key: cord-0845520-dq28i7gu
authors: Vicario, José; Alenda-Asensi, Raquel; Lucea-Gallego, Irene; Sánchez-Bonilla, Javier; Saez, Rebeca; Clayton, Reginald; Robson, David; Robb, Janine; Barea-Garcia, Luisa
title: Evaluation of the Performance of the Quotient MosaiQTM COVID-19 Professional Use Microarray Assay for the Detection of Antibodies to SARS CoV-2 in a Clinical Setting
date: 2021-05-24
journal: J Clin Virol Plus
DOI: 10.1016/j.jcvp.2021.100022
sha: 2c46cd32f91d62ec5593d17494be537f30e2d134
doc_id: 845520
cord_uid: dq28i7gu

nan

As the SARS-CoV-2 pandemic progresses and public and patient health providers seek advice and guidance to inform health policy, the value of serological testing in establishing infection rates and correlates of immune protection will become elucidated. The detection of antibody responses may enable post-diagnosis management, the identification of asymptomatically infected individuals, and assessment of pandemic spread on a population or cohort basis. Serological testing may also be useful in the formulation of lock-down or exit strategies, safe return to work, assessing previous infection rates, confirmation of infection, use of convalescent (passive antibody therapy) plasma and assessment of vaccination programmes.

Antibody testing can provide valuable information to guide healthcare providers and inform the management of patients. In an ELISA based method, SARS-CoV-2-specific antibodies were detectable in approximately 40% of COVID-19 patients at seven days after the onset of symptoms, and seroconversion rates exceed >90% at the comparatively early time of day 14 [1] . Recent studies demonstrated antibody testing to be more sensitive than nucleic acid detection by PCR after approximately eight days of COVID-19 illness duration [2] . In combination, both PCR and antibody tests will enable for accurate diagnosis [3] at and around the time of primary infection, with antibody detection more pertinent for the later stages of the disease where viral replication or persistence is reduced, or the virus has been cleared [4] .

Furthermore, antibody testing enables the identification of individuals demonstrating a detectable adaptive immune response subsequent to infection that may prevent re-infection or reduce the severity of a re-infection [5] , and also enable an assessment of epidemiological exposure and developing herd immunity in the population.

This field utility study, performed in an active hospital setting, was conceived and conducted to demonstrate the operational performance of the assay (sensitivity and specificity): the MosaiQ TM COVID-19 Antibody Microarray for testing of human serum, plasma or blood samples for the presence or absence of antibodies to SARS-CoV-2. We also report the assay sensitivity and specificity in specific cohorts over time, where samples were obtained post PCR diagnosis and time to clearance of the virus were considered.

The study protocol was reviewed and approved by appropriate senior management at the test facility before commencement of the study.

In order to assess sensitivity, 100 samples were obtained from convalescent patients for which infection with the SARS-CoV-2 virus had been proven by PCR real time detection kits from Thermo Fisher Scientific, Roche and CerTest Biotec. This testing was conducted at health centres within Comunidad de Madrid, Spain. Samples were obtained from patients across the spectrum of disease; none of the patients had been hospitalised. Criteria for eligibility of sample inclusion was a PCR test (typically nasopharyngeal swab) positive result indicating infection. Samples where no PCR data was available were excluded.

Ethical approval: Written informed consent was obtained from all participants for apheresis procedures.

COVID-19 convalescent plasma banking had been approved by the ethical committee of the Hospital Ramon y Cajal in Madrid, Spain.

To test the specificity of the assay, 404 EDTA plasma samples from healthy blood donors, presumed negative for SARS-CoV-2, were tested. These samples were collected and cryopreserved in 2015, and therefore the negative criterion was they were from a population cohort prior to the pandemic and consequently negative for antibodies to SARS-CoV-2.

All samples were brought to room temperature at least one hour prior to testing. Plasma and serum samples were gently mixed and centrifuged prior to testing.

Testing was performed over three days. All samples were tested on the Quotient MosaiQ system as per instructions from the manufacturer. All positive sample testing was completed in one day, with all other samples tested over a further two days; sample loading was dependent on operator availability and not a limitation of the system. The time to first result on the system was 35 minutes, with each subsequent result being generated every 24 seconds, processing 125 samples per hour. Microarray. An incubation step enables antibodies (if present) to bind to the printed antigens on the glass and is followed by removal of unbound antibodies using a wash buffer. This is then followed by addition of a detection reagent containing a gold-conjugated secondary antibody formulation that binds to human IgG and IgM. Excess detection reagent is removed by washing and an additional wash precedes the final stages of the assay. Finally, the addition of enhancement reagent provides a soluble silver solution, which nucleates on the gold nanoparticles and increases the diameter of the gold nanoparticle. Subsequent washing with purified water reveals a readily detectable reaction by reflection of incident light during image acquisition and analysis by the instrument.

The instrument optical system imaged and analysed each microarray and delivered a positive result (reactive) or negative result (non-reactive) accordingly.

Qualitative detection of IgG/IgM antibodies against recombinant nucleocapsid SARS-CoV 

SARS-CoV-2 spike protein antigen was coated onto a 96-well microplate through passive adsorption. The microplate was then blocked using a commercial blocking buffer containing BSA. Samples and positive and negative controls were then added to the microplate and incubated. Following a wash step to remove any unbound material, the microplate was incubated with a secondary antibody labelled with horse radish peroxidase (HRP). The secondary antibody bound to any antibody that has attached to the microplate well. Anti-IgG and Anti-IgM secondary antibodies were used separately. Following another wash step, tetramethylbenzidine (TMB) substrate was added. Any bound secondary antibody HRP reacts with the TMB to generate a signal. The microplate was read using a standard spectrophotometer.

Positive and negative control samples were provided in kit form by Quotient and were used with the MosaiQ COVID-19 Antibody Microarray assay. Acceptable control results were achieved before proceeding with testing of samples.

Data analysis -sensitivity 7 | P a g e Sensitivity was defined as the percentage of expected positive samples reactive on the test system, samples being from patients previously diagnosed with SARS-CoV-2 infections using nucleic acid detection (RT-PCR) of SARS-CoV-2.

Specificity was defined as the percentage of all negative samples from 2015 (pre-SARS-CoV-2) identified as non-reactive for SARS-CoV-2.

Confidence interval was determined using an online tool. 95% was used as the desired level of confidence.

For the determination of assay specificity, an initial panel of 404 negative samples were tested. 401 samples returned an expected negative result, and the remaining three samples were investigated for potential discordance or other factors. Two samples in the ostensibly negative cohort were excluded from the study due to sample quality issues and are therefore not reported, reducing the expected negative cohort to 402 samples. The third reported sample was considered potentially discordant following a positive result in the MosaiQ COVID-19 Antibody Microarray assay; yet this sample was obtained from a negative donor (sample collected in 2015, pre-SARS-CoV-2). Upon testing in the ELISA assay, this sample elicited a negative result and therefore it was concluded that the true serological status of the sample is negative for both IgM and IgG antibodies to SARS-CoV-2. As the sample reported as positive in the MosaiQ COVID-19 Antibody Microarray, one false positive on the MosaiQ COVID-19 assay is reported in this study.

For the determination of assay sensitivity, 94 samples from a total of 100 expected positives returned a positive result. Four samples were removed from the dataset as full investigative resolution was not 8 | P a g e performed and therefore could not be included, with the remaining two samples investigated for potential discordance ( Table 1) . One of the potentially discordant samples (Table 1 

The summary performance of 100% sensitivity (Positive Percent Agreement) and 99.8% specificity (Negative Percent Agreement) represents excellent performance for an in vitro diagnostic serological assay. Furthermore, in the context of required levels of sensitivity and specificity required for licensure of an in vitro diagnostic assay in the EU and the US, this level of performance is suitable [6].

There were two samples which returned initial unexpected non-reactive results, and following a detailed investigation, were shown to be seronegative; the samples were obtained from patients that did not undergo seroconversion. In the event of non-seroconversion, an antibody negative result would be expected and the test system therefore showed the correct result as no antibodies to SARS-CoV-2 are present. This interpretation is consistent with data from studies conducted in the Wuhan and Guangdong areas in the initial outbreak; there is now significant data to show that a small percentage of patients do not seroconvert for IgM and IgG, and therefore, this result would be consistent with that lower level of unusual immune responses in the cohort [1] , and is at a comparable level to that seen in the Chinese studies [2] . Of course, the true serological status of a patient must be interpreted based on presence of antibodies, not previous evidence of infection by PCR assay, and the removal of non-seroconverted samples from the expected positive sample cohort is therefore justified. The heterogenous immunological response across patient cohorts should be considered in establishing a test panel, and here carefully selected samples were used where infection was supported by PCR diagnosis, but the indication of IgM and IgG status still requires discordance management exercises in cases where ostensibly negative or positive samples elicit discordant results. This highlights the rigour required in establishing performance of assays to detect serological status in patients with emerging viral infections.

This study has succeeded in showing early detection of IgG and IgM antibodies to SARS-CoV-2, utilising a high throughput system. At the time of the study, a relatively small sample set was available for testing, and future studies would include larger sample sets to develop understanding of assay performance. In addition, further studies would ideally include samples from patients with new variants emerging to confirm similar performance.

In summary, we demonstrate that the Quotient MosaiQ COVID-19 Antibody Microarray test is suitable for accurate testing of blood samples for IgM and IgG antibodies to SARS-CoV-2 in an active field setting, regardless of sample age since time of presenting a positive RT-PCR result. The system demonstrated suitable sensitivity and specificity to be considered fit for its intended purpose and useful in the assessment of immunity to this virus.

Corresponding Authors

Profiling Early Humoral Response to Diagnose Novel Coronavirus Disease (COVID-19)

Temporal profiles of viral load in posterior oropharyngeal saliva samples and serum antibody responses during infection by SARS-CoV-2: an observational cohort study, The Lancet Infectious Diseases

SARS-CoV-2 viral load in upper respiratory specimens of infected patients

Reinfection could not occur in SARS-CoV-2 infected rhesus macaques

The authors thank Matilde Ruiz Tovar of the Blood Donor Laboratory, Centro de Transfusión de Madrid, for collaboration in this work. The authors also thank all nurses, technicians and MD/PhD staff not mentioned as co-authors from the Madrid Transfusion Center. 

14 | P a g e