Abstract
-
Background/Aims
- Diagnosis of intestinal tuberculosis (ITB) is challenging. Histopathology and microbiological examination remain the gold standard, but previous studies show varied diagnostic performance. We aimed to systematically evaluate the accuracy of tests to diagnose ITB in both conventional and novel methods.
-
Methods
- We searched MEDLINE and EMBASE from inception to October 2023. All studies enrolling at least 10 patients with reported information regarding the diagnosis of ITB based on endoscopic biopsy specimens, stool tests, and blood tests were included. We performed a meta-analysis using a random-effects model to estimate the performance of each test.
-
Results
- Of 3,308 abstracts reviewed, 55 studies with 6,072 participants met the inclusion criteria. Endoscopic tissue biopsy for acid-fast bacilli, the presence of caseous granuloma on histopathology, polymerase chain reaction (PCR) for tuberculosis, mycobacterial culture, and Xpert MTB/RIF showed pooled sensitivity of 12% (95% confidence interval [CI], 8%–17%), 18% (95% CI, 12%–27%), 58% (95% CI, 44%–72%), 23% (95% CI, 12%–40%) and 29% (95% CI, 17%–46%), respectively. The liquid medium culture showed higher sensitivity than conventional Lowenstein-Jensen medium (25% [95% CI, 13%–43%] and 6% [95% CI, 3%–13%]). Pooled sensitivity and specificity of stool PCR for TB were 73% (95% CI, 43%–90%) and 95% (95% CI, 79%–99%), respectively. Additionally, the pooled sensitivity and specificity of interferon-gamma release assay (IGRA) were 86% (95% CI, 79%–91%) and 86% (95% CI, 81%–89%).
-
Conclusions
- Endoscopic tissue biopsy samples had limited sensitivity in diagnosing ITB. IGRA showed good accuracy and may be combined with other methods to improve the diagnostic yield. Stool PCR demonstrated a good performance but based on a few studies.
-
Keywords: Gastrointestinal tuberculosis; Diagnosis; Predictive value of tests; Meta-analysis
INTRODUCTION
Tuberculosis (TB) remains one of the major healthcare problems, especially in developing countries that are considered endemic [1]. About 10.6 million people suffered from TB in 2022, with an increasing incidence rate of 133 cases per 100,000 population and 1.3 million deaths globally, second only to COVID-19 as a cause of death from infectious diseases [2]. Extrapulmonary TB accounts for 20%–36% of cases, and 10%–30% of patients have gastrointestinal involvement, with a higher prevalence in immunocompromised individuals [3-5].
Diagnosis of intestinal TB (ITB) is challenging. It tends to be delayed, as a recent report showed a median time to diagnosis of 70 days [6], potentially leading to complications such as intestinal stricture, bleeding, or even perforation [7]. This is due to its non-specific symptoms and radiographic and endoscopic findings. ITB must be differentiated from any other causes of chronic enterocolitis, especially Crohn’s disease (CD), where misdiagnosis occurs frequently [8]. Some endoscopic features can suggest the diagnosis of ITB, such as the presence of transverse ulcers or a patulous ileocecal valve, but none are pathognomonic [9-12]. Therefore, endoscopy with biopsy remains essential. Supporting histopathologic findings for ITB diagnosis include the presence of caseous granuloma; however, the sensitivity is reported to be only 13%–33% [13-15]. Therefore, increasing diagnostic yield with microbiological examination plays an essential role in the diagnosis. The well-known conventional methods include acid-fast bacilli (AFB) staining, culture, and polymerase chain reaction (PCR) for TB. Previous studies show varied diagnostic accuracy with the sensitivity range from 17%–45% for AFB staining, 6%–48% for Lowenstein-Jensen (LJ) solid media culture, 50%–76% for MGIT BACTEC liquid media culture, and 60%–70% for PCR. Combining these methods with histopathology may heighten the diagnostic accuracy by 17%, with a reported sensitivity of up to 80% [11,14,16]. Furthermore, there are novel tests used to diagnose ITB currently, such as multiplex PCR targeting more primer or Xpert MTB/RIF which can provide the information regarding drug resistance simultaneously, and interferon-gamma release assay (IGRA) which can identify individuals with both active and latent TB infection. Newer culture media or antigen testing methods are also currently being studied [11,16]. In addition to endoscopic biopsy specimens, stool has been increasingly recognized as a specimen for TB diagnosis, especially in pediatric populations [17]. Recent study also suggests its potential role in the diagnosis of ITB with promising diagnostic performance [18].
Although many studies have investigated ITB diagnosis methods, the reported results are varied, and no standard systematic review with meta-analysis has been available. Therefore, we aimed to systematically review and perform meta-analyses evaluating the diagnostic performance of each method in diagnosing ITB.
METHODS
This systematic review and meta-analysis were conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines [19]. The study was registered in the PROSPERO database with the registration number CRD42024583515.
1. Search Strategy
We systematically searched the Ovid MEDLINE and EMBASE databases for relevant articles regarding the diagnosis of ITB from inception until October 2023. We identified the studies with the terms (“gastrointestinal tuberculosis” OR “intestinal tuberculosis” OR “tuberculous colitis” OR “tuberculous ileitis” OR “colonic tuberculosis” OR “ileal tuberculosis”) AND (“diagnosis” OR “sensitivity and specificity” OR “predictive value of test”) The search terms of both databases were summarized in the Appendix 1. Additional articles were identified by searching the references cited in the selected study. Small case series, case reports, review articles, commentaries, and duplications were excluded. The included studies were restricted to only those in English.
2. Study Selection
Two reviewers (P.W. and T.G.) independently screened the titles and abstracts from the search results to determine the eligibility of the study. The inclusion criteria were as follows: (1) adult patients aged at least 18 years; (2) case-control, cohort, cross-sectional study, or case series with more than 10 patients; or (3) diagnosis of ITB based on AFB staining, PCR for TB, mycobacterial culture, caseous necrosis or caseous granuloma from histopathology and compatible clinical and endoscopic findings with response to anti-TB treatment. Articles with esophageal, gastric, perianal, and peritoneal TB were excluded. Studies utilizing surgical specimens were also excluded. Discrepancies were resolved through consensus or discussion with the third reviewer (J.L.).
3. Data Extraction
The data were independently extracted from each study and transferred to a customized data extraction sheet by 2 reviewers (P.W. and T.G.). The author’s name, site of study, year, and duration of the study, study design, mean age and gender, criteria and tests used for ITB diagnosis, number of participants, and diagnostic performance of the tests, including sensitivity and specificity, were recorded.
4. Quality Assessment
Quality assessment was separately carried out on each study by 2 reviewers (P.W. and T.G.). We used the Quality Assessment of Diagnostic Accuracy Studies Version 2 (QUADAS-2) checklist, which categorizes quality into “risk of bias” and “concerns regarding applicability.” The checklist identifies key 4 domains: patient selection, index test, reference standard, and flow and timing [20]. The quality assessment was graded into low, high, or unclear risk of bias and concerns regarding applicability. Discrepancies were resolved by consensus or discussion with a third senior reviewer (J.L.).
5. Statistical Analysis
A meta-analysis was performed on the diagnostic performance of all tests used to diagnose ITB, employing a bivariate random-effects model. The data were presented as pooled sensitivity, specificity, positive predictive value, negative predictive value, positive and negative likelihood ratios (PLR and NLR), diagnostic odds ratio (DOR), and the area under the curve (AUC) of summary receiver operating characteristic (SROC), along with the 95% confidence intervals (CIs). Heterogeneity across the study was assessed using the I2 method. Subgroup analysis was performed according to the culture medium, PCR technique, and IGRA type. Sensitivity analysis was conducted according to the type and quality of the study and included populations. Publication bias was evaluated by Deek’s asymmetrical funnel plot [21]. The diagnostic performance of all tests was analyzed using STATA version 17.0 (StataCorp, College Station, TX, USA).
RESULTS
The PRISMA diagram for the study selection is shown in Fig. 1. After removing duplicates, 3,308 abstracts were reviewed, and 280 full-text articles were further assessed for eligibility. After full-text screening, 223 articles were further excluded for reasons listed in Fig. 1. Finally, we included 55 studies in this review, comprising 6,072 participants and 2,392 patients with ITB. The information is detailed in Supplementary Table 1. All of the included studies utilized composite reference standards that comprised histological and microbiological parameters, the presence of extraintestinal TB, and response to anti-TB treatment in diagnosing ITB.
1. Quality Assessment of Included Studies
The results of the QUADAS-2 assessment of study quality are summarized in Fig. 2. The results of each study are demonstrated in Supplementary Table 2. Most of the studies were classified as having a low or unclear risk of bias. The high risk of bias in the patient selection domain came from the nonconsecutive patient selection method. Some studies with vague or no standard diagnostic criteria for ITB were considered to have a high risk of bias regarding reference standards.
2. Results of Diagnostic Test Accuracy
1) Endoscopic Tissue Biopsy Samples
Table 1 presents the overall diagnostic performance of all tests, along with the number of studies and participants included in each test. The heterogeneity of the pooled analysis is shown separately. Endoscopic tissue biopsy samples for AFB, the presence of caseous granuloma or caseous necrosis on histopathology, mycobacterial culture, and Xpert MTB/RIF showed a pooled sensitivity of 12% (95% CI, 8%–17%; I2=66.4%), 18% (95% CI, 12%–27%; I2=77.8%), 23% (95% CI, 12%–40%; I2=84.9%) and 29% (95% CI, 17%–46%; I2=70.6%), respectively with considerable heterogeneity in all outcomes. The pooled specificity of these tests was 100% with low heterogeneity (I2=0.0%–0.6%). Of the 18 studies evaluating the diagnostic performance of tissue mycobacterial culture, subgroup analysis was performed according to the type of culture medium. Four studies with 308 participants used the conventional LJ medium, and 8 studies with 406 participants evaluated the liquid medium culture method, which showed higher sensitivity than the LJ medium (25% [95% CI, 13%–43%; I2=82.9%] and 6% [95% CI, 3%–13%; I2=25.2%]). The heterogeneity was substantially low when limited to studies utilizing LJ culture medium.
The diagnostic accuracy of endoscopic tissue biopsy samples for PCR was assessed in 17 studies. Among these, 6 studies evaluated the use of the multiplex PCR technique, while 6 studies employed conventional PCR targeting the IS6110 sequence, and 1 study assessed the diagnostic value of fluorescent quantitative PCR. The pooled sensitivity, specificity, PLR, NLR, and DOR were 58% (95% CI, 44%–72%; I2=86.8%), 98% (95% CI, 94%–99%; I2=20.7%), 23.5 (95% CI, 9.2–59.6), 0.43 (95% CI, 0.30–0.60), and 55 (95% CI, 18–169) respectively (Fig. 3). Subgroup analysis revealed a higher pooled sensitivity of 79% (95% CI, 66%–87%; I2=73.3%) in studies utilizing multiplex PCR technique compared to studies using single-target PCR for the IS6110 sequence, which had sensitivity of 44% (95% CI, 28%–61%; I2=82.0%). The heterogeneity remained moderate to high across the analyses.
Effect of the number of tissue biopsy samples on diagnostic yield
Systematic search yielded 2 studies regarding the number of biopsy samples in diagnosing ITB. Pulimood et al. [22] retrospectively evaluated the segmental colonoscopic biopsies from different parts of the colon in diagnosing ITB and CD. All ITB patients with 3 or more biopsy samples showed histopathological involvement with distinguishing features. In contrast, 2 of the patients who underwent 1 or 2 colonoscopic biopsies had non-diagnostic results. The number of positive biopsies increased with the number of biopsies taken. The ileocecal segment showed the highest diagnostic yield; 21 out of 29 patients with cecal involvement had distinguishing features present in histopathology. Later, Mehta et al. [23] prospectively collected 8 colonic biopsies from 70 patients subsequently diagnosed as ITB. The biopsy specimens were divided into 2 containers, each containing 4 specimens. The AFB culture positivity rate increased by 11.4%–14.3% when comparing 8 with 4 colonoscopic biopsies, with a total culture positive rate of 52.8% using BACTEC MGIT method.
2) Stool Samples
Four studies comprising 403 participants assessed the diagnostic value of stool PCR for TB; 3 studies used conventional PCR targeting the IS6110 sequence, and 1 used the fluorescent quantitative PCR technique. The meta-analysis demonstrated pooled sensitivity and specificity of 73% (95% CI, 43%–90%; I2=75.9%) and 95% (95% CI, 79%–99%; I2=24.9%), respectively, with considerable heterogeneity (Fig. 4).
3) Interferon-Gamma Release Assay
Twenty-three studies comprising 3,001 participants evaluated the diagnostic performance of IGRA in diagnosing ITB. Most of the studies used CD as the control group. The meta-analysis showed pooled sensitivity, specificity, PLR, NLR, and DOR of 86% (95% CI, 79%–91%; I2=74.3%), 86% (95% CI, 81%–89%; I2=72.4%), 6.0 (95% CI, 4.4–8.1), 0.16 (95% CI, 0.11–0.25), and 37 (95% CI, 20–68) respectively. The SROC-AUC was 0.92 (95% CI, 0.89–0.94; I2= 97.0%) (Fig. 5). Subgroup analysis demonstrated slightly higher sensitivity and lower specificity of T-SPOT TB compared to the QuantiFERON-TB GOLD test. The pooled sensitivity and specificity of the T-SPOT TB test were 90% (95% CI, 84%–94%; I2=57.3%), and 83% (95% CI, 74%–90%; I2=85.5%) with the SROC-AUC of 0.93 (95% CI, 0.91–0.95; I2=81.0%) while the QuantiFERON-TB GOLD test had the pooled sensitivity and specificity of 71% (95% CI, 53%–84%; I2=79.3%), and 88% (95% CI, 82%–92%; I2=28.4%) with the SROC-AUC of 0.89 (95% CI, 0.86–0.92; I2=74.0%). The heterogeneity was reduced but remained moderate to high after subgroup analysis.
4) Combination of Diagnostic Methods
Many studies evaluated the incremental diagnostic performance of combined microbiologic and histologic tests, as summarized in Table 2 [23-32]. Most of the studies revealed increased diagnostic yield of up to 25% when combining mycobacterial culture and histopathology. The effect is more prominent when using BACTEC as culture media, while studies using LJ media showed little to no benefit. In addition, 3 studies evaluated the combination other than culture and histopathology. Sekine et al. [31] retrospectively reviewed the data of 182 patients with suspected ITB who underwent colonoscopy with biopsy for histopathological and microbiological testing, including culture and PCR from tissue biopsies, stool, and intestinal fluid. Eventually, 50 patients were diagnosed with TB. The combination of granuloma from histopathology and culture yielded the highest sensitivity of 77%, significantly higher than the combination of histopathology and intestinal fluid culture, stool culture, tissue PCR for TB, intestinal fluid PCR for TB and histopathology alone, which had the sensitivity of 67%, 71%, 61%, 65%, and 51% respectively. Paulose et al. [33] also retrospectively collected the data of patients with clinical suspicion of ITB in a referral center in India and found that a combination of histopathology and Xpert MTB/RIF produced a sensitivity of 97.1% and increased to 100% when further combined with culture. However, in this study, histopathology and mycobacterial culture independently yielded very high sensitivities of 94.3% and 91.4%, respectively. A Previous study from our center demonstrated that the sensitivity of tissue biopsy was 40.7% for AFB staining, 13.9% for histopathology of caseous necrosis, 25.7% for PCR for TB, and 53.4% for culture for TB using either LJ or BACTEC media. Notably, combining all modalities increased diagnostic sensitivity to 68% [18].
3. Sensitivity Analysis
To address substantial heterogeneity, especially the pooled sensitivity outcomes, a sensitivity analysis was conducted based on study design and quality assessment. The diagnostic performance results, restricted to prospective studies, those with a low risk of bias, and studies designed to differentiate ITB from CD, are detailed in Supplementary Tables 3, 4, and 5, respectively. Overall, the diagnostic performances of tests utilizing endoscopic tissue biopsy samples were comparable to those in the primary analysis, with notably reduced heterogeneity, except for the PCR tests. The pooled analysis of 3 studies, which evaluated stool PCR tests using a prospective design, demonstrated a higher pooled sensitivity of 83% (95% CI, 72%–91%; I2=6.2%), with comparable specificity of 93% (95% CI, 83%–98%; I2=25.2%). There was no significant heterogeneity. However, given the high risk of bias associated with stool samples as an index test, further analysis based on the quality of studies was not performed. Diagnostic performance of IGRA tests remained consistent with the primary analysis, with substantially decreased heterogeneity, especially when restricted to prospective studies.
To further address the high heterogeneity of the PCR tests in endoscopic tissue biopsy samples, the analysis, which was restricted to studies using fresh specimens rather than paraffinbased specimens, was conducted. The results revealed reduced heterogeneity, with a pooled sensitivity of 78% (95% CI, 69%–85%; I2=56.0%) and a pooled specificity of 97% (95% CI, 94%–99%; I2=14.9%). Lower level of heterogeneity observed when limited to studies using multiplex PCR (I2=2.7% for sensitivity and I2=4.0% for specificity) (Supplementary Fig. 1).
4. Publication Bias
Using Deek’s asymmetrical funnel plot, only the endoscopic tissue biopsy of AFB showed publication bias (P=0.08). There was no publication bias for all other tests (Fig. 6).
DISCUSSION
This meta-analysis of 55 studies represents the first comprehensive assessment of all diagnostic tests for ITB. However, interpretation of some results requires caution due to substantial heterogeneity observed among the included studies. Our analysis reveals that conventional microbiological examinations, including AFB staining, culture, and the presence of caseous necrosis or caseous granuloma on histopathology, have limited sensitivity in diagnosing ITB. These findings comply with the paucibacillary nature of Mycobacterium tuberculosis (MTB). The highest sensitivity among these methods is achieved through culture using a liquid medium, with a sensitivity of 25%. The diagnostic performance of some tests in this meta-analysis shows lower sensitivity than in previous reviews [11,14,16]. The AFB staining shows a sensitivity of 12% compared with 17%–45% in previous reviews, and the presence of caseous necrosis or caseous granuloma on histology shows a sensitivity of 18% compared with previously reported of 30%–50%. This could be explained by the inclusion of only endoscopic biopsy specimens in this study, rather than surgical specimens, which may have a higher diagnostic yield [34]. The exclusion of surgical specimens means that the reported diagnostic performance in this study represents current clinical practice. This approach enhances the applicability and generalizability of our study findings since the diagnosis of ITB depends largely on endoscopy with biopsy rather than surgical resection.
PCR-based methods are beneficial for more rapid diagnosis compared with culture and histopathology. In this study, PCR for TB demonstrates substantial diagnostic performance with a pooled DOR of 55. However, the pooled sensitivity is low compared to previous reports in the literature (58% vs. 75.7%–93.1%) [16]. This is possibly owing to the inclusion of both conventional and multiplex PCR in this study, as the pooled sensitivity increased to 79% when restricted to the studies using the multiplex PCR technique. The result aligns with a previous meta-analysis of intestinal tissue PCR targeting only the IS6110 sequence, which revealed a pooled sensitivity of 47% [35]. In contrast, a recent study using Truenat MTB Plus, a chip-based real-time multiplex PCR, demonstrated a sensitivity of 70% in diagnosing ITB [36]. Furthermore, Xpert MTB/RIF has a pooled sensitivity of 29%, which is lower than the sensitivity reported in the previous meta-analysis by Ding et al. [37], which showed a pooled sensitivity of 48%. However, the pooled analysis of Ding’s study included studies evaluating stool and surgical specimens, which demonstrated high sensitivities of 60% and 71%, respectively. This explains the discrepancy compared with our findings. Furthermore, substantial heterogeneity may limit the applicability of this meta-analysis. The other meta-analysis by Sharma et al. [38] evaluating Xpert MTB/RIF in the diagnosis of abdominal TB shows a sensitivity of 64%, which decreases to 23% when limited to the ITB subgroup, comparable to our meta-analysis.
Stool PCR for TB demonstrates remarkable diagnostic performance with a pooled sensitivity of 73% and a pooled specificity of 95%. The results are consistent between studies, which all show high diagnostic accuracy. There was no significant heterogeneity after limiting the analysis to 3 studies with a prospective design, emphasizing the reliability of the results. However, these results need to be interpreted cautiously because stool PCR for TB can yield false-positive results from concomitant pulmonary TB. A recent study demonstrated a strong agreement of 85.6% between stool quantitative PCR for TB and sputum culture for TB in patients with pulmonary TB [39]. Nonetheless, one of the studies included in the meta-analysis, which excluded the presence of pulmonary TB with positive AFB staining from sputum, still shows a sensitivity of 89% and specificity of 100% [40]. Based on these results, stool PCR could be a promising diagnostic tool to facilitate ITB diagnosis in the context of clinical suspicion, as it is noninvasive and provides rapid results, leading to early diagnosis.
Focusing on immunological tests, IGRA demonstrated excellent diagnostic performance, with a pooled DOR of 37 and an SROC-AUC of 0.92. T-SPOT and QFT tests demonstrate outstanding diagnostic performance, with T-SPOT exhibiting slightly higher sensitivity and lower specificity compared to QFT. Some cautions remain when interpreting this test, as it cannot clearly distinguish between active and latent TB infection. The most recent study in India included in the meta-analysis showed only 40.7% and 75.5% sensitivity and specificity with an AUC of 0.66 [41]. This study highlights the high false-positive rate of IGRA, especially in the endemic area for TB. However, with a high negative predictive value, IGRA could rule out ITB when CD is a compelling diagnosis, and it has been integrated into many models differentiating CD from intestinal TB [12,42-45].
The result of this meta-analysis highlighted the unmet needs in the diagnosis of ITB, specifically the low diagnostic yield and time-consuming nature, particularly when using histopathology and culture, which can lead to delayed diagnosis. Several approaches have been proposed to overcome these limitations. A combination of tests, particularly culture and histopathology, has consistently proven to increase the diagnostic yield across studies. Due to the feasibility and availability of tests, this approach should be adopted. Increasing the number of endoscopic biopsies also tends to enhance sensitivity in both culture and histopathology, although further studies are needed to determine the optimal number of biopsies required to achieve the best results.
Our meta-analysis demonstrated considerable heterogeneity, particularly in the pooled sensitivity estimates. Subgroup analysis stratified by culture media types and IGRA methods notably reduced heterogeneity. Further sensitivity analyses by study design, study quality and included participants also resulted in a substantial decrease in heterogeneity, suggesting a contribution. Nonetheless, substantial heterogeneity remained. The possible explanations included variation in study populations and inconsistencies in diagnostic testing protocols across studies. Other uncontrollable factors including variations in the number of biopsy samples, the instruments used for tissue acquisition and subsequent tissue processing method may also contribute to the residual heterogeneity. This is supported by the marked reduction in heterogeneity when the PCR analysis was limited to studies using fresh specimens, as tissue fixation and paraffin embedding are known to damage DNA and potentially compromise PCR result [46]. While heterogeneity limits definitive conclusions, the comparable diagnostic performance with the primary analysis underscores the robustness and reliability of our findings.
To our knowledge, this is the first meta-analysis to evaluate all tests used for diagnosing ITB. However, there are some limitations. First, the diagnosis for ITB in this meta-analysis is based on multiple clinical criteria and tests, which would normally affect the accuracy of the results. However, this composite reference standard is considered standard in the diagnosis of ITB and complies with current clinical practice. The adoption of these criteria in this meta-analysis may enhance the generalizability of the results. Second, there is high heterogeneity in the analysis of some tests. Subgroup and sensitivity analyses partially reduced the heterogeneity, but a substantial degree remained in some tests. Third, only studies in English were included, although only a few studies in other languages were found during our systematic search. Fourth, we were unable to retrieve the full text of some studies, but most of these were conducted before 2000, which should have had a minimal impact on our study results, given the differences in the techniques used in the tests. Fifth, this meta-analysis only included studies with a control group, potentially excluding some studies. Nonetheless, publication bias was assessed. Only endoscopic tissue biopsy for AFB staining showed significant publication bias.
In conclusion, conventional tests such as AFB staining, culture, and histopathology from endoscopic tissue biopsy samples yield limited sensitivity in diagnosing ITB. In contrast, newer tests, such as PCR for TB, showed higher diagnostic sensitivity. Stool PCR may provide a rapid, noninvasive diagnosis, especially where limitations exist for endoscopy. IGRA remains helpful, especially in ruling out intestinal TB due to its high negative predictive value, while positive results must be interpreted cautiously due to the high rate of false positives in TB-endemic areas. The combined diagnostic performance of histopathology and microbiologic tests, along with an adequate number of tissue biopsies, should be further explored to improve the diagnostic yield of intestinal TB.
NOTES
-
Funding Source
The authors received no financial support for the research, authorship, and/or publication of this article.
-
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
-
Data Availability Statement
Data, analytic methods, and study materials are available to other researchers upon request.
-
Author Contributions
Conceptualization: Katchamart W, Limsrivilai J. Data curation: Weeranawin P, Geeratragool T. Formal analysis: Geeratragool T. Investigation: Weeranawin P, Geeratragool T, Limsrivilai J. Methodology: Weeranawin P, Katchamart W, Limsrivilai J. Supervision: Katchamart W, Limsrivilai J. Writing–original draft: Weeranawin P, Limsrivilai J. Writing–review & editing: all authors. Approval of final manuscript: all authors.
Supplementary Material
Supplementary materials are available at the Intestinal Research website (https://www.irjournal.org).
Supplementary Fig. 1.
Forest plots for endoscopic tissue biopsy samples for polymerase chain reaction (PCR) for tuberculosis limited to studies with fresh specimens of all studies (A), limited to multiplex PCR (B). CI, confidence interval.
ir-2025-00098-Supplementary-Fig-1.pdf
Fig. 1.Study disposition.
Fig. 2.Summary of the risk of bias assessment according to Quality Assessment of Diagnostic Accuracy Studies Version 2 (QUADAS-2).
Fig. 3.Forest plot and summary receiver operating characteristic (SROC) curve demonstrate the pooled sensitivity and specificity of endoscopic tissue biopsy samples for polymerase chain reaction testing for tuberculosis. CI, confidence interval.
Fig. 4.Forest plot for stool for polymerase chain reaction for tuberculosis. CI, confidence interval.
Fig. 5.Forest plots and summary receiver operating characteristic (SROC) curve demonstrate the pooled sensitivity and specificity of the interferon-gamma release assay. CI, confidence interval.
Fig. 6.Publication bias assessment with Deek’s asymmetrical funnel plot for tissue AFB (A), tissue PCR for TB (B), stool PCR for TB (C), and IGRA (D). AFB, acid-fast bacilli; PCR, polymerase chain reaction; TB, tuberculosis; IGRA, interferon-gamma release assay; ESS, effective sample size.
Table 1.Pooled Diagnostic Performance of All Tests
|
Diagnostic test |
No. of study |
No. of patients |
Pooled diagnostic performance
|
|
Pooled sensitivity |
I2
|
Pooled specificity |
I2
|
SROC-AUC |
I2
|
|
Endoscopic tissue biopsy sample |
|
|
|
|
|
|
|
|
|
AFB |
22 |
2,250 |
0.12 (0.08–0.17) |
66.4 |
1.00 (0.00–1.00) |
0.0 |
NA |
NA |
|
Pathology |
26 |
2,252 |
0.18 (0.12–0.27) |
77.8 |
1.00 (0.97–1.00) |
0.6 |
NA |
NA |
|
Culture |
|
|
|
|
|
|
|
|
|
Total |
18 |
2,082 |
0.23 (0.12–0.40) |
84.9 |
1.00 (0.00–1.00) |
0.0 |
NA |
NA |
|
LJ medium |
4 |
308 |
0.06 (0.03–0.13) |
25.2 |
1.00 (0.00–1.00) |
0.0 |
NA |
NA |
|
Liquid medium |
8 |
735 |
0.25 (0.13–0.43) |
82.9 |
1.00 (0.00–1.00) |
0.0 |
NA |
NA |
|
PCR |
|
|
|
|
|
|
|
|
|
Total |
17 |
1,281 |
0.58 (0.44–0.72) |
86.8 |
0.98 (0.94–0.99) |
20.7 |
0.96 (0.94–0.98) |
80.0 |
|
IS6110 |
6 |
530 |
0.44 (0.28–0.61) |
82.0 |
0.96 (0.87–0.99) |
13.8 |
0.95 (0.92–0.96) |
1.0 |
|
Multiplex |
6 |
487 |
0.79 (0.66–0.87) |
73.3 |
0.98 (0.93–1.00) |
23.7 |
0.97 (0.95–0.98) |
0.0 |
|
Xpert MTB/RIF |
7 |
837 |
0.29 (0.17–0.46) |
70.6 |
1.00 (0.75–1.00) |
0.0 |
NA |
NA |
|
Stool sample |
|
|
|
|
|
|
|
|
|
PCR |
4 |
205 |
0.73 (0.43–0.90) |
75.9 |
0.95 (0.79–0.99) |
24.9 |
0.95 (0.93–0.97) |
51.0 |
|
Serum sample: IGRA |
|
|
|
|
|
|
|
|
|
Total |
23 |
3,001 |
0.86 (0.79–0.91) |
74.3 |
0.86 (0.81–0.89) |
72.4 |
0.92 (0.89–0.94) |
97.0 |
|
T-SPOT |
11 |
1,645 |
0.90 (0.84–0.94) |
57.3 |
0.83 (0.74–0.90) |
85.5 |
0.93 (0.91–0.95) |
81.0 |
|
QFT |
8 |
595 |
0.71 (0.53–0.84) |
79.3 |
0.88 (0.82–0.92) |
28.4 |
0.89 (0.86–0.92) |
74.0 |
Table 2.Diagnostic Yield of Combination Between Histopathology and TB Culture
|
Study |
No. of patients with ITB |
Histology, No. (%) |
Culture positivity, No. (%) |
Culture medium |
Combine diagnostic yield (%) |
|
Bhargava et al.[24] |
29 |
12/29 (41.4) |
6/15 (40.0) |
LJ medium |
48.3 |
|
Vij et al.[25] |
28 |
21/28 (75.0) |
13/28 (46.4) |
LJ medium |
75.0 |
|
Shah et al.[26] |
50 |
40/50 (80.0) |
3/50 (6.0) |
LJ medium |
80.0 |
|
Lee et al.[27] |
225 |
52/225 (23.1) |
52/177 (29.3) |
Not stated |
38.7 |
|
Leung et al.[28] |
23 |
3/23 (13.6) |
17/23 (73.9) |
Liquid |
82.6 |
|
Krisch et al.[29] |
18 |
14/18 (77.7) |
14/18 (77.7) |
Liquid |
77.8 |
|
Samant et al.[30] |
61 |
48/61 (78.7) |
31/61 (50.8) |
Liquid |
91.8 |
|
Sekine et al.[31] |
50 |
25/50 (50.0) |
17/34 (50.0) |
Liquid |
76.5 |
|
Mehta et al.[23] |
70 |
42/70 (60.0) |
37/70 (52.9) |
Liquid |
77.1 |
|
Dhoble et al.[32] |
224 |
64/224 (28.6) |
30/224 (13.4) |
Liquid |
29.2 |
REFERENCES
- 1. Chakaya J, Petersen E, Nantanda R, et al. The WHO Global Tuberculosis 2021 Report: not so good news and turning the tide back to End TB. Int J Infect Dis 2022;124 Suppl 1:S26–S29.ArticlePubMedPMC
- 2. World Health Organization (WHO). Global tuberculosis report 2023. Geneva: WHO, 2023.
- 3. Mohammed H, Assefa N, Mengistie B. Prevalence of extrapulmonary tuberculosis among people living with HIV/AIDS in sub-Saharan Africa: a systemic review and meta-analysis. HIV AIDS (Auckl) 2018;10:225–237.ArticlePubMedPMCPDF
- 4. Abu-Zidan FM, Sheek-Hussein M. Diagnosis of abdominal tuberculosis: lessons learned over 30 years: pectoral assay. World J Emerg Surg 2019;14:33.ArticlePubMedPMCPDF
- 5. Holty JE, Gould MK, Meinke L, Keeffe EB, Ruoss SJ. Tuberculosis in liver transplant recipients: a systematic review and meta-analysis of individual patient data. Liver Transpl 2009;15:894–906.ArticlePubMedPDF
- 6. Tanoglu A, Erdem H, Friedland JS, et al. Clinicopathological profile of gastrointestinal tuberculosis: a multinational ID-IRI study. Eur J Clin Microbiol Infect Dis 2020;39:493–500.ArticlePubMedPDF
- 7. Malikowski T, Mahmood M, Smyrk T, Raffals L, Nehra V. Tuberculosis of the gastrointestinal tract and associated viscera. J Clin Tuberc Other Mycobact Dis 2018;12:1–8.ArticlePubMedPMC
- 8. Seo H, Lee S, So H, et al. Temporal trends in the misdiagnosis rates between Crohn’s disease and intestinal tuberculosis. World J Gastroenterol 2017;23:6306–6314.ArticlePubMedPMC
- 9. Cho JK, Choi YM, Lee SS, et al. Clinical features and outcomes of abdominal tuberculosis in southeastern Korea: 12 years of experience. BMC Infect Dis 2018;18:699.ArticlePubMedPMCPDF
- 10. Patel B, Yagnik VD. Clinical and laboratory features of intestinal tuberculosis. Clin Exp Gastroenterol 2018;11:97–103.ArticlePubMedPMCPDF
- 11. Al-Zanbagi AB, Shariff MK. Gastrointestinal tuberculosis: a systematic review of epidemiology, presentation, diagnosis and treatment. Saudi J Gastroenterol 2021;27:261–274.PubMedPMC
- 12. Limsrivilai J, Shreiner AB, Pongpaibul A, et al. Meta-analytic Bayesian model for differentiating intestinal tuberculosis from Crohn’s disease. Am J Gastroenterol 2017;112:415–427.ArticlePubMedPMC
- 13. Cheng W, Zhang S, Li Y, Wang J, Li J. Intestinal tuberculosis: clinico-pathological profile and the importance of a high degree of suspicion. Trop Med Int Health 2019;24:81–90.ArticlePubMedPDF
- 14. Mehta V, Desai D, Abraham P, Rodrigues C. Making a positive diagnosis of intestinal tuberculosis with the aid of new biologic and histologic features: how far have we reached? Inflamm Intest Dis 2019;3:155–160.ArticlePubMedPMCPDF
- 15. Dasgupta A, Singh N, Bhatia A. Abdominal tuberculosis: a histopathological study with special reference to intestinal perforation and mesenteric vasculopathy. J Lab Physicians 2009;1:56–61.ArticlePubMedPMC
- 16. Maulahela H, Simadibrata M, Nelwan EJ, et al. Recent advances in the diagnosis of intestinal tuberculosis. BMC Gastroenterol 2022;22:89.ArticlePubMedPMCPDF
- 17. Yenew B, de Haas P, Babo Y, et al. Diagnostic accuracy, feasibility and acceptability of stool-based testing for childhood tuberculosis. ERJ Open Res 2024;10:00710–2023.ArticlePubMedPMC
- 18. Sudcharoen A, Ruchikajorndech G, Srisajjakul S, et al. Clinical characteristics and diagnosis of intestinal tuberculosis in clinical practice at Thailand’s largest national tertiary referral center: an 11-year retrospective review. PLoS One 2023;18:e0282392.ArticlePubMedPMC
- 19. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71.ArticlePubMedPMC
- 20. Whiting PF, Rutjes AW, Westwood ME, et al. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011;155:529–536.ArticlePubMedPDF
- 21. Deeks JJ, Macaskill P, Irwig L. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol 2005;58:882–893.ArticlePubMed
- 22. Pulimood AB, Peter S, Ramakrishna B, et al. Segmental colonoscopic biopsies in the differentiation of ileocolic tuberculosis from Crohn’s disease. J Gastroenterol Hepatol 2005;20:688–696.ArticlePubMed
- 23. Mehta V, Desai D, Abraham P, et al. Do additional colonoscopic biopsies increase the yield of Mycobacterium tuberculosis culture in suspected ileo-colonic tuberculosis? Indian J Gastroenterol 2018;37:226–230.ArticlePubMedPDF
- 24. Bhargava DK, Kushwaha AK, Dasarathy S, Chopra P. Endoscopic diagnosis of segmental colonic tuberculosis. Gastrointest Endosc 1992;38:571–574.ArticlePubMed
- 25. Vij JC, Malhotra V, Choudhary V, Jain NK, Prasad G, Choudhary A. A clinicopathological study of abdominal tuberculosis. Indian J Tuberc 1992;39:213–220.
- 26. Shah S, Thomas V, Mathan M, et al. Colonoscopic study of 50 patients with colonic tuberculosis. Gut 1992;33:347–351.ArticlePubMedPMC
- 27. Lee YJ, Yang SK, Myung SJ, et al. The usefulness of colonoscopic biopsy in the diagnosis of intestinal tuberculosis and pattern of concomitant extra-intestinal tuberculosis. Korean J Gastroenterol 2004;44:153–159.PubMed
- 28. Leung VK, Law ST, Lam CW, et al. Intestinal tuberculosis in a regional hospital in Hong Kong: a 10-year experience. Hong Kong Med J 2006;12:264–271.PubMed
- 29. Kirsch R, Pentecost M, Hall Pde M, Epstein DP, Watermeyer G, Friederich PW. Role of colonoscopic biopsy in distinguishing between Crohn’s disease and intestinal tuberculosis. J Clin Pathol 2006;59:840–844.ArticlePubMedPMC
- 30. Samant H, Desai D, Abraham P, et al. Acid-fast bacilli culture positivity and drug resistance in abdominal tuberculosis in Mumbai, India. Indian J Gastroenterol 2014;33:414–419.ArticlePubMedPDF
- 31. Sekine K, Nagata N, Shindo T, et al. Combined identifying granuloma and biopsy culture is useful for diagnosing intestinal tuberculosis. Int J Colorectal Dis 2015;30:939–945.ArticlePubMedPDF
- 32. Dhoble P, Desai D, Abraham P. Is the rise in Crohn’s disease in India accompanied by a fall in intestinal tuberculosis? A single-center experience. Indian J Tuberc 2021;68:210–214.ArticlePubMed
- 33. Paulose RR, Kumar VA, Sharma A, et al. An outcome-based composite approach for the diagnosis of intestinal tuberculosis: a pilot study from a tertiary care centre in South India. J R Coll Physicians Edinb 2021;51:344–350.ArticlePubMedPDF
- 34. Shi XC, Zhang LF, Zhang YQ, Liu XQ, Fei GJ. Clinical and laboratory diagnosis of intestinal tuberculosis. Chin Med J (Engl) 2016;129:1330–1333.ArticlePubMedPMC
- 35. Jin T, Fei B, Zhang Y, He X. The diagnostic value of polymerase chain reaction for Mycobacterium tuberculosis to distinguish intestinal tuberculosis from Crohn’s disease: a meta-analysis. Saudi J Gastroenterol 2017;23:3–10.ArticlePubMedPMC
- 36. Sharma K, Sharma M, Sharma V, et al. Evaluating diagnostic performance of Truenat MTB Plus for gastrointestinal tuberculosis. J Gastroenterol Hepatol 2022;37:1571–1578.ArticlePubMedPDF
- 37. Ding YL, Yan RC, Chen HK, Guo XG. A meta-analysis of the accuracy of Xpert MTB/RIF in diagnosing intestinal tuberculosis. Lab Med 2024;55:238–244.ArticlePubMedPDF
- 38. Sharma V, Soni H, Kumar-M P, et al. Diagnostic accuracy of the Xpert MTB/RIF assay for abdominal tuberculosis: a systematic review and meta-analysis. Expert Rev Anti Infect Ther 2021;19:253–265.ArticlePubMed
- 39. Kay A, Vasiliu A, Carratala-Castro L, et al. Performance of a stool-based quantitative PCR assay for the diagnosis of tuberculosis in adolescents and adults: a multinational, prospective diagnostic accuracy study. Lancet Microbe 2024;5:e433–e441.ArticlePubMedPMC
- 40. Balamurugan R, Venkataraman S, John KR, Ramakrishna BS. PCR amplification of the IS6110 insertion element of Mycobacterium tuberculosis in fecal samples from patients with intestinal tuberculosis. J Clin Microbiol 2006;44:1884–1886.ArticlePubMedPMCPDF
- 41. Sachdeva K, Kumar P, Kante B, et al. Interferon-gamma release assay has poor diagnostic accuracy in differentiating intestinal tuberculosis from Crohn’s disease in tuberculosis endemic areas. Intest Res 2023;21:226–234.ArticlePubMedPMCPDF
- 42. Huang X, Liao WD, Yu C, et al. Differences in clinical features of Crohn’s disease and intestinal tuberculosis. World J Gastroenterol 2015;21:3650–3656.ArticlePubMedPMC
- 43. Bae JH, Park SH, Ye BD, et al. Development and validation of a novel prediction model for differential diagnosis between Crohn’s disease and intestinal tuberculosis. Inflamm Bowel Dis 2017;23:1614–1623.ArticlePubMed
- 44. Wu X, Huang H, Hou H, et al. Diagnostic performance of a 5-marker predictive model for differential diagnosis between intestinal tuberculosis and Crohn’s disease. Inflamm Bowel Dis 2018;24:2452–2460.ArticlePubMed
- 45. He Y, Zhu Z, Chen Y, et al. Development and validation of a novel diagnostic nomogram to differentiate between intestinal tuberculosis and Crohn’s disease: a 6-year prospective multicenter study. Am J Gastroenterol 2019;114:490–499.ArticlePubMed
- 46. Gilbert MT, Haselkorn T, Bunce M, et al. The isolation of nucleic acids from fixed, paraffin-embedded tissues-which methods are useful when? PLoS One 2007;2:e537.ArticlePubMedPMC
Appendix
Citations
Citations to this article as recorded by
KASID
, Tanawat Geeratragool1
PubReader
ePub Link
Cite
XML Download
