A recent meta-analysis looks at all studies calculating the risk of developing either myocarditis (inflammation of the heart muscle) or pericarditis (inflammation of the lining around the heart) following each of the various COVID-19 vaccines. We already knew that there was an association, but the authors wanted to look at all the data currently available to nail down the statistics of the risk as much as possible.
The meta-analysis included 11 studies and a total of 58,620,611 subjects. Here are the main findings:
- The risk of developing either myocarditis or pericarditis was increased following COVID-19 vaccination, with a relative risk of 2.04.
- Risk was statistically significant for myocarditis alone, but not pericarditis, which may be because there is no increased risk for pericarditis, or because fewer studies looked at this outcome and therefore there was less statistical power.
- Risk was increased in men more than women, and in people <40 years old vs > 40.
- Risk was increased more following the second dose than the first.
- Risk was significant only in the mRNA vaccines, not the viral vector vaccines.
- Most cases of myocarditis were mild and self-limiting.
The authors did a good job of fleshing out the statistics, showing good homogeneity of results (the studies generally agreed with each other) and minimal publication bias (a symmetrical funnel plot). However, I was disappointed that they did not calculate the absolute risk or the number needed to harm. Only presenting the relative risk can often be misleading.
This comes up frequently on SBM – it is critical for consumers of health information to understand the difference between relative risk and absolute risk. If, for example, exposure to a drug cause the risk of side effect X to increase from 1% to 2%, that is a relative risk of 2, or a 100% increase, or twice the risk. That sounds significant. But the absolute risk increase is only 1%, which doesn’t sound as significant.
In this case I searched for some mention of absolute risk, but could not find it. That information may be buried in a table somewhere, but was nowhere in the write up itself. I would have preferred that it was right there in the abstract, or at least the discussion. It wasn’t even in the text of the results section. The authors might argue that this was not the purpose of the meta-analysis, but I would reject that explanation. Scientists must realize (and this is fairly explicit with academia now) that they are essentially communicating with the public when they publish studies, not just their colleagues. What they write influences the press release, and influences reporting on their findings.
This is exactly why “number needed to treat” or “number needed to harm” statistics are encouraged and becoming more common in medical reporting. They are more intuitive for the public, and even scientists and practitioners (i.e., humans), to understand.
Further, the authors have to know that the safety of COVID-19 vaccinations is a hot political topic. This makes it necessary, in my opinion, for them to make their write-up as bullet-proof as possible, to anticipate and cut-off motivated misinterpretation, and to put their data in as much context as possible.
Looking at discussions of the individual studies elsewhere, I was able to find that the risk of myocarditis following one of the mRNA vaccines is about 40 cases per million doses among the highest risk group – young males, or an absolute risk of 0.004%. The rate was 2.4 per million doses in older males and 1 per million in women. Even in the high-risk group:
Per million second doses of mRNA COVID-19 vaccine administered to males aged 12–29 years, 11,000 COVID-19 cases, 560 hospitalizations, 138 ICU admissions, and six deaths due to COVID-19 could be prevented, compared with 39–47 expected myocarditis cases after COVID-19 vaccination.
Given that >90% of cases of myocarditis will completely recover, that means in young men the vaccines prevent six deaths per million doses while causing <4 cases of myocarditis that have less than complete recovery. In all other groups the results are much more dramatic, saving hundreds of lives for every case of myocarditis (including the mild cases). Also, the data generally looks only at hospitalizations, ICU admissions, and deaths from COVID. They don’t look at long-COVID morbidity. These statistics are becoming alarming, with the CDC estimating that 7.5% of American adults having long COVID symptoms following infection.
To their credit the authors of the current meta-analysis point out that the risk vs. benefit of the vaccine is still favorable, and people should get their recommended vaccines. But they did not present the data to make this explicit.
We also have to keep in mind that COVID-19 infection itself causes myocarditis. The relative risk here is 16 times – with the absolute risk increasing from 9/100,000 to 150/100,000. That’s 1,500 cases of myocarditis per million COVID-19 infections, vs 40 per million in the high-risk group of young men from mRNA vaccines, and 1-2 per million doses in lower risk groups. The risk is literally 1-2 orders of magnitude (10-100x) greater from getting infected than from the vaccine.
These are the kinds of statistics that the public needs to hear, and they should be prominent in any study of a public health intervention, such as vaccines.
The bottom line from this data is that the small increased risk of myocarditis, especially in young males, following 1-2 doses of an mRNA vaccine for COVID appears to be real. However, the risk is small in absolute terms, most cases are self-limiting, and the benefit of the vaccines vastly outweigh this small risk (even just for myocarditis itself).