COVID-19 Vaccine Effectiveness: A Review of the First 6 Months of COVID-19 Vaccine Availability (1 January–30 June 2021)

Observational studies are needed to demonstrate real-world vaccine effectiveness (VE) against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) outcomes. Our objective was to conduct a review of published SARS-CoV-2 VE articles, supplemented by preprints, during the first 6 months of COVID-19 vaccine availability. This review compares the effectiveness of completing the primary COVID-19 vaccination series against multiple SARS-CoV-2 disease presentations and disease severity outcomes in three population groups (general population, frontline workers, and older adults). Four hundred and seventy-one published articles and 47 preprints were identified. After title and abstract screening and full article review, 50 studies (28 published articles, 22 preprints) were included. VE results were reported for five COVID-19 vaccines and four combinations of COVID-19 vaccines. VE results for BNT162b2 were reported in 70.6% of all studies. Seventeen studies reported variant specific VE estimates; Alpha was the most common. This comprehensive review demonstrates that COVID-19 vaccination is an important tool for preventing COVID-19 morbidity and mortality among fully vaccinated persons aged 16 years and older and serves as an important baseline from which to follow future trends in COVID-19 evolution and effectiveness of new and updated vaccines.


Introduction
The coronavirus disease 2019 (COVID-19) pandemic has caused significant morbidity, mortality, and economic loss globally. As a result, scientists around the world have been working tirelessly to develop, produce, and test COVID-19 vaccines that limit the spread of SARS-CoV-2 and prevent the adverse health effects of SARS-CoV-2 infection. Clinical trials have shown COVID-19 vaccines to be safe and immunogenic, with an efficacy against symptomatic infection in randomized controlled trials (RCTs) ranging from 95.0% and 94.1% for the messenger RNA (mRNA) vaccines BNT162b2 (Pfizer-BioNTech) [1] and mRNA-1273 (Moderna) [2], respectively, to 50.7% for the inactivated whole-virion vaccine CoronaVac (Sinovac) [3]. Other vaccines included in this review had intermediate efficacies of 77.8%, 67.1%, and 66.9%, for Covaxin ® (Bharat Biotech) [4], ChAdOx1 (AstraZeneca) [5], and Ad26.COV2.S (Janssen/Johnson & Johnson) [6], respectively. The first vaccine authorized and used in the United States (US) was BNT162b2; first doses were administered on 14 December 2020, and the first individuals completed the two-dose primary vaccination series in January 2021. Since vaccine trials, including the above-mentioned RCTs, are conducted in controlled settings with healthy individuals or those with stable medical conditions [1][2][3][4][5][6][7][8][9], observational studies are needed to demonstrate real-world vaccine effectiveness (VE) against all severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) outcomes, including asymptomatic infections and severe outcomes including hospitalizations and deaths, in field settings across the globe. It is also important to determine VE in subsets of the population who may be at higher risk of being infected with SARS-CoV-2 (e.g., frontline and healthcare workers) or having more severe outcomes (e.g., older adults and persons with underlying illnesses). Finally, it is important to monitor VE over time to assess changes in effectiveness, which may occur following waning immunity or the dissemination of SARS-CoV-2 variants that are associated with increased transmissibility or more severe illness.
We conducted a review of published (i.e., peer-reviewed) SARS-CoV-2 VE articles, supplemented by preprints posted on preprint servers and reports published on websites of public health agencies during the first 6 months of COVID-19 vaccine availability. While other VE reviews have been published [10][11][12][13][14], our review is unique in that we (1) provided VE results for the first 6 months of global vaccine use and for only fully vaccinated participants, (2) examined VE for three population groups separately, and (3) plotted VE results to allow for direct comparison across disease presentation and disease severity categories by vaccine and by days after full vaccination.
The objective of our review is to compare the effectiveness of completing the primary COVID-19 vaccine series (i.e., "fully vaccinated," as defined during the period of this review) against multiple SARS-CoV-2 outcomes (i.e., infection, asymptomatic infection, symptomatic infection, hospitalization, severe disease, intensive care unit [ICU] admission, death) by vaccine product, study population, number of days after full vaccination, and variant. VE information assists physicians and public health officials with identifying which vaccines are most effective for which population subgroup and with monitoring trends to inform the need for subsequent vaccine doses.

Materials and Methods
A literature search was conducted in PubMed to identify articles published between 1 January and 30 June 2021, written in English, and describing observational studies that assessed VE against SARS-CoV-2 outcomes in real-world settings. This 6-month period was chosen to focus our review on VE among fully vaccinated persons aged 16 years and older without having to factor in the influence waning immunity or subsequent vaccine doses. In addition, this early time period provides an important baseline from which to follow future trends in COVID-19 evolution and effectiveness of new or updated vaccines.
The literature search terms are described in the Supplementary Methods. Separately, Pfizer investigators searched the medRxiv and bioRxiv COVID-19 SARS-CoV-2 preprint server and the SSRN preprint server daily for preprints of articles related to COVID-19 VE with the term "BNT162b2" or "effectiveness" in the title to identify preprints describing COVID-19 VE studies. Following a cursory review for appropriateness, preprint servers post scientific articles that have not yet been peer reviewed; such servers have been a vital mechanism for timely dissemination of scientific results during the rapidly evolving SARS-CoV-2 pandemic. Pfizer also monitored media reports and websites of national public health agencies daily to identify both published articles and preprints. These included reports from government agencies (e.g., Public Health England) that included COVID-19 VE information; for the purposes of this review, such reports are also considered as preprints. Published articles that were identified by Pfizer's daily monitoring of media reports and websites of national public health agencies but were not identified through the PubMed search are referred to as "Published articles identified by Pfizer." Published articles and preprints identified by the PubMed search and by Pfizer were included in the title and abstract screening and full article review process described below and summarized in Figure 1. Although we performed a comprehensive search of available literature as a Published articles and preprints identified by the PubMed search and by Pfizer were included in the title and abstract screening and full article review process described below and summarized in Figure 1. Although we performed a comprehensive search of available literature as a part of our methods, we did not conduct a quality assessment of published articles and preprints. Thus, our review should not be considered a systematic literature review. Figure 1. PRISMA Flow Chart. VE = vaccine effectiveness. † Economic or cost-effectiveness (n = 4); vaccine side effects (n = 4); case report or series (n = 3); surveillance study (n = 2); nutrition (n = 1); risk-benefit analysis (n = 1); symptoms (n = 1). ‡ Image or audio clip with no article (n = 2); news article (n = 2); author reply (n = 1). § Five identified by through sources other than PubMed. * One study did not distinguish between one-and two-dose VE.
Published articles and preprints eligible for inclusion were observational studies that reported the effectiveness of any COVID-19 vaccine for fully vaccinated persons. The primary series for Ad26.COV2.S is one dose; all other vaccines are two doses. Two investigators (L.M.B. and S.M.H. or S.M.E.-D.) independently screened the titles and abstracts of all published articles, where available, to identify studies for a full article review. Published articles with no abstract or those with titles and abstracts that did not provide sufficient context to exclude them at the abstract review stage were included in the full article review. All preprints were included in the full article review. Among the 82 published articles assessed for eligibility, 20 were excluded because they did not present VE or a measure from which VE could be directly calculated, 14 were vaccine side effects (n = 4); case report or series (n = 3); surveillance study (n = 2); nutrition (n = 1); risk-benefit analysis (n = 1); symptoms (n = 1). ‡ Image or audio clip with no article (n = 2); news article (n = 2); author reply (n = 1). § Five identified by through sources other than PubMed. * One study did not distinguish between one-and two-dose VE.
Published articles and preprints eligible for inclusion were observational studies that reported the effectiveness of any COVID-19 vaccine for fully vaccinated persons. The primary series for Ad26.COV2.S is one dose; all other vaccines are two doses. Two investigators (L.M.B. and S.M.H. or S.M.E.-D.) independently screened the titles and abstracts of all published articles, where available, to identify studies for a full article review. Published articles with no abstract or those with titles and abstracts that did not provide sufficient context to exclude them at the abstract review stage were included in the full article review. All preprints were included in the full article review. Among the 82 published articles assessed for eligibility, 20 were excluded because they did not present VE or a measure from which VE could be directly calculated, 14 were excluded because they were a review or commentary, 13 were excluded because they only presented VE for a single vaccine dose for vaccines with two-dose regimens, 2 were excluded because they contained data that were updated in a more recently published article (source data duplicate), and 5 were excluded for another reason detailed in Figure 1. Among the 47 preprints assessed for eligibility, 15 were excluded because they only presented VE for a single vaccine dose for vaccines with two-dose regimens, 9 were excluded because they did not present VE or a measure from which VE could be directly calculated, and 1 was excluded because it was removed from the preprint server prior to submission of this manuscript.
Abstracted information included country, study design, study period, study population, number of participants, participant age in years (mean, median, or category), number of participants vaccinated and unvaccinated, vaccine, number of days after being fully vaccinated, identified or circulating variant, and VE and 95% confidence intervals (CIs) by SARS-CoV-2 outcome. VE and 95% CIs were rounded to the nearest whole number. For published articles or preprints that provided VE for >1 vaccine and reported both combined and individual VE estimates, we reported only the individual results unless the combined VE estimates included additional stratification (e.g., by variant, disease presentation, disease severity) not provided for the individual VE estimates. A variant was considered "identified" if the study authors performed laboratory testing to identify the variant detected from each infected person, or a sample of infected persons, that contributed to the VE estimate. A variant was considered "circulating" if the study presented background information or other evidence of the dominant strain(s) circulating in the population during the study period but did not perform laboratory testing to identify variant(s) detected from infected persons. This detailed information is presented in Table 1 for each study. To compare VE results between populations with distinct disease or exposure risks, study populations were classified into three broad categories: general population aged ≥16 years, adult frontline workers, and older adults aged ≥65 years. One study [15] included in the older adults' category persons aged >60 years. The link between the detailed study populations presented in Table 1 and the broad categories used in Table 2 and Figures 2-4 is provided in Supplementary Table S1. To compare VE results by time after full vaccination, days after full vaccine dose were grouped into two categories: ≥7 days and ≥14 days (Figures 2-4). For completeness, estimations of VE at <7 days are provided for two-dose regimens in Table 1. (n = 1). e Includes studies that calculated VE at 7-13 days, 7-28 days, >8 days, >11 days, and ≥7 days (BNT162b2) or >14 days (mRNA-1273). f Includes studies that calculated VE at >15 days and ≥15 days. g Includes studies that calculated VE at ≥0 days, >1 day, 0-13 days. h Study provided VE estimate for multiple variants together (did not stratify VE by variant). Includes published studies that provided a VE for B.1.1.7, B.1.351, B.1.617, and "wildtype strains" (n = 1); B.1.1.7, B1.177, P.1, and B.1.351 (n = 1); B.1.1.7 and B1.525 (n = 1); and B.1.1.7, B.1.427, B.1.429, and P.2 (n =1). Includes preprint studies that provided a VE for B.1.1.7, B.1.427, and B.1.429 (n = 1); B.1.351 and P.1 (n = 1); and B.1.1.7, P.1, and other VOC (n = 1). i Includes one article that provides VE for "earlier variants."       Death 99 (87-100) CI = confidence interval; HCW = health care worker; HS = health system; IBD = inflammatory bowel disease; ICU = intensive care unit; LTCF = long-term care facility; NA = not applicable; SNF = skilled nursing facility; UK = United Kingdom; US = United States; VHA = Veterans Health Administration; VOC = variant of concern. ++ Article not identified by PubMed search criteria, identified by Pfizer; ** >1 dose; * Crude VE; @ Preprint published prior to submission of this manuscript. a Individual studies sometimes included more than one of the listed categories; categories will not sum to total N. b Includes one study that reported VE among persons aged >60 years. c Other study populations in published articles include airline passengers (n = 1), hospital patients (n = 1), and veterans with IBD/immunosuppression (n = 1). Other study populations in preprint articles include LTCF residents, older adults, HCWs, and severe risk individuals (n = 1); severe risk individuals (n = 1); hospitalized COVID patients (n = 1); and symptomatic cases (n =1). d Other study designs among published articles include cross-sectional (n = 1), outbreak investigation (n = 1), surveillance study (n = 1), and test-negative cohort (n =1  Vaccine effectiveness estimates and 95% CIs were abstracted for both SARS-CoV-2 outcomes (infection, asymptomatic infection, or symptomatic infection) and disease severity outcomes (hospitalization, severe disease, ICU admission, or death). Adjusted VE results are presented unless otherwise specified. Studies are categorized and presented separately based on their source: "published articles" or "preprints." For preprints that were published before manuscript submission, results were updated to reflect the published version of the article.

Results
Four hundred and seventy-one published articles and 47 preprints were identified. After title and abstract screening and full article review, 50 studies were included in this review, of which 28 were published articles and 22 were preprints (Figure 1). Of the 22 preprints, 12 were published prior to submission of this review. There was a change in the VE estimates between the preprint and published article for three articles [48,50,57] due to increases in participant numbers.
Characteristics of abstracted published articles and preprints included in the review are described in Table 2. Most studies were conducted in the US (26.0%), United Kingdom

Results
Four hundred and seventy-one published articles and 47 preprints were identified. After title and abstract screening and full article review, 50 studies were included in this review, of which 28 were published articles and 22 were preprints (Figure 1). Of the 22 preprints, 12 were published prior to submission of this review. There was a change in the VE estimates between the preprint and published article for three articles [48,50,57] due to increases in participant numbers.
Characteristics of all 50 published articles and preprints that assessed VE are provided in Table 1. Results are presented in alphabetic order by study author  under the headings Published Articles and Preprints. Vaccine effectiveness estimates for identified and circulating variants are also provided. In Figures 2-4, VE estimates are presented separately by population group for disease presentation and disease severity, and are stratified by vaccine (BNT162b2, mRNA-1273, BNT162b2 and mRNA-1273, ChadOx1, Ad26.COV2.S, CoronaVac) and days after final dose (≥7 and ≥14). Where available, variant specific VE results are shown for variants of concern (VOCs) [67]: Alpha, Beta (B.1.351), Delta (B.1.617.2), and other recorded variants, including B.1.617, R.1, and multiple variants.

General Population Aged ≥16 Years
Vaccine effectiveness results for the general population aged ≥16 years by disease presentation are presented in Figure 2a.

Frontline Workers
Vaccine effectiveness results for frontline workers by disease presentation are shown in Figure 3. The only VE estimates for disease severity were reported for the combined ChAdOx1 and Covaxin vaccines [44] (see Table 1).

Older Adults Aged ≥65 Years
Disease presentation VE estimates for older adults are illustrated in Figure 4a.

Discussion
This review included 50 real-world studies encompassing both published peer-reviewed articles and preprints conducted among participants aged 16 years and older during the first 6 months of COVID-19 vaccine use worldwide. Including preprints in the review enabled the capture of timelier COVID-19 research in this rapidly evolving field. Of the 23 preprints initially identified, 12 were published prior to submission of this article, and in only a few instances were minor updates to the VE results required. While other VE reviews have been published [10][11][12][13][14], our review was unique in that we (1) provided VE results for the first 6 months of global vaccine use and for only fully vaccinated participants, (2) examined VE for three population groups separately, and (3) plotted VE results to allow for direct comparison across disease presentation and disease severity categories by vaccine and by days after full vaccination. For a global population that was immunologically naïve to SARS-CoV-2, our focus on VE among fully vaccinated persons aged 16 years and older is valuable because it provides a baseline to compare the effectiveness of full vaccination in various populations around the world without having to factor in the influence of waning immunity [68,69], novel variants [70], or subsequent vaccine doses. Future reviews of VE over longer time frames and within the context of VOC and subsequent vaccine doses will help address these important topics and help guide public health recommendations.
The real-world studies in our review indicate that among fully vaccinated persons, the mRNA vaccines BNT162b2 and mRNA-1273 were highly effective, particularly in preventing severe outcomes of SARS-CoV-2 infection. For example, among the general population aged ≥16 years BNT162b2 VE estimates were ≥82%, ≥81%, and ≥94% against hospitalization, severe disease or ICU admission, and death, respectively, and mRNA-1273 VE estimates were ≥86% against all disease severity categories. Among older adults, BNT162b2 VE was ≥75% and ≥69% against hospitalization and death, respectively, and combined mRNA vaccines VE was ≥88% and ≥97% against hospitalization and death, respectively. Although most VE estimates were similar for the two time periods ≥7 and ≥14 days after full vaccination, a large study by Aran et al. [15] noted significantly higher VE estimates for the BNT162b2 vaccine in the general population aged ≥16 years and in adults aged ≥65 years for infection, hospitalization, and severe disease for vaccination >14 days compared with 7-13 days after full vaccination. A possible reason why these time period differences were observed only by Aran et al. [15] is their use of distinct time periods (i.e., 7-13 days and ≥14 days rather than ≥7 days and ≥14 days).
ChAdOx1 VE values against SARS-CoV-2 infection (≥60%) were generally lower than those for the mRNA vaccines; however, ChAdOx1 provided similarly strong protection against severe disease compared with the mRNA vaccines, including among frontline workers. In contrast, CoronaVac effectiveness against infection (38%) and symptomatic infection (37%) were substantially lower than mRNA vaccines and ChAdOx1. This lower VE is not unexpected given the relatively lower efficacy reported in CoronaVac RCTs in Turkey and Brazil [3,7]. Our review included only one study that reported VE for Ad26.COV2.S. This is likely because the US Food and Drug Administration and World Health Organization did not authorize Ad26.COV2.S for emergency use until 27 February 2021 [71], and 11 March 2021 [72], respectively.
Although we collected variant-specific information, our review predominantly included studies from Israel, the UK, and the US during the time period when the Alpha variant was the only VOC that broadly disseminated in these countries. Alpha was first identified in the UK in September 2020 and was the predominant variant globally between January and May 2021 [57,65]. Alpha became the dominant variant in Israel in December 2020 and in the US in April 2021. Although Beta, Gamma, and Delta variants were detected in South Africa, Brazil, and India during our study period, their prevalence in Israel, the UK, and the US during our study period was low [57,65].
This review was subject to several limitations. We did not perform a rigorous evaluation of study quality; as a result, some errors in study design or analysis may not have been identified. We did not attempt to perform meta-analyses due to the heterogeneity in study design, study analysis methods, study populations, circulating variants, time of VE assessment, and other variables that limit VE comparison among studies [73]. For example, there was variation in the timetable that the vaccine was available in each subgroup (i.e., frontline workers were offered the vaccine ahead of the general population) and the type of vaccine available in each country (e.g., ChAdOx1 vaccine is not authorized for use in the US). Information about the effectiveness of vaccination among previously infected persons was not abstracted, and we did not evaluate whether studies included previously infected persons in their vaccinated or unvaccinated groups. Some studies estimated VE by pooled analysis of two or more vaccines, and the proportion of the study population receiving each vaccine was often unevenly distributed; in these cases, pooled estimates might underestimate or overestimate the VE for one or more vaccines. Although we categorized study populations based on exposure or disease severity risk, heterogeneity with respect to exposure and disease severity risk within our population groups exists. We were unable to assess the impact of waning immunity on VE because our study focuses on the first 6 months of vaccine use; thus, few study participants had been fully vaccinated for >4 months. Finally, 10 studies included in this review have not yet been published and thus have not been certified by peer review.
Despite these limitations, this review of the effectiveness of COVID-19 vaccines in the first 6 months of vaccine use demonstrates that COVID-19 vaccination is an important tool for preventing COVID-19 morbidity and mortality. We found that mRNA vaccines are highly effective at preventing severe outcomes of SARS-CoV-2 infection, including among vulnerable populations such as older adults. As we limited our review to studies that reported VE among fully vaccinated persons aged 16 years and older, it serves as an important baseline from which to follow future trends in COVID-19 evolution and effectiveness of new or updated vaccines. To better understand the broader vaccine landscape, future reviews should include observational studies from a wider range of countries of new or updated vaccines as they become more widely available, and of adolescents and children. They should also include studies that evaluate the impact of VOC, comorbidities, waning immunity, and subsequent vaccine doses on VE.

Conclusions
This comprehensive review of 50 real-world studies conducted during the first 6 months of COVID-19 vaccine use worldwide demonstrates that COVID-19 vaccination, particularly with the mRNA vaccines, is an important tool for preventing COVID-19 morbidity and mortality among fully vaccinated persons aged 16 years and older, including among vulnerable populations. This review also serves as an important baseline from which to follow future trends in COVID-19 evolution and effectiveness of new and updated vaccines. Funding: Pfizer provided support through a contract with Research Triangle Institute to conduct this study. Pfizer and RTI approved the decision to submit the manuscript for publication.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.