key: cord-0801325-041lq5n1
authors: Gillot, Constant; Douxfils, Jonathan; Cadrobbi, Julie; Laffineur, Kim; Dogné, Jean-Michel; Elsen, Marc; Eucher, Christine; Melchionda, Sabrina; Modaffarri, Élise; Tré-Hardy, Marie; Favresse, Julien
title: An Original ELISA-Based Multiplex Method for the Simultaneous Detection of 5 SARS-CoV-2 IgG Antibodies Directed against Different Antigens
date: 2020-11-21
journal: J Clin Med
DOI: 10.3390/jcm9113752
sha: ffb94ce53e93ef8b949314afe59a3ea6621e2e06
doc_id: 801325
cord_uid: 041lq5n1

Strategies to detect SARS-CoV-2 are increasingly being developed. Among them, serological methods have been developed. Nevertheless, although these may present an interesting clinical performance, they are often directed against only one antigen. This study aims at evaluating the clinical performance of an innovative multiplex immunoassay (i.e., CoViDiag assay) detecting simultaneously the presence of antibodies directed against N, S1, S2, RBD and NTD antigens. Sensitivity was evaluated in 135 samples obtained from 94 rRT-PCR confirmed coronavirus disease 2019 (COVID-19) patients. Non-SARS-CoV-2 sera (n = 132) collected before the COVID-19 pandemic with potential cross-reactions to the SARS-CoV-2 immunoassay were included in the specificity analysis. The antibody signature was also studied in hospitalized and non-hospitalized patients. The specificity of the CoViDiag assay was excellent for all antibodies (99.2 to 100%) using adapted cut-offs. None of the false positive samples were positive for more than one antibody. The sensitivity obtained from samples collected 14 days since symptom onset varied from 92.0 to 100.0% depending on the antibody considered. Among samples collected more than 14 days after symptom onset, 12.8, 66.3, 3.5, 9.3, 5.8 and 2.3% were positive for 5, 4, 3, 2, 1 or 0 antibodies, respectively. A trend toward higher antibody titers was observed in hospitalized patient in the early days since symptom onset. However, no significant difference was observed compared to non-hospitalized patients after 14 days since symptom onset. The clinical performance of the CoViDiag 5 IgG assay is sufficient to recommend its use for the detection and the characterization of the antibody signature following SARS-CoV-2 infection. The combination of several antigens in the same test improves the overall specificity and sensitivity of the test. Further research is needed to investigate whether this strategy may be of interest to identify severe disease outcome in patients with SARS-CoV-2 infection.

The CoViDiag kit (Innobiochips ® , Loos, France) is a qualitative miniaturized and parallel-arranged enzyme-linked immunosorbent assay (ELISA) intended for the detection of multiple IgG antibodies against SARS-CoV-2 in human serum or plasma. The wells of the microtiter plates are printed with 5 different antigens of the SARS-CoV-2 to bind the corresponding specific antibodies present in the patient sample. The antigens included the N protein (nucleocapsid protein, internal to the virus), the S1 protein (Subunit 1 of the Spike protein), the RBD domain of the S1 protein (receptor binding domain of the cell), the NTD domain of the S1 protein (N-terminal S1 subunit), the S2 protein (C-terminal S2 subunit) (Supplemental Figure S1 ) [30] . After washing the wells to remove excess sample, an enzyme-labelled conjugate is added. This conjugate binds to the captured antibodies. In a second washing step, the unbound conjugate is removed. The complex formed by the bound conjugate is visualized by the addition of substrate: tetramethylbenzidine which gives locally an insoluble, blue-colored product. The intensity of this product is proportional to the quantity of specific antibodies in the sample. The cut-offs for positivity are proposed by the manufacturer (Table 1 ) and an algorithm is provided for the definition positive, negative, or borderline samples: -Sample considered positive: S1 and/or RBD and/or NTD is positive or S2 and/or N is >40 AU or S2 and/or N is positive and S1 and/or RBD and/or NTD and/or S2 and/or N is borderline or S2 and/or N is borderline and S1 and/or RBD and/or NTD and/or S2 and/or N is borderline. -Sample considered doubtful (or equivocal): S2 and/or N is positive or S1 and/or RBD and/or NTD is borderline. -Sample considered negative: all other cases not listed above. Additionally, all samples were also tested on 3 other platforms for the detection of antibodies against SARS-CoV-2 antigens: Roche (total antibodies directed against the N antigen), Euroimmun (IgG antibodies directed against the N antigen) and DiaSorin (IgG antibodies directed against the S1/S2 antigen).

One-hundred thirty-two non-SARS-CoV-2 sera were analyzed to determine the cross-reactivity and establish the specificity. Among these sera some of them were found positive for multiple antigens selected for the specificity. Potential cross-reactive samples included positive antibodies against HKU1 coronavirus (n = 78), positive antibodies against OC43 coronavirus (n = 118), positive antibodies against 229E coronavirus (n = 32) and positive antibodies against NL63 coronavirus (n = 19). The positivity against the four common coronaviruses was assessed using the BlueDiver COVIDOT 11 IgG assay (RUO assay; D-Tek, Mons, Belgium). Potential cross-reactive samples also included positive antinuclear antibodies (n = 5), anti-treponema pallidum antibodies (n = 3), anti-thyroid peroxidase antibodies (n = 3), antibodies RAI+ (search for irregular agglutinins) (n = 5), chikungunya antibody (n = 1), direct coombs (n = 1), hepatitis B antigen (n = 7), hepatitis C antibodies (n = 7), hepatitis E antibodies (n = 4), human immunodeficiency virus antibodies (n = 2), IgA Chlamydia pneumoniae (n = 1), IgM Borrelia + IgA Helicobacter pylori (n = 1), IgM Chlamydia pneumoniae (n = 1), IgG Chlamydia trachomatis (n = 1), IgG Coxiella burnetii (n = 2), IgM Coxiella burneti (n = 1), IgM cytomegalovirus (n = 13), IgM Epstein-Barr virus viral capsid (n = 5), IgM Mycoplasma pneumoniae (n = 6), IgM parvovirus B19 (n = 8), IgM Toxoplasma gondii (n = 11), influenza A antibodies (n = 4), influenza A and B (n = 1), high level of total IgG (17.40 g/L) (normal range: 7.00-16.00 g/L) (n = 1), both high levels of total IgM (5.26 g/L; normal range: 0.4-2.3 g/L) and total IgG (28.67 g/L) (n = 1), rheumatoid factor (n = 6), urinary infection with Escherichia coli (n = 1), urinary infection with Klebsiella oxytoca (n = 1). All these samples were collected before the COVID-19 pandemic and were stored at −20 • C.

The 135 specimens used for the sensitivity were divided into five different groups based on days since symptom onset: 0-6 days: 23 sera; 7-13 days; 26 sera; 14-20 days; 24 sera; 21-28 days; 25 sera; and over 28 days; 37 sera. The latest sample was taken 62 days after the onset of symptoms. The evaluation of clinical sensitivity was evaluated using the cut-offs provided by the manufacturer and were optimized by ROC curve analyses.

Descriptive statistics were used to analyze the data. Sensitivity was defined as the proportion of correctly identified COVID-19 positive patients initially positive by rRT-PCR SARS-CoV-2 determination in respiratory samples and with COVID-19 symptoms. Specificity was defined as the proportion of naïve patients classified as negative. A ROC curve analysis was performed to propose adapted cut-offs for each antigen to improve clinical performance. Samples included for ROC curves analyses were sera obtained from at least 14 days since symptom onset (n = 86) and sera from the specificity study (n = 132). A Mann-Whitney test was used to assess potential differences in critical and non-critical patients. Data analysis was performed using GraphPad Prism ® software (version 8.2.1, San Diego, CA, USA) and MedCalc ® software (version 14.8.1, Ostend, Belgium).

One-hundred and thirty-two non-SARS-CoV-2 sera were analyzed to determine the cross-reactivity and establish the specificity. When using the lower cut-off (i.e., borderline positive) provided by the manufacturer (i.e., N, S2 > 15 AU and S1, RDB, NTD > 10 AU), the specificity for the 5 different IgG antibodies varies from 92.4% (95% CI: 87.9-97.0%) to 100% [95% CI: 97.2-100%]. When considering the antigens independently, a specificity of 100% [95% CI: 97.2-100%] was obtained for IgG directed against N and S1, 97.7% (95% CI: 95.2-100%) for IgG directed against RBD and 98.5% (95% CI: 96.4-100%) for IgG directed against NTD. The lowest specificity was obtained for IgG antibodies against the S2 antigen with 92.4% (95% CI: 87.9-97.0%) (10 false positive results out of 132 samples) (Figure 1 ). Using the adapted cut-offs, the specificity for S2 (optimized cut-off: AU > 39), RBD (optimized cut-off: AU > 18), NTD (optimized cut-off: AU > 16) increased from 92.4 to 99.2% (95% CI: 97.7-100%), from 97.7 to 99.2% (95% CI: 97.7-100%) and from 98.5 to 99.2% (95% CI: 97.7-100%), respectively. The specificity for S1 (optimized cut-off: AU > 7) decreased from 100 to 99.2% (95% CI: 97.7-100%). The specificity for N (optimized cut-off: AU 12) remains unchanged (Figure 1 ). The description of false-positive samples according to the positive antigen, the AU value of the antigen concerned and the potentially cross-reactive antibody, is presented in Supplemental Table S1 . None of these false positive samples were positive for more than one antigen (Figure 1 ). 100%), respectively. The specificity for S1 (optimized cut-off: AU > 7) decreased from 100 to 99.2% (95% CI: 97.7-100%). The specificity for N (optimized cut-off: AU 12) remains unchanged (Figure 1 ). The description of false-positive samples according to the positive antigen, the AU value of the antigen concerned and the potentially cross-reactive antibody, is presented in Supplemental Table  S1 . None of these false positive samples were positive for more than one antigen (Figure 1 ).

The calculated specificities using manufacturer's cut-off for the Roche, Euroimmun and DiaSorin assays were 100% (95% IC: 97.2-100%), 96.2% (95% IC: 92.9-99.5%) and 97.7% (95% IC: 85.9-95.9%) respectively. Using ROC-curve adapted cut-offs, the calculated specificities were 100% (95% IC: 97.2-100%), 93.9% (95% IC: 89.8-98.1%) and 90.9% (95% IC: 84.0-99.7%), respectively. The manufacturer's cut-offs for Roche, Euroimmun and DiaSorin were ratio ≥1.0, ratio ≥0.8 and ≥12.0 AU/mL, respectively. The adapted cut-offs for Roche, Euroimmun and DiaSorin were ratio > 0.165, ratio >0.4 and >3.94 AU/mL, respectively.

If at least one antigen out five was considered sufficient to determine positivity, the specificity was 89.4% (95% CI: 84.1-94.7%) with the manufacturer's cut-offs and 95.2% (95% CI: 91.8-99.0%) with adapted cut-offs. When using the algorithm provided by the manufacturer, the specificity was 94.7% (95% CI: 90.8-98.5%) (7 false positives out of 132 samples) (Figure 1 ). The manufacturer's algorithm provided a similar specificity than the adapted cut-off (i.e., 94.7 vs. 95.4%). 

One-hundred and thirty-five samples of rRT-PCR-positive patients for SARS-CoV-2 were included in the sensitivity analyses. The sensitivity in the first week after symptom onset ranged from 0% (95% CI: 0-14.8%) (IgG direct against NTD) to 21.7% (95% CI: 3.5-40%) (IgG directed against N or S2) ( Figure 2 ). It increased in the second week (7-13 days) with values ranging from 11.5% (95% CI: 0-25.7%) for IgG directed against NTD to 65.4% (95% CI: 45.8-85%) for IgG directed against N and S2. In the third week (14-20 days) the sensitivity peaked at 100% (95% CI: 85.7-100%) for IgG directed against S2. Positivity for IgG directed against NTD was 8.3% (95% CI: 0-18.7%). After 28 days, sensitivity increased to 100% for IgG against S2 and also increased for all other IgG against the different antigens except for IgG against N antigen, which remained lower than in week 3. The calculated specificities using manufacturer's cut-off for the Roche, Euroimmun and DiaSorin assays were 100% (95% IC: 97.2-100%), 96.2% (95% IC: 92.9-99.5%) and 97.7% (95% IC: 85.9-95.9%) respectively. Using ROC-curve adapted cut-offs, the calculated specificities were 100% (95% IC: 97.2-100%), 93.9% (95% IC: 89.8-98.1%) and 90.9% (95% IC: 84.0-99.7%), respectively. The manufacturer's cut-offs for Roche, Euroimmun and DiaSorin were ratio ≥1.0, ratio ≥0.8 and ≥12.0 AU/mL, respectively. The adapted cut-offs for Roche, Euroimmun and DiaSorin were ratio > 0.165, ratio >0.4 and >3.94 AU/mL, respectively.

If at least one antigen out five was considered sufficient to determine positivity, the specificity was 89.4% (95% CI: 84.1-94.7%) with the manufacturer's cut-offs and 95.2% (95% CI: 91.8-99.0%) with adapted cut-offs. When using the algorithm provided by the manufacturer, the specificity was 94.7% (95% CI: 90.8-98.5%) (7 false positives out of 132 samples) ( Figure 1 ). The manufacturer's algorithm provided a similar specificity than the adapted cut-off (i.e., 94.7 vs. 95.4%).

One-hundred and thirty-five samples of rRT-PCR-positive patients for SARS-CoV-2 were included in the sensitivity analyses. The sensitivity in the first week after symptom onset ranged from 0% (95% CI: 0-14.8%) (IgG direct against NTD) to 21.7% (95% CI: 3.5-40%) (IgG directed against N or S2) (Figure 2 ). It increased in the second week (7-13 days) with values ranging from 11.5% (95% CI: 0-25.7%) for IgG directed against NTD to 65.4% (95% CI: 45.8-85%) for IgG directed against N and S2. In the third week (14-20 days) the sensitivity peaked at 100% (95% CI: 85.7-100%) for IgG directed against S2. Positivity for IgG directed against NTD was 8.3% (95% CI: 0-18.7%). After 28 days, sensitivity increased to 100% for IgG against S2 and also increased for all other IgG against the different antigens except for IgG against N antigen, which remained lower than in week 3. If at least one of the five positive antibodies is considered sufficient to determine positivity towards SARS-CoV-2 (a different approach than the algorithm provided by the manufacturer), the sensitivity increased ( Figure 2 ). Maximum sensitivity ranged from 21.7% (95% CI: 3.5-40%) to 30.4% (95% CI: 10.1-50.8%) during the first week, from 65.4% (95% CI: 45.8-85%) to 76.9% (95% CI: 59.7-94.3%) during the second week, is maximal (i.e., 100%) during the third weeks, ranged from 88.0% (95% CI: 74.3-100%) to 92.0% (95% CI: 80.6-100%) during the fourth week, and finally remained maximal after 4 weeks ( Figure 2 ). The use of the adapted cut-offs slightly reduced the global sensitivity i.e., 100% (95% CI: 90.5-100%) after 28 days using manufacturer cut-offs vs. 97.3% (95% CI: 91.8-100%) after 28 days when using the adapted cut-offs (Supplemental Table S2 ). According to the manufacturer's algorithm, the sensitivity was 21.7% (95% CI: 3.5-40%) during the first week (0-6 days), 73.1% (95% CI: 54.8-91.3%) during the second week (7-13 days), 100% (95% CI: 85.7-100%) during the third week (14-20 days), 92,0% (95% CI: 80.6-100%) in the fourth week and 100% (95% CI: 90.5-100%) over 28 days (Figure 2 A dynamic of increase in the different antibody titers (AU) and the positivity rate (%) has been observed over time since the onset of symptoms for all IgG antibodies (Figure 3 ). Among samples collected more than 14 days after symptom onset, 11 of the 86 samples were positive for the 5 antibodies (12.8%), 57 out of 86 were positive for 4 antibodies (66.3%), 3 out of 86 were positive for 3 antibodies (3.5%), 8 out of 86 were positive for 2 antibodies (9.3%), 5 out of 86 were positive for 1 antibody (5.8%) and 2 were not positive (2.3%). The highest titers of antibodies (AU) were detected for the N antigen and the lowest levels for the NTD antigen. There was a delay in the increase in IgG antibody positivity rate to RBD and S1 compared to N and S2 (Figure 3) . If at least one of the five positive antibodies is considered sufficient to determine positivity towards SARS-CoV-2 (a different approach than the algorithm provided by the manufacturer), the sensitivity increased ( Figure 2 ). Maximum sensitivity ranged from 21.7% (95% CI: 3.5-40%) to 30.4% (95% CI: 10.1-50.8%) during the first week, from 65.4% (95% CI: 45.8-85%) to 76.9% (95% CI: 59.7-94.3%) during the second week, is maximal (i.e., 100%) during the third weeks, ranged from 88.0% (95% CI: 74.3-100%) to 92.0% (95% CI: 80.6-100%) during the fourth week, and finally remained maximal after 4 weeks ( Figure 2 ). The use of the adapted cut-offs slightly reduced the global sensitivity i.e., 100% (95% CI: 90.5-100%) after 28 days using manufacturer cut-offs vs. 97.3% (95% CI: 91.8-100%) after 28 days when using the adapted cut-offs (Supplemental Table S2 ). According to the manufacturer's algorithm, the sensitivity was 21.7% (95% CI: 3.5-40%) during the first week (0-6 days), 73.1% (95% CI: 54.8-91.3%) during the second week (7-13 days), 100% (95% CI: 85.7-100%) during the third week (14-20 days), 92,0% (95% CI: 80.6-100%) in the fourth week and 100% (95% CI: 90.5-100%) over 28 days (Figure 2 A dynamic of increase in the different antibody titers (AU) and the positivity rate (%) has been observed over time since the onset of symptoms for all IgG antibodies (Figure 3 ). Among samples collected more than 14 days after symptom onset, 11 of the 86 samples were positive for the 5 antibodies (12.8%), 57 out of 86 were positive for 4 antibodies (66.3%), 3 out of 86 were positive for 3 antibodies (3.5%), 8 out of 86 were positive for 2 antibodies (9.3%), 5 out of 86 were positive for 1 antibody (5.8%) and 2 were not positive (2.3%). The highest titers of antibodies (AU) were detected for the N antigen and the lowest levels for the NTD antigen. There was a delay in the increase in IgG antibody positivity rate to RBD and S1 compared to N and S2 (Figure 3) . 

Even if we observed a trend to have higher signals for N, S1, S2 and RBD antibodies from 14 days since symptom onset in critical patients, the differences were not statistically different compared to non-critical patients in our cohort (p > 0.05) (Figure 4) . Interestingly, higher antibody levels in hospitalized vs. non-hospitalized patients were observed in samples obtained early (0-6 days and 7-13 days) since the onset of COVID-19 symptoms ( Figure 5 ). 

Even if we observed a trend to have higher signals for N, S1, S2 and RBD antibodies from 14 days since symptom onset in critical patients, the differences were not statistically different compared to non-critical patients in our cohort (p > 0.05) (Figure 4) . Interestingly, higher antibody levels in hospitalized vs. non-hospitalized patients were observed in samples obtained early (0-6 days and 7-13 days) since the onset of COVID-19 symptoms ( Figure 5 ). 

Even if we observed a trend to have higher signals for N, S1, S2 and RBD antibodies from 14 days since symptom onset in critical patients, the differences were not statistically different compared to non-critical patients in our cohort (p > 0.05) (Figure 4) . Interestingly, higher antibody levels in hospitalized vs. non-hospitalized patients were observed in samples obtained early (0-6 days and 7-13 days) since the onset of COVID-19 symptoms ( Figure 5 ). 

This study is the first to evaluate the clinical performance of the CoViDiag IgG test kit, which is designed to detect the presence of IgG antibodies to 5 SARS-CoV-2 antigens. It includes a total of 135 serum samples obtained from 94 rRT-PCR-positive patients and a total of 132 Non-SARS-CoV-2 serum samples. The sensitivity of the CoViDiag test kit after 14 days since symptom onset varies from 92.0 to 100.0%. However, it remains too low in the first 6 days after the onset of symptoms to correctly identify all positive patients, as confirmed by other studies [31, 32] . Nevertheless, detection of early seroconversion may be useful to understand the heterogeneity of clinical presentations and the interactions between antibody isotypes and viral proteins [33] [34] [35] . Many of the current serological assays for the detection of SARS-CoV-2 antibodies on the market are mono-antigenic, with some of them up to two antigens [36] . Currently, most of the tests available on the market use the N antigen (Roche, Euroimmun, Abbott) [5, 9, 10, 25] , the S1 antigen (Euroimmun) [9, 12, 37] , the combination of S1 + S2 antigens (DiaSorin) [9, 12, 38] , some tests also use the RBD antigen (Siemens, Wantai) [21, 22, 39] , or the combination of S and N antigens (iFlash, Maglumi, Mikrogen) [10, 38] . Attempts to correlate the results obtained with different tests using the same antigens are often unsatisfactory. This demonstrates the lack of overall harmonization that exists within these tests [10, 22, 40, 41] . In addition, most tests focus on N and S [9, 16] . At present, a limited number of studies have evaluated the possibility of using the multiplex technique in the detection of anti-SARS-CoV-2 IgG antibodies. Several techniques have been used; Luminex-based assay (N and S antigens) [27] , protein micro-array assay (N and S antigens) [26] , bead-based immune assay (S1, RBD and N antigens) [25, 26] and a solidphase chemiluminescent assay (trimeric S, S1, RBD and N antigens) [42] .

According to the manufacturer's algorithm, the specificity of the CoViDiag was evaluated at 94.7%. The adapted cut-off permit to obtain a slightly higher specificity, i.e., 95.4%, compared to the algorithm of the manufacturer (Figure 1) . These specificities were not optimal but, taken separately, the specificity of each antigen was excellent (i.e., 99.2 to 100%) when using adapted cut-offs, 

This study is the first to evaluate the clinical performance of the CoViDiag IgG test kit, which is designed to detect the presence of IgG antibodies to 5 SARS-CoV-2 antigens. It includes a total of 135 serum samples obtained from 94 rRT-PCR-positive patients and a total of 132 Non-SARS-CoV-2 serum samples. The sensitivity of the CoViDiag test kit after 14 days since symptom onset varies from 92.0 to 100.0%. However, it remains too low in the first 6 days after the onset of symptoms to correctly identify all positive patients, as confirmed by other studies [31, 32] . Nevertheless, detection of early seroconversion may be useful to understand the heterogeneity of clinical presentations and the interactions between antibody isotypes and viral proteins [33] [34] [35] . Many of the current serological assays for the detection of SARS-CoV-2 antibodies on the market are mono-antigenic, with some of them up to two antigens [36] . Currently, most of the tests available on the market use the N antigen (Roche, Euroimmun, Abbott) [5, 9, 10, 25] , the S1 antigen (Euroimmun) [9, 12, 37] , the combination of S1 + S2 antigens (DiaSorin) [9, 12, 38] , some tests also use the RBD antigen (Siemens, Wantai) [21, 22, 39] , or the combination of S and N antigens (iFlash, Maglumi, Mikrogen) [10, 38] . Attempts to correlate the results obtained with different tests using the same antigens are often unsatisfactory. This demonstrates the lack of overall harmonization that exists within these tests [10, 22, 40, 41] . In addition, most tests focus on N and S [9, 16] . At present, a limited number of studies have evaluated the possibility of using the multiplex technique in the detection of anti-SARS-CoV-2 IgG antibodies. Several techniques have been used; Luminex-based assay (N and S antigens) [27] , protein micro-array assay (N and S antigens) [26] , bead-based immune assay (S1, RBD and N antigens) [25, 26] and a solid-phase chemiluminescent assay (trimeric S, S1, RBD and N antigens) [42] .

According to the manufacturer's algorithm, the specificity of the CoViDiag was evaluated at 94.7%. The adapted cut-off permit to obtain a slightly higher specificity, i.e., 95.4%, compared to the algorithm of the manufacturer (Figure 1 ). These specificities were not optimal but, taken separately, the specificity of each antigen was excellent (i.e., 99.2 to 100%) when using adapted cut-offs, suggesting that improvements of the algorithm could be done (Figure 1) . Additionally, none of the false positive samples were positive for more than one antigen suggesting that a sample with two positive antibodies is very unlikely to be a false positive result (Figure 1 ). The utilization of several antigens in one assay is in line with CDC's recommendations about the orthogonal testing strategy [43] . Among our cohort, 14 days after symptom onset and using the adapted cut-offs, only two samples were not positive at all (2.3%) and five were positive for only one antigen (5.8%). Interestingly, for samples with only one antigen, one was positive for the N antigen and 4 were positive for the S2 antigen. As the specificity of N and S2 is 100 and 99.2% with adapted cut-offs, the risk of having a false positive is very low. The calculated specificities were similar to those obtained on Roche assay (i.e., N antigen) but higher compared to those obtained on the Euroimmun (i.e., N antigen) and DiaSorin (i.e., S1/S2 antigen) assays, whatever the use of the manufacturer's or the adapted cut-offs.

Current data on the immune response to SARS-CoV-2 estimate that seroconversion occurs between 7 and 14 days from the onset of symptoms [4, 5, 14, 44] . Our results confirm these observations but bring this evidence to several antigens, i.e., N, RBD, S1 and S2. When considering the individual sensitivity of the test for each antigen, the positivity rate remained low before 14 days and was higher for S2 and N antigens. After 14 days, the sensitivity increased for all antigens even though the data obtained for the NTD remains very low. The most sensitive antigens were N, S2 and RBD but interestingly, RBD is only positive when the samples were also reactive for the N and S2 antigen, except for one sample which has been reactive to NTD instead of RBD. Considering the final result of the test, i.e., positive or negative for SARS-CoV-2 according to the manufacturer's algorithm, the sensitivity after day 14 varies between 92.0 and 100%. The adapted cut-offs obtained from the ROC curves do not significantly change the sensitivity (from 100 to 97.3%) (Figure 2 and Supplemental Table S1 ). These sensitivities were comparable to those obtained on Roche, Euroimmun and DiaSorin assays. The results obtained for the different antigens are in agreement with the data from the literature on other serological assays [45] . Nevertheless, the CoViDiag is able to provide in one single test information on all these antigens, dispensing the use of other analyses to confirm each positive patient and is relevant for seroprevalence studies. Amongst others, this approach might improve the turnaround time and decrease overall costs in a routine testing workflow. Furthermore, because most vaccines will use the S protein or S-domains as immunogen, [46] assays targeting the N-protein would therefore not be a good candidate to evaluate the vaccine response [47] . However, having only assays targeting the S-protein might, be misleading in some situations. If a patient develops COVID-19-related symptoms following the vaccination, an assay with multiple targets might differentiate neo-COVID-19 infection from side effects due to the vaccination (i.e., flu-like syndrome).

In our cohort, the disease severity and the antibody titers were not related in samples obtained 14 days since symptom onset (Figure 4 ). To et al. also reported a lack of correlation between serum SARS-CoV-2 antibody levels and clinical severity of disease [48] . Van Elslande et al. reported no significant difference in seroconversion times for IgG between critical and non-critical patients [10] . Similar results were also obtained in previous studies [49, 50] . However, other research groups reported that the early detection of anti-SARS-CoV-2 antibodies, and high antibody titers, were associated to disease severity [8, 39, 50, 51] . Interestingly, in our study, higher mean antibody titers were observed in the hospitalized cohort compared to the non-hospitalized cohort when samples were obtained early, i.e., from 0 to 6 days and from 7 to 13 days, after the onset of COVID-19 symptoms ( Figure 5 ). However, due to the retrospective design of this study and the heterogeneity between the hospitalized versus the non-hospitalized cohorts within these timeframes i.e., from 0 to 6 days and from 7 to 13 days, definitive conclusions cannot be drawn. Such finding deserves further investigations which currently fall outside the scope of this study.

This study evaluates the performance of a multiplex assay for the simultaneous detection of IgG antibodies to 5 different antigens of SARS-CoV-2. The specificity of the CoViDiag assay was excellent for all antigens (99.2 to 100%) using adapted cut-offs. Moreover, none of the false positive samples were positive for more than one antigen. The sensitivity obtained from samples collected 14 days since symptom onset varied from 92.0 to 100.0% depending on the antibody considered. The role of the NTD antigen in this test remains unclear based on our results. Nevertheless, the multiplex serological test shows a definite advantage over single gene tests. The combination of several antigens in the same test improves the overall specificity and sensitivity and permits us to counterbalance the lack of sensitivity or specificity of one of the antigens. We also observed a trend toward higher antibody levels in hospitalized patients in the first two weeks following the onset of symptoms. Non-statistical differences were, however, observed between hospitalized and non-hospitalized patients after 14 days since symptom onset. Future research and development of this multiplex approach is needed for better prediction of disease severity, for monitoring the immune response in patients who have suffered from COVID-19 and for the design of future vaccination campaigns.

Supplementary Materials: The following are available online at http://www.mdpi.com/2077-0383/9/11/3752/s1. Figure S1 : Schematic representation of the spiking of the antigens at the bottom of each well. CTRL+: positive control; N: nucleocapsid protein; S1: subunit 1 of the spike protein; RBD: S1 protein receptor binding domain; NTD: N-terminal subunit of protein S1; S2: subunit 2 of the spike protein. Table S1 : Description of false-positive samples according to the positive antigen, the AU value of the antigen concerned and the potentially cross-reactive antibody. * = not positive when adapted cut-offs are used. # = positive when adapted cut-offs are used. Table S2 : Evolution of the positivity rates according to the number of days since the onset of symptoms for each antigen (at least 1 positive antigen based on the adapted cut-offs). (Put them in the SI as just one paragraph). 

COVID-19) Situation Report-Weekly Evaluation Update

In Vitro Diagnostic Assays for COVID-19: Recent Advances and Emerging Trends

Potential preanalytical and analytical vulnerabilities in the laboratory diagnosis of coronavirus disease 2019 (COVID-19)

Molecular, serological, and biochemical diagnosis and monitoring of COVID-19: IFCC taskforce evaluation of the latest evidence

Analytical performances of a chemiluminescence immunoassay for SARS-CoV-2 IgM/IgG and antibody kinetics

SARS-CoV-2 Serology: Much Hype, Little Data

IFCC interim guidelines on serological testing of antibodies against SARS-CoV-2

Review of Current Advances in Serologic Testing for COVID-19

Performance of six SARS-CoV-2 immunoassays in comparison with microneutralisation

Antibody response against SARS-CoV-2 spike protein and nucleoprotein evaluated by four automated immunoassays and three ELISAs

Performance Characteristics of Four High-Throughput Immunoassays for Detection of IgG Antibodies against SARS-CoV-2

Analytical and clinical validation of an ELISA for specific SARS-CoV-2 IgG, IgA, and IgM antibodies

High clinical performance and quantitative assessment of antibody kinetics using a dual recognition assay for the detection of SARS-CoV-2 IgM and IgG antibodies

Response of anti-SARS-CoV-2 total antibodies to nucleocapsid antigen in COVID-19 patients: A longitudinal study

Clinical Performance of the Elecsys Electrochemiluminescent Immunoassay for the Detection of SARS-CoV-2 Total Antibodies

The receptor binding domain of the viral spike protein is an immunodominant and highly specific target of antibodies in SARS-CoV-2 patients

Heterogeneous antibodies against SARS-CoV-2 spike receptor binding domain and nucleocapsid with implications for COVID-19 immunity

Clinical Outcomes in 55 Patients with Severe Acute Respiratory Syndrome Coronavirus 2 Who Were Asymptomatic at Hospital Admission in Shenzhen

Structure of the SARS-CoV-2 spike receptor-binding domain bound to the ACE2 receptor

Structure, Function, and Antigenicity of the SARS-CoV-2 Spike Glycoprotein

An evaluation of COVID-19 serological assays informs future diagnostics and exposure assessment

Evaluation of three fully-automated SARS-CoV-2 antibody assays

The important role of serology for COVID-19 control

SARS-CoV-2 antibodies titration: A reappraisal

SARS-CoV-2-Specific Antibody Detection for Seroepidemiology: A Multiplex Analysis Approach Accounting for Accurate Seroprevalence

Accurate serology for SARS-CoV-2 and common human coronaviruses using a multiplex approach

Multiplex detection and dynamics of IgG antibodies to SARS-CoV2 and the highly pathogenic human coronaviruses SARS-CoV and MERS-CoV

Evaluation of Orthogonal Testing Algorithm for Detection of SARS-CoV-2 IgG Antibodies

Distinct Early Serological Signatures Track with SARS-CoV-2 Survival

Information for Use CoViDiag: Version 1.3

Comparative assessment of multiple COVID-19 serological technologies supports continued evaluation of point-of-care lateral flow assays in hospital and community healthcare settings

SARS-CoV-2 seroprevalence and neutralizing activity in donor and patient blood

Ultrasensitive high-resolution profiling of early seroconversion in patients with COVID-19

Antibody seroconversion in asymptomatic and symptomatic patients infected with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)

Ultra-sensitive Serial Profiling of SARS-CoV-2 Antigens and Antibodies in Plasma to Understand Disease Progression in COVID-19 Patients with Severe Disease

Validation of a chemiluminescent assay for specific SARS-CoV-2 antibody

Diagnostic performance of seven rapid IgG/IgM antibody tests and the Euroimmun IgA/IgG ELISA in COVID-19 patients

Diagnostic performances and thresholds: The key to harmonization in serological SARS-CoV-2 assays?

Antibody Profiles in Mild and Severe Cases of COVID-19

Clinical evaluation of five different automated SARS-CoV-2 serology assays in a cohort of hospitalized COVID-19 patients

Immune response to SARS-CoV-2 in health care workers following a COVID-19 outbreak: A prospective longitudinal study

Evaluation of a novel multiplexed assay for determining IgG levels and functional activity to SARS-CoV-2

Interim Guidelines for Covid-Antibody Testing

Unexpected kinetics of anti-SARS-CoV-2 total antibodies in two patients with chronic lymphocytic leukemia

Performance characteristics of five immunoassays for SARS-CoV-2: A head-to-head benchmark comparison

The starting line for COVID-19 vaccine development

SARS-CoV-2 Serology-Need for Quantitative Testing and Interpretive Reporting

Temporal profiles of viral load in posterior oropharyngeal saliva samples and serum antibody responses during infection by SARS-CoV-2: An observational cohort study

Profile of IgG and IgM antibodies against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)

Antibody responses to SARS-CoV-2 in patients of novel coronavirus disease 2019

Viral Kinetics and Antibody Responses in Patients with COVID-19

Funding: This research received no external funding.

Among the authors, J.D. is CEO and founder of QUALIblood s.a., a contract research organization manufacturing the DP-Filter, is co-inventor of the DP-Filter (patent application number: PCT/ET2019/052903) and reports personal fees from Daiichi-Sankyo, Mithra Pharmaceuticals, Stago, Roche and Roche Diagnostics outside the submitted work. The other authors have no conflict of interest to disclose.