Evaluation of RdRp & ORF-1b-nsp14-based real-time RT-PCR assays for confirmation of SARS-CoV-2 infection: An observational study

Sir,

COVID-19 caused by SARS-CoV-2 has spread to most countries across the globe including India1. Laboratory diagnosis depends on the detection of viral RNA in nasopharyngeal and/or oropharyngeal swabs using real-time reverse transcription polymerase chain reaction (qRT-PCR)2. In India, the Indian Council of Medical Research-National Institute of Virology (ICMR-NIV) at Pune, adopted a two-step strategy for the diagnosis of COVID-19 using qRT-PCR. Primers and probes from two different protocols were combined, and initial screening was performed for E (envelope) gene specific to Sarbeco sub-genus. Samples positive in the screening test were further subjected to a confirmatory test targeting two genes, one SARS-CoV-2 specific RdRp (RNA dependent RNA polymerase) gene and other Sarbeco sub-genus ORF-1b-nsp14 gene3,4. The samples positive for either of the two genes were confirmed as positive for SARS-CoV-2.

To declare a sample positive, four reactions, each for the genes, E, RdRp, ORF-1b-nsp14 and RNaseP (internal control), are run. This strategy ensures the quality of the clinical sample as well as the testing process and identification of true positives. This two-step diagnostic protocol for SARS-CoV-2 detection is followed by the Virus Research and Diagnostic Laboratory Network (VRDLN)5,6. However, with increase in the number of suspected cases, testing for two confirmatory genes would utilize more time, consumables and workforce. Thus, there is an urgent need to revisit the strategy of using two confirmatory genes. In this context, we analyzed the qRT-PCR data of 313 SARS-CoV-2-positive cases tested at ICMR-NIV to find out the sensitivity of RdRp and ORF-1b-nsp14b gene-based assays to confirm SARS-CoV-2 infection.

All the 313 cases were positive by E gene screening [cycle threshold (Ct) values for E gene were ≤35] and were also positive by qRT-PCR for either RdRp or ORF-1b-nsp14 or for both genes (Ct value cut-off ≤36). Among the 313 samples, 79.2 per cent (n=248) were positive for both RdRp and ORF-1b-nsp14 genes. ORF-1b was exclusively positive in 8.2 per cent (n=57) samples, whereas RdRp was exclusively positive in 2.6 per cent (n=8) samples. The sensitivity with 95 per cent confidence interval (95% CI) for the detection of SARS-CoV-2 by ORF-1b-based assay was 97.4 per cent (95.0-98.7). The sensitivity with 95 per cent CI for the detection of SARS-CoV-2 by RdRp-based assay was 81.8 per cent (77.1-85.7). For comparing the Ct values of the RdRp and ORF-1b-nsp14 gene assays of all 313 positive samples, those showing undetermined Ct values were assigned a Ct value of 45, which was the maximum cycle number of qRT-PCR. The mean Ct value of ORF-1b-nsp14 gene assay was 28.8 and that of RdRp gene assay was 32.6. The mean Ct value was significantly lower in ORF-1b assay (P <0.001) using Student's t test.

The results of this observational study suggested that ORF-1b-nsp14-based assay performed well as a confirmatory assay as compared to RdRp-based assay. A recent study has also reported that the RdRp-based assay has missed 35 per cent of SARS-CoV-2 positive cases compared to a novel RdRp/helicase-based qRT-PCR assay7. A couple of studies posted in preprint servers observed that the primer-probe set reported by Corman et al3 for RdRp (with SARS-CoV-2-specific probe) assay had lower sensitivity compared to that of ORF-1b-nsp14-based assay, suggesting that it might be due to the presence of a degenerate base at the 12^th position of reverse primer8,9. The results of these studies are summarized in the Table I7,8,9,10,11,12. With these observations, negative results from commercial qRT-PCR kits which use primers and probe targeting RdRp gene as described earlier3 should be treated with caution and supported with confirmatory assays. We suggest the following diagnostic algorithm for SARS-CoV-2 detection, first screening with E gene-based assay3 followed by confirmation with ORF-1b-nsp14 gene-based assay4. Though both assays would detect all viruses from the Sarbeco sub-genus, as there is no current circulation of SARS CoV-1, positive results in both assays would mean SARS-CoV-2 positivity. Further, when a sample tests positive in E gene assay and negative in ORF-1b-nsp14 assay, testing in a third assay can be considered. Under the current situation, the third assay can be the existing RdRp-based assay. Alternatively, highly conserved regions in the S (spike glycoprotein) and N (nucleocapsid) genes can be explored for developing a confirmatory assay and used for the samples that test negative in the ORF-1b-nsp14-based confirmatory assay. The modified algorithm involving three-stage assay strategy will lead to a reduction in the number of reactions required for a positive sample. Instead of the four reactions required for a positive sample (screening, internal control and 2 confirmatory assays), only three might be required. Only the samples positive by E gene screening and negative by ORF-1b-nsp14 assay would require the fourth reaction. Such a strategy will save cost and time. Before implementing the strategy pan India during the current pandemic, data from all the testing centres may be analyzed to make an informed decision.

Table.

Summary of different published studies and studies in preprint servers comparing the RdRp based assay3 with other assays

Reference	Primer probe sets	Analytical sensitivity	Clinical sensitivity
Nalla et al10, 2020	Corman et al3, RdRp	63 viral genomic equivalents/reaction	-
	Corman et al3, E	6.3 viral genomic equivalents/reaction
	CDC N212	6.3 viral genomic equivalents/reaction
Chan et al7, 2020	Corman et al3, RdRp	-	Detected 28% positivity from 273 specimens from 15 patients
	RdRp/Hel7	-	Detected 43% positivity from 273 specimens from 15 patients
Barra et al11, 2020	Corman et al3, RdRp	350 viral genomic equivalents/reaction	-
	Corman et al3, RdRp modified primer/probe concentrations	33.7 viral genomic equivalents/reaction	-
Vogels et al8, 2020	ORF-1b-nsp14	Detected 10 viral genomic equivalents/µl	-
	Corman et al3, RdRp	Could not detect ≤100 genomic equivalents/µl	-
Lim et al9, 2020	Corman et al3, RdRp	7-43 times less sensitive than CDC

As the ORF-1b-nsp14-based assay detects both SARS-CoV-1 and SARS-CoV-2, it is hypothesized that even within the 132 base pair region amplified by the primers reported by Chu et al4, there is a possibility of presence of SARS-CoV-2 specific region to be used as a probe. When the 132 base pair region was analysed, a 24 base pair region downstream of the forward primer was identified, which could be targeted for designing a probe in reverse orientation, specific to SARS CoV-2. This region is highly conserved among SARS-CoV-2 genomes reported in Genbank and has multiple mismatches with SARS-CoV-1 genome. Because ORF-1b-nsp14 primer set has been reported to be more sensitive, adding an additional probe specific to SARS-CoV-2 tagged with a fluorescent dye (which is different from the dye used for tagging the already available probe which can detect both SARS-CoV-1 and CoV-2) might provide more information. Thus, a modified duplex ORF-1b-nsp14-based assay with an additional probe might be able to add more specificity to the assay in detecting SARS-CoV-2.

To conclude, the present study suggests a modified diagnostic algorithm for qRT-PCR-based diagnosis of SARS-CoV-2 in public health laboratories in India which will be cost-effective. This study also reports the design of a modified duplex ORF-1b-nsp14-based assay for discriminating SARS-CoV-1 and SARS-CoV-2 based on in silico analysis.