The short version: Even though the vaccines for SARS-CoV-2/COVID-19 were made at an unbelievably fast rate, we have no basis to believe that there were corners cut from the perspective of the science or the efficacy. The cutting of the time in research and development is the result of two things: the existence of an extensive body of research that indicated which strategies for a vaccine and targets were viable and inviable (as well as considerations for which kinds of vaccines could be made rapidly, given the pandemic), financial limitations that are normally present for pharmaceutical research were not applicable, there was an enormous pool of willing subjects for the vaccines, and there was extensive streamlining of the regulatory process on the part of the FDA. If these vaccine candidates make it through EUA and then licensure, we should have every reason to believe that these are safe and effective vaccines and should not be worried about getting them.
I’m seeing a lot of people very concerned about the pace at which the vaccines for COVID-19 have been produced, which I get. We were told it could take years, we were told vaccines normally take decades of research and development to work out all the kinks and make it to licensure, and now in less than a year, it looks as though we will have effective vaccines being distributed in the US and licensed some time thereafter. Even the mumps vaccine, in an era with far more lax regulations, took 4 years to make. Surely some chicanery is afoot? This could not have been done without cutting some corners! And my answer to that is: Yes, but actually no.
Yes, the vaccines were expedited, in that it was made a huge priority to get out an effective vaccine as quickly as possible and every reasonable step was taken to make that happen. But no, this does not in any way mean that corners were cut in the process of getting us there.
Let’s talk about the process of making a vaccine and why it takes so long. Also the WHO has a really nice document on the subject if you want a detailed look.
At this point most people have at least heard of the phases of the clinical trial process, but they are worth going over. Under normal, non-pandemic circumstances it goes something like this:
- Phase 0, Pre-Clinical (and Exploratory) Work: Studying the pathogen in vitro and in animal models, identifying viable leads for antigens, considering choice of adjuvant, considering the type of vaccine that would be appropriate. Testing the vaccine in a relevant animal model. This typically involves challenging the animal model with a dose of the pathogen (typically something much higher than what would be encountered in nature) to evaluate efficacy. There is nothing going to humans at this point. It is also at the preclinical stage that toxicity and safety testing is done. There are two phases for this: pre-trial requirements and in-trial studies. Pre-trial studies examine the safety, immunogenicity, reactivity, and toxicity. Once clinical trials begin, population-specific studies can be conducted including fertility and embryofetal development.
- Phase I Clinical Trials: The manufacturer files an investigational new drug application (IND) with the regulatory body at this point (in the US, the FDA). These are the first time the vaccine gets put into humans. These studies typically have on the order of 10–100 healthy volunteers who are given the vaccine candidate. The phase exists mainly to figure out the pharmacokinetics and pharmacodynamics of the pharmaceutical. Doses get calibrated to balance the adverse effect profile with the therapeutic (or prophylactic) benefit, typically by progressive escalation to the limit of tolerability in the form of SAD (single ascending dose, where subjects receive one dose of a drug) and MAD trials (multiple ascending dose, where subjects receive multiple doses of the drug), usually with about 10 subjects per group. For a vaccine, samples may be taken from patients to evaluate immunogenicity. Placebos may or may not be used, but as efficacy is not being measured, tend to be less important. The point of this phase is to function as a preliminary safety evaluation. The principal question being answered is essentially “Is this vaccine acutely toxic to the vaccinee?” and we really hope the answer is “No.” This is also where considerations start emerging for a dosing schedule.
- Phase II Clinical Trials: Similar to Phase 1 trials, the primary goal here is to evaluate safety and immunogenicity again in a larger group, usually a few hundred people. They have largely the same goals as far as figuring out how to dose the vaccine and the schedule for it. If this seems somewhat redundant with phase I, that’s because it is (more on this shortly).
- Phase III Clinical Trials: After confirming that your vaccine is safe and stimulates an immune response, it’s time to see if it actually works. You recruit thousands of people (usually at least 30,000, but the trials for the polio vaccine had 2 million people before it was approved) and randomize them to receive placebo or vaccine. The placebo depends on the type of vaccine. For example, today we give MMR-II, which is a second-generation MMR vaccine. MMR-II was compared with the first generation vaccine as placebo. This is because of a fundamental principle in ethics: you cannot withhold the standard of care. If already an effective therapy (or prophylaxis) exists, you cannot give someone placebo. The safety profile of the first-generation vaccine is at that point well understood because it has made it through clinical licensure and has been subject to post-marketing surveillance. I only expound upon this point because some people have an obsession with vaccines being compared to “true placebo.” Additionally, an important aspect of this stage is blinding. Subjects in the trial might modify their behavior if they know they are in the vaccine group (for instance, being less careful about risky behaviors), which they might infer from their reactions (e.g. if they get a fever), so it’s very valuable to have a control that has a similar side effect profile so you can get accurate information on efficacy. If you want further information, here’s a great read on that. Even though the fundamental goal of Phase III is an efficacy evaluation, safety is still being considered at every step. Participants in the trial are expected to report their experiences back to the manufacturer and that goes to a data safety monitoring board (DSMB), an independent organization responsible for evaluating the data in the trial to judge safety and efficacy. If the vaccine manages to pass this stage (which happens ONLY if it meets its primary endpoint), the manufacturer can petition for licensure.
- Petitioning for licensure via BLA (Biologics License Application): At this point the manufacturer compiles all the data and sends it to the regulatory body (in the case of the US, the FDA) who evaluates it. If approved, the vaccine can be added to the national vaccination program. After FDA approval, in the US it is also considered by the Advisory Committee on Immunization Practices (ACIP), an independent body of 15 experts, and
It’s also worth noting that vaccines are treated with substantially MORE rigor than prescription drugs are as far as their clinical development goes. Because they are given to healthy people (often babies and young children in particular, and society places an especially high value on their life compared with other members) to prevent disease, the risk-benefit gradient margin of tolerability is much narrower than for therapeutics. In other words, vaccines are judged much more harshly in terms of how safe they are than prescription medications are because vaccines are given to healthy people with the promise of preventing a disease, as opposed to sick people with the hopes of treating their disease.
So, given how things are in normal times, how can I claim that corners have not been cut in the production of COVID-19 vaccines when it’s taken less than a year? Simple.
One thing that doesn’t get discussed enough is how risky vaccine development is, in the financial sense. It is extraordinarily costly, easily billions of dollars, for a vaccine to be taken from the preclinical setting to be put into effect in a vaccination program. More than that, only 1 in 15 vaccines that pass Phase II clinical trials actually make it to licensure. There are a few reasons for that (I have a post that explains in greater depth why it’s hard to make an effective vaccine here):
- Even though vaccinology is transitioning into a more rigorous science where scientists work to understand the interactions of the pathogen with its host in detail, this is still very hard to do. Sometimes the incorrect antigens are targeted. Sometimes the incorrect correlate of protection is selected for. I discuss
- What works in a laboratory animal doesn’t always work the same way, or at all, in a human. Immune systems in particular (as compared to other body systems) are highly variable across species.
- Sometimes when all the components of a vaccine are combined, unexpected things happen (but that’s why we put them through clinical trials as an entire vaccine rather than test every single component). This tends to be most relevant, however, for combination vaccines e.g. MMR, MMRV, DTaP, DTaP-IPV-Hib-Hep B.
This was circumvented by the infusion of cash into vaccine manufacturers’ endeavors to produce a COVID-19 vaccine. The manufacturers had already begun producing vaccine before they knew it worked (at risk) because it was not THEIR money that was to be lost should the vaccine fail to meet its endpoints (note that some vaccine manufacturers did not accept money from the US government for this endeavor e.g. Pfizer). This Viewpoint in JAMA summarizes the role of Operation Warp Speed (still an utterly horrific name). It’s notable however that production of a vaccine will always occur at-risk because it usually begins while under review for BLA approval, and purchase agreements are put in place to protect the assets in the event that the BLA is rejected.
However, resources other than money played a role here. For this part, I am going to quote my friend Andrea Bailey (her qualifications to comment on this are stated at the end of the article):
It can take a considerable amount of planning to run a single small-scale clinical trial. There are a number of critical and time-consuming processes involved to set up a trial. This includes scheduling biolot production (and associated quality control), signing agreements and contracts, obtaining import/export permits to ship unlicensed product across borders, protocol development, agency approval, CRO scheduling, etc. A simple small-scale study can take months to arrange. Much of this time has been cut by prioritizing staff and resources
Basically every single project that needed the FDA’s attention was put on the back burner as everything to do with a vaccine for COVID-19 was pushed immediately to the front of the line. On top of this, the FDA allowed for seamless trials, meaning that if a vaccine met its phase I or II endpoints, it was permitted to immediately proceed to phase III, rather than deal with the regulatory rigamarole and paperwork required for normal progression.
There’s also another more subtle factor directly relating to the pandemic that allowed for expedited research and development. Normally, the regulatory process would require in-person meetings which can only occur if everyone can be present at a given time, and that’s tough with how busy these people are. The pandemic meant that people did not physically have to be in the same place for this: it could be done virtually. Furthermore, it can be very difficult to assemble a subject pool for these vaccine trials, but the exigency of the situation created circumstances in which there were a huge number of volunteers willing to
Earlier I said that phase II trials tend to be fairly redundant with Phase I trials. Phase II examines the vaccine for safety in a larger group but formally, it’s not adding new information, per se, beyond that the vaccine is safe in that larger group (you would get to see some of the rarer side effects but even then the sample size here doesn’t allow you to see the very rare ones), and thus more likely to be safe in the general population. The point is to increase the confidence of the vaccine manufacturer in the product so that they are willing to spend the millions of dollars on the production and approval. As Professor Paul Offit (an inventor of the rotavirus vaccine and member of VRBPAC, a regulatory agency associated with the FDA that will also evaluate the data of the vaccine candidates that are seeking licensure) stated in an interview about the subject:
As you move forward with these kinds of trials, you go from phase I to phase II to phase III, you just try to reduce uncertainty. That’s the goal. How much uncertainty do you need to reduce? Do we need to wait, for example, to see what the 1-year follow-up is, knowing that during that time hundreds of thousands of people may die? Should we wait 2 years, which William Haseltine, PhD, actually reasons: Why don’t we wait 2 years and get all the data as we would do for a typical vaccine? The answer is because with the rotavirus trials, 60 children died per year [as opposed to the 225,000 Americans who had died at the time of the interview from COVID-19]; that’s a little different from this.
Basically, from the perspective of the science and only the science, you could plug a vaccine into phase I and go straight to phase III after determining it isn’t acutely toxic. Or you could even go directly to a large Phase III trial (assuming you could recruit people for it).
Many pharmaceutical companies did a combined Phase 1 and 2 study e.g. Sanofi and GSK, which is an excellent compromise to accelerate production.
Another critical point that is worth mentioning is that the preclinical work of identifying leads for a vaccine normally takes a huge amount of time. This time it didn’t. Why? We knew from the past what strategies were likely to work with a vaccine against a coronavirus, what risks were probable, and what targets should be included in the vaccine. For instance, we knew that the spike protein was probably the best target for a vaccine (I imagine anyone who has taken a first course in immunology could have told you that), but we also knew from research on MERS-CoV that a double proline substitution locks the spike protein in a prefusion conformation. What this means is that the antigen that the immune system is intended to target, the pre-fusion form of the spike protein, is exposed for much longer, and indeed, this double proline substitution is used in the design of every vaccine candidate I could find data on.
It’s also worth noting that not all vaccine platforms are going to be suitable to respond to a pandemic. The platforms have to be rapidly scalable to meet production demands for example, and it was known in advance what type of vaccines would be effective for this and what kind weren’t practical. For example, setting aside the risks of a live attenuated vaccine for a coronavirus for a moment, it takes a long time to generate an attenuated strain. On the other hand in vitro transcribed RNA can be generated extremely quickly. There was some regulatory expediting at this stage too. Andrea told me:
To further expedite the development process the FDA allowed manufacturers to cross-reference to past BLAs for data that also supported their COVID vaccine development. For example, the manufacturer was permitted to cross-reference (by linking to the specific documents on file with the FDA) submitted safety data or proof of concept data to the FDA for another vaccine that used a similar platform, or any pre-developement data on new vaccine platforms that the agency had already reviewed. This means that they could avoid repeating studies that had already been conducted and reviewed.
So yes, the vaccines were rushed in that they are getting here much faster than anyone anticipated and via a different path, but there weren’t corners cut from the perspective of the science nor the safety. In fact, in the trial protocols for the phase 3 trials of the candidates currently in them, patients are being followed for 2 years after the receipt of the vaccine to confirm long-term safety (though I should point out this really isn’t the the risk people are making it).
So, we will have COVID-19 vaccines in record time, and we have every reason to believe that they will be effective (though for how long isn’t entirely clear yet) and safe.
I’d also like to include this short clip from the Oxford Vaccine Group:
I would like to credit and extend my sincerest gratitude to my dear friend, Andrea Bailey PhD, Senior Regulatory Affairs Specialist, for her thorough vetting of the content of this blog post for accuracy, and insightful suggestions for how to improve it. It would not have been possible without her.
I also need to thank Magdalen Rose Wind-Mozley for her immeasurably valuable insights on this process outside of the US context and valuable edits. The US accounts for < 5% of the world’s population, but this is a vaccine that everyone needs, which means it’s important that confidence be high across the entire world.
- Bailey, Andrea, Written Correspondence, 11/27/20
- Center for Biologics Evaluation, Research. 2019. Vaccines and Related Biological Products Advisory Committee. Fda.gov. https://www.fda.gov/advisory-committees/blood-vaccines-and-other-biologics/vaccines-and-related-biological-products-advisory-committee.
- Center for Biologics Evaluation, Research. 2020. Vaccine Development — 101. Fda.gov. https://www.fda.gov/vaccines-blood-biologics/development-approval-process-cber/vaccine-development-101.
- Cohen J. 2020. The $1 billion bet: Pharma giant and U.S. government team up in all-out coronavirus vaccine push. Science. doi:10.1126/science.abc0056. https://www.sciencemag.org/news/2020/03/1-billion-bet-pharma-giant-and-us-government-team-all-out-coronavirus-vaccine-push.
- Development and Licensure of Vaccines to Prevent COVID-19 Guidance for Industry. Fda.gov. https://www.fda.gov/media/139638/download.
- Ellis R. PPI advisory group. Nihr.ac.uk. 2020 Nov 12. https://oxfordbrc.nihr.ac.uk/ppi/ppi-advisory-group/
- Fact sheet: Data safety monitoring boards. 2014 Feb 13. Avac.org. https://www.avac.org/data-safety-monitoring-boards.
- Griffin R, Armstrong D. 2020 Nov 9. Pfizer vaccine’s funding came from Berlin, not Washington. Bloomberg News. https://www.bloomberg.com.
- Hagai, T., Chen, X., Miragaia, R.J. et al. Gene expression variability across cells and species shapes innate immunity. Nature 563, 197–202 (2018). https://doi.org/10.1038/s41586-018-0657-2
- Jiang S, He Y, Liu S. 2005. SARS Vaccine Development. Emerg Infect Dis. 11(7):1016.
- O’Callaghan KP, Blatz AM, Offit PA. 2020. Developing a SARS-CoV-2 vaccine at warp speed. JAMA. 324(5):437–438.
- Pallesen J, Wang N, Corbett KS, Wrapp D, Kirchdoerfer RN, Turner HL, Cottrell CA, Becker MM, Wang L, Shi W, et al. 2017. Immunogenicity and structures of a rationally designed prefusion MERS-CoV spike antigen. Proc Natl Acad Sci U S A. 114(35):E7348–E7357.
- Plotkin SA, Orenstein W, Offit DPA, Edwards KM. 2017. Plotkin’s Vaccines. 7th ed. Elsevier.
- Principles and considerations for adding a vaccine to a national immunization programme. Who.int. https://apps.who.int/iris/bitstream/handle/10665/111548/9789241506892_eng.pdf;jsessionid=E3B4D2C2110AA558F00D517543C10358?sequence=1.
- Sanofi and GSK initiate Phase 1/2 clinical trial of COVID-19 adjuvanted recombinant protein-based vaccine candidate. Sanofi.com. https://www.sanofi.com/en/media-room/press-releases/2020/2020-09-03-07-00-00.
- Singh K, Mehta S. 2016. The clinical development process for a novel preventive vaccine: An overview. J Postgrad Med. 62(1):4–11.
- Topol EJ. 2020 Oct 27. For COVID-19 vaccines, ACIP will be a critical gatekeeper. Medscape.com. https://www.medscape.com/viewarticle/938494.
- Vaccine Platforms: State of the Field and Looming Challenges. Centerforhealthsecurity.org. https://www.centerforhealthsecurity.org/our-work/pubs_archive/pubs-pdfs/2019/190423-OPP-platform-report.pdf.
- Wind-Mozley, Magdalen Rose, Written Correspondence, 11/27/20