"Covid vaccines protect against cardiovascular disease" - no, no no. nope

Summary:

A study examining Swedish medical data finds that unvaccinated adults of various ages had certain rates of cardiac disease events in 2021 and 2022, while corresponding adults who just received a Covid vaccine dose (within six weeks) had rates which were lower than would be expected based on age etc. The study fails to disentangle this difference from cardiac events merely prevented (postponed) by short-term immunity to infection with SARS-CoV-2, and consequently, fails to demonstrate that Covid vaccines did not cause more cardiac disease in the vaccinated.

Contents:

1. Background

2. The claims

3. The context

4. This post’s subject study

5. The results

6. What we want to know, but don’t

7. Also the whole thing’s a fraud anyway (The third dose problem)

8. Conclusion

i. Background

(Skip to see current study results.)

Perhaps no theme more consistently appears in posts at Unglossed as the disconnect between ambiguous, unimpressive study findings and the heavily slanted claims about those findings made in the corresponding study abstracts and attached “science journalism.” Whatever reality is depicted at the business end of a “Study says” headline has no resemblance to what the study in question says; and by looking at the results and asking a few questions about context and methodology, a curious reader can unearth why this is so.

Typically such slanted claims, in the wild, are in alignment with mainstream public health dogma; though as a reader service I have devoted just as much (or more) ink to reporting on studies which incorrectly seem to affirm the more alarmist counter-mainstream narratives of the day.

Today is a classic example of an error from the mainstream; it’s¹ much rarer to see such a gem these days than it was three years ago, yet this one is a real prize canine.

As already framed in the previous ‘graph, some science-doers have conducted a review of Swedish medical records producing mixed or ambiguous results about Covid vaccine safety; they have adorned their work with a misleadingly vaccine-triumphalist abstract and discussion segment; and finally, a second online publication has “reported” these triumphant claims in complete disconnect from the actual, mixed and ambiguous findings of the study. (The reader should understand that my use of “mixed” reflects that the study in fact reaffirmed what is already well-known regarding the Covid vaccines and harms to the heart in the form of myo/pericarditis and excessively fast heart-beats.)

ii. The claims

“Reduced risk of serious cardiovascular disease after COVID vaccination.” This appears in reporting of the new study in question at “EurekAlert!,” and has made its way to twitter in recent days.

source: eurekalert.org

The study in question overtly posits similar good news within the abstract: “full vaccination substantially reduced the risk of several more severe COVID-19-associated cardiovascular outcomes, underscoring the protective benefits of complete vaccination.”

Xu, Y. et al. (the study this post is about)

Back to the summary at EurekAlert!, it is also disclosed in the first paragraph that the protection reported by this new study regards cardiac harms associated with infection by SARS-CoV-2 — the viral infection which Covid vaccines are well-known to temporarily blunt. It is also reported that the Covid vaccines increase other, different cardiac disease outcomes.

People who have been fully vaccinated against COVID-19 have a significantly lower risk of developing more severe cardiovascular conditions linked to COVID-19 infection, according to a nationwide study at the University of Gothenburg. At the same time, some cardiovascular effects are seen after individual doses of the vaccine.

iii. The context

Fair enough, except not. The reason this somewhat balanced-seeming presentation of the study data still counts as such an egregious misrepresentation is built into the context: SARS-CoV-2 infections can cause cardiovascular complications like stroke or heart failure; Covid vaccines temporarily prevent infections with SARS-CoV-2. This set of facts somewhat “rigs” the game in favor of making Covid vaccines look like magic heart protection spells and obscuring more nuanced questions; in other words, it is immediately necessary to know how any study looking at this question looked at it before talking about what they saw when they did.

We should expect that Covid vaccines temporarily prevent infection with SARS-CoV-2 and as such, temporarily prevent corresponding post-infection cardiac complications. Here I have modeled what is “normal” without the vaccines, and how the vaccines effect a difference.

I insert “for individuals here” to emphasize that on a population level, in 2021 and 2022, Covid infections are still happening, typically among the unvaccinated but to some extent among those whose Covid vaccines have “worn off,” if they haven’t received their rush-authorized boosters. For this reason in the US, it is still observable that reported infections correspond to excess deaths ascribed to cardiovascular disease in these years. This is just to affirm the first part of the model above: SARS-CoV-2 infection results in cardiac complications.

From previous post. In age groups over 35, increased deaths ascribed to cardiovascular disease correspond to spikes in reported infections w/ SARS-CoV-2. Since these are not recorded as “Covid-19” deaths, they are likely from cardiac complications which occur some time after infections.

Meanwhile “IRL,” this study is somewhat provocatively timed. It is published just as I have begun to observe some movement to consensus among zoomers that the Covid vaccines have (rare it still would seem, but serious) issues. I won’t belabor the reader with my own second-hand anecdotes, but it seems like there is a “vibe shift” at play.

iv. Our study

Notwithstanding the reality that no study is perfect, Xu, et al. is especially narrow. Ostensibly because it was designed to look for Covid vaccine harms, it only examines results within the first weeks after a given dose throughout 2021 and 2022. While other studies with such a narrow focus have used the prior health history of vaccinated individuals as matched “controls,” this study simply makes a big bucket out of everyone who hasn’t had a vaccine yet (which is then somewhat divided by age and sex, mostly in the supplemental materials). What’s valuable about the self-matched control method is it allows the observation of health records from before the virus; what’s flawed about this study’s approach is that it specifically includes the virus and limits it to the un-, not-yet, and not-recently vaccinated. The design:

“Buckets” of time (person-days) are created out of these individual statuses, and reported cardiovascular events of interest are counted and then rated by the corresponding time-bucket they took place within.

Because of high vaccine uptake in 2021, most of the overall “unvaccinated” bucket (especially in the younger age groups) will consist of the truncated “green” observation windows given to those who receive their first dose sometime mid-2021. Meanwhile, all the buckets for dose 1, 2, and 3 are deliberately limited to the first weeks after vaccination.

Returning to the problem intrinsic to this study’s topic, this design ensures that the “Covid vaccinated” (who only “exist” for a few weeks after vaccination) are protected from infection, while the “unvaccinated” are not; thus for the “unvaccinated” the virus is driving some additional risk of reported cardiovascular events; these events virtually have to fall in the unvaccinated time-bucket. (What, some readers may ask, about all the extra infections “caused by vaccination,” due to the “worry-window”? These infections don’t happen; the worry window is not real.)

Somewhat perversely, this design even ensures that most breakthrough infections that did occur in 2021 and 2022 are not even included on the ledger. Here, for example, is how this approach would typically record outcomes for an individual who was infected sometime after their second vaccine waned, and then went on to get a booster some time later: The study does not report whether they had a cardiac complication after infection.

v. The results

Now that we understand the study set-up, we can conclude a priori that it is conditionally capable of revealing cardiac harms from the Covid vaccines, if and only if those harms are not camouflaged by events driven by the virus itself; and further, that it isn’t really capable of demonstrating how protective the Covid vaccines are from the virus, since it only looks at outcomes in the windows when the vaccines truly “work” against infection.

We can toss out the triumphant reporting of the abstract and EurekAlert! profile: This study doesn’t “see” whether Covid vaccines protect against cardiovascular complications after infections. And even to report that the study strictly affirmed some types of cardiac vaccine harms is misleading, because the unvaccinated control is “rigged” by added harms from the virus.

The “time buckets” created by this study: SARS-CoV-2 driven cardiac events weigh down the unvaccinated bucket.

Rather, the study sees those harms which the virus doesn’t obscure, myo/pericarditis and rapid heartbeat.

Note that this depicts the results after the authors’ adjustments for demographics. In fact, raw rates for all cardiovascular events are higher in all vaccinated “buckets”; but this is to be expected due to the unvaccinated being more young.

vi. What we want to know, but don’t

This study focuses on Swedish adults. A previous study from Sweden examines outcomes among teens. Early last year, A Certain Vaccine Skeptic Wont to Make Cash Offers asked if anyone could produce an example where the Covid vaccinated showed as better off for cardiovascular outcomes than the unvaccinated. Simply because I had seen this Swedish teens study and remembered the outcome, I replied to send the link. The person who made this offer had me call him, and asked why outcomes were better among the Covid vaccinated. I answered that it was probably just because the vaccine was preventing infections, at least during the time window of this Swedish study.

I never received the cash prize for this find. Anyway.

Since we know that infection with SARS-CoV-2 can cause cardiovascular complications, measuring the questions of “do Covid vaccines increase complications, or prevent them” becomes contingent on when and how many times the unvaccinated and vaccinated are infected.

Xu, et al. makes no accounting for the confounding factor of the virus. As examined in section (iv), in fact, this study overtly excludes almost all breakthrough infections from the “vaccinated” by only observing results for a few weeks after each dose. Therefore it cannot claim to tell the reader that the vaccine really prevented “serious cardiovascular diseases” — rather than rescheduled them to a more “convenient” time. Nor can it say that the Covid vaccines did not cause some number of these diseases.

vii. Also the whole thing’s a lemon anyway (The third dose problem)

Having established that the unvaccinated “control” is “spiked” by exposure to the virus itself, it is revealing to tackle the question of why most of the “protection” against cardiovascular diseases observed by this study resides predominately in the window after the 3rd dose, rather than dose 1 or 2. For a glimpse, this plot from supplemental materials combines minor and serious strokes:

As established, Covid vaccine-receivers are only observed in this study for a few weeks after each dose. Therefore the major confounder of this study, SARS-CoV-2-infection, shouldn’t be driving any differences between specific doses. It is thus a sign of bad work that the authors would report such an “increase” in protection immediately after the 3rd dose, one not found after the first or more significantly the second. And if this bad work is in effect the driver of all the “protection” reported by this study in the first place, well… that doesn’t really look good for the Covid vaccines.

I will acknowledge first that this discrepancy lends itself to an alternate theory, which is that the Covid vaccines are driving cardiac events after the initial course, which are camouflaged (made to appear as a “1” hazard ratio) by virus-induced events in the unvaccinated, but not driving these events by the time dose 3 hits the market.

However, I am not convinced that the Covid vaccines caused as many immediate strokes and etc. as were driven by the virus, based once again on my overlay of CDC-reported deaths in (iii.). But neither are the authors compelling in their explanation for this discrepancy. They suggest that survivor or “healthy user” bias might be at work, with individuals experiencing cardiac events after the first two doses reluctant to get dose 3. Well, sure — but this is potentially irrelevant if events after the six-week window are excluded, since the study cannot “see” second-dose-havers “hesitating.” The alternate theory above, which posits that the first and second doses caused a heap of extra events, would be necessary to make a survivor bias theory plausible.

But instead of such a complicated model, it is more parsimonious to take this discrepancy as a glaring red flag over the study’s own math. As stated, the study adjusts the observed “event per time bucket” rates for demographic factors. The “reduction” of events observed for the 3rd dose is simply an artifact of this adjustment. It is in fact clear from the raw rates that more, not fewer of these events took place after the third dose.

The raw numbers and rates. “Lower” rates (right) after the third dose are actually “higher” in real numbers (center) , showing that the “lower” numbers are a magic artifact of the authors’ demographic adjustments.

So the authors have merely over-adjusted the hazard ratio for the older-ness of third-dose-getters; which means they probably also have done this for dose one and dose two (which are also higher in real rates than for the unvaccinated).

viii. In conclusion

If you received a Covid vaccine in 2021 or 2022, congratulations are in order. Your experimental gene therapy possibly came with weeks and weeks of free extra heartbeats, courtesy of Pfizer and/or one of the other fine makers of experimental transfections.

If you derived value from this post, please drop a few coins in your fact-barista’s tip jar.

click button or visit https://ko-fi.com/unglossed

1. Sadly, in a way.