Since January 2020 Elsevier has created a COVID-19 resource centre with

free information in English and Mandarin on the novel coronavirus COVID-

19. The COVID-19 resource centre is hosted on Elsevier Connect, the
company's public news and information website.

Elsevier hereby grants permission to make all its COVID-19-related

research that is available on the COVID-19 resource centre - including this
research content - immediately available in PubMed Central and other
publicly funded repositories, such as the WHO COVID database with rights
for unrestricted research re-use and analyses in any form or by any means
with acknowledgement of the original source. These permissions are
granted for free by Elsevier for as long as the COVID-19 resource centre
remains active.
 Vaccine 40 (2022) 3294–3297

Contents lists available at ScienceDirect

Vaccine
journal homepage: www.elsevier.com/locate/vaccine

Short communication

Effectiveness of Covishield vaccine in preventing Covid-19 –

A test-negative case-control study
Stuti Pramod, Dhanajayan Govindan, Premkumar Ramasubramani, Sitanshu Sekhar Kar ⇑, Rakesh Aggarwal,
JIPMER vaccine effectiveness study group1
Jawaharlal Institute of Postgraduate Medical Education and Research, Puducherry, India

a r t i c l e i n f o a b s t r a c t

Article history: Introduction: This study aimed at assessing the vaccine effectiveness (VE) of Covishield, which is identical
Received 29 July 2021 to AstraZeneca vaccine, in preventing laboratory-confirmed Covid-19.
Received in revised form 9 January 2022 Methods: Using test-negative case-control design, information on vaccination status of cases with Covid-
Accepted 2 February 2022
19 among healthcare workers in our institution in Puducherry, India, and an equal number of controls
Available online 9 February 2022
matched for age and date of testing, was obtained. The groups were compared using multivariable con-
ditional logistic regression to calculate odds ratios (OR). VE was calculated as 100*(1–adjusted odds
Keywords:
ratio)%.
Covishield
Vaccine effectiveness
Results: Using data from 360 case-control pairs, VE of one dose and of two doses, in providing protection
Test-negative design against Covid-19 was 49% (95% CI: 17%-68%) and 54% (27%-71%), respectively. Among cases with moder-
ately severe disease that required oxygen therapy, VE following any number of vaccine doses was 95%
(44%-100%).
Conclusion: Covishield vaccine protected significantly against Covid-19, with a higher protection rate
against severe forms of disease.
Ó 2022 Elsevier Ltd. All rights reserved.

1. Introduction [CI]: 57%-74%) for preventing symptomatic and laboratory-

proven Covid-19 and of nearly 100% (72–100%) for preventing
Vaccination is an important measure for preventing Covid-19. hospitalization and severe infection, beginning 21 days after the
India started vaccination against Covid-19 on 16th January 2021 second dose [4].
in a phased manner, prioritising first the health care workers Since clinical trials include selected individuals, it is important
(HCWs) and other frontline workers, extending it to those over for the protection offered by any vaccine to be studied in real-
60 years of age and those aged 45–60 years with comorbidities, world settings. A commonly used method for evaluating
then to all those aged 45 years and above, and finally to all adults population-level effectiveness of covid vaccines has been to assess
[1]. Two vaccines were authorized by the Indian drug regulator for their effect in preventing infection, which is defined as detection of
emergency use, namely Covishield (a recombinant, replication- viral RNA or antigen in a respiratory specimen collected from a
deficient chimpanzee adenovirus vector that encodes SARS-CoV-2 person, after a specified period has elapsed after receipt of all rec-
spike glycoprotein) and Covaxin (inactivated whole virions grown ommended doses, using a case-control design [5].
in Vero cells). Covishield, which is identical to the Oxford- Our study was aimed at determining the effectiveness of the
AstraZeneca (ChAdOx1 nCoV-19) vaccine in composition and Covishield vaccine in preventing laboratory confirmed Covid-19,
immunogenicity [2], has accounted for nearly 88% of all doses in separately for those who had received a single dose and for those
the country to date, and has been the sole vaccine used in some who had received two doses of this vaccine.
areas, including our city [3]. In pooled data from four trials, this
vaccine had a protective efficacy of 67% (95% confidence interval
2. Methods

⇑ Corresponding author. We designed a test-negative case-control study in our institu-

E-mail address: [email protected] (S.S. Kar).
tion, a large teaching hospital, located in the Puducherry district
1
Nandhini Manoharan, Palanivel Chinnakali, Mahalakshmy Thulasingam, Sonali in Southern India, on the East coast, with nearly 8700 healthcare
Sarkar, Molly Mary Thabah workers (HCWs). Vaccination with Covishield had started in our

https://doi.org/10.1016/j.vaccine.2022.02.014
0264-410X/Ó 2022 Elsevier Ltd. All rights reserved.
S. Pramod, D. Govindan, P. Ramasubramani et al. Vaccine 40 (2022) 3294–3297

institution on 16th January 2021, though some of our HCWs could In the analysis of data from 203 case-control pairs where both
receive the other vaccine by travelling to other areas. Assuming a cases and controls had symptoms, after adjustment for various fac-
vaccination coverage of 50%, effectiveness of 70%, and a matched tors, the VE associated with one dose and two doses was 58% (28%-
case-control design, the sample size was calculated to be 346 75%) and 64% (38%-78%), respectively.
case-control pairs [12]. The study was approved by our institu- In a subgroup analysis of the 10 cases (2.8%) who had a moder-
tion’s ethics committee (JIP/IEC/2021/233) and all subjects pro- ately severe disease that required oxygen therapy, nine were found
vided an oral informed consent. to be unvaccinated. By contrast, only 3 of their 10 matched controls
A case was defined as a HCW in our institution who had tested were unvaccinated (p = 0.019, Fisher’s exact test; VE = 95% [44%–
positive for active SARS-CoV-2 infection using RT-PCR during 100%]).
March 1-May 31, 2021. Students, whether medical, nursing and
other allied health science, were not considered as HCWs. All the
4. Discussion
consecutive cases identified were contacted, and were enrolled if
they agreed. In persons who had more than one positive test result,
Our data show that vaccination was associated with a reduction
the date of first positive report was used.
in the risk of Covid-19 and, in particular, of moderately severe dis-
A control was a HCW aged within 3 years of the particular case,
ease needing hospital care, among HCWs in our institution.
and who had tested negative for SARS-CoV-2 by RT-PCR within
Another test-negative case-control study from Vellore, India
3 days of the particular case testing positive. Persons with negative
showed VE among HCWs who had received two doses of a
SARS-CoV-2 antigen test alone were not included.
Covid-19 vaccine to be 65% (95% CI: 61–68) [6], which was some-
Data on vaccination status, type of vaccine, test positivity, pres-
what higher than that in our study. This difference in VE can be
ence of symptoms (fever, cough, myalgia, sore throat etc.) and
related to many factors, such as differences in the prevalent virus
comorbidities were collected using a telephonically administered
strains, overall disease endemicity and vaccination coverage.
questionnaire, captured using EpiCollect5 application and analysed
Though Vellore is situated fairly close to our city, the study there
using STATA V14.0. To calculate vaccine effectiveness (VE), the pro-
included cases that occurred between mid-January 2021 and April
tective effect was taken as appearing 21 days after vaccination if
2021, whereas our study included those from March to May 2021.
only one dose had been administered and 14 days after the second
It is well known that the number of cases with the delta variant
dose if two doses had been administered. As the healthcare work-
(B.1.617.2), a variant of concern, of SARS-CoV-2 surged in India
ers in our institution were posted in the Covid wards on a rota-
during March to May 2021. Thus, during the period of our study,
tional basis, the risk of infection was comparable across the
over 70% of cases in Puducherry were caused by this variant [7].
groups. Matched-pair analysis was done, and univariate and multi-
Further, a higher overall disease rate in our area and a lower vac-
variable conditional logistic regression was done to calculate unad-
cine coverage rate among our HCWs than those in the Vellore
justed and adjusted odds ratios. Factors used for adjustment
study could also explain the observed difference. The fact that
included gender, occupational role, presence of any comorbidity
some (nearly 7%) of subjects in the Vellore study and none in our
and presence of symptoms at the time of RT-PCR testing. Percent
study had received Covaxin, the other vaccine available in India,
VE was calculated as 100*(1-odds ratio). A subgroup analysis was
is unlikely to have made a difference.
done for cases with moderately severe disease and their matched
There has been only one other published report from India on
controls, to look specifically at the VE against such disease.
the protection afforded by Covid-19 vaccines [8]. It showed that
In addition, a separate analysis was done for the pairs where
the proportion of those who had received one or two doses of a
both the case and the respective control were symptomatic.
Covid-19 vaccine had a lower risk of having Covid-19; however,
since no data were provided on the time interval between vaccine
3. Results doses and disease, a formal calculation of VE was not possible.
Several studies from other parts of the world have assessed the
Our database showed that around 2200 tests had been per- VE of the AstraZeneca vaccine, which Covishield is identical to. A
formed in our HCWs from March to May 2021. Of these, 795 were cohort study conducted in Chile between February 2021 through
positive. To recruit 360 cases and 360 matched controls, we con- May 2021 showed VE of 65.9% (95% CI: 65.2–66.6) among the fully
tacted 547 test-positive HCWs (65.8%) and 963 test-negative immunized [9]. In a cohort study conducted in Scotland, the vac-
HCWs (37.3%), respectively (Table 1). Their median (interquartile cine effect for this vaccine was 88% (95% CI: 75–94) between
range) age was 34 (28–43) and 33 (28–42) years, respectively. December 2020 to February 2021 [10]. In the United Kingdom,
The distribution of gender and comorbidities was comparable VE against B.1.617.2 variant was estimated to be 32.9 (95% CI:
between cases and controls. Among cases, 15% (n = 54) had one 19.3–44.3) after only the first dose and 67.0% (95% CI: 28.9–77.3)
or more comorbidities, of which the most common were hyperten- after two doses of this vaccine [11]. Real-life data for Covid-19 vac-
sion (n = 29; 8.1%) and diabetes mellitus (n = 25; 7.0%). Most of the cines based on mRNA platform have also shown similarly high VE.
cases had mild disease requiring only home isolation (n = 350; The test-negative case-control study design is efficient and
97.2%). eliminates bias stemming from differences in healthcare-seeking
All the vaccinated subjects among both cases and controls had behaviour and community-level variations in vaccine access and
received Covishield, and none had received another Covid-19 vac- disease risk [12]. However, since we found a difference in the fre-
cine. Analysis was done ensuring similar exposure assessment for quency of symptoms among our cases and controls, we undertook
cases and controls, by basing the vaccination status of controls an additional analysis to control for this factor. In this analysis
on the date of onset of disease in the respective case. Considering restricted to only those case-control pairs where both the case
the onset of protection as 21 days after the first dose or 14 days and the control had symptoms, the VE estimates for one as well
after the second dose, and after adjustment for gender, occupa- as two doses were somewhat better than those in our primary
tional role, comorbidity, symptomatology, the effectiveness of analysis. This may indicate that the vaccine may in fact have a bet-
one dose and two doses, in protecting against Covid-19 was found ter efficacy than suggested in our initial analysis above. Further-
to be 49% (95% CI: 17%-68%) and 54% (27%-71%), respectively more, as ours is a teaching institute and government hospital,
(Table 2). RT-PCR testing was mandated for certain conditions like interstate

3295
S. Pramod, D. Govindan, P. Ramasubramani et al. Vaccine 40 (2022) 3294–3297

Table 1
Characteristics of cases with Covid-19 and controls matched for age and day of onset of illness, March–May 2021, Puducherry, India.

Characteristic Cases Matched controls P value

(n = 360) (n = 360)
Age categories, number (%)

29 108 (30.0) 115 (31.9)
30 to 39 130 (36.1) 135 (37.5)
40 to 49 84 (23.3) 75 (20.8)
50 38 (10.6) 35 (9.7)
Gender, number (%)
Men 178 (49.5) 181 (50.3) 0.819*
Women 182 (50.5) 179 (49.7)
Occupational roley
Doctors 49 (13.6) 71 (19.7) <0.001^
Nursing staff 138 (38.3) 173 (48.1)
Paramedical or Support staff 136 (37.8) 84 (23.3)
Administrative staff 37 (10.3) 32 (8.9)
Vaccination status at enrollment
Unvaccinated 148 (41.1) 57 (15.8)
Received one dose 114 (31.7) 140 (38.9)
Received two doses 98 (27.2) 163 (45.3)
Comorbidity§
Yes 54 (15.0) 42 (11.7) 0.188*
No 306 (85.0) 318 (88.3)
Covid symptoms at the time of RT-PCR test
Yes 343 (95.3) 212 (58.9) <0.001*
No 17 (4.7) 148 (41.1)
Time of RT-PCR test
March 2021 19 (5.3) 19 (5.3)
April 2021 95 (26.4) 93 (25.8)
May 2021 246 (68.3) 248 (68.9)
*
McNemar test,
^
Chi squared test.
§
Chronic lung disease, malignancy, heart disease, chronic liver disease, chronic renal disease, diabetes, hypertension or immunocompromised state.
y
Paramedical or support staff include laboratory technicians, operation theatre technicians, radiographers, ward attendants, drivers, security staff, sanitation worker.

Table 2
Comparison of vaccination status between cases and matched controls, and adjusted vaccine effectiveness against Covid-19.

Type of analysis and vaccination status Cases Matched Controls Crude OR Adjusted OR@ Vaccine effectiveness§,
(95% CI) (95% CI) % (95% CI)
For all cases and controls (360 pairs)
Unvaccinated 202 (56.1) 145 (40.3) Ref Ref
Received one dose 81 (22.5) 100 (27.8) 0.55 (0.37–0.81) 0.51 (0.32–0.83) 49 (17–68)
Received two doses 77 (21.4) 115 (31.9) 0.45 (0.31–0.67) 0.46 (0.29–0.73) 54 (27–71)
For symptomatic cases and controls (203 pairs)
Unvaccinated 121 (59.6) 76 (37.4) Ref Ref
Received one dose 41 (20.2) 58 (28.6) 0.43 (0.25–0.72) 0.42 (0.25–0.72) 58 (28–75)
Received two doses 41 (20.2) 69 (33.9) 0.36 (0.22–0.60) 0.36 (0.22–0.62) 64 (38–78)
For cases with moderately severe disease and matched controls (10 pairs)
Unvaccinated 9 (90) 3 (30) Ref Ref
Received any dose > 3 weeks ago **
1 (10) ***
7 (70) 0.05 (0.00–0.56) -y 95 (44–100)
@
Adjusted for gender, occupational role, comorbidity, symptomatology in conditional logistic regression analysis; VE§ = (1-adjusted OR)x100%.
§
Considering onset of protection as 21 days after the first dose, or 14 days after the second dose.
**
Had received 2 doses, with 2nd dose >14 days ago.
***
Of the 7, 3 had received both doses >14 days ago, and the remaining four had received one dose >21 days ago.
y
Adjusted odds ratio was not possible, in view of small numbers.

travel and new employees reporting for duty. Therefore, the tenets with disease even when only a few cases are available, and thus
of a test-negative case-control study are not fully met and this may allowed us to detect this effect. Though our analysis did show a
be revisited in future studies. statistically significant protection against moderately severe dis-
Importantly, our study showed that the VE of Covishield against ease after Covishield, the confidence intervals of the estimate are
moderately-severe disease was much higher than that against dis- relatively broad and further data on this association may be
ease of any severity. This is an important finding since the primary needed to improve our confidence in this observation.
aim of Covid-19 vaccination is to prevent serious disease needing Our study has two key limitations. First, the study design used
hospitalization so that healthcare facilities are not overwhelmed relies heavily on reporting for RT-PCR testing. Thus, it may overes-
and lives are not lost. It has been difficult to reliably assess the effi- timate the benefit of vaccination if the vaccinated HCW, whether
cacy of Covid-19 vaccines in clinical trials against moderate to sev- asymptomatic or having symptoms suggestive of Covid-19, were
ere disease because of relative infrequency of the latter. The case- to believe that they were unlikely to have Covid-19 and decide
control design that we used, despite its several limitations, has the not to report for testing. Second, since genomic sequencing of
advantage of permitting assessment of association of intervention SARS-CoV-2 had not been done in our study participants, we were

3296
S. Pramod, D. Govindan, P. Ramasubramani et al. Vaccine 40 (2022) 3294–3297

unable to assess the VE separately for the ancestral strain and vari- [2] World Health Organization. 2021. Interim recommendations for use of the
ChAdOx1-S [recombinant] vaccine against COVID-19 (AstraZeneca COVID-19
ant strains of the virus.
vaccine AZD1222, SII Covishield, SK Bioscience). Available from: https://apps.
In conclusion, our data show that Covishield vaccination, either who.int/iris/rest/bitstreams/1343289/retrieve. Last accessed 16 July 2021.
as one dose or in a 2-dose schedule, was effective in halving the [3] Ministry of Health and Family Welfare, Government of India. CoWIN
frequency of Covid-19 disease among HCWs in a period when Dashboard [Internet]. [cited 2021 July 16]. Available from:
https://dashboard.cowin.gov.in/
B.1.617.2 strain of SARS-CoV-2 was the dominant strain circulating [4] Voysey M, Costa Clemens SA, Madhi SA, Weckx LY, Folegatti PM, Aley PK, et al.
in our area, and had an even greater effect on preventing a more Single-dose administration and the influence of the timing of the booster dose
severe and clinically relevant form of this disease. on immunogenicity and efficacy of ChAdOx1 nCoV-19 (AZD1222) vaccine: a
pooled analysis of four randomised trials. Lancet 2021;397(10277):881–91.
[5] Hacisuleyman E, Hale C, Saito Y, Blachere NE, Bergh M, Conlon EG, et al.
Funding Vaccine breakthrough infections with SARS-CoV-2 variants. N Engl J Med
2021;384(23):2212–8. https://doi.org/10.1056/NEJMoa2105000.
[6] Victor PJ, Mathews KP, Paul H, Mammen JJ, Murugesan M. Protective Effect of
This research did not receive any specific grant from funding COVID-19 Vaccine Among Health Care Workers During the Second Wave of the
agencies in the public, commercial, or not-for-profit sectors. Pandemic in India. Mayo Clin Proc 2021;96(9):2493–4.
[7] Priya A. Indian SARS-CoV-2 Genomics Consortium (INSACOG) [Internet]. [cited
2021 May 31]. Available from: https://dbtindia.gov.in/sites/default/files/Final%
Declaration of Competing Interest 20MoHFW_Mutations%20in%20SARS-CoV2_NCDC_Final.pdf.
[8] Rana K, Mohindra R, Pinnaka L. Vaccine Breakthrough Infections with SARS-
The authors declare that they have no known competing finan- CoV-2 Variants. N Engl J Med. 2021;385(2):e7.. https://doi.org/10.1056/
NEJMc2107808. Epub 2021 Jun 2.
cial interests or personal relationships that could have appeared [9] Jara A, Undurraga EA, González C, Paredes F, Fontecilla T, Jara G, et al.
to influence the work reported in this paper. Effectiveness of an inactivated SARS-CoV-2 vaccine in Chile. N Engl J Med.
2021;385(10):875–84. https://doi.org/10.1056/NEJMoa2107715.
[10] Vasileiou E, Simpson CR, Shi T, Kerr S, Agrawal U, Akbari A, et al. Interim
Acknowledgement findings from first-dose mass COVID-19 vaccination roll-out and COVID-19
hospital admissions in Scotland: a national prospective cohort study. Lancet
We thank the members of our survey team, data management 2021;397(10285):1646–57.
[11] Lopez Bernal J, Andrews N, Gower C, Gallagher E, Simmons R, Thelwall S, et al.
group and the study participants for their help. Effectiveness of Covid-19 Vaccines against the B.1.617.2 (Delta). Variant. N
Engl J Med. 2021;385(7):585–94.
References [12] World Health Organization. 2021. Evaluation of COVID-19 vaccine
effectiveness: interim guidance, 17 March 2021. Available from: https://
www.who.int/publications/i/item/WHO-2019-nCoV-vaccine_effectiveness-
[1] Ministry of Health and Family Welfare, Government of India. COVID-19
measurement-2021.1
vaccines operational guidelines. Updated December 28, 2020. [Internet]. [cited
2021 Jul 14]. Available from: https://main.mohfw.gov.in/sites/default/files/
COVID19VaccineOG111Chapter16.pdf

3297