Optimization and Clinical Evaluation of a Multi-Target Loop-Mediated Isothermal Amplification Assay for the Detection of SARS-CoV-2 in Nasopharyngeal Samples

Abstract

SARS-CoV-2 is the coronavirus responsible for COVID-19, which has spread worldwide, affecting more than 200 countries, infecting over 140 million people in one year. The gold standard to identify infected people is RT-qPCR, which is highly sensitive, but needs specialized equipment and trained personnel. The demand for these reagents has caused shortages in certain countries. Isothermal nucleic acid techniques, such as loop-mediated isothermal amplification (LAMP) have emerged as an alternative or as a complement to RT-qPCR. In this study, we developed and evaluated a multi-target RT-LAMP for the detection of SARS-CoV-2. The method was evaluated against an RT-qPCR in 152 clinical nasopharyngeal swab samples. The results obtained indicated that both assays presented a "good concordance" (Cohen's k of 0.69), the RT-LAMP was highly specific (99%) but had lower sensitivity compared to the gold standard (63.3%). The calculated low sensitivity was associated with samples with very low viral load (RT-qPCR Cq values higher than 35) which may be associated with non-infectious individuals. If an internal Cq threshold below 35 was set, the sensitivity and Cohen's k increased to 90.9% and 0.92, respectively. The interpretation of the Cohen's k for this was "very good concordance". The RT-LAMP is an attractive approach for frequent individual testing in decentralized setups.

Keywords: N; ORF3a; ORF8; RT-LAMP; SARS-CoV-2; clinical evaluation.

Figure 1

Circular representation of the complete genome of SARS-CoV-2 (NC_045512) corresponding to the isolate Wuhan-Hu-1, highlighting in green the different viral genes and in orange the fragments amplified by the primers selected in this study.

Figure 2

Typical amplification plots (a) and melt curves (b) obtained with the synthetic control 1 for ORF8, ORF3a, and N. ΔRn is the magnitude of normalized fluorescence signal, relative to the baseline fluorescence, generated by the reporter at each cycle during the amplification.

Figure 3

Detection probability of the fluorescence-based RT-LAMP for all the targets (a). The results of the probit analysis is shown in (b) for ORF8, (c) for ORF3a, and (d) for N. Graphs generated with MedCalc.

Figure 4

RT-qPCR Cq values obtained with the Allplex 2019-nCoV Assay for all the positive samples identified in this study. In each column, the average along with the standard deviation is preScheme 49. For positive samples identified by the reference RT-qPCR, 34 were positive for all the screened genes (E, RdRp/S, and N). One was positive for the combination E- RdRP/S, 5 for RdRP/S- N, 2 were positive for the combination E/N, and regarding samples with one single positive target by RT-qPCR, 1 sample was positive only for RdRP/S, 3 only for N, and another 3 only positive for E.