Hi, all! Been a little while. Sorry about that. I had some really critical projects to tend to that couldn’t wait and now I have a bunch of backlogged posts that are unfinished but I think will be of value which I am progressively working on getting out. I think this is a good note to start on though for the end of my hiatus. Today (October 26, 2021) was the 170th VRBPAC meeting concerning the use of the Pfizer vaccine in children aged 5 to 11 years, and I did something insane: I live tweeted it- the entire 8:09:15. My thread can be found below. STATNews’s Helen Branswell and Matthew Herper also live-blogged it here.

I won’t begrudge you if you do not wish to read through 278 of my tweets though, so below are what I think are the most salient points for people to understand. For the following, the source of all claims are the VRBPAC presentations unless stated otherwise.

The vaccine for 5 to 11 year-olds is slightly different.

The vaccine differs from the vaccines for other age groups in a few ways.

• The first is the dose: the mRNA content for this age group is 10 mcg instead of 30 mcg (1 mcg = 1 millionth of 1 gram). The reason for this is because Pfizer conducted a dose-finding study and found that the 10 mcg dose was less reactogenic (caused fewer side effects) but equally immunogenic (elicited an immune response as good as higher doses) compared with higher doses in this age group. Immunogenicity was broadly similar within age groups of the 5- to 11-year-olds. Theoretically, this may also offer some safety advantages as there are suggestions that the risk of serious adverse events, with special interest in myocarditis risk (more on that shortly), may be related to the dose of the mRNA in the product, but this is still speculative. In other words, if you have an 11-year-old, there is no need to wait for them to turn 12 to get this vaccine (provided it all gets approved for them). The vaccine showed 90.7% efficacy in this study (16 cases in the placebo group, 3 in the vaccine group, all due to the delta variant). This is a small number of cases, but the focus of the study was immunobridging- demonstrating that the immune response to the vaccine is as good as for other groups in whom the vaccine has already demonstrated efficacy. Thus, this is acceptable.

• The buffer formulation is slightly different. A buffer is an inactive ingredient whose job is to resist pH changes (changes in acidity or alkalinity). Biological systems are filled with buffers because the large biological molecules that comprise them are very sensitive to pH changes and can only function in a narrow range of pHs. In the case of RNA, alkalinity of any kind is very poorly tolerated and can cause the RNA to degrade in a reaction called hydrolysis. The buffer solution ensures that the vaccine vial never gets too alkaline. This is also helped by the very low temperature storage conditions. The Pfizer/BioNTech COVID-19 vaccine (aka Comirnaty) has used a buffer called phosphate-buffered saline before (a mixture of common salts occurring in our own bodies- potassium chloride, monobasic potassium phosphate, dibasic potassium phosphate, and sodium chloride aka table salt). However, this is a bit nonstandard for storage of nucleic acids like the mRNA vaccines are made of; tris, also called tromethamine, is much more typical and preferred (for long term storage of RNA, tris-EDTA buffer is typically used, sometimes called TE buffer. The EDTA is a chelating agent that sucks up metal ions like magnesium, which can rapidly degrade the RNA. No EDTA is contained within the updated Pfizer vaccine however). Incidentally, this component of the buffer is used in the Moderna vaccine. The vaccine will no longer contain table salt (sodium chloride) or potassium chloride. The vaccine was tested in the 5 to 11 year old age group using the old buffer, but as this is an inactive ingredient, this isn’t concerning for any issues of safety or efficacy, and Pfizer conducted standard studies on shelf life to confirm the integrity of the RNA vaccine over time and found that now it can be stored at standard refrigeration for 10 weeks! That’s really excellent because it means it can be given in pediatricians offices, and it opens up possibilities for its use in lower and middle income countries as this shift happens with the other Pfizer/BioNTech mRNA vaccines.

• Because the dose of the vaccine is smaller in this age group, there are some changes to the vials. Firstly, it’s hard to accurately draw up very small volumes, and the vaccine which had triple the dose was just 0.3 mLs, so this one would have to be 0.1 mLs if you used the same concentration. Thus, the vials are more diluted to allow for accurate dosing (they are to have 0.9% sodium chloride solution added to a volume of 1.3 mLs so that a concentration of 0.1 mg/mL of RNA is achieved). Because the vial composition is slightly different, it will also come with an orange cap and now 10 doses can be extracted instead of 6.

COVID-19 is a very serious public health issue in this age cohort.

CDC’s Fiona Havers gave an excellent presentation summarizing the public health data for this age group. It’s true that of all members of our population, 5-11 year-olds actually have the lowest risk of hospitalization due to COVID-19 according to COVID-NET data published in the MMWR. However, even if you look at this fact, based on historical estimates, COVID-19 would be the 8th leading cause of death in this age group. This, however, likely understates the seriousness of the problem. In this age group, influenza and pneumonia is the 7th leading cause of death. COVID-19 burden per COVID-NET data was similar to that of 2017-2018 influenza as far as hospitalization based on data from the Flusurv-NET surveillance system that was used to establish COVID-NET. In that dataset, COVID-19 attained a burden in this age group analogous to seasonal flu, an even more prominent cause of death in the age group. 9 hospitalizations due influenza occurred in total in that age group over the same time period, probably because of non-pharmaceutical interventions. Furthermore, COVID-19 hospitalizations are more likely to require ICU admission, invasive mechanical ventilation, and have longer hospitalizations than those for influenza in this age group. It is therefore probable that the burden of COVID-19 in children of this age group is substantially greater than that of influenza if not for non-pharmaceutical interventions implemented to control COVID-19 that also substantially restricted other respiratory viruses. Prominent risk factors for hospitalization include obesity, which is rising across the US in this age group, and feeding tube dependence. However, one-third of hospitalized children have no prior medical conditions. I mention this not because the children who do are less important but to underscore that a severe case of COVID-19 is something that no one should presume to be exempt from, even if some groups are at more risk than others. The burden of COVID-19 among this age group is also disproportionate with Hispanic and Black children accounting for 68% of hospitalized cases. Hospitalizations surged with the rise of the delta variant, though, as an absolute proportion of cases, remained a rare occurrence. Data from the UK also suggest that despite recovery, a significant proportion of children experience persistent symptoms for more than 12 weeks -7 to 8%- even after mild infections. The nature of the symptoms is similar to the PASC experienced by adults with significant effects on performance including in school. This is particularly concerning for how it can disrupt children’s success long-term.

• We are not overestimating COVID-19 hospitalizations in this age group. The data presented earlier are from COVID-NET, which covers about 10% of the US population and includes 14 states and 250 medical centers. A major concern raised several times by members of the public and VRBPAC members is whether or not COVID-19 cases included in the COVID-NET data could reflect incidental infections as patients have to be screened as part of infection control measures. Dr. Havers of the CDC presented the data and explained that this is not likely to be a significant issue because the CDC has officers who review charts to see what the reasons for hospitalizations were, and thus incidental COVID-19 is not included. It is actually possible that they may underestimate it because sometimes the child comes in for an unrelated condition, is found to have COVID-19, and then that becomes the principal reason for their illness. So, the unfortunate and very sobering data we have here are likely accurate.

Myocarditis

CDC’s Matthew Oster, a pediatric cardiologist, presented on myocarditis, largely summarizing the findings from the preprint he recently co-authored with some additional important notes. Broadly, when it comes to myocarditis here, we can divide it into 4 types: classic myocarditis (which here refers to all non-COVID-19-related forms of myocarditis, most of which are related to infection but not always), COVID-19 myocarditis (myocarditis resulting from acute COVID-19), MIS-C myocarditis (myocarditis resulting from MIS-C), and vaccine myocarditis. Though these are all myocarditis, they differ in very important ways. Vaccine myocarditis does resemble classic forms of myocarditis in terms of the symptoms and presentation- but the outcomes are dramatically different. Among children in the US, there has been no confirmed mortality associated with vaccine myocarditis. Classic myocarditis on the other hand is fatal in 4-7% of children acutely, and requires a heart transplant in 7-9% of children, though children do tend to do well after the acute episode. Outcomes in adults for classic myocarditis are far worse. Ejection fraction (EF) is a parameter of cardiac function which describes the proportion of blood ejected from the heart with each beat, with normal values being between 55 and 75% depending on the source you consult. Both MIS-C and vaccine myocarditis show a rapid normalization in EF. In contrast, classic myocarditis can cause persistent deficits in EF after recovery. COVID myocarditis is not shown in the figure because it appears to be too rare for the center the study was done to identify any cases. Preliminary data suggest that myocarditis from COVID-19 in kids is almost always due to MIS-C, occurring in about 8-9% of cases in the preliminary dataset described (though other sources report values as high as 75%). Myocarditis does have an associated risk of sudden death in later life among adults, usually noted on autopsy, typically when it fails to be diagnosed. Right now, long-term outcome data are unavailable but in the short and medium term (3-6 months after the episode) most outcomes look very good with normalization of function and return to normal ECG. However, there is the question of late gadolinium enhancement (LGE) on cardiac MRI noted with some cases of vaccine myocarditis. This finding is indicative of scarring and inflammation in the myocardium, and in adults, has a robust association with bad outcomes. In children however, its meaning is very controversial and unclear, especially after everything else normalizes. Other factors that are relevant regarding vaccine myocarditis are the symptoms it presents with as they are prognostic indicators. Vaccine myocarditis basically universally presents with chest pain, which is associated with better outcomes, and also commonly shows ST wave abnormalities on ECG which is also associated with better outcomes. The finding of LGE is prognostically concerning, however. Overall, the available evidence suggests that MIS-C myocarditis and vaccine myocarditis recover quite well after the acute episode but the long-term consequences remain to be determined.

• A few aspects of classic myocarditis’s epidemiology are really critical and can lend clues for what to expect in this younger age group. First is the sex disparity: it biases males significantly and peaks among younger adult and adolescent males, dropping at the prepubescent period. This is thought to reflect a role for sex hormones, especially testosterone. When looking at females, the risk across age group is relatively constant. Notably, with the vaccine-associated cases, the risk drops when you look at 12-15 year olds relative to 16-19 year olds. Excess risk of vaccine myocarditis in females at virtually any age is very small, and though there is a significant increase in the relative risk of the condition, the absolute increase in terms of case numbers is also small for males. Given the natural history and epidemiology of classic myocarditis, and potentially the role of the smaller dose, the risk of myocarditis for 5 to 11-year-old kids, in particular males, is likely to be quite low. Indeed, no cases of myocarditis were observed in Pfizer’s trial despite approximately 3000 children in the age group receiving it.

• Myocarditis may sometimes be asymptomatic, and the significance of such cases is controversial. Sudden cardiac death in adults may occur following resolution of acute myocarditis even with functional normalization (but in kids the picture is a lot murkier) and it is noted on 5 to 10% of autopsies on young adults who die suddenly. For this reason, work is underway to define the burden of asymptomatic myocarditis in this age group. The issue, however, is that there really isn’t a good way to do that. The principal mechanism one might consider is measuring blood troponin levels. Troponin is a marker of myocardial injury and is released when the heart experiences damage. The issue is, it is not specific to myocarditis and is released into the blood with any kind of damage. In fact, it can become elevated from vigorous exercise alone. Because the background rate of the condition is so rare, if you go by troponins, it is probable that most of your positive myocarditis cases are going to be false positives, rendering finding an accurate rate impossible. Pfizer is currently trying to work out a way to study this in more detail, however.

“We won’t know about the safety of the vaccine until we roll it out”

I can’t not discuss this one because this comment was made offhandedly by several VRBPAC members and if you don’t understand exactly what it means it sounds very alarming when it’s just not, but translated into less alarming language: any concerning issues (e.g. something requiring hospitalization like myocarditis) are too rare for a trial that included thousands of children to detect and so to see those very rare adverse events, we need to roll out to a much larger group.

Firstly, there was a lot of discussion about the decisions made by VRBPAC being used to mandate the vaccine from this age group, even among VRBPAC members, which is just not how anything works (truthfully multiple members of the committee seemed confused about their job description). Their decision has absolutely nothing to do with that. VRBPAC isn’t the be-all end-all for vaccines in the US, and, in general, mandates (at the state or federal level in the US) cannot be issued for vaccines that are under EUA only. They advise the FDA on what indications the vaccine product should be authorized for and the FDA considers their non-binding recommendations, almost always going with them, and choosing whether to grant an EUA or a BLA (full licensure) depending on the occasion. From there, the CDC’s Advisory Committee on Immunization Practices deliberates on what to recommend as far as using the vaccine. Broadly, the task of VRBPAC is to judge whether or not the known and potential benefits of the vaccine product exceed the known and potential risks, essentially in a vacuum, for the given proposed indication (in this case, prevention of COVID-19 in 5- to 11-year-olds). ACIP ultimately decides for whom to recommend the vaccine and under what circumstances. From there, the CDC director makes a decision based on their non-binding recommendations, also almost always going along with it.

Anyway, what’s this now about us not knowing the safety of the vaccines? I mean… we have a giant briefing document and data from the Sponsor that has been reviewed by an independent data safety monitoring board (DSMB) for accuracy looking at the safety of this vaccine 2 months after ~1500 children in this age group received it, and another ~1500 children with about 2.4 weeks of follow-up data (enough time to see almost all cases of myocarditis, which is the principal risk we are concerned with). No adverse events of special interest (e.g. myocarditis) nor serious adverse events related to the vaccine were noted (there were a few serious adverse events in the vaccine group- for instance, one child swallowed a penny; a good reminder that an adverse event gets reported even if it’s unrelated to the intervention- although I am positive there are some who would try to rationalize that a child eating a penny is somehow the fault of the vaccine). So basically, we have data on about ~3000 children that shows no apparent safety issues, and we have a bag of biostatistical tools that includes the rule of threes. Basically, because this is a sample of about 3000 kids, we can be 95% confident that the risk of any of the adverse events that did not occur in the trial is at most 1 per 1000 kids in this age group (and in the case of myocarditis, which is the main one we are interested in, because of all the factors discussed earlier, it’s probably quite a bit lower than that because myocarditis risk even from the highest estimates in the highest risk age groups and male sex does not approach 1 per 1000). I think that’s good news. The issue is, the trial cannot detect very rare adverse events. For example, anaphylaxis occurs in about 5 per million doses, so to reliably capture it, you need a sample of 300,000 people getting 2 doses of vaccine. Hence, the trial cannot estimate the risk of very rare adverse events, just like all pre-licensure trials. This is what VRBPAC members mean when they say that the safety cannot be known until the vaccine is rolled out: we cannot detect very rare adverse events in this age group until hundreds of thousands of children in this age group get the vaccine, and that’s likely where we will see myocarditis, if it is to occur at a rate beyond the background. It was also noted in the FDA briefing documents and during the meeting that according to the COVID-19 data tracker, 205,046 individuals less than 12 years of age have received at least one dose and 125,656 are fully vaccinated, and Dr. Cohn reported that no excess adverse events were noted in this group during the meeting. However, many caveats apply there because some of these may be misclassification (e.g. someone made a typo entering a date of birth), and detailed data were not presented. Additionally, it’s unknown what formulation of the vaccine they received (i.e. 10 mcg vs 30 mcg vs some other dose) as of the time of writing this. This might be discussed more at the ACIP meeting.

In short, the VRBPAC meeting deliberations and clinical trial data did not reveal any tangible safety issues for this vaccine in this age group and any problems that will be given special attention are based on both precautionary principles and the history of the vaccine’s use for months in hundreds of millions of people globally. Note that for all of the concerns they raised during the meeting, 17 of the 18 members voted in favor of granting the vaccine EUA in this age group and the remaining member who did not abstained from the vote.

Miscellaneous FAQs

• What about the public comments?

  • Anti-vaccine zealots love the public comment period at ACIP and VRBPAC meetings because they can appear as official-looking experts and spread lies and distort facts about the vaccines. Unfortunately, the bulk of comments at this VRBPAC meeting were issued by such zealots. Most of what was claimed was shown explicitly to be false by the presentations during the VRBPAC meeting, so I recommend either watching the presentations or reading the Twitter thread to address them in detail. I will make note of two particular individuals, however. One is Steve Kirsch, who has previously offered his own fortune to explore early treatments for COVID-19 which is an extremely noble and commendable thing to do. Unfortunately, things soured quite a bit since that, and he decided he knew better than the investigators he himself was funding about what did and didn’t work, and has become a tireless source of anti-vaccine falsities. Kirsch holds no relevant expertise about the subjects he opines on and simply has more stamina than many of his rhetorical opponents. Kirsch has previously collaborated with Jessica Rose who claims the title of viral immunologist in this meeting for her comment, and then proceeds to shamelessly lie (though the possibility exists that her grasp of immunology is so poor that she makes the errors unintentionally). Specifically, Rose insinuates that the vaccines induce a Th2-driven immunopathology in the form of myocarditis (the use of the mRNA platform was designed specifically to avoid the risk of Th2 immunopathology, which myocarditis in general is not regarded as being part of), when in reality the vaccines were demonstrated in preclinical and early clinical trials to produce a strongly Th1-biased response with “minute or undetectable levels of the T-helper-2 (TH2) cytokines”, and detailed profiling of myocarditis patients shows no apparent evidence for a Th2-driven response. It’s also worth noting that she and Peter McCullough, acolyte of hydroxychloroquine and grifter currently engaged in a lawsuit with Baylor College of Medicine for claiming an affiliation there that he does not hold, currently have a publication that is currently experiencing a temporary removal (from a skim of the abstract though, it appears to be a classic example of VAERS dumpster diving, which definitely should not meet publication standards for… well… any journal). My honest advice is to ignore public comments. Occasionally, good and honest people raise valid concerns or points about the subject matter, but overwhelmingly it is an opportunity for anti-vaccine activists to espouse lies that, even if easy enough to spot as untrue, can create inappropriate doubts that can make you question good decisions.

• Can we give the vaccine by a different route e.g. intradermally vs intramuscularly?

  • There is no data on what, if anything, this would accomplish. I think this is a reference to the recent mouse study showing intravenous administration of the mRNA vaccines caused myocarditis in mice but few things to note: (1) the doses the mice received are far higher than the equivalent dose in humans, (2) intradermal route for vaccination is used only with the BCG vaccine. This isn’t something that should be attempted without at least a study showing that the immune response is just as good and there aren’t obvious safety issues.

• Can the dosing interval be extended?

  • It’s not currently recommended (though you don’t need to restart the series if you go past 3 weeks), but going more than 3 weeks seems to be fine, and may even have immunological advantages for the durability of the response and quality of the antibodies elicited. However, it’s unclear how this might affect the safety profile e.g. regarding myocarditis. Moderna for instance is given 4 weeks apart (Pfizer/BioNTech’s vaccine is given 3 weeks apart) and seems to have a slightly higher risk of myocarditis, so a longer interval might not reduce it. It’s also possible this is related to the dose. Alternatively, data from Israel shows a much lower risk of myocarditis with dose 3 than dose 2, which is given 6 months later. Is this because of the interval, or because the vast majority of people susceptible to myocarditis from the vaccine already got it with dose 2?

• What is happening pathophysiologically to cause vaccine myocarditis?

  • This is currently unknown but is the subject of active investigation.

• Can we give just 1 dose of the vaccine?

  • Hard to say. To previously infected people? Maybe but this wasn’t really tested. Data suggests that the second dose is really critical for both the T cell and neutralizing antibody response, and seems to be important for maintaining durability of the protection from the vaccine. The issue with giving 1 dose to infected people is part logistical and part scientific. One question is how you define previously infected. For instance, a transient positive PCR may have been an infection but may not have resulted in significant immunity that the vaccines can build on for protection. Using antibody testing is one option, but is logistically difficult, expensive, and runs into epidemiological complexities (i.e. false positives and false-negatives), mostly limiting the use of these tests to serological surveillance. What’s more is that anti-N antibodies, which demonstrate prior infection, wane more quickly than those directed against spike protein, but anti-spike cannot distinguish between vaccine-elicited and those from prior infection. In that sense, it’s possible that anti-spike antibodies might be used, but again, this is subject to the same issues of sensitivity and specificity mentioned earlier. Furthermore, no data currently available suggest that the vaccine is more dangerous in recovered individuals, though there is a consistent theme that the first dose is more intense, but the second is milder.

• Will we need booster doses every 6 months?

  • No evidence currently suggests that. Many vaccines on the childhood immunization schedule require multiple doses as part of a primary series and then may not need boosters or may need them only infrequently. As additional doses of COVID-19 vaccine are given, immunological memory is fortified further and takes longer to decline to baseline.

One more question of importance did arise regarding the possible effects of vaccines on transmission. I have decided to discuss the evidence regarding this in a separate post because of the thoroughness it demands.