SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

Background:

Mucosal ischaemia may contribute to the pathogenesis of Crohn’s disease. Microvascular abnormalities have been found in colonic resection specimens, and mucosal levels of constitutive nitric oxide synthase are reduced.

Aim:

To assess the efficacy of a novel, enteric-release formulation of the nitric oxide donor, glyceryl trinitrate, aimed at increasing the mucosal circulation and relaxing smooth muscle in the affected bowel.

Methods:

The trial was randomized, double-blind and placebo-controlled. Baseline disease activity was assessed by a structured symptom diary, with blood tests and a quality of life assessment. Patients with a Crohn’s disease activity index of ≥ 150 and < 450 were randomized to receive 12 weeks of either glyceryl trinitrate (initially 6 mg twice daily, increasing to 9 mg twice daily after 6 weeks) or an identical placebo. Assessments were repeated at 6 and 12 weeks.

Results:

Seventy patients (22 male) entered the study; 34 were given glyceryl trinitrate and 36 placebo. At 12 weeks, there were no differences between the treatment groups in terms of Crohn’s disease activity index, pain, stool frequency, inflammatory markers or quality of life scores.

Conclusions:

Enteric-release glyceryl trinitrate did not benefit patients with mild to moderately active Crohn’s disease. Whilst ischaemia may contribute to the pathogenesis of Crohn’s disease, our results fail to provide supportive evidence for this hypothesis.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

Mucosal ischaemia is one of several mechanisms which may contribute to the pathogenesis of Crohn’s disease. Microvascular abnormalities, termed multifocal gastrointestinal infarctions, were first shown in resected Crohn’s bowel by combined histopathological, electron microscopic and resin casting techniques.1 Similar findings of acute intestinal mucosal damage were subsequently reproduced in a ferret model, in which blood supply to the submucosal collateral plexus was interrupted by intra-arterial injection of styrene microspheres.2 Microvascular ischaemia may also contribute to early anastomotic recurrence after surgery.3

Smoking has adverse effects on the prevalence, clinical progress and surgical recurrence of Crohn’s disease, and this may constitute a further link between ischaemia and the disease process.4 The mechanisms involved may be similar to those in Buerger’s disease,5, 6 or involve thrombotic factors which increase the risk of cardiovascular events.

It has also been observed at the time of open surgery for Crohn’s disease that handling the bowel may provoke spasm with prolonged blanching (Prof. L. E. Hughes, personal communication).

Nitric oxide is produced by a number of different enzyme systems,7 some of which are physiological or ‘constitutive’; they produce low concentrations of nitric oxide, which can diffuse locally over a limited range, exerting beneficial effects on vasodilatation, platelet function, neurotransmission and gut motility.7–9 The constitutive production of nitric oxide by the endothelium is deficient in a number of pathologies associated with ischaemia.7, 10 In patients with Crohn’s disease, both constitutive and inducible nitric oxide synthase activity are low in the colonic mucosa and muscle of surgical resection specimens. This contrasts with the high levels of nitric oxide synthase activity found in patients with ulcerative colitis.8

Our work addresses the hypothesis that mucosal ischaemia may contribute to the pathogenesis of Crohn’s disease. We therefore formulated a preparation of the nitric oxide donor, glyceryl trinitrate (GTN), designed for release in the ileum and colon, with the intention of both improving the mucosal circulation and relaxing the smooth muscle in the gut wall. Suitable patients were randomized to receive active or placebo treatment over 12 weeks.

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

Seventy patients (22 male) were admitted to the trial between November 1998 and November 2000 from four centres within the UK: Cardiff (33), Swansea (22), Llandough (8) and Leicester (7). Patients with a Crohn’s disease activity index of ≥ 150 and < 450 were identified after an initial interview and completion of a diary card for symptoms over 2 weeks. All were more than 16 years of age, with Crohn’s disease involving colon proximal to the rectosigmoid junction and/or distal small intestine (verified by colonoscopy, barium or white cell scan studies). No patients were on maintenance doses of prednisolone greater than 20 mg, and all treatment, including 5-aminosalicylic acid, azathioprine and metronidazole, had remained unchanged for at least 6 weeks before they entered the trial. Exclusion criteria were a previous surgical resection of ileum > 100 cm, the presence of an ileostomy, pouch or colostomy and a history of significant cardiac (especially hypertrophic obstructive cardiomyopathy (HOCM), aortic or mitral valve stenosis, constrictive pericarditis or hypotensive episodes), respiratory, endocrine, renal, hepatic, neurological or psychiatric disease. Other exclusion criteria included significant coexisting gastrointestinal disease, anaemia, glaucoma, total parenteral nutrition and patients already taking nitrates or calcium antagonists. Pregnancy, breast feeding, participation in an interventional study within the past 3 months or any reason likely to result in poor compliance with the study protocol were also reasons for exclusion.

A diary card was used to record daily the number of bowel motions in 24 h and their consistency, with semiquantitative scores for abdominal pain (0=none to 3=severe), urgency (1=normal to 9=incontinent), general well-being (0=generally well to 4=terrible) and whether a headache had been present. This was completed during a 2-week baseline period after taking blood for measurement of the full blood count, urea and electrolytes, liver function tests, erythrocyte sedimentation rate and C-reactive protein; the information was used to calculate a baseline Crohn’s disease activity index. A disease-specific quality of life questionnaire was also completed.11

Patients were randomized according to a computer-generated sequence to receive 12 weeks of treatment. Those in the active treatment group were given GTN, 6 mg twice daily for the first 6 weeks; if this was well tolerated, the dose was increased to 9 mg twice daily for the remaining 6 weeks. An identical placebo was used to ensure that both the trial investigator and patient remained blind to the treatment allocation.

Daily scores were recorded using the structured diary card throughout the trial, with details of any adverse effects, their duration, severity and possible relationship to medication. The same routine blood tests and quality of life questionnaire were repeated at 6 and 12 weeks.

The study medication contained 3 mg of GTN dispersed in a polyglycolized triglyceride (Gelucire, Gattefosse) wax matrix, designed to produce both a delayed and sustained release of the drug. The matrix was contained in a size 1 hard gelatin capsule coated with Eudragit-L (Rohm Pharma), an acrylic resin with an optimal dissolution at pH 6.8; this allowed the release of GTN to commence in the distal small bowel (pH 6.8) and continue through to the large bowel (pH 7.2). Previous studies have demonstrated the reliable in vivo release pattern of similar enteric-coated capsules.12–14In vitro dissolution studies, conforming to the standard British Pharmacopoeia method,15 with high pressure liquid chromatography, confirmed the linear release profile for GTN from Gelucire (Figure 1). Matching placebo capsules contained only the wax matrix.

image

Figure 1.  The in vitro dissolution profile of the glyceryl trinitrate formulation used in the study (determined by high pressure liquid chromatography). Mean values are plotted over 8 h.

Download figure to PowerPoint

The inclusion of 60 patients, 30 in each group, was calculated to give the study 80% power, based on a reduction in Crohn’s disease activity index of 10% and 40% for placebo and GTN, respectively. Inclusion of 70 patients allowed for a 15% drop-out rate. The primary outcome measure was a change in the Crohn’s disease activity index at the termination visit. Secondary outcome measures included changes in the pain score, stool frequency, inflammatory markers and quality of life score. Statistical analysis was performed with SPSS software, version 10. To assess the efficacy of treatment, two-sided analysis of covariance (ANCOVA) was used for continuous variables, and the Mann–Whitney U-test for non-parametric data on both an intention-to-treat and per protocol basis. Statistical significance was set at the P=0.05 level.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

Of the 70 patients who entered the study, 34 were given active treatment and 36 placebo treatment. The demographics of the two groups are shown in Table 1. Randomization produced groups that were similar with regard to age, sex, disease site, previous surgery, current medication, complications and the presence or absence of an abdominal mass. However, there was a preponderance of current smokers with a deficit of lifelong non-smokers in the active compared to the placebo group. A comparison of the proportion of current smokers in each group using the chi-squared test gave χ2=7.02 (2 d.f.), P=0.03, for the difference.

Table 1.   Demographics of the active treatment and placebo groups Thumbnail image of

Fifty-three patients completed the trial protocol, and 17 were withdrawn prematurely (all before the week 6 visit) (Figure 2). The reasons for withdrawal are given in Table 2. Complete data were therefore available for analysis in 53 patients, 24 of whom had received active GTN and 29 placebo. Outcome data, available in 66 patients, were used in the intention-to-treat analysis; only baseline observations were obtained in the remaining four patients.

image

Figure 2.  Outcomes by treatment group following randomization.

Download figure to PowerPoint

Table 2.   Reasons for premature withdrawal in both groups Thumbnail image of

No significant differences were identified between the two treatment groups in terms of the Crohn’s disease activity index or any of the secondary outcome measures (changes in pain score, stool frequency, inflammatory markers or quality of life score). The results of intention-to-treat and per protocol analysis are given in Tables 3 and 4, respectively. Individual changes in the Crohn’s disease activity index of the two groups completing the study are also shown in Figure 3a,b. Ten patients in the active group went on to receive 9 mg GTN twice daily; no significant benefit was associated with the higher dose.

Table 3.   Differences in the primary and secondary outcome measures in active and placebo groups at termination (intention-to-treat analysis) Thumbnail image of
Table 4.   Differences in the primary and secondary outcome measures for patients completing the study per protocol in active and placebo groups at termination Thumbnail image of
image

Figure 3.  Crohn’s disease activity index (CDAI) for the 53 patients completing the protocol at baseline, week 6 and week 12: (A) active treatment; (B) placebo. Full lines indicate patients taking at least 10 mg prednisolone.

Download figure to PowerPoint

In view of the relative excess of non-smokers in the placebo group, subanalysis by smoking status was performed. An inverse correlation between the number of cigarettes smoked daily and the improvement in the Crohn’s disease activity index was demonstrated for both current and ex-smokers (previous levels of consumption) (Spearman rank correlation coefficient, 0.607; P=0.03). However, this effect was independent of the treatment given, and there were no significant differences between the two groups for the major outcome variables when stratified by smoking status. There were also no differences after allowance for the effect of variable maintenance doses of steroid. Of the 34 patients taking steroids, 19 on active treatment and 15 on placebo, the dose was increased during the trial in seven (five active and two placebo) and decreased in two (both taking placebo).

The side-effects recorded in the two groups are given in Table 5. Headache was the commonest problem with both the GTN and placebo preparations. In most cases, this was worse in the first week of treatment and was only severe enough to cause withdrawal from the study in four cases. Headache was accompanied by flushes or dizziness in four patients, and there was one syncopal event in a subject who had mistakenly taken a 15-mg dose of the GTN formulation. No serious or life-threatening events occurred.

Table 5.   Reported side-effects Thumbnail image of

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

The results failed to show any benefit from the enteric-release GTN, compared with placebo, in this group of patients with mild to moderately active Crohn’s disease. The improvements which occurred after 6 and 12 weeks in both groups illustrate the regression in symptoms which often occurs in patients continuing to take maintenance therapy, which included steroids in almost half. A further analysis of patients who appeared to benefit from active treatment failed to identify any particular subgroups or patient characteristics for which the GTN preparation may have been beneficial.

Prior to this clinical trial, we had carried out a pilot study lasting 6 weeks in 14 patients with active Crohn’s disease, using a dose of 6 mg twice daily of the GTN formulation. After 6 weeks, the Crohn’s disease activity index was lower in 10 patients and significant clinical improvement occurred in six. The response to the GTN therapy in these patients was the basis for proceeding with the controlled clinical trial.

The criteria for admission to the trial were strictly applied and limited the numbers admitted. In addition to those included, a further 120 patients with appropriate symptoms had the initial baseline measurements, but were not admitted to the trial because their Crohn’s disease activity index failed to meet the required levels. The trial design continued maintenance therapy, which included modest doses of prednisolone, because it was considered unethical to proceed otherwise in these patients with active disease. Although a substantial number of those who entered the study were withdrawn (17, 24%), five of these because of a deterioration in symptoms and four of them on active therapy, the study still has a 75% power to detect the anticipated changes in Crohn’s disease activity index, and it is therefore unlikely that this affected the outcome of the trial. Although headache was the most commonly reported side-effect, which involved 12 and eight of the active and placebo groups, respectively, ‘blinding’ was not unmasked by this as the majority of patients were able to take between 12 and 18 mg of GTN daily without adverse effects.

GTN was chosen to achieve a topical effect in the gut because of its established role as a nitric oxide donor, releasing nitric oxide in the vascular endothelium to produce vasodilatation and relaxation of smooth muscle in the gut wall. Furthermore, as the compound is largely removed by ‘first pass’ liver metabolism, only 3% remains for a systemic effect, with the associated side-effects.16 We were encouraged to consider what may be an analogous clinical response of anal fissures to the topical application of GTN: the healing which follows treatment is accompanied by an increase in blood flow in the peri-anal region, and a fall in resting pressure within the anal canal itself, due to the action of GTN on the internal sphincter.17–19 If the increased blood flow with a fall in smooth muscle tone is relevant to the healing of fissured lesions in the bowel wall of patients with Crohn’s disease, clinical benefit might be expected from topical release of the preparation.

Our failure to show any clinical benefit with GTN does not discount the possibility that ischaemia may play a role in the pathogenesis of Crohn’s disease, but it does not provide supporting evidence for the hypothesis. Whilst it is possible that a different dose or release profile for GTN in a more suitable group of patients may have produced a positive result, there is little encouragement from our results for this possibility. We feel that our cautious approach of delaying publication of the initial pilot study, and waiting for the results of a formal controlled trial, has been vindicated. Whilst uncontrolled observations on new treatments may appear to be beneficial, the onus is on those who make the observations to try and follow through with controlled studies where this is possible.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References

Ethical approval for the study was given by the Local Research Ethics Committee for each of the participating centres. Dr Robert Newcombe, Senior Lecturer in Medical Statistics, University of Wales College of Medicine, gave advice on the study design, statistical analysis and presentation of the results. The trial capsules were produced by MW Encap Ltd, Livingston, UK. Drs Hawkes and Richardson were supported by the Gastrointestinal Foundation Trust.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. References