Abstract
The COVID-19 pandemic illustrated the important role of diagnostic tests, including lateral flow tests (LFTs), in identifying patients and their contacts to slow the spread of infections. INSTAND performed external quality assessments (EQA) for SARS-CoV-2 antigen detection with lyophilized and chemically inactivated cell culture supernatant of SARS-CoV-2 infected Vero cells. A pre-study demonstrated the suitability of the material. Participants reported qualitative and/or quantitative antigen results using either LFTs or automated immunoassays for five EQA samples per survey. 711 data sets were reported for LFT detection in three surveys in 2021. This evaluation focused on the analytical sensitivity of different LFTs and automated immunoassays. The inter-laboratory results showed at least 94% correct results for non-variant of concern (VOC) SARS-CoV-2 antigen detection for viral loads of ≥ 4.75 × 106 copies/mL and SARS-CoV-2 negative samples. Up to 85% had success for a non-VOC viral load of ~ 1.60 × 106 copies/mL. A viral load of ~ 1.42 × 107 copies/mL of the Delta VOC was reported positive in > 96% of results. A high specificity was found with almost 100% negative SARS-CoV-2 antigen results for HCoV 229E and HCoV NL63 positive samples. Quantitative results correlated with increasing SARS-CoV-2 viral load but showed a broad scatter. This study shows promising SARS-CoV-2 antigen test performance of the participating laboratories, but further investigations with the now predominant Omicron VOC are needed.
Similar content being viewed by others
Introduction
Since the onset of COVID-19 in 2019, quantitative reverse transcription polymerase chain reaction (RT-qPCR) methods have been considered the gold standard for the detection of SARS-CoV-2 in infected individuals as a result of their high sensitivity and specificity1. However, these methods are challenging to handle as they require specialized laboratory equipment and trained professionals, have relatively long processing, and reaction times2,3,4,5. A lack of sufficient infrastructure means that it can take up to a week after sample collection before the test results are obtained6,7,8. Antigen-based lateral flow tests (LFTs) provide the opportunity to diagnose symptomatic individuals earlier, thereby potentially enabling interruption of virus transmission sooner9. In addition, antigen-LFTs could be a more cost-effective method than nucleic acid amplification techniques (NAAT)10, which is an important factor in low- and middle-income countries. As a result, the WHO has defined minimum performance criteria for antigen-LFTs in comparison to NAAT-based methods, and recommends performing the antigen-LFT within the first five to seven days after the onset of symptoms9.
A previous comparative evaluation of the sensitivity of 122 CE-marked SARS-CoV-2 antigen LFTs revealed a wide range of sensitivity for an evaluation panel consisting of pooled clinical specimens with high, medium and moderate viral loads. The study concluded that most of the SARS-CoV-2 antigen-LFTs analyzed appeared to be suitable in detecting acute infections with high viral loads; however, 26 out of the 122 LFTs indicated a lower sensitivity and some LFTs failed completely11. In addition, a meta-analysis conducted by Dinnes et al., which compared various published antigen-LFT study results, showed that, while many LFTs met the WHO specificity criteria, only seven were able to meet the sensitivity criterion in the pooled results of all of the analyzed studies12. A few LFTs were found to meet the criteria in individual studies but not in the pooled analysis. These results highlight the need for good quality control procedures to enable the best possible diagnostic performance.
The Society for Promoting Quality Assurance in Medical Laboratories e.V. (INSTAND) has been designated a German reference institution by the German Medical Association. In March 2021 INSTAND became one of the first institutions worldwide to introduce an external quality assessment (EQA) scheme, also known as a proficiency test, for SARS-CoV-2 antigen detection.
In this paper, we present the qualitative antigen results for both antigen-LFTs and automated immunoassays from the three first EQA surveys performed in 2021 as well as the quantitative results of the automated immunoassays from March 2021. This report is the first to evaluate consecutive EQA surveys for SARS-CoV-2 antigen detection.
Materials and methods
Sample materials—properties and preparation
Each survey consisted of five different samples. SARS-CoV-2 positive samples contained supernatants of cell cultures infected with a non-VOC SARS-CoV-2 or the SARS-CoV-2 Delta variant of concern (VOC) B.1.617.2 (Table 1). For preparation of cells and viruses see13. Briefly, different strains of SARS-CoV-2 and alpha coronaviruses were cultivated, and the supernatants were collected and further processed as described previously. Vero cells (ATCC CRL-1586) were used for propagation of the SARS-CoV-2 viruses, Huh-7 cells (CVCL_0336) for the propagation of alpha coronavirus hCoV 229E and LLC-MK2 cells (ATCC CCL-7) for the propagation of alpha coronavirus hCoV NL63 (Table 1). The two alpha coronaviruses hCoV229E (March 2021) and hCoV NL63 (June 2021) were used as specificity controls to get an impression, whether the detection antibodies in LFT or automated immunoassays might be sufficiently specific for coronavirus SARS-CoV-2. High test specificity is needed in terms of quick and selective isolation of SARS-CoV-2 infected individuals, as the acute respiratory symptoms of the different coronaviruses are similar. In each survey, one sample was virus negative and contained cell culture lysate from non-infected MRC-5 cells (ATCC-CCL-171). See Table 1 for sample properties and distribution in the EQA surveys.
All materials containing the SARS-CoV-2 virus were chemically inactivated with beta-propiolactone (BPL). A 10 mL aliquot of SARS-CoV-2 positive cell culture supernatant was treated with 0.05% BPL for 14 h at 4 °C. Subsequently, the BPL was hydrolyzed for 2 h at 37 °C. This treatment reduced the number of plaque-forming units (PFU) from 1.21 × 106 PFU/mL to 0 PFU/mL.
The samples containing the alpha coronavirus were not inactivated. All viruses were provided by the National Coronavirus Consultant Laboratory, Charité—University Medicine Berlin, Institute of Virology, Berlin, Germany.
The samples were generated as described previously13. In brief, the supernatants from cell cultures were diluted using a cell culture medium (Minimal Essential Medium, PanBioTech, Aidenbach, Germany), supplemented with non-essential amino acids (PanBioTech), HEPES buffer (PanBioTech) and fetal bovine serum (PanBioTech, gamma irradiated; 15% v/v for supplemented cell culture medium). Finally, 0.5 mL of the materials were aliquoted in screw cap micro tubes (2.0 mL; Sarstedt, Nümbrecht, Germany) and lyophilized. Randomly selected vials of each of the EQA samples were analyzed for stability during the period of the EQA survey and for homogeneity according to DIN EN ISO/IEC 17043:2010-0514 in order to show the stability of the control materials. Prior to each of the EQA surveys in March, June and September 2021, samples were proficiency tested by up to four INSTAND expert laboratories which confirmed that the samples were fit for purpose.
Pre-study on EQAS sample suitability
To ensure the suitability of the samples for the March 2021 EQA survey, an independent pre-study was conducted to identify the most appropriate dilution of SARS-CoV-2 positive supernatant as well as to assess the effect of lyophilization on antigen detectability. A tenfold dilution series of a supernatant of a non-VOC SARS-CoV-2 was prepared (dilution from 1:75 to 1:750,000). Half of the samples were lyophilized and half were kept in liquid form. These samples were characterized by four expert laboratories and the sample manufacturer using different SARS-CoV-2 antigen-LFTs. The LFTs of Abbott—Panbio COVID-19 Ag Rapid Test and Roche (SD Biosensor) 1—SARS-CoV-2 Rapid Ag Test were used by several laboratories. Automated immunoassays were applied by two laboratories with either the SARS-CoV-2 Ag Roche Cobas [COI] or the SARS-CoV-2 Ag-CLIA Diasorin [TCID50/mL]. In the case of a positive result, the expert laboratories were asked to compare the intensity of the test line with the intensity of the control line and to classify the visible difference in intensity of both lines. A positive test result was classified as either ‘test line stronger than control line’ (5), ‘test line as strong as control line’ (4), ‘test line weaker than control line’ (3), ‘test line much weaker than control line’ (2) and ‘test line faint, just visible’ (1). A negative result was defined as (0). This classification was used purely for analytical evaluation and does not in any way suggest that such an assessment could be used on a patient specimen to provide any additional clinical information. Four laboratories performed additional RT-qPCR to further characterize the samples.
Determination of SARS-CoV-2 RNA loads by RT-dPCR
Reverse transcription digital PCR (RT-dPCR) was used to analyze the SARS-CoV-2 RNA load of all SARS-CoV-2 positive EQA samples in order to correlate the EQA antigen testing results of the selected dilution levels with the determined viral loads.
RT-dPCR was used for quantification purposes by three National Metrology Institutes (NMIs): the National Measurement Laboratory (NML at LGC, UK), the National Institute of Standards and Technology (NIST, USA) and the Physikalisch-Technische Bundesanstalt (PTB, Germany).
The lyophilized samples were reconstituted in 0.5 mL molecular biology grade water (PCR grade), extracted using the Qiagen QIAamp Viral RNA mini kit, and eluted (volumes used by each laboratory as described previously13). These eluates were analyzed by RT-dPCR on the Bio-Rad QX200 platform using the Bio-Rad one-step RT-digital droplet (dd)PCR supermix using the CDC N1, CDC N215 and China N assay16. The results were analyzed by NML using R version 3.6.1 and RStudio version 1.2.5001. In Table 1, the assigned values for SARS-CoV-2 RNA load are shown for the five SARS-CoV-2 positive EQA samples.
EQA procedure
INSTAND conducted a total of three EQA surveys: March (263 participants), July (148 participants) and September 2021 (168 participants) (see Supplementary Figure S1). Participants received five samples per EQA survey. See Table 1 for details on the properties of each EQA sample.
The samples had to be reconstituted with 0.5 mL double distilled water (sterile, pyrogen-free, PCR-grade) for 20 min at room temperature. In order to standardize the results of the three EQA surveys, the laboratories were instructed to use 100 µL of the reconstituted EQA sample for the testing regardless of the LFT or automated immunoassay used, representing the entire liquid volume contained in a standard swab at the time of the study. This volume of 100 µL was directly added to the buffer/elution solution specified by the manufacturer. Then the LFT or automated immunoassay was supposed to be performed according to the manufacturers’ instructions. The laboratories were asked to report back to INSTAND with their results via the RV-Online platform (https://rv-online.instandev.de). Multiple results for each sample, obtained by different test systems, could be entered. Furthermore, the laboratories were asked to provide detailed information on the test system(s) used for each analysis, including the test kit supplier(s) and test kit(s) (see Supplementary Table S2 online). The tests used by the laboratories for qualitative detection are summarized in Supplementary Table S3.
If EQA participants reported qualitative LFT results, they were further asked to compare the visible intensity of the test line with the visible intensity of the corresponding control line according to the criteria from the pre-study (Section "Pre-study on EQAS sample suitability"). If the same sample material was used for more than one EQA survey (see Table 1), the results from the participants were aggregated and evaluated together. This was purely for the purposes of understanding how the different dilutions of EQA material may compare between laboratories and LFTs or automated immunoassays and should not be construed as a clinically appropriate practice.
In the case of automated immunoassays, quantitative values were reported in pg/mL (ScheBo SARS-CoV-2 Antigen ELISA and Lumipulse G SARS-CoV-2 Antigen assay) or Tissue Culture Infection Dose50 (TCID50)/mL (LIAISON SARS-CoV-2 Ag assay). Only the quantitative results of the first EQA survey in March 2021 were evaluated. The quantitative results from the other surveys were not evaluated in detail due to the low number of quantitative values and thus the insufficient statistical significance.
Data evaluation
We evaluated a total of 711 qualitative datasets from all three EQA surveys provided by 344 laboratories, as well as 36 quantitative datasets from the first survey in March provided by 36 laboratories. The evaluation was carried out in a test- and sample-dependent way (Supplementary Table S3).
Basic statistical analyses were performed using JMP 16.0 from SAS Institute (Cary, North Carolina, USA).
Generation of images
The overlay images were generated using the GIMP–GNU Image Manipulation Program 2.10.3.
Results
Our study comprised the validation of suitable SARS-CoV-2 antigen EQA samples and the evaluation of the interlaboratory results of the qualitative and quantitative detection of SARS-CoV-2 antigens from three EQA surveys (March, June and September 2021) with five samples each (Table 1).
Pre-study results
Since the cell culture supernatant samples where chemically inactivated their suitability for LFTs was initially assessed. The SARS-CoV-2 RNA quantity was evaluated as a surrogate for viral, and thus epitope quantity, for the positive samples used in these EQA surveys. In the pre-study described in Section "Pre-study on EQAS sample suitability", four expert laboratories and the sample manufacturer confirmed that lyophilization did not affect the detectability of the SARS-CoV-2 antigen in the LFTs and automated immunoassays. For the March EQA, a dilution of 1:750 was selected for preparation of one of the five EQA samples since, in the pre-study, mainly band intensities between ‘test line much weaker than control line’ (2) and ‘test line weaker than control line’ (3) were observed at this dilution, corresponding to a test sample that was reliably detectable without excessively high viral load. The 1:7,500 and 1:75,000 dilution levels produced lower band intensities and were used in the March 2021 EQA to test the detectability of samples that showed borderline results (Fig. 1, see Supplementary Table S1 online).
The differences of the determined SARS-CoV-2 RNA loads corresponded very well to the expected gradations of the individual dilution levels for the four non-VOC SARS-CoV-2 EQA samples.
EQA results
Since the qualitative and quantitative EQA results were similar for March, June and September, only the data for the March survey are presented graphically (Figs. 2, 3). The qualitative results for June and September can be found in the supplements (Supplementary Figures S2 and S3). The qualitative results and assay/test kit-specific results of the participants of all surveys were summarized (see Supplementary Tables S4 and S5 online).
In all three surveys, the qualitative results showed almost 100% correct results for samples with non-infected SARS-CoV-2 control cells (MRC-5 cells) as well as for samples with human alpha coronaviruses 229E and NL63, both for the LFTs (see Fig. 2a for the March EQAS and Table 2a for all three surveys) and for the automated immunoassays (see Fig. 2b for the March EQAS and Table 2a for all three surveys). In the case of the human alpha coronaviruses, used as test specificity control, 0.7% false positive results were observed for LFTs for HCoV 229E in March and 1.3% false positive results for HCoV NL63 in June (see Table 2a and Supplementary Tables S4 and S5 online). No false positive antigen results were reported for the automated immunoassays.
For the LFTs, samples containing the same non-VOC SARS-CoV-2 with a viral load of 1.65 × 107 copies/mL were correctly identified as positive in at least 97.5% of the analyses (see Table 2b).
In terms of the distribution of the qualitative results for lower viral loads of the same non-VOC SARS-CoV-2, high virus detection success rates continued to be observed down to a viral load of 4.75 × 106 copies/mL (Table 2b, 94.3% positive results in September). The antigen results became divergent at a lower viral load of 1.60 × 106 copies/mL. The detection rate was between 53.6 and 65.8% for the LFTs and between 68.4 and 85.5% (borderline results of 2.6% to 9.4%) for the automated systems (Table 2b and Fig. 4). The percentage of positive antigen LFT or automated immunoassay results declined distinctly for the lower SARS-CoV-2 viral load of 1.55 × 105 copies/mL in March. Still, 6.0% of the LFT results and 14.5% of the automated immunoassay results were reported to be positive (Table 2b and Fig. 2).
In the case of the sample containing the SARS-CoV-2 Delta VOC with an assigned viral load of 1.42 × 107 copies/mL (Table 2b), 96.2% of the results were reported to be positive for SARS-CoV-2 antigen. For the automated systems, success in SARS-CoV-2 antigen detection was nearly 100%, with only one borderline result in the case of the Delta VOC of the virus (Table 2b).
In addition to the qualitative result, between 66 and 80% of the EQA participants using LFTs compared test line intensities to the visible intensity of the control line. 49.6% of the results for the samples with the highest non-VOC SARS-CoV-2 RNA load of 1.65 × 107 copies/mL showed a ‘test line weaker than control line’ (3) (Fig. 5). For non-VOC SARS-CoV-2 samples containing ~ 1.60 × 106 copies/mL, 58.6% of the participants estimated weaker test line intensities of 1–2 (test line faint, just visible or test line much weaker than control line). For this sample, 34.2% of the results were found to be negative. The sample with an intermediate viral load of ~ 4.75 × 106 copies/mL showed that 47.5% of the results displayed a ‘test line much weaker than the control line’ (2) (Fig. 5). For the sample with the lowest SARS-CoV-2 RNA load of ~ 1.55 × 105 copies/mL, 4.0% of the participants were able to visually detect the SARS-CoV-2 antigen successfully (Fig. 5). The extent to which these values are influenced by differences in an individual’s visual ability cannot be assessed.
The distribution of the quantitative antigen results of the automated systems in March is shown in Fig. 3. The quantitative results in pg/mL or TCID50/mL correlated well with the estimated viral load of the samples. The value distribution of the reported TCID50/mL results showed an overlap between the different viral loads, for example between the samples with the lowest viral load of 1.55 × 105 copies/mL and the SARS-CoV-2 negative samples (March). For the SARS-CoV-2 negative samples, which contained either non-infected MRC-5 cells or HCoV 229E, antigen concentrations were below 0.15 pg/mL or scattered around the median value of 40 TCID50/mL.
Discussion
This study investigates the interlaboratory performance of SARS-CoV-2 antigen testing for both antigen-LFTs and automated immunoassays. Although the RT-qPCR methods are considered to be the gold standard for diagnosing COVID-19 through the detection of SARS-CoV-2 RNA, antigen testing provides and alternative solution that is rapid, inexpensive and—in many cases—laboratory-independent diagnostic detection of SARS-CoV-29,10.
The pre-study demonstrated that the EQA materials were suitable for SARS-CoV-2 antigen testing by both LFTs and automated immunoassays for both lyophilized and liquid samples (Fig. 1). The qualitative EQA results for SARS-CoV-2 negative samples showed overall good analytical sensitivity and specificity.
All methods were able to detect the positive samples with an estimated SARS-CoV-2 viral load of ~ 1.42 × 107 copies/mL (Fig. 2). LFTs had a high, 94% success rate for detecting the virus when viral load was ~ 4.75 × 106 copies/mL. This success rate decreased to between 54 and 66% when there was an estimated viral load of 1.17 × 106 copies/mL. For the automated systems, the analytical sensitivity was generally better, with positive rates of 68% to 86% for an estimated viral load of ~ 1.17 × 106 copies/mL (Fig. 4, Table 2b). Unexpectedly, the percentage of positive results for a viral load of ~ 1.60 × 106 copies/mL declined notably for the automated immunoassays. One reason for this is most likely the small sample size, which is more easily impacted by false-negative results.
So, in this study, a concentration of 1.17 × 106 copies/mL seems to be the viral load, where the rapid tests begin to struggle in the detection of SARS-CoV-2.
The commercial tests used by the laboratories could not be directly ranked according to quality based on the individual results, presented in Supplementary Tables S4 and S5 online, since (i) laboratory-intrinsic influences might have predominated, (ii) the influence of swabs used in routine diagnostics was deliberately omitted and (iii) some LFTs or automated immunoassays were only used by a handful of laboratories.
For the samples used in the EQA survey in March 2021, median antigen concentrations reported for the automated immunoassays correlate well with the different estimated viral loads (Fig. 3). However, scatter in reported TCID50/mL antigen concentrations led to an overlap in the result distribution between the EQA samples with a tenfold difference in the estimated viral load. An overlap in the reported concentrations was also found between the SARS-CoV-2 negative sample and the sample with a SARS-CoV-2 RNA load of ~ 1.60 × 105 copies/mL (Fig. 3). This indicates that the antigen concentration of this sample is close to the manufacturer’s cut off of 200 TCID50/mL.
Encouragingly, a high analytical specificity was found with respect to the other human coronaviruses HCoV 229E and HCoV NL63, with almost 100% negative rate for SARS-COV-2 antigen detection using both LFTs and automated immunoassays. Thus, the detection antibodies of the LFTs and immunoassays appeared to be sufficiently specific to detect SARS-CoV-2 antigens. This will be helpful for a quick and selective isolation of SARS-CoV-2 infected individuals using LFTs or automated immunoassays, since the acute respiratory symptoms of the different coronaviruses are similar. Other reports comparing different LFTs describe similar test performances3,17.
One possible explanation for the differences in analytical performance of the automated immunoassays and the LFTs in SARS-CoV-2 antigen detection at low viral load may be caused by the different workflows. The difference in the signal detection, generally visual and person-dependent for LFTs and using photodetectors in a qualified and validated test systems, might influence the interpretation of the results. Furthermore, LFTs often lack proper low positive control samples per lot which is particularly important when assessing weak line intensities.
In the September EQA, one sample containing the SARS-CoV-2 Delta VOC was included and showed similar results to the non-VOC SARS-CoV-2 virus. This observation is consistent with the results of other comparative studies18,19 which have shown that antigen assays that use cell culture samples were generally just as sensitive to other VOCs, such as the Alpha, Beta or Gamma VOC, as they were to the non-VOC SARS-CoV-2 virus. Contrary, reduced test sensitivities were described in antigen detection of the Omicron VOC20,21. Due to different antibody characteristics and various antigen binding sites, the sensitivity and specificity of the LFTs or automated immunoassays could be influenced by mutations in the antigen22,23. Thus, the sensitivity and specificity of each test has to be validated for each variant. Research is still ongoing24 and the introduction of an International Standard preparation is expected to produce better comparability.
Clinical parameters, like the severity or the course of disease, are essential factors for successful SARS-CoV-2 diagnostics25,26,27, but were not addressed by the EQA surveys presented here. Furthermore, preanalytical issues, including specimen type, collection, storage and transportation, are not covered in the concept of this EQA scheme, even though sample collection is another crucial aspect for SARS-CoV-2 testing9. One could get an indication of the influence of the sample collection by using a swab instead of a fixed sample volume pipetted for sample application. This could result in a further decrease in the concentration of the SARS-CoV-2 antigen in the LFT or automated system, which could have a negative impact on antigen detection, as also shown by a study by Puyskens et al.28 as well as by Cubas-Atienzar et al.29. To address this problem, the INSTAND EQA procedure was changed starting in November 2021 to include the use of a swab for sample collection. The respective results will be evaluated after two to three EQA surveys.
The strength of our INSTAND study was the use of well-defined EQA samples containing different viral loads of quantified SARS-CoV-2 non-VOC and Delta VOC and additional alpha coronaviruses. This enabled an assessment to be made of the general analytical performance of a broad number of laboratories worldwide which applied different antigen detection methods. The chemical inactivation of SARS-CoV-2 did not seem to have a negative effect on the analytical performance of any of the antigen LFTs presented in this study, which corresponds to the results previously reported by Zhou et al.30.
Our study does have some limitations: Since the EQA materials were derived from cell culture supernatants and, in contrast to routine diagnostics, no swabs were used in the analyses, the initial steps of sample preparation differ from the instructions in the test kit manuals. Furthermore, the sample origin leads to a discrepant matrix compared to patient material, so that EQA survey samples might contain different interfering components. Unfortunately, few of quantitative results for automated immunoassays were reported, which means limited statistical significance of the analysis. So more data is needed to confirm the findings of the quantitative results.
The fact that the analysis is based on EQA data, derived from artificial samples, one has to be careful with clinical conclusions on the quality and suitability of individual tests. Rather, they mirror the test-dependent performance of laboratories for a certain medical laboratory detection and thus provide an insight about the general quality of diagnostic testing. Interlaboratory assessments have therefore a key role in monitoring and promoting accurate and reliable diagnostic testing.
As virus detectability by the different LFTs and automated systems might be impaired depending on the virus variant, further analyses of antigen test performance for other SARS-CoV-2 VOCs, such as the Omicron VOC, would be of interest, especially at lower RNA loads (< 1.65 × 107 copies/mL), and should be conducted as soon as possible.
To sum up, this study shows satisfactory analytical sensitivities of at least 94% for LFTs and automated immunoassays in case of non-VOC SARS-CoV-2 antigen detection for viral loads of ≥ 4.75 × 106 copies/mL based on interlaboratory data. The success in non-VOC antigen detection by LFTs decreased to at least 53.4% for lower viral loads of ~ 1.60 × 106 copies/mL, but the EQA results were still up to 85% positive for automated immunoassay detection. In the case of the Delta VOC the analytical sensitivity was found to be almost the same compared to this for a similar viral load of the non-VOC SARS-CoV-2, but the Delta VOC was only used in a high virus load of ~ 1.42 × 107 copies/mL in one of the three EQA surveys analyzed. Analysis of HCoV 229E- and HCoV NL63-positive samples showed high specificity for antigen detection of SARS-CoV-2, with nearly 100% negative results. Thus, the EQA results for SARS-CoV-2 antigen detection by LFTs and automated immunoassays showed overall very promising performances of the participating laboratories. As reduced sensitivity of rapid antigen detection was described for the SARS-CoV-2 Omicron VOC in clinical performance21, further investigations of interlaboratory test data for the Omicron VOC will be of interest.
NIST disclaimer
Points of view in this document are those of the authors and do not necessarily represent the official position or policies of the U.S. Department of Commerce. Certain commercial software, instruments, and materials are identified in order to specify experimental procedures as completely as possible. In no case does such identification imply a recommendation or endorsement by NIST, nor does it imply that any of the materials, instruments, or equipment identified are necessarily the best available for the purpose. All work has been reviewed and approved by the U. S. National Institute of Standards and Technology Research Protections Office. This study was determined to be “not human subjects research” (often referred to as research not involving human subjects) as defined in U. S. Department of Commerce Regulations, 15 CFR 27, also known as the Common Rule (45 CFR 46, Subpart A), for the Protection of Human Subjects by the NIST Human Research Protections Office and therefore not subject to oversight by the NIST Institutional Review Board.
Data availability
All data supporting the results of this article are provided in the supplementary information as raw data of the pre-study and anonymized raw data of the EQA survey.
References
WHO, Diagnostic Testing for SARS-CoV-2: Interim Guidance, 11 September 2020. (2020), World Health Organization.
Chaimayo, C. et al. Rapid SARS-CoV-2 antigen detection assay in comparison with real-time RT-PCR assay for laboratory diagnosis of COVID-19 in Thailand. Virol. J. 17(1), 177 (2020).
Corman, V. M. et al. Comparison of seven commercial SARS-CoV-2 rapid point-of-care antigen tests: A single-centre laboratory evaluation study. Lancet Microbe 2(7), e311–e319 (2021).
Pillonel, T., Scherz, V., Jaton, K., Greub, G. & Bertelli, C. Letter to the editor: SARS-CoV-2 detection by real-time RT-PCR. Euro Surveill. https://doi.org/10.2807/1560-7917.ES.2020.25.21.2000880 (2020).
Sethuraman, N., Jeremiah, S. S. & Ryo, A. Interpreting diagnostic tests for SARS-CoV-2. JAMA 323(22), 2249–2251 (2020).
Caruana, G. et al. Diagnostic strategies for SARS-CoV-2 infection and interpretation of microbiological results. Clin. Microbiol. Infect 26(9), 1178–1182 (2020).
Kobia, F. & Gitaka, J. COVID-19: Are Africa’s diagnostic challenges blunting response effectiveness?. AAS Open Res. 3, 4 (2020).
Mendelson, M. & Madhi, S. South Africa’s coronavirus testing strategy is broken and not fit for purpose: It’s time for a change. S. Afr. Med. J. 110(6), 429–431 (2020).
WHO. Antigen-detection in the diagnosis of SARS-CoV-2 infection - Interim Guidance. 2021 06.10.2021 08.10.2021]; Available from: file:///C:/Users/N2944~1.WOJ/AppData/Local/Temp/WHO-2019-nCoV-Antigen-Detection-2021.1-eng.pdf.
Ricks, S. et al. Quantifying the potential value of antigen-detection rapid diagnostic tests for COVID-19: A modelling analysis. BMC Med. 19(1), 75 (2021).
Scheiblauer, H. et al. Comparative sensitivity evaluation for 122 CE-marked rapid diagnostic tests for SARS-CoV-2 antigen, Germany, September 2020 to April 2021. Euro Surveill. https://doi.org/10.2807/1560-7917.ES.2021.26.44.2100441 (2021).
Dinnes, J. et al. Rapid, point-of-care antigen tests for diagnosis of SARS-CoV-2 infection. Cochrane Database Syst. Rev 7(7), Cd013705 (2022).
Vierbaum, L. et al. RNA reference materials with defined viral RNA loads of SARS-CoV-2-A useful tool towards a better PCR assay harmonization. PLoS ONE 17(1), e0262656 (2022).
DIN EN ISO/IEC. Konformitätsbewertung-Allgemeine Anforderungen an Eignungsprüfungen (ISO/IEC 17043: 2010); Deutsche und Englische Fassung EN ISO/IEC 17043: 2010. Conformity assessment-General requirements for proficiency testing (ISO/IEC 17043: 2010); German and English version EN ISO/IEC 17043: 2010, (2010)
CDC. CDC N1 assay, CDC N2 assay [CDC. Research use only 2019-novel coronavirus (2019-nCoV) real-time RT-PCR primers and probes. 2020; Available from: https://www.cdc.gov/coronavirus/2019-ncov/lab/rt-pcr-panel-primer-probes.html.
WHO. China N Assay Specific Primers and Probes for Detection of 2019 Novel Coronavirus. (2020).
Andreani, J. et al. Evaluation of six commercial SARS-CoV-2 rapid antigen tests in nasopharyngeal swabs: Better knowledge for better patient management?. J. Clin. Virol. 143, 104947 (2021).
Bekliz, M. et al. Analytical performance of eleven SARS-CoV-2 antigen-detecting rapid tests for Delta variant. medRxiv https://doi.org/10.1101/2021.10.06.21264535 (2021).
Jungnick, S. et al. In vitro rapid antigen test performance with the SARS-CoV-2 variants of concern B.1.1.7 (Alpha), B.1.351 (Beta), P.1 (Gamma), and B.1.617.2 (Delta). Microorganisms 9(9), 1967 (2021).
Osterman, A. et al. Impaired detection of omicron by SARS-CoV-2 rapid antigen tests. Med. Microbiol. Immunol. 211(2–3), 105–117 (2022).
Wagenhäuser, I. et al. Virus variant-specific clinical performance of SARS coronavirus two rapid antigen tests in point-of-care use, from November 2020 to January 2022. Clin. Microbiol. Infect. 29(2), 225–232 (2023).
Bjerner, J., Børmer, O. P. & Nustad, K. The war on heterophilic antibody interference. Clin. Chem. 51(1), 9–11 (2005).
Ascoli, C. Could mutations of SARS-CoV-2 suppress diagnostic detection?. Nat. Biotechnol. 39(3), 274–275 (2021).
STANDARDIZATION, W.-E.C.O.B. Requests to initiate new WHO reference material projects for biologicals. WHO Reference Number: WHO/BS/2020.2404 2020 04.01.2021]; Available from: https://cdn.who.int/media/docs/default-source/biologicals/ecbs/bs-2020-2404-1st-is-sars-cov-2-ag-sj-ecbs-2020_bb193c28-1dab-4c44-a30c-94b674f2c2e4.pdf?sfvrsn=206d1405_9&download=true.
Bohn, M. K. et al. IFCC interim guidelines on rapid point-of-care antigen testing for SARS-CoV-2 detection in asymptomatic and symptomatic individuals. Clin. Chem. Lab. Med. 59(9), 1507–1515 (2021).
Eshghifar, N., Busheri, A., Shrestha, R. & Beqaj, S. Evaluation of analytical performance of seven rapid antigen detection kits for detection of SARS-CoV-2 virus. Int. J. Gen. Med. 14, 435–440 (2021).
Dinnes, J. et al. Rapid, point-of-care antigen and molecular-based tests for diagnosis of SARS-CoV-2 infection. Cochrane Database Syst. Rev. 3(3), Cd013705 (2021).
Puyskens, A. et al. Establishment of a specimen panel for the decentralised technical evaluation of the sensitivity of 31 rapid diagnostic tests for SARS-CoV-2 antigen, Germany, September 2020 to April 2021. Euro Surveill. https://doi.org/10.2807/1560-7917.ES.2021.26.44.2100442 (2021).
Cubas-Atienzar, A. I. et al. Limit of detection in different matrices of 19 commercially available rapid antigen tests for the detection of SARS-CoV-2. Sci. Rep. 11(1), 18313 (2021).
Zhou, H. et al. The impact of sample processing on the rapid antigen detection test for SARS-CoV-2: Virus inactivation, VTM selection, and sample preservation. Biosaf. Health 3, 238–243 (2021).
Funding
The work from SF and AK was partially funded by EURAMET in the project 18HLT03 SEPTIMET. The funder had no role in study design, data collection, and analysis, decision to publish, or preparation of the manuscript.
Author information
Authors and Affiliations
Contributions
H.Z., M.K., I.S., P.K., H.P.G., C.D., V.C., H.F.R., M.O., A.N. and J.M. conceptualized the EQA. scheme design, the sample preparation and the respective research goals. S.G. ensured the EQA. data collection via the INSTAND RV-Online platform for the participants and thus the data transfer to the evaluation software. H.P.G., U.D., D.N., H.F.R., M.O., A.N., J.M., A.P., P.M.V. and M.H.C. performed the suitability testing of the E.Q.A. material in the pre-study. J.H., D.O.S., E.B., S.C., S.F. and A.K. performed the quantification of SARS-CoV-2 viral load. Data analyses for this study, including the pre-study and the EQA results, and data interpretations were done by L.V., N.W., H.P.G., A.Z., A.S., U.D., J.H., D.O.S., S.C., M.K. and H.Z. Responsible for the visualization and data presentation were L.V., N.W., I.S., M.K. and H.Z. L.V., N.W., P.K., I.S., M.K. and H.Z. wrote the original draft. All authors reviewed and revised the original draft of this work.
Corresponding author
Ethics declarations
Competing interests
HZ declares that he was majority owner of GBD Gesellschaft fuer Biotechnologische Diagnostik mbH, Berlin, and owner and managing director of IQVD GmbH - Institut fuer Qualitaetssicherung, Berlin. HPG declares that he was minority owner of GBD Gesellschaft fuer Biotechnologische Diagnostik mbH, Berlin. HZ and HPG declare that they were managing directors of GBD Gesellschaft fuer Biotechnologische Diagnostik mbH, Berlin, during the study. This does not alter our adherence to Nature Scientific Reports policies on sharing data and materials. All other authors have declared that no competing interests exist. The work from SF and AK was partially funded by EURAMET in the project 18HLT03 SEPTIMET. The funder had no role in study design, data collection, and analysis, decision to publish, or preparation of the manuscript. All other authors have no conflict of interest.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Vierbaum, L., Wojtalewicz, N., Grunert, HP. et al. Results of German external quality assessment schemes for SARS-CoV-2 antigen detection. Sci Rep 13, 13206 (2023). https://doi.org/10.1038/s41598-023-40330-2
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-023-40330-2
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.