Article Text

This article has a correction. Please see:

Download PDFPDF

Long-term outcome of endoscopic dilatation in patients with Crohn's disease is not affected by disease activity or medical therapy
  1. G Van Assche2,
  2. C Thienpont2,
  3. A D'Hoore1,
  4. S Vermeire2,
  5. I Demedts2,
  6. R Bisschops2,
  7. G Coremans2,
  8. P Rutgeerts2
  1. 1Division of Abdominal Surgery, University Hospital Gasthuisberg, Leuven, Belgium
  2. 2Division of Gastroenterology, University Hospital Gasthuisberg, Leuven, Belgium
  1. Correspondence to Dr G Van Assche, Division of Gastroenterology. University Hospital Gasthuisberg, 49 Herestraat, 3000-Leuven, Belgium; gert.vanassche{at}


Background Endoscopic dilatation of Crohn's disease-related strictures is an alternative to surgical resection in selected patients. The influence of disease activity and concomitant medical therapy on long-term outcomes is largely unknown.

Aim and methods To study the long-term safety and efficacy of stricture dilatation in a single centre cohort.

Results Between 1995 and 2006, 237 dilatations where performed in 138 patients (mean age 50.6±13.4, 56% female) for a clinically obstructive stricture (<5 cm, 84% anastomotic). Immediate success of a first dilatation was 97% with a 5% serious complication rate. After a median follow-up of 5.8 years (IQR 3.0–8.4), recurrent obstructive symptoms led to a new dilatation in 46% or surgery in 24%. Niether elevated levels of C-reactive protein nor endoscopic disease activity predicted the need for new intervention. None of the concomitant therapies influenced the outcome.

Conclusion This largest series ever reported confirms that long term efficacy of endoscopic dilatation of Crohn's disease outweighs the complication risk. Neither active disease at the time of dilatation nor medical therapy afterwards predict recurrent dilatation or surgery.

  • Crohn's disease
  • endoscopic dilatation
  • complications
  • medical therapy
  • colonoscopy
  • inflammatory bowel disease

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Fibrotic strictures are a common complication in patients with Crohn's disease resulting in repeated surgery, which may ultimately result in chronic diarrhoea and short bowel syndrome. Strictures are a result of chronic transmural inflammation leading to tissue remodelling, mesenchymal cell hypertrophy, hyperplasia and fibrosis.1–4 Strictures can appear at any segment of the gastrointestinal tract involved in Crohn's disease but predominate in the terminal ileum, ileocolonic anastomosis and the rectum.5

In practice, a clinically relevant stricture is usually defined as a constant luminal narrowing with a pre-stenotic dilatation and obstructive symptoms. Estimating the ratio between the inflammatory and the fibrotic component of the wall thickening in a Crohn's disease stricture is impossible if classical barium imaging is used to investigate the origin of obstructive symptoms. On the other hand, endoscopy assesses the mucosa but provides no information on transmural inflammation or bowel wall thickening.

Despite advances in medical therapy with immunosuppressives including anti-tumour necrosis factor (TNF) agents, the need for surgery to treat fibrostenotic disease has not been eliminated. Ileal resection is a good treatment for symptomatic strictures, but disease recurrence is a certain outcome and after 15 years more than 50% of patients need a second resection.5 Therefore, bowel-conserving strategies including surgical stricturoplasty and endoscopic balloon dilatation have been developed as alternatives to resection. Uncontrolled observational studies indicate that endoscopic balloon dilatation is a safe and efficacious alternative to bowel resection in selected patients.7–18 In a recent meta-analysis of these trials short-term success rates were estimated at 78%.6 Long-term outcome of endoscopic balloon dilatation is less clear since in most cohorts follow-up was limited to less than 3 years and the individual cohorts report on a relatively small number of patients. More importantly, the influence of endoscopic and serological disease activity at the time of dilatation, and the influence of concomitant medical therapy on the outcome of stricture dilatation, is largely unknown.19 We therefore aimed to study the long-term safety and efficacy of endoscopic dilatation of Crohn's disease strictures in a large referral centre cohort incorporating the influence of disease activity and concomitant immunosuppression.


Patient identification and endoscopic procedure

Patients were identified by a search of our hospital central billing registry with the unique identifier of endoscopic dilatation as the search item. Retrospective chart analysis including case verification was performed to exclude dilatations for non-Crohn's disease indications. All these patients were also part of our VLECC database, which contains all patients with inflammatory bowel disease (IBD) followed at our hospital. Patients were only entered in the VLECC database after they had given informed consent. The VLECC register has been approved by the ethics committee of our hospital.

At our institution endoscopic dilatation is reserved for patients with obstructive symptoms including nausea, vomiting and abdominal cramps and a stricture less than 5 cm long, documented by a barium small bowel follow through or abdominal computed tomography (CT). This practice has not changed over the last 15 years. All dilatations were performed by regular colonoscopy, with a Through-The-Scope Boston Scientific Rigiflex® balloon (18 mm diameter, 8 cm long; Rigiflex, Cork, Ireland). The balloon was filled with water under visual control, and was insufflated by a multistep inflation with 2 min of inflation time at the maximal diameter (15–16.5–18 mm). Passage of the scope through the stricture was attempted in all patients and was used to define therapeutic success. An additional inflation of the balloon to 18 mm for 2 min. was performed if the first dilatation did not allow passage of the endoscope through the stenosis. Most of the procedures were performed under general anaesthesia. After dilatation patients were observed for at least 1 h or until resolution of post-procedure abdominal pain in our recovery room. In the case of persisting pain or abdominal distension, a plain abdominal x-ray was obtained and patients were hospitalised or referred for surgery in case of overt perforation.

Endpoints and statistical analysis

Long-term efficacy of endoscopic balloon dilatation, defined as no need for new dilatation or surgery in all patients undergoing their first dilatation, was chosen as the primary endpoint. Secondary endpoints included the impact of endoscopic disease activity, baseline CRP, and medical therapy on long-term re-dilatation rates or resection rates. Complication rates were recorded from the charts and classified as serious when they necessitated or prolonged hospitalisation, or when they resulted in surgery within 1 week from the procedure or resulted in death. Descriptive statistics were used to analyse patient demographics. For differences in proportions of patients experiencing a given outcome χ2 tests with correction for multiple comparison were applied. Kaplan–Meyer survival analysis with log rank statistics was used to assess event-free interval and Cox forward conditional proportional hazards regression analysis regression to assess predictors of recurrence including endoscopic and biochemical disease activity. For continuous variables such as age and disease duration, data were dichotomised using the median value as the cut-off. Kruskall–Wallis tests were used to compare lengths of follow-up.


Patient characteristics

Between March 1995 and May 2006, 237 dilatations were performed in 138 patients (77 women) with a mean age of 50.6±13.4. Data were extracted from the medical records until July 2008. For 96% of patients complete data were available and 4% (n=5) were lost to follow-up. During follow-up, four patients died (two suicides, one cholangiocarcinoma and one unknown). None of these were related to dilatation or subsequent surgery. Patients' demographics are listed in table 1. The majority of dilatations (84%) were performed at anastomotic sites and for ileal or ileocolonic disease locations (85%). Eighty-seven per cent of patients received medical therapy at the time of dilatation, and in 47% this included anti-metabolites or anti-TNF agents. Median follow-up was 5.8 years (IQR, 3.0–8.4).

Table 1

Patient demographics

Efficacy and safety of endoscopic dilatation

A high immediate success rate (defined as the ability to pass an adult colonoscope through the stricture after dilatation) of 97.1% or 134/138 was obtained for the first dilatation. Forty-four per cent (44.2%, 61/138) of patients remained dilatation-free and surgery-free until the end of follow-up (primary endpoint). Fifty-four per cent (54.3%) (74/138) of the patients included in the original cohort required no new dilatation for the entire duration of follow-up. Also, 76.1% (105/138) had no need for surgery after a first dilatation (figure 1).

Figure 1

Immediate and long-term success rates of the first and second endoscopic dilation in the full cohort of patients after a median follow-up 5.8 years.

A second dilatation was necessary in 56 out of the 138 patients. Of those 56 patients, 29 (51.8%) needed no further dilatation or surgery until the end of the follow-up (83.0% were surgery-free). Median time to new dilatation or surgery after the first dilatation was 12.5 months (IQR, 6–21.5) and after the second dilatation 8 months (IQR, 3–14.5; p<0.05).

For patients with anastomotic strictures (84%) absence of or mild endoscopic activity defined as a score of i0 or i1 (no more than five aphtous ulcerations at the site of anastomosis) on the Rutgeerts score18 for post-operative recurrence, was not associated with the need for re-intervention in Kaplan–Meyer analysis (figure 2a). Also, increased serum C-reactive protein (CRP; >5 mg/l) at the time of first dilatation was not associated with an early or overall increased need for dilatation or surgery (figure 2b). Other potential predictors, such as age, sex, and disease duration, were not inferring a significant risk of early recurrence (table 2).

Figure 2

a) Impact of endoscopic disease activity at the dilatation site on the redilatation or surgery rate. Defined by the Rutgeerts score. (b) Impact of C-reactive protein (CRP) at the moment of dilatation on the long-term re-dilatation or resection rates. ASA; 5-aminosalicylic acid; AZA, azathioprine; 6-MP, 6-mercaptopurine; MTX, XXXXX; TNF, tumour necrosis factor.

Table 2

Risks associated with potential predictors of endoscopic/surgical relapse after the first endoscopic dilation

The proportion of patients needing a new dilatation and/or surgery by the end of follow-up was 71.4% for those treated with 5-aminosalicylic acid (5-ASA), 44.2% for azathioprine, 47.8% for anti-TNF therapy, 50% for budesonide treated patients and 47.4% for patients without therapy (p=0.01 for 5-ASA vs other groups). The follow-up of patients treated with 5-ASA (median: 8.3 years (IQR, 6.7–9.4) was longer than that of the other patient groups (median 5.1 (IQR, 3.2–8.4), p<0.001) (figure 3). When expressed per patient year of follow-up, the proportion of patients needing dilatation and/or surgery was 11.3%/PY (patient year) for those treated with 5-ASA, 10.3% PY for azathioprine, 13.8%/PY for anti-TNF therapy, 9.8%/PY for budesonide treated patients and 6.0%/PY for untreated patients (all values are not significant). Also, in the Cox regression analysis none of the therapies were associated with an increased risk of relapse (table 2).

Figure 3

Influence of medical therapy after dilation on the proportion of patients experiencing surgical or endoscopic relapse. The median follow-up, in years, for every subpopulation is listed below the x-axis.

Of the 237 dilatations 12 resulted in a serious adverse event of which six were perforations. If all dilatations are considered the serious adverse event rate is 5.1%; in a per patient analysis (intention to treat) this rate rises to 8.7%. Surgery with intestinal resections was required in all perforations. Five out of six perforations were documented during the endoscopic procedure, and one was diagnosed within 1 h. One perforation occurred when a second ileal stricture proximal from the ileocolonic anastomosis was dilated, but otherwise the appearance of the strictures or the inflation procedure was not different in the patients with a perforation as compared to those without. All six surgeries were open procedures, involving limited ileocolonic resection with a new ileocolonic anastomosis in five and segmental ileal resection in one patient. One patient received a temporary, protective ileostomy. Within 3 months from surgery, 3/6 patients had a postoperative bleed requiring second-look laparatomy with splenectomy in one and segmental ileal resection in the other patient. Median postoperative hospital stay in the six patients was 11 days (IQR, 9–18). As a comparison the median hospital stay for patients with an elective open primary ileocecal resection at our institution is 7.5 days. Other serious adverse events included gastrointestinal bleeding (5/237) and abdominal pain requiring hospitalisation (1/237) (figure 4).

Figure 4

Serious adverse events, total and perforations after all 237 dilations. Data expressed per procedure (A) and per patient (B).


Strictures are a common cause of morbidity for patients with Crohn's disease and often occur at the site of the ileocolonic anastomosis.5 Bowel-conserving procedures have emerged as valuable tools to limit repeated intestinal resections and the incidence of short bowel syndrome. In this large cohort study we demonstrate that, in carefully selected patients, endoscopic stricture dilatation has a good long-term benefit to risk profile. Also our data indicate that disease activity assessed with endoscopic and serological criteria does not influence long-term outcomes.

Endoscopic balloon dilatation of Crohn's strictures is an alternative to bowel resection and can be performed at the same time as a diagnostic colonoscopy. In this series patients were highly selected and the vast majority had a stricture at the site of the ileocolonic anastomosis. Given the good long-term outcomes in our cohort, with only 24% of patients needing surgery with resection after a median of 5.8 years follow-up, this technique should be considered as an alternative in patients with a short stricture accessible with the endoscope. A surgical resection would not have resulted in a short bowel in most of these patients due to the expected limited resection, but 84% of patients in our series had already been operated on and would have required a new laparotomy if a surgical resection would have been opted for without an attempt to dilate endoscopically.

Several other reports of case series indicate a long-term benefit of endoscopic dilatation with high initial success rates and acceptable risks. Most of the reports available in the literature, however, have included a limited number of patients and follow-up has usually been limited to 3 years or less 7–18

All the studies, like our study, included data based on a retrospective chart analysis, without a control group. The number of patients varied between 17 and 59 with a mean follow-up per patient of 33 months.6 Most of the strictures were localised at the level of ileocolonic anastomosis and most were strictures shorter than 5 cm, as was the case in our larger cohort. In all the studies, obstructive signs were an essential criterion for dilatation. Dilatation techniques differed between the different studies and the immediate success rate varied between 73% and 100%. In a recent meta-analysis, it was found to be 78% if all studies were taken together.6 In all reported cohorts, as in our data analysis, most of the dilatations where performed in patients who had previously undergone surgery. In our study we performed a multistep inflation with a gradually increasing diameter, under visual control. This is possibly the reason for the lower complication rates in our study versus the important complication rates in the study by Cockuyt et al10 also from our centre, where larger balloons were still used. The risk of a severe adverse event associated with a dilatation in our hands was 5% (2.5% perforations). If expressed on an intention to treat basis the risk of a severe adverse event over the full extent of the follow-up was 8.7%.

In the systematic review, 58% of the patients where surgery free for the duration of follow-up. In our study we showed an excellent (97%) immediate success without need for repeat dilatation in 54% of patients or surgery in 76% of patients after a first dilatation. Consecutive dilatations followed more rapidly after a second dilatation but outcomes were generally comparable. Therefore, the benefit to risk ratio to avoid a laparotomy still favours an attempt to repeat endoscopic dilatation since the risk of perforation was not increased in consecutive procedures. We have identified our patients through a search of the central registration system of our hospital, and may therefore have overestimated the immediate success rate, since some unsuccessful dilatations may not have been billed. However, we feel that the long-term outcomes are more important. In our hands endoscopic dilatation was able to avoid the risks associated with surgery in the majority of patients, at least for the follow-up period of this study.

We did not explore accessory techniques to increase the efficacy of dilatation long term. Local steroid injection has been attempted, to improve the outcome of the dilatation in uncontrolled cohorts.8–12 16 17 However, a recent placebo controlled trial although with a limited number of patients, failed to show a benefit of local steroid injection on short and medium term recurrence rates after endoscopic dilatation for Crohn's disease strictures.20 We have never adopted the strategy of local steroid injection at out centre. Larger controlled studies exploring the added benefit of accessory techniques such as steroid injection or stenting are needed.

Sabaté et al7 found a higher recurrence rate for active smokers in their cohort of endoscopic dilatation, but only for evolution towards new surgery rather than for recurrent dilatation. Hoffman et al identified smoking and local ulceration as prognostic factors for an increased risk of recurrence.18 In contrast, in our cohort, we found no association between endoscopic activity, mainly determined by the size and number of ulcerations, and of an increased CRP on the need for surgery or new dilatation. Colombel et al21 have found that in a proportion of patients with transmural ileal Crohn's disease CRP levels correlated better with findings on CT enterography than with endoscopic lesions. However, in our study neither an elevated CRP nor endoscopic activity was associated with a worse outcome. Data on smoking status were only prospectively collected in our records after 2004 and were therefore not analysed.

We did find an apparent higher need for endoscopic or surgical re-intervention in patients treated with 5-ASA immediately after their first dilatation. However, given the scarce evidence supporting the use of 5-ASA in maintenance therapy for ileocecal Crohn's disease, we generally abandoned this therapy in the last 5 years. Therefore, patients treated with 5-ASA had a significantly longer follow-up. When we corrected for the patient years of follow-up, the difference between treatment groups disappeared. Also, in the multivariate Cox regression analysis 5-ASA treatment after dilatation was not associated with an increased risk of surgical or endoscopic relapse. Finally, this is a retrospective cohort study and the severity of the disease course may have influenced treatment decisions at the time of dilatation. Whether, the substantially higher relapse rate in the 5-ASA treatment group, heralds higher relapse rates with extended follow-up beyond the median of 6 years in the present study, needs to be investigated further.

In summary, our data demonstrate that the long-term benefit to risk ratio of endoscopic dilatation in selected patients with Crohn's-related strictures is clearly positive. Active inflammation at the site of the anastomosis and medical therapy following dilatation do not affect long term outcomes of endoscopic dilatation



  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the University of Leuven Hospitals.

  • Patient consent Obtained.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles