select search filters
roundups & rapid reactions
before the headlines
Fiona fox's blog

expert reaction to preprint looking at incidence of rare cerebral venous thrombosis (CVT) following COVID-19 infection compared to incidence after vaccination or influenza

A preprint, an unpublished non-peer reviewed study, looks at incidence of Cerebral Venous Thrombosis (CVT) after COVID-19, compared to incidence after receiving an mRNA COVID-19 vaccine (Pfizer or Moderna), or following influenza infection.

This Roundup accompanied an SMC Briefing.


Dr Peter English, Retired Consultant in Communicable Disease Control, Former Editor of Vaccines in Practice Magazine, Immediate past Chair of the BMA Public Health Medicine Committee, said:

“Following the concerns that some Covid-19 vaccines may be associated with a slightly increased incidence of a rare clotting disorder, it is important to better understand both the background rate of the disorder (in order to understand whether there is a causal link with vaccination), and the rate of the disorder in people who have Covid-19 disease.  It is important to be able to put into context the risks, to help people decide if they wish to be vaccinated, and whether to accept vaccination with a particular vaccine.

“One of the reasons postulated for the increased risk of this disorder postvaccination is that people who develop the disorder were actually already infected with Covid-19, and did so as a result of the infection, rather than the vaccine.  The better we understand the links between vaccination, disease, and the disorder, the clearer we will be about the likelihood that hypotheses such as this are correct.

“It has long been established that other viral illnesses, like influenza, can cause clotting disorders: the rate of heart attacks and strokes is raised considerably by influenza.  So it is not surprising that Covid-19 also has this effect.  This paper attempts to quantify the risk from Covid-19, from some Covid-19 vaccines, and from influenza.

“The study used electronic record linkage – using electronic databases to link people with a diagnosis of Covid-19, influenza, or receipt of Pfizer-BionTech (BNT162b2) or Moderna (mRNA-1273) COVID-19 vaccines with subsequent diagnoses (within the following two weeks) of portal vein thrombosis (PVT) or cerebral venous thrombosis (CVT – also know as cerebral venous sinus thrombosis (CVST)), as these are conditions that may be linked to use of the AstraZeneca vaccine.  (This study was undertaken in the USA, where the AstraZeneca vaccine is not used, which is why this vaccine was not included in the study.)

“The study used quite large numbers of people: 513,284 who had had Covid-19 disease, 172,742 who had had influenza, and 489,871 who had received one of the two vaccines.  Because such large numbers were used, a reasonable number of CVT cases were found.  The estimate of 39 cases per million people who had had Covid-19 presumably equates to roughly half that number of cases in the study group of roughly half a million people, so the data are fairly robust, as shown by the relatively small confidence intervals (25-60 cases of CVT per million cases of Covid-19).

“The study was also able to use the same methods to estimate the incidence of CVT (and also of PVT and some related conditions) in people who had not been recorded as having been vaccinated or having a diagnosis of Covid-19 to provide the background rate of these conditions.

“The diagnosis of CVT is not easy to make, so it is possible that cases were missed.  But using record linkage methodology will have reduced reporting bias, as this was proactively seeking information, rather than relying on passive reporting of cases.  There may still be a possibility that, given the publicity about this possible link, the diagnosis of CVT may have been sought more assiduously in people who had been vaccinated, and therefore be more likely to have been recorded in the records compared to other people with the condition, leading to some recording bias.  If there was no such bias, then the use of identical methodology to identify cases in the various different groups will mean that any e.g. underestimation of case numbers will likely apply equally to all the groups or “cohorts”.  As the authors point out, however, the cohorts (the study and comparison groups) were not matched for age, sex, or other demographic factors; and much larger samples would be required to be confident before we could conclude that the mRNA vaccines studied are associated with an increased risk of CVT.  The rarity of CVT means that very large samples will be required to be confident that any difference in incidence between any two groups is a true association, rather than a chance effect.

“The study was able to conclude, however, that the risk of CVT is increased, compared to the general population, in both people who have Covid-19 disease, and in people receiving the vaccines in the study.  It found the following rates of disease:

  • Background rate of CVT: 0.41 per million people over any two-week period.
  • Rate of CVT following diagnosis of Covid-19: 39.0 per million people (95% CI, 25.2–60.2).
  • Rate of CVT following diagnosis of influenza (0.0 per million people, 95% CI 0.0–22.2, adjusted relative risk compared to the baseline population (RR)=6.73, P=.003)  [The finding that no cases were identified among the 172,742 people in this cohort is not too surprising, given the low background rate of CVT.  I am not entirely clear why this gives a RR of 6.73.]
  • Rate of CVT after receiving BNT162b2 or mRNA-1273 vaccine (4.1 per million people, 95% CI 1.1–14.9, adjusted RR=6.36, P<.001).  [I am not clear why this gives a RR of 6.73, rather than 10, given that 4.1 is exactly 10 times 0.41.]
  • Rate of PVT following Covid-19 disease: 436.4 per million people (382.9-497.4).
  • Rate of PVT following influenza: 98.4 (61.4-157.6) after influenza.
  • Rate of PVT following vaccination: 44.9 (29.7-68.0).

“The authors summarise that: “…the risk after COVID-19 is approximately 8-10 times that reported for the vaccines, and about 100-fold increased compared to the population rate. The increased rate of CVT in COVID-19 is notable, being much more marked than the increased risks for other forms of stroke and cerebral haemorrhage.”

“Note that the references to risks associated with the AstraZeneca vaccine are based on the European Medicines Agency findings, rather than the findings from this study, which did – because it is not used in the USA – did not include that vaccine.

“The study did not distinguish between PVT and CVT overall, and these conditions associated with the particular pattern of “vaccine-induced thrombotic thrombocytopaenia” that it is thought may have caused the increase of cases associated with the AstraZeneca vaccine.  Indeed, the paper uses record linkage to attempt to quantify the rates of PVT and CVT in different cohorts; but it is not designed to look in detail at the causes – the pathophysiology or aetiology – of these conditions.  It did consider some other findings, but it cannot tell us why these conditions arise as complications of Covid-19.

“The key thing to take from this study is that Covid-19 is a potent cause of clotting disorders, including (but not limited to) the very rare diagnosis of CVT; and any low risk associated with the vaccine has to be put in the context of the relatively high risk associated with Covid-19 disease.

“I would expect similar studies to be underway in the UK, in Europe, and elsewhere.  As more data are published, we will be able to be increasingly confident about the magnitude of risks associated with the disease, with any vaccine, and with specific vaccines.”


Prof David Werring, Professor of Clinical Neurology, UCL Institute of Neurology, UCL, said:

“COVID-19 is known to be associated with blood venous thrombosis – due to sticky (hypercoagulable) blood – in about 1 in 5 patients.  The risk of venous clots (thromboses) is highest in those with severe COVID-19 disease treated on an intensive care unit; these can occur in the lungs (Pulmonary Embolism, PE), legs (deep vein thrombosis, DVT) or, less often, the brain (cerebral venous sinus thrombosis, CVST).  CVST has also been reported in a very small proportion of people following the AstraZeneca vaccine, thought this seems to be related to a different mechanism associated with unusual antibodies to blood platelets.  This risk of CVST after the AstraZeneca vaccine, though very low indeed (probably about 5 per million people), has led to a lot of discussions about the risks and benefits of vaccination in countries using this vaccine.

“The present study, using information from a US database, suggests that the risk of CVST in the two weeks after vaccination with the Moderna and Pfizer vaccines is also very low (about 4 per million people), while the incidence of CVST in the two weeks after COVID-19 diagnosis is about 8-10 times as high as this (39 per million people).  Both of these risks were higher than in the general population over a two week period (0.4 per million).  However, there were some methodological limitations; in particular, the authors could not verify the accuracy of the diagnosis of CVST, which we know can be challenging, requiring appropriate clinical expertise and timely scanning of the veins in the brain.  Furthermore, they were not able to look at the risks of CVST associated with the AstraZeneca vaccine in the same population.”


Prof Adrian Newland, Professor of Haematology, Queen Mary University of London, said:

“While there has been much interest and speculation regarding the incidence and relationship of Cerebral Sinus Vein thrombosis (CSVT) and portal vein thrombosis following the Covid-19 vaccinations there has been less emphasis on the numbers seen following Covid-19 infection.  The association between covid-19 infection and thrombotic problems has been long recognised and this review from Oxford also confirms the high rate of CSVT in over 500,000 patients reported in the States, occurring in 39 per million and particularly affecting those under 30.  They reviewed at the same time nearly 500,000 receiving the mRNA vaccines (Pfizer and Moderna) and found a significantly lower incidence of 4 per million.  As the AZ-Oxford vaccine was not used in the States they reviewed the reports to the EMEA in Europe where the rate was 5 per million.  So the overall rates following infection were 8-10 times greater than following vaccination.

“While such studies are always difficult to do and to report as the various groups cannot be matched and the completeness of the records and the diagnostic approach is uncertain the numbers are such that the trends give considerable strength to the conclusions and suggest that the emphasis given to the risks following vaccination, which run at a rate not dissimilar to that seen in the normal population for CSVT, may have been overemphasised.  We do not know whether the causes of clotting in the various groups is similar, and we do know that Covid-19 infection is in general associated with increased numbers of thrombotic episodes, but the figures do support the importance of pushing on with the vaccination initiative.”


Prof Beverley Hunt OBE, Medical Director of Thrombosis UK, and Professor of Thrombosis & Haemostasis, King’s College London, said:

“This paper looks at rates of CVST after patients in the United States had COVID-19 infection or the Moderna and Pfizer vaccines.  They show increased rates of CVST after COVID-19 infection, these were greater than after vaccination.  However the paper is based on ICD coding which is often entered late, so the data on rates after Moderna and Pfizer vaccines may be incomplete: thus the results need to be interpreted with caution.

“There is no new data on what is happening after AstraZeneca vaccine because they are not using this vaccine in USA.”

What do we know in general about the potential mechanism for CVST?

“Patients who are hospitalised with COVID-19 have very prothrombotic (sticky) changes in their blood, which persist after they have been discharged.  This will lead to increased rate of blood clots.  The mechanism for the very rare blood clots and low platelet counts seen  after AstraZeneca vaccine is different.  It is associated with an immune response – an antibody against platelet factor 4 is found, this provokes blood clotting and  consumes platelets.

Is CVT (used in this preprint) and CVST the same thing?



Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“A huge amount has been written and said recently about blood clot risks from vaccines that protect against Covid-19.  In particular, it’s been pointed out again and again that the adverse effects that are of particular concern, certain types of blood clots occurring alongside an atypical pattern of low platelet counts in the blood (thrombocytopenia), are very, very rare after vaccination.  But exactly how rare they are after vaccination, in comparison with how rare they are in unvaccinated people, has been difficult to establish.  The issue is that the condition is so rare that really good data on the prevalence in unvaccinated people, before the pandemic, does not really exist.  Various other comparisons with blood clotting rates have been made, in people who have not been vaccinated but are doing something else that may affect clotting risk (like taking long flights), and also in people who have Covid-19, but usually these are not about the specific and very rare kind of blood clots that are causing concerns about the vaccines to regulators.

“This new piece of research throws a bit more light, though the researchers make it clear that they do not have all the answers, and that more data and more analysis and interpretation are necessary.  Also, it’s important to remember that the research report has not yet been peer reviewed.  The new research does not look at the risk of all blood clotting disorders, but only at the particular type that has been of most concern in relation to the vaccines – cerebral venous thrombosis (or CVT, also known as cerebral venous sinus thrombosis or CVST).  The researchers filled in one bit of the picture that has largely been missing, by looking at the risk of CVT in people who have been diagnosed with Covid-19, and comparing it with the risk in people vaccinated with three different vaccines (the two mRNA vaccines (Pfizer/BioNTech and Moderna), and Oxford/AstraZeneca).  An issue is that the main comparison still does not quite refer only to the cases of particular concern, where one of these blood clotting disorders occurred with low platelets, but it’s much close than most of the comparisons we have seen.  The researchers did also look separately at another type of clotting in a vein in the abdomen.  Overall the main finding is that these CVT events are very rare – a few in every million people involved – in Covid-19 patients and in people who had one of the vaccines – but they were very much rarer in the people who had a vaccine than in people who had Covid-19.  For people who had had one of the mRNA vaccines (Pfizer/BioNTech or Moderna), about 4 people in every million had a CVT in the two weeks after vaccination, but the rate in people who had Covid-19 is almost 10 times that, at 39 in every million.  That’s still not common enough to be a major concern in people with Covid-19, compared to all the other risks and problems, but it indicates yet again how low the risk is from vaccines.  The researchers are not claiming that vaccines do not increase the risk at all compared to the risk in people who have not been vaccinated and have also not had Covid-19 – but they say the CVT risk in people who have had Covid-19 is about 100 times the risk in the general population.  I do think this puts things into context.

“The preprint also makes comparisons with the risk of CVT after the Oxford/AstraZeneca vaccine.  The researchers found that risk to be only very, very slightly higher than for the two mRNA vaccines, at 5 cases in every million, but that’s still very much less than the 39 per million in people with Covid-19.  They caution, rightly, that the comparisons with the Oxford/AZ vaccine are rather less certain, because the Oxford/AZ data come from a different data source.  (All the other data come from TriNetX Analytics, which runs very predominantly in the US, and the Oxford/AZ vaccine does not figure in that data source.)”


Prof Paolo Madeddu, Professor of Experimental Cardiovascular Medicine at Bristol Medical School, University of Bristol, said:

“The major issue here is that the comparison showing the higher risk after Covid-19 does not exclude the possibility the pathogenesis is the same and therefore some common denominator should be searched.

“For instance, if the mechanism is the same, one can speculate that the high occurrence in covid-19 vs. vaccination is because the whole virus is more thrombogenic than the spike protein alone.  These studies are important but seem to be focused on demonstrating the minor risk of vaccination instead of making efforts to explain the cause of complications, taking advantage of the similarities of the events in the two populations.  I have previously expressed my opinion that the thrombogenic interactor CD147 of the spike protein could be involved.”



Preprint (not a paper): ‘Cerebral venous thrombosis: a retrospective cohort study of 513,284 confirmed COVID-19 cases and a comparison with 489,871 people receiving a COVID-19 mRNA vaccine’ by Maxime Taquet et al.  This work is not peer-reviewed.



All our previous output on this subject can be seen at this weblink:



Declared interests

Dr Peter English: “No COI to declare.”

Prof Adrian Newland: “I have no conflicts of interest.”

Prof Kevin McConway: “I am a Trustee of the SMC and a member of its Advisory Committee.  I am also a member of the Public Data Advisory Group, which provides expert advice to the Cabinet Office on aspects of public understanding of data during the pandemic.  My quote above is in my capacity as an independent professional statistician.”

None others received.

in this section

filter RoundUps by year

search by tag