Crohn's disease is a granulomatous inflammatory disease characterized by transmural and segmental involvement of the intestinal wall [1]. Complications of Crohn's disease, such as extension to the adjacent mesentery and organs, fistulas, abscesses, and stenoses, are found in approximately 40% of these patients [2, 3]. Patients with inflammatory and some-times even obstructive small-bowel disease require prompt and accurate treatment to relieve their symptoms and to minimize the risk of potential complications [4].

Conventional small-bowel enteroclysis is currently regarded as the radiologic technique of choice in evaluating small-bowel disease. Radiographic findings of mucosal abnormalities of the terminal ileum and the distribution pattern throughout the small bowel are indicative of Crohn's disease [5-8].

Much interest has been focused on the use of cross-sectional imaging in patients with small-bowel disease because enteroclysis fails to show important extraintestinal manifestations of small-bowel disease [9-11]. Recently, CT enteroclysis was introduced as an alternative imaging method to overcome the individual deficiencies of CT (no distention of the small bowel) and conventional enteroclysis (no extraluminal information) and to combine the advantages of both in one technique. This method has been described as highly accurate in depicting mucosal abnormalities and extraintestinal complications in patients with Crohn's disease [12-16].

To our knowledge, the diagnostic value of CT enteroclysis has not yet been established in comparison with the current gold standard (i.e., conventional enteroclysis) in evaluating small-bowel disease [17, 18]. In this study, we prospectively compared CT enteroclysis and conventional enteroclysis with regard to the presence and extent of radiologic findings in patients with histopathologically proven Crohn's disease.

In this prospective study, 50 patients (26 women, 24 men; mean age, 36.3 years; age range, 18-52 years) underwent CT enteroclysis and conventional enteroclysis within a mean of 21.7 days (median, 18.5 days; minimum, 1 day; maximum, 47 days). The study was approved by the institutional review board, and written informed consent was obtained from all patients. All patients had histopathologically proven Crohn's disease and were evaluated during a symptomatic stage of their disease. Fifteen patients (30%) previously underwent gastrointestinal surgery, with 13 patients (26%) having undergone ileocecal resection and two patients (4%) having undergone segmental small-bowel resection. The 35 remaining patients (70%) were referred from the gastroenterology department, with clinical symptoms of Crohn's disease, for evaluation of the extent of their disease.

For both investigations, after local anesthesia was established using a lidocaine hydrochloride 2% gel (Xylocaine, AstraZeneca), a nasogastric tube was advanced beyond the duodenojejunal junction under fluoroscopic guidance. Enteroclysis was performed according to the technique established by Herlinger [6] and Maglinte et al. [11]. A standard amount of barium (300 mL) and 0.5% methylcellulose solution (1,500 mL) was infused through the nasogastric tube positioned at the duodenojejunal flexure, achieving optimal double-contrast and small-bowel distention. Standardized compression views were obtained in all patients for evaluation of the small bowel, especially of the terminal ileum.

CT enteroclysis was performed in a standardized fashion: 1,500 mL of 0.5% methylcellulose solution was administered via a roller pump at an injection rate of 150 mL/min to provide adequate and uniform small-bowel distention. Patients who had previous small-bowel resection received up to 1,300 mL of methylcellulose solution. The methylcellulose infusion was stopped when patients reported any kind of discomfort during the instillation. Immediately after the administration of the designated or tolerated volume of methylcellulose solution, patients were transferred from the fluoroscopy suite to the CT unit, at which point 20 mg of hyoscine-N-butylbromide (Buscopan, Boehringer Ingelheim) was injected IV for minimization of potential artifacts by peristaltic bowel movement. An additional 500 mL of contrast material was administered while the patient was lying on the CT table at a flow rate of 80 mL/min to maintain optimal distention of the upper small jejunal loops during CT examination.

CT scans were obtained on a Tomoscan SR 8000 scanner (Philips Medical Systems) using the following parameters: 5-mm collimation, 10-mm table speed, and a reconstruction increment of 3 mm. Scanning was started 25 sec after IV application of 120 mL of nonionic contrast medium (300 mg I/mL iopamidol [Iopamiron, Bracco]), with a flow of 4 mL/sec during one breath-hold, from the dome of the liver to the lower margin of the symphysis pubis.

Conventional enteroclysis and CT enteroclysis images were analyzed in random order and were interpreted by two experienced gastrointestinal radiologists in consensus. Radiologists were blinded to patient data to ensure objective interpretation of the respective imaging technique. Hard-copy images of enteroclysis and CT enteroclysis were mounted on a light box with the appropriate collimation capability.

Enteroclysis and CT enteroclysis were specifically analyzed for the findings listed in Table 1. Mucosal abnormalities, such as thickening of mucosal folds and erosive or aphthous lesions on enteroclysis and bowel wall thickening and increased segmental contrast enhancement on CT enteroclysis, were rated as mucosal changes. Wall thickening was diagnosed when the bowel wall measured at least 3 mm on an adequately distended loop [19]. Increased enhancement was assessed subjectively by comparing the suspected segment with unaffected small-bowel loops. Stenosis was characterized as a lack of bowel loop distention below a minimum diameter of 1.5 cm, and the criterion for the diagnosis of prestenotic dilatation was a diameter exceeding 2.5 cm or the presence of segments wider than more proximal bowel loops. A fistula was defined as abnormal communication between two epithelial surfaces or from the bowel wall to the skin (Figs. 1A, 1B, and 1C). Abscesses were diagnosed as a circumscribed, round, or oval fluid collection with a contrast-enhancing wall. Skip lesions were defined as additional segmental inflammatory small-bowel abnormalities in locations other than the terminal ileum (Figs. 2A, 2B, 2C, 2D, 2E, 3A, 3B, 3C, 3D, and 3E). Additional abnormalities included mesenteric lymphadenopathy and conglomeration of bowel loops, defined as a clump of small-bowel loops, matted due to transmural inflammatory changes.

Both methods were compared with regard to the assessment of disease extent and distribution. Conventional enteroclysis findings served as the gold standard. Results were given as means ± SD. Standard descriptive statistics, a binominal analysis, and a McNemar test were used [20].

Placement of the nasogastric tube was successful in all patients. The mean fluoroscopy time for placing the tube was 4.1 min; there were no failed tube placements.

Conventional enteroclysis was successfully performed in all 50 patients (100%). Crohn's disease-associated radiographic changes could be found in 42 patients (84%). In the remaining eight patients (16%), no radiographic abnormalities were seen at enteroclysis, although in two of these patients (4%), double-contrast imaging was not considered sufficient for an exact radiologic diagnosis because of constant superposition of small-bowel loops in the region of the terminal ileum and the pelvis.

CT enteroclysis also was successfully performed in all 50 patients (100%), and all investigations were considered diagnostic with sufficient small-bowel distention. Crohn's disease-associated abnormalities were found in 44 patients (88%), and the remaining six patients (12%) presented with normal CT enteroclysis findings. The total number of diagnosed abnormalities was significantly higher with CT enteroclysis (p < 0.01).

Both imaging methods showed small-bowel stenosis in 34 patients (68%), and prestenotic dilatation was diagnosed with enteroclysis in 15 patients (30%) and with CT enteroclysis in 20 patients (40%). Fistulas were seen in eight patients (16%) using enteroclysis and in 18 patients (36%) using CT enteroclysis. Three patients (6%) showed skip lesions with conventional enteroclysis, and using CT enteroclysis, reviewers could identify skip lesions in 17 patients (34%). Conglomerate tumors were diagnosed with enteroclysis in three patients (6%) and with CT enteroclysis in 13 patients (26%). CT enteroclysis showed abscesses in eight patients (16%) and mesenteric lymphadenopathy in 26 patients (52%). Neither of these abnormalities could be identified in any patient with conventional enteroclysis (0%).

In the two patients for whom enteroclysis was considered of insufficient quality, CT enteroclysis was performed successfully and showed minimal changes in one patient and an ileal stenosis with prestenotic dilatation in the other.

The number of detected fistulas, abscesses, skip lesions, conglomeration of bowel loops, and lymphadenopathy was significantly higher with CT enteroclysis (p < 0.01).

For both imaging techniques, no procedure-related complications were observed.

Enteroclysis has been considered the gold standard imaging method for evaluation of the small bowel in patients with Crohn's disease. The results of this study clearly show the superiority of CT enteroclysis over conventional enteroclysis in the evaluation of patients with Crohn's disease. Although there was no significant difference in characterizing mucosal and mural small-bowel abnormalities between these two imaging methods, CT enteroclysis provided important information about the extraintestinal extension of the disease—in particular, by depicting more fistulas and abscesses. In addition, the detection rate of skip lesions (17 with CT enteroclysis vs three with enteroclysis) and the detection rate of conglomeration of bowel loops (13 vs three, respectively) were significantly higher using CT enteroclysis. Identification of these two known complications of Crohn's disease is crucial for initiation of further therapy. Patients who develop abscesses, conglomeration of bowel loops, and skip lesions are often candidates for elective gastrointestinal surgery. Considering the consequence of these Crohn's disease-associated complications on a patient's further management again emphasizes the diagnostic value of CT enteroclysis in the assessment of the extent of disease and in the detection of extraluminal complications in patients with Crohn's disease.

To date, enteroclysis has been the diagnostic gold standard for evaluating patients with small-bowel Crohn's disease [2]. It is a sensitive technique in detecting and characterizing small-bowel abnormalities, especially minimal mucosal changes, that are well depicted by enteroclysis [21, 22]. In the current study, both imaging methods showed minor inflammatory changes equally well, and CT enteroclysis depicted even more cases, but this was basically due to individual anatomy (overlying bowel loops) rather than technical factors. In this regard, it should be mentioned that all patients included in this study had well-known and documented Crohn's disease for several years, and none of our patients had extremely early mucosal changes—that is, lymphoid hyperplasia or scattered aphthous lesions. In these cases, conventional enteroclysis might still be superior, but with the widespread use of ileoendoscopy, we rarely see these patients anymore in our radiology department.

The value of enteroclysis has been questioned by some authors who assert that conventional small-bowel follow-through is superior to enteroclysis [23]. Those results might be explained by the fact that the failure rate for placement of the nasogastric tube was 20% in that study [23]. In the present study, the nasogastric tube could be placed in all patients undergoing CT and conventional enteroclysis. A weak point of the present study might be the mean fluoroscopy time of 4.1 min for tube placement. All examinations were performed in a teaching hospital, and nasogastric tube placement was, in most patients, performed by radiology residents under the supervision of a board-certified radiologist.

In experienced hands, enteroclysis has been considered superior to conventional small-bowel follow-through. Enteroclysis produces better delineation of the entire small bowel than follow-through study. Not only does enteroclysis produce better distention of individual small-bowel loops, but the biphasic approach allows all loops, including the distal ileum, to be visualized on both single- and double-contrast studies [24].

CT enteroclysis was introduced to overcome the individual deficiencies of both CT and conventional enteroclysis and to combine the advantages of both in a single technique [25].

The feasibility and accuracy of CT enteroclysis in patients with Crohn's disease have been well documented in the literature [11-17, 23]. This technique enables detailed morphologic characterization of inflamed bowel segments and reveals mural and extramural manifestations of enteroclysis. Entities, such as fistulas, abscesses, or skip lesions, that emphasize clearly the need for elective gastrointestinal surgery can easily be detected.

Rollandi et al. [18] evaluated CT enteroclysis versus enteroclysis only in patients in an advanced stage of disease, and they concluded that CT enteroclysis should be used as a complementary imaging method to conventional enteroclysis. They did not include patients with minimal inflammatory changes and suggested additional studies.

There are some limitations to our study. First, there is a selection bias: Only a small percentage of patients with early stage disease could be included in this study.

Patients were referred to our institution from a gastroenterologic center specializing in Crohn's disease, and these patients previously underwent detailed clinical and serologic examination and ileoendoscopy or endoscopy, which explains the high percentage of patients with positive Crohn's disease-associated findings. The remaining patients without small-bowel findings on enteroclysis and CT enteroclysis had proven involvement of the colon. Because of the high prevalence of disease in our study population, the sensitivity and specificity rates are skewed.

Another limitation might be the mean interval of 21.7 days between the two examinations. The natural course of this particular disease is chronic and recurrent; thus, the status of the inflammation might have been changed within the 3 weeks—for better or worse.

CT scans were obtained on a single-detector scanner with a lowest possible collimation of 5 mm and were analyzed in the axial plane only. Coordinated reformatting in the coronal plane was not performed and must be stated as a limitation of this study. MDCT allows thinner collimation—for example, 4 × 1 mm, and multiplanar reconstructions allow better detection and localization of Crohn's disease-associated intestinal changes.

Both imaging methods require the instillation of a distention medium via a nasoenteric tube. Adequate distention of the small-bowel loops is crucial, especially for depicting minor inflammatory changes. This has been shown in several studies [18, 26]; without bowel distention, image quality and consequent results are considerably poorer, and the value of such studies with regard to establishing the true diagnostic value of this method is limited [17].

In conclusion, CT enteroclysis proved to be significantly superior compared with conventional enteroclysis in depicting Crohn's disease-associated intra- and extraluminal abnormalities. CT enteroclysis is the imaging method of choice and should replace conventional enteroclysis in patients with known Crohn's disease.

Address correspondence to J. Sailer ([email protected]).