Crohn’s disease (CD) is a chronic inflammatory disease which involves in about 70% of cases the small bowel, alone or in association with other segments of the gastrointestinal tract, mainly the colon [1, 2]. Evaluation of the small bowel in CD may be performed for several reasons: in the initial assessment of disease, in the differential diagnosis between ulcerative colitis and CD, in the diagnosis of complications such as strictures and fistulas, and, as suggested recently, in monitoring the therapeutic response to therapy or in assessing postsurgical relapse [3, 4].

Small bowel evaluation is notoriously difficult, especially when the lesions are subtle and minute and are located deep in the small bowel, beyond the reach of standard endoscopy. Small bowel enteroclysis or follow-through has been demonstrated to be useful in detecting advanced lesions and complications such as strictures and fistulas, but has a very low sensitivity in recognizing early mucosal changes of CD [5]; similarly, the more recently introduced cross-sectional imaging methods, i.e., magnetic resonance and computed tomography enteroclysis, are useful for indicating transmural inflammation and fistulas, but also show poor sensitivity to demonstrate early lesions of disease [6].

Capsule endoscopy (CE) has been proposed as a first-line test for visualizing the mucosa of the small intestine: CE is useful for depicting early mucosal changes in CD, being superior to any other imaging techniques [7], but great concern exists with regard to capsule retention because of intestinal strictures [8]. Moreover, uncertainties exist regarding the clinical implications of the lesions depicted by CE in small bowel. These doubts are similar to those existing for the endoscopic assessment of the colon in CD: the clinical utility of assessing the extent and severity of endoscopic disease in CD is still a matter of debate due to interobserver variability and because of a poor correlation between the mucosal findings and the clinical features in a disease characterized mainly by transmural inflammation [9–11].

Double-balloon enteroscopy (DBE), introduced in 2004, has made possible the endoscopic examination of the entire small bowel; moreover, biopsies and therapeutic interventions are possible during DBE [12]. CD lesions are found in about 12% of patients undergoing DBE for different clinical reasons [13]; however, only a few studies have examined specifically the use of DBE in CD patients, and the available data suggest the superiority of DBE in comparison with barium studies in detecting aphthae, erosions, and ulcers in the ileum [14, 15].

DBE is nowadays routinely used in the management of patients with CD, but its use is likely to be empiric since no data exist with regard to the indications for the procedure, as well as to the accuracy and the diagnostic yield of DBE in the different clinical pictures of CD. Other unanswered questions are whether and how DBE could alter the management of CD, and what the role of DBE should be in the diagnostic approach to patients with CD relative to the other procedures that are available today.

The aim of the present study is to determine the impact of utilizing DBE in the management of suspected or established CD.

Patients and methods

From October 2006 to June 2008, 165 DBEs were performed in 130 patients observed in the Gastroenterology Department of the University Hospital “Luigi Sacco”, Milan, a third-level center where about 3,500 ambulatory patients with IBD are cared. For the purpose of the present study we prospectively considered all patients undergoing DBE with an indication related to suspected or established CD:

  1. 1.

    To make the initial diagnosis in patients with signs or symptoms suggestive of CD;

  2. 2.

    To distinguish between CD and ulcerative colitis in patients with colitis;

  3. 3.

    To evaluate disease extent and activity in uninvestigated CD patients

  4. 4.

    To define the cause of anemia or of overt gastrointestinal bleeding in patients with obscure bleeding;

  5. 5.

    To diagnose a stricture or to define its nature;

  6. 6.

    In case of symptoms suggesting complications or malignancy;

  7. 7.

    To evaluate postsurgical recurrence.

All these clinical problems indicated the use of DBE when the conventional endoscopic and radiological methods failed to solve them.

Pre-enteroscopy assessment

The following data were recorded before DBE by interviewing the patients and analyzing the patient’s chart: age, sex, extension of the disease if already known, indication for DBE, and previous endoscopic and radiologic examinations performed with the same indication and their findings.

Double-balloon enteroscopy characteristics

The double-balloon enteroscope used in the present study consists of a high-resolution endoscope with a working length of 200 cm and a working channel 2.2 mm in diameter, and a flexible overtube with a length of 145 cm (Fujinon EN-450 P5/P20, Fujinon Inc., Japan). The principle of the double-balloon technique has been already described elsewhere [12], as well as the method to measure the insertion depth [16].

For the oral approach with DBE no specific preparation was prescribed, but bowel cleansing as in colonoscopy was indicated for the anal route. The insertion route was chosen according to the estimated location of the suspected lesion, mainly based on the results of the procedures performed previously to visualize the small bowel. If no data were available, patient’s clinical presentation was the basis for starting DBE from either approach; if a diagnosis was not obtained with the primary access route, the alternative route was used. Stopping of an approach from one side was based on the identification of a diagnostic lesion, or if further advancement of the enteroscope was not possible due to technical difficulties.

The examinations were carried out with the patients under conscious sedation; general anesthesia was used in only two patients: a 13-year-old girl with suspicion of CD and a 17-year-old boy with known CD and a suspected ileal stricture.

All of the patients provided written consent to undergo DBE, after having received extensive information regarding the nature of the procedure. The study was approved by the local institutional review board.

Study design

The study was designed with the aim of assessing the following features of DBE performed in patients with CD: (1) feasibility and success rate, (2) ability of DBE to detect small bowel lesions in comparison with other imaging procedures, and (3) diagnostic yield of DBE in comparison with the other procedures performed with the same indication.

According to the aim of the study, a diagnosis was regarded as relevant if able to influence the management of the disease or to give information regarding disease characteristics such as disease extension, activity, and complications, as follows:

  1. (a)

    Altering the medical therapy of CD;

  2. (b)

    Indicating further testing or operative procedures not otherwise hypothesized before endoscopy;

  3. (c)

    Offering new findings which modified the management of the disease: a change in the initial diagnosis or in previously reported extension of the disease; a different disease activity than otherwise hypothesized; the presence of complications such as strictures or fistulas;

  4. (d)

    Achieving a diagnosis of malignancy or dysplasia.

Descriptive statistics consisted of t tests for continuous variable and χ 2 tests for categorical variables. Statistical significance was defined as p < 0.05.

Results

In the study period 50 DBEs were performed in 37 patients (18 males and 19 females; mean age 42 years, range 13–77 years) with known or suspected CD. Demographic and clinical characteristics of patients according to the indication to DBE are reported in Table 1.

Table 1 Demographic and clinical characteristics of the study patients. Patients are grouped according to the indication for double-balloon enteroscopy (see also the text)
Full size table

All considered patients have previously undergone upper and lower gastrointestinal endoscopy and intestinal ultrasonography. Twenty-eight further investigations were performed with the same indication as DBE: 12 small bowel follow-through, 7 computed tomography enteroclysis, and 9 CE (the capsule was retained at the level of a stenosis in one patient). Overall, an average of 3.7 procedures have been performed in each patient prior to perform DBE.

DBE technical performance

Of the 50 procedures, 32 were performed from the oral route and 18 from the anal route (5 in patients with an ileo-colon anastomosis); in 6 patients DBE was performed from both routes, obtaining a complete exploration of the small bowel in 4 (10.8% of the whole series). Endoscopy was repeated in seven cases due to either the impossibility to intubate the ileocecal valve (four cases) or to visualize the suspected lesion (three cases). No complication occurred. Access to small bowel from either route was possible in all patients but one, a patient with a colonic disease and an ileocecal valve stenosis, in whom DBE was performed only from the anal route under suspicion of a terminal ileum involvement. The mean length of endoscope insertion was 266.5 ± 100 cm beyond the Treitz’s ligament from the oral route and 72.5 ± 60 cm beyond the colon from the anal route. The ileocecal valve intubation was possible in 8 of 13 patients (61.5%), but in 4 of 8 patients DBE was able to explore less than 50 cm of the ileum, due to the presence of stenosis in 3 patients and excessive looping of the endoscope in 1 patient.

Diagnostic yield of DBE

Access to small intestine with DBE was possible in 36 of 37 patients. In 18 patients (50%) DBE did not find any lesions. The negative finding was considered relevant in nine cases: in four patients DBE allowed to exclude a small bowel involvement of CD which was suspected on the basis of mucosal abnormalities detected by computed tomography enteroclysis in one, small bowel follow-through in one, and CE in two patients; in three patients DBE excluded the presence of a stenosis suspected by small bowel follow-through; in two further patients DBE excluded a malignancy suspected on the basis of clinical symptoms. In the remaining nine patients, four patients with anemia, four with suspected small bowel lesions, and one who had recently undergone an ileum resection, the negative finding was considered a false-negative result: in the four patients with anemia and negative upper and lower gastrointestinal endoscopy DBE did not find any lesions; two of these patients had undergone CE, which was negative in one case and able to identify blood in the proximal ileum at the level of a surgical anastomosis in the other; the four patients with suspected CD involvement of the small bowel have previously undergone small bowel follow-through (three patients) and computed tomography enteroclysis (one patient) showing minimal changes of the mucosa; also in these patients DBE was negative; in the remaining patient DBE was performed to assess recurrence of disease at the level of a surgical anastomosis but was not able to reach it.

DBE revealed at least one small bowel lesion in 18 patients: ulcers in 5 patients, aphthae in 12, strictures in 5, a nonbleeding angiodysplasia in 1, a bleeding stricturoplasty in 1, and a capsule retained at the level of a stricture in 1. In 13 of these patients the endoscopic diagnosis was judged to be clinically relevant: in 4 patients DBE demonstrated the cause of bleeding or anemia, one hemorrhagic oozing stricturoplasty and 3 ulcers; in 4 it revealed a stenosis and/or permitted the histological definition of its nature; in 4 it confirmed the suspected diagnosis of CD and/or defined the extension of the small bowel involvement, recognizing and characterizing histologically different lesions such as aphthae, ulcer, or flogosis; in 1 patient the recurrence of disease at the level of an anastomosis was recognized. In five patients the DBE findings did not affect the management of disease: two patients with anemia in whom DBE was not able to find the bleeding lesion, and three patients in whom DBE was not able to define the extension of small bowel involvement due to the incomplete exploration of the organ.

Argon plasma coagulation was performed in three patients. In one patient with a stenosis and a retained capsule DBE reached the stenosis but was unable to remove the capsule. The capsule was removed surgically some days after.

The overall diagnostic yield of DBE in CD was 59.4% (22/37 patients). The diagnostic yield of DBE was different according to the indication for the procedure, being significantly higher when performed to detect a stricture (7/7, 100%) then a bleeding lesion (4/10, 40%) or to diagnose or stage the small bowel involvement of disease (8/16, 50%) (Fig. 1). Interestingly, the majority of DBEs (17/22, 77.8%) achieving a relevant diagnosis were performed on the basis of one or more previous investigations able to identify a lesion and then to determine the proper DBE introduction route; in comparison, 60% (9/15) of DBEs that did not achieve a relevant diagnosis were performed without having information on the location of the presumptive lesion [odds ratio (OR)5.1, 95% confidence interval (CI)1.2–21.4, p = 0.037].

Fig. 1
figure 1

Diagnostic yield of double-balloon enteroscopy according to the indication for the procedure and in all patients

Full size image

CE was performed in nine patients prior to perform DBE and in eight cases it demonstrated at least one lesion (88.9%). In comparison, only 18 of 37 DBEs (48.6%) were able to demonstrate at least one lesion (p = 0.05). DBE was able to reach the lesions visualized by CE only in four of eight patients (50%). In a fifth patient where CE demonstrated the presence of aphthae in the mid jejunum, DBE reached the presumptive site of the lesion without revealing any abnormalities. We have considered this a false-positive result of CE. Nevertheless, while CE visualized the whole small bowel in 8 of 9 patients (88.9%), this was possible in only 4 of 37 patients with DBE (10.8%, p < 0.001).

Discussion

The usefulness of endoscopy in the management of patients with CD has been always a matter of debate. There are several reasons for this: only the colonic and terminal ileum involvement of CD can be assessed by means of conventional endoscopy; symptoms may correlate poorly with endoscopic activity; an extreme interobserver variability has been reported in the endoscopic assessment of mucosal lesions; and therapies such as steroids usually do not determine a significant mucosal healing and are unlikely to influence the prognosis of disease [9–11]. The consequence is that clinical and laboratory indices have been considered more suitable than endoscopy to monitor the disease activity and course. In the last years, however, two facts have significantly changed the role of endoscopy in the evaluation of patients with CD: first, the recent introduction of biologic therapy for CD, potentially able to obtain a complete mucosal healing and thus to change the clinical history of the disease, makes endoscopy a possibly necessary tool to assess the mucosal response to therapy [17, 18]; second, the introduction of CE and, more recently, of DBE makes the complete visualization of the small bowel potentially possible.

CE is usually considered a precise and at the same time easy and well-accepted method to visualize the entire small bowel mucosa; great concern, however, exists with regard to the risk of capsule retention in CD [7, 8]; moreover, CE does not permit to take biopsies to define histopathologically the nature and clinical significance of the depicted lesions. From a theoretical point of view DBE is able to overcome all the limitations of CE and, because of this, is likely to be of great utility in the assessment of patients with CD. Only a few data, however, exist on the use of DBE in CD, so that we still have only sparse information on its feasibility and safety, as well as on the impact of the procedure on the clinical management of the disease. Oshitani et al. [14]. studied by means of retrograde DBE 40 patients with CD. They found involvement of the terminal ileum in 7 patients and ileal involvement proximal to the terminal ileum in 27 patients. Interestingly, 24 of these 27 patients showed no involvement of the terminal ileum and thus CD could not be diagnosed by conventional ileocolonoscopy. According to the data, DBE was superior to radiological study to detect aphthae, erosions, and small ulcers in the ileum. Seiderer et al. [15] compared the efficacy of DBE and magnetic resonance enteroclysis in detecting different small bowel lesions in a sample of ten patients with CD. In both of these studies DBE was performed in selected groups of patients with selected indications which are unlikely to reproduce the whole spectrum of the clinical use of DBE in patients with CD. In our study, we have considered a number of different clinical conditions, typical of the polymorph clinical pictures of CD, which may indicate to perform DBE from either anal or oral route.

The first information arising from our CD series is that DBE is safe, but that it could be more technically demanding than for other indications, especially when the procedure is performed by the anal route: the rate of ileocecal valve intubation was about 50%, significantly lower than the 80% rate reported in the literature by some authors for other indications [12, 19, 20]. The main reason for this is the frequent occurrence of ileocecal valve stenosis in CD. Nonetheless, once the ileocecal valve is intubated, the progression of the endoscope in the ileum may be difficult and frequently limited by the presence of stenosis and adhesions.

The second information is the diagnostic yield of DBE in patients with CD, which resulted to be about 50%, similar to that reported in the literature for other indications, mainly obscure bleeding [21]. Interestingly, the diagnostic yield of DB was higher in patients where previous investigations were able to recognize a lesion. In these patients the insertion route of the endoscope was decided on the basis of this information. Notably, the lowest diagnostic yield was observed in patients undergoing DBE for the definition of the cause of a bleeding: in these patients the bleeding lesions are likely to be subtle and not recognized by radiologic procedures; the highest diagnostic yield was conversely registered for patients with strictures which are relatively easy to detect by enteroclysis and cross-sectional imaging methods. This observation has two important corollaries: First, in CD DBE is likely to be highly efficient when performed with the correct indication and on the basis of correct information; second, also in CD the major limit of DBE is the frequent incomplete visualization of the small bowel and the consequent high rate of false-negative results [21].

To overcome the respective limits of CE and DBE the integrated use of the two procedures has been proposed in patients with obscure bleeding [22]: CE should be performed immediately after a negative upper and lower endoscopy as a screening method; results of CE should guide the use of DBE which aims either at the confirmation or at the treatment of the lesions detected by CE. In CD such an approach is not always possible due the frequent occurrence of small bowel stenosis, and even using the patency capsule a significant number of patients would not be suitable to undergo CE [23]. In CD, the use of DBE frequently follows other procedures such as ultrasonography, enteroclysis, and cross-sectional imaging methods. These methods are very precise in the detection of advanced lesions but show poor sensitivity in recognizing early lesions of disease [5, 6]. It is thus possible that, in a significant number of CD patients, i.e., patients with a suspicion of CD or with obscure bleeding, DBE could be theoretically performed as a first-line diagnostic approach.

In conclusion, in the present study we have shown that DBE, although technically demanding, is likely to be a feasible, safe, and in some selected cases useful method in the assessment of patients with suspected or diagnosed CD. Efficacy of DBE is, however, strictly related to the clinical context in which DBE is performed, as well as to the procedures which are performed before DBE and the information that these procedures offer. Larger series are needed to better determine the exact collocation of DBE in the diagnostic and therapeutic flow charts of the different clinical questions that CD presents to the gastroenterologists.

What is current knowledge

  • Endoscopic assessment of the entire small bowel in Crohn’s disease is potentially possible by means of capsule endoscopy and double-balloon enteroscopy.

  • Double-balloon enteroscopy is likely to be more sensitive than current imaging methods in recognizing aphthae, erosions, and ulcers in the ileum in Crohn’s disease.

What is new here

  • Double-balloon enteroscopy is safe and feasible, but technically demanding in Crohn’s disease.

  • Double-balloon enteroscopy has a diagnostic yield of about 50%, but its efficacy is strictly related to the clinical context in which DBE is performed, as well as to the procedures which are performed before DBE and the information that these procedures offer.