Intended for healthcare professionals

Rapid Response:

AstraZeneca covid-19 vaccine is not “overwhelmingly the safest option” in all population groups

Dear Editor

In their editorial, Perera and Fletcher [1]⁠ report on a new study published in the BMJ [2]⁠ which demonstrates the link between AstraZeneca vaccination and thromboembolism. In their editorial subtitle they stress that, however, “vaccination remains overwhelmingly the safest option [compared to no vaccination]”. This may be true at the wider population level, but not in certain subsets of the population.

Perera and Fletcher quote explicitly from the study only the AstraZeneca vaccine induced cerebral venous thrombosis (CVT) rate of 2.5 in 100,000 – incidentally still 20 times higher than in the general population. They do not quote from the study the rate of all types of vaccine induced thrombosis events, which stands at 11 in 100,000 (13 in 100,000 for the 18-45 age group), a dramatic revision of the 1 in 250,000 reassuring figure given by the Medicines and Healthcare products Regulatory Agency (MHRA) [3]⁠, which once more is revealed as an underestimation [4]⁠.

Unfortunately, when considering specific population age groups, such as those under 45 years of age, a 13 in 100,000 thrombosis risk is hardly overwhelmingly lower than the risk of being infected and than dying from covid-19. In fact, the opposite is true. A healthy, say, 35-year-old white British woman has around 0.6 in 100,000 chances of getting and then dying from a covid-19 infection [5,6]⁠. Assuming a 100% effective vaccine, the woman’s risk of dying from covid-19 is also her chance of benefit from vaccination. Hence, the woman’s chance of benefit from vaccination is one order of magnitude lower than the CVT risk, and two order of magnitudes lower than all thrombosis events risk.

Even in the general 30-39 age group, including those with any relevant comorbidities, the risk of dying from covid-19 is only 7.4 in 100,000 [7]⁠, which has the same order of magnitude of CVT risk, and is one order of magnitude lower than all thrombosis events risk. Importantly, according to MHRA data [3]⁠, 1 in 4 thrombosis events lead to death. Unfortunately, there seem to be no data around the rate and nature of life-long disabilities associated with these events.

It is surprising that in most pieces mentioning “vaccine risks/benefits ratio” – which is in fact logically and linguistically wrong , as it should read “vaccine harms/benefits ratio” – the actual absolute value of the chance of benefit from covid-19 vaccines is almost routinely omitted. This gap is partially filled below with a brief harms-benefits analysis for the 30-39 age group.

Assuming a 100% effective AstraZeneca covid-19 vaccine, to prevent 1 covid-19 death in the 30-39 age group we would need to vaccinate around 13,514 (100,000/7.4) individuals who, even without the vaccine, would have not died form covid-19, but who would be nonetheless exposed to the vaccine’s known and unknown short and long term harms. Taking thrombosis as a currently known AstraZeneca vaccine harm, and using its 13 in 100,000 risk, for 1 covid-19 death prevented, nearly 2 (13 x 13,514 / 100,000) of those 13,514 individuals will have a thrombosis event because of the vaccine. Hence, it can be estimated that, for every 2 covid-19 deaths prevented through AstraZeneca vaccination, nearly 4 individuals, who would have survived a covid-19 infection without the vaccine, will get a thrombosis, and, using the MHRA rate of death from thrombosis, 1 of the 4 individuals will die from it. The other 3 of the 4 individuals might end-up with some degree of life-long disabilities.

One individual dying and three possibly ending up with disabilities from vaccine harms for every two deaths from covid-19 prevented with the vaccine does not seem a recipe for “first do no harm”. On this basis, would the MHRA consider it appropriate to have the AstraZeneca vaccine administered to healthy 30-39 year-old individuals?


[1] R. Perera, J. Fletcher, Thromboembolism and the Oxford-AstraZeneca vaccine, (2021) 1–2.
[2] A. Pottegård, L.C. Lund, Ø. Karlstad, J. Dahl, M. Andersen, J. Hallas, Ø. Lidegaard, G. Tapia, H.L. Gulseth, P.L. Ruiz, S.V. Watle, A.P. Mikkelsen, L. Pedersen, H.T. Sørensen, R.W. Thomsen, A. Hviid, Arterial events , venous thromboembolism , thrombocytopenia , and bleeding after vaccination with Oxford-AstraZeneca ChAdOx1-S in Denmark and Norway : population based cohort study, (2021) 1–10.
[3] Medicines and Healthcare products Regulatory Agency, Press release: MHRA issues new advice, concluding a possible link between COVID-19 Vaccine AstraZeneca and extremely rare, unlikely to occur blood clots, (2021). (accessed May 6, 2021).
[4] M.T. Suadoni, Pharmacovigilance needs improvement (Rapid response to ), BMJ. (2021).
[5] University of Oxford, Welcome to the QCovid® risk calculator, (2021). (accessed April 9, 2021).
[6] University of Oxford, Oxford leads development of risk prediction model for smarter COVID-19 shielding advice, (n.d.). (accessed April 9, 2021).
[7] Public Health England, COVID-19 confirmed deaths in England (to 31 January 2021): report, (2001).

Competing interests: No competing interests

06 May 2021
Marco Tullio Suadoni
Practice Development Nurse and Vaccination Centre Supervisor
Cambridgeshire Community Services NHS Trust
The Poynt, Poynters Road, Luton, LU4 0LA, UK