Yet, even though the speed of this process was unprecedented, obtaining this sort of safety and efficacy data takes time. In the case of both vaccines, these phase 3 studies were only carried out following the prior accumulation of phase 1 and phase 2 data. Moreover, the primary efficacy end-point of these trials was confirmed, symptomatic COVID-19, the efficacy analysis could only be carried out once a sufficient number of participants had been infected with the virus in the real world.
Consider now the later regulatory approval of using these vaccines as boosters. The MHRA authorised the use of Pfizer and Astra Zeneca vaccines as ‘safe and effective’ booster doses on 9th September 2021. Although the MHRA announcement did not specify the particular safety and effectiveness data that informed their decision, the JCVI guidance on booster programmes (issued on 14th September) suggested that their guidance was informed by “the epidemiology of COVID-19 in the UK, mathematical modelling, data on vaccine safety and vaccine effectiveness, and data from trials undertaken to understand the immunological impact of booster vaccination”.
More specifically, whilst the JCVI guidance (understandably) referred to a number of works that were unpublished at the time, they also referred to some studies that either were or now are published. These include Flaxman et al.’s (2021) randomised controlled study on reactogenicity and immunogenicity of late second doses or third ‘booster’ doses; this trial was a sub-study of participants enrolled in COV001, one of the phase 1/2 trials for the Astra Zeneca vaccine. The JCVI also referred to the COV-BOOST trial (Munro et al., 2021), a blinded, multicentre, randomised, controlled, phase 2 trial investigating the safety and immunogenicity of seven vaccines as a potential booster. Flaxman et al.’s study enrolled 90 participants in its third dose cohort in mid-March, whilst the COV-BOOST trial screened 3’498 participants in June 2021, with a data lock date of August 19th.
Finally, it is also worth acknowledging that regulators frequently have to consider the authorisation of variant-specific vaccines in the context of the seasonal influenza virus. For instance, as Weir and Gruber detail, US vaccine manufacturers have to first obtain licensure for their vaccines by demonstrating their safety and efficacy (amongst other things). They must then submit supplements to their vaccine license prior to distributing ‘updated’ strain-dependent versions of their vaccines. Supplements for inactivated and recombinant protein seasonal vaccines do not require additional clinical data specific for the new strain. In contrast, live vaccines require data from small scale studies (around 300 participants) to establish that the vaccine adequately attenuates the virus.
Regulating Variant-Specific Vaccines
There are various degrees of evidence that regulators might require prior to approving variant-specific vaccines. Regulators could treat such vaccines as an entirely novel product, and subject it to the same high regulatory requirements of phase III safety and efficacy data. As detailed above, even when this process is expedited, this sort of regulatory approval would take a considerable amount of time. This is only justifiable if the benefit of the additional knowledge such trials would give us is proportionate to the very significant costs of taking time in a pandemic.
Conversely, it might be argued that regulators need not view variant specific vaccines as an entirely novel product if the alterations are sufficiently minor. If so, they might adopt a regulatory approach akin to the approach that they take to flu variant vaccines. As detailed above, they might not require additional clinical data specific for the new variant vaccine and rely solely on preclinical data. Alternatively, they might require only small-scale studies on correlates of protection (rather than prevention of confirmed infections) – as briefly described above, this is the approach that the FDA adopts for live flu variant vaccines.
The Access Consortium – a coalition of regulatory authorities from the UK (i.e. the MHRA), Australia, Canada, Singapore and Switzerland – has also roughly adopted this latter approach in its guidance on modified COVID variant-specific vaccines. Their guidance suggests that clinical efficacy studies are not required prior to approval, but that regulators should request ‘bridging immunogenicity data from a sufficient number of individuals’.
This approach balances the need for speed with a degree of scientific rigour. However, some have disputed whether immunogenicity studies, targeting correlates of clinical protection, will provide sufficiently robust evidence of a variant-specific vaccine’s clinical efficacy – even if the vaccine is effective in generating neutralising antibodies, there is still a concern that a given variant may be resistant to neutralisation.
In view of these concerns about immunogenicity studies, and the time-consuming nature of clinical efficacy trials, Rohrig and Eyal have recently argued that challenge trials could have a significant role to play in developing variant-specific vaccines. These trials, which would involve intentionally infecting participants with the variant of concern, are not as time-consuming as standard clinical efficacy trials: they do not require potentially lengthy periods of waiting for participants to encounter the virus strain in the real world. Furthermore, such trials could help to establish more robust correlates of protection for specific variants – this recall, is a potential obstacle to standard immunogenicity trials for virus variants. Indeed, challenge trials could avoid the issue of clinical proxies altogether by directly testing for clinical efficacy itself.
Challenge trials potentially provide an evidential avenue for achieving the best of both worlds in the development and regulatory approval of variant-specific vaccines: scientifically rigorous evidence of clinical efficacy achieved in a timely fashion. In this sense, they have advantages over standard clinical efficacy trials and immunogenicity trials. However, this is achieved at some cost, given the significant ethical debate about challenge trials over the course of the pandemic – concerns have been raised about exposing individuals to the risks of challenge trials, and the possibility of valid consent for such trials. Perhaps the most significant question here though is whether regulators should stick with the compromise between rigour and speed afforded by immunogenicity studies, or whether the additional benefit of the evidence that challenge trials can obtain is proportionate to the putative moral costs of these trials. If so, and even if this were not true earlier in the pandemic, it may be that now is the time to endorse challenge trials.
Authors: Jonathan Pugh and Dominic Wilkinson.
Affiliations: Oxford Uehiro Centre for Practical Ethics, University of Oxford
Social media accounts of post authors: @NeonatalEthics