1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Background  Some reports have suggested that infliximab may induce obstructive symptoms and, although there is no firm evidence, it is usually contra-indicated in-patients with Crohn’s disease (CD) and strictures.

Aims  To evaluate the effect of infliximab on symptomatic strictures of the small intestine in CD and to identify predictive factors of clinical response.

Methods  This retrospective study included symptomatic patients treated with infliximab after conventional treatment had failed. The short-term (week 8) and long-term results were classified according to predefined criteria as complete, partial response, or failure.

Results  Before infliximab, 18 patients had complete obstruction or intermittent chronic abdominal pain. Fourteen patients were treated by corticosteroids and 13 received immunosuppressive drugs. At week 8, complete, partial response and failure were observed in 10, 7 and 1 patients, respectively. Fourteen patients continued maintenance infliximab treatment after week 8. During the most recent evaluation (median follow-up: 18 months), 8 patients were on maintenance infliximab treatment; only eight were still on prednisone; there were five complete responses, 10 partial responses and three failures. Initiating prednisone or increasing its dosage was the only factor associated with a short-term complete response.

Conclusions  Infliximab may be effective in patients with symptomatic strictures from CD, and should be tested before considering surgery.


  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Seventy per cent of patients with Crohn’s disease (CD) have small bowel lesions (CESAME cohort).1 Strictures are a common complication and occur most often in the ileum or on surgical anastomoses.2–4 Studies have suggested that CD spontaneously progresses from inflammation to stenosis over a long period.2, 5 Factors determining the development of strictures are not fully understood, but chronic and trans-mural inflammation probably plays a major role.6–8 Despite advances in the treatment of CD, the incidence of strictures and surgical resection has not changed in the last 25 years.2 Although surgery is the standard treatment for symptomatic stenosis, it does not prevent anatomic or symptomatic relapses (40% at 4 years for symptoms).9, 10 There are also functional sequellae after large resections. Endoscopic dilatation may be suggested for short symptomatic strictures, but the efficacy is of a limited duration.11

Infliximab has changed the management of inflammatory CD and decreased the need for surgery in fistulae.12, 13 However, because of reports of complete obstruction after infliximab in-patients with or without initial stenosis, it is contra-indicated in stenotic forms of CD.3, 14 The theory is that rapid infliximab-induced tissue healing may result in marked architectural changes in the intestinal wall, in particular in excess scar tissue leading to stenosis. However, stenosis does not occur without inflammation and chronic inflammation per se may lead to strictures.7 So the effect of infliximab on stenosis cannot be predicted on a theoretical basis because of its anti-inflammatory and healing effect.

The aim of this study was to determine whether infliximab is a valid option in-patients with obstructive symptoms caused by ileal stenosis in CD. We also identified predictive factors of response.

Patients and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Patients included in the study had ileal CD with clinically symptomatic strictures. CD was defined according to the clinical history, typical radiological and endoscopic features.

Obstructive symptoms were either abdominal cramps induced or increased by food, gurgling during pain, resolved by gas or stools. Spontaneously self limited obstruction was defined as an obstructive episode, which was resolved by alimentary restriction without hospitalization. Complete obstruction required emergency hospitalization.

Strictures were defined by constant luminal stenosis as assessed by small bowel follow-up, which was chosen as the standard approach.15, 16 Tight stenosis was defined by the ring sign with a constant diameter on each negative, the presence of a dilated small bowel and fluid level was also noted; loose stenosis was defined as a constant but variable decrease in the loop diameter. The number, length, calibre and ulcerations in each stenosis were assessed as well as bowel distension above the stenosis. C reactive protein (CRP) was measured the week before inclusion, during the first evaluation and the last evaluation.

Demographics (age, gender) and disease characteristics (disease location, duration, history of intestinal surgery, hospitalization and other CD drug medications) were noted for each patient. Infliximab was prescribed in-patients with obstructive symptoms caused by at least one ileal stenosis. Complications requiring short term surgery such as peritonitis or abscesses were excluded before the first injection. Infliximab (5 mg/kg) was administered at day 0 and weeks 2 and 6. Associated treatment was not standardized and was administered on a case by case basis for clinical reasons based on the physician’s decision. Prednisone was decreased after week 4, if there was clinical response. Assessment of short-term clinical efficacy was performed at week 8.

Efficacy at week 8 was classified into three categories. Complete response was the total disappearance of symptoms with no abdominal cramps in the last 15 days and for the last evaluation; prednisone was also not required in these patients. Partial response was defined according to the initial symptoms as either the disappearance of complete obstruction with persistent intermittent obstructive symptoms or less frequent and less severe abdominal pains. Failure was defined as persistent or recurrent obstruction that was as severe as initially, and/or an indication for surgery within 15 days before evaluation.

Further management of patients after week 8 was decided as needed in relation to: stopping or prolonging infliximab, associated treatment or surgery. A second evaluation, including for symptoms of intestinal obstruction and treatment, was performed during the last visit in each patient. Assessment of the efficacy of infliximab was based on the same criteria as those used at week 8 (complete, partial response or failure).

Eighteen patients were included (11 males, 7 females). The median age at the first symptoms of CD was 23.5 years (range: 14–41), age at the first infliximab infusion was 30.5 years (range: 24–66).

The small bowel location of CD was as follows: ileocolonic (n = 9), jejuno-ileal (n = 5), ileal (n = 3) and extensive disease (n = 1). Some patients had had complications before the study: abscess (n = 4), fistula (n = 4) and severe haemorrhage (n = 5). Four patients had had one bowel resection, five 2 and one 3. Types of obstructive symptoms before inclusion are provided in Table 1; results of small bowel follow-up in Table 2. The median time between small bowel follow-up and inclusion was 2 months (0–24). The results of CRP levels are found in Table 3.

Table 1.   Type and duration of obstructive symptoms at day 0. The same patient could have different symptoms (n > 18)
 nDuration/D0 months (median; range)
  1. * Between day 0 and the last episode of complete obstruction.

Crampy abdominal pain1312 (2–120)
Spontaneously self limited obstruction 913 (0.3–72)
Complete obstruction 42.25 (0.3–2.4)*
Table 2.   Barium contrast X-ray results before day 0
Number of stenoses per patient (median, range) 3 (1–7)
Number of patients with tight stenosis13
Number of patients with loose stenosis 5
Number of patients with small bowel distension14
Number of patients with tight stenosis and fluid level 6
Number of patients with ulcerated strictures 9
Length (cms) of the longest stricture – median (range) 5 (1–60)
Table 3.   C reactive protein for each patient during the study
PatientsCRP before inclusionResults at first evaluationCRP at first evaluationLast evaluationCRP at the last evaluation
  1. * Two Patients received only one Infliximab infusion.

  2. CRP, C reactive protein (mg/L).


Other medications before infliximab infusion were: corticosteroids [prednisone (n = 10), budenoside (n = 4)], azathioprine/mercaptopurine (n = 13). Twelve patients had received both corticosteroids and azathioprine. The median dosage of prednisone at inclusion was 25 mg (15–50).

Some treatments were begun at day 0: total parenteral nutrition (n = 2), prednisone (n = 6) (including two patients who had received budenoside), azathioprine (n = 5). Doses of prednisone were increased in three patients.

The following predictive factors associated with the efficacy of infliximab were investigated: age at first symptoms, age at diagnosis of CD, time between initial symptoms and inclusion, duration of obstructive symptoms, type of clinical symptoms (abdominal cramps, spontaneously self limited obstruction, complete obstruction), degree of radiological stenosis, associated treatment, new treatment started at day 0, median time between the last infliximab infusion and the last evaluation, results at the first evaluation, treatment chosen after the first evaluation, treatment at the last evaluation and CRP level.

Statistical analysis

Predictive factors of a complete response (vs. partial response or failure) at the first evaluation; of complete and partial success (vs. failure) at the last evaluation were determined by the Wilcoxon Chi 2 test for categorical data and U Mann–Whitney test for numerical data. P < 0.05 was considered significant.


  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Sixteen patients received infliximab infusions (5 mg/kg) at day 0 and weeks 2 and 6. Two patients only received one infusion. Median follow-up was 18 months (range: 1.5–56).

Treatment results of initial treatment with infliximab

Obstructive symptoms were evaluated at week 8. There was a complete response in ten patients (55.6%), a partial response in seven (38.9%) and treatment failed in one patient. Four of the ten patients with a complete response had stopped prednisone or were receiving a lower dosage. Therapeutic decisions about further management in-patients during the first evaluation at week 8 are provided in Table 4.

Table 4.   Therapeutic decisions at week 8 of infliximab
Clinical evaluationTreatmentn
Patients with a complete response (n = 10)Maintenance infliximab infusion7
Infliximab interruption3
Partial response (n = 7)Maintenance infliximab infusion6
Infliximab interruption and surgical resection1
Failure (n = 1)Increased prednisone dosage and maintenance infliximab infusion1

Follow-up of patients receiving maintenance infliximab treatment (every 8 weeks) (n = 14)

Eight patients were still on maintenance infliximab treatment during the last evaluation after a median follow-up of 18 months (range: 3.5–41.5). In 6 patients, maintenance infliximab treatment was interrupted before the last evaluation after a follow-up of 18.5 months (range: 10–41.5). In those patients, the median time between the last infliximab infusion and the final evaluation was 13.2 months (range: 6.5–38).

Clinical evaluation during the final visit

There was a complete response in five patients, a partial response in ten and a failure in three. Seven patients had no symptoms of intestinal obstruction during the final evaluation (five with a complete response and two classified as a partial response because prednisone was begun again during follow-up) (Table 5). Five of those seven patients were on maintenance treatment and two had stopped infusions.

Table 5.   Obstructive symptoms according to infliximab treatment during the last evaluation
 nDelay between symptom recurrence and end of treatment* or follow-up** (months, range)
  1. * Delay between symptoms and the end of treatment in patients with no maintenance treatment or those who had stopped it.

  2. ** Median follow-up when patients were still treated.

No obstructive symptoms (n = 7)
 No maintenance of infliximab treatment0 
 Maintenance infliximab treatment5 
 Interruption of infliximab treatment2 
Obstructive symptoms (n = 11)
 No maintenance infliximab treatment48 (0.75–11)*
 Maintenance infliximab treatment28 (5–13)**
 Interruption of infliximab treatment513.5 (8–24)*

During the final evaluation, 8 of the 16 patients who were receiving prednisone on day 0 had stopped it. The median time between weaning and the final evaluation was 12 months (5.5–35). Eight patients were still being treated with prednisone at a median daily dose of 15 mg (5–60 mg). Five patients were receiving budesonide (6 mg per day in two and 9 mg per day in three). Sixteen patients were receiving azathioprine. Four patients had undergone a surgical resection including one patient with a partial response at week 8 who had chosen to undergo surgery.

Six patients had a CRP above the upper limit of normal (10 mg/L) before inclusion; including three (50%) with a complete response and three (50%) with a partial response during the first evaluation. Six (55%) of eleven patients with a normal CRP had a complete response, four (36%) a partial response and one (9%) a failure. Clinical results were not different between patients with normal or increased CRP’s. Only two patients had a CRP above 10 mg/L during the first evaluation. One had a complete response and one a partial response which is more difficult to analyse. Among patients with a normal CRP, four (22%) had a complete response, nine (50%) a partial response and three (18%) a failure.

Analysis of factors associated with the efficacy of infliximab

The only factor associated with a short-term complete response was initiating prednisone or increasing its previous dosage in association with infliximab (P = 0.015). Initial CRP was clearly not associated with the type of clinical response. During the final evaluation, none of the factors was found to influence significantly the response to treatment.


  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

The efficacy of infliximab for CD strictures has rarely been studied and the results are conflicting. The only open label pilot study with infliximab in-patients with symptomatic CD with strictures was prematurely discontinued because more than two of the first five patients had to undergo surgery less than 2 months after starting infliximab treatment.3

Although 67 of 76 patients treated with infliximab for refractory luminal CD showed a clinical improvement, seven (10.4%) of them developed bowel obstruction requiring hospitalization and surgery. Pre-existing strictures were known in five out of seven patients.17

Toy et al. reported four patients with documented strictures treated by a short course of infliximab infusion. The clinical symptoms improved in each patient but secondary complete bowel obstruction requiring surgery occurred 5 weeks after infusion.18 Whether obstruction was related to this short-term infliximab administration or to the natural course of the disease is not clear.

On the other hand, other studies do not show that infliximab plays a role in the occurrence of obstruction. In the first major study with infliximab (ACCENT 1), patients with high-grade fixed stenosis were excluded.12 During follow-up, about 6% of patients had symptomatic stenosis with no difference between those with episodic or maintenance treatment, despite differences in the amount of infliximab. In the observational TREAT register, patients receiving infliximab were compared with a control group receiving other treatment (more than 3000 patients in each group).19 Patients receiving infliximab had more severe disease and 1.95 symptomatic stenoses/100 patient/years occurred compared to 0.99 in the control group. The difference is significant (P < 0.001), but after adjustment for other factors (severity and ileal location, duration of CD and new corticosteroid use), infliximab was not found to be associated with stenosis.19

In a smaller study, 9 out of 11 patients with inflammatory stenosis responded well to infliximab and were symptom-free with remission up to 18 months later. Most patients were treated concomitantly with immunosuppressors and most of them had an ileal stricture. Only one patient underwent surgery for an abscess.20

Marrache et al. performed a retrospective study in 22 CD patients with a stenosis before infliximab infusion.21 Half of them were asymptomatic. After a follow-up of 16 months, obstructive symptoms had improved in 5 out of 11 symptomatic patients and 10 out 11 patients who had no obstructive symptoms initially remained symptom free. Clinical obstruction occurred in three patients during treatment.

Compared to our study, these results included heterogeneous populations. We selected only symptomatic patients with an ileal stenosis. Azathioprine or corticosteroids had failed in all of them. At week 8, 17 out of 18 patients had a complete or partial response. At the end of follow-up (median 18 months), seven patients had no symptoms of obstruction (five were on maintenance treatment). The short-term success was shown to be associated with prednisone treatment.

Several factors could explain these different clinical results. On one hand, a long-term inflammatory process sustained by increased cytokine production leads to an excess fibrotic response, which could play a role in stricture pathogenesis.22 This is mainly caused by fibroblasts and smooth cell activation, resulting in trans-mural fibrosis.22 Thus, the process from inflammation to fibrosis provides a potential explanation for the poor response to anti-inflammatory therapies in-patients who already have fibrotic stenosis.23

On the other hand, substantial thickening of the mesenchymal layers is observed during mucosal repair (secondary to infliximab, for example).7 There is less information about the latter theory than about the role of the inflammatory process in the pathogenesis of fibrotic stenosis. In our opinion, the control of chronic inflammation to prevent fibrosis and stenosis seems more important than the risk of fibrosis induced by treatment, thus justifying infliximab infusions. The two processes are probably associated. Infliximab decreases the inflammatory process and results in mucosal healing, which could have a beneficial effect on fibrogenesis. However, it also induces rapid mucosal healing that could favour fibrogenesis and strictures. So, it is difficult to predict the effect of this drug on strictures based on current physiopathological information.

Data about the efficacy of other drugs on active lesions such as ulcerations, and the occurrence or progress of stenosis might help predict infliximab efficacy in strictures.24 Although azathioprine has been shown to promote healing of active lesions and to maintain good long term endoscopic results,25–27 one study provided indirect evidence that this immunosuppressor did not prevent stenosis; while the probability of receiving immunosuppressors has increased from 0% to 56% between 1978 and 2002, the cumulative risk of strictures has remained similar over this period.2 However, surgery was performed early in the course of the disease in most of these patients and most of them had either not yet received immunosuppressors or had received them for less than 3 months before surgery.2

Corticosteroids have been shown not to be effective in the treatment of acute lesions,28, 29 while they are effective in clinical stenosis. These results clearly show that mucosal healing is not necessarily linked to decreased inflammation of the deeper layers of the intestinal wall.5, 30, 31 So, results with immunosuppressors and corticosteroids and hypothetical data on the physiopathology of stenosis do not help predict the efficacy of infliximab in the prevention or control of stenosis. Only clinical reports with infliximab in CD patients with symptomatic stenosis can guide the physician. We obtained similar good results in our patients with symptomatic stenosis, as an ACCENT 1 study, which did not include patients with obstructive symptoms.12 At the end of follow-up (median duration 18 months), seven patients (39%) no longer had obstructive symptoms. This proportion is similar to that of patients (41%) receiving combined scheduled strategies who were in remission at week 54 in the ACCENT 1 study. Moreover, the use of infliximab obviously had a corticosteroid sparing effect because at the end of follow-up, 10 of 18 patients had been weaned from prednisone and five had a complete response (symptom-free and no corticosteroids). The results of this retrospective study with a small number of patients must be extrapolated for clinical use with caution. As we did not identify any clinical, biological or radiological predictive factors of response, there is no way to select patients who are likely to respond well. Nevertheless, in our opinion, infliximab can be administered in CD patients with symptomatic strictures who are poor candidates for surgery. Although a prospective double blind randomized placebo controlled study is necessary to clarify further the role of infliximab, it would be difficult to design.


  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References

Declaration of personal and funding interests: None.


  1. Top of page
  2. Summary
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgement
  8. References
  • 1
    Beaugerie L, Carat F, Bouvier A, et al. Cohorte nationale CESAME (cancers et sur-risque associé aux maladies inflammatoires chroniques intestinales en France). données démographiques et médicales 2004–2005 pour 20919 patients 2006; 30: A154.
  • 2
    Cosnes J, Nion-Larmurier I, Beaugerie L, Afchain P, Tiret E, Gendre JP. Impact of the increasing use of immunosuppressants in Crohn’s disease on the need for intestinal surgery. Gut 2005; 54: 23741.
  • 3
    Louis E, Boverie J, Dewit O, Baert F, De Vos M, D’Haens G. Treatment of small bowel subocclusive Crohn’s disease with infliximab: an open pilot study. Acta Gastroenterol Belg 2007; 70: 159.
  • 4
    Oostenbrug LE, Van Dullemen HM, Te Meerman GJ, Jansen PL, Kleibeuker JH. Clinical outcome of Crohn’s disease according to the Vienna classification: disease location is a useful predictor of disease course. Eur J Gastroenterol Hepatol 2006; 18: 25561.
  • 5
    Vermeire S, Van Assche G, Rutgeerts P. Review article: Altering the natural history of Crohn’s disease – evidence for and against current therapies. Aliment Pharmacol Ther 2007; 25: 312.
  • 6
    Gelbmann CM, Mestermann S, Gross V, Kollinger M, Scholmerich J, Falk W. Strictures in Crohn’s disease are characterised by an accumulation of mast cells colocalised with laminin but not with fibronectin or vitronectin. Gut 1999; 45: 2107.
  • 7
    Graham MF. Pathogenesis of intestinal strictures in Crohn’s disease: an update. Inflamm Bowel Dis 1995; 1: 2207.
  • 8
    Van Assche G, Geboes K, Rutgeerts P. Medical therapy for Crohn’s disease strictures. Inflamm Bowel Dis 2004; 10: 5560.
  • 9
    Olaison G, Smedh K, Sjodahl R. Natural course of Crohn’s disease after ileocolic resection: endoscopically visualised ileal ulcers preceding symptoms. Gut 1992; 33: 3315.
  • 10
    Rutgeerts P, Geboes K, Vantrappen G, Beyls J, Kerremans R, Hiele M. Predictability of the postoperative course of Crohn’s disease. Gastroenterology 1990; 99: 95663.
  • 11
    Hassan C, Zullo A, De Francesco V, et al. Systematic review: endoscopic dilatation in Crohn’s disease. Aliment Pharmacol Ther 2007; 26: 145764.
  • 12
    Hanauer SB, Feagan BG, Lichtenstein GR, et al. Maintenance infliximab for Crohn’s disease: the ACCENT I randomised trial. Lancet 2002; 359: 15419.
  • 13
    Rutgeerts P, Feagan BG, Lichtenstein GR, et al. Comparison of scheduled and episodic treatment strategies of infliximab in Crohn’s disease. Gastroenterology 2004; 126: 40213.
  • 14
    Treton X, Stefanescu C, Attar A, Alves A, Zappa M, Bouhnik Y. [Ileal stenosis]. Gastroenterol Clin Biol 2007; 31: 38792.
  • 15
    Bernstein CN, Greenberg H, Boult I, Chubey S, Leblanc C, Ryner L. A prospective comparison study of MRI versus small bowel follow-through in recurrent Crohn’s disease. Am J Gastroenterol 2005; 100: 2493502.
    Direct Link:
  • 16
    Schreyer AG, Golder S, Seitz J, Herfarth H. New diagnostic avenues in inflammatory bowel diseases. Capsule endoscopy, magnetic resonance imaging and virtual enteroscopy. Dig Dis 2003; 21: 12937.
  • 17
    Vasilopoulos S, Kugathasan S, Saeian K, et al. Intestinal strictures complicating initially successful Infliximab treatment for luminal Crohn’s disease. Am J Gastroenterol 2000; 2503: A315.
  • 18
    Toy LS, Scherl J, Kornbluth A, et al. Complete bowel obstruction following initial response to infliximab therapy for Crohn’s disease: a series of newly described complications. Gastroenterology 2000; 118: A569.
  • 19
    Lichtenstein GR, Olson A, Travers S, et al. Factors associated with the development of intestinal strictures or obstructions in patients with Crohn’s disease. Am J Gastroenterol 2006; 101: 10308.
    Direct Link:
  • 20
    Holtmann MH, Neurath MF. Anti-TNF strategies in stenosing and fistulizing Crohn’s disease. Int J Colorectal Dis 2005; 20: 18.
  • 21
    Marrache F, Gornet JM, Pacault V, et al. Effet de l’infliximab au cours de la maladie de Crohn chez les patients ayant une sténose digestive. Gastroenterol Clin Biol 2005; 29: A188.
  • 22
    Geboes K, Dalle I. Influence of treatment on morphological features of mucosal inflammation. Gut 2002; 50(Suppl. 3): III3742.
  • 23
    D’Haens G. Mucosal healing in pediatric Crohn’s disease: the goal of medical treatment. Inflamm Bowel Dis 2004; 10: 47980.
  • 24
    Rutgeerts P, Diamond RH, Bala M, et al. Scheduled maintenance treatment with infliximab is superior to episodic treatment for the healing of mucosal ulceration associated with Crohn’s disease. Gastrointest Endosc 2006; 63: 43342.
  • 25
    Colombel JF, Bouhnik Y, Soulé JC, Lerebours E, groupe GETAID. Endoscopic healing of Crohn’s ileo-colitis with azathioprine. Gastroenterology 2003; 124: A196.
  • 26
    Soulé JC, Etieney I. L’imurel peut-il guérir la maladie deCrohn (MC)? Gastroenterol Clin Biol 2003; 27: A148.
  • 27
    Mantzaris JG, Petraki K, Chadio-Lordanides H, et al. Maintenance therapy with azathioprine is superior to budesonide in healing endoscopic lesions and improving histology in clinically quiescent Crohn's disease. Gastroenterology 2008;122:A81.
  • 28
    Landi B, Anh TN, Cortot A, et al. Endoscopic monitoring of Crohn’s disease treatment: a prospective, randomized clinical trial. The Groupe d’Etudes Therapeutiques des Affections Inflammatoires Digestives. Gastroenterology 1992; 102: 164753.
  • 29
    Modigliani R, Mary JY, Simon JF, et al. Clinical, biological, and endoscopic picture of attacks of Crohn’s disease. Evolution on prednisolone. Groupe d’Etude Therapeutique des Affections Inflammatoires Digestives. Gastroenterology 1990; 98: 8118.
  • 30
    Hellers G, Cortot A, Jewell D, et al. Oral budesonide for prevention of postsurgical recurrence in Crohn’s disease. The IOIBD Budesonide Study Group. Gastroenterology 1999; 116: 294300.
  • 31
    Rutgeerts P, Vermeire S, Van Assche G. Mucosal healing in inflammatory bowel disease: impossible ideal or therapeutic target? Gut 2007; 56: 4535.