Introduction

Intestinal tuberculosis (ITB) is a chronic infectious disease which is complicated by intestinal strictures in about 50% patients [1]. Patients with intestinal strictures present with recurrent subacute intestinal obstruction (SAIO) as their major symptom which leads to frequent hospitalizations and poor quality of life. The stricture in ITB patients can be inflammatory or fibrotic depending upon the disease activity. The resolution of these strictures with anti-tuberculous therapy (ATT) has not been well studied, and there are few studies, mostly case reports and case series, on this aspect. The earliest reports on this aspect published in the 1940s documented an excellent symptomatic response, but a poor stricture healing after streptomycin [2,3,4]. Further reports published which used a combination regimen of ATT again showed lower frequency of stricture resolution, and it was even hypothesized that rifampicin could lead to an increase in incidence of cicatrisation and obstruction [5, 6]. However, these were retrospective studies with varying follow-up duration, and because of this lack of systematic evidence of ATT on stricture response, Anand et al. [7] in a prospective study of 34 patients showed that with streptomycin-based ATT, 91% had significant clinical improvement and stricture resolution on barium studies was seen in 70% patients after completion of ATT, which was much higher than previously published studies. In another study of 30 patients with colonic tuberculosis and luminal narrowing, stricture response after ATT (evaluated endoscopically) was seen only in 53% patients [8]. Endoscopy directly visualizes the mucosa and would be more sensitive than radiology in detecting and following up strictures in ITB patients. This discrepancy can explain the difference in the stricture healing rates in latter 2 studies. Therefore, the exact status of stricture resolution after ATT in ITB patients remains unclear. Off late barium studies are being replaced by cross-sectional imaging (CT/MR enterography: CTE/MRE) for evaluation of small bowel, and CTE/MRE has been shown to be equally sensitive/better than barium studies in evaluation of small intestine [9, 10]. In a recent study, MRE was more sensitive that small bowel follow through (SBFT) in the diagnosis of ITB and was also able to show higher number of strictures than SBFT [11]. However, no study has evaluated the stricture healing after ATT with CTE/MRE and the endoscopic study included only patients with colonic tuberculosis. The baseline factors which can predict the stricture resolution after ATT also remain largely unknown. Therefore, the present study evaluated the frequency of symptomatic, endoscopic, and radiologic resolution of tubercular intestinal strictures to ATT and predictive factors of response.

Materials and Methods

Patients

Patients of intestinal tuberculosis (ITB) following up at Department of Gastroenterology at the All India Institute of Medical Sciences, New Delhi, India, from January 2004 to December 2015 were screened for inclusion in the study. The patients were included if they had intestinal strictures prior to starting ATT, remained on regular follow-up, and had repeat evaluation for stricture after completion of ATT. Written informed consent was obtained from the patients, and permission from the Institute Ethics Committee was obtained (Ref No.: IESC/T-252/2010).

Study Design

This was an ambispective cohort study in which of the 106 patients included, 36 were followed in a prospective manner, while the data for 70 patients were collected in a retrospective manner. At our center, all patients suspected to have ulceroconstrictive disease of intestine are subjected to a uniform evaluation including baseline complete blood count, liver function tests, renal function tests, cross-sectional imaging (CT/MR enterography), endoscopy (UGI endoscopy/ileocolonoscopy according to suspected site of involvement) and biopsy for histopathological examination, acid-fast bacillus (AFB), and culture and polymerase chain reaction (PCR) for Mycobacterium tuberculosis. CT/MR enterography is done according to a pre-defined protocol followed at our center.

For prospective patients, the data were compiled on a pre-defined proforma on demographics, mode of diagnosis of ITB, disease duration, duration of ATT, clinical features, clinical and radiologic/endoscopic response to ATT, and follow-up duration of ATT. For retrospective patients, the data were extracted from patient files and by clinical follow-up. Our medical records comprise a paper file system where each patient is assigned an IBD number, mentioned on the face of the file. Each file contains all dated information concerning the patient’s disease, including a detailed history, medical examination, relevant test findings, and follow-up symptom assessment. A team of physicians maintains the files, and the parameters used for the assessment are consistent between physicians.

Patients were divided into two groups on the basis their stricture response to ATT for further analysis.

Diagnosis of Intestinal TB [12, 13]

The diagnosis of ITB was made in an appropriate clinical setting with the demonstration of necrotizing granulomas on histopathology or demonstration of acid-fast bacilli (AFB) on histolopathology or culture of intestinal tissue. In patients who did not fulfill the above definitions, therapeutic trial with anti-tubercular treatment was given. Diagnosis of ITB was made if patient had clinical and endoscopic/radiologic response to anti-tubercular treatment. These defining criteria for diagnosis of ITB have been validated earlier in another study at our center [14].

Treatment and Follow-Up

Anti-tuberculous therapy [15]: induction regimen of isoniazid 5 mg/kg, rifampicin 10 mg/kg, pyrazinamide 20–25 mg/kg, and ethambutol 15–20 mg/kg for 2 months followed by maintenance therapy of isoniazid and rifampicin for duration of 4–7 months, was given. The total treatment duration was 6–9 months depending upon the response to ATT. Patients were routinely followed up at 2 monthly intervals and as and when required to assess symptom response and also to monitor drug toxicity by the assessment of liver function tests. Repeat endoscopy/cross-sectional imaging was done to assess mucosal healing (on endoscopy) or radiologic resolution.

Blinding For patients whose data were evaluated retrospectively, the endoscopy reports were retrieved from the records file and follow-up radiology images were evaluated by the radiologists (RS and KSM) blinded to the diagnosis and treatment response.

For the patients who were included prospectively, colonoscopy was done by the investigator who was blinded to the clinical details and radiology images were reviewed by the radiologists in a blinded fashion.

Definitions

Stricture On endoscopy, the presence of luminal narrowing (with or without surrounding ulceration) was defined as stricture. It was further classified into two types depending upon whether scope could be negotiated across the stricture or not. On cross-sectional imaging, the presence of bowel wall thickening accompanied by pre-stenotic dilatation of the bowel wall (>3 cm) was defined as stricture [1].

Stricture was further classified as inflammatory (active) upon the presence of one or more of following features on endoscopy or radiology

  • Inflammatory stricture

    • Endoscopy: the presence of stricture with evidence of active ulceration and/or erythema

    • Radiology [16, 17]: enhancement of the involvement segments in comparison with surrounding segments, the presence of mural stratification, vascular engorgement of surrounding mesentery (comb sign), and diffusion restriction on diffusion-weighted MRI.

  • Fibrotic stricture: the absence of any of these features classified the stricture as fibrotic.

Ulceroconstrictive disease GI disease which occurs due to ulcers or stricture in the small or large intestine

Mucosal healing Defined as the absence of any evidence of ulceration, erythema (suggestive of active inflammation) on follow-up colonoscopy. Radiologically absence of any signs of inflammatory stricture (defined above) also classified as mucosal healing.

Symptom response pertaining to structuring disease Resolution of symptoms attributed to stricture (pain abdomen/recurrent subacute intestinal obstruction)

Resolution of stricture The absence of any evidence of stricture on either endoscopy or radiology

Location of stricture Divided into following categories

  • Distal ileal/ileocecal: stricture located in distal/terminal ileum with or without involvement of adjacent ileocecal valve or cecum

  • Colonic: stricture located in colon

  • Ileocolonic: stricture located in small as well as large bowel

  • Proximal small bowel: stricture involving jejunum and/or proximal ileum

  • Gastroduodenal: stricture located in stomach (antropyloric area) or duodenum

Length of stricture

  • Short: ≤3 cm

  • Long: >3 cm

Statistical Analysis

Continuous variables were expressed as median with interquartile range (IQR), and categorical variables were expressed as percentages. Baseline factors were compared between patients who had stricture resolution versus those who did not have. Student’s t test was used to compare continuous variables, and Chi-square test was used to compare categorical variables. p value <0.05 was considered as statistically significant. SPSS software (v. 17) was used for analysis.

Results

Of 286 patients diagnosed as intestinal tuberculosis, 128 were found to have a stricturing lesion. Of these 128 patients, 14 were lost to follow-up or were not willing to attend the clinic for post-ATT response evaluation. Three patients underwent a surgery prior to starting ATT, while 5 patients had to be operated for the SAIO prior to completing ATT. Therefore, 106 patients were finally included for final evaluation (Fig. 1).

Fig. 1
figure 1

Consort flowchart showing number of patients finally included in the study

Full size image

Baseline Clinical and Demographic Characteristics

Clinical Features

The median age of entire cohort was 35 (27–45) years (63 males). Median disease duration was 12 (6–24) months. Fever was present in 44 (41.4%) patients, pain abdomen in 99 (93.4%), recurrent SAIO in 80 (75%), and weight loss in 86 (81%) patients (Table 1). Symptoms pertaining to stricture (pain abdomen/recurrent SAIO) were seen in 104/106 (98%) patients.

Table 1 Baseline clinical characteristics of entire cohort
Full size table

Endoscopic/Radiologic Features

More than one stricture was seen in 16 (15.1%) patients with 5 patients having multiple ileal strictures, 4 having multiple strictures in the colon, 4 having both ileal and colonic strictures, and 3 patients having multiple jejunal strictures. 16.7% patients had long strictures (>3 cm). Stricture was located in distal ileum/ileocecal area in 52 (49.1%), colon in 37 (34.9%), ileum and colon in 4 (3.8%), proximal small bowel in 10 (9.4%), and gastroduodenal area in 4 (3.8%) patients. The evidence of stricture was both endoscopic and radiologic in 74/106 patients (Table 1).

Diagnosis of ITB

Intestinal TB was diagnosed in 9 patients on the basis of caseating granuloma, 3 had positive AFB smear, and 5 had positive culture. More than one investigation was positive in 3 patients. Diagnosis in rest of the patients (n = 93) was made on the basis of clinical and endoscopic/radiologic response to ATT (Table 1).

Response to ATT

Clinical

The median duration of ATT was 6 (6–9) months, and median follow-up duration after ATT was 13.5 (5–30) months. Fever resolved in all the patients after ATT. Weight gain after ATT was seen in 68/86 (79.5%) patients. Resolution of SAIO was seen in 46/80 (58%) patients, and pain abdomen improved in 68/99 (68.7%) patients. Resolution of symptoms pertaining to stricture (both pain abdomen/recurrent SAIO) was seen in 52/104 (50%) patients only (Table 2).

Table 2 Follow-up characteristics in patients after ATT
Full size table

Endoscopic/Radiologic

All patients had evidence of mucosal healing on follow-up endoscopic/radiologic imaging. However, complete resolution of stricture was seen only in 25/106 (23.6%) patients (Table 2).

Comparison of Features in Patients Who Had Stricture Resolution Versus Those Who Did Not Have (Table 3)

Clinical Features

There was no difference in age at presentation, disease duration, duration of ATT, gender, and duration of follow-up between the patients who had stricture resolution versus those who did not have (Table 3). Fever resolved in all the patients irrespective of the persistence of stricture. All except 1 patient (21/22, 95.4%) who had resolution of stricture had weight gain after ATT as compared to 47/64 (73.4%) patients with persistent stricture (p = 0.029). Similarly, significantly higher number of patients with stricture resolution had improvement in pain abdomen (87 vs 63%, p = 0.031) and symptoms related to SAIO (100 vs 46.8%, p < 0.001). Therefore, the resolution of stricture symptoms (pain abdomen/recurrent SAIO) was also significantly higher in patients who had resolution of stricture (88 vs 38%, p < 0.001).

Table 3 Comparison of baseline and follow-up features between those who had stricture resolution on ATT versus those who did not have
Full size table

Endoscopic/Radiologic

There was significant variation in stricture resolution with respect to disease location (Figs. 2, 3, 4). Only 5.4% colonic strictures resolved after ATT (Fig. 4) as compared to 36.5% stricture resolution seen in patients with distal ileal/ileocecal strictures (Figs. 2, 3). The frequency of colonic stricture was significantly higher (43 vs 8%, p = 0.001), and distal ileal/ileocecal stricture was significantly lower (40.7 vs 76%, p = 0.004) among patients with persistent stricture as compared to those with stricture resolution. Though not statistically significant, stricture resolution was less frequent among patients with multiple (4 vs 18.5%, p = 0.076) and longer strictures (10 vs 22.9%, p = 0.217), and more frequent in patients with less tighter strictures (scope negotiable prior to ATT) (28.6 vs 14.0%, p = 0.193) than single strictures, shorter strictures, and tighter strictures (scope not negotiable prior to ATT) (Table 3).

Fig. 2
figure 2

MR enterographic image in a patient with ileocecal TB showing resolution of stricture after treatment with ATT

Full size image
Fig. 3
figure 3

MR enterographic image in a patient with ileocecal TB showing persistent stricture after treatment with ATT

Full size image
Fig. 4
figure 4

Colonoscopic image of patients with a ascending colon stricture with evidence of active inflammation (before ATT), b mucosal healing after ATT (without any evidence of active inflammation) in a patient with persistent ascending colonic stricture

Full size image

Discussion

The goals of treating any disease not only include symptom resolution but also the correction of associated structural or functional abnormality. Strictures are the structural consequence of intestinal tuberculosis and major cause of morbidity associated with ITB. Effective treatment of intestinal TB not only includes clinical response but also the resolution of strictures, which according to the current literature is not considered as an end point. In the background of limited and contradictory evidence, the present study presents the largest evidence till date on this aspect.

The evidence on the response of intestinal strictures to ATT dates back to 1948 when in a study of 32 patients treated with streptomycin, almost all had a symptomatic response, but the radiologic improvement was described as “leisurely” in appearing and normal picture was not usually observed [2]. In another study of 30 patients, only 6/17 (35.3%) patients with abnormal barium findings before treatment had evidence of complete radiologic healing after treatment with streptomycin [3]. Another study published in the same year showed that again only 6/17 (35.3%) patients had complete radiologic healing, with 8/17 having partial improvement after treatment [4]. However, these were retrospective reports published more than 60 years from now and used only streptomycin as anti-tubercular therapy which could have led to low stricture response after ATT. This was followed by a prospective study from India by Anand et al. [7] which revealed a better (70%) stricture healing rate (assessed radiologically) after streptomycin-based combination ATT. Another recent study showed an intermediate (53%) improvement (assessed endoscopically) in luminal narrowing among 30 patients with colonic tuberculosis and luminal narrowing [8]. Although the evidence is limited and variable across all studies, except for the study by Anand et al., the stricture healing rate after ATT has not been more than 50%. The present study also matched the consensus and revealed a poor stricture resolution rate of only 23.6%. More than 85% strictures in the present study were diagnosed on endoscopy which is much more sensitive than radiology in diagnosing as well as assessing the resolution of stricture after treatment. Secondly, stricture has both an inflammatory and a fibrotic component, and with effective therapy, the inflammatory component gets resolved leading to healing and scarring which may lead to persistence of fibrotic stricture. This is evidenced by the fact that in spite of low frequency of stricture resolution all patients had evidence of mucosal healing indicating that inflammatory component of stricture had resolved. These factors could account for the lower frequency of stricture resolution in the present and other studies. This was in spite of the fact that all patients received standard 4 drug ATT in the present study as compared to the use of rifampicin, isoniazid, and streptomycin in the study by Anand et al. Intestinal strictures may lead to poor absorption of orally administered drugs, and the use of parenteral streptomycin in the study by Anand et al. could be one of the other factors accounting for higher stricture resolution in their study.

Predictors of non-resolution of strictures in the study by Anand et al. [7] were younger age, female gender, longer duration of symptoms, and multiple and longer strictures, although none of these factors were statistically significant. In the present study also patients with persistent strictures were relatively younger, had multiple, longer, and tighter (scope not negotiable prior to ATT) strictures, although the difference was again not statistically significant. In the present study, stricture location was the most important predictor of stricture resolution with colonic strictures having the least resolution rate (5.4%) followed by proximal small intestinal (20%) and distal ileal/ileocecal (36.5%) strictures. However, Mukewar et al. [8] reported 53% stricture healing rate after ATT in their patients with colonic TB which was quite higher than the present study. Possible factors responsible for this discrepancy could be higher frequency of non-negotiable stricture in the present study (83 vs 60%) as compared to that study. In the study by Anand et al., extending the ATT for 2 years in 2 patients had led to stricture resolution in both indicating that longer duration of ATT may lead to better stricture healing rate. Lowest resolution rate in the colonic strictures could also be because of better accessibility of colon to endoscopy, leading to better detection even of minor strictures which could have been missed on CT/MRI.

This study lays the platform for further research into exploring ways for improving the stricture healing in patients of ITB with intestinal strictures. These strictures are the result of excessive intestinal fibrosis that occurs as a result of response to injury/inflammation of the intestine [18]. Therefore, co-administration of steroids, other anti-inflammatory agents, or anti-fibrotic agents with ATT needs to be studied. Increasing the duration of ATT to 9 months or more could also improve the stricture healing as evidenced in the study by Anand et al. However, a recent multicenter RCT from India had shown similar response rates between 6 and 9 months of ATT [19], but there was no subgroup analysis for stricturing intestinal TB. So whether extending ATT beyond 6–9 months would lead to better stricture healing remains unclear. Adding streptomycin to the 4 drug regimen in stricturing disease could also be one of the strategies as evidenced from study by Anand et al. [7]. The use of early endoscopic dilatation in accessible areas, especially when the strictures are more pliable, could also lead better stricture resolution rate [20].

There are few limitations with this study. Definite diagnosis of ITB was possible only in 12% patients, and in rest the diagnosis was made after response to ATT. However, with poor sensitivity of definitive diagnostic techniques one has to resort to a therapeutic trial of ATT in majority of circumstances. One could also doubt whether the included cases could be patients of CD, but all patients had mucosal healing after ATT which confirms the diagnosis of ITB [14]. We have recently documented that endoscopic and clinical response to ATT can differentiate CD and ITB, with all patients of ITB having mucosal healing as compared to only 5% in patients with CD [14]. The same study demonstrated more than 90% symptomatic response in ITB patients, which is quite high as compared to the present study. However, only 40% patients in the previous study had strictures, and in the present study, 88% patients who had resolution of strictures had complete symptomatic response. The postoperative histology in none of the operated patients was suggestive of CD which again confirms the diagnosis against CD.

Therefore, stricture resolution occurs only in one-fourth ITB patients after ATT and depends upon the disease location. Tightness of stricture, number, and length of strictures may also affect the resolution rate. The study provides an insight into further research for ways of improving stricture healing in patients with ITB.