Montelukast in the Treatment of Duodenal Eosinophilia

I. Background/Rationale

Recurrent abdominal pain is a common complaint among school-age children, being present in up to 15% at any given time. It represents the most common chronic pain entity in pediatric patients. These patients frequently have dyspepsia defined as upper abdominal pain or discomfort. (1) Eighty-three percent of children referred to our institution for recurrent pain have dyspepsia. Shaffer, et al, found dyspepsia to be present in 73% of 154 children with recurrent abdominal pain. (2)

We have found mucosal eosinophilia in 71% of children undergoing endoscopy with mucosal biopsy for dyspepsia. However, eosinophil density does not necessarily correlate with eosinophil activation and many eosinophil-derived mediators are bioactive in a concentration-dependent fashion. (3) In previous investigations, we have found evidence of moderate to extensive eosinophil degranulation in the biopsies of dyspeptic children who had normal mucosal eosinophil densities. (4)

Montelukast is a competitive antagonist of the cys LT1 receptor with an affinity that is similar, but lower than that of leukotriene D4. (5) We and others have previously reported good clinical response to montelukast in patients with eosinophilic gastroenteritis. (6-9) Recently, Vanderhoof and Young (8) reported eight patients with dysphagia, diarrhea, and/or constipation associated with tissue eosinophilia who had prolonged remission of symptoms with montelukast therapy. This experience prompted us to undertake a double-blinded placebo-controlled cross-over trial of montelukast in dyspeptic children with duodenal eosinophilia. We were able to demonstrate the superiority of montelukast as compared to placebo in the relief of pain. (9) Despite an average duration of pain of nearly 22 months prior to study enrollment, approximately one-half of the patients became pain free or nearly pain-free during the two week course of therapy with montelukast. Also of interest from this particular study was the finding that montelukast pharmacokinetics, and thus exposure, were different than previously observed in children receiving the drug. Specifically, the average population elimination t-1/2 for montelukast in our subjects (1.8 hours) was substantially shorter than mean values for this parameter (3.4 hours) determined from children without concurrent intestinal disease. (5, 10) While the reasons for this apparent disparity are not clear, it is possible that local montelukast metabolism (ie., in the small intestine) may vary as a consequence of disease state. Nonetheless, what remains to be determined is whether there is a link between systemic and tissue levels and whether an exposure-response relation can be established for montelukast in pediatric patients with eosinophilic duodenitis.

Finally, the mechanism responsible for the demonstrated clinical efficacy of montelukast in this disorder has not been established. It is possible that the therapeutic effect might result consequent to a lowering in eosinophil density, alteration of the eosinophil activation state, blocking leukotrienes released by eosinophils (or other cells) at their site of action, or any combination of the aforementioned effects. The clinical effectiveness of medications for eosinophilic esophagitis has been previously found to be associated with a decrease in eosinophil density, but the activation state was not assessed. (11-12) As well, no previous investigations of therapeutic interventions for this disorder have addressed exposure-response relationships.

II. Hypothesis

The beneficial therapeutic effect of montelukast in patients with eosinophilic duodenitis is determined by the activation state and local density of eosinophils, and by attainment of sufficient systemic and local exposure.

III. Objectives / Specific Aims

A. To determine the effect of montelukast on mucosal eosinophil density and activation state in pediatric patients with eosinophilic duodenitis presenting as dyspepsia.

B. To evaluate the relationship between clinical response, systemic drug exposure, local tissue drug concentration, and the eosinophil density and activation state.

IV. Methods

A. Study Population

A total of 24 patients evaluated at the Children's Mercy Hospitals and Clinics will be enrolled in the study.

1. Inclusion Criteria

A subject will be eligible for inclusion in this study if all of the following criteria are met:

1. Patients with dyspepsia defined as upper abdominal pain or discomfort of at least two months duration.

2. Male and female subjects 6 - 17 years of age

3. Endoscopic mucosal biopsies revealing at least 20 eosinophils in a high power field on specimens from the duodenum in the absence of other defined duodenal pathology.

4. Written informed consent signed by the subject's legal guardian and by the investigator and patient assent.

2. Exclusion Criteria

Subjects will not be eligible for inclusion in this study if any of the following criteria apply:

1. A subject who has previously been treated with montelukast.

2. Treatment with steroids or oral cromolyn in the four weeks prior to endoscopy or during the study.

3. Any history, clinical, and/or biochemical evidence of clinically significant renal or liver disease.

4. Treatment with any inducers or inhibitors of CYP3A4.

5. Intake of grapefruit juice or oral vitamin E supplementation

- Patients meeting all criteria except 1.c. will be approached for enrollment prior to the initial endoscopy. With consent, two duodenal biopsies will be obtained and processed for EM evaluation as below pending histologic evaluation to determine whether the patient meets criteria 1.c. If the patient does not meet criteria 1.c., EM specimens will be discarded.

3. Methods

1. Treatment: All patients will receive an open-label, 21 day course of montelukast at a dose of two 5 mg tablets daily, given between 0700 and 0800.

2. Assessment of treatment response

1. . Global evaluation - A five point global pain relief evaluation will be conducted by a single investigator on day 21. The evaluation will employ a Likert-type scale adapted to assess change in pain or discomfort as follows:

Grade 1: Worse - clinical deterioration with increasing pain intensity and/or frequency

Grade 2: No change - no increase or decrease in pain intensity or frequency.

Grade 3: Moderate improvement - partial clinical response with definite improvement in pain, but not meeting the criteria for a Grade 4 response.

Grade 4: Good - nearly complete relief of symptoms with minimal residual pain or discomfort. Relief denoted as adequate by patient. Pain not interfering with daily activities.

Grade 5: Excellent - complete relief of pain or discomfort.

2. . Mucosal Eosinophils

On day 21 of therapy, all patients will undergo a follow up endoscopy to obtain mucosal biopsies. All endoscopies will be performed in the usual fashion, which includes obtaining standard biopsies (2 esophageal, 2 antral, and 4 duodenal). Patients will also have two additional duodenal biopsies obtained for electron microscopic (EM) evaluation.

Both peak and mean eosinophil density will be determined by a single observer on both pre- and post-treatment specimens. The observer will be blinded to which specimens are pre- and which are post-treatment.

3. . Laboratory

On day 21, at the time of endoscopy, a blood sample will be obtained on all patients for the determination of serum eosinophil cationic protein concentration. Serum will be frozen and saved for a period not to exceed 24 months from study completion to be used for determination of other serum protein concentrations which may be determined to have possible clinical or physiologic relevance.

3. Pharmacokinetic analysis

On day 21, all patients will report to the GI procedure room three hours prior to their endoscopy. An intravenous catheter will be placed for the procedure. At 2.5 hours prior to the procedure, patients will receive two, 5 mg montelukast tablets with 2 oz of water and the time will be recorded. Blood samples will be obtained at 1, 2.5, and 6 hours post-dosing. Plasma will be moved following centrifugation and stored at -70ºC. At the time of endoscopy (approximately 2.5 hours post-dosing), multiple duodenal mucosal biopsies (appx. aggregate tissue yield = 0.5 gm) will be obtained for determination of tissue drug concentration. All serum and tissue drug determinations will be determined by the manufacturer (Merck) using a validated HPLC method that has been used to support two previous pharmacokinetic studies of montelukast (9,10) conducted in pediatric patients, both of which have successfully used the aforementioned approach for pharmacokinetic data analysis.

V. Statistical Analysis

Pre- and post-treatment eosinophil densities, degranulation indexes, and ECP concentrations, respectfully, will be compared using paired t-tests. Changes in density, degranulation index, and ECP from pre- to post-treatment will be compared between responders (Grade > 3) and non-responders (Grade 1 or 2), respectively, by repeat measures ANOVA. Statistical analysis of pharmacokinetic data will utilize a combination of approaches to assess the potential interactions between montelukast disposition and response. Relationships between continuous variables (eg., dose, ECP concentrations) will be evaluated by linear and nonlinear regression techniques. A two-tailed, Student's t test will be used to evaluate differences in disposition (eg., plasma vs. tissue concentrations) and response associated with gender and also, to examine exposure-response relationships for subjects classified by pain relief assessment. Finally, logistic regression techniques will be used to examine the interaction between multiple covariates (eg., pharmacokinetic and pharmacodynamic parameters, biochemical parameters reflective of eosinophil activation, demographic parameters, etc.). All statistical analyses will be completed using the SSPS software package (version 9.0, SPSS Inc., Chicago, IL) and will use a significance limit of α = 0.05.

VI. Sample Size Estimate/Power Calculation

Calculations were completed based on tissue eosinophil density estimates. The initial (pre-treatment) eosinophil counts were obtained from data obtained in the initial montelukast trial (9) and we assumed the standard deviation would remain an equal percent of the mean post-treatment. Calculations were based on achieving 80% power performing a two-sided paired t-test and a significance level of α = 0.05.

1 2 3 4 Pre-treatment mean 34 34 34 34 Post-treatment mean 10 20 25 27 N = 5 10 20 30

We chose a sample size of 24 patients to allow for a patient drop out rate of approximately 15% (twice the rate of the original study).

VII. Projected Timeline for Study Completion

It is estimated that it will take 16 weeks to enroll the entire cohort of subjects. An additional 8 to 12 weeks is allocated for sample analysis by Merck, followed by an additional 4 to 6 weeks for data analysis and report generation. All data (in the form of a final study report) will be submitted for review and comment by the study sponsor (appx. 4-6 weeks). Manuscript generation and submission will then follow (appx. 6-8 weeks).

References Available