The Phase 1/2 Trial of Indomethacin in Chronic Pancreatitis (The PAIR Trial): Protocol for a Parallel Multi-Center Randomized Controlled Trial

Abstract

Background/Objectives:

Current treatments for chronic pancreatitis focus on symptom management and therapeutics targeting disease reversal are lacking. Given the role of the cyclooxygenase-2 (COX-2) enzyme in producing prostaglandin E₂ (PGE₂), a key component in the inflammatory pathway of chronic pancreatitis, this study evaluates the physiologic effect of oral indomethacin, a COX-2 inhibitor, on PGE₂ levels in pancreatic fluid.

Methods:

This pilot two-center randomized controlled trial seeks to examine 32 subjects with chronic pancreatitis who have no contraindications to indomethacin. Subjects will be randomized to either oral indomethacin 50 mg twice a day or placebo twice a day for a total of 28 days. Baseline (pre-treatment) assessment of pain and quality of life will be performed using the Brief Pain Inventory and the PROMIS-10 questionnaires, respectively. Biological specimens including blood, urine, and saliva will be collected at pre-treatment and post-treatment(day 28). Endoscopic pancreatic function testing with concomitant pancreatic fluid collection will also be performed pre- and post-treatment to assess the change in pancreatic fluid PGE₂ levels. The relationship between pancreatic fluid PGE₂ levels with blood and saliva PGE₂ levels will be examined.

Conclusions:

This study will elucidate the effect of oral indomethacin on PGE₂ levels in the pancreas to assess its role in the inflammatory pathway of chronic pancreatitis. Should indomethacin significantly reduce PGE₂ levels, this may represent a potential disease-altering treatment for chronic pancreatitis.

INTRODUCTION

Chronic pancreatitis represents a chronic progressive fibro-inflammatory disease of the pancreas often marked by parenchymal atrophy, fibrosis, pain, pancreatic duct strictures, calcifications, exocrine dysfunction, and endocrine dysfunction.² The resulting chronic pain and loss of pancreatic function can lead to poor quality of life and impairment and place a large burden on our healthcare system.^{3, 4} Current treatments focus on managing symptoms associated with chronic pancreatitis with the primary limitation being the inability to reverse or slow disease progression.⁵ Therefore, one of the man objectives of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) is to develop new treatments that target the restoration of pancreatic acinar cells and/or ductal cells and ultimately improve the pain and quality of life of patients with this disease.⁷

While much about the dynamic nature of the inflammatory pathway in chronic pancreatitis remains to be elucidated, several studies have implicated cyclooxyegenase-2 enzyme (COX-2) induced prostaglandin E₂ (PGE₂) at sites of inflammation in the pancreas.^8–11 Specifically, induction of the COX-2 enzyme leads to production of thromboxanes and prostaglandins, including PGE₂. As shown in Figure 1, PGE₂ acts as a potent mediator of chronic inflammation and stellate cell activity in the pancreas.^{8, 9} In patients with chronic pancreatitis, COX-2 activity is over-expressed in both acinar and ductal cells and within a murine model, pancreatic inflammation is COX-2 dependent.^{8, 10, 11} Based upon this pathway, COX-2 inhibition may alter the inflammatory pathway in chronic pancreatitis, as seen when administration of rofecoxib, a selective COX-2 inhibitor, resulted in lower tissue levels of PGE₂ and inhibition of pancreatic inflammation and fibrosis in rats.¹² Similarly, in mouse models of chronic pancreatitis, sulindac, a nonselective COX-1 and COX-2 inhibitor, decreased inflammation and fibrosis and the selective COX-2 inhibitor NS-398 blocked the response of pancreatic stellate cells to proinflammatory cytokines.^{13, 14}

Figure 1: Cyclooxygenase-2 enzyme (COX-2) and prostaglandin E₂ (PGE₂) in chronic pancreatitis.

COX-2 in monocytes (A) produces PGE2, which stimulates acinar cells (B) to release monocyte chemoattractant protein-1 (MCP-1), recruiting additional monocytes and macrophages into the pancreas and propogating chronic pancreatic inflammation.¹ In addition, stromal PGE2 stimulates human pancreatic stellate cell (C) activation via the EP4 receptor, inducing stellate cell COX-2 activity, PGE2 release, and stellate cell propogation, activation, collagen production, and pancreatic fibrosis.⁶

Within this perspective, indomethacin has garnered interest as a non-steroidal anti-inflammatory drug (NSAID) that inhibits COX-2. The protective effect of indomethacin in preventing post-ERCP pancreatitis is well-known, but the effect of indomethacin in chronic pancreatitis remains unstudied.¹⁵ Although rectal indomethacin is commonly used for post-ERCP pancreatitis prophylaxis, oral indomethacin may be more appropriate for long-term use as it has greater peak plasma concentrations, higher likelihood for compliance and greater absorption compared to rectal administration.¹⁶ Therefore, this study aims to investigate the physiologic effect of oral indomethacin in patients with chronic pancreatitis with the long-term goal of investigating its use as a potential therapy for this disease with limited treatment options.

Objectives

The primary objective of this study is to assess the physiologic effect of oral indomethacin on pancreatic juice PGE₂ concentrations in patients with chronic pancreatitis. Our primary hypothesis postulates that the mean pancreatic juice PGE₂ concentration will decrease after indomethacin to less than 66% of baseline and will be significantly lower compared to placebo treatment. This would justify future large-scale clinical trials of oral indomethacin in chronic pancreatitis.

The secondary objectives of this study include assessing the correlation of drug-induced changes in pancreatic juice PGE₂ levels with changes in salivary and blood PGE₂ levels, blood indomethacin levels, and changes in patient-reported outcomes including pain and quality of life. Our corresponding exploratory hypotheses are that: 1) the peak blood indomethacin level will correlate with the degree of change in pancreatic juice PGE₂ concentrations, 2) baseline salivary and blood PGE₂ concentrations, and changes in these concentrations after treatment, will correlate with pancreatic juice PGE₂ concentrations, and 3) improvement in pain and quality of life will be associated with changes in pancreatic juice PGE₂ concentrations. These findings would provide further insight into the pancreatic effects of oral indomethacin, determine whether measurement of PGE₂ in fluids collected noninvasively correlate with pancreatic juice PGE₂ concentrations, and provide pilot data regarding patient-centered outcomes for planning subsequent clinical trials.

METHODS

Design

The PAIR Study is a phase 1/2 randomized, placebo-controlled, double-blinded study comparing oral indomethacin with placebo in patients with chronic pancreatitis being conducted at Mayo Clinic, Rochester and The Ohio State University Wexner Medical Center.

Study Population

This study will include adult subjects with chronic pancreatitis meeting diagnostic criteria as established by American Pancreatic Association guidelines, irrespective of the etiology.⁵ Subjects will need to have normal renal function and cannot have any contraindications to NSAID use or regularly use NSAIDs (including daily low-dose aspirin). Furthermore, an episode of acute pancreatitis (within 6 weeks prior to enrollment) will preclude eligibility into the study. Specific inclusion and exclusion criteria are specified in Table 1.

Table 1:

Inclusion and Exclusion Criteria

Inclusion Criteria

Any gender, age ≥ 18 years and < 80 years

Diagnosed with chronic pancreatitis per American Pancreatic Association guidelines (pancreatic calcifications and/or Cambridge 3–4 changes on CT, MRI, and/or ERCP)

Scheduled for an upper GI endoscopic procedure (EGD or EUS) for clinical or research indications (not conflicting with current investigation)

Able to provide written informed consent

Serum creatinine within normal laboratory range, as measured within 30 days of the baseline study endoscopy

For females of reproductive potential: willing to use highly effective contraception while taking study medication and for an additional 5 days after completing study medication

Exclusion Criteria

Diagnosed with acute pancreatitis requiring hospitalization within the 6 weeks prior to study enrollment

Habitual use of aspirin or non-steroidal anti-inflammatory medications (NSAIDs), defined as use more than once per week

Any use of aspirin or NSAIDs within 1 week of baseline study endoscopy procedure

Allergy to secretin, indomethacin or NSAIDs

History of known chronic renal insufficiency or cirrhosis

History of coronary artery disease, angina pectoris, myocardial infarction, cerebrovascular accident (stroke), or transient ischemic accident (TIA)

History of peptic ulcer or gastrointestinal bleeding

Incarcerated

Found to have active gastrointestinal ulceration at the time of baseline endoscopy

Hospitalized for acute pancreatitis while participating in this research protocol. Participants who are hospitalized for an episode of acute pancreatitis during study participation will be withdrawn from the study, and considered non-accrued

Active alcohol use

Study Interventions and Measures

Subjects will be randomized to receive either oral indomethacin (50 mg) or a placebo. Both will be taken orally twice a day (BID). Commercially available indomethacin will be utilized and will be over-encapsulated to match the placebo capsules.

Upon enrollment, patients will be scheduled for 2 endoscopic pancreatic function tests (ePFT). These will be performed at baseline and on day 28. This procedure entails the administration of intravenous secretin (0.2 mcg/kg) over 1 minute to stimulate pancreatic juice secretion. This pancreatic juice will then be aspirated from the duodenum using either a standard gastroscope or echoendoscope (if an endoscopic ultrasound is being performed). Fluid will be aspirated from the duodenum pre-secretin (to assess basal levels of PGE₂) and then at time intervals of 0–10 minutes, 10–20 minutes, 20–30 minutes, and 30–45 minutes after secretin administration. The fluid will then be processed in accordance with the standard operating procedures of the Chronic Pancreatitis, Diabetes, and Pancreatic Cancer (CPDPC) Consortium.¹⁷

The pancreatic juice bicarbonate and lipase concentrations will be measured at each time point using a previously validated method.¹⁸ PGE₂ concentrations will be measured in saliva, blood, and pancreatic juice at baseline and at day 28 using a tandem mass spectroscopy assay developed and validated at Mayo Clinic, Rochester. We had previously demonstrated the increased pancreatic fluid PGE₂ levels in patients with chronic pancreatitis.¹⁹ Peak indomethacin drug levels will be measured in the blood on day 28, which will serve as an assessment of participant compliance with the study drug compliance. We will also monitor the time interval between the last dose of the study treatment and the final ePFT (day 28).

The Brief Pain Inventory (BPI) and the PROMIS-10 Global Health are validated scales that quantify pain and quality of life, respectively.^{20, 21} Subjects will complete these two questionnaires at baseline and at day 28. Additionally, subjects will complete a daily medication and pain diary during the treatment period that details all opiate and over-the-counter medication (including proton pump inhibitors) use. Opiates will be expressed in oral morphine equivalents. Baseline data will include demographics data including co-morbidities, substance use (i.e. smoking, alcohol, marijuana) and baseline disease characteristics including CP etiology and pancreatic morphology.

Up to 40 subjects will be enrolled in order to accrue 32 participants in the study protocol. As shown in Figure 2, biospecimen collection will occur at Day 1 and Day 28 (start and end of study drug intake). Half of the subjects will receive oral indomethacin 50 mg BID and the other half will receive placebo BID.

Figure 2:

Study Flowchart

Randomization and Blinding

A computer-generated randomization list will compose randomized blocks of size 2 or 4 to randomize subjects to receive either indomethacin or placebo. Two separate sets of blocks will be created, one for subjects with chronic pancreatitis of alcohol etiology, and another for those with non-alcohol etiologies. Within each block, an equal number of subjects will be assigned to both study medications. A separate set of randomization blocks will be provided for the two study sites.

Both study medications will be identical in appearance. The investigators, study coordinators, and clinical caregivers will remain blinded to subject allocation throughout the study. The study statistician (LL) will have access to the randomization list and will perform data analysis, but will not be involved in any other aspect of the study.

Recruitment

Study subjects will be recruited from the two study sites utilizing the existing infrastructure at both sites for the PROspective Evaluation of Chronic Pancreatitis for EpidEmiologic and Translational Studies (PROCEED) study as part of the CPDPC Consortium.²² This entails use of the existing methods for subject enrollment, ePFT, and biospecimen handling.

Statistical Analysis

The primary analysis will utilize a linear random intercept models that incorporates pancreatic juice PGE₂ data from both baseline and day 28 from all randomized subjects. Transformation will be used as appropriate to reduce the skewness of data and satisfy modeling assumption. The model includes randomized groups, time, and their interactions as fixed effect covariates, from which the mean change in pancreatic fluid PGE₂ will be estimated for each randomized group. An intention-to-treat analysis will be performed for the primary outcome. Effects of additional covariates, such as sex and sites, will be explored. The intra-subject correlation will be estimated from the variances of the random intercepts and residuals. Other outcome variables, such as other biomarkers and questionnaire data, will be analyzed in a similar way. The estimates about the variability of data and intrasubject correlations will be useful for designing a subsequent larger study.

For the secondary analyses, we will use a multivariate analysis of variance model (MANOVA) with repeated measures. The multivariate outcomes include patient-related outcomes (BPI and PROMIS-10 scores) and pancreatic fluid PGE₂ concentrations at baseline and day 28. Fixed effect covariates include randomized groups, time, and their interactions. Additional secondary analyses include descriptive analysis and statistical comparison of pancreatic fluid PGE₂ concentrations at baseline and after secretin administration between the various intervention arms, and correlations between baseline pancreatic fluid PGE₂ levels, clinical features, and changes in plasma and saliva PGE₂ levels. Changes in daily pain ratings and narcotic use (in morphine equivalents) will be analyzed in comparison to changes in pancreatic fluid PGE₂ concentrations. Any dropout before day 28 will be documented and descriptively summarized. Sensitivity analysis on dropout will be conducted in an exploratory manner according to standard missing data analysis procedure for longitudinal data.²³ As a pilot study, we will collect and descriptively analyze data on feasibility, patient compliance and satisfaction, and attrition, in order to inform methods for a possible future study.

We propose a sample size of 16 subjects per treatment arm (32 subjects total). Since this is a pilot study, this sample size is largely limited by recruitment feasibility. However, this sample size is adequate for studying the primary hypothesis with a two-sample two-sided t-test on the mean PGE₂ level at post-treatment (day 28) between the randomized groups, if the effect size is large, as discussed below We expect to show that the treatment group has significantly lower mean PGE₂ than the placebo group. In our previous publication on pancreatic fluid PGE₂,²⁴ we reported that median pancreatic fluid PGE₂ level in 10 CP patients was 307 pg/ml with IQR of 249–362 pg/ml and standard deviation of 84 pg/ml. The boxplot of those data suggests that normality is a reasonable distributional assumption for the PGE₂ data. That study also found that mean pancreatic fluid PGE₂ in a healthy control group was 104 pg/ml, about one third of the mean level of the CP group. If we conservatively assume that the control group in the proposed trial has 50% decrease in mean, with a Type I error of 0.05, the power of the t-test is 0.998. If we use an even more conservative mean for the control group (placebo arm), with 1/3 reduction in mean PGE₂, the power is 0.94. If we additionally assume that there is some data loss (due to consent withdrawal, dropout, or missing specimens) so that the sample size per treatment group is 12 instead of 16, the statistical power decreases to 0.85, but is still reasonable for a proof-of-concept study. Note that the large statistical power in the calculation above is a result of the relatively small standard deviation of the PGE₂ compared with our working assumption on the mean reduction.²⁴ If this pilot study is successful, demonstrating a statistically significant indomethacin-related decrease in pancreatic fluid PGE₂ by at least 1/3^rd , we will conduct a future study that focuses on patient centered outcomes; that study is expected to have a larger sample size because those outcomes are expected to be less sensitive to the treatment than pancreas fluid-based biomarkers.

DISCUSSION

This study represents the first study assessing the effect of oral indomethacin in human subjects with chronic pancreatitis. We aim to investigate targeting the fibroinflammatory pathway via COX-2 inhibition to decrease PGE₂ levels. The ability to obtain pancreatic fluid PGE₂ levels during ePFT enables direct measurement of pancreatic PGE₂ levels, which we hope to identify as a marker for response to treatment. Given the high burden of chronic pancreatitis on quality of life and pain in patients in the context of limited treatment options, this study attempts to open the door for therapies that will potentially reverse or slow disease progression.

Indomethacin as a medication has gained much interest in the field of pancreatology due to its potential prophylactic effect in post-ERCP pancreatitis. As a NSAID, indomethacin does have several associated adverse events, including most notably allergic reactions, gastrointestinal ulceration and bleeding, and renal impairment, which informed our exclusion criteria. As data are lacking regarding the effect of indomethacin on human pancreatic COX-2 activity, we selected a study dose of 50 mg BID of oral indomethacin, a standard clinical dosing for the treatment of other inflammatory conditions, to provide a steady-state administration of the drug. Prior studies have demonstrated similar blood levels achieved in oral vs. rectal administration of the medication and given that daily NSAID dosing is required to maintain suppression of COX-2, the oral form was chosen given its higher likelihood of compliance.^{16, 25} A single-dose regimen was avoided as it would be difficult to repeatedly assay the pancreatic fluid at various time points after single dose administration within each patient to create composite time curves. Should this study demonstrate successful reduction in pancreatic PGE₂ levels, chronic inhibition of pancreatic COX-2 activity may provide a therapeutic strategy for the treatment of chronic pancreatitis.

This study protocol leverages the resources of the NIH-supported CPDPC Consortium. An overriding objective of the Consortium is “to encourage translational research focusing upon elucidating the pathogenesis that will provide a basis for understanding the natural history and developing means of diagnosis, treatment, and clinical management of chronic pancreatitis.” A primary benefit of the Consortium’s resources is seen in the recruitment of subjects where enrollment of subjects with chronic pancreatitis into the PROCEED study is nearly complete, increasing the feasibility of this study.

Related to the CPDPC, a critical component of this study rests in the collection of biospecimens before and after treatment. These will be stored in a biorepository with the goal of potentially identifying other biomarkers that may help elucidate the inflammatory pathway in chronic pancreatitis and serve as indicators for response to potential therapies. As both study sites have operating biorepositories for the CPDPC, biospecimen handling and storage for this study can easily be integrated.

Lastly, should this study provide the physiologic and pharmacologic rationale and preliminary patient-reported outcomes data for the use of long-term oral indomethacin for the treatment of chronic pancreatitis, the study data will be utilized for the development and execution of a much larger clinical trial. The remaining study sites of the CPDPC Consortium would provide the natural study environment for conducting this larger trial and facilitate obtaining pre- and post-treatment imaging if subjects are part of the PROCEED study.²² Furthermore, the Consortium would facilitate sharing of the biospecimens for further exploratory research.

Limitations and Pitfalls

Performing endoscopy carries inherent risk related to the endoscopy itself as well as sedating the patient. This protocol, therefore, utilized the least invasive method of assessing human pancreatic COX-2 activity in performing pancreatic fluid collection via aspiration of duodenal fluid. The collection of other biospecimens including blood and saliva will also help identify whether PGE₂ levels can be measured less invasively.

Barriers to recruitment and retention include the lack of immediate benefit to subjects during the study period. Even if indomethacin does suppress pancreatic PGE₂ levels, long-term administration is likely required to improve symptoms and delay progression of chronic pancreatitis. Therefore, the study duration of 28 days was chosen for the primary purpose of generating preliminary data which does not currently exist, while minimizing the cost, time, and number of research interventions required. In line with this, as we were recruiting subjects who had an indication for an endoscopy or were receiving pancreatic function testing as part of their care, this reduces the recruitment pool and introduces sample bias. Furthermore, while the adverse event rate of indomethacin and placebo are relatively low, subjects are always at risk for drop-out due to adverse events or perceived lack of benefit with the treatment. Potential confounders include use of other medications for which subjects are asked to keep a medication diary and the natural history of chronic pancreatitis, which can fluctuate in regard to symptoms and pancreatic inflammation. Randomization and blinding are therefore performed to minimize the risk of confounding. Lastly, the small sample size limits sub-group analyses and heterogeneity within the study population may introduce bias.

In summary, the PAIR trial represents a first step in the examination of oral indomethacin as a novel therapeutic agent for the long-term treatment of chronic pancreatitis. This study will provide crucial pharmacologic and clinical data regarding oral indomethacin and preliminary data needed to develop a large-scale randomized study in our mission to develop therapies for the treatment of chronic pancreatitis. Furthermore, we aim to utilize the various biospecimens to identify noninvasive measures of inflammation and predict response to therapies such as NSAIDs in patients with chronic pancreatitis.

Institutional Review Board

This study received approval from the respective Institutional Review Boards at both study sites.

Dissemination policy

Throughout and upon completion of the study, deidentified data will be stored securely in a REDCap database. In compliance with the Dissemination of NIH-Funded Clinical Trial Information policy, this study has been registered on ClinicalTrials.gov and results from this trial will be submitted to ClinicalTrials.gov. In addition, every attempt will be made to publish results in peer-reviewed journals. Lastly, all researchers may request deidentified data from this study 5 years after the completion of the primary endpoint by contacting the Principal Investigators (PIs).

Safety Oversight

Safety oversight will be under the direction of a Data and Safety Monitoring Board (DSMB) composed of individuals with the appropriate expertise, including one gastroenterologist from each study site. Members of the DSMB will be independent from the study conduct and free of conflict of interest. The DSMB will meet via video conference on a quarterly basis to assess safety and efficacy data of the study, and will meet as needed when informed of study serious adverse events unanticipated problems. The DSMB will provide its input to the study PIs and IRB, and generate a quarterly DSMB report.

Trial Status

This trial was registered on ClinicalTrials.gov (NCT04207060). The 1^st subject was enrolled on 9/30/20.