key: cord-1003462-c2kkjw43 authors: Ricarte, D.; Gador, A.; Mendiola, L.; Gonzales, I. C. title: Diagnostic Performance of Pooled RT-PCR Testing for SARS-CoV-2 Detection date: 2021-02-19 journal: nan DOI: 10.1101/2021.02.17.21251961 sha: b8af97a5fe7efd2262b99bc293f22ca80f9fd024 doc_id: 1003462 cord_uid: c2kkjw43 Background: With the high number of COVID-19 cases, a need to optimize testing strategy must be regarded to obtain a timely diagnosis for early containment measures. With this, several studies have employed pooled RT-PCR testing for SARS-CoV-2 as this could potentially conserve laboratory resources while has the capacity to test several individuals. However, this was recommended to firstly validate the method as different laboratory reagents and equipment vary with its diagnostic performance. Objective: The aim of this study was to determine the diagnostic performance of pooled SARS-CoV-2 nasopharyngeal/oropharyngeal swabbed samples using the RT-PCR technique. Methods: A records review of two-staged pooled RT-PCR testing data from August 10, 26, 30, and September 5, 2020, was utilized from CHDNM TB Regional Center. For the first stage, using known samples, a total of 30 pools were made for each of the pooling sizes, 5- and 10-pooled, on both pooling phase, pre-and post-RNA extraction. One positive individual was used to represent each of the Cycle threshold values given (<24, 25-28, 29-32, 33-36, and 37-40) while the rest of the samples were negative. For the second stage, 54 pools of five from 270 random unknown samples were used to validate the results. Target gene performance of N gene and RdRp was also determined. Key Results: Results show that 5-pooled sample has higher sensitivity (SN), specificity (SP), positive predictive value (PPV), and negative predictive value (NPV) of 100% (95% confidence interval (CI) 88.97-100), 66.95% (95% CI, 60.75-72.6), 28.18% (95% CI, 20.62-37.22), and 100% (95% CI, 97.66-100) compared to 10-pooled sample that has 87.1% (95% CI, 71.15-94.87), 56.9% (50.57-63.02), 20.77% (95% CI, 14.68-28.53) and 97.14% (95% CI, 92.88-98.88). Further, these Ct values were only from the N gene, emphasizing its higher diagnostic performance as well to detect SARS-CoV-2 compared to RdRp as only a few samples were detected, thus, no analysis was made. Conclusion: This study found out that a 5-pooled sample has better diagnostic performance compared to 10-pooled samples. Specifically, all positive individual samples were detected in 5-pooled samples in the pre-RNA extraction phase which these results are evident and consistent on both known and unknown samples. N gene was found out to detect more SARS-CoV-2 samples compared to RdRp. control strategies of the government such as aid for early detection of positive cases while assisting policymakers' decisions for their implementation of containment measures. While Northern Mindanao is experiencing a delay in supply logistics that resulted in a backlog of samples, hence, this pooling method can possibly conserve the resources and continue testing individuals. A retrospective research design was employed to obtain secondary data on SARS-CoV-2 swabbed samples that were collected from nasopharyngeal and oropharyngeal sites of individuals ranging from asymptomatic, mild with no comorbidities, mild with comorbidities, severe and critical cases in Northern Mindanao and that were sent by the trained Medical Technologists to the Northern Mindanao Medical Center COVID-19 Satellite Laboratory (formerly CHDNM TB Regional Center) last August 10, 26, 30, 2020 and September 5, 2020. • Known and unknown SARS-CoV-2 sample results within the valid control were included for data collection where a positive control had a Ct value of ≤22 for RdRp and N gene while a Ct value of ≤21 for RNAse P (Internal control) and no detection cycles of the genes mentioned for a negative control as also based on the Instruction for Use (IFU) of GeneFinder™ COVID-19 Plus RealAmp Kit. For further validation, the pathologist, two Medical Technologists as well as the Head of the Infectious Disease Cluster verified the control results. • Individual and pooled samples that were received and tested within 24 hours of receipt to Northern Mindanao Medical Center COVID-19 Satellite Laboratory (formerly CHDNM TB Regional Center) were only included for data collection. • GeneFinder™ COVID-19 Plus RealAmp Kit uses three target genes: N gene, E gene and RdRp and RNAse P as the internal control gene. However, E gene is only interpreted as PRESUMPTIVE POSITIVE if it is detected, thus, this gene was not included for the data collection. This study had two stages. First, the pooling data on known (determined) samples was used to identify what ideal sample size and pooling phase could detect the positive samples in each of the Ct value ranges given. Its target genes and diagnostic performance were analyzed. Second, the pooling data on arbitrarily chosen unknown (clinical) samples was then used to validate the ideal results found on the pooled known samples in which its target genes and diagnostic performance were analyzed. The methodology on how the pooled samples were done prior and after RNA extraction were based on the published literature ( 2020) having high diagnostic performance and reliable results, thus, its pooling procedures were synthesized and adopted in this study. • Combined nasopharyngeal and oropharyngeal swabs were collected by the trained DOH CHDNM Medical Technologists, transported at 2 to 8 degree Celsius in a 2 ml Sanli Virus specimen collection tube; Liuyang, Sanli Medical Technology Devt. Co., China (RITM, 2020), a valid and compatible for the SARS-CoV-2 detection kit used where it states that it is intended for use with RNA extracted from nasopharyngeal swabs, and oropharyngeal swabs. However, there was no ideal transport medium indicated in its Instruction for Use (GeneFinder™ COVID-19 Plus RealAmp Kit, OSANG Healthcare Co., Ltd, South Korea), thus, valid VTM is acceptable. The samples that were sent to the said laboratory were directly processed. While the results of the individual samples were yet to be determined for a turnaround time of 8 to 12 hours, VTM were stored at 4 degree Celsius. Once known, these were then pooled after. Hence, no freeze-thaw cycles were done. • From the time the samples were received at the laboratory, these were then processed to avoid RNA degradation. Samples were labeled and assigned in a designated pool for each of the Ct value ranges given. After which, these were prepared in pool of 5 and pool of 10 and were pooled in two phases, pre-RNA extraction and post-RNA extraction in three replicates to ensure reproducibility of the results. For a pool of five, one positive and four negative samples were used in each of the Ct value ranges: <24, 25-28, 29-32, 33-36 and 37-40. For a pool of ten, one positive and nine negative samples were used in each of the Ct value ranges: <24, 25-28, 29-32, 33-36 and 37-40. (Supplementary Figure S1 and S2). • For the first stage, pooling of known samples was used. Samples were pooled right after the results of the individual samples were determined or about 12 hours from the time individual samples were received at the laboratory and tested until these were then pooled. Previously tested samples were stored at 4 degree Celsius prior to pooling. For the second stage, pooling of unknown samples was used. Individual and pooled samples were tested in parallel. An automated extractor, Genolution NextractorⓇ NX-48S with its nucleic acid reagents, was used for the purification of RNA. The extraction underwent the following steps: lysis of the sample, DNA binding to the magnetic beads, washing and elution. • Individual samples A total volume of 200 microliter from each of the universal transport medium buffers was dispensed onto the pre-filled sample well of the automated Genolution NextractorⓇ NX-48S plate. After loading all the individual samples, these were placed into the auto extractor for RNA purification. . CC-BY-NC-ND 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) • Pooling of samples prior to RNA extraction An initial volume of 100 microliter aliquot from each viral transport medium buffer was initially transferred onto a conical tube for the pooling of samples prior to dispensing onto the pre-filled sample well of the automated Genolution Nextractor NX-48S plate. A total volume aliquot of 500 microliter for the 5-pooled sample and 1,000 microliters for the 10-pooled samples. A final volume of 200 microliter, either of the 5-or 10-pooled samples, was dispensed into the pre-filled sample well of the automated Genolution Nextractor NX-48S plate. The samples were pooled in three replicates. Eluted samples were then transferred to a new tube for RT-qPCR test. • Pooling of samples after RNA extraction Eluted RNA samples were firstly pooled prior to mixing with the master mix reagent onto the 96-well plate. An aliquot volume of 10 microliter each eluted RNA sample was dispensed initially onto the tube for pooling. A total volume of 50 microliter for the 5-pooled samples and 100 microliters for the 10-pooled samples. A final volume of 5 microliter, either of the 5or 10-pooled samples, was dispensed onto the 96-well plate for RT-PCR testing. After the RNA extraction, RT-PCR test was performed to detect the presence of SARS-CoV-2 RNA using GeneFinder™ COVID-19 Plus RealAmp Kit in Applied Biosystems® 7500 Real-Time PCR Instrument (ABI 7500) machine (Thermo Fisher Scientific Inc.) based on the Instruction for Use (IFU) provided. A 15 microliter of GeneFinder™ master mix reagent was dispensed onto the 96-well plate. After which, a 5 microliter of eluted samples, either of the individual, 5-or 10-pooled samples, were dispensed. A 96-well plate containing mastermix reagent and eluted RNA sample was placed onto the RT-PCR machine to detect 3 SARS-CoV-2 target genes: N, E, and RdRP. For quality control, there were three controls provided on the Instruction for Use (IFU) of GeneFinder™ COVID-19 Plus RealAmp Kit. For internal control, RNAse P gene with Ct value of ≤22 was used while for positive external control, N gene and RdRp with Ct value of ≤22. No detection cycles of the genes mentioned was set for a negative control. Reactions were heated to 50℃ for 20 minutes for reverse transcription, denatured at 95℃ for 15 minutes and then 45 cycles were carried for amplification set at 95℃ for 15 seconds and 58℃ for 1 minute. Fluorescence was measured using four channels: FAM (RdRp gene), JOE (N gene), TEXAS RED (E gene), and CY5 (RNAse P). After which, the data on the individual and pooled SARS-CoV-2 NP/OP swabbed samples were validated and will be obtained for analysis. . CC-BY-NC-ND 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) A total of 450 known samples were utilized for this stage. A positive sample was used to represent each of the cycle threshold values given. There were 150 known samples used for pooling of five while 300 samples were used for pooling of 10 that consist of one positive sample and four negative samples in each of the five Ct value ranges and were pooled in two phases, prior to RNA extraction or after RNA extraction, in three replicates. i. Pooling of 5-pooled samples . CC-BY-NC-ND 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) (Table 3) . Meanwhile, no analysis for RdRp Ct values was made as it was undetected for most of the samples (Table 4) . Table S1 ). . CC-BY-NC-ND 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. The copyright holder for this preprint this version posted February 19, 2021. ; https://doi.org/10.1101/2021.02.17.21251961 doi: medRxiv preprint To compare the difference between 5-pooled and 10-pooled known nasopharyngeal/oropharyngeal swabbed samples that is pooled in each RNA extraction phase in detecting SARS-CoV-2 in Ct value ranges of <24, 25-28, 29-32, 33-36, 37-40, Open Epi was used to analyze the set of data. Table 5 shows that there is no significant difference between 5-and 10-pooled samples in Ct value range of 29-32 (95 CI: -0.68, 2.16) and 33-36 (95 CI: -1.65, 0.37) in post-RNA extraction phase. Further, a known positive sample in Ct value range of 37-40 is negative in 5-pooled sample of post-RNA extraction phase and is undetected in 10-pooled samples for both RNA extraction phases. To compare the diagnostic performance between the target genes, N gene and RdRp, in detecting unknown 5-pooled and 10-pooled SARS-CoV-2 nasopharyngeal/oropharyngeal swabbed samples that is pooled in pre-RNA extraction phase, Microsoft Excel and Open Epi was used to analyze the set of data. Using 2x2 table analysis (Table 6 , 7, 9, and 10), results show that N gene of 5-pooled sample has higher diagnostic performance in terms of its sensitivity, positive predictive value, and negative predictive value. On the other hand, RdRp of 5-pooled sample is more specific. However, there is no significant difference between the target genes, N gene and RdRp, for both pooling sizes, 5pooled and 10-pooled samples, in detecting SARS-CoV-2 nasopharyngeal/oropharyngeal swabbed samples (Table 8 and 11) . . CC-BY-NC-ND 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. The copyright holder for this preprint this version posted February 19, 2021. ; https://doi.org/10.1101/2021.02.17.21251961 doi: medRxiv preprint . CC-BY-NC-ND 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) To compare the diagnostic performance of the pooling method of unknown 5-pooled and 10pooled SARS-CoV-2 nasopharyngeal/oropharyngeal swabbed samples that is pooled in pre-RNA extraction phase, Microsoft Excel and Open Epi was used to analyze the set of data. Using 2x2 table analysis (Table 12 and 13), Table 14 shows that 5-pooled samples have higher diagnostic performance compared to 10-pooled samples. However, there is no significant difference between the two pooling sizes. CC-BY-NC-ND 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. The copyright holder for this preprint this version posted February 19, 2021. This study found out that 5-pooled sample in pre-RNA extraction phase has higher diagnostic performance as results are evident and consistent on both known and unknown samples. To further elaborate, it has a sensitivity (SN), specificity (SP), positive predictive value (PPV), and negative predictive value (NPV) of 100% (95% confidence interval (CI) 88. on pooling of ten samples were higher than the results of this study. This may probably be due to higher sensitivity of the viral detection kit used as it potentially contribute on the diagnostic performance of the pooling method. In this study, the higher diagnostic performance of 5-pooled sample is expected as these were diluted less than 10-pooled which tends to increase the cycle threshold value and lowers the viral load. This is also apparent when pooling was firstly done on known samples as only 5-pooled in pre-RNA extraction phase could be able to detect all known individual positive samples, specifically the N gene, on all the cycle threshold value ranges given (Table 1 & As shown in this study, only N gene was detected on all known individual positive samples when pooled in five in pre-RNA extraction phase. When validated and analyzed on unknown samples, it was found out that this gene is more sensitive and has higher positive predictive value and negative predictive value, however, RdRp is more specific. To the best of our knowledge, target gene performance was only evaluated on individual samples. According to Benrahma et. al. (2020) , N gene is persistently detected in samples with Ct value of ≤ 40 in RT-PCR having similar viral detection kit used in this study. This may explain due to its abundant production within infected cells having multiple functions such as replication, transcription, translation and specifically binding to viral RNA to form the ribonucleocapsid (McBride, Van Zyl and Fielding, . CC-BY-NC-ND 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) and Nalla et. al. (2020) also showed that N gene is found to be more sensitive than RdRp when tested on individual nasopharyngeal and oropharyngeal samples, and noticeably more when it is pooled as illustrated in this study. In addition, this is also emphasized when an evaluation of SARS-CoV-2 detection kit was done by Zhou et. al. (2020) and recommended by the study of Chu et. al. (2020) as a screening assay as N gene is ten times more sensitive than Orf1b in detecting positive samples. On the other hand, as RdRp is a viral enzyme with no host cell homolog that functions for viral RNA replication, this might explain why it has higher diagnostic performance (Zhu, et al., 2020) . The scope of this study is the diagnostic performance of the reagents and equipment used. Further, its limitation includes no analysis on the pooling of SARS-CoV-2 swabbed samples of the symptomatic and asymptomatic individuals as only de-identified samples were utilized. Overall, this study serves as a proof of concept of the pooling method, illustrating the results of the different pooling sizes on both RNA extraction phases, of SARS-CoV-2 using RT-PCR technique. In conclusion, this study found out that 5-pooled sample has better diagnostic performance compared to 10-pooled samples as, specifically, all positive individual samples were only detected in 5pooled samples in pre-RNA extraction phase which these results are evident and consistent on both known and unknown samples. Further, N gene was found out to detect more SARS-CoV-2 samples compared to RdRp. Based on the results of this study, the following are hereby recommended for future research and development work: For research institutes: • Given the scope of this study, a validation on pooling method using other diagnostic reagents and equipment must also be done as its performance may vary and may not be similar with the results shown in this study. • To address the limitation of this study, consider symptom profile and its duration when pooling of SARS-CoV-2 swabbed samples. • If incidence or surveillance of COVID-19 is aimed for the study, a pooled testing method is suggested to lessen the workload and resources used while at the same time having a reliable result compared to using rapid diagnostic kits. The results may also be a significant basis for policy-making in the respective areas. For policymakers: • Based on the results of the study, pooling of five samples prior to RNA extraction must be used as it could possibly detect much positive samples even for a high cycle threshold value or low viral load compared to pooling of 10 samples that has a higher chance of missing out possible positive samples. Moreover, this could be a potential tool to hasten the process of tracking and isolating possible COVID-19 individuals. • For pooling of SARS-CoV-2, the following reagents and equipment are hereby suggested as these shows satisfactory performance: Genolution NextractorⓇ NX-48S for RNA extraction, GeneFinder™ COVID-19 Plus RealAmp Kit for SARS-CoV-2 viral detection and Applied Biosystems® 7500 Real-Time PCR Instrument (ABI 7500; Thermo Fisher Scientific Inc.) for PCR machine. . CC-BY-NC-ND 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. The copyright holder for this preprint this version posted . CC-BY-NC-ND 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. The copyright holder for this preprint this version posted . CC-BY-NC-ND 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. The copyright holder for this preprint this version posted February 19, 2021 . CC-BY-NC-ND 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. The copyright holder for this preprint this version posted February 19, 2021. ; https://doi.org/10.1101/2021.02.17.21251961 doi: medRxiv preprint Supplementary Table S2 . Comparison of mean Ct value of 5-and 10-pooled samples in preextraction phase with a Ct value of <24. Table S3 . Comparison of mean Ct value of 5-and 10-pooled samples in preextraction phase with a Ct value of 25-28. . CC-BY-NC-ND 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. The copyright holder for this preprint this version posted February 19, 2021 . CC-BY-NC-ND 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. The copyright holder for this preprint this version posted February 19, 2021 . CC-BY-NC-ND 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) . CC-BY-NC-ND 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) . CC-BY-NC-ND 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) . CC-BY-NC-ND 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) . CC-BY-NC-ND 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) . CC-BY-NC-ND 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) . CC-BY-NC-ND 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) . CC-BY-NC-ND 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) The copyright holder for this preprint this version posted February 19, 2021. ; https://doi.org/10.1101/2021.02.17.21251961 doi: medRxiv preprint Assessment of Specimen Pooling to Conserve SARSCoV-2 Testing Resources Increasing SARS-CoV-2 RT-qPCR testing capacity by sample pooling Epidemiological description and analysis of RdRp, E and N gene dynamic by RT-PCR of SARS-CoV-2 in Moroccan population: Experience of the National Reference Laboratory (LNR)-UM6SS. medRxiv Epidemiology and transmission of COVID-19 in 391 cases and 1286 of their close contacts in Shenzhen, China: a retrospective cohort study Interim Guidance for Use of Pooling Procedures in SARS-CoV-2 Diagnostic, Screening, and Surveillance Testing 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 COVID-19 Tracker August) Philippines Now Has 100 Licensed Testing Labs-DOH Sample Pooling for SARS-COV-2 RT-PCR screening The Detection of Defective Members of Large Population Methodology for estimating point prevalence of SARS-CoV-2 infection by pooled RT-PCR testing DOH: No shortage in COVID-19 test kits in the Philippines Pooled RNA sample reverse transcriptase real time PCR assay for SARS CoV-2 infection: A reliable, faster and economical method Sample Pooling as a Strategy to Detect Community Transmission of SARS Pooling of samples for testing for SARS-CoV-2 in asymptomatic people Optimizing SARS-CoV-2 Pooled Testing for Low Resource Setting The Coronavirus Nucleocapsid Is a 333 Multifunctional Protein Evaluation of Sample Pooling for Screening of SARS-CoV-2. Research Square Comparative Performance of SARS-CoV-2 Detection Assays Using Seven Different Primer-Probe Sets and One Assay Kit When COVID-19 Testing Supply Shortages Hit, This Behind-the-Scenes Lab Improvised and Innovated Fast