ABSTRACT
Specific serological tests are mandatory for reliable SARS-CoV-2 seroprevalence studies but assay specificity may vary considerably between populations due to interference of immune responses to other pathogens. Here, we assess the false positive rates obtained with four commercially available IgG ELISAs in serum panels originating from three different African countries.
Article summary line Several commercially available SARS-CoV-2 ELISAs show limited specificity when applied to serum panels of African origin
By August 16th 2020, 945,165 laboratory-confirmed SARS-CoV-2 infections and 18,476 deaths caused by COVID-19 have been reported from the WHO Africa region (1). Currently, community transmission is observed in 33 African countries including Tanzania, Ghana, Madagascar, Nigeria, Kenya, and South Sudan; in 14 other African countries (e.g. Uganda, Burundi, Rwanda) clusters of cases or sporadic cases have been described (1). To correctly determine the actual exposure of the population to SARS-CoV-2 and to draw reliable conclusions on morbidity and case fatality rates, highly sensitive and specific serological tests are mandatory.
Up to now, a plethora of serological tests for detection of anti-SARS-Cov-2 antibodies has been developed and commercialized (2, 3). Although performance data for several of these assays have been rapidly communicated by different laboratories (4-8), to our knowledge no reports are available yet on the applicability of these tests on African serum panels. Here, assay specificity may be challenged by previous or current infections with other pathogens; in particular, hypergammaglobulinemia induced by Plasmodium infection may cause false positive results in serological tests (9, 10).
THE STUDY
To assess the specificities of commercially available SARS-CoV-2 IgG ELISA tests in serum panels of different origin, a priori SARS-CoV-2 IgG negative serum panels (Appendix Table 1) collected before 2019 in African countries (Ghana: n=79, Madagascar: n=79, Nigeria: n=40, comprising samples from both febrile patients and symptom-free donors), South America (Colombia: n=40, symptom-free donors), Asia (Lao PDR: n=20, symptom-free donors), and Europe (Germany: n=75, symptom-free donors) were analyzed with the Euroimmun Anti-SARS-CoV-2-NCP IgG ELISA (Euroimmun, Lübeck, Germany), the Euroimmun Anti-SARS-CoV-2 IgG ELISA (Euroimmun, Lübeck, Germany), the EDI™ Novel Coronavirus COVID-19 IgG ELISA (Epitope Diagnostics, San Diego, US), and the Mikrogen recomWell SARS-CoV-2 IgG ELISA (Mikrogen, Neuried, Germany) (Appendix Table 2). Assays were performed and evaluated according to the manufacturers’ instructions. To quantify variation of index values determined on different days/plates (inter-assay/inter-lot variation), two sera from German patients with a previous PCR-confirmed SARS-CoV-2 infection which were collected on day 19 post onset of symptoms and one commercially available human negative control serum (Merck Millipore) were included in each test run/plate. In addition, subsets of serum panels were assayed using the Wantai SARS-CoV-2 Ab ELISA (Wantai, Beijing, China) detecting total anti-SARS-CoV-2 Ig (Appendix Table 2). BNITM in-house SARS-CoV-2 IgG indirect immunofluorescence testing (IIFT) was performed as described previously (11).
While IgG ELISA specificities where good to excellent for pre-COVID-19 serum panels originating from Colombia, Lao PDR, Madagascar, and Germany, increased false positive rates were observed in a priori SARS-CoV-2 IgG negative sera from Ghana and Nigeria (Figure 1, Table 1). False positive rates were comparable in serum samples originating from febrile vs. asymptomatic donors from Ghana; no immediately apparent correlation of false positive ELISA results with P. falciparum parasitemia was observed (Appendix Table 3). Nevertheless, a sound statistical analysis of possible correlations with the donors’ health status would require larger subgroups/sample numbers than were available for the current study.
Index values obtained for serum/plasma samples collected before 2019 in three different African countries (Ghana (n=79), Madagascar (n=79), Nigeria (n=40)), Colombia (n=40), Lao PDR (n=20), and Germany (n=75) with (A) the Euroimmun Anti-SARS-CoV-2-NCP IgG ELISA, (B) the Euroimmun Anti-SARS-CoV-2 IgG ELISA, (C) the EDI Novel Coronavirus COVID-19 IgG ELISA, and (D) the Mikrogen recomWell SARS-CoV-2 IgG ELISA. Dotted lines represent negative and positive cut-off values, respectively. Grey shading indicates index values rated as “borderline” according to the manufacturers’ instructions. Diamonds represent index values obtained for two IgG positive COVID-19 patient sera sampled on day 19 post onset of symptoms (dark grey: SARS-CoV-2 IgG IIFT titer 1:640, light grey: SARS-CoV-2 IgG IIFT titer 1:160) and one negative control serum; error bars represent standard deviation of n=5 independent measurements.
In contrast to the four indirect IgG ELISA tests evaluated in this study, the Wantai SARS-CoV-2 Ab ELISA, detecting total antibodies employing a sandwich antigen procedure, showed a high specificity when applied to a subset of sera from Ghana, Madagascar, and Nigeria (Appendix Figure 1). Further testing with this assay was not possible due to the large sample volume needed per well (100 ! of undiluted sample) and scarcity of stocked pre-2019 sera. No SARS-CoV-2-specific fluorescence pattern could be observed in any of the African samples by BNITM in-house SARS-CoV-2 IgG IIFT; non-specific staining was observed in low dilutions of 20/79 (25.3%), 6/79 (7.6%), and 7/40 (17.5%) serum samples from Ghana, Madagascar, and Nigeria, respectively.
While the index values obtained with the Euroimmun SARS-CoV-2-NCP IgG ELISA and the EDI™ Novel Coronavirus COVID-19 IgG ELISA, both employing recombinant SARS-CoV-2 nucleoprotein as antigen, showed a clear correlation (Table 2, Figure 2A, 2C, 2E), only few serum samples were concordantly classified as positive by the nucleoprotein-based Euroimmun SARS-CoV-2-NCP IgG ELISA and the spike-based Euroimmun Anti-SARS-CoV-2 IgG ELISA (Table 2, Figure 2B, 2D, 2F). Thus, testing specificity can be increased by combination of two serological tests featuring different antigens (spike and nucleoprotein).
Index values obtained for the Ghanaian (A, B), Madagascan (C, D), and Nigerian (E, F) serum samples with the Euroimmun Anti-SARS-CoV-2-NCP IgG ELISA and the nucleoprotein-based EDI Novel Coronavirus COVID-19 IgG ELISA (A, C, E) or the spike-based Euroimmun Anti-SARS-CoV-2 IgG ELISA (B, D, F). Dotted lines represent negative and positive cut-off values, respectively. Grey shading indicates index values rated as “borderline” according to the manufacturers’ instructions.
CONCLUSIONS
In accordance with evaluation studies recently published by other authors (4-6, 8), the commercially available SARS-CoV-2 IgG ELISAs displayed a good to excellent specificity when applied to serum panels originating from European or US donors. In contrast, significantly increased false positive rates were observed in African pre-COVID-19 serum panels originating from countries with a high malaria burden (Ghana and Nigeria (12)), calling into question the eligibility of these assays for seroprevalence studies in these regions. Therefore, the following recommendations should be considered when planning and performing SARS-CoV-2 seroprevalence studies (not only) in Africa: 1) Prior to performance of seroprevalence studies, carefully assess background/false positive signals obtained with the chosen serological test(s) in the target population (using a priori SARS-CoV-2 IgG negative serum samples which were stocked before 2019). Under certain circumstances, adjustment of the assay cut-off originally proposed by the manufacturer may suffice to improve/restore test specificity, but this may compromise assay sensitivity. 2) If necessary, combine information from two independent serological tests (employing different antigens and/or test principles). 3) If possible, re-evaluate samples generating a positive ELISA result with IIFT on virus-infected cells (when evaluated by experienced personnel, the obtained staining pattern contains additional information about signal specificity) and/or neutralization testing.
Further studies will be necessary to assess sensitivity of the commercially available assays in detecting anti-SARS-CoV2 IgG antibodies induced by SARS-CoV-2 infection in African COVID-19 patients and to thoroughly identify causes for the reduced assay specificity in sera from African donors.
Data Availability
All data relevant for the study are presented in the manuscript and its appendix.
FUNDING
The study was conducted within the framework of the project “East African Community (EAC) regional network of public health reference laboratories for communicable diseases” (https://mobilelabs.eac.int/) funded by the Federal Republic of Germany through the German Development Bank (KfW). The study was supported by the German Research Foundation (DFG, GU 883/4-1 and GU 883/5-1), by the German Federal Ministry of Health through support of the WHO Collaborating Centre for Arboviruses and Hemorrhagic Fever Viruses at BNITM (agreement ZMV I1-2517WHO005), through the Global Health Protection Program (agreement ZMV I1-2517GHP-704), and through the COVID support agreement ZMVI1-2520COR001. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
AUTHOR BIOGRAPHY
Dr. Emmerich is a laboratory group leader at the Bernhard Nocht Institute for Tropical Medicine in Hamburg, Germany. Her research is focused on diagnostics of viral infectious diseases.
ACKNOWLEDGMENTS
The authors thank J. Blessmann and S. Kann for providing symptom-free donor serum samples from Lao PDR and Colombia, respectively. We thank C. Ehmen and N. Pekarek for expert technical assistance.
