Diagnostic evaluation of PanBio, and standard Q COVID-19 rapid antigen tests for the detection of SARS-CoV-2: a cross-sectional study from Nepal

Introduction: The diagnosis of COVID-19 infection was time-consuming and costly, contributing to rampant transmission, particularly in developing countries like Nepal. This study aimed to compare the diagnostic performance of two rapid antigen diagnostic tests (RDTs, PanBio and Standard Q) against real-time reverse transcriptase PCR (rRT-PCR). Methods: This retrospective cross-sectional study was conducted among 1171 suspected COVID-19 patients at a provincial hospital in Nepal. Each participant provided two nasopharyngeal swabs, one for RDT and the other for rRT-PCR, spanning a total duration of four months. Sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), positive predictive value (PPV), negative predictive value (NPV), and diagnostic accuracy of each RDT, as well as the combined antigen-RDT, were determined in reference to the rRT-PCR status. Results: The sensitivity and specificity of PanBio were 60.87% (95% CI: 55.84–65.74) and 98.43% (95% CI: 96.80–99.37), respectively, while that of Standard Q was 59.83% (95% CI: 53.16–66.23) and 96.15% (95% CI: 90.44–98.94), respectively. Positive and negative predictive values of PanBio were 97.14 and 74.20%, respectively, and that of Standard Q were 97.16 and 52.08%. The positive likelihood ratio was higher for PanBio (38.87, 95% CI: 18.56–81.41) than for Standard Q (15.55, 95% CI: 5.92–40.90). Meanwhile, the negative likelihood ratio was more than 0.40 for both RDTs. The accuracy for PanBio, Standard Q, and combined RDTs were 80.91% (95% CI: 78.08–83.52), 71.17% (95% CI: 65.98–75.98), and 78.14% (95% CI: 75.66–80.48), respectively. Conclusion: RDTs exhibited unsatisfactory sensitivity and positive predictive value, rendering them ineffective as a screening tool. Nonetheless, they demonstrated excellent specificity and negative predictive value. Hence, a negative test result almost entirely excludes the possibility of infection.


Introduction
Severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) is a highly contagious and causative pathogen of the COVID-19 pandemic, leading to significant morbidity and mortality worldwide.Despite the vaccination rollout, multiple waves of COVID-19 have occurred globally due to mutations in the SARS-CoV-2 strain [1] .Early detection and isolation of victims, along with regular standard precautions by all individuals, are crucial to curb the transmission of the virus [2] .
A nasopharyngeal swab is the preferred specimen and is recommended over an oropharyngeal swab in the diagnosis of COVID-19 [3] .There are two accepted techniques for detecting SARS-CoV-2 in the swab: viral genome detection by RT-PCR and antigen/antibody detection by immunochromatography-based rapid diagnostic tests (RDTs) [4] .
RT-PCR-based assays are considered the gold standard due to their high sensitivity and specificity, allowing detection even in asymptomatic individuals [5] .However, the fallacies of genomebased techniques include needing trained personnel, specialized laboratories, long result awaiting time (hours to days), and increased cost [6,7] .Although the accuracy and reliability of RDTs are lower than PCR's, they are widely used as they provide rapid results (within minutes), are inexpensive, simple to perform, and do not need specialized expertise (self-testing is possible) [8] .The rapidity of test results allows for early case detection with timely patient management and the institution of appropriate public health measures [9] .The importance of RDTs in early detection is further accentuated by the fact that the virus can spread from symptomatic as well as asymptomatic and presymptomatic carriers.
This study was done to compare the diagnostic performances of two rapid antigen diagnostic tests (RDTs, PanBio, and Standard Q) against real-time reverse transcriptase PCR (rRT-PCR).

Methodology
Setting, study design, and participants Using purposive sampling, a hospital-based cross-sectional study of retrospectively collected data from 21st April 2021 to 21st August 2021 was done from a government-run hospital in Nepal.The education about adequate sampling and the sample handling procedures of all included physicians and nurses was conducted before the start of the research and lasted for 7 days.We included all symptomatic patients suspected of SARS-CoV-2 infections presenting with one or more signs or symptoms of it: fever, cough, malaise, sore throat, myalgia, arthralgia, headache, diarrhea, nausea, vomiting, anosmia, ageusia, regardless of their age and sex.Asymptomatic patients or those with missing data in the hospital register book over this period were excluded from the analysis.The final sample size became 1171 after excluding 42 cases.A semi-structured questionnaire was collected to obtain age, sex, report RDTs and rRT-PCR, and cycle threshold (Ct) values of each E, N, and RdRp gene.The data was collected from the hospital register book.The work has been reported in line with the strengthening the reporting of cohort, cross-sectional, and case-control studies in surgery (STROCSS) 2021 criteria [10] .

Specimen collection and laboratory testing
Adequate personal protective equipment was used to collect the nasopharyngeal swabs and perform the RDTs.Two NPS were collected for each participant, the first for RDTs and the second for viral genome detection by rRT-PCR, a standard flocked swab placed in viral transport media (VTM).The RDT was performed with either Panbio M or Standard Q at the sample collection site.
The two antigen-based RDTs viz., Panbio (Abbott Rapid Diagnostics Jena GmbH, Jena, Germany, www.abbott.com/poct)and Standard Q (SD BIOSENSOR, distributed by Roche, Switzerland, www.sdbiosensor.com)were used with the materials provided in the kit and as recommendations of the manufacturers.Both assays were manually read, with two individuals reading the results separately.In case of discordant results, the two validators sought consensus.All visible bands were considered positive results.
All participants were tested with a second sample by rRT-PCR assay for SARS-CoV-2 (cobas1 SARS-CoV-2 Test, Cobas 6800) using NPS in 3 ml VTM (VTM, Noble Bio, Hwaseong-si, Gyeonggi-do).The target genes were E, N, and RdRP genes.A sample was considered 'SARS-CoV-2 positive' when the cycle threshold (Ct) values of all genes in its RNA were ≤ 35.If only one or two tested genes had Ct ≤ 35, the test results were interpreted as 'inconclusive' according to local guidelines.The sample was considered 'negative' when the Ct values of all the genes were > 35.

Data analysis
Data were entered into Microsoft Excel 2019 v16.0 (Microsoft), and appropriate commands were used for data cleaning.Data was analyzed using Statistical Packages for Social Sciences version 21 (IBM SPSS Corporation) and MedCalc for Windows version 12.3.0(MedCalc-Software).We expressed the categorical data as frequency and percentages and continuous data as mean SD.The χ 2 was used to test for group differences.The P-value of < 0.05 was taken as statistically significant.The diagnostic sensitivity of each RDT and for each gene at different Ct values were determined.Simulated positive predictive values (PPVs) and negative predictive values (NPVs) were calculated according to various prevalence rates of COVID-19.Similarly, the overall sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), positive predictive value (PPV), negative predictive value (NPV), and diagnostic accuracy of each RDT as well as combined antigen-RDT were determined in reference to rRT-PCR status.

Results
Out of 1171 participants in this study, females had a significant difference of higher odds of rRT-PCR positivity than males (P-value: 0.004).Similarly, out of 838 participants tested with PanBio RDT, females had a considerable difference in high odds of positivity than males (P-value: 0.042).However, there was no significant difference in positivity with Standard Q RDT across females and males (P-value: 0.552).
The average age of all the participants was 40.82 18.64 years.The participants of the age group 26-50 years had higher odds of rRT-PCR positivity than other groups, viz., those less than 26 and more than 50 years (P-value: 0.028).With PanBio RDT, there was a significant difference of high positivity in the age group 26-50 years compared to those less than 26 and more than 50 years (P-value: 0.011).However, when testing positive with Standard Q RDT, there was no significant difference across the three age groups (P-value: 0.851) (Table 1).
For the setting of 1% of prevalence, PPVs of PanBio and Standard Q for the RdRP gene were 13.92 and 5.81%, respectively.If the prevalence increased to 10%, The values of the assays were 64.01 and 40.43%, respectively.NPVs of both assays for each gene were higher than 99.00% for disease prevalence of 0.5-5.0%(Table 3).

Discussion
Our study showed higher detection rates among females with the rRT-PCR test and PanBio RDT but no differences with Standard Q RDT.In contrast, a study in Canada showed that the number of laboratory-confirmed cases of COVID-19 was consistently high among males of all ages, even though the proportion of men being tested was lower than females [11] .A study by Fortunato et al. [12] suggests that the detection rate was similar in both sexes, but the hospitalization rate was higher among males.A study by Ristic et al. [6] showed that sex had no significant relationship with the diagnostic performance of PCR and standard COVID antigen tests.Our study showed that the COVID-19 positivity rate was higher in the middle age group (26-50 years) in PanBio and rRT-PCR tests.A study by Setiadi et al. [13] showed that the COVID-19 positivity rate was highest among individuals older than 40 years and lowest among children younger than 5 years.The sensitivity of PanBio was slightly higher (60.87 vs 56.8%), while the sensitivity of Standard Q RDT was almost similar (59.83 vs. 60.4%)compared to a similar study conducted in Germany [7] .However, the sensitivities were markedly lower compared to a similar study conducted in Switzerland, which showed a sensitivity of 85.5 and 89% for PanBio and Standard Q, respectively [14] .There is a marked variation in the sensitivity of Standard Q with a reported sensitivity ranging from 17.5 to 98.73% [15,16] .The specificity of PanBio and Standard Q in our study were 98.43 and 96.15, respectively.Most studies have shown that the specificity of both these tests were above 99% [15,17,18] .The minimum sensitivity and specificity of COVID antigen tests as recommended by WHO are 80 and 97%, respectively [19] .In reference to the WHO standards, the sensitivity of our test kits is low, while the specificity is excellent.Our findings suggest that an individual with a positive result on RDT is almost always infected due to high specificity.At the same time, a genuine case can be missed due to relatively low sensitivity.
Previous studies and our study showed that the sensitivity of RDTs declines with increasing ct-values in PCR.This is a justifiable finding since ct values bear an inverse relationship with viral load and period of infectivity.A study in the Netherlands showed that all false-negative cases from antigen tests were seen in individuals with a lower viral load corresponding to a ct value of higher than 32 [20] .A large-scale study has demonstrated that the sensitivity of Standard Q was high (93.75%) in lower ct values (ct less than 25) and considerably low (29.41%) in higher ct values (ct ≥ 25) [21] .
RDTs were found to have excellent positive predictive values in our study, with a PPV of 97.14 and 97.16% in PanBio and Standard Q, respectively.This is according to the study by Lopera et al. [22] , where the PPV of antigen tests is close to 100% concerning RT-PCR.NPV in that study was 99.9%, implying that an individual with a negative antigen test result is almost noncontagious.This contrasts with our study, where NPV is markedly low (74.2 and 68.79% for PanBio and Standard Q, respectively).A study conducted among emergency department attendees in Korea showed that the PPV and NPV of standard COVID antigen test (Biosensor) were 98.1 and 96.4%, respectively [23] .The positive and negative predictive values of BioSensor were found to be 100 and 87.4% in a study conducted in Italy [24] .PPV and NPV of a test are properties dependent on the prevalence of the disease.The variation in the predictive values in different geographical regions could be explained by the variation in prevalence in these regions.
The ability of antigen tests to detect Sars-cov-2 infection is also affected by the number of days since the symptom onset, which needs to be considered in our study.RDTS are deemed reliable, especially during the first week of illness [25] .A study showed that the sensitivity of rapid antigen tests was 71.4% in patients with a duration of illness less than 5 days and 85.9% when the illness was more than 5 days [26] .PPVs and NPVs depend on the prevalence, with PPV being higher at higher prevalence and NPV being higher at lower prevalence.Our study has certain limitations.Firstly, we did not take into account the number of days since the symptom onset.Secondly, we conducted this study during the time of high prevalence setting, thus extrapolating the findings of our study to low prevalence time must be done with caution.We need further studies to determine the gender-based higher detection in females using the PanBio RDT, to assess the detection rate of COVID-19 concerning the number of days of symptom onset, and to compare diagnostic statistics with other available RDTs.

Figure 1 .
Figure 1.Sensitivity of both RDTs for each gene and Ct value out of all positive COVID-19 rRT-PCR cases (n = 620).
HIGHLIGHTS• Early detection of COVID-19 decreases the spread of disease and reduces disease burden.• Rapid Diagnostic tests (RDTs) are helpful in early diagnosis.• RDTs show high specificity and negative predictive values.(95% CI: 76.69-89.45),respectively, for specimens with Ct value <25.Similarly, the sensitivity of PanBio and Standard Q was 88.33% (95% CI: 83.58-92.11)and 86.82% (95% CI: 79.74-92.13),respectively, for specimens with Ct value <25 for the RdRp gene.The sensitivity of both RDTs across three categories of Ct values for E, N, and RdRp gene was comparable and in decremental order except for the sensitivity of the RdRp gene at Ct value 25-30, which is more significant for Standard Q than PanBio (Table

Table 1
Basic demographic characteristics and their diagnostic status with the rapid point-of-care SARS-CoV-2 antigen assays and real-time RT-PCR assay for SARS-CoV-2.

Table 2
Diagnostic sensitivity of the PanBio and the Standard Q COVID-19 Ag tests, the rapid point-of-care SARS-CoV-2 antigen assays compared to real-time RT-PCR assay for SARS-CoV-2 detection.

Table 3
Simulated positive and negative predictive values of the PanBio and the Standard Q COVID-19 Ag assays according to the disease prevalence for the SARS-CoV-2 positive nasopharyngeal swab specimens with cycle threshold value less than 25.