The size and pace of development of the COVID-19 vaccine product pipeline is unprecedented, expanding to more than 200 candidates in less than a year since the discovery of the pathogen in humans. The scale of the research efforts to date are commendable, but the absence of benchmarks on the durability, safety, and immunogenicity of emerging candidates makes it challenging to ascertain the quality of the COVID-19 vaccine portfolio. By reviewing the early-stage clinical evidence generated by the front running investigational COVID-19 vaccines, we aim to provide a snapshot of the performance of the eight vaccines with published clinical trial results, helping to standardise and decipher the varying presentations (and interpretations) of the findings, and analyse how candidates are progressing in the context of an accelerated development pathway.

*Data in table as of 8 September 2020

What are the critical parameters for understanding the progress of COVID-19 vaccine development efforts?

There are a lot of firsts when you look at the COVID-19 vaccine pipeline, including a never-before-tried approach to rapidly developing health technologies in the paradigm of an ongoing global pandemic. In contrast to the traditional pathway with sequential milestones, the accelerated development of COVID-19 vaccine candidates is seeing clinical trials, manufacturing scale-up and procurement negotiations all taking place in parallel. To capture the complete picture, our COVID-19 vaccines clinical trial results table above provides an overview of multiple facets of the COVID-19 vaccine development lifecycle, including:

• The current and future manufacturing scale, related investments, and advanced purchase commitments (including price estimates and global access commitments)
• Characteristics of the clinical trials including size and age of the participants
• Safety and immunogenicity profile of the front running vaccines based on the published early-stage clinical trial results

How to interpret the data

The data in the table below is structured in a cross-tabulation format with the individual vaccine candidates listed as column headers and the different development parameters as the row headers.

For ease of reading the data, the table has 5 pages:

• Page 1: vaccine characteristics and current status
• Page 2: safety and immunogenicity
• Page 3: all parameters
• Page 4: publication source
• Page 5: glossary

Data curation methodology

The manufacturing and access-related information is gathered through announcements made by uni- and multi-lateral initiatives such as the U.S. Government’s Operation Warp Speed (OWS) and COVAX (the vaccines pillar of the WHO, Gavi, and CEPI-led Access to COVID-19 Tools (ACT) Accelerator); developer websites; press releases and security and exchange commission filings. Clinical trial design-related information is sourced from reputable trial registries such as ClinicalTrials.gov.

All information related to the clinical trial results is taken directly from the original publications, including peer-reviewed and advanced manuscripts. Wherever possible, we have included originally-presented data analysis from the published trial results in our table. However, for the sake of an apple-to-apple comparison, in some instances, we carried out minor calculations to the safety data based on the values presented in the original publications. All self-calculated data points are highlighted with an asterisk symbol in the table.

The current state of clinical evidence based on the early-to-mid-stage vaccine trial results

As of 9 September, results from eight investigational COVID-19 vaccines have been shared through either a peer-reviewed publication (6) or in the form of an advanced manuscript (2). Data readouts for three vaccines are from Phase II, three from Phase I/II and the remaining two from Phase I, and the depth of the data shared ranges from complete to a preliminary analysis. For the five (out of eight) vaccines that have moved into pivotal Phase III trials, we have limited our analysis (and the accompanying data) to the dose and schedule chosen by the developer for the Phase III trial.

The vaccine concepts adopted by the developers are quite diverse – three candidates use more traditional, tried-and-tested concepts, including two inactivated and one protein subunit approach. The remaining candidates utilise vaccine concepts with limited real-world experience such as replicating and non-replicating viral vector or approaches yet to be approved, including DNA and RNA platforms. Similarly, the choice of target antigen ranged from full-length spike to whole cell. When combined, the safety and immunogenicity data reported is from a total of 1,361 participants; at the individual trial level, the cohort size ranged from 42 (Moderna) to 472 (Sinovac) participants. For the most part, the age group of the trial participants ranged between 18 to 60, with only one trial reporting data on a cohort over 65 years (Pfizer/BioNTech), the highest-risk age group for severe COVID-19 infection.

According to WHO, a single dose COVID-19 vaccine is ideal, however only one (CanSino’s adeno-vectored vaccine) of the seven candidates uses a prime-only regimen, and the remaining candidates are either investigated as a prime-boost or a combination. For both safety and immunogenicity endpoints, the reported data is based on a follow-up ranging from 14 to 56 days after the last dose. The safety profile of all vaccines appears to be within an acceptable limit, as the vast majority of adverse events were reported as mild-to-moderate. For all vaccines, seroconversion took place between day 14 and 28, with 92% (Sinovac’s inactivated vaccine) to 100% (CNBG Wuhan’s inactivated vaccine, Moderna and Pfizer’s mRNA vaccine, Gamaleya’s adeno-vectored vaccine, Novavax’s sub-unit candidate and Oxford’s chimpanzee adeno-vectored vaccine) of the participants mounting a binding antibody response. The neutralising antibody response matched the binding antibody response for all but one candidate, with the CanSino’s adeno-vectored candidate showing a low rate ranging between 47-59%. For all vaccines with a prime-boost regimen, the overall antibody response supports a need for a two-dose vaccination schedule.

The only trial (Pfizer/BioNTech) to report on any data from an over-65 age group showed a significantly lower antibody response in participants between the ages of 65-85 compared to those in the 18-55 age group. Given those over 70 years of age are a high-risk group for COVID-19 related morbidity and mortality, more data from this age group cohort is urgently needed. Based on the published data, not all developers measured or reported on corresponding cellular immune responses. For those that did, the T-cell response appears to be biased towards T1 helper cells, which is reassuring as the T2 helper cells are most often associated with an antibody-dependent enhancement phenomenon.

The correlates of protection for a COVID-19 vaccine are yet to be defined. Some of the preliminary research, including non-human primates (NHP) studies, suggests neutralising antibodies (nAb) play a critical role in preventing severe forms of the disease, if not providing a sterilising immunity. At least six vaccine candidates, three of which are included in this analysis, have undergone NHP challenge studies. The pooled evidence from these studies ranged from complete protection in the lower respiratory tract to near-complete protection at the level of the nasopharynx, suggesting nAb titre could serve as a useful biomarker of efficacy. If one uses nAb as a surrogate of protection, then judging by the published data, and taking into consideration the guidance provided by WHO and the US FDA, most vaccines (including Oxford’s single-dose regimen) should at least prevent an infected person progressing to severe COVID-19 disease.

As mentioned above, at best, the evidence generated so far is from a maximum follow-up of 56 days; therefore, the durability of protection remains unknown. Data from the human adenovirus vectored vaccine (CanSino’s Ad-5 based vaccine) also strongly suggests pre-existing immunity to the viral vector can potentially have a negative impact on the vaccine-induced immune response, which could call into question the universal efficacy of vaccine candidates utilising this approach. That said, no such correlation was observed in the Gamaleya’s heterologous prime-boost adeno-vectored vaccine. Finally, the data from the only sub-unit vaccine investigation published so far indicates adjuvants may be necessary for protein vaccines.

Vaccine development is typically a long and costly process with no guarantee of success and high risk of failure. Vaccine development for epidemic infectious agents can be even more unpredictable, especially if the infectious agent is new to science. Even if a vaccine candidate appears suitable, it must remain safe and efficacious from early development through to large clinical trials and post-registration surveillance. As COVID-19 vaccine candidates move towards the pivotal efficacy stage of clinical development (seven out of the eight vaccines included in this analysis have either begun or are about to start recruiting participants for efficacy trials) and as we set our sights towards potential authorised use, it will be vital to ensure that outcomes for at-risk populations are adequately measured, as well as to understand the expectations of the regulators. For a vaccine to receive the US FDA’s emergency use authorisation or WHO prequalification, it must, at a minimum, be at least 50% effective, with the efficacy endpoint including protection against severe disease along with preventing infection.

How do the front-running candidates fare against the WHO COVID-19 vaccine target product profile (TPP)?

In April 2020, WHO published a COVID-19 vaccine target product profile (TPP) outlining the preferred and minimum acceptable characteristics for a potential future vaccine, including programmatic suitability. The TPP includes both outbreak and routine-use scenarios covering a wide range of attributes, including efficacy, target population and safety. As the vaccines included in this analysis have not finished all development milestones, we are only able to comment on their performance against some of the TPP attributes (e.g. information on efficacy is not yet available).

It is quite likely, one or more of these eight vaccines will be among the first COVID-19 vaccines to be approved for use in the general population. A significant proportion of them fail to meet a few key TPP attributes. At least three of them – Moderna, Pfizer/BioNTech and Oxford – do not meet the minimum product stability and storage criteria (shelf life of at least 6-12 months as low as -60 to -70°C and demonstration of at least 2-week stability at 2- 8°C), and all but one (CanSino) fail to meet the preferred single-dose regimen criteria for use in outbreak settings. Therefore, a word of caution is needed – for a broader public health impact, and to effectively mitigate the COVID-19 pandemic will require a vaccine which is appropriate in diverse contexts, including resource-limited settings and different age groups; efficacy as the sole criteria most probably will not be sufficient.

What are the next steps for the front-running candidates

Five candidates, including CNBG Wuhan, Moderna, Oxford, Pfizer/BioNTech and Sinovac are currently in Phase II/III or III trials, and CanSino and Gamaleya candidates are expected to start enrolling participants for a pivotal efficacy trial soon. Recruitment for a Phase II trial of Novavax’s candidate is currently ongoing.

As noted above, taking the pandemic approach for development, most developers have scaled up manufacturing in advance of the efficacy trial results. The Oxford candidate leads the way with an estimated capacity of 2.94 billion doses by the end of 2021, followed by Novavax (1.35b), Pfizer/BioNTech (1.3b), Moderna (0.95b), and Sinovac (350m).

To provide eventual access to one or multiple successful candidates, several advanced market commitments are in progress involving national, regional and international initiatives, including the US government’s Operation Warp Speed (800m), the Gavi/CEPI led COVAX facility (300m) and the European Union’s Emergency Support Instrument (300m). As more and more agreements between individual countries and developers are signed each day, there is a genuine fear of nationalist priorities negatively impacting fair and equitable access to COVID-19 vaccines. As an example, wealthy nations, such as the UK, have pre-secured approximately five doses per person, whereas in comparison, 92 low- and middle-income countries are projected to have access to less than half-a-dose per person through the COVAX facility.

Clinical development phases of vaccine development lifecycle

Phase I

• A phase of research to describe clinical trials that focus on the safety of a vaccine and its ability to induce an immune response (immunogenicity). They are usually conducted with a small number of healthy volunteers over a short period of time. The researchers will record any adverse events (often across different dosages) experienced during and after administration of the vaccine by the volunteers.

Phase II

• A phase of research to describe clinical trials that expand the evaluation of the safety and immunogenicity of the vaccine and aims to identify the optimal dose and schedule. They usually involved hundreds or more volunteers in different populations at risk of acquiring the infection. These trials will often be double blind (from the researcher and volunteer) with a placebo group included. Short term adverse events are studied as with phase I trials.

Phase III

• A phase of research to describe clinical trials that assess how effective the vaccine is at preventing infection in a wide range of populations. They involve thousands of volunteers and record safety information for a longer period allowing for detailed study of adverse events. Phase III trials are usually conducted in a double- or single-blind, placebo-controlled and randomized manner. Ordinarily these trials last 1-4 years.

Regulatory approval

• Various national and international medical regulators will review applications to license the production of a vaccine once enough safety and efficacy data has been collated from the clinical development process. Medical regulators may interpret the information within the vaccine application differently leading to some vaccines or medicines being approved in one country but not another. In certain circumstances such as the COVID-19 pandemic regulators may allow medical products that are not fully approved to be used in specific situations based on the safety and efficacy data available. An example if the FDA Emergency Use Authorization (EUA).

Vaccine concepts

Traditional

Inactivated whole cell

• These vaccines contain the whole virus, but in a form that will not be able to replicate and cause disease in the vaccine recipient. This makes the vaccine safe for a broad range of people, but multiple doses may be required to maintain an immune response that can prevent infection over a lifetime. The polio vaccine is an example of an inactivated whole cell vaccine.

Live attenuated

• These vaccines use the whole virus, but in a modified (attenuated) form that that prevents them causing disease once a person is vaccinated. As the whole virus is used the immune system can clearly recognise the pathogen and mount a strong and long-lasting immune response. As these vaccines contain “live” virus it means they may not be suitable for those with weakened immune systems due to the risk of adverse reactions from viral replication. An example live attenuated vaccine is the MMR vaccine used against Measles, Mumps and Rubella.

Protein sub-unit

• These vaccines consist of fragments (proteins) of the virus which have been chosen to best stimulate the immune response against the pathogen. When presented to the immune system these fragments are known as antigens. As the vaccine only contains components of the virus, the virus cannot replicate and cause diseases. Multiple doses or “boosters” are often needed to maintain immunity.

Novel

DNA

• DNA vaccines are a newer type of vaccine, with none currently approved for human use. A section of DNA is created which encodes for one or more virus antigens. The DNA is typically in a plasmid (circular) format and needs to enter host (vaccine recipient) cells where it will eventually be converted into (protein) antigens and expressed on the surface of the cells. This expression by the host cells should in principle offer a robust and specific protective immune response. As this is a newer technology the risks are still unclear in humans.

Replication deficient and competent viral-vector

• These vaccines use a section of viral DNA (gene) delivered into the host cells via a specific virus (different to the virus that is being vaccinated against). Using a virus (known as a vector) as a vehicle to deliver the DNA into the host cells means specific host immune cells are targeted. Antigens can then be produced and recognised by the immune system in a similar manner to DNA vaccines. Vectors are either deficient (not able to multiply in the host) or competent (able to replicate in the host). The choice of vector will affect both the immune response generated and any possible disadvantages to the vaccine such as having pre-existing immunity to the viral vector. Two Ebola vaccines – one each based on replicating and non-replicating viral vector approaches – are the only licensed products based on these concepts.

RNA

• RNA vaccines are related to DNA vaccines as they also contain encoded information that is converted into protein antigens by host cells. These antigens are recognised by the immune system in a similar manner to DNA vaccines. RNA vaccines can also be used in different ways to DNA vaccines by producing full virus-like particles or whole antibodies specific to the targeted virus. RNA vaccines can be delivered to host cells in several ways such as using viral vectors, lipid nanoparticles or even direct injection without a vector. There are no RNA vaccines currently approved for human use.

Back to the COVID-19 R&D Tracker