According to an FDA Advisory Committee Meeting of experts it is ok to vaccinate children 5-11 years of age against Covid-19. Note that FDA makes judgements about the benefit-risk of drugs but not public health policy. Unfortunately questions, common sense and analysis of publicly available data is always worth your time before you make a decision for your children. Do parents have their kids’ best interest in mind when they make such decisions? According to Terry McAuliffe probably not, he is the one who has their best interest. In the meantime let us review data and think for ourselves. We are citizens, not free citizens now, but we do not pay homage to our authorities, they are at our service.
- Unmet medical need
Excerpt from the Briefing Book (BB) by Pfizer ( publicly available)
“Based on CDC data, among children 5 to <12 years of age, there have been approximately 1.8 million confirmed and reported COVID-19 cases and 143 COVID-19-related deaths in the US through 14 October 2021. In this age group, there have been 8,622 COVID-19-related hospitalizations through 18 September 2021.This translates to cumulative incidence rates of approximately 6000 and 30 per 100,000 for confirmed COVID-19 cases and COVID-19-related hospitalizations, respectively, among children 5 to <12 years of age. The pediatric burden of COVID-19 likely exceeds that of seasonal influenza (recently estimated at 0.8 per 100,000 influenza hospitalizations).”
Data published by CDC includes more than a year worth of data, CDC data describes 78 weeks of data, more than a year. Incidence rates are always described within a time frame reference ( a week, a month, a year). Seasonal flu data are always given in a year period (season 2017-2018 for example) so the is the first caveat of what we are being told. Let us compare incidence rates of flu and covid in 5-11 yeears old children in the same units:
In a population of 28,384,878 children 5-11 there were 8,622 Covid-19 related hospitalizations in 78 weeks. The incidence rate of hospitalization is 30 cases per 100,000 children but (big one) the incidence rate per year is in fact 20.25 cases of hospitalization per 100,000 (78 weeks of data vs 52 weeks in a year). So now let us compare these numbers with the number of hospitalizations for flu in the year 2017-2018 (again cdc data, publicly available). The reported 19,636 hospitalization among 5-17 years of age children produce an incidence rate of 36.5 hospitalizations per 100,000 children 5-17 years of age per season.
If we compare the hospitalization rate for Covid-19 vs that for regular flu, in fact regular flu produced more hospitalizations in the 2017-2018 season than Covid-19 during the pandemic (even not adjusting for the year correction). The briefing book mentions that the hospitalization rate for flu is around 0.8 cases per 100,000. This website reports an incidence rate of 0.8 hospitalizations per 100,000 in the 5-17 years of age children. This web monitors cases of laboratory confirmed cases of flu, and in the fine print of their tables the following statement reads: “laboratory confirmed is dependent on clinician ordered influenza testing. Therefore, the rates are likely underestimated as influenza-related hospitalizations can be missed due to test availability and provider and facility testing practices”.
That is the reason why there is a big discrepancy between data reported by CDC and data reported by this web. However, the authors of the briefing book did not provide all data and the source of the argument to support the statement that Covid-19 had a much higher burden of hospitalization compared to flu in this age group is simply not true.
As per death rate in children 5-11 years of age, Covid-19 death rate is estimated at 0.5 cases per 100,000 in the 78 weeks reported. Adjusting to a yearly rate the incidence is 0.33 deaths per 100,000 children 5-11 years old followed for a year. The death rate per season in 2017-2018 reported by cdc in children 5-17 years of age is 0.77 cases per 100,000. Using the incorrect or the correct incidence rate, flu produced more deaths in kids than Covid-19 with the data at hand.
- Study design
Excerpt from the Briefing Book (BB) by Pfizer (publicly available)
As per the BB: “The disposition of Phase 2/3 pediatric participants 5 to <12 years of age is summarized in Table 2. In total, 1,528 participants were randomized to receive BNT162b2 10 µg and 757 participants were randomized to placebo, reflecting the 2:1 randomization ratio.“
They exposed 1,528 children in their pivotal trial and they followed up study participant 2 months after the second dose. As reference, the pivotal trial for approval in adults, same vaccine, included over 21,000 adult patients exposed to the two injections of the vaccine.
When you design a clinical trial you decide how many patients you need to show your drug is efficacious and also (very important) to show an acceptable safety profile. You need enough patients to detect safety signals that occur with low frequency so that you can have an accurate evaluation of the benefit-risk. Assuming a Poisson distribution for rare events, with over 21,000 adult patients you can detect events that happen with a frequency of 17 cases per 100,000 patients or higher (in summary, pretty rare). If you expose only 1,528 children you can detect events that happen with a frequency 241 cases per 100,000 patients or higher, so you are going to miss events that happen with lower frequency. Those rare events will only be detected when hundreds of thousands of children have already been inoculated. As a reminder, and again data from FDA Briefing Book, the frequency of myocarditis Reporting rates for medical chart-confirmed myocarditis and pericarditis in VAERS (Vaccine Adverse Event Reporting System) have been (~71.5 cases per million second primary series doses among males age 16-17 years and 42.6 cases per million second primary series doses among males age 12-15 years as per CDC presentation to the ACIP on August 30, 2021)”
Translating to our units: the frequency of myocarditis in 12-15 years of age was 4.26 cases per 100,00 patients and 7.15 per 100,000 in the 16-17 years of age. With the current sample size for children 5-11 the possibility of detecting any such low frequency event is impossible (by orders of magnitude). Let us remind ourselves that the frequency of the event we want to prevent ( hospitalization due to Covid-19) has been reported with a frequency between 20 to 30 cases per 100,000 children 5-11 years.
- Population included
Excerpt from the Briefing Book (BB) by Pfizer (publicly available)
As per the BB: “Efficacy endpoints are confirmed COVID-19 incidence from 7 days after Dose 2 per 1000 person-years of follow-up in participants (1) without or (2) with or without serological or virological evidence (prior to 7 days after receipt of Dose 2) of past SARS-CoV-2 infection.”
In summary, they included children with previous infection by SARS-CoV-2, ”baseline positive status for prior evidence of SARS-COV-2 infection was reported for 8.7% of the BNT162b2 group and 8.4% of the placebo group.”
If you exclude those who had the disease already, the sample size decreases to 1,385 exposed to the vaccine and 688 patients in the placebo arm. The difference in results is statistically significant in this sub-group analysis also.
As per the BB: “children with baseline comorbidities …made up approximately 20% of the evaluable efficacy population in this study. No cases in the BNT162b2 group and 3 cases in the placebo group occurred in children with reported baseline comorbidities”
Going back to the numbers it means 303 patients with baseline comorbidities in the treatment arm and 150 in the placebo arm. The difference is statistically significant in this comparison though the sample size is extremely small. But let us do some math here.
This means that there were 1,214 children (healthy overall) in the treatment arm and 601 in the placebo arm. Among those healthy children (80% of the sample) the cases were distributed as follows: 13 cases in placebo vs 3 cases in Vaccine group. The difference continues to be statistically significant.
Of note, the BB says “No cases of COVID-19 were observed in either the vaccine group or the placebo group in participants with evidence of prior SARS-CoV-2 infection”. There is not much evidence to vaccinate children who have already had the disease, something commonly known by the scientific community for decades now.
- Dose
Excerpt from the Briefing Book (BB) by Pfizer (publicly available)
As per the BB: “The similarity in post-vaccination immunogenicity as reflected in Day 7 post-Dose 2 GMTs across 10 µg and 20 µg dose levels, along with the more favorable reactogenicity profile observed in the 10-µg dose level, led to the selection of BNT162b2 at the 10-µg dose level to proceed to Phase 2/3 evaluation for participants 5 to12 years of age “
Let us remind ourselves that the dose approved for 12 years and older is 30-µg dose level, each dose of the two-injection schedule. If we take a look at a growth chart of the US and focus on boys (as the signal for myocarditis was reported mainly in males 12-18 years. In the US ( CDC data) the following numbers can be drawn:
The 50th percentile of weight for boys 5 years of age is 19 kg, 11 years is 38 kg, 12 years is 40 kg and 18 years if 67 kg.
That means that the dose per kg provided by the vaccine in different ages is as follows: 5 years of age 0.52µg/Kg; 11 years of age 0.26 µg/Kg; 12 years of age 0.75 µg/Kg; 18 years of age 0.44 µg/Kg.
The dose was selected based on the probability of achieving a threshold of antibody increase expected to be protective, very reasonable. But in light of the safety signal aforementioned it is worth reviewing the dosing. In fact, as you can see there is an overlap in the exposure per Kg: between 0.26 to 0.52 µg/Kg in the 5-11 age-group and 0.44 to 0.75 µg/Kg in the 12-18 age-group. I wonder if the massive vaccination effort that is coming will also produce a signal in the 5-11 age group.
Of note the same Briefing Book provided by Pfizer described a study in Israel (page 67) that “found an overall standardized incidence ratio of 5.34 after a second dose using a 30-day risk window compared to 2017-2019 rates, driven mostly by the standardized incidence rate (SIR) in males under 30 years of age “. But if you look into the table referenced in fact in males 16-19 years of age the increased risk was close to 9 times (8.96 with a 95% CI between 4.50 to 17.83).
- Results
Excerpt from the Briefing Book (BB) by Pfizer (publicly available)
As per the BB: “No severe COVID-19 cases (per protocol definition, or per CDC definition) were reported in children 5 to <12 years of age as of the data cutoff date (08 October 2021). No cases of MIS-C were reported as of the data cutoff date.”
The vaccine prevented Covid-19 cases but it did not prevent severe cases or hospitalizations. It is not surprising because the study was extremely underpowered to assess that event. In addition, did they measure the end point? How do you know the % of patients who were infected by Covid-19 if you did not do Covid-19 tests? It is known by now that vaccinated patients may have the disease but be asymptomatic, which is a really good thing by the way, but that does not mean the frequency of infection is lower. Right now , with this study, we do not know. Why is this relevant? One of the arguments for vaccinating children in school age is to prevent the spread of the disease. The spread is likely not to be diminished, may be the severity in other age groups. Along those lines, let us review a Study from Sweden published in early 2021 from Sweden. One of the good things about the health care databases in Nordic countries is that they link primary care with prescription and hospital care, and population in general is stable so they have long and reliable follow-up data. Sweden was one of the few countries that decided to keep preschools (generally caring for children 1 to 6 years of age) and schools (with children 7 to 16 years of age) open. Data on severe Covid-19, as defined by intensive care unit (ICU) admission, were prospectively recorded in the nationwide Swedish intensive care registry. In the pre–Covid-19 period of November 2019 through February 2020 was 3.3 deaths per 100,000 children 1-16 years of age and 3.5deaths per 100,000 children 1-16 years of age in the Covid-19 period March through June 2020. It is relevant to note that schoolteachers in Sweden had an incidence rate of intensive care admission of 19 per 100,000. As compared with other occupations (excluding health care workers), this corresponded to a relative risks of 1.10 (95% confidence interval [CI], 0.49 to 2.49) among preschool teachers ( non statistically significant) and 0.43 (95% CI, 0.28 to 0.68) among schoolteachers ( this is a statistically significant lower risk). SO these data underscore that Covid-19 is not causing a big burden or a bigger burden of disease that flu in tow completely different countries.
Conclusions
- Hospitalizations in children 5-11 years of age due to covid-19 seem to be similar or lower than those due to Influenza
- Children with comorbidities might benefit from the vaccine
- Phase 3 trial was extremely underpowered to draw any conclusion on whether the vaccine has any impact on preventing severe forms of disease in children 5-11
- The safety database is extremely inadequate in light of the massive population vaccination that may occur in the coming months and the safety signal identified in children 12-18 years of age.
- Parents, not governments or school boards, should be the decision makers in the vaccination or lack thereof of children 5-11 years of age
- All data shown are publicly available
References
https://www.fda.gov/media/153447/download
https://gis.cdc.gov/grasp/fluview/fluhosprates.html
https://www.cdc.gov/growthcharts/data/set1clinical/cj41l021.pdf
https://www.fda.gov/VRBPAC-10.26.21-Meeting-Briefing-Document-Sponsor-Pfizer-508-Waiver.pdf
https://www.nejm.org/doi/full/10.1056/NEJMc2026670
El Quijote by Miguel de Cervantes, always a good read