Evaluation of the MTBDRplus 2.0 assay for the detection of multidrug resistance among persons with presumptive pulmonary TB in China

We have conducted a multicenter study of the diagnostic accuracy of the MTBDRplus 2.0 assay in compared with conventional and molecular reference standard in four tuberculosis (TB)-specialized hospitals of China. A total of 5038 patients were enrolled in this study. The overall sensitivity of the assay for the diagnosis of TB was 92.7% [1723/1858, 95% confidence interval (95% CI): 91.5–93.9]. In smear-positive/culture-positive cases the sensitivity was 97.7% (995/1018, 95% CI: 96.6–98.6), whereas in smear-negative/culture-positive cases it was 86.7% (728/840, 95% CI: 84.2–88.9). The agreement rate between MTBDRplus 2.0 and Xpert MTB/RIF was 97.7% (1015/1039, 95% CI: 96.6–98.5) for smear-positive cases and 97.0% (3682/3794, 95% CI: 96.5–97.6) for smear-negative cases. As compared with phenotypic drug susceptibility testing, the MTBDRplus 2.0 correctly identified 298 of 315 patients (94.6%, 95% CI: 91.5–96.8) with rifampicin-resistance. As noted previously, isoniazid resistance is associated with many different mutations and consequently the sensitivity compared to phenotypic testing was lower (81.0%, 95% CI: 76.8–84.7). In conclusion, this assay is a rapid, accurate test in terms of increased sensitivity for detecting smear-negative TB patients, as well as an alternative for detecting both RIF and INH resistance in persons with presumptive TB, whereas the absence of a mutation in the specimens must be interpreted cautiously.

We also compared the agreement rate for detecting TB between MTBDRplus 2.0 and Xpert MTB/RIF. As shown in Table 2, the overall agreement rate between the two assays was 97.2% (4697/4833, 95% CI, 96.7-97.6), with a kappa value of 0.94. When separating participants into smear-positive and smear negative cases, the agreement rate was 97.7% (1015/1039, kappa value: 0.83) for smear-positive cases and 97.0% (3682/3794, kappa value: 0.92) for smear-negative cases. Statistical analysis revealed that there was no significant difference between smear-positive and smear-negative groups (P = 0.267).

Discussion
The low detection rate of MDR-TB is one of the major challenges influencing the gap between diagnosis and treatment in China (WHO). There are several reasons responsible for this issue: (1) the implementation of rapid molecular assays for the detection of drug resistance is unsatisfactory in resource-poor settings with a high MDR-TB burden; (2) at present, all the molecular assays approved by the Chinese Food and Drug Administration are only approved for use on smear-positive clinical samples, including the endorsed Xpert MTB/RIF. Hence, more than 50% of MDR-TB cases may be missed due to lack of accessibility to molecular laboratory examinations for smear negative patients 17 . In this study, we first conducted a multicenter study to assess the diagnostic accuracy of the new version 2 of the GenoType MTBDRplus test for the detection of multidrug resistance among persons with presumptive TB. Our results suggest that the MTBDRplus 2.0 assay provides an accurate option for the diagnosis of MDR-TB among both smear-positive and smear-negative TB cases. The MTBDRplus 2.0 assay outperformed smear microscopy because it detected a significant proportion of smear-negative TB cases. The proportion of rapidly diagnosed cases relative to smear microscopy significantly improved from 53.6% to 92.7%. According to our observations in this study, nearly one third of culture-positive patients were smear-negative. Considering that the four city hospitals in this study were top level TB specialized hospitals, most of the patients seeking care there were likely associated with more severe symptoms, thereby   accounting for more smear-positive patients. Hence, there may be a higher proportion of smear-negative culture-positive patients in the county and prefectural level dispensaries. Consistent with our hypothesis, the national survey conducted in 2010 revealed that smear-negative patients contributed half of the bacteriologically positive tuberculosis burden in China in 2010 17 . Taking into account the high prevalence of smear-negative patients, there is an urgent need to scale up use of a rapid molecular assay, such as MTBDRplus 2.0, to improve the early detection of TB in smear-negative patients in China. In addition, we also found that the sensitivity (97.7%) of MTBDRplus 2.0 for detection of MTB in smear-positive patients was significantly higher than that (86.7%) for the smear-negative group, indicating that the bacterial load correlated strongly with test performance. Similar results were observed by Theron and colleagues, showing that the sensitivity of Xpert MTB/RIF in smear-negative TB was limited by bacterial load 9 . Sample concentration may serve as an option for improving the performance of the molecular assays. Based on our previous experience, sample concentration could increase the positivity rate of MTBDRplus in smear-positive sputum samples, especially for sputum at the smear scanty level. Further studies are required to evaluate the effectiveness of sample concentration for the detection of MTB in smear-negative clinical samples. Secondly, the MTBDRplus 2.0 assay demonstrated a good performance across all four sites for the detection of resistance to RIF, which is consistent with numerouspublications 15,18,19 . In contrast, the sensitivity of the assay for detection of resistance to INH was moderate (81.0%). A recent meta-analysis showed that the sensitivity of MTBDRplus was 88.7% for the diagnosis of INH resistance for smear-positive TB patients, which was higher than in our findings 18 . However, several evaluations of new diagnostic tools in China revealed that the sensitivity for detecting resistance to INH was 80.3% for Genechip 8 and 80.2% for MTBDRplus 20 , respectively, similar to our observations. In addition, a molecular epidemiological study by Pang and other colleagues found that the combination mutations in the katG gene and the promoter of inhA gene only identify less than 75% of INH-resistant isolates in the MDR population in China, whereas 5.1% of MDR isolates only harbored point mutations in the oxyR-ahpC region 21 . Thus, this lower sensitivity for INH resistance was likely influenced by the high prevalence of oxyR-ahpC region mutations in China, which are not detected by the MTBDRplus 2.0 assay. Similarly, our results exhibited diverse sensitivities across the four sites, which may be also attributed to the heterogeneity of INH-resistant isolates circulating in different local regions in China.
There were several obvious limitations of the MTBDRplus 2.0 assay. First, despite novel modifications introduced into the DNA extraction procedure, the MTBDRplus 2.0 assay also requires extensive manual work, and a relatively long turn-around time when compared with Xpert MTB/RIF. Second, the common mutations in the oxyR-ahpC region need to be included in the MTBDRplus 2.0 assay to improve the sensitivity for INH resistance, which would then be more suitable for use in China. Third, about 15% of smear negative culture positive TB patients could not be detected by the MTBDRplus 2.0 assay, which highlights the urgent need to optimize the procedures to improve the detection sensitivity. Nevertheless, MTBDRplus 2.0 assay provides an ideal solution to simultaneously rule out both RIF and INH resistance from clinical sputum samples. With the support of the Global Fund, 125 prefectural laboratories have been equipped with MTBDR detection equipments. Fortunately, the MTBDR assay platform is also suitable for use with the MTBDRplus 2.0 assay. Hence, it is easy to integrate this novel tool with the existing platform, which will provide a promising solution at the prefectural level.
To the best of our knowledge, this study is the first multicenter evaluation of the MTBDRplus 2.0 assay in a high TB prevalence setting. Our data reveal that this assay is an accurate rapid test in terms of increased sensitivity for detecting smear-negative TB patients. In addition, it established a diagnosis with excellent rule-out value for detecting both RIF and INH resistance from persons with presumptive TB, especially for regions with a high prevalence of mono RIF resistance, as occurs in China. Considering that a small portion of drug-resistant clinical  specimens harbor no detectable mutations, especially for INH-resistant cases, the results of the molecular assay must be interpreted cautiously.

Materials and Methods
Study Population. From April 2014 through April 2015, we conducted this study at four TB specialized hospitals: Guangdong Chest Hospital, Shanghai Pulmonary Hospital, Xi'an Chest Hospital, and Anhui Chest Hospital. All patients with symptoms suggestive of pulmonary tuberculosis were consecutively enrolled in this evaluation. We collected sputum specimens of 1.5 mL each from patients. All patients gave written informed consent before they were included in this study. Ethical approval was granted by the Ethics Committee of the Chinese Center for Disease Control and Prevention. The methods were performed in accordance with the approved guidelines. Laboratory methods. Direct smears from each sputum specimen were examined using Auramine O staining for acid fast bacilli (AFB) (National guidelines for TB laboratories, China Center for Disease Control and Prevention) 22 . The specimens were further digested with N-acetyl-L-cysteine and sodium hydroxide (NALC-NaOH) for 15 minutes. PBS buffer was added up to a total volume of 45 mL, and the suspension was centrifuged for 15 min at 3,000 × g. Following centrifugation, the supernatant was discarded and the sediment was suspended in 1.5 mL of PBS buffer. A0.5 mL portion of suspension was inoculated into a BACTEC MGIT tube (BD Microbiology Systems, USA); a second 0.5 mL of suspension was used to perform the MTBDRplus 2.0 test according to the manufacturer's instructions; the final 0.5 mL of suspension was mixed with 1.5 mL of Xpert sample reagent for a 15-minute period at room temperature, then 2 mL of the inactivated material was transferred to a cartridge for GeneXpert MTB/RIF analysis (Cepheid, USA). Positive cultures from specimens were preliminarily identified as M. tuberculosis complex with the MPT64 antigen kit (Genesis, Hangzhou, China) 23 . Indirect drug-susceptibility testing with the MGIT 960 SIRT was performed to detect the susceptibility of M. tuberculosis isolates to RIF and INH as per manufacturer's instructions 24 . All the participating laboratory workers received standardized training, and were further approved by the National Reference Laboratory of Tuberculosis, China. The four clinical laboratories have participated in DST proficiency tests of the National TB Reference Laboratory since 2007. DNA Sequencing. Crude genomic DNA was extracted from the positive cultures as previously described 21 .

TB case definitions.
Briefly, colonies were scraped from the surface of Lowenstein-Jensen medium and heated for 1 h at 95 °C in 500 μl Tris-EDTA (TE) buffer. Followed by centrifugation at 13,000 rpm for 2 min, the supernatant was used as a template for amplification and was sequenced for the corresponding gene fragments (rpoB for RIF resistance; katG and inhA promoter for INH resistance) by the Sanger method with the designed internal primers 21 . The PCR mixture was prepared in a final volume of 50 μL containing 25 μL 2×PCR Mixture (Genestar, Beijing, China), 2 μL of DNA template and 0.2 μM of each primer set. The PCR was carried out under the following conditions: initial denaturation at 94 °C for 5 min, and then 35 cycles of denaturation at 94 °C for 1 min, annealing at 58 °C for 1 min, and extension at 72 °C for 1 min, followed by a final extension at 72 °C for 10 min. The PCR products were sent to Tsingke Biotech Company for DNA sequencing service (Beijing, China). The sequencing results were analyzed by alignment with the corresponding reference strain (M. tuberculosis H37Rv; GenBank accession no. AL123456).
Data analysis. The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated to assess the performance of MTBDRplus 2.0 when using liquid culture and conventional phenotypic DST as reference standards, respectively. All of the data were evaluated by the National Reference Laboratory of Tuberculosis, then enrolled in SPSS15.0 as a data base and analyzed by the software. The kappa statistic was used to gauge the strength of agreement between MTBDRplus 2.0 assay and other methods. Values of the kappa coefficient higher than 0.75 indicated excellent agreement; values between 0.4 and 0.75 fair to good agreement; values lower than 0.4 poor agreement. In addition, the chi square test was performed to compare the performance of MTBDRplus 2.0 assay between different sample groups.