019736 S470 M5 c4591012 protocol track

Pfizer Documents (PHMPT/FDA)

Pfizer Bla Submission

Pfizer 16 Plus Documents

237

Document text

BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 1 of 237  
NON -INTERVENTIONAL (NI) STUDY CONCEPT PROTOCOL  
Title Post-Emergency Use Authorization  Active 
Safety Surveillance Study among Individuals 
in the Veteran’s Affairs Health System 
Receiving Pfizer -BioNTech  Coronavirus 
Disease 2019 (COVID -19) Vaccine  
Protocol number  C4591012  
Protocol version identifier  Final Version 12.0 
Date of last version of protocol  27 January 2021 
EU Post Authori zation Study (PAS) 
register number  To be registered before the start of data 
collection EUPAS39779  
Active substance  COVID-19 mRNA Vaccine is single -stranded, 
5’-capped messenger RNA (mRNA) produced 
using a cell -free in vitro transcription from the 
corresponding DNA templates, encoding the 
viral spike (S) protein of SARS -CoV-2. 
 
Medicinal product  Pfizer-BioNTech COVID -19 Vaccine  
Research question and objectives  Research question: what are the incidence rates 
of safety events of interest (based on adverse  
events of special interest [AESI]) among 
individuals vaccinated with the Pfizer-
BioNTech  COVID-19 vaccine within the US 
Veterans Health Administration (VHA) system  
overall and in sub -cohorts of interest, as 
compared to expected rates of those events?  
Primary study objectives:  
• To assess whether individuals in the 
VHA system experience increased risk 
of safety e vents of interest following 
receipt of the Pfizer -BioNTech 
COVID-19 vaccine;  
• To assess whether sub -cohorts of 
interest (i.e., immunocompromised, 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077536
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 2 of 237 elderly, individuals with specific 
comorbidities, individuals receiving 
only one dose of the Pfizer -BioNTech 
COVID-19 vaccine, and individuals 
with prior SARS -CoV-2 infection ) in 
the VHA system experience increased 
risk of safety events of interest 
following receipt of the Pfizer -
BioNTech COVID -19 vaccine.  
Secondary study objective : 
• To characterize utilization patt erns of 
the Pfizer-BioNTech COVID -19 
vaccine among individuals within the 
VHA, including estimating the 
proportion of individuals receiving 
vaccine, 2 -dose vaccine completion 
rate, and distribution of time gaps 
between the first and second dose, 
demographi cs and health histories of 
recipients, overall and among the sub-
cohorts of interest.  
Authors Yinong Young -Xu, ScD, MA, MS  
Director, Clinical Epidemiology Program  
Veterans Affairs Medical Center  
White River Junction, VT  
 
Cynthia de Luise, PhD, MPH  
Senior Epidemiologist / Safety Surveillance 
Research Scientist ; Risk Management and 
Safety Surveillance Research  
Pfizer, Inc.  
New York, NY  
 
Mei Sheng Duh, ScD, MPH  
Managing Principal and Chief Epidemiologist  
Analysis Group, Inc.  
Boston, MA  
 
This document contains confidential information belonging to Pfizer. Except as otherwise agreed to in writing, 
by accepting or reviewing this document, you agree to hold this information in confidence and not copy or 
disclose it to others (except where required  by applicable law) or use it for unauthorized purposes. In the event 
of any actual or suspected breach of this obligation, Pfizer must be promptly notified.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077537
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
PFIZER CONFIDENTIAL  
Page 3 of 237 1.TABLE OF CONTENTS
1. TABLE OF CONTENTS  ................................ ................................ ................................ 3
2. LIST OF ABBREVIATIONS  ................................ ................................ .......................... 5
3. RESPONSIBLE PARTIES ................................ ................................ .............................. 8
4. ABSTRACT  ................................ ................................ ................................ ................... 9
5. AMENDMENTS AND UPDATES  ................................ ................................ ............... 26
6. MILESTONES  ................................ ................................ ................................ .............32
7. RATIONALE AND BACKGROUND ................................ ................................ ...........33
8. RESEARCH QUESTION AND OBJECTIVES  ................................ ............................. 34
9. RESEARCH METHODS  ................................ ................................ .............................. 35
9.1. Study Des ign ................................ ................................ ................................ ....35 
9.1.1. Self -Controlled Risk Interval (SCRI) Design with Post -Vaccination 
Control Interval  ................................ ................................ .......................... 35 
9.1.2. Active Comparator Design ................................ ................................ .....39 
9.1.3. Additional Study Designs in the Signal Evaluation Phase  ....................... 40 
9.1.4. Study Period ................................ ................................ .......................... 40 
9.2. Setting  ................................ ................................ ................................ ..............40 
9.2.1. Inclusion Criteria ................................ ................................ ................... 40 
9.2.2. Exclusion criteria  ................................ ................................ .................. 40 
9.2.3. Subgroups  ................................ ................................ ............................. 40 
9.3. Variables  ................................ ................................ ................................ ..........42 
9.3.1. Exposure of Interest ................................ ................................ ............... 42 
9.3.1.1. Pfizer -BioNTech COVID -19 Vaccine Groups of Interest ........43 
9.3.2. Baseline Characteristics  ................................ ................................ .........43 
9.3.3. Outcomes  ................................ ................................ .............................. 45 
9.4. Data Source  ................................ ................................ ................................ ......67 
9.5. Study Size ................................ ................................ ................................ .........68 
9.5.1. Power  ................................ ................................ ................................ ...68 
9.6. Data Management  ................................ ................................ ............................. 71 
9.6.1. Case report forms (CRFs)/Electronic data record  ................................ ....71 
9.6.2. Record retention  ................................ ................................ .................... 71 
9.7. Data Analysis ................................ ................................ ................................ ....72 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077538
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 4 of 237 9.7.1. Baseline Characteristics  ................................ ................................ .........72 
9.7.2. Vaccine Utilization Patterns ................................ ................................ ...73 
9.7.3. Safety Signal Analyses  ................................ ................................ ..........73 
9.7.3.1. Signal Detection ................................ ................................ ....75 
9.7.3.2. Signal Evaluation  ................................ ................................ ..79 
9.7.3.3. Signal Verification  ................................ ................................ 83 
9.7.4. Seasonality -Adjusted Cases -Centered Method  ................................ ........84 
9.7.5. End -of-Season and End -of-Surveillance Analyses  ................................ ..84 
9.7.6. Subgroup Analysis  ................................ ................................ ................ 86 
9.7.7. Incidence Rates and Time to Safety Event of Interest Analysis ................ 86 
9.7.8. Prioritized Safety Analysis of Myocarditis/Pericarditis  ........................... 86 
9.8. Quality Control  ................................ ................................ ................................ .88 
9.9. Strengths and Limitations of the Research Methods ................................ ............88 
9.10. Other Aspects ................................ ................................ ................................ ..90 
10. PROTECTION OF HUMAN SUBJECTS ................................ ................................ ....90 
10.1. Patient Information ................................ ................................ .......................... 90 
10.2. Patient Consent  ................................ ................................ ............................... 91 
10.3. Institutional Review board (IRB)/Independent Ethics Committee (IEC)  ............91 
10.4. Ethical Conduct of the Study  ................................ ................................ ...........91 
11. MANAGEMENT AND REPORTING OF ADVERSE EVENTS/ADVERSE  
REACTIONS  ................................ ................................ ................................ ............... 91 
12. PLANS FOR DISSEMINATING AND COMMUNICATING STUDY RESULTS  .......93 
13. REFERENCES  ................................ ................................ ................................ ...........94 
14. LIST OF TABLES  ................................ ................................ ................................ ....100 
15. LIST OF FIGURES ................................ ................................ ................................ ...100 
16. ANNEX 1. LIST OF STAND ALONE DOCUMENTS  ................................ ..............101 
17. ANNEX 2. ENCEPP CHECKLIST FOR STUDY PROTOCOLS  ............................... 101 
18. ANNEX 3. ADDITIONAL INFORMATION  ................................ ............................ 101 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077539
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 5 of 237 2. LIST OF ABBREVIATIONS  
Abbreviation  Definition  
ACIP Advisory Committee on Immunization Practices  
ADEMACOS Acute disseminated encephalomyelitis Associate Chief of Staff  
AE Adverse event  
AEM Adverse event monitoring  
AESI Adverse event of special interest  
AIDS Acquired immunodeficiency syndrome  
AMI Acute myocardial infarction  
BMI Body mass index  
CAD Coronary artery disease  
CBER Center for Biologics Evaluation and Research  
CI Confidence Interval  
CCI Charlson comorbidity index  
CDC Centers for Disease Control and Prevention  
CDW Corporate Data Warehouse  
CEP Clinical Epidemiology Program  
CIDP Chronic inflammatory demyelinating polyneuropathy  
CMA Conditional Marketing Authorization  
CMS Centers for Medicare & Medicaid Services  
COPD Chronic obstructive pulmonary disease  
COVID-19 Coronavirus Disease 2019  
CPT Current Procedural Terminology  
CRADA Cooperative Research and Data Agreement  
CRFs Case report forms  
DIC Disseminated intravascular coagulation  
DVT Deep vein thrombosis  
TDapTdap Diphtheria, tetanus and (acellular) pertussis 
Td Diphtheria and tetanus  
ED Emergency department  
EMA European Medicines Agency  
EMR Electronic medical records  
EU European Union  
EUA Emergency Use Authorization  
EU PAS European Union Post -Authorization Safety  
FDA Food and Drug Administration  
GBS Guillain-Barré syndrome  
GEP Good Epidemiological Practice  
GPP Good Pharmacoepidemiology Practices  
H0 Null hypothesis  
Ha Alternative hypothesis  
HBV Hepatitis B virus  
 HCPCS Healthcare Common Procedure Coding System  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077540
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 6 of 237 Abbreviation  Definition  
HCV Hepatitis C virus  
HIV Human immunodeficiency virus  
HPV Human papillomavirus  
ICD-10-CM International Classification of Diseases, Tenth Revision, Clinical 
Modification  
ICD-10-PCS International Classification of Diseases, Tenth Revision, Procedure 
Coding System  
IEA International Epidemiological Association  
IEC Independent Ethics Committee  
IPTW Inverse probability of treatment weighting  
IQR Interquartile range  
IRB Institutional Review Board  
KD Kawasaki disease  
LLR Log-likelihood ratio  
MaxSPRT  Maximized sequential probability ratio test  
MenACWY  Meningococcal conjugate  
 MenB Serogroup B meningococcal  
MIS-A Multisystem inflammatory syndrome  in adults 
mRNA Messenger RiboNucleic Acid  
MS Multiple sclerosis  
NDC National Drug Codes  
NIS Non-interventional  study 
NNERC VAMC  Northern New England Research Consortium VA Medical Centers  
NSAID Non-steroidal anti -inflammatory drug  
ON Optic neuritis  
PASS Post-Authorization Safety Study  
PE Pulmonary embolism  
PRISM Post-Licensure Rapid Immunization Safety Monitoring  
PS Propensity score  
R&D Research and Development  
RCA Rapid cycle analysis  
RR Relative risk  
SAP Statistical analysis plan  
SARS-CoV-2 Severe acute respiratory syndrome coronavirus 2  
SAS SAS Institute  
SCCS Self-controlled case series  
SCRI Self-controlled risk interval  
SD Standard deviation  
SJS Stevens-Johnson syndrome  
SPEAC Safety Platform for Emergency vACcines  
SRSS Subcommittee on Research Safety and Security  
TEN Toxic epidermal necrolysis  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077541
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 7 of 237 Abbreviation  Definition  
TM Transverse myelitis  
TTS Thrombosis with thrombocytopenia syndrome  
UK United Kingdom  
US United States  
VA Department of Veterans Affairs 
VAIRRS VA Innovation and Research Review System  
VAERS Vaccine Adverse Event Reporting System  
VHA Veterans Health Administration  
VINCI VA Informatics and Computing Infrastructure  
VINNE Veteran’s IRB of Northern New England  
VISN Veterans Integrated Service Networks  
VSD Vaccine Safety Datalink  
VTE Venous thromboembolism  
WHO World Health Organization  
WOC Without compensation  
YRR Your Reporting Responsibilities  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077542
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 8 of 237 3. RESPONSIBLE PARTIES  
Name, degree(s)  Job Title  Affiliation  Address  
Principal Investigator:  
Yinong Young -Xu, 
ScD, MA, MS  Director, Clinical Epidemiology 
Program Veterans Affairs 
(VA) Medical 
Center 163 Veterans Drive, 
White River Junction, 
VT 05009  
Cynthia de Luise, 
PhD, MPH  Senior Epidemiologist /Safety 
Surveillance Research Scientist ; 
Risk Management and Safety 
Surveillance Research  Pfizer, Inc.  235 East 42nd Street, 
New York, NY 10017  
Mei Sheng Duh,  
ScD, MPH  Managing Principal and Chief 
Epidemiologist  
 
 
Visiting Scientist, Department of 
Biostatistics  Analysis Group, 
Inc. 
 
 
Harvard T. H. 
Chan School of 
Public Health  111 Huntington Ave  
14th Floor 
Boston, MA 02199  
 
677 Huntington Ave  
Boston, MA 02115  
Maral DerSarkissian, 
PhD  
 Vice President and Senior 
Epidemiologist  
 
 
 
Adjunct Assistant Professor  Analysis Group, 
Inc. 
 
 
 
Fielding School of 
Public Health, 
University of 
California, Los 
Angeles 333 South Hope Street  
27th Floor 
Los Angeles, CA  
90071 
 
650 Charles E Young 
Drive South 
Los Angeles, CA 
90095 
Rachel Bhak, MS  
 Manager and Senior Biostatistician  
 Analysis Group, 
Inc. 
 111 Huntington Ave  
14th Floor 
Boston, MA 02199  
 
 
  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077543
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 9 of 237 4. ABSTRACT  
Title: Post-Emergency Use Authorization  Active Safety Surveillance Study among 
Individuals in the Veteran’s Affairs Health System Receiving Pfizer -BioNTech Coronavirus 
Disease 2019 (COVID -19) Vaccine  
Protocol Version: 12.0; Date of Protocol : 27 January31 Aug 2021 
Authors: Yinong Young Xu, ScD, MA, MS , Veterans Affairs Medical Center; Cynt hia de 
Luise, PhD, MPH, Pfizer, Inc.; Mei Sheng Duh, ScD, MPH, Analysis Group, Inc.  
Rationale and b ackground :  
In March 2020, the World Health Organization (WHO) declared a global pandemic for the 
coronavirus disease 2019 (COVID -19) due to the severe acute  respiratory syndrome 
coronavirus 2 (SARS -CoV-2), which was first identified by public health officials in China 
in December 2019.1 The COVID -19 pandemic presents an unprecedented public health 
crisis. As of January 7 , 2021, over 21.4 million COVID -19 cases and 364,000 deaths have 
been reported in the United States (US) alone.2   
Pfizer and BioNTech have partnered to develop a novel messenger RiboNucleic Acid 
(mRNA) vaccine against SARS -CoV-2 for the prevention of COVID -19 (Candidate 
BNT162b2). Pfizer is conducting a Phase 1/2/3, randomized, placebo -controlled, observer -
blind, dose -finding, vaccine candidate -selection, and efficacy study among healthy 
individuals (NC T04368728). The Food and Drug Administration (FDA) reviewed the 
available safety data from 37,586 participants 16 years of age and older and did not identify 
any specific safety concerns. In addition, the analysis of available efficacy data from 36,523 
participants 12 years of age and older without evidence of prior SARS -CoV-2 infection at 
least 7 days after receiving the second dose demonstrated 95% efficacy of the vaccine in the 
prevention of COVID -19 (as confirmed by 8 vs. 162 COVID -19 cases in the vacci ne and 
placebo groups, respectively).3,4 Based on these safety and efficacy data, as well as a review 
of manufacturing information regarding product quality and consistency, the FDA 
determined that the known and potential benefits of the vaccine outweighed the known and 
potential risks for the prevention of COVID -19 in individuals 16 years of age and older .4 
Therefore on December 11, 2020, the Pfizer -BioNTech COVID-19 vaccine was granted an 
Emergency Use Authorization (EUA) by the FDA to prevent COVID -19 in individuals 16 
years of age and older.5   
With respect to geographic regions other than the US, on December 2, 2020, the United 
Kingdom (UK) was the first country in the world to grant temporary authorization for 
emergency use of the Pfizer -BioNTech COVID -19 vaccine.6 On December 21, 2020, the 
European Medicines Agency (EMA) granted the Pfizer -BioNTech COVID -19 vaccine a 
conditional marketing authorization (CMA) for use among individuals 16 years of age and 
older throughout all of the European Union’s (EU) 27 member st ates.7  
As required by the EUA, post-authorization observational studies using real -world data are 
needed in order to assess the association between Pfizer -BioNTech COVID -19 vaccine and 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077544
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 10 of 237 pre-determined safety events of in terest (including deaths , hospitalizations, and severe 
COVID-19) among individuals administered the vaccine in both the population at large and 
in populations of interest ( e.g., immunocompromised individuals, elderly, and those with 
specific comorbidities ).4 Pfizer in collaboration with  the US Veterans Health Administration 
(VHA) and Analysis Group  herein propose post -EUA active safety surveillance of  safety 
events of interest based primarily  on the Priority List of Adverse Events of Special Interest 
from the Brighton Collaboration’s Safety Platform for Emergency vACcines (SPEAC) 
Project, and from the FDA and  thepreliminary list of safety events of interest  presented at the 
September 22 , 2020, meeting of Centers for Disease Control and P revention’s ( CDCCDC’s) 
Advisory Committee on Immunization Practices (ACIP)  on the enhanced safety monitoring 
recommendation .of COVID -19 vaccines .8,9 This safety surveillance study will  identify and 
evaluate rapid, near real-time potential safety signals associated with the Pfizer -BioNTech 
COVID-19 vaccine  in the large -scale VHA electronic medical record (EMR) database. The 
observed safety event of int erest rates will be compared to expected rates derived from self -
controls and active comparators receiving seasonal influenza vaccination. Part of the 
methodologies used in this study are constructed based on approaches previously used by the 
Post-Licensure Rapid Immunization Safety Monitoring (PRISM) program  for the H1N1 
vaccine.810 This non-interventional study is designated as a Post -Authorization Safety Study 
(PASS) commitment to the US FDA  and is a Category 3 commitment in th e EU Risk 
Management Plan . 
Research question and objectives :  
Research question: what are the incidence rates of safety events of interest (based on adverse 
events of special interest [AESI])  among individuals vaccinated with the Pfizer-BioNTech 
COVID-19 vaccine within the US VHA  system overall and in sub -cohorts of interest, as 
compared to expected rates of those events?  
Primary study objectives:  
• To assess whether individuals in the VHA system experience increased risk of safety 
events of interest followin g receipt of the Pfizer -BioNTech COVID -19 vaccine;  
• To assess whether sub -cohorts of interest (i .e., immunocompromised, elderly, 
individuals with specific comorbidities, individuals receiving only one dose of the 
Pfizer-BioNTech COVID -19 vaccine, and indivi duals with prior SARS -CoV-2 
infection) in the VHA system experience increased risk of safety events of interest 
following receipt of the Pfizer -BioNTech COVID -19 vaccine.  
Secondary study objective:  
• To characterize utilization patterns of  the Pfizer-BioNTech COVID -19 vaccine 
among individuals  within the VHA , including estimating the proportion of 
individuals  receiving vaccine, 2-dose vaccine completion  rate, and distribution of 
time gaps between the first and second dose , demographics and healt h histories of 
recipients, overall and among the sub-cohorts of interest .  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077545
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 11 of 237 Study design : This post-EUA active safety surveillance program will employ a rapid-cycle, 
longitudinal, observational cohort study design  to provide early real -world safety 
information.  
• The self-controlled risk interval (SCRI) design will be used to sequentially monitor 
occurrence of safety events of interest while controlling for time -invariant 
confounders.  The SCRI design uses data from cases ( i.e., individuals who experience 
safety events of interest following vaccination) to compare the risk interval following 
vaccination t o pre or post-vaccination non -risk intervals (“prevaccination control 
interval” and “ post-vaccination control interval”) in the same individual.   
• An active comparator design will be used to sequentially m onitor occurrence of safety 
events of interest with Pfizer -BioNTech COVID -19 vaccinations as compared to 
recipients of influenza vaccine in the VHA during 2014/2015 through 2018/2019 flu 
seasons. Data in peri -COVID time periods from January 2020 to present  are excluded 
because of pandemic -associated under -utilization of health resources and under -
reporting of medical events.   
There will be additional study designs conducted during the signal evaluation phase if a 
signal is detected from the above analyses.  These include self -controlled case series (SCCS) 
and comparison to  unvaccinated contemporary control s. Additionally, s ignal evaluation 
analyses may also be conducted based on signa ls detected in external sources  or based on 
regulatory request.    
Population : The exposed population will be kept as broad as possible in order to capture 
safety events of interest that occur among all individuals  receiving the Pfizer -BioNTech 
COVID-19 vaccine  in the period from December 11, 2020 t o present. Individuals  will be 
included if they have a record of at least one dose of Pfizer -BioNTech COVID -19 vaccine . 
Individuals who receive at least one dose of COVID -19 vaccine from a manufacturer other 
than Pfizer -BioNTech will be identified and reportedsummarized, but they will be excluded 
from further analysis . All individuals will be required to be enrolled in and not disenrolled 
from VHA benefits during the 1 year prior to vaccination date ( i.e., baseline period).  
Depending on the attrition rate, the leng th of the baseline period may be modified to 6 
months. 
The influenza vaccine comparator cohort will be identified based on a record of at least one 
dose of seasonal influenza vaccine during prior flu seasons, from 2014/2015 through 
2018/2019 . 
Variables : 
• Exposures: Administration of Pfizer -BioNTech COVID -19 vaccine post -EUA 
approval will be identified based on the following : (see Appendix  Table 3 for 
additional details) :  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077546
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 12 of 237 o Current Procedural Terminology (CPT) code 91300 (Severe acute respiratory 
syndrome coronavirus 2 (SARS CoV2) (coronavirus disease [COVID 19]) 
vaccine, mRNALNP, spike protein, preservative free, 30 mcg/0.3mL dosage, 
diluent reconstituted, for intramuscular use) and associated vaccine 
administration Healthcare Common Procedure Coding System ( HCPCS) 
codes corresponding to the first dose: 0001A (ADM SARS CoV2 30 
mcg/0.3mL 1st), and the sec ond dose: 0002A (ADM SARS CoV2 30 
mcg/0.3mL 2nd);9,10; OR 
o 10 and 11 -digit National Drug C odes (NDCs ) 5926710001 (corresponds to 
first dose), 59267 100001 (corresponds to second dose);9); OR 
o Immunization records that contain data on vaccine code descriptor, vaccine 
manufacturer ( i.e., Pfizer), lot number, injection site, and date(s) of 
immunization;911 
Relevant codes will be continuously reviewed and amended if new codes are added.  
• Administration of the seasonal influenza vaccine during 2014/2015 through 
2018/2019 flu seasons will be identified based on the following : (see 
Appendix  Table 3 for additional details) : 
o CPT codes   
o 90654 (Influenza virus and associated  vaccine, trivalent (IIV3), spli t virus, 
preservative free, for intradermal use);  administration HCPCS codes ; OR 
▪ 90656 (Influenza virus vaccine, trivalent (IIV3), split virus, 
preservative free, 0.5 mL dosage, for intramuscular use); OR  
▪ 90658 (Influenza virus vaccine, trivalent (IIV3), split virus, 0.5 mL 
dosage, for intramuscular use); OR  
o 10 and 11 -digit NDCs; OR 
o Immunization records that contain data on vaccine code descriptor, vaccine 
manufacturer, lot number, injection site, and date(s) of immunization.  
• Outcomes : Safety events of interest for active surveillance (see  Table 1 and Appendix 
Table 2Appendix  Table 2) are based on the Priority List of Adverse Events of Special 
Interest from the Brighton Collaboration’s Safety Platform for Emergency vACcines 
(SPEAC) Project, the FDA and the Centers for Disease Control and Prevention’s 
(CDC) Advisory Committee on Immunization Practices ( CDC’s ACIP) enhanced 
safety monitoring recommendations .  
The list of safety events of interest may be revised over the course of the study, and i f 
unanticipated potential safety events  of interest are identified during the course of 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077547
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 13 of 237 surveillance, they will be added to the list and included in the analyses . The risk and 
control intervals for each safety event of interest are based on biological plausibility 
and precedents in the literature (see Table 1). Outpatient  (including , emergency 
department [(ED])), and/or inpatient settings will be used to identify safety events of 
interest depending on the type of event. The specific encounter setting to be 
considered for each safety event of interest is summarized in Table 1 and can be 
assigned to 1) the risk interval following vaccination Pfizer-BioNTech COVID -19 
vaccination, 2) the prevaccination self control interval, 3) the post-vaccination self -
control interval, or 43) risk interval for the active comparators of receiving seasonal 
influenza vaccine . Events outside the intervals will not be counted.     
Only the individual ’s first instance  of a safety event  of interest followi ng a specified 
clean window ( i.e., the occurrence -free baseline period  used to define incide nt 
outcomes during which individuals enter the study cohort only if the safety event of 
interest did not occur during this period) will be captured; this means that  if a safety 
event of interest is identified but diagnosis codes corresponding to the safety event of 
interest are also observed during the clean window, it will not be counted . The 
duration of the pre-specified clean window will differ by safety event of interest (see 
Appendix  Table 2) in order to rule out pre -existing events.  
 
• Key Covariates : Baseline demographic ( i.e., age, sex, race/ethnicity, stateservice 
region) and clinical characteristics ( i.e., smoking, body mass index [ BMI], history of 
anaphylaxis/allergic reactions, previous anaphylaxis to vaccine component, history of 
hospitalizations, frailty index, Charlson Comorbidity Index [CCI], select ed 
comorbidities , and concurrent immunizations )1112 will be assessed based on available 
data (i.e., during 1-year baseline ) prior to the date of vaccination with Pfizer -
BioNTech COVID -19 vaccine and date of seasonal influenza vaccination  for active 
comparators .  
• Subgroups : Immunocompromised individuals , elderly, in dividuals with specific 
comorbidities,  (i.e., individuals diagnosed with symptomatic human 
immunodeficiency virus  (HIV)/acquired immunodeficiency syndrome ( AIDS), 
hematologic malignancy, or other immune conditions; individuals diagnosed with 
solid malignancy, organ transplant, or rheumatologic/inflammatory conditions, all of 
whom were administered chemotherapy or immune modulators; individuals 
diagnosed with rheum atologic/inflammatory conditions and administered systemic 
corticosteroids; individuals who were administered chemotherapy, immune 
modulators, or systematic steroids for at least 14 days) ,13 elderly, individuals with 
specific comorbidities,12 those receiving only one dose of Pfizer -BioNTech COVID -
19 vaccine, those with prior SARS -CoV-2 infection , those with regular use of VHA 
medical care, and VA priority group 1 veterans  will be identified . Analyses will also 
be performed among individuals enrolled in the VHA with dual coverage who are 
also identified in linked Centers for Medicare & Medicaid Services (CMS) Medicare 
administrative claims data.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077548
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 14 of 237 Data source : The VHA is the largest integrated health ca re system in the US, providing both 
inpatient and outpatient clinical care to over 9 million Veterans enrolled at more than 170 
medical centers and 1,074 community -based outpatient clinics.1214 This study will use data 
from VHA’s Corporate Data Warehouse (CDW), which is an integrated electronic medical 
record (EMR) system with a centralized data warehouse that is updated on a daily basis.  The 
CDW does not include information on any care received outside of a VHA facility . The VA 
offers eligible Veterans long -term care services ranging from nursing homes and assisted -
living centers to caregiver support in the Veterans’ own homes.13 15 In a subgroup analysis of 
individuals with both VHA and Medicare coverage,  CDW data will be supplemented and 
linked with Medicare administrative claims data at the patient level to ensure a more 
comprehensive evaluation of the care an indivi dual receives.  
Study size : The sample size achieved will depend on the number of recipients of Pfizer -
BioNTech COVID -19 vaccine within the VHA database, which will increase over time with 
subsequent analyses. As of January 21, 2021, 112,201 doses of Pfizer -BioNTech COVID -19 
vaccine have been administered within the VHA (based on CPT code 91300) to a total of 
107,458 patients.  
Data analysis : A stepwise approach, illustrated in the diagram , will be performed for signal 
detection, evaluation, and verification.     
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077549
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 15 of 237 
 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077550
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 16 of 237  
Notes:  
[1] List of safety events of interest and corresponding definitions may be refined as the study progresses based on additiona l 
available information  
[2] The risk and control intervals selected for the SCRI analysis for each safety event of i nterest are based on biological 
plausibility and precedents in the literature  Only the individual’s first instance during the specified clean window ( i e, the 
interval used to define incident outcomes) will be included   Note that only the first inpatien t or outpatient occurrence of a 
safety event of interest following the clean window will be used to identify incident events ( e g, if an inpatient safety event 
of interest occurs in the clean window, a repeat occurrence will not be counted in the risk interval)  However, event 
worsening will be counted as a safety event of interest  For example, if an outpatient safety event of inte rest occurs in the 
clean window and an inpatient occurrence for the same type of safety event of interest occurs in the risk interval, the 
inpatient occurrence will be counted as a safety event of interest   
1) Signal detection : The goal is to provide rapi d-cycle, near real -time safety surveillance. In 
the signal detection phase, the SCRI analysis will only include prevaccination control 
intervals as the post-vaccination control intervals  willfor certain safety events of interest that 
require a longer time to accumulate and will be used COVID-19 diagnosis (i.e., severe 
COVID-19, multisystem inflam matory syndrome  in the signal evaluation  phase.adults [MIS-
A]). To account for multiple testing and bi -weekly review of the data, the maximized 
sequential probability ratio test (MaxSPRT ) using a binomial probability model will be 
applied. For comparison with individuals  who received  seasonal influenza vaccination, the 
Poisson-based MaxSPRT will be applied . for all other safety events of interest .  
Sequential analyses for each safety event of interest will commence once at least 3 events 
occur. This approach is consistent with the  FDA’s COVID -19 Vaccine Safety Surveillance 
Project to avoid spurious signals from a few early events .116 Signals will be detected if the 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077551
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 17 of 237 critical values  are reached via  the SCRI or active comparator analysis.  Critical values wil l be 
determined for each safety event of interest based on historical incidence rate, expected upper 
limit of the number of events under the null hypothesis, and pre -specified significance level 
and power. Incidence rates will also be calculated and Kaplan Meier methods wi ll be used to 
analyze time to safety event of interest  
2) Signal evaluation: If signals are detected for  safety events of interest based on the analysis 
described above, further evaluation will be conducted to refine and confirm such detections.  
This will include comprehensive quality assurance ( for example, check for possible 
duplications of claims or medi cal records, checking for unusual clustering in claim or 
medical record accrual by service date for potential coding issues, check for geographical 
distribution of cases that may be related to lot numbers or diagnostic practice ) and 
multivariate adjustment using Poisson regression to account for baseline differences between 
Pfizer-BioNTech COVID -19 vaccinated and active comparator cohorts. SCRI analyses using 
the post-vaccination control intervals  and SCCS using post -vaccination contr ol time periods  
will be conducted as an additional inferential analysis once enough post -vaccination time has 
accumulated. To address potential period effects, a  comparison to contemporary 
unvaccinated controls will also be performed , with adjustment  using inverse probability of 
treatment weighting (IPTW). The assessment of temporal clustering will also be conducted. 
Incidence rates will also be calculated and Kaplan -Meier methods wi ll be used to analyze 
time to safety event of interest. Lastly, the assessment of temporal clustering will also be 
conducte d.  Signal evaluation a nalyses will be conducted every six months.  
3) Signal verification: diagnostic validation of the detected safety events  of interest via 
adjudication  of medical records by VHA clinicians for outcome verification will be 
conducted in a representative sample of cases. For rare events, potentially all cases may be 
adjudicated.  
End-of-season analyses (over the course of the 30 -month period) and an end -of-surveillance 
analysis ( i.e., at 30 months) will be conducted. Various subgroup analyses  will also be 
conducted, examining different age groups, immunocompromised individuals ,,13 individuals 
with specific comorbidities ,,12 those who only received one dose of the Pfizer -BioNTech 
COVID-19 vaccine, those with prior SARS -CoV-2 infection based on medical history or pre -
vaccination serology, those receiving care regularly at VA facilities , and lastly those with VA 
Priority group 1 status, which determines these individuals are of highest priority for VHA 
care and likely receive all of their care within the VHA system , and lastly, those with 
additional Medicare coverage whose Medicare data can be linked to the CDW . 
Notably, CDC recently investigated myocarditis/pericarditis following mRNA COVID -19 
vaccinations .17 To provide additional context to the investigation conducted by CDC, 
separate safety analyses will be prioritized and performed to assess the risk of 
myocarditis/pericarditis following Pfizer -BioNTech COV ID-19 vaccination . These analyses 
will be conducted to align with the rapid -cycle analysis performed by the Vaccine Safety 
Datalink (VSD).18 The number of myocarditis/pericarditis events in the risk interval will be 
identified , and incidence rates per million doses will be summarized . Subgroup analyses will 
also be performed, stratified by age ( e.g., 12-39 years, 40-49 years, 50 -64 years, 65+ years ), 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077552
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 18 of 237 gender, and race/ethnicity, respectively.  Incidence rate ratios will be summarized to compare 
the rate of myocarditis/pericarditis events between vaccinated individuals who se event occurs 
in a pre-specified risk interval versus vaccinated individuals who se event occurs in a 
comparison interval on the same calendar day. Myocarditis/p ericarditis events will also be 
adjudicated via chart review and validated using the Brighton Collaboration’s case 
definitions.19 Risk factor analysis may also be conducted among confirmed cases. Lastly, 
additional da ta surrounding risk factors, clinical course, and sequelae of identified 
myocarditis/pericarditis event up to 365 days following the event will be collected and 
summarized .  
Milestones :  
• Registration in the EU PAS register : To be registered before the start of data 
collection;  
• VHA Cooperative Research and Data Agreement ( CRADA) and  execution: 8 
January 2021 ; 
• Determination of Institutional Review Board ( IRB) approvals  (estimated) :exemption: 
10 February 2021 ; 
• Determination of Research Safety and Security exemption: 17 February 2021 ; 
• Approval by Designated Member Review: 26 February 2021 ; 
• Registration in the EU PAS register: 5  MarchApril 2021; 
• Start of data collection  (estimated planned date for starting data extraction for 
analysis):: 11 May 2021; 
• Interim reports: 30 June 2021; 31 December 2021; 30 June 2022, 31 December 2022;  
• End of data collection  (estimated planned date for final data  cut): 10: 30 June 2023; 
• Final study report: 31 December  2023 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077553
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 19 of 237 SUMMARY  
 
Objective  Primary 1  Primary 2  Secondary  
Aim To assess whether individuals in the 
Veterans Health Administration 
(VHA) system experience increased 
risk of safety events of interest 
following receipt of the Pfizer -
BioNTech COVID -19 vaccine.  To assess whether sub -cohorts of 
interest (i.e. immunocompromised, 
elderly, with specific comorbidities, 
individuals receiving only one dose 
of the Pfizer -BioNTech COVID -19 
vaccine, and individuals with prior 
SARS-CoV-2 infection ) in the VHA 
system experience increased risk of 
safety events of interest following 
receipt of the Pfizer -BioNTech 
COVID-19 vaccine.   To characterize utilization patterns of  
the Pfizer-BioNTech COVID -19 
vaccine among individuals within the 
VHA including estimating the 
proportion of individuals receiving 
vaccine, 2 -dose vaccine completion 
rate, and distribution of time gaps 
between the first and second dose, 
demographics and health histories of 
recipients, overall and among the 
sub-cohorts of interest.  
Study design  This post-EUA active safety surveillance program will employ a rapid -cycle, longitudinal, observational cohort study 
design to provide early real -world safety information.  
• The self-controlled risk interval (SCRI) design to sequentially monitor occurrence of safety events of interest 
while controlling for time -invariant confounders. This design allows inclusion of either a pre vaccination 
control interval or a post-vaccination control interval , depending on the safety event of interest (e.g., post -
vaccination control intervals are used for outcomes where there is concer n for bias due to indication or 
contraindication); .  
• An active comparator design will be used to sequentially monitor occurrence of safety events of interest with 
Pfizer-BioNTech  COVID-19 vaccinations as compared to recipients of influenza vaccine in the VHA during 
2014/2015 through 2018/2019 flu seasons. Data in peri -COVID time periods from January 2020 to present are 
excluded because of pandemic -associated under -utilization of h ealth resources and under -reporting of medical 
events. 
There will be additional study designs conducted during the signal evaluation phase if a signal is detected from the 
above analyses. These include self -controlled case series (SCCS) and comparison of v accinated to unvaccinated 
contemporary control s. Additionally, signal evaluation analyses may also be conducted based on signals detected in 
external sources or based on regulatory request  (e.g., myocarditis/pericarditis) .    
Study population  The study will be kept as broad as possible in order to capture safety events of interest that occur among vaccinated 
individuals.  
 
Inclusion criteria:  
• Record of at least one dose of Pfizer -BioNTech COVID -19 vaccine  in the period of December 11, 2020 to 
present, or 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077554
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 20 of 237  
Objective  Primary 1  Primary 2  Secondary  
• Record of at least one dose of seasonal influenza vaccine during prior flu seasons, from 2014/2015 through 
2018/2019 ( applies to active comparator s only); and 
• At least 1 year of enrollment in and no disenrollment from VHA benefits ( i.e., the base line period) prior to 
Pfizer-BioNTech COVID -19 or seasonal influenza vaccination date.  
 
Exclusion criteria:  
• Individuals who receive at least one dose of Pfizer -BioNTech COVID -19 vaccine in addition to a COVID -19 
vaccine from a manufacturer other than Pfiz er-BioNTech will be identified and reportedsummarized , but they 
will be excluded from further analysis. . 
Study Period  The study will be conducted for a period of 30  months, starting on December 11, 2020 onward, with data collection 
concluding on June 10, 2023. 
Exposure  Administration of Pfizer -BioNTech COVID -19 vaccine post -EUA approval will be identified based on records of the 
following : (see Appendix  Table 3 for additional details) : 
• Current Procedural Terminology (CPT) code 91300 (Severe acute respiratory syndrome coronavirus 2 
(SARSCoV2) (coronavirus disease [COVID 19]) vaccine, mRNALNP, spike protein, preservative free, 30 
mcg/0.3mL dosage, diluent reconstituted, for intramuscular use) and associated vaccine administration 
HCPCS codes  corresponding to the first dose: 0001A (A DM SARS -CoV-2 30 mcg/0.3mL 1t), and the second 
dose: 0002A (ADM SARS -CoV-2 30 mcg/0.3mL 2d);; OR 
• 10 and 11 -digit National Drug Codes (NDCs ) 59267-1000-1 (corresponds to first dose), 59267 -1000-01 
(corresponds to second dose ); OR 
• Immunization records that contain data on vaccine code descriptor, vaccine manufactur er (i.e., Pfizer), lot 
number, injection site, and date(s) of immunization;  
Administration of the seasonal influenza vaccine during 2014/2015 through 2018/2019 flu seasons will be identified 
based on records of the following : (see Appendix  Table 3 for additional details) : 
• CPT codes   
• 90654 (Influenza virus and associated  vaccine, trivalent (IIV3), split virus, preservative -free, for intradermal 
use); administration HCPCS codes ; OR 
o 90656 (Influenza virus vaccine, trivalent (IIV3), split virus, preservative free, 0.5 mL dosage, for 
intramuscular use); OR  
o 90658 (Influenza virus vaccine, trivalent (IIV3), split virus, 0 5 mL dosage, for intramuscular use); 
OR 
• 10 and 11 -digit NDCs; OR 
• Immunization records that contain data on vaccine code descriptor, vaccine manufacturer, lot number, 
injection site, and date(s) of immunization.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077555
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 21 of 237  
Objective  Primary 1  Primary 2  Secondary  
Safety Events of Interest  Safety events of interest for active surveillance were identified based on the Priority List of Adverse Events of Special 
Interest from the Brighton Collaboration’s Safety Platform for Emergency vACcines (SPEAC) Project, the FDA and 
the Centers for Disease Control and Prevention’s (CDC) Advisory Committee  on Immunization Practices (ACIP) 
enhanced safety monitoring recommendations. The list of safety events may be revised over the course of the study, 
and if unanticipated potential safety events of interest are identified during the course of surveillance, they will be 
added to the list and included in the analyses. The risk and control intervals for each safety event of interest are based 
on biological plausibility and precedents in the literature. Outpatient  (including , emergency department ), and/or 
inpatient settings  will be used to identify safety events of interest depending on the type of event. Safety events of 
interest can be assigned to 1) the risk interval following vaccination Pfizer -BioNTech COVID -19 vaccination, 2) the 
pre-vaccination self -control interval, 3) the post-vaccination self -control interval,  or 43) risk interval for the active 
comparators of receiving seasonal influenza vaccine. Events outside the intervals will not be counted. Only the 
individual’s first instance of a safety event of interest following a specified clean window ( i.e., the occu rrence-free 
baseline period used to define incident outcomes during which individuals enter the study cohort only if the safety 
event of interest did not occur during this period) will be included;  this means that if a safety event is identified but 
diagnosis codes corresponding to the safety event are also observed during the clean window, it will not be counted . 
The duration of the pre -specified clean window will differ by type of safety event of interest in order to rule out pre -
existing events.  
Neurologic : 
• Generalized convulsions/seizures;  
• Guillain-Barré syndrome (GBS);  
• Aseptic meningitis ; 
• Encephalitis/encephalomyelitis ; 
• Other acute demyelinating diseases ; 
• Transverse myelitis (TM);  
• Multiple sclerosi s (MS); 
• Optic neuritis (ON);  
• Bell’s palsy 
• Cerebrovascular non -hemorrhagic stroke  
• Convulsions/seizures in individuals with controlled epilepsy 
• Encephalitis/encephalomyelitis  
• Guillain -Barré Syndrome (GBS)  
• Generalized convulsion/seizures  
• Multiple sclerosis  (MS)  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077556
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 22 of 237  
Objective  Primary 1  Primary 2  Secondary  
• Optic neuritis (ON)  
• Other acute demyelinating diseases  
• Transverse myelitis (TM)  
 
Immunologic : 
 
• Anaphylaxis ; 
• Vasculitides ; 
• Arthritis and arthralgia/joint pain ; 
• Multisystem inflammatory syndrome in adults (MIS -A); 
• Kawasaki disease (KD);  
• Fibromyalgia ; 
• Autoimmune thyroiditis  
• Fibromyalgia  
• Kawasaki disease (KD)  
• Multisystem inflammatory syndrome in adults (MIS -A) 
• Vasculitides  
 
Cardiac : 
• Myocarditis;  
• Pericarditis;  
• Acute myocardial infarction (AMI)  
• Arrhythmia  
• Coronary artery disease (CAD)  
• Heart failure and cardiogenic shock  
• Microangiopathy  
• Myocarditis  
• Pericarditis  
• Stress cardiomyopathy  
 
Hematologic : 
 
• Cerebrovascular hemorrhagic stroke  
• Chilblain -like lesions  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077557
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 23 of 237  
Objective  Primary 1  Primary 2  Secondary  
• Th o bocytope i ; 
• Disseminated intravascular coagulation (DIC)  
• COVID19 (for all COVID 19related safety events of interest listed below, an inpatient diagnosis of COVID
19 will be re quired in combination with the codes or laboratory values specified in Appendix Table 2 ; in 
addition, COVID -19 related safety events of interest will only be evaluated using data from 2020 onward 
using the SCRI design ): 
• Se e e COVID 19 dise se ; 
• Microangiopathy ; 
• Heart failure and cardiogenic shock ; 
• Stress cardiomyopathy ; 
• Coronary artery disease (CAD);  
• Arrhythmia ; 
• Deep vein thrombosis (DVT );) 
• Pulmonary embolus;  
• Cerebrovascular hemorrhagic stroke; Hemolytic anemia  
 
• Ce eb o sc l  non-hemorrhagic stroke;  
• Limb ischemia ; 
• Hemorrhagic disease ; 
• Limb ischemia  
• Pulmonary embol ism (PE) 
• Acute kidney injury ; 
• Liver injury ; 
• Chilblain like lesions ; 
• Single organ cutaneous vasculitis ; 
• Thrombocytopenia  
• Thrombosis with thrombocytopenia syndrome (TTS)  
 
Other: 
• Acute kidney injury  
• Appendicitis  
• Death  
• Erythema multiforme  
• Liver injury  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077558
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 24 of 237  
Objective  Primary 1  Primary 2  Secondary  
• Other 
• De th; 
• Narcolepsy / and cataplexy  ; 
• Non-anaphylactic allergic reactions ; 
• Appendicitis Severe COVID -19 disease  
• Stevens-Johnson syndrome (SJS)/Toxic epidermal necrolysis (TEN)  
 
Data source  The VHA Corporate Data Warehouse (CDW) database will be used . and may be supplemented with Medicare 
administrative claims data from the Centers for Medicare & Medicaid Services (CMS) . 
Data analysis  A stepwise approach will be performed for signal detection, evaluation, and verification.  
1) Signal detection: The goal is to provide rapid -cycle, near real -time safety surveillance. In the signal detection phase, 
the SCRI analysis will only include pre -vaccination control intervals as SCRI analyses using  the post-vaccination 
control intervals will be conducted for certain safety events of interest  that require a longer time to accumulate and will 
be used in the signal evaluation phase. COVID-19 diagnosis  (i.e., severe COVID -19 illness, MIS -A). To account for 
multiple testing and bi -weekly review of the data, the m aximized sequential probability ratio test (MaxSPRT) using a 
binomial probability model will be applied. For comparison with individuals who received seasonal influenza 
vaccination, the Poisson -based MaxSPRT will be applied . for all other safety events of i nterest.  
Sequential analyses for each safety event of interest will commence once at least 3 events occur. This approach is 
consistent with the FDA’s COVID -19 Vaccine Safety Surveillance Project to avoid spurious signals from a few early 
events.  Signals will be detected if the critical values are reached via the SCRI or active comparator analysis. Critical 
values will be determined for each safety event of interest based on historical incidence rate, expected upper limit of 
the number of events under the null hypothesis, and pre -specified significance level and power. Incidence rates will 
also be calculated and Kaplan -Meier methods will be used to analyze time to safety event of interest.  
2) Signal evaluation: If signals are detected for safety events of i nterest based on the analysis described above, further 
evaluation will be conducted to refine and confirm such detections. This will include comprehensive quality assurance 
(for example, check for possible duplications of claims or medical records, checkin g for unusual clustering in claim or 
medical record accrual by service date for potential coding issues, check for geographical distribution of cases that may 
be related to lot numbers or diagnostic practice) and multivariate adjustment using Poisson regre ssion to account for 
baseline differences between Pfizer -BioNTech COVID -19 vaccinated and active comparator cohorts. SCRI analyses 
using the post -vaccination control intervals  and the SCCS design with post -vaccination control  time period will be 
conducted as an additional inferential analysis once enough post -vaccination time has accumulated. Lastly, theTo 
address potential period effects, a  comparison to contemporary unvaccinated controls will also be performed with 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077559
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 25 of 237  
Objective  Primary 1  Primary 2  Secondary  
adjustment using inverse probability of treatment w eighting (IPTW).  The assessment  of temporal clustering will also 
be conducted. Signal evaluation analyses will be conducted very six months.  
3) Signal verification: diagnostic validation of the detected safety events of interest via adjudication of medical  records 
by VHA clinicians for outcome validation will be conducted in a representative sample of cases. For rare events, 
potentially all cases may be adjudicated.  
End-of-season analyses (over the course of the 30 -month period) and an end -of-surveillance a nalysis (i.e., at 30 
months) will be conducted. Various subgroup analyses will also be conducted, examining different age groups, 
immunocompromised individuals, individuals with specific comorbidities, those who only received one dose of the 
Pfizer-BioNTech COVID -19 vaccine, those with prior SARS -CoV-2 infection based on medical history or pre -
vaccination serology, those receiving care regularly at VA facilities, and lastly those with VA Priority group 1 status, 
which determines these individuals are of highest  priority for VHA care and likely receive all of their care within the 
VHA system ., and lastly, those with additional Medicare coverage whose Medicare data can be linked to the CDW . 
Notably, CDC recently investigated myocarditis/pericarditis following mRNA COVID-19 vaccinations. To provide 
additional context to the investigation conducted by CDC, separate safety analyses will be prioritized and performed to 
assess the risk of myocarditis/pericarditis following Pfizer -BioNTech COVID -19 vaccination . These anal yses will be 
conducted to align with the rapid -cycle analysis performed by the Vaccine Safety Datalink (VSD).  The number of 
myocarditis/pericarditis events in the risk interval will be identified, and incidence rates per million doses will be 
summarized . Subgroup analyses will also be performed, stratified by age ( e.g., 12-39 years, 40 -49 years, 50 -64 years, 
65+ years ), gender, and race/ethnicity, respectively. Incidence rate ratios will be summarized  to compare the rate of 
myocarditis/pericarditis events between vaccinated individuals who se events occur in a pre-specified risk interval 
versus vaccinated individuals who se events occur  in a comparison interval on the same calendar day. 
Myocarditis/pericarditis events will also be adjudicated via chart review  and validated using the Brighton 
Collaboration’s case definitions. Risk factor analysis may also be conducted among confirmed cases. Lastly, additional 
data surrounding the risk factors, clinical course, and sequelae of the identified myocarditis/pericard itis event up to 365 
days following the event will be collected and summarized . 
 
 
  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077560
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 26 of 237 5. AMENDMENTS AND UPDATE S 
None. 
 
Amendment number  Date Protocol 
section(s) 
changed Summary of amendment(s)  Reason 
1 31 August 
2021 6 Updated the Milestones section.  To add additional information that 
became available after the initial 
protocol was submitted to FDA  
regarding IRB review, EU PAS 
registration, and data collection 
dates.  
1 31 August 
2021 9.1.1 Added clarification on the self-
controlled risk interval ( SCRI) design, 
including a description of the 
measurements when there is a gap 
between risk intervals for the first and 
second dose and an illustration (new 
Figure 2B) . To respond to a request from Center 
for Biologics Evaluation and 
Research (CBER) to demonstrate 
how the period after the risk interval 
for dose 1 and prior dose 2 will be 
handled in the analysis if there is no 
overlap between the risk intervals 
for the two doses.  
1 31 August 
2021 9.1.1 Added that a dditional doses of the 
Pfizer-BioNTech COVID -19 vaccine 
may be included in the analysis.   To address the potentia l approval of 
additional doses. Details for this 
analysis will be further described in 
the statistical analysis plan.  
1 31 August 
2021 9.1.1, 9.3.3, 
9.7.3, 9.7.5, 9.9  Removed SCRI design with pre -
vaccination control interval and added 
SCRI design with post -vaccination To address CBER request to 
remove the pre -vaccination control 
interval as its comparison to the risk 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077561
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 27 of 237 Amendment number  Date Protocol 
section(s) 
changed Summary of amendment(s)  Reason 
control interval for 2 safety events of 
interest (severe COVID -19, 
multisystem inflammatory syndrome 
in adults [MIS-A]) that could not be 
evaluated with the seasonal influenza 
vaccinated comparators.  
Revised Figures 1 -5 to remove pre -
vaccination control interval and 
provide examples for post -vaccination 
interval.  interval may introduce bias and 
reduce the probability of subsequent 
vaccination. Note additional and 
more robust analyses were added to 
signal evaluation phase (see new 
sections under 9.7.3.2.5 and 
9.7.3.2.6). SCRI with post -
vaccination control intervals was 
included in the signal detection 
phase to evaluate severe COVID -19 
and MIS-A as they require COVID -
19 diagnosis, which would not be 
observed in a seasonal influenza 
comparator.  
1 31 August 
2021 9.2.3, 9.4 Added clarification for the 
identification of subgroups who are 
immunocompromised s and 
individuals with specific 
comorbidities . 
Added one additional subgroup of 
interest (individuals with Medicare 
coverage for whom Veterans Health 
Administration [VH A] records can be 
linked to their Medicare claims).  To provide additional detail 
regarding how subgroups who are 
immunocompromised and 
individuals with specific 
comorbidities  will be defined and 
operationalized.  
To respond to a query from CBER 
regarding the potential for 
incomplete data for healthcare 
encounters not received at VHA, an 
additional subgroup of individuals 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077562
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 28 of 237 Amendment number  Date Protocol 
section(s) 
changed Summary of amendment(s)  Reason 
with linked Medicare data has been 
added. 
1 31 August 
2021 9.3.1 All measurement details concerning 
how Pfizer -BioNTech COVID -19 
vaccine and seasonal influenza 
vaccine will be identified in the data 
to an Appendix Table 3 in Section 18. 
Appendix Table 3 includes all specific 
CPT/HCPCS/NDC codes previously 
listed in Section 9.3.1 as wel l as 
additional codes identified at the time 
of the data analysis.  To update the protocol with all 
relevant CPT/HCPCS/NDC codes, 
while maintaining concise language 
in the main text.   
1 31 August 
2021 9.3.1, 18 Added Appendix Table 4 in Section 
18 regarding the LOINC codes used to 
identify COVID -19 RT-PCR Test 
among the study population and 
corresponding reference.  To provide additional details on 
how individuals with prior SARS -
CoV-2 infection will be identified  
in the data.  
1 31 August 
2021 9.3.2, 18 Added frailty index as a baseline 
characteristic of interest.  To describe the identification of 
frailty in the Pfizer -BioNTech 
COVID-19 and seasonal influenza 
cohorts during the 1 -year baseline 
period prior to vacc ination as frailty 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077563
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 29 of 237 Amendment number  Date Protocol 
section(s) 
changed Summary of amendment(s)  Reason 
may be a prognostic factor for 
safety events of interest.  
1 31 August 
2021 9.3.3, 18 Added four additional  safety events of 
interest: thrombosis with 
thrombocytopenia syndrome, 
convulsions/seizures in individuals 
with controlled epilepsy, Steven -
Johnson syndrome/Toxic epidermal 
necrolysis, and hemolytic anemia 
(increasing the number of safety 
events of interest from 42 to 46).  
Reclassified COVID -19-related safety 
events of interest to be measured 
independently of the pa tient’s 
COVID-19 infection status ; this 
change had no impact on the number 
of safety events of interest (reflected 
both in the revised text and revised 
Table 1). 
Added that the clean window may be 
extended (e.g., 2 years).  To consider new safety events 
based on emerging research and 
align with codes from the FDA 
CBER COVID-19 Vaccine Safety 
Surveillance : Active Monitoring 
Master Protocol.20,21,22 
The COVID -19-related safety 
events were reclassified to more 
closely align with the FDA CBER 
COVID-19 Vaccine Safety 
Surveillance : Active Monitoring 
Master Protocol. COVID-19-related 
safety events that were previously 
listed may not necessarily be related 
to COVID -19 infection (e.g., 
coronary artery disease), and 
therefore are defined independent of 
a COVID -19 diagnosis, with the 
exception of “s evere COVID -19 
disease” and “MIS -A” which 
requires a concurrent COVID -19 
diagnosis.   
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077564
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 30 of 237 Amendment number  Date Protocol 
section(s) 
changed Summary of amendment(s)  Reason 
Extending the clean window will 
address the reduction in healthcare 
resource utilization during the 
pandemic to more accurately 
identify incident events.  
1 31 August 
2021 9.7.3.2 Clarified in the Signal Evaluation 
section that the signal evaluation 
analyses will be conducted every six 
months. To provide additional detail on the 
timing of the signal evaluation 
analyses.  
1 31 August 
2021 9.1.3, 9.7.3.2.5  Added self-controlled case series 
(SCCS) design with  full post-
vaccination peri od as an additional  
analysis in the Si gnal Evaluation 
analysis. Added that  Signal 
Evaluation analyses may also be 
conducted based on signa ls detected 
in external sources  or based on 
regulatory request  (e.g., 
myocarditis/pericarditis) .    
  To further align with the CBER 
Master Protocol: Assessment of 
Risk of Safety Outcomes Following 
COVID-19 Vaccination ( March 23, 
2021).23 SCCS analysis has 
increased power compared to SCRI 
design using post -vaccination 
control interval and has been added 
to complement the SCRI design. In 
addition, clarified that Signal 
Evaluation analyses may also be 
conducted based on signals detected 
in external sources or based on 
regulatory request even if such 
analyses were not first id entified in 
the Signal Detection phase of this 
study. 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077565
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 31 of 237 Amendment number  Date Protocol 
section(s) 
changed Summary of amendment(s)  Reason 
1 31 August 
2021 9.7.3.2.6 Added a comparison group of 
contemporary unvaccinated controls 
in the Signal Evaluation analysis. To address the recommendation 
from CBER to include a 
contemporary control group of 
unvaccinated individuals due to 
potential period effects of an active 
comparator design that uses 
historical controls of influenza 
vaccinated individuals.  
1 31 August 
2021 9.7.8 Added new section on 
myocarditis/pericarditis safety 
analysis and risk factor analysis.  To include a separate analys is 
focused on myocarditis/pericarditis 
based on emerging evidence 
regarding th is event in association 
with mRNA COVID -19 vaccines.17  
1 31 August 
2021 9.9 Added strengths and limitations 
associated with the addition of the 
SCCS design, contemporaneous 
unvaccinated controls, and subgroup 
analysis of individuals with linkage to 
Medicare claims data. To further describe the rationale for 
these additional analyses.  
 
 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077566
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 32 of 237 6. MILESTONES  
Milestone  Planned date  
VHA CRADA  execution, Determination of IRB & 
Research Safety and Security exemptions, 
Approval by Designated Member Review[1-
3]Registration in the EU PAS register  To be registered before the start of 
data collection January - February 
2021 
Registration in the EU PAS register VHA CRADA 
and IRB approvals (estimated)  5 March  April 2021  
Start of data collection (estimated)  11 May 2021[14] 
Interim reports 30 June 2021  
31 December 2021  
30 June 2022  
31 December 2022  
End of data collection (estimated)  1030 June 2023[25] 
Final study report  31 December 2023  
Abbreviations:  ACOS, Associate Chief of Staff; COVID -19, Coronavirus disease 2019; CRADA, 
Cooperative Research and Data Agreement; IRB, Institutional Review Board; EUA, Emergency Use 
Authorization; FDA, Food and Drug Administration; NNERC VAMC, Northern New England Research 
Consortium VA Medical Centers; R&D, Research and Development; SRSS, Subcommittee on Research 
Safety and Security; VA, Veterans Affairs; VAIRRS, VA Innovation and Research Review System; VINNE, 
Veteran’s IRB of Northern New England; VHA, Veterans Health Adm inistration ; US, United States .  
Notes:  
[1] Start of data collection is the planned [1] IRB exemption determination was granted in accordance with 38 
CFR 16 by the Veteran’s IRB of Northern New England (VINNE), White River Junction VA Medical 
Center, White River Junction, VT for the signal detection  and signal evaluation phases. Prior to progressing 
to the signal verification phase for chart review, a second IRB review application will be submitted for an 
expedited or full review. The two -stage IRB application process is to expedite the initiation of  the project.  
[2] Research Safety and Security exemption determination was granted by the Subcommittee on Research 
Safety and Security (SRSS), VA Innovation and Research Review System (VAIRRS).  
[3] Approved by Associate Chief of Staff for Research and Deve lopment (ACOS/R&D) and R&D 
Committee of the Northern New England Research Consortium VA Medical Centers (NNERC VAMC).  
[4] Start of data collection is the  date for starting data extraction for the purposes of the study analysis. The 
initial data analysis will include the includes Pfizer-BioNTech COVID -19 vaccine exposure since exposures 
from December 11, 2020 ,  (the EUA approval date by the US FDA .) to March 12, 2021 (the data cutoff date).  
[25] End of data collection is the planned date on which after the Pfizer -BioNTech COVID -19 vaccine 
exposure data reached 30 mont hs post-EUA approval . 
 and the last day of the month that the study will be completed.  
  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077567
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 33 of 237 7. RATIONALE AND BACKGROUND  
In March 2020, the World Health Organization (WHO) declared a global pandemic for the 
coronavirus disease 2019 (COVID -19) due to the severe acu te respiratory syndrome 
coronavirus 2 (SARS -CoV-2), which was first identified by public health officials in China 
in December 2019 .1 The COVID -19 pandemic presents an unprecedented public health 
crisis. As of January 7, 2021, over 21.4 million COVID -19 cases and 364,000 deaths have 
been reported in the United States (US) alone.2 To date, the incidenc e of COVID -19 has 
continued to rise, largely affecting the elderly and middle -aged individuals, with worsening 
clinical sequelae linked to increasing age and comorbid conditions ( e.g., cardiovascular 
disease, active cancer, obesity, diabetes and chronic lu ng disease).15,1624,25 SARS-CoV-2 is a 
well-adapted highly infectious human pathogen with a case fatality rate that ranges between 
0.5% and 20%, based on the individual’s age, gender, race, and como rbidites.1726  
Pfizer and BioNTech have partnered to develop a novel messenger RiboNucleic Acid 
(mRNA) vaccine against SARS -CoV-2 for the prevention of COVID -19 (Candidate 
BNT162b2). To this end, Pfizer is conducting a Phas e 1/2/3, randomized, placebo -controlled, 
observer-blind, dose -finding, vaccine candidate -selection, and efficacy study among healthy 
individuals (NCT04368728). In their Phase 1 trial evaluating safety and immunogenicity of 
two mRNA vaccine candidates ( i.e., BNT162b1, BNT162b2) at various dose levels, 
candidate BNT162b2 was selected for advancement to a pivotal Phase 2/3 safety and efficacy 
evaluation due to its milder systemic reactogenicity profile, especially in older adults.1827 The 
study was initiated in July 2020 with a target enrollment of 43,998 individuals .1928  
The US Food and Drug Administration (FDA) announced that regulatory emergency use 
authorization (EUA ) as well as full approval of any COVID -19 vaccine will require 
demonstrating prevention of the disease or decrease in its severity in at least 50% of the 
individuals who receive it. In addition, data from Phase 3 studies are required to include a 
median follow-up duration of at least 2 months after completion of the full vaccination 
regimen to assess the vaccine’s benefit -risk profile, especially adverse events and cases of 
severe COVID -19 in vaccinated study subjects.20,2129 The FDA reviewed the available safety 
data of the Phase 1/2/3 trial from 37,586 participants 16 years of age and older and did not 
identify any specific safety concerns. In addition, the analysis of available efficacy data from 
36,523 participants 12 years o f age and older without evidence of prior SARS -CoV-2 
infection at least 7 days after receiving the second dose demonstrated  95% efficacy of the 
vaccine in the prevention of COVID -19 (as confirmed by 8 vs. 162 COVID -19 cases in the 
vaccine and placebo grou ps, respectively).3,4 Based on these safety and efficacy data, as well 
as a review of manufacturing information regarding product quality and consistency, the 
FDA determined that the known a nd potential benefits of the vaccine outweighed the known 
and potential risks for the prevention of COVID -19 in individuals 16 years of age and older .4 
Therefore on December 11, 2020, the Pfizer -BioNTech COVID -19 vaccine was granted an 
Emergency Use Authorization ( EUA) by the FDA to prevent COVID -19 in individuals 16 
years of age and older .5   
With respect to geographic regions other than the US, on December 2, 2020, the United 
Kingdom (UK) was the first country in the wo rld to grant temporary authorization for 
emergency use of the Pfizer -BioNTech COVID -19 vaccine.6 On December 21, 2020, the 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077568
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 34 of 237 European Medicines Agency (EMA) granted the Pfizer -BioNTech  COVID-19 vaccine a 
conditional marketing authorization (CMA) for use among individuals 16 years of age and 
older throughout all of the European Union’s (EU) 27 member states.7   
As required by the EUA, post-authorization observational studies using real -world data are 
needed in order to assess the association between Pfizer -BioNTech COVID -19 vaccine and 
pre-determined safety events of interest (including deaths , hospitalizations, and severe 
COVID-19) among individuals administered the vaccine in both the population at large and 
in populations of interest ( e.g., immunocompromised individuals, elderly, and those with 
specific comorbidities ).4 Post-authorization  safety evalua tions are important for identifying 
rare, serious safety events of interest in larger populations that may not have been detected 
during clinical trials (either due to sample size or selected study populations), and ensure a 
favorable benefit -risk ratio po st-trial. Pfizer in collaboration with the US Veterans Health 
Administration ( VHA) of the Department of Veterans Affairs (VA)  and Analysis Group 
herein propose post -EUA active safety surveillance of safety events of interest based 
primarily  on the Priority  List of Adverse Events of Special Interest from the Brighton 
Collaboration’s Safety Platform for Emergency vACcines (SPEAC) Project , and from the 
FDA and  thepreliminary list of safety events of interest  presented at the September 22 , 2020, 
meeting of Centers for Dise ase Control and Prevention’s ( CDCCDC’s) Advisory Committee 
on Immunization Practices (ACIP)  on the enhanced safety monitoring recommendation .of 
COVID-19 vaccines .8,9 This safety surveillance study will  identify and evaluate rapid, near 
real-time potential safety signals associated with the Pfizer -BioNTech COVID -19 vaccine in 
the large-scale VHA electronic medical record ( EMR) database. The observed rates of safety 
event of interest will be compared to expected rates derived from self -controls and active 
comparators. Part of the methodologies used in this study are constructed based on 
approach es previously used by the Post -Licensure Rapid Immunization Safety Monitoring  
(PRISM) program  for the H1N1 vaccine .810 This non-interventional study is designated as a 
Post-Authorization Safety Study (PASS) commitment to the US FDA  and is a Category 3 
commitment in the EU Risk Management Plan .  
8. RESEARCH Q UESTION AND OBJECTIVES  
Research question: what are the incidence rates of safety events of interest (based on adverse 
events of special interest [AESI]) among individuals vaccinated with the Pfizer-BioNTech 
COVID-19 vaccine within the US VHA  system overall  and in sub -cohorts of interest as 
compared to expected rates of those events?  
Primary study objectives:  
• To assess whether individuals in the VHA  system experience increased risk of safety 
events of interest following receipt of the Pfizer -BioNTech COVID -19 vaccine;  
• To assess whether sub -cohorts of interest ( i.e., immunocompromised, elderly, with 
specific comorbidities, individuals  receiving only o ne dose of the Pfizer-BioNTech 
COVID-19 vaccine, and individuals with prior SARS -CoV-2 infection ) in the VHA 
system experience increased risk of safety events of interest following receipt of the 
Pfizer-BioNTech COVID -19 vaccine.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077569
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 35 of 237 Secondary study objective s: 
• To characterize utilization patterns of the Pfizer-BioNTech COVID -19 vaccine 
among individuals  within the VHA including estimating the proportion of individuals  
receiving vaccine, 2-dose vaccine completion  rate, and distribution of time gaps 
between the  first and second dose , demographics and health histories of recipients, 
overall and among the sub-cohorts of interest.  
9. RESEARCH METHODS  
9.1. Study Design  
This post-EUA active safety surveillance program will employ a rapid-cycle, longitudinal, 
observational cohort study design  to provide early real -world safety information . The self-
controlled risk interval (SCRI) design will be used to sequentially monitor occurrence of 
safety events of interest while controlling for time -invariant confounders (such as sex, race, 
chronic illness, and state). In addition, safety events of interest associated with Pfizer -
BioNTech COVID -19 vaccinations will be sequentially monitored and compared to 
recipients of influenza vaccine in the VHA between 2014/2015 to 2018/2019 .8,22 10,30 
9.1.1.  Self-Controlled Risk Interval (SCRI) Design with Post -Vaccination Control 
Interval  
The SCRI design uses data from cases ( i.e., individuals who experience safety events of 
interest following vaccin ation) to compare the risk interval following vaccination to pre or 
post-vaccination non -risk intervals (“prevaccination control interval” and “ post-vaccination 
control interval”) in the same individual.23 Whether a pre  or postvaccination control 
interval is used will depend  on the clinical nature, seasonality, and frequency of the safety 
event of interest , as described in greater detail below.31 A length of 42 days has been used to 
define the risk interval i n SCRI design studies for signal detection to ascertain the safety 
profile of the H1N1 vaccine.8,2210,30 The same length of risk interval is proposed here, subject 
to further modification based on clinica l input, clinical trial data, biologic plausibility, and 
published literature. The day of vaccination will only be included in the risk period for those 
safety events  of interest for which a same -day occurrence  is biologically plausible 
(e.g., anaphylaxis) . 
As some individuals may choose to decline or delay Pfizer BioNTech COVID 19 
vaccination soon after an illness (known as the “healthy vaccinee effect”),24 the 
prevaccination control interval will exclude the 14 day period before vaccination.25 While 
using a prevaccination control period allows for timely analysis, especially pertinent for 
rarer  safety events of interest , aA post-vaccination control interval would be more 
appropriate and will be used for certain safety events of interest for the following reasons: (1) 
a recent prior safety event of interest might preclude vaccination ( i.e., anaphylaxis), (2) 
individuals might have an underlying conditi on that is also a contraindication for vaccination 
(i.e., seizure disorder), or (3) safety event of interest and vaccination may be seasonal in 
nature.2632 The time between the risk and control intervals will be determined based  on the 
biological mechanism of action for each safety events  of interest assessed, and may be 
subject to change based on further clinical input . Examples  of the SCRI design with a pre
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077570
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 36 of 237 vaccination control interval and a post-vaccination control interval  (in an individual  who only 
receives the first dose of vaccine) is presented in Figure 1 below. 
Figure 1. Example of SCRI Design for Assessment of  a Safety Event of Interest with a 
42-day Risk Interval in an Individual who Receives Only One Vaccine Dose, 
Showing Both Pre  andwith Post-vaccination Control Intervals * 
 
 
*The risk interval may include day 0, date of Pfizer -BioNTech COVID -19 vaccination, for some of the safety 
events of interest assessed ( e.g., anaphylaxis). The length of the risk interval will vary across each safety event 
of interest  and may be subject to change based on clinical input . Note that some individuals  may not receive the 
complete course of vaccination, and thus may only receive the first dose of vaccine. This is represented in 
Figure 1 while Figure 2 represents an example where the complete course with 2 doses are received.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077571
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 37 of 237 Two doses of the Pfizer -BioNTech COVID -19 vaccine are recommended 3 weeks apa rt.  
This study program will monitor safety events of interest that occur after dose 1 before dose 
2 (i.e., during risk interval 1), after dose 2 ( i.e., during risk interval 2), and aggregate for 
doses 1 and 2 ( i.e., risk interval 1 + risk interval 2), respectively, for individuals receiving 
both doses. Additional doses of the Pfizer -BioNTech COVID -19 vaccine may be included in 
the analysis should they be approved , and those details will be described in the  statistical 
analysis plan  (SAP). 
For Given the risk intervals for specific safety events of interest  range from 1 day to 90 days 
(please see Table 1 in Section 9.3.3), the time between the first and second dose may be 
longer or shorter than the recommended risk interval for a given safety event  after the first 
dose.  See Figure 2 below for SCRI design examples  where a safety event with a 42 risk 
interval window (e.g., Bell’s palsy ; Table 1 in Section 9.3.3 ) is assessed in hypothetical  
individual s who receive two doses of the vaccine, two separate control intervals will be 
defined to correspond to the risk interval associated with each dose (regardless of whether 
pre or postvaccination control intervals are used). See Figure 2 below for an example in an 
individual who receives two doses of Pfizer-BioNTech  COVID-19 vaccine ,: Figure 2A 
shows the SCRI design  with the second dose received 21 days after the first. Safety (i.e., the 
risk interval for dose 1 overlaps wi th the risk interval for dose 2) , while Figure 2B shows the 
SCRI design with the second dose received 60 days after the first (i.e., there is a gap between 
the end of the risk interval  for dose 1  and dose 2  initiation) . For the first scenario ( Figure 
2A), the risk interval for  dose 1 will be censored at the time of dose 2; further, s afety events 
of interest that occur during the overlapping period of risk interval 1 and risk interval 2 
(shown in gray shading in Figure 2)Figure 2A) may be flagged for separate analyses to 
discern the additive effect of Pfizer -BioNTech COVID -19 vaccine dose 1 and dose 2.  For the 
second scenario ( Figure 2B), events will only be measured during the risk intervals, ignoring 
the gap between the end of the risk interval for dose 1 and dose 2 initiation.  
 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077572
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 38 of 237 For each analysis, control intervals  corresponding to the risk interval s will be defined either 
at end of the risk interval for dose 1 (for individuals with only one dose observed) or after the 
risk interval for dose 2 (for individuals with two doses observed), regardless of  whether of 
the analyses focus on safety events after  dose 1, after dose 2, or aggregated for doses 1 and 2  
(Figure 2A and Figure 2B). 
 
Figure 2. Example of SCRI Design with Overlapping for Assessment of a Safety Event 
of Interest with a  42-day Risk Intervals when Two Doses of Pfizer BioNTech 
COVID 19 Interval in an Individual who Receives Two  Vaccine are 
Administered, Showing a Pre  andDose s, with  Post -vaccination Control 
Interval Intervals  
 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077573
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 39 of 237  
 
 
 
9.1.2.  Active Comparator  Design  
In the active comparator  design, the frequency of safety events of interest among individuals 
who received Pfizer -BioNTech COVID -19 vaccine  from December 11, 2020 onward  will be 
compared with the event frequency amo ng recipients of the seasonal influenza vaccination in 
five prior seasons, between 2014/2015 t hrough 2018/2019. Data in peri -COVID time periods 
from January 2020 to present are excluded because of pandemic -associated under -utilization 
of health resources a nd under-reporting of medical events. The same risk interval length 
(e.g., 42 days) will be used to evaluate safety events  of interest following vaccination with 
Pfizer-BioNTech COVID -19 vaccine and to assess safety events  of interest occurring after 
vaccination for seasonal influenza in prior seasons. The observed number of safety events  of 
interest for Pfizer -BioNTech COVID -19 vaccine will be compared to the expected number 
calculated for the influenza vaccine in past seasons.8.10 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077574
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 40 of 237 9.1.3.  Additional Study Design s in the Signal Evaluation Phase  
There will be additional study designs conducted during the signal evaluation phase if a 
signal is detected from the above analyses  in the signal detection phase . These include 
analyses using self-controlled case series (SCCS) and comparison of vaccinated unvaccinated 
contempora ry controls. Additionally, signal evaluation analyses may also be conducted for 
signals detected in external sources or based regulatory request  (e.g., 
myocarditis/pericarditis) . These analyses are further detailed in Section 9.7.3.2. 
9.1.3. 9.1.4.  Study Period  
The study will be conducted for a period of 30 months , starting on December 11, 2020 
onward, with data collection concluding on June 10,  2023.  
9.2. Setting  
The study population will be kept as broad as possible in order to capture safety events  of 
interest that occur among all vaccinated individuals.  
9.2.1.  Inclusion Criteria  
• Record of at least one dose of Pfizer -BioNTech COVID -19 vaccine  in the per iod of 
December 11, 2020 to present , or 
• Record of at least one dose of seasonal influenza vaccine during prior flu seasons, 
from 2014/2015 to 2018/2019 (applies to active comparator s only); and 
• At least 1 year of enrollment in and no disenrollment from VHA benefits ( i.e., the 
baseline period) pr ior to Pfizer -BioNTech COVID -19 or seasonal influenza 
vaccination date.  
9.2.2.  Exclusion criteria  
• Individuals who receive at least one dose of Pfizer -BioNTech  COVID-19 vaccine in 
addition to a COVID -19 vaccine from a manufacturer other than Pfizer -BioNTech 
will be identified and reportedsummarized , but they will be excluded from further 
analysis. 
9.2.3.   Subgroups  
Safety surveillance may be conducted for subgroups of interest , including, but not limited to:  
• Immunocompr omised individuals ;, defined as  individuals diagnosed with 
symptomatic human immunodeficiency virus (HIV)/acquired immunodeficiency 
syndrome ( AIDS), hematologic malignancy, or other immune conditions; individuals 
diagnosed with solid malignancy, organ transplant, or rheumatologic/inflammatory 
conditions, all of whom were administered chemotherapy or immune modulators;  
individuals diagnosed with rheumatologic/inflammatory conditions and administered 
systemic cortico steroids; or individuals who were administered chemotherapy, 
immune modulators, or systematic steroids for at least 14 days ;13  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077575
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 41 of 237 • Different age groups , with a focus on the elderly  (e.g., < 35, 35 - < 45, 45 - < 55, 55 - 
< 65, 65 - < 75, > 75); 
• Individuals with specific comorbidities;  
• Individuals with specific comorbidities  identified as high risk for COVID -19 by the 
CDC (i.e., cancer, chronic kidney disease, chronic obstruction pulmonary disease 
[COPD], Down Syndrome, cardiovascular conditions [e.g., heart failure, coronary 
artery disease, or cardiomyopathies], immunocompromised state from solid organ 
transplant, obesity [body mass index (BMI) of 30 kg/m2 or higher but < 40 kg/m2], 
severe obesity [BMI of 40 kg/m2or higher], sickle cell disease, smoking, type  1 and 2 
diabetes mellitus) ;12 
• Individuals receiving only one dose of Pfizer-BioNTech  COVID-19 vaccine;  
• Individuals with prior SARS -CoV-2 infection based on medical history or 
pre-vaccination serology ; (Appendix Table 4); 
• Individuals with regular use of VHA medical care, defined as at least two outpatient 
(excluding  emergency department  [ED,], as ED visits may not be considered regular) 
or inpatient encounters in the one year prior to vaccination . The encounters must be 
separated by >  30 days (for inpatient, by admission date), and at least one must be 
within six months prior to the date of vaccination . This will ensure that individuals  
have ongoi ng health care encounters, particularly near the vaccination date, and 
regularly receive their healthcare from VHA facilities, rather than outside facilities  
that would not be captured in the  CDW;VHA’s Corporate Data Warehouse ( CDW); 
• Individuals  who are in the  VA priority group 1 Veteran. These individuals have either 
the highest levels of service connected disability  (> (≥50% disabling), are considered 
unemployable, or have received the medal of honor .27.33 Individuals categ orized as 
priority group 1 are the highest priority for VHA care. This will ensure that the 
individual  is more likely to receive all of their care from a VA facility .  
• Individuals enrolled in the VHA  with dual coverage who  are also identified in the 
Centers for Medicare & Medicaid Services (CMS) Medicare administrative claims 
data, which will be linked to the CDW,  in order to supplement CDW data for a more 
complete evaluation of health care encounters .  
Additional subgroups of interest will be assessed as ad ditional information becomes 
available from ongoing clinical trials, Vaccine Adverse Event Reporting System (VAERS), 
and other sources that will inform the Pfizer -BioNTech COVID -19 vaccine safety profile.  
Given that VA population has a median age of over 46 years for females and is comprised of 
approximately 90% males , the evaluation of the Pfizer-BioNTech COVID -19 vaccine safety 
during pregnancy, including fetal death and infant outcomes, may have poor feasibility  and 
will therefore not be conducted.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077576
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 42 of 237 9.3. Vari ables  
9.3.1.  Exposure of Interest  
Administration of Pfizer -BioNTech COVID-19 vaccine  post-EUA approval will be identified 
based on the following : (see Appendix  Table 3 for additional details) : 
• Current Procedural Terminology (CPT) code 91300 ( Severe acute respiratory 
syndrome coronavirus 2 (SARS CoV2) (coronavirus disease [COVID 19]) vaccine, 
mRNALNP, spike protein, preservative free, 30 mcg/0.3mL dosage, diluent 
reconstituted, for intramuscular use ) codes and associated vaccine administration 
HCPCS codes  corresponding to the first dose: 0001A (ADM SARS CoV2 30 
mcg/0.3mL 1st), and the second dose: 0002A (ADM SARS CoV2 30 mcg/0.3mL 
2nd);9,10; OR 
 
• 10 and 11 -digit National Drug Codes (NDCs ) 5926710001 (corresponds to first 
dose), 59267 100001 (corresponds to second dose ); OR 
• Immunization records that contain data on vaccine code descriptor, vaccine 
manufacturer ( i.e., Pfizer), lot number, injection site, and date(s) of immunization.911  
Relevant codes will be continuously reviewed and amended if new codes are added.   
Person-time at-risk exposure to the  first dose only, overlapping first and second doses, and 
second dose only will be analyzed separately.  
Administration of the seasonal influenza vaccine during 2014/2015 through 2018/2019 flu 
seasons wil l be identified based on the following : (see Appendix  Table 3 for additional 
details): 
• CPT codes   
o 90654 (Influenza virus vaccine, trivalent (IIV3), split virus, preservative free, 
for intradermal use) ; OR 
o 90656 (Influenza virus vaccine, trivalent (IIV3), split virus, preservative free, 
0.5 mL dosage, for intramuscular use) ; OR 
• 90658 (Influenza virus vaccine, trivalent (IIV3), split virus, 0.5 mL dosage, for 
intramuscular use );; OR 
• 10 and 11 -digit NDCs; OR  
• Immunization records that contain data on vaccine code descriptor, vaccine 
manufacturer, lot number, injection site, and date(s) of immunization.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077577
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 43 of 237 9.3.1.1.  Pfizer -BioNTech COVID -19 Vaccine Groups of Interest  
While the primary vaccination group of interest is all in dividuals receiving Pfizer -BioNTech 
COVID-19 vaccine (irrespective of receipt of seasonal influenza vaccination), additional 
subsets of the study population will be studied, similar to the PRISM safety surveillance 
program of H1N1 vaccine safety:810  
Cohort A:  Individuals vaccinated with Pfizer -BioNTech COVID -19 vaccine who did not 
receive the influenza vaccine during the flu season in which COVID -19 vaccination 
occurred; 
Cohort B:  Individuals vaccinated with Pfizer -BioNTech COVID -19 vaccine who received 
the seasonal influenza vaccine at least 42 days prior to COVID -19 vaccination  during the 
same flu season in which COVID -19 vaccination occurred;  
Cohort C:  Individuals vaccinated with Pfizer -BioNTech COVID -19 vaccine who received 
the seasonal influenza vaccine within 42 days before or any time after COVID -19 
vaccination  during the same flu season in which COVID -19 vaccination occurred;  
Cohort D:  Individuals vaccinated with both Pfizer -BioNTech COVID -19 vaccine and the 
seasonal influe nza vaccine on the same day.  
The following sub -cohorts will be assessed for each of the Cohorts A -D: 
• Individuals vaccinated with only 1 dose ( i.e., incomplete course) of Pfizer -BioNTech 
COVID-19 vaccine ; 
• Individuals vaccinated with 2 doses ( i.e., complete course) of Pfizer -BioNTech 
COVID-19 vaccine.  
9.3.2.  Baseline Characteristics  
The following data elements regarding baseline demographic and clinical characteristics will 
be assessed based on  a 1-year baseline period  prior to the date of vaccination with Pfizer -
BioNTech COVID -19 vaccine and date of seasonal influenza vaccination for active 
comparator s. Depending on the attrition rate, the length of the baseline period may be 
modified to 6 months.  All diagnoses, procedures, and medications will be identified by the 
International Classification of Diseases, Tenth Revision, Clinical Modification ( ICD-10-CM) 
diagnosis codes, ICD -10-PCS (procedure coding system) codes, ICD10CM Current 
Procedural Terminology (CPT), CPT, or Healthcare Common Procedure Coding System 
(HCPCS) procedure codes, and generic drug names, as appropriate  (Appendix Table 1). The 
following demographic and clinical characteristics will be assessed:  
Demographic s: 
• Age  
• Sex 
• Race/ethnicity  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077578
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 44 of 237 • State 
• VHA service area  
 
Clinical characteristics:  
• Smoking status  
• Body mass index ( BMI) 
• History of anaphylaxis/ allergic reactions  
• Previous anaphylaxis  of vaccine component  
• History of hospitalizations  
• Frailty index  
• Charlson comorbidity index (CCI)  
• Selected c omorbidities   
o Autoimmune d isease 
o Asthma 
o Bleeding diathesis or condition associated with prolonged bleeding  
o Cancer 
o Cardiovascular conditions  
o Chronic kidney disease/dialysis  
o Chronic obstructive pulmonary disease (COPD)/COPD/interstitial lung 
disease 
o Diabetes mellitus  
o Down syndrome  
o Sickle cell disease  
o Hepatitis B virus (HBV) 
o Hepatitis C virus (HCV) 
o Human immunodeficiency virus (HIV) 
o HIV 
o Hyperlipidemia  
o Hypertension  
o Liver disease  
o Neurological disease  
o Other immune deficiencies  
o Solid organ transplant  
o Venous thromboembolism (VTE) 
• Concurrent immunizations   
o Seasonal influenza vaccine  
o Tetanus di phtheria and pertussis (Tdap or Td)  
o Chickenpox (varicella)  
o Shingles (herpes zoster recombinant and/or live)  
o Human papillomavirus (HPV)  
o Pneumococcal conjugate  
o Pneumococcal polysaccharide  
o Hepatitis A  
o Hepatitis B  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077579
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 45 of 237 o Meningococcal conjugate (MenACWY) and serogrou p B meningococcal 
(MenB) 
o Haemophilus influenza type b  
 
Specific covariates of interest for the prioritized analysis of myocarditis/pericarditis are 
described in Section 9.7.8 . 
9.3.3.  Outcomes  
The safety events of interest for active surveillance were identified based on  the Priority List 
of Adverse Events of Special Interest from the Brighton Collaboration’s Safety Platform for 
Emergency vACcines ( SPEAC) Project, the FDA and Centers for Disease  Control and 
Prevention ( CDC) enhanced safety monitoring recommendations.288,299 Endpoints of special 
interest in signal detection, as noted by the FDA and CDC’s Advisory Committee on 
Immunizati on Practices ( ACIP) are denoted in italics .29. 9 If unanticipated potential safety 
events of interest are identified during the course of  surveillance, they will be added to the 
list and included in the analyses . See Appendix  Table 2 for the operational definitions of the 
outcome variables based on ICD -10-CM diagnosis codes, which may be refined as the study 
progresses based on additional available information  and the published literature  (e.g., 
frequency of ICD -10-CM codes). Outpatient  (including , ED), and/or inpatient setting s will be 
used to identify safety events of interest, depending on the type of event. The specific 
encounter setting considered for each safety event of interest is summarized in  Table 1. Any 
record of death will be captured, regardless of whether the individual died in a healthcare or 
non-healthcare setting. The following safety events of interest will be assessed:  
Neurologic : 
• Generalized convulsions/seizures  
• Guillain Barré syndrome (GBS)  
• Aseptic meningitis  
 Encephalitis/encephalomyelitis  
• Bell’s palsy 
• Cerebrovascular non -hemorrhagi c stroke  
• Convulsions/seizures in individuals with controlled epilepsy 
• Encephalitis/encephalomyelitis  
• Guillain -Barré Syndrome (GBS)  
• Generalized convulsion/seizures  
• Multiple sclerosis (MS)  
• Optic neuritis (ON)   
• Other acute demyelinating diseases  
• Transverse  myelitis (TM)  
 
Immunologic : 
 
• Anaphylaxis  
• Arthritis and arthralgia/joint pain  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077580
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 46 of 237 • Autoimmune thyroiditis  
• Fibromyalgia  
• Kawasaki disease  (KD)  
 Multiple sclerosis (MS) Optic neuritis (ON) Bell’s palsy 
 
Immunologic : 
 
 Anaphylaxis  
• Vasculitides  
• Arthritis and arthralgia/joint pain  
• Multisystem inflammatory syndrome in adults  (MIS -A) 
• Vasculitides  
 Kawasaki disease  (KD)  
• Fibromyalgia  
 Autoimmune thyroiditis  
 
Cardiac : 
• Myocarditis  
• Pericarditis  
• Acute myocardial infarction (AMI)  
• Arrhythmia  
• Coronary artery disease (CAD)  
 
Hematologic : 
 
 Thrombocytopenia  
• Disseminated intravascular coagulation (DIC)  
 
COVID 19 (for all COVID 19related safety events  of interest  listed below, an inpatient 
diagnosis of COVID 19 will be required in combination with the codes or laboratory values 
specified in  Appendix Table 2; in addition, COVID 19 related safety events of interest will 
only be evaluated using data from 2020 onward using the SCRI design ): 
 
• Severe COVID 19 disease  
 Microangiopathy  
• Heart failure and cardiogenic shock  
• Microangiopathy  
• Myocarditis  
• Pericarditis  
• Stress cardiomyopathy  
 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077581
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 47 of 237 Hematologic : 
 
• Cerebrovascular hemorrhagic stroke  
• Chilblain -like lesions  
• Disseminated intravascular coagulation (DIC)  
 Coronary artery disease (CAD)  
 Arrhythmia  
• Deep vein thrombosis (DVT)  
• Pulmonary embolus  
• Hemolytic anemia  
 Cerebrovascular hemorrhagic stroke  
• Cerebrovascular non hemorrhagi c stroke  
• Limb ischemia  
• Hemorrhagic disease  
• Limb ischemia  
• Pulmonary embol us (PE) 
• Acute kidney injury  
 Liver injury  
 Chilblain like lesions  
• Single organ cutaneous vasculitis  
• Thrombocytopenia  
• Thrombosis with thrombocytopenia syndrome (TTS)  
 Erythema multiforme  
 
Other : 
• Acute kidney injury  
• Appendicitis  
• Death  
• Erythema multiforme  
• Liver injury  
 Death  
• Narcolepsy / and cataplexy  
• Non-anaphylactic allergic reactions  
• Appendicitis  
• Severe COVID -19 disease   
• Stevens-Johnson syndrome (SJS)/Toxic epidermal necrolysis (TEN ) 
 
The risk and control intervals selected  for the SCRI analysis  for each safety event of interest 
are based on biological plausibility a nd precedents in the published literature ( Table 1). A 
safety event of interest will only be counted if it can be assigned to 1) the risk interval 
(following Pfizer -BioNTech COVID -19 vaccination , (all designs) , 2) the prevaccination 
control interval,  3) the post-vaccination control interval , (self-controlled designs) , or 43) the 
risk interval for the active co mparators receiving seasonal influenza vaccine . (active 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077582
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 48 of 237 comparator design) . Events outside the intervals will not be counted. Only the individual’s 
first instance of a safety event of interest following a specified clean window ( i.e., the 
occurrence -free baseline period used to define incident outcomes  during which individuals 
enter the study cohort only if the safety event of interest did not occur during this period) will 
be included; this means that if a safety event of interest is identified but diagnosi s codes (or 
laboratory values in the case of select safety events of interest) corresponding to the safety 
event of interest are also observed during the clean window, it will not be counted. The 
duration of the pre -specified window will differ by safety e vents of interest in order to rule 
out pre-existing events. This approach is consistent with the FDA’s COVID -19 Vaccine 
Safety Surveillance Project.116 Additionally, the length of the clean window may be 
extended (e.g., 2 ye ars) given the reduction in healthcare resource utilization since the start of 
the pandemic.  By way of example, a safety events of interest for the SCRI design can be 
considered in the following ways:  
• If a safety event of interest occurs in the individual’ s prevaccination control risk 
interval and there are no other diagnosis codes for the same safety event of interest in 
the clean window (e.g., 1 -year prior to thatvaccination  date), the safety event of 
interest should be assigned to the prevaccination control risk interval. 
o If a safety event of interest occurs in the pre vaccination control interval but 
another diagnosis code for the same safety event of interest is identified 
during the risk interval, then the safety event of interest will not be assig ned to 
the risk interval and will only be assigned to the pre vaccination control 
interval as it will have occurred in the required clean window preceding the 
risk interval. However, if an outpatient safet y event of interest occurs in the 
clean window and an inpatient occurrence for the same type of safety event of 
interest occurs in the risk interval, the inpatient occurrence will be counted in 
order to capture event exacerbation.  
o If a safety event of int erest occurs in the risk interval and another diagnosis 
code for the same safety event of interest is identified during the post -
vaccination control interval, then the safety event of interest will only be 
assigned to the risk interval  
o If a safety event o f interest occurs in the post -vaccination control interval and 
there are no other diagnoses for the same safety event of interest in the  risk 
interval and clean window  (e.g., oneyear prior to this date), which also 
includes the pre -vaccination control interval , then the safety event of interest 
will be assigned to the riskpost-vaccination  control interval.  
• The same appr oach will be applied for the post vaccination control intervals.  
• The risk intervals for outcome evaluation for the active comparators who received 
seasonal influenza vaccination will be the same as for the individuals who received 
Pfizer-BioNTech COVID -19 vaccine .  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077583
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 49 of 237 • However, it is possible that some safety events  of interest do not have a precise time 
interval from which to evaluate risk, for example if biological plausibility is unknown 
or the diagnostic time window is more delayed than anticipated. In these cases, 
misspecification of the risk (and control) intervals could result in misclassification 
and introduce bias, often toward the null. For instance, the assumption of a longer risk 
interval than is true may result in “washing out” the signal, and an erroneously short 
risk interval may similarly result in underestimation of effect when using post -
vaccination time intervals for self -control. To address this, sensitivity analyses may 
be conducted with varying risk intervals (longer as well as shorter) in order to 
increase the likelihood that the safety risk is detected accurately. Additionally, if 
further refinement and evaluation is necessary, temporal scan statistics may be used 
to empirically identify the at-risk time interval by evaluating clusters of safety even ts 
of interest. This will be further described in the SAP.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077584
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 50 of 237 Table 1. Outcome algorithms for SCRI analysis, with risk and control intervals  
Safety Event of Interest * Setting  
(Inpatient [IP], 
Outpatient [OP ], 
Emergency 
Department [ED ]) Clean 
window  Prevaccination 
control interval 
(days)  Ri
sk 
int
er
val 
(d
ay
s) Post-vaccination control 
interval (days)  
Neurologic  
Aseptic meningitis  IP only16 6 months16 1-
42
34 43-8434 
Bell’s palsy16Ge e li ed 
co lsio /sei es8 IP or OP8 6 months  N/
A1
-
42 43-840
14 1529 
GBS8,22 Cerebrovascular non -
hemorrhagic stroke16 IP onlyIP, primary 
position only14 1 year  N/A 1-
42
28 438429-56 
Aseptic 
meningitis30Convulsions/seizures in 
individuals with controlled 
epilepsy35 IP or OP-EDIP 
only14 1 year N/A 1-
42
90 91-1804384 
Encephalitis/encephalomyelitis816 IP only14 6 months   1 
year 56 through 15 1-
42 43-84N/A 
Other acute demyelinating 
diseases8Guillain-Barré Syndrome 
(GBS)16  IP or OP8IP, primary 
position on 1 year 98 through 15 1-
42 43-84N/A 
Generalized convulsion/seizures10 IP or OP-ED 6 months  0-
14 15-29 Deleted Cells
Deleted Cells
Deleted Cells
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077585
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 51 of 237 Table 1. Outcome algorithms for SCRI analysis, with risk and control intervals  
Safety Event of Interest * Setting  
(Inpatient [IP], 
Outpatient [OP ], 
Emergency 
Department [ED ]) Clean 
window  Prevaccination 
control interval 
(days)  Ri
sk 
int
er
val 
(d
ay
s) Post-vaccination control 
interval (days)  
TMaMultiple sclerosis (MS)10,30 IP only14or OP 1 year  98 through 15 1-
42 43-84N/A 
MS8,22Optic neuritis (ON)10,30  IP or OP8 1 year  98 through 15 1-
42 43-84N/A 
ON8,22Other acute demyelinating 
diseases10,30  IP or OP8 1 year  98 through 15 1-
42 43-84N/A 
Bell’s palsy8,22Transverse myelitis 
(TM)16  IP or OP1-ED 1 year 56 through 15 1-
42 43-84N/A 
Immunologic  
Anaphylaxis8,22 IP or OP1-ED16 6 months 1 
month16 N/A 0
20
-
116 797-810,30  
VasculitideseArthritis and 
arthralgia/joint pain0 IP onlyor OP  1 year  N/A 1-
28
42 43-842956 
Arthritis and arthralgia /joint 
paincAutoimmune thyroiditis0 IP or OP 1 year  N/A 1-
42 43-84 
MISAbFibromyalgia0 IP or OPIP only14 1 year  N/A 1-
42 43-84 
KD31Kawasaki disease (KD)36 IP only31 1 year  N/A 1-
28 29-56 Deleted Cells
Deleted Cells
Deleted Cells
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077586
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 52 of 237 Table 1. Outcome algorithms for SCRI analysis, with risk and control intervals  
Safety Event of Interest * Setting  
(Inpatient [IP], 
Outpatient [OP ], 
Emergency 
Department [ED ]) Clean 
window  Prevaccination 
control interval 
(days)  Ri
sk 
int
er
val 
(d
ay
s) Post-vaccination control 
interval (days)  
FibromyalgiacMultisystem 
inflammatory syndrome in adults 
(MIS-A)16  IP or OP-ED 1 year  N/A 1-
42 43-84 
Autoimmune 
thyroiditiscVasculitides0 IP or OPonly 1 year  N/A 1-
42
28 29-564384 
Cardiac 
Acute myocardial infarction (AMI)16  IP only 1 year  1-
28 29-56 
Arrhythmiac IP only 1 year  1-
42 43-84 
Coronary artery disease (CAD)c IP only 1 year  1-
42 43-84 
Heart failure and cardiogenic shockc IP only 1 year  1-
42 43-84 
Microangiopathy0 IP only 1 year  1-
28 29-56 
Myocarditis8,2216 IP or OP14 1 year  56 through 15 1-
42
d1
42 43-84N/A Deleted Cells
Deleted Cells
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077587
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 53 of 237 Table 1. Outcome algorithms for SCRI analysis, with risk and control intervals  
Safety Event of Interest * Setting  
(Inpatient [IP], 
Outpatient [OP ], 
Emergency 
Department [ED ]) Clean 
window  Prevaccination 
control interval 
(days)  Ri
sk 
int
er
val 
(d
ay
s) Post-vaccination control 
interval (days)  
Pericarditis8,2216 IP or OP14 1 year  56 through 15 1-
42
d1
42 N/A43-84 
AMIdStress cardiomyopathyc IP only14 1 year  56 through 15 1-
42 N/A43-84 
Hematologic  
Thrombocytopenia30 IP or OP14 1 year N/A 142 4384 
DICeCerebrovascular hemorrhagic 
stroke16 IP only14 1 year  N/A 1-
42
28 29-564384 
COVID 19 (for all COVID 19related safety events of interest listed below, an inpatient diagnosis of COVID 19 will be required in 
combination with the codes or laboratory values specified in Appendix Table 2; in addition, COVID 19 related safety events of 
interest will only be evaluated using data from 2020 onward using the SCRI design ): 
Severe COVID 19 
diseasebChillblain -like lesions0 IP onlyor OP 1 year  N/A 1-
42
28 29-564384 
MicroangiopathyeDisseminated 
intravascular coagulation (DIC)16  IP onlyor OP-ED 1 year  N/A 1-
42
28 29-564384 Deleted Cells
Deleted Cells
Deleted Cells
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077588
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 54 of 237 Table 1. Outcome algorithms for SCRI analysis, with risk and control intervals  
Safety Event of Interest * Setting  
(Inpatient [IP], 
Outpatient [OP ], 
Emergency 
Department [ED ]) Clean 
window  Prevaccination 
control interval 
(days)  Ri
sk 
int
er
val 
(d
ay
s) Post-vaccination control 
interval (days)  
Heart failure and cardiogenic 
shockdDeep vein thrombosis 
(DVT)16  IP onlyor OP 1 year  1-
28 29-56 
through 
15 142 N
/
A 
Stress cardiomyopathydHemolytic 
anemiae IP onlyor OP 1 year 56 through 15 1-
42 43-84N/A 
CADdHemorrhagic disease0 IP only 1 year  1-
28 29-56 
through 
15 142 N
/
A 
ArrhythmiadLimb ischemia0 IP only 1 year  1-
28 29-56 
through 
15 142 N
/
A 
DVTe IP or OP1 1 year N/A 142 4384 
Pulmonary emboluse (PE)16  IP or OP14OP 1 year  N/A 1-
42
28 29-564384 
Cerebrovascular hemorrhagic stroke8 IP only14 1 year N/A 142 4384 
Cerebrovascular non hemorrhagic 
stroke8 IP only1 1 year N/A 142 4384 
Limb ischemiaeSingle organ 
cutaneous vasculitis0 IP only 1 year  N/
A1
-
28 29-561
42 4384 Deleted Cells
Inserted Cells
Deleted CellsDeleted Cells
Deleted Cells
Inserted Cells
Deleted Cells
Deleted Cells
Deleted Cells
Deleted Cells
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077589
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 55 of 237 Table 1. Outcome algorithms for SCRI analysis, with risk and control intervals  
Safety Event of Interest * Setting  
(Inpatient [IP], 
Outpatient [OP ], 
Emergency 
Department [ED ]) Clean 
window  Prevaccination 
control interval 
(days)  Ri
sk 
int
er
val 
(d
ay
s) Post-vaccination control 
interval (days)  
Hemorrhagic diseasee IP only 1 year N/A 142 4384 
Acute kidney injuryg IP only 6 months  N/A 142 4384 
Liver injurygThrombocytopenia16 IP or OP 1 year  N/A 1-
42 43-84 
Chillblain like lesionseThrombosis 
with thrombocytopenia syndrome 
(TTS)e IP or OP 1 year N/A 1-
42 43-84 
Single organ cutaneous vasculitise IP only 1 year N/A 142 4384 
Erythema multiformef IP only 6 months  N/A 12 89 
Other 
Acute kidney injuryf IP only 6 months     1-
42 43-84 
Appendicitis16 IP or OP-ED  1 year  1-
42 43-84 
Death  IP or OP 1 year 0-
42 43-85 
Erythema multiformeg IP only 6 months     1-
2 8-9 
Liver injuryf IP or OP 1 year  1-
42 43-84 
Narcolepsy and cataplexya IP or OP1416 1 year16 98 through 15 1
42N/A43-84 Deleted Cells
Deleted Cells
Deleted Cells
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077590
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 56 of 237 Table 1. Outcome algorithms for SCRI analysis, with risk and control intervals  
Safety Event of Interest * Setting  
(Inpatient [IP], 
Outpatient [OP ], 
Emergency 
Department [ED ]) Clean 
window  Prevaccination 
control interval 
(days)  Ri
sk 
int
er
val 
(d
ay
s) Post-vaccination control 
interval (days)  
1-
42
16 
Non-anaphylactic allergic 
reactions810,2230  IP or OP8 6 months     N/A 1-
2 8-9 
Appendicitis32Severe COVID -19 
diseaseh IP only14 1 year 6 
months N/A 01
-
42 43-84 
Stevens-Johnson syndrome 
(SJS)/Toxic epidermal necrolysis 
(TEN)g IP only 6 months     1-
2 8-9 
*Safety events  of interest  are based on the Priority List of Adverse Events of Special Interest from the Brighton Collaboration’s Safety Platform for 
Emergency vACcines (SPEAC) Project, the FDA and the Centers for Disease Control and Prevention’s (CDC) Advisory Committee on Immunization 
Practices (ACIP) enhanced safety monitoring recommendations.  
Notes:  
a Published risk and control intervals for demyelinating diseases and cranial disorders were applied to TM and narcolepsy/catap lexy 
b Published setting, clean window, and As severe COVID 19 ranges from severe pneumonia, acute respiratory distress syndrome, and multisystem organ 
failure/MIS A, a 142 day risk interval was appli ed in order to capture the 14 day incubation period of the disease and 4 5 day period from exposure to 
symptom onset  
a. c Published  risk and control intervals for autoimmune disorders were applied to similar autoimmune rheumatic conditions (i.e., arthritis and arthralgia/joint 
pain, fibromyalgia and autoimmune thyroiditis).  
d Published setting, clean window, and risk and control intervals for myocarditis DVT, pulmonary embolus  and pericarditis DIC were applied to other  
cardiovascular  conditions (ie, heart failure and cardiogenic shock, stress cardiomyopathy, CAD, arrhythmia, AMI)  Deleted Cells
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077591
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 57 of 237 Table 1. Outcome algorithms for SCRI analysis, with risk and control intervals  
Safety Event of Interest * Setting  
(Inpatient [IP], 
Outpatient [OP ], 
Emergency 
Department [ED ]) Clean 
window  Prevaccination 
control interval 
(days)  Ri
sk 
int
er
val 
(d
ay
s) Post-vaccination control 
interval (days)  
b. e Similar risk and control intervals were applied to all  cardiovascular and hematological disorders characterized by damage to the blood vessels and/or 
arteries and clotting (i.e., microangiopathy, DVT, pulmonary embolus, limb ischemia, hemo rrhagic disease, DIC, chilblain-like lesions , single organ 
cutaneous vasculitis and vasculitides ). The published risk and control intervals for KD were applied to vasculitides given that KD is a type of medium and 
small-vessel vasculitis.  
c. f Published Published setting, cle an window, and risk and control intervals for myocarditis and pericarditis were applied to other cardiovascular conditions 
(i.e., heart failure and cardiogenic shock, stress cardiomyopathy, CAD, arrhythmia).  
d. For the prioritized safety analysis of myocardi tis/pericarditis, additional risk intervals (i.e., 1 -7 days and 1 -21 days) will be examined and are described in 
Section 9.7.8.  
e. Published setting, clean window and risk and control intervals for thrombocytopenia were applied to hem olytic anemia and TTS.  
f. Risk intervals of 42 days were applied for acute kidney injury and liver injury to be consistent with other similar safety ev ents of interest.  
c.g. Published setting, clean window, and  risk and control intervals for non -anaphylactic aller gic reactions were applied to hypersensitivity disorders (i.e., 
erythema multiforme  and SJS/TEN ).  
g Risk intervals of 42 days were applied for acute kidney injury and liver injury to be consistent with other COVID -19 related safety events  of interest . 
d.h. As severe COVID -19 ranges from severe pneumonia, acute respiratory distress syndrome, and multisystem organ failure/MIS -A, a 1-42 day risk 
interval was appli ed in order to capture the 14 -day incubation period of the disease and 4 -5 day period from exposure to symptom onset  
 Deleted Cells
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077592
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 58 of 237 9.4. Data Source  
The VHA  is the largest integrated health care system in the US, providing both inpatient and 
outpatient clinical care to over 9 million Veterans enrolled at more than 170 medical centers 
and 1,074 community -based outpatient clinics.1214 VHA’s health care delivery system is 
organized regionally around 18 Veterans Integrated Service Networks (VISNs) across the 
US. Each VISN is responsible for health care planning and resource allocation in a particular 
geographical region. For example, the VA New England Healthcare System (VISN 1) covers 
VHA facilities in Massachusetts, Connecticut, New Hampshire, Maine, and Rhode Island, 
while the VA Heart of Texas Health Care Network (VISN 17) oversees the facilities in 
Texas.  
The VHA also maintains i ts own mortality data where 99% of enrollees’ deaths are reported 
within one month  of occurrence . As of January 7, 2021, the VHA has had over 174,000  
confirmed COVID -19 cases.3337 Among active and convalescent cases , approximately 
145,000 are Veterans and approximately 15,000  are employees  (with an estimated 630 as 
Veteran employees ).33).33 While African American Veterans make up approximately 12% of 
the VHA,3438 the burden of COVID -19 cases are skewed, with African American Veterans 
comprising approximately  20% of all COVID -19 cases.3337 Approximately 7,099 COVID-
19-infected VA patients  have died, an estimated 2,738 in VHA hospitals.3337  
The objectives of this study will be addressed using data from VHA’s Corporate Data 
Warehouse ( CDW),, which is an integrated EMR system with a centralized data warehouse 
that is updated on a daily basis.  The CDW stores data in separate databases, one for each type 
of clinical information ( e.g., inpatient medication, inpatient admission, outpatient medication, 
outpatient visit). Individual  demographic information such as  date of birth and gender are 
also available. Immunization records include information on manufacturer, lot number, 
injection site, and concurrent immunizations. The CDW does not include information on any 
care received outside of a VHA facility.  
Each individual  is assigned a unique identif ication number to allow for longitudinal 
follow-up as well as to cross -reference to the various separate databases. For example, in 
each inpatient admission record, there is information on the primary discharge diagnosis (and 
as many as 15  secondary diagno ses), date of admission, date of discharge, and length of stay. 
This record can then be linked to other information of that inpatient stay located in other 
files, including procedures that the patient underwent during the hospitalization, medical 
specialty of the provider, and prescriptions dispensed. Other files are similarly structured, 
and therefore may be linked together to provide comprehensive information about the patient 
and his/her medical encounters.  
The VHA database is an appropriate data source  to evaluate the safety of the Pfizer -
BioNTech  COVID-19 vaccine for the following reasons. First, as the vaccine will be 
distributed through government facilities (including VHA) as part of initial distribution , 
analysis of VHA data will provide early data  on the safety of the vaccine.  Veterans living in 
long-term care facilities and Veteran s who are healthcare workers  will be prioritized in the 
first wave of Pfizer -BioNTech COVID -19 vaccinations.3539 The VA offers eligible Veterans 
long-term care services ranging from nursing homes and assisted -living centers to caregiver 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077593
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 59 of 237 support in the Veterans’ own homes.1315 Secondly, and relatedly, VH A data are refreshed 
daily and would thus enable early and rapid data analysis. Third, the VHA population is on 
average older than the general US population.3640 Of these, about 30% (roughly 
1,000,000  individuals) use VHA health se rvices almost exclusively ( i.e., those with a priority 
group of 1 or 4 ; Veterans assigned to Priority group 4 are either accepting VA assistance or 
housebound benefits, or have been determined to be “catastrophically disabled” by the 
VA.2733), which lends itself to having complete, longitudinal healthcare data for such 
individuals  who may be at higher risk of COVID -19 due to older age.37,3841,42 These priority 
groups include Veterans with the highest levels of service -connected disability and are 
therefore, the highest priority for VHA care.2733 Finally, the VHA population ha s, on average, 
more comorbid conditions than the general population, which also indicat es that these 
individuals  may be at higher risk of COVID -19.3943 While the VHA pop ulation is 
predominantly male (approximately 90%), and thus lacks generalizability to females, it will 
still provide a useful setting to examine re al-world vaccine safety.   
Since it is possible that individuals may not have all their health encounters within the VHA, 
(especially older veterans who are also covered by Medicare ), additional subgroup analyses 
will be conducted in which the CDW  data will be supplemented with data from CMS, linking 
Medicare administrative claims data at the patient level to ensure a more comprehensive 
evaluation of the care an individual receives.  Medicare data will include eligibility files and 
claims for servic es received in the inpatient and outpatient setting, as well as skilled nursing 
facilities, hospice, and home health agencies, and will cover the US primarily among those 
aged 65 years or older.  
9.5. Study Size  
The sample size achieved will depend on the numbe r of recipients of Pfizer -BioNTech 
COVID-19 vaccine within the VHA database  during the study period , which will increase 
over time with subsequent analyses. The population size will increase with each bi -weekly 
analysis as the Pfizer -BioNTech COVID -19 vaccine becomes more readily available and a 
greater number of individuals are vaccinated. Specifically, the data will be refreshed on a 
biweekly basis and a continuous sequential test procedure will be used to reevaluate data 
according to this schedule. As of January 21, 2021, 112,201 doses of Pfizer -BioNTech 
COVID-19 vaccine have been administered within the VHA (based on CPT code 91300) to a 
total of 107,458 patients.  
As a result of the ability to perform near-real-time analysis, the risk interval (and post -
vaccination control interval, for applicable safety events of interest ) may have only partially 
elapsed in some cases. To account for this, we will use methods adopted in previous 
studies,810,2544,045 whereby risk intervals will be scaled (or truncated) in order to ensure an 
equivalent length (or a fixed ratio) of time is assessed between the control and risk intervals.  
9.5.1.  Power  
Power calculations for the rapid cycle analysis (RC A) approaches proposed for safety event 
of interest  signal detection will be conducted according to the methods of Kulldorff et 
al.41,4246,47 Table 2 illustrates the estimated power for the RCA approach using the Poisson -
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077594
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 60 of 237 based maximized sequential probability ratio test (MaxSPRT), and provides an overview of 
the power required to detect varying relative risk (RR ) estimates with an alpha level of 0.01. 
T denotes the expected number of safety events of interest to occur during the risk interval of 
interest (Table 2 and Table 3). Power of  ≥ 80% is typically desirable in drug safety research. 
Usually the FDA views a RR of >  3 as meaningful, so this has been used for power 
calculations here. 4348 As an example, as shown in Table 2, the surveillance system would 
have sufficient power (80.0%) to detect an increased risk of safety e vents of interest 
associated with the Pfizer -BioNTech COVID -19 vaccine by 3 fold when the expected 
number of safety events of interest  reaches 6 events.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077595
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 61 of 237 Table  2. Estimated Statistical Power for the Poisson -based MaxSPRT1MaxSPRT46 
 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077596
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 62 of 237 9.6. Data Management  
Data for this study wil l be stored and extracted from the VHA database (p reviously described 
in Section 9.4) that contain information about patient demographics, vaccinations, 
procedures, diagnoses, and death.   
9.6.1.  Case report forms (CRFs)/Electronic data record  
As used in this protocol, the term CRF should be understood to refer to either a paper form or 
an electronic data record or both, depending on the data collection method used in this study.  
A CRF is required and should be completed for each included patient in the signal 
verification phase that  requires EMR and chart review ( see Section 9.7.3.3). The completed 
original CRFs should not be made available in any form to third parties, except for 
authorized rep resentatives of Pfizer or appropriate regulatory authorities, without written 
permission from Pfizer. The CRF will consist of two parts: (1) a database CRF that will be 
populated based on a direct extraction of data from the VA CDW for review by the 
adjudicators; (2) an adjudication page that will be completed by an adjudicator after 
reviewing data in the completed CRFs.  Analysis Group shall ensure that the CRFs are 
securely stored on VHA servers in an encrypted electronic and/or paper]  form and will be 
password protec ted or secured in a locked room to prevent access by unauthorized third 
parties. 
Analysis Group has ultimate responsibility for the collection and reporting of all clinical, 
safety, and laboratory data entered on the database CRFs and any other  data collection forms 
(source documents) and ensuring that they are accurate, authentic/original, attributable, 
complete, consistent, legible, timely (contemporaneous), enduring, and available when 
required.  The adjudication page  must be signed by the ad judication committee members to 
attest that the data contained on the forms  are true and accurate based on their review of the 
EMR data . Any corrections to entries made in the CRFs or source documents must be dated, 
initialed, and explained (if necessary) and should not obscure the original entry.  
The source documents are the hospital or the physician's chart. In these cases, data collected 
on the CRFs must match those charts.  
9.6.2.  Record retention  
To enable evaluations and/or inspections/audits from regulatory  authorities or Pfizer, 
Analysis Group agrees to keep all study -related records, which includes study documents and 
deliverables such as the protocol, SAP, aggregated results tables, SAS Institute (SAS) 
programming files , and study report. The records shou ld be retained by Analysis Group 
according to local regulations or as specified in the vendor contract, whichever is longer. 
Analysis Group must ensure that the records continue to be stored securely for so long as 
they are retained.  
If Analysis Group beco mes unable for any reason to continue to retain study records for the 
required period, Pfizer should be prospectively notified. The study records must be 
transferred to a designee acceptable to Pfizer.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077597
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 63 of 237 Study records must be kept for a minimum of 15 years a fter completion or discontinuation of 
the study, unless Analysis Group and Pfizer have expressly agreed to a different period of 
retention via a separate written agreement. Record must be retained for longer than 15 years 
if required by applicable local re gulations.   
Analysis Group must obtain Pfizer's written permission before disposing of any records, even 
if retention requirements have been met.  
9.7. Data Analysis  
Detailed methodology for summary and statistical analyses of data  analyzed in this study will 
be documented in a statistical analysis plan ( SAP),, which will be dated, filed , and 
maintained by the sponsor. The SAP may modify the plans outlined in the protocol; any 
major modifications of primary endpoint definitions or their analyses would be ref lected in a 
protocol amendment. T he SAP will also provide additional detail regarding the evaluation of 
a threshold of excess risk for each of the safety events of interest. Consistent with the 
approach of Kulldorff et al ., this will be determined based on background incidences for each 
event (e.g., based on historical influenza vaccinated active comparator cohort data to be 
evaluated during the study), in addition to pre -specified significance level ( e.g., alpha = 0.01 
or 0.05) and power.4146 This information, in conjunction with a clinically meaningful RR 
(e.g., 2 or 3) and the expected upper limit of events under the null hypothesis will allow for 
the calculation of critical values of each safety event of in terest using the MaxSPRT method. 
Greater power ( e.g., 80%) is also a natural criterion to use when selecting the upper limit on 
the length of surveillance, and in turn, the expected number of events to occur, although there 
is ultimately a tradeoff between  that power and the time allowed to identify the expected 
number of events to occur.  
 
Data analyses will be conducted using SAS Enterprise Guide version 7.1 (SAS Institute Inc., 
Cary, NC) or R Version 3.5.3 or its latest version (R Core Team, Vienna, Austr ia). In 
addition, SaTScan will also be used to conduct specific temporal analyses.  
9.7.1.  Baseline Characteristics  
Baseline demographics and clinical characteristics for individuals receiving Pfizer -BioNTech 
COVID-19 vaccine and individuals who received seasonal influenza vaccination will be 
summarized using descriptive statistics, consisting of the mean and standard deviation (SD) 
and median (interquartile range [IQR]) values for continuous variables and frequency 
distributions for categorical variables.  Incidence rates (i.e., per-patient per -month) for prior 
hospitalizations may be calculated as the number of events divided by person -time of 
observation since the length of the baseline period may vary between individuals. 
Standardized differences will be calculat ed between Pfizer  BioNTech COVID -19 vaccine 
recipients and active comparators who received seasonal influenza vaccination to evaluate 
whether there are any major differences in individuals’ baseline characteristics.  Standardized 
differences <  10% will indi cate that matching  has appropriately balanced the charac teristics 
between recipients of the Pfizer -BioNTech COVID -19 vaccine and seasonal influenza 
vaccine. 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077598
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 64 of 237 9.7.2.  Vaccine Utilization Patterns  
Descriptive statistics will also be used to summarize vaccine utilizat ion patterns, including 
proportion of individuals receiving vaccine, 2 -dose completion rate , distribution of time gaps 
between the first and second dose , and care setting where immunization was received ( e.g., 
outpatient clinic, pharmacy, inpatient ward).  Counts of individuals who received a COVID -
19 vaccine from a different manufacturer in addition to the Pfizer -BioNTech COVID -19 
vaccine will be reportedsummarized . 
9.7.3.  Safety Signal Analyses  
Several analyses corresponding to the designs discussed previously will be co nducted to 
detect safety signals associated with Pfizer -BioNTech COVID -19 vaccine. Analyses will be 
conducted among all individuals receiving the vaccine, individuals who received 
Pfizer-BioNTech  COVID-19 vaccine without seasonal flu vaccine (Cohort A will be used for 
SCRI; Cohort B +C will be used for active comparator  analyses), and individuals receiving 
Pfizer-BioNTech COVID -19 vaccine and seasonal flu vaccine on the same day (Cohort D), 
along with sub-cohorts receiving only one dose vs. two doses.  
A stepwise process, illustrated below, will be performed for signal detection, evaluation, and 
verification ( Figure 3). This approach has been adapted from the Active Monitoring Protocol 
of the FDA’s COVID -19 Vaccine Safety Surveillance Project.1416 The statistical approach 
described below may be modified further based on data availability, additional clinical input, 
and for consistency or to complement similar studies of Pfizer -BioNTech COVID -19 
vaccine. 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077599
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 65 of 237 Figure 3. Steps in Signal Detection, Evaluation , and Verification  
 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077600
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 66 of 237  
Notes:  
[1] List of safety events of interest and corresponding definitions may be refined as the study progresses based 
on additional available information.  
[2] The risk and control intervals selected for the SCRI analysis for each safety event of interest are based on 
biological plausibility and precedents in the literature. Only the individual’s first instance during the specified 
clean window ( i.e., the interval used to define incident outcomes) will be included.  Note that only the first 
inpatient or outpatient occurrence  of a safety event of interest following the clean window will be used to 
identify incident events ( e.g., if an inpatient safety event of interest occurs in the clean window, a repeat 
occurrence will not be counted in the risk interval). However, event wor sening will be counted as a safety event 
of interest. For example, if an outpatient safety event of interest occurs in the clean window and an inpatient  
occurrence for the same type of safety event of interest occurs in the risk interval, the inpatient occ urrence will 
be counted as a safety event of interest.  
9.7.3.1.  Signal Detection  
Signal detection will rely on SCRI design with comparison to post-vaccination control 
intervals for the two safety events that require COVID -19 diagnosis (i.e., severe COVID -19 
disease, MIS-A) and active comparator design for the remaining safety events. While the 
active comparator design will be the main analysis for signal detection because it can be 
performed the fastest, it cannot be used for safety events that require COVID -19 diagnosis 
because historical controls would not meet the criteria of having a COVID -19 diagnosis.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077601
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 67 of 237 9.7.3.1.1.  Sequential Testing - SCRI Design using the Binomial -based MaxSPRT  for 
Comparison to PrePost-vaccination Control Intervals  For the Two Safety Events 
Requiring COVID -19 Diagnosis  
The goal is to provide rapid -cycle, near real -time safety surveillance. In the signal detection 
phase, the SCRI analysis will only include pre vaccination control intervals as the with post-
vaccination control intervals will require a longer time to accumulate and thus will not 
allowbe used for timely analysis. certain safety events of interest ( i.e., severe COVID -19 
disease, MIS-A). All other safety events of interest will be assessed in the signal detection 
phase using the active comparator design . The post-vaccination control period will be 
assessed during the signal evaluation phase (see Section 9.7.3.2), to allow for additional 
observation time to accrue as well as to more deeply investigate potential signals. This will 
allow for timely RCA without the need to wait for data to accumulate for  safety events of 
interest with post vaccination control intervals once enough post -vaccination time has 
accumulated .  
To account for multiple testing and bi -weekly review of the data, the MaxSPRT using a 
binomial probability model will be applied. The null hypothesis (H 0) assumes that the risk of 
a safety event of interest  during the risk interval is equivalent to the risk of the same safety 
event of interest developing during the control interval, accounting for differences in interval 
duration as needed ( e.g., for safety events of interest  such as demyelinating disease), meaning 
a RR of 1 is specified under H 0.2230 The one-sided composite alternative hypothesis (H a) 
assumes that the risk of a safety event of interest  during the risk interval is greater than the 
risk of the same  safety event  of interest  developing during the control interval, accounting for 
differences in interval duration ( i.e., RR > 1, Ha is applicable across a range of RRs).146  
Specifically, for the Pfizer -BioNTech COVID -19 vaccine, let x represent the total count of 
safety events of interest in the control in terval (Figure 4), let y represent the total count of 
safety events of interest in the risk interval, and let r represent the ratio of y to x under the 
null hypothesis. Thus, when the total control interval du ration and total risk interval duration 
are equal, r will be 1. The RR is estimated by 𝑦𝑟
𝑥.2544 The RR and corresponding 99% 
confidence intervals (CIs) will be calculated.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077602
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 68 of 237 Figure 4. Example of SCRI Design for a Safety Event of Interest with a 42-day Risk 
Interval and a PrePost-vaccination Control Interval  
 
For the binomial model, the log -likelihood ratio (LLR ) is calculated as the log probability of 
observing this distribution of y under H a, divided by the probability of this occurring under 
H0.146 This ratio is calculated whenever new data are received to account for the conti nuous 
data stream until the full 42 -day risk period is complete.  
𝐿𝐿𝑅 =ln𝑃(𝑦 | 𝐻𝑎)
𝑃(𝑦 | 𝐻0) 
Once the LLR test statistic reaches a pre -specified critical value , a signal is detected.  
Specifically, the null hypothesis will be rejected if the LLR exceeds  the critical value. The 
null hypothesis will not be rejected if the LLR does not reach or exceed the critical value, if 
the total number of safety events of interest  reaches a pre -specified upper limit, or if 
surveillance ends without reaching this upper limit.2544   
For each safety event of interest (and specific to each age group, if age -stratified analyses are 
conducted), the critical value of the LLR will be determined based on the safety event of 
interest specific upper limit of expected safety events of interest and alpha level.2544 Upper 
limits will be determined based on the expected number of safety events of interest under the 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077603
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 69 of 237 null hypothesis, assuming the risk after Pfizer -BioNTech COVID -19 vaccination is no 
greater than the risk of safety events of interest after seasonal influenza vaccination. 
Therefore, upper limits will be chosen such that they would not usually be reached.  
9.7.3.1.2.  Sequential Testing  - Poisson -based MaxSPRT  for Comparison to Active 
Comparators who Received Seasonal Influenza Vaccination   
For compariso n with active comparators who received seasonal influenza vaccination, the 
Poisson-based MaxSPRT will be applied, following the same statistical approach as 
described above, but using a Poisson probability distribution. In the Poisson MaxSPRT 
approach, the  event frequency of safety events of interest in the risk interval after Pfizer -
BioNTech COVID -19 vaccination will be compared to a background  rate of safety events of 
interest in the risk interval after seasonal influenza vaccination in five prior seasons , ranging 
from 2014/15 through 2018/19. This approach is particularly important for extremely rare 
safety events of interest  (i.e., less than 50 anticipated based on historical influenza vaccine 
rates of safety events of interest ).2230 Poisson MaxSPRT is used to monitor very rare safety 
events of interest  as binomial MaxSPRT may not detect a signal, despite a clinically 
meaningful RR.2544 This will also allow for more timely analysis using historical da ta, as well 
as improved power and sample size.  
GBS is of particular interest relative to  the safety profile of Pfizer -BioNTech 
COVID-19 vaccine. As GBS is an extremely rare safety event of interest , the primary RCA 
proposed will focus on Poisson MaxSPRT and apply an alpha of 0.05. The Poisson 
MaxSPRT has increased power to detect a signal with fewer occurrences of the safety event 
of interest. However, this method cannot fully control for confoundin g by indication.  
9.7.3.1.3.  Critical Values and Alpha Spending  
Critical values for the LLR test statistic are shown below in Table 3 based on calculations 
conducted by Kulldorff et al 2011.4146 For example, assuming T  = 6 (number of expected 
events under the null) and RR  = 3, which corresponds to a power of 80.0% (See 
Section 9.5.1), the critical value would be 5.14 using alpha of 0.01 for the Poisson -based 
MaxSPRT. As noted previously, each safety event of interest will be evaluated separately to 
determine a critical value based on background incidence, alpha, power, and clinically 
meaningful RR. These details will be addressed in the SAP.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077604
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 70 of 237 Table 3. Critical Values for Poisson -based MaxSPRT  
 
Multiple types of alpha spending fun ctions can be employed to calculate the cumulative rate 
at which Type 1 error (alpha) probability is spent during sequential testing.4449 To achieve 
optimal expected time -to-signal, especially when historical Poisson data are used with 
surveillance data, a power-type convex alpha spending shape will be used based on published 
literature.49 Additionally, ρ = 1.5 is referenced as a “rule of thumb” as it i s suggested to be 
appropriate in most applications.  
9.7.3.2.  Signal Evaluation  
Signals are detected when the event frequency of a  safety event  of interest during the risk 
interval following vaccination with Pfizer -BioNTech COVID -19 vaccine is significantly 
increased compared to the event frequency of the same safety events of interest in the control 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077605
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 71 of 237 comparator  (i.e., the critical value is achieved and surpassed). If signals are indeed detected 
for safety events of interest based on the analysis described above, furt her evaluation is 
warranted to refine and confirm such detections.  This will includeconsist of  the following 
additional analyses to assess the robustness of  the findings describe in the following sections, 
which will be conducted every six months .  
9.7.3.2.1.  Post-Signal Quality Assurance  
Quality assurance will first be con ducted in order to assess the quality of the data and 
analysis that produced the signal. While quality control measures  will be conducted during 
the signal detection phase  (see Section 9.8), post-signal quality assurance will also be 
performed during the signal evaluation phase. This will include a comprehensive quality 
assurance (for example, check for possible duplications of claims or medical record s, 
checking for unusual clustering in claim or medical record accrual by service date for 
potential coding issues, check for geographical distribution of cases that may be related to lot 
numbers or diagnostic practice) . In addition, for signals detected via active comparison, 
additional analyses comparing to prepost-vaccination control intervals may be formed to 
check for consistency. Signals will also be confirmed across all of the safety studies planned 
to be performed (i.e., C4591008, C4591011, C 4591012) to confirm that specific data sources 
are not biased.  
9.7.3.2.2.  Multivariate Adjustment using Poisson Regression  
If signals are detected and persist after conducting quality assurance, further evaluation via 
statistical measures are warranted. Specifically , to investigate whether potential signals 
identified via Poisson MaxSPRT for the comparison to active comparators with seasonal 
influenza vaccination are not confounded ( i.e., to take into account baseline differences 
between the Pfizer BioNTech COVID -19 vaccinated and active comparator populations), a 
multivariate Poisson regression analysis will be conducted to compare the incidence rates of 
the safety events of interest occurring within the risk intervals. T he predictor would be 
whether the individual had received the Pfizer -BioNTech COVID -19 vaccine or had received 
the influenza vaccine during historical seasons. Analyses will be adjusted for relevant 
baseline and/or clinical characteristics ( e.g., age, sex, race, CCI and/or specific comorbidities 
of interest, state, etc.).8 10 
If the signal remains, based on an IRR >  3 with a p -value < 0.01 from the adjusted Poisson 
regression, further evaluation may be considered via signal verification.  
9.7.3.2.3.  Assessment of Temporal Clusters  
Vaccine safety surveillance must allow for sufficient type I error probability for rapid 
detection  of safety events of interest , and statistically significant signals must be studied 
further to ensure that a true association is present.4550 Therefore, the presence of temporal 
clusters will be assessed using the software SaTScan to calculate temporal scan statistic in 
order to further refine safety signals detected from the signal detection analyses.2230 A 
temporal scan statistic accounts for multiple testing present during overlapping risk intervals. 
The null hypothesis assumes that there is no association between the safety events  of interest 
and immunization, and safety events of interest  are assumed to be di stributed independently 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077606
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 72 of 237 and uniformly during a period of time subsequent to Pfizer -BioNTech COVID -19 
vaccination.2230  A temporal scan statistic will be generated by moving a time interval of 
fixed length across the risk interval, comparing the number of observed versus expected 
safety events of interest  within the time interval under the null hypothesis.646  
9.7.3.2.4.  Sequential Testing - SCRI Design using the Binomial MaxSPRT for 
Comparison with Post -Vaccination Control Intervals  
Similar to the SCRI design Any safety events of interest with signals detected and not 
already analyzed during the signal detection phase with the SCRI design using the binomial -
based MaxSPRT will be analyzed during the signal evaluation phase using SCRI design 
using the binomial -based MaxSPRT method for prevaccination control intervals, sequential 
testing analyses will be conducted using the post-vaccination control intervals  as appropriate 
for specific safety events of interest . This will be conducted during the signal evaluation 
phase in order to allow time to accumulate during the post -vaccination control period. The 
same statistical methodology as described for the pre above will be applied.  
9.7.3.2.5.  SCCS Design using Conditional Poisson Regression for Comparison with 
Post-Vaccination Control Time Period  
Similar to the SCRI design with post-vaccination control intervals, SCCS design with post -
vaccination control time period will include cases (i.e., individuals vaccinated with the 
Pfizer-BioNTech COVID -19  vaccine who experience safety events of interest following 
vaccination) to compare the  incidence of safety events occurring in the risk interval 
following vaccination with the incidence of safety events occurring during all other times 
post-vaccination  in the same individual  until the earliest of 183 days af ter the Pfizer-
BioNTech COVID -19 vaccination, disenrollment, death, end of data availability . This 
analysis will be conducted for all safety events of interest with signals detected  in the signal 
detection phase . The SCCS design differs from the SCRI desig n in that instead of having 
fixed post-vaccination control intervals will be applied. of the same duration as the risk 
interval, it has a time-varying post-vaccination control time period that include s all-non risk 
interval time from Pfizer-BioNTech COVID-19 vaccination date until the earliest of 183 
days after Pfizer-BioNTech COVID-19 vaccination , disenrollment, death, end of data 
availabilit y.23 
For individuals who receive two doses of the vaccine, the post-vaccination control time 
period may include time before and after Pfizer-BioNTech COVID -19 vaccine dose 2 or 
solely include time after Pfizer-BioNTech COVID -19 vaccine dose 2. See Figure 5 below for 
an example of an individual who receives two doses of Pfizer -BioNTech COVID -19 vaccine , 
where the safety event of interest has a 42 -day risk interval window (e.g., Bell’s palsy ; Table 
1 in Section 9.3.3 ). Figure 5A demonstrates the SCCS design with the second dose received 
21 days after the first  (i.e., the risk interval for dose 1 overlaps with the risk interval for dose 
2), while Figure 5B demonstrates the SCCS design with the second dose received 60 days 
after the first (i.e., with gaps between the end of dose 1 risk interval and dose 2 ). The post-
vaccination control time period is  displayed below as shading with gray lines . 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077607
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 73 of 237 Figure 5. Example of SC CS Design  for Safety Event of Interest  with a 42 -day Risk 
Interval  with Post -vaccination Control Intervals when Two Doses of Pfizer -
BioNTech COVID -19 Vaccine are Administered  
 
 
Compared to the SCRI design, the SCCS design with post -vaccination control time period 
will have increased statistical power, which  is especially useful for the study of rare safety 
events of interest.  A conditional Poisson regression model will be used to compare the rates 
of safety even ts of interest in the risk interval vs post -vaccination control time period.  From 
this model we will report rate ratios and 95% CIs that will be interpretated as the rate ratio 
for the safety event of interest  in the risk interval compared to the control interval. 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077608
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 74 of 237 9.7.3.2.6.  Comparison with Contemporary Unvaccinated Controls  
To address period effects that could impact the appropriateness of using the historical 
comparator cohort, a nalyses will also be performed comparing individuals who received the 
Pfizer-BioNTech CO VID-19 vaccine to individuals who were not vaccinated at that point in 
time. The unvaccinated controls will be assigned an index date matched to a corresponding 
Pfizer-BioNTech COVID -19 vaccinee’s vaccination date ; these individuals can later receive 
the Pfizer-BioNTech COVID -19 vaccine and enter the vaccination group if all inclusion and 
exclusion criteria are met . To address possible selection bias due to health seeking behaviors, 
the unvaccinated controls will be selected from a population of patients wh o have regular use 
of VHA medical care, defined as at least two outpatient (excluding ED, as ED visits may not 
be considered regular) or inpatient encounters in the one year prior to vaccination. The 
encounters must be separated by > 30 days (for inpatient , by admission date), and at least one 
must be within six months prior to index date.  This approach is consistent with the Center for 
Biologics Evaluation and Research (CBER) Surveillance Program, Draft Master Protocol 
Assessment of Risk of Safety Outcomes  Following COVID -19 Vaccination.23  
Inverse probability treatment weighting (IPTW)  will be used to ensure comparability 
between the Pfizer -BioNTech  COVID-19 vaccinated cohort and contemporary unvaccinated 
controls. The IPTW approach uses weights to create a “pseudo -population” in which the 
distribution of covariates is on average  the same in each cohort.51 IPTW is defin ed as the 
inverse of the individual’s probability of receiving the  first dose of  Pfizer-BioNTech 
COVID-19 vaccine, conditional on their demographic and clinical characteristics. This 
approach assumes that an individual’s probability of receiving Pfizer -BioNTech COVID -19 
vaccination is constant for the first and second doses of the vaccine , as the weight will be 
applied for both doses .23 Initial inverse probability weights will be calculated as 1 / 
propensity score (PS) for individuals who received the Pfizer -BioNTech COVID -19 vaccine 
and 1/ (1-PS) for individuals with no record of COVID -19 vaccination. To avoid extreme 
weights, each individual’s weight will be stabilized by the marginal probability of being in 
their assigned  cohort. Therefore, the stabilized weights will be calculated as Pr (Pfizer -
BioNTech COVID -19 = 1) / PS for individuals who received the Pfizer-BioNTech COVID -
19 vaccine and 1 - Pr (Pfizer-BioNTech COVID -19 = 1) / (1 -PS) for the contemporary 
unvaccinated c ontrols. The distribution of weights will be examined to check for extreme 
values, and truncation will be considered if necessary.  
Weighted Cox regression with robust standard errors to account for within -subject 
correlation will be conducted to compare t he risk of safety events of interest between cohorts. 
Hazard ratios and corresponding 95% CIs will be summarized . 
9.7.3.3.  Signal Verification  
If a signal persists after conducting signal evaluation , signal verification through medical 
records review may be conducted . 
9.7.3.3.1.  Medical Records Review  
As part of the signal evaluation process, diagnostic validation of the detected safety events of 
interest (i.e., cases) via a djudication  of patient medical records by VHA clinicians for 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077609
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 75 of 237 outcome verification in a representative sample of cases will be conducted. The total number 
of charts to be reviewed will depend on the number of safety events of inter est detected, such 
that all cases may be reviewed for safety events of interest where a small number of events 
result in signal detection and a representative sub -sample may be reviewed for safety events 
of interest  where a larger number of events results in signal detection .4752 For rare events, 
potentially all cases may be adjudicated. An adjudication charter will be developed to govern 
signal evaluation and medical records review. Specifically, validation of detected safety 
events of interest will be performed through patient medical chart review in collaboration 
with an adjudication committee comprised of  the treating or trained healthcare 
professionals.7 52 
9.7.4.  Seasonality -Adjusted Cases -Centered Method  
A case-centered analysis for specific safety events of interest  for which signals were detected 
may also be conducted in order to a ccount for bias caused by seasonality of safety events of 
interest and vaccination.23 3053 
This method will use data on all safety event of interest  cases that occur after vaccination 
with Pfizer -BioNTech COVID -19 vaccine. Logistic regression will be used to compare the 
number of safety event of interest  cases that were vaccinated inside versus outside a pre -
specified risk interval , as of the date of the safety events,  where the total number of 
vaccinations given inside versus outside the risk interval (in the population of all vaccinees) 
is used as the offset term.2544 Specifically, the association of vaccination with risk of safety 
events of interest will be estimated from  a logistic regression model that includes 
summarized data with one record per risk set. The key independent variable will be the 
proportion of the risk set who were in the risk interval on the date of the safety event  of 
interest occurrence. In this way, risk sets are anchored to calendar dates, and confounding by 
seasonality of the safety events  of interest  and vaccination is addressed .853 Note that other 
confounders may also be adjusted for by restri cting risk sets to vaccinees similar with respect 
to select characteristics ( i.e., through stratification).  
9.7.5.  End-of-Season and End -of-Surveillance Analyses  
For any safety event of interest with signals detected, end -of-season analyses (over the course 
of the 30-month period) and an end -of-surveillance analysis ( i.e., at 30 months , after the end 
of surveillance) will be conducted. Similar methodology will be applied for the end -of-
surveillance analysis and end -of-season analysis conducted for seasonal influe nza vaccine in 
order to adjust for the seasonality of both disease and vaccine administration.810 This 
approach will be able to define the true risk intervals after each dose and estimate the risk for 
potential safety events of i nterest after both dose 1 and 2 of the Pfizer -BioNTech COVID -19 
vaccine, as well as the ability to discern whether or not one or two doses of seasonal 
influenza vaccine were administered during the same period.   
The number of events in the sum of three di stinct risk intervals will be compared to the 
control interval, adjusting for potential differences in interval length, to estimate the RR of 
Pfizer-BioNTech COVID -19 vaccine  compared to the influenza vaccine . In order to monitor 
the safety after the first  and full course of the vaccine, the number of potential safety events 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077610
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 76 of 237 of interest occurring in three separate risk intervals (P 1, P2, P3) will be estimated ( Figure 
5).Figure 6). P1 represents the risk interval after the first dose only, excluding any overlap in 
risk intervals with the second dose. P 2 represents the overlapping risk intervals for first and 
second dose of the vaccine. P 3 represents the risk interval of the second do se of the vaccine, 
excluding the overlapping risk interval already captured in P 2. This design will allow for the 
assessment of risk during the appropriate periods, regardless of the time interval between 
vaccine doses. As multiple endpoints will be assessed, 99% CIs will be calculated around the 
RR in order to ascertain whether the Pf izer-BioNTech COVID -19 vaccine is associated with 
safety events of interest.  
Figure 6. Example of Risk (P1, P2, P3) and PreAggregate Post-vaccination Control 
Intervals for the SCRI End -of-surveillance Analyses of 1 or 2 Doses of  
Pfizer-BioNTech COVID -19 Vaccine  
 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077611
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 77 of 237  
 
In Figure 6A, P1 + P2 + P3 represent the risk intervals where a safety event of interest may 
occur. In Figure 6B, there is no overlapping risk interval so that P 1 + P3 represent the risk 
intervals where a  safety event of interest  may occur.  The timing of the risk and control 
intervals may be adjusted for in order to  control for the effect of seasonality across the 
intervals assessed.  
9.7.6.  Subgroup Analysis  
Separate analyses of baseline characteristics, vaccine utilization patterns, signal detection, 
signal evaluation, and signal verification  in subgroups of interest may be conducted based on 
feasibility, sample size, and data available.  
9.7.7.  Incidence Rates and Time to Safety Event of Interest  Analysis  
Incidence rates (and corresponding CIs) will be calculated from safety event of interest  signal 
detection analyses. Kaplan -Meier methods will be used to analyze time -to-event (i.e., time to 
safety event of interest ). If individuals do not experience the  safety events of interest , they 
will be censored at the end of the risk interval. Median time to safety event of interest and 
corresponding CI s will be reportedsummarized . 
9.7.8.  Prioritized Safety Analysis of Myocarditis /Pericardi tis 
Notably, CDC recently investigated the occurrence of myocarditis/pericarditis following 
mRNA COVID -19 vaccinations.17 Therefore, separate safety analyses will be prioritized and 
performed to assess the risk of myocar ditis/pericarditis following Pfizer -BioNTech COVID -
19 vaccination, to provide additional context to the CDC investigation and address regulatory 
requests for further information on this safety event . Therefore, separate analyses will be 
prioritized and conducted to better understand the risk of myocarditis/pericarditis following 
Pfizer-BioNTech COVID -19 vaccination in the VHA. This analytical approach is intended to 
align with the methodology used by  the Vaccine Safety Datalink  (VSD) and preliminary 
findings of myocarditis/pericarditis published by ACIP on June 23, 2021.17,18  The VSD 
protocol defines myocarditis/pericarditis (ICD-10-CM codes B33.22, B33.23, I30, I40) 
events as the first ev ent in 60 days identified through an ED or inpatient encounter, without a 
first diagnosis of COVID -19 (i.e., COVID -19 diagnosis code or positive COVID -19 lab test) 
in the 30 days prior to or on the day of the event.  This analysis will follow the outcome 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077612
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 78 of 237 definition used in the VSD and uses three distinct risk intervals following vaccination (i.e., 1 -
7 days, 1-21 days, and 1 -42 days). This definition and the statistical approach differ from the 
primary analysis described in this protocol , but will facilitate comparison with the results 
presented by ACIP.16,17 
This analysis will include all individuals in the primary analysis who were vaccinated with 
the Pfizer -BioNTech COVID -19 vaccine. The n umber of myocarditis/pericarditis events in 
the risk interval will be identified, and incidence rates per million doses will be summarized . 
Subgroup analyses will also be performed, stratified by age ( e.g., 12-39 years, 40 -49 years, 
50-64 years, 65+ years ), gender, and race/ethnicity, respectively.  
In addition, vaccinated concurrent comparators will be selected among individuals who 
received the Pfizer -BioNTech COVID -19 vaccine, and then events will be compared 
between vaccinees who are in their risk interv al and vaccinees who are concurrently, on the 
same calendar date, in their comparison interval. Poisson regression will then be used to 
calculate incidence rate ratios and 95% CIs to compare the rate of myocarditis/pericarditis 
events between those individ uals who were in a risk interval versus those individuals who 
were in a comparison interval on the same calendar day. Data will be analyzed at the stratum 
level for each calendar day and will include strata for the independent variable of interest 
(i.e., risk vs. comparison interval) and for adjustment variables (i.e., age group, sex, 
race/ethnicity, and VHA service area). Thus , the number of myocarditis/pericarditis events in 
a risk or comparison interval on a calendar day will be modeled as a function of whether the 
stratum’s vaccinees are in a risk versus comparison interval on that calendar day, controlling 
for age, sex, race/ethnicity, and VHA service area. The log of the number of individuals 
contributing data to each stratum on each calendar day will be included as an offset term in 
the Poisson model.  Additionally, if it is suggested that calendar time may be associated with 
risk of post -vaccination myocarditis/pericarditis, to account for changes COVID -19 and other 
viruses circulating and other ecologic factors, analyses may also be stratified by calendar 
time, for example in 6 months increments.  
In addition to analyzing codified data, Case confirmation for myocarditis/pericarditis events 
identified in the codified data will be conducted based on medical chart review. 
Myocarditis/pericarditis cases will be confirmed and validated using the Brighton 
Collaborati on’s case definitions .19 Risk factor analysis will also be conducted via logistic 
regression among confirmed cases of myocarditis/pericarditis to further evaluate variables 
associated with the event ; additional detail s will be provided in the SAP . 
Additional data surrounding risk factors, clinical course, and sequelae of identified 
myocarditis/pericarditis events up to 365 days following the event will be collected and 
summarized . These will include an examination of other possible etiologies/risk factors (i.e., 
prior COVID -19 infection, prior Coxsackie infection, other prior viral infections, other 
vaccines received, comorbid immunocompromising conditions and systemic immune -
mediated diseases, demographics, and medicat ion history); time between Pfizer -BioNTech 
COVID-19 dose (first and second) and onset of myocarditis/pericarditis; echocardiogram 
information; lab troponin information; symptoms (e.g., chest pain, shortness of breath, 
weakness or fatigue, arm or shoulder p ain, heart palpitations cough, swelling in abdomen or 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077613
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 79 of 237 legs, fever); treatments received for myocarditis/pericarditis (e.g., non-steroidal anti -
inflammatory drugs ( NSAIDs), colchicine, corticosteroids, pericardectomy); healthcare 
resource utilization follow ing the event, and long -term sequelae for up to one year following 
the event (for myocarditis: recovery, sudden cardiac death, heart failure cardiogenic shock, 
fulminant myocarditis, inflammatory cardiomyopathy, heart transplant, arrhythmia; for 
pericardit is: recovery, chronic pericarditis, restrictive pericarditis, recurrent pericarditis).  
9.8. Quality Control  
Data for the study will be extracted from electronic databases in the CDW of the VHA. Each 
data content area in the CDW is subjected to similar checks, f rom high level variable 
name/type checks, to detailed trending comparisons. As an example, the diagnostic data is 
subject to the following checks:  
• Referenced table exists  
• Diagnosis type is correctly assigned by codes defining the diagnosis  
• Percentages, rat es, are as expected (check ranges and for missing)  
• Both inpatient and outpatient diagnosis codes are captured. Referenced variables exist 
and are of appropriate length and type  
Data retrieval will be coordinated by an experienced programmer/analyst. The an alyst will 
write programming for retrieval of each data element from the electronic databases. Double 
programming will be performed for the first iteration of the analyses; results/datasets will be 
compared, and if any discrepancies are identified, both pr ogrammers will determine a 
resolution, bringing in a third programmer if needed. Subsequent iterations of analyses (i.e., 
re-runs of the analyses) will be audited by a senior programmer. All tables will be reviewed 
by the project manager and the principal investigator to evaluate for internal consistency of 
counts and totals. All calculated variables will be checked against the component variables 
(cross tabs) to ensure accuracy. For example, categorical age would be compared with 
continuous age to confirm that each category of age contained only individuals of the 
expected age ranges within that category.  
9.9. Strengths and Limitations of the Research Methods  
To identify individuals  who experienced safety events of interest associated with Pfizer-
BioNTech COVID -19 vaccine, the SCRI method of signal detection offers some key 
advantages. The SCRI approach inherently adjusts for within-individual  confounders, such as 
age, sex, and confounding by indicat ion. Additionally While control intervals can be defined 
both pre- and post- vaccination , the inclusion of current study will only use  a post-vaccination 
control period will account because individual s may be more vigilant for increased detection 
bias from stimulated the reporting of  possible safety events of interest due to heightened 
vigilance on COVID 19 vaccines after they receive a vaccine  than before vaccination, which 
may bias the comparison between a post -vaccine risk interval with a pre -vaccine control 
interval.4954 Specifically, safety events of interest may be more likely to be reported or sought 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077614
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 80 of 237 care for after vaccination with Pfizer -BioNTech COVID -19 vaccine than before  (i.e., during 
the prevaccination control interval) , which may result in bias against the Pfizer -BioNTech 
COVID-19 vaccine . Lastly, SCRI allows for near real -time monitoring of safety risks 
associated with the Pfizer -BioNTech COVID -19 vaccine. Similar considerations apply to the 
SCCS design with post -vaccination control interval that will be used in the signal evaluation 
phase.  
 
The comparison  of vaccinated  to contemporary unvaccinated controls yield s a more 
interpretable result than other planned analyses using SCRI and active comparators who 
receive seasonal influenza vaccination (i.e., the increased risk of experiencing a specific 
safety event due to Pfiz er-BioNTech COVID -19 vaccination). The potential for selection bias 
(i.e., confounding by indication, healthy user bias) will be mitigated by comparing baseline 
demographic and clinical characteristics  among the unvaccinated controls. Unvaccinated 
controls will be required to have similar healthcare -seeking behaviors as Pfizer-BioNTech 
COVID-19 vaccinees, including a t least 1 year of enrollment in and no disenrollment from 
VHA benefits prior to their match date. This design is also not limited to assumption s 
required by SCCS and SCRI, and can also be completed rapidly as it does not require post -
vaccination control intervals. However, it is noted that the mass vaccination campaign in the 
past year has provided various channels to receive vaccination, and the refore unvaccinated 
controls may be misclassified if they are vaccinated outside of the VHA.  
 
The VHA CDW provides a range of benefits, including its comprehensive structure, large 
number of variables, and electronic accessibility. The VHA CDW also includ es EMR data 
that include structured fields (which will be used for signal detection) and open fields (such 
as physician notes, which will be used for signal evaluation verification  and case validation, 
as needed). Importantly, the VHA CDW retains electronic immunizat ion records that include 
manufacturer name and lot numbers, facilitating the identification of brand -specific vaccines, 
such as the Pfizer -BioNTech COVID -19 vaccine. Moreover, the VHA CDW data are 
updated on a daily basis, enabling near real -time rapid mon itoring of potential safety signals.  
 
However, there are several limitations when relying on VHA that should be noted. First, 
there could be gaps in the data since individuals may receive healthcare services outside of 
VHA facilities. As such, if individu als receive the Pfizer -BioNTech COVID -19 vaccine 
outside of a VHA facility, this information will not be captured in the VHA EMR system. 
Similarly, individuals may have also received past seasonal influenza vaccinations outside of 
the VHA system, and thus would be misclassified as not having received vaccine in the 
current analysis. For example, veterans with secondary insurance or veterans who are 
65 years of age or older who have Medicare may receive health care services outside of VHA 
facilities. One stu dy on VHA enrollees in seven different states found that of all individuals 
admitted to VHA hospitals in 2007, one fifth also had a non VHA hospitalization during that 
year.5055 Another study reported that about 53% of Veterans 65 years of age and older who 
were dually eligible for VHA and Medicare services in 2003 2004 used both.5156 Hence, it is 
important to note that data on vaccination status ma y be incomplete. However, this limitation 
will be addressed by examining subgroups of individuals who receive care regularly at VHA 
facilities, as well as those with Priority group 1 status, to ensure that their healthcare data are 
complete to the extent p ossible in the CDW. SecondThe results from these  subgroup  analyses 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077615
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 81 of 237 will be compared to the overall population results from the VHA CDW to confirm consistent 
findings such that if there are missing data for individuals in the overall population, the 
missing data can be assumed to be  missing at random and not biasing  the results in any 
direction. This will be evaluated in the context of evaluating the relative risk of safety events 
of interest in the comparative analyses. However, if there are discrepancies that su ggest data 
are not missing at random and could bias results, subgroup analyses will be conducted for 
individuals with dual coverage in the VHA and Medicare. T he CDW data will be  
supplement ed and linked with Medicare administrative claims data at the patien t level to 
ensure a more comprehensive evaluation of the care an individual receives. Linking variables 
are available in the data to allow for patient -level linking of the two data sources. Given the 
older age of many veterans, it is likely that these indi viduals have secondary coverage with 
Medicare.  
 
Lastly, to the extent that the individuals in the VHA database are different from individuals 
outside of the VHA, the results may not be generalizable to the broader US population. For 
example, since the VHA  includes predominantly male Veterans (approximately 90% male), 
findings from this study may not be generalizable to women in the US.  
9.10. Other Aspects  
Not applicable.  
 
10. PROTECTION OF HUMAN SUBJECTS  
10.1. Patient Information  
All parties will comply with all applicab le laws, including laws regarding the implementation 
of organizational and technical measures to ensure protection of patient personal data. Such 
measures will include omitting patient names or other directly identifiable data in any 
reports, publications,  or other disclosures, except where required by applicable laws.  
To protect the rights and freedoms of natural individuals with regard to the processing of 
personal data, when study data are compiled for transfer to Pfizer and other authorized 
parties, any  patient names will be removed and will be replaced by a single, specific, 
numerical code. All other identifiable data transferred to Pfizer or other authorized parties 
will be identified by this single, patient -specific code. In case of data transfer, Pfi zer will 
maintain high standards of confidentiality and protection of individuals’ personal data 
consistent with the vendor contract, and applicable privacy laws.  
No personal data is planned to be transferred off the VA servers. Specifically, the Clinical  
Epidemiology Program (CEP) at White River Junction VA Medical Center will conduct this 
safety surveillance study with sponsorship from Pfizer and assistance from Analysis Group, 
Inc. The project will be led by the VA, with Dr. Yinong Young -Xu, Director of  CEP, serving 
as the Principal Investigator. Data access will be granted through VA Informatics and 
Computing Infrastructure (VINCI). VHA data will not be provided to Pfizer or Analysis 
Group. Rather, only VA employees, including those with research servic e without 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077616
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 82 of 237 compensation (WOC) employee status, who have completed necessary VA training and have 
proper clearance will access and analyze data on secure VA servers and behind necessary 
firewalls, under the direction and supervision of Dr. Young -Xu. Given th e sensitive nature of 
healthcare data, comprehensive security measures will be implemented to ensure the 
confidentiality, integrity, and protection of Veterans’ privacy and healthcare data.  
10.2. Patient Consent  
As this study does not involve data subject to pri vacy laws according to applicable legal 
requirements, obtaining informed consent from individuals by Pfizer is not required.   
10.3. Institutional Review board (IRB)/Independent Ethics Committee (IEC)  
There must be prospective approval of the study protocol, pro tocol amendments, and their 
relevant documents from the relevant IRBs/IECs. All correspondence with the IRB/IEC must 
be retained. Copies of IRB/IEC approvals must be forwarded to Pfizer. The study protocol 
will be reviewed by the IRB of the VA Medical Cent er, White River Junction, VT.  
10.4. Ethical Conduct of the Study  
The study will be conducted in accordance with legal and regulatory requirements, as well as 
with scientific purpose, value and rigor and follow generally accepted research practices  
described in G uidelines for Good Pharmacoepidemiology Practices (GPP) issued by the 
International Society for Pharmacoepidemiology,5357 the FDA Guidance for Industry and 
FDA Staff: Best Practices for Conducting and Reporting, Pharmacoepidemiol ogic Safety 
Studies Using Electronic Healthcare Data5258 and Good Epidemiological Practice (GEP) 
guidelines issued by the International Epidemiological Association (IEA).5459  
11. MANAGEMENT AND REPORTING OF ADVERSE EVENTS/ADVERSE 
REACTIONS  
Signal Detection and Signal Evaluation  
This study involves data that exist as structured data by the time of study start or a 
combination of existing structured data and unstructured data, which will be conver ted to 
structured form during the implementation of the protocol solely by a computer using 
automated/algorithmic methods, such as natural language processing. In these data sources, 
individual patient data are not retrieved or validated, and it is not pos sible to link (i.e., 
identify a potential association between) a particular product and medical event for any 
individual.  Thus, the minimum criteria for reporting an adverse event (AE) (i.e., identifiable 
patient, identifiable reporter, a suspect product,  and event) cannot be met.  
Signal Verification  
This study protocol requires human review of patient -level unstructured data; unstructured 
data refer to verbatim medical data, including text -based descriptions and visual depictions 
of medical information, such as medical records, images of physician notes, neurological 
scans, X-rays, or narrative fields in a database.  The reviewer is obligated to report adverse 
events (AEs) with explicit attribution to any Pfizer drug that appear in the reviewed 
informatio n (defined per the patient population and study period specified in the protocol).  
Explicit attribution is not inferred by a temporal relationship between drug administration 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077617
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 83 of 237 and an AE, but must be based on a definite statement of causality by a healthcar e provider 
linking drug administration to the AE.  
The requirements for reporting safety events of interest on the non -interventional study 
(NIS) adverse event monitoring (AEM) Report Form to Pfizer Safety are as follows:  
• All serious and non -serious AEs wit h explicit attribution to any Pfizer drug that 
appear in the reviewed information must be recorded on the data collection tool (e.g., 
chart abstraction form) and reported, within 24 hours of awareness, to Pfizer Safety 
using the NIS AEM Report Form.  
• Scenarios involving drug exposure, including exposure during pregnancy, exposure 
during breast feeding, medication error, overdose, misuse, extravasation, lack of 
efficacy, and occupational exposure associated with the use of a Pfizer product must 
be reported, w ithin 24 hours of awareness, to Pfizer Safety using the NIS AEM 
Report Form.  
For these AEs with an explicit attribution or scenarios involving exposure to a Pfizer 
product, the safety information identified in the unstructured data reviewed is captured in the 
Event Narrative section of the report form, and constitutes all clinical information known 
regarding these AEs.  No follow -up on related AEs will be conducted.  
All the demographic fields on the NIS AEM Report Form may not necessarily be completed, 
as the form designates, since not all elements will be available due to privacy concerns with 
the use of secondary data sources. While not all demographic fields will be completed, at the 
very least, at least one patient identifier (e.g., gender, age as captur ed in the narrative field of 
the form) will be reported on the NIS AEM Report Form, thus allowing the report to be 
considered a valid one in accordance with pharmacovigilance legislation. All identifiers will 
be limited to generalities, such as the stateme nt “A 35-year-old female...” or “An elderly 
male...”  Other identifiers will have been removed.       
Additionally, the onset/start dates and stop dates for “Illness”, “Study Drug”, and “Drug 
Name” may be documented  in month/year (mmm/yyyy) format rather than identifying the 
actual date of occurrence within the month /year of occurrence in the day/month/year 
(DD/MMM/YYYY ) format.  
All research staff members must complete the following Pfizer training requirements:    
• Your Reporting Responsibilities ( YRR) Training for Vendors Working on Pfizer 
Studies  
These trainings must be completed by research staff members prior to the start of data 
collection.  All trainings include a “Confirmation of Training Certificate” (for signature by 
the trainee) as a record o f completion of the training, which must be kept in a retrievable 
format.  Copies of all signed training certificates must be provided to Pfizer.  
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077618
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 84 of 237 Re-training must be completed on an annual basis using the most current Your Reporting 
Responsibilities train ing materials.  
12. PLANS FOR DISSEMINATING AND COMMUNICATING STUDY RESULTS  
This protocol will be posted on publicly available registers following its finalization. The 
final study results will be made publicly available via the European Union Post Authorisati on 
Safety (EU PAS) Register and may be submitted for publication in a peer reviewed medical 
journal. 
In the event of any prohibition or restriction imposed (e.g., clinical hold) by an applicable 
competent authority in any area of the world, or if the inves tigator is aware of any new 
information which might influence the evaluation of the benefits and risks of a Pfizer 
product, Pfizer should be informed immediately.   
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077619
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 85 of 237 13. REFERENCES  
1. World Health Organization (WHO). WHO Director -General’s opening remarks at th e 
media briefing on COVID -19 -11 March 2020.  
https://www.who.int/dg/speeches/detail/who -director-general-s-opening-remarks-at-
the-media-briefing-on-covid-19---11-march-2020. Accessed November 11, 2020.  
2. Johns Hopkins University. Coronavirus Resource Cen ter. Accessed November 10, 
2020. https://coronavirus.jhu.edu/. Accessed November 10, 2020.  
3. U.S. Food and Drug Administration (FDA). Vaccines and Related Biological 
Products Advisory Committee Meeting: FDA Briefing Document Pfizer -BioNTech 
COVID-19 Vaccine. Accessed December 10, 2 020. 
4. U.S. Food &and Drug Administration (FDA). Pfizer COVID -19 Vaccine EUA Letter 
of Authorization reissued 12 -23-20. In:December 23, 2020.  
5. U.S. Food and Drug Administration (FDA). FDA Takes Key Action in Fight Against 
COVID-19 By Issuing Emergency Use Authorization for First COVID -19 Vaccine.  
https://www fda.gov/news -events/press -announcements/fda -takes-key-action-fight-
against-covid-19-issuing-emergency -use-authorization -first-covid-19. 
6. Pfizer Inc. Pfizer and BioNTech Achieve First Au thorization in the World for  a 
Vaccine to Combat COVID -19. Accessed January 4, 2021.  
 https://www.pfizer.com/news/press -release/press -release-detail/pfizer -and-biontech-
achieve-first-authorization -world. Accessed January 4, 2021.  
7. Pfizer Inc. Pfizer and BioNTech Receive Authoriz ation in the Europe an Union for 
COVID-19 Vaccine. Accessed December 28, 2020.  
            https://www.pfizer.com/news/press -release/press -release-detail/pfizer -and-biontech-
receive-authorization -european-union. Accessed December 28, 2020.  
8. Law B. SO2 -D2.1.2 Priority List of COVID-19 Adverse Even ts of special interest: 
 Quarterly update. Brightoncollaboration.us . April 23, 2020. 
 https://brightoncollaboration.us/wp -
 content/uploads/2021/01/SO2 D2.1.2 V1.2 COVID -19 AESI-update-23Dec2020 -
 review final.pdf  
9. Shimabukuro T.  Enhanced sa fety monitoring for COVID -19 vaccines in early phase  
vaccination . National Center for Immunization & Respiratory Diseases. 8September 
22, 2020. https://www.cdcgov/vaccines/acip/meetings/downloads/slides -2020-
09/COVID -03-Shimabukuro.pdf  
10. Yih WK, Lee GM, Lieu TA, et al. Surveillance for adverse events following receipt 
of pandemic 2009 H1N1 vaccine in the Post -Licensure Rapid Immunization Safety  
Monitoring (PRISM) System, 2009 -2010. Am J Epidemiol. 2012;175(11):1120 -1128. 
911. American Medical Association (AMA). Appendix Q: Severe Acute Respiratory 
Syndrome Coronavirus 2 (SARS -CoV-2) (coronavirus  disease [COVID -19]) 
Vaccines.  Accessed January 12, 2021. https://www.ama -assn.org/system/files/2020 -
11/covid-19-immunizations -appendix-q-table.pdf. Accessed January 12, 2021.  
10. Centers for Medicare & Medicaid Services. COVID 19 Vaccines and Mon oclonal 
Antibodies. 2020; https://www.cms.gov/medicare/medicare partbdrugaverage
salesprice/covid 19vaccinesandmonoclonal antibodies. Accessed January 14, 
2021. 
090177e197e9b182\Final\Final On: 26-Aug-2021 15:45 (GMT)
FDA-CBER-2021-5683-1077620
BNT162b2 (COVID -19 vaccine)  
C4591012 NON-INTERVENTIONAL  STUDY PROTOCOL   
Final Version 12.0, 27 January 31 Aug 2021 
 
 
PFIZER CONFIDENTIAL  
Page 86 of 237 1112. Center for Disease Control (CDC). People with Certain Medical Conditions. 
December 29, 2020 ;. Accessed January 4, 2021.  
https://www.cdc.gov/coronavirus/201 9-ncov/need -extra-precautions/people -with-
medical-conditions html. Accessed January 4, 2021.  
12. U.13. Patel M, Chen J, Kim S, et al. Analysis of MarketScan Data for 
Immunosuppressive Conditions and Hospitalizations for Acute Respiratory Illness, 
United States. Emerging Infectious Disea ses. 2020;26(8):1720 -1730. 
doi:10.3201/eid2608.191493.  
14. U.S. Department of Veterans Affairs. Veterans Health Administration. 2020 ; 
https://www.va.gov/health/. . Accessed  November 10, 2020.  
https://www.va.gov/health/  
15. 13. U.S. Department of Veterans Affai
…[truncated]