This study is among the first independent on-field evaluations of the HG COVID-19 RT-LAMP assay. Our findings confirmed the high diagnostic performance of this assay in detecting SARS-CoV-2 in NP samples with a wide (15–39) range of Ct values. We also demonstrated that the recent predominance of the Alpha variant of concern did not affect the analytical performance of the kit. Indeed, the HG COVID-19 assay targets the highly conserved N region, while the novel variants of concern present key mutations in the S region. These mutations may affect the diagnostic performance of RT-PCR; indeed, so-called “S-gene target failure” is very common in samples positive for the Alpha variant of concern [21, 22]. Finally, our study is the first to validate the HG COVID-19 assay on an alternative extraction platform, which shows that RT-LAMP can be run successfully on different types of laboratory equipment.
According to the manufacturer’s instructions for use, the relative (vs RT-PCR) sensitivity and specificity values of the HG COVID extracted format in symptomatic patients are 100% (95% CI: 84.5–100%) and 100% (95% CI: 89.1–100%), respectively. In asymptomatic individuals, these parameters are 88.5% (95% CI: 75.9–95.2%) and 97.1% (95% CI: 92.9–98.9%), respectively. Considering that our sample included both symptomatic and asymptomatic subjects, our findings are consistent with the on-label analytical performance and testify to the generalizability of our results.
Our results are also in line with the meta-analytical estimates obtained by Subsoontorn et al. [14]. In their systematic review, the pooled (N = 26 studies) sensitivity and specificity values of the extracted RT-LAMP format versus RT-PCR were 94% (95% CI: 90–96%) and 100% (95% CI: 99–100%), respectively. When the analysis was restricted to high viral load samples (Ct <30), the pooled estimates (N = 10 studies) were 100% (95% CI: 89–100%) and 100% (95% CI: 99–100%), respectively [14]. In our study, while some false negatives with high RT-PCR Ct values (>30) were expected, some comments on the “false positives” should be made. It has been suggested that some non-specific amplifications may occur during RT-LAMP [23]. We documented a total of three false positive results; of these, however, two subjects had proved positive on RT-PCR 1–3 weeks earlier. On the one hand, it is well-known that the RT-LAMP technique can amplify up to 100 times more RNA copies than RT-PCR [12]. On the other hand, the RT-PCR assay used in this study (Allplex 2019-nCoV; Seegene Inc., South Korea) considers samples with Ct ≤ 40 to be positive [17]. It is therefore likely that the two above-mentioned subjects with false positive results could have had Ct values over 40. In turn, this may also mean that the reported specificity of the HG COVID-19 assay might have been underestimated. There is an ongoing debate on the association between infectiousness and RT-PCR Ct values. It is generally believed that infectiousness is significantly lower for clinical specimens with Ct values >30 [24, 25]. However, a substantial number of samples with Ct >35 may still produce a viable virus [26–28]. Moreover, Ct values across different RT-PCR protocols may vary significantly [19].
The possibility of rapidly obtaining a precise result makes RT-LAMP technology an attractive point-of-care or near-the-patient tool. This feature of RT-LAMP resembles that of Ag-RDTs. The World Health Organization [29] recommends that SARS-CoV-2 Ag-RDTs should have sensitivity and specificity values of at least 80% and 97%, respectively, while the European Centre for Disease Control and Prevention [24] has suggested that the sensitivity of point-of-care tests should be at least 90%. The HG COVID assay satisfies these criteria. An assessment of several Ag-RDTs recently performed at our laboratory [30] showed overall sensitivity of 78.7% (95% CI: 73.2–83.3%); this, however, varied significantly by Ct value and brand. We concluded that Ag-RDTs are convenient for screening purposes in moderate-to-high intensity settings [30]. In the present study, the analytical performance of the HG COVID-19 assay was comparable to that of RT-PCR; the assay may therefore be an alternative molecular diagnostic tool for SARS-CoV-2 detection, independently of the purpose and the viral epidemiology. On the other hand, it is likely that not all commercially available RT-LAMP kits perform well. For instance, a real-world evaluation of the diagnostic performance of the Isopollo COVID-19 RT-LAMP assay (M Monitor, South Korea) showed a sensitivity of only 61.9% [31]. Independent and setting-specific pilot evaluations would therefore be beneficial before the widespread implementation of RT-LAMP assays.
Like that of most available Ag-RDTs, the HG COVID readout is qualitative. However, the time to threshold is a good proxy of viral load. Indeed, we observed a strong linear association between the HG COVID-19 time to result and RT-PCR Ct values. Results that are available in less than 15–20 min are highly suggestive of high viral loads (Ct <30).
Our study is not without limitations. First, for ethical reasons, we were not able to link the RT-PCR readout to the clinical characteristics of patients (e.g., presence of symptoms, days after the onset of symptoms, etc). According to the manufacturer’s instructions for use, the performance of the HG-COVID assay is better in symptomatic cases, especially when the test is performed soon after the onset of symptoms. Second, we cannot completely rule out misclassification bias. Although RT-PCR is currently considered the “gold standard” assay for the laboratory diagnosis of both symptomatic and asymptomatic cases [1, 4, 5], its sensitivity is not perfect [7]. In our study, two samples from patients with previously confirmed SARS-CoV-2 infection were positive on RT-LAMP but negative on RT-PCR. This means that, in some instances, RT-LAMP may be more sensitive than RT-PCR, and that the reported relative specificity of RT-LAMP might have been underestimated. Finally, the study was carried out in a period when Delta variant of concern did not circulate in Italy. We, however believe that this possible limitation has a limited impact on the study conclusions for two reasons. First, the HG COVID assay targets the highly conserved N region, while the key mutations of the Delta variant are located in the S region. Second, our subsequent routine use of the HG COVID assay was able to detect isolates belonging to the Delta variant (results not shown).
In conclusion, HG COVID-19 RT-LAMP is a reliable assay for the molecular diagnosis of SARS-CoV-2 in NP samples, and yields a final diagnosis in less time than RT-PCR. RT-LAMP technology is promising for use in small and medium-sized hospitals, emergency departments and general practices without sophisticated laboratory equipment.