Waning effectiveness of BNT162b2 and ChAdOx1 COVID-19 vaccines over six months since second dose: a cohort study using linked electronic health records

Background The rate at which COVID-19 vaccine effectiveness wanes over time is crucial for vaccination policies, but is incompletely understood with conflicting results from different studies. Methods This cohort study, using the OpenSAFELY-TPP database and approved by NHS England, included individuals without prior SARS-CoV-2 infection assigned to vaccines priority groups 2-12 defined by the UK Joint Committee on Vaccination and Immunisation. We compared individuals who had received two doses of BNT162b2 or ChAdOx1 with unvaccinated individuals during six 4-week comparison periods, separately for subgroups aged 65+ years; 16-64 years and clinically vulnerable; 40-64 years and 18-39 years. We used Cox regression, stratified by first dose eligibility and geographical region and controlled for calendar time, to estimate adjusted hazard ratios (aHRs) comparing vaccinated with unvaccinated individuals, and quantified waning vaccine effectiveness as ratios of aHRs per-4-week period. The outcomes were COVID-19 hospitalisation, COVID-19 death, positive SARS-CoV-2 test, and non-COVID-19 death. Findings The BNT162b2, ChAdOx1 and unvaccinated groups comprised 1,773,970, 2,961,011 and 2,433,988 individuals, respectively. Waning of vaccine effectiveness was similar across outcomes and vaccine brands: e.g. in the 65+ years subgroup ratios of aHRs versus unvaccinated for COVID-19 hospitalisation, COVID-19 death and positive SARS-CoV-2 test ranged from 1.23 (95% CI 1.15-1.32) to 1.27 (1.20-1.34) for BNT162b2 and 1.16 (0.98-1.37) to 1.20 (1.14-1.27) for ChAdOx1. Despite waning, rates of COVID-19 hospitalisation and COVID-19 death were substantially lower among vaccinated individuals compared to unvaccinated individuals up to 26 weeks after second dose, with estimated aHRs <0.20 (>80% vaccine effectiveness) for BNT162b2, and <0.26 (>74%) for ChAdOx1. By weeks 23-26, rates of SARS-CoV-2 infection in fully vaccinated individuals were similar to or higher than those in unvaccinated individuals: aHRs ranged from 0.85 (0.78-0.92) to 1.53 (1.07-2.18) for BNT162b2, and 1.21 (1.13-1.30) to 1.99 (1.94-2.05) for ChAdOx1. Interpretation The rate at which estimated vaccine effectiveness waned was strikingly consistent for COVID-19 hospitalisation, COVID-19 death and positive SARS-CoV-2 test, and similar across subgroups defined by age and clinical vulnerability. If sustained to outcomes of infection with the Omicron variant and to booster vaccination, these findings will facilitate scheduling of booster vaccination doses.


Background
The rate at which COVID-19 vaccine effectiveness wanes over time is crucial for vaccination policies, but is incompletely understood with conflicting results from different studies.

Methods
This cohort study, using the OpenSAFELY-TPP database and approved by NHS England, included individuals without prior SARS-CoV-2 infection assigned to vaccines priority groups 2-12 defined by the UK Joint Committee on Vaccination and Immunisation. We compared individuals who had received two doses of

Interpretation
The rate at which estimated vaccine effectiveness waned was strikingly consistent for COVID-19 hospitalisation, COVID-19 death and positive SARS-CoV-2 test, and similar across subgroups defined by age and clinical vulnerability. If sustained to outcomes of infection with the Omicron variant and to booster vaccination, these findings will facilitate scheduling of booster vaccination doses.
Examination of waning COVID-19 vaccine effectiveness is difficult. The success of vaccine rollouts in many countries means that only a small and selected proportion of the population remains unvaccinated.
Continuing uptake of vaccination further depletes this unvaccinated group over time. Vaccines were offered in priority order determined by age and clinical vulnerability, so that the longest follow up is in people at highest risk of severe COVID-19. Rapid changes in rates of SARS-CoV-2 infection over time, related to pandemic control measures and introduction of new variants, make it essential to account for the calendar date on which events occurred.
Many studies of waning COVID-19 vaccine effectiveness have used "test-negative case-control" (TNCC) designs, restricted to people tested for infection with SARS-CoV-2 and comparing those testing positive (cases) and negative (controls), 5,9,17,18 or reported indirect evidence such as changing rates of COVID-19 with time since vaccination. 11,12 The extent to which TNCC designs control bias due to confounding, or are biased because of the restriction to people who were tested, remains unclear. 19,20 We conducted a cohort study within the OpenSAFELY-TPP database (https://opensafely.org), which includes detailed linked data on 24 million people registered with an English general practice (GP) using TPP SystmOne electronic health record (EHR) software. We compared rates of COVID-19 hospitalisation, COVID-19 and non-COVID-19 mortality, and infection with SARS-CoV-2, between adults fully vaccinated with the Pfizer-BioNTech BNT162b2 mRNA vaccine (BNT162b2) and the Oxford-AstraZeneca ChAdOx1 nCoV-19 AZD1222 (ChAdOx1), and those who were unvaccinated.

Data source
OpenSAFELY-TPP includes detailed pseudonymized primary care data linked (via National Health Service (NHS) number) with accident and emergency attendance, inpatient hospital spell records (NHS Digital's Hospital Episode Statistics dataset), national SARS-CoV-2 testing records (Second Generation Surveillance . CC-BY 4.0 International license It is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

Study design
Individuals eligible for this cohort study, which was approved by NHS England, were adults who had been assigned to UK Joint Committee on Vaccination and Immunisation (JCVI) priority groups 2 to 12 (Supplementary Table 1) and registered with a primary care GP for ≥1 year before eligibility for their first vaccine dose (the "eligibility date", based on JCVI group and age, Supplementary Table 2). Individuals were assigned to JCVI groups based on information in their linked EHR. They were excluded if their sex, geographical region, ethnicity or index of multiple deprivation were unknown; or they were resident in a care home at six weeks after their eligibility date. Full details are in Supplementary Figure 1.
We defined three groups who: (1) received two doses of BNT162b2; (2) received two doses of ChAdOx1; (3) were unvaccinated. Eligibility for the vaccinated groups was restricted to those who received their second vaccine dose during a 4-week "second vaccination period" (SVP) within analysis strata defined by JCVI group, eligibility date (for groups within which eligibility was based on age (Supplementary Table 2)), and English NHS region, defined using individuals' GP address). The SVP was defined as the 28-day period during which the greatest number of individuals in the stratum received their second dose. Individuals were excluded from the vaccine groups if they: received their first dose before their eligibility date; had an interval between first and second dose of <6 or >14 weeks; or were flagged as a healthcare worker on their vaccination record. Individuals were assigned to the unvaccinated group if they had received no COVID-19 vaccine at the start of the SVP for their analysis stratum. Individuals were excluded from any group if they had: evidence of previous SARS-CoV-2 infection by the start of their SVP; ever been recorded as being resident in a care home; or evidence of having started an end-of-life care pathway. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 23, 2022. ; https://doi.org/10.1101/2022.03.23.22272804 doi: medRxiv preprint SVP, and split into the three consecutive 8-week calendar periods during which vaccinated individuals were followed in each comparison period. Unvaccinated individuals assigned to start at 2 weeks were followed during comparison periods 1, 3 and 5, while those assigned to at 6 weeks were followed during comparison periods 2, 4 and 6.

Outcomes
The outcomes were COVID-19 hospitalisation (identified using HES inpatient hospital records), COVID-19 death, positive SARS-CoV-2 test, and non-COVID-19 death. We also investigated test-seeking behaviour by comparing rates of testing for SARS-CoV-2 between the vaccine groups. COVID-19 and non-COVID-19 deaths (death certificates with and without a COVID-19 code) were based on death registry data from the Office for National Statistics. SARS-CoV-2 tests were identified using SGSS records and based on swab date.
Both polymerase chain reaction (PCR) and lateral flow tests were included, without differentiation between symptomatic and asymptomatic infection. All outcomes were defined by the date of their first occurrence during the comparison follow-up period. Where there was no positive SARS-CoV-2 test, but a record of COVID-19 hospitalisation and/or COVID-19 death, the date of positive SARS-CoV-2 test was imputed as the date of COVID-19 hospitalisation or date of COVID-19 death.

Potential confounding factors
The following potential confounders were defined at a single time (typically the day before the start of comparison period 1; full details in Supplementary Table 3 shielding; history of chronic heart disease, kidney disease, liver disease, neurological disease or respiratory disease; history of diabetes, learning disability, serious mental illness; current immunosuppression or history of permanent immunosuppression; and number of conditions in the clinically "at risk" classification (according to national prioritisation guidelines). After exclusions for missing values in demographic variables, there were no missing values in the remaining variables as they were defined as presence or absence of codes in the EHR.

Statistical analysis
For each comparison period we estimated hazard ratios (HRs) comparing: (1) BNT162b2 vs unvaccinated; is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 23, 2022. ; https://doi.org/10.1101/2022.03.23.22272804 doi: medRxiv preprint up ended at the "outcome of interest" only when the test result was positive. Fully vaccinated individuals who received a booster dose, and unvaccinated individuals who received a first dose, were excluded from subsequent comparison periods, but follow-up within comparison periods was not censored after these events.
To estimate hazard ratios, we fitted Cox regression models with baseline hazards stratified by JCVI group, eligibility date and region used to define the SVPs, and with the covariates described above. To avoid issues with model convergence, binary covariates were excluded from the model if there were <3 events in any cell of the table defined by cross tabulating the covariate with vaccine group and comparison period. For categorical covariates with more than two levels, levels were merged until either all levels had greater than three events, or there was only one level, in which case the variable was excluded. This process was carried out independently for each outcome. Age within strata was modelled as linear, with quadratic terms additionally included for strata with age range >5 years. We used meta-regression to quantify waning effectiveness as estimated ratios of HRs per comparison period.
All analyses were done independently in four vaccine priority subgroups: (A) aged ≥65 years and in JCVI This study followed STROBE-RECORD reporting guidelines. Any counts below six were redacted or rounded for disclosure control. The funders had no role in the study design, collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the article for publication.

Results
Of 13,923,580 individuals satisfying initial eligibility criteria (Supplementary Figure 2), 4,780,020 received second doses of BNT162b2 or ChAdOx1 during the SVP for their stratum (Supplementary Figures 3-17) and 2,596,920 were unvaccinated at the start of their SVP. Of these, 1,773,970, 2,961,011 and 2,433,988 were included in the first comparison period BNT162b2, ChAdOx1 and unvaccinated groups respectively. Table 1 and Supplementary Table 4 show summary statistics for these three groups by subgroup. Compared to vaccinated individuals, unvaccinated individuals were less likely to be white; live in a more affluent area; have had a flu vaccine in the previous five years or have tested for SARS-CoV-2 before their eligibility date.
The distribution of other characteristics between vaccine groups differed according to subgroup. For example, in the 65+ subgroup, those vaccinated with BNT162b2 were older than those vaccinated with . CC-BY 4.0 International license It is made available under a perpetuity.
is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The cumulative incidence of first dose by weeks 23-26 in previously unvaccinated individuals was 17%, 27%, 12% and 13% in the 65+, 16  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint  Table 8).
There were 7,318, 7,646, and 3,237 non-COVID-19 deaths in BNT162b2 recipients, ChAdOx1 recipients and unvaccinated individuals, respectively (Supplementary Table 5). Across subgroups, estimated aHRs during  is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint  Table 5). Across subgroups, rates of testing during weeks 3-6 were broadly similar in When quantified as ratios of aHRs, the waning of estimated effectiveness was strikingly similar across risk groups, except that estimated effectiveness waned fastest in the 18-39 subgroup (those at lowest risk of severe COVID-19, all vaccinated with BNT162b2). In those subgroups in which the two vaccines could be compared, estimated effectiveness was initially greater for BNT162b2 than for ChAdOx1, but effectiveness waned somewhat faster for BNT162b2 than ChAdOx1, so that the two brands' comparative effectiveness became more similar over time. Estimated aHRs for COVID-19 hospitalisation and COVID-19 death is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint  , and by 21 percentage points (95% CI 14-30) for SARS-CoV-2 infection, from 1 to 6 months after full vaccination. Estimates varied substantially between studies with reductions comparable to those in our study reported by only two studies. 9,14 The review concluded that the decline in vaccine effectiveness against severe COVID-19 was less than for SARS-CoV-2 infection and symptomatic disease, whereas we found that rates of waning, quantified as ratios of HRs, were similar for both outcomes across the four subgroups and that estimated vaccine effectiveness against COVID-19 hospitalisation and COVID-19 death remained high 23-26 weeks after receipt of the second dose.
As in this study, Andrews and colleagues analysed NHS England EHR data to investigate of the duration of protection by COVID-19 vaccines against symptomatic and severe COVID-19. 5 They used a TNCC design, restricted to individuals who were tested for SARS-CoV-2 infection. In the 65+ and 40-64 CV subgroups, identically defined in the two studies, estimated vaccine effectiveness against COVID-19 hospitalisation and COVID-19 death was consistently 3-6% lower in this study than that of Andrews and colleagues, who concluded (by contrast with our study) that waning was greater for ChAdOx1 than BNT162b2, and greater among older adults and those in a clinical risk group. While those authors found continuing vaccine effectiveness against symptomatic COVID-19 up to 26 weeks since receipt of the second vaccine dose, we found rates of SARS-COV-2 infection to be similar to or higher than those in unvaccinated individuals by that time.
Our study is based on whole-population data analysed within the OpenSAFELY Trusted Research Environment, which has stringent disclosure controls to protect patient privacy. The large study size and large numbers of outcome events led to precise estimates of vaccine effectiveness according to vaccine brand and time since second vaccine dose. We accounted for risk-dependent vaccine allocation by separating the cohort into subgroups based on JCVI group, 26 and conducting analyses within strata defined by JCVI group, eligibility date for primary vaccination and geographical region. Our analyses also excluded individuals with a pre-vaccine rollout record of SARS-CoV-2 infection and accounted for rapid changes in COVID-19 incidence with calendar time, censoring due to occurrence of outcome events, and attenuation of comparison groups because receipt of receipt of first vaccine dose by unvaccinated individuals and third dose by fully vaccinated individuals. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint Our study has several limitations. First, as in any observational study, our estimates could be affected by confounding by unmeasured factors. However, the detailed linked data analysed permitted adjustment for a wide range of potential confounding factors. Second, patients registered with a GP who have moved or emigrated (or whose death was not recorded) 25 may contribute person-time but not events. Because the BNT162b2 and ChAdOx1 groups are defined by recent vaccination, these "ghost" patients are more likely to be present in the unvaccinated group, leading to bias in estimates of waning. Also, healthcare workers could be identified and excluded from the vaccinated groups because this information was recorded at the time of vaccination, but not from the unvaccinated group. This limitation should not affect results for the 65+ subgroup, most of whom are retired, or comparisons between BNT162b2 and ChAdOx1. Third, consistent with an Australian survey, 24 we found that unvaccinated individuals had tested less frequently than vaccinated individuals during the pre-vaccine rollout period when widespread testing was available, and were considerably less likely to be tested during follow-up. Fourth, differential depletion of susceptible people in the unvaccinated groups over time may lead to attenuation of HRs even when true vaccine effectiveness does not change. However, such bias is likely to be minimal when vaccine effectiveness is high. 27 Our results have immediate implications for COVID-19 vaccination policies. When quantified as ratios of HRs, the rate at which estimated vaccine effectiveness waned was strikingly consistent (there was little variation around the fitted rates of waning displayed in Figure 2) and (by contrast with other studies) similar across subgroups defined by age and clinical vulnerability. If sustained to outcomes of infection with the Omicron variant and to booster vaccination, these findings will facilitate scheduling of booster vaccination doses. By 26 weeks after second dose, rates of infection with SARS-CoV-2 were similar to or higher in fully vaccinated than unvaccinated individuals, implying that vaccination has only transient impacts on transmissibility of SARS-CoV-2. This may arise partly because a desirable consequence of vaccination is to facilitate greater social mixing because of the reduced risk of severe COVID-19. Protection against COVID-19 hospitalisation and COVID-19 death was substantial up to 26 weeks after second vaccination, even in older and clinically vulnerable individuals. Finally, cessation of freely-available population-based testing programmes is likely to limit applications of the TNCC design, which have to date provided rapid estimates of vaccine effectiveness. By contrast, cohort approaches based on detailed linked EHR data, such as were used in this study, should remain feasible for severe COVID-19 outcomes. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint

Contributors
The copyright holder for this this version posted March 23, 2022. ; https://doi.org/10.1101/2022.03.23.22272804 doi: medRxiv preprint and BG acquired funding for the study. WJH, JM, AM, BG contributed to project administration. SB and ID contributed resources for the study. SB, and ID created and maintained software. SB, DE, PI, CEM, GH, SD, TW, and ID, created and maintained software. JACS, WJH, JM, AM, and BG supervised the study. AM, and BG were responsible for work relating to information governance. EMFH and JACS wrote the first draft of the manuscript. All authors contributed to reviewing and editing of the manuscript. All authors had final responsibility for the decision to submit for publication. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint We are very grateful for all the support received from the EMIS and TPP Technical Operations team throughout this work, and for generous assistance from the information governance and database teams at NHS England / NHSX.

Information governance and ethical approval
NHS England is the data controller for OpenSAFELY-TPP; TPP is the data processor; all study authors using OpenSAFELY have the approval of NHS England. This implementation of OpenSAFELY is hosted within the TPP environment which is accredited to the ISO 27001 information security standard and is NHS IG Toolkit compliant. 28,29 Patient data has been pseudonymised for analysis and linkage using industry standard cryptographic hashing techniques; all pseudonymised datasets transmitted for linkage onto OpenSAFELY are encrypted; access to the platform is via a virtual private network (VPN) connection, restricted to a small group of researchers; the researchers hold contracts with NHS England and only access the platform to initiate database queries and statistical models; all database activity is logged; only aggregate statistical outputs leave the platform environment following best practice for anonymisation of results such as statistical disclosure control for low cell counts. is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint (COPI) to require organisations to process confidential patient information for the purposes of protecting public health, providing healthcare services to the public and monitoring and managing the COVID-19 outbreak and incidents of exposure; this sets aside the requirement for patient consent. 31 Taken together, these provide the legal bases to link patient datasets on the OpenSAFELY platform. GP practices, from which the primary care data are obtained, are required to share relevant health information to support the public health response to the pandemic, and have been informed of the OpenSAFELY analytics platform.
This study was approved by the Health Research Authority (REC reference 20/LO/0651) and by the LSHTM Ethics Board (reference 21863).    is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint The copyright holder for this this version posted March 23, 2022. ; https://doi.org/10.1101/2022.03.23.22272804 doi: medRxiv preprint