Interferon gamma release assays for diagnosis of osteoarticular tuberculosis: A systematic review and meta-analysis

Abstract

Background

Although the Interferon Gamma Release Assays (IGRA) is often used to identify latent tuberculosis, it also plays a crucial role in diagnosing active extrapulmonary tuberculosis. Some studies have assessed the use of IGRA as a biomarker for osteoarticular tuberculosis (OATB), which is elevated following TB infection. Still, conclusive results about its effectiveness have not been reported.

Method

We searched PubMed, Embase, and Cochran databases. We obtained literature related to the diagnosis of OATB by IGRA, and the retrieval period was from the establishment of the database to June 2021. The bivariate random effect model was used to summarize the sensitivity, specificity, and accuracy of other indicators in diagnosing OATB by IGRA, and the forest plot and receiver operating characteristic (ROC) curve were used for testing.

Results

We included seven studies involving 643 subjects in diagnosing OATB by IGRA. The comprehensive sensitivity and specificity were 0.84 (95% CI, 0.70–0.92) and 0.78 (95% CI, 0.66–0.87), respectively. The area under the curve (AUC) was 0.87.

Conclusion

In blood samples, the diagnostic accuracy of IGRAS is poor in patients with suspected OAT. We conclude that IGRA may not be appropriate for patients with OATB.

Introduction

On a global scale, tuberculosis is an infectious disease with high morbidity and mortality, causing two to three million deaths annually and presenting approximately nine million new cases [1, 2]. Extrapulmonary tuberculosis accounted for 1 in 8 of all tuberculosis, of which the incidence of osteoarticular tuberculosis (OATB) was 11.3%-34.5% [3, 4]. The early symptoms of OATB are not obvious, and the signs and imaging features are not typical, leading to difficulties in diagnosis[5, 6]. At present, the diagnosis of OATB is based on the pathogeny or pathology of surgical (biopsy) specimens. However, due to limited medical conditions or unwillingness of patients, or inability to undergo invasive examination, the diagnosis becomes a dilemma, leading to misdiagnosis and missed diagnosis. Besides, traditional methods of culturing tuberculosis bacteria take weeks to obtain positive results and have a low detection rate [7]. The incidence of OATB with atypical clinical manifestations is increasing [8]. Early diagnosis of OATB is essential for controlling this infectious disease and further treatment. Nowadays, the number of studies on the diagnosis of OATB with Interferon Gamma Release Assays (IGRA) is climbing each year. However, due to the small sample size of individual studies, differences in study design, and source of patients, the accuracy of IGRA varies, and there is no evidence-based medicine evidence on its clinical diagnostic value.

The tuberculin skin test (TST), which uses tuberculin pure protein derivatives as diagnostic reagents, can trigger specific skin allergic reactions to organisms that have been infected with tuberculosis or have been vaccinated with BCG. TST failed to be widely deployed because, for one thing, the population inoculated with BCG-vaccinated, and the people in high-incidence countries may be tested positive, and it has poor specificity. In recent years, interferon release assay (IGRAs) has been used as an immune diagnostic tool for tuberculosis. IGRA is an immune-based blood detection method that can quantify the interferon-γ released by T cells under the stimulation of two mtb-specific antigens (culture filter protein (CFP) -10) and early secretory antigen target (ESAT) -6) [9–11]. In contrast to TST, IGRA did not cross-react with BCG strains and most nontuberculous mycobacteria (NTM) species [12]. QuantiFERON-TB Gold In-Tube test (QFT-GIT; QIAGEN, Hilden, Germany) and Enzyme-linked immunosorbent SPOT (ELISA) assay (T-SPOT.TB; Oxford Immunotec Limited, United Kingdom) are two of the most common testing methods for IGRAs in the market. QFT-GIT is a test that uses a mixture of Mycobacterium tuberculosis-specific antigen (ESAT-6 and CFP-10) to stimulate specific T-cells in peripheral blood in vitro to detect the amount of IFN-γ released. T-SPOT Tuberculosis is an enzyme-linked immunoblot assay that measures the number of cells in the peripheral blood that respond to ESAT-6 and CFP-10 antigens and produce IFN-Y [13, 14]. Although IGRAs can not distinguish active from latent tuberculosis, they use a blood or humoral samples to diagnose pulmonary and extrapulmonary tuberculosis [15, 16]. IGRAs are simple and noninvasive in a quicker diagnosis of OATB. We conducted this meta-analysis to obtain more critical evidence about the diagnostic accuracy of IGRA in diagnosing OATB.

Materials and methods

The methodology followed for the systematic review is the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [17]. We evaluated the quality of studies with the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) checklist [18].

Search strategy

We systematically searched all published studies in English in electronic databases PubMed, Embase, and Cochrane Library from 1999 through Jun 2021. The search strategy terms included as follows: (interferon-gamma OR interferon-gamma release assay OR gamma-interferon OR IFN interferon-gamma OR interferon OR IGRA OR interferon release assay OR interferon-gamma assay OR T-SPOT OR enzyme-linked immunosorbent spot OR ELISpot OR QuantiFERON OR T cell response OR T cell-based assay) AND (Tuberculosis, Bone OR Tuberculoses, Bone OR Tuberculoses, Osteoarticular OR Bone Tuberculosis OR Bone Tuberculoses OR Osteoarticular Tuberculosis OR Osteoarticular Tuberculosis OR Joint Tuberculosis OR Joint Tuberculoses OR Tuberculoses, Joint OR Tuberculosis, Joint OR Tuberculosis, Osteoarticular). We also identified additional studies from the references in these papers selected.

Study selection

We do a preliminary reading of titles and abstracts to filter out articles that might be suitable for a more detailed evaluation (Fig 1). Two reviewers (Jie Tang and Liangfeng Xia) independently conducted a second review of the full text of all the selected articles to determine suitable studies for inclusion. If there are differences, resolve them by consensus. Inclusion criteria (1) Original data, including true negative value (TN), true positive value (TP), false negative value (FN), and a false positive value (FP) of IGRA in peripheral blood were obtained directly or indirectly in the diagnosis of OATB; (2) Patients in the experimental group were diagnosed by etiology, pathology or clinical diagnosis, or after anti-tuberculosis treatment, symptoms were significantly relieved, and various auxiliary examinations tended to be expected; Patients in the control group included diseases that needed to be differentiated from bone and joint tuberculosis during a clinical or pathological examination, including nonspecific infection and tumor; (3)The number of cases of bone and common tuberculosis was more than 10, and the difference was statistically significant compared with the control group. Exclusion criteria (1) literature and data repeatedly published in each database; (2) it is not possible to extract values for diagnostic accuracy or allow sensitivity and specificity to be calculated from observations reported by numerical data or dot plots; (3) Meeting abstract, case report, and review.

[Figure omitted. See PDF.]

The flowchart shows the procedure used to select the articles for the qualitative and quantitative synthesis.

The Diagnostic reference standard for OATB is (1) positive culture or smear of acid-fast bacilli (AFB), (2) histopathological evidence of tuberculous granuloma (proliferative or caseous necrosis), and (3) radiological features consistent with ultrasound or CT scan of the osteoarthrosis. Patients are classified as probable to have tuberculosis if they respond well to anti-tuberculosis treatment and have radiographic features consistent with OATB.

Data extraction

According to the preset inclusion and exclusion criteria, two reviewers (Jie Tang and Liangfeng Xia) independently evaluated the full texts for eligibility. The following details were extracted from all included studies: first author, year of publication, country of analysis, sample size, the age range of participants, the proportion of HIV infection, IGRA method, number of positive, negative, and non-conclusive results. Indeterminate results may occur in IGRA tests, excluding these patients may change the sensitivity and specificity. They were excluded before specificity and included as false negatives when calculated sensitivity [15, 19].

Quality assessment

The quality of these studies was evaluated by diagnostic accuracy research quality assessment-2 (QUADAS-2). We evaluated the risk of bias and applicability of the study. The risk of bias is mainly divided into four parts, including case selection, diagnostic gold standard, indicator test, flow, and timing. At the same time, we conduct an applicability assessment on the first three fields. Each field is classified as high, low, and unknown, corresponding to 0, 1, and 0. An overall score greater than or equal to 4 is considered high quality. Any differences are resolved by consensus between the two reviews [20].

Statistical analysis

The sensitivity and specificity of each study were collected and calculated, and the corresponding 95% confidence interval (95% CI) was given. The results of sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), and diagnostic advantage ratio (DOR) were summarized using the bivariate random effect model. The receiver operating characteristic curve (ROC) is the coordinate graph composed of a false-positive rate on the horizontal axis and a valid positive rate on the vertical axis. The curve is drawn by the different results obtained by the receiver under specific stimulus conditions using different judgment criteria. Summarize the receiver operating characteristic (SROC): The meta-analysis of multiple different trials with the same indicator can be expressed by a ROC curve named SROC according to the weight of their odds ratio (OR). We can obtain the specificity and sensitivity of this group of studies [21]. The I2 test assessed heterogeneity and I² was more than 25%, 50%, and 75%, defined as low, moderate, and high, respectively [22]. By Spearman rank correlation, P < 0.05 was the significant threshold. Assess potential publication bias using funnel plots (Deeks method). All statistical tests were analyzed using Stata (intercooler version 14.0; Stata Corp., College Station, TX).

Results

Study characteristics

We searched 437 articles and eventually included 7 studies that met our criteria (Table 1) [1, 23–28]. These seven studies were conducted between 2009 and 2020, six of which were in China (589 patients) and one in South Korea (54 patients). The IGRA method used in all studies was the T-SPOT.TB (Oxford Immunology Technologies, Abington, UK).

[Figure omitted. See PDF.]

Quality of studies included

All studies prospectively or retrospectively included suspected tuberculosis patients. Three studies excluded patients with an uncertain diagnosis, which may exaggerate the diagnostic accuracy of the index test, so the risk of bias was higher [1, 23, 24]. In addition, a study excluded patients without IGRA results and had an increased risk of selection bias [25]. In the only study, researchers adopted the blind method to diagnose patients in the laboratory [23], while other studies did not report the use of the blind approach. In addition to two studies, all studies used manufacturer-specified thresholds to classify the results. One uses the point noted in the previous literature, which is not far from the threshold recommended by the manufacturer [26]. Another study did not mention the use of threshold in the article [27]. Only one study reported uncertainty in the test, but We excluded the uncertainty from the analysis [23]. In addition to one study, all studies provided detailed information on the reference standard for diagnosing OATB [25]. Three studies have shown that each OATB patient has clear pulmonary tuberculosis [1, 23, 24]. Some patients had similar definite diagnoses [1, 25–27].

The quality of seven studies was listed, showing that only two studies had a low bias risk (Fig 2). Four studies with patient selection bias were classified as high-risk studies without clear diagnostic criteria or excluding patients with uncertain clinical diagnoses. Only one study has a low-risk bias on the index test because it is explained without understanding other tests. Only one study showed a high risk of applicability.

[Figure omitted. See PDF.]

According to the QUADAS-2 (Quality Assessment of Diagnostic Accuracy Studies-2), the summary of the methodological quality of studies.

Diagnostic performance

The sensitivity range of all studies was 0.45 to 0.96, and the total estimation was 0.84 (95% CI, 0.70 to 0.92). The specificity range of all studies was 0.58–1.0, and the comprehensive estimation was 0.78 (95% CI, 0.66–0.78) (Fig 3). The combined estimates of PLR, NLR and DOR were 3.9 (95% CI, 2.4–6.4), 0.21 (95% CI, 0.11–0.39) and 19 (95% CI, 8–46), respectively. The corresponding ROC curve is shown in Fig 4. The position and shape of the curve indicate that the total sensitivity/specificity is a suboptimal discriminator. I² of sensitivity and specificity was 84.20% and 71.03%, respectively, with moderate and severe heterogeneity between studies.

[Figure omitted. See PDF.]

Filled squares represent sensitivity/specificity estimates from studies done using. interferon-γ release assay. See references 24 to 30 for details. Thresholds for the interpretation of I2 are as follows: 0% to 40%: might not be essential; 30% to 60%: may represent moderate heterogeneity; 50% to 90%: may represent substantial heterogeneity; 75% to 100%: considerable heterogeneity.

[Figure omitted. See PDF.]

Each study is represented by an open circle whose size is proportional to the inverse standard error of sensitivity and specificity. The filled square represents the summary estimate of the test accuracy, with the surrounding dashed zone outline denoting the 95% confidence region around this estimate.

In subgroup analysis, only the study included patients with immunosuppressive conditions with lesser heterogeneity insensitivity. However, those not excluded patients with an uncertain diagnosis, patients with or without previous tuberculosis infection history, patients with the immunosuppressive condition, and those included with patients with other diseases were associated with lesser heterogeneity in specificity (Table 2). Nevertheless, estimated variations in pooled specificity and sensitivity between subgroups were slight. The threshold effect of statistical results showed no significance, and the Spearman correlation coefficient was 0.28 (P = 0.54).

[Figure omitted. See PDF.]

A Deeks’ funnel plot was symmetric and showed a lack of publication bias (P = 0.50) (Fig 5). There was no evidence of publication bias in our evaluation.

[Figure omitted. See PDF.]

Deeks funnel plot assessment test evaluates the potential publication bias for interferon-gamma release assays on OATB. The plot shows the symmetric distribution of the log of diagnostic odds ratios against the inverse root of effective sample sizes (ESS), indicating the absence of any publication bias.

Discussion

As an in vitro immunodiagnostic test, IGRA is an excellent diagnostic tool for LTB and is increasingly being used to detect Mycobacterium tuberculosis infection [29]. This emerging technology is applied not only to latent tuberculosis but also to examine the diagnostic value of IGRA for extrapulmonary tuberculosis based on different sites of infection [30]. There has been no meta-analysis of the role of IGRA in the diagnosis of OATB. Therefore, we conducted this meta-analysis to comprehensively evaluate the overall diagnostic accuracy of IGRA for OATB. The Pooled sensitivity and specificity estimates of IGRA were 0.84 and 0.78, respectively. The pooled area under the ROC curve (AUC) was 0.87, indicating poor diagnostic accuracy. AUC is a comprehensive indicator of acute sensitivity and specificity. An AUC greater than 0.9 is generally considered as high accuracy in diagnostic tests [30]. Meta-analysis is a collection of the results of previous individual studies. Still, the heterogeneity of the combined studies is high, and no apparent cause of heterogeneity is found through subgroup analysis. More high-quality studies are required to obtain more reliable results. Several possible factors contribute to the high heterogeneity. Some non-OATB patients may have latent tuberculosis status because some studies included patients previously suffering from tuberculosis, and the false positives caused by them may reduce the specificity. Different sites of OATB may generate different amounts of interferon-gamma release; the various affected areas of included studies may cause heterogeneity. If subgroup analysis of different places is available, We may obtain more accurate results. Besides, only one study showed inconclusive results from the IGRA, and other studies were not clear; blind exclusion may result in inaccurate specificity. Most of the studies included in this meta-analysis are Chinese studies, and only one Korean study, which cannot analyze for country differences. In general, the results of this meta-analysis suggested that the diagnostic accuracy of IGRA in OATB was poor; It was also mentioned in other systematic reviews [16, 31]. This study focused on osteoarticular tuberculosis and found similar results as other tuberculosis.

There are some limitations to our meta-analysis. Because the quality of some studies was poor, the selection of patients was not fully randomized, and patient screening conditions were inconsistent. The small number of patients included in some studies also contributed to the inaccurate results. Although most studies used OATB diagnostic reference standards, the diagnostic criteria of some studies were not perfect. Because in the clinic, it is impossible to diagnose thoroughly by microbiological/pathological. The majority of patients included in this study were immunocompetent, and patients included in other studies included immunocompromised, which contributed to the heterogeneity. Sensitivity and specificity, as we know, can be influenced by the prevalence of the disease. The studies included in this study were all based in Asia. Therefore, our conclusions may only be applicable apply to the Asian region.

Conclusion

To sum up, IGRA(whole blood) exhibits poor diagnostic accuracy in OATB. Although IGRA is a noninvasive diagnostic method for OATB, its poor diagnostic accuracy does not benefit the diagnosis. We suggest that it should not be used as a standard diagnostic tool.

Supporting information

S1 Checklist.

https://doi.org/10.1371/journal.pone.0269234.s001

(DOCX)

S1 Dataset.

https://doi.org/10.1371/journal.pone.0269234.s002

(XLSX)

Citation: Ren C, Tang J, Xia L (2022) Interferon gamma release assays for diagnosis of osteoarticular tuberculosis: A systematic review and meta-analysis. PLoS ONE 17(6): e0269234. https://doi.org/10.1371/journal.pone.0269234

About the Authors:

Chunnian Ren

Roles: Conceptualization, Data curation, Formal analysis, Project administration, Software, Writing – review & editing

Affiliation: Department of Cardiothoracic Surgery, Chengdu Women’s and Children’s Central Hospital, School of Medicine, University of Electronic Science and Technology of China, Chengdu, China

Jie Tang

Roles: Conceptualization, Data curation, Funding acquisition, Investigation, Methodology, Validation, Visualization, Writing – review & editing

E-mail: [email protected] (LX); [email protected] (JT)

Affiliation: Department of Biostatistics and Epidemiology, School of Public Health, Shenyang Medical College, Shenyang, China

Liangfeng Xia

Roles: Conceptualization, Formal analysis, Investigation, Methodology, Project administration

E-mail: [email protected] (LX); [email protected] (JT)

Affiliation: Department of Cardiothoracic Surgery, Chengdu Women’s and Children’s Central Hospital, School of Medicine, University of Electronic Science and Technology of China, Chengdu, China

References

1. Tang Y, Yin L, Tang S, Zhang H, Lan J. Application of molecular, microbiological, and immunological tests for the diagnosis of bone and joint tuberculosis. J Clin Lab Anal. 2018 Feb;32(2):e22260. Epub 2017 May 22. pmid:28543876; PMCID: PMC6816924.

2. Du F, Xie L, Zhang Y, Gao F, Zhang H, Chen W, et al. Prospective Comparison of QFT-GIT and T-SPOT.TB Assays for Diagnosis of Active Tuberculosis. Sci Rep. 2018 Apr 12;8(1):5882. pmid:29651163

3. Misra SP, Misra V, Dwivedi M, Gupta SC. Colonic tuberculosis: clinical features, endoscopic appearance and management. J Gastroenterol Hepatol. 1999 Jul;14(7):723–9. pmid:10440219

4. Kulchavenya E. Extrapulmonary tuberculosis: are statistical reports accurate? Ther Adv Infect Dis. 2014 Apr;2(2):61–70. pmid:25165556

5. Lesić A, Bumbasirević M, Savić B, Cobeljić G, Pesut D. Savremena dijagnostika i lecenje kostnosglobne tuberkuloze [Current diagnosis and treatment of the osteoarticular tuberculosis]. Srp Arh Celok Lek. 2004 Sep-Oct;132(9–10):345–51. Serbian. pmid:15794059

6. Sequeira W, Co H, Block JA. Osteoarticular tuberculosis: current diagnosis and treatment. Am J Ther. 2000 Nov;7(6):393–8. pmid:11304648

7. Boehme CC, Nicol MP, Nabeta P, Michael JS, Gotuzzo E, Tahirli R, et al. Feasibility, diagnostic accuracy, and effectiveness of decentralised use of the Xpert MTB/RIF test for diagnosis of tuberculosis and multidrug resistance: a multicentre implementation study. Lancet. 2011 Apr 30;377(9776):1495–505. pmid:21507477

8. Rajasekaran S, Soundararajan DCR, Shetty AP, Kanna RM. Spinal Tuberculosis: Current Concepts. Global Spine J. 2018 Dec;8(4 Suppl):96S–108S. pmid:30574444

9. Sester M, Sotgiu G, Lange C, et al. Interferon-γ release assays for the diagnosis of active tuberculosis: a systematic review and meta-analysis [published correction appears in Eur Respir J. 2012 Mar;39(3):793]. Eur Respir J. 2011;37(1):100–111. pmid:20847080

10. Yakin M, Kesav N, Cheng SK, Caplash S, Gangaputra S, Sen HN. The Association between QuantiFERON-TB Gold Test and Clinical Manifestations of Uveitis in the United States. Am J Ophthalmol. 2021;230:181–187. pmid:33945821

11. Pourakbari B, Mamishi S, Benvari S, Mahmoudi S. Comparison of the QuantiFERON-TB Gold Plus and QuantiFERON-TB Gold In-Tube interferon-γ release assays: A systematic review and meta-analysis. Adv Med Sci. 2019;64(2):437–443. pmid:31586819

12. Pai M, Denkinger CM, Kik SV, Rangaka MX, Zwerling A, Oxlade O, et al. Gamma interferon release assays for detection of Mycobacterium tuberculosis infection. Clin Microbiol Rev. 2014 Jan;27(1):3–20. pmid:24396134

13. Hill PC, Brookes RH, Fox A, Fielding K, Jeffries DJ, Jackson-Sillah D, et al. Large-scale evaluation of enzyme-linked immunospot assay and skin test for diagnosis of Mycobacterium tuberculosis infection against a gradient of exposure in The Gambia. Clin Infect Dis. 2004 Apr 1;38(7):966–73. pmid:15034828

14. Nicol MP, Pienaar D, Wood K, Eley B, Wilkinson RJ, Henderson H, et al. Enzyme-linked immunospot assay responses to early secretory antigenic target 6, culture filtrate protein 10, and purified protein derivative among children with tuberculosis: implications for diagnosis and monitoring of therapy. Clin Infect Dis. 2005 May 1;40(9):1301–8. pmid:15825033

15. Metcalfe JZ, Everett CK, Steingart KR, Cattamanchi A, Huang L, Hopewell PC, et al. Interferon-γ release assays for active pulmonary tuberculosis diagnosis in adults in low- and middle-income countries: systematic review and meta-analysis. J Infect Dis. 2011 Nov 15;204 Suppl 4(Suppl 4):S1120–9. pmid:21996694

16. Sester M, Sotgiu G, Lange C, Giehl C, Girardi E, Migliori GB, et al. Interferon-γ release assays for the diagnosis of active tuberculosis: a systematic review and meta-analysis. Eur Respir J. 2011 Jan;37(1):100–11. pmid:20847080

17. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JP, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med. 2009 Jul 21;6(7):e1000100. pmid:19621070

18. Whiting P, Rutjes AW, Reitsma JB, Bossuyt PM, Kleijnen J. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol. 2003 Nov 10;3:25. pmid:14606960

19. Santin M, Muñoz L, Rigau D. Interferon-γ release assays for the diagnosis of tuberculosis and tuberculosis infection in HIV-infected adults: a systematic review and meta-analysis. PLoS One. 2012;7(3):e32482. pmid:22403663

20. Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, et al.; QUADAS-2 Group. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011 Oct 18;155(8):529–36. pmid:22007046

21. Arends LR, Hamza TH, van Houwelingen JC, Heijenbrok-Kal MH, Hunink MG, Stijnen T. Bivariate random effects meta-analysis of ROC curves. Med Decis Making. 2008;28(5):621–638. pmid:18591542

22. Diel R, Goletti D, Ferrara G, Bothamley G, Cirillo D, Kampmann B, et al. Interferon-γ release assays for the diagnosis of latent Mycobacterium tuberculosis infection: a systematic review and meta-analysis. Eur Respir J. 2011 Jan;37(1):88–99. pmid:21030451

23. Cho OH, Park SJ, Park KH, Chong YP, Sung H, Kim MN, et al. Diagnostic usefulness of a T-cell-based assay for osteoarticular tuberculosis. J Infect. 2010 Sep;61(3):228–34. Epub 2010 Jun 30. pmid:20600294.

24. Jia H, Pan L, Qin S, Liu F, Du F, Lan T, et al. Evaluation of interferon-γ release assay in the diagnosis of osteoarticular tuberculosis. Diagn Microbiol Infect Dis. 2013 Jul;76(3):309–13. Epub 2013 May 4. pmid:23647965.

25. Li X, Wang J, Yang Z, Song Q. Diagnostic values of peripheral blood T-cell spot of tuberculosis assay (T-SPOT.TB) and magnetic resonance imaging for osteoarticular tuberculosis: a case-control study. Aging (Albany NY). 2021 Mar 19;13(7):9693–9703. Epub 2021 Mar 19. pmid:33744856; PMCID: PMC8064225.

26. Liao CH, Chou CH, Lai CC, Huang YT, Tan CK, Hsu HL, et al. Diagnostic performance of an enzyme-linked immunospot assay for interferon-gamma in extrapulmonary tuberculosis varies between different sites of disease. J Infect. 2009 Dec;59(6):402–8. Epub 2009 Oct 9. pmid:19819258.

27. Wu X, Ma Y, Li D, Yang Y, Liang Y, Xue H, et al. Diagnostic value of ELISPOT technique for osteoarticular tuberculosis. Clin Lab. 2014;60(11):1865–70. pmid:25648028.

28. Zhou Z, Zheng Y, Wang L. A comparative study on the value of Xpert MTB/RIF and T-SPOT.TB tests in the diagnosis of bone and joint tuberculosis. Clin Chim Acta. 2020 Jan;500:115–119. Epub 2019 Oct 22. pmid:31654631.

29. Cho OH, Park KH, Kim SM, Park SJ, Moon SM, Chong YP, et al. Diagnostic performance of T-SPOT.TB for extrapulmonary tuberculosis according to the site of infection. J Infect. 2011 Nov;63(5):362–9. pmid:21781986

30. Deeks JJ. Systematic reviews in health care: Systematic reviews of evaluations of diagnostic and screening tests. BMJ. 2001 Jul 21;323(7305):157–62. pmid:11463691

31. Fan L, Chen Z, Hao XH, Hu ZY, Xiao HP. Interferon-gamma release assays for the diagnosis of extrapulmonary tuberculosis: a systematic review and meta-analysis. FEMS Immunol Med Microbiol. 2012 Aug;65(3):456–66. pmid:22487051