A multiple-center clinical evaluation of a new real-time reverse transcriptase PCR diagnostic kit for SARS-CoV-2

Aim: The outbreak of severe acute respiratory syndrome-coronavirus 2 (SARS-CoV-2) has had serious repercussions worldwide. This study was aimed to evaluate the accuracy of a new kit for detection of SARS-CoV-2 compared with similar detection kit. Materials & methods: A total of 500 subjects were included and tested with both the new test and control kits. Clinical diagnosis results were taken as the reference standard. Results: Compared with clinical diagnosis, the sensitivity of the test kit was 82.64%, specificity was 98.45% and total coincidence rate was 90.80%. The total coincidence rate, sensitivity and specificity between control kit and clinical diagnosis were 89.20%, 78.10% and 99.61%, respectively. Conclusions: The new kit was comparable to the similar detection kit for detection of SARS-CoV-2 in sensitivity, specificity and total coincidence rate.

similar detection kit (control kit, Shanghai ZJ Bio-Tech Co. Ltd, Shanghai, China). Specimens were collected between February 14 and February 18, 2020, and were held at 4 • C if all testing could be completed within 24 h. 500 subjects (395 oropharyngeal swab specimens; 167 sputum specimens) with suspected or clinical diagnosis of SARS-CoV-2 or with history of close contact from three centers in China were enrolled in this study. Among these subjects, 62 underwent double sampling. Patients with fever, cough, chest discomfort/tightness and other clinical symptoms were also included in the study. The results of clinical diagnosis were taken as the reference standard. The clinical diagnosis of SARS-CoV-2 infection was according to the latest guideline of Diagnosis and Treatment of Pneumonitis Caused by SARS-CoV-2 (trial seventh version) published by the Chinse government [9]. Diagnosis was based on clinical history, laboratory, chest radiographic findings and nucleic acid-based assays. The remaining specimens for clinical testing were collected and tested with two RT-PCR diagnostic kits. This study was conducted in accordance with Declaration of Helsinki and Good Clinical Practice Guidelines and approved by the ethics committee institutional review board of each participating center.

Performance verification of the test kit
The performance verification of the test kit was conducted according to the Measures for the Administration of Registration of In-Vitro Diagnostic Reagents, including lowest detection limit, freeze-thaw stability, cross-reactivity and anti-interference ability. The lowest detection limit for the test kit was determined as the lowest concentration with 90-95% of the tested samples in positive. The three batches of test kits were frozen (-20 ± 5 • C) and thawed 3, 6, 9, 10 or 11 times at 25 ± 2 • C, respectively, to verify the freeze-thaw stability. To evaluate potential crossreactivity, 56 pathogens with the same infection site or similar symptoms as SARS-CoV-2 were detected. These organisms included human coronavirus (229E, OC43, HKU1 and NL63), severe acute respiratory syndromerelated coronavirus (SARSr-CoV), seasonal influenza A (H1N1, H3N2, H5N1 and H7N9) viruses and adenovirus (type 1, 2, 3, 4, 5, 7, and 55), among others. Anti-interference ability was measured in the presence of various interfering substances. For this purpose, two independent oropharyngeal swabs and sputum specimens were spiked with potential interfering substances that could be collected with the sample. Interfering substances included 2.5%, 5.0% and 7.5% blood (v/v), mucoprotein (0.45, 0.9 or 1.35 mg/ml), beclomethasone (2.5 mg/ml), dexamethasone (5 μg/ml), phenylephrine (50%, v/v) and oxymetazoline (50%, v/v), for example.
Real-time PCR assay for screening of SARS-CoV-2 RNA was extracted from different specimens according to the manufacturer's instructions. Both RT-PCR kits targeted the ORF1ab and nucleoprotein gene regions. If two targets tested positive, the subject was considered to be laboratory confirmed. For the test kit developed by Beijing Applied Biological Technologies Co., a cycle threshold value (Ct-value) of 38 or less was considered a positive test, whereas a Ct-value greater than 40 was defined as a negative test. Specimens with a Ct-value of 38-40 need to be retested. If the repeated Ct-value was less than 40 and an obvious peak was observed, the retest was considered as positive. For the control kit developed by Shanghai ZJ Bio-Tech Co., a Ct-value less than 37 was defined as a positive test, whereas a Ct-value of 40 or more was considered as a negative test. Specimens with a Ct-value of 37-40 need to be retested. If the repeated Ct-value was less than 40 and an obvious peak was observed, the retest was considered positive. When both sputum and oropharyngeal swab specimens were collected from the same subject, sputum specimen was included in the final statistic.

Statistical analysis
All statistical analyses were performed by using SPSS version 22.0 (SPSS Institute, Chicago, IL, USA). Quantitative data were expressed as means ± SD. Qualitative data were expressed as number and percentage. Sensitivity (positive coincidence rate), specificity (negative coincidence rate), total coincidence rate, positive predictive value (PPV), negative predictive value (NPV) and Kappa index were used for degree of agreement of the two kits. Statistical significance was set at p < 0.05.

Results
Performance verification of the test kit As shown in Table 1, the positive rates of the three batches of kits were all 100% when the sample concentration were 1.0 × 10 3 , 5.0 × 10 2 and 2.0 × 10 2 , respectively. Therefore, the lowest detection limit was 2.0 × 10 2 copies/ml. The results obtained after 11 freeze-thaw cycles of the test kit showed no significant difference between the results obtained after 0 freeze-thaw cycles ( Table 2). In addition, there was no cross-reaction with other pathogens, and  Negative (N1-N10) 10/10 10/10 10/10 10/10 10/10 10/10

L1-10
ORF1ab no interference was observed with the interfering substances.

Baseline characteristics
A total of 500 subjects (258 males, 242 females) were enrolled in this study from February 14 to February 18, 2020. The baseline characteristics of included subjects are shown in Table 3. A total of 562 specimens were collected, including 395 oropharyngeal swabs specimens and 167 sputum specimens. A total of 62 subjects underwent double sampling. The age range of subjects was 0.75 (9 months)-93 years old. The majority of the subjects were aged 20-80 years, of which 40.8% subjects aged 20-40 years, 30.6% subjects aged 40-60 years, and 16.8% subjects aged 60-80 years. Among the 500 subjects, 94 had symptoms. The main symptoms of included subjects were fever (9.0%), pneumonia (2.6%), fatigue (2.0%), pulmonary infection (1.6%), cough (1.4%) and chest discomfort/tightness (1.4%).  Table 4, test kit results of 542 specimens were consistent with control kit (215 were positive, 327 were negative) and 20 specimens were inconsistent. The total coincidence rate between two kits was 96.44% (95% CI: 94.56 to 97.81), and the Kappa index was 0.9259 (p < 0.05), indicating the accuracy of test kit for detection of SARS-CoV-2 was comparable to control kit. Among the 20 inconsistent specimens, 18 were positive and 2 were negative by test kit. In addition, the experimental kit results of 15 specimens were consistent with clinical diagnosis (14 were positive, 1 was negative).

Subgroup analysis by age
The results of subgroup analysis by age are shown in Table 5