Performance of a novel diagnostic assay for rapid SARS-CoV-2 antigen detection in nasopharynx samples

To the Editor,

Among the laboratory testing methods developed for identifying patients with acute infection due to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)—the aetiological agent of coronavirus disease 2019 (COVID-19)—viral RNA amplification using real-time PCR (RT-PCR) is to date the standard method in many clinical virology laboratories [1]. However, RT-PCR-based assays are labour intensive and, when not completely automated, take hours to yield results. Conversely, rapid antigen detection assays—intrinsically less laborious and requiring a few minutes to results—have the potential to satisfy the pressing demand for an early SARS-CoV-2 infection diagnosis [[2], [3], [4]].

Here, we evaluated the performance of the STANDARD F COVID-19 Ag fluorescent immunoassay (FIA) (SD Biosensor, Suwon, South Korea), an assay detecting SARS-CoV-2 nucleoprotein antigen, on nasopharynx swab samples. It consists of a test device on which a pre-extracted sample is allowed to react with a monoclonal anti-SARS-CoV-2 antibody and, after a 30-min incubation, a Standard FX 2400 analyser reads the intensity of fluorescence following antibody–antigen complex formation. Initially, we determined the limit of detection (LOD) of the STANDARD F COVID-19 Ag FIA by analysing replicates of a dilution series containing Vero E6 cell-cultured SARS-CoV-2 (INMI-1 strain) spiked in RT-PCR-negative nasopharynx swab samples at a 1.0 × 10³ 50% tissue culture infective dose (TCID₅₀)/mL (4 × 10⁶ RNA copies/mL) to 62.5 TCID₅₀/mL (2.5 × 10⁵ RNA copies/mL) concentration range. The LOD was 5 × 10² TCID₅₀/mL (2 × 10⁶ RNA copies/mL) at 95% detection probability (Fig. S1), and the value moderately matched the LOD (1.2 × 10² TCID₅₀/mL) claimed by the SD Biosensor manufacturer. Then, we tested 359 samples (104 positive and 255 negative), previously characterized with the Altona Diagnostics RealStar® SARS-CoV-2 RT-PCR, the Seegene Allplex™ 2019-nCoV, the DiaSorin Simplexa™ COVID-19 Direct or the Roche Diagnostics Cobas® SARS-CoV-2 test, in virology laboratories from two COVID-19 reference hospitals in Rome. Tests were performed within 24 hr after collection on samples kept at 4°C until testing, according to the SD Biosensor manufacturer's recommendations. The median RT-PCR cycle threshold (Ct) value of positive samples was 31.3 (range, 15.3–39.7). Secondly, we compared the results obtained with the STANDARD F COVID-19 Ag FIA with those from any of mentioned RT-PCR assays (Table 1 ). Among negative samples, the STANDARD F COVID-19 Ag FIA detected 251 of 255 samples as negative, resulting in a negative per cent agreement of 98.4% (95% confidence interval (CI), 96.0–99.6). Repeat STANDARD F COVID-19 Ag FIA testing was performed on four false-positive samples, and again three (all with blood contamination) were positive. Among positive samples, the STANDARD F COVID-19 Ag FIA detected only 49 of 104 samples, resulting in a positive per cent agreement of 47.1% (95% CI, 37.1–57.1). According to RT-PCR Ct values, the sample per cent positivity of the STANDARD F COVID-19 Ag FIA ranged from 100.0% (5/5; Ct, <18), 93.8% (15/16; Ct, ≥18–<25), 42.0% (23/55; Ct, ≥25–<35) to 21.0% (6/28; Ct, ≥35). Assuming samples as from individuals diagnosed (n = 104) or not diagnosed (n = 255) with COVID-19, positive and negative predictive values were 92.5% (49/53; 95% CI, 81.8–97.9) and 82.0% (251/306; 95% CI, 77.3–86.2), respectively. At 10% SARS-CoV-2 infection prevalence, anticipated positive and negative predictive values were 76.9% (95% CI, 55.3–90.0) and 94.4% (95% CI, 93.3–95.3), respectively.

Table 1.

Agreement between the STANDARD F COVID-19 Ag FIA and the RT-PCR assay results for 359 nasopharynx samples tested

	Detected	Not detected	Positive per cent agreement (95% CI)a	Negative per cent agreement (95% CI)b
Positive samples (n = 104)
STANDARD F COVID-19 Ag FIA	49	55	47.1 (37.1–57.1)	—
RT-PCR assay	104	0	NA	—
Negative samples (n = 255)
STANDARD F COVID-19 Ag FIA	4	251	—	98.4 (96.0–99.6)
RT-PCR assay	0	255	—	NA

CI, confidence interval; NA, not applicable.

^aThis value corresponds to the assay's sensitivity assuming all 104 samples being from individuals diagnosed with COVID-19.

^bThis value corresponds to the assay's specificity assuming all 255 samples being from individuals not diagnosed with COVID-19.

Our study shows that the STANDARD F COVID-19 Ag FIA had a good specificity for SARS-CoV-2 detection in nasopharynx swab samples but had a good sensitivity only for samples with Ct values lower than 25 (corresponding to higher viral loads). Thus, we believe that the STANDARD F COVID-19 Ag FIA might be reliably used in the early phases of acute SARS-CoV-2 infection, e.g. within the first days after infection when Ct values are likely to still be below 25. This would be consistent with the viral load kinetics within the first days of SARS-CoV-2 infection, showing that nasopharynx swab samples obtained on day 7 may be persistently positive (Ct values, 23–24) for SARS-CoV-2 [5]. However, considering the current epidemiological scenario, a non-negligible proportion of symptomatic or, most commonly, asymptomatic patients, whose nasopharynx swab samples display Ct values of ≥25–<35 or ≥35, might be negative with STANDARD F COVID-19 Ag FIA (or similar) assays. This scenario would also encompass “new” patients who begin their SARS-CoV-2 infection course with a low viral load (resulting in Ct values of ≥35). In the light of these observations, it is presently difficult to envisage the correct, fruitful and safe use of these assays unless they are integrated in laboratory diagnostic algorithms based on both molecular and serological testing for SARS-CoV-2 infection.

Author contributions

Conceptualization, F.M.L., G.M., M.S., P.C. and M.R.C.; formal analysis, F.M.L.., G.M. and E.L.; investigation, I.P., S.M. and F.C.; resources, M.S. and M.R.C.; data curation, M.L.S., G.S. and L.B.; writing—original draft preparation, B.P.; writing—review and editing, E.L., F.C. and P.C.; supervision, M.S., M.R.C. and B.P. All authors have read and agreed to the published version of the manuscript.

Transparency declaration

This study was supported by funds to the Istituto Nazionale per le Malattie Infettive (INMI) Lazzaro Spallanzani IRCCS, Rome, Italy, from the Ministero della Salute (Ricerca Corrente, linea 1; COVID-2020-12371817), the European Commission – Horizon 2020 (EU project 101003544 – CoNVat; EU project 101003551 – EXSCALATE4CoV; EU project 12371675 – EXCALATE4CoV; EU project 101005075 – KRONO) and the European Virus Archive – GLOBAL (grants no. 653316 and no. 871029). All authors report no conflicts of interest relevant to this letter.

Editor: F. Allerberger

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Fig. S1

Determination of limit of detection (LOD) of the STANDARD F COVID-19 Ag FIA. Probit analysis revealed a LOD of 5 × 10² TCID₅₀/mL (2 × 10⁶ RNA copies/mL) at 95% detection probability.