key: cord-0883028-xp0jc7lr
authors: Villarreal, A.; Rangel, G.; Zhang, X.; Wong, D.; De La Guardia, C.; Britton, G.; Llanes, P.; Restrepo, C. M.; Perez, A.; Oviedo, D.; Carreira, M. B.; Skildsen, G.; Sambrano, D.; Zaldivar, Y.; Franco, D.; Lopez Verges, S.; Zhang, D.; Fan, F.; Wang, B.; Saez Llorens, X.; DeAntonio, R.; Torres-Atencio, I.; Ortega-Barria, E.; Kosagisharaf, R.; Lleonart, R.; Chong, L.; Goodridge, A.; GROUP, COVID-19 SEROLOGY COLLABORATOR
title: Performance of a point of care test for detecting IgM and IgG antibodies against SARS-CoV-2 and seroprevalence in blood donors and health care workers in Panama
date: 2020-09-25
journal: nan
DOI: 10.1101/2020.09.25.20201459
sha: 0e83647f60986b6b6b6c4be7d1898dfe08ed308a
doc_id: 883028
cord_uid: xp0jc7lr

Novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the etiologic agent of the ongoing coronavirus disease 2019 (COVID-19) pandemic, which has reached 28 million cases worldwide in eight months. The serological detection of antibodies against the virus will play a pivotal role in complementing molecular tests to improve diagnostic accuracy, contact tracing, vaccine efficacy testing and seroprevalence surveillance. Here, we aimed first to evaluate a lateral flow assay ability to identify specific IgM and IgG antibodies against SARS-CoV-2 and second, to report the seroprevalence of these antibodies among health care workers and healthy volunteer blood donors in Panama. We recruited study participants between April 30th and July 7th, 2020. For the test validation and performance evaluation, we analyzed serum samples from participants with clinical symptoms and confirmed positive RT-PCR for SARS-CoV-2, participants with other confirmed infectious diseases, and a set of pre-pandemic serum samples. We used two by two table analysis to determine the test sensitivity and specificity as well as the kappa agreement value with a 95% confidence interval. Then, we used the lateral flow assay to determine seroprevalence among serum samples from COVID-19 patients, potentially exposed health care workers, and healthy volunteer donors. Our results show this assay reached a positive percent agreement of 97.2% (95% CI 84.2-100.0%) for detecting both IgM and IgG. The assay showed a kappa of 0.898 (95%CI 0.811- 0.985) and 0.918 (95% CI 0.839-0.997) for IgM and IgG, respectively. The evaluation of serum samples from hospitalized COVID-19 patients indicates a correlation between test sensitivity and the number of days since symptom onset; the highest positive percent agreement (87% (95% CI 67.0-96.3%)) was observed at [≥]15 days post-symptom onset. We found an overall antibody seroprevalence of 11.6% (95% CI 8.5-15.8%) among both health care workers and healthy blood donors. Our findings suggest this lateral flow assay could contribute significantly to implementing seroprevalence testing in locations with active community transmission of SARS-CoV-2.

In Panama, the CAST was evaluated independently by the Gorgas Memorial Institute of Health Studies (GMI), the National Reference Public Health laboratory responsible for COVID-19 diagnostic test validation, as well as for national molecular SARS-CoV-2 diagnoses. The real-time reverse-transcriptionpolymerase chain-reaction (RT-qPCR) assay for detecting SARS-CoV-2 was used as a non-reference standard [17] . All RT-qPCR assays were performed at GMI by trained technicians following best clinical laboratory practices and quality control assurance programs. The diagnostic accuracy of the CAST was evaluated as indicated below in the statistical analysis section.

Field evaluation of CAST: Study participants and sample distribution: This study was conducted between April 30 th and July 7 th , 2020 in four private and public hospitals and two donation centers located in Panama and Colon cities ( Figure 1 ). The sample size was calculated using an estimated sensitivity of at least 80% and a specificity of at least 90% for the CAST. Based on the target population of the study, which included positive cases and contacts, we assumed a prevalence of at least 15%. Thus, the sample size was estimated at a minimum of 650 participants, aiming for a 95% level of accuracy. The inclusion criteria were being an adult over 18 years old and providing written informed consent. All study participants completed a clinical screening survey for COVID-19-related symptoms and consented to submit samples for screening of other infections. Only healthy blood donors (HD) that tested negative for other infectious diseases, including Chagas disease, HIV, HBV, HAV, and HTLV1, were invited to participate in our study. The exclusion criteria comprised those with missing data and patients in intensive and semi-intensive care units. Health care workers (HCW) were asked to provide an additional blood sample 15 days after the first sample was taken. All HCW in contact with confirmed COVID-19 cases were considered high risk. This study was registered with the Panama Ministry of Health (No. 1462) and was approved by the National Research Bioethics Committee (CNBI; No. EC-CNBI-2020-03-43).

Specimen collection, demographic and clinical data: Predesigned questionnaires related to COVID-19 from the World Health Organization (WHO) were completed by trained interviewers. This questionnaire was given to all study participants. Information on sociodemographic factors, medical history, current COVID-19 symptoms, and epidemiological data were collected through personal interviews. Questions related to anosmia and ageusia symptoms were not included in the interview.

Venous blood samples were collected from all the study participants for CAST analysis. The blood collection tubes were kept at room temperature to allow clot formation and then centrifuged for 10 min at 250 x g to obtain serum specimens. All rapid test analyses were conducted with fresh serum samples.

We followed a step-wise protocol for conducting the CAST.

Briefly, we added one drop of serum (approximately 20-25 μl) into the cassette sample well followed by . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 25, 2020. . https://doi.org/10.1101/2020.09.25.20201459 doi: medRxiv preprint two drops of the provided buffer (approximately 70 μl). If IgM and/or IgG anti-SARS-CoV-2 antibodies are present in the sample, they will bind to the colloidal gold conjugate, forming an immunocomplex. This immunocomplex is then captured by the respective pre-coated band containing either anti-IgM or anti-IgG antibodies, forming a red colored IgM and/or IgG line. The presence of one red lines indicates the sample is positive for specific IgM or IgG anti SARS-VoV-2 antibodies, while the presence of two lines indicates the sample is positive for both IgM and IgG antibodies. A third line functions as a positive control, indicating that the kit is working properly. All analyses were interpreted by two independent technicians at 15 minutes after the serum was added. If there were disagreements, a third trained technician evaluated the result and provided the final decision. 

This study was conducted between April 30 th and July 7 th , 2020 in four private and public hospitals located in Panama and Colon cities, as well as the blood donation center in Panama City ( Figure 1 ). A total of 702 participants were recruited for the field study: 255 (35%) were HD, while the remaining 65% of the sample comprised 351 HCW and 96 COVID-19 patients (confirmed by RT-qPCR) ( Figure 2 ). Table 1 summarizes the age, sex, COVID-19 exposure, and presence of comorbidities across participants in the COVID-19 patient, HCW, and HD groups. Among the participants from COVID-19 group, 67 (69.9%) reported a pre-existing chronic disease; whereas 90 (25.6%) HCW and 28 (11.0%) HD reported a preexisting chronic disease. All hospitalized patients reported at least one COVID-19-compatible symptom previous to or during their hospitalization (Supplementary Table) .

CAST test diagnostic performance using panel of reference sera . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. We proceeded to evaluate the CAST's performance in the field during the current COVID-19 pandemic in Panama. We recruited 96 COVID-19 ward patients ( Figure 2 ). All participants from this group were RT-PCR-confirmed positive cases and developed moderate COVID-19 symptoms. Analysis of the COVID-19-confirmed patient group showed a PPA of 67.7% (95% CI 57.8-76.2%) for IgM and IgG anti-SARS-CoV-2 antibodies (Table 2 ). In order to investigate seroconversion over the course of COVID-19 evolution in patients, the data from these 66 sera samples were divided into three groups according to the time of sample collection after illness onset. The CAST results showed a PPA of 36.4% (95% CI 19.6-57.1%) for either or both IgM and IgG in patients whose samples were collected from 0-7 days after RT-PCR diagnosis (Table 3) . PPA scores of 76.2% (95% CI 54.5-89.8%) and 71.4% (95% CI 49.8-86.4%) for IgM and IgG, respectively, were found for patients whose samples were collected from 8-14 days after positive RT-PCR results. The highest PPA score of 87.0% (95% CI 67.0-96.3%) for both IgM and IgG antibodies was found for samples collected more than 15 days after diagnosis (Table 3) .

To determine seroprevalence among a potentially exposed population and a population of healthy donors, we applied the CAST to participants with a high (HCW) and low (HD) risk of exposure to the virus. We found that forty-five out of 351 HCW tested positive for both IgM and IgG SARS CoV-2 antibodies, which corresponds to a prevalence of 11.61% (95% CI 8.6-15.4%) (data not shown). In contrast, 86.97% (95% CI 83.0-90.1%) of the HCW samples were non-reactive, while 0.28% (95 C.I 0-1.7%) and 1.42% (95% CI 0.5-3.4%) of the HCW samples were positive for only IgM or only IgG, respectively ( Figure 3 ). Next, we determined the seroprevalence among a group of HD and found that 11.72% (95% CI 8.3-16.3%) of the samples from this group were positive for both IgM and IgG anti-SARS-CoV-2 antibodies (Figure 3) .Thus, 85.94% (95% CI 81.1-89.7%) of the HD samples were non-. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 25, 2020. . https://doi.org/10.1101/2020.09.25.20201459 doi: medRxiv preprint reactive, while 0.78% (95% CI 0-3.0%) of the HD samples were positive for only IgM or only IgG antibodies. We also found that HCW with clinical responsibilities (nurses and physicians) had a similar seroprevalence compared to those without clinical responsibilities (administrators, laboratory technicians, etc.) (12.9% vs. 14.1%, respectively). In order to determine the risk of exposure during interactions with hospitalized COVID-19 patients, we asked HCW if they had been in contact with those patients. We found that those HCW that reported having close contact with confirmed COVID-19 cases demonstrated a higher seroprevalence than HCW that did not report close contact (12.4% vs. 1.8%, respectively, data not shown).

We also analyzed differences in the age and gender of seropositive participants. In both the HCW and HD groups, we found the highest seropositivity among participants in the age range of 20-39 years (48.9% and 64.7%, respectively). Among seropositive HCW, 77.8% were male and 22.2% female ( Table   1 ). The majority of HCW (75.6%) reported having contact with a confirmed COVID-19 case, while most of the HD participants reported no contact (70.6%).

.

Here we report COVID-19 antibody seroprevalence in HCW and HD in Panama. We also report the performance of a rapid test kit for detecting anti-SARS-CoV-2 IgM/IgG antibodies with an LFIA. We When we stratified the COVID-19 patient samples according to when they were collected in terms of number of days after symptom onset, we observed differences in the prevalence of positive results. A . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Regarding the results in HCW and HD participants, we found a seroprevalence of IgM and IgG antibodies of 11.61% (95% CI 8.6-15.4%) and 11.72% (95% CI 8.3-16.3%) in the HCW and HD groups, respectively. Given that HCW are a high-risk population [26, 27] , it was quite surprising to find that the seroprevalence was nearly equal between the two groups, similar to a previous report [28] . It is worth mentioning that special precautions for blood donations are under investigation. To date, previous reports suggest no direct threat to blood safety itself [29-31]. Likewise, it is important to highlight the ability of this rapid test to detect antibodies in mild or asymptomatic COVID-19 populations since there are indications that less severe illness is associated with lower antibody titers [32-35]. As none of the members of this group reported being hospitalized or having symptoms clearly indicating recent infection, it is tempting to conclude that these represent asymptomatic cases that were infected while exposed to COVID-19 patients or in the community. Our study has several limitations. First, we were not able to use samples from individuals with other respiratory tract infections to rule out cross reactivity with human coronaviruses causing common . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this this version posted September 25, 2020. . https://doi.org/10.1101/2020.09.25.20201459 doi: medRxiv preprint seasonal colds. However, a set of 55 pre-pandemic samples was used to validate the test, and all but one tested negative for anti-SARS-CoV-2 antibodies. Second, this rapid test is based on a colorimetric evaluation of the IgG and IgM bands determined by an operator, which implies the limitations that a qualitative inter-intra-operator evaluation might produce in terms of variability. In our study, this limitation was addressed by resorting to double operator evaluation and taking photographs of all test results to be re-analyzed by a third party in the case of first level evaluation disagreement. Third, the CAST is a qualitative detection method; thus, the antibody levels in COVID-19 patients were not measured in this study. Also, the sample size was calculated for the minimum sample required to validate the test. It is yet to be determined if the CAST produces the same results in a point-of-care setting using fresh blood samples since we used serum after centrifugation. Moreover, we have not evaluated if the CAST produces comparative results with ELISA or immunochemiluminescent tests. Ongoing work by our research team will allow us to establish the CAST's limits of detection by comparing results with an ELISA test.

The CAST (rapid test) has some advantages compared to other more complex laboratory-based tests. Compared to automated ELISA and immunochemiluminescent assays, CAST is economical and time efficient, does not require advanced equipment, is simple to perform, and requires minimal training.

The CAST can be used for seroprevalence studies in primary health care settings as well as in specific contexts outside of hospitals, such as high-prevalence areas. Due to its low cost and short turnaround time, CAST is suitable for large-scale sample screening. In addition, using blood samples as opposed to nasal swabs could eliminate the need for operational steps that may produce aerosols and place technicians at higher risk. Some groups have attempted to compare serological tests with RT-PCR platforms. These tests have different targets and applications. The RT-PCR is intended for acute phase diagnosis, while the serology tests are intended for antibody seroprevalence studies.

In conclusion, the findings of this cross-sectional study demonstrate the value of the CAST for the detection of specific IgM and IgG antibodies at the population level, including health care personnel, healthy blood donors and other community members. The use of a rapid test among both healthy individuals and patients to conduct surveillance in outbreak areas could provide critical information about the status of the COVID-19 pandemic. Such a rapid test would allow the characterization of the pandemic's behavior at the community level and the identification of transmission hot spots in the community, which, in turn, would help us to better understand the situation and establish optimal strategies within quickly changing epidemic scenarios [15] . In addition, it will facilitate the massification of diagnostic methods allowing us to determine the seroprevalence of the Panamanian population and the true extent of SARS-CoV-2 community spread.

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 25, 2020. . https://doi.org/10.1101/2020.09.25.20201459 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 25, 2020. . Renal failure 13 (13.5%) 6 (9.2%) 1 (0.3%) 1 (2.2%) 0 (0%) 0 (0%)

3 (3.1%) 1 (1.5%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)

3 (3.1%) 2 (3.1%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)

Cancer 5 (5.2%) 2 (3.1%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)

Diabetes 29 (30.2%) 21 (32.3%) 18 (5.1%) 2 (4.4%) 0 (0%) 0 (0%) Asthma 1 (1.0%) 1 (1.5%) 11 (3.1%) 4 (8.9%) 7 (0%) 1 (2.9%) Other 1 7 (7.3%) 6 (9.2%) 23 (6.5%) 3 (6.7%) 5 (0%) 2 (5.8%)

. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 25, 2020. . https://doi.org/10.1101/2020.09.25.20201459 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 25, 2020. . https://doi.org/10.1101/2020.09.25.20201459 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint

The copyright holder for this this version posted September 25, 2020. . https://doi.org/10.1101/2020.09.25.20201459 doi: medRxiv preprint

Outbreak of pneumonia of unknown etiology in Wuhan, China: The mystery and the miracle

COVID-19) as a Multi-Systemic Disease and its Impact in Low-and Middle-Income Countries (LMICs)

Organization PAH: COVID-19 -PAHO/WHO Response

SARS-CoV-2 antibodies up to 25 April 2020 and public health implications

Rapid roll out of SARS-CoV-2 antibody testing-a concern

Diagnostic accuracy of serological tests for covid-19: systematic review and meta-analysis

Figure 3. Percentage of positive SARS-CoV2 IgM and IgG in healthcare workers and healthy