Intended for healthcare professionals

CCBY Open access

Rapid response to:


Risk of thrombocytopenia and thromboembolism after covid-19 vaccination and SARS-CoV-2 positive testing: self-controlled case series study

BMJ 2021; 374 doi: (Published 27 August 2021) Cite this as: BMJ 2021;374:n1931

Linked Editorial

Strengthening international surveillance of vaccine safety

Rapid Response:

The problem with combined exposures and the authors’ choice

Dear Editor,

The article combines an impressive number of data points but it does not say very much about the quality of a data point. There is reason to question the data quality of this data set.

The article does not explain how the pre-risk period of 1 month (28 days) comes about. I.e. the authors observed an increased risk of adverse outcomes up to 28 days before the exposure. What should be the reason for that except that there is a considerable problem with the exposure dates and/or the outcome dates?

As mentioned in the text and shown in supplementary figure 2a, adverse outcomes for vaccinations preceded the exposure up to 1 month (28 days). This indicates that the exposure dates were accurate only up to plus/minus 1 month. Also, the authors give no explanation how this pronounced 2-peaked form in supplementary figure 2a comes about.

An exposure date accurate only up to plus/minus 1 month is significant because the total observation period is only 5 months. If the data quality of the exposure dates is poor it is difficult to discriminate between positive test and vaccination when it comes to the combined exposure of a positive SARS-CoV2 test and a vaccination. This is what the article is reporting when it comes to test related events, a combined exposure to test and vaccination. A clear differentiation between those events could be possible, if they differ by more than 2 month.

That this is a serious problem can be seen in table 3 of the article where the incidence rate ratio (IRR) of an adverse outcome in the pre-risk phase after (i.e. before) a positive test is up to 5 times higher than baseline. The same elevated pre-risk does not appear after (i.e. before) a vaccinations, for none of the outcomes. That happened because the authors attributed the adverse effect to the positive test exposure only though it was a combined exposure of test and vaccination that relates to an adverse effect. In this sense the authors decided what caused the outcome.

This data set does say something about the combined exposure of vaccination and positive test and the outcome thrombosis like adverse reaction and this gives cause for concern. But it says very little if not nothing about the positive test event alone and the outcome thrombosis like event.

In view of the recent deaths related to vaccinations it would have been desirable if the authors had interpreted their data more conservatively. Caution is advised to conclude from this data set that the risk of vaccination related adverse outcomes is smaller than the risk of positive test related adverse outcomes.

Best regards,
Johannes Kreis

Competing interests: No competing interests

01 September 2021
Johannes Kreis