Evaluation on Diagnostic Efficiency of Different Methods in Detecting COVID-19

Objective: To evaluate the diagnostic efficiency of different methods in detecting COVID-19 to provide preliminary evidence on choosing favourable method for COVID-19 detection. Methods: PubMed, Web of Science and Embase databases were searched for identifing eligible articles. All data were calculated utilizing Meta Disc 1.4, Revman 5.3.2 and Stata 12. The diagnostic efficiency was assessed via these indicators including summary sensitivity and specificity, positive likelihood ratio (PLR), negative LR (NLR), diagnostic odds ratio (DOR), summary receiver operating characteristic curve (sROC) and calculate the AUC. Results: 18 articles (3648 cases) were included. The results showed no significant threshold exist. EPlex: pooled sensitivity was 0.94; specificity was 1.0; PLR was 90.91; NLR was 0.07; DOR was 1409.49; AUC=0.9979, Q*=0.9840. Panther Fusion: pooled sensitivity was 0.99; specificity was 0.98; PLR was 42.46; NLR was 0.02; DOR was 2300.38; AUC=0.9970, Q*=0.9799. Simplexa: pooled sensitivity was 1.0; specificity was 0.97; PLR was 26.67; NLR was 0.01; DOR was 3100.93; AUC=0.9970, Q*=0.9800. Cobas ®: pooled sensitivity was 0.99; specificity was 0.96; PLR was 37.82; NLR was 0.02; DOR was 3754.05; AUC=0.9973, Q*=0.9810. RT-LAMP

COVID-19 has caused a worldwide epidemic and most countries have implemented a containment strategy as advised by the WHO to control further transmission (17,18).It is urgent to develop rapid, accurate diagnosis methods to effectively identify these early infected patients, treat them in time and control the disease spreading (19).
Based on the rapid characterization of the full genome of the virus, many molecular diagnostic methods have been developed rapidly.Real-time reverse transcription-polymerase chain reaction (rRT-PCR) is currently the standard of the diagnosis of acute coronavirus disease (COVID-19) (10,21,22).However, rRT-PCR assay has many limitations, such as required high purity samples, trained personnel and time-consuming.This test did not meet the rapidly growing need for virus testing in patients with COVID-19 infection, suspected infection or close contact with confirmed cases, which have significantly hampered public health efforts to contain the outbreak.To resolve this contradiction, the U.S. Food and Drug Administration (FDA) issued Emergency Use Authorization (EUA) for multiple SARS-CoV-2 rapid tests beginning in March 2020.Up to now, some studies have been published to comparing the diagnostic features of these methods (23)(24)(25), whereas the results were inconsistent.Therefore, it is urgent to performed this meta-analysis to comprehensively summarize the characteristics of the diagnostic assay in detecting COVID-19 to provide the preliminary evidence to support further guide clinical routine through evidence-based medicine.

Study selection
Pertinent articles were identified by two reviewers after screening titles or abstracts.Duplicated articles were removed by checking duplication.
Besides, references of selected articles were screened to find more relevant articles.Disagreement was resolved by the third reviewer.All eligible articles were evaluated by Quadas-2 for systematic reviews of intervention (Version 5.3.0).

Inclusion and exclusion criteria
Articles meeting the following criteria were included in this study: i) the articles studied on the method for detecting COVID-19; ii) patients recruited in articles should include COVID-19 and non-COVID-19; iii) patients defined as COVID-19 or non-COVID-19 were proved by gold standard; iv) data in articles were sufficient to analyze the pooled sensitivity and specificity; v) articles published in English.Incomplete data, reviews and case reports were excluded.

Data extraction
Two reviewers independently pooled the following data from eligible articles: author, year, Country, sample size, methods, true positive (TP), false positive (FP), false negative (FN), true negative (TN).When the extracted data were inconsistent, discrepancies were resolved by discussion or the third reviewer.

Statistical method
All statistical analyses were conducted by employing Meta-Disc 1.

Results
After searching three databases, out of 1094 duplicate articles, 3874 articles were identified.Then, 3837 articles were removed by screening the title or abstract.Afterwards, 19 articles were further excluded via thoroughly reading the full-text.Eventually, 18 articles of 3648 cases were included in this meta-analysis.Table 1

Quality evaluation
All eligible articles were assessed through the Quadas-2 tool.The result suggested that all articles exhibited good quality (Supplementary Figure 2).

Publication bias
The funnel plot exhibited that no significant publication bias in analysis on ePlex, Panther Fusion, Simplexa, RT-LAMP, Cobas ® and Xpert Xpress, because of the relatively symmetrical in this funnel plot (all P-value >0.1)(Supplementary Figure 3).

Discussion
The COVID-19 pandemic is putting enormous pressure on clinical and public health laboratories.The COVID-19 pandemic may be caused by widespread transmission, viral detection is key to isolate positive patients and stop viral transmission (44)(45)(46)(47)(48). rRT-PCR is the most reliable and widely used technology to diagnose viruses including coronaviruses (44)(45)(46)(47)(48).However, rRT-PCR has some limitations, which can not meet the huge demand for the global pandemic of COVID-19.Recently, the US FDA has been authorized multiple rapid molecular tests to meet the huge diagnostic need.Different diagnostic methods have proposed different virus targets for detection of the virus, which would detect SARS-CoV-2 and other related beta coronaviruses such as SARS-CoV, including RNA-dependent RNA polymerase (RdRp), envelope (E), spike (S), Open Reading Frame (ORF) 1a and nucleocapsid (N) (21,47,51,52).This article discusses some of the diagnostic methods include EUA-granted assays.
To date, this study was the first meta-analysis on systematically evaluating the diagnostic efficiency of different method for detecting COVID-19.In this study, we analyzed the pooled sensitivity, specificity, PLR, NLR, DOR, AUC and Q* on each methods (ePlex, Panther Fusion, Simplexa, Cobas ® , Xpert Xpress and RT-LAMP), respectively.The results demonstrated that these above methods bear higher sensitivity and specificity, and might be efficient methods complement to the gold standard.
The ePlex assay targets the N gene of SARS-CoV-2, which is an in vitro diagnostic test.The ePlex has a relatively short turnaround time, simple operational flow, but a shortage of supply and inventory limits its full implementation.
The Panther Fusion SARS-CoV-2 assay targets two conserved regions of ORF1ab in the same fluorescence channel.This platform is automated, high-throughput systems that can process >1,000 specimens in 24 hours, which is met the huge diagnostic need.RT-LAMP methodology is regarded as a new generation diagnostics (57), which was developed by Notomi et al.in 2000(58).This method has high sensitivity, high specificity, simple method, low cost and time saving, which has been widely applied for the detection of influenza virus, MERS-CoV, West Nile virus, Ebola virus, Zika virus, yellow fever virus, and a variety of other pathogens (59)(60)(61)(62)(63)(64).

4 ,
RevMan 5.3.2,Stata 12 software.The threshold effect was evaluated by computation of Spearman correlation coefficient between the logit of sensitivity and logit of 1-specificity.A strong positive correlation suggested the threshold effect.Heterogeneity among included articles was assessed via Chi-square and I-square tests.I 2 >50% or P<0.1 were deemed as significant heterogeneity existing.The random-effects or fixed-effects model was chosen to analyze the pooled data depending on the presented heterogeneity.The summary sensitivity and specificity, positive likelihood ratio (PLR), negative LR (NLR), diagnostic odds ratio (DOR), summary receiver operating characteristic curve (sROC) and calculate the AUC were utilized to evaluate the diagnostic efficiency of the aboved methods in detecting COVID-19.The publication bias was evaluated by Deek's funnel plot asymmetry test.P>0.1 suggested no significant publication bias in this study.
Simplexa COVID-19 Direct assay targeted two distinct regions of the SARS-COV-2 genome, the surface (S) gene and the open reading frame 1AB(ORF1ab), distinguish with FAM and JOE fluorescent probes.Compared with RT-PCR, which requires nucleic acid extraction from clinical samples, Simplexa COVID-19 Direct assay can be detected directly from clinical samples.This detection method is easy to operate and does not require additional equipment such as a centrifuge or extraction system, which indicated is promising for laboratory diagnosis of COVID-19 and field application.Cobas ® SARS-CoV-2 is a qualitative dual target assay, includes the ORF1/a nonstructural regional unique to SARS-CoV-2 and a region on the E gene, which is conserved across the sarbecovirus subgenus.Cobas ® SARS-CoV-2 is based on fully automated sample preparation (nucleic acid extraction and purification) followed by PCR amplification and detection(53).Automated solutions for molecular diagnostics can help process large numbers of samples, save testing time, allow non -professional operators.The assay has passed clinical evaluation and received EUA from the U.S. FDA(54-56).The Xpert Xpress SARS-CoV-2 assay targets two genes, the E-gene (Sarbeco specific) and N2-gene (SARS-CoV-2 specific), which received EUA status on March 20, 2020.The Xpert test platform integrates specimen processing, nucleic acid extraction, reverse transcriptase polymerase chain reaction amplification of SARS-CoV-2 RNA, and amplicon detection in a single cartridge, which improves actionability overall.This assay is simple to operate with the least technical interventions, and FDA has authorized trained non-laboratorians to test.

Table 1 .
the characteristics of included studies.