Efficacy and safety of four COVID-19 vaccines in preventing SARS-CoV-2 infection: A rapid review

Abstract Introduction: Since the emergence of the SARS-CoV-2, there have been efforts to develop vaccines to control the COVID-19 pandemic. Objective: The present study assessed the efficacy and safety of the BNT162b2, mRNA-1273, ChAdOx1/AZD1222 and Gam-COVID-Vac rAd26-S/rAd5-S vaccines against the SARS-CoV-2. Materials and methods: We searched PubMed/MEDLINE, Google Scholar, Cochrane, and the WHO International Clinical Trials Registry Platform on March 15, 2021. The search terms used were: "vaccine" OR "vaccination" AND "covid19" OR "coronavirus" OR "sarscov2" AND "bnt162b2" OR "chadox1-S" OR "azd1222" OR "sputnik" OR "Gam-COVID-Vac" OR "mrna" OR "mRNA-1273" . We measured the risk of bias of the studies and the quality of the evidence using GRADE profiles. A qualitative and quantitative analysis of the results of clinical trials is presented. Results: Of the 74 identified studies, 4 were finally included in this review. The efficacies of the BNT162b2, mRNA-1273, ChAdOx1/AZD1222 and Gam-COVID-VacrAd26-S/rAd5-S vaccines against symptomatic COVID-19 were 95,0% (CI95% 90,3-97,6), 94,1% (CI95% 89,3-96,8), 66,7% (CI95% 57,4-74,0), and 91,1% (CI95% 83,8-95,1), respectively. There°was moderate certainty of the evidence due to serious indirectness, when we measured the risk of bias of the studies and the quality of the evidence using GRADE profile. The safety profiles were acceptable, and data on serious adverse events (summary RR=0,93; CI95% 0,77-1,12; p=0,16) and deaths from all causes (summary RR=0,70; CI95% 0,33-1,50; p=0°90) showed no significant differences. Conclusion: The results of this review support the level of evidence for the efficacy and safety of the COVID-19 vaccines analysed.

Containment measures failed to stop the spread of the virus, and there are no specific treatments against COVID-19. Vaccines prepare the immune system to detect and counteract viruses with greater success than other interventions. Successful cases of mass vaccination have controlled or eliminated smallpox, poliomyelitis, measles and rubella in America, which avoided illness, premature death and health costs (4).
Efforts to develop SARS-CoV-2 vaccines to control the pandemic have been underway since the virus emerged, with more than 500 candidate vaccines in clinical trials (5). Some vaccine development are more advanced than others. More than 80 candidate vaccines are being tested in humans (6). The impact of COVID-19 vaccines on the pandemic depends on several factors: the effectiveness of the vaccines; the speed of vaccine approval, manufacture, and delivery; the possible development of other variants; individual factors; and the number of people vaccinated (7).
Due to these public policy decisions on health, besides the promising preliminary results of these vaccines against the virus SARS-CoV-2 and the initiation of mass vaccination campaigns, it is necessary to evaluate the available evidence on its efficacy and safety.

Materials and methods
We adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines throughout the manuscript (12). The results were submitted to the International Register of Systematic Perspective Reviews (PROSPERO) and approved with registration number CRD42021229802. Two researchers performed the screening process independently and applied pre-established inclusion and exclusion criteria to select studies for a complete reading.

Search strategy
On March15, 2021, we searched in PubMed/MEDLINE, Google Scholar, Cochrane, and the WHO International Clinical Trials Registry Platform. The search terms used were: "vaccine" OR "vaccination" AND "covid19" OR "coronavirus" OR "sarscov2" AND "bnt162b2" OR "chadox1-S" OR "azd1222" OR "sputnik" OR "Gam-COVID-Vac" OR "mrna" OR "mRNA-1273" . These terms could be found anywhere in the article, title, or abstract. The search equations used for each of the databases are provided in online supplement 1. The Rayyan® web application was used to organise the list of references, remove duplicates, and obtain the full document for review.

Inclusion criteria
According to the PICOT question-model, population, intervention, comparison, and outcomes are shown in Table 1.
Studies: Randomized clinical trials (RCT) phases II/III and III in humans.
Results report: Studies that reported individual effect estimates for each primary research study that was attributable to the comparison of interest and at least one outcome.
-No language or country restrictions were applied.

Exclusion criteria
We did not include studies that were only available on abstract format, because the information reported is insufficient to evaluate methodological quality.

Screening and selection of studies
Two reviewers screened the total number of references identified in the search by examining the titles and abstracts against independently predefined eligibility criteria. From the group of pre-selected references, a smaller number of studies were selected. The reviewers verified that each study met the eligibility criteria by reading the full text on each publication. Disagreements were resolved via consensus. To extract the information, a standardized Excel tool was used, which was tested by the reviewers before use. The structure was based on collecting information on the basic characteristics of each study, such as participants, intervention (vaccine), comparators and outcomes. The data extraction was performed in duplicate and subsequently verified by the researchers involved, who compared the extracted data with the studies.
Reviewers selected the effect estimates of for the comparison and critical outcomes from reported values in the studies.

Assessments of risk of bias and certainty of the evidence
The Cochrane risk of bias tool was used to assess the following domains: randomization, deviations from intended intervention, missing outcome data, measurement of the outcome, selection of the reported results, and overall risk of bias (13). Studies were categorized as: "high risk" , "low risk" or "some concerns" . The certainty of the evidence was measured using GRADE profiles (14). A single-first reviewer author rated the risk of bias and certainty of the evidence for each study, and a second reviewer author checked the ratings.

Statistical analysis
Abstracted data were aggregated in tables. Risk ratios (RR), risk differences and corresponding 95% confidence intervals (CI 95% ) were calculated or extracted from the selected publications. We computed the RR for the outcomes of the adverse events and serious adverse events. Data on the vaccine efficacy were extracted from the publications and calculated ([1-(risk ratio or rate ratio comparing vaccine and placebo recipients)] × 100) for the outcome asymptomatic or unknown infection of the vaccine ChAdOx1-S.
We performed inverse variance-weighted random-effects metaanalyses using the Paule and Mandel τ 2 estimator for heterogeneity (15). Heterogeneity across the RCTs was described using the I 2 and τ 2 metrics (16). Data were analysed using the statistical package Stata™, version 15 (Stata Corporation, College Station, Texas, United States). Results Figure 1 shows the search results, screening, and selection of evidence for this rapid review. The search of the identified databases detected 74 documents. After exclusion of duplicates, 68 articles remained for title and abstract screening. Fifty-six of these papers were deemed irrelevant according to the titles and abstracts and were later excluded. We assessed the full text of the remaining 12 articles: seven of these articles were excluded due to RCT phase I (n=1), phase I/II (n=4), phase II/III (n=1, no outcomes of interest), and wrong vaccine and phase I/II (n=1). Bibliographic data of these 7 studies were excluded after full-text assessment (17)(18)(19)(20)(21)(22)(23). Of the 5 included articles,  (24)(25)(26)(27)(28). The age of the participants at study entry ranged from 16 to 85 years. We summarized the characteristics of the studies included in Table 2.

Risk of bias
The overall risk of bias was classified as having some concerns in all 4 RCTs due to deviations in the interventions. In general, all trials assessed the main efficacy outcomes using per-protocol and not intention-to-treat analyses. Details of the risk of bias assessment are provided in online supplement 2.

Certainty of the evidence
For the efficacy outcome evaluated via symptomatic SARS-CoV-2 infection confirmed by a laboratory, the certainty of the evidence (using GRADE) was moderate due to serious indirectness (data from an interim analysis of the trial, with a short follow-up duration; estimates may change during longer follow-up; population included in RCTs may not represent all persons aged 16 years in BNT162b2; and 18 years in mRNA-1273, ChAdOx1-S, Gam-COVID-VacrAd26-S/rAd5-S). For the severe or critical disease due to COVID-19 in the RCTs regarding BNT162b2 and ChAdOx1 vaccines, the certainty of the evidence was very low, and moderate for mRNA-1273 and Gam-COVID-VacrAd26-S/rAd5-S vaccines. In the four RCTs, the certainty of the evidence for the outcomes of all-cause mortality and serious adverse events was very low due to serious indirectness and very serious imprecision based on the wide CI 95% for RR, which is consistent with substantial benefit or harm (see GRADE tables for more details in online supplement 4).

Discussion
Available data indicated that the vaccines evaluated in this systematic review effectively prevented symptomatic laboratory-confirmed COVID-19 (range 69.7-95%) (24)(25)(26)(27)(28), with moderate certainty of evidence. For severe cases of the disease, the four evaluated vaccines were between 90and 100% effective with moderate to very low certainty of evidence. The subgroup analysis in each RCT revealed that the global efficacy of the vaccine was similar to the efficacy showed in different populations according to age, sex, comorbidities, and race/ethnicity (24)(25)(26)(27)(28).
The current concern related to asymptomatic infections is based on people who is able to continue transmitting the virus to others despite being vaccinated. Although asymptomatic infections are not a direct measure of disease transmission, researchers have viewed the information on this result as an indicator of how vaccines would help reducing the risk of the spread of SARS-CoV-2. The reports included in this review revealed information on this outcome from the ChAdOx1-S vaccine, which provided preliminary data on the prevention of asymptomatic SARS-CoV-2 infection. Showing an efficacy of 49.3% (CI 95% 7.4-72.2) in the group that received a half dose followed by a standard dose, compared to only 2% (CI 95% -50.7-36.2) in the group that received two standard doses (26,27). These results suggest that COVID-19 vaccines also reduce asymptomatic infection and potentially transmission.
The evidence shows that vaccination prevents a person from getting COVID-19 in its severe form and reduces the possibility of transmission to others (29). Substantial reductions in SARS-CoV-2 infections (symptomatic and asymptomatic) will help to reduce overall levels of disease and the transmission of the virus worldwide.
For adverse events and death, the analysed vaccines showed an adequate safety profile. However, the level of certainty of evidence was very low. Therefore, surveillance and follow-up of vaccinated cohorts is very important for the identification of possible adverse effects, especially rare events of death or disability.
Our review synthesized the evidence in a simple and methodical way to inform health providers and readers about the general efficacy and safety of four vaccines against COVID-19. These vaccines exhibited different degrees of effectiveness in the prevention of SARS-CoV-2 infection. However, all were very effective in preventing serious forms of illness and secondary death from COVID-19. The BNT162b2, mRNA-1273, ChAdOx1-S, and Gam-COVID-VacrAd26-S/rAd5-S vaccines showed high efficacy based on the minimum criteria established by the WHO to recommend a vaccine against COVID-19, which consisted of an estimate of 50% and with a clear demonstration of efficacy in the base population (30).
Global, national and regional regulatory procedures evaluating the suitability of new medical devices for public health emergencies, are responsible for performing a rigorous process to decide the administration of a vaccine with specified prioritisation for the earliest use (31). Other vaccines that were not the subject of this review have also received approval for emergency use by health regulatory agencies in different countries, such as CanSino, Sinopharm, Sinovac and Johnson & Johnson's vaccines (32)(33)(34)(35). This information is encouraging due to the main challenge producing safe and effective vaccines in sufficient quantity for equitable distribution worldwide. More vaccines provide greater hope of ending the pandemic.

Strengths and limitations
Rapid review is a methodology that may be particularly important in the COVID-19 pandemic because the evidence is rapidly emerging and verified information is needed to make policy or practical decisions. However, there are some limitations to our study, which primarily relate to the performance of this review in a limited time frame. First, while a comprehensive search was performed in three databases, there was insufficient time to search other sources (grey literature) to ensure comprehensiveness. However, making these methodological trade-offs is consistent with the time-limited approach taken in other rapid review methodologies (36,37).
Second, this review included only 4 phase II/III and III clinical trials with preliminary published data because they involve ongoing research. We did not include other study designs that analysed real-life data, which provide more reliable and representative conclusions for the population. Third, there is no evidence of the long-term effectiveness and safety of the vaccine.
These trials had a short follow-up of up to 28 days after vaccination and unsolicited serious adverse events through 6 months after the second dose. It is necessary to highlight that the level of evidence of the outcomes evaluated in this review ranged between moderate and very low. This is primarily because the analysis was performed per-protocol, as planned for the interim analysis, which generates uncertainty in the results because the estimates may change during longer follow-up. The results of an RCT acquire greater validity when the patients are analysed according to the group in which they were assigned, i.e., application of the intention-to-treat principle. This application makes it possible to maintain the advantages of randomization, avoiding overestimates of the effects of the therapy under study, and admitting the non-adherence of some patients, which is a situation that is closer to reality. Finally, we found no major limitations in study designs in relation to randomization sequence, blinding of investigators, personnel involved in the study, or significant losses to follow-up. However, the sample sizes of the intervention group (n=14,964) and the control group (n=4,902) ( Table 3) in the RCT of the Gam-COVID-VacrAd26-S/rAd5-S vaccine were striking, but the distributions observed of the baseline characteristics between the study groups (28) were compatible with chance.

Implications for practice
Health care providers and professionals should communicate consistent, complete, and clear information on the benefits, risks, contraindications, safety issues, warnings and follow-up recommendations to persons receiving COVID-19 vaccines and their caregivers. It should be explained that there is limited evidence on how much COVID-19 vaccines reduce transmission ingeneral population and for how long protection lasts. Also, prevention and biosecurity guidelines should be strictly followed. This includes the use of masks, hand washing, social distance greater than 2 meters, avoiding crowdsand enclosed unventilated areas (38,39).
We should generate pharmacovigilance systems to identify and respond quickly to any adverse events in a recipient after vaccination, including vaccine administration errors, serious adverse events, multisystemic inflammatory syndrome cases, and COVID-19 cases resulting in hospitalization and death. It is recommended that any other clinically significant adverse event should be reported, even if its association with the administration of the vaccine is not clear.

Implications in research
It is necessary to keep track of the studies and reports that evaluate the beneficial and harmful effects of the different vaccines, which may be possible using living systematic reviews periodically informing the best practices in vaccine prevention and clinical research of this highly prevalent disease.
The results of this review support the level of evidence for the efficacy and safety of the COVID-19 vaccines that were analysed. The information presented in this manuscript has not been presented elsewhere.

Ethical approval and other permissions
Ethical approval for this study was not necessary because no patient data were collected.