Document text
Department of Health and Human Services
Food and Drug Administration
Center for Biologics Evaluation and Research
MEMO
RANDUM
Date: September 13, 2021
To: Ramachandra Naik
From: Hong Yang, Ph.D., Patrick Funk, Ph.D. , and Osman N. Yogurtcu , Ph.D.
Analytics and Benefit -Risk Assessment Team
OBE
Through: Richard Forshee, Ph.D.
Acting Deputy Office Director
OBE
Re: STN 125742/0: review memo for benefit -risk assessment
Page 2 – Hong Yang – 125742/0
Review Memo on Benefit -Risk Assessment of Pfizer Vaccines for Age 16-17 yrs
Reference submission: BLA 125742/0
Reviewers : Hong Yang, Patrick Funk, and Osman Yogurtcu
Date: August 20, 2021
Contents
1. Executive Summary ......................................................................................................................... 3
2. Background and regulatory questions ............................................................................................. 4
3. Methods ........................................................................................................................................... 4
3.1. Model Overview ...................................................................................................................... 4
3.2. Benefits .................................................................................................................................... 5
3.3. Risks ......................................................................................................................................... 6
4. Results .............................................................................................................................................. 9
4.1. Scenario 1: Base c ase .............................................................................................................. 9
4.2. Scenario 2: Most likely scenario .............................................................................................. 9
4.3. Scenario 3: Worst case scenario ............................................................................................. 9
5. Conclusions and discussion ............................................................................................................ 10
6. Limitations ..................................................................................................................................... 11
7. Acknowledgements ....................................................................................................................... 12
References ............................................................................................................................................. 13
List of Tables .......................................................................................................................................... 14
List of Figures ......................................................................................................................................... 14
Supplementary Materials ...................................................................................................................... 26
Page 3 – Hong Yang – 125742/0
1. Executive Summary
FDA conducted a benefit -risk assessment to inform the review of the Biologics License Application (BLA)
for use of the Pfizer -BioNTech COVID -19 mRNA vaccine (also referred to as BNT162b2) among ages 16
years and older. We assessed the ben efits and risks per million individuals who are vaccinated with two
complete doses of BNT162b2 . The analysis was conducted for the groups stratified by combinations of
sex and age (12 -15, 16-17, 18-24, and 25-29 years). The model assesses the benefits of v accine -
preventable COVID -19 cases, hospitalizations, intensive care unit (ICUs) visits, and deaths, and the risks
of vaccine -related excess myocarditis/pericarditis cases, hospitalizations, and deaths. The major sources
of data include age/sex specific COV ID-19 case and hospitalization incidences reported on COVID NET on
July 10, 2021, the myocarditis/pericarditis case rate attributable to vaccine obtained from the OPTUM
health claims database, and the vaccine related myocarditis/pericarditis deaths reporte d through VAERS.
We constructed scenarios for both the most likely short -term moving direction of the pandemic and the
worst case, which used the most conservative assumptions for all model inputs.
The most likely scenario:
We assumed vaccine protectio n duration of 6 -months, 10x COVID -19 case incidence and 4x COVID -19
hospitalization incidence as of July 10, 70% vaccine efficacy against COVID -19 cases, 80% vaccine efficacy
against hospitalization, and no vaccine- related myocarditis death. The model resu lts indicate that, for all
age/sex groups and across all model outcomes, the benefits clearly outweigh the risks. For males 16 -17
years old— the group with the highest risk of myocarditis/pericarditis —the model predicts that
prevented COVID cases, hospitalizations, ICUs, and deaths are 135,771, 506, 166 , and 4, respectively. The
excess myocarditis/pericarditis cases, associated hospitalizations, and deaths attributable to vaccine are 196, 196, and 0, respectively.
The worst -case scenario:
We used the most conservative assumptions for all the model inputs in this scenario. We assumed 6 -
months vaccine protection, the COVID -19 case and hospitalization incidence as of July 10, 2021, 70%
vaccine efficacy against COVID-19 case, 80% vaccine efficacy against COVID -19 hospitalization, and
0.002% myocarditis/pericarditis death rate.
For males 16 -17 years old, the model predicts that prevented COVID cases, hospitalizations, ICUs, and
deaths are 13,577, 127, 41, and 1, respectively. The excess myocarditis/pericarditi s cases and associated
hospitalizations and deaths attributable to the vaccine are 196, 196, and 0, respectively. Even with the conservative assumption on the myocarditis/pericarditis death rate, the model predicted 0 deaths
associated with myocarditis/pericarditis. The model predicted a higher number of
myocarditis/pericarditis related hospitalizations compared to prevented COVID -19 hospitalizations.
However, considering the differential clinical outcomes of the hospitalization from two difference causes,
we consider the benefits of the vaccine still outweigh the risks for the highest risk group, males 16 -17
years old, under this worst -case scenario.
Our results demonstrate that the benefits of BNT162b2 clearly outweigh its risks for all age and sex
Page 4 – Hong Yang – 125742/0
groups we analyzed. However, the benefit -risk estimates are highly uncertain due to the dynamics of
pandemics. Other major uncertainties in benefits are vaccine efficacy and duration of protection in the
face of emerging virus variants. The major risk uncertainty is the data on vaccine -related myocarditis
cases and deaths.
2. Background and regulatory questions
The Pfizer-BioNTech COVID -19 mRNA vaccine (also referred to as BNT162b2) has been recommended for
persons 12 years of age and older in the United States under FDA’s Emergency Use Authorization (EUA) .
Since authorization of mRNA COVID -19 vaccin es (Pfizer-BioNTech and Moderna ), real-world evidence
has indicated the vaccines are effective in preventing COVID -19 cases and related hospitalizations and
deaths . However, increased cases of myocarditis and pericarditis have been re ported in the United States
associated with mRNA COVID -19 vaccination, particularly in adolescents and young adults (Marshall et
al. 2021; Shay et al. 2021; Watkins, et al., 2021) . FDA conducted a benefi t-risk assessment to inform
regulatory decisions related to the Biologics License Application (BLA) for use of BNT162b2 vaccines
among ages 16 years and older . The regulatory question to be answered is whether the benefits of
vaccination outweigh the risks among various age and sex subgroups being considered for approved use
of the vaccine (and in particular , males , age 16 -17 years old), considering the potentially elevated
myocarditis/pericarditis risk after vaccination suggested by post- authorization safety surveillance.
3. Methods
3.1. Model Overview
We assessed the benefits and risks per million individuals who are vaccinated with two complete
doses of BNT162b2 . The analysis was conducted for the groups stratified by combinations of sex
and age (12 -15, 16-17, 18 -24, and 25 -29 years). The m odel assesses the benefits of vaccine -
preventable COVID -19 cases, hospitalizations, intensive care unit (ICUs) visits and deaths, and
the risk s of vaccine related excess myocarditis/pericarditis cases, hospitalizations , and deaths
(Figure 1). The key model inputs include duration of vaccine protection , vaccine efficacy against
COVID -19 case s and hospitalizations , age/sex specific COVID-19 case and hospitalization
incidence rates , age/sex specific vaccine-attributable myocarditis case rate , and myocarditis
death rate (Table 1 ). To evaluate the impact of uncertainty of these key model inputs on the
benefit s and risks , low and high values of these model inputs are used for sensitivity analysis.
Our model generates benefit-risk outcomes for seven scenarios (Table 2 and Supplement Table
S1) with different combinations of the input values. T he three most important scenarios are
presented in the main body of this report : Scenario 1, a base scenario using the COVID -19
incidence on July 10; S cenario 2, the most likely scenario; and Scenario 3 , the worst -case
scenario . Other scenarios are summarized in the supplementary materials of the report .
Page 5 – Hong Yang – 125742/0
3.2. Benefits
3.2.1. Calculation of benefits
Our benefit-risk model has four benefit endpoints (Figure 1): preventable COVID -19 cases,
hospitalizations, Intensive Care Unit admissions ( ICUs ), and deaths. To calculate the potential
COVID -19 cases preventable by vaccine ( 𝐶𝐶𝑃𝑃), we use Equation 1
𝐶𝐶𝑃𝑃= 𝐼𝐼𝐶𝐶
𝑃𝑃𝑈𝑈𝐿𝐿 𝐷𝐷 𝐸𝐸𝐶𝐶 Eq. 1
where 𝐼𝐼𝐶𝐶 is the COVID -19 case incidence rate, 𝑃𝑃𝑈𝑈 is the proportion of the population that is at
risk (i.e. , unvaccinated), L is the duration of vaccine protection , D is the number of second vaccine
doses administered (fixed at 1 million) , and E is the vaccine efficacy against COVID -19 cases. For
preventable COVID -19 hospitalizations ( 𝐻𝐻𝑃𝑃), we use a similar equation (Equation 2) in which we
consider the COVID -19 hospitalization incidence rate ( 𝐼𝐼𝐻𝐻) and vaccine efficacy against
hospitalization ( 𝐸𝐸𝐻𝐻).
𝐻𝐻𝑃𝑃= 𝐼𝐼𝐻𝐻
𝑃𝑃𝑢𝑢𝐿𝐿 𝐷𝐷 𝐸𝐸𝐻𝐻 Eq. 2
The pr
eventable COVID -19 ICUs (𝐼𝐼𝑃𝑃) and preventable COVID -19 deaths (𝐷𝐷𝑃𝑃) are fractions of 𝐻𝐻 𝑃𝑃,
such that 𝐼𝐼 𝑃𝑃=𝑓𝑓𝐼𝐼𝐻𝐻 𝐻𝐻𝑃𝑃 and 𝐷𝐷𝑃𝑃=𝑓𝑓𝐷𝐷𝐻𝐻 𝐻𝐻𝑃𝑃.
We
perform these calculations over the individual age and sex groups and combined male and
female groups.
3.2.2. Data and assumptions
3.2.2.1. Duration of vaccine protection
We assume the vaccine has at the least 6 month s of protection since this is the period
examined by Pfizer in their ongoing study (Thomas et al., 2021) . The model assesses
the benefits for a period of 6 month s post 2nd dose of vaccination. For the sensitivity
analysis in the supplement, we use a protection period of 12 months as an upper
bound. For simplicity, t he model does not account for the benefits of partial
vaccination (protection between the first and second dose ) or the second order
benefits of reducing the risk of transmission of COVID -19.
3.2.2.2. Incidences of COVID -19 case, hospitalization, ICU , and death
We assume the incidence rates of COVID -19 case and hospitalization remain constant
over the assessment period (next 6 or 12 months). The incidence rates of COVID -19
cases as of week July 10 are obtained from COVID NET for all sex /age groups. Four -
week average s of incidence (6/26-7/10) are used due to the variability in rates given
Page 6 – Hong Yang – 125742/0
the small numbers of hospitalizations per age/sex groups . The percent of
hospitalizations going to ICU and the percent of hospitalized patients who die are
estimated based on cumulative rates of hospitalizations, ICUs, and deaths for each
sex/age groups reported on COVID NET since March 2020. All the incidence data for
these factors are summarized in Table 3. Comparing the incidence from the 2nd week
of August with those reported at the lowest point in the summer, we find a 10 -time
incidence and 4- time hospitalization increase over a 6 -week period. Considering the
great uncertainty in COVID -19 incidence during the pandemic, we conduct a
sensitivity analysis using a 10-time s multiplier for case incidence and 4-time s
multiplier for ho spitalization incidence in the sensitivity analysis. The m ultipliers
were derived from the public data in COVID data tracker and COVID NET, respectively,
to project the increase in COVID -19 infections/hospitalizations.
3.2.2.3. Unvaccinated population
We estimate the unvaccinated population among each age/sex groups using US
census data and “Age groups of people with at least one dose” from COVID data
tracker . Data for Texas is not contained in COVID data tracker so we impute
proportional vaccination counts based on population averages from the census data. The incidence of COVID -19 cases and hospitalization , described in section 2.2.2.2
“Incidences of COVID -19 case , hospitalization, ICU and death ,” are converted into the
incidence of COVID -19 cases and hospitalization s among unvaccinated individuals of
each age/sex group.
3.2.2.4. Vaccine efficacy
We use vaccine efficacy rates of protection against COVID -19 cases of 70% and 90%
and vaccine efficacy rates of protection against COVID-19 hospitalization s of 80% and
90% in different scenarios. The high efficacy of 90% represents the lower bound of
the confidence interval from the clinical trial data (Oliver et al . 2020). The low efficacy
of 70% for cases and 80% for hospitalization represents a conservative efficacy rate
given the uncertainty o f the vaccine’s protection against the Delta variant. Early
studies on the vaccine’s efficacy against cases from the Delta variant suggest 79% in
Scotland (Sheikh et al ., 2021), 87% in Canada (Nasreen et al ., 2021), and 88% in India
(Lopez Bernal et al. , 2021). To remain conservative in the face of uncertainty in this
rapidly changing pandemic , we use a lower bound of vaccine efficacy from these early
reports.
3.3. Risks
3.3.1. Calculation of risks
Our benefit-risk model has three risk endpoints (Figure 1): excess myocarditis/pericarditis cases,
hospitalizations , and deaths. Estimates of e xcess cases of myocarditis/pericarditis are calculated
by subtracting the background rate of myocarditis in Optum’s sample population from 2019 from
Page 7 – Hong Yang – 125742/0
the rate of myocarditis in the study window from 12/10/2020 – 07/10/2021. We use Equation 3
to calculate excess cases of myocarditis/pericarditis (M Exc) per one million fully vaccinated
individuals.
M𝐸𝐸𝐸𝐸𝐸𝐸=�𝑀𝑀𝑂𝑂𝑂𝑂𝑂𝑂1−𝑀𝑀𝐸𝐸𝐸𝐸𝐸𝐸1+𝑀𝑀𝑂𝑂𝑂𝑂𝑂𝑂2−𝑀𝑀𝐸𝐸𝐸𝐸𝐸𝐸2�∗𝐹𝐹 Eq. 3
Mobs1 and MExp1 are observed and expected myocarditis/pericarditis case rate s post dose 1 , Mobs2
and MExp2 are corresponding case rates post dose 2, and F is a multipl ier for unit conversion .
Expected myocarditis/pericarditis case rates are the predicted background case rate
unassociated with vaccine.
The number of myocarditis hospitalization (M H) and deaths ( MD) are fractions of excess
myocarditis/pericarditis cases (M Exc), such that M H = M EXC* F HM and MD = M exc * fDM .
3.3.2. Data and assumptions
3.3.2.1. Myocarditis /pericarditis attr ibutable to vaccine
We use myocarditis/pericarditis reports data provided by Acumen LLC that are
derived from the Optum health claims database (Table 5) . Acumen reports cases of
myocarditis in 7-, 21-, and 42 -day risk windows from each vaccine dose . Our analysis
focuses on the 7 -day risk window where most cases are found for all groups . The
database contains rates of expected ( MExp 1 and MExp 2) and observed ( MObs1 and MObs2)
myocarditis/pericarditis in 100k person -years for the 1st and 2nd dose of the vaccine .
Converting from 100k person-years in the risk window to one million vaccinated
individual s’ daily risk, we multiply the rates by a factor F = (7*10)/365. This factor is
used to convert the rate per 100k person years to a n expected case count for one
million full vaccinations assuming a 7 -day risk window. Confidence intervals for the
myocarditis cases are calculated using the chi -square method for Poisson distribution
of rare events ( Garwood, 1936) .
3.3.2.2. Myocarditis /pericarditis hospitalization and death rate
Almost all adolescent and young adult patients with suspected
myocarditis/pericarditis cases are hospitalized and monitor ed for the condition. In
this mo del, w e assume all myocarditis/pericarditis cases are hospitalized, but Vaccine
Adverse Events Reporting System (VAERS) data show median stay lengths of one day
for observation .
A total of 1,0 61 myo carditis cases among US <30 year s old after vaccination with
BNT162b2 are reported through VAERS . Among them , two deaths are reported. The
search terms used for query and t he narratives for two death s who had vaccination
with BNT162b2 are included in the supplement. Review of the available data by FDA
Page 8 – Hong Yang – 125742/0
and CDC indicates that both cases are unlikely to be related to the vaccine. In our
model, we assume the death rate related to vaccine is most likely to be zero in the
base case and most likely scenarios (Scenario 1 and 2). However, we use 2/1 ,061 as
the death rate for the worst -case scenario (Scenario 3) to account for the very unlikely
outcome of these two deaths being attributed to vaccine related myocarditis/pericarditis.
Page 9 – Hong Yang – 125742/0
4. Result s
This section summarizes the results for three major model scenarios .
4.1. Scenario 1 : Base case
Our model scenario s start with the base case that is using the most recent available incidence
data on July 10, 2021 and assume a 6-month vaccine protection period, 90% vaccine efficacy
against both COVID -19 case and hospitalization, and zero myo carditis/pericarditis death rate .
Figures 2 , 3, and 4 summarize the results for analyses of combined male/female, male only, and
female only, respectively. The results indicate that benefit-risk is more favorable for male and
female co mbined , female only, and male >18 years old. The model predicted far more prevented
COVID -19 cases com pared to excess myocarditis/pericarditis for male 12-15 and 16 -17 years
old, but the model predicted 142 prevented COVID -19 hospitalizations vs. 196
myocarditis/pericarditis hospitalizations f or male age 16 -17 years old and 122 prevented COVID -
19 hospitalizations vs. 179 myocarditis/pericarditis hospitalizations for male 12-15 years old.
However, hospitalizations associate d with COVID -19 have more sever e clinical outcomes than
those associated with myocarditis/pericarditis. For this reason, we consider that the benefits of
the vaccine outweigh the risks in this scenario even for male age 12-15 and 16-17 years old. See
Table 5 for details and the benefit -risk results for 16-17 year olds.
4.2. Scenario 2: Most l ikely scenario
We constructed a scenario that most likely represen ts the short -term moving direction of the
pandemic . W e assume 6-mo nth vaccine protection , and 10X higher COVID -19 case incidence
and 4X higher COVID -19 hospitalizations incidence compar ed to the incidence on July 10 . We
also assume lower vaccine efficacy (70% against COVID -19 case , 80% against hospitalization)
against new ly emerging virus variants such as D elta strain . We assumed zero myo carditis death
rate based on our best knowledge on the vaccine related myocarditis .
Figures 5, 6, and 7 summarize the results for analyses of combined male/female, male only, and
female only , respectively. For all age/sex groups and across all attributes, the benefits clear ly
outweigh the risks in this scenario . See Table 5 for details and benefit-risk results for 16-17 year
olds.
4.3. Scenario 3: Worst c ase scenario
We also constructed the worst -case scenario using the most conservative assumption s for all
the model inputs. We assumed 6-month vaccine protection , the COVID -19 incidence as of July
10, 2021, 70% vaccine efficacy against COVID -19 case , 80% vaccine efficacy against COVID -19
hospitalization, and 0.002% ( 2/1,061) myo carditis /pericarditis death rate .
Figures 8, 9, and 10 summarize the results for analyses of combined male/female, male only,
and female only , respectively. Even with the conservative assumption on
Page 10 – Hong Yang – 125742/0
myocarditis/pericarditis death rate, the model predicted 0 death s associated with
myocarditis/pericarditis compared to 1 prevented COVID -19 death for both male 12-15 and 16-
17 year old groups. The model predicted 127 prevented COVID -19 hospitalization s vs 196
myocarditis/pericarditis hospitalizations for male age 16 -17 years old and 1 09 prevented COVID -
19 hospitalizations vs 179 myocarditis/pericarditis hospitalizations for male 12 -15 years old.
Considering the differential clinical outcomes of the hospitaliz ation from two different causes ,
we consider the benefits of the vaccine still outweigh the risks in this “worst case scenario” . See
Table 5 for details and the benefit -risk results for 16-17 year olds.
5. Conclusions and d iscussion
Our results demonstrate that the benefits of BNT162b2 clearly outweigh its risks for all age and sex
groups we analyzed. Under the base case scenario and the worst -case scenario (Scenario 1 and 3), we
predicted a higher number of myocarditis hospitalizations than the COVID -19 hospitalizations among
male 16 -17 years old; however, considering the different ial clinical implications of COVID -19 and
myocarditis hospitalization, we consider the benefits of the vaccine still outweigh its risks. Moreover ,
under all other scenarios including the most likely ( Scenario 2), our model predicted that preventable
COVID -19 cases, hospitalizations, and deaths exceed the myocarditis cases and related hospitalizations
and deaths for all age and sex groups.
We note that COVID -19 incidence highly influences the predicted benefits of the vaccine. If the disease
incidence is higher, the benefits of the v accine will be greater , and vice versa. Therefore, the bene fit-risk
conclusion may change if the COVID -19 incidence rate becomes very low in the fut ure. Also, “the worst -
case scenario” presented here is the worst only among the modelled scenarios. Scenarios worse than
Scenario 3 could occur if the data fall outside the ranges of model inputs we used, such as lower COVID -
19 incidence than those reported on July 10 , lower vaccine effectiveness against COVID-19 cases ( <70%)
and against hospitalizations ( <80%), and shorter vaccine protection duration (< 6 months).
Page 11 – Hong Yang – 125742/0
6. Limitations
• (BENEFIT) The constant COVID -19 incidence rate assumption in our model generates high uncertainty
on the estimate of benefits considering the uncertain dynamic s of the pandemic . Additionally ,
estimated b enefits of the vaccine would decrease if the vaccine become s less effective against novel
variants of COVID -19. The d urability of vaccine protection is another source of uncertainty for the
model . Any significant w aning of vaccine -induced immunity before 6 or 12 months would reduce the
benefit of the vaccine .
• (RISK) There is uncertainty in the myocarditis case and death rates attributable to the vaccine. In the
US, two deaths among those less than 30 years old occurred following the administration of
BNT162b2 and were evaluated by FDA and CDC. Based on the review of the available clinical
information, the cause s of death for b oth cases are not thought to be related to vaccination. To
estimate myocarditis/pericarditis risk attributable to the vaccine, health claims data are used, which
have inherent limitations such as small sample sizes for these rare outcomes. The cases have not
been validated by medical chart review. The crude myocarditis rate in our model was adjusted using myocarditis 2019 background rate, which did not account for COVID -19 infection related risk of
myocarditis/pericarditis and may lead to overestimating the risk attributed to the vaccine .
• (BENEFIT -RISK BALANCE ) Some benefit -risk endpoints in our assessment are difficult to compare
directly, for example, hospitalizations from COVID -19 and myocarditis hospitalizations. This benefit-
risk assessment does not co nsider the potential long -term health impacts of COVID -19
or myocarditis . Also, it does not include secondary benefits and risks , such as any potential impact
on the public trust in COVID -19 vaccines and the benefit of the vaccine in reducing the
viral transmission in the population . In this analysis, we did not investigate the benefit s and risks of
subpopulation s with comorbidity due to limited information . The benefit -risk profile could be
different depending on the individual ’s health condition.
Page 12 – Hong Yang – 125742/0
7. Acknowledgements
We thank J. Rosser Matthews , Ph.D., and Katherine Scott , M.D., for editing th is Benefit -Risk review
memo.
Page 13 – Hong Yang – 125742/0
References
1. Garwood, F. "Fiducial limits for the Poisson distribution." Biometrika 28.3/4 (1936): 437 -442.
2. Lopez Bernal, Jamie, et al. "Effectiveness of Covid -19 vaccines against the B. 1.617. 2 (delta) variant." New
England Journal of Medicine (2021).
3. Marshall, Mayme, et al. "Symptomatic acute myocarditis in seven adolescents following Pfizer- BioNTech
COVID -19 vaccination." Pediatrics (2021): 2.
4. Nasreen, Sharifa, et al. "Effectiveness of COVID -19 vaccines against variants of concern,
Canada." Medrxiv (2021).
5. Oliver, Sara E., et al. "The advisory committee on immunization practices’ interim recommendation for use of
Pfizer-BioNTech COVID -19 vaccine —United States, December 2020." Morbidity and Mortality Weekly Report
69.50 (2020): 1922.
6. “Population -Level Risk -Benefit Analysis.” Centers for Disease Control and Prevention, Centers for Diseas e
Control and Prevention, 5 May 2021, www.cdc.gov/vaccines/covid-19/info -by-product/janssen/risk -benefit -
analysis.html .
7. Shay, David K., Tom T. Shimabukuro, and Frank DeStefano. "Myocarditis occurring after immunization with
mRNA -based COVID -19 vaccines." JAMA cardiology (2021).
8. Sheikh, Aziz, et al. "SARS-CoV -2 Delta VOC in Scotland: demographics, risk of hospital admission, and vaccine
effectiven ess." The Lancet (2021).
9. Thomas, Stephen J., et al. "Six Month Safety and Efficacy of the BNT162b2 mRNA COVID- 19
Vaccine." medRxiv (2021).
10. Watkins, Kevin, et al. "Myocarditis after BNT162b2 vaccination in a healthy male." The American Journal of
Emergency Medicine (2021).
Page 14 – Hong Yang – 125742/0
List of Tables
Table 1. Low and high values for model input parameters considered in our sensitivity analysis ........... 15
Table 2. The three main model scenarios with different combinations of model input values that are
shown on Table 1 ....................................................................................................................................... 15
Table 3. Vaccine coverage and COVID incidences by sex and age groups ............................................... 15
Table 4. Optum-reported myocarditis cases for 1 million fully vaccinated individuals by age and sex.
95% confidence intervals for the rates are shown in brackets ................................................................. 15
Table 5. Model predicted benefit -risk outcomes of Scenarios 1 -3 for the 16-17-year-old groups ........... 16
List of Figures
Figure 1. Benefits -risks value tree. ............................................................................................................ 16
Figure 2. Results of Scenario 1 for combined male and female populations ............................................ 17
Figure 3. Results of Scenario 1 for the male population ........................................................................... 18
Figure 4. Results of Scenario 1 for the female population ........................................................................ 19
Figure 5. Results of Scenario 2 for combined male and female populations ............................................ 20
Figure 6. Results of Scenario 2 for the male population ........................................................................... 21
Figure 7. Results of Scenario 2 for the female population ........................................................................ 22
Figure 8. Results of Scenario 3 for combined male and female populations ............................................ 23
Figure 9. Results of Scenario 3 for the male population ........................................................................... 24
Figure 10. Results of Scenario 3 f or the female population ...................................................................... 25
Page 15 – Hong Yang – 125742/0
Table 1 . Low and high values model input parameters considered in our sensitivity analysis
Model inputs Low High
Vaccine protection period 6 months 12 months
Vaccine efficacy against cases 70% 90%
Vaccine efficacy against hospitalization 80% 90%
COVID -19 case incidence rate July 10 rate 10X July 10 rate
COVID -19 hospitalization case incidence rate July 10 rate 4X July 10 rate
Myocarditis death rate 0% 0.002%
Table 2. The three main model scenarios with different combinations of model input values that are shown on
Table 1
Scenario Protection
period Efficacy
against
cases Efficacy against
hospitalization COVID -19
case
incidence COVID -19
hospitalization
incidence Vaccine
attributable
myocarditis
death rate
Scenario 1 Low High High Low Low Low
Scenario 2
(Most Likely) Low Low Low High High Low
Scenario 3
(Worst Case) Low Low Low Low Low High
Table 3 . Vaccine coverage and COVID incidences by sex and age groups
Sex Age group Population1 Vaccinated
population2 COVID -19
cases/100k
persons3 Hospitalizati
ons/100k
persons3 Percent of
hospitalized
going to ICU3 Percent of
hospitalized
who die3
Female 12-15 8,183,216 2,886,252 37.3 0.671 23.9 0
16-17 4,119,686 1,985,672 47.9 1.593 19.5 0.7
18-24 14,923,948 8,033,040 64.6 2.025 8.1 1
25-29 11,428,122 5,918,524 68.6 2.45 5.9 0.3
Male 12-15 8,535,307 2,815,693 33.1 0.35 31.8 0.9
16-17 4,300,731 1,826,299 42.9 0.35 32.7 0.7
18-24 15,633,953 7,217,945 53.3 0.8 22.2 0.6
25-29 12,036,982 5,592,473 57.8 0.875 22.7 1.5
Source: 1-CDC Wonder, 2-COVID Data Tracker , 3-COVID NET
Table 4 . Estimated excess number of myocarditis /pericarditis cases for 1 million fully vaccinated individuals
with Pfizer BNT162b2 by age and sex. 95% confidence intervals for the rates are shown in brackets
Sex Age (years) Rate of excess myocarditis /pericarditis
per 1 million full y vaccinat ed and 95%
confidence intervals
Male 12-15 179 [38, 332]
16-17 196 [36, 424]
18-25 131 [27, 224]
26-35 49 [0, 123]
Female 12-15 32 [0, 235]
16-17 36 [0, 298]
18-25 57 [9, 147]
26-35 2 [0, 80]
Source: Optum Database pre-adjudicated claims 12/11/2020 – 07/10/2021
Page 16 – Hong Yang – 125742/0
Table 5 . Model predicted benefit -risk outcomes of Scenarios 1- 3 for the 16- 17-year -old groups
Benefits Risks
Scenario Prevented
COVID -19
Cases Prevented
COVID -19
Hospitalizations Prevented
COVID -19
ICUs Prevented
COVID -19
Deaths Excess
Myocarditis
Cases Excess
Myocarditis
Hospitalizations Excess
Myocarditis
Deaths
Males & Females
Scenario 1 19,425 241 59 2 116 116 0
Scenario 2 151,080 855 210 6 116 116 0
Scenario 3 15,108 214 52 1 116 116 0
Males only
Scenario 1 17,456 142 47 1 196 196 0
Scenario 2 135,771 506 166 4 196 196 0
Scenario 3 13,577 127 41 1 196 196 0
Females only
Scenario 1 21,657 350 68 2 36 36 0
Scenario 2 168,443 1245 243 9 36 36 0
Scenario 3 16,844 311 61 2 36 36 0
Figure 1 . Benefits -risks value tree.
Source : Reviewer Analysis
Page 17 – Hong Yang – 125742/0
Figure 2. Results of Scenario 1 for combined male and female populations
Source: Reviewer Analysis
Page 18 – Hong Yang – 125742/0
Figure 3. Results of Scenario 1 for the male population
Source : Reviewer Analysis
Page 19 – Hong Yang – 125742/0
Figure 4 . Results of Scenario 1 for the female population
Source : Reviewer Analysis
Page 20 – Hong Yang – 125742/0
Figure 5 . Results of Scenario 2 for combined male and female populations
Source : Reviewer Analysis
Page 21 – Hong Yang – 125742/0
Figure 6 . Results of Scenario 2 for the male population
Source : Reviewer Analysis
Page 22 – Hong Yang – 125742/0
Figure 7 . Results of Scenario 2 for the female population
Source : Reviewer Analysis
Page 23 – Hong Yang – 125742/0
Figure 8 . Results of Scenario 3 for combined male and female populations
Source : Reviewer Analysis
Page 24 – Hong Yang – 125742/0
Figure 9 . Results of Scenario 3 for the male population
Source : Reviewer Analysis
Page 25 – Hong Yang – 125742/0
Figure 10. Results of Scenario 3 for the female population
Source : Reviewer Analysis
Page 26 – Hong Yang – 125742/0
Supplementary Materials
Text S1. Narratives of Myocarditis Death cases and Terms Used in VAERS Se arch
Table S1. Additional model scenarios with different combinations of model input values that are shown
on Table 1 in the main text.
Figure S1: Results of Scenario 4 for combined male and female populations.
Figure S2: Results of Scenario 4 for the male population.
Figure S3: Results of Scenario 4 for the female population.
Figure S4: Results of Scenario 5 for combined male and female populations.
Figure S5: Results of Scenario 5 for the male population.
Figure S6: Results of Scenario 5 for the female population.
Figure S7: Results of Scenario 6 for combined male and female populations.
Figure S8: Results of Scenario 6 for the male population.
Figure S9: Results of Scenario 6 for the female population.
Figure S10: Results of Scenario 6 for combined male and female populations.
Figure S11: Results of Scenario 6 for the male population.
Figure S12: Results of Scenario 6 for the female population.
Page 27 – Hong Yang – 125742/0
Text S1. Narratives of Myocarditis Death cases and Terms Used in VAERS Search
Narratives of Myocarditis Death C ases
Pfizer BioNTech Vaccine
VAERS ID 1406840: 13 years old male with attention deficit hyperactivity disorder and developmental
coordination disorder experienced flu -like symptoms for days and then was found deceased; onset of
symptoms 1 -day post -vaccination. The preliminary autopsy report revealed cardiomegaly with
biventricular dilatation, bilateral serous pulmonary effusions and serous pericardial effusion, marked pulmonary edema and congestion, and moderate degree of diffuse cerebral edema. COVID -19 and
influenza A/B tests were negative. Additional testing on autopsy found this patient died of sepsis due
Clostridium septicum.
VAERS ID 1486852: 21 years old female who experienced fever, confusion, seizure, cardiac arrest days
post vaccination. Autopsy revealed histology with extensive lymphocytic/plasmocytic myocarditis with
rare eosinophils no granuloma. Additional review by pathologists at CDC found t he patient had severe
myocarditis with intravascular leukocytosis suggestive of sepsis.
Reviewer Comment: Both of these death cases had alternate etiologies likely related to non-COVID -19
infections and were not attributed to vaccine.
VAERS Sear ch Terms
Atypical mycobacterium pericarditis, Autoimmune myocarditis, Autoimmune pericarditis, Bacterial
pericarditis, Coxsackie myocarditis, Coxsackie pericarditis, Cytomegalovirus myocarditis,
Cytomegalovirus pericarditis, Enterovirus myocarditis, Eosinophilic myocarditis, Hypersensitivity myocarditis, Immune -mediated myocarditis, Myocarditis, Myocarditis bacterial, Myocarditis helminthic,
Myocarditis infectious, Myocarditis meningococcal, Myocarditis mycotic, Myocarditis post infection,
Myocarditis sept ic, Pericarditis, Pericarditis adhesive, Pericarditis constrictive, Pericarditis helminthic,
Pericarditis infective, Pericarditis mycoplasmal, Pleuropericarditis, Purulent pericarditis, Viral
myocarditis, Viral pericarditis.
(b)
(6)
(b)
(6)
Page 28 – Hong Yang – 125742/0
Table S1. Additional model scenarios with different combinations of model input values that are shown on
Table 1 in the main text.
Scenario Protection
period Efficacy
against
cases Efficacy against
hospitalization COVID -19 case
incidence COVID -19
hospitalization
incidence Vaccine
attributable
myocarditis
death rate
4 Low High High High High Low
5 High High High Low Low Low
6 Low Low Low Low Low Low
7 Low Low Low High High High
Note: Additional scenarios are used to examine the impact of specific changes to model inputs.
Page 29 – Hong Yang – 125742/0
Figure S1: Results of Scenario 4 for combined male and female populations
Page 30 – Hong Yang – 125742/0
Figure S2: Results of Scenario 4 for the male population
Page 31 – Hong Yang – 125742/0
Figure S3: Results of Scenario 4 for the female population
Page 32 – Hong Yang – 125742/0
Figure S4: Results of Scenario 5 for combined male and female populations
Page 33 – Hong Yang – 125742/0
Figure S5: Results of Scenario 5 for the male population
Page 34 – Hong Yang – 125742/0
Figure S6: Results of Scenario 5 for the female population
Page 35 – Hong Yang – 125742/0
Figure S7: Results of Scenario 6 for combined male and female populations
Page 36 – Hong Yang – 125742/0
Figure S8: Results of Scenario 6 for the male population
Page 37 – Hong Yang – 125742/0
Figure S9: Results of Scenario 6 for the female population
Page 38 – Hong Yang – 125742/0
Figure S10: Results of Scenario 7 for combined male and female populations
Page 39 – Hong Yang – 125742/0
Figure S11: Results of Scenario 7 for the male population
Page 40 – Hong Yang – 125742/0
Figure S12: Results of Scenario 7 for the female population