The effectiveness of primary series CoronaVac vaccine in preventing COVID‐19 illness: A prospective cohort study among healthcare workers in Azerbaijan, May–November 2021

Abstract Background Healthcare workers (HCWs) have suffered considerable morbidity and mortality during the COVID‐19 pandemic. Few studies have evaluated the CoronaVac vaccine effectiveness (VE), particularly in Eastern Europe, where the vaccine has been widely used. Methods We conducted a prospective cohort study among HCWs in seven hospitals in Baku, Azerbaijan between May 17 and November 30, 2021, to evaluate primary series (two‐dose) CoronaVac VE against symptomatic SARS‐CoV‐2 infection. Participants completed weekly symptom questionnaires, provided nasopharyngeal swabs for SARS‐CoV‐2 RT‐PCR testing when symptomatic, and provided serology samples at enrollment that were tested for anti‐spike and anti‐nucleocapsid antibodies. We estimated VE as (1 – hazard ratio)*100 using a Cox proportional hazards model with vaccination status as a time‐varying exposure, adjusting for hospital and previous SARS‐CoV‐2 infection status. Results We enrolled 1582 HCWs. At enrollment, 1040 (66%) had received two doses of CoronaVac; 421 (27%) were unvaccinated. During the study period, 72 PCR‐positive SARS‐CoV‐2 infections occurred; 36/39 (92%) sequenced samples were classified as Delta variants. Primary series VE against COVID‐19 illness was 29% (95% CI: −51%; 67%) for the entire analysis period. For the Delta‐only period (July 1–November 30, 2021), primary series VE was 19% (95% CI: −81%; 64%). For the entire analysis period, primary series VE was 39% (95% CI: −40%; 73%) for HCWs vaccinated within 14–149 days and 19% (95% CI: −81%; 63%) for those vaccinated ≥150 days. Conclusions During a period in Azerbaijan characterized by mostly Delta circulation, VE point estimates suggested that primary series CoronaVac protected nearly 1 in 3 HCWs against COVID‐19, but 95% confidence intervals were wide, with lower bounds that crossed zero, reflecting the limited precision of our VE estimates. Our findings underscore the need to consider booster doses for individuals who have received the primary series of CoronaVac.


| INTRODUCTION
COVID-19 vaccination has been shown to be a critical intervention to reduce morbidity and mortality from COVID-19. 1 Protecting healthcare workers (HCWs) through COVID-19 immunization is essential for effective control of the COVID-19 pandemic; HCWs are highly exposed to infection, have frequent contact with vulnerable patients, and are essential to the ongoing function of health services. 2though many observational studies have evaluated the effectiveness of mRNA and viral vector vaccines, much less is known about the real-world effectiveness of whole inactivated virus vaccines.Inactivated whole-virus vaccines, which have the potential advantages of being easier to produce and store and also presenting a wider range of viral antigens to the immune system, 3 accounted for nearly half of the 7.3 billion COVID-19 vaccine doses delivered globally as of October 2021. 4Most of these doses have been used in low-and middle-income countries (LMICs), where few studies have evaluated the effectiveness of COVID-19 vaccines in general. 5In LMICs, differences in population demographics and differences in operational aspects of the vaccination campaign could potentially impact realworld vaccine effectiveness (VE). 6ronaVac (Sinovac, Beijing), a newly-developed inactivated vaccine authorized for emergency use by the World Health Organization (WHO) in June 2021, has been widely used in countries in Central Asia and other countries in the eastern part of the WHO European Region 7 ; however, only one study from the region-a study of HCWs in Turkey conducted during a period of Alpha variant predominancehas evaluated the effectiveness of this vaccine. 5,7 Azerbaijan, an upper-middle-income country in the WHO European Region with a population of approximately 10 million people, the COVID-19 vaccination campaign began on January 18, 2021.
Populations considered most at risk, including HCWs, were prioritized for early vaccination.CoronaVac was the main vaccine available at the beginning of the vaccination campaign.
We conducted a prospective cohort study of COVID-19 VE in HCWs in Azerbaijan.In this interim analysis of the study, we aimed to estimate the early VE of primary series (two-dose) CoronaVac against PCR-confirmed SARS-CoV-2 illness during May-November 2021.

| Study setting and population
The study design and analysis were guided by the WHO European Region HCW VE guidance document, 6 and the study was conducted within the framework of WHO's Unity platform. 8om May 3 to July 17, 2021, we enrolled HCWs at seven hospitals in the Baku United Hospital network in Baku, the capital city of Azerbaijan, into a prospective cohort.We selected hospitals based on their accessibility to the study team in the context of quarantine-related travel restrictions at the time.At the time of enrollment and during the study period, all sites admitted patients with COVID-19.
We offered enrollment to all HCWs employed by participating hospitals for whom COVID-19 vaccination was not contra-indicated by previous allergies to vaccinations.National guidance at the time required HCWs to wait 6 months from a PCR-confirmed infection before receiving the first dose.The national vaccination schedule recommended the administration of the second dose of CoronaVac 14-21 days after the first. 9sts of all HCWs employed at study sites were provided by hospital directors to guide recruitment.We invited all HCWs to participate, including physicians, nurses, clinical support staff, and custodial workers, regardless of their COVID-19 vaccination status, intention to get vaccinated in the future, or history of previous infection with SARS-CoV-2.At the time of enrollment, and throughout the analysis period, HCWs were required to have COVID-19 vaccination in order to work in hospitals in Azerbaijan; however, because of the limited availability of the vaccine in much of the first half of 2021, this requirement was not strictly enforced.

| Data collection and management
At enrollment, participants completed a questionnaire that included questions about demographics, comorbidities, previous SARS-CoV-2 infection, and SARS-CoV-2 vaccination history.Study staff then contacted participants weekly, using a standard questionnaire, to ask participants if they had experienced any symptoms in the past week (fever, cough, general weakness, fatigue, headache, muscle ache, sore throat, runny nose, shortness of breath, lack of appetite, nausea, vomiting, diarrhea, altered mental status, loss of taste, or loss of smell), and to ask about the details of any new COVID-19 vaccines participants had received.
We advised all participants who became symptomatic to attend pants to resolve discrepancies between data from questionnaires and those from national databases, and to complete questions from study questionnaires that had not been answered.

| Laboratory testing
Blood was collected from all participants at enrollment, and serum was stored at À20 C until testing.Sera samples were tested for anti-spike antibodies with the Wantai SARS-CoV-2 total antibody ELISA (Beijing Wantai Biological Pharmacy, Beijing, China) and for anti-nucleocapsid antibodies using the Platelia SARS-CoV-2 total antibody ELISA (Bio-Rad Laboratories, Hercules, CA) in the InterDiagnostic Clinic Laboratory, Baku.The manufacturer's recommended controls and thresholds were used to define seropositivity.For participants who had hemolyzed or insufficient specimens, repeat blood draws were performed within 30 days of enrollment.
We selected a convenience sample for sequencing from PCRpositive samples from participants at all study sites.SARS-CoV-2 PCR-positive study samples were sent to the Charite University Institute of Virology laboratory in Berlin, Germany, for whole genome sequencing (WGS).Only samples with low cycle threshold values were sequenced.

| Sample size estimation
We estimated that 1500 participants needed to be enrolled in the study to identify VE against symptomatic infection of 80% with an incidence of SARS-CoV-2 of 0.05 and vaccine coverage among participants of 80% to reach an 80% power level with alpha = 0.05.This estimate accounted for a likely drop-out rate of roughly 10%.

| VE analysis
For our primary study outcome, we measured two-dose CoronaVac VE against symptomatic PCR-confirmed COVID-19 for the overall cohort.We considered a symptomatic COVID-19 illness to be an event where the participant reported having symptoms anytime between 14 days before and 4 days after the swab date of the PCR-positive test.We calculated unadjusted and adjusted HRs and also calculated VE estimates.We categorized hospitals into two groups based on geographical location: three hospitals in central Baku were considered "central," whereas four hospitals located on the outskirts of Baku were considered "peripheral."Both unadjusted and adjusted HR estimates included hospital groups and prior infections as fixed effects.

| Statistical model
We assessed other prespecified potential confounders (e.g., month, age, sex, occupation, hands-on care [which we defined as clinical HCWs who reported providing hands-on medical care to patients], smoking, household size, any chronic condition, and BMI).In order to avoid overfitting the model, we included variables that changed the VE estimate by an absolute percentage of more than 5% using step-wise backward selection.4) the day of the last weekly questionnaire before complete loss to follow-up, withdrawal from the study, transfer or retirement from their hospital of employment, death, or censor date for the analysis period (November 30, 2021).Participants were also censored from the primary analysis upon receiving a dose of any COVID-19 vaccine other than CoronaVac.Because of the high level of completeness in the available data, we opted for a complete case approach.

| Further analyses and sensitivity analyses
We only conducted VE analyses for CoronaVac because few other vaccines were used in the study population.We evaluated VE during the overall study period (May 17-November 30, 2021) and separately for the period in which only SARS-CoV-2 B.1.617.2 viruses (the Delta variant) were sequenced (July 1, 2021-November 30, 2021), which we defined using WGS data from study samples along with publicly available data from Global Initiative on Sharing All Influenza Data (GISAID). 11,12We also performed these two analyses excluding participants who had a PCR-confirmed SARS-CoV-2 infection prior to enrollment.In addition, we examined VE since time since vaccination by comparing VE in the period from 14-149 days since the second vaccine dose to VE ≥ 150 days since the second vaccine dose.
We performed three sensitivity analyses.For the first, because reinfection can occur earlier than 90 days, 13 we changed the definition of "time at risk" from 90 days after infection to 60 days after infection.For the second, because individuals may be protected as early as 7 days after their second dose, 14 we defined "fully vaccinated" as 7 days after the second dose rather than 14 days after the second dose.Finally, we conducted a sensitivity analysis to evaluate the potential impact of unmeasured confounding using the E-value approach, assuming both possible directions of bias-overestimation and underestimation. 15

| RESULTS
We enrolled 1582 HCWs, which comprised 38.5% of eligible HCWs in participating hospitals.Three participants were excluded because no follow-up data were obtained after enrollment (Figure S1).
At enrollment, 1040 (66%) participants had received the primary series of CoronaVac, 121 (8%) participants had received only one dose of CoronaVac, and 421 (27%) were unvaccinated (Table 1).No participants had received other COVID-19 vaccines at enrollment.Among those vaccinated with the primary series before enrollment, the median time since receipt of the second dose was 99 days (IQR: 74-112).Overall, 248 HCWs (16%) had a PCR-positive COVID-19 infection documented prior to enrollment, and 204/248 (84%) prior infections occurred before HCWs had received their first vaccine dose.Compared with unvaccinated participants, participants who had received the primary series vaccine before enrollment worked more commonly in central hospitals (39% vs. 29%), had more chronic conditions (38% vs. 48%) and had fewer PCR-confirmed infections prior to enrollment (6% vs. 40%).Differences in participants by vaccination status at enrollment have been previously reported. 16 total, 963/1040 (93%) participants who had received the primary series vaccine were seropositive at enrollment for either anti-nucleocapsid or anti-spike protein antibodies; 903 (90%) were seropositive for anti-spike antibodies, whereas 827 (83%) were seropositive for anti-nucleocapsid antibodies.Among unvaccinated participants, 363/421 (86%) were seropositive to at least one of the two antibodies (Table 1).Of the 197 unvaccinated participants who did not report a previous PCR infection at enrollment, 136 (69%) were seropositive at enrollment by one of the two assays.
During the follow-up period, 10 (<1%) participants received primary series vaccination with vaccines other than CoronaVac (9 received Pfizer and 1 received Sputnik) and were excluded from the analysis.At their exit from analysis, 1485 (95%) participants had received the primary series of CoronaVac, 41 (3%) had received one dose of CoronaVac, and 43 (3%) remained unvaccinated.Changes in the vaccination status of the enrolled population over the course of for which WGS data were available, 36/39 (92%) were Delta variants (Figure S2).
During the course of the analysis period, in the 30 days following For the overall cohort, two-dose VE was 29% (95% CI: À51%-67%) (Table 2, Figure 3A).For the Delta-only period, two-dose VE was 19% (95% CI: À81%-64%).VE was adjusted for previous infection only; no other potential confounders changed the VE estimates by more than 5% (Table S2).For the overall cohort analysis, vaccinated participants had received their second dose a median of 90 days (IQR: 75-112) prior to the beginning of the study period, while during the period of Delta circulation, vaccinated participants  had received their second dose a median of 108 days (IQR: 106-132) prior to the analysis period.
In sensitivity analyses, when we decreased the period after infection that participants were considered not at risk from 90 to 60 days, results were very similar (<1% difference) to the 90-day analysis (Table 4).When we considered a participant to be fully vaccinated at 7 days instead of 14 days, VE differed by <2% from the results of the primary analysis.
Our calculation of E-values indicated that an unmeasured confounder would have had to have had at least associations of ≥3 with both the exposure and the outcome to fully explain away our observed VE of the primary analysis if the true VE was outside of 0%-50% (Figure S3).T A B L E 4 Two-dose CoronaVac effectiveness against symptomatic PCR-confirmed COVID-19 infection for full cohort for entire study period (a) when re-infection could occur 60 days (rather than 90 days) after infection; and (b) when participants were considered fully vaccinated 7 days (rather than 14 days) after their second dose.Delta-predominant period. 18In contrast, two studies from Asia found higher two-dose VE against illness and infection during periods of Delta-predominant circulation, but these studies included individuals who had mostly been recently vaccinated.A study from Thailand that used the test-negative design reported two-dose CoronaVac effectiveness of 60% (95% CI: 49-69) against infection; vaccinated participants in that study had received their last dose a mean of 81 days (range: 60-91 days) prior to the analysis period. 19A study related to an outbreak of the SARS-CoV-2 Delta variant in China reported a two-dose CoronaVac VE against illness rate of 73.0 (95% CI: 22.3-90.6)among individuals who had mostly received their second vaccine dose within the previous 3 months. 20r individuals who received the primary vaccine series, VE against infection during Delta-predominant periods has been shown to be mostly higher for mRNA and viral vector vaccines compared with CoronaVac. 21However, against Omicron, two-dose VE has been much lower against symptomatic infection and, to a lesser extent, severe disease across vaccine products. 22Homologous and heterologous monovalent booster doses have been shown to increase protection against both mild and severe COVID-19 illnesses in Delta and Omicron. 22 63.9-80.7])for those vaccinated 8-59 days prior to the analysis period. 18Differences persisted for those boosted >59 days prior.
However, in Hong Kong Special Administrative Region, also during a period of Omicron BA.2 circulation, VE against mortality and severe complications was mostly similar for individuals who had received a primary series of CoronaVac followed by Comirnaty as a booster compared with those who had received three doses of CoronaVac. 23Both studies showed increased VE against all endpoints for booster doses compared with the primary series.In Azerbaijan, where less than 10% of the adult population has received a booster vaccine, 24 conveying the important benefits of booster doses to the public is critical.
Our study population likely had high rates of previous infection at enrollment.Because vaccination with inactivated vaccine leads to seroconversion in both anti-spike and anti-nucleocapsid antibody tests, we could not use antibody testing to determine previous infection among participants who had been vaccinated prior to enrollment.
However, 69% of unvaccinated HCWs who did not report a previous infection were seropositive by at least one of the two antibody tests at enrollment, and these findings likely reflect the extent of previous infection in the overall study population.Despite more than two of every three participants likely having been previously infected, we still found some benefit, albeit not statistically significant, to primary series vaccination with CoronaVac.The added benefit of primary COVID-19 vaccination and booster doses in previously infected individuals (hybrid immunity) has been widely demonstrated in other studies. 25 our study, we found a trend towards decreased VE among participants for whom more than 5 months had passed since their second CoronaVac vaccine.Waning VE with increased time since COVID-19 vaccination has been described for CoronaVac and other COVID-19 vaccines. 26,27The waning effectiveness of primary series vaccination again underscores the importance of booster doses to increase protection.
Our study has a number of strengths.Because we enrolled and systematically followed a discrete cohort of HCWs, we were able to obtain data about vaccination status, SARS-CoV-2 test results, and clinical outcomes, information that would not have been discernible from routinely collected data.The protocol was followed rigorously; participants completed more than 95% of the weekly symptom questionnaires during the study period.Only 34 (2%) participants were lost to follow-up.Finally, the use of serology at enrollment allowed us to estimate the prevalence of prior infections among unvaccinated participants and also provided information on seroconversion rates among HCWs who received the inactivated CoronaVac vaccine.
testing facilities at the study hospitals, where trained nurses collected nasopharyngeal swabs, which were tested for SARS-CoV-2 by RT-PCR at on-site government-accredited laboratories or other Ministry of Health (MoH) laboratories.Symptomatic participants completed an additional survey that included information about the date of symptom onset, clinical care-seeking, and details of PCR testing and results.PCR-positive participants were interviewed again 30 days after their positive test, at which time further details about their course of illness, medical care, hospitalization, and complications were collected.Data from interviewer-led questionnaires and laboratory records were entered and stored securely using the Sorgular.azplatform (Azerbaijan Public Health Reform Center).Participants' reports of positive PCR results were verified using the two national SARS-CoV-2 laboratory databases-the Etabib electronic medical records database 10 and the MoH/Mandatory Health Insurance database-to which all SARS-CoV-2 PCR tests performed in public and private laboratories in the country are required to be reported.Participants' COVID-19 vaccination history was verified using the national vaccine registry (Rendezvous, Azerbaijan MoH).Study staff contacted partici- VE was estimated as (1hazard ratio [HR])*100.HRs comparing vaccinated and unvaccinated individuals were estimated using Cox proportional hazards models with vaccination as a time-varying exposure; the vaccination status of some individuals changed over time from unvaccinated to vaccinated, and therefore the same participant could contribute person-time to both exposure categories.Study time was used as the underlying time scale in the Cox regression model.Participants were considered fully vaccinated 14 days after they received their second vaccine dose.Age was collected as years and successively grouped into five categories (20-29 years, 30-39 years, 40-49 years, 50-59 years, and 60+).
We defined previous infection as aPCR-confirmed infection prior to enrollment documented in either the Etabib database or the MoH/Mandatory Health Insurance database.Participants who were unvaccinated at enrollment and had PCR-confirmed SARS-CoV-2 infection prior to enrollment began to contribute person-time on the date they became eligible for vaccination (6 months after their last positive test).Participants who were vaccinated at enrollment and who had PCR-confirmed COVID-19 infection prior to enrollment were included in the analysis at the time point they were considered "at risk" of reinfection, which we defined as 90 days after their most recent positive PCR test.Participants contributed person-time from enrollment, or, for individuals with prior PCR-confirmed SARS-CoV-2 infection, from the start of time at risk until the earliest outcome or exit from the study.Person-time ended at whichever came first of the following outcomes: (1) the day of the first SARS-CoV-2 infection, designated by the date of symptom onset (for participants with symptomatic COVID-19 infection) or by the swab date of the PCR test (for asymptomatic participants); (2) the day of receipt of a second COVID-19 vaccination if it occurred before the recommended interval between the first and second dose; (3) the day of receipt of a third vaccine dose, or ( The study was approved by theWHO Research Ethics Review Committee (protocol: CERC.0097C) and the Ethics Committee of Azerbaijan State Academy of Physical Culture and Sport (March 3, 2021; Protocol #3/21).All participants provided informed, written consent.The study is registered in the clinicaltrails.govregistry (NCT050694).
their positive test, 44 participants with PCR-confirmed COVID-19 illness sought medical care (4 unvaccinated and 40 fully vaccinated), and 29 participants went to an emergency room (3 unvaccinated and 26 fully vaccinated).Three participants with PCR-confirmed COVID-19 illnesses were hospitalized (1 unvaccinated and 2 fully vaccinated).No deaths occurred among participants with PCR-confirmed COVID-19 illnesses.VE could not be calculated against these more severe outcomes because of the small number of events among unvaccinated participants.

F
I G U R E 1 COVID-19 vaccine coverage among study participants, by epidemiologic week, Azerbaijan, 2021.F I G U R E 2 Epicurve showing COVID-19 symptomatic cases by vaccination status in the study population (left y-axis), and SARS-CoV-2 cases nationally (right y-axis), by epidemiologic week, Azerbaijan, May 17-November 30, 2021.*Data on weekly SARS-CoV-2 incidence in Azerbaijan were taken from WHO COVID-19 tracker.28 occupational, and clinical characteristics of participants at enrollment, by site,Azerbaijan, 2021.
*"Hands-on care" refers to clinical HCWs who reported providing "hands-on medical care to patients."**Chronic conditions include: cancer, chronic heart disease, high blood pressure/hypertension, chronic kidney disease, chronic liver disease (such as cirrhosis, hepatitis, and fatty liver disease), chronic lung disease (such as asthma and COPD), diabetes, immunocompromised (including solid organ transplant and HIV), neurologic disease (including cerebrovascular disease, epilepsy, and multiple sclerosis), obesity, and autoimmune disorder.
Two-dose CoronaVac effectiveness against symptomatic PCR-confirmed COVID-19 infection for entire cohort and for entire cohort excluding participants who had a PCR-confirmed SARS-CoV-2 infection prior to enrollment, during the total study period, and for Delta-predominant period only, Azerbaijan, 2021.Note: The total number of participants is not equal to the sum of the groups because one participant could have contributed person-time to more than one group during the course of the analysis period.Abbreviations: HR, hazard ratio; VE, vaccine effectiveness.
T A B L E 2 *Adjusted by hospital group.**Adjusted by hospital group and previous infection status.***Adjusted VE could not be calculated.
Two-dose CoronaVac effectiveness against symptomatic PCR-confirmed COVID-19 infection for entire cohort for total study period, and for Delta-predominant period only, by time since receipt of second vaccine, Azerbaijan, 2021.Note: The total number of participants is not equal to the sum of the groups because one participant can contribute to the person-time simultaneously in several groups.Abbreviations: HR, hazard ratio; VE, vaccine effectiveness.*Adjusted by hospital group and previous infection status.