Multiple-Centre Clinical Evaluation of Rapid Recombinase-Aided Amplification Assays for Five Pathogens


 Background: Recombinase-aided amplification(RAA) is a new, simple, and ultrafast isothermal molecular diagnostic technique performed within 30min at 39°C–42°C.In this study, we evaluated the clinical performance of four duplex RAA kits for hepatitis B virus(HBV), human adenovirus 3(HAdV3), human adenovirus 7(HAdV7), and Bordetella pertussis and one duplex reverse-transcription RAA (RT-RAA) kit for respiratory syncytial virus (RSV).Methods: A total of 392 sera and 374 respiratory tract samples were collected from five institutions in four China regions. Each RAA kit’s sensitivity and specificity were compared with those of real-time quantitative polymerase chain reaction(qPCR),real-time quantitative reverse-transcription polymerase chain reaction(qRT-PCR), or sequencing. Results: Compared with qPCR or qRT-PCR, the sensitivities of HBV RAA,RSV RT-RAA, and B.pertussis RAA were 97.55%,96.67%, and 100%,respectively,and all of the specificities were 100%.The total coincidence rates were 97.78%(383/392,95%CI:95.63%–98.85%),97.70%(212/217, 95%CI:94.57%–99.16%), and 100%(60/60,95%CI:92.80%–100%),respectively.The Kappa values were 0.977,0.947, and 1,respectively(P<0.05).Regarding the sequencing, the sensitivities of HAdV3 RAA and HAdV7 RAA were 100% and 97.37%, respectively,and all specificities were 100%.The total coincidence rates were 100%(97/97,95%CI:91.58%–100%) and 98.97%(96/97,95%CI:94.39%–99.82%),and the Kappa values were 1 and 0.978 (P<0.05),respectively.Conclusions: With comparable clinical performance, these RAA kits are suitable assays for rapidly detecting pathogens in resource-limited laboratories.


Background
Infection with pathogens remains a wide spread problem globally and places a severe disease burden on society and individuals [1][2][3] .For example,the human hepatitis B virus(HBV) is a blood-borne pathogen that can cause serious complications, such as liver cirrhosis, liver cancer, and other chronic liver diseases 2 .Respiratory tract infection is a common condition caused by various pathogens in people of all ages,especially in infants and young children. Such infection can involve respiratory syncytial virus(RSV), human adenovirus(HAdV), and Bordetella pertussis,which is di cult to distinguish only by the clinical symptoms [3][4][5][6] .Therefore,rapid etiological identi cation in the early stage is essential for treating, preventing, and controlling the disease.The pathogen isolation and culture and immunological detection methods have limitations in the rapid diagnosis of pathogens in the early stage of the disease because of being time-consuming or having a low detection rate [7][8] .Real-time quantitative polymerase chain reaction (qPCR) with high sensitivity and speci city increases clinical laboratories' diagnostic accuracy [7][8][9][10][11] .Though commercial PCR-based detection kits and thermal cycling equipment are widely used and improve work e ciency in centralized laboratories,they are challenging to use in resource-limited laboratories because of the need for costly instruments and highly skilled professionals [12][13][14][15] .The simple and rapid isothermal nucleic acid ampli cation technology overcomes the di culty of applying complex PCR technology and instruments and thus is more suitable for grassroots units 16 .
Recombinase-aided ampli cation(RAA),a novel isothermal nucleic acid ampli cation technique,was reported to detect various pathogens within 30min at 39-42°C in vitro [17][18][19][20] .Under the activation of magnesium acetate,the primers and recombinase complex searches for and complements the homologous sequence of double-stranded DNA with the help of a single-strand binding protein. The extension of double-stranded DNA is completed under the action of DNA polymerase.With the introduction of reverse transcriptase and a 46-52bp probe, RAA can simultaneously perform reverse transcription and real-time uorescence detection in a single closed tube 19,21 .
In previous studies,we reported rapid duplex real-time RAA detection assays for HBV 22 ,HAdV3 23 , HAdV7 23 , and B. pertussis 24 , and duplex reverse-transcription RAA(RT-RAA)assay for RSV 25 .These assays show high sensitivity and speci city and incorporate a non-competitive internal control into the system to prevent false-negative results and increase clinical sample detection accuracy. However,a comprehensive evaluation of the clinical performance of these methods has not yet been carried out.We freeze-dried the primers,probes, and enzymes in reaction unit tubes to make ready-to-use kits.These kits have passed internal quality assessment at the Department of Facility,National Institute for Viral Disease Prevention and Control,the Chinese Center for Disease Control and Prevention (CCDC).In this paper,we report the evaluation of these ve RAA kits using a large number of clinical samples from ve institutions in four regions of China: Hunan Center for Disease Control and Prevention (CDC) and Hunan Provincial People's Hospital,Beijing Capital Institute of Pediatrics,Suizhou CDC in Hubei,and Tangshan Gongren Hospital in Hebei.

Samples
From January 2019 to January 2020,we collected clinical samples from patients in ve institutions in four regions of China(Hubei, Hebei, Hunan, and Beijing) for multicenter clinical evaluation.As shown in Table 1,these samples were divided into four groups.Group A: Atotal of 392 serum samples were collected,of which 223 were from Suizhou CDC in Hubei Provinceand 169 were from Tangshan Gongren Hospital in Hebei Province.Group B: A total of 217 sputum and bronchoalveolar lavage uid samples were collected,of which 121 samples were from Hunan Province CDC and 96 samples were from Capital Institute of Pediatrics in Beijing.Group C: A total of 97 sputum and bronchoalveolar lavage uid samples were collected from Hunan Province CDC and Hunan People's Hospital.Group D: A total of 60 nasopharyngeal swab samples were collected from the Capital Institute of Pediatrics in Beijing.All aspects of the study were conducted as per the national code of ethics and approved by the institutional review committees of the abovementioned medical institutions and hospitals. Detection of clinical samples using RAA and RT-RAA kits The serum samples in group A were detected using 2µL of extracted DNA/RNA with the HBV RAA kit, in accordance with a previous report but with a slight modi cation to the DNA extraction method 22 .Samples in group B were detected using 5µL of extracted DNA/RNA with the RSV RT-RAA kit in accordance with a previous report 25 .Samples from group C were detected using 2µL of extracted DNA/RNA with the HAdV3 and HAdV7 RAA kits in accordance with a previous assay 23 .Nasopharyngeal swabs in group D were detected using 2µL of extracted DNA/RNA with the B. pertussis RAA kit in accordance with a previous report 24 .Positive controls(recombinant plasmids) of the above ve pathogens and negative controls(DNase-free water) were included in each run to ensure the reliability of the experimental results.The primers and probe sequences of the above ve RAA kits are shown in Table 2.The FAM channel was used to detect the ampli cation of the target gene,and the HEX channel was used to detect the ampli cation of the internal control gene.If both channels were positive or the FAM channel was positive and the HEX channel was negative,the results were considered to be positive.If the FAM channel was negative but the HEX channel was positive,the result was negative.If both channels were negative, the result was considered invalid and the RAA assay was redone.If the samples had discordant results,these samples were retested with the corresponding RAA kit by optimizing the reaction conditions,such as increasing or reducing the amount of input template, or increasing the premixing time of the RAA reaction to fully mix the RAA reaction system. Among the 60 samples in group D,50 positive samples and 10 negative samples were detected by the B.pertussis RAA kit,which were consistent with the qPCR results. The sensitivity and speci city were both 100%.The PPV was 98.63% (95%CI:91.11-100%),the NPV was 100%(95%CI:65.55-100%),and the total coincidence rate was 100%(95%CI:92.80-100%). The Kappa value was 1(P < 0.05).    27 , and RAA 17,22,25 ,have demonstrated great potential to be used in low-income countries and regions with limited resources and di cult conditions.LAMP has been successfully applied to the nucleic acid detection of the above ve pathogens, with good sensitivity and speci city 26,28−30 .However,it is easy to produce false positive results due to cross-reactions associated with the four to six primers used in the method 31 .RPA has also been successfully applied to the detection of RSV and HBV,whose sensitivity and speci city are comparable to those of our RAA methods 27,32−33 .However, few RPA methods introduce an internal reference to monitor the reaction system, preventing clari cation of the authenticity of the nucleic acid detection.
In our work,the introduction of non-competitive internal controls in these ve kits greatly reduced the false-negative rate caused by experimental operation errors and system errors 19,25 .Two uorescent probes complementary to the target genes and internal reference quality controls were added to monitor the whole reaction in real time, and the results can be observed within 15-30min 17,22−25 .It turned out that the HEX channel had a steady positive curve in all of the experiments in our study, thus con rming the reliability of our results.
We retested and veri ed the samples whose RAA results were inconsistent with qPCR or qRT-PCR and sequencing results by optimizing the reaction system.We extended the premixing time to 8 min. As a result, ve false-negative samples became positive in group A and four false-negative samples became positive in group B.The premixing step is necessary and critical to fully oscillate and mix the RAA reaction system and maximize the likelihood of inter-molecular contact before uorescence signal detection 19 .Therefore, our work con rmed that increasing the premixing time is bene cial to improve the repeatability and rate of nucleic acid detection.In addition,we increased the amount of input template in the HBV RAA system,and two HBV false-negative samples (2.73 × 10 2 IU/ml and 3.45 × 10 2 IU/ml,respectively)became positive,indicating that more templates might improve the detection rate of the samples with low viral load.In group A,we found that nine negative samples missed by RAA had CT values in the range of 25.29-37.80.Considering that an inhibitor might be present in these samples,we diluted these nine samples and repeated duplex RAA experiments and single RAA experiments(without an internal quality control),yet there were no positive results.We suspect that this might have been due to gene mutation in the RAA primer or probe region, although the exact reasons need to be explored in further work.In the case of group B, ve samples were still missed by RAA after retesting, of which three had Ct values greater than 38,one had a Ct value of 33,and another one 35,suggesting that the RSV RT-RAA kit exhibits slightly low clinical sensitivity.Further improvement should be performed to optimize the RAA system and working conditions.In group C,HAdV3 and HAdV7 kits maintained very high speci city as no cross-reaction was observed with the other four types of HAdV samples(HAdV1, 2, 4, and 55).No false positive or false-negative results were found with the RAA HAdV3 and HAdV7 kits, except for one false-negative sample being misdiagnosed by RAA HAdV7.
We conducted clinical veri cation in clinical samples from multiple regions in China to test the reliability of the ve RAA kits.Among them, the RAA kit for HBV and the RT-RAA kit for RSV were evaluated using clinical samples from two different regions(one in a southern city and the other in a northern city) in China.RAA kits for HAdV3 and HAdV7 were evaluated using clinical samples from two different institutions.No signi cant difference in the testing results was observed among the different sites of sample collection,indicating wide applicability in testing.Moreover,four different sample types were included in the evaluation(serum, sputum,bronchoalveolar lavage uid,and nasopharyngeal swab),suggesting that the nucleic acid extraction method in this study works well with these sample types.
Local coworkers carried out all of the tests with the portable RAA uorescence detector and vortex mixer in the specimen preservation sites.The total turn-around time(from sample in to results out) is 45 min for 16 samples per run per person,while the detection time of the PCR kits used in this experiment using the same samples is more than 70 min.All of the RAA reactions were performed in closed tubes, reducing the likelihood of laboratory contamination.In addition,RAA kits preserve the reaction reagents(enzymes, probes, and primers)in the form of freeze-dried powder,which is easy to transport at room temperature and helpsto reduce the follow-up operation steps 22,24 . The costs of RAA kits are lower than those of the LAMP and PCR assays 21,24 .These advantages of RAA kits promote their application in primary laboratories.However,the step of conventional nucleic acid extraction is not skipped in this study, which hinders the eld use of RAA kits.Hence, we attempted to simplify the nucleic acid extraction by using a DNA releasing agent atroom temperature and demonstrated that RAA kits for HAdV 3 and 7,HBV,and B.pertussis using DNA releasing agent achieved similar results to those obtained using the conventional nucleic acid extraction method(data not shown), which implies the potential for eld use of these kits.Another alternative is to integrate nucleic acid extraction with the RAA reaction in a new portable device, which is under development.

Conclusions
With clinical performance comparable to that of commercial qPCR or qRT-PCR assays,the RAA kits in this study are suitable tools for the rapid detection of HBV,RSV,HAdV3,HAdV7, and B. pertussis in resource-limited laboratories and also have potential for eld use.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.

Con ict of interests
All the authors approved the nal manuscript and they have no con ict of interest to declare. Shen, Yuan Gao wrote the paper. All authors have read and agreed to the published version of the manuscript.