Is the Oxford/AZ COVID-19 Vaccine causing blood clots? — Deplatform Disease

The short version: I am skeptical that the blood clots that have been associated with the vaccine are caused by the vaccine (though I do think that vaccination cannot be excluded as a cause based on the information available). However, they are worth investigating further and we should have the patience to await the results of those investigations. Additionally, these clots are exceptionally rare and the benefits of the vaccine with respect to preventing severe COVID-19, preventing hospitalization, and preventing death vastly outweigh the potential risks of the clots: of 20 million individuals who received the vaccine in European countries, 8 cases of disseminated intravascular coagulation and 18 cases of cerebral venous sinus thrombosis in association with the vaccines have been reported to the European Medicines Agency (EMA). The risks posed by getting COVID-19 are substantially greater, and this vaccine does prevent COVID-19 and for that reason I feel that suspension of vaccination with the AZ vaccine was inappropriate. Please do read the EMA and Science articles on the subject as they make the issue very clear. The MHRA also has a discussion. For anyone who is still concerned and wondering whether or not the benefits of the vaccine outweigh the risks, Prof Sir David Spiegelhalter has contributed an excellent analysis (note that though he is focusing on the cases of disseminated intravascular coagulation you could easily observe a similar calculation for the cases of venous sinus thrombosis):

“The MRHA are investigating these five serious events that happened after 11,000,000 AZ vaccines.

“Let’s suppose the link is causal, although this is so far unproven. So assume there is around 1 in 2,000,000 chance of this severe event.

“Suppose 2,000,000 AZ vaccines are given in a week to middle-aged people, say 50-year-olds who are not at high risk from COVID-19. We would expect one such serious adverse event.

“But suppose these 2,000,000 people did not get their vaccine that week. At current UK rates, we might expect 2,000 to catch the virus — this would be higher elsewhere in Europe. As these are at lower risk, we might expect 20–30 to be hospitalised, and around 5 to die. So even for the week of the vaccination, the benefits far outweigh the risks.

“But that’s just one week. These unvaccinated people continue to have that risk for every subsequent week that their vaccination is delayed.

“So the benefit/risk balance seems firmly in favour of the vaccine, even if these events do turn out to be tied to the vaccine itself.”

If you are offered this vaccine and you have no contraindications, you should take it.

The EMA has recently reported that there have been a cluster of unusual blood clots in association with the COVID-19 vaccines, and this has understandably sparked concern. This is a very complex issue to analyze for several reasons, and truthfully I don’t find the hypotheses exploring the mechanisms for how this could be occurring to be very convincing (as I will explain). Firstly, it’s important to understand here that we have a significant epidemiological dilemma: we have a very rare outcome with a small signal in a near universal exposure (20 million individuals have received the vaccine and of these, 18 cases of cerebral sinus thrombosis and 7 cases of disseminated intravascular coagulation have been associated with these doses, wherein unfortunately 9 people died) which means we need a very high level of certainty to conclude a causal relationship. That is not to say that a causal relationship does NOT exist- I don’t know and I think no one currently can say with certainty and I do hope the investigation is able to give a definitive answer- but if it does, this still represents an extraordinarily rare outcome and in general, it does not affect whether or not most people should receive Oxford/AstraZeneca’s COVID-19 vaccines with the possible exception of certain individuals who have a significant predisposition to clots (discuss it with your care team). The Science article on the subject has an exceptional summary of the issue and I refer everyone to read theirs regarding the facts of the situation because hearing the soundbites only is not an effective way to understand the facts of the situation. Firstly, some brief words on the conditions in question:

While the incidence of cerebral venous sinus thromboses (CVSTs) is generally thought to be very low (up to 7 cases per million people per year; the EMA report notes that in the interval specified 1.35 cases may have been expected to follow vaccination but there were 7 reported), Capecchi et al do note that this is likely to be an underestimate because of the lack of well-designed epidemiological studies examining the incidence and also cite studies estimating incidence to be 13.5 and 15.7 per million people per year. It, therefore, cannot be excluded that the apparent spike in cases in this interval may just reflect enhanced surveillance of vaccinees. Note that this is significantly rarer however than clots (venous thromboembolism) in general; the CDC estimates 1–2 cases per 1000 people per year of blood clots generally (and for this reason I am very frustrated that these were reported as being just blood clots by the popular press, though it is understandable). Additionally, there was a single case of transverse sinus thrombosis reported for the Johnson and Johnson/Janssen vaccine (which is a hAd26 vaccine- this vaccine uses sAd25 and though they sound similar that’s a pretty important distinction) in a 25-year-old male vaccine recipient 21 days after vaccination which resolved; it is the only case in the entire trial and it is very difficult (if not impossible) to make a conclusion of causality based off of this single case but a relationship is not impossible. In my view, the most probable explanation here is in fact superior surveillance of a condition that is likely to be underestimated in prevalence because of the huge variation in the way it presents and the delay in symptoms following onset.

The epidemiology of disseminated intravascular coagulation (DIC) is complex and DIC is very serious; it is in fact a very important independent predictor of mortality and in fact platelet levels generally are a better predictor of mortality than either APACHE II or MODS scores in sepsis. It is frequently a complication of a number of conditions (see Box 139.1) and thus fairly common among hospitalized patients but among the general population it is very hard to estimate an incidence. Singh et al do find an overall incidence of 21.2 (95% CI, 10.6 37.9) per 100,000 persons per year, which is relatively rare. However, this risk depends on many factors and thus without knowing the precise composition of the individuals experiencing the DIC, it’s hard to provide more detailed commentary (and that would be protected under patient privacy). The EMA does report that in this interval 1 case of DIC would be expected while 5 were noted, but based on the statistic from Singh et al, the incidence of DIC would actually be well below the background rate. I will defer to the EMA data on this as they have more detailed information than me, however, and accept that the incidence is truly above the background rate. Again though, this is still an exceptionally rare event.

More recently, scientists in Germany have reported that these clotting abnormalities are due to heparin-induced thrombocytopenia (HIT). Heparin-induced thrombocytopenia is a condition in which individuals develop antibodies against the anticoagulant drug heparin (but typically due to unfractionated heparin, and the risk is lower but not zero with low molecular weight heparin, and even lower for heparin analogues) which complexes with a protein called platelet factor 4 (PF4). In this condition, platelets release platelet factor 4 (PF4) from their α-granules which bind the heparin (but also potentially other negatively charged polyanions like bacterial cell wall components) and antibodies can recognize the complex, leading to the activation of certain immune cells, including platelets. The platelets as a result promote clotting and get depleted. This condition is known to be a risk factor for the development of CVSTs. However, little information is given in the report other than that anti-PF4 antibodies were noted in these patients. These antibodies are common in the population and while their absence indicates a high probability that HIT is not present, their presence is not sufficient because these antibodies may not be strong activators of the platelets and they are common in the general population. The test here that would likely help provide a definitive answer is a serotonin-release assay wherein the patient’s platelets would be incubated with the antibodies to see if the platelets become activated, but in the NPR article at least this is not discussed (and I would go to the original source but I don’t speak German). The major issue I have here is understanding how this reaction could occur from a vaccine that lacks heparin or heparin-related molecules in it, but hopefully investigation will give some answers. HIT has been documented to occur after exposure to heparin following certain bacterial infections because PF4 binds some bacterial cell wall components potently and this can result in sensitization, but to the best of my knowledge this has not been documented with adenoviruses (and while it has been observed that HIV patients have an increased risk of HIT, HIV tends to be the exception in many of these cases rather than the rule so I would not use that as evidence that it can be caused by the vectors). The other issue here is the question of what the trigger for HIT is- normally it’s heparin or one of the closely related analogues, but as far as I am aware, there is nothing in the vaccines that resembles heparin that could act as a trigger. There are still many unanswered questions.

The trouble here ultimately comes down to this: people want an option that is risk-free and has only benefit, but that doesn’t exist. Vaccination generally comes pretty close to that because the risks of a serious adverse event are exceptionally rare, but I am willing to bet that a sizable number of individuals reading this have made a medical decision after considering the balance of risks and benefits to do something where the risks of clots is numerically much greater than that from any of these vaccines. For instance, consider the case of oral contraceptive pills. Combined OCPs raise the risk of a blood clot from 1–5 events per 10,000 person-years to 3–9 per 10,000 person-years- doubling to tripling that risk from baseline (to be clear, though this is a large relative increase, it’s not a massive absolute increase; it’s not as though the advent of OCPs sparked an epidemic of blood clots, though their use is a risk factor for the cerebral venous sinus thromboses that some vaccine recipients experienced). How then can we justify the use of OCPs given this risk? The risk has to be compared to the risk of not taking them. In this case, that’s pregnancy. Admittedly, for some people the small increased risk of a blood clot would be worth it to avoid pregnancy alone, but let’s pretend for the moment that the patient in question is value-neutral with respect to pregnancy and just wants to minimize their risk of clots while also being sexually active. The risk of a clot in pregnancy is 5–20 per 10,000 person-years, which is even higher than that from OCPs by more than double. Furthermore, in the post-partum period, that rises to as high as 40–65 clots per 10,000 person-years. Clearly then, if you have the potential to become pregnant and want to be sexually active but minimize your risk of blood clots, OCPs are a viable option. So you see, even if these events are real, a huge proportion of the population has already taken on what amounts to a much greater risk, and that’s entirely okay (this is not an attempt to dissuade people from the use of OCPs).

I will say, however, that these events have essentially no effect on my personal enthusiasm for the vaccine. I would happily roll up my sleeve for it for the simple reason that even if they are real: they are extremely rare. The risk of thromboembolic complications from COVID-19 is dramatically higher. I do think that individuals, especially those with an increased risk for developing these conditions, should be counseled on the possible association between the vaccine and these very rare clotting events. I would take the Oxford/AZ vaccine in a heartbeat if it were offered to me.


  1. Vogel, G. ‘It’s a very special picture.’ Why vaccine safety experts put the brakes on AstraZeneca’s COVID-19 vaccine. Science (2021) doi:10.1126/science.abi5259.
  2. Stam, J. Thrombosis of the cerebral veins and sinuses. N. Engl. J. Med. 352, 1791–1798 (2005).
  3. Warkentin, T. E. & Greinacher, A. Thrombosis of the cerebral veins and sinuses. N. Engl. J. Med. 353, 314–315 (2005).
  4. Capecchi, M., Abbattista, M. & Martinelli, I. Cerebral venous sinus thrombosis. J. Thromb. Haemost. 16, 1918–1931 (2018).
  5. CDC. Data and Statistics on Venous Thromboembolism. (2020).
  6. Salter, B. S. et al. Heparin-induced thrombocytopenia: A comprehensive clinical review. J. Am. Coll. Cardiol. 67, 2519–2532 (2016).
  7. Ferro, J. M. & Canhão, P. UpToDate | Cerebral venous thrombosis: Treatment and prognosis.
  8. Kunutsor, S. K. & Laukkanen, J. A. Incidence of venous and arterial thromboembolic complications in COVID-19: A systematic review and meta-analysis. Thromb. Res. 196, 27–30 (2020).
  9. Risk of venous thromboembolism among users of drospirenone-containing oral contraceptive pills.
  10. COVID-19 Vaccine AstraZeneca: benefits still outweigh the risks despite possible link to rare blood clots with low blood platelets.
  11. Medicines and Healthcare products Regulatory Agency. UK regulator confirms that people should continue to receive the COVID-19 vaccine AstraZeneca. GOV.UK (2021).
  12. expert reaction to MHRA statement on the Oxford-AstraZeneca vaccine and blood clots.
  13. Gando, S., Levi, M. & Toh, C.-H. Disseminated intravascular coagulation. Nat. Rev. Dis. Primers 2, 16037 (2016).
  14. Hunt, B. J. Bleeding and coagulopathies in critical care. N. Engl. J. Med. 370, 847–859 (2014).
  15. Hoffman, R. et al. Hematology: Basic principles and practice. (Elsevier — Health Sciences Division, 2017).
  16. Singh, B. et al. Trends in the incidence and outcomes of disseminated intravascular coagulation in critically ill patients (2004–2010): a population-based study. Chest 143, 1235–1242 (2013).
  17. Yu, J. et al. Direct activation of the alternative complement pathway by SARS-CoV-2 spike proteins is blocked by factor D inhibition. Blood 136, 2080–2089 (2020).
  18. Vaccines and Related Biological Products Advisory Committee Meeting February 26, 2021 FDA Briefing Document Janssen Ad26.COV2.S Vaccine for the Prevention of COVID-19. 20.Gavriilaki, E. & Brodsky, R. A. Complementopathies and precision medicine. J. Clin. Invest. 130, 2152–2163 (2020).
  19. Tonya Villafana VP Global Franchise Head, I. AstraZeneca COVID-19 Vaccine (AZD1222) ACIP COVID-19 Emergency Meeting January 27, 2021.
  20. Shimabukuro, T. COVID-19 vaccine safety update Advisory Committee on Immunization Practices (ACIP) March 1, 2021. 26.Anthes, E. The W.H.O. grants emergency authorization to the Johnson & Johnson vaccine, a boost for global vaccination efforts. The New York times (2021).
  21. Arepally, G. M. Heparin-induced thrombocytopenia. Blood 129, 2864–2872 (2017).
  22. Greinacher, A. Heparin-induced thrombocytopenia. N. Engl. J. Med. 373, 252–261 (2015)

Originally published at on March 23, 2021.

I write about vaccines here. You can find me on Twitter @enirenberg and at (where I publish the same content without a paywall)

I write about vaccines here. You can find me on Twitter @enirenberg and at (where I publish the same content without a paywall)