Development and validation of a highly sensitive and specific electrochemical assay to quantify anti-SARS-CoV-2 IgG antibodies to facilitate pandemic surveillance and monitoring of vaccine response.

Amperial™ is a novel assay platform that uses immobilized antigen in a conductive polymer gel followed by an electrochemical detection. A highly specific and sensitive assay was developed to quantify levels of IgG antibodies to SARS-CoV-2 in saliva. After establishing linearity and limit of detection we established a reference range of 5 standard deviations above the mean. There were no false positives in 667 consecutive saliva samples obtained prior to 2019. Saliva was obtained from 34 patients who had recovered from documented COVID-19 or had documented positive serologies. All of the patients with symptoms severe enough to seek medical attention had positive antibody tests and 88% overall had positive results. We obtained blinded paired saliva and plasma samples from 14 individuals. The plasma was analyzed using an EUA-FDA cleared ELISA kit and the saliva was analyzed by our Amperial™ assay. All 5 samples with negative plasma titers were negative in saliva testing. Eight of the 9 positive plasma samples were positive in saliva and 1 had borderline results. A CLIA validation was performed as a laboratory developed test in a high complexity laboratory. A quantitative non-invasive saliva based SARSCoV-2 antibody test was developed and validated with sufficient specificity to be useful for population-based monitoring and monitoring of individuals following vaccination.

CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint Introduction 41 A novel corona virus, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), 42 has caused a global pandemic causing major disruptions world-wide (1). Multiple high-43 throughput PCR based tests have been developed that are reasonably sensitive and specific, 44 however the same cannot be said for antibody testing, prompting The Center for Disease Control 45 (CDC) to issue guidelines entitled "Interim Guidelines for COVID-19 Antibody Testing" (2). 46 This publication describes the variability of in-home antibody tests and the lack of specificity 47 required to make home-based antibody testing a valuable tool for epidemiologic surveillance. 48 Having a reliable self-collection antibody test may be of enormous help in epidemiologic 49 studies of background immunity, testing symptomatic individuals without RNA based testing 50 during their acute illness, and screening health care providers and first responders to establish 51 prior COVID-19 infection. Such a test may also be valuable in following vaccinated patients to 52 assess the kinetics of anti-SARS-CoV-2 antibody production following inoculation. Multiple 53 serological tests based on serum or plasma have been developed and marketed, with ELISA and 54 lateral flow methods predominating. However, many methods suffer from low sensitivities and 55 specificities (2-6). 56 Antibodies begin appearing in the first week following the development symptoms. IgG, 57 IgM, and IgA are detectable with IgA appearing somewhat earlier than IgG and IgM. Most 58 patients seroconvert by 2 weeks following symptoms. Unlike IgA and IgM, IgG persists for 59 several months following infection (7)(8)(9). 60 In a published study of 1,797 Icelandic individuals recovered from qPCR documented 61 COVID-19 disease, 91% were IgG seropositive and antibody levels remained stable for 4 months 62 after initial symptoms (10). Notably 2.3% of individuals quarantined due to exposure but 63 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ;https://doi.org/10.1101https://doi.org/10. /2020 Page 4 of 33 untested for virus, with negative qPCR results, tested positive for IgG antibodies. Of 18,609 64 patients who were both unexposed and asymptomatic, the seropositivity rate was 0.3% (11). 65 Since health care systems are burdened with care for COVID-19 patients, having a test 66 that does not require phlebotomy would be extremely beneficial. To that end, investigations have 67 been carried out using home finger prick blood sampling and even some home blood spot testing 68 lateral flow strips (5-7). However, home finger stick is invasive and not acceptable to some 69 individuals, and requires a health care professional to administer the test to vulnerable 70 individuals such as the elderly and children. In addition, home blood collection tests are less 71 accurate than phlebotomy, with specificities less than 98%. In a low prevalence disease, the 72 positive predictive value for a test with 98% specificity is less than 50% (7, 11). 73 Saliva is an oral fluid that is obtained easily and non-invasively. Proteomic studies show 74 that the immunoglobulin profile in saliva is nearly identical to that of plasma (12). Therefore, 75 saliva is an excellent medium for COVID-19 antibody measurement. There are several 76 commercially available collection devices to facilitate saliva collection, stabilization of IgG, and 77

transport. 78
A recently published study demonstrated excellent correlation between levels of COVID-79 19 antibodies in serum and saliva (13). In order to be useful in population-based screening and 80 to determine individual immunity in exposed populations, a SARS-CoV-2 antibody test must be 81 highly specific because of the low seroprevalence rate in the population (2,14). In addition, the 82 ability to quantify antibody levels is important for vaccine development and in monitoring for 83 waning immunity (2,14). The only published saliva based assay for SARS-CoV-2 antibodies 84 had only 89% sensitivity with 98% specificity (13), leading to a positive predictive value of only 85 49% in a population with a 2% prevalence of COVID-19 exposure. 86 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ;https://doi.org/10.1101https://doi.org/10. /2020 Page 5 of 33 Our goal was to develop a non-invasive saliva based quantitative test for COVID-19 87 antibodies with exquisite sensitivity. We reviewed existing literature to find the SARS-CoV-2 88 antigen domain with the highest specificity and the ability to distinguish between the COVID-19 89 virus and other related Coronaviruses. The S1 domain is the most specific in terms of cross 90 reactivity with other Corona and other respiratory viruses. As recombinant S1 antigen is readily 91 available from at least 2 vendors, we chose the S1 antigen for our assay development. 92 Levels of IgM and IgA deteriorate rapidly following recovery from COVID-19 infection; 93 IgG levels remain detectable for several weeks to months (10). Since the intended use of our 94 assay is for population-based screening and vaccine efficacy monitoring, we chose to assay IgG 95 only. 96 The Amperial™ technology, formerly known as Electric Field Induced Release and 97 Measurement (EFIRM™), is a novel platform capable of performing quantitation of target 98 molecules in both blood and saliva (13-16). We developed quantitative Amperial™ assays for 99 IgG, IgM, and IgA antibodies to the S1 spike protein antigen of SARS-CoV-2. This test is highly 100 sensitive (>88%) and specific (>99.85%) for patients with COVID-19 infections and correlates 101 well with plasma ELISA analysis. The unique assay described in this article is completely non-102 invasive, allows home-collection, is quantitative, and has shown no false positives in 667 103 unexposed individuals, leading to a specificity of at least 99.6%. The widespread use of this test 104 may be of great value in identifying individuals with prior exposure to SARS-CoV-2, to follow 105 patients longitudinally to determine the kinetics of diminishing antibody concentration, and may 106 be of special value in the longitudinal monitoring of vaccinated individuals to assess continued 107 serologic immunity. 108 109 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10. 1101/2020

Materials and Methods 110
The schematic of the Amperial™ SARS-CoV-2 IgG antibody is shown in Figure 1. The 111 principle of the Amperial™ platform is that a biomolecule (in this case SARS-CoV-2 Spike 112 protein S1 antigen) is added to a liquid pyrrole solution that is then pipetted into the bottom of 113 microtiter wells containing a gold electrode at the bottom of each well. After the solution is 114 added to each well, the plate is placed into the Amperial™ Reader and subjected to an electric 115 current leading to polymerization. This procedure results in each well becoming coated with a 116 conducting polymer gel containing the S1 antigen. Following the polymerization, diluted saliva, 117 plasma, or serum is added to the well. Specific anti-S1 antibodies bind to the S1 antigen in the 118 polymer. After rigorous washing procedures, the bound antibody is detected by using 119 biotinylated anti-human IgG and then the signal is amplified by a standard streptavidin / 120 horseradish peroxidase reaction that produces an electric current measured by the Amperial™ 121 Reader in the nanoampere (nA) scale. The instrument is capable of accurately measuring current 122 in the picoampere (pA) range, so the measurement is well within the ability of the instrument 123 (13)(14)(15)(16). The measurement of current rather than optical absorbance, as is done in the typical 124 ELISA, has two important advantages over standard ELISA. Firstly, it allows precise 125 quantitation of the amount of bound antibody and secondly, the measurement of current rather 126 than optical absorbance allows increased sensitivity. Since antibody levels in saliva are lower 127 than in plasma (13,15), this increased sensitivity is crucial. The precise details of the assay are 128 described in the next paragraph. 129 COVID-19 Spike-1 Antigen (Sanyou-Bio, Shanghai, China) was diluted to a 130 concentration of 6.25 µg / mL, added to each well of the microtiter plate, and co-polymerized 131 with pyrrole (Sigma-Aldrich, St. Louis, MO) onto the bare gold electrodes by applying a cyclic 132 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10. 1101/2020 Page 7 of 33 square wave electric field at 350 mV for 1 second and 1100 mV for 1 second. In total, 133 polymerization proceeded for 4 cycles of 2 seconds each. Following this electro-polymerization 134 procedure, 6 wash cycles were performed using 1x PBS with 0.05% Tween-20 (PBS-T) using a 135 96-channel Biotek 405LS plate washer programmed to aspirate and dispense 400 µL of solution 136 per cycle. 137 Following the application of the polymer layer, 30 µL of saliva diluted at a 1:10 ratio in 138 Casein/PBS (Thermo-Fisher, Waltham, MA) was pipetted into each well and incubated for 10 139 minutes at room temperature. Unbound components were removed by performing 6 wash cycles 140 of PBS-T using the plate washer. 141 Biotinylated anti-human IgG secondary antibody (Thermofisher, Waltham, MA) at a 142 stock concentration of 1.5 mg / mL was diluted 1:500 in Casein/PBS and 30 µL pipetted to the 143 surface of each well and incubated for 10 minutes at room temperature followed by 6 wash 144 cycles using PBS-T. Subsequently, 30 µL of Poly-HRP80 (Fitzgerald Industries, Acton, MA) at 145 a stock concentration of 2 µg / mL was diluted 1:25 in Casein/PBS, added to the wells, and 146 incubated at 10 minutes at room temperature. Following a final wash using 6 cycles of PBS-T, 147 current generation is accomplished by pipetting 60 µL of 1-Step Ultra TMB (Thermofisher,148 Waltham, MA) to the surface of the electrode and placing the plate into the Amperial™ reader 149 where current is measured at -200 mV for 60 seconds. The current in nA is measured 3 times for 150 each well. The process for reading the entire 96 well plate requires approximately 3 minutes. 151

Plasma Quantitative Amperial™ Assay for SARS-CoV-2 IgG 152
The protocol is similar to the Amperial™ SARS-CoV-2 IgG antibody for saliva samples. 153 Following the application of the polymer layer, 30 µL of plasma diluted at a 1:100 ratio in 154 Casein/PBS (Thermo-Fisher, Waltham, MA) was pipetted into each well and incubated for 10 155 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10.1101/2020.11.12.20230656 doi: medRxiv preprint minutes at room temperature. The standard curve for plasma contains the following points: 300 156 ng / ml, 150 ng / ml, 75 ng / ml, 37.5 ng / ml, 18.75 ng / ml, and 0 ng / ml. 157

Plasma SARS-CoV-2 ELISA Assay 158
We purchased FDA EUA ELISA kits EUROIMMUN Anti-SARS-CoV-2 ELISA Assay 159 for detection of IgG antibodies (EUROIMMUN US, Mountain Lakes, NJ). We processed 160 samples exactly as described in the package insert. consists of an absorbent pad on the end of a scored plastic wand. The individual places the pad 178 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10. 1101/2020 Page 9 of 33 between cheek and gum for a period of 2 -5 minutes. Subsequently the wand and pad are placed 179 into a tube containing transport medium, the top of the stick is broken off, and the tube is sealed 180 for transport. The sealed tube is placed into a zip-lock bag and shipped by any standard method. 181 According to the package insert, samples are stable at ambient temperature for 21 days (see 182 results below and Orasure™ website). An alternate sample collection method involves the 183 individual swabbing the pad 4 times in the gingival tooth junction prior to placing the pad 184 between the cheek and gum. This method has been shown to improve IgG yield in some patients 185 with low antibody levels (personal communication). 186 Samples collected pre-2012 were used as controls. Saliva was collected from healthy 187 individual volunteers at meetings of the American Dental Association between 2006 and 2011. 188 Consent was obtained under IRB approval UCLA IRB #06-05-042. Both male and females, 189 mostly non-smokers, 18-80 years of age, and a differing ethnicities were included. All subjects 190 were consented prior to collection. Each subject expectorated ~ 5 mL of whole saliva in a 50cc 191 conical tube set on ice. The saliva was processed within 1/2 hour of collection. Samples were 192 spun in a refrigerated centrifuge @ 2600 X g for 15 minutes at 4°C. The supernatant (cell-free 193 saliva) was then pipetted into two-2 mL cryotubes and 1.1 µL Superase-In (Ambion, Austin, TX) 194 was added as a preservative. Each tube was inverted to mix. The samples were frozen in dry ice 195 and later stored in -80°C. 196

Results 197
Linearity 198 Figure 2 demonstrates the dynamic range and linearity of the assay. In these experiments 199 varying amounts of monoclonal human anti-S1 IgG was added to a saliva sample from a healthy 200 volunteer and subjected to the assay. Figure 2A shows a range of 0.2 to 6 ng/ml. The Y-axis 201 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10. 1101/2020 Page 10 of 33 shows amperage measured in nA. The X-axis represents spike-in concentrations of IgG. The 202 assay begins to become saturated at about 3 ng / ml. Panel 2B shows dilutions down to 0.03 ng / 203 ml to 0.6 ng / ml and shows linearity in that range. This allows us to create a standard curve 204 containing the following points: 3 ng / ml, 1.5 ng / ml, 0.75 ng / ml, 0.375 ng / ml, 0.1875 ng / 205 ml, and 0 ng / ml. 206

Inhibition Assay 207
In order to demonstrate the specificity for the assay on actual clinical samples, we used 208 the saliva from 3 recovered patients who had high levels of SARS-CoV-2 antibodies and added 209 exogenous S1 antigen in varying amounts prior to analysis on the Amperial™ assay. The 210 exogenous S1 antigen should compete for binding sites and therefore extinguish the nA signal. 211 Figure 3 shows the results of this experiment. The red, purple, and green represent 3 different 212 patients. The X-axis demonstrates increasing concentration of exogenous S1 added to the saliva 213 before subjecting it to the assay. As shown, saliva pre-incubated with S1 antigen extinguishes the 214 detectable IgG signal proportionately, therefore demonstrating the specificity of the assay to S1 215 antigen in clinical samples. 216

Matrix Effects 217
Since we are be comparing samples collected by various methods, it is vital to determine 218 if any significant matrix effects could interfere with data interpretation. We examined the 3 219 different collection methods used in this study: Expectoration/centrifugation, Orasure™ without 220 swabbing and Orasure™ with swabbing. 221 Two methods of collection using the Orasure™ Oral Fluid Collection Device were tested. 222 The first method (non-swabbing) collects saliva by placing an absorbent pad into the lower gum 223 area for 2-5 minutes and then placing the saturated collection pad into a preservative collection 224 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10. 1101/2020 Page 11 of 33 tube. The second method (swabbing) adds the step of first gently rubbing the collection pad 225 along gum line, between the gum and cheek, 5 times, before placing the device in the lower gum 226 area for 2-5 minutes, and then immersing the saturated collection pad into the collection tube. 227 Healthy donors (n=5) collected their saliva using these two different methods. The control pre-228 2012 samples were collected with an expectoration protocol for whole saliva collection (falcon 229 tubes), processing (centrifuge), stabilization, and storage. Five samples collected by each of the 3 230 methods and were analyzed in duplicate. The results are shown in Figure 4 under the heading 231 "No spike in." There are no differences among 3 sample types. We then added monoclonal 232 human anti-S1 IgG to each sample and again ran them in duplicate ( Figure 4) above caption 233 Spike-in 1.5 ng / ml IgG. A non-parametric Student t-test was performed with no significant 234 differences between any of the collection methods. 235

Stability 236
The Orasure™ collector is an FDA-cleared device for the analysis of anti-HIV IgG. The 237 package insert describes a 21-day stability at ambient temperature. We wished to establish the 238 stability of anti COVID-19 IgG using this collector. Passive whole saliva was collected from 239 four healthy individuals using 50 mL falcon tubes and spiked with anti-Spike S1 IgG to reach a 240 final concentration of 300 ng / ml. Aliquots of 1.75 mL of saliva were placed into 50 mL tubes 241 and then the sponge of the Orasure™ collector was submerged into the saliva for five minutes 242 and processed as described in Methods. The collected saliva was then aliquoted into PCR tubes 243 and left at ambient temperature (21 o C) for 0, 1, 3, 7, and 14 days before storage at -80 o C. After 244 14 days, samples were thawed and assayed using the anti-Spike S1 IgG Amperial™ assay to 245 assess stability. At 14 days, 95% of the original signal remained, demonstrating the 14-day 246 stability of anti-SARS-CoV-2 antibodies collected in Orasure™ containers (data not shown). 247 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Specificity and Reference Range 248
Once we established no significant differences between the tube collection method and 249 the Orasure™ collector method, we analyzed a series of 667 samples collected between 2006 and 250 2009 at the annual meeting of the American Dental Association. Scatter plots of these data for 251 both nA and ng / ml are shown in Figure 5A and 5B. We established the mean and standard 252 deviation for both raw nA values and concentration in ng / ml. In order to maximize specificity, 253 we selected a reference range > 5 SD above the mean. A 5 sigma level would lead to a specificity 254 of 99.9994%. In fact, we have never seen a healthy sample above the 5 sigma level. As will be 255 seen, the sensitivity of the assay remains greater than 88% even with this rigorous specificity. 256  Figure 7 demonstrates the relationship of anti-S1 IgG levels to severity of symptoms. 267 Table 1 is a tabular summary of these data. All patients who had severity indexes ≥3 (sought 268 medical attention, admitted to hospital, admitted to ICU, on ventilator) had positive antibody 269 levels. Although 4 patents with mild symptoms had antibody levels in the normal range, both 270 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review) preprint
The copyright holder for this this version posted November 16, 2020. ;https://doi.org/10.1101https://doi.org/10. /2020 Page 13 of 33 asymptomatic patients had appreciable antibody levels. These patients were close contacts of 271 more severely affected patients. The highest antibody level recorded is severity index level 2 272 patient (moderate symptoms, did not seek medical care). It is important to note that both 273 asymptomatic patients had easily detectable antibody levels in saliva, suggesting this test may be 274 useful in general population screening. 275

Paired Saliva and Plasma Samples 276
We obtained 14 paired, blinded plasma and saliva samples. The plasma was analyzed by 277 an FDA EUA-cleared ELISA test purchased from EUROIMMUN (see Methods). The saliva 278 samples, collected in Orasure™ buffer, were analyzed by the Amperial™ assay described in 279 Methods. After unblinding, we discovered 8 recovered COVID patients and 5 healthy patients in 280 this series. All 5 healthy patients were negative in both the saliva and plasma assays. In 7 of the 8 281 recovered patients, both plasma and saliva tests were positive. There was one sample with a 282 discrepancy between saliva and plasma, with the plasma positive and the saliva in the 283 indeterminate range. 284 The EUROIMMUNE ELISA assay is a semi-quantitative assay and yields an absorbance 285 ratio rather than a quantity. Figure 8 demonstrates the relationship between the saliva 286 quantitative results and plasma absorbance ratio for the paired plasma and saliva samples. There 287 is a clear relationship between the 2 levels, with the higher plasma absorbance ratios associated 288 with higher saliva quantitation. 289 We developed a research quality assay to quantify anti-SARS-CoV-2 IgG levels in 290 plasma (see Methods). We analyzed the 13 plasma samples using this assay. The results of this 291 experiment are shown in Figure 9. Panel A shows a log / log plot of plasma versus saliva levels 292 showing a clustering with high plasma levels associated with high saliva levels. Panel B shows 293 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review) preprint
The copyright holder for this this version posted November 16, 2020. ;https://doi.org/10.1101https://doi.org/10. /2020 Page 14 of 33 the box plot of these values, demonstrating that plasma levels are approximately 50X those of 294 saliva. This observation explains the necessity for an extremely sensitive assay such as the 295 Amperial™ assay in order to detect antibodies in saliva. Of note, the publication regarding saliva 296 SARS-CoV-2 IgG detection reports levels of 25 -60 mcg / ml, 1000 times less sensitive than our 297 assay. 298

Longitudinal Tracking of Antibody Levels 299
Three of our volunteers supplied samples at weekly intervals so we could determine the 300 stability of their antibody levels. Results appear in Figure 10. The 5 standard deviation cutoff is 301 again shown with the dashed green line. All 3 patients continued to have detectable levels for 302 more than 12 weeks, with the longest interval of 15 weeks. All tests were positive in all patients 303 and antibody levels in all 3 patients remained clearly positive during the time interval studied. 304 Patients C1 and C3 seem to have a rise in antibody level between 11 and 12 weeks post initial 305 symptoms followed by a return to baseline level. Patient C2 might also have had a spike in 306 antibody levels at 10 weeks. This may be result of the amnestic B-cell population becoming 307 established. There is insufficient data at this time to determine if this is a generalized pattern. 308

CLIA Evaluation 309
We performed a full CLIA laboratory developed test evaluation for the Amperial™ 310 COVID-19 IgG Antibody test. The validation assayed 72 unaffected patients and 30 recovered 311 patients and demonstrated 100% sensitivity and specificity. The intra-assay and inter-assay 312 variability were 9.28% and 16.2% respectively. 313

Discussion 314
We have developed an exquisitely specific, sensitive, non-invasive saliva based 315 quantitative assay for anti-SARS-CoV-2 IgG antibodies. Our goal was to create a quantitative 316 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10. 1101/2020 Page 15 of 33 assay with sufficient positive predictive value to be useful to inform individuals regarding 317 previous infection with COVID-19. By establishing a reference range of 5 sigma above than the 318 mean we have a theoretical analytical specificity of 99.9999994%. We plan to repeat the analysis 319 of all positive samples to further increase analytical specificity. Since our test is non-invasive 320 with home-collection we can also offer repeat testing on a second sample to further increase 321 specificity. These procedures will minimize the false positives due to purely technical issues. 322 There is still the possibility of biological false positives, however, due to cross reactivity with 323 other infectious or environmental agents. The S1 antigen appears to be specific for SARS-CoV-2 324 (2, 3, 10) and in our series of 667 samples collected prior to 2019 we observed no false positive 325 results. 326 Many investigations of neutralizing antibodies use antibodies directed to a different 327 epitope, the Ribosomal Binding Domain (RBD). Therefore, we tried to assay the RNA binding 328 domain (RBD) but found a false positive in the initial 10 unaffected controls indicating 329 significant cross reactivity between the RBD and other viral species, disqualifying RBD for our 330 purposes. 331 We cannot predict the eventual clinical specificity of this assay. At a minimum, the 332 specificity is 667 / 668 or 99.985% assuming the next control sample tested would be a false 333 positive, but the specificity is likely to be higher. Our current sensitivity is 100% for patients 334 with symptoms severe enough to seek medical care. For all patients, including mildly 335 asymptomatic patients, our clinical sensitivity is 88%. Since the Amperial™ assay only requires 336 6 µL of collection fluid, several assays can be performed from the same sample. This allows all 337 positives to be repeated to confirm the positive results, This further increasing the specificity of 338 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ;https://doi.org/10.1101https://doi.org/10. /2020 Page 16 of 33 the assay. We will offer testing of a second, independent sample for all patients testing positive. 339 Since saliva collection is easily be performed at home, obtaining a second sample is not difficult. 340 For any laboratory test, the PPV is proportional to the prevalence of positivity in the 341 population. A recent study demonstrated a prevalence of between 4.4% to 6% in Britain (16). 342 Using the minimum specificity of 99.985% and a prevalence of 6% the Amperial™ saliva assay 343 would have a minimum PPV of 96%. In contrast, a published saliva antibody detection assay 344 reported a specificity of 98% with a similar sensitivity (89%). This specificity leads to PPV of 345 only 69% making it an ineffective tool for population screening. 346 Our data demonstrate that the Imperial™ assay is appropriate for longitudinal screening 347 of antibody levels, a particular utility in vaccine trials and in population monitoring following 348 mass immunization. Since this assay is quantitative and levels appear to be stable with time, 349 patients may be monitored from home at frequent intervals. If antibodies raised in response to 350 vaccination do not include IgG antibodies to S1 antigen, it is easy to rapidly develop Amperial™ 351 antibody tests to any antigen. This requires adding the new antigen to the pyrrole solution and 352 does not require significant alteration of assay conditions. 353 A particular advantage of this assay is convenience. The Orasure™ collector is simple 354 and easy to use and does not require professional monitoring for adequate collection. Home 355 collection relieves the burden to an already stressed health care system. Vulnerable populations 356 such as children and the elderly can be guided through the collection process by parents or other 357 adults. It is possible to obtain repeat samples to confirm positives and to perform longitudinal 358 testing since the only requirement for testing is shipping th collecting kit. 359 The Amperial™ IgG test is plate-based and high-throughput. An entire plate is easily 360 processed in 2 hours, leading to rapid turnaround time once the sample enters the laboratory. 361 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ;https://doi.org/10.1101https://doi.org/10. /2020 Page 17 of 33 There is no pre-processing of the sample required; samples are taken directly from the collection 362 vial and placed into the assay. With standard liquid handlers, the assay may be easily automated 363 allowing for extremely high-throughput since the Amperial™ reader is only required for the 364 polymerization step of less than a minute at the beginning of the assay and 3 minutes for the 365 measurement phase at the end of the assay. 366 Published data (13) and our own demonstrate a correlation between blood results and 367 saliva results indicating that the IgG present in saliva is most likely derived from the plasma 368 through filtration. Our data shows that saliva IgG levels are approximately 50-fold less than 369 those in plasma necessitating a highly sensitive assay in order to detect the IgG levels in saliva. 370 There is some discussion in the literature of the role antibody testing may have in 371 managing the COVID-19 epidemic. Alter and Seder published an editorial in the New England 372 Journal of Medicine arguing, "Contrary to recent reports suggesting that SARS-CoV-2 RNA 373 testing alone, in the absence of antibodies, will be sufficient to track and contain the pandemic, 374 the cost, complexity, and transient nature of RNA testing for pathogen detection render it an 375 incomplete metric of viral spread at the population level. Instead, the accurate assessment of 376 antibodies during a pandemic can provide important population-based data on pathogen 377 exposure, facilitate an understanding of the role of antibodies in protective immunity, and guide 378 vaccine development. (14)" 379 In this article, we describe the development of a non-invasive, home collection based, 380 exquisitely specific, and acceptably sensitive test for the presence of anti-SARS-CoV-2 381 antibodies in saliva. This may be an important tool in controlling the pandemic and facilitating 382 and understanding of the role of antibody production in COVID-19 immunity. Longitudinal 383 monitoring of anti-SARS-CoV-2 IgG levels could also play a valuable role in vaccine 384 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint  CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10.1101/2020.11.12.20230656 doi: medRxiv preprint CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10. 1101/2020 Page 25 of 33 487 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10.1101/2020.11.12.20230656 doi: medRxiv preprint CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10.1101/2020.11.12.20230656 doi: medRxiv preprint CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10.1101/2020.11.12.20230656 doi: medRxiv preprint CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10.1101/2020.11.12.20230656 doi: medRxiv preprint Figure 6. Amperial™ detection of anti-Spike S1 IgG in saliva of COVID-19 (n=34) and healthy 504 subjects (n=667). Green dashed line indicates 5 SD reference range cutoff. 505 506 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10.1101/2020.11.12.20230656 doi: medRxiv preprint Figure 7. Clinical severity index and anti-Spike S1 IgG level in saliva. 508 509 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted November 16, 2020. ; https://doi.org/10. 1101/2020