Background and Rationale
SARS-CoV-2 has spread rapidly and globally since its emergence, causing Coronavirus Disease 2019 (COVID-19). The World Health Organization (WHO) declared that the outbreak constituted a public health emergency of international concern on 30 January 2020 and declared the outbreak to be a pandemic on 11 March 2020. Due to the global spread of COVID-19 pandemic, rapid development of a COVID-19 vaccine is a worldwide priority.
Following the primary analysis of the Phase 3 study COV3001, the United States (US) Food and Drug Administration (FDA) granted an emergency use authorization (EUA) for the Ad26.COV2-S vaccine for use in individuals 18 years of age and older on 27 February 2021. However, following a successful efficacy analysis that supports issuance of an EUA, further evaluation of the Ad26.COV2-S vaccine is still needed, including observational studies that leverage health insurance claims databases and more precise estimation of vaccine effectiveness. Active surveillance of vaccines through additional pharmacovigilance activities such as observational studies should also be considered.
Per communication by the European Commission on October 15, 2020. the Commission has entered into agreements with individual vaccine producers on behalf of the Member States, purchasing and/or reserving the right to purchase vaccine doses under Advance Purchase Agreements. Member States and public health authorities should prepare to undertake studies of vaccine effectiveness and safety via coordination by the European Medicines Agency (EMA) and the European Centre for Disease Prevention and Control, and specifically to prepare for participation in large-scale EU-wide effectiveness and safety monitoring studies. On 11 March 2021, the EMA granted a conditional marketing authorization (CMA) for the Ad26.COV2.S vaccine for use in individuals 18 years of age and older.
To address these regulatory requests, the sponsor plans to initiate a post-authorization safety study (PASS) aiming to characterize and evaluate the safety profile of Ad26.COV2.S in a large population sample size and to inform the scientific community on AESIs that could be associated with the use of Ad26.COV2.S.
Research Question and Objectives
This study has 2 chronologically consecutive aims: 1) to conduct a feasibility assessment aiming to inform the safety evaluation study and 2) to assess the risk of developing pre-specified and newly identified AESIs following administration of Ad26.COV2.S.
Objectives
Feasibility Assessment
The feasibility assessment will include the following objectives (not all objectives may be assessed in all analyses):
?To provide a comprehensive overview of the methods for identification of COVID-19 vaccine exposure including provenance of data and linkage to vaccination registry
?To monitor the number of individuals unexposed to any COVID-19 vaccine and to investigate the risk of misidentification of the COVID-19 vaccine exposure. Both missing vaccination instances and misclassification of the type of vaccine will be assessed with the benchmark provided by: https://vaccinetracker.ecdc.europa.eu/public/extensions/COVID-19/vaccine-tracker.html#uptake-tab)
?To compare individuals receiving and not receiving different COVID-19 vaccines with respect to demographics and clinical characteristics of recipients and assess potential confounders and risk factors for the AESI
?To conduct time-to-onset analyses for the AESIs (including exacerbation of asthma) with respect to time since vaccination.
Secondary objectives
The secondary objectives for this study are to assess the potential association between the occurrence of predefined and newly identified AESIs and vaccination with Ad26.COV2.S in the following specific subpopulations (part of the main study population):
?Immunocompromised individuals,
?Pregnant women,
?Individuals who have a prior history of thrombotic events and/or thrombocytopenia,
?Prior COVID-19 infection,
?Individuals with a prior history (ever) of the specific event more than a year before start of follow-up.
These results will be compared to COVID-19 vaccine unexposed individuals and, to individuals exposed to other types of COVID-19 vaccines (mRNA or adenovirus-based platforms), or compared to a control window within the same individual.
Endpoints
Feasibility Assessment
The endpoints for the feasibility assessment, including a monitoring phase, are:
?Quality assessment of data that has been extracted from the source data banks (completeness, logics and benchmarking between data sources and against external data)
?The number of doses and uptake of different COVID-19 vaccines by calendar time
?The demographic and morbidity characteristics of individuals receiving the different COVID-19 vaccines as well as non-COVID-19 vaccinated individuals
?Algorithms for AESI identification (including exacerbation of asthma)
?Risk factors for AESI.
Safety Evaluation Study
The primary endpoints for the safety evaluation study are to estimate the risk of the selected AESIs listed below among individuals vaccinated with the Ad26.COV2.S vaccine and in corresponding unexposed individuals and individuals exposed to other types of COVID-19 vaccines (split between mRNA platform-based vaccines and adenovirus-based platforms), or during a control window within the same individual.
The contractual agreements with data partners will stipulate the ability to allow newly identified AESIs to be incorporated into the study objectives, and the current protocol that will be used for governance approvals will state that additional AESI may be added. Each data access provider is responsible to inform the appropriate ethics boards of these amendments.
Acute events (events expected to be recorded within 60 days of vaccination).
?The incidence of anaphylaxis within 0-2 days.
?The incidence of generalized convulsion, arrhythmia, acute kidney failure and acute hepatic failure within 1-14 days.
?The incidence of the following events within 1-28 days.
*thrombotic events (microangiopathy (including capillary leak syndrome), disseminated intravascular coagulation, deep vein thrombosis, pulmonary embolism, cerebral vein thrombosis, peripheral thrombosis, thrombosis with thrombocytopenia and ischemic strokes, acute coronary syndrome (acute myocardial infarction, unstable angina, and new episode of angina)
*bleeding events including hemorrhagic stroke (no subarachnoid)
*composite endpoint: Venous thrombosis (including pulmonary embolism and deep vein thrombosis)
*composite endpoint: Arterial thrombosis (including acute coronary syndrome and ischemic stroke)
*composite endpoint: stroke (including hemorrhagic and non-hemorrhagic stroke)
*heart failure, and stress cardiomyopathy.
?The incidence of immune/inflammatory events: encephalitis (including acute demyelinating encephalomyelopathy and meningoencephalitis), Guillain-Barré Syndrome, Bell’s palsy, immune thrombocytopenia, thrombocytopenia, transverse myelitis and cardiac inflammatory disorders (including myocarditis and pericarditis) within 1-42 days.
?The incidence of sensorineural hearing loss within 1-60 days.
Non-acute events (events expected to be recorded >60 days after vaccination)
?The incidence of autoimmune thyroiditis, multiple sclerosis, acute aseptic arthritis, and type 1 diabetes mellitus within 1-365 days.
Newly identified AESI may be incorporated to the list of AESI to be investigated by the current study if a safety signal occurs.
Study Design
The study will comprise a feasibility assessment for a period of 18 months and a safety evaluation study.
This study is a retrospective observational study using electronic health care databases of various types in Europe. Eligible individuals will be included in the study from 01 December 2020, and the study will end at the last date of data availability in each database. The AESIs included in this study are considered as potential or identified safety risks following administration of Ad26.COV2.S. The selected AESIs represent a heterogeneous group including multiple organ systems, and acute and chronic conditions.
Feasibility: A feasibility assessment, focusing on the availability of Ad26.COV2.S vaccine data (vaccine uptake), characteristics of individuals vaccinated with different COVID-19 vaccines and non-vaccinated as well as measurement of AESIs will be conducted for each participating data source. From the start of the feasibility assessment, participating data sources will extract data every 4 months (where possible, it is known already this will not be possible in Norway, where only one extract is done per year). Summary reports will be produced at each extraction (up to 3 rounds of monitoring analyses are foreseen). Primary design for the feasibility assessment will be a cohort study including all individuals with at least one day of follow-up after 01 December 2020.
Safety Evaluation: Evaluation of safety concerns will be conducted using a retrospective observational study in electronic health care databases of various types in Europe. Eligible individuals will be included in the study from 01 December 2020 and the study will end at the last date of data availability in each database.
For each AESI, the study design of the safety evaluation study will depend on whether the AESI is an acute or non-acute event and follow the decision framework described in the ACCESS template protocols for evaluation of safety events in electronic health care databases.
The primary study design for evaluation of acute events (events expected to occur within 2, 14, 28, 42, or 60 days of vaccination) will be a self-controlled risk interval (SCRI) design with a pre-vaccination control window and for non-acute events (events expected to occur within 365 days) will be a cohort design with concurrent unexposed comparators. Additionally, the cohort analysis will also include two active comparator groups: one viral vector vaccine comparator group (ie, Vaxzevria® [AZD1222] by Oxford/AstraZeneca) and one mRNA vaccine comparator group (ie, Comirnaty® [tozinameran] by Pfizer/BioNTech and/or Spikevax® [elasomeran] by Moderna), where possible (pending on the feasibility analysis with a maximum follow-up time of one year).
In sensitivity analyses, an equivalent cohort design will also be used to assess acute events for data sources. When it is established that data sources do not capture vaccinations well during the feasibility assessment, we will conduct a sensitivity analysis excluding these data sources from the cohort analysis with non-exposed comparators for the non-acute events, to avoid misclassification of exposure.
Individuals start follow-up at time zero (cohort entry) and end follow-up at the earliest occurrence of latest data availability of the databank, individual exit from the database, completion of the study period, or death. At least one year of enrollment prior to time zero (cohort entry) will be required to determine whether individuals meet the study criteria and to define baseline characteristics. If more historic data is available, this will be included.
Self-controlled risk interval design (primary study design for acute events)
The SCRI design will compare the risk of the AESI in a post-vaccination risk window to a pre vaccination (interim analyses) and post-vaccination control window (sensitivity analysis) within the same individual. For the final analysis, when enough follow-up is accrued, the main analysis will be the post-vaccination control window.
The SCRI design will include only individuals in the primary analysis who received at least one first ever dose of the Ad26.COV2.S vaccine during the study period. Vaccinated individuals enter study at the date of the start of the pre-vaccination control window (time zero). The SCRI design will compare the risk of each outcome during the post-vaccination risk window following the COVID-19 dose with a self-matched control interval that may be prior to vaccination, or after the vaccination risk window, to assess the baseline risk of the outcome. The control window will have the same maximum length as the risk window to minimize time-varying confounding that cannot be measured. In case the follow-up time does not capture the maximum post-vaccination risk or control -window due to right censoring, all available follow-up time will be utilized. Follow-up will be censored upon vaccination with another COVID-19 vaccine or end of follow-up. A washout period between the control and risk window will be applied to minimize capturing of prevalent events during the risk window. A post vaccination control window may induce a bias towards the null when an AESI that was identified during the vaccination risk window, is captured again during the control window due to a re-admission instead of reoccurrence of the AESI. This situation is more likely to be encountered for acute potentially recurring events such as cardiac events (including acute coronary syndrome, myocarditis, pericarditis, arrhythmia, and stress cardiomyopathy). Moreover, due to lag-times in updating the relevant databanks there is a higher risk that the post-vaccination control window may not capture all events and underestimate the rate of events. Use of a pre-vaccination control window, will deal with these issues. For the interim analyses we will therefore use a pre-vaccination control window as primary analysis, for the final analysis a pre- and post-vaccination control window will be used and analyzed separately.
Cohort design (primary study design and non acute events and sensitivity analysis for acute events)
For the primary analysis of non acute events, a retrospective cohort design will be used to estimate the rate of vaccination with Ad26.COV2.S vaccine, describe the characteristics of these vaccinated individuals and subsequently to estimate the incidence of autoimmune thyroiditis, multiple sclerosis, acute aseptic arthritis, and type 1 diabetes mellitus after receipt of the vaccine dose and to compare this incidence with that occurring in an unvaccinated matched comparator group, and in two groups of individuals exposed to other types of COVID-19 vaccines (ie, one viral vector vaccine comparator group vaccinated (ie, Vaxzevria® [AZD1222] by Oxford/AstraZeneca) and one mRNA vaccine comparator group (ie, Comirnaty® [tozinameran] by Pfizer/BioNTech and Spikevax® [elasomeran] by Moderna).
As a sensitivity analysis for acute events, a cohort analysis (following the same approach used in the main analysis for non-acute events) will be conducted.
?Exposed cohort: individuals will have received at least 1 dose of Ad26.COV2.S.
?Concurrent unexposed cohort: individuals that have not been vaccinated with Ad26.COV2.S vaccine or any other COVID-19 vaccine at any time prior to time zero matched to the vaccinated individual for important characteristics.
?Cohort exposed to other COVID-19 vaccines: individuals will have received at least 1 dose of a viral vector COVID-19 vaccine (ie, Vaxzevria® [AZD1222] by Oxford/AstraZeneca) or an mRNA COVID-19 vaccine (Comirnaty® [tozinameran] by Pfizer/BioNTech and Spikevax® [elasomeran] by Moderna, respectively).
In this retrospective cohort design, time zero (cohort entry) will be defined as the time at which the exposure status is assigned, when selection criteria are applied and when study outcomes start to be counted. Time zero in the exposed cohorts (ie, recipients of the vaccines) will be the day the first dose of the corresponding COVID-19 vaccination was received. Time zero in the unexposed group will be a day when an individual did not receive a COVID-19 vaccine dose and randomly chosen by calendar matching to the time zero of the corresponding Ad26.COV2.S exposed individual (ie, a random day during the same week that the matching individual in the exposed cohort receives the Ad26.COV2.S vaccine).
Individuals in the Ad26.COV2.S exposed cohort will be individually matched to one individual in the concurrent unexposed cohort and to one individual in each active comparator cohorts on key clinical variables (exact age, sex, and presence of one or more risk factors for severe COVID-19 [eg, cancer, sickle cell disease, obesity, chronic kidney disease, chronic respiratory disease, human immunodeficiency virus infection], and month of vaccination (using caliper if needed)) at time zero. Additional details on the matching process will be specified in the statistical analysis plan (SAP). A single individual may contribute person-time to the exposed and unexposed groups at different time points (details will be described in the SAP). Individuals will be classified into exposure groups that are compatible with their data at time zero. Follow-up under unexposed status is censored if an individual receives a COVID-19 vaccine. In a sensitivity analysis, vaccination to other non-COVID 19 vaccines will also be an exclusion and censoring event.
Setting and Study Population
For the implementation of the feasibility assessment, electronic health care databases in Northern, Southern and Western Europe that have showed interest and are member of the Vaccine Monitoring Collaboration for Europe (VAC4EU) will be used.
Feasibility Assessment
For the feasibility assessment, the study period will start on 1 December 2020 or the latest date when the individual is registered in the data source (plus 365 days), or born and will end a maximum of 18 months thereafter. Individuals will be followed until the earliest of the following dates: death, end of data availability, individual exit from the data source, or the completion of the period
Safety Evaluation Study
For the SCRI design, the study period will start at 01 December 2020 and last until the end of follow-up.
For the cohort design, the study period will start at 01 December 2020 and will end at the end of follow-up.
?For the SCRI design, follow-up starts at the start of the pre-vaccination control window following an AESI specific washout period. Follow-up ends at the earliest of the following: death, end of data availability, individual withdrawal of the study, end of the post-vaccination risk or control window, or receipt of another COVID-19 vaccine.
?For the cohort design follow-up starts at time zero (Ad26.COV2.S or other COVID-19 vaccination date or the matched date in the non-exposed individual). Follow-up ends at occurrence of each AESI, death, end of data availability, individual exits the database, after one year of follow-up, receipt of a COVID-19 vaccine (unexposed cohort only), receipt of a different COVID-19 vaccine from the one granting access to the study (exposed cohorts only). For the vaccine exposed comparator groups, a second dose of the same corresponding vaccine will be allowed (more details will be provided in the SAP). In a sensitivity analysis, follow-up in the cohort study will be censored upon receipt of other vaccines (non-COVID-19 vaccine). For the subgroup analysis of subjects without COVID-19 at time zero, patients will be censored if they develop COVID-19 during follow-up.
Variables
Exposure assessment
Exposure will be based on available prescription, dispensing, or administration of the Ad26.COV2.S and other COVID-19 vaccines. Vaccine receipt and date of vaccination will be obtained from all possible sources that capture COVID-19 vaccination, such as pharmacy dispensing records, general practice records, immunization registers, vaccination records, medical records, or other data banks. During the feasibility assessment, the completeness of information will be assessed and described. Depending on the data source, vaccines may be identified via nationally used product codes (including batch numbers) where possible. The exposure of interest for the safety evaluation study is the receipt of Ad26.COV2.S vaccine.
Study outcomes
AESIs, as listed below and in line with the definitions and code lists that for most of the AESI have been created for the ACCESS project (https://zenodo.org/communities/vac4eu/), will be identified, with a date of diagnosis, using predefined validated algorithms (where available), based on diagnosis codes (with procedure and/or pharmacy dispensing codes and/or limited to specific medical care settings if applicable to the outcome). The impact of different provenance of data (hospital, GP diagnoses) and algorithms on the outcome frequency will be assessed and described in the feasibility assessment.
Evaluation of Safety Outcomes
The sponsor has created a list of AESIs based on current knowledge of the Ad26.COV2.S vaccine. Background incidence rates for most of the data sources and AESIs are available on the VAC4EU dashboard https://vac4eu.org/covid-19-tool/. Definitions and codes are available on Zenodo (https://zenodo.org/communities/vac4eu/).
Covariate definition
Feasibility Assessment
In the feasibility assessment covariates will be assessed at the time of COVID-19 vaccine administration or time matched non-vaccinated individuals.
Safety Evaluation Study
Covariate assessment for the description of the patient population at baseline (time zero) will use all available time of enrollment in the corresponding data source (minimum required enrollment period is 1 year).
Covariates will be assessed at time zero (for the cohort design) or the date of vaccine dose (for the SCRI design) to be used to define patient populations of special interest or priority vaccination groups, to define subgroups of interest for sensitivity analyses, or to control for confounding.
We will consider the following time-varying covariates and corresponding period of evaluation for the SCRI measured at start of the risk window and the control window. To capture use of medications of interest during the month prior to start of the study periods (ie, risk control and control risk), we will consider the drug supply covering the month prior to the start of the two periods of interest.
Data Sources
The study will use data from secondary electronic health record databases that are population based. All data sources will have the ability to provide data on COVID-19 vaccines, outcomes (diagnoses, procedures, laboratory results, and treatments), and important covariates. It is not currently known the extent to which COVID-19 vaccines, product types, and batch numbers will be captured in data sources. To be included in the study, data sources should preferably be updated at a minimum once every 3 months. At the proposal stage, members of VAC4EU were offered the option to participate in the study. Seven data sources from 5 countries (Italy, Norway, Spain, The Netherlands, UK) countries will be included in the study. A more detailed description will be included in the VAC4EU FAIR Catalogue.
When establishing the agreements with the data sources to conduct the study, it will be emphasized that the current list of AESIs may be expanded during the course of the study to accommodate newly identified AESIs.
Study Size
The study will be conducted in a source population of approximately 43 million individuals, although children will not be vaccinated. It is assumed that up to 10% will be vaccinated with the Ad26.COV2.S vaccine.
Main Statistical Methods
A general description of the planned statistical methods to be used to analyse the data collected in this study is presented in the main body of the document. Additional details will be provided in the SAP.
For the feasibility analysis the utilization patterns of Ad26.COV2.S and other COVID-19 vaccines will be characterized and monitored over time. Description of demographics and clinical characteristics will be reported for different groups of vaccine recipients and non-recipients, overall and among sub-cohorts of interest, such as individuals who are immunocompromised, pregnant, or have specific comorbidities.
The primary analysis will focus on the calculation and comparison of the incidence rates of each non-acute AESI between individuals exposed to Ad26.COV2.S and
1.unexposed individuals
2.individuals exposed to another viral vector COVID-19 vaccine (ie, Vaxzevria® [AZD1222] by Oxford/AstraZeneca); and
3.individuals exposed receiving a mRNA COVID-19 vaccine (cohort).
For acute events, the relative risk between risk window and control window will be estimated (SCRI) among individuals exposed to Ad26.COV2.S. All analyses will be conducted within each data source and pooled across data sources using a random-effects model.
For the COVID-19 vaccines with a 2-dose schedule, AESI-specific risk windows after each dose of the comparator will be considered for the analyses.
Analyses within the primary study designs (cohort design for non-acute events and SCRI design for acute events) will be stratified by the clinically relevant subgroups below. Stratification will be done also for matching variables when considered relevant.
?Selected comorbidities, including risk factors for severe COVID 19 (by presence or absence of each comorbidity)
?Frailty score (categorized)
?Age ([<18], [18-39]; [40-59]; [60-79], [80+])
?Sex.
Sensitivity analyses will be performed to:
?Assess the risk of AESI within extended disease-specific risk windows for events for which the risk interval is not well known or documented in the vaccine safety literature.
?Assess the impact of data sources that are shown not to capture vaccinations well during the feasibility assessment. In the sensitivity analyses with non-exposed comparators those data sources from the cohort analysis for the acute and non-acute events will be excluded, to avoid misclassification of exposure.
?Excluding and censoring upon occurrence of other (non-COVID-19) vaccinations in the SCRI and cohort study.
?Conduct a SCRI analysis using a post-vaccination control window for all acute AESIs.
?Interim Analysis.
Interim analyses may be conducted for all included individuals to generate data for disease-specific time periods (ie, every 4 months).
