Comparison of four commercial RT‐PCR diagnostic kits for COVID‐19 in China

Abstract We compared the sensitivity and specificity of four commercial coronavirus disease (COVID‐19) diagnostic kits using real‐time reverse transcription–polymerase chain reaction (RT‐PCR). Kits I‐IV approved by the State Drug Administration of China were selected, and the detection targets were ORF1ab gene and N gene. Specificity was evaluated by detecting other respiratory viruses. The sensitivity and batch effect of each kit were evaluated by testing 10‐fold dilutions of RNA. Clinical application was verified by testing nasopharyngeal swab and sputum specimens from COVID‐19 patients. Among the 78 cases infected by other respiratory viruses, no amplification curve was observed using these four COVID‐19 RT‐PCR kits. The minimum detection limits of kits I‐IV were 10−6, 10−5, 10−5, and 10−6 dilutions, respectively, and concentrations were 10 copies/mL (10−5 dilution) and 1 copies/mL (10−6 dilution). The sensitivities of kits I‐IV detected using 142 nasopharyngeal swab specimens from COVID‐19 patients were 91.55%, 81.69%, 80.28%, and 90.85%, respectively, while they were 92.68%, 85.37%, 82.93%, and 93.90%, respectively, for the 82 sputum samples. The specificity of each kit was 100.00% (77/77). The total expected detection rate using sputum samples was 88.59% (691/780) higher than 86.15% (672/780) of nasopharyngeal swabs. Comparison of nasopharyngeal swab and sputum samples from the same COVID‐19 patient led to the detection of ORF1ab and N genes in 16 (100%) sputum samples; only ORF1ab and N genes were detected in 12 (75%) and 14 (87.5%) nasopharyngeal swab specimens, respectively. In conclusion, comparison of commercial COVID‐19 RT‐PCR kits should be performed before using a new batch of such kits in routine diagnostics.


| INTRODUC TI ON
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) emerged among humans during the final months of 2019, causing severe acute respiratory diseases, multiple organ injuries, and fatal outcomes. [1][2][3][4] The resulting disease, therefore, has been named coronavirus disease (COVID-19). 1-3 SARS-CoV-2 is a human coronavirus (HCoV). HCoVs are enveloped viruses with a single-stranded, positive-sense RNA and belong to the order Nidovirales. 5,6 The length of HCoVs is approximately 27-32 kilobases, and these viruses are divided into seven species, including HCoV-229E, HCoV-NL63, HCoV-OC43, HCoV-HKU1, SARS-CoV, MERS-CoV, and SARS-CoV-2. 7,8 Real-time reverse transcription-polymerase chain reaction (RT-PCR) is the most sensitive and specific assay that can provide crucial etiological evidence for COVID-19 diagnosis. 9,10 The coronavirus nucleocapsid (N) protein is expressed through the production of subgenomic messenger RNAs, and the number of N proteins markedly exceeds that of genomic RNAs in several stages of the replication cycle. 6,8,11,12 The RNA-dependent RNA polymerase, called ORF1ab, is the main region for virus replication and transcription. 6,8,11,12 Therefore, both ORF1ab and N genes are the crucial targets used for RT-PCR-based SARS-CoV-2 detection. 13,14 Recently, the efficacy of RT-PCR for COVID-19 diagnosis has been questioned. 15 Although several COVID-19 RT-PCR diagnostic kits are commercially available, the detection rates of SARS-CoV-2 infection have been unsatisfactory, and several cases have been detected following negative detection results obtained from repeated RT-PCR laboratory diagnostic tests and COVID-19 features already observed on computed tomography images. [16][17][18][19][20] Currently, there is no better diagnostic method for COVID-19 than RT-PCR. 21 The most significant steps for SARS-CoV-2 detection in a RT-PCR diagnostic laboratory are to identify and use RT-PCR kits with high sensitivity and specificity. [16][17][18][19][20] Comparisons of sensitivity and specificity among different commercial RT-PCR diagnostic kits are still limited. Moreover, there is a dearth of information on the comparison methods. Therefore, this study aimed to compare the sensitivity and specificity among four commercial COVID-19 RT-PCR diagnostic kits from different manufacturers and suggest comparison methods that may be employed to identify efficient kits for routine diagnostics. Bioscience Co., Ltd (listed alphabetically). All the kits were suitable for ABI 7500 real-time PCR system, and the detection targets were ORF1ab gene, N gene, and ribonucleoprotein (RNP).

| Clinical specimens
Clinical specimens, including nasopharyngeal swab and sputum specimens, were collected from confirmed or excluded COVID-19 patients experiencing acute respiratory tract infection. The criteria for confirmed COVID-19 cases included the following: (a) already-

| Nucleic acid extraction
Total nucleic acids (RNA and DNA) were extracted from the clinical specimens using Thermo Scientific ™ KingFisher ™ Flex Magnetic Particle Processors (cat no. KFR-805496; Thermo Fisher: Waltham, MA). Approximately 60 µL of total nucleic acid eluates was recovered in nuclease-free tubes and tested immediately.

| Specificity assessment
Nasopharyngeal swab specimens containing other respiratory viruses, including influenza virus A, influenza virus B, respiratory syncytial virus, parainfluenza virus, human adenovirus, human rhinovirus, other HCoVs (NL63, OC43, 229E, and HKU1), human metapneumovirus, and human bocavirus, were also collected to identify the specificity of the assessed kits. A commercial RT-PCR kit (cat. no. CN12-33-100; Jiangsu Uninovo Biological Technology Co. Ltd., Jiangsu, China) was used to detect these viruses.

| Sensitivity assessment
Ten fold serial dilutions of nucleic acid eluates of the COVID-19 case were tested in duplicate to evaluate the sensitivity of the kits. The concentrations of the nucleic acid eluates (10 −1 to 10 −6 dilutions) were 27 230 copies/mL, 5399 copies/mL, 1395 copies/mL, 437 copies/mL, 10 copies/mL, and 1 copies/mL, respectively, as detected by digital

| Clinical application of nasopharyngeal swab specimens
The types and corresponding patients of the 189 specimens are shown in Table 2

| Detection rates of sputum and nasopharyngeal swab specimens
Direct standardization was performed using SPSS software, version 20.0, to select sample type between the nasopharyngeal swab and sputum specimens for SARS-CoV-2 detection. The expected detection numbers of all the specimens are shown in Table 4, and the total expected detection rate of the sputum specimens was 88.59% (691/780), which was higher than that of the nasopharyngeal swab specimens (86.15%; 672/780).  Table S1. The head-to-head comparison of sputum and nasopharyngeal swab specimens collected from the same COVID-19 patient is shown in Figure 2. It revealed that ORFlab and N genes were detected in all (16; 100%) sputum specimens; only ORFlab and N genes were detected in 12 (75%) and 14 (87.5%) nasopharyngeal swab specimens, respectively. The C t values of 12 sputum specimens tested for ORFlab and N genes were higher than those of the nasopharyngeal swab specimens .
TA B L E 2 Detection of targets in the nasopharyngeal swab specimens from COVID-19 patients using four commercial kits

| D ISCUSS I ON
COVID-19 outbreaks have become a global public health concern. 10 Early detection of the disease, quarantine of patients, and diagnosis have been reported to be crucial for controlling its spread, with RT-PCR being a significant method for detection and diagnosis. 16,22 The sensitivity of COVID-19 RT-PCR diagnostic kits is not only related to the types, sampling, transportation, and preservation of the viral specimens but also to the quality of the kits, which is considered the most important factor. 17 COVID-19 RT-PCR diagnostic kits with high sensitivity and specificity could help reduce the rate of false-negative detection and significantly improve the identification of COVID-19 patients. 23 This study showed that detection rates in the sputum specimens from the lower respiratory tract of COVID-19 patients using kits I to IV were higher than those in the nasopharyngeal swab specimens from the upper respiratory tract of the patients. Several studies have shown that the sampling quality of specimens obtained from the upper respiratory tract cannot be guaranteed, and specimens with low RNA concentration from the initial stage of COVID-19 cases also lead to false-negative detection. 18,19,24 Detection rates in sputum specimens from the lower respiratory tract are expected to be better than those in nasopharyngeal swab specimens. 18,19,24 However, patients with weak constitution cannot cough up sputum from the lower respiratory tract and only cough up a small amount of it from the upper respiratory tract, which ultimately leads to false-negative detection. 18,19,24 According to the manufacturers' instructions of the different RT-PCR kits for COVID-19 diagnosis available commercially, minimum detection limits are 100-1000 copies/mL, and a few kits could

F I G U R E 2
Head-to-head comparisons of severe acute respiratory syndrome coronavirus 2 detection among nasopharyngeal swab and sputum specimens from the same patients using kit I. The sputum specimens are indicated in red and nasopharyngeal swab specimens in yellow. A, ORF1ab gene detection in nasopharyngeal swab and sputum specimens; (B) N gene detection in nasopharyngeal swab and sputum specimens reach 20 copies/mL. 14 Although these minimum detection limits are sufficiently low, this study found that the sensitivities of the four assessed kits were slightly different for different targets. N gene and ORF1ab gene targets of SARS-CoV-2 could not be concurrently detected in high-dilution samples containing low RNA concentration. Thus, it is possible that positive cases could be falsely identified to be negative and thereby missed. Thus, we suggest that two or three kits should be used in the attempts to identify COVID-19 patients to improve the efficacy of the identification process.
Additionally, the results of batch effects among the four kits showed that the abilities to detect the same gene target were slightly different between different batches of the same kit, and overall batch effects were slightly different between the tested commercial kits.
Although there might be differences in the raw materials and production lines employed in producing COVID-19 diagnostic kits, manufacturers should ramp-up supervision to ensure product quality from batch to batch, which may result in only a slight detection difference within an allowable error range. Further, criteria of comparison should be formulated and comparisons should be made to confirm the sensitivity and specificity of the kits prior to using new batches.
In conclusion, the sensitivities and batch effects of the assessed kits were slightly different for different targets, and sputum specimens were more applicable for SARS-CoV-2 detection than nasopharyngeal swab specimens. Therefore, these data suggest that suspected COVID-19 cases with low RNA concentration or at the initial stages of the disease should be examined using different COVID-19 kits or sampling sputum specimens to a feasible extent and that comparison of commercial COVID-19 RT-PCR kits should be performed prior to using new batches of the kits in routine diagnostics. Additionally, with the increasing number of commercial COVID-19 kits, it is necessary for researchers to share information such as multi-center kit comparison methods and the detection abilities of various commercial RT-PCR diagnostic kits among different specimens.