The Sensitivity of Rapid Tests for SARS-CoV-2 Antigen: Association With Clinical Manifestations, Viral Load, and Sublines

Association with clinical manifestations, viral load, and sublines

Rapid diagnostic tests for SARS-CoV-2 antigen (RDTs) are an established point-of-care solution for diagnosis, screening and self-testing. On the other hand, the use of the reference standard in SARS-CoV-2 testing—the reverse transcription-quantitative polymerase chain reaction (RT-qPCR)—is limited by diagnostic capacities and infrastructural requirements. However, since the establishment of RDTs, the SARS-CoV-2 wild-type strain has been replaced by SARS-CoV-2 variants of concern (VOC). For this reason, the primary focus of our study is on the impact of the omicron sub-lineage BA.4/5 on RDT performance (1).

Long-term evaluation of the performance of SARS-CoV-2 antigen rapid diagnostic tests in the screening setting

In this prospective study, we compared RDTs (NADAL, PANBIO or MEDsan) and RT-qPCR for SARS-CoV-2 carried out in parallel in a tertiary care hospital between 12 November 2020 and 30 September 2022. 602 multiple tests on the same day, 32 invalid RDT results and 70 patients with recently ended isolation for SARS-CoV-2 infection were excluded. Altogether 54 740 parallel RDT/RT-qPCR oropharyngeal swabs of 38 373 study participants aged 18 years or older were included. One of the three RDTs was used in each instance of parallel testing being performed. The medical history for symptoms of COVID-19 was obtained from the hospital information system (HIS), but symptoms were only documented in cases of positive PCR results. Here, a distinction was made between typical symptoms (fever, respiratory symptoms), atypical symptoms (e.g. diarrhea, vomiting, deterioration in general condition) and asymptomatic. In 164 test tandems, no symptom history was available. Confidence intervals were calculated using the Wilson-Brown method (test performance) or Baptista-Pike method (Odds Ratio). The study protocol, the RT-qPCR methods and the determination of VOCs have been described in detail and published in pilot studies, including preliminary results from a total of 35 479 RDTs included (2, 3). ChatGPT (OpenAI, San Francisco, California, USA) was used for language optimization of the manuscript. Subsequently, the authors reviewed and edited the text.

Overall, the results showed an RDT sensitivity of 36.4% (433/1189; 95% confidence interval: [33.7; 39.2]) and an RDT specificity of 99.7% (53 375/53 551; [99.6; 99.7]) compared to RT-qPCR. The positive predictive value was 71.1% (433/609; [67.4; 74.6]) and the negative predictive value 98.6% (53 375/54 131; [98.5; 98.7]).

Factors associated with rapid test sensitivity

Univariate analyses showed higher RDT sensitivity when testing typical and atypical symptomatic individuals compared to asymptomatic individuals (Table). The viral load in individuals infected with omicron BA.4/5 VOC was higher compared to wild type (p = 0.0004) and omicron BA.1/2 (p = 0.002).

Table

The sensitivity of rapid diagnostic tests (RDTs) compared to RT-qPCR, grouped by potential influencing factors

In the LASSO regression analysis, viral load (odds ratio [OR]: 2.26; [1.97; 2.57]; p<0.001) and symptoms (fever, respiratory symptoms; OR: 2.49; [1.67; 3.72]; p<0.001) were the primary factors correlating with RDT performance. The factors age (p = 0.45), atypical COVID-19 symptoms (p = 0.09), infection with the omicron sub-lineages BA.1/2 VOC (p = 0.06) and Omicron BA.4/5 VOC (p = 0.36) showed no significant correlation with RDT performance (Table).

It was shown that both symptoms and viral load have a decisive impact on RDT sensitivity. For the Omicron VOC, especially the sublinages BA. 1/2, decreased RDT sensitivity was described earlier (1, 2). This can be explained by the fact that patients infected with Omicron VOC display less symptoms. One potential explanation is a smaller ratio of nucleocapsid protein to SARS-CoV-2 RNA in patients with milder symptoms, as it is the case with infections with Omicron VOC, for example.

Our study has a number of limitations: Due to the use of the RDTs as a screening method and testing being performed independent of symptoms, there is a low prevalence in the overall study cohort of only 2.2%. Another factor contributing to the comparatively low sensitivity of 36.4% is that many of the individuals tested were asymptomatic, some even in a late stage of the infection. This sensitivity is comparable to the findings of other studies conducted in a screening setting (2, 4).

However, these data provide a realistic scenario for the use of RDTs as a screening method and allow to investigate the influence of symptoms on sensitivity. VOC testing was only performed on a routine basis between January 2021 and January 2022. For this reason, a relevant proportion of wild-type SARS-CoV-2 and Omicron VOC samples were only categorized based on epidemiological considerations. Furthermore, the majority of samples are attributable to Omicron VOC, primarily due to the general increase in infected persons at that time. However, no molecular biological differentiation between BA.1/2 and BA.4/5 was made. Despite the fact that RT-qPCR is globally recognized as the reference method for SARS-CoV-2 testing, this method itself does not offer 100% sensitivity and specificity (5). The study design excluded potential repetitions in the case of the 32 invalid RDT results. One limitation is the use of multiple tests on the same patient, as this could have led to an underestimation of the confidence intervals.

Rapid diagnostic tests of limited utility for SARS-CoV-2 screening

In conclusion, RDTs are a reliable diagnostic tool for the rapid detection of symptomatic individuals with high SARS-CoV-2 viral loads. Thus, RDTs can be expected to provide reliable results shortly after the first symptoms appear at the onset of the disease when the viral load is high. However, RDTs show significant limitations when used as a screening method, particularly in asymptomatic patients with SARS-CoV-2. The reduced sensitivity of rapid diagnostic tests in patients infected with a SARS-CoV-2 Omicron VOC appears to be explained by their less symptomatic course rather than by structural changes in the virus (1, 2).

Kerstin Knies*¹, Isabell Wagenhäuser*¹, Daniela Hofmann, Vera Rauschenberger, Michael Eisenmann, Julia Reusch, Sven Flemming, Oliver Andres, Nils Petri,
Max S. Topp, Michael Papsdorf, Miriam McDonogh, Raoul Verma-Führing, Agmal Scherzad, Daniel Zeller, Hartmut Böhm, Anja Gesierich, Anna Katharina Seitz,
Michael Kiderlen, Micha Gawlik, Regina Taurines, Thomas Wurmb, Ralf-Ingo
Ernestus, Johannes Forster, Dirk Weismann, Benedikt Weißbrich, Johannes Liese, Ulrich Vogel^†, Oliver Kurzai, Lars Dölken, Alexander Gabel*², Manuel Krone*²

Institute for Virology and Immunobiology, University of Würzburg, Würzburg, Germany (Knies, Hofmann, Weißbrich, Dölken)

University Hospital Würzburg, Würzburg, Germany (Wagenhäuser, Rauschenberger, Eisenmann, Reusch, Flemming, Andres, Petri, Topp, Papsdorf, McDonogh, Verma-Führing, Scherzad, Zeller, Böhm, Gesierich, Seitz, Kiderlen, Gawlik, Taurines, Wurmb, Ernestus, Weismann, Liese, Vogel, Gabel, Krone), [e-mail]

Institute for Hygiene and Microbiology, University of Würzburg, Würzburg, Germany (Rauschenberger, Forster, Vogel, Kurzai, Krone)

*¹ The authors share first authorship.

*² The authors share last authorship.

Financial support
The German Federal Ministry of Education and Research (BMBF; Network University Medicine on COVID-19, B-FAST; grant no. 01KX2021), Free State of Bavaria, Bavarian State Ministry of Health and Care (Bay-VOC)

Conflict of interest statement
MaKr has received consulting fees from Abbott, GSK and Pfizer.

The remaining authors declare that no conflict of interest exists.

Manuscript received on 4 May 2023, revised version accepted on 2 August 2023.

Translated from the original German by Ralf Thoene, MD.

Cite this as:
Knies K, Wagenhäuser I, Hofmann D, Rauschenberger V, Eisenmann M, Reusch J, Flemming S, Andres O, Petri N, Topp MS, Papsdorf M, McDonogh M, Verma-Führing R, Scherzad A, Zeller D, Böhm H, Gesierich A, Seitz AK, Kiderlen M, Gawlik M, Taurines R, Wurmb T, Ernestus RI, Forster J, Weismann D, Weißbrich B, Liese J, Vogel U, Kurzai O, Dölken L, Gabel A, Krone M: The sensitivity of rapid tests for SARS-CoV-2 antigen—association with clinical manifestations, viral load, and sublines. Dtsch Arztebl Int 2023; 120: 763–4. DOI: 10.3238/arztebl.m2023.0185

References

1. Osterman A, Badell I, Basara E, et al.: Impaired detection of omicron by SARS-CoV-2 rapid antigen tests. Med Microbiol Immunol 2022; 211:105–117

2. Wagenhäuser I, Knies K, Hofmann D, 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 2023; 29: 225–32

3. Wagenhäuser I, Knies K, Rauschenberger V, et al.: Clinical performance evaluation of SARS-CoV-2 rapid antigen testing in point of care usage in comparison to RT-qPCR. EBioMedicine 2021; 69: 103455

4. Dinnes J, Sharma P, Berhane S, et al.: Rapid, point-of-care antigen tests for diagnosis of SARS-CoV-2 infection. Cochrane Database Syst Rev 2022; 7: CD013705

5. Krüger S, Leskien M, Schuller P, et al.: Performance and feasibility of universal PCR admission screening for SARS-CoV-2 in a German tertiary care hospital. J Med Virol 2021; 93: 2890–8