Head-to head comparison of anterior nares and nasopharyngeal swabs for SARS-CoV-2 antigen detection in a community drive-through test centre in the UK

SARS-CoV-2, COVID-19, Nasal swabs, nasopharyngeal swabs, antigen detection, 17 RDT, LFA, head-to-head comparison. 18 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.

19

To conduct a head-to-head diagnostic accuracy evaluation of professionally taken 20 anterior nares (AN) and nasopharyngeal (NP) swabs for SARS-CoV-2 antigen detection using 21 rapid diagnostic tests (Ag-RDT). 22

NP swabs for SARS -CoV-2 reverse transcription quantitative polymerase chain 23 reaction ( RT-qPCR) testing and paired AN and NP swabs for the antigen detection were 24 collected from symptomatic participants enrolled at a community drive-through COVID-19 25 test centre in Liverpool. Two Ag-RDT brands were evaluated: Sure-Status (PMC, India) and 26 Biocredit (RapiGEN, South Korea) . The visual read out of the Ag -RDT test band was 27 quantitative scored and the 50% and 95% limit of detection (LoD) of both Ag-RDT brands using 28 AN and NP swabs was calculated using a probabilistic logistic regression model. 29

A total of 604 participants were recruited of which 241 (40.3%) were SARS-CoV-2 30 positive by RT-qPCR. Sensitivity and specificity of AN swabs was equ ivalent to the obtained 31 with NP swabs : 83.2% (75.2 -89.4%) and 98.8% (96.5 -99.6%) utilising NP swabs and 84.0% 32 (76.2-90.1%) and 99.2% (97.0-99.8%) with AN swabs for Sure-Status and; 81.2% (73.1-87.7%) 33 and 99.0% (94.7-86.5%) with NP swabs and 79.5% (71.3 -86.3%) and 100% (96.5 -100%) with 34 AN swabs for Biocredit. The agreement of the AN and NP swabs was high for both brands with 35 an inter-rater relatability ( κ) of 0.918 and 0.833 for Sure-Status and Biocredit, respectively. 36 The overall 50% LoD and 95% LoD was 0.9-2.4 × 104 and 3.0-3.2 × 108 RNA copies/mL for NP 37 swabs and 0.3- 1.1 x 105 and 0.7-7.9 x 107 RNA copies/mL and for AN swabs with no significant 38 difference on LoD for any of the swabs types or test brands. Quantitative read-out of test line 39 intensity was more often higher when using NP swabs with significantly higher scores for both 40 Ag-RDT brands. 41 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint

the diagnostic accuracy of the two SARS-CoV-2 Ag-RDTs brands evaluated in this 42 study was equivalent using AN swabs than NP swabs. However, test line intensity was lower 43 when using AN swabs which could influence negatively the interpretation of the Ag -RDT 44

for lay users. Studies on Ag-RDT self-interpretation using AN and NP swabs are needed 45 to ensure accurate test use in the wider community. 46

word count: 345 47 48 49 50 51 52 53 54 55 56 57 58 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint

59 To meet the immense diagnostic demand of the COVID -19 pandemic, the development of 60 rapid diagnostic tests for the detection of SARS -CoV-2 antigens (Ag-RDTs) became a priority 61 [1]. Nasopharyngeal (NP) swabs are considered the standard of care for SARS-CoV-2 62 detection[2] and thus the majority of Ag-RDT kits are developed for NP swabs exclusively [1]. 63 However, the use of anterior nasal (AN) swabs has been increasing as a less invasive 64 alternative to promote access to testing in the community and facilitate mass testing 65 programmes particularly in the UK [3]. 66 For Ag-RDTs, studies on Ag-RDTs comparing sensitivity on AN swabs and NP swabs are very 67 limited, with only two reported studies performed on the same Ag-RDT brand, Standard-Q 68 (SD Biosensor, Inc., Korea) , one study on professional taken swabs [6] and another in self -69 taken [7]. Sensitivity obtained with AN swabs was comparable (although 3% to 5% lower) than 70 with NP swabs sensitivity but neither of the swab types fulfilled WHO target product profile 71 (TPP) standards in any of the two studies [8]. AN swabs are considered accurate and clinically 72 acceptable alternatives to NP swabs in outpatient settings for SARS -CoV-2 reverse 73 transcription polymerase chain reaction (RT -PCR) testing [4]. However, an in depth 74 metanalysis on SARS -CoV-2 RT-PCR testing found that anterior nares specimens were 12% -75 18% less sensitive than NP swabs [5]. 76 The aim of this study was to perform a head -to-head evaluation on two World Health 77 Organisation ( WHO) approved or under assessment for Emergency Use listing (WHO-EUL) 78 SARS-CoV-2 Ag-RDT brands marketed for AN and NP swabs : Sure-Status COVID-19 Antigen 79 Card Test (Premier Medical Corporation, India) and Biocredit COVID-19 Antigen Test 80 (RapiGEN, South Korea) respectively. This study is of particular interest in the UK as the use 81 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint of home Ag-RDTs on AN swabs has been integral to combatting the spread of COVID-19 during 82 the pandemic [3],as on the 1 st of April 2022 free national RT -PCR COVID-19 testing was 83 suspended, with the purchase of Ag -RDTs using AN-swabs online or in pharmacies the only 84 approach to access COVID-19 testing in a non-clinical setting. 85

86 Clinical evaluation 87 This was a prospective evaluation of consecutive participants enrolled at a community 88 National Health Service (NHS) drive-through COVID-19 test centre located at the Liverpool 89 John Lennon Airport. Two Ag-RDT brands were evaluated; Sure-Status COVID-19 Antigen Card 90 Test (Premier Medical Corporation India) and Biocredit COVID -19 Antigen Test (RapiGEN, 91 South Korea) referred as Sure -Status and Biocredit thereafter. The study progressed until at 92 least 100 Ag-RDT positives using AN swabs in line with WHOs requirements for evaluation of 93 alternative sample type [10]. 94 All adults over the age of 18 who attended the drive -through test centre with symptoms of 95 COVID-19 were asked to participate in the study. The symptoms included fever, cough, 96 shortness of breath, tight chest, chest pain, runny nose, sore throat, anosmia, ageusia, 97 headache, vomiting, abdominal pain, diarrhoea, confusion, rush, or tiredness. Participants 98 were recruited under the Facilitating Accelerated COVID-19 Diagnostics (FALCON) study using 99 verbal consent. Ethical approval was obtained from the National Research Ethics Service and 100 the Health Research Authority (IRAS ID:28422, clinical trial ID: NCT04408170). 101 Swabs were collected following the same process with the NP swab collected first in one 102 nostril and placed in Universal Transport Media (UTM) (Copan Diagnostics Inc, Italy) for the 103 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint

RT-qPCR test. This was followed by the collection of two swabs to evaluate the Ag-104 RDTs, first an NP swab in the other nostril and finally a AN swab in both nostrils following the 105 manufacturer’s instructions for use (IFU) . Samples were given a unique identification code 106 and transported within cooler bags to the Liverpool School of Tropical Medicine (LSTM) where 107 samples were processed in category level 3 (CL3) containment laboratory upon arrival. 108 Sure-Status and Biocredit Ag-RDTs were carried out following their instructions for use (IFU). 109 The protocol for both Ag-RDT was the same when using AN and NP swabs. Results were read 110 by two operators, blinded to one another and if a discrepant result occurred, a third operator 111 acted as a tiebreaker. The visual read out of the Ag-RDT test band was scored on a quantitative 112 scale from 1 (weak positive) - 10 (strong positive) . Ag-RDT results were classified as invalid 113 when the control line was absent. 114 RNA was extracted using the QIAamp® 96 Virus QIAcube ® HT kit (Qiagen, Germany) on the 115 QIAcube® (Qiagen, Germany) and screened using TaqPath COVID -19 (ThermoFisher, UK) on 116 the QuantStudio 5 TM thermocycler (ThermoFisher, UK) , an internal extraction control was 117 incorporated before the lysis stage, as recommend ed by the manufacturer . SARS-CoV-2 RT-118 qPCR result was considered (1) positive if any two of the three SARS-CoV-2 target genes (N 119 gene, ORF1ab and S gene) amplified with cycle threshold (Ct) ≤ 40 , (2) indeterminate if only 120 one SARS-CoV-2 gene amplified and (3) negative if the internal extraction control amplified 121 and the SARS -CoV-2 target genes did not . Samples with i nvalid RT -qPCR results (no 122 amplification of the internal extraction control) were re-extracted and re-run once. Viral loads 123 in UTM swabs were measured with a ten -fold serial dilution standard curve of quantified 124 specific in vitro-transcribed RNA using five replicates for each standard curve point [11]. 125 Statistical Analysis 126 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint Sensitivity, specificity, positive predicted value (PPV) and negative predictive values (NPV) 127 were calculated with 95% confidence intervals (CIs) by comparing the Ag -RDT results to the 128 RT-qPCR, as the reference standard. Sub-analyses of diagnostic performance were performed 129 by swab type (AN and NP), Ct -value ranges, onset of symptoms and vaccination status using 130 nonparametric statistics. The level of agreement between AN and NP swabs was determined 131 using Cohen's kappa (κ) [10]. The correlation between test line intensity and viral loads were 132 measured by Person correlation, coefficient ( rP) [12] and t o further analyse Ag-RDT 133 sensitivities, we used logistic regression, with RNA copy numbers of the RT-qPCR NP swab and 134 swab type (AN and NP) as independent variables and test outcomes as the dependent 135 variable, yielding detection probabilities for each viral load level . Statistical analyses were 136 performed using SPSS V.28.0, Epi Info V3.01 and R scripts. Statistical significance was set at P 137 < 0.05. 138

139 140 Participant demographics 141 A total of 60 4 participants were recruited for this study, 37 2 recruited between August and 142 October 2021 were enrolled for the Sure -Status Ag-RDT evaluation and 232 recruited 143 between December 2021 and March 2022 were enrolled for the Biocredit Ag-RDT evaluation. 144 Details of the demographics of the population of study are found in Table 1. Our study 145 population had a mean age of 43 years (range 18-81, interquartile range [IQR] 33.0-50.0), 348 146 (58%) were female and 566 were British ( 94%), with the remaining 36 participants being of 147 other ethnic groups ( n = 14), white background ( n = 9), Asian ( n = 8), mix white and black 148 backgrounds (n = 2) and Arab (n = 1). Three hundred and fourteen participants of the 372 149 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint enrolled for the Sure-Status evaluation (84.4%) and 217 participants of the 232 recruited for 150 Biocredit (93.5%) received complete SARS -CoV-2 vaccination (2 doses) . Additionally, 143 of 151 the participants enrolled from December 2021 (61.6%) for the Biocredit evaluation received 152 a third dose as part of the UK booster roll out [13]. All participants were symptomatic with a 153 median onset of symptoms of 2 days (IQR 1-3). The most common symptoms were cough 154 (387, 64.3%), sore throat (232, 38.5%), headache (203, 33.7%), fever (160, 26.6%), body aches 155 (80, 13.3%) and runny nose (80, 13.3%) (Table 1). 156 Overall, 241 participants (40.3%, CI95% 36.3-44.4%) were SARS-CoV-2 positive by RT-qPCR, 6 157 had indeterminate RT -qPCR results and the remaining 355 were negative. Participants with 158 indeterminate RT-qPCR results were excluded from further analysis. 159 RT-qPCR positivity was significantly higher (p<0.05) among the participants enrolled for the 160 Biocredit evaluation cohort (53.7%, CI95% 47-60.4%) during December 2021 and March 2022 161 which coincided with the Omicron wave in the UK [14] than among the participants enrolled 162 between August and October 2021 (32.1%, CI95% 27.4-37.1%) when Delta was the dominant 163 SARS-CoV-2 variant. 164 Diagnostic evaluation 165 Sure Status 166 The overall sensitivity and specificity for the Sure-Status Ag-RDT compared to RT -qPCR was 167 83.2% (CI95% 75.2-89.4%) and 98.8% (CI95% 96.5-99.6%) utilising NP swabs and 84.0% (CI95% 168 76.2-90.1%) and 99.2% (CI95% 97.0-99.8%) with AN swabs. For individuals with Cts < 25, the 169 sensitivity was 91.8% (CI95% 84.5-96.4%) and 93.8% (CI95% 87.2-97.7%) for NP and AN-swabs 170 respectively. Seven Ag-RDTs gave invalid results, one NP swab (0.03%) sample and six AN 171 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint swab samples (1.6%). Participants with invalid Ag -RDTs results were excluded from further 172 analysis. Four SARS-CoV-2 positive cases were detected by NP only (3.4%) and six cases were 173 detected by AN only (5.0%) but this discrepancy on sensitivity between swab types was not 174 significant (P = 0.43). The percentage of agreement of NP and AN swab using Sure-Status was 175 96.7% (95% CI 94.3-98.3%) and inter-rater reliability was almost perfect (κ = 0.918). Inter-176 rater reliability was strong for both NP (κ = 0.871) and AN (κ = 0.852) swabs when compared 177 to RT-qPCR. 178 Biocredit 179 For the Biocredit Ag -RDT the sensitivity and specificity were 81.2% (CI95%73.1-87.7%) and 180 99.0% ( CI95%94.7-86.5%) with NP swabs and 79.5% ( CI95%71.3-86.3%) and 100% 181 (CI95%96.5-100%) with AN sampling compared to RT-qPCR. Sensitivity was 92.2% (CI95%84.6-182 96.8%) and 95.5% ( CI95%89.0-98.8%) using NP and AN swabs among participants with Ct < 183 25. Ten SARS-CoV-2 positive cases were detected solely by NP (8.2%) and eight cases were 184 detected only by AN (6.6%) but no significance on sensitivity was observed between NP and 185 AN swabs for this brand of Ag-RDTs either (P = 0.43). The percentage of agreement of NP and 186 AN swab for Biocredit was 91.6% (95% CI 87.2-94.9%) and inter-rater reliability was strong (κ 187 = 0. 833). I nter-rater reliability was moderate for both NP ( κ = 0. 790) and AN ( κ = 0. 782) 188 sampling compared to RT -qPCR. Diagnostic accuracy for both Sure-Status and Biocredit is 189 displayed in Table 2. 190 Head to head comparison of Sure Status and Biocredit 191 rWe report nosignificant difference in the diagnostic accuracy among participants with 192 symptoms irrespective of days since onset, or vaccination status for all Ag-RDTs and swabbing 193 combination (all P values > 0.05). Both Biocredit and Sure -Status Ag-RDTs using both swab 194 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint types had better sensitivities on detecting SARS-CoV-2 antigens on individuals with Ct values 195 30 (P = 0.029 in NP and P = 0.047 in AN for Sure-Status and P = 0.018 and P = 0.001 196 for Biocredit). 197 The RNA copy numbers per mL (RNA copies/mL) of RT -PCR NP swab s was calculated and 198 statistically higher viral loads were obtained for the Sure -Status cohort than Biocredit (Fig 1) 199 measured by Kruskal–Wallis (P= 0.006). We determined the 50% and 95% limits of detection 200 (LoD) for both Ag-RDT and swab types based on a logistic regression model (Fig 2). For Sure-201 Status, the RNA copies/mL for 50% LoD and 95% LoD were 2.4 × 104 and 3.16 × 108 for NP 202 specimen and 3.4 x 10 4 and 7.94 x 10 7 for AN swabs. For Biocredit, the RNA copies/mL for 203 LoD50 and LoD95 were 9.12 × 103 and 3.02 × 108 for NP specimen and 1.12 x 105 and 6.76 x 204 106 for AN swabs. Although the LoD95 was better for AN swabs for both Ag-RDT brands (3.98 205 for Sure-Status and 44.67 for Biocredit), there was no statistical difference on LODs neither 206 by swab type and Ag-RDT brand (all P values > 0.05). 207 Quantitative read-out analysis 208 Quantitative read-out in paired positive AN and NP was more often higher for the NP (40 209 instances higher on NP and four higher on AN in Sure-Status; and 35 instances higher on NP 210 and 12 higher on AN in Biocredit) and gave significantly higher scores for both Ag-RDT, Sure-211 Status (P = 0.007) and Biocredit (P = 0.013) (Figure 1) measured by Kruskal–Wallis. 212 Additionally, test lines scores were analysed by RNA copies/mL and these had a positive 213 correlation. For Biocredit, strong correlation using AN swabs (rP = 0.727) but moderate using 214 NP swabs (r P = 0.591). For Sure-Status, both swab types had a moderate correlation to viral 215 loads (NP swab rP = 0.614 and AN swab rP = 0.661). 216

217 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint This is the first diagnostic clinical evaluation of Sure -Status Ag-RDT and results have shown a 218 satisfactory performance for both AN and NP swabs fulfilling the sensitivity (≥80%) and 219 specificity (≥97%) outlined in the target product profile (TPP) WHO standards [8]. For Biocredit 220 Ag-RDT, th ere are five studies to date that have evaluated the performance of NP swabs 221 reporting varied sensitivities from 52% to 85% [15]. In this study we report a higher sensitivity 222 (81.2% ,CI95%73.1-87.7%) and specificity (99.0% ,CI95%94.7-86.5%) of the Biocredit Ag-RDT 223 fulfilled the WHO standards using NP swab but underperformed in the sensitivity (79.5%, 224 CI95% 71.3-86.3%) criteria when using AN. 225

presented here demonstrate that AN swabs are equivalent to NP swabs for SARS-CoV-226 2 Ag-RDT testing giving comparable sensitivities, 50% LoD and 95 % LoD for both Ag -RDTs 227 brands evaluated here. Our results supports previous findings where AN and NP swabs were 228 compared for the Ag -RDT Standard-Q (SD Biosensor, Inc., Korea) in Lesotho [6], but we 229 reported a higher sensitivity compared to the 67.3% and 70.2% for AN and NP swabs 230 previously described [6]. Studies on RT -qPCR have found lower sensitivity using AN swabs 231 compared to NP swabs consistently [5]. However, the difference i n sensitivity was only 232 significant for patients with viral loads < 103 copies/mL [16] and this threshold is not relevant 233 to Ag-RDTs of which the limit of detection ranges between 10 4- 108 RNA copies/mL in swabs 234 [11]. 235 Quantitative assessment of the test line scores showed that test line intensity was 236 significantly higher on NP swabs than AN swabs. The line intensity is an important component 237 of home tes ting as studies have shown fainter lines are more difficult to interpret for a lay 238 person, likely due to lower signal intensity [17]. In a n user experience home based study , 239 77.1% of the cases that the participants interpreted wrongly as negative being positive, were 240 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint weak and moderate positives while only 22.9% were strong positives [17]. The significantly 241 lower intensity of the AN swab compared to NP swab is likely attributed to the differences of 242 SARS-CoV-2 viral loads in the respiratory tract. Studies have found lower viral loads on AN 243 swabs compared to NP swabs [16]. Statistical analysis supported this hypothesis wh ere a 244 positive correlation between viral loads and Ag -RDT test line scores was shown Further 245 implementation studies on Ag-RDT test results interpretation by patients or within a home 246 testing setting are urgently needed to drive self-testing to scale. 247 This study has several strengths, the use of standardised sampling methods, independent 248 blinded readers, robust statistical analysis, quantitative assessment of Ag-RDT test line results 249 and the evaluation of one approved WHO-EUL Ag-RDT test brands (Sure-Status) and under 250 review (Biocredit). Qualifying it to have high global public health relevance [18]. 251 The main limitation of this study is that the AN swabs were always taken last . The order of 252 sample collection could have negatively biased the results obtained for AN swabs caused by 253 a possible sample depletion. However, in the two studies that compared Ag -RDT using AN 254 swabs, the AN swab was collected first and our reported sensitivity and specificity for AN 255 swabs are greater than the previous studies [6,7]. Further, studies on RT -qPCR found lower 256 sensitivity using AN swabs compared to NP swabs [5], even when AN swabs were collected 257 first [15,18,19]. Thereby it is unlikely that the order of the swabs impacted sample availability 258 for AN and NP sampling. 259 In conclusion, this study demonstrates the sensitivity of two SARS-CoV-2 Ag-RDTs using AN-260 sampling are comparable to that of NP-sampling. AN-sampling can be performed with less 261 training, reduces patient discomfort, and enables scaling up of antigen testing strategies. 262 Test line intensity however is lower when using AN swabs which could influence negatively 263 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint the interpretation of the Ag-RDT results. Additional studies on Ag-RDTs using AN swabs on 264 self-interpretation by a lay person are needed and further education around how to 265 interpret a positive Ag-RDT to the wider community. 266

267 The work was funded by the Foundation of Innovative Diagnostics (FIND) 268

269 1 FIND. SARS-CoV-2 diagnostic pipeline. https://www.finddx.org/covid-19/pipeline/ 270 2 Center of Disease Control and Prevention. Interim Guidelines for Collecting, Handling, and 271 Testing Clinical Specimens for COVID-19. May 22. 2020. 272 3 GOV.UK. New campaign urges public to get tested twice a week. Published Online First: 273 2021.https://www.gov.uk/government/news/new-campaign-urges-public-to-get-tested-274 twice-a-week 275 4 Péré H, Péré H, Péré H, et al. Nasal swab sampling for SARS-CoV-2: A convenient alternative 276 in times of nasopharyngeal swab shortage. J. Clin. Microbiol. 2020. doi:10.1128/JCM.00721-277 20 278 5 Zhou Y, OLeary TJ. Relative sensitivity of anterior nares and nasopharyngeal swabs for initial 279 detection of SARS-CoV-2 in ambulatory patients: Rapid review and meta-Analysis. PLoS One. 280 2021. doi:10.1371/journal.pone.0254559 281 6 Labhardt, Niklaus D González Fernández L, Katende, Bulemba Muhairwe J, Bresser M, et al. 282 Head-to-head comparison of nasal and nasopharyngeal sampling using SARS-CoV-2 rapid 283 antigen testing in Lesotho. medRxiv Published Online First: 2022. 284 doi:10.1101/2021.12.29.21268505 285 7 Lindner AK, Nikolai O, Kausch F, et al. Head-to-head comparison of SARS-CoV-2 antigen-286 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint detecting rapid test with self-collected anterior nasal swab versus professional-collected 287 nasopharyngeal swab. Eur Respir J 288 2020;57:2003961.http://erj.ersjournals.com/lookup/doi/10.1183/13993003.03961-2020 289 8 (WHO) WHO, R&D Blue Print WH (HQ). Target product profiles for priority diagnostics to 290 support response to the COVID-19 pandemic v.1.0. Published Online First: 291 2020.https://www.who.int/publications/m/item/covid-19-target-product-profiles-for-292 priority-diagnostics-to-support-response-to-the-covid-19-pandemic-v.0.1 293 9 GOV.UK. People with a positive lateral flow test no longer required to take confirmatory PCR 294 test. Published Online First: 2021.https://www.gov.uk/government/news/people-with-a-295 positive-lateral-flow-test-no-longer-required-to-take-confirmatory-pcr-test 296 10 World Health Organization (WHO). Prequalification Teams D. Instructions and requirements 297 for Emergency Use Listing (EUL) Submission: In vitro diagnostics detecting SARS-CoV-2 nucleic 298 acid or antigen. 299 https://extranet.who.int/pqweb/sites/default/files/documents/220317_PQDx_347_Version6300 _NAT-Ag.pdf 301 11 Cubas-Atienzar, Ana I., Kontogianni K, Edwards T, et al. Limit of detection in different 302 matrices of 19 commercially available rapid antigen tests for the detection of SARS-CoV-2. Sci 303 Rep 2021;11:1–8. doi:https://doi.org/10.1038/s41598-021-97489-9 304 12 Schober P, Schwarte LA. Correlation coefficients: Appropriate use and interpretation. Anesth 305 Analg Published Online First: 2018. doi:10.1213/ANE.0000000000002864 306 13 GOV.UK. Vaccinations in the UK. Coronavirus UK. 307 2022.https://coronavirus.data.gov.uk/details/vaccinations 308 14 Mahase E. Covid-19: Is the UK heading for another omicron wave? BMJ 2022;376. 309 doi:https://doi.org/10.1136/bmj.o738 310 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint 15 Brümmer LE, Katzenschlager S, Gaeddert M, et al. Accuracy of novel antigen rapid diagnostics 311 for SARS-CoV-2: A living systematic review and meta-analysis. PLoS Med. 2021. 312 doi:10.1371/journal.pmed.1003735 313 16 Callahan C, Lee R, Lee G, et al. Nasal-Swab Testing Misses Patients with Low SARS-CoV-2 Viral 314 Loads. medRxiv 2020. 315 17 Jing M, Bond R, Robertson LJ, et al. User experience of home-based AbC-19 SARS-CoV-2 316 antibody rapid lateral flow immunoassay test. Sci Rep 2022;12:1173. doi:10.1038/s41598-317 022-05097-y 318 18 World Health Organization (WHO). Coronavirus disease (COVID-19) Pandemic — Emergency 319 Use Listing Procedure (EUL) open for IVDs. Prequalification Med. Prod. (IVDs, Med. Vaccines 320 Immun. Devices, Vector Control. https://extranet.who.int/pqweb/vitro-321 diagnostics/coronavirus-disease-covid-19-pandemic-—-emergency-use-listing-procedure-eul-322 open 323 19 Hanson KE, Barker AP, Hillyard DR, et al. Self-collected anterior nasal and saliva specimens 324 versus health care worker-collected nasopharyngeal swabs for the molecular detection of 325 SARS-CoV-2. J Clin Microbiol Published Online First: 2020. doi:10.1128/JCM.01824-20 326 20 Tu Y-P, Jennings R, Hart B, et al. Swabs Collected by Patients or Health Care Workers for 327 SARS-CoV-2 Testing. N Engl J Med Published Online First: 2020. doi:10.1056/nejmc2016321 328 329 330 331 332 333 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint 334 335 Figure 1. Boxplot of the SARS-CoV-2 viral load distribution of the RT-qPCR NP swabs used as 336

standard for the participants enrolled for Sure-Status and Biocredit Ag-RDT evaluation. 337 The whiskers show the maximum and minimum values and the vertical line the median. Asterisks 338 indicate statistical significance between AN and NP swab types. 339 340 341 342 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint 343 Figure 2. Limit of detection analyses of upper-respiratory samples positive by RT-qPCR for Sure-344 Status and Biocredit using AN and NP swabs. The log10 RNA copies on the x axis were plotted 345 against a positive (1.0) or negative (0.0) Ag-RDT result on the y axis. Green (Sure-Status) and purple 346 (Biocredit) curves show logistic regressions of the viral load on the Ag-RDT result; vertical dashed 347 lines indicate log10 RNA copies subjected to the test at which 50% and 95% LoD of the samples are 348 expected positive based on the regression results. No significant differences were observed. 349 350 351 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint 352 Figure 3. Boxplot of the scores of the test lines for both Ag-RDT Sure-Status and Biocredit using AN 353 and NP swabs. The whiskers show the maximum and minimum values and the vertical line the 354 median. Asterisks indicate statistical significance between AN and NP swab types. 355 356 357 358 359 360 361 362 363 364 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint 365 Table 1. Demographics of the population of study for Sure-Status and Biocredit cohorts 366 Sure-Status Biocredit All Total 372 232 604 Age [mean (min-max), IQR] 43 (18-81), 33-53 43 (18-78), 33-51 43 (18-81), 33-52 Gender [%F, (n/N)] IQR] 57%, (211/372) 59%, (137/232) 58%, (348/602) Triple vaccinated (n, %) NA* 143 (61.6%) 143 (23.8%) Double vaccinated (n, %) 314 (84.4%) 74 (40%) 388 (64.4%) Partially vaccinated (n, %) 29 (7.8%) 4 (1.7%) 33 (5.5%) Not vaccinated (n, %) 27 (7.3%) 10 (4.3%) 37 (6.2%) Vaccination not disclosed (n, %) 2 (0.5) 1 (0.3%) 3 (0.5%) Days symptoms onset [median (IQR); N] 2 (1-3), 371 2 (1-3), 232 2 (1-3), 601 Days 0-3 (n, %) 304, 81.7% 186, 80.2% 490, 81.1% Days 4-7 (n, %) 56, 15.1% 41, 17.7% 97, 16.1% Days 8+ (n, %) 10, 2.7% 5, 2.2% 15, 2.5% RT-qPCR SARS-CoV-2 Positivity [%, (n/N)] 32.1%, (119/371) 53.7%, (122/227) 40.3% (241/598) Symptom [total n (%), in RT-qPCR positive n (%)] Cough 248 (66.7%), 73 (61.3%) 139 (60.0%), 71 (58.2%) 387 (64.3%), 144 (60.0%) Sore throat 129 (34.7%), 34 (28.6%) 103 (44.4%), 56 (45.9%) 232 (38.5%), 90 (37.4%) Headache 123 (33.1%), 57 (47.9%) 80 (34.5%), 45 (36.9%) 203 (33.7%), 102 (42.3%) Fever 106 (28.5%), 30 (25.2%) 54 (23.3%), 28 (22.9%) 160 (26.6%), 58 (24.1%) Body aches 41 (11.0%), 21 (17.7%) 39 (16.8%), 29 (23.8%) 80 (13.3%), 51 (21.2%) Runny nose 39 (13.2%), 20 (16.8%) 41 (17.7%), 31 (25.4%) 80 (13.3%), 51 (21.2%) Loss taste 48 (12.9%), 19 (16.0%) 19 (8.2%), 10 (8.2%) 67 (11.1%), 29 (12.0%) Loss smell 29 (7.8%), 9 (7.6%) 14 (6.0%), 7 (5.7%) 43 (7.1%), 16 (6.6%) Chest pain 18 (4.8%),7 (5.9%) 12 (5.2%), 8 (6.6%) 30 (5.0%), 15 (6.2%) Fatigue 13 (3.5%), 4 (3.4) 17 (7.3%), 10 (8.2%) 30, (5.0%), 14 (5.8%) Shortness of breath/tight chest 13 (3.5%), 3 (2.5%) 9 (3.9%), 5 (4.1%) 22 (3.6%), 15 (6.2%) Vomiting 11 (3%), 5 (4.2%) 2 (8.6%), 2 (1.6%) 13 (2.2%), 7 (2.9%) Diarrhoea 9 (2.4%), 3 (2.5%) 3 (13%), 3 (2.5%) 12 (2.0%), 6 (2.5%) Abdominal pain 6 (1.6%), 3 (2.5%) 1 (0.4%), 1 (0.8%) 7 (1.2%), 4 (1.7%) Rash 3 (0.8%), 0 (0.0%) 1 (0.4%), 1 (0.8%) 4 (0.6%), 1 (0.4%) Confusion 1 (0.3%), 0 (0.0%) 0 (0%) 1 (0.2%), 0 (0%) Other 159 (42.7%), 68 (57.4%) 134 (57.8%), 85 (69.7%) 293 (48.7%), 153 (63.5%) *Participants were enrolled before booster rolled out in the UK 367 368 369 370 371 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint Table 2. Clinical sensitivity and specificity of Sure-Status and Biocredit using NP and Nasal Swab 372 All Ct values TP FP TN FN Sensitivity Specificity NPV PPV Sure-Status 200 5 494 38 83.9 (78.3-88.1) 99.0 (97.7-99.7) 92.7(90.5-94.4) 97.6 (94.3-98.9) NP swab 99 2 249 20 83.2 (75.2-89.4) 98.8 (96.5-99.6) 91.7 (88.4-94.2) 97.1 (91.44-99.03) AN swab 101 2 245 18 84.0 (76.2-90.1) 99.2 (97.0-99.8) 92.8 (89.5-95.1) 98.0 (92.62-99.5) Biocredit 196 1 209 48 80.3 (74.8-85.1) 99.1 (95.0-99.9) 81.3(77.2-84.9) 99.5 (96.5-100) NP swab 99 1 104 23 81.2 (73.1-87.7) 99.0 (94.7-86.5) 81.9 (73.99-85.2) 99.0 (93.4-99.9) AN swab 97 0 105 25 79.5 (71.3-86.3) 100 (96.5-100) 80.8 (74.8-85.6) 100 (100-100) All NP 198 4 353 43 82.2 (76.7-86.8) 98.9 (97.2-99.7) 89.1 (86.2-91.5) 98.0 (94.9-99.2) All AN 198 2 350 43 82.2 (76.7-86.8) 99.4 (97.9-99.9) 89.1 (86.2-91.5) 99.0 (96.1-99.8) 373 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted September 9, 2022. ; https://doi.org/10.1101/2022.09.06.22279637doi: medRxiv preprint