Emerging viral infections continue to pose a major threat to global public health. In the past decades, different viral emergencies have been reported including the severe acute respiratory syndrome coronavirus (SARS-CoV), H1N1 influenza, Middle East respiratory syndrome coronavirus, Ebola vírus, Zika virus, and most recently, the new coronavirus has been described, which cause coronavirus disease 2019 (COVID-19; Wang et al., 2020; Zhu et al., 2020). COVID-19’s etiologic agent is SARS-CoV-2, which belongs to the Coronaviridae family, Betacoronavirus genus (Gorbalenya et al., 2020; Rambaut et al., 2020). People with COVID-19 have a wide range of symptoms reported such as fever, cough, anosmia, ageusia, headache, fatigue, muscle or body aches, sore throat, and shortness of breath or difficulty breathing. Some of these symptoms help spread the virus; however, human-to-human transmission from infected individuals with no or mild symptoms has been extensively reported (Bai et al., 2020; Rothe et al., 2020). This outbreak has spread rapidly; as of September 2021, there were more than 230 million confirmed COVID-19 cases with more than 4.7 million deaths recorded worldwide¹. Isolation and quarantine of infected individuals are essential to viral spread and community dissemination of airborne pathogens and require an accurate, fast, affordable, readily available tests for massive population testing. In contrast to antibody detection, which may take weeks after the onset of the infection, detection of viral RNA is the best way to confirm the acute infection phase, the most important phase for viral shedding, so that rationally managed social distancing and lockdown can be implemented (Long et al., 2020; Wang et al., 2020).

Reverse transcriptase–quantitative polymerase chain reaction (RT-qPCR) is considered the gold-standard method for SARS-CoV-2 RNA detection, mainly targeting combinations of viral genome regions that codes for nucleocapsid protein (N), envelope protein (E), RNA-dependent RNA polymerase (RdRp), and other targets on the open reading frame (ORF1ab; Esbin et al., 2020). Although RT-qPCR assays have played an important role in the SARS-CoV-2 diagnosis, the technique has limitations for massive population testing such as processing time; it requires sophisticated equipment, infrastructure, and highly trained staff, as well as costly reagents with high demand and shortages around the world. Thus, developing complementary, inexpensive point-of-care (PoC) methods that are rapid and simple and allowing the use of alternative reagents for COVID-19 diagnosis test are urgently needed. Methods gathering these features can make affordable massive testing campaigns, including contact tracing strategies in highly dense countries, saving lives (Baek et al., 2020; Dudley et al., 2020; Song et al., 2021; Godfrey et al., 2020; Park et al., 2020; Wang, 2020; Yan et al., 2020; Yu et al., 2020; Anahtar et al., 2021). In this regard, reverse transcription loop-mediated isothermal amplification (RT-LAMP) has been shown to be an affordable technique applied to detect different pathogens (Mori and Notomi, 2009; Li et al., 2017). RT-LAMP has been used during Ebola outbreak (Kurosaki et al., 2016a,b) and for tracking Zika virus (da Silva et al., 2019) or Wolbachia (Gonçalves et al., 2014) in Brazilian mosquitoes. The method relies on specific DNA amplification at constant temperature without the need for sophisticated thermal cyclers (Zhang et al., 2020a). The amplified products can be visually detected through magnesium pyrophosphate precipitation, fluorescence emission from DNA intercalating dyes, agarose gel electrophoresis, lateral flow immunochromatography, magnesium chelating color indicators (Bhadra et al., 2021), and pH-dependent colorimetric reaction that changes from fuchsia (pink) to yellow (positive result) due to proton release during nucleic acid amplification (Tanner et al., 2015; Figure 1). The possibility of accessing results by the naked eye made RT-LAMP an exciting alternative that facilitates the use of COVID-19 molecular testing. Simple, scalable, cost-effective RT-LAMP–based alternatives for SARS-CoV-2 detection have emerged during pandemics including protocols for viral inactivation, quick run, RNA extraction–free and LAMP-associated CRISPR/Cas strategies (Baek et al., 2020; Broughton et al., 2020; Chow et al., 2020; Dudley et al., 2020; Song et al., 2021; Godfrey et al., 2020; Joung et al., 2020; L’Helgouach et al., 2020; Park et al., 2020; Rabe and Cepko, 2020; Bektaş et al., 2021; Bokelmann et al., 2021). On April 14, 2020, the RT-LAMP received the emergency use authorization from the United States Food and Drug Administration (FDA) for SARS-CoV-2 detection in COVID-19 diagnostics.

In this study, we optimized and validated a colorimetric RT-LAMP assay to detect SARS-CoV-2 RNA in clinical samples collected in different parts of Brazil, including samples with known SARS-CoV-2 variants of interest (VOIs) and concern (VOCs). After testing different primer sets for SARS-CoV-2 RNA detection by RT-LAMP, best results were achieved when using N gene or N/E genes-based strategies. 367 nasopharyngeal swabs collected in a guanidine-containing viral transport medium (VTM; Faria et al., 2021) from suspect patients were tested. The clinical validation revealed a sensitivity of 98% [95% confidence interval (CI) = 95.3–99.5%] with samples of cycle-threshold (Ct) values ranging from 15 to 32 with 100% specificity. We also demonstrated that RT-LAMP is affordable for the detection of more transmissible SARS-CoV-2 variants encompassing a number of genomic nucleotide changes. Part of the results presented here is the research basis of OmniLAMP^® SARS-CoV-2 kit, which was approved by the Brazilian Heath Regulatory Agency for COVID-19 molecular testing (Anvisa no: 10009010368) as an alternative for massive decentralized diagnostic in Brazil, which records the third-highest number of COVID-19 cases worldwide (see text footnote 1). Together with vaccination, RT-LAMP for COVID-19 diagnosis could help to improve better life quality during the pandemic, offering an alternative molecular testing for monitoring lockdown measures; traveling restrictions; the return of universities, schools, kindergartens; and sport league activities with worldwide impact.

The COVID-19 pandemics demanded a rapid global response in massive diagnostic solution to face the worldwide crisis. In this context, the RT-qPCR—considered the gold-standard technique for SARS-CoV-2 RNA detection—requires high-cost equipment, trained staff, and specialized laboratory infrastructure. In addition, during the COVID-19 pandemic, several health care centers and private laboratories competed for RT-qPCR kits and related products to meet the high diagnostic demand. In order to overcome the lacking of molecular testing and provide affordable alternatives, RT-LAMP had become one of the main hopes. Because of its simplicity, accuracy comparable with RT-qPCR to detect SARS-CoV-2 RNA, the fact that it does not require PCR machine, and for offering a naked eye readable colorimetric output, RT-LAMP is the focus of massive testing campaigns (Chow et al., 2020; Dudley et al., 2020; Song et al., 2021). This screening strategy is compatible with home, primary care clinics, point of entry (borders), schools, universities, sport leagues, and companies and can help to achieve a safe back-to-work and quarantine monitoring (Chow et al., 2020; Dudley et al., 2020; Song et al., 2021; Godfrey et al., 2020; Bokelmann et al., 2021). Since April 14, 2020, the United States FDA issued the emergency use authorization of Color SARS-CoV-2 RT-LAMP Diagnostic Assay from Color Health, Inc. (EUA no. EUA200539).

In order to provide an affordable SARS-CoV-2 detection tool, we validate a colorimetric RT-LAMP for the COVID-19 diagnosis using clinical samples collected from different parts of Brazil. The country has a flawed screening performance, testing fewer than 220 individuals per 1,000 people (May 2021)² where the majority of tests rely on antibodies-based rapid tests, which are not the most reliable and recommended for mass screening and decision making to control local outbreaks. The test sensitivity of RT-LAMP is comparable to the gold standard RT-qPCR and clearly relies on the target choice, incubation time, viral load (asymptomatic patients, days of symptoms, and correct sampling), output reading, sample integrity, and quality (viral transport media, sample storage condition, preanalytical treatments, extraction procedure, and crude RNA extraction–free samples), and sample type (nasal, nasopharyngeal, saliva, sputum, and gargle lavage; Table 2).

Upon RNA extraction from nasopharyngeal swab–derived clinical samples, we found an LoD of 20 viral genomic copies/μL, confirming previous studies based on N SARS-CoV-2 target (Anahtar et al., 2021; Bokelmann et al., 2021; González-González et al., 2021). It is worth noting that when using nonclinical SARS-CoV-2 extracted RNA or synthetic target, the LoD reaches less than 0.5 copies/μL. This can be explained by the presence of interferents such as VTM, host cells, and enzymes that could reduce the yield (Dudley et al., 2020; Nawattanapaiboon et al., 2021). In this regard, we have to be careful when interpreting LoD calculated using nonclinical samples. Nevertheless, extracted samples are rich enough in viral genomic copies to meet SARS-CoV-2 clinically relevant levels.

Clinical validation of RT-LAMP for COVID-19 diagnosis relies on calculating parameters, such as sensitivity, specificity, positive predictive value, negative predictive value, and accuracy compared to the gold standard RT-qPCR. We have to be careful when associating the RT-LAMP sensitivity, and indirect assumption on RT-qPCR viral load is not straightforward because of some technical concerns. It is well accepted that Ct values can be representative of viral load. However, this parameter could lead to misinterpretation when comparing different kits, targets, and nonstandardized samples. A survey conducted by the College of American Pathologists on more than 700 laboratories, reported a variation as much as 14 cycles among different methods on the same batch material. Single laboratories using different platforms and targets in SARS-CoV-2 molecular testing can represent a potential variability on Ct values (Rhoads et al., 2020). Considering previous convergent reports and presuming different targets and platforms, the data from the literature show that with an RT-qPCR Ct 30 cutoff, RT-LAMP sensitivity for SARS-CoV-2 detection is close to 100% (Schermer et al., 2020; Thi et al., 2020; de Oliveira Coelho et al., 2021; García-Bernalt Diego et al., 2021) and eventually with a higher threshold Ct 35 as well (L’Helgouach et al., 2020; de Oliveira Coelho et al., 2021). Indeed, we confirm that up to Ct 30 RT-LAMP returned 100% sensitivity for SARS-CoV-2 detection, reaching 98 and 94% when considering Ct values up to 32 and 34, respectively. Curiously, Kim et al. (2021) found that samples from hospitalized patients presenting Ct value of 28.4 or less were infective to human cell culture (Kim et al., 2021), an evidence based on in vitro extrapolation that RT-LAMP sensitivity is compatible with the threshold of infectivity. This reinforces that simple, robust, and reliable RT-LAMP meets clinical requirements, presenting similar COVID-19 diagnostic accuracy as RT-qPCR (Österdahl et al., 2020; Inaba et al., 2021; Jang et al., 2021).

The choice of SARS-CoV-2 genomic target plays an important role when selecting the RT-LAMP method for COVID-19 diagnosis. Several research groups have tested different regions on SARS-CoV-2 genome with the potential to generate RT-LAMP primers. Once the majority of primers were designed using the open source software Primer Explorer, it is expected that at some point, the default algorithm returned the same result or overlapping regions, independently identified in a context where molecular biology scientists everywhere in the world are working to tackle COVID-19 (Figure 9). According to our data compilation, N gene and ORF1ab regions (overlapping NSP3, As1e, and RdRp-coding sequences) were the most frequent targets chosen for SARS-CoV-2 RT-LAMP (Table 2 and Figure 9). Ganguli et al. (2020) and Zhang et al. (2020b) arrived at the same conclusion when selecting the SARS-CoV-2 N gene-targeting primer set after confirming better performances for RNA viral detection when compared to other targets (Ganguli et al., 2020; Zhang et al., 2020b). When testing N, E, and RdRp genes in true-positive, previously RT-qPCR–characterized clinical samples, we observed more false-negative outputs from assays using E and RdRp genes, corroborating what was previously reported. We also highlight that primer subsets within the same N target gene can contribute differentially to RT-LAMP test sensitivity (Supplementary Table 2). Furthermore, multiplexing different primer sets is encouraged in order to increase sensitivity (Figure 9; Kim et al., 2019; Mautner et al., 2020; Zhang et al., 2020b).

Another important (almost neglected) point is the fact that, although inspired by RT-qPCR target selection, few SARS-CoV-2 RT-LAMP approaches reported an internal control target to confirm the presence of human RNA and monitor sampling or extraction process (de Oliveira Coelho et al., 2021). Wilson-Davies et al. (2021) pointed out that the lack of amplification can happen for different reasons concerning the whole reaction, a specific well, or due to inhibitory substances, highlighting the importance of including internal control even before nucleic acid extraction, in order to be considered a reliable SARS-CoV-2 LAMP assay (Wilson-Davies et al., 2021). In this study, all clinical samples were previously characterized by RT-qPCR, including human RNAse P as housekeeping gene (internal control). In the current OmniLAMP^® assay, we included human b-actin RNA (rACTB) as internal control. Other constitutive targets for SARS-CoV-2 RT-LAMP include BPIFA1 (Bektaş et al., 2021), human 18S RNA (de Oliveira Coelho et al., 2021), and Statherin RNA (Sherrill-Mix et al., 2021a; Table 2).

Similar to its high sensitivity, obtained in this work and by other studies, the SARS-CoV-2 RT-LAMP specificity is undoubtedly high and is frequently reported as 100% without any cross-reactivity with other respiratory or SARS-CoV–unrelated viruses (Chow et al., 2020; Mohon et al., 2020; Park et al., 2020; Aoki et al., 2021; Nawattanapaiboon et al., 2021). We also confirm that the SARS-CoV-2 RT-LAMP solution presented here is highly specific and does not cross-react with Brazilian occurring seasonal influenza A and B, hRSV, or arboviruses.

Despite the advantages presented by purified and nucleic acid–enriched samples for SARS-CoV-2 RT-LAMP, RNA extraction–free protocols have attracted attention as they can be noninvasive (saliva-based), do not require additional steps and equipment, and fulfill point-of-sampling requirements. Indeed, the preanalytical phase on RT-LAMP is the bottleneck for PoC applications. For this reason, several studies highlighted the feasibility of primary RNA extraction–free approaches for SARS-CoV-2 RNA detection (Asprino et al., 2020; Ben-Assa et al., 2020; Dudley et al., 2020; Esbin et al., 2020; L’Helgouach et al., 2020; Schermer et al., 2020; Srivatsan et al., 2020; Anahtar et al., 2021; Lalli et al., 2021; Wei et al., 2021). Pretreatment of saliva samples includes heat sample inactivation, and the use of lysis/stabilizing buffers that can contain proteinase K, TCEP, EDTA, and DTT could help the viral RNA assessment maintaining its integrity (Ben-Assa et al., 2020; Lamb et al., 2020; L’Helgouach et al., 2020; Smyrlaki et al., 2020; Lalli et al., 2021; Newman et al., 2021; Yang et al., 2021). Caution must be taken when running colorimetric RT-LAMP as pretreatment could interfere on result outputs. One of the main limitations for direct sample test by colorimetric RT-LAMP based on pH-sensing is the false-positive result upon input sample addition (previous to amplification) because of naturally acidic samples (Thi et al., 2020; Bokelmann et al., 2021). To prevent spurious amplification due to the presence of DNA from oral microbiome, food, or host cells on primary samples, Bokelmann et al. (2021) treated samples with λ exonuclease that acts by preferentially digesting 5′-phosphorylated DNA, leaving nonphosphorylated primers or LAMP products intact (Bokelmann et al., 2021). Here we showed the preliminary results on RNA extraction–free (also pretreatment free) diluted 10× in hydrochloride guanidine-containing VTM nasopharyngeal samples directly accessed to compare colorimetric results. Three of five RT-qPCR true-positive, directly accessed samples returned positive yellow output on colorimetric RT-LAMP for SARS-CoV-2 detection. This provides clues on the use of unextracted samples for massive COVID-19 testing campaigns with a trade-off on cost-benefits for LoD and test sensitivity. A recent study on 559 swabs and 86,760 saliva samples performed a sample preparation method for RNA extraction–free and found diagnostic sensitivity of 70.35 and 84.62%, respectively, for swab and saliva samples (Kidd et al., 2021). Most of the high and medium viral load samples will be detected on unextracted protocols. However, to meet RT-qPCR detection sensitivity levels, this requires some type of purification step and RNA concentration (Broughton et al., 2020; Park et al., 2020; Zhang et al., 2020a).

We are currently observing rapid converging evolution of SARS-CoV-2 during the COVID-19 pandemic worldwide. Several reports alert for the emergence of VOIs and VOCs such as the alpha (B.1.1.7), first detected in England (ECDC threat assessment brief on December 20, 2020; European Centre for Disease Prevention and Control, 2020); Beta (B.1.351), initially reported from South Africa (Tegally et al., 2021); gamma (P.1 or B.1.1.28.1), which was identified in Japan but obtained from a traveler from Brazil (Faria et al., 2021); and more recently, the VOI kappa (B.1.617.1) and VOC delta (B.1.617.2) detected in India, responsible for the majority of new COVID-19 cases in many countries in different parts of the world. The regional selection of SARS-CoV-2 VOC is associated with higher transmissibility, mortality and reduced neutralizing antibody response (Shah et al., 2020; Davies et al., 2021a,b; Li et al., 2021). In Brazil, we observed the emergence of different SARS-CoV-2 VOCs and VOIs, including gamma (P.1), zeta (P.2; Resende et al., 2021a; Voloch et al., 2021), B.1.1.33.9 (N.9; Resende et al., 2021c), and B.1.1.33.10 (N.10; Resende et al., 2020, 2021b). A plethora of mutations is observed in these variants, including N501Y, E484K/Q, K417N/T, A570D, and the Δ69–70 at the SARS-CoV-2 S protein sequence, which was associated with detection failures by S-target RT-qPCR methods (Brown et al., 2021). For SARS-CoV-2 RT-LAMP detection, few studies selected S-coding protein region as a target (Figure 9). In addition, isothermal amplification for SARS-CoV-2 RNA detection strategies is commonly addressed as multiplex targeted, making RT-LAMP a good choice even for SARS-CoV-2 variant detection. Indeed, here we reported that singleplex N gene-based or multiplex N2/E1-based RT-LAMP was able to perfectly detect VOCs and VOIs circulating in Brazil such as gamma (P.1), zeta (P.2), B.1.1.374, and B.1.1.371 (Figure 8 and Supplementary Figure 1D), the two latter first detected in Finland and Saudi Arabia³. Recent efforts made by Sherrill-Mix et al. (2021a,b) showed a beacon-based RT-LAMP strategy designed to precisely identify alpha (B.1.1.7) SARS-CoV-2 variant (Sherrill-Mix et al., 2021a,b), a promising tool not only for massive screening but also to monitor VOC/VOI SARS-CoV-2 spreading.

The colorimetric RT-LAMP is a reliable molecular tool for detecting SARS-CoV-2, providing rapid and easy-to-read results, compatible with high-throughput screenings and PoC requirements. This test is especially important for nations with poor diagnostic conditions, such as Brazil, where RT-qPCR COVID-19 diagnostic is far from ideal to control disease spreading. The RT-LAMP sensitivity can be equivalent to those reported from the gold standard RT-qPCR method and also present 100% specificity. Results are commonly obtained after 30-min reaction and if needed, additional 20 min was not associated with spurious unspecific amplification. Sample collection in guanidine-containing VTM has been described as a useful strategy to avoid contamination of health care workers during sample manipulation. RT-LAMP primer selection can directly interfere on sensitivity, being N genes the best target for SARS-CoV-2 RNA detection with fewer false-negative results, especially in low viral load samples, which is improved upon multiplexing E/N targets. Colorimetric RT-LAMP is also compatible with detecting SARS-CoV-2 VOIs and VOCs, being robust to cope with the monitoring of emerging new SARS-CoV-2 variants and that can be easily adapted. We thus reinforce and recommend the use of RT-LAMP for massive testing as a decentralized PoC alternative to avoid SARS-CoV-2 spread and to tackle COVID-19.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.gisaid.org/, The genomes of SARS-CoV-2 VOI and VOCs generated are deposited on GISAID according to the following accession codes: EPI_ISL_2221860, EPI_ISL_2221850, EPI_ISL_2221873, EPI_ISL_2221890, EPI_ISL_2221902, EPI_ISL_2221885, EPI_ISL_2221844, and EPI_ISL_2221866.

The studies involving human participants were reviewed and approved by Research Ethics Committee involving human beings at Instituto René Rachou, Fundação Oswaldo Cruz, under license protocol number: 4084902 and CAAE (certificate of presentation for ethical appreciation): 31984720300005091. The ethics approval was issued on June 12, 2020. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

PA, IB, and RM-N: conceptualization and experimental design. EO, AF-L, LA, AG, IB, and FR: investigation and performed experiments. AF-L, LA, AG, EO, PA, and RM-N: analyzed the data. MB, PM, FC, HM, GW, ST, and GLW: contributed to reagents, materials, and analysis and tools. PA and RM-N: supervision. RM-N, EO, AF-L, and RR: writing—original draft. PA, ST, GLW, GW, MB, and RM-N: writing, review, and editing. All authors discussed the results and contributed to the final manuscript.

HM is part of Visuri company. Results presented here are the basis of a COVID-19 RT-LAMP diagnostic test offered by Visuri named OmniLAMP^® SARS-CoV-2 kit. PA and RM-N are co-founders and scientific advisors at CEPHA Biotech. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.