Published online Nov 13, 2023.
https://doi.org/10.3947/ic.2023.0077
Usefulness of the Rapid Antigen Test in Detecting SARS-CoV-2 for Infection Control in Hospitals
,1,*
Jiwon Jung,1,2,*
Jae Lim Jo,1
Young-Ju Lim,1
Sun-Kyung Kim,1
Hyejin Yang,1
Soyeon Park,1
Eun Ok Kim,1
and Sung-Han Kim1,2
Abstract
Background
We aimed at evaluating the diagnostic performance of rapid antigen test (RAT) compared to polymerase chain reaction (PCR) for severe acute respiratory syndrome coronavirus 2 and the possible transmission of infection to close contacts from patients with negative RAT and positive PCR results.
Materials and Methods
Patients/guardians urgently requiring admission to the ward on the same day had been hospitalized with RAT-negative result before the PCR results were available. We performed an epidemiologic investigation of the close contacts of those with negative RAT but positive PCR results after hospitalization.
Results
A total of 4,237 RATs were performed from March to August 2022. When the PCR test was used as the reference, RAT had a sensitivity of 28.8% (17/59; 95% confidence interval [CI], 17.8 - 42.1), a specificity of 100% (4,220/4,220; 95% CI, 99.9 – 100.0), a positive predictive value of 100.0% (17/17; 95% CI, 100.0 - 100.0), and a negative predictive value of 99.0% (4,178/4,220; 95% CI, 99.3 - 99.8). The epidemiologic investigation revealed that among the 32 patients with negative RAT and subsequent positive PCR results after admission into multi-patient room, two (6.3%) showed secondary coronavirus disease 2019.
Conclusion
The secondary transmission rate from patients with negative RAT and positive PCR results was low. Our data suggest that RAT may be useful for rapid exclusion of high transmissible cases. However, further evaluation using whole genome sequencing is needed to determine the potential for transmissibility in cases showing a negative RAT but a positive PCR result.
Graphical Abstract
Introduction
Polymerase chain reaction (PCR) testing is the gold standard for diagnosis of coronavirus disease 2019 (COVID-19). It is a highly sensitive and specific test. However, it has slow turnaround time, and can remain positive for up to 90 days after testing positive, as severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) viral RNA may continue to shed [1]. Therefore, it cannot differentiate between acute infection and a persistent positive result without infectivity, leading to difficulties in infection control in hospitals when relying solely on PCR testing.
Rapid antigen test (RAT) is immunoassay that detect the presence of viral proteins and positive test result indicates current infection. Although it is less sensitive compared to PCR [2, 3], but has the advantage of faster test results [4]. Despite the widespread use of RAT, data are limited for its usefulness for infection control in hospitals. Patients with a negative RAT result but a subsequently positive PCR result could be a source of SARS-CoV-2 transmission and outbreak in the early course of infection. At our hospital, pre-admission PCR for SARS-CoV-2 was mandatory for patients and guardians. However, for patients and guardians who required hospitalization before pre-admission PCR results could be obtained, RAT and PCR were performed simultaneously. Subsequently, these patients were allowed hospitalization with negative RAT results. Thus, we had a unique opportunity to evaluate how many patients had discordant RAT and PCR results and whether or not these patients transmitted the virus to their contacts. This study was aimed at evaluating the usefulness of RAT as a screening tool in asymptomatic pre-hospitalization patients and secondary transmission risk in a ward from patients with negative RAT but positive PCR results.
Materials and methods
1. Setting
This retrospective study was performed in a 2,743-bed tertiary care hospital in Seoul, Korea, from March to August 2022. During the study period, the average number of confirmed cases in Korea per day for one week (per 100,000 population) was follows; 11.7 in January, 139.5 in February, 598.0 in March, 248.9 in April, 57.4 in May, 17.3 in June, 91.5 in July, and 210.0 in August. All patients and guardians at our hospital were screened for SARS-CoV-2 using real-time PCR at pre-admission (≤ 2 days before admission). From March 3, RAT and PCR were performed for patients and guardians who required prompt hospitalization. Those necessitating immediate hospitalization encompassed patients who had made prior reservation for admission and were suddenly assigned to rooms on the same day along with their guardians. This category also comprised individuals admitted to the general ward directly from the emergency room, as well as patients who transitioned from the intensive care unit to the general ward, along with their guardians. We performed RAT on asymptomatic patients/guardians. For symptomatic patients, hospitalization was postponed until 5 days after the onset of symptoms. In cases where urgent hospitalization was necessary for symptomatic patients, they were required to undergo PCR testing, and were placed in a shared patient room after receiving negative PCR results. Alternatively, they were isolated in a single-patient room if PCR results were not yet available.
Patients who tested positive in RAT were discharged if they did not have any urgent medical concerns; otherwise, they were placed within a single room. Guardians who tested positive RAT were not permitted to remain in the hospital. In cases where patients or guardians were admitted after receiving negative RAT results, but later received positive PCR results, we conducted an epidemiologic investigation to identify the roommates (both patients and guardians) who had stayed before the positive PCR results were reported. Exposed roommates were subjected to periodic SARS-CoV-2 PCR testing following their exposure or in case symptoms developed. We categorized individuals exposed based on the length of the time they shared a patient room with the index cases. Roommates who had spent more than 2 calendar day with the index cases were placed under quarantine for a period of 5 days, and PCR testing was conducted on the day of exposure and 5 day after exposure. Exposed roommates who had spent less than 2 calendar day were not subjected to quarantine, and PCR testing was carried out 5 days after exposure (Supplementary Table 1). A secondary case was defined as an additional confirmed COVID-19 cases occurring 1 to 7 days after exposure to the index case. We divided the subjects into 2 groups according to the RAT test result, as follows; Group A comprised the patients with positive RAT results, and Group B with negative RAT but positive PCR results.
Other infection control measures for preventing SARS-CoV-2 transmission in the wards were follows; universal masking was recommended, and aerosol generating procedure was conducted in a single room. Visitor access was restricted, except for registered single guardians. Symptom screening was performed for both patients and their guardian. Hospital’s heat, ventilation and air-conditioner (HVAC) system was configured to maintain an air change rate exceeding 6 per hour.
2. Ethics statement
The Institutional Review Board of the Asan Medical Center evaluated and approved the medical, scientific, and ethical aspects of the study protocol and waived the requirement of obtaining informed consent (IRB No.: 2023-0161).
3. RAT & PCR
A nasopharyngeal swab specimen was collected by healthcare workers and delivered to the microbiological laboratory. There, a technician conducted PCR (AllplexTM SARS-CoV-2 Assay, Seegene, Seoul, Korea) and RAT (Wonmed Covid-19 Ag, Wonmed, Gyeonggi, Korea). The RAT result was reported as positive or negative. The PCR result was reported as positive, inconclusive, or negative, with the cycle threshold (Ct) was reported in the case of a positive and inconclusive result. The cut-off of cycle threshold was 40. The test evaluated three SARS-CoV-2 genes (E, N, and RdRp gene), and the result was reported as positive only if all three genes were below the cut-off.
4. Statistical analysis
For evaluation of the usefulness of RAT, we compared the Ct value of the E gene between group A, and group B. We excluded patients with a history of COVID-19. In addition, to determine the diagnostic usefulness of RAT, we evaluated the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) when PCR, which served as the reference test. If the PCR result was inconclusive, it was categorized as part of the PCR-negative group. We identified the time difference between the RAT and PCR results. Continuous variables were analyzed using the Mann–Whitney U test, as appropriate. All tests of significance were two-tailed, and P-values <0.05 were considered to be statistically significant. Data were analyzed using R software version 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria) with the RStudio (version 2022.07.1+554) platform [5].
Results
A total of 4,237 RATs (mean 23 cases per day) were performed during the study period (Fig. 1). Of these, 0.4% (17/4,237, 17 guardians) were positive, with all of them showing positive PCR results (group A). RAT-negative but PCR-positive cases were 1.0% (42/4,237, 42 guardians; group B), while 1.2% (51/4,237) cases showed inconclusive PCR results.
Figure 1
Flowchart of the study.
PCR, polymerase chain reaction; RAT, rapid antigen test.
When PCR was used as the reference test, the sensitivity, specificity, PPV, and NPV of RAT were 28.8% (17/59; 95% confidence interval [CI], 17.8 - 42.1), 100% (4,220/4,220; 95% CI, 99.9 - 100.0), 100% (17/17; 95% CI, 100.0 - 100.0), and 99.0% (4,178/4,220; 95% CI, 98.8 - 99.3; Supplementary Table 2). The average time interval between obtaining RAT test results and receiving PCR test results was 7 h 38 min ± 3 h 57 min. The mean Ct value of E gene of group A was 19.4 ± 2.6, and that of group B was 31.0 ± 3.9 (P <0.001).
Of the 42 guardians in group B, 10 had not been admitted to the hospital before the PCR results were reported. The remaining 32 guardians resulted in 80 individuals being exposed (mean exposed individuals per index, 2.5 ± 1.9). Out of the 32 guardians, two (6.3%) had close contacts who were diagnosed with COVID-19; therefore, they could be presumed to be the index patient who transmitted the virus. Case #1 was of the guardian of a patient transferred from the intensive care unit to the ward on the same day, with Ct of 35.2. Case #1 did not have any symptom for COVID-19, and there was no follow-up test as she was a guardian who left the hospital due to a positive test. She stayed in the patient room for 10 h, and one of the eight exposed individuals had positive PCR results on day 3 after exposure (Fig. 2). However, this patient had a history of exposure from other patients (P1 and P2) with COVID-19 in the same room. Case #2 was of the guardian of a patient transferred from the dedicated COVID-19 ward to the general ward on the same day, with a Ct of 29.7. As the duration of stay in the ward was 4 h, we did not perform testing immediately for roommates. Of the five roommates, three exposed patients and one guardian had positive results 2–6 days after exposure. However, all were exposed to the symptomatic guardian with COVID-19 (E1) diagnosed the next day after case #2. E1 had stayed in the ward 1 day before case #2 entered the ward, and she underwent PCR because of cough with an unknown onset. The roommates of the E1 guardian underwent PCR testing on the same day that the positive PCR test result for E1 guardian was obtained, and this led to the discovery of 4 additional COVID-19 cases (Fig. 3).
Figure 2
Epidemiologic investigation of case #1. Case #1 was guardian. Cycle threshold, day of diagnosis, and spatiotemporal relationship were compared among cases. Exposed case of case #1 (E1) showed exposure histories from P1 and P2.
Ct, cycle threshold; COVID-19, coronavirus disease 2019; RAT, rapid antigen test; PCR, polymerase chain reaction.
Figure 3
Epidemiologic investigation of case #2. Case #2 was guardian. Cycle threshold, day of diagnosis, and spatiotemporal relationship were compared among cases. Of the five roommates, 4 subjects had positive results from two to six days after exposure. However, all were exposed to the symptomatic guardian with COVID-19 (E1) diagnosed the next day after case #2.
Ct, cycle threshold; COVID-19, coronavirus disease 2019; RAT, rapid antigen test; PCR, polymerase chain reaction.
Discussion
When analyzing the results of over 4,000 COVID-19 RATs and PCRs, the sensitivity of RAT was approximately 30% compared to that of PCR and showed high specificity, PPV, and NPV. Of the 32 patients with negative RAT but positive PCR (group B) who stayed in the ward, two (6.3%) may have been the source of infection to the roommates. In addition, of the 4,237 test results, 4,195 (99.0%) were concordant between RAT and PCR. These subjects were admitted to the hospital and received adequate medical care without any delay.
Two (6.3%) patients in group B may have been responsible for transmission to other roommates; exposed roommates had positive results in the follow-up PCR. However, possible secondary cases from case #1 had other exposure history (Fig. 1), and both case #2 and the other guardian (E1 in Fig. 2) may be the index cases, as E1 was symptomatic and had a longer exposure time with secondary cases other than case #2. Determining the index case based solely on epidemiologic investigation is limited [6]. We did not perform whole-genome sequencing, and the possibility of transmission because of other exposures could not be excluded. Therefore, the proportion of transmission from RAT-negative but PCR-positive cases in this study may have been overestimated. The proportion of transmissible cases with negative RAT and positive PCR results may be ≤6.3%.
The sensitivity of RAT ranges from 16.6% to 30.2% depending on the sample preparation and collection [7, 8]. In other meta-analyses, RAT showed pooled sensitivity and specificity of 0.76 (95% CI, 0.73 - 0.79) and 1.00 (95% CI, 1.00 - 1.00), respectively [9]. The pooled sensitivity decreased as Ct increased. For Ct <20 and >30, the pooled sensitivities were 1.00 (95% CI, 0.70 - 1.00) and 0.24 (95% CI, 0.16 - 0.33), respectively [10]. The area under the receiver operating characteristic curve considering the pooled positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio was 0.97 (95% CI, 0.96 - 0.98). In the present study, RAT showed sensitivity and NPV of 28.8% and 99.0%, respectively. Although RAT has low sensitivity, it is helpful when urgent hospitalization is required because of rapid availability of test results in case of the simultaneous use of PCR. In addition, because we analyzed the usefulness of RAT in the setting of simultaneous use of PCR, further evaluation will be needed when RAT is used alone.
This study has several limitations. First, although epidemiologic investigations were extensively performed by experienced infection control practitioners, multiple exposures from unknown index cases during Omicron variant surge in the community might have occurred. In the instance of case #2, another symptomatic guardian could have been the index responsible for the emergence of 4 secondary cases. Nonetheless, the potential for high transmissibility in cases showing a negative RAT but a positive PCR result (such as case #2) cannot be entirely ruled out, as whole genome sequencing was not carried out. As a result, we cannot definitively establish the relationship of transmission dynamics. Second, we performed concurrent RAT and PCR for asymptomatic subjects. For symptomatic patients, only PCR was performed. Therefore, our data is difficult to extrapolate for RAT with PCR analyses of symptomatic subjects. Third, we did not perform cell culture of the respiratory specimens of patients with negative RAT but positive PCR results. Fourth, as we did not compare the transmission risk between negative RAT cases and positive RAT cases, we could not draw a firm conclusion that the secondary transmission rate is low in negative RAT cases. In addition, even though the rates are low, transmission to high-risk individuals could have significant implications. Finally, as this study was performed during the Omicron surge period, and PPV and NPV can vary according to the incidence of COVID-19, PPV can be lower and NPV higher in lower incidence. However, even in lower incidence, RAT can still be useful as it has a high NPV. This limits our understanding of the theoretic difference in transmissibility between those with negative RAT but positive PCR results and those with positive RAT and PCR results. Despite these limitations, this study provided important information pertaining to the real-world experience of RAT-based screening of subjects with a high risk of transmission in terms of the low transmissibility of those with negative RAT but positive PCR results, with ≤6.3% of them transmitting the virus to the close contacts.
In conclusion, out of the 32 subjects with negative RAT but positive PCR results who were admitted the hospital, two (6.3%) possibly contributed to secondary transmission in the multi-patients’ room in the general ward. RAT exhibited high PPV and NPV compared to PCR. Our data suggest that RAT may be useful for rapid exclusion of highly transmissible cases until definite PCR results are available, making it a helpful tool for infection control.
SUPPLEMENTARY MATERIALS
COVID-19 exposure investigation guidelines in hospitalsSupplementary Table 1
Comparison of rapid antigen test and polymerase chain reaction resultsSupplementary Table 2
Funding:This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government. (MSIT) (RS-2023-00219002).
Conflict of Interest:No conflict of interest.
Author Contributions:
Conceptualization: SHK.
Data curation: JYL, JLJ, HY.
Formal analysis: JYL.
Funding acquisition: SHK.
Investigation: JYL.
Methodology: JYL.
Software: JYL.
Supervision: YJL, SKK, SP, EOK.
Visualization: JYL.
Writing - original draft: JYL, JJ.
Writing - review & editing: SHK.
References
-
Centers for Disease Control and Prevention (CDC). Overview of testing for SARS-CoV-2, the virus that causes COVID-19. [Accessed 4 September, 2023].Available at: https://www.cdc.gov/coronavirus/2019-
ncov/hcp/testing- overview.html.
-
-
Yamayoshi S, Sakai-Tagawa Y, Koga M, Akasaka O, Nakachi I, Koh H, Maeda K, Adachi E, Saito M, Nagai H, Ikeuchi K, Ogura T, Baba R, Fujita K, Fukui T, Ito F, Hattori SI, Yamamoto K, Nakamoto T, Furusawa Y, Yasuhara A, Ujie M, Yamada S, Ito M, Mitsuya H, Omagari N, Yotsuyanagi H, Iwatsuki-Horimoto K, Imai M, Kawaoka Y. Comparison of rapid antigen tests for COVID-19. Viruses 2020;12:1420.
-
-
Scheiblauer H, Filomena A, Nitsche A, Puyskens A, Corman VM, Drosten C, Zwirglmaier K, Lange C, Emmerich P, Müller M, Knauer O, Nübling CM. Comparative sensitivity evaluation for 122 CE-marked rapid diagnostic tests for SARS-CoV-2 antigen, Germany, September 2020 to April 2021. Euro Surveill 2021;26:2100441
-
-
The R Foundation. The R project for statistical computing. [Accessed 31 October 2022].Available at: https://www.R-
project.org/.
-
MeSH Terms
Figures
1 / 3
ORCID IDs
Funding Information
-
National Research Foundation of Korea
RS-2023-00219002
