Is AstraZeneca vaccine associated with more blood clot than Pfizer’s mRNA vaccine?

A friend of mine is hesitant to take an AstraZeneca (AZ) jab, because he believes that the vaccine is less effective and has higher risk of blood clot than Pfizer’s mRNA vaccine. He is not alone, many Australians, including the educated ones, hold the same view. I beg to differ. In this note, I point out that the two vaccines have similar incidence of blood clot and almost identical effectiveness.

In Australia, the AZ vaccine has a rather bad publicity. After a few days of rolling out the vaccine in April 2021, a 48 yr old woman had died from blood clot after receiving the first dose of the vaccine. Since then, there have been occasional reports of blood clot and even deaths after vaccination. Most newspapers use the phrase of ‘linked to the AstraZeneca vaccine’, and I consider that the word link has been incorrectly used [1].

However, in reality, the incidence of blood clot is very rare. According to TGA (Australian equivalent FDA), since the roll out (April, 2021) until now (mid July, 2021) there have been 5 deaths that were associated with blood clot after 10.1 million doses of Covid-19 vaccines. Do the math yourself to see the rarity. Although TGA does not break down the number of deaths according to vaccine, the number clearly shows that the incidence is very rare.

Nevertheless, the frequency of media reports accompanied by a deterministic language tends to cloud our judgment. Many people in the community, including the highly educated ones, believe that AZ vaccine is less effective and has more side effects than mRNA vaccine (Pfizer and Moderna). One of my friends, a doctor, pointed out that the efficacy of Pfizer vaccine is 95% which is numerically higher than that of AZ vaccine (72%), and he decided to wait for the mRNA vaccine. Other friends argue that there are more reports of blood clot linked with AZ vaccine than to Pfizer vaccine. Sound convincing?

Well, no.

First, let us consider the issue of blood clot. According to the European Medicines Agency, until 4/4/2021, approximately 34 million European people have received an AZ jab, and among whom, 169 cases of blood clot were reported [2]. So, the incidence of blood clot associated with AZ vaccine is 5 per million.

What about blood clot associated with Pfizer vaccine? An Oxford University group examined the data pertaining to Covid-19 from 20/1/2020 to 25/3/2021 that included 366,869 people who had received a Pfizer vaccine jab [3]. They found that the incidence of blood clot associated with Pfizer vaccine was 6 per million.

So, as you can see, the risk of blood clot is very low. Moreover, there was no real difference between the two vaccines in terms of blood clot risk. Actually, one can argue that the blood clot risk associated with AZ vaccine was lower than Pfizer vaccine, but I don’t encourage that kind of comparison.

Talking about comparison, let me now explain why it is not possible to compare vaccine efficacy between studies. The efficacy of each vaccine is tested in a randomized controlled trial (RCT), in which participants are randomly assigned into either a control or a vaccine group. After following up the participants for a pre-determined period, researchers record the incidence of new Covid-19 cases. Let the proportion of infection among those in the vaccinated group and placebo be R1 and R0, respectively, the vaccine efficacy (VE) is estimated as:

VE = 1 — (R1 / R0)

As can be seen from the above formulation, in order to work out VE, researchers need to have a control group. And, for vaccine efficacy that we are talking about, different trials were based on different control groups. The incidence of infection in a control group tells us something about the background risk of that testing population. The table below shows that there was a substantial difference in the infection rate across control groups for the Pfizer (0.9%), Moderna (1.3%), JJ (1.6%) and AZ (1.7%) trial.

Because VE estimate is, as a norm, dependent on a control group, it is not possible to say that VE of AZ vaccine (70%) is lower than that of Pfizer vaccine (95%). Comparing VE in such a way is like comparing apple and orange.

One way to validly compare VE between AZ and Pfizer vaccines is to conduct a trial with 3 groups: control, AZ vaccine, and Pfizer vaccine. In this trial, VE of AZ and Pfizer can be estimated by comparing with a common control, and as a result, their difference in VE can be statistically assessed. However, such a ‘head-to-head’ trial is unlikely to be carried out, because no drug company wants to see their vaccine has a numerically lower efficacy than the competitor’s vaccine, no matter how statistically trivial the difference is.

A more practical approach to know something about the degree of protection from vaccination is to look at real world data (also called ‘vaccine effectiveness’). Another Oxford group has done such a study (actually it was a survey), in which they analyzed data of 373,402 participants who had received AZ and Pfizer vaccines [4]. They did a lot of sophisticated analyses, but I summarize here a relevant and key finding:

· AZ vaccine reduced the odds of infection by 64% (95% confidence interval [CI], 55% to 70%) 21 days after the 1st dose.

· Pfizer vaccine reduced the odds of infection by 67% (95% confidence interval [CI], 61% to 72%) 21 days after the 1st dose.

There was no statistically significant (P = 0.90) between the two effectiveness figures (64% vs 67%).

So, in summary, these data allow me to say that: (a) the incidence of blood clot is very rare, and that there was no difference between AstraZeneca (5 per million) and Pfizer vaccines (6 per million); and (b) the effectiveness of AstraZeneca vaccine is almost identical to that of Pfizer vaccine.

You may ask what vaccine have I received. Well, I am glad to say that I have already had two jabs of AstraZeneca vaccine, and I have been fine.

____

[1] In scientific parlance, ‘link’ is a strong word, implying a causal relationship between exposure and outcome. In reality, scientists until now have no idea on the association between AZ vaccine and blood clot. Therefore, the word link is wrong in this context. We can only say that AZ vaccine may be associated with blood blot.

[2] https://www.ema.europa.eu/en/news/astrazenecas-covid-19-vaccine-ema-finds-possible-link-very-rare-cases-unusual-blood-clots-low-blood

[3] https://osf.io/a9jdq

Interestingly, in this study, the investigators found that the incidence of blood clot associated with Covid-19 was 42.8 per million individuals, 6.7-fold higher than the incidence associated with Pfizer vaccine.

[4] https://www.medrxiv.org/content/10.1101/2021.04.22.21255913v1.full.pdf

osteoporosis | epidemiology | genetics | biostatistics | data enthusiast