Diagnostic performance of oral swab specimen for SARS-CoV-2 detection with rapid point-of-care lateral flow antigen test

We evaluated the performance of oral swab specimen both health-care worker (HCW) collected and self-collected for severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) detection with rapid antigen test (RAT) as compared to reverse transcriptase polymerase chain reaction (RT-PCR). Of the 529 participants enrolled, 121 (22.8%) were RT-PCR positive. Among the RT-PCR positives, 62 (51.2%) were RAT positive using oral swab. When compared with RT-PCR, RAT with oral swab had sensitivity and specificity of 63.3 and 96.8% respectively among symptomatic individuals. No statistically significant difference was observed in RAT positivity with HCW collection and self-collection, p = 0.606. Ct values were significantly lower in RT-PCR and RAT positive samples (ORF gene: 18.85 ± 4.36; E gene: 18.72 ± 4.84) as compared to RT-PCR positive and RAT negative samples (ORF gene: 26.98 ± 7.09; E gene: 26.97 ± 7.07), p < 0.0001. Our study demonstrated moderate sensitivity of RAT with oral swab in symptomatic individuals. Oral swab was the preferred sampling by almost all participants in terms of convenience and comfort as compared to nasopharyngeal swab. Oral swabs have utility for SARS-CoV-2 antigen detection among symptomatic individuals residing in remote rural areas and can serve as an initial screening tool during COVID-19 spikes when cases rise exponentially and laboratory capacities for RT-PCR testing become overwhelmed.


Study phases and sample collection.
The study was conducted from 19th June to 1st July 2021 in two phases. In Phase I, oral swabs were collected by HCW (trained dentists collected samples at the OBH COVID-19 facility), while in Phase II, oral swabs were self-collected by the participants; flocked nylon swabs were used. The workflow of specimen collection and testing is shown in Fig. 1. In the first phase, HCWs rubbed the tip of the swab on the dorsal and ventral surface of the tongue and the buccal mucosal surface of each cheek for 10 s. The swab was put in the buffer tube provided with the RAT kit. In the second phase, participants were provided with pictorial pamphlets illustrating the procedure for oral swab collection. Each participant self-collected the oral swab and put in the buffer tube provided with the RAT kit.
In both phases of the present investigation, HCWs collected two NP swab specimens from each participant, one for RAT and the other for RT-PCR (along with an oropharyngeal swab in 3 ml viral transport medium) as per standard procedure which were transported to ICMR-NARI COVID-19 laboratory for RT-PCR testing 14 . Sample processing. RAT was performed on-site directly after sampling. RAT used in the study was Coviself, Pathocatch (Mylab Discovery Solutions Pvt. Ltd., Maharashtra, India), a qualitative immunochromatographic test for detection of SARS-CoV-2 specific antigens. This kit is validated by ICMR for use with NP and nasal swabs. The test procedure followed in the present investigation with oral swab was as per manufacturer's instructions. The visual read-outs were available within 20 min of the test.
For SARS-CoV-2 RT-PCR, viral RNA was isolated from the VTM using the MDS Viral RNA Extraction kit (MetaDesign Solutions, Gurgaon, India) and tested for SARS-CoV-2 with the Covidsure Multiplex RT-PCR kit (Trivitron Healthcare Labsystems Diagnostics, Chennai, India) on the CFX96 Real-Time Detection System (Bio-Rad, Hercules, CA, USA). The kit targets the E and ORF genes of SARS-CoV-2 and uses RPP30 human gene as internal control. The test was considered positive if cycle threshold (Ct) value was less than 35 for the genes tested.
Staff performing RAT were blinded to the results of RT-PCR tests and vice versa.
Statistical analysis. Descriptive statistics are presented as means for continuous and as proportions for categorical variables. The independent t-test or Mann-Whitney test was applied to continuous variables and Fisher's exact test to the categorical variables for comparisons. RAT results of oral swab were compared to RT-PCR and sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) with 95% confidence intervals (95% CI) were calculated. Statistical analysis was done using GraphPad statistical software   (Fig. 2). We observed a trend of decreasing RAT positivity with increasing Ct values was observed, p < 0.0001.
RAT positivity with oral swab by duration of symptoms. RAT positivity by duration of symptoms (in days) with self-collection and HCW collected oral swabs is presented in Fig. 3. COVID-19 duration of ≤ 5 days corresponded to RAT positivity rate of 31.3%, while COVID-19 duration of more than 5 days had RAT positivity rate of 21.1%.

Feedback from participants and HCW with regards to oral swab sampling.
Almost all participants (524/529, 99.1%) chose oral swab as the most convenient and comfortable sample for SARS-CoV-2 testing as compared to NP sampling. We collected feedback from the HCW involved with oral swab collection. HCW opined that patients were more co-operative during oral swab collection compared to NP swab. There was no coughing and sneezing during the sample collection procedure. During the phase 2 of the study, the participants were willing for self-collection and followed the instructions in the pamphlet carefully. The drawback with regards to oral RAT testing as told by HCW were that, while performing the test the material from the swab at times settled at bottom of the buffer tube leading to delay in flow and that the readings have to be taken meticulously as some positive results appeared as light bands.

Discussion
Easy sampling alternatives will enhance the uptake of rapid antigen testing and accelerate COVID-19 detection. Oral swabs have been previously used for detection of infectious etiologies [15][16][17] and disease biomarkers 18,19 , though, their performance for SARS-CoV-2 detection with RAT has not been reported. In the present study we evaluated the performance of oral swab specimen for detection of SARS-CoV-2 antigen. Similar results When compared with RT-PCR, RAT with oral swab sampling had a moderate sensitivity of 63.3% among symptomatic individuals, which was reduced to 28.6% among asymptomatic individuals. The sensitivity of RAT is influenced by the viral load in the specimen as evidenced by decrease in RAT positivity with increasing Ct values. RAT performed with oral swab detected 71.3% cases with Ct values < 25 that are considered to be highly infectious. However, our data also revealed that some individuals with low Ct values were missed by RAT. Hence, RAT with oral swab sampling cannot be a reliable replacement for SARS-CoV-2 detection.
The performance of RAT varies by number of factors, primarily, the clinical status of the patient (symptomatic versus asymptomatic), the RAT kit used, the specimen type used and time of testing (fresh versus stored specimen). NP swab is the recommended specimen for most RAT kits. A systematic review and meta-analysis conducted for the accuracy of SARS-CoV-2 rapid antigen test kits reported a pooled sensitivity of 76.3% with NP swabs 23 . However, NP swab collection is invasive, limited by need of a trained HCW for sample collection and poses risk of infection transmission to HCW. Among the alternative samplings, the sensitivity reported with anterior nasal/mid-turbinate swabs (75.5%) is comparable with NP swabs, while comparatively lower sensitivities with oropharyngeal (53.1%) and saliva (37.9%) specimens are reported 23 . Currently there is no literature with regards to use of oral swab for SARS-CoV-2 antigen detection.
NP swabs will remain as favored approach for testing with RAT. The oral swab with its ability to detect majority of the infectious cases, will have utility for testing among symptomatic individuals, especially for those residing in remote rural areas without the availability of trained HCWs and RT-PCR testing facilities. In such settings, conducting RAT test with self-administered oral swab will have the potential to enable detection and timely isolation of SARS-CoV-2 infected individuals in order to forestall local spread of infection. Oral swab also has the potential to serve as an initial screening tool in situations of COVID-19 surge when the cases rise exponentially and testing with RT-PCR becomes overwhelming for the laboratories. However, development of additional strategies to minimize false negative results will have to be laid down in these settings. These can include referral of individuals with symptoms compatible with COVID-19 but a negative antigen result for RT-PCR testing.
RAT with oral swab sampling had an overall specificity of 97.3%. We acknowledge that the RAT kit used in the study was designed for testing with NP and nasal swabs. As per the experience of HCWs, the viscous material from oral swab at times settled at the bottom of the buffer tube and false positive results were particularly observed in these samples. Similar observation was made by Chaimayo et al. while evaluating the rapid Standard Q COVID Ag test with respiratory specimens 24 .
Our study had strengths and limitations. It was conducted in a real world setting and included both symptomatic and asymptomatic individuals. We demonstrated participants' acceptability for oral swab and the feasibility of its use for RAT testing. However, we did not include pediatric age-group and individuals with nasal anomalies, who we feel could have benefited more with this sampling approach, warranting further studies in these population groups and using different RAT kits. One time antigen test may miss a few positive cases, while serial testing will identify individuals who are likely to become infectious in the following few days due to viral proliferation and viral loads go high enough to be detected by RAT, specifically among asymptomatic individuals 25,26 . Thus, future studies should incorporate a component of serial testing to see whether it can compensate for the lower sensitivity observed.
To conclude, our study demonstrated moderate sensitivity of oral swab specimen for SARS-CoV-2 detection antigen detection in symptomatic individuals. Oral swab, was the preferred sampling by almost all participants in terms of convenience and comfort as compared to NP swab. Oral swabs will have utility for SARS-CoV-2 antigen detection among symptomatic individuals residing in remote rural areas and can also serve as an initial