Clinical Biochemistry 85 (2020) 33–37

Contents lists available at ScienceDirect

Clinical Biochemistry
journal homepage: www.elsevier.com/locate/clinbiochem

Performance of conventional histopathology and GeneXpert MTB/RIF in the T

diagnosis of spinal tuberculosis from bone specimens: A prospective clinical
study
Yali Yua,1, Yiyi Konga, Jing Yeb,1, Aiguo Wangc,
⁎

a
Department of Clinical Laboratory, Zhengzhou Orthopaedics Hospital, Zhengzhou, Henan 450000, PR China
b
Department of Pathology, Zhengzhou Orthopaedics Hospital, Zhengzhou, Henan 450000, PR China
c
Department of Orthopaedics, Zhengzhou Orthopaedics Hospital, Zhengzhou, Henan 450000, PR China

ARTICLE INFO ABSTRACT

Keywords: Objective: Xpert MTB/RIF is recommended to detect pulmonary tuberculosis; however, there is insufficient data
Spinal tuberculosis on its utility for bone samples. This study aimed to assess the accuracy of Xpert MTB/RIF compared with
Xpert MTB/RIF conventional histopathology in diagnosing spinal tuberculosis (STB) based on bone specimens in high burden
Histopathology settings.
Diagnosis
Materials and methods: Totally, 128 suspected STB participants were enrolled into this study. The bone speci-
mens were obtained through puncture or operation for histological and Xpert MTB/RIF analyses, so as to
compare their accuracy in diagnosing STB by the composite reference standard (CRS).
Results: Finally, 106 subjects with suspected STB were recruited into the analysis, including 27 confirmed and 33
clinically diagnosed STB patients. Relative to histopathology, Xpert MTB/RIF achieved a 86.7% sensitivity, and
12 out of 30 STB patients were positive, while the negative results in them were obtained upon histopathology.
Based on CRS, Xpert MTB/RIF yielded a 63.3% sensitivity, which significantly elevated relative to that ob-
tained upon histopathological test (50.0%, p < 0.001). In addition, the pooled sensitivity obtained using the
above 2 approaches was as high as 95.0%, which was higher than that of any of the 2 approaches alone. The
pooled specificity was 97.8%. Moreover, the area under the curve (AUC) value was 0.75 for Xpert MTB/RIF and
0.81 for histopathology, with no statistical significance. The two methods showed moderate concordance in the
diagnosis of STB.
Conclusions: The Xpert MTB/RIF test achieves superior specificity and fair sensitivity, which can not be re-
commended to replace the conventional examinations for the diagnosis of STB. The combined application of
these 2 approaches can improve the pooled diagnostic sensitivity and accuracy for STB.

1. Introduction irreversible paralysis and spinal deformity [4,5]. Therefore, the rapid
diagnosis of spinal tuberculosis (STB) is greatly significant to achieve
Tuberculosis (TB), one of the oldest diseases in human history, is early STB diagnosis and treatment, and to control the STB epidemic
also a disease with the highest mortality caused by one individual mi- situation in countries with high TB burden, such as China.
crobiological pathogen, and there are approximately 10 million newly At present, the bacteriological diagnosis of STB mainly depends on
diagnosed patients in 2018 [1,2]. China is one of the 22 countries with the acid-fast staining and mycobacterial cultures, but smear acid-fast
high TB burdens worldwide, which ranks the second place only after staining only achieves a low positive rate, and the process is tedious and
India [1]. Osteoarticular TB accounts for 3%-5% of all TB cases and technically demanding [6]. In addition, the solid and liquid culture of
15% of the extrapulmonary TB cases, while spine stands for the most Mycobacterium TB requires a long culture cycle (usually 6–8 weeks); as
susceptible osteoarticular TB site, occupying about 50% osteoarticular a result, it cannot meet the requirements of early clinical diagnosis and
TB cases [3]. Patients who are not treated effectively in time are as- treatment.
sociated with a greatly increased risk of sequelae, such as nerve injury, GeneXpert MTB/RIF (Xpert MTB/RIF, Cepheid, Sunnyvale USA) is a

⁎
Corresponding author.
E-mail address: [email protected] (A. Wang).
1
These authors contributed equally to this work.

https://doi.org/10.1016/j.clinbiochem.2020.08.010
Received 15 April 2020; Received in revised form 10 July 2020; Accepted 19 August 2020
Available online 25 August 2020
0009-9120/ © 2020 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
Y. Yu, et al. Clinical Biochemistry 85 (2020) 33–37

fully automatic and integrated semi-nested fluorescent PCR detection resistance of rpoB gene was also detected by Xpert MTB/RIF using five
system, in which the rpoB gene is utilized as the target gene [7,8]. It is overlapping probes (A, B, C, D and E) in combination with the MTB
designed to rapidly diagnose TB with rifampin (RIF) resistance within results.
2 h, and it achieves high sensitivity and specificity for pulmonary and
various extrapulmonary TB samples, like pus, cerebrospinal fluid (CSF) 2.5. Pathological examination
and other tissues. In 2013, the Xpert MTB/RIF test has been re-
commended by WHO to be used for rapidly diagnosing extrapulmonary Firstly, tissue samples were embedded in paraffin and subsequently
TB [9,10,11,12]. This molecular biological method is simple and rapid, sliced. Later, tissue sections were subjected to hematoxylin&eosin (HE)
which greatly improves the microbiological diagnostic status for STB in staining, and the characteristic manifestations, such as caseous necrosis,
clinical practice. However, in view of the limited data on the diagnostic Langhans cells and tuberculous granuloma, were observed using the
value of Xpert MTB/RIF in bone specimens reported in published lit- microscope. Subsequently, statistical analysis was carried out, with CRS
erature, these guidelines are not suitable for bone specimens [13]. In as the reference standard.
this study, the clinical diagnostic reference standard (CRS) was adopted
as the gold standard for evaluating the clinical value of Xpert MTB/RIF 2.6. Statistical analysis
in rapidly diagnosing STB based on bone specimens.
The SPSS17.0 software was employed to analyze the specificity,
2. Materials and methods sensitivity, positive/negative predictive value (PPV/NPV), likelihood
ratio (LR), AUC, and the corresponding 95% confidence interval (CI) for
2.1. thical approval Xpert MTB/RIF assay and Histopathology, respectively. Histopathology
and CRS were selected as the gold standards. The mean value, standard
This study was approved by the Ethics Committee of Zhengzhou deviation (SD), and the median along with interquartile range (IQR)
Orthopedic Hospital. Each participant had provided the written in- were utilized to analyze the data. Additionally, the categorical variables
formed consent prior to participation into this study. were analyzed using chi-square test. A difference of p < 0.05 indicated
statistical significance.
2.2. Patient enrollment
3. Results
The study was conducted at Zhengzhou Orthopedic Hospital from
July 2018 to November 2019. Altogether 128 suspected STB cases 3.1. Patient features
undergoing CT-guided biopsies or surgical treatments were enrolled for
analysis. The patient demographic characteristics, clinical details and 128 cases with suspected STB were consecutively enrolled into the
results of each test were obtained from the electronic medical record present study. Among these patients, 2 were culture-contaminated, 5
system. were lost to follow-up, 3 were infected with non-tuberculous myco-
The diagnostic criteria of STB were as follows, (1) the symptoms and bacteria (NTM), 10 showed undefined Xpert MTB/RIF test results, while
signs were consistent with abscesses and dead bones of local TB lesions 2 had insufficient samples. Thus, 106 patients were enrolled finally,
found by imaging examinations (including X, CT and MRI); (2) the among them, 27 (25.5%) were confirmed with STB, 33 (31.1%) were
pathological examination of lesion specimens showed caseous necrosis clinically diagnosed, and 46 (43.4%) were non-STB cases based on CRS
and Langerhans granuloma; (3) the microbiological examination of (Fig. 1). Table 1 and 2 present the demographic and clinical features for
bone specimens (traditional Roche culture) confirmed Mycobacterium the 106 studied cases of different categories.
TB infection; and (4) TB symptoms and local symptoms were improved
after regular anti-TB treatment for over 6 months. STB was confirmed 3.2. Accuracy of histopathological and Xpert MTB/RIF assays in diagnosing
when any two of the above criteria were met. STB compared with CRS
The inclusion criteria were shown below, (1) all patients who had
local symptoms such as back pain, joint swelling, tenderness, deformity, When CRS was used as the reference, histopathological test
limitation of motion (LOM), and systematic symptoms like fever achieved the specificity, sensitivity, PPV, NPV, negative likelihood ratio
for > 2 weeks, loss of weight/appetite, cough, night sweats; (2) each (NLR), positive likelihood ratio (PLR), and area under the curve (AUC)
patient was followed up for at least 3 months; and (3) patients had of 100.0% (100.0–100.0%), 50.0% (40.8–56.5%), 100.0%
definite results for Xpert MTB/RIF test. (100.0–100.0%), 65.1%(56.8–73.4%), 0.75 (0.66–0.84), and max 0.5
(0.39–0.61), respectively. Among those 60 STB patients identified ac-
2.3. Smear as well as TB cultures cording to CRS, Xpert MTB/RIF assay confirmed 38 definite STB pa-
tients, yielding a 63.3% sensitivity (95% CI: 52.4–73.5%), and such
First of all, the bone samples were grounded and decontaminated figure markedly elevated compared with that achieved by histopatho-
using 4% sodium hydroxide (NaOH) (Petroff method). Afterwards, the logical test (χ2 = 41.812, P < 0.001). Meanwhile, the specificity,
sediments of treated samples were directly stained by the Ziehl–Neelsen PPV, NPV, PLR, NLR and AUC were 97.8% (93.4–100.0%), 97.4%
technique and cultured using the mycobacterial detection system for (94.3–100.0%), 67.2% (60.7–77.9%), 2.9 (2.1–3.8), 0.37 (0.26–0.52),
42 days. Thereafter, smear was examined using the 100 × oil lens. and 0.81 (0.72–0.89) (Table 3), respectively.
The pooled diagnostic accuracy of these two approaches was mea-
2.4. Xpert MTB/RIF assay sured as well, and a positive result was affirmed if any one or both of
these 2 tests were positive (n = 57). Meanwhile, negative results in
The tissue specimens collected surgically or through puncture were both of these 2 tests indicated negative results (n = 3). Moreover, the
cut into the 2–3 mm fragments. Then, 2 ml PBS was added and grinded combined application of these 2 tests achieved the specificity and
thoroughly until a homogeneous suspension was obtained. Later, 1 ml sensitivity of 96.7% and 95.0%, respectively, which were higher than
suspension was mixed with 2 ml sample treatment solution (SR), fol- those achieved by each method alone, with statistically significant
lowed by 10 s of vortexing, 10 min of standing under ambient tem- difference (χ2 = 12.065, P = 0.024 and χ2 = 34.439, P = 0.000 ,
perature, further 10 s of oscillation, and 10 min of incubation at room separately).
temperature. Finally, 2 ml digested sample was collected into the Xpert The consistency between the 2 assays was further measured by the
MTB/RIF cartridge and loaded to the instrument. Meanwhile, the RIF use of Kappa index, in which a k-value < 0.4 indicated poor

34
Y. Yu, et al. Clinical Biochemistry 85 (2020) 33–37

Fig. 1. Enrollment of suspected spinal tuberculosis patients.

agreement, 0.75 ≥ k ≥ 0.4 was indicative of fair-good agreement, and the sensitivity, specificity, PPV, NPV, PLR and NLR of 86.7%
k > 0.75 suggested superb agreement. Among those 106 participants (78.4–95.6%), 60.0% (50.1–71.8%), 68.4% (58.2–91.4%), 81.8%
enrolled, 26 had positive results and 63 had negative results in both (76.8–91.4%), 2.17 (1.36–3.43), and 0.22 (0.08–0.57), respectively.
assays, and medium concordance was detected between these 2 assays When histopathology was set as the golden standard, pathological
(k = 0.638). Upon further subgroup analysis, the agreements were and Xpert MTB/RIF tests revealed moderate consistency (k = 0.467),
0.634, 0.574 and 0.584, respectively, among the 3 groups(Table 4). and the AUC value of the diagnostic performance of Xpert MTB/RIF for
STB was 0.734 (Table 5).
In Xpert MTB/RIF test, the resistance to RIF was detected in 5 cases
3.3. Xpert MTB/RIF performance in the diagnosis of STB compared with only, and these patients were consistent with susceptibility test to
histopathological test phenotypic drug, yielding a 100.0% sensitivity. Thereafter, the patients
were treated with appropriate second-line anti-TB drugs and they re-
Of those 60 STB cases, 30 had positive pathological results, and 38 sponded well to the therapy.
had positive results in Xpert MTB/RIF assay. Upon additional subgroup
analyses, of those 30 cases with positive histopathological results for
TB, 26 (88.68%) had positive results upon Xpert MTB/RIF test, while 4 4. Discussion
(11.32%) had negative results. Although 4 histopathology-positive
participants were missed in Xpert MTB/RIF test, it found 12 additional STB stands for one of the most frequently seen extrapulmonary TB,
patients who were negative in histopathological test but were defined apart from tuberculous pleurisy and lymph node nucleus, which occu-
as STB patients upon CRS. Notably, Xpert MTB/RIF test demonstrated pies approximately 1%-3% all TB patients [1,3]. It will result in

Table 1
Classification of patients with suspected bone and joint tuberculosis.
Group Results

Culturea Signs and symptomsb Histopathologyc Imagingd Anti-TB treatmente Following-upf

Conformed STB + + +/- +/- + +

Probable STB – + + – + +
– + – + + +
– + + + + +
Non-STB – + – – – –

a
Culture grew M tuberculosis.
b
Common symptoms of STB: back pain and swelling, fever, night sweats, body weight loss, neurological abnormality such as paraplegia.
c
Pathological manifestations involve caseous necrosis, Langhans cells and tuberculous granuloma.
d
Imaging findings included soft tissue swelling, vertebral destructionor vertebral body collapse, dead bone formation and extensive paraspinal abscess.
e
Positive: Anti-TB treatment was effectivet; Negative: No-response to anti-TB treatment.
f
A follow-up duration of at least 6 months made reliable diagnoses and improved the objectivity of the conclusions.

35
Y. Yu, et al. Clinical Biochemistry 85 (2020) 33–37

Table 2
Baseline characteristics of study cases.
Characteristics All (n = 96) Confirmed STB (n = 15) Probale TB(n = 35) Non-TB(n = 46)

Age, years, mean 49.3 ± 13.9 52.3 ± 20.7 44.6 ± 16.4 56.0 ± 18.3
Male (n,%) 69 (71.9) 12 (80.0) 25 (78.1) 32 (69.6)
Lesion locations (n,%)
Thoracic 40 (41.67) 9 (60.00) 20 (57.14) 11 (23.91)
Lumbar 35 (36.46) 3 (20.00) 9 (25.71) 23 (50.00)
Lumbosacral vertebrae 14 (14.58) 2 (0.13) 4 (11.43) 8 (17.39)
Chest waist 4 (4.17) 0 (0) 1 (2.86) 3 (6.52)
Cervical vertebra 3 (3.12) 1 (6.67) 1 (2.86) 1 (2.17)
Laboratory results
WBC (1012/l, median, IQR) 6.79 (3.85–13.09) 6.74 (4.52–12.20) 7.02 (4.36–13.09) 6.74 (4.52–12.20)
ESR (mm/h, median, IQR) 27 (1–78) 32 (12–78) 35 (8–63) 29 (13–56)
CRP (mg/l, median, IQR)) 10.03 (< 0.05–69.74) 13.62 (2.15–24.07) 6.82 (4.66–47.64) 19.03 (< 0.05–56.3)
Albumin (g/l, median, IQR,) 30.2 (10.2–35.0) 36.8 (25.9–40.2) 42.2 (20.3–43.2) 35.1 (20.3–36.9)

Table 3
Performance of Xpert MTB/ RIF and Histopathology compared to the composite reference standard.
Gold standard Method Sensitivity Specificity PPV NPV PLR NLR AUC
95% CI 95% CI 95% CI 95% CI (95%CI) (95%CI) (95%CI)

Composite Histopathology 50.0% 100.0% 100.0% 65.1% max 0.5 0.75

reference (40.8–56.5%) (100.0–100.0%) (100.0–100.0%) (56.8–73.4%) – 0.39–0.61 0.66–0.84
standard GeneXpert 63.3% 97.8% 97.4% 67.2.0% 2.9 0.37 0.81
(52.4–73.5%) (93.4–100.0%) (94.3–100.0%) (60.7–77.9%) 2.1–3.8 0.26–0.52 0.72–0.89
XpertMTB/RIF + Histopathology (if Xpert 95.0% 97.8% 98.3% 93.4% 43.7 0.05 0.96
MTB/RIF or Histopathology positive) (88.2–98.7%) (93.4–100.0%) (94.4–100.0%) (88.2–98.7%) (40.5–49.3) (0.0–0.15) 0.92–1.0

Table 4 practice, can diagnose TB and the resistance to RIF simultaneously

Concordance of Gene Xpert Results Compared With Histopathology for STB. within 2 h. WHO has recommended to use it in detecting kinds of ex-
Groups Number Histopathology k value
trapulmonary samples, such as lymph node samples. Biadglegne F
showed that Xpert MTB/RIF produced a sensitivity of 93.5% and a
Positive negative specificity of 69.2%, whereas for pleural TB, the sensitivity was 22.5%
and the specificity was 98%. For the diagnosis of bone TB, it achieved
Comrimed STB 27 Xpert Positive 25 0 0.634
Negative 2 0
the sensitivity and specificity of 66.7%-86.0% and 94%-100%, respec-
Probable BJTB 33 Xpert Positive 1 12 0.574 tively, using pus samples [17,18,19,20].
Negative 2 18 This study aimed to confirm the Xpert MTB/RIF specificity and
Non-BJTB 46 Xpert Positive 0 1 0.584 sensitivity for the diagnosis of STB patients using bone samples com-
Negative 0 45
pared with CRS and histopathological test, which has not been reported
TOTAL 106 Xpert Positive 26 13 0.638
Negative 4 63 in published literature so far. The prospective study verified that Xpert
MTB/RIF achieved the specificity of 97.8% and the result was con-
k value > 0.75, excellent agreement; 0.75 ≥ k ≥ 0.4, fair to good agreement; sistent with previous reports presenting the specificity of 94%-
k < 0.4, poor agreement. 100%[21,22,23,24], relative to CRS.
Xpert MTB/RIF of our assay yielded a lower sensitivity of 63.7%
neurological sequela and other adverse consequences if early diagnosis than that of other studies (66.7%-86.0%) using pus samples. Such result
and effective treatment is not implemented. However, in high-pre- was consistent with a study in Iran using bone tissues. The inferior
valence countries and areas, the traditional diagnostic tests are still the sensitivity of Xpert MTB/RIF of our study might either be related to the
cornerstone for case discovery, which give rise to the low positive rate, low efficiency of DNA extraction from bone specimens or the lowered
ranging from 20% to 60%[14] . The duration before a diagnosis is made bacillary load due to smashing and homogenization of bone tissues.
ranges from weeks to several years; besides, the timely judicious Xpert MTB/RIF achieved a high specificity, and it performed excellently
medical intervention is always delayed or improper, causing the re- in the exclusion of STB. However, its undesirable sensitivity may hinder
quirement of DR, local recrudescence and serious consequences, like its performance in the rapid elimination of STB, especially for culture-
paraplegia. In some severe case, it can threaten the lives of patients. negative cases. If the diagnosis of STB is ambiguity or only the Xpert
Therefore, it is important to develop the convenient and highly efficient MTB/RIF result is negative, other examinations should be carried out
methods for the accurate diagnosis of STB [15,16]. before the diagnosis of STB is excluded.
The Xpert MTB/RIF test, an automated and closed system that de- STB is classified into exudative, caseous or hyperplastic necrotic TB
mands little technical expertise, less biosafety instruments and minor based on the TB pathological alterations in spine lesions. These

Table 5
Performance of Xpert MTB/ RIF compared to Histopathology.
Method Sensitivity Specificity PPV NPV PLR NLR AUC
95% CI 95% CI 95% CI 95% CI (95%CI) (95%CI) (95%CI)

Xpert MTB/RIF 86.7% 60.0% 68.4% 81.8% 2.17 0.22 0.734

(78.4–95.6%) (50.1–71.8%) (58.2–79.1%) (76.8–91.4%) 1.36–3.43 0.08–0.57 0.66–0.84

36
Y. Yu, et al. Clinical Biochemistry 85 (2020) 33–37

subtypes can coexist, resulting in the diversity of STB pathological References

morphology. Granuloma with caseous necrosis is seen under micro-
scope, meanwhile, Langerhans giant cells, fibrous connective tissue and [1] World Health Organization, Global tuberculosis report 2019, World Health
Organization, Geneva, Switherland, 2019.
chronic inflammatory cell infiltration in the periphery are also ob-
[2] Y. Qian, Q. Han, W. Liu, W.E. Yuan, C. Fan, Characteristics and management of
served. There is no doubt that STB can be diagnosed. However, there bone and joint tuberculosis in native and migrant population in Shanghai during
are more cases of granulomatous inflammation in clinical cases, which 2011 to 2015, BMC Infect. Dis. 18 (1) (2018) 543.
[3] L. Maynard-Smith, N. Larke, J.A. Peters, S.D. Lawn, Diagnostic accuracy of the
can also be detected in non-infectious and other non-tuberculosis in- Xpert MTB/RIF assay for extrapulmonary and pulmonary tuberculosis when testing
fectious diseases, leading to the limited routine pathological diagnosis. non-respiratory samples: a systematic review, BMC Infect. Dis. 31 (14) (2014) 709.
Hetty [25]reported 66 STB cases, and 47 of them had tuberculous [4] Shanmuganathan Rajasekaran, Kyphotic deformity in spinal tuberculosis and its
management, Int. Orthop. 36 (2012) 359–365.
granuloma (71%) as the characteristic pathological changes. In this [5] M. Turgut, Spinal tuberculosis (pott's disease): Its clinical presentation, surgical
study, among the 60 patients diagnosed with STB based on CRS, 30 management, and outcome. A survey study on 694 patients, Neurosurg. Rev. 24
(2001) 8–13.
(50.0%) showed typical pathological changes, while the other 30 [6] S. Iram, A. Zeenat, S. Hussain, N. Wasim Yusuf, M. Aslam, Rapid diagnosis of tu-
(50.0%) had chronic inflammation or granulomatous inflammation in berculosis using Xpert MTB/RIF assay - Report from a developing country, Pak. J.
fibrous connective tissues, which disabled the definite diagnosis of TB. Med. Sci. 31 (1) (2015) 105–110.
[7] C.C. Boehme, P. Nabeta, D. Hillemann, M.P. Nicol, S. Shenai, F. Krapp, et al., Rapid
In our study, the results of histopathological test revealed the specificity molecular detection of tuberculosis and rifampin resistance, N. Engl. J. Med. 363
and sensitivity of 100% and 50.0%, respectively, with the PPV and NPV (11) (2010) 1005–1015.
of 100% and 65.1%, separately. In other words, TB was creditably ex- [8] S. Chakravorty, M.K. Sen, J.S. Tyagi, Diagnosis of extrapulmonary tuberculosis by
smear, culture, and PCR using universal sample processing technology, J. Clin.
cluded by the negative result. Due to the STB characteristics at each Microbiol. 43 (9) (2005) 4357–4362.
stage, insufficient specimens were obtained from patients who under- [9] Automated Real-Time Nucleic Acid Amplification Technology for Rapid and
Simultaneous Detection of Tuberculosis and Rifampicin Resistance: Xpert MTB/RIF
went surgery due to intervertebral space destruction. Xpert MTB/RIF Assay for the Diagnosis of Pulmonary and Extrapulmonary TB in Adults and
can not be performed when < 1 ml specimen is available, while his- Children: Policy Update. Geneva2013.
topathological examination is not affected by this. Moreover, in this [10] G. Yu, B. Ye, D. Chen, F. Zhong, G. Chen, J. Yang, et al., Comparison between the
diagnostic validities of Xpert MTB/RIF and interferon-gamma release assays for
study, results of histopathological and Xpert MTB/RIF assays demon- tuberculous pericarditis using pericardial tissue, PLoS ONE 12 (12) (2017)
strated moderate consistency, with the k-value of 0.467. When these 2 e0188704.
[11] J. Patel, M. Upadhyay, V. Kundnani, Z. Merchant, S. Jain, N. Kire, Diagnostic
tests were used in combination, the pooled specificity and sensitivity
Efficacy, Sensitivity, and Specificity of Xpert MTB/RIF Assay for Spinal Tuberculosis
were 97.8% and 95.0%, separately, which outperformed any one of the and Rifampicin Resistance, Spine 45 (3) (2020) 163–169.
2 tests used alone (P = 0.000), yielding a surprising diagnostic value in [12] C.M. Denkinger, S.G. Schumacher, C.C. Boehme, N. Dendukuri, M. Pai,
K.R. Steingart, Xpert MTB/RIF assay for the diagnosis of extrapulmonary tubercu-
STB. losis: a systematic review and meta-analysis, Eur. Resp. J. 44 (2) (2014) 435–446.
We suggested that the STB specimens should be sent for both pa- [13] S.D. Lawn, M.P. Nicol, Xpert(R) MTB/RIF assay: development, evaluation and im-
thological examination and molecular biological tests, such as Gene plementation of a new rapid molecular diagnostic for tuberculosis and rifampicin
resistance, Future Microbiol. 6 (9) (2011) 1067–1082.
Xpert MTB/RIF test, so as to improve the STB diagnostic rate in drug- [14] Fluorescent Light-Emitting Diode (LED) Microscopy for Diagnosis of Tuberculosis:
resistant patients at the same time. For cases who were revealed with Policy Statement. Geneva2011.
[15] M.N. Massi, K.T. Biatko, I. Handayani, M.Y. Pratama, S. Septriani, G.M. Nurdin,
positive results upon Xpert MTB/RIF assay with no resistance to RIF, et al., Evaluation of rapid GeneXpert MTB/RIF method using DNA tissue specimens
suitable first-line ATD should be carried out at once. However, Culture of vertebral bones in patients with suspected spondylitis TB, J. Orthopaed. 14 (1)
is still recommended to obtain a complete susceptibility profile unless (2017) 189–191.
[16] G. Wang, W. Dong, T. Lan, J. Fan, K. Tang, Y. Li, et al., Diagnostic accuracy eva-
molecular detection for antimicrobials beyond rifampin is performed. luation of the conventional and molecular tests for Spinal Tuberculosis in a cohort,
Certain limitations should be noted in this study. For instance, the head-to-head study, Emerging Microbes Infect. 7 (1) (2018) 109.
latest molecular pathological method was not used to evaluate the di- [17] Y. Shen, G. Yu, F. Zhong, X. Kong, Diagnostic accuracy of the Xpert MTB/RIF assay
for bone and joint tuberculosis: A meta-analysis, PLoS ONE 14 (8) (2019)
agnostic accuracy for STB; for instance, the performance of Xpert Ultra. e0221427.
Studies on the effect of Xpert MTB/RIF Ultra on the diagnosis of EPTB [18] Zibo Zhou, Yan Zheng, Leiming Wang, A comparative study on the value of xpert
mtb/rif and t-spot.Tb tests in the diagnosis of bone and joint tuberculosis, Clin
show that Xpert MTB/RIF Ultra outperforms culture and Xpert MTB/ Chim. Acta; Int. J. Clin. Chem. 500 (2020) 115–119.
RIF assay in the diagnosis of EPTB, but it can yield some false-positive [19] Guomei Wei, Mu Jing, Guirong Wang, Fengmin Huo, Lingling Dong, Yunxu Li,
results [26–28]. In the years of our study period, there was no Xpert et al., The reliability analysis of xpert-positive result for smear-negative and culture-
negative specimen collected from bone and joint tuberculosis suspects, J. Thoracic
MTB/RIF Ultra at our hospital, so we were unable to analyze its ap- Dis. 8 (2016) 1205–1209.
plication value in bone specimens. Only the traditional method was [20] D. Perez-Risco, D. Rodriguez-Temporal, I. Valledor-Sanchez, F. Alcaide, Evaluation
of the Xpert MTB/RIF Ultra Assay for Direct Detection of Mycobacterium tu-
used for specimen detection, which underestimated the development
berculosis Complex in Smear-Negative Extrapulmonary Samples, J. Clin. Microbiol.
potential and value of histopathological test in diagnosing STB. 56 (9) (2018).
[21] S.D. Lawn, P. Mwaba, M. Bates, A. Piatek, H. Alexander, B.J. Marais, et al.,
Advances in tuberculosis diagnostics: the Xpert MTB/RIF assay and future prospects
5. Conclusions for a point-of-care test, Lancet. Infect. Dis. 13 (4) (2013) 349–361.
[22] C. Broderick, S. Hopkins, D.J.F. Mack, W. Aston, R. Pollock, J.A. Skinner, et al.,
To sum up, pathological examination and culture are still the gold Delays in the diagnosis and treatment of bone and joint tuberculosis in the United
Kingdom, Bone Joint J. 100-B (1) (2018) 119–124.
standards in diagnosing STB. The Xpert MTB/RIF test on bone tissues [23] M. Held, M. Laubscher, H.J. Zar, R.N. Dunn, GeneXpert polymerase chain reaction
achieves superior specificity and fair sensitivity, and it is advantageous for spinal tuberculosis: an accurate and rapid diagnostic test, Bone Joint J. 96-B
(10) (2014) 1366–1369.
in diagnosing STB at the early stage and in identifying the RIF resistant [24] J. Arockiaraj, J.S. Michael, R. Amritanand, K.S. David, V. Krishnan, The role of
mutation, as long as there are sufficient specimens. The combined ap- Xpert MTB/RIF assay in the diagnosis of tubercular spondylodiscitis, Eur. Spine J.
plication of these 2 approaches can improve the pooled diagnostic 26 (12) (2017) 3162–3169.
[25] A.P. Shetty, V.K. Viswanathan, R.M. Kanna, R. Shanmuganathan, Tubercular
sensitivity and accuracy for STB. Both pathological and Xpert MTB/RIF spondylodiscitis in elderly is a more severe disease: a report of 66 consecutive
tests are recommended in medical centers with relevant hardware patients, Eur. Spine J. 26 (12) (2017) 3178–3186.
conditions. [26] G. Wang, S. Wang, X. Yang, et al., Accuracy of Xpert MTB/RIF Ultra for the
Diagnosis of Pleural TB in a Multicenter Cohort Study, Chest 157 (2) (2020)
268–275.
Declaration of Competing Interest [27] R.R. Atherton, F.V. Cresswell, J. Ellis, et al., Detection of Mycobacterium tubercu-
losis in urine by Xpert MTB/RIF Ultra: A useful adjunctive diagnostic tool in HIV-
associated tuberculosis, Int. J. Infect. Dis. 75 (2018) 92–94.
The authors declare that they have no known competing financial [28] S. Chakravorty, A.M. Simmons, M. Rowneki, et al., The New Xpert MTB/RIF Ultra:
Improving Detection of Mycobacterium tuberculosis and Resistance to Rifampin in
interests or personal relationships that could have appeared to influ- an Assay Suitable for Point-of-Care Testing, mBio 8 (4) (2017) e00812–e00817
ence the work reported in this paper. Published 2017 Aug 29.

37