key: cord-1028795-09uriow8
authors: Heaney, Katy; Ritchie, Allyson V.; Henry, Rowan; Harvey, Adam J.; Curran, Martin D.; Allain, Jean-Pierre; Lee, Helen H.
title: Evaluation of sample pooling using the SAMBA II SARS-CoV-2 Test
date: 2021-10-22
journal: J Virol Methods
DOI: 10.1016/j.jviromet.2021.114340
sha: 848396ca681b66b5bf40f05f11ba0a1bf2b185f6
doc_id: 1028795
cord_uid: 09uriow8

BACKGROUND: Screening of infectious asymptomatic or pre-symptomatic individuals for SARS-CoV-2 is at present a key to controling the COVID-19 pandemic. In order to expand testing capability and limit cost, pool testing of asymtomatic individuals has been proposed, provided assay performance is not significantly affected. METHODS: Combined nose and throat (N/T) swabs collected from COVID-19 infected or non-infected individuals were tested using SAMBA II individually and in pools of four (one positive and 3 negative). The evaluation was conducted by the manufacturer and an independent NHS site. Ct cycles of individual positives and pooled positives were determined by qRT-PCR. RESULTS: In 42 pools containing a single positive sample with Ct values ranging between 17 and 36, 41 pools (97.6%) were found positive by the SARS-CoV-2 SAMBA II test. The false-negative pool by SAMBA was also negative by both reference methods used in this evaluation.The individual positive sample in this pool was positive by SAMBA (Orf only) and by one of the reference methods (S gene only, Ct 35) but negative by the second reference method indicating that the sample itself was very low viral load. All 78 pools containing 4 negative swabs were negative (100% specificity). DISCUSSION: The preliminary data of the evaluation indicated a high level of performance in both sensitivity and specificity of the SAMBA II assay when used to test pools of 4 patient samples. The implementation of this pooled protocol can increase throughput and reduce cost/test when the prevalence of COVID is low.

Until recently in the UK, emphasis has been placed on testing individuals with COVID-like symptoms (fever, cough, anosmia) in order to identify infection and subsequently isolate the individuals and their close contacts. This strategy has been applied in the community, in hospitals for triage of patients and in schools in order to decide on isolation of bubbles. However, the main drivers of the pandemic are asymptomatic and pre-symptomatic infections that remain undetected despite infectiousness similar to symptomatic cases (Arons et al., 2020) (Rivett et al., 2020) . A recent study estimated that at least 50% of COVID-19 cases may have been contracted from asymptomatic individuals (Johansson et al., 2021) . Therefore, screening of pre-symptomatic and asymptomatic carriers is crucial for SARS-CoV-2 infection prevention and in a hospital setting to diagnose SARS-CoV-2 infection in incoming patients regardless of symptoms. The SAMBA II SARS-CoV-2 Test is an accuracte point-of-care (POC) test for diagnosis of SARS-CoV-2 infection with a limit of detection of 250 cp/ml and high clinical sensitivity and specificity Assennato et al., 2020 , Collier et al., 2020 . Rapid J o u r n a l P r e -p r o o f POC tests, such as SAMBA II, with fast results are useful, so that those who are positive can be promptly isolated and attended. This cannot be achieved with centralised testing, with turnaround times of 24 hours or more.

In order to limit cost and expand testing capability, Public Health England (PHE) recommended pooling to increase testing capacity and reduce reagent consumption when there is a low background prevalence in target groups where there is need and benefit from identifying positive individuals, eg asymptomatic patients and professionals, at a time of low positive prevalence (Hogan et al., 2020 , Lohse et al., 2020 , Mastrianni et al, 2020 (https://www.england.nhs.uk/coronavirus/wpcontent/uploads/sites/52/2020/09/C0777-sample-pooling-sop-v1.pdf]). DRW, the manufacturer of the SAMBA II test, investigated the performance of the SAMBA SARS-CoV-2 Test in pooled clinical samples using pools of four.

The study was intended to examine the feasibility and reliability of the SAMBA II SARS-CoV-2 Test using pools of 4 samples in order to expand the availability of the assay without compromising its performance. The first phase was conducted in the manufacturer's facilities in collaboration with the Clinical Microbiology and Public Health Laboratory, Addenbrooke's Hospital, Cambridge (CMPHL) using surplus frozen samples from the COVIDx study and additional fresh negative samples. SAMBA and PHE positive, wellcharacterised, individual samples were thawed and mixed with three negative swab samples by one operator to make positive pools and four negative J o u r n a l P r e -p r o o f samples were mixed to make negative pools. SAMBA testing was carried out by a second operator in a blinded fashion.

The second phase was carried out in the Royal Berkshire Foundation Trust POC testing laboratory. This pilot study consisted of selecting 10 patient samples which previously tested by the SAMBA-SARS-CoV-2 and deemed to be either strong positive (both Orf and N detected) 

The SAMBA II platform and the SAMBA II SARS-CoV-2 Test kit are CE IVD marked for diagnosis of SARS-CoV-2 infection. The system and test have been previously described in detail . The assay specifically amplifies two regions of the SARS-CoV-2 genome in the ORF1ab and nucleocapsid gene (N) with a visual readout on a test strip. The uppermost line detects the internal control, which ensure adequate test Repeated samples were run using the original swab sample that had been stored in the refrigerator at 4-8°C and brought to room temperature before running.

The PHE reference test was performed at CMPHL. Samples used for the DRW pooling study were collected in April 2020 as part of the COVIDx study and tested using the Cambridge RdRp gene assay on the Rotor gene Q realtime PCR assay routinely used by CMPHL as previously described (Sridhar et al., 2020) but modified by switching the enzyme master mix used to Taqpath™ 1-Step RT-q PCR from Life Technologies (Cat No A15300). The J o u r n a l P r e -p r o o f samples used for the Royal Berkshire phase 2 pooling study were tested using an upgraded assay, which also amplifies the S gene target in addition to the RdRp gene as previously detailed (Skittrall et al., 2020) . A reactive result for either or both genes below Ct 36 was considered a positive result on both assays run at CMPHL.

Combined nose and throat (N/T) swab samples were re-suspended in 2 mL of SAMBA SCoV buffer, provided with the kit. The SAMBA SCoV buffer inactivates SARS-CoV-2 within 10 minutes (Collier et al., 2020) (Welch et al., 2020) . It is therefore recommended that samples be incubated at room temperature for 10 minutes to inactivate the sample before loading it into the machine. The input volume for the SAMBA test is 300 l of which 250 l is used by the SAMBA II machine as input into the sample processing. For the purpose of this study four samples were pooled together by pipetting 75 l of each of 4 samples into the SAMBA input tube to give a total volume of 300 l. This pooled sample was then run in the SAMBA II and the result recorded. All pooled samples were also tested individually and the results recorded and compared to the pooled result.

Thirty-one (31) frozen surplus SARS-CoV-2 positive combined N/T swab samples from a previous evaluation (Collier et al., 2020) were used to prepare positive pools. These samples had previously been tested as positive for SARS-CoV-2 by both the SAMBA SARS-CoV-2 test and the PHE reference J o u r n a l P r e -p r o o f laboratory assay by CMPHL. These samples were all collected from symptomatic individuals and the Ct values ranged from 17-34 by the reference method. Fifty-two frozen surplus SARS-CoV-2 positive combined N/T swab samples from the same evaluation were used to prepare the positive and negative pools. These samples were negative individually by both SAMBA and the standard PHE test . In addition, 103 SAMBA negative N/T swab samples from healthy individuals without symptoms were used to prepare pools. In total 32 positive pools (containing 1 positive and 3 negative samples) and 44 negative pools (containing 4 negative samples) were generated. The content of each pool was blinded from the operator running and interpreting the SAMBA and results.

In phase 1, samples already run by the operational SAMBA point-of-care service were selected and randomised by an operator into groups of 4. The aim was to test 50 pools, with at least 10 containing a positive sample. Of these 10 positives, 5 to contain a strong positive sample, defined as positive for both SAMBA targets (ORF1ab and N) and 5 to contain a sample with low level positive result, defined as a single target line only detected (ORF1ab).

These patient samples were kept at 2-8°C for 0 -3 days before they were used to constitute pools with fresh negative samples. In Phase 2 10 positive pools (5 high and 5 low as described above) were run on the same day as the individual samples ranging from 1.5 to 7 hours between the individual result and the pooled result. Aliquots of the ten positive individual samples and pools were frozen and sent to the RSML and CMPHL for reference testing.

The samples used in this study were surplus volume from the sample taken as part of the patient's standard care, once all analysis had been completed.

No additional tests were carried out on patients or clinical decisions made as a result of this work. This work is classified as a service evaluation.

In total 32 positive pools (containing one positive and three negative samples) were tested along with 44 negative pools (containing four negative samples).

The Ct value of the positive samples ranged from 17 to 34 according to the PHE Cambridge method described in 2.2.3 (Table 1) . All 32 positive pools tested and all 44 negative pools were negative. SAMBA results were recorded by visually reading the test strip and by recording the camera results reported on the tablet. On the basis of this limited study, the sensitivity and specificity of the pool of four testing was 100% respectively and camera and visual results were 100% concordant. Table 2 ). If any individual was not detected, the pool was also not detected by Genesig. (Table 3) .

Four of the five pools containing a low level positive were positive and one (1LP-02) was negative ( Table 3) .

The five high level positives were also positive by the Genesig and PHE tests both individually and in pools (Table 3 ). Of the five low level positives four were negative by Genesig (ND or Ct>37) individually and all were negative in pools (Table 3) . All five low level positives were negative by the PHE RdRp gene (Neg or Ct>36) both individually and in pools (Table 3) 

Pooling techniques enable screening of greater numbers of individuals while preserving testing resources. Numerous publications recently pointed out sample pooling as a method to reduce cost and maximising efficiency (10-12) and investigated an optimum balance between test performance and number of samples in a pool. Studies reporting on SARS-CoV-2 testing in pools were conducted in pools of 5, 8 or 10 samples (Praharaj et al., 2020) (Chhikara et al., 2021) (Torres et al., 2020) . In pools of five (the closest to the pools of four described here), the expected number of tests performed was 57% less than in individual testing at a 5% prevalence rate and a 100% sensitivity with LOD of 1,000-3,000 copies/ml (Abdalhamid et al., 2020) . Another study using the Cepheid Xpert® Xpress SARS-CoV-2 assay could detect positive samples of Ct 20-28 when run in a pool or 4 or 6 with a median change of Ct value of 2.0 and 2.9 respectively but samples with higher Ct values were not tested (Graham et al., 2021) . Another report examined the options of pooling before or after nucleic acid extraction and did not find a significant difference between the two (Chhikara et al., 2021) . In our case, for simplicity and time saving, pooling swab samples prior to extraction was adopted and samples of With regards to sensitivity, the SAMBA II SARS-CoV-2 Test has a limit of detection of 250 copies/ml and therefore pooling four samples will likely raise this to 1000 copies/ml. Viral loads below such levels appear to be a small minority and to be far below the infectivity level estimated around 100,000 copies/ml (Wölfel et al., 2020) . This evaluation was carried out in two distinct settings: one internal to the SAMBA manufacturer, the other external in a NHS point of care testing setting. The combined specificity between the 2 sites was Data interpretation however needs to take into consideration for the Royal Berkshire data that the SAMBA testing was performed on fresh individual swabs and pools. The individual and pooled samples were immediately frozen after SAMBA testing and were tested by Genesig at RSML after 8-18 days and by PHE at CMPHL after 11-21 days. In the DRW study the SAMBA individual samples were tested fresh and samples frozen the same day. The pools were tested by SAMBA after around 4 months. Therefore, samples used for pooling by SAMBA must be tested on the same day of collection (within 7 hours) or frozen and tested at a later date to ensure that the low level positive samples are detected.

The utility of a diagnostic strategy using pooled samples also holds close relation to the prevalence of infection in the proposed setting. In high prevalence scenarios, a greater number of pools need to be retested, rendering the strategy more costly and time-consuming. Several mathematical models indicated that pool testing was cost-effective below an acute infection prevalence of 10-30% (Mutesa et al., 2020) (Mallapaty, 2020) (Aragón-Caqueo et al., 2020) . Although the exact cost-benefit of the pooling approach J o u r n a l P r e -p r o o f needs to be individually assessed based on circumstances, it appears reasonable to pool samples of 4 individuals if the SARS-CoV-2 if prevalence in the target population is below 10% with predicted test reduction in test numbers and hence cost of 35, 55 and 71% at prevalenceof 10, 5, and 1% respectively.

Both arms of the evaluation study concluded the high performance of the SAMBA II SARS-CoV-2 Test using pools of 4 samples. Funding source. DRW funded the study and lent SAMBA-II machines to Royal Berkshire NHS Foundation Trust and provided the SAMBA-II SARS-CoV-2 test kits for the duration of the study. Assays performed at the CMPHL were a generous contribution to the study. SAMBA results can be interpreted visually and also reported electronically. The Orf target is more sensitive than the N target. Detection of either target results in a positive diagnosis. Tests with both Orf and N detected were categorised a "strong positives" for this study and samples with Orf but no N were categorised as "weak positives" for this study. 2 ND = not detected 3 With Gensig Ct 35-37 is considered equivocal. Ct >37 is negative J o u r n a l P r e -p r o o f 

Assessment of Specimen Pooling to Conserve SARS CoV-2 Testing Resources

Optimization of group size in pool testing strategy for SARS-CoV-2: A simple mathematical model

Infections and Transmission in a Skilled Nursing Facility

Performance evaluation of the point-of-care SAMBA II SARS-CoV-2 Test for detection of SARS-CoV-2

Validation of SARS CoV-2 detection by real-time PCR in matched pooled and deconvoluted clinical samples before and after nucleic acid extraction: a study in tertiary care hospital of North India

Point of Care Nucleic Acid Testing for SARS-CoV-2 in Hospitalized Patients: A Clinical Validation Trial and Implementation Study

SARS-CoV-2 Transmission From People Without COVID-19 Symptoms

Sample Pooling as a Strategy to Detect Community Transmission of SARS-CoV-2

Viral RNA load as determined by cell culture as a management tool for discharge of SARS-CoV-2 patients from infectious disease wards

Pooling of samples for testing for SARS-CoV-2 in asymptomatic people

The mathematical strategy that could transform coronavirus testing

Pooled Testing for SARS-CoV-2 in Hospitalized Patients

A strategy for finding people infected with SARS-CoV-2: Optimizing pooled testing at low prevalence

Pooled testing for COVID-19 diagnosis by real-time RT-PCR: A multi-site comparative evaluation of 5-& 10-sample pooling

Elife 9

Duration of infectiousness and correlation with RT-PCR cycle threshold values in cases of COVID-19

Specificity and positive predictive value of SARS-CoV-2 nucleic acid amplification testing in a low-prevalence setting

A blueprint for the implementation of a validated approach for the detection of SARS-Cov2 in clinical samples in academic facilities

Pooling of nasopharyngeal swab specimens for SARS-CoV-2 detection by RT-PCR

Analysis of Inactivation of SARS-CoV-2 by Specimen Transport Media, Nucleic Acid Extraction Reagents, Detergents, and Fixatives

Covid-19: Safety of lateral flow tests questioned after they are found to miss half of cases

Virological assessment of hospitalized J o u r n a l P r e -p r o o f patients with COVID-2019

The authors wish to thank Prof R Gupta, Dept of Medicine, Addenbrooke's hospital Trust, Cambridge, UK for providing some of the samples used in this study and Dr Nadine Collins from the Royal Surrey Molecular Laboratory for support with testing samples.J o u r n a l P r e -p r o o f