Another Misleading Look at Post-Vaccine Myocarditis

The CDC stays on brand but off the mark

Photo by CDC on Unsplash

I plan to branch off from writing about Covid in the near future; the world of medicine is so full of interesting topics to research and discuss, and life is all too short. Alas, my plans to dive into a recent article questioning the benefits of widespread statin prescribing were derailed by a recent CDC report, aptly published on April Fool’s Day.

In it, the CDC asserts that myocarditis rates in the wake of a SARS-CoV-2 infection are 2-8X as high as the rate of post-vaccination myocarditis, even in the highest risk sub-groups (young men ages 12-29). This argument is presumably well-meaning, intended to counter arguments overstating the risks of mRNA vaccines. We have heard it before, from another CDC MMWR — the “Covid increases risk of myocarditis by 16X” report from last year — and from the American Academy of Pediatrics:

The problem is: the CDC’s argument is factually incorrect. More importantly, the goal of science is to reach toward truth, even when unwanted or inconvenient, and here the majority of my colleagues seem to keep failing.

Allow me to explain my confidence that this latest report from the CDC gets it all wrong. I will try to keep this short.

The most important point ignored by this common narrative is that we already have a rather good idea of the frequency of myocarditis after Covid-19, and an excellent idea of the rate of myocarditis in young men after their second Pfizer shot. The latter has been extensively studied; while the range in high-quality reports has varied from about 1 in 2,500 to 1 in 7,000, the best studies (especially this carefully curated report from Hong Kong) tend to cluster around the lower figure. We can state with some confidence that post-Pfizer-second-dose myocarditis runs in the 1/3000 range. While still “rare”, it’s a serious condition which usually requires hospitalization and, while not typically as severe as a post-viral myocarditis, potentially causes long-term changes to the hearts of a significant minority of its sufferers. I think it’s a fair proxy in severity for an average hospitalization for Covid pneumonia in a young man.

Now, what about the risk of myocarditis from a case of Covid-19? Since everyone who chooses to be vaccinated is exposed to the risk of myocarditis, but only some proportion of those choosing not to be vaccinated will be exposed to Covid, the meaningful number should be the annual rate for the entire demographic in question.

This one is harder to calculate, but not that hard to approximate in good faith. Using the CDC’s COVID-NET system which monitors regions in 24 states comprising about 10% of the U.S. population, we can extract an annualized hospitalization rate for an “average” 12-17 year old. The past 12 month period includes the spring alpha, summer delta, fall delta, and winter delta/omicron waves - likely a “worse-than-representative” period. The cumulative hospitalization rate was 1 in 1,410 over the entire year. Of course, about half of this age cohort was vaccinated for the second half of this time span.

We can look at the year prior to observe the rates prior to adolescent vaccination; here, the figure is more like 1/2,170, but this number presumably understates the more recent risk to an unvaccinated child given the divergence both in vaccination rates and overall case numbers in this population between the pandemic’s first and second year. Even if we forget about this pre-vaccine rate, and attribute all the adolescent hospitalizations in the past year to those unvaccinated (which, given the remarkable effectiveness of the Pfizer vaccine in this age group, might be nearly true), we still end up with a number around 150 hospitalizations per 100,000 (about 1/650-700). How many of these hospitalizations involved myocarditis?

Per Dr William Oster’s presentation to the CDC on this subject, we see that a small proportion of children hospitalized for Covid in the massive Children’s Hospitals PHIS database were diagnosed with myocarditis — under 0.1%.

Fewer than 1 in a 1000! Even if we allow for boys developing Covid-induced myocarditis about twice as often as girls, and the 12-17 age group about 4X as prone as younger kids, we land around a 0.6% rate of hospitalized boys ages 12-17 being diagnosed with myocarditis. That translates to about 1 in 100,000 unvaccinated 12-17 year old boys likely to have both been hospitalized with Covid-19 and then developed myocarditis in the past year.

This is why, when I see figures like this, the part of me that demands there be some plausibility to a presented statistic just sees red (instead of CDC blue):

To be fair, as those careful readers who looked closely at the table from Dr Oster surely noted, we must consider Multisystem Inflammatory Syndrome (MIS), which causes a lot of myocarditis. These cases are probably lost in the COVID-NET database due to the delay between positive PCR test and hospitalization for MIS-C. As both the huge Children’s Hospital and even larger COSMOS databases show, MIS-C is the greater driver of myocarditis, although still with a little under 10% of cases yielding a myocarditis diagnosis. Since one study implied MIS-C hospitalizations can equal the number of Covid admissions in younger kids (although MIS-C is perhaps 3 times more common in the 5-11 age group than 12-17), it’s fair to add in another third to our 150 per 100,000 hospitalization rate, bringing it up to 200/100K, or about 1 in 500 unvaccinated 12-17 year old males in this past year. If 10% of our male MIS-C patients end up with myocarditis, the overall Covid-19 triggered rate of myocarditis would be around 6 cases per 100,000 per annum, or a little under 1 in 17,000.

Again: our highball estimate of 1 in 17,000 is a much lower rate than 1 in 3,000, or even 1 in 7,000. Portraying the risk of myocarditis from Covid-19 as being greater than the risk from vaccination in this particular demographic is simply disingenuous.

So, how does a study by intelligent, well-meaning researchers end up with so flawed a result? First off, looking only at myocarditis after Covid cases, and not for the entire unvaccinated population, is an apples-to-oranges comparison; as I said earlier, everyone who gets vaccinated assumes the risk of an adverse reaction, but not everyone eschewing vaccination will get Covid. Moreover, as others have discussed, there was a denominator problem; only reported cases were used in the denominator, and we all know that only a fraction of actual Covid-19 cases ever see the ink of a lab printer. Moreover, counting on claims data to report complications in timely fashion (i.e., myocarditis within 30 days of a PCR test) is liable to be a case study in under-reporting, as this thoughtful study detailed.

Why does this irk me so much? Well, it’s bad science, which is problematic in its own right. This on-going CDC endorsement of a false narrative — that “Covid is more likely to cause myocarditis than a Pfizer vaccine even in teenage boys” — gives fodder for those inclined to distrust the CDC. (Increasingly, I find myself among those ranks.) It also distracts from the more valid comparison: while 1 in 3,000 is a concerning rate for a serious vaccine side effect, it is still quite a bit lower than a 1 in 500 rate for overall hospitalization for Covid-19 or MIS-C, which can be nearly eliminated, at least for a time, via vaccination.

What’s more, focusing on this debate in an effort to squelch any rational discussion of individual cost:benefit might serve the public health community, but not every individual. An extremely healthy adolescent boy might look at the crashing rates of MIS-C with Omicron, and, especially if he has already developed some degree of immunity from a confirmed Covid infection, might decide that his overall hospitalization risk might be higher with a second Pfizer shot. This logic is especially apt when considering a booster shot in this demographic, for whom the myocarditis rates might not be quite as low as originally hoped.

I recall the Ottawa pre-print on vaccine-induced myocarditis which came out last year. The data collection was superior, but, alas, the authors made a mistake in calculating their denominator of vaccinate recipients, leading to a rather outrageous lead claim: about 1 in 1000 people of all ages and genders experienced post-vaccine myocarditis! Social media influencers like Steve Kirsch and Robert Malone ran with this headline before the study could be redacted.

That study frustrated me for similar reasons as this recent CDC effort: when preparing to publish results that absolutely do not make sense, it is imperative to stop and seek peer input before publishing! The flaws in this study were obvious, and led to conclusions that were pointless.

I hope the “restructured” CDC which emerges someday will hew to this advice. It is a minimal first step towards restoring trust in this once-revered agency.