key: cord-0812001-z6cdyy0w authors: Goldberg, Y.; Mandel, M.; Woodbridge, Y.; Fluss, R.; Novikov, I.; Yaari, R.; Ziv, A.; Freedman, L.; Huppert, A. title: Protection of previous SARS-CoV-2 infection is similar to that of BNT162b2 vaccine protection: A three-month nationwide experience from Israel date: 2021-04-24 journal: nan DOI: 10.1101/2021.04.20.21255670 sha: c76a82e47da36304cc511ce192e7c20790cf630c doc_id: 812001 cord_uid: z6cdyy0w Worldwide shortage of vaccination against SARS-CoV-2 infection while the pandemic is still uncontrolled leads many states to the dilemma whether or not to vaccinate previously infected persons. Understanding the level of protection of previous infection compared to that of vaccination is critical for policy making. We analyze an updated individual-level database of the entire population of Israel to assess the protection efficacy of both prior infection and vaccination in preventing subsequent SARS-CoV-2 infection, hospitalization with COVID-19, severe disease, and death due to COVID-19. Vaccination was highly effective with overall estimated efficacy for documented infection of 92.8% (CI: [92.6, 93.0]); hospitalization 94.2% (CI: [93.6, 94.7]); severe illness 94.4% (CI: [93.6, 95.0]); and death 93.7% (CI: [92.5, 94.7]). Similarly, the overall estimated level of protection from prior SARS-CoV-2 infection for documented infection is 94.8% (CI: [94.4, 95.1]); hospitalization 94.1% (CI: [91.9, 95.7]); and severe illness 96.4% (CI: [92.5, 98.3]). Our results question the need to vaccinate previously-infected individuals. 3 less motivation to get tested. In addition to voluntary testing, Israel conducts routine testing of all nursing-home workers. Recent results based on aggregated data 3-5 and individual level data [6] [7] [8] [9] [10] The prospective observational analysis that we present faced several challenges. The first challenge was self-selection of treatment, which implies differences in potential risk factors between vaccinated and non-vaccinated individuals. These include age, sex, sociodemographic level, 17 level of infection in the immediate environment, and possibly other behavioral variables that could affect level of exposure to the virus. The second challenge was detection bias: willingness to undergo vaccination can be associated with trust in the healthcare system, which may also imply a tendency to comply with testing regulations. On the other hand, vaccinated individuals may feel more protected and may ignore mild symptoms indicative of the disease, and have less motivation to get tested as they are not required to self-quarantine after a contact with a positive individual. The third challenge was the variation in infection risk throughout the vaccination campaign, mainly due to varying lockdown levels, relative prevalence of viral mutants, and local outbreaks. Lastly, the status of individuals (i.e., unvaccinated, partially vaccinated, or fully vaccinated) was All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. ; https://doi.org/10.1101/2021.04.20.21255670 doi: medRxiv preprint 4 dynamic: with time, individuals move from one cohort to another, and between risk groups. In the Methods Section we explain how we designed the analysis to address these challenges. The database included two main tables. The definition of hospitalization, severe disease, and death due to COVID-19 is based on international recommendations. 18 Specifically, hospitalization is defined as being admitted due to COVID-19. Disease is considered severe when a patient has >30 breaths per minute, oxygen saturation on room air <94%, or ratio of arterial 148 partial pressure of oxygen to fraction of inspired oxygen <300mm mercury. Data on symptoms were also available but we found them less reliable and thus did not include symptomatic COVID-19 as an outcome. Thus, the table contained an entry for every adult (age 16) in Israel who had at least one PCR test or had received at least the first dose of the vaccine (with a total of 5,682,928 entries). Adults with no PCR test and no vaccination (668,975) were added to the table using data from the Israel Central Bureau of Statistics. Thus, this second table included All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. To account for environmental risk, we calculated a municipality daily risk index by the number of cases newly confirmed in the past seven days per 10,000 residents. We used a 7day moving average since the number of PCR tests typically drops at weekends. The index was categorized into four risk levels (up to one , one to four, four to ten, and more than ten daily cases per 10,000) to yield the municipality daily risk category, and was used as a covariate in the risk model. Behavioral differences among people may result in different levels of exposure to infection and compliance with PCR testing guidelines. We partially accounted for this by counting the number of PCR-test clusters that an individual underwent from March 1, 2020, to December 20, 2020 (i.e., prior to the vaccination program). Here, a PCR-test cluster comprised all consecutive test performed within 10 days of each other. We then defined three individualized background risk levels: no PCR tests, one cluster, and two or more clusters, and this covariate was also included in the risk model. For previously-infected individuals, we set the level to one cluster and checked sensitivity to this value. Note that the time interval for defining this variable (up to December 20, 2020) did not overlap with the follow-up period. In addition to estimating vaccine efficacy, we estimated the protection of prior SARS-CoV-2 infection against a recurrent infection. Thus, we also included in the dataset individuals who had recovered from COVID-19. Recovery from SARS-CoV-2 infection is not welldefined, and individuals may continue to show traces of the virus weeks and sometimes even months after the infection. 14 We defined as a recurrent infection only cases occurring three months or more after the first diagnosis. We also considered only individuals for whom the first infection was diagnosed between June 1 and September 30, 2020, as the PCR results before June 1 are considered less reliable. Hence, individuals infected before All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. ; https://doi.org/10.1101/2021.04.20.21255670 doi: medRxiv preprint 6 June 1, 2020 or between October 1, 2020 and December 20, 2020 were excluded from the analysis. To estimate the efficacy of the Pfizer BNT162b2 vaccine in reducing documented SARS-CoV-2 infection and other COVID-19 events, we considered four dynamic sub-populations or cohorts: • Cohort 0: Unvaccinated and not previously infected with SARS-CoV-2; • Cohort 1: Vaccinated and followed from the day of first vaccination to 6 days after the second dose; • Cohort 2: Vaccinated and followed from a week after the second dose onwards; • Recovered: Unvaccinated and previously diagnosed with SARS-CoV-2 between June 1 and September 30, 2020. On any given calendar day, each individual included in the analysis belongs to a single cohort, but cohort membership is dynamic. Moreover, individuals may not only move between cohorts over time (for example, from cohort 0 to cohort 1 after first vaccination, or from cohort 1 to cohort 2 at 7 days after the second vaccination), but also exit from the follow-up (for example, on infection with SARS-CoV-2 or death (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. ; https://doi.org/10.1101/2021.04.20.21255670 doi: medRxiv preprint 7 and high). We refer to each combination of possible covariate values (age group, sex, background risk level, and municipality risk level) as the risk profile. Our analysis model falls within the framework of multi-state survival models, where each cohort represents a separate state; 19 see Figure Our model assumes that for a given cohort and risk profile, the hazard was constant and did not depend on the time from the second dose (Cohort 2). Obviously, the hazard of individuals who have never received the first dose (Cohort 0) cannot depend on the time of the first dose, but we also assumed that the time elapsed from the second vaccination did not affect the hazard in Cohort 2. In other words, we assumed that the protection level did not change with time after the "completion" of the vaccination protocol. While protection by vaccination is expected to decrease in the long run, our assumption is reasonable given the time frame of only three months after first vaccination, where waning immunity is not expected to play a role. We split Cohort 1 into two sub-cohorts: Cohort 1A from the first dose to two weeks after the first dose, and Cohort 1B from 15 days after the first dose to six days after the second dose. Following Skowronski and De Serres, 22 we considered, as a crude approximation, a constant hazard for each of these two sub-cohorts for every risk profile. To estimate the level of protection among the Recovered Cohort, we made a similar assumption, that the time elapsed from SARS-CoV-2 infection did not affect the hazard ratio. The formal definition of vaccine efficacy adopted was as follows. Consider any particular risk profile. Let (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Israel need to self-quarantine for 14 days after contacting SARS-CoV-2 infected persons, which can be shortened to ten days if they present two negative PCR tests. This is not required for fully vaccinated and recovered persons unless they develop symptoms. We first investigated the dynamics of the vaccination program, disease outcomes, PCR testing, and municipality risk as a function of calendar time. Figures S4 and S5 present the proportion of vaccinated over time among different age and municipality risk groups, respectively. As can be seen from Figure S4 , the Israeli vaccination policy was initially to immunize the older population, and as time progressed, younger age groups. Figure S5 shows the association between environmental risk and vaccination. Figure S6 shows the All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. ; https://doi.org/10.1101/2021.04.20.21255670 doi: medRxiv preprint rates over time of the different age groups among those tested, infected, hospitalized, having severe disease, and dying. Table 4 shows, by age group, the estimated vaccine efficacy for the main outcomes for Cohort 2 (fully vaccinated) adjusted for sex, municipality risk, and past PCR. Note that for age groups below 60 years, there were, fortunately, none or very few events of severe illness and death, and thus estimates were omitted for these groups. The table shows that vaccine efficacy was quite similar in all age groups with some decrease in efficacy for the 80+ age category. Table 5 presents the results for the Recovered Cohort when the past PCR-based individualized risk was set to one PCR cluster. Again, the protection was quite similar in all age groups with some decrease in efficacy for the 80+ age category, and quite similar to the results in Table 4 . The overall estimated protection of prior SARS-CoV-2 infection for documented recurrent infection was 94·8% (CI: [94·4, 95·1]); hospitalization 94·1% (CI: [91·9, 95·7]); and severe illness 96·4% (CI: [92·5, 98·3]). As there were only 1 death cases in the Recovered Cohort, protection against death was not estimated. As described above, we assigned the recovered individuals to the middle PCR risk group, so that the estimated protection of a prior infection is compared to unvaccinated individuals having a single PCR cluster in the past. The protection levels afforded by a prior infection compared to unvaccinated persons who had no or 2+ past PCR tests are given in a sensitivity analysis shown in Table S1 . In addition, Table S1 presents results of a model without PCR, which can be interpreted as the overall protection of a prior infection. As expected, the protection of a prior infection compared to unvaccinated persons who did not have past PCR tests is estimated to be smaller and compared to those who had 2+ tests is larger. The results when omitting the PCR variable are very similar to the figures in Table 5 . All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Tables S4-S7 . This population-based observational study demonstrates the high efficacy of the BNT162b2 vaccine and prior SARS-CoV-2 infection against both subsequent SARS-CoV-2 infection and other COVID-19-related outcomes. There are a few characteristics that make this study unique. First, it was a nationwide study and thus represented the real-world effectiveness of vaccination and prior infection on the full population. Second, it used individual-level data that enabled, at least to some degree, to mitigate biases caused by selection to get vaccinated, selection to undergo PCR testing, and time-changing level of risk, via adjustment for between-cohort differences in individuals' characteristics and municipality risk level. Third, the study included follow-up of the population for a period of three months, allowing follow-up of the fully vaccinated cohort over an extended duration. Fourth, this is the first large-scale study that has explored the protection due to prior SARS-CoV-2 infection compared to the Pfizer BNT162b2 vaccine. There are some limitations to this observational study. One major source of confounding is related to possible population differences between individuals who were vaccinated compare to those who were not. This confounding is partially addressed by controlling for risk factors. Specifically, for each individual we adjusted for sex, age group, number of past PCR tests and the time-dependent environmental exposure. Another major source of potential bias is related to detection of SARS-CoV-2 infection. As apparent from the PCR test counts in Table 3 , individuals who are fully vaccinated or were previously infected get tested less often than the unvaccinated cohort. Our results for the outcomes of All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. ; https://doi.org/10.1101/2021.04.20.21255670 doi: medRxiv preprint hospitalization, severe disease, and death do not suffer from this bias and thus are more reliable. The vaccine protection against infection might be biased upward as explained above, nevertheless the remarkable curtailing of the outbreak in Israel which followed the high vaccine uptake by the Israeli population further suggest that the vaccine is efficient in blocking transmission, see Figure 1 . The efficacy estimates of the BNT162b2 vaccine in this study are similar to those reported by previous large-scale studies. For the severe disease outcome, the randomized trial of BNT162b2 1 reported 89% efficacy for severe disease. A study by the Israeli Ministry of Health using aggregated data 5 reported 96% efficacy for people as defined in our Cohort 2. A study on data from Israel's largest HMO 6 split people as defined in our Cohort 1B and reported an efficacy of 62% and 80% for the third and fourth weeks after the first vaccine, respectively, and of 92% for their Cohort 2. In comparison, our analysis showed efficacy of 66% for Cohort 1B and 94% for Cohort 2. For other outcomes, the estimated vaccine efficacy for Cohort 2 in our study were 93% and 94%, for documented infection and hospitalization, respectively. These estimates are similar to previous studies 5,6 that estimated efficacy of 92% and 96% for documented infection, and of 87% and 96% for hospitalization. Our findings are based on a longer follow-up and a larger number of event than in the previous individual-level data reports. For example, the analysis of severe cases in the randomized clinical trial is based on only 10 cases, and that of Israel's largest HMO on 229. 6 In comparison, the analysis in our study is based on 8,463 cases, including 2,240 cases from Cohort 1 and 319 cases from Cohort 2. On the other hand, the other two studies 1,6 have the respective advantages of randomization and a detailed matching process which help in bias reduction. The estimated protection against reinfection in this study is similar to that of the BNT162b2 vaccine. For documented SARS-CoV-2 reinfection, these results are similar to the results obtained in a large study from Qatar of 95% protection, 13 and suggest higher protection than reported by other previous studies. A large study from Denmark 14 suggested 80% protection against reinfection. A study on healthcare workers in the United Kingdom 16 reported that previous infection was associated with an 83% lower risk of infection. These All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. The study was approved by the Institutional Review Board of the Sheba Medical Center. All authors declare no competing interests. The data used in this study are sensitive and will not be made publicly available. The authors would like to thank Guy Katriel for fruitful discussions. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. We assume that for Thus, the effect of sex, past PCR test, and municipal risk on efficacy is multiplicative and identical among cohorts. However, efficacy may vary between different age groups. The constant hazard assumption implies underlying exponential event-free models for these cohorts, with time-dependent covariates. The analysis can be carried out by performing All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. , and the efficacy is defined as . Under the assumption of equal coefficients for sex, past PCR tests and municipality risk, the relative efficacy depends only on the age group. Technically, in order to estimate the coefficients in the model, we create a working dataset as follows. For each combination of cohort, age group, sex, municipality risk level, and This person contributes: 1. 11 days (Dec-20 to Dec-31) and 0 events to the group: cohort_0/50-60/male/mun_risk=1/past_pcr=1 2. 14 days (Jan-1 to Jan-14) and 0 events to the group: cohort_1A/50-60/male/mun_risk=1/past_pcr=1 3. 6 days (Jan-15 to Jan-20) and 0 events to the group: cohort_1B/50-60/male/mun_risk=1/past_pcr=1 4. 9 days (Jan-21 to Jan-29) and 0 events to the group: cohort_1B/50-60/male/mun_risk=2/past_pcr=1 5. 10 days (Jan-30 to Feb-8) and 1 event to the group: cohort_2/50-60/male/mun_risk=2/past_pcr=1 All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. ; https://doi.org/10.1101/2021.04.20.21255670 doi: medRxiv preprint Figure S1 : The dynamics of the cohort model. Solid arrows indicate possible transitions between cohorts. Dashed arrows indicate possible disease outcomes. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. ; https://doi.org/10.1101/2021.04.20.21255670 doi: medRxiv preprint Figure S2 : Length of follow-up for Cohort 2. Length of follow-up for Cohort 2 of the fully vaccinated, according to age group. Vaccination became available first to the 60+ age groups and then gradually to younger age groups as can be seen from the follow-up counts. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. ; https://doi.org/10.1101/2021.04.20.21255670 doi: medRxiv preprint Figure S3 : Length of follow-up for the Recovered Cohort. Length of follow-up from first positive PCR test for the Recovered Cohort, according to age group. This cohort included individuals that had a positive PCR test between June 1 and September 30, 2020. Note the sharp decrease in counts as a function of the follow-up. Note that each subfigure has a different scale. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. ; https://doi.org/10.1101/2021.04.20.21255670 doi: medRxiv preprint Figure S6 : Events over time. Cases per 100,000, smoothed using seven-day moving average for the different age groups and the outcomes: PCR tests, documented infection cases, hospitalized cases, severe cases, and deaths. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. ; https://doi.org/10.1101/2021.04.20.21255670 doi: medRxiv preprint All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. ; https://doi.org/10.1101/2021.04.20.21255670 doi: medRxiv preprint All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted April 24, 2021. ; https://doi.org/10.1101/2021.04.20.21255670 doi: medRxiv preprint Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine Covid-19: Pfizer-BioNTech vaccine is rolled out in US Patterns of COVID-19 pandemic dynamics following deployment of a broad national immunization program Nationwide Vaccination Campaign with BNT162b2 in Israel Demonstrates High Vaccine Effectiveness and Marked Declines in Incidence of SARS-CoV-2 Infections and COVID-19 Cases, Hospitalizations BNT162b2 mRNA Covid-19 Vaccine in a Nationwide Mass Vaccination Setting The effectiveness of the first dose of BNT162b2 vaccine in