Persistent Selection Biases in Pro-modRNA Literature
Risk of Myocarditis and other AESI > Risk of COVID Hospitalizations (Part 20)
As a noticer of trends and patterns, one of the patterns I tend to recognize with observational studies in particular claiming to find that cardiomyopathy is more prevalent after infection or Long COVID after fewer boosters is that they come to these conclusions based on data gathered from Electronic Medical Records of hospitals and/or start counting unvaccinated cases before any vaxx hit the market and/or confine the study parameters to pre-omicron variants when infection rates were much lower and far fewer unvaccinated had mucosal immunity from prior infections. Asymmetrical case counting windows are also prevalent along with exclusions of recipients of one dose and/or doses received outside the designated medical system in the study parameters but those are not nearly as confounding as the selection biases for older and more chronically diseased samples created by relying on hospital records and this one published in Nature Communications which relied on data from the medical records of the Hong Kong hospital authority to generalize about the VE of booster doses reducing the risk of long COVID. The study authors acknowledge that their research cannot establish causality due to several critical variables they never accounted for that would predict the course of infection outcome on their own:
Several important unmeasured confounders, namely obesity, smoking, socioeconomic status, educational level and strains of SARS-CoV-2 virus found in individual patients, could not be accounted for in this study owing to data availability, which may have introduced bias to our results.
This one published in Journal of the American College of Cardiology also relied on electronic health records (i.e. the records of hospitalized patients) to observe Major Adverse Cardiovascular events among 0.7% of unvaccinated patients and 0.5% of vaccinated hospital patients, but unlike the prior study actually accounted for comorbidities and found larger differences between patients with and without Major Adverse Cardiovascular events than vaccinated and unvaccinated patients including that the latter were more likely to have T2 diabetes, hyperlipidemia, ischemic heart disease, liver disease, obesity and 66 years of age or older. The study also starts in March 2020, thus counting all patients infected with the most virulent SARS-COV-2 variants as unvaccinated for a full 9 months before any shot was approved. They subtly but incompletely acknowledge this bias in the last statement: we could not account for different SARS-CoV-2 variants underlying infection.
In Total:
The study sample is limited to hospitalized with COVID19 which is not representative of the general population that has been infected, the vast majority of whom have mild or asymptomatic infection that require staying home for maybe a week and study parameters once again conspicuously cut off 2 years of omicron in which the vast majority of infections have occurred and the start for counting events is set 9 months before there is any vaccine cohort (December 2020) to compare to the infection cohort (March 2020). These are the same selection biases found in every pro-modRNA study.
To add to the selection biases you also have pro-modRNA observational studies that only consider VE for 4-6 months after the last booster despite the fact that a new booster is only available annually. I recounted an example of that last year:
Refutation of a Highly Bias Risk Benefit Analysis of Omicron Era Boosters
This staggered cohort study, published in BMJ, also relied on electronic health records from 3 countries and notes that mild and asymptomatic cases (the vast majority of cases btw) might not be recorded and only includes ‘a small number of young men and male teenagers, who were the main population concerned with increased risks of myocarditis/pericarditis following vaccination.’ If you’ve followed this series for any substantial length you’ll know that I’ve repeatedly noted the inverse relationship between age, a proxy for risk of severe disease from infection, and risk of cardiomyopathy from injection (i.e. males with the lowest risk of severe disease have the highest risk of shot induced cardiomyopathy).
The Canadian multiprovincial retrospective analysis I covered in (Part 16) used similar parameters to claim cardiac complications were more frequent after infection than injection but still had to concede the age specific caveat for young adult and adolescent males. The study recorded 686 cases of myocarditis/pericarditis following 18.8 million primary series (1 case per 27,493 recipients) complete and 160 following 860,000 infections (1 case per 5,377 infections) . The first obvious limitation imposed on the external validity of these findings by the study’s parameters is that its time frame, January to September 2021, excludes the vast majority of infections that occurred in Canada after the emergence of Omicron. The National Seroprevalence study for Canada found that only 9% of Canadians had been infected by SARS-COV-2 by November 2021, when omicron first emerged, while 76% had been infected by March 2023 and almost a year later that number is likely even closer to 100%. Aside from excluding all myo/pericarditis cases arising from any of the booster doses, based on the time frame, the study also excluded individuals, from the vaccination cohort, who had received either of their primary series doses outside their province and only received the first dose which accounts for about 20% of shot induced myopericarditis. The study also applied a one year asymmetrical case counting window to the infection cohort by counting myocarditis/pericarditis from infection (from the wildtype variant) going all the way back to January 26, 2020, a whole 11 months before any vaccine was even available. Obviously, if you count cardiomyopathy arising from infection, and infection from much more virulent variants,for twice as long as those from the shot, excluding vaccinated who got even one shot outside their province you’re going to find shot induced myocarditis and pericarditis at a lower rate. Despite designing the study to select for data that would reinforce the ‘safe and effective’ narrative the study authors still had to concede that the age stratified risk of myocarditis/pericarditis from modRNA products is inverse with 77% of post-infection myocarditis events occurring among those aged 40 years or older while the opposite was true for post shot myocarditis/pericarditis with 38% of cases occurring among recipients 18-29 years old, the least susceptible cohort to severe disease requiring hospitalization. And even though it finds a higher incident rate after infection than injection overall they admit some rather glaring exceptions such as a higher incidence rate of myocarditis and pericarditis following the Pfizer primary series for adolescents 12-17 years of age (rate ratio 1.16) and a higher rate of myocarditis and pericarditis following the Moderna primary series than SARS-COV-2 infection for persons 18-29 years of age (ratio 1.68). In the past I’ve pointed out that these results have been found consistently across the world even in studies designed to minimize the actual risk of cardiomyopathy from modRNA products such as (Part 1) and (Part 5).
The Canadian study was at least honest about the higher risk of cardiomyopathy for male adolescents and young adults following the shot rather than infection. A widely circulated CDC study that is used to claim that teenage boys have a 2-6x higher risk of myocarditis and pericarditis from infection than injection and young men have a 7-8x higher risk of myocarditis and pericarditis from infection than injection, actually has very low external validity to the general population for the same reasons the other observational research covered here does. Like all of the other pro-modRNA observational research, the study only includes infections captured in electronic health record data of 40 hospital systems. As the study authors admit later in the discussion section, their design omits the vast majority of infections that are not recorded in hospital systems. The electronic health record data also excludes the vast majority of vaccinations in persons five years of age or older.
only SARS-CoV-2 test results and mRNA COVID-19 vaccinations documented in EHRs were available for assessment. SARS-CoV-2 infections were not captured if testing occurred in homes, schools, community sites, or pharmacies. Similarly, EHR data in this study captured ≥1 dose of mRNA COVID-19 vaccine for 28% of persons aged ≥5 years. Nationally, 82% of persons aged ≥5 years were reported to have received any COVID-19 vaccination; 97% of all vaccinations administered were mRNA COVID-19 vaccines.
The study also excludes cases of myocarditis and pericarditis if the patient doesn’t have a previously documented infection in the electronic health records : this excludes vaccine injured patients who had a previous mild or asymptomatic infection(s) or virus naive patients in any age cohort of the study. The study also excludes recipients of booster doses from their criteria.
Thus the conclusions of the study aren’t generalizable beyond patients with documented infections in the electronic health records of hospitals. Even though they have a considerably large sampling size of 15 million patients the study’s authors acknowledge that the obvious sampling bias makes it far from conclusive and greatly reduces the generalizability of the results.
Underascertainment of SARS-CoV-2 infections and mRNA COVID-19 vaccinations reduced sample size and might have introduced bias if capture of infection or vaccination within the EHR occurred differentially for those with cardiac outcomes. Fourth, case definitions for myocarditis, pericarditis, or MIS were ICD-10-CM code–based; diagnoses were not confirmed with chart review and are subject to misclassification. Fifth, cases of MIS among persons without documented SARS-CoV-2 infection were not included (9). Finally, some overlap might have occurred in risk windows for persons who had a SARS-CoV-2 infection soon after vaccination or a vaccination soon after infection. Exclusions were made for persons who received COVID-19 vaccine doses ≤30 days before infection or who had infections ≤30 days before vaccination.
Most importantly, the study’s timeline, January 1, 2021–January 31, 2022, indicates that the vast majority of its data is from the pre-omicron variants (i.e. Alpha through delta) that were more virulent but much less transmissible than the ones actually in circulation like every other observational study touting the efficacy of modRNA. Thus, we cut out 2+ years of COVID to arrive at this predetermined conclusion.
It's not that studies summarized here are wrong and need to be debunked. Most of the time its just that their parameters make them more or less generalizable to fewer people. This one tells me nothing about the reduced relative risk of cardiovascular events of people with prior mild and asymptomatic infections as they are not even in the data set.
If you’ll recall in (Part 12) where I covered the literature on subclinical myocardial injury one of the studies referenced, the Swiss Hospital Study, found that boosted hospital workers developed myocardial injury post booster at an alarming rate of nearly 3% of recipients (22 out of 777). Despite finding such a high injury rate the authors claimed:
Before the COVID-19 vaccine was available, the incidence and extent of myocardial injury associated with COVID-19 infection was much higher than observed in this active surveillance study after booster vaccination.
However, as Klement and Walach note in a recent reanalysis published in The Egyptian Heart Journal, this can only be asserted of persons ever HOSPITALIZED with COVID19 not anyone who has had a SARS-COV-2 infection because all three studies cited in support of that statement consider only hospitalized COVID19 patients which excludes the vast majority of the general public especially younger age groups who are more likely to be infected than their elders but far less likely to even develop symptoms or progress to severe disease requiring hospitalization. The three studies were also published prior to omicron and thus limited their purview to pre-omicron variants with the latest one, also published in the same European Heart Journal, being revised as a final draft at the beginning of the Delta variant dominance. Their study parameters not only make their findings inadequate to generalize to the entire population of infected persons like the rest of the observational research reviewed here; the time they were conducted also makes them extremely anachronistic given how rapidly the virus has mutated from the ancestral variant and other pre-omicron variants.
This CDC study of Oregon death certificates for persons aged 16-30 was touted by USA Today as proof positive that modRNA is not associated with any cardiac arrest deaths in that age range but as even the authors admit the sample size is entirely too small to detect such a signal.
Unfortunately, none of these observational studies evaluated any other risk factor such as vitamin D serum levels which as I’ve noted in two separate posts (here and here) significantly reduce or elevate the risk of infection, severe disease, hospitalization, ICU admission, intubation and death. Vitamin D sufficiency has been found to reduce infection risk by 28-80%, cut the relative risk of severe disease in half, reduce ICU admission and intubation by 60-70%, and COVID mortality by 33-70%. It appears they only want to compare modRNA to the standard of care or no treatment and only stratify risk by general demographic characteristics.
They also did not compare the unvaccinated rates of cardiac compilations to the pre-pandemic baseline rates within the study populations but only the rates of cardiomyopathy following the much smaller number infections that occurred between between the beginning of 2020 and end of 2021 with the rates found after the much larger number of modRNA doses.
An Israeli observational study of the incidence of myocarditis and pericarditis in Clalit Health Services (n = 787,968), published in the Journal of Clinical Medicine, compared rates of myocarditis and pericarditis in patients infected with covid19 before infection (196,992) between March 7, 2020 and January 31, 2021 and a test negative control with no prior PCR tests and at least one negative PCR test between March 7, 2020 and 15 December 2020 (n = 590,976). The study found 9 cases of myocarditis (1 in 21,888 patients) and 11 cases of pericarditis (1 in 17,902 patients) in the infected cohort. In the control group the study found 27 cases of myocarditis (1 in 21,888 patients) and 52 cases of myocarditis (1 in 11,365 patients). This indicates that 1) rates of myocarditis and pericarditis following a covid infection is not higher than the baseline rate of myocarditis and pericarditis in the general population and 2) the baseline rate of pericarditis is actually higher than the rate of pericarditis following a covid19 infection although the scientists conclude: No statistical difference in the incidence rate of both myocarditis (p =1) and pericarditis (p =0.17) was observed between the COVID-19 cohort and the control cohort (at an alpha level of 0.05). A multivariable proportional regression found that being male slightly elevated the risk of myocarditis (p = 0.045) but a prior covid19 infection did not (p = 0.869). Being male (p = 0.025) and having peripheral vascular disease (p = 0.006) are statistically significant predictors of pericarditis.
Or evaluate the incidence and predictors of cardiomyopathy after the introduction of modRNA therapies and the role of a previous immunization against COVID-19 in modifying the risk of myocardial injury associated with SARS-CoV-2 infection.
A retrospective cohort study (CardioCOVID-Gemelli) conducted using data from all patients over 18 years of age admitted to Fondazione Policlinico A. Gemelli IRCCS Hospital in Rome for COVID19 disease between March 2021 and February 2022 found that out of 1,019 patients treated for COVID19 145 patients suffered myocardial injury. Out of the 352 vaccinated patients admitted 61 suffered myocardial injury with high sensitivity cardiac troponin levels greater than 56 nanograms/Liter, with the majority of recipients (35) receiving one or more Pfizer doses. Among the several positive predictors of elevated risk of myocardial injury including advanced age, hypertension, renal impairment, COPD, and Chronic Kidney disease vaccination status at zero doses was not one of them. In fact, between the 3 tertiles the study divided patients into by age, the COVID shots were only found to reduce the risk of myocardial injury from infection in the last tertile of seniors over 76 years of age and elevated risk of myocardial injury after infection in the lower tertiles ( 60-75 years and less than 60 years) although it was only statistically significant in the first tertile at the 0.05 alpha level. However, the authors conclude that vaccination status was not a statistically significant predictor of myocardial injury following hospitalization in the overall study population.