Evaluating novel COVID-19 vaccines in the current chapter of the pandemic

The field of vaccine development against COVID-19 has rapidly evolved over the past 2 years. Different vaccine delivery platforms were used in different geographical areas, of which mRNA-based vaccines (BNT162b2 [Pfizer–BioNTech] and mRNA-1273 [Moderna]) and vector-based vaccines (Ad26.COV2.S [Johnson & Johnson] and ChAdOx1-S [Oxford–AstraZeneca]) were initially approved for use in Europe, Australia, and the USA. Later, a subunit S-protein-based vaccine (NVX-CoV2373 [Novavax]) was approved, mainly to be used as a booster vaccine. The adenovirus-based Gam-COVID-Vac (also known as Sputnik V, Gamaleya National Centre of Epidemiology and Microbiology, Moscow, Russia) was predominantly used in Russia and South America, whereas whole-virus inactivated adjuvanted vaccines (CoronaVac [Sinovac Biotech], BBIBP-CorV [Sinopharm], and Covaxin [Bharat Biotech]) were first approved in Asia and South America, and were used throughout those continents. Whole-virus inactivated vaccines have the advantage that they are relatively easy to produce (without the need for genetic modification) and are stable at refrigerated temperatures.

On June 24, 2022, the European Medicines Agency (EMA) granted full market authorisation to the whole-virus inactivated adjuvanted vaccine VLA2001 (Valneva), which is the first inactivated vaccine to be approved for use in Europe and was additionally approved in the UK, United Arab Emirates, and Bahrain. The approval was based on the interim results (up to day 43) of the randomised, controlled, phase 3 trial by Rajeka Lazarus and colleagues,¹ published in The Lancet Infectious Diseases. In this trial, the safety and immunogenicity of primary vaccination with two doses of VLA2001 was assessed in an immunobridging study including 4017 adult participants, with ChAdOx1-S as a comparator. Vaccination with VLA2001 led to significantly fewer solicited local or systemic adverse events than did ChAdOx1-S. Based on seroconversion rates on day 43 in adults aged 30 years and older, VLA2001 was non-inferior to ChAdOx1-S (both led to >95% seroconversion), but VLA2001 induced superior neutralising antibody titres (geometric mean titre [GMT] 803·5 [95% CI 748·5–862·6] in the VLA2001 group vs 576·6 [543·6–611·7] in the ChAdOx1-S group; GMT ratio 1·39 [95% CI 1·25–1·56]; p<0·0001).

Because of the geographical differences in the use of mRNA-based or vector-based vaccines compared with inactivated vaccines, clinical trials making direct comparisons between multiple vaccine platforms are extremely rare, making this study by Lazarus and colleagues unique. By contrast with Lazarus and colleagues' study, cross-sectional studies comparing primary regimens of ChAdOx1-S with another inactivated vaccine (Covaxin), or the mRNA-based BNT162b2 with BBIBP-CorV, showed that whole-virus inactivated vaccines were inferior to vector-based or mRNA-based vaccines when assessing antibody levels.2, 3 Interestingly, the difference in binding antibody levels between the two vaccine platforms in the study by Lazarus and colleagues was less pronounced than the difference between neutralising antibody levels, indicating that VLA2001 might induce more functional antibodies than ChAdOx1-S. Clear differences in T-cell responses were not observed, with the exception that whole-virus inactivated vaccines in general, and VLA2001 in the discussed study, induce broader responses than do vaccines that exclusively encode for the S protein, activating T cells that additionally target the nucleocapsid (N) and matrix (M) proteins.

In the current phase of the COVID-19 pandemic with many vaccine options now available, defining required endpoints in upcoming clinical trials that assess novel vaccines will be crucial. In our opinion, depending on the intended use of the vaccine, it is important to study the following factors: (1) immunogenicity in populations with pre-existing immunity, either induced by previous vaccination, natural infection, or a combination of both; (2) cross-reactivity of induced (neutralising) antibodies with novel, antigenically distinct SARS-CoV-2 variants; and (3) the breadth of the virus-specific T-cell response after (booster) vaccination. Notably, polyclonal T-cell responses do not seem to be affected by the mutations detected to date in the S protein, whereas these mutations do lead to at least partial escape from neutralising antibodies, making standardised T-cell assessments even more important.4, 5

Unfortunately, to date, clinical trials addressing the crucial endpoints we propose have not been performed for whole-virus inactivated vaccines. For example, the report of a phase 2 trial in which a third dose of CoronaVac was administered to CoronaVac-primed individuals clearly showed immunological recall responses but did not include an analysis of variant-specific antibodies or virus-specific T cells.⁶ In a direct comparison between BNT162b2 and CoronaVac booster vaccination in CoronaVac-primed individuals, restoration of omicron (B.1.1.529) BA.1 neutralisation was observed after BNT162b2 booster immunisation but not after a third dose of CoronaVac; again, virus-specific T-cell responses were not measured.⁷

Taken together, VLA2001 can be regarded a promising addition to the arsenal of COVID-19 vaccines. However, despite the positive findings of Lazarus and colleagues, it is important to note that the bridging with ChAdOx1-S might not be an optimal choice. ChAdOx1-S was shown to induce less virus-specific immune responses than the mRNA-based vaccines.⁴ Additionally, the usefulness of VLA2001 in the current phase of the pandemic remains to be determined through critical studies with VLA2001 in the intended target populations, thereby defining its position in the landscape of available vaccines.

We declare no competing interests. We thank Marion Koopmans for a critical review of the manuscript.