A prospective data collection diagnostic validation study was conducted from May 5 to June 6, 2020 in the COVID-19 screening units and the COVID-19 patient management centre in Brazzaville. The COVID-19 test authorized in Congo is PCR. Brazzaville is the epicentre of the new coronavirus disease in the Congo and at this time has two analysis laboratories (National Public Health Laboratory and the private laboratory of the Congolese Foundation for Scientific Research) and several sampling sites and 3 dedicated management sites: the University Hospital of Brazzaville, the Leyono Clinic for the management of symptomatic cases and the Hotel Hospital la Concorde of Kintélé for the management of asymptomatic cases in Brazzaville. The study took place at the National Public Health Laboratory, the Leyono municipal clinic, the Kintélé Hotel-Hospital and the screening centre of the Public Health Emergency Operations Centre. During the study period, individuals with a negative or positive PRC test were asked to participate in the study. Individuals who tested negative were included at the time of their result withdrawal. Covid-19 patients were recruited from two others study sites. Only COVID-19 positive cases who had not yet started treatment were included in the study.

We adopted as the sample size formula: N = [Uα² * p(1 − p)]/i², with: N = sample size, P = prevalence; i = margin of error: 5%. The unknown prevalence was estimated to be P=0.5. Using a proportion of 10% of the study population, recommended for pilot studies [24], we obtained a minimum sample size of N = [(1.962 * 0.52)/0.052] * 10/100 = 38 persons.

Two categories of samples were collected: blood samples from patients with a negative PCR test and blood samples from individuals with a positive PCR test. The collection of blood samples from PCR-negative subjects took place at the collection unit of the Centre des Opérations des Urgence de Santé Publique in Brazzaville, when the test results were withdrawn. For confirmed positive cases, samples were taken at the COVID-19 patient management sites. Blood samples were collected and placed in EDTA tubes by qualified biologists wearing personal protective equipment, accompanied by epidemiologists to fill out the survey forms, then transferred and stored in the molecular biology department of the National Public Health Laboratory (LNSP). Only the first reading of the test result, within the reading time recommended by the manufacturer, was considered a valid result. Information on the sociodemographic and clinical characteristics of the participants and the results of the diagnostic tests used were collected using a standardized questionnaire.

Epi-info software was used for statistical analysis and Excel for the production of tables. The qualitative variables were summarised in numbers and percentages. Quantitative variables were summarised as minimum, maximum, mean and standard deviation. We compared the results obtained with the serologic tests with those obtained by PCR. Sensitivity, specificity and likelihood values were used as indicators of test accuracy. The definitions are as follows:

- Sensitivity (Se) is the proportion of COVID-19 positive “sick” subjects by serologic test, among all positive “sick” subjects by PCR.

- Specificity (Sp) is the proportion of COVID-19 negative “non-diseased” subjects by serologic test, among all PCR negative subjects.

- Positive predictive value (PPV) is the probability of being ill when the serologic test is positive.

- Negative predictive value (NPV) is the probability of not being ill when the serologic test is negative.

Table 1 shows the details of the calculation of these four parameters, using the classification of participants as sick and not sick.

A positive serological test was considered as the presence of at least 1 of the two immunoglobulins tested: IgG or IgM. We considered as “true positives” all subjects in both serological and PRC tests were positive. Patients with a negative serological test and have a positive PCR result were considered as “false negatives”.

The Laboratory and Research Commission of the National Technical Committee for the Response to Covid-19 gave its approval for the study. All participants gave informed consent prior to enrolment in the study. For reasons of confidentiality, each of the serological tapes was marked with a registration number. The identity of the study subjects and their PCR results were linked to an identifier and known only to the study investigators. The read cassettes were kept for archiving at the LNSP.

The study included five (5) tests, the main characteristics of which are described below:

- IchromaTM COVID-19 Ab (ICHROMA), Boditech Med Incorporated

This is a test that detects IgM and IgG antibodies in serum, plasma and whole blood. It consists of an automatic result reader, a cassette with a 25-test ID, a diluent, and a tube containing a pellet. The procedure is as follows: introduce 150 µl of the diluent into the tube containing the granule, mix until it is completely

dissolved, then add 10 µl of the whole blood/serum/plasma to the solution. Then initialise the cassette in the reader, place 75 µl of the mixture in the “sample” well and press the “start” icon on the reader to initiate the reading procedure. The result obtained after 15 minutes is both quantitative and qualitative in three modalities, according to the following threshold: <0.9 for “Negative” [0.9 - 1.1] for “Indeterminate” and [1.1 - 200] for “Positive”.

- One Step Test for Novel Coronavirus (2019-nCoV) IgM/IgG Antibody (GB), Getein Biotech, Inc, China

Immuno-chromatographic test for IgM and IgG antibodies in serum/plasma/ total blood. It consists of a cassette and a diluent. The procedure is to place 20 µL of whole blood or 10 µL of serum/plasma into the sample well of the cassette and immediately add 3 drops of diluent. The test is positive when a band appears near the IgG or IgM area. Qualitative result obtained between 10 and 20 minutes, and validated by the presence of a band at the “control” zone. IgG/IgM zones do not appear clearly on the cassette: need to refer to the manufacturer’s instructions.

- Wantai SARS-CoV-2 Ab Rapid Test (WANTAI), Beijing Wantai Biological Pharmacy Enterprise Co, Ltd, China

An immuno-chromatographic test that detects antibodies in serum/plasma/ total blood, without differentiation between IgM and IgG. It consists of a cassette and diluent. The procedure is as follows: place 10 µL of whole blood/serum/plasma in the “sample” well of the cassette and immediately add 2 drops of diluent. The positivity of the test is similar to that of GP, the qualitative result obtained after 15 minutes, and validated by the presence of a band at the “control” zone

- Rapid IgG/IgM test 2019-nCoV (MULTI-G), Multi-G bvba, Belgium

Immuno-chromatographic test for IgM and IgG antibodies in serum/plasma/ total blood. It consists of a cassette and a diluent. The procedure is as follows: place 20 µL of whole blood or 10 µL of serum/plasma in the “sample” well of the cassette and immediately add 2 drops of diluent. The positivity of the test is similar to that of GP, the qualitative result obtained after 10 minutes, and validated by the presence of a band at the “control” area. Manufacturer:

- SARS-CoV-2 Antibody Test Strip (Colloidal Gold Method) (SINOCARE), Changsha Sinocare Inc.

This is an immuno-chromatographic test that detects IgM and IgG antibodies in serum/plasma/total blood. It consists of a cassette with a diluent and a micropipette. The procedure is as follows: place a drop of serum or plasma in the “sample” well of the cassette and immediately add 2 to 3 drops of diluent. The positivity of the test is similar to that of GP, the qualitative result obtained after 10 minutes, and validated by the presence of a band in the “control” area.

A total of 130 people were surveyed, 70% were male and the mean age was 39.4 ± 13.0 years, 41 (31%) had a positive Covid-19 PCR test and 89 (69%) 41 a negative PCR test, 86% of subjects were asymptomatic. The participants in the survey came from the nine districts of Brazzaville. Details of the main characteristics of the participants are presented in Table 2.

The detailed results regarding the diagnostic performance of the five tests assessed are presented in Table 3. Below is the key information on the performances of each test:

Sensitivity (Se)

- WANTAI gives 77.8% true positives and 22.2% false negatives;

- ICHROMA gives 62.5% true positives and 37.5% false negatives;

- GB gives 55.0% true positives and 45.0% false negatives;

- MULTI-G gives 65.0% true positives and 35.0% false negatives;

- SINOCARE gives 70.0% true positives and 30.0% false positives.

Specificity (Sp)

- WANTAI gives 86.2% true negatives and 13.8% false positives;

- ICHROMA gives 67.2% true negatives and 32.8% false positives;

- GB gives 81.4% true negatives and 18.6% false positives;

- MULTI-G gives 81.2% true negatives and 18.8% false positives;

- SINOCARE gives 86.3% true negatives and 13.7% false positives.

Positive predictive value (PPV)

The probability of being ill when the test is positive is 77.8% for WANTAI, 53.2% for ICHROMA, 66.7% for MULTI-G, 62.9% for GB and 80.0% for Sinocare.

Negative predictive value (NPV)

The probability of having a negative test when not ill is 86.2% for WANTAI, 75.0% for ICHROMA, 80.0% for MULTI-G, 76.0% for GB and 78.6% for Sinocare.

Specific case of symptomatic persons

Analyses restricted to symptomatic cases with negative or positive PCR show much better results with sensitivitie and negative predictive values all above 80%. Only the WANTAI test has optimal diagnostic performance (Se = 100% and NPV = 100%). The detailed results are presented in Table 4.