Capsule endoscopy in small bowel Crohn’s disease and Tuberculosis

Introduction

Small bowel tuberculosis (SBTB) and Crohn’s disease (CD) are chronic granulomatous disorders of the gastrointestinal tract which are difficult to differentiate.^1–7 This difficulty arises because of great similarity in clinical, endoscopic, radiological and even histopathological features.^1–5 This is not only an important diagnostic dilemma in tropical countries such as India where TB is still widely prevalent and the frequency of CD is gradually increasing, but also in Western countries where the incidence of TB is increasing. An accurate way of differentiating between CD and SBTB is to document Mycobacterium tuberculosis in the intestinal tissue. Acid-fast bacilli (AFB) staining of intestinal tissue lacks sensitivity and specificity; culture of the biopsy specimen is time-consuming with results being frequently negative. ⁸

The natural history as well as treatment of these two diseases is different and repercussions of wrong treatment are disastrous. Treating SBTB as CD with immunosuppressive agents would make the disease worse and widespread. Similarly, treating CD with anti-TB drugs may delay the appropriate treatment of CD, put the patient to risk of hepatotoxicity of anti-tubercular therapy (ATT) and thus impair the quality of life as well as increase the health burden.^7,9

Various endoscopic, serological, radiological and histopathological criteria have been proposed to differentiate between CD and SBTB.^1–8 However, all these investigations and criteria have limitations and, in many situations, even after extensive investigations, a diagnostic dilemma persists. Interferon gamma release assay (IGRA) and polymerase chain reaction (PCR) for detection of M. tuberculosis in the intestinal biopsy specimen have been reported to be useful investigational modalities for differentiation between SBTB and CD but experience of their use is limited.^10,11

Capsule endoscopy (CE) can effectively diagnose subtle small intestinal (SI) lesions missed by radiological investigations and therefore has become the investigation of choice for evaluating suspected small bowel lesions.^12–14 CD, because of being a widespread chronic inflammatory disease, may have subtle asymptomatic involvement of the gastrointestinal tract (GIT) which may be missed by radiological investigations. Aphthous ulceration, an early morphological change of CD, can be missed on radiological investigations but is easily detected on CE.^12–14 Therefore, CE helps in better mapping of the morphological changes in SI in patients with CD.

On the other hand, SBTB is usually a focal, localised disease of the GIT commonly involving the ileocaecal area with morphological changes in SI described previously either in autopsy or radiological studies. The mucosal changes in SI in patients with SBTB have not been previously studied by a sensitive diagnostic modality such as CE. Therefore, we conducted this prospective study to evaluate the small bowel findings on CE in patients with SBTB. We also hypothesised that there may be subtle difference in small bowel mucosal changes between the two diseases that may be better detected by CE and therefore compared the small bowel mucosal changes detected by CE between these two diseases.

Patients and methods

All consecutive patients with suspected SBTB or CD (based on clinical findings, radiological investigations and ileocolonoscopy findings) seen between March 2012 and December 2014 were prospectively enrolled and followed up till December 2015. All enrolled patients were symptomatic, had ileocaecal thickening with or without intra-abdominal lymphadenopathy on computed tomography (CT) examination, had terminal ileal involvement on ileocolonoscopy and histological examination of the endoscopic mucosal biopsies of terminal ileum was non-diagnostic. Patients having diagnostic endoscopic ileal biopsy were excluded from the study. The presence of caseating granulomas or AFB in the ileal biopsy was considered as diagnostic of SBTB and these patients were excluded. Informed consent was obtained from all patients prior to CE and the study protocol was approved by the institute ethics committee. Patients with suspected intestinal obstruction, intestinal diverticulum, gastroparesis, stricture or fistula, pregnancy, previous gastric or small bowel surgery were excluded, as well as patients with neurological disorders.

Following enrolment, patency of the GIT was confirmed by ingestion and subsequent passage of a patency capsule (PillCam® patency capsule; Given Imaging Ltd, Yoqneam, Israel) within 30 h of ingestion as visualised by the patient and confirmed by the attending physician. In a few patients, the patency was assessed by absence of stricture on Barium enteroclysis or CT enterography. Radiological tests were considered positive for non-patent GIT if there was small bowel dilation >3 cm proximal to stricturing in the small bowel. However, in the absence of small bowel dilatation, any visualised narrowing was ignored and the small bowel was considered to be patent. Patients with non-patent GIT were excluded and patients with patent GIT with no significant stricture were further studied by CE within seven days of passage by patency capsule or radiological study.

Capsule endoscopy procedure

CE was performed using Pillcam SB capsule (Given Imaging Ltd, Israel) after overnight fasting and bowel preparation with 2 L of polyethylene glycol solution. Patients fasted during the first 4 h after ingestion of the capsule, then received a clear liquid and a meal 5 h after beginning CE recording. The recording device was returned in the evening for analysis and the patient was sent home. Data were downloaded to a workstation (RAPID workstation, Given Imaging) and analysed. The small bowel mucosal findings were recorded and analysed later.

Further work-up and follow-up

Based on the CE findings, patients were further evaluated by push enteroscopy or double balloon enteroscopy with biopsies from the suspected lesions. Any further investigations required in order to arrive at the final diagnosis were done at the discretion of the treating physician. Since TB is common in our country, the majority of patients with indeterminate findings were labelled as SBTB after initial work-up and started on ATT. However, the diagnosis of SBTB was confirmed by documenting both clinical as well as endoscopic response to ATT. Clinical response was determined by subjective disappearance of symptoms and endoscopic response was determined by disappearance of the ulcerations.^2,15

The diagnosis of CD was made on the combination of clinical, endoscopic, radiological and histological features. ¹⁶ Patients who were suspected of SBTB after initial work-up and did not respond to ATT were considered as having CD. The final diagnosis of CD was made after a positive response to therapy for CD had been documented.

Clinical response was defined as complete resolution of all clinical symptoms and signs with therapy, while endoscopic response was defined as complete resolution of all ulcerations and mucosal nodularity visualised in the terminal ileum on baseline ileocolonoscopy. This ileocolonoscopy was repeated in the following situations: after three months of appropriate therapy or at any time if symptoms worsened.

Once the final diagnosis was established on follow-up, the CE findings recorded before institution of treatment between SBTB and CD were compared.

Statistical analysis

Results were presented as percentages for categorical variables and mean ± SD for continuous variables. CE findings were compared between patients with final diagnoses of SBTB or CD. Differences in the findings between the two groups were examined by using the χ ² test. Fisher’s exact test was used for comparison of categorical variables in small sample sizes.

Results

A total of 52 patients with suspected SBTB or CD with terminal ileal involvement were prospectively enrolled: 32 (20 men; age range, 18–42 years) with a final diagnosis of SBTB and 20 (14 men; age range, 20–54 years) with CD. The chest radiograph was normal in all patients with CD whereas five (16%) with SBTB had radiological evidence of pulmonary TB. None of the patients with CD had tuberculin positivity whereas 14 (44%) with SBTB had tuberculin positivity. IGRA was not done in any patient.

The patency of the GIT was evaluated with patency capsule in 16 (80%) patients with CD and nine (28%) with SBTB, respectively. The remaining patients underwent barium enteroclysis or CT enterography for assessment of patency. The patency capsule did not pass out intact within 30 h of ingestion in two patients with CD and four patients with SBTB and these patients were excluded. Sixteen patients with SBTB had a tight stricture on radiological evaluation and four patients had confirmatory histopathology with presence of AFB on ileal biopsy and therefore these 20 patients were also excluded. Of the 52 consecutive enrolled patients, 2/20 with a final diagnosis of CD and 24/32 with SBTB had an ileal stricture precluding CE (P = 0.0001).

After exclusion, CE was done in 18 patients (14 men; mean age, 31.4 ± 10.3 years) with CD and eight patients (5 men; mean age, 36.2 ± 10.2 years) with SBTB. Non-caseating granulomas were seen in 4/18 (22%) patients of CD and 1/8 (13%) patients of SBTB on histopathological examination of terminal ileal endoscopic biopsies. The remaining patients had non-specific inflammation. The CE traversed the whole length of the small bowel during the battery life period of the capsule in all eight patients with SBTB and in 15 (83%) patients with CD. All three patients with CD with incomplete small bowel examination had prolonged small bowel transit time with no identifiable stricture. The mean small bowel transit time was 261 ± 111 min in SBTB patients and 277 ± 66 min in CD patients (P > 0.05). None of these 26 patients who underwent CE had capsule retention.

The CE findings were compared between eight SBTB and 15 CD patients who had complete small bowel examination by CE (Table 1). All patients with SBTB had terminal ileal involvement along with involvement of the ileocecal valve (ICV) in the form of ulcerations and nodularity and these findings were similar to the findings observed on ileocolonoscopy. Six patients (75%) had large ulcers (Figure 1) and two (25%) patients had aphthous ulcers (Figure 1). The large transverse or oblique ulcers were located in the ileum and had irregular borders. The remaining small bowel was normal in all eight patients of SBTB.

OPEN IN VIEWER

Table 1.

CE findings in SBTB and CD.

CE finding	SBTB (n = 8)	CD (n = 15)	P value
Ileocecal valve involvement	8 (100%)	5 (33%)	0.002
Large ulcers	6 (75%)	7 (47%)	0.37
Aphthous ulcers	2 (25%)	15 (100%)	0.007
Proximal small bowel involvement	0 (0%)	5 (33%)	0.12

OPEN IN VIEWER

Figure 1.

Capsule endoscopy findings. (a) Patient with SBTB: large ulcer with narrowing; (b) patient with SBTB: aphthous ulcer in ileum; (c) Patient with CD: large ulcer in ileum; (d) patient with CD: aphthous ulcer in jejunum; (e) patient with CD: nodularity in jejunum.

All patients with CD also had terminal ileal involvement but ICV involvement was seen in 5/15 and these findings were similar to those observed on ileocolonoscopy. All 15 patients with CD evaluated with CE had aphthous ulcers and seven (47%) also had large ulcers (Figure 1). The morphology of large ulcers in CD was similar to that of ulcers seen in SBTB. Involvement of small bowel proximal to terminal ileum in the form of aphthous ulcers and nodularity (Figure 1) was seen in four (27%) and one (7%) patients, respectively.

On comparison with CD, patients with SBTB had increased frequency of ICV involvement (100% versus 33%; P = 0.002) and lesser frequency of aphthous ulcers (25% versus 100%; P = 0.007). The frequency of large ulcers was similar between patients of SBTB and CD (75% versus 47%; P = 0.37). Involvement of segments of the small bowel other than the terminal ileum was observed in 33% of patients with CD but in none with SBTB, although this difference was not statistically significant.

Follow-up

All eight patients with a final diagnosis of SBTB were successfully treated with ATT and none developed stricture or required surgical resection. Of the 18 patients of CD who underwent CE, 13 (72%) have been regularly followed up in our clinic. Of these, two developed an ileal stricture which required surgical resection in one and was successfully managed with endoscopic balloon dilatation in the other. Both these patients were initially diagnosed as SBTB and treated with ATT.

Discussion

Differentiation of SBTB from CD is an important and difficult diagnostic dilemma that clinicians, especially in tropical countries, face in their clinical practice. ¹⁷ The rise in incidence of CD in tropical countries such as India where TB is endemic further complicates the issue.^17,18 Both diseases have similar clinical, radiological, endoscopic and even histopathological features, and this makes their differentiation very difficult. CE has been established as a safe and effective modality for the diagnosis of CD as it has been shown to detect small or subtle bowel lesions missed by conventional endoscopy or radiologic studies.^{12–14,19,20} However, there is a paucity of data on the diagnostic role of CE in SBTB. The data are limited to the description of CE findings in case reports of patients with SBTB or reports of few cases of SBTB in studies describing the diagnostic role of CE in patients of obscure gastrointestinal bleeding in tropical countries.^21–23 The paucity of data may be attributed to non-affordability of CE in tropical countries and a majority of patients with SBTB already having stricturing disease that precludes the use of CE, ¹⁷ this figure being 50% in our study. CE is an expensive procedure and a majority of patients with TB in tropical countries are treated based on radiological and ileocolonoscopic findings rather than undergoing further evaluation by CE.

CE can visualise aphthous ulcers and early inflammatory lesions missed by other investigational modalities and therefore we hypothesised that it can help differentiating between SBTB and CD by detecting such changes in other segments of the small bowel in CD, a widespread inflammatory disease. In our study, we indeed found that CE detected asymptomatic involvement of the small bowel proximal to the terminal ileum in 33% of patients with CD, while none with SBTB had involvement of any other segment of the small bowel except the terminal ileum. Also, all CD patients had aphthous ulcers present on CE, while only 25% of SBTB patients had them. The majority of SBTB patients had large ulcers that were oriented in an oblique or transverse manner with irregular borders, as described previously.^21,22 However, the morphology of large mucosal ulcers was found to be similar in SBTB and CD.

We also found that involvement of the ICV was significantly more common in SBTB as compared to CD. This observation is in accordance with previous radiological studies in CD which typically involves the ileum with conspicuous sparing of the ICV.^18,24,25

The prospective study design and strict diagnostic criteria, including response to appropriate treatment, are important strengths of our study. However, its small sample size, especially of patients with TB, and our exclusion of large numbers of patients are important limitations. Nonetheless, our comparison of SBTB and CD has brought out important significant differences which will aid our diagnostic conundrum.