Misleading: The claim omitted the fact that one group of individuals used as a reference group to calculate the risk of ischemic stroke experienced a lower than usual number of strokes, which biased the calculation.
FULL CLAIM: “An increased risk of stroke events has been identified with the Pfizer COVID-19 bivalent vaccine”; “Despite the identified risk, the CDC continues to recommend booster shots
REVIEW
On 13 January 2023, the U.S. Centers for Disease Control and Prevention (CDC) and the U.S. Food and Drug Administration (FDA) released a joint statement explaining that one of their pharmacosurveillance systems called Vaccine Safety Database (VSD) had detected a “safety signal” with the Pfizer bivalent COVID-19 vaccines. Specifically, VSD data suggested an elevated occurrence of ischemic stroke among older people within the three weeks following receipt of the bivalent vaccine.
Numerous news outlets reported on the FDA/CDC statement, with many internet users spinning the information on social media to imply that COVID-19 vaccines were causing strokes (see also here and here). Health Feedback fact-checked these claims before and found them to be unsupported by the evidence.
Health Feedback explained that vaccine safety surveillance systems need to be sensitive enough to detect any possible risks, which can sometimes lead to false alarms due to statistical flukes. Because this safety signal occurred in only one pharmacosurveillance database and wasn’t observed in other countries using the same Pfizer bivalent vaccine, the CDC and the FDA concluded that “the data currently suggests that it is very unlikely that the signal in VSD represents a true clinical risk”.
On 26 January 2023, CDC representatives provided additional data during an FDA Vaccines and Related Biological Products Advisory Committee meeting. Reporting on the results, the Epoch Times claimed that the Pfizer bivalent vaccine was indeed dangerous and associated with an elevated risk of stroke: “another life-threatening side effect has been associated with the Pfizer-BioNTech vaccine”.
The Epoch Times acknowledged that “no other safety databases had detected this [safety] signal” and that “adverse stroke event[s]” would only occur after vaccination, “at a very rare rate”. Still, the outlet maintained that “An increased risk of stroke events has been identified with the Pfizer COVID-19 bivalent vaccine”. The article further implied that the CDC was putting the health of vaccine recipients in danger by saying “Despite the identified risk, the CDC continues to recommend booster shots for all people over six months of age”.
Health Feedback looked at the CDC data. Some results indeed show a higher occurrence of strokes in the early days following booster shot administration compared to later time points. However, the Epoch Times’ take on the analysis overlooked other results that depicted a much more nuanced situation. By doing so, the journal misled readers into a distorted view of the vaccine’s safety.
To understand why we need to take a closer look at the data. The CDC analyzed VSD data collected from 28 August 2022 to 7 January 2023 on more than 12 million people, around 800,000 of whom received a booster shot with the Pfizer bivalent vaccine.
The analysis compared the incidence of ischemic stroke between different groups of people. Here, it is important to introduce the notion of risk ratio. A risk ratio measures the relative risk of a given health event—an ischemic stroke in this case—between one group of people under study and a control group used as a reference.
A risk ratio of one means that there is no risk difference between the two groups. A risk ratio above one means that there is a higher risk in the group under study relative to the control group. Conversely, a risk ratio below one would mean that there is a smaller risk in the group under study. A corollary of this is that a risk ratio not only depends on the risk in the population under study but also on the risk in the group chosen as reference. This point will prove important for the rest of this review.
The CDC first compared the incidence of ischemic stroke during the first 21 days following a booster dose and within 22 to 42 days after boosting. It yielded a risk ratio of 1.47, meaning that more people were having a stroke in the first three weeks following vaccination than in the three weeks after that. In other words, there seemed to be a time window close to the booster shot date where people had an elevated risk of stroke. This is the result on which the Epoch Times based its claims.
However, the CDC also calculated a risk ratio between people who were within the three weeks following their booster shot and people who didn’t get the booster shot. In this case, the risk ratio was very close to one (1.07). An analysis found that there was no statistically significant difference between boosted and unboosted people. Results are deemed not statistically significant when the data cannot prove that there’s a difference of risk or outcomes between the two groups being compared. This result indicates that people who received a booster shot less than 21 days earlier weren’t at a higher risk of ischemic stroke than those who didn’t, contradicting the Epoch Times’ claim.
In the transcript of the 26 January meeting, the CDC representative offered a possible explanation for the apparent contradiction between the two risk ratios: “the question now becomes this, is this increased r[isk] ratio? Is it because there’s a higher rate within the risk interval, or is there [an] unexpectedly lower rate during the comparison interval?”
What this means is that a risk ratio can be higher than one because the number of strokes is unusually high in the group under study or because the number of strokes is unusually low in the control group used for comparison.
As the CDC data show, the risk ratio is lower than one—0.76—among those who received the booster shot within 3 to 6 weeks following vaccination compared to unboosted individuals. This means that the risk of stroke in people within three to six weeks following the booster shot is unusually lower compared to a reference population.
Recall that this group of people—who’d received a booster shot 22 to 42 days earlier—is the control group chosen as the reference for the first risk ratio calculation that yielded a value of 1.47. Because this group has an unusually low risk of stroke, it partly explains why we obtain a high risk ratio when it’s used as a control group.
In other words, the risk ratio of 1.47 obtained in this analysis doesn’t necessarily mean that people who got boosted less than three weeks earlier are actually at a higher risk of stroke than the general population. Instead, it may simply be due to the fact that the reference group is at a lower risk of stroke than the general population.
The CDC didn’t offer an explanation for the lower-than-expected number of strokes in the 22 to 42 days post-vaccination group. However, the CDC presentation and the meeting transcript highlight that data are still uncertain, with a low number of strokes overall that limits the power of the statistical analysis.
In summary, the data on ischemic strokes following the Pfizer bivalent booster shot so far don’t provide reliable evidence for the claim that the booster shot causes a higher risk of stroke. People who received a COVID-19 vaccine booster aren’t at a higher ischemic stroke in the next three weeks following vaccination compared to people who didn’t get a booster dose. While it is true that the CDC found a higher risk of ischemic stroke during the first three weeks following vaccination compared to 22 to 42 days after receiving the booster dose, an unusually low number of strokes in this latter category affected the comparison between the two and the calculation of the risk ratio. The CDC and the FDA highlighted that more data are required and that they will continue to monitor the situation closely.
