Healthcare workers (HCWs) are at risk of coronavirus disease 2019 (COVID-19) due to high levels of exposure. When compared to the general population, frontline HCWs have >10 times the risk of testing positive for severe acute respiratory coronavirus virus 2 (SARS-CoV-2), and those reporting inadequate access to personal protective equipment (PPE) have a 23% higher risk. Reference Mutambudzi, Niedwiedz and Macdonald1,Reference Nguyen, Drew and Graham2 In addition, when compared to HCWs reporting adequate access to PPE and who were not caring for patients with COVID-19, HCWs caring for patients with documented COVID-19 had a nearly 5-times higher risk of testing positive if they had adequate access to PPE and a nearly 6-times higher risk if they had inadequate access to PPE. Reference Nguyen, Drew and Graham2 These reports emphasize the need for effective vaccines, especially among frontline HCWs.
Over the last few months, multiple studies have yielded a large amount of data from different institutions that provided real-world data on short-term vaccine effectiveness. Reference Dagan, Barda and Kepten3,Reference Tenforde, Olson and Self4 The great majority of these studies examined COVID-19 mRNA vaccines that significantly prevented symptomatic and asymptomatic severe acute respiratory coronavirus virus 2 (SARS-CoV-2) infection among HCWs. Reference Marra, Kobayashi and Suzuki5 However, data regarding the vaccine effectiveness of other COVID-19 vaccines (eg, viral vector or inactivated virus) are limited.
We investigated real-world vaccine effectiveness for Oxford-AstraZeneca (ChAdOx1), and CoronaVac against laboratory-confirmed SARS-CoV-2 infection among HCWs.
Methods
Population and setting
This retrospective cohort study was conducted between January 1, 2021, and August 3, 2021, in Brazil. We included all adult HCWs (aged ≥18 years) working at the Hospital Israelita Albert Einstein (HIAE). The HIAE is a Brazilian, nonprofit, healthcare, education, and research organization, with its headquarters in the city of São Paulo. HIAE manages a diverse healthcare system, including primary healthcare to tertiary-care services, in the public and private healthcare sectors. This hospital operates 40 healthcare units, mainly in the state of São Paulo, and in 2020 it had ∼700,000 emergency department visits, 900,000 outpatient visits, and 70,000 hospital discharges overall. Since the beginning of the COVID-19 pandemic, HCWs with COVID-19 symptoms had access to free-of-charge SARS-CoV-2 RT-PCR testing conducted by the institution’s laboratory. Vaccination with 2 doses of CoronaVac vaccine, 21 days apart, and with 2 doses of the ChAdOx1 vaccine, 12 weeks apart, were evaluated for COVID-19 vaccine effectiveness. HCWs were considered unvaccinated if no COVID-19 vaccine doses were received. Individuals who tested positive for SARS-COV-2 prior to the first vaccine dose, between vaccine doses, or before 14 days after the second vaccine dose, and individuals who had been vaccinated before the study period, were excluded from the study. We also excluded HCWs who received the Pfizer COVID-19 vaccine because the sample size (∼130 HCWs) was too small to obtain an estimate of vaccine effectiveness (Supplementary Appendix 1 online).
Real-time polymerase chain reaction (RT-PCR) methodologies for SARS-CoV-2 detection
The diagnostic confirmation of COVID-19 was performed using RT-PCR on specimens obtained via nasopharyngeal swab, according to the protocol instituted at the hospital. The following RT-PCR kits were utilized: XGEN MASTER COVID-19 (Mobius, Pinhais, Paraná, Brazil); cobas SARS-CoV-2 Test (Roche Molecular Systems, Branchburg, NJ); Xpert Xpress SARS-CoV-2 (Cepheid, Sunnyvale, CA); and Abbott RealTime SARS-C0V-2 (Abbott Molecular, Des Plaines, IL).
Next-generation sequencing of viral full-length genome
We extracted total nucleic acid from the naso-oropharyngeal (NOP) swab samples with QIAamp Viral RNA Mini kit (QIAGEN, Hilden, Germany). After purification and concentration, DNAse I treatment, and depletion of human ribosomal RNA, the samples were subjected to random amplification. Reference Greninger, Naccache and Federman6 The preparation of sequencing libraries for the Illumina platform was carried out with DNA Prep (Illumina, San Diego, CA) using the random 2-step PCR amplification product as input. The libraries were quantified with the Qubit instrument (Thermo Fisher Scientific, Waltham, MA) and were loaded on the NextSeq 550 equipment (Illumina) for sequencing with MID 300 paired-end reads (Illumina).
Outcome measures and statistical analyses
Laboratory-confirmed COVID-19 was considered the primary outcome to calculate vaccination effectiveness after 2 doses of a COVID-19 vaccine (CoronaVac or ChAdOx1). RT-PCR testing for the diagnosis of COVID-19 was performed only on symptomatic HCWs. Hospitalization related to COVID-19, length of stay, ICU admission, necessity of mechanical ventilation and death were secondary outcomes. Vaccination status and SARS-CoV-2 RT-PCR results of all study participants were obtained from institutional electronic records. We excluded those with a positive COVID-19 diagnosis before January 15, 2021, because it corresponded to data from the first vaccine (January 1, 2021) plus 14 days. For those vaccinated, the initial follow-up date was 14 days after the second vaccine dose. The last date was defined as the date COVID-19 was diagnosed, or up to August 3, 2021, for the censored cases without a positive diagnosis of COVID-19.
The qualitative variables were characterized using absolute and relative frequencies in general and by interest groups; for comparisons, we used the χ Reference Nguyen, Drew and Graham2 or the Fisher exact test. The quantitative variables have been reported as medians, interquartile range (IQR, first and third quartiles), minimum and maximum values due to the asymmetry observed in the variables, Reference Ashby7 and comparisons were performed via nonparametric Kruskal-Wallis tests. Vaccine effectiveness was calculated as 1 − rate ratio (RR), Reference Nauta8 with RR obtained by adjusting Poisson models with SARS-CoV-2 infection confirmed by RT-PCR as the outcome and vaccination status as the main exploratory variable, in addition to 95% confidence intervals. We used the logarithmic link function to estimate unadjusted models and models adjusted for sex, age, and HCW job type (ie, direct patient contact vs no direct patient contact). The cumulative incidence curves of COVID-19 for the vaccinated and unvaccinated groups were estimated using the Kaplan-Meier method Reference Klein and Moeschberger9 and the cumulative incidence estimated at 30 and 90 days with unadjusted models. All analyses were performed with the R software environment for statistical computing and graphics version 4.1.0. 10 All reported tests were 2-sided, and P < .05 was considered significant. The study was approved by the Hospital Israelita Albert Einstein Ethics Committee (no. CAAE 47110421.7.0000.0071), which waived the need for informed consent.
Results
By the end of the study period, 18,359 individuals were screened for eligibility to evaluate the COVID-19 vaccination effectiveness after the second vaccine dose. Overall, 13,813 HCWs met inclusion criteria (Supplementary Appendix 1 online). Among this cohort, 6,385 (46.2%) received the CoronaVac vaccine, 5,916 (42.8%) received the ChAdOx1 vaccine, and 1,512 (11.0%) were unvaccinated. Most were female (71.0%), and the median age of the entire study population was 35 years. Unvaccinated workers were younger, and women more frequently received 2 doses of CoronaVac (Table 1). The proportions of HCWs with direct patient contact were ∼20% among the unvaccinated, 80% among those with CorovaVac, and ∼20% among those with ChAdOx1. Of 13,813 HCWs, past medical history was available for 10,786 HCWs (78.1%). Among them, 2,783 (25.8%) had at least 1 comorbidity: obesity (n = 1,009, 9.4%), hypertension (n = 843, 7.8%), dyslipidemia (n = 592, 5.5%), asthma (n = 478, 4.4%), and diabetes mellitus (n = 250, 2.3%).
Baseline Characteristics of Study Participants, Hospital Israelita Albert Einstein, São Paulo, Brazil, from January 1, 2021, to August 3, 2021
Note. ChAdOx1 vaccine, Oxford-AstraZeneca vaccine; IQR, interquartile range; ICU, intensive care unit.
a χ Reference Nguyen, Drew and Graham2 test.
b Kruskal-Wallis test.
c Information available for 10,786 participants, 943 unvaccinated, 4,749 with 2 doses of CoronaVac vaccine and 5,094 with 2 doses of ChAdOx1 vaccine.
d Mann-Whitney test.
e Follow-up was initiated 15 days after the second dose for those vaccinated.
f Fisher exact test.
During the study period, 325 HCWs (2.4%) were diagnosed with COVID-19. For the unvaccinated HCWs, the cumulative incidences of COVID-19 were 0.73% at 30 days of follow-up and 3.57% at 90 days of follow-up. The cumulative incidences were 0.85% at 30 days and 2.32% at 90 days for those vaccinated with the CoronaVac vaccine. The cumulative incidences were 0.58% in 30 days and 0.73% in 90 days for those vaccinated with the ChAdOx1 vaccine (Fig. 1). None of the HCWs (vaccinated or unvaccinated) died during the study period.
Cumulative incidence of COVID-19 infection (by RT-PCR) among vaccinated (2 doses of CoronaVac vaccine, and 2 doses of ChAdOx1 [Oxford-AstraZeneca] vaccine) and unvaccinated healthcare workers.
The estimated vaccine effectiveness for CoronaVac after 2 doses was 50.3% (95% CI, 36.2%–61.3%), and the vaccine effectiveness for ChAdOx1 vaccine after 2 doses was 87.9% (95% CI, 82.6%–91.6%). After controlling for sex, age, and professional category, the estimated vaccine effectiveness rates were 51.3% (95% CI, 34.6%–63.7%) for CoronaVac after 2 doses and 88.1% (95% CI, 82.8%–91.7%) for ChAdOx1 vaccine after 2 doses (Table 2).
Observed Rate Ratios and Vaccine Effectiveness Among Healthcare Workers after COVID-19 Vaccine Second Dose and COVID-19 Infection by RT-PCR, Hospital Israelita Albert Einstein, São Paulo, Brazil, from January 1, 2021, to August 3, 2021
Note. RT-PCR, real-time polymerase chain reaction; RR, rate ratio; CI, confidence interval; ChAdOx1, Oxford-AstraZeneca vaccine; HCW, healthcare worker.
Whole-genome sequencing analysis
From March to June 2021, 19 SARS-CoV-2 samples from 19 HCWs were screened for mutations of interest. Of those, 18 were the P1 strain (γ variant), and 1 B.1.1.7 strain (ie, the α [alpha] variant) was identified (Table 3).
Characteristics of Participants With SARS-CoV-2 Variants of Concern (n=19), Hospital Israelita Albert Einstein, São Paulo, Brazil, from January 1, 2021, to August 3, 2021 a
Note. P1, γ (gamma) variant; B.1.1.7, α variant; ChAdOx1, Oxford-AstraZeneca vaccine; DPC, direct patient facing; NDPC, no direct patient contact.
a The whole-genome sequencing was performed from March to June 20
Discussion
This retrospective study revealed that the estimated vaccination effectiveness among HCWs against symptomatic COVID-19 were 51.3% for CoronaVac and 88.1% for AstraZeneca after adjusting for age, sex, and job type. Both a viral vector vaccine (ChAdOx1) and an inactivated viral vaccine (CoronaVac) reduced the number of COVID-19 cases, the number of hospitalizations, the length of hospital stay, and the need for mechanical ventilation. These vaccines were even effective against a new variant of concern in Brazil, the γ (gamma) variant.
Based on a recent published systematic literature review evaluating short-term vaccination effectiveness between December 2020 and April 2021, COVID-19 vaccines (primarily the mRNA vaccines) decrease symptomatic COVID-19 infection with vaccine effectiveness of 92.8%. Reference Marra, Kobayashi and Suzuki5 Our study showed that the estimated vaccine effectiveness rates after 2 doses of CoronaVac and ChAdOx1 among HCWs were lower than the vaccine effectiveness rates of mRNA COVID-19 vaccines among the general population reported in the randomized trials Reference Polack, Thomas and Kitchin11,Reference Baden, El Sahly and Essink12 and also in a noncontrolled setting. Reference Dagan, Barda and Kepten3 A randomized clinical trial evaluating vaccine effectiveness of CoronaVac among HCWs in Brazil reported vaccine effectiveness after 2 doses of 50.7%. Reference Palacios, Batista and Albuquerque13 Another randomized clinical trial of CoronaVac in Turkey reported an estimated vaccine effectiveness after 2 doses of 83.5%. Reference Tanriover, Doğanay and Akova14 Both clinical trials were conducted in 2020, prior to the emergence of the variants of concern. More recent observational studies evaluating the vaccine effectiveness of the CoronaVac vaccine did not include genomic surveillance for SARS-CoV-2 virus but reported the circulation of at least 2 viral lineages considered to be variants of concern: B.1.1.7 (α variant) Reference Jara, Undurraga and González15 and P.1 (γ variant). Reference Jara, Undurraga and González15,Reference Hitchings, Ranzani and Torres16 Results from these studies after 2 doses, in a prospective national cohort study from Chile, demonstrated that the estimated vaccine effectiveness of CoronaVac in a general population was 65.9%. Reference Jara, Undurraga and González15 Results from a test-negative case–control study of the vaccine effectiveness of CoronaVac vaccine among HCWs in Manaus, Brazil, where the γ variant was also predominant, showed that the estimated vaccine effectiveness after 2 doses was low (36.8%) against COVID-19. Reference Hitchings, Ranzani and Torres16 In terms of the ChAdOx1 vaccine, a randomized clinical trial conducted in Brazil, South Africa, and the United Kingdom showed that the vaccine effectiveness of ChAdOx1 after 2 doses was 62% against COVID-19. Reference Voysey, Clemens and Madhi17 Results from a test-negative case–control study evaluating the vaccination effectiveness of ChAdOx1 vaccine after 2 doses for the B.1.1.7 (α variant) and the δ (delta) variant were 74.5% and 67.5%, respectively. Reference Lopez Bernal, Andrews and Gower18 Although it is not clear why the vaccine effectiveness rates in our present study were higher compared to vaccine effectiveness rates in previously published studies. Reference Voysey, Clemens and Madhi17,Reference Lopez Bernal, Andrews and Gower18 Possible explanations are the strict infection control policies in our institution and adequate PPE throughout the pandemic while many other institutions suffered critical shortages of PPE. Reference Mutambudzi, Niedwiedz and Macdonald1,Reference Nguyen, Drew and Graham2 Although a peak of COVID-19 was observed in March–June 2021, the community incidence of COVID-19 was relatively stable and the γ (gamma) variant was dominant during the study period.
ChAdOx1 (Oxford-AstraZeneca) and CoronaVac were the first COVID-19 vaccines authorized by the Brazilian Heath Surveillance Agency, 19 and HCWs were considered the priority group to receive them as of January 2021. Reference Mehrotra, Janes and Fleming20 We observed that HCWs who received 2 doses of CoronaVac were more likely to provide direct patient care in comparison to HCWs that received 2 doses of the ChAdOx1 vaccine. This difference can be explained by the CoronaVac vaccine being the first available COVID-19 vaccine in our institution, and for that reason, the frontline HCWs were prioritized to receive the COVID-19 vaccine. Later in the pandemic, our institution started using ChAdOx1 vaccine, which was mainly given to nonclinical persons. The duration of our study (8 months) among HCWs is justified, particularly to understand the short-term vaccination effectiveness in the context of a global pandemic with a novel pathogen. Reference Hodgson, Mansatta, Mallett, Harris, Emary and Pollard21 We collected data during a rapid vaccination campaign during a period with one of the highest community transmission rates of the pandemic, which allowed for a relatively short follow-up period and the estimation of the prevention of COVID-19 cases, related hospitalization, necessity of mechanical ventilation, and ICU stay. Both CoronaVac and ChAdOx1 vaccines were effective at preventing COVID-19 and serious illness (hospitalizations, necessity of mechanical ventilation and ICU care).
During the HCW COVID-19 vaccine campaign, the dominant variant in circulation was P.1 (γ variant), and both COVID-19 vaccines (CoronaVac and ChAdOx1) showed effectiveness against this variant. More studies are needed regarding the SARS-CoV-2 variants of concerns (VOC) that have multiple spike-protein mutations and appear to be more infectious or cause more disease than other circulating SARS-CoV-2 variants. Reference Challen, Brooks-Pollock, Read, Dyson, Tsaneva-Atanasova and Danon22 Some deletions in the spike-protein mutations can alter the shape of the spike and may help it evade some antibodies. Reference Wang, Nair and Liu23 No COVID-19 vaccine is 100% effective against SARS-CoV-2 infection, consistent with COVID-19 breakthrough infections reported in HCWs after COVID-19 vaccination. Reference Hacisuleyman, Hale and Saito24,25 We detected a clear effect of the vaccines against the new variants (mainly P.1).
Our study had several limitations. First, this was an observational study, subject to multiple biases Reference Harris, Lautenbach and Perencevich26 ; however, this is the most common study design in the infection prevention literature. Reference Harris, Lautenbach and Perencevich26 Second, we estimated vaccine effectiveness based on short-term duration, and longer-term observational studies are needed to assess sustained immune response and vaccine effectiveness. Third, due to the uncertainty related to the number of days required to develop immunity postvaccination, we decided to adopt the CDC definition for the CoronaVac vaccine and for the ChAdOx1 vaccine, which defines people fully vaccinated as being ≥14 days after the second dose in a 2-dose series (Pfizer/BioNTech or Moderna), or ≥14 days after a single-dose vaccine (Johnson & Johnson/Janssen). 27 Other studies adopted different definitions of a fully vaccinated person. Reference Marra, Kobayashi and Suzuki5 Currently, no postvaccination time limit on fully vaccinated status has been established. In addition, the CDC defines unvaccinated people as individuals of all ages including children who have not completed a vaccination series or have not received a single-dose vaccine. 27 Fourth, we have not reported nonneutralizing viral antigen-binding antibody levels in our HCW cohort study. However, the US Food and Drug Administration (FDA) does not recommend antibody testing for SARS-CoV-2 to determine immunity or protection from COVID-19, especially among those who are vaccinated. 28 Lastly, since our study focused only the short-term vaccination effectiveness among HCWs, we could not evaluate the need for a third dose. Consolidated knowledge indicates that each HCW needs to get 2 doses of CoronaVac or 2 doses of ChAdOx1 vaccine; thus, we decided to not report the analysis of vaccine effectiveness after 1 dose only. Considering our data regarding vaccine effectiveness for both COVID-19 vaccines, our institution began administering a third dose to HCWs in October 2021, after authorization from the Ministry of Health.
In conclusion, both COVID-19 vaccines (viral vector and inactivated virus) can significantly prevent COVID-19 among HCWs. The 2 COVID-19 vaccines were also effective among HCWs even after an emergence of a new variant (ie, the γ variant). More observational studies are needed to evaluate vaccination effectiveness of other COVID-19 vaccines (eg, other types of viral vector or inactivated virus). Studies are also needed to evaluate the impact of COVID-19 vaccines on personal protective equipment among HCWs, on vaccine effectiveness, and on COVID-19 breakthrough infection. Also, the vaccine effectiveness of unmatched COVID-19 vaccines as a third dose should be evaluated. Further genomic surveillance is needed for better understanding of vaccine effectiveness against the new SARS-CoV-2 variants.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/ice.2022.50
Acknowledgments
We thank all the participants for their contributions to this study.
Financial support
The sequencing reactions carried out to characterize the circulating SARS-CoV-2 described in this study were supported by Chamada MCTIC/CNPq/FNDCT/MS/SCTIE/Decit 07/2020 in Brazil (grant no. 402669/2020-7).
Conflicts of interest
All authors report no conflict of interest relevant to this article.
Open access
