key: cord-1053366-c3jawu22
authors: Bruemmer, L. E.; Katzenschlager, S.; McGrath, S.; Schmitz, S.; Gaeddert, M.; Erdmann, C.; Bota, M.; Grilli, M.; Larmann, J.; Weigand, M. A.; Pollock, N. R.; Mace, A.; Erkosar, B.; Carmona, S.; Sacks, J. A.; Ongarello, S.; Denkinger, C. M.
title: Accuracy of rapid point-of-care antigen-based diagnostics for SARS-CoV-2: an updated systematic review and meta-analysis with meta regression analyzing influencing factors
date: 2022-02-15
journal: nan
DOI: 10.1101/2022.02.11.22270831
sha: 56a98beb86a355962a17e221863d873093001e6f
doc_id: 1053366
cord_uid: c3jawu22

Background Comprehensive information about the accuracy of antigen rapid diagnostic tests (Ag-RDTs) for SARS-CoV-2 is essential to guide public health decision makers in choosing the best tests and testing policies. In August 2021, we published a systematic review and meta-analysis about the accuracy of Ag-RDTs. We now update this work and analyze the factors influencing test sensitivity in further detail. Methods and findings We registered the review on PROSPERO (registration number: CRD42020225140). We systematically searched multiple databases (PubMed, Web of Science Core Collection, medRvix, bioRvix, and FIND) for publications evaluating the accuracy of Ag-RDTs for SARS-CoV-2 until August 31, 2021. Descriptive analyses of all studies were performed, and when more than 4 studies were available, a random-effects meta-analysis was used to estimate pooled sensitivity and specificity with reverse transcription polymerase chain reaction (RT-PCR) testing as a reference. To evaluate factors influencing test sensitivity, we performed 3 different analyses using multivariate mixed-effects meta-regression models. We included 194 studies with 221,878 Ag-RDTs performed. Overall, the pooled estimates of Ag-RDT sensitivity and specificity were 72.0% (95% confidence interval [CI] 69.8 to 74.2) and 98.9% (95% CI 98.6 to 99.1), respectively. When manufacturer instructions were followed, sensitivity increased to 76.4% (95%CI 73.8 to 78.8). Sensitivity was markedly better on samples with lower RT-PCR cycle threshold (Ct) values (sensitivity of 97.9% [95% CI 96.9 to 98.9] and 90.6% [95% CI 88.3 to 93.0] for Ct-values <20 and <25, compared to 54.4% [95% CI 47.3 to 61.5] and 18.7% [95% CI 13.9 to 23.4] for Ct-values [≥]25 and [≥]30) and was estimated to increase by 2.9 percentage points (95% CI 1.7 to 4.0) for every unit decrease in mean Ct-value when adjusting for testing procedure and patients symptom status. Concordantly, we found the mean Ct-value to be lower for true positive (22.2 [95% CI 21.5 to 22.8]) compared to false negative (30.4 [95% CI 29.7 to 31.1]) results. Testing in the first week from symptom onset resulted in substantially higher sensitivity (81.9% [95% CI 77.7 to 85.5]) compared to testing after 1 week (51.8%, 95% CI 41.5 to 61.9). Similarly, sensitivity was higher in symptomatic (76.2% [95% CI 73.3 to 78.9]) compared to asymptomatic (56.8% [95% CI 50.9 to 62.4]) persons. However, both effects were mainly driven by the Ct-value of the sample. With regards to sample type, highest sensitivity was found for nasopharyngeal (NP) and combined NP/oropharyngeal samples (70.8% [95% CI 68.3 to 73.2]), as well as in anterior nasal/mid-turbinate samples (77.3% [95% CI 73.0 to 81.0]). Conclusion Ag-RDTs detect most of the individuals infected with SARS-CoV-2, and almost all when high viral loads are present (>90%). With viral load, as estimated by Ct-value, being the most influential factor on their sensitivity, they are especially useful to detect persons with high viral load who are most likely to transmit the virus. To further quantify the effects of other factors influencing test sensitivity, standardization of clinical accuracy studies and access to patient level Ct-values and duration of symptoms are needed.

We developed a study protocol following standard guidelines for systematic reviews [8, 9] , which 95 is available in the supplements (S1 Text). We also completed the PRISMA checklist (S1 PRISMA Check-96 list), and registered the review on PROSPERO (registration number: CRD42020225140). 97 98 Search strategy 99

We performed a search of the databases PubMed, Web of Science, medRxiv, and bioRxiv. The 100 search terms were developed with an experienced medical librarian (M. Grilli), using combinations of 101 subject headings (when applicable) and text-words for the concepts of the search question. The main 102 search terms were "Severe Acute Respiratory Syndrome Corona-virus 2", "COVID-19", "Betacorona-103 virus", "Coronavirus", and "Point of Care Testing" with no language restrictions. The full list of search 104 terms is available in S2 Text. We also manually searched the website of FIND, the global alliance for 105 diagnostics (https://www.finddx.org/sarscov2-eval-antigen/), for additional relevant studies. We per-106 formed the search bi-weekly through August 31, 2021. The last manual search of the FIND website was 107 performed on September 10, 2021. In addition to conducting the present review, we updated our 108 website www.diagnosticsglobalhealth.org weekly with the latest search results based on the methods 109 outlined below. 110 111

We included studies evaluating the accuracy of commercially available Ag-RDTs to establish a di-113 agnosis of SARS-CoV-2 infection at the point-of-care (POC), against reverse transcription polymerase 114 chain reaction (RT-PCR) or cell culture as reference standard. We included all study populations irre-115 spective of age, presence of symptoms or study location. We considered cohort studies, nested cohort 116 studies, case-control or cross-sectional studies, and randomized studies. We included both peer-re-117 viewed publications and preprints. 118 . CC-BY-NC-ND 4.0 International license It 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 this this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 We excluded studies, in which patients were tested for the purpose of monitoring or ending quar-119 antine. Also, publications with a population size smaller than 10 were excluded (although the size 120 threshold of 10 is arbitrary, such small studies are more likely to give unreliable estimates of sensitivity 121 and specificity). Analytical accuracy studies, where tests are performed on spiked samples with a 122 known quantity of virus, were also excluded. 123 124 Index tests 125

Ag-RDTs for SARS-CoV-2 aim to detect infection by recognizing viral proteins (typically the SARS-126

CoV-2 nucleoprotein). Most Ag-RDTs dedicated for point-of-care deployment use specific labeled an-127 tibodies attached to a nitrocellulose matrix strip (lateral flow assay), to capture and detect the viral 128 antigen. Successful binding of the antibodies to the antigen either is detected visually by the appear-129 ance of a line on the matrix strip or through a specific reader instrument for fluorescence detection. 130

Other point-of-care instrument-based tests use chips or cartridges that enable an automated immu-131 noassay testing procedure. Ag-RDTs typically provide results within 10 to 30 minutes [10] . 132 133 Reference standard 134 Viral culture detects viable virus that is relevant for transmission but is only available in research 135 settings. Since RT-PCR tests are more widely available and SARS-CoV-2 RNA (as reflected by cycle threshold [Ct] value) highly correlates with SARS-CoV-2 antigen quantities [11], we considered 137 RT-PCR an acceptable reference standard for the purposes of this systematic review. Where an inter-138 national standard for the correlation of the viral load to the Ct-values was used, we also report the 139 viral load [12] . 140

The study selection, data extraction, assessment of the quality of studies and their independence 142 from manufacturers were performed as per prior description [4] . The data items extracted can be 143 found in the Supplement (S1 Table) . 144 7 145 Statistical analysis and data synthesis 146

We extracted raw data from the studies and recalculated performance estimates where possible 147 based on the extracted data. Also, some primary studies reported the median Ct-value along with the 148 first and third interquartile range (IQR) and/or minimum and maximum values rather than the sample 149 mean and standard deviation. To incorporate these studies in our analyses, we applied the Quantile 150

Estimation approach [13] to estimate the mean and standard deviation of the Ct-values. In an effort to 151 use as much of the heterogeneous data as possible, the cutoffs for the Ct value groups were relaxed 152 by 2-3 points within each range. The <20 group included values reported up to ≤20, the <25 group 153 included values reported as ≤24 or <25 or 20-25, and the <30 group included values from ≤29 to ≤33 154 and 25-30. The ≥25 group included values reported as ≥25 or 25 to 30, and the ≥30 group included 155 values from ≥30 to ≥35. For the same reason, when categorizing by age, the age group <18 years (chil-156 dren) included samples from persons whose age was reported as <16 or <18 years, whereas the age 157 group ≥18 years (adults) included samples from persons whose age was reported as ≥16 or ≥18 years. 158

Also, for the symptom duration groups, the ≤7 days group included ≤4, ≤5, ≤6, 6 to 7, ≤7 and ≤9 days, 159 and the >7 days group >5, 6 to 10, 6-21, >7 and 8 to 14 days. The raw data can be found in the supple-160 ments (S2 Table) and with more details online (https://doi.org/10.11588/data/T3MIB0). We prepared 161 forest plots for the sensitivity and specificity of each test and visually evaluated the heterogeneity 162 between studies. 163

If 4 or more datasets were available with at least 20 RT-PCR-positive samples per dataset for a 164 predefined analysis, a meta-analysis was performed. We report pooled estimates of sensitivity and 165 specificity for SARS-CoV-2 detection along with 95% confidence intervals (CIs) using a bivariate model 166 (implemented with the "reitsma" command from the R package "mada," version 0.5.10). In subgroup 167 analyses (below), where papers presented data only on sensitivity, a univariate random effects inverse 168 variance meta-analysis was performed (using the "metagen" command from the R package "meta", 169 version 5.1-1, and the "rma" command from the R package "metafor", version 3.0-2). When there were 170 . CC-BY-NC-ND 4.0 International license It 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 this this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 fewer than 4 studies for an index test, only a descriptive analysis was performed, and accuracy ranges 171 are reported. 172

We predefined subgroups for meta-analysis based on the following characteristics: Ct value 173 range, testing procedure in accordance with manufacturer's instructions as detailed in the instructions 174 for use (IFU) (henceforth called IFU-conforming) versus not IFU-conforming, age (<18 versus ≥18 175 years), sample type, presence or absence of symptoms, symptom duration (<7 days versus ≥7 days), 176 viral load, and type of RT-PCR assay used. We also provide mean Ct-value across true positive and false 177 negative test results. For categorization by sample type, we assessed (1) nasopharyngeal (NP) alone or 178 combined with other (e.g., oropharyngeal [OP]), (2) OP alone, (3) anterior nasal (AN) or mid-turbinate 179 (MT), (4) a combination of bronchoalveolar lavage and throat wash (BAL/TW), or (5) saliva. 180

We applied multivariate linear mixed-effect meta-regression models to explore factors that affect 181 diagnostic test sensitivity. Based on our previous analysis [4], we a priori defined an individual's time 182 since infection and sample type and condition as underlying factors, influencing test sensitivity through 183 an individual's symptom status (symptomatic vs. asymptomatic), the samples' viral load (estimated by 184 the mean Ct-value as presented in the study for the sub cohort of interest) and the testing procedure 185 (IFU-vs. not IFU-conforming). We performed 3 different analyses, each of which obtained unadjusted 186 and adjusted estimates (i.e., an estimate of the association between a factor and test sensitivity, hold-187 ing the other covariates in the model constant) of the effect of factors on test sensitivity. 188

In the first analysis, we estimated the direct effect of symptom status, viral load and testing pro-189 cedure on test sensitivity. For the second and third analysis, we restricted the meta-regression models 190 to data sets of symptomatic persons due to a lack of data. Specifically, the second analysis assessed 191 the effect of time since infection (estimated as the sample mean of symptom duration), viral load, and 192 testing procedure on test sensitivity. The third analysis also assessed the effect of time since infection, 193 viral load, and testing procedure on test sensitivity, but depicted the time since infection as a binary 194 covariate of the symptom duration subgroup (≤7 vs >7 days). Further details on the implementation of 195 the meta-regression models and the underlying casual diagrams are available in the Supplement (S3 196 . CC-BY-NC-ND 4.0 International license It 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 this this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 Text, S3a to S3c). Data sets with less than 5 RT-PCR positives were excluded. We considered an effect 197 to be statistically significant when the regression coefficient's 95% confidence interval did not include 198 0. The analyses were performed using the "metafor" R package, version 3.0-2 [14]. 199

As recommended to investigate publication bias for diagnostic test accuracy meta-analyses, we 200 performed the Deeks test for funnel-plot asymmetry [15] (using the "midas" command in Stata, ver-201 sion 15); a p-value < 0.10 for the slope coefficient indicates significant asymmetry. 202 203 Sensitivity analysis 204 3 sensitivity analyses were performed: estimation of sensitivity and specificity excluding case-205 control studies, estimation of sensitivity and specificity excluding not peer-reviewed studies, and esti-206 mation of sensitivity and specificity excluding studies that were potentially influenced through test 207 manufacturers. We compared the results of each sensitivity analysis against the overall results to as-208 sess the potential bias introduced by case-control, not peer-reviewed, and manufacturer-influenced 209 studies. 210

The systematic search resulted in 31,254 articles. After removing duplicates, 11,462 articles were 214 screened and 433 papers were considered eligible for full-text review. Of these, 259 were excluded 215 because they did not present primary data or the Ag-RDT was not commercially available, leaving 174 216 studies to be included in the systematic review . A list of the studies excluded and their reason 217 for exclusion can be found in the Supplement (S4 Text). Further 20 studies were included from the 218 FIND website [190] [191] [192] [193] [194] [195] [196] [197] [198] [199] [200] [201] [202] [203] [204] [205] [206] [207] [208] [209] (Fig 1) . 219 220 . CC-BY-NC-ND 4.0 International license It 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 this this version posted February 15, 2022. ; https://doi.org/10.1101/2022.02.11.22270831 doi: medRxiv preprint At the end of the data extraction process, 21 studies were still in preprint form [16, 17, 21, 47, 50, 226 55, 58, 65, 69, 74, 84, 100, 116, 121, 129, 160, 167, 168, 173, 174, 186] . All studies were written in 227 English, except for 3 in Spanish [53, 62, 134] , 1 in Turkish [95] , and 1 in French [153] . Out of the 196 228 studies, 26 conducted a case-control study [21, 32, 34, 66, 67, 72, 81, 84, 88-90, 93, 94, 96, 103, 108, 229 135, 140, 144, 145, 151, 156, 166, 168, 182, 184] , while the remaining 168 were cross-sectional or 230 cohort studies. The reference method was RT-PCR in all except 1 study, which used viral culture [135] . 231

The 194 studies were divided into 333 data sets. Across these, 76 different Ag-RDTs were evalu-232 ated (75 lateral flow assays, of which 63 are interpreted visually and 12 required an automated, pro-233 prietary reader; 1 assay is an automated immunoassay). The most common reasons for testing were 234 

Records sought for retrieval (n = 24)

Records not retrieved (n = 0)

. CC-BY-NC-ND 4.0 International license It 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 this this version posted February 15, 2022. ; https://doi.org/10.1101/2022.02.11.22270831 doi: medRxiv preprint the occurrence of symptoms (98 data sets; 29.4% of data sets) and screening of asymptomatic persons 235 with (3; 0.9%) or without (22; 6.6%) close contact to a SARS-CoV-2 confirmed case. In 142 (42.6%) of 236 the data sets, individuals were tested due to more than 1 of the reasons mentioned and for 68 (20.4%) 237 the reason for testing was unclear. 238

In total, 221,878 Ag-RDTs were performed, with a mean number of samples per study of 666 239 (Range 15 to 22,994). The age of the individuals tested was specified for only 90,986 samples, of which 240 84,124 (92.5%) were from adults (age group ≥18) and 6,862 (7.5%) from children (age group <18 The findings on study quality using the QUADAS 2 tool are presented in Fig 2a and 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 this this version posted February 15, 2022. ; https://doi.org/10.1101/2022.02.11.22270831 doi: medRxiv preprint chosen (i.e., they avoided inappropriate exclusions or a case-control design and enrollment occurred 261 consecutive or randomly). The conduct and interpretation of the index tests was considered to have low risk of bias in 176 268 (52.9%) data sets (e.g., through appropriate blinding of persons interpreting the visual read-out). How-269 ever, for 155 (46.5%) data sets sufficient information to clearly judge the risk of bias was not provided. 270

In only 151 (45.3%) data sets the Ag-RDTs were performed according to IFU, while 138 (41.4%) were 271 not IFU-conforming, potentially impacting the diagnostic accuracy; for 44 (13.2%) data sets the IFU 272 status was unclear. The most common deviations from the IFU were (1) use of samples that were pre-273 diluted in transport media not recommended by the manufacturer (113 data sets; 12 unclear), (2) use 274 of banked samples (103 data sets; 12 unclear), and (3) a sample type that was not recommended for 275

Ag-RDTs (8 data sets; 11 unclear). 276

In 126 (37.8%) data sets the reference standard was performed before the Ag-RDT, or the opera-277 tor conducting the reference standard was blinded to the Ag-RDT results, resulting in a low risk of bias. 278 . CC-BY-NC-ND 4.0 International license It 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 this this version posted February 15, 2022. ; https://doi.org/10. 1101 In almost all other data sets (206; 61.9%) this risk could not be assessed, due to missing information 279 and for 1 data set (0.3%) intermediate concern was raised. The applicability of the reference test was 280 judged to be of low concern for all data sets, as viral culture or RT-PCR are considered to adequately 281 define the target condition for the purpose of this study. 282

In 327 (98.2%) data sets, the sample for the index test and reference test were obtained at the 283 same time, while this was unclear in 6 (1.8%). In 226 (67.9%) data sets, the same RT-PCR assay was 284 used as the reference of all included samples, while in 85 (25.5%) data sets multiple RT-PCR assays 285 were used as the reference. For 22 (6.6%) data sets, the RT-PCR used as reference standard was un-286 clear. Furthermore, for 19 (5.7%) data sets, there was a concern that not all selected study participants 287 were included in the analysis. Overall, 38 data sets were excluded from the meta-analysis, as they included fewer than 20 RT-296 PCR positive samples. An additional 28 data sets were missing either sensitivity or specificity and were 297 only considered for univariate analyses. The remaining 267 data sets, evaluating 198,584 tests, pro-298 vided sufficient data for bivariate analysis. The results are presented in Fig 3a- e. Detailed results for 299 the subgroup analysis are available in the Supplement (S3 Fig, S4 Fig, S5 Fig, S6 Fig and S7 Fig) . 300

Including any test and type of sample, the pooled estimates of sensitivity and specificity were 301 72.0% (95%CI 69.8 to 74.2) and 98.9% (95%CI 98.6 to 99.1), respectively. If testing was performed 302 according to IFU, sensitivity increased to 76.4% (95%CI 73.8 to 78.8) compared to not IFU-conforming 303 . CC-BY-NC-ND 4.0 International license It 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 Ag-RDT (Standard Q) reported equivalent performance [112, 113] . In contrast, saliva swabs (4 data 328 sets, 1,216 samples) showed the lowest pooled sensitivity with only 50.1% (95% CI 7.7 to 92.3) (Fig 3c) . 329 . CC-BY-NC-ND 4.0 International license It 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 this this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 In 3 of the data sets utilizing a saliva sample, saliva was collected as whole mouth fluid [20, 88, 150] . 330

The fourth used a cheek swab for sample collection [51] . 331 Due to only 3 data sets with 3586 samples, we were not able to estimate pooed sensitivity and 332 specificity for OP samples. Median sensitivity and specificity were 59.4% (range 50.0% to 81.0%) and 333 99.1% (range 99.0% to 100.0%), respectively. We were also not able to perform a subgroup meta-334 analysis for BAL/TW due to insufficient data, with only 1 study with 73 samples evaluating the Biocredit 335 

Within the data sets possible to meta-analyze, 55,186 (43.2%) samples were from symptomatic 341 and 72,457 (56.8%) from asymptomatic persons. The pooled sensitivity for symptomatic persons was 342 markedly higher compared to asymptomatic persons with 76.2% (95% CI 73.3 to 78.9) vs. 56.8% (95% 343 CI 50.9 to 62.4). Specificity was above 98.6% for both groups (Fig 3d) . 344 345 Subgroup analysis comparing symptom duration 346 Data were analyzed for 9,470 persons from 26 data sets with symptoms less than 7 days, while 347 for persons with symptoms  7 days fewer data were available (620 persons, 13 data sets). The pooled 348 sensitivity estimate for individuals with symptoms <7 days was 81.9% (95% CI 77.7 to 85.5), which is 349 markedly higher than the 51.8% (95% CI 41.5 to 61.9) sensitivity for individuals tested  7 days from 350 onset of symptoms (Fig 3d) . 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 this this version posted February 15, 2022. ; https://doi.org/10.1101/2022.02.11.22270831 doi: medRxiv preprint overlapping confidence intervals in sensitivity with 74.8% (95% CI 71.5 to 77.8) and 69.8% (95% CI 61.0 356 to 77.3) for the adult and pediatric group, respectively. For those data sets that reported a median Ct-357 value per age group, the Ct-value was slightly lower in the adult (median 22.6; Q1 = 20.5; Q3 = 24.6; 358 48 data sets) compared to the pediatric group (median 23.2; Q1 = 20.3; Q3 = 25.2; 3 data sets). Speci-359 ficity was similar in both groups with over 99% (Fig 3e) . is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint In the first analysis, we found viral load (as estimated by Ct-value) to be the driving factor of 375 sensitivity. Sensitivity was estimated to increase by 2.9 percentage points (95% CI 1.7 to 4.0) for every 376 unit the mean Ct-value decreased (Fig 4, Supplement Text S3d) , after adjusting for symptom status and 377 testing procedure. In addition, sensitivity was estimated to be 20.0 percentage points (95% CI 13.7 to 378 26.3) higher for samples from symptomatic compared to asymptomatic participants. However, when 379 controlling for testing procedure and mean Ct-value, this difference declined to only 11.1 percentage 380 points (95% CI 4.8 to 17.4). The difference between IFU-conforming vs. not IFU-conforming testing 381 procedure was not significant (5.2 percentage points [95% CI -2.6 to 13.0] higher for IFU-conforming) 382 after controlling for symptom status and mean Ct-value. 383 384 . CC-BY-NC-ND 4.0 International license It 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 this this version posted February 15, 2022. ; https://doi.org/10.1101/2022.02.11.22270831 doi: medRxiv preprint When assessing only symptomatic participants, test sensitivity was estimated to decrease by 3.2 392 percentage points (95% CI -1.5 to 7.9) for every 1 day increase in average duration of symptoms (mean 393 duration of symptoms ranged from 2.75 to 6.47 days). However, with the confidence interval including 394 the value zero, this effect was not statistically significant. When controlling for mean Ct-value and test-395 ing procedure, the estimated effect of the average duration of symptoms was close to 0 (0.7 percent-396 age points [95% CI -5.0 to 6.4], Supplement Text S3e). 397

Concordantly, for samples collected after 7 days of symptom onset sensitivity was estimated to 398 is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint Based on 179 data sets with 143,803 tests performed, we were able to perform bivariate meta-405 analysis of the sensitivity and specificity for 12 different Ag-RDTs (Fig 5) . Across these, pooled estimates 406 of sensitivity and specificity on all samples were 71.6% (95%CI 69.0 to 74.1) and 99.0% (95% CI 98.8 to 407 99.2), which were very similar to the overall pooled estimate across all meta-analyzed data sets (72.0% 408 and 98.9%, above). is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint were similar with 57.7% (95% CI 40.9 to 74.4) and 56.6% (95% CI 26.9 to 86.3), respectively (Fig S11) . 430

Specificity was only possible to calculate for the Nadal, which was lowest throughout the per test anal-431 ysis with 91.1% (95% CI 80.2 to 100). For the remaining 62 Ag-RDTs, there were insufficient numbers 432 of data sets for a uni-or bivariate meta-analysis, but performance estimates for each of these tests are 433 descriptively compared in the Supplement (S3 Table) 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 this this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 available in the Supplement (S4 Fig A-F) as well. In addition, for 8 tests it was possible to calculate 445 pooled sensitivity and specificity estimates only including data sets that conformed to the IFU. These 446 are also listed in the Supplement (S4 Table) . 447

In total, 31 studies accounting for 106 data sets conducted head-to-head clinical accuracy evalu-448 ations of different tests using the same sample(s) from the same participant. These data sets are out-449 lined in the Supplement (S2 Table) . 9 studies performed their head-to-head evaluation as per IFU and 450 on symptomatic individuals. Across 4 studies, the Standard Q nasal (sensitivity 80.5% to 91.2%) and 451 the Standard Q (sensitivity 73.2% to 91.2%) showed a similar range of sensitivity [112, 126, 211] is the author/funder, who has granted medRxiv a license to display the preprint in (which was not certified by peer review) preprint performed according to manufacturers' instructions. Viral load as estimated by Ct-value was the de-474 ciding factor for sensitivity, confirming prior work [11] . There was a significant correlation between 475 test sensitivity and Ct-value, with sensitivity increasing by 2.9 percentage points for every unit decrease 476 in mean Ct-value when controlling for symptom status and testing procedure. The pooled Ct-value for 477 TP was on average over 8 points lower than for FN results (Ct-value of 22.2 for TP compared to 30.4 478 for FN results). A decrease in performance when not following manufacturer's instructions was sug-479 gested but was not significant (-5.2 percentage points [95% CI -13.0 to 2.6]). 480

Furthermore, sensitivity was found to be higher when samples were from symptomatic (76.2% 481 sensitivity) compared to asymptomatic participants (56.8% sensitivity). This was confirmed in the re-482 gression model, estimating sensitivity to be 20.0 percentage points higher in samples that originated 483 from symptomatic participants. In our previous analysis, we assumed that the increase in sensitivity is 484 not due to the symptom status as such, but results from the fact that in symptomatic study populations 485 chances are higher to include participants at the beginning of the disease with high viral load [4] . In 486 the present analysis, this assumption shows to be largely true. Controlling for Ct-value, the RT-PCR 487 correlate for viral load, the effect of symptomatic vs. asymptomatic participants on test sensitivity 488 strongly decreased to 11.1 percentage points. As others found symptomatic and asymptomatic indi-489 viduals to have the same viral load when at the same stage of the disease [6], we would have expected 490 the regression coefficient to have decreased even further to 0. This non-zero difference in sensitivity 491 between symptomatic and asymptomatic participants may be due to the lack of access to individual 492 participant Ct-values, which required our analyses to control for the mean Ct-value over all participants 493 in a data set rather than the individual Ct-values. Furthermore, some variability is likely introduced by 494 the testing for the Ag-RDT and the RT-PCR not to occur from the sample. Therefore, some degree of 495 residual confounding is likely present. 496 . CC-BY-NC-ND 4.0 International license It 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 this this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 We also found sensitivity to be higher when participants were tested within 7 days of symptom 497 onset (81.9% sensitivity) compared to > 7 days (51.8% sensitivity). Concordantly, our regression model 498 estimated that sensitivity decreases by 3.2 percentage points for every 1 day increase in mean symp-499 tom duration. Again, this decrease in sensitivity is driven by viral load as was seen when controlling for 500

Ct-value. Importantly, it is not yet clear how the emergence of new SARS-CoV-2 variants of concern 501 and the growing vaccination coverage will affect Ag-RDTs sensitivity in the early days after symptom Looking at specific tests, LumiraDx and Standard Q nasal showed the highest sensitivity, perform-510 ing above the 80% sensitivity target defined by the World Health Organization. However, while the 511 Standard Q nasal was 99.1% (95% CI 98.4 to 99.5) specific, the LumiraDx only had a specificity of 96.9% 512 (95) (96) (97) (98) 3), which is just below the WHO target of 97%. The reason for the lower specificity is 513 unclear, particularly as independent analytical studies also confirmed the test had no cross-reactivity 514 [102] . Sample to sample variability must be considered, particularly as the sensitivity of the index tests 515 approaches that of the reference test. The 2 most often evaluated tests, namely Panbio (32,370 sam-516 ples; sensitivity of 71.9%) and Standard Q (35,664 samples; sensitivity of 70.9%), performed slightly 517 below the overall average. Similarly, Panbio and Standard Q were also the most extensively evaluated 518

Ag-RDTs in the prior analysis, and with a sensitivity slightly above average [4] . Nonetheless, this up-519 dated analysis indicates that there is limited added value is to be expected from any further analysis 520 of Ag-RDTs' overall sensitivity or the sensitivity of the most widely-evaluated tests. However, it will be 521 important to continue to reassess tests' analytical sensitivity for detection of new specific variants 522 . CC-BY-NC-ND 4.0 International license It 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 this this version posted February 15, 2022. ; https://doi.org/10.1101 . Also, it would be of interest to evaluate the characteristics of specific formats of Ag-523

RDTs, such as Ag-RDTs that are self-performed or those that are instrument-based. 524

Furthermore, sensitivity strongly differed between studies that conducted the Ag-RDTs as per 525 manufacturer's instructions and those that did not (sensitivity of 66.7% for not IFU-conforming vs. 526 76.4% for IFU-conforming). With regards to sample types, saliva showed a markedly lower sensitivity 527 of 50.1%, compared to NP or AN/MT samples, confirming what we found in our previous analysis [4] . 528

Especially in light of the current debate on whether saliva or throat swabs might be a more sensitive 529 sample to detect the SARS-CoV-2 Omicron variant than NP or AN/MT samples [216] [217] [218] , further re-530 search is urgently needed to quantify the difference in viral load resulting from different sample types 531 and thus the effect of sample type on test sensitivity. 532

In concordance with the above, many studies reporting an unusually low sensitivity performed 533 the Ag-RDT not as per IFU [26, 28, 40, 66, 97, 108, 133, 184] or used saliva samples [20, 150, 155, 219] . 534

However, 2 studies with IFU-conforming testing procedure on NP or AN/MT sample, still showed a low 535 sensitivity. This quite likely results from the on average low viral load in 1 study [49] , and the asymp-536 tomatic study population in the other [175] . On the contrary, compared to the other studies unusual 537 high sensitivity was found in studies where average viral load was high [45, 84, 144, 145] or participants 538 were mainly within the first week of symptom onset [42, 54, 135] . 539

The main strength of our study lies in its living approach. The ability to update our methodology 540 as the body of evidence grows has enabled more improved analysis. For example, while data were too 541 heterogenous for a meta-regression during the prior analysis, with additional datasets we are now able 542 to analyze the relationship between an Ag-RDT's sensitivity, the samples' Ct-value, and the partici-543 pants' symptom status in depth. Similarly, we decided to focus on clinical accuracy studies for POC RDTs in this current review as analytical accuracy studies require a dedicated approach to be compa-545 rable. Furthermore, the main results of our latest extractions are publicly available on our website. 546

This has not only equipped public health professionals with an up-to-date overview on the current 547 . CC-BY-NC-ND 4.0 International license It 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 this this version posted February 15, 2022. ; https://doi.org/10.1101/2022.02.11.22270831 doi: medRxiv preprint research landscape [220, 221] , but also led other researchers and the test manufacturers to check our 548 data, improving the quality of our report through continuous peer-review. 549

Nonetheless, our study is limited in that we use RT-PCR as a reference standard to assess the 550 accuracy of Ag-RDTs, which might be a less appropriate reference standard than viral culture [135, 551 222 ]. However, viral culture is available in research settings only and its validity as a true proxy of actual 552 transmissibility is not proven, therefore we find RT-PCR a suitable reference standard for the clinical 553 accuracy studies included in this review. Furthermore, we fully acknowledge that Ct-value is not equiv-554 alent to viral load, and that the correlation between Ct-value and viral load varies between RT-PCR 555 assays; nonetheless, we believe that the analysis of pooled Ct-value data across a very large data set 556 is a useful strategy to understand the overall contribution of viral load to Ag-RDT performance. Finally, 557

we are aware that the test specific sensitivities and specificities can be influenced by differences in 558 study design. However, we aimed to counterbalance this effect by assessing relevant aspects in study 559 design for each study and analyzing outliers. To enhance comparability in between clinical accuracy 560 studies, future studies should include individuals at a similar stage in the disease, use the same sample 561 types and adhere to the WHO standard for measuring SARS-CoV-2 viral load [12] . 562 563 CONCLUSION 564

In summary, Ag-RDTs detect most of the persons infected with SARS-CoV-2 when performed ac-565 cording to the manufacturers' instructions. While this confirms the results of our previous analysis, the 566 present analysis highlights that the sample's viral load is the most influential factor underlying test 567 sensitivity. Thus, Ag-RDTs can play a vital role in detecting persons with high viral load and therefore 568 likely to be at highest risk of transmitting the virus. This holds true even in the absence of patient 569 symptoms or differences in the duration of symptoms. To foster further research analyzing specific Ag-570

RDTs and the factors influencing their sensitivity in more detail, standardization of clinical accuracy 571 studies and access to patient level Ct-value and duration of symptoms are essential. 572 . CC-BY-NC-ND 4.0 International license It 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 this this version posted February 15, 2022. ; https://doi.org/10.1101/2022.02.11.22270831 doi: medRxiv preprint . CC-BY-NC-ND 4.0 International license It 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 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 this this version posted February 15, 2022. ; https://doi.org/10.1101 https://doi.org/10. /2022 

World Health Organization, Recommendations for national SARS-CoV-2 testing strategies and 574 diagnostic capacities

Rapid, point-of-care antigen and molecular-based tests for diagnosis of 577 SARS-CoV-2 infection

Accuracy of novel antigen rapid diagnostics for SARS-CoV-2: A living 581 systematic review and meta-analysis

COVID-19 Symptoms: Longitudinal Evolution and Persistence in Outpatient 583 Settings

Viral dynamics of acute SARS-CoV-2 infection and applications to 585 diagnostic and public health strategies

Kinetics of viral load and antibody response in relation to COVID-19 severity

Systematic reviews of diagnostic test accuracy

World Health Organization, Antigen-detection in the diagnosis of SARS-CoV-2 infection: 593 interim guidance

Correlation of SARS-CoV-2 Nucleocapsid Antigen and RNA Concentrations in 595

Nasopharyngeal Samples from Children and Adults Using an Ultrasensitive and Quantitative 596

World Health Organization, First WHO International Standard for SARS-CoV-2 RNA

Estimating the sample mean and standard deviation from commonly 600 reported quantiles in meta-analysis

Investigation of publication bias in meta-analyses of diagnostic test 604 accuracy: a meta-epidemiological study

Field evaluation of specificity and sensitivity of a standard SARS-CoV-606 2 antigen rapid diagnostic test: A prospective study at a teaching hospital in Northern Ghana. 607 medRxiv

Comparison of SARS-COV-2 nasal antigen test to nasopharyngeal RT-609 PCR in mildly symptomatic patients. medRxiv

Time scale performance of rapid antigen testing for SARS-COV-2: 611 evaluation of ten rapid antigen assays

David vs. Goliath": A simple antigen detection test with potential to 613 change diagnostic strategy for SARS-CoV-2

Evaluation of the rapid antigen test Panbio COVID-19 in saliva and nasal 616 swabs: A population-based point-of-care study

Evaluation and clinical implications of the time to a positive results of 618 antigen testing for SARS-CoV-2. medRxiv

Field performance evaluation of the PanBio rapid 620 SARS-CoV-2 antigen assay in an epidemic driven by 501Y.v2 (lineage B.1.351) in the Eastern 621 Cape, South Africa

Field evaluation of a rapid antigen test

Device) for COVID-19 diagnosis in primary healthcare centres. Clinical Microbiology and 624 Infection

Analytical and Clinical Performance of the Panbio

Detecting Rapid Diagnostic Test

A Real-World Comparison of SARS-CoV-2 Rapid Antigen vs

Polymerase Chain Reaction Testing in Florida

Evaluation of Three Immunoassays for the Rapid Detection of SARS-CoV-2 631

Clinical validation of an open-access SARS-COV-2 antigen detection 633 lateral flow assay, compared to commercially available assays

Performance characteristics of five antigen-detecting rapid diagnostic test 636 (Ag-RDT) for SARS-CoV-2 asymptomatic infection: a head-to-head benchmark comparison

Comparison of Quidel Sofia SARS FIA Test to Hologic Aptima SARS-CoV-2 639 TMA Test for Diagnosis of COVID-19 in Symptomatic Outpatients

Diagnostic accuracy of two commercial SARS-CoV-2 Antigen-detecting rapid 642 tests at the point of care in community-based testing centers

Evaluation of an antigen-based test for hospital point-of-care diagnosis of 645 SARS-CoV-2 infection

Efficacy comparison of three rapid antigen tests for SARS-CoV-2 and how 647 viral load impact their performance

Real-life evaluation of the Sofia SARS-CoV-2 antigen assay in a large 649 tertiary care hospital

Analytical performances of the point-of-care SIENNA™

Antigen Rapid Test for the detection of SARS-CoV-2 nucleocapsid protein in nasopharyngeal 652 swabs: A prospective evaluation during the COVID-19 second wave in France. International 653 Journal of Infectious Diseases

Diagnostic Performance of an Antigen Test with RT-PCR for the Detection of 655 SARS-CoV-2 in a Hospital Setting

Comparative diagnostic performance of different rapid antigen detection 658 tests for COVID-19 in the real-world hospital setting. International Journal of Infectious 659 Diseases

Panbio™ rapid antigen test for SARS-CoV-2 has acceptable accuracy in 661 symptomatic patients in primary health care

Are antigenic tests useful for detecting SARS-CoV-2 infections in patients 663 accessing to emergency departments? Results from a North-West Italy Hospital

Evaluation of a rapid antigen detection test

Ag Rapid Test Device) as a point-of-care diagnostic tool for COVID-19 in a Pediatric 667 Emergency Department

ImplemeNting SARS-CoV-2 Rapid antigen testing in the Emergency wArd 669 of a Swiss univErsity hospital: the INCREASE study. microorganisms

Evaluation of a SARS-CoV-2 antigen-detecting rapid diagnostic test as a 673 self-test: diagnostic performance and usability

Frontline Screening for SARS-CoV-2 Infection at Emergency Department 675 Admission by Third Generation Rapid Antigen Test: Can We Spare RT-qPCR? Viruses-Basel

Urgent need of rapid tests for SARS CoV-2 antigen detection: Evaluation of 678 the SD-Biosensor antigen test for SARS-CoV-2

Rapid SARS-CoV-2 antigen detection assay in comparison with real-time 681 RT-PCR assay for laboratory diagnosis of COVID-19 in Thailand

Evaluation of the INDICAID COVID-19 Rapid Antigen Test in symptomatic 684 populations and asymptomatic community testing. Microbiology Spectrum

Clinical evaluation of BD Veritor™ SARS-CoV-2 and Flu A+B Assay for 687 point-of-care (POC) System. medRxiv

Performance of a rapid antigen test in the diagnosis of SARS-CoV-2 infection

Comparison of SARS-CoV-2 Detection by Rapid Antigen and by Three 691

A Study from Martin University Hospital in Slovakia. International 692 Journal of Environmental Research and Public Health

Emergence of N antigen SARS-CoV-2 genetic variants escaping 694 detection of antigenic tests. medRxiv

Diagnostic Accuracy of a New Antigen Test for SARS-CoV-2 Detection. 696 International Journal of Environmental Research and Public Health

Diagnosing SARS-CoV-2 with Antigen Testing, Transcription-Mediated 698 Amplification and Real-Time PCR

Experience with Panbio rapid antigens test device for the 700 detection of SARS-CoV-2 in nursing homes. Enfermedades Infecciosas y Microbiología Clínica

High-Sensitivity SARS-CoV-2 Nucleocapsid Immunoassay to Aid 703 Diagnosis of Acute COVID-19 at the Point of Care: A Clinical Performance Study. Infectious 704 Diseases and Therapy

The sensitivity of SARS-CoV-2 antigen tests in the view of large-scale 706 testing. medRxiv

Real-life evaluation of a COVID-19 rapid antigen detection test in 708 hospitalized children

The effectiveness of rapid antigen test-based for SARS-CoV-2 detection in 710 nursing homes in Valencia

Evaluation of the Panbio™ COVID-19 Ag Rapid Test at an Emergency 712 Room in a Hospital in São Paulo, Brazil. medRxiv

Head-to-Head Comparison of Rapid and Automated Antigen Detection 714 Tests for the Diagnosis of SARS-CoV-2 Infection

Evaluation of the Panbio Covid-19 rapid antigen detection test device for 717 the screening of patients with Covid-19

Validation testing to determine the sensitivity of lateral flow testing for 720 asymptomatic SARS-CoV-2 detection in low prevalence settings: Testing frequency and public 721 health messaging is key

Usefulness of the LumiraDx SARS-COV-2 antigen test in nursing home

Enfermedades Infecciosas y Microbiología Clínica

Validation of a rapid antigen test as a screening tool for SARS-725

CoV-2 infection in asymptomatic populations. Sensitivity, specificity and predictive values

Accuracy of COVID-19 rapid antigenic tests compared to RT-PCR in a student 728 population: The StudyCov study

COVID-19 Rapid Antigen Test at hospital admission associated to the 730 knowledge of individual risk factors allow overcoming the difficulty of managing suspected 731 patients in hospitals COVID-19 Rapid Antigen Test facilitates the management of suspected 732 patients on hospital admission. medRxiv

Performance of six rapid diagnostic tests for SARS-CoV-2 antigen detection 734 and implications for practical use

Multidisciplinary assessment of the Abbott BinaxNOW SARS-CoV-2 point-736 of-care antigen test in the context of emerging viral variants and self-administration. Sci Rep

Performance of the Innova SARS-CoV-2 antigen rapid lateral flow 739 test in the Liverpool asymptomatic testing pilot: population based cohort study

Differences in detected viral loads guide use of SARS-CoV-2 antigen-742 detection assays towards symptomatic college students and children. medRxiv

Real-life validation of the Panbio COVID-19 Antigen Rapid Test (Abbott) 745 in community-dwelling subjects with symptoms of potential SARS-CoV-2 infection

Rapid chromatographic immunoassay-based evaluation of COVID-19: A 748 cross-sectional, diagnostic test accuracy study & its implications for COVID-19 management 749 in India

An optimized stepwise algorithm combining rapid antigen and RT-qPCR for 751 screening of COVID-19 patients

SARS-CoV-2 Rapid Antigen Testing of Symptomatic and Asymptomatic 753 Individuals on the University of Arizona Campus. Biomedicines, 2021

Testing of Healthcare Workers Exposed to COVID19 with Rapid Antigen 755 Detection. medRxiv

SARS-CoV-2 rapid antigen test: Fast-safe or dangerous? An analysis in the 757 emergency department of an university hospital

Five Antigen Tests for SARS-CoV-2: Virus Viability Matters. Viruses

Covid-19 antigen testing: better than we know? A test accuracy study. 762 Infectious Diseases

Diagnostic accuracy and utility of SARS-CoV-2 antigen lateral flow assays 764 in medical admissions with possible COVID-19

Use of Immunochromatographic SARS-CoV-2 Antigen Testing in Eight 767 Long-Term Care Facilities for the Elderly. Healthcare

Clinical evaluation of the Roche/SD Biosensor rapid antigen test with 769 symptomatic, non-hospitalized patients in a municipal health service drive-through testing 770 site. Emerging Infectious Diseases

Evaluation of three rapid lateral flow antigen detection tests for the 772 diagnosis of SARS-CoV-2 infection

Performance of the BinaxNOW COVID-19 Antigen Card test relative to the 774 SARS-CoV-2 real-time reverse transcriptase polymerase chain reaction assay among 775

Evaluation of accuracy, exclusivity, limit-of-detection and ease-of-use of 828 LumiraDx: An antigen-detecting point-of-care device for SARS-CoV-2

Comparison of the SARS-CoV-2 Rapid antigen test to the real star Sars-830

CoV-2 RT PCR kit

Pre-Operative SARS CoV-2 Rapid Antigen Test and Reverse Transcription 832

Polymerase Chain Reaction: A conundrum in surgical decision making

The evaluation of a novel digital immunochromatographic assay with 835 silver amplification to detect SARS-CoV-2

Diagnostic accuracy of SARS-CoV-2 rapid antigen detection testing in 838 symptomatic and asymptomatic children in the clinical setting

Evaluation of a Rapid Diagnostic Assay for Detection of SARS-CoV-2 841

Antigen in Nasopharyngeal Swabs

Clinical Performance of the Standard Q COVID-19 Rapid Antigen Test and 844 Simulation of its Real-World Application in Korea

Evaluation of the AMP SARS-CoV-2 rapid antigen test in a hospital setting

Performance of a SARS CoV-2 antigen rapid immunoassay in patients admitted 849 to the Emergency Department

Panbio antigen rapid test is reliable to diagnose SARS-CoV-2 infection in 852 the first 7 days after the onset of symptoms

Head-to-head comparison of SARS-CoV-2 antigen-detecting rapid test with 855 professional-collected nasal versus nasopharyngeal swab

Head-to-head comparison of SARS-CoV-2 antigen-detecting rapid test 858 with self-collected nasal swab versus professional-collected nasopharyngeal swab

Diagnostic accuracy and feasibility of patient self-testing with a SARS-861

Performance of a novel diagnostic assay for rapid SARS-CoV-2 antigen 863 detection in nasopharynx samples

SARS-CoV-2 Antigen Detection to Expand Testing Capacity for COVID-867 19: Results from a Hospital Emergency Department Testing Site. Diagnostics

Clinitest rapid COVID-19 antigen test for the diagnosis of SARS-869

CoV-2 infection: A multicenter evaluation study

Detection Test for the diagnosis of SARS-CoV-2 infection

Development and Potential Usefulness of the COVID-19

Diagnostic Assay in a Pandemic Context. Frontiers in Medicine

Point of Care Testing using rapid automated Antigen Testing for SARS-877 COV-2 in Care Homes -an exploratory safety, usability and diagnostic agreement evaluation

SARS-CoV-2 Antigen Rapid Immunoassay for Diagnosis of COVID-19 in the 880 Emergency Department

Multi-site assessment of rapid, point-of-care antigen testing for the diagnosis 882 of SARS-CoV-2 infection in a low-prevalence setting: A validation and implementation study. 883 Lancet Regional Health Western Pacific

Diagnostic Test in Uganda using Nasopharyngeal Samples

Anterior nasal versus nasal mid-turbinate sampling for a SARS-CoV-2 890 antigen-detecting rapid test: does localisation or professional collection matter?

SARS-CoV-2 rapid antigen test: High sensitivity to detect infectious virus

Performance Characteristics of BinaxNOW COVID-19 Antigen Card for 895

Screening Asymptomatic Individuals in a University Setting

Field performance of NowCheck rapid antigen test for SARS-CoV-2 in 898

Accuracy of a Novel SARS-CoV-2 Antigen-Detecting Rapid Diagnostic 902 Test from Standardized Self-Collected Anterior Nasal Swabs

Evaluation of two rapid antigen tests to detect SARS-CoV-2 in a hospital 905 setting

CoV-2 variants of concern

Usefulness of the antigen for diagnosing SARS-CoV-2 infection in 909 patients with and without symptoms. Enfermedades Infecciosas y Microbiología Clínica

Antigen-based testing but not real-time PCR correlates with SARS-CoV-2 912 virus culture

Performance of SARS-CoV-2 rapid antigen test compared with real-time RT-914 PCR in asymptomatic individuals

Diagnostic performance of CerTest and Panbio antigen rapid 917 diagnostic tests to diagnose SARS-CoV-2 infection

Comparative evaluation of Panbio and SD Biosensor antigen rapid 920 diagnostic tests for COVID-19 diagnosis

COVID-19: Rapid antigen detection for SARS-CoV-2 by lateral flow assay: A national 923 systematic evaluation of sensitivity and specificity for mass-testing. EClinicalMedicine

Comparative performance of SARS-CoV-2 lateral flow antigen tests and 926 association with detection of infectious virus in clinical specimens: a single-centre laboratory 927 evaluation study

Performance characteristics of a rapid SARS-CoV-2 antigen detection 929 assay at a public plaza testing site in San Francisco

Performance and Implementation Evaluation of the Abbott BinaxNOW 932 Rapid Antigen Test in a High-throughput Drive-through Community Testing Site

Performance and Operational Evaluation of the Access Bio CareStart 935 Rapid Antigen Test in a High-Throughput Drive-Through Community Testing Site

Open Forum Infectious Diseases

Evaluation of two fluorescence immunoassays for the rapid detection of SARS-938

CoV-2 antigen -new tool to detect infective COVID-19 patients

Evaluation of a novel antigen-based rapid detection test for the diagnosis of 940 SARS-CoV-2 in respiratory samples

SARS-CoV-2 rapid antigen testing for departing passengers at 943 Vancouver international airport

Rapid antigen detection testing for universal screening for severe 947 acute respiratory syndrome coronavirus 2 in women admitted for delivery

Clinical assessment of the Roche SARS-CoV-2 rapid antigen test. 950 Diagnosis

Immunochromatographic test for the detection of 952 SARS-CoV-2 in saliva

Diagnostic accuracy of rapid antigen tests in asymptomatic and 958 presymptomatic close contacts of individuals with confirmed SARS-CoV-2 infection: cross 959 sectional study

Diagnostic du Covid-19 en milieu ambulatoire. Revue médicale suisse

Low performance of rapid antigen detection test as frontline testing for 963 COVID-19 diagnosis

Evaluation of rapid antigen tests based on saliva for the detection of SARS-965

Evaluation of Two Rapid Antigenic Tests for the Detection of SARS-CoV-2 967 in Nasopharyngeal Swabs

Performance of a Rapid SARS-CoV-2 Antigen Detection Assay in 972 Symptomatic Children

Evaluation of the Abbott Panbio COVID-19 Ag rapid antigen test for the 974 detection of SARS-CoV-2 in asymptomatic Canadians. Diagnostic Microbiology and Infectious 975 Disease

The Value of Rapid Antigen Tests to Identify Carriers of Viable SARS-977

Sensitivity and specificity of lateral flow antigen test kits for covid-19 in 979 asymptomatic population of quarantine centre of province 3. Kathmandu University Medical 980

Clinical Evaluation of Sofia Rapid Antigen Assay for Detection of Severe 982

Acute Respiratory Syndrome Coronavirus 2 among Emergency Department to Hospital 983 Admissions

Usefulness of rapid antigen testing for SARS-CoV-2 screening of 985 healthcare workers: a pilot study

Self-testing for the detection of SARS-CoV-2 infection with rapid antigen tests 987 for people with suspected COVID-19 in the community

Clinical performance of the Abbott Panbio with nasopharyngeal, throat, 990 and saliva swabs among symptomatic individuals with COVID-19

Microbiology & Infectious Diseases

Performance of a Point-of-Care Test for the Rapid Detection of SARS-CoV-2 993

Diagnostic performance of a novel digital immunoassay

CoV-2): a prospective observational study with 1,127 nasopharyngeal samples

SARS-CoV-2 qRT-PCR Ct value distribution in Japan and 998 possible utility of rapid antigen testing kit. medRxiv

The evaluation of a newly developed antigen test

Ag) for SARS-CoV-2: A prospective observational study in Japan

Diagnostic performance and characteristics of anterior nasal collection for 1003 the SARS-CoV-2 antigen test: a prospective study

Antigen Detection Kit (Colloidal Gold) in the Detection of COVID-19. Diagnostics

Utility of Antigen-Based Rapid Diagnostic Test for Detection of SARS-CoV-2 1008 Virus in Routine Hospital Settings

Evaluation of a novel, rapid antigen detection test for the diagnosis of SARS-1010

Evaluation of a rapid antigen test for SARS-CoV-2 in symptomatic 1012 patients and their contacts: a multicenter study. medRxiv

Evaluation of a SARS-CoV-2 rapid antigen test: potential to help reduce 1016 community spread

Point-of-care evaluation of a rapid antigen test (CLINITEST) Rapid

Antigen Test) for diagnosis of SARS-CoV-2 infection in symptomatic and asymptomatic 1019 individuals

Real-life evaluation of a rapid antigen test (Panbio COVID-19 Ag Rapid Test 1021

Device) for SARS-CoV-2 detection in asymptomatic close contacts of COVID-19 patients. 1022 Clinical Microbiology and Infection

Real world clinical performance of SARS-CoV-2 rapid antigen tests in 1024 suspected COVID-19 cases in Taiwan

Clinical application of a rapid antigen test for the detection of SARS-CoV-2 1027 infection in symptomatic and asymptomatic patients evaluated in the emergency 1028 department: a preliminary report

Comparison of five SARS-CoV-2 rapid antigen detection tests in a 1033 hospital setting and performance of one antigen assay in routine practice: a useful tool to 1034 guide isolation precautions

Diagnostic Accuracy of the Panbio SARS-CoV-2 Antigen Rapid Test 1036 Compared with Rt-Pcr Testing of Nasopharyngeal Samples in the Pediatric Population. The 1037 Journal of Pediatrics

Comparative evaluation of four rapid SARS-CoV-2 antigen detection tests 1039 using universal transport medium. Travel Medicine and Infectious Disease

SARS-CoV-2 rapid antigen detection tests

Diagnostic Efficacy of Rapid Antigen Testing for SARS-CoV-2: The 1044

SARS-CoV-2 Diagnostic Tests: Algorithm and Field Evaluation From the Near 1046

Clinical evaluation of BD Veritor SARS-CoV-2 point-of-care test performance 1050 compared to PCR-based testing and versus the Sofia 2 SARS Antigen point-of-care test

Foundation for Innovative New Diagnostics, FIND Evaluation of RapiGEN Inc. BIOCREDIT 1055 COVID-19 Ag. External Report Version 2.1

Foundation for Innovative New Diagnostics, FIND Evaluation of Guangzhou Wondfo Biotech 1066 Co., Ltd Wondfo 2019-nCoV Antigen Test (Lateral Flow Method)

Foundation for Innovative New Diagnostics, FIND Evaluation of Mologic Ltd, COVID 19 RAPID 1075 ANTIGEN TEST

Foundation for Innovative New Diagnostics

ActivXpress+ COVID-19 Antigen Complete Testing Kit

Foundation for Innovative New Diagnostics, FIND Evaluation of Hotgen; Novel Coronavirus 1086 2019-nCoV Antigen Test (Colloidal Gold)

Foundation for Innovative New Diagnostics, FIND Evaluation of SD Biosensor, Inc.; STANDARD 1097 Q COVID-19 Ag Test, nasal swab

SARS-CoV-2 patient self-testing with an antigen-detecting rapid test: a 1101 head-to-head comparison with professional testing. medRxiv

Assessment of the analytical sensitivity of ten lateral flow devices against 1104 the SARS-CoV-2 omicron variant

Viral Dynamics of SARS-CoV-2 Variants in Vaccinated and Unvaccinated 1108 Persons

Viral dynamics and duration of PCR positivity of the SARS-CoV-2 Omicron 1110 variant. medRxiv

Discordant SARS-CoV-2 PCR and Rapid Antigen Test Results When 1112 Infectious: A December 2021 Occupational Case Series. medRxiv

Saliva swabs are the preferred sample for Omicron detection

Direct Comparison of SARS CoV-2 Nasal RT-PCR and Rapid Antigen Test 1117 (BinaxNOW(TM)) at a Community Testing Site During an Omicron Surge. medRxiv

Data sharing in public health emergencies: a call to researchers. Bull World 1122 Health Organ

Developing Global Norms for Sharing Data and Results during Public 1124 Health Emergencies

Rethinking Covid-19 Test Sensitivity -A Strategy for 1126 Containment