CD and ITB share many similarities in clinical manifestations and signs, which make them difficult to differentiate [18]. Some studies have proposed certain criteria for identifying CD and ITB, but these criteria have been shown to have great limitations [19-21]. In this study, common symptoms of CD and ITB, such as abdominal pain, diarrhea and ascites, had no significant differences between two groups. Perianal lesions in CD patients were significantly higher than in ITB patients, suggesting that perianal lesions had certain value for the differentiation of the two diseases. Perianal lesions are often the first symptom of CD and are closely related to disease activity, which is why clinicians should pay close attention to them. The data in our study showed that the longitudinal ulcer and stenosis in CD patients were significantly higher than that in ITB patients, and the annular ulcer in ITB patients was significantly higher than that in CD patients. These data were consistent with previous reports [22, 23], and they suggest that the longitudinal ulcer and stenosis are the characteristic manifestations of CD, and that the annular ulcer is the characteristic manifestation of ITB. The pathological changes of CD and intestinal tuberculosis usually occur in the submucosa of the intestinal wall. Because of mucosal swelling, the endoscopic biopsy tissues tend to be too small and the sampling too superficial, resulting in low positive rates. In the current study, there were only three cases of caseous granuloma in ITB endoscopic biopsy, while the pathological positive rate was much lower than that of surgical specimens, suggesting that the differential diagnostic value of endoscopic biopsy specimens in the two diseases was much lower than that of surgical specimens. Non-caseous granuloma was more common in ITB patients than in CD patients, but there was no statistically significant difference. Non-caseous granuloma alone are not enough to completely exclude ITB and should not be considered as a CD-specific manifestation.
At present, WHO recommends Xpert MTB/RIF for rapid diagnosis of pulmonary tuberculosis and some forms of extrapulmonary tuberculosis [24, 25]. A total of 738 clinically suspected extrapulmonary tuberculosis specimens were detected by Xpert MTB/RIF, and the sensitivity and specificity of Xpert MTB/RIF for detecting extrapulmonary tuberculosis were 84.9% and 86.7% respectively. This suggested that Xpert MTB/RIF detection was a promising method for rapid diagnosis of extrapulmonary tuberculosis [26]. Currently, there are few reports on Xpert MTB/RIF detection in intestinal mucosa in China and abroad [27, 28].40 patients with ileocecal ulcers were studied and the intestinal tissue biopsy was evaluated for GeneXpert-Mtb/Rif for the diagnosis of ITB [27]. The research demonstrated that the sensitivity of GeneXpert-Mtb/Rif for diagnosis of ITB was 32%. In a retrospective study from India ,of 37 intestinal TB patients, the Xpert MTB/RIF assay was positive in three of 37 (8.1%) [28].In the present study, the biopsy and surgical specimens from 42 cases of intestinal tuberculosis and 46 cases from Crohn’s disease were examined by Xpert MTB/RIF. The results showed that the positive rate of M. tuberculosis detected by Xpert MTB/RIF in intestinal tissue samples of ITB patients was 33.33%, which was higher than that of CD patients (0%), and the specificity was 100%. This suggested that the detection of Xpert MTB/RIF in intestinal mucosa might have an important role in the differential diagnosis of intestinal tuberculosis and Crohn’s disease.
In this study, the positive rate of M. tuberculosis detected by Xpert MTB/RIF in endoscopic biopsy specimens was much lower compared to surgical specimens, which might be because the endoscopic biopsy specimens were superficial and small. MTB was not evenly distributed in intestinal tissues, and low levels of MTB extracted from mucosal biopsy specimens may lead to false negative results of Xpert MTB/RIF in tissues.
The gold standard for the etiological diagnosis of tuberculosis is M. tuberculosis culture. However, the growth of M. tuberculosis usually takes 3-8 weeks, and the positive rate of culture is low, which often causes difficulties in diagnosis. Although the routine acid-fast staining method takes only 2-3 hours, this approach has very low sensitivity and specificity of strain identification. Even though it can detect the acid-fast Mycobacterium, it cannot distinguish M. tuberculosis from other mycobacterium. The sensitivity and accuracy of PCR for pathogen detection provides a useful and new method for the diagnosis of M. tuberculosis [29-31].
In this study, the detection of M. tuberculosis by Xpert MTB/RIF in the intestinal mucosa was compared with traditional methods, and the results showed that the sensitivity of Xpert MTB/RIF was 33.3%, which was significantly higher than that of acid-fast staining (11.9%) and higher than that of bacillus culture (21.4%). The specificity of the three methods was 97.9%, 100% and 100% respectively, without significant difference. The above results indicated that Xpert MTB/RIF had superior sensitivity compared to acid-fast staining in detection of Mycobacterium tuberculosis, it also had superior detection time compared with M. tuberculosis culture, and it had higher specificity. Therefore, the Xpert MTB/RIF was better approach for the early diagnosis and treatment of intestinal tuberculosis.
TB culture with Mycobacterium Growth Indicator Tube (MGIT) system was reported to be positive in 20% to 42% in colonoscopic biopsy specimens[32-33]. In our study,the positive rate of MTB detected by tissue culture in ITB patients was 21.43%( 7.14% in biopsy specimens and 50% in 14 surgical specimens,LJ medium). Lower yield of AFB culture in biopsy specimens may be associated with culture medium and the number of colonoscopic biopsies.The yield of AFB culture on LJ medium is lower than that on MGIT system. however it should be noted that false-positive rates for MGIT system was higher than LJ methods[34].In a recently report, additional four (total eight) colonoscopic biopsies improved the yield of TB culture positivity over four biopsies by 11.4% to 14.3%[33]. If for improving AFB culture positivity, Increasing the number of colonoscopic biopsy specimens to eight should be considered,although it cost more time.
Over recent years, following the development of immune technology, an increasing attention has been paid to the study of MTB antigen [13, 35]. At present, the main samples used for MTB antigen diagnosis are body fluid and bacterial culture medium. Reports on the application of tuberculosis antigen in ITB tissues are very rare. There are only two reports on 38KDa antigen with limited number of ITB cases (≤ 10 cases). Ihama et al. have performed immunohistochemical detections of 38KDa antigen in intestinal paraffin specimens of 10 ITB patients, revealing that 40% of the patients were positive, which in turn suggested that immunohistochemical detection with MTB monoclonal antibody might be an effective and simple diagnostic method for ITB [36]. Ince and colleagues have performed immunohistochemical detection of 38KDa antigen in 45 tuberculosis tissue specimens (including 8 colon tissue specimens, 8 skin specimens, 5 lung tissue specimens and 24 lymph node specimens), and found that 73% of tuberculosis patients were positive, while only 2 out of 28 patients with CD were positive, suggesting that immunohistochemical detection of 38KDa antigen could be used to distinguish the early diagnosis of MTB from CD [37]. Currently, there are no reports on the expression of pstS1 (38KDa) and East-6 antigens in intestinal tissues. The results of this study showed that the expression of Ag85B, pstS1 (38KDa) and East-6 antigen proteins in intestinal tuberculosis and Crohn’s disease had no significant difference. The positive expression rates of Ag85B, pstS1 (38KDa) and East-6 antigen in intestinal tissues were 9.52%, 23.8% and 26.2%, respectively. The detection sensitivity of these three proteins in intestinal tissues was low, and could not be improved even by combined detection. Future studies should examine whether other detection methods such as PCR could improve the detection sensitivity.
Jørstad et al. have used immunohistochemistry to detect MPT64 antigen in 132 cases of extrapulmonary tuberculosis and non-tuberculosis specimens, and found that compared with the diagnostic composite reference standard, the sensitivity, specificity, positive predictive value and negative predictive value of MPT64 were 69%, 95%, 94%, 75% and 82%, respectively. MPT64 detection had the best performance in TB lymphadenitis patients and children with TB, suggesting that MPT64 detection could be used for routine diagnosis under low resource allocation to improve the diagnosis of extrapulmonary tuberculosis, especially for tuberculous lymphadenitis and children with TB [38]. Purohit et al. have studied 51 cases of pulmonary tuberculosis and 38 control specimens of non-pulmonary tuberculosis, including fine-needle aspirations biopsies and formalin fixed biopsy specimens. The sensitivity, specificity, positive predictive value and negative predictive value of anti-MPT64 immunostaining detection were 100%, 97%, 97%, 100% and 82% respectively, suggesting that anti-MPT64 immunostaining detection was a rapid and sensitive method for early specific diagnosis of Mycobacterium tuberculosis infection [39]. This technique can easily be incorporated into routine pathological laboratories. Currently, there are no reports on the expression of MPT64 antigen in intestinal tissues.
In this study, the positive rates of MPT64 antigen in intestinal tuberculosis and Crohn’s disease were 40.48% and 19.58%, respectively. The observed difference was statistically significant, which suggested that MPT64 antigen has certain value in the differential diagnosis of intestinal tuberculosis and Crohn’s disease. The sensitivity of combined detection with Xpert MTB/RIF in the diagnosis of pulmonary tuberculosis was 50.0%, which was higher compared to the two methods used individually; however, the specificity decreased.