key: cord-0702667-kgcukmk5
authors: Glatman-Freedman, Aharona; Bromberg, Michal; Dichtiar, Rita; Hershkovitz, Yael; Keinan-Boker, Lital
title: The BNT162b2 vaccine effectiveness against new COVID-19 cases and complications of breakthrough cases: A nation-wide retrospective longitudinal multiple cohort analysis using individualised data
date: 2021-09-17
journal: EBioMedicine
DOI: 10.1016/j.ebiom.2021.103574
sha: 7827709b9e610c20357fa135bc9d30360c9f9636
doc_id: 702667
cord_uid: kgcukmk5

BACKGROUND: The rapid vaccination campaign against COVID-19 in Israel relied on the BNT162b2 vaccine. We performed a longitudinal analysis of multiple cohorts, using individual data, to evaluate the effectiveness of the vaccine against new and breakthrough cases. METHODS: We estimated vaccine effectiveness (VE) for 27 consecutive cohorts, each comprised of individuals vaccinated on specific days. VE against new COVID-19 cases was evaluated for five SARS-CoV-2-related outcomes: infection, symptomatic disease, hospitalisation, severe/critical disease and death. For breakthrough cases, rate reduction was evaluated for hospitalisation, severe/critical disease and death. Outcomes were evaluated at predetermined time-periods after vaccination, the last one dedicated to individuals who became SARS-CoV-2-positive 22–28 days after the second dose. FINDINGS: The highest VE estimates against new cases in ≥16 year old individuals, for all outcomes, were reached at the 15–21 day period after the second dose, ranging between 97.7% (95% CI: 95.9–98.7%) for deaths and 98.6% (95% CI: 97.8–99.1%) for severe/critical disease. VE estimates of the 14–20 day period after the first dose ranged between 54.3% (95% CI: 50.6–57.8%) for infection and 77.3% (95% CI: 71.2–82.1%) for severe/critical disease. VE rose more slowly among ≥80 year old individuals. Rate reductions of breakthrough complications were highest at the 22–28 day period after the second dose, ranging between 47.4% (95% CI: 4.3–71.2%) for death and 66.2% (95% CI: 44.2–79.6%) for severe/critical disease. INTERPRETATION: The BNT162 vaccine is highly effective in preventing new SARS-CoV-2 cases. Among ≥80 year old individuals, high effectiveness develops more slowly. In breakthrough cases, vaccination reduces complications and death. FUNDING: None.

Several vaccines against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) were granted Emergency Use Authorisation (EUA) by the Food and Drug administration (FDA) and Conditional Marketing Administration (CMA) by the European Medicines Agency (EMA) [1, 2] .

Israel initiated its public vaccination campaign against SARS-CoV-2 on December 20, 2020, prioritising first health care workers, individuals 60 years old and above, and individuals residing in long-term care facilities. As the vaccination campaign progressed, additional age groups and priority groups were gradually vaccinated as well. The Public vaccination campaign relied on the BNT162b2 vaccine. By May 21, 2021, a total of 5,439,734 individuals received the first dose, and 5,112,516 individuals received the second dose of the BNT162b2 vaccine. These constitute 58.4% and 54.9% of the entire Israeli population and 83.2% and 78.2% of the Israeli population 16 years of age, respectively, placing Israel as a leading country in population vaccine coverage [3] .

The efficient vaccination rollout in Israel led to rapid changes in vaccine coverage, as well as changes in viral circulation. As such, monitoring vaccine effectiveness (VE) estimates over time, among individuals that were vaccinated on particular dates and were subjected to similar external conditions, such as vaccine coverage and virus circulation, is important. Although a study based on the Israel national database evaluating VE following the administration of two vaccine doses, was recently reported, it was performed using aggregate rather than individual data, and did not evaluate discrete time periods following each vaccine dose [4] .

Moreover, a comprehensive analysis of the effect of the vaccine on individuals who became SARS-CoV-2-positive despite being vaccinated (breakthrough cases), in terms of hospitalisations and death reduction, has not been reported to date.

In randomised clinical trials the BNT162b2 vaccine demonstrated 52% and 95% vaccine efficacy after the first and second vaccine doses, respectively, in individuals 16 years old [5] . Characterising the effectiveness of the BNT162b2 vaccine is of utmost importance, because it provides information on the performance of the vaccine under real-life circumstances. Multiple factors can potentially affect the performance of the BNT162b2 vaccine, among them, demographics, usage in patients with multiple or unstable chronic medical conditions, inadequate vaccine storage or transport conditions [6] , varying degrees of SARS-CoV-2 in the population, variations in vaccine coverage and the appearance of new genomic variants [7À12] .

The purpose of this study was to evaluate the effectiveness of the BNT162b2 vaccine against new cases and against complications of breakthrough SARS-CoV-2 cases by performing a retrospective longitudinal multiple cohort study, using national data of the mass BNT162b2 vaccination campaign. We analysed the dynamics of vaccine effectiveness against multiple outcomes over time, for 16 year old individuals and by age groups.

Israel is a country with a population of 9.3 million [13] . It has universal health coverage provided to all residents by four Health Maintenance Organisations (HMOs). The BNT162b2 vaccine has been administered in Israel by these four HMOs, by medical centres, longterm care facilities, emergency medical services and the armed forces. Individuals who had a documented positive SARS-CoV-2 test or an MOH-approved SARS-CoV-2 serology test confirming past infection, were not eligible to receive the vaccine initially. Each SARS-CoV-2 vaccine administration and all SARS-CoV-2 Polymerase Chain Reaction (PCR) test results have been registered electronically into two separate designated national databases. Both SARS-CoV-2 PCR testing and vaccines have been provided free of charge.

We performed a retrospective longitudinal cohort study, using two MOH national databases: the SARS-CoV-2 vaccine database and the SARS-CoV2 tests database. The national SARS-CoV-2 vaccine database includes the name and lot number of the vaccine administered and the date of administration for each individual. The national SARS-CoV2 PCR tests database includes the results of each test performed, the date of testing, and the date of results for each individual. It also included the presence or absence of symptoms, date of hospitalisation, severity of illness and date of death, if applicable.

Each resident in Israel has a unique personal identity number (UPIN), which was used in both databases. Data retrieved from both databases, were cross-referenced by using individuals' twiceencrypted UPINs; thus, the study database was fully-de-identified to the researchers.

The analyses of this study were performed on 27 consecutive cohorts. Analyses for the first vaccine dose were performed for cohorts that received the first vaccine dose from 21 December 2020 to 16 January 2021 (Fig. 1A) . Analyses for the second vaccine dose were performed for individuals who received the second dose from January 11 to February 6, 2021 (Fig. 1A) .

Our analysis included two main steps. The first step consisted of determining VE against specific outcomes for 27 cohorts, during specific time-periods following the receipt of the first and second doses. The second step included analysis of rate reductions of specific outcomes for cohort individuals who became SARS-CoV-2-positive following the first or second vaccine dose (breakthrough cases).

Vaccine effectiveness estimation. Individuals included in the analysis were Israeli residents without a documented PCR-confirmed

Evidence before this study A rapid BNT162b2 mRNA vaccination campaign started in Israel on December 20, 2021. The vaccine was reported to be highly effective against symptomatic laboratory-confirmed COVID-19 in a randomised controlled trial and was approved initially for use in individuals aged 16 years and older. We searched Pubmed and preprint servers for articles published between December 11, 2020 (date of first Emergency Use Authorisation by the FDA) and July 21, 2021, to identify reports of the effect of the BNT162b2 vaccine in prevention of new COVID-19 cases and on the outcome of disease diagnosed among individuals who had received the BNT162b2 vaccine. While studies estimating the effectiveness of the BNT162b2 vaccine were reported, a nationwide evaluation of the precise dynamics of vaccine effectiveness estimates over time following the administration of one or two vaccine doses, across multiple age groups and multiple outcomes, using multiple cohorts at welldefined one-week evaluation periods was not reported to-date. Furthermore, no studies evaluated the effect of SARS-CoV-2 vaccines on the outcome of disease diagnosed among individuals who were vaccinated.

Our analysis, which was based on nationwide individualised data recorded in designated national repositories, followed multiple cohorts. It demonstrated modest effectiveness with high variability after the first vaccine dose and high effectiveness with low variability after the second vaccine dose against five SARS-CoV-2-related outcomes: infection, symptomatic disease, hospitalisation secere/critical disease and death. Moreover, our analysis demonstrated for the first time, the reduction in SARS-CoV-2-related hospitalisations, severe/critical disease and death among individuals who became SARS-CoV-2-positive following the receipt of one or two BNT162b2 vaccine doses. Finally, our study demonstrates for the first time, that vaccine effectiveness estimates rise more slowly among individuals 80 years old and above. Each cohort was followed-up for up to 68 days after the second dose.

The high effectiveness of the BNT162b2 vaccine against SARS-CoV-2-related outcomes, coupled with the growing number of SARS-CoV-2 hotspots around the world, highlights the urgency of making the COVID-19 vaccines available worldwide. The reduction of severe SARS-CoV-2-related outcomes among breakthrough cases provides further support for the importance of vaccination. The slower rise in vaccine effectiveness among individuals 80 years old and older, suggests considering them fully vaccinated later than younger individuals. SARS-CoV-2 test prior to the evaluation periods. VE was determined for 16 year old individuals and for pre-determined age group (16À39, 40À59, 60À79 and 80 years old).

VE was evaluated against the following in laboratory-confirmed SARS-CoV-2 outcomes: infection, symptomatic disease, hospitalisations, severe/critical disease severity and death. This analysis was performed on each of the selected 27 cohorts and for all of them combined (combined VE). VE against laboratory-confirmed SARS-CoV-2 related outcomes were estimated for individuals who became SARS-CoV-2-positive during the following four evaluation periods: days 14-20 after receipt of the first dose (first dose evaluation period) (Fig. 1B) and days 8À14, 15À21 and 22À28 after receipt of the second dose (second dose evaluation periods) (Fig. 1C) . Days 1-13 after the first dose and 1-7 after the second dose were designated 'waiting periods' (Fig. 1B, 1C) .

For VE estimation against hospitalisations, severe/critical disease and death, the time-period allotted for their occurrence following the first positive PCR test were determined based on frequency histograms (Fig. S4) .

Disease severity was determined based on the National Institutes of Health guidelines [14] .

The number of individuals in the 'unvaccinated group' for each cohort was derived by omitting the following individuals from the total number of Israeli residents [15] : individuals who had a documented PCR-confirmed SARS-CoV-2 test prior to the evaluation period, individuals who were vaccinated before and on the relevant vaccination date, and during the relevant waiting period. Individuals who were vaccinated on a cohort vaccination date, during the cohort waiting period and the cohort evaluation period were moved to the 'vaccinated group' of another cohort based on their vaccination date.

The number of Israeli residents (total, by age, by sex) was based on the 2020 Central Bureau of Statistics (CBS) statistical abstract [15] .

Inclusion and criteria for VE analysis are summerised in Tables S1-S4 of the supplementary index.

Hospitalisations, severe/critical disease and death among SARS-CoV-2-positive vaccinated individuals. Rates of SARS-CoV-2-related hospitalisations, severe/critical disease and death were determined for vaccinated individuals belonging to the 27 cohorts of the first vaccine dose and the second vaccine dose who became SARS-CoV-2-positive by PCR during the evaluation periods detailed above (breakthrough cases), and were compared to those of unvaccinated individuals. The time-periods allotted for the appearance of these outcome following the first positive PCR test were based on frequency histograms (Fig. S4) as described above, and adjusted rate reductions (1-RR) were calculated for each.

VE and 95% confidence interval was estimated using (1-IRR)x100 where IRR, the Incidence Rate Ratio, represents the ratio of PCR-confirmed SARS-CoV-2 cases rate in the vaccinated group to the equivalent rate in the unvaccinated control group.

For individuals with more than one positive SARS-CoV2 test during the evaluation period, only the first one was considered in the analysis.

Individuals who had a positive SARS-CoV-2 PCR test prior to the evaluation periods were excluded from analysis, irrespective of whether or not they were vaccinated with the BNT162b2 vaccine or not. The number of unvaccinated controls for each cohort was calculated by omitting the number of Israeli residents who received the relevant BNT162b2 vaccine dose on the cohort vaccination date from the total number of Israeli residents that did not have a documented positive SARS-CoV-2 test by that date. In addition, the number of person-days each individual contributed as unvaccinated during each evaluation period was calculated. Thus, individuals who belonged to the 'unvaccinated' group on the cohort vaccination date, but received the BNT162b2 vaccine during the evaluation period, were transferred to the 'vaccinated group' of a new cohort based on their vaccination date. The person-days they contributed to the evaluation period prior to their vaccination date were included in those of the unvaccinated group.

The number of Israeli residents (total, by age, by gender) was based on the 2020 Central Bureau of Statistics (CBS) statistical abstract [15] .

The VE was calculated for each cohort for the evaluation periods determined in the study design.

To calculate combined VE, the following steps were taken: A. The number of vaccinated/unvaccinated SARS-CoV-2 positive cases for each outcome of each evaluation period was summed. B. The number of person-days each individual was vaccinated/unvaccinated, without becoming SARS-CoV-2-positive, were counted once for all 27 cohorts. The number of days in which each individual was vaccinated/unvaccinated without becoming SARS-CoV-2-positive for each cohort during an evaluation period were summed to give the number of person-days in the vaccinated/unvaccinated status of individuals included in all 27 cohorts.

IRR was then determined for the 27 cohorts combined. Reduction in SARS-CoV2 hospitalisations, illness severity during hospitalisations and mortality in individuals who received the BNT162b2 vaccine were evaluated using 1-IRR.

IRR and 95% confidence interval (95% CI) were adjusted for agegroup (16À19, 20À29, 30À39,40À49,50À59, 60À69, 70À79 and 80 years old), sex, and calendar week of vaccination by using Poisson regression. The offset of the Poisson regression was the number of vaccinated individuals for the SARS-CoV-2 positive vaccinated cases, and number of unvaccinated individuals for the SARS-CoV-2positive unvaccinated individuals, included in the analysis. In cases of over dispersion, negative binomial regression was used. Adjustment for fewer variables was performed as required by the data size.

Statistical analysis was performed using SAS Enterprise Guide 7.1 (SAS Institute Inc.) software.

Our study followed the Strengthening the Reporting of Observational studies in Epidemiology guidelines [16] .

The study was approved by the superior ethical committee of the Israel MOH, under Helsinki protocol number CoR-MOH-081-2021, and included exemption from informed consent.

Not relevant.

The progress of two-dose vaccine coverage by age group is demonstrated in Fig. S1 . It is consistent with age prioritisation by the MOH, with individuals 60 years old being the first to reach and surpass vaccine coverage of 60 %. Fig. 2 demonstrates the adjusted VE against outcomes of interest for each of the 27 cohorts for the first three pre-determined evaluation periods. The highest VE estimates with the lowest variability were reached 15À21 days after the second dose. Similarly high VE estimates were observed 22À28 days after the second dose (data not shown). The greatest variability in VE estimates was observed on days 14À20 after the first vaccine dose.

Tables 1À4 demonstrate the adjusted VE against outcomes of interest of all 27 cohorts combined, for the four evaluation periods. The highest combined VE estimates for all outcomes of 16 year old individuals were reached 15À21 days after the second dose (Table 3) , and were maintained at similar levels 22À28 days after the second dose ( Table 4 ). The largest change in VE against all outcomes was observed between days 14À20 after the first dose and days 8À14 after the second dose, for all age groups (Tables 1À4). The VE differences between days 14À20 after the first dose and days 8À14 after the second dose, and the VE differences between days 8À14 and 15À21 after the second dose were statistically significant (Tables 1À3).

The combined adjusted VE by age group for all 27 cohorts against the five outcomes are demonstrated in Tables 1À4. VE estimates against laboratory-confirmed SARS-CoV2Àrelated outcomes for age groups 16À39, 40À59 and 60 to 79 years old reached or surpassed 90%, 8À14 days following the second vaccine dose ( Table 2 ). The highest VE estimates for the above age groups were observed on days 15 to 21 (Table 3) , and were maintained at similar levels on days 22À28 (Table 4 ) following the second vaccine dose.

For 80 year old individuals, VE estimates were lower than those of other age groups, for the first three evaluation periods (Tables 1À3). These differences were statistically significant for most estimates of the 8À14 and 15À21 days evaluation periods (Tables 2  and 3 ). VE estimates against laboratory-confirmed SARS-CoV2Àrelated outcomes for 80 year old individuals reached or surpassed 90% 15À21 days following the second vaccine dose ( Table 3 ). The highest VE estimates of 80 year old individuals were observed 22À28 days after the second dose (Table 4 ) and were then closest to VE estimates of the other age group ( Table 4) .

The VE differences between days 14-20 after the first dose and days 8À14 after the second dose, and the VE differences between days 8-14 and 15À21 after the second doses were statistically significant for most age groups and for most outcomes (Tables 1À3). Table 5 demonstrates the results of hospitalisations, severe/critical disease and death among vaccinated SARS-CoV-2-positive individuals.

A total of 7,166 individuals from the first dose 27 cohorts turned SARS-CoV-2-positive 14À20 days after receipt of the first dose. A total of 1,639, 542 and 430 from the second dose 27 cohorts became positive on days 8À14, 15À21 and 22À18 after the receipt of the second dose (Table 5) .

Due to the relatively small number of individuals who became SARS-CoV-2-positive, the analysis was performed for three age groups only: 16 years old, 60 years old and 80 years old.

The rate reductions for hospitalisations, severe/critical disease and deaths for 16 year old individuals who became SARS-CoV-2-positive on days 14À20 after the first vaccine dose were 44.2% (95% CI: 27.3À57.3%), 46.8% (95% CI: 32.9À57.9%) and 36.4% (95% CI: 18.6À50.4%), respectively (Table 5) . , hospitalisations (panels c, h, m), severe/critical disease (panels d, i, n) and death (panels e, j, o). Panels a-e represent outcomes that occurrred on days 14À20 after the first vaccine dose; panels fÀj represent outcomes that occurred 8À14 after the second dose; panels kÀo represent outcomes that occurred 15À21 after the second vaccine dose dose. Statistical adjustment was performed for age and sex. Asterix (*) represents abrreviation for severe/critical disease. The rate reductions for hospitalisations, severe/critical disease and deaths for individuals who became SARS-CoV-2-positive on days 22À28 after the first vaccine dose were 56.1% (95% CI: 35.0À70.4%), 66.2% (95% CI: 44.2À79.6%) and 47.4% (95% CI: 4.3À71.2%), respectively (Table 5) .

Analysis by age groups demonstrated that the rate reductions for hospitalisations and severe/critical disease among 80 year old individuals were lower than other age categories during the first three evaluation periods. However, these differences were not statistically significant (Table 5 ).

In this study, we evaluated the effect of the BNT162b2 vaccine on both prevention of SARS-CoV-2-related outcomes and on outcomes that occurred despite vaccination. Individuals that are vaccinated at different times can be subjected to various external factors such as, inadequate vaccine storage or transport conditions [6] , varying degrees of SARS-CoV-2 in the population, variations in vaccine coverage and the appearance of new genomic variants [7À12]. Since such external factors cannot be readily adjusted for, we elected to estimate VE against SARS-CoV-2 for multiple cohorts, each representing individuals who were vaccinated on a separate date.

Our findings demonstrated high VE against laboratory-confirmed SARS-CoV-2 infection, symptomatic disease, hospitalisations, severe/ critical disease and death after the second BNT162b2 vaccine dose, with the highest VE point estimates reaching well over 90% during the 15À21 and the 22À28 day evaluation periods after the second vaccine dose. VE point estimates on days 14À20 after the first vaccine dose against SARS-CoV-2-related outcomes ranged between 54.3% and 77.3%. Our results indicate consistently that a high level of protection, with little variability, is provided by the BNT162b2 vaccine against COVID-19 after the second vaccine dose. We also demonstrated that VE estimates achieved after the first vaccine dose were more modest and had substantial variability.

Our VE estimates are consistent with the vaccine efficacy results published by the manufacturer [5] . Several other VE assessments also suggested that the two-dose regimen of the BNT162b2 vaccine is highly effective [17À19] and a single vaccine dose is less effective [18, 20, 21] . However, none of these assessments relied on nationallevel data. Although several VE studies from Israel were recently published, few important differences exist between those studies and our study. Two studies relied on a substantially smaller number of individuals as compared to our study, all belonging to specific HMOs [18, 19] . These studies demonstrated VE point estimates of 90%-92% against SARS-CoV-2 infection and 94% against symptomatic disease 7 days after the second vaccine dose [18, 19] . Although the HMO studies provided co-morbidity-related VE estimates, they did not provide age group-specific VE estimates against hospitalisations, severe/critical illness or death, nor did they provide VE analysis for specific timeintervals after the second vaccine dose. Another study from Israel relied on the national databases; however, it used aggregated data and assessed VE following the second vaccine dose only. Furthermore, it presented VE results of 7 days and 14 days after the second dose. As such, VE results of 7 days, included also the VE results of 14 days and could not present the real difference between VE results of 7-13 days and 14 days [4] . Lastly, our study is the first study to assess the effect of the vaccine on breakthrough cases.

The large number of vaccine recipients in our study, the careful determination of the necessary follow up time for each outcome and the ethical approval to examine individualised data, allowed us to perform a detailed analysis of VE in specific cohorts.

Our results demonstrated, for the first time, a delayed rise in VE estimates among 80 year old individuals, against all outcomes. Specifically, while VE estimates in <80 years old individuals reached VE of 90% on days 8À14 after the second vaccine dose, individuals 80 years of age reached that level of VE a week later (days 15À21 after the second dose). In fact, VE estimates among 80 year old individuals became closest to those of the other age groups, only for the 22À28 day evaluation period after the second vaccine dose.

This estimation is of particular importance since the oldest age group (75 years old), was substantially smaller than other age groups in the vaccine efficacy study conducted by the manufacturer [5] . Furthermore, the 80 year old age-group was not compared to other age groups over time in other VE studies [4, 18, 19, 22] .

Based on the manufacturer's vaccine efficacy study [5] , Israeli individuals have been considered fully vaccinated if seven days passed from the receipt of the second BNT162b2 vaccine dose [23] . However, the U.S. CDC recommends waiting two weeks before full vaccination can be assumed [24] . Our results support the U.S. CDC recommendation, and we suggest that full vaccination is defined once 14 days have passed since the second vaccine dose for all individuals, with special emphasis on 80 year old individuals.

Our study demonstrates, for the first time, the beneficial effects of the BNT162b2 vaccine on all relevant outcomes of interest in individuals 16 years old who became SARS-CoV-2-positive despite being vaccinated.

The rapid pace of the vaccination campaign in Israel allowed for the analysis of multiple cohorts, each representing a specific vaccination date. This form of analysis demonstrated the consistent nature of protection elicited by the vaccine through the rapidly changing circumstances of vaccine coverage and level of viral circulation that occurred during the vaccination campaign. Furthermore, each cohort is expected to have somewhat different population mix in terms of age, sex and socio-economic status.

Such analysis is of particular importance in view of the penetration of new genomic variants [7À12], their potential for rapid transmission [25] , and the concern that some may evade the immunity elicited by existing vaccines and lower VE. In this regard the Alpha variant (B.1.1.7) was the most prevalent genomic variant until May 2021 [4] . Although other variants of concerns (VOCs) were discovered in Israel [26À28] prior to May 2021, they did not spread widely in the population by the end of May 2021 [29] . A wide spread of VOCs in the population will require re-evaluation of VE. The worldwide spread of the highly transmissible Delta variant (B.1.617.2) [30] , underscores this issue.

Re-evaluation of VE may also be required because of the concern of waning vaccine-induced immunity. In this regard, waning of spike protein antibody levels was observed over-time following the administration of a second dose of SARS-CoV-2 vaccines [31] .

Our study has several limitations. Due to the nature of our research, the size of control unvaccinated population group was computed based on the Israel Bureau of statistics data, which included information regarding population size, age and sex. Specific individual data was available in our databases for unvaccinated individuals who became positive for SARS-CoV-2. Thus, despite the control group size computation, we were able to adjust age, sex and epidemiologic week, and our results are consistent with the manufacturer's vaccine efficacy results [5] .

The unavailability of information regarding underlying morbidity presents another limitation. However, the consistent VE results among multiple cohorts and multiple age groups, and the marked decrease in the incidence of SARS-CoV-2 cases in Israel indicate that the vaccine is highly effective for the population at large.

The prevalent SARS-CoV-2 PCR testing in hospitals, could have led to the inclusion of some individuals who were admitted to the hospital due to reasons other than COVID-19, but were found to be asymptomatic carriers of SARS-CoV-2. However, critical/severe disease was registered in our database only for hospitalised COVID-19 patients.

In this study, we used the date of positive SARS-CoV-2 PCR test, rather than the date of symptom onset, due to the fact that it was the most reliable and consistent datum in our database. Using the date of positive PCR test is valuable since a significant subset of individuals are asymptomatic. Furthermore, individuals can become symptomatic at different time-points after they were determined to be SARS-CoV-2-positive, thus their symptom onset date may not be available or could introduce other biases.

Since this is an observational study, PCR testing for SARS-CoV-2 could differ between vaccinated and unvaccinated individuals. Comparing the percent of PCR testing among fully vaccinated individuals and among unvaccinated individuals, demonstrated lower testing frequency among vaccinated individuals (not shown). Such differences can stem from reduced symptomatology among vaccine recipients and/or behavioural changes resulting from confidence in the vaccine. However, since SARS-CoV-2 PCR testing was widely performed among hospitalised individuals, VE against hospitalisations, severe/critical disease and death, were probably not affected by individuals' behaviour.

Cohort studies with active regular PCR testing regardless of the appearance of symptoms can eliminate testing differences. In this regard, a cohort study from Italy demonstrated VE of 93.7% (95%CI: 50.8À99.2%) of the BNT162b2 vaccine against laboratory confirmed SARS-CoV-2 infection 7 days after the second vaccine dose among health care workers that were PCR tested regularly [17] . Another study of US healthcare and emergency workers who received one of two mRNA vaccines found VE of 90% (95%CI: 68À97%) 14 days after the second vaccine dose against laboratory-confirmed SARS-CoV-2 infection [32] . A study from a large medical centre in Israel found an adjusted incidence rate ratio of 0.02 (95% CI: 0À0.07) for symptomatic disease and 0.09 (95% CI: 0.03À0.25) for asymptomatic infection following full BNT162b2 vaccination [33] .

The finding that the BNT162b2 vaccine is beneficial in terms of lowering hospitalisations, severe/critical disease and death among breakthrough SARS-CoV-2 cases, is important. However, a close examination of the breakthrough cases with regard to co-morbidities and genomic SARS-CoV-2 variants is required and should be the subject for future studies. Furthermore, the last evaluation period of our study included individuals who became SARS-CoV-2-positive 22À28 days after the second vaccine dose. Analysis of additional evaluation periods after vaccination is important in order to evaluate the long-term effect of the BNT162b2 vaccine.

Our results of high BNT162b2 VE against new COVID-19 cases and the beneficial effect among breakthrough cases, coupled with the reports of major COVID-19 hotspots around the world, strongly support the need to widen and to intensify vaccination efforts throughout the world in order to control the spread of SARS-CoV-2.

AG-F and LK-B conceived the study. AG-F designed the study, wrote the protocol, led data analysis and wrote the first draft of the manuscript. YH and RD retrieved the data and performed data analysis. AG-F, MB and LK-B interpreted the data and edited the final manuscript. YH and RD verified the underlying data. All authors revised the manuscript critically for important intellectual content and approved the final manuscript.

The authors declare no conflict of interest

Individual-level data used in this study cannot be publicly available due to material sensitivity. Requests for data can be made to the Ministry of Health of Israel.

Supplementary material associated with this article can be found in the online version at doi:10.1016/j.ebiom.2021.103574.

Food and Drug Administration. COVID-19 vaccines

Treatments and vaccines for COVID-19: authorised medicines

Our World in Data. Coronavirus (COVID-19) vaccination

Alroy-Preis S. Impact and effectiveness of mRNA BNT162b2 vaccine against SARS-CoV-2 infections and COVID-19 cases, hospitalizations and deaths followin nationwide vaccination campaign in Israel: an observational study using national surveillance data

Safety and efficacy of the BNT162b2 mRNA covid-19 vaccine

Centers for Disease Control and Prevention. Pfizer-BioNTech COVID-19 vaccine

Genetic variants are identified to increase risk of COVID-19 related mortality from UK Biobank data

The D614G mutation in the SARS-CoV2 Spike protein increases infectivity in an ACE2 receptor dependent manner. bioRxiv: the preprint server for biology

Introduction of the South African SARS-CoV-2 variant 501Y.V2 into the UK

Introduction of Brazilian SARS-CoV-2 484K.V2 related variants into the UK

SARS-CoV-2 viral variants-tackling a moving target

Emergence and rapid transmission of SARS-CoV-2 B.1.1.7 in the United States

COVID-19) treatment guidelines: clinical spectrum of SARS-CoV-2 infection

Population -statistical abstract of Israel -No. 71

The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies

Effectiveness of the comirnaty (BNT162b2, BioNTech/Pfizer) vaccine in preventing SARS-CoV-2 infection among healthcare workers, Treviso province

BNT162b2 mRNA covid-19 vaccine in a nationwide mass vaccination setting

The effectiveness of the TWO-DOSE BNT162b2 vaccine: analysis of real-world data

Interim findings from first-dose mass COVID-19 vaccination roll-out and COVID-19 hospital admissions in Scotland: a national prospective cohort study

Assessment of effectiveness of 1 dose of BNT162b2 vaccine for SARS-CoV-2 infection 13 to 24 days after immunization

Effectiveness of the Pfizer-BioNTech and Oxford-AstraZeneca vaccines on covid-19 related symptoms, hospital admissions, and mortality in older adults in England: test negative case-control study

What is a vaccination certificate?

Centers for Disease Control and Prevention. When you've been fully vaccinated. 2021

Efficacy and safety of the mRNA-1273 SARS-CoV-2 vaccine

44 confirmed cases of the South African variant have been detected so far

41 Indian variant cases detected, 24 of them among returning international passengers

For the first time: Brazilian and Chilean variants have been detected among returning international passengers

Share of COVID sequences that are the Delta variant

Confronting the delta variant of SARS-CoV-2, summer 2021

Spike-antibody waning after second dose of BNT162b2 or ChAdOx1

Interim estimates of vaccine effectiveness of BNT162b2 and mRNA-1273 COVID-19 vaccines in preventing SARS-CoV-2 infection among health care personnel, first responders, and other essential and frontline workers -eight U.S. locations

Association between vaccination with BNT162b2 and incidence of symptomatic and asymptomatic SARS-CoV-2 infections among health care workers