Sero-prevalence of anti-SARS-CoV-2 Antibodies in Addis Ababa, Ethiopia

Background Anti-SARS-CoV-2 antibody tests are being increasingly used for sero-epidemiological purposes to provide better understanding of the extent of the infection in the community, and monitoring the progression of the COVID-19 epidemic. We conducted sero-prevalence study to estimate prior infection with with SARS-CoV-2 in Addis Ababa. Methods A cross-sectional study was done from April 23 to 28, 2020 among 301 randomly selected residents of Addis Ababa; with no known history of contact with confirmed COVID-19 person. Interviews on socio demographic and behavioural risk factor followed by serological tests were performed for SARS-CoV-2 IgM, and IgG antibodies, using COVID-19 IgG/IgM Rapid Test Cassette. The test has sensitivity of 87·9% and specificity of 100% for lgM; and a sensitivity of 97·2% and specificity of 100% for IgG. RT-PCR test was also done on combined nasopharyngeal and oropharengeal swabs as an important public health consideration. Findings The unadjusted antibody-based crude SARS-CoV-2 prevalence was 7·6% and the adjusted true SARS-CoV-2 prevalence was estimated at 8·8% (95% CI 5·5%-11·6%) for the study population. Higher sero-prevalence were observed for males (9.0%), age below 50 years (8.2%), students and unemployed (15.6%), those with primary education (12.1%), smokers (7.8%), alcohol consumers (8.6%), chatt-chewers (13.6%) and shish smokers (18.8%). Seroprevalence was not significantly associated neither with socio-demographic not behavioral characteristics. According to the findings, possibly more individuals had been infected in Addis Ababa than what was being detected and reported by RT-PCR test suggestive of community transmission. The use of serological test for epidemiological estimation of the extent of SARS-CoV-2 epidemic gives a more precise estimate of magnitude which would be used for further monitoring and surveillance of the magnitude of the SARS CoV-2 infection.


Introduction
Testing SARS-CoV-2 is made in two ways, by detecting the virus itself (RT-PCR) and by detecting the host's response to the virus (serology) 1 . The World Health Organisation (WHO) recommends RT-PCR test as gold standard for COVID-19 case identification [2][3][4] . The clinical course of SARS-CoV-2 infection ranges from asymptomatic to fatal 5,6 making the identification of "cases" very complex. The response plan to the pandemic requires epidemiological data that shows the true magnitude of the disease and the serology test can be used for such purposes 7 . The specificity of the RT-PCR test is almost 100%. However, sensitivity is reported as low as 64% [8][9][10] . Differences in the detection limits of the test kits used, low initial viral load due to timing as well as improper specimen collection [11][12][13] may account for such low sensitivity.
The lower sensitivity or "false negative" RT-PCR results and/or the "immune passport" led to the rapid development of serologic assays [14][15][16] . The serologic tests in the market have different formats (lateral flow immunoassays, ELISA, and chemiluminescent immunoassays), detect different classes of antibodies (total antibody, IgG, IgM, and IgA), target different antigens (recombinant nucleocapsid protein [NP], subunit 1 of the spike glycoprotein [S1], and the Spike glycoprotein receptor-binding domain [RBD]) and accept different types of the specimen (whole blood, serum, and plasma) 16 . Data generated from population-based serology can be used for various purposes including estimation of community transmission rates and assess the impact of nonpharmacological interventions 14 . Despite the potential of the serology test, equally, the availability of a serology test with excellent sensitivity and specificity is a discussion point.
Since the report of the first confirmed SARS-CoV-2 inflection in Ethiopia on March 13, 2020, the rate of increment has been largely slow at the beginning. This required a more precise estimation of the magnitude of SARS CoV-2 infection on top of the estimates based on routine RT-PCR testing. To this end, we set a goal to conduct a series of sero-epidemiological studies. Here were reporting our first findings on anti-SARS-CoV-2 antibody among permanent residents in Addis Ababa conducted to estimate what extent SARS-CoV-2 infection has been circulating in the population.

Study design and participants
A cross-sectional community-based study was conducted from April 23-28, 2020. The study was undertaken in Addis Ababa, the capital city of Ethiopia, with an estimated total population of 4,793,699 17 .
The study population were adult (≥ 18 years), residents of Addis Ababa, with no history of known contact with confirmed COVID-19 cases and no recent history of travel out of Ethiopia in the past four months prior to the period of data collection. Individuals who were unable to consent, those with unstable mental state and suspected to have acute SARS-CoV-2 infection were excluded.
Sample size was calculated using a samples size formula for single population proportion, assuming the proportion of people with immunity among community members. with no close contact with SARS-CoV-2 Infected individuals to be14% based on earlier European studies about the time of the study 18 , 80% power and 99% confidence level The Epi Info/Open Epi program is used for the sample size calculation 19 . Accordingly, the sample size calculated was 319.
Study participants were recruited from the community level randomly through sub-city health extension focal personnel. All randomly selected individuals who fulfilled study inclusion criteria were invited to the study.
They were approached and received an invitation through health extension workers (HEWs) from the respective sub-city. All study participants were briefed about the study and gave written consent before proceeding to interview and sample collection.

Data collection
Socio-demographic and clinical data were collected through interviewer administered questionnaire. Whole blood sample for antibody testing and combined nasopharyngeal and oropharengeal swab for RT-PCR testing was collected from each participant. A questionnaire comprising socio-demographic characteristics such as age, sex, marital status, and place of residence; a history of illness in the past three to five months like signs and symptoms of pneumonia and/ or influenza-like illness; and perceived severity of illness, etc was administered.
Prior to data collection, the questionnaire was pre-tested and necessary adjustments made. Peripheral blood (4ml) was collected from each participant by experienced phlebotomist by venipuncture using Ethylenediaminetetraacetic acid (EDTA) tubes and transported to the Ethiopian Public Health Institute (EPHI) Influenza laboratory for analysis. Plasma was extracted following centrifugation at 3000rpm for 10 minutes in level-II bio-safety cabinet. Plasma was stored using cryovials in refrigerator at 2-8  C until the next day. The leftover plasma was stored at -80  C for future use at EPHI bio-bank. A combined nasopharyngeal and oropharengeal swab was collected from each participant using a Dacron or polyester flocked swabs (KANGJIAN Medical Apparatus Co., Ltd.) by trained personnel, and in Viral Transport Medium (VTM) (Longsee, LAKEbio) transported to EPHI for RT-PCR analysis of SARS-CoV-2 in cold chain (2-8  C) and processed immediately or stored for 1 to 3 days at -80  C.

Serology
Analysis for the presence of plasma IgM and/or IgG antibodies was done using a commercially available immunolateral flow immunoassay kit (Zhejiang Orient Gene Biotech Co Ltd, Huzhou, Zhejiang, China) following manufacturers instruction. The interpretation of the test was made by two experts (microbiologists and laboratory technologist). According to the manufacturer the test has a sensitivity and specificity of 87.9 and 100 % respectively. Another validation study has reported the test has lower sensitivity (69% for IgM and 93.1% for IgG and the specificity was found to be 100% for IgM and 99.2% for IgG using RT-PCR assay as a comparator 20 .

RNA extraction and RT-PCR analysis of COVID-19
The combined swab (Nasopharyngeal and oropharyngeal) was transferred into lysis buffer that contains a guanidinium-based inactivating agent and viral RNA was extracted using Nucleic Acid Isolation Kit (Da'an Gene Corporation) following manufacturer's instruction. Briefly a 200 μL of combined swab in VTM was used for viral RNA extraction and viral RNA was eluted with 60 μL elution buffer. Real-time reverse transcriptional polymerase chain reaction (RT-PCR) reagent of Da'an Gene cooperation was employed for SARS-CoV-2 detection following manufacturer's protocol. Briefly, two PCR primer and probe sets, which target the open reading frame 1a/b (ORF 1a/b) (FAM reporter) and nucleocapsid protein (N) genes and N (VIC reporter) genes were added in the same reaction mixture. In each run, positive and negative controls were included. Samples were considered to be positive when both sets gave reliable signal (<40CT value).

Data processing and Analysis
All demographic, epidemiologic and laboratory data were cleaned and entered into EPI DATA using a controlled and programmed data entry format. The data was coded and anonymized data were merged to laboratory data which were exported into SPSS for windows. Descriptive statistics was used to summarise key findings. Estimation of true values for sero-prevalence for the given levels of sensitivity and specificity were computed via web-based statistical software and epidemiologic calculator including EPI Tools (https://epitools.ausvet.com.au/trueprevalence) 21 . Binary logistic regression was used to identify factors associated with the outcome variable. Variables whose p value less than 0.05 at bivariate analysis were included in multivariable analysis.

Ethical Considerations
The research proposal is approved by the Institutional Review Boards (IRB) of both the College of Health Sciences of the Addis Ababa University and EPHI. Written informed consent was obtained from all study participants. All study participants were informed of their RT-PCR test results as per the national testing protocol.

Funding
While we had no any official sponsor, costs of the research were met by resources from Addis Ababa University and the Ethiopian Public Health Institute (EPHI).

Discussion
As far as our literature review during the writing of this paper goes, this was the first SARS-CoV-2 sero-

Sero-prevalence:
The study identified an overall SARS-CoV-2 sero-prevalence of 7·6% which is found to be higher than the estimates based on RT-PCT testing in Addis Ababa, during the period of the study.  20 . This reflects the fact that the actual estimates of seropositivity in the study subjects could even be higher, when adjusted to the lower values of sensitivity.
In the presence of such a big burden of cases in a big city like Addis Ababa, why did not we see an increase or surge in the number of severe cases in the health facilities? The same factors i.e. predominantly asymptomatic cases could explain this owing to the predominantly young population and lower levels of non-communicable diseases in the Ethiopian population. However, our study did not get any supporting evidence to the looming theory of the protective association between BCG Vaccination and SARS-CoV-2 infection.
The implications of the study: As the first antibody testing among the general public in Addis Ababa, the study signified the need for extensive serological evidence to make comprehensive estimates of the true extent of SARS-CoV-2 infection in the population especially in a situation where we have the limited testing capacity for RT-PCR. Comparative advantages of antibody testing include cost-effectiveness, easy technology, and point of care testing. This makes serological estimates very much appealing to the low-resource settings. However, in the given circumstances, serological tests are taken as complementary rather than independent in generating epidemiologic evidence for SARS-CoV-2 infection. This is indicative of the need for more precise populationbased sero prevalence and surveillance to have a better sense of the extent of the epidemic and to monitor its progress over time, which further guides the design and implementation of targeted interventions. As this is the first of series of studies we planned, we plan to conduct nationwide sero-prevalence study and possibly set-up sentinel sero-prevalence longitudinal studies.
Policy implications: Such serological estimates provide a substantial and comprehensive input to the epidemiologic projections of SARS-CoV-2 especially in a setting where extensive and reliable viral antigen tests are a limitation. In addition, the current and subsequent data will provide evidence for policy makers to make decisions on locally generated evidence.
While this possibly the first SARS-CoV-2 sero-prevalence study in Ethiopia (possibly in Africa, there are a number of limitations. The test kit utilized was not validated in our population and the target antigen of the kit is not described by the company. In addition the relatively smaller number subjects included in the study may affect generalisability. The test kit has been validated in Sweden and used for population based studies in Spain and China. In-country validation of the study is currently being done by EPHI but not completed. The serological tests were not primarily done to detect a concurrent active infection with no contribution to SARS-CoV-2 case detection and management purposes.  It is recommended that a more extensive sero-prevalence study be conducted with a more representative population sample to further establish more substantial evidence on the extent of SARS-CoV-2infection in Addis Ababa and nationally.  Establish a system for longitudinal sero-prevalence studies in Addis Ababa and other parts of the country to have a continued precise estimate of the extent of SARS-CoV-2infection

Conclusion
 The national estimate of cases should be complemented with appropriate serology tests in the community to diagnose and identify levels of community transmission and guide containment strategies among different population groups while maintaining non-pharmacological interventions