Which COVID vaccine is the best? Pfizer vs Moderna vs Covishield vs Covaxin?

 

Which COVID vaccine should I take? Pfizer vs Moderna vs Covishield vs Covaxin? Which is the best? 

In this post, I have tried to share the relevant information that I have accumulated which might help someone better think about and answer these questions for themselves.

The first section of this post is a primer to immunology for those of us who might be unfamiliar with it, followed by a little discussion about the process of vaccine development. Finally, I summarize information that I gathered after going through phase 3 clinical trial data of COVID vaccines from four different companies and also provide decision trees for each of these, constructed from their phase 3 data. It is a long post, and if you have limited time, then you can get the gist of this post by going through only figure 2 and section 3. However, I encourage you to read it in full to better understand section 3.

1. Understanding vaccines

A: Punching bag (vaccine) used to train the fighter (immune cells)

B: Fighter (immune cells) defeating the opponent (virus)

A fighter goes through A in the hope of B; similarly, we vaccinate ourselves in the hope that immune cells prevent us from getting infected if and when we get exposed to infecting agent or the  virus in case of COVID.

A vaccine is a weakened version of the infecting agent (virus) which is non-infecting but is capable of providing training to our immune cells and prepare the immune system for an attack by the virus. I must again highlight that vaccines are non-infecting and thus, do not warrant any concerns, just like the punching bag or other training equipment which can provide a similar resistance to the fighter as the opponent but can never really defeat our fighter in the above example.

What does this 'training of the immune cells' really mean? 

The job of the immune cells in our body is to attack (kill) any outside agents. 💡What is the first and foremost requirement to carry this out? Immune cells should be able to correctly identify these outside agents, otherwise, they might start attacking the cells of our own body (which is what happens in auto-immune diseases such as rheumatoid arthritis). The immune cells distinguish between our cells and foreign cells based on the proteins attached to the cell surface of these cells, known as 'antigens'. For an immune cell to attack a foreign cell, it should have a receptor that can bind to the foreign cell's antigen and thus, identify it. 'Training of immune cells' against an infection, say COVID means familiarizing our immune system with the antigens of the virus leading to the production of memory cells and antibodies which are capable of identifying the virus if and when it attacks. If you are interested in knowing further about our immune system, I will encourage you to read about T cells and B cells from other sources. To understand this post, you do not need to know any further details.

2. Vaccine development

Discovery --> Pre-clinical trials --> Clinical trials

Figure 1: Stages of vaccine discovery and development

We are all currently eager to get vaccinated for COVID. However, there are multiple vaccines available in the market and different governments are opting for different vaccines. On what basis do the governments (or if given a choice, you) make this decision?  It is based on the clinical trial data. The clinical trial data informs us about the efficacy and safety of the vaccine based on which governments or independent entities choose the vaccine. Let us now delve into what this 'efficacy' means and what relevance does the debate about the difference between target and tested population for the vaccine has in this context.

What happens in the clinical trial?

Several people consent to participate in the clinical trial. Once we have enough participants, they are divided into two groups. One group is injected with the vaccine and we call it the experimental group. The other group is not injected with vaccine and we call it the control group. These are called Randomized Controlled Trials (RCT). After some time (months), we look at how many people in experimental and control groups got infected. The comparison of the number of infections in the experimental group and control group gives us an idea of how effective the vaccine is in fighting the infection.

What is the need of the control group? 

Consider 100 healthy individuals today and that we want to look at the number of infections in this group after one month. As a result of day-to-day activities and random processes, a fraction of these individuals would get infected, say 5 individuals while the rest would remain healthy, that is 95 individuals, by the end of 1 month period. Now, had we vaccinated these people and even if the vaccine did nothing (that is vaccine was ineffective), we would still observe 95 people as uninfected by the end of 1 month period. Thus, for a fair assessment of the vaccine's efficacy, we need to compare the number of infected individuals in the vaccinated group with that in the unvaccinated group. 

Refer to the following figure for a better understanding of the clinical trial:

Figure 2: Randomized control trial in the clinical trial stage of vaccine development

Here, Nvac: Total number of people who got vaccinated

Ncon: Total number of people who did not get vaccinated

In this example, Nvac = Ncon = N

Here, E: Efficacy

The probability of getting infected after getting vaccinated is (1 - E) times the probability of getting infected if you don't get vaccinated. It should be noted that the efficacy of the vaccine is a measure that is dependent on the baseline infection rate of a population. So, if a vaccine has 95% efficacy and the infection rate in a country is 2% (i.e. baseline probability of getting infected in 0.02), then the probability that you would get infected even after getting vaccinated is (1-0.95)*(0.02) = 0.001. 

This means that if 20 in every 1000 individuals are getting infected, if you vaccinate everyone using a vaccine of efficacy 95%, then only 1 person in every 1000 individuals would get infected.

As stated earlier the likelihood of getting infected even after getting vaccinated is dependent on baseline infection rate in a population. Thus, for gauging the utility of the vaccine for a particular population, it is important that we consider whether the population on which the vaccine was tested (that is, the population based on whose data the efficacy was calculated) and the population to which vaccine is to be administered is comparable. The efficacy calculated based on testing on a particular population might not be a good indicator of how useful the vaccine might be in bringing down the infection rates in a different population with a much lower or much higher baseline infection rate as compared to the test population. The infection rate would depend not only on the baseline immunity of the population but also on the lifestyle of people (mask compliance and hand washing).

3. Efficacy and safety of different COVID vaccines available in the market

3.1 Covishield (AZD1222) : Oxford/AstraZeneca vaccine [2]

Test population: UK, Brazil, and South Africa

Size of test population ('2N' in the figure 2) = 23848

Efficacy = 70.4%

Baseline infection rate in the test population = 14.92%

So for a population comparable to the test population, the probability of getting infected by COVID even after getting vaccinated with Covishield is (1 - 0.704)*0.1492 = 0.044

(So, if in a population, 149 in every 1000 individuals are getting infected, then if everyone is vaccinated with Covishield, only 44 in every 1000 individuals would get infected)

Serious adverse events related to the vaccine: Transverse myelitis in one individual

If you identify with the above test population, the decision tree for you is:

Figure 3: Decision tree for getting vaccinated with Covishield

3.2 Pfizer vaccine (BNT162b2) [3]

Test population: USA, Brazil, South Africa, Argentina, Germany, Turkey

Size of test population ('2N' in the figure 2) = 40137

Efficacy = 94.6% 

Baseline infection rate in the test population = 0.838%

So for a population comparable to the test population, the probability of getting infected by COVID even after getting vaccinated with Pfizer vaccine is (1 - 0.946)*0.00838 = 0.00045

(So, if in a population, 8 in every 1000 individuals are getting infected, then if everyone is vaccinated with Pfizer vaccine, only 0.0045 in every 1000 individuals would get infected.)

Serious adverse events related to the vaccine: They do not report the analysis of adverse events that can be attributed to vaccine. Overall, 21% of individuals in vaccinated group and 5% of individuals in control group had side-effects related to injection, such as pain at injection site, swelling etc. The probability of serious adverse events in the decision tree below is based on 4 events that they report as relating to vaccine but it is not clear if it is based on any analysis that was done to figure if vaccination can be deemed as the cause of the events.

If you identify with the above test population, the decision tree for you is:

Figure 4: Decision tree for getting vaccinated with Pfizer vaccine

3.3 Moderna vaccine [4]

Test population: USA

Size of test population ('2N' in the figure 2) = 28207

Efficacy = 94.1% 

Baseline infection rate in the test population = 1.315%

So for a population comparable to the test population, the probability of getting infected by COVID even after getting vaccinated with Moderna vaccine is (1 - 0.941)*0.01315 = 0.00077

(So, if in a population, 13 in every 1000 individuals are getting infected, then if everyone is vaccinated with Moderna vaccine, only 0.77 in every 1000 individuals would get infected.)

Serious adverse events related to the vaccine: They do not report the analysis of adverse events that can be attributed to vaccine. The serious adverse events reported in both, experimental and control groups can found in table S15 in [6]. 

If you identify with the above test population, the decision tree for you is:

Figure 5: Decision tree for getting vaccinated with Moderna vaccine

3.4 Covaxin: Developed by Bharat Biotech [5]

Test population: India

Size of test population ('2N' in the figure 2) = 25800

Efficacy = 80.6% 

(this is from the interim analysis shared on their website; they are yet to publish their results in a peer-reviewed journal)

Baseline infection rate in the test population = 0.279%

So for a population comparable to the test population, the probability of getting infected by COVID even after getting vaccinated with Covaxin vaccine is (1 - 0.806)*0.00279 = 0.00054

(So, if in a population, 3 in every 1000 individuals are getting infected, then if everyone is vaccinated with Covaxin vaccine, only 0.54 in every 1000 individuals would get infected.)

Serious adverse events related to the vaccine: Not yet published.

If you identify with the above test population, the decision-tree for you is:

Figure 6: Decision tree for getting vaccinated with Covaxin

References:

1. Immunology courses in my undergraduate career at Indian Institute of Technology, Madras and my experience as a research scientist at a consultancy firm providing modeling and simulation services to pharmaceutical companies

2. Voysey, M., Clemens, S. A. C., Madhi, S. A., Weckx, L. Y., Folegatti, P. M., Aley, P. K., ... & Bijker, E. (2021). Safety and efficacy of the ChAdOx1 nCoV-19 vaccine (AZD1222) against SARS-CoV-2: an interim analysis of four randomised controlled trials in Brazil, South Africa, and the UK. The Lancet, 397(10269), 99-111.

3. Polack, F. P., Thomas, S. J., Kitchin, N., Absalon, J., Gurtman, A., Lockhart, S., ... & Gruber, W. C. (2020). Safety and efficacy of the BNT162b2 mRNA Covid-19 vaccine. New England Journal of Medicine, 383(27), 2603-2615.

4. Baden, L. R., El Sahly, H. M., Essink, B., Kotloff, K., Frey, S., Novak, R., ... & Zaks, T. (2021). Efficacy and safety of the mRNA-1273 SARS-CoV-2 vaccine. New England Journal of Medicine, 384(5), 403-416.

5. https://www.bharatbiotech.com/images/press/covaxin-phase3-efficacy-results.pdf

6. https://www.nejm.org/doi/suppl/10.1056/NEJMoa2035389/suppl_file/nejmoa2035389_appendix.pdf

7. Punch bag illustration by Emma Preston: https://emma_preston.artstation.com/

8. Fighters match illustration by Pa Trashu:  https://www.pinterest.com/pin/372743306635724531/,