It doesn’t matter what I believe, it only matters what I can prove!
Emphasizing nuanced and detailed data analysis, contextualized interpretation, and scientific integrity
MAKING SENSE OF NEW PREPRINT STUDY OF ADOLESCENTS IN THAILAND AFTER BNT126b2 (Pfizer’s mRNA COVID-19 VACCINATION)
Vaccine associated myocarditis (VAM) has been a topic of much vigorous discussion on social media since early reports from Israel first surfaced in April 2020. Many studies have since been published that demonstrate the risk of VAM after mRNA COVID vaccination is far greater than CDC’s estimate using VAERS data alone, and that the risk seems to be highest after dose 2 in males under 30 years old. Most recently a preprint study from Thailand evaluating myocarditis after BNT126b2 COVID-19 vaccination (Pfizer) has made the rounds on social media, garnering passionate opinions from supporters and critics alike. As an interventional cardiologist, I will provide a focused critical analysis of the preprint study. This is a focused post reviewing this study and not intended to be a comprehensive review of VAM. For a more detailed review see Myocarditis After Vaccination Against COVID-19 (June 2021). For a deep dive analysis of some recent studies on the rate of VAM after COVID vaccination see my recent post here. Finally, for some of the ethical considerations regarding COVID vaccination in adolescents in view of the rapidly waning immunity and grossly inadequate CDC follow-up of VAERS reports of myocarditis see my analysis here.
Depending upon one’s a priori scientific and ethical concerns of VAM (and subclinical myocarditis) and the consequent risks, one may summarily accept the results and interpretations of this study. Prima facie the results are stunningly exactly what many of us feared (i.e, that the rate of VAM is higher than what public health officials concede and the clinical myocarditis may just be the tip of the ice berg with far greater rate of subclinical myocarditis). Our task as scientists and physicians, however, is to evaluate the data objectively and see what conclusions naturally arise from such analysis. We may rightfully have concerns, even grave concerns, about COVID-19 vaccine mandates for healthy children and young adults given the risk of severe adverse reactions and the overall low risk of severe COVID-19 complications in this cohort. We may yearn, nay demand, that a properly designed prospective study better assess the true risk of subclinical myocarditis in adolescents and young adults. But lest we succumb to the very unscientific dogma of the advocates of COVID-19 vaccine mandates in this young population, we sill have to prove it.
So, let’s commence our dissection of the study itself.
OVERVIEW
One of the main strengths of the study is that it is “a prospective cohort study focused on adolescent students… who received a second dose of the BNT162b2 mRNA COVID-19 vaccine.” This is indeed the first of its kind, and ought to have been mandated in the US prior to granting EUA, let alone full BLA approval of the mRNA COVID-19 vaccinations. CDC has been mostly reliant on VAERS data (passive reporting) which has repeatedly been proven to be under estimating the risk of VAM by 3–4x (discussed here in much greater detail). As many on social media have commented, this is the study design that ought to have been conducted in the US.
The study included 13–18 year old students from two schools who had already received dose 1 without any severe adverse reactions. This itself is one of the limitations of the study when considering its generalizability (although the greatest risk is after dose 2, dose 1 also carries a small increased risk of VAM). Excluded from the study were any child who had prior history of cardiomyopathy, tuberculous pericarditis or constrictive pericarditis, or severe adverse reaction to dose 1. All study participants were assessed on Day 0, Day 3, and Day 7. Those with abnormal findings were also assessed on Day 7. This scheduled periodic assessment included complete history and physical exam, vital signs, echocardiography, EKG, and labs (Troponin-T, CK-MB). Some also had inflammatory markers (ESR, CRP). As noted above, this prospective design to specifically assess risk of myocarditis is unique and commendable. No other study is designed this methodically. Even the Phase III protocol for adolescents did not have such rigorous design, despite the fact that risk of VAM after mRNA vaccination was already established.
Interestingly, of the 314 adolescent students enrolled, 11 were lost to follow-up after baseline testing. The authors do not seem to clarify if these 11 received dose 2 or were lost to follow-up before receiving dose 2. An additional 2 were lost to follow-up after Day 3 testing. This is a serious limitation of the study. These are not random volunteers from the general community. These are students enrolled in school. Why were they lost to follow-up? What attempts were made to reach out to them. In a worst case scenario, did they suffer severe illness, injury, or death? The authors make no mention of the significance of these 13 lost to follow-up, let alone any potential explanation.
Study definition of cardiac manifestations is seen below (2.3) Here is where the study gets interesting (and where many readers not familiar with nuances of cardiology testing seem to be seriously misinterpreting the study’s findings and implications). Nuances and scientific integrity matter. Among the most fundamental questions in Journal Club in medical training is: Did the study assess what it aimed to asses? Are the conclusions valid interpretations of the results? Are the results generalizable? Hopefully, the following will elucidate why this study is fatally flawed.
STUDY DEFINITIONS
ABNORMAL OR PHYSIOLOGIC?
First, a very minor but important point of contention which by itself is not a fatal flaw, but does set the tone of repeated lack of precision by the researchers. Sinus tachycardia (fast heart rate but normal rhythm) and sinus bradycardia (slow heart rate but normal rhythm) are by themselves not ‘abnormal’ in an otherwise asymptomatic volunteer. We must remember that all “participants were healthy and without any abnormal symptoms before receiving the second dose of vaccine”. These are children and that must be the broader context throughout. It is not uncommon for children or even adults to be a little anxious or restless in a physician’s exam room. Many are even rushing to their appointment. Therefore, the authors should have explicitly clarified that the sinus tachycardia was noted after some period of requisite rest. For example, I frequently have patients with elevated BP in my cardiology practice because the medical assistant assessed the BP immediately upon bringing the patient back to the office (and the patient was rushing to make it to the office in time). After a small glass of cold water and waiting 5–15 minutes, most patients will have normalized heart rate and blood pressure. In addition to medical pathologies (e.g., infection, asthma exacerbation, vaccine side effects), sinus tachycardia can also occur from normal physiologic triggers like stress, exercise, anxiety, being outdoors in the heat, being dehydrated, etc. While an arrhythmia such as nonsustained ventricular tachycardia is never ‘normal’, sinus tachycardia may be an indicator of underlying pathology or a normal response to physiologic stress. If one is to include sinus tachycardia as indicative of “abnormal” finding, one must explicitly explain study protocol to ensure the phyiologic triggers have been eliminated. Sinus bradycardia can be a normal vagal response to emotional stress or fear (vasovagal response). Did these children have their EKG performed before or after blood work? Blood draws are a very common source of vasovagal response (i.e., slower heart rate in response to emotional stress). Did they have the EKG before or after the echocardiogram? One could argue that such reaction should have been present at baseline also. Such a dismissal is indicative of someone who has never been around patients who are anxious with repeated medical testing (even what seems otherwise harmless to others). A properly conducted echocardiogram can take 30–40 min with a probe pushing on the rib cage at different pressure points. It is not uncommon for even many adults, let alone adolescents, to find this to be an uncomfortable procedure. Perhaps the sinus tachycardia and sinus bradycardia are indicative of vaccine injury. Or perhaps they are indicative of normal physiologic response to stress of all the testing in the study. The point is we cannot be sure based upon the lack of details.
BASELINE CHARACTERISTICS
Baseline characteristics and aggregate study findings are summarized in Table 1 below. Of note, 14.6% of the adolescent students had underling medical conditions (we will return to this below).
SYMPTOMS
Adverse cardiovascular events amongst study participants were tachycardia (7.64%), shortness of breath (6.64%), palpitation (4.32%), chest pain (4.32%), and hypertension (3.99%). Make no mistake about it: despite my emphasis on precision and nuances of research definitions (to differentiate common physiologic variations versus pathologic abnormalities), it is by no means normal for children to have chest pain, shortness of breath, palpitations, or hypertension in response to a prophylactic medical intervention. These findings unfortunately are not new and sadly consistent with all the EUA trial data on presumed safety and transient side effects.
EKG FINDINGS
During EKG surveillance, the authors note “Fifty-four patients had abnormal electrocardiograms (predominantly sinus tachycardia or sinus arrhythmia) after vaccination.” This is where the study begins to fall apart (and further degenerates in the section on laboratory results). Sinus arrhythmia is by no stretch of the imagination an “abnormal EKG”. As explained in the American Heart Association’s website on Children and Arrhythmia:
“Some arrhythmias, or irregular heartbeats, are normal. For example, in many children, the heart rate speeds up while breathing in, then slows back down when exhaling. This heartbeat variation with breathing is called sinus arrhythmia, and it’s no cause for concern.”
I have discussed sinus tachycardia and sinus bradycardia above. Premature atrial contractions (PAC) and premature ventricular contractions (PVC) are also normal. They do not indicate an “abnormal EKG”. These ‘extra beats’ normally occur up to 1% of the total heart beats in a 24-hour period. An EKG is a 10-second snapshot in time out of approximately 86,400 seconds in a day. On any given EKG, these ectopic beats (PAC’s and PVC’s) may or may not be present for an individual. The same individual may have PAC’s and PVC’s on one EKG and then not on another EKG without any change in medical condition or intervention. Excessive PAC’s and PVC’s can be concering and warrant further clinical evaluation. However, one does not assess change in frequency of PAC’s and PVC’s with a static EKG. One orders a 24 hour Holter monitor to assess PAC or PVC burden (percent ectopic beats out of total beats in a 24 hour period). If the percent (of total beats) increases from one study to the next, then perhaps there is a clinical change worthy of further investigation. In the absence of such evaluation, a static EKG (representing 10-second snapshot in time) is completely meaningless. Using the wrong methodology (and not understanding cardiac electrophysiology), many on social media have erroneously accepted the conclusion that “18% had abnormal EKG”. Whether or not these results reached statistical significance is irrelevant: garbage in garbage out (the wrong test was used to assess electrical change over time and many of the normal electrophysiologic findings were misclassified as “abnormal EKG”).
Some have dismissed the concerns about the “abnormal EKG” rate as less important and not central to “the main finding of this study”. I respectfully but strongly disagree. On the contrary, understanding this failure of misclassfication lays the foundation for a more complex issue of the laboratory values (cardiac troponin). The same lack of precision and inappropriate study definition of “abnormal” carries forth in the subsequent section.
CARDIAC TROPONIN
This is a more complex issue to explore and one that quite frankly most physicians (who are not cardiologists) seem to misunderstand (i.e., the accept the definitions in the study at face value). I daresay, some cardiologists don’t even fully appreciate the nuances of cardiac troponin testing, especially the specific type used in this study. Furthermore, the issue is further complicated when applying this to the pediatric population to assess myocarditis and subclinical myocarditis.
In basic terms, “Troponin is a type of protein found in the muscles of your heart. Troponin isn’t normally found in the blood. When heart muscles become damaged, troponin is sent into the bloodstream. As heart damage increases, greater amounts of troponin are released in the blood.” (source: Medline). In essence, troponin is a marker of heart muscle cell necrosis (cellular injury and death). One of the main roles of troponin testing is in evaluation of potential acute coronary syndrome (ACS, or heart attacks). In the context of this study, troponin can also be used to assess for myocarditis. Recently, over a few years, it is now more commonplace to use what is known as high sensitivity troponin and this study also utilized these newer tests (“High-sensitivity cardiac troponin-T assay (HS-cTnT) and CK-MB isoenzyme levels were determined for all participants”). A more detailed discussion of HS-cTnT can be found here. However, the diagnostic utility of high sensitivity troponin in pediatric population for the diagnosis of clinical and subclinical myocarditis is replete with much uncertainty.
A Scientific Statement from the American Heart Association on Diagnosis and Management of Myocarditis in Children notes: “However, available biomarkers lack specificity, and no biomarker can differentiate myocarditis from other causes of acute myocardial dysfunction, injury, or ischemia… Troponin does not appear to be a sensitive or specific enough marker of biopsy-proven myocarditis. When elevated, it is often detected at very high levels.30 Troponin elevation has not been correlated with cardiac dysfunction or arrhythmias, but higher troponin levels have been associated with the use of extracorporeal membrane oxygenation (ECMO) and mortality.”
In fact, in the absence of endomyocardial biopsy, “Currently, the non-invasive gold-standard method for the diagnosis of myocarditis is cardiac MRI (class I recommendation, level of evidence C in the 2012 ESC guidelines161)”, not troponin alone (as noted in Myocarditis and inflammatory cardiomyopathy: current evidence and future directions). Why? Because troponin can have false positive results (ie., be elevated for other reasons) and thus suspicion for myocarditis must be confirmed by CMR (or at the very least new abnormal findings on echocardiogram). According to the Brighton Collaboration, in the absence of such imaging, symptoms plus troponin yields at best suspected myocarditis. Subclinical myocarditis is better assessed with asymptomatic CMR surveillance, not troponin alone. I shall explain why below.
The authors used 14 ng/L as the cutoff for ‘abnormal’ HS-cTnT assay. This is indeed the standard laboratory cutoff…. for adults… for the evaluation of acute coronary syndrome (heart attacks).
Huh?
Yes, you read that correctly. Normal value for HS-cTnT is <14 ng/L for adults in the evaluation of ACS. But this paper is about clinical and subclinical myocarditis, right? Correct. The authors never discuss the validity of using HS-cTnT in evaluating myocarditis nor the validity of defining normal as <14 ng/L. A very quick search on the topic demonstrates that once again the authors misclassified what is truly ‘normal’ or ‘abnormal’.
So, what is the appropriate range to use for HS-cTnT for evaluation of myocarditis in pediatric population?
“The two most important limitations of studies on references intervals of hs-cTnI and hs-cTnT, particularly in neonates, infants and children, are the volume of blood usually collected (about 0.5–1.0 mL) and the number of subjects needed to accurately measured the distribution values of biomarkers. Indeed, almost 300 apparently healthy individuals are needed for each age (i.e., neonates, infants or children) and sex group (boys or girls) to calculate the URL with a 99% confidence interval [18, 39]. Unfortunately, even larger studies do not satisfy the criteria of at least 300 cases for different age and sex groups” (from High-sensitivity cardiac troponins in pediatric population). Review of some of the studies (Table 1 below) demonstrates the 99th percentile varies significantly by age, has been estimated by studies with small sample size, and the confidence interval can extend the 99th percentile to upper teens rather than 14 ng/L.
In short, the 99th percentile (for upper limit of normal in pediatric population) is not well established. The authors of the Thailand study never justify the utilization of 14 ng/L in pediatric population (which may represent 99th percentile in adults, but has not been robustly validated in children for the purposes of this study). Furthermore, troponin, especially HS-cTnT is utilized to evaluate disease. Surveillance testing of all study participants (even asymptomatic testing) is fraught with diagnostic confounding. “The high false positive rate for troponin testing is largely due to indiscriminate ordering practices.” (A Brief Review of Troponin Testing for Clinicians). Troponin elevation must be in clinical context and surveillance testing in asymptomatic people is not well defined. False positive troponin has been previously described also. While these may or may not apply to this study from Thailand, the authors make no mention of the possiblity.
Spurious HS-cTnT elevation has been noted in hemolysis (see table above). This is important because the incidence of hemolysis in pediatric blood draws is much higher in pediatric blood draws than adult blood draws (21% versus 5% as reported here). With all four cases of apparent “subclincial myocarditis” having normal CRP and CK-MB levels, the authors of the Thailand study should have acknowledged this limitation and discussed the potential implications. Ideally, they would have a second blood draw for every elevated asymptomatic student with elevated HS-cTnT to ensure it was not spuriously elevated. In clinical practice we do this all the time, for example, with significant hyperkalemia in someone not taking any medications that could cause hyperkalemia. The study should have confirmed every case of elevated HS-cTnT with CMR to answer the question definitely.
So, what HS-cTnT value is to be used for upper limit of normal in pediatric population for myocarditis? It may surprise many that the value may be much higher than 14 ng/L . In the study High sensitive troponin T as gatekeeper for cardiac magnetic resonance imaging in patients with suspected acute myocarditis, the authors aim was to “to evaluate hsTNT as a gatekeeper for CMR with a lower cut off-value for exclusion and an upper cut-off value for confirmation of acute myocarditis as defined by CMR.” Note, CMR is the standard by which myocarditis is defined, not troponin alone, especially for asymptomatic individuals. HS-cTnT <4 yields 98.7% sensitivity (low false negative) and a value >343 yields 99.4% specificity. However, 14 ng/L (which is between 4 and 343) would yield a specificity somewhere between 33.7% and 99.4% (i.e., a fair amount would not have abnormal CMR findings for myocarditis and thus would be false positive).
In the study from Thailand, those deemed to have “subclinical myocarditis” do not seem to have had CMR to corroborate the suspicion of suspected subclinical myocarditis. Furthermore, the CRP (marker of inflammation) was normal in all cases of subclincial myocarditis. The CK-MB (another marker of cardiac injury) was also normal (CK-MB>6.22 is abnormal by the study definition). The HS-cTnT values are also in th range of poor specificity (as per the table above). So, with poor specificity of HS-cTnT values, normal CRP and CK-MB values, and no CMR to corroborate the suspected subclinical myocarditis, there is no scientific justification to conclude that 4/301 (1.3%) had subclinical myocarditis. None. ZIP. While it may be the case, the available data just do not support such a definitive conclusion.
TROPONIN AS PROGNOSTIC INDICATOR
Is elevated troponin a concern in these patients even if CMR is negative for myocarditis? Previous studies have indeed demonstrated that regardless of the cause (true positive for ACS, true positive for myocarditis, or false positive due to elevated troponin from other causes like sepsis, congestive heart failure, chronic kidney disease, etc.) elevated troponin carries a worse long term prognosis. The studies conducted on troponin as a prognostic indicator involved patients who were already sick for other reasons (hospitalized for other serious illnesses). Extrapolating that data to asymptomatic surveillance testing in the general population is not scientifically justified. Fore example, troponin can also be elevated after strenuous exercise (endurance training) without any evidence of ACS or myocarditis. Would it be scientifically appropriate to test all asymptomatic athletes after an athletic event and use troponin levels as a prognostic indicator of long term adverse outcomes without any actual research in this subset? Most certainly not.
SAMPLE SIZE
Finally, the issue of sample size cannot be ignored. The study found 1 case of clinical myocarditis out of 301 included in final analysis. Most other large scale studies (not based upon VAERS, but large scale healthcare and insurance database studies) found the rate to be about 300 per million (or 1 in 3,333). Even with boosters and heterologous dosing (dose 2 is from different manufacturer than dose 1), the highest rate thus far reported is 777 per million (or 1 in 1,287). [See my previous post reviewing some major studies on the rate of VAM here]. So, if the preponderance of data suggests a rate of approximately 1 in 3,333, why is any credible scientist or physician accepting 1 in 301 as valid and generalizable? Perhaps confirmation bias is a factor. Let’s say the study found zero in 301, would the same people accept zero percent risk of VAM as valid? No! In fact that is the very same criticism made about the pediatric ‘trials’ used to justify granting EUA of mRNA COVID-19 vaccination for children: SMALL SAMPLE SIZE. If this study had doubled the sample size, they may very well have found zero in the next 300.
Another example of how small sample size skews results is the follow-up CMR in this study. The follow-up CMR on the child with clinical myocarditis found no residual inflammation 5 months later. So, one could conclude 100% (1 of 1) had no residual inflammation on CMR 5 months after clinical VAM. However, a larger CMR study of 35 patients found 75% had persistent CMR abnormalities 3–8 months after initial diagnosis of VAM. In statistics, sample size does matter.
SUBCLINICAL MYOCARDITIS
Very briefly, subclinical myocarditis has been studied with prior vaccines (small pox and influenza). In this study elevated cTnT (prior to high sensitivity troponin) without symptoms without was deemed “possible subclinical myocarditis” and found to be occur 7 times more than clinical myocarditis. However, CMR was not performed to corroborate the subclinical myocarditis.
We must also be mindful that even with CMR proven subclinical myocarditis, there are no data to help us understand the clinical significance, long term consequences, or prognosis. Subclinical myocarditis may or may not be worrisome. It definitely warrants further research with better designed study protocols since the overall risk of severe COVID complications in this age group is very low.
SUMMARY
Many physicians, scientists, and general community members are scientifically, clinically, and ethically deeply worried about the risk of VAM (clinical and subclinical). We are morally justified in these concerns as the preponderance of evidence suggests for healthy children and young adults the risk may outweigh the benefits of vaccination against COVID-19. The risk to this cohort of COVID-19 infection (risk of hospitalization and death from COVID-19 infection) is ultra low. School and college mandates in healthy children and young adults are absolutely not warranted and may be imposing preventable serious and irreversible harm. To date no prospective study has been conducted to evaluate the risk of clinical, let alone subclinical, myocarditis after COVID-19 vaccination. Given the potential risk of fatal arrhythmia and sudden cardiac death, the potential long term effects of VAM are deeply troublesome. A prospective study is conspicuously missing from CDC recommendations and FDA post market surveillance requirements. However, no matter how much we may yearn, nay demand, such a study, this study from Thailand fails to meet that objective. There are fatal flaws in EKG classification and utilization of HScTnT without MRI corroboration to define subclinical myocarditis. After all, “it doesn’t matter what I believe, it only matters what I can prove!”
