Summary
Background:
The combination of rectally administered indomethacin and placement of a prophylactic pancreatic stent is recommended to prevent pancreatitis after endoscopic retrograde cholangiopancreatography (ERCP) in high-risk patients. Preliminary evidence suggests that the use of indomethacin may eliminate or substantially reduce the need for stent placement, a technically complex, costly, and potentially harmful intervention.
Methods:
In this multicenter, randomized, non-inferiority trial, patients at high risk for post-ERCP pancreatitis were randomly assigned to receive rectal indomethacin alone or the combination of indomethacin plus a prophylactic pancreatic stent. Patients, treating clinicians, and outcomes assessors were blinded to study group assignment. To declare non-inferiority, the upper bound of the two-sided 95% confidence interval for the difference in post-ERCP pancreatitis (indomethacin alone – indomethacin plus stent) would have to be less than 5% (non-inferiority margin).
Findings:
A total of 1950 patients were randomized. Post-ERCP pancreatitis occurred in 145 of 975 patients (14.9%) in the indomethacin alone group and in 110 of 975 (11.3%) in the indomethacin plus stent group (risk difference 3.6%; 95% confidence interval, 0.6%−6.6%; p=0.18 for non-inferiority). A post hoc intention-to-treat analysis of the risk difference between groups showed that indomethacin alone was inferior to the combination of indomethacin plus prophylactic stent (p=0.01). The relative benefit of stent placement was generally consistent across study subgroups but appeared more prominent among patients at highest risk for pancreatitis.
Interpretation:
For preventing post-ERCP pancreatitis in high-risk patients, a strategy of indomethacin alone was not as effective as a strategy of indomethacin plus prophylactic pancreatic stent placement.
Funding:
National Institutes of Health grant U01DK104833.
Introduction:
Endoscopic retrograde cholangiopancreatography (ERCP) – which is performed an estimated 500,000 times annually in the US alone – remains among the most dangerous endoscopic procedures, primarily due to the associated risk of acute pancreatitis. Despite recent advances in the prevention of this complication, post-ERCP pancreatitis continues to be a major problem in clinical practice, with rising mortality (1, 2).
For many years, the only effective preventive strategy in patients at high risk for this complication was the placement of a temporary, prophylactic stent in the pancreatic duct during ERCP (3–5). The prosthesis ensures adequate drainage in the face of pancreatic outflow tract edema that can develop during or after ERCP, reducing the potential for intraductal hypertension and consequent inflammation. More recently, nonsteroidal anti-inflammatory drugs (NSAIDs) – namely indomethacin and diclofenac administered rectally around the time of ERCP – have emerged as the first consistently effective pharmacologic prophylaxis for this condition (5,6). While their exact mechanism of action for preventing pancreatitis remains unclear, NSAIDs are believed to interfere with various components of the pancreatic inflammatory cascade (7–8). Rectal NSAIDs and pancreatic stent placement are the two most evidence-based prophylactic approaches and their combined use – in addition to aggressive intravenous fluid hydration – is guideline-recommended to prevent post-ERCP pancreatitis in high-risk cases (9–11).
Rectal NSAIDs are widely available, safe, and easy to administer. In contrast, prophylactic stent placement is technically complex, time consuming, and costly, partly because every patient requires an abdominal radiograph to ensure spontaneous stent passage and, in up to 20% of patients, a follow up endoscopy is necessary to retrieve a retained stent (12–14). If these small-caliber stents are left in situ for a long period of time, they may induce clinically relevant strictures of the pancreatic duct (15). Moreover, prophylactic stent placement is potentially harmful as the process can be traumatic in some patients and an unsuccessful placement attempt is associated with a significant increase in the risk of pancreatitis (16–17). Beyond this, the perceived benefit of prophylactic stent placement may be exaggerated as prior trials were un-blinded in design and conducted by a small number of expert centers, limiting generalizability to broader practice settings (18–20).
Preliminary studies have suggested that the administration of rectal NSAIDs may obviate the need for prophylactic stent placement (14,21). Considering the potential disadvantages of stents, and remaining uncertainty around their effectiveness, a strategy of pharmacoprevention alone could have major clinical and cost implications in ERCP practice (14). To reappraise the need for prophylactic stent placement among patients receiving rectal NSAIDS, we performed a comparative effectiveness trial of rectal indomethacin alone vs. the combination of indomethacin plus a prophylactic pancreatic stent in patients at high risk for developing post-ERCP pancreatitis.
Methods:
Study design
In this randomized, non-inferiority trial, we enrolled eligible patients at 20 referral centers in the United States and Canada after approval from the human studies review committee at each institution. An independent data and safety monitoring board (DSMB), established by the National Institutes of Health, provided regulatory and scientific oversight. A statistical and data coordinating center and a clinical coordinating center provided comprehensive study direction and monitoring. The complete study protocol is available in the Supplementary Appendix.
Patients
The eligibility criteria aimed to select patients at elevated risk for post-ERCP pancreatitis who required pancreatic stent placement for the sole purpose of pancreatitis prevention. Patients could be included if they met one or more of the following major criteria: a history of post-ERCP pancreatitis, difficult cannulation (defined as ≥6 cannulation attempts or ≥6-minute duration of cannulation), precut sphincterotomy (a procedure to facilitate biliary access when standard cannulation techniques are unsuccessful), pancreatic sphincterotomy, short duration (≤1 minute) balloon dilation of an intact biliary sphincter, or clinical suspicion of sphincter of Oddi dysfunction (as defined in the Supplementary Appendix). Patients were also eligible for inclusion if they met two or more minor criteria, including age <50 years and female sex, history of recurrent pancreatitis (≥2 episodes), or ≥3 pancreatic duct injections (Supplementary Appendix). The exclusion criteria are also listed in the Supplementary Appendix and were intended to exclude patients who are unfit for ERCP, undergo an ampullectomy, require a pancreatic stent for therapeutic purposes (e.g., treatment of a pancreatic leak or stricture), have an allergy to NSAIDs, have experienced acute pancreatitis within 1 week, or have an anticipated low risk of post-ERCP pancreatitis (e.g., biliary stent exchange through a pre-existing biliary sphincterotomy).
Intervention
All procedure-related interventions except for the administration of rectal indomethacin and placement of a prophylactic stent were left to the discretion of the endoscopist. During ERCP, once eligibility was confirmed, patients were randomized in a 1:1 ratio to receive a prophylactic stent or not. In patients assigned to the indomethacin plus stent group, endoscopists were expected to attempt stent placement, although the technical approach, including the amount of time and effort expended for this purpose, was not standardized. All patients received two 50 mg indomethacin suppositories. In the first 211 patients, indomethacin was administered immediately after the conclusion of the ERCP. However, in July of 2016, the timing of indomethacin administration was changed in response to the publication of a large-scale trial showing improved outcomes associated with pre-procedure rather than post-procedure drug delivery (22). Since reliably predicting the timing of an ERCP procedure in real-world clinical practice is challenging and to reconcile emerging evidence with existing data supporting the use of indomethacin post-procedurally (6), the protocol was amended after DSMB approval to administer the drug at the time of randomization, which was intra-procedural by protocol and in close proximity to the time of cannulation.
Blinding and post-ERCP care
To ensure unbiased care, anyone present during the ERCP and potentially aware of study group assignment (including the endoscopist, trainee, nursing staff, and research coordinator) could not be involved in the clinical care of the patient for at least 48 hours after the procedure, at which point the primary endpoint would very likely have developed. By protocol, the endoscopy procedure report did not state whether or not a stent was placed.
Patients were observed in the recovery area for as long as clinically appropriate after the procedure. Those with concern for a complication were admitted to the hospital (or for current inpatients, kept in the hospital). Clinical care during the observation period and subsequent hospitalization (if necessary) was directed by clinicians who were blinded to treatment group for at least 48 hours. Endoscopists had the option of implementing a post-procedure order set, as long as it was activated prior to randomization (to ensure that it was executed uniformly regardless of study group assignment). Endoscopists were also permitted to communicate with patients and caregivers after the procedure but could not provide any direct or indirect information about whether a pancreatic stent was placed.
Patients were assessed either in person or by telephone 3–5 days and then 30 (+/−5) days after randomization to capture outcome and safety events. Patients who received a prophylactic stent were contacted by un-blinded clinical personnel at least 5 days after randomization (when the primary outcome would have already developed) to arrange abdominal radiography to assess stent passage.
Outcomes
The primary outcome measure was post-ERCP pancreatitis, based on a widely-validated consensus definition that was applied as a diagnostic framework (23). In this consensus definition, post-ERCP pancreatitis is diagnosed if there was new onset (or increase) of pain in the upper abdomen, elevation in pancreatic enzymes of at least three times the upper limit of the normal 24 hours after the procedure, and hospitalization for at least 2 nights. The outcome was independently adjudicated by 3 ERCP experts at non-enrolling centers based on review of the medical records for study participants who were hospitalized with any adverse event within 2 days of the ERCP. To ensure unbiased assessment of the primary outcome, medical records were redacted of all information that could potentially reveal study group assignment, including radiology reports. The consensus definition was applied as a diagnostic framework (rather than a strict definition) so that adjudicators could use their best judgment in cases that did not strictly satisfy the criteria (e.g., pancreatic enzymes not collected until 2 days after ERCP). The secondary outcome measure was moderate-severe post-ERCP pancreatitis, also based on the consensus definition as a diagnostic framework (Supplementary Appendix). For the severity assessment, radiographic information was made available to the adjudicators. The primary and secondary outcomes were declared if there was agreement between at least 2 of the 3 adjudicators.
Statistical analysis
Given the potential disadvantages of prophylactic stent placement, we judged that indomethacin alone would become the dominant strategy in clinical practice if it resulted in no more than a 5% greater absolute risk of pancreatitis (non-inferiority margin) compared with the combination of indomethacin plus prophylactic stent. Additional justification for the non-inferiority margin is presented in the protocol and statistical analysis plan (Supplementary Appendix). Based on our prior large-scale trial in a similar patient population (6), we assumed that 9.7% of patients who receive indomethacin plus a prophylactic stent would experience post-ERCP pancreatitis. We thus estimated that a sample of 1430 patients would be required to give the trial at least 85% power to rule out a 5% higher rate of post-ERCP pancreatitis in the indomethacin alone arm, at a one-sided error rate of 0.025, and assuming 5% treatment crossover. The statistical plan included a prespecified blinded review of the sample size based on the observed percentage of pancreatitis in the first 1/3 of enrollment. The study design also included two planned interim analyses for futility (unlikely to show non-inferiority) using conditional power. In September of 2017, on the basis of higher than assumed overall pancreatitis and crossover rates in the first 472 participants, the DSMB recommended increasing the sample size to 2180 patients. In November of 2022, based on a requested blinded sample size and power re-estimation, the DSMB recommended termination of the study after enrollment of 1950 patients.
The primary analysis – an estimation of the risk difference in post-ERCP pancreatitis between the two groups – was performed in both the intention-to-treat population (all randomized patients) and the per-protocol population (patients without eligibility violation, indomethacin not administered, treatment crossover, or missing primary outcome). To declare non-inferiority, the upper bound of the two-sided 95% confidence interval for the difference in post-ERCP pancreatitis (indomethacin alone – indomethacin plus stent) needed to be less than 5% in both analysis populations. Summary measures with 95% confidence intervals were calculated for the secondary outcome and safety endpoints.
Exploratory subgroup analyses were performed to examine the risk difference in pancreatitis between indomethacin alone and the combination of indomethacin plus prophylactic stent across various pre-specified subgroups of patients in the trial, listed in the statistical analysis plan (Supplementary Appendix). A subgroup analysis of interest evaluated the treatment effect of indomethacin plus stent according to patients’ pretreatment risk for post-ERCP pancreatitis (24). Individual patient risk scores were determined by assigning one point for each major inclusion criterion and 0.5 points for each minor inclusion criterion (6). The risk score was examined as a categorical and continuous variable.
The funding agency (United States National Institutes of Health) was involved in the design and conduct of this trial, in the interpretation of data, and in writing and approval of the manuscript.
Results:
Patients
Between September of 2015 and January of 2023, a total of 1950 patients were randomized (Figure 1). Nine hundred and seventy-five patients were assigned to receive indomethacin alone and 975 to receive indomethacin and a prophylactic stent. Crossover occurred in 188 patients (19.3%) in the combination arm (who did not receive a stent because of inability to advance a wire into the pancreatic duct or to place the prosthesis) and in 16 patients (1.6%) in the indomethacin alone arm (who did receive a stent despite study group assignment). Seven patients (2 in the indomethacin group and 5 in the combination group) did not receive indomethacin. Primary and secondary outcome data were missing for two patients at the end of the study; one in each arm. Baseline characteristics were similar in the two groups (Table 1).
Figure 1:
Enrollment.
Table 1:
Characteristics of the patients at baseline.*
| Indomethacin + Stent N=975 | Indomethacin Alone N=975 | ||
|---|---|---|---|
| Age, Mean (SD) | 55.8 (16.3) | 55.6 (16.4) | |
| Male Sex, n (%) | 379 (38.9%) | 376 (38.6%) | |
| Race, n (%) | Asian | 19 (1.9%) | 17 (1.7%) |
| Black/African American |
102 (10.5%) | 115 (11.8%) | |
| White | 825 (84.6%) | 809 (83.0%) | |
| Other | 29 (3.0%) | 34 (3.5%) | |
| Hispanic or Latino, n (%) | 99 (10.2%) | 101 (10.4%) | |
| BMI1, Mean (SD) | 29.6 (7.2) | 28.6 (6.8) | |
| Antibiotic use in prior 3 months, n (%) | 366 (41.8%) | 363 (41.8%) | |
| Clinical suspicion or known sphincter of Oddi dysfunction, n (%) | 252 (25.8%) | 262 (26.9%) | |
| History of post-ERCP pancreatitis, n (%) | 24 (2.5%) | 36 (3.7%) | |
| History of recurrent pancreatitis, n (%) | 126 (12.9%) | 128 (13.1%) | |
| Difficult cannulation, n (%) | 823 (84.4%) | 795 (81.5%) | |
| Pre-cut (access) sphincterotomy, n (%) | 100 (10.3%) | 112 (11.5%) | |
| Number of pancreatic injections2, Median (IQR; Range) | 1 (0, 2; 0, 15) | 0 (0, 1; 0, 20) | |
| Pancreatic sphincterotomy, n (%) | 66 (6.8%) | 58 (5.9%) | |
| Pancreatic acinarization, n (%) | 8 (0.8%) | 6 (0.6%) | |
| Biliary sphincterotomy, n (%) | 869 (89.1%) | 864 (88.6%) | |
| Trainee involvement, n (%) | 546 (56.0%) | 555 (56.9%) | |
| Total intravenous fluid (in milliliters) received during peri-procedural period, mean (SD) | 1852 (938) | 1796 (935) | |
| Total intravenous lactated Ringer’s fluid (in millileters) received during peri-procedural period, mean (SD) | 1606 (1085) | 1554 (1052) | |
| Prophylactic pancreatic stent caliber in those who received a stent, n/N (%) | 3 French | 18/787 (2.3) | 0 |
| 4 French | 165/787 (21) | 4/16 (25) | |
| 5 French | 586/787 (74.5) | 9/16 (56) | |
| > 5 French | 18/787 (2.3) | 0 | |
| Missing | 0 | 3/16 (19) | |
| Prophylactic pancreatic stent length in those who received a stent, n/N (%) | 2–5 cm | 455/787 (57.8) | 7/16 (44) |
| 6–8 cm | 95/787 (12.1) | 3/16 (19) | |
| > 8 cm | 237/787 (30.1) | 3/16 (19) | |
| Missing | 0 | 3/16 (19) | |
Baseline characteristics were similar in the two treatment groups.
Precut sphincterotomy is performed to facilitate biliary access when standard cannulation techniques are unsuccessful.
Pancreatic acinarization occurs when excessive injection of contrast agent into the pancreatic duct results in opacification of pancreatic acini.
BMI missing in 3 subjects (indomethacin alone)
Number of pancreatic injections missing in 3 subjects (1 in indomethacin + stent, 2 in indomethacin alone)
Study outcomes
Post-ERCP pancreatitis occurred in 145 of 975 patients (14.9%) in the indomethacin alone group and in 110 of 975 (11.3%) in the indomethacin plus stent group (risk difference, 3.6%; 95% confidence interval, 0.6%−6.6%; p=0.18 for non-inferiority) (Figure 2). Results were similar in the per protocol population (Table 2). Because the upper bound of the 95% confidence interval around the risk difference in post-ERCP pancreatitis between the groups exceeded 5%, non-inferiority was not demonstrated. Since the confidence interval around the risk difference excluded zero for the intention-to-treat population, a post hoc one-sided test of the comparison of the two proportions indicated that indomethacin alone was inferior to indomethacin plus a prophylactic stent (p=0.01). The indomethacin alone group had a 32% higher risk of developing post-ERCP pancreatitis relative to the combination (RR 1.32; 95% CI 1.05–1.66). Moderate-severe post-ERCP pancreatitis (secondary outcome) occurred in 78 patients (8.0%) in the indomethacin alone group and in 58 patients (6.0%) in the combination group (risk difference, 2.1%; 95% confidence interval, −0.2%−4.3%). Severe pancreatitis occurred in 20 patients (2.1%) in the indomethacin group and 14 patients (1.4%) in the combination group. Pancreatitis-related death occurred in 3 patients in the indomethacin alone group and in none of the patients in the combination group. Non-pancreatitis safety outcomes did not differ between groups (Table 2). Among the 188 patients assigned to receive a prophylactic stent but in whom stent placement was unsuccessful, the rate of post-ERCP pancreatitis was 10.1% (95% CI: 5.8%−14.4%).
Figure 2:
Risk difference in post-ERCP pancreatitis between indomethacin alone and indomethacin plus a prophylactic stent.*
*Because the upper bound of the 95% confidence interval around the risk difference in post-ERCP pancreatitis between the groups was more than 5%, non-inferiority was not demonstrated. Since the lower bound of the risk difference favoring the combination group was >0, indomethacin alone was found to be inferior to indomethacin plus a prophylactic stent.
*Pi-Pi+s represents the difference in the primary outcome (the proportion of patients with post-ERCP pancreatitis) between patients assigned to indomethacin alone and those assigned to indomethacin plus stent; CI denotes confidence interval.
Table 2:
Incidence of the primary, secondary, and safety outcomes.*
| Intent-to-Treat Analysis N=1950 | Per Protocol Analysis N=1728 | ||||
|---|---|---|---|---|---|
| Primary Outcome | Indomethacin + Stent N=975 | Indomethacin Alone N=975 | Indomethacin + Stent N=777 | Indomethacin Alone N=951 | |
| Presence of Post-ERCP | N (%) | 110 (11.3%) | 145 (14.9%) | 90 (11.6%) | 137 (14.4%) |
| Acute Pancreatitis | 95% CI | (9.3%, 13.3%) | (12.6%, 17.1%) | (9.3%, 13.8%) | (12.2%, 16.6%) |
| RD (95% CI) | 3.6% (0.6%, 6.6%) | 2.8% (−0.3%, 6.0%) | |||
| Secondary and Safety Outcomes | |||||
| Moderate-Severe Post-ERCP Pancreatitis1 | 58 (6.0%) | 78 (8.0%) | 45 (5.8%) | 74 (7.8%) | |
| RD (95% CI) | 2.1% (−0.2%, 4.3%) | 2.0% (−0.4%, 4.4%) | |||
| Severe pancreatitis − n (%) | 14 (1.4%) | 20 (2.1%) | 12 (1.5%) | 19 (2.0%) | |
| RD (95% CI) | 0.6% (−0.5%, 1.8%) | 0.5% (−0.8%, 1.7%) | |||
| Pancreatitis-related Death − n (%) | 0 (0%) | 3 (0.3%) | 0 (0%) | 3 (0.3%) | |
| RD (95% CI) | 0.3% (0%, 0.7%) | 0.3% (0%, 0.7%) | |||
| ICU Admission− n (%) | 39 (4.0%) | 29 (3.0%) | 27 (3.5%) | 29 (3.0%) | |
| RD (95% CI) | −1% (−2.7%, 0.6%) | −0.4% (−2.1%, 1.3%) | |||
| Hospital length of stay | N | 975 | 975 | 777 | 951 |
| Mean (SD) | 2.9 (6.7) | 3.2 (7.2) | 2.8 (6.7) | 3.2 (7.1) | |
| Median (Min, Max) | 0.0 (0.0, 71.0) | 0.0 (0.0, 86.0) | 0.0 (0.0, 62.0) | 0.0 (0.0, 86.0) | |
| Difference in Means (95% CI) | 0.4 (−0.3, 1.0) | 0.3 (−0.3, 1.0) | |||
The intention-to-treat analysis included all randomized subjects; RD denotes risk difference for the indomethacin alone group minus the indomethacin plus stent group; CI denotes confidence interval.
The per protocol analysis excluded randomized subjects with any of the following predefined types of protocol deviations: eligibility violations, indomethacin not administered, treatment crossover, or missing outcome data.
Moderate-severe post-ERCP pancreatitis missing for 2 subjects (1 indomethacin + stent, 1 indomethacin alone) and excluded from percentage denominator.
Subgroup analyses
The favorable effect of indomethacin plus stent on the primary outcome was generally consistent across the prespecified subgroups, including those who experienced a difficult cannulation (Figure 3a). The combination appeared most beneficial in patients with a prior history of post-ERCP pancreatitis and when pancreatic sphincterotomy was performed. The relative benefit of indomethacin plus stent appeared to be greatest for the highest risk score category although the difference from the other risk categories was not statistically significant (Figure 3b). The absolute risk reduction associated with indomethacin plus stent ranged from a number needed to treat (NNT) to prevent one case of ERCP-related pancreatitis of 44 when the risk score was 1–1.5, to 38 when the risk score was 2–2.5, to 7 (95% CI: 3.9–56) when the risk score was 3+. The finding of increasing benefit with rising risk of post-ERCP pancreatitis was also observed when risk score was treated as a continuous variable (Supplement Figure 1).
Figure 3:
Exploratory subgroup analyses.
3a. The primary outcome was generally consistent across the prespecified subgroups. *The following subgroups had a statistically significant interaction with indomethacin + stent: biliary sphincterotomy (p=0.01), pancreatic sphincterotomy (p=0.006), double wire technique (p=0.004), and prior post-ERCP pancreatitis (p=0.04).
3b. The relative benefit of indomethacin plus stent increased with rising pretreatment risk score categories. The absolute risk reduction associated with indomethacin plus stent ranged from a number needed to treat (NNT) to prevent one case of ERCP-related pancreatitis of 44 when the risk score was 1–1.5, to 38 when the risk score was 2–2.5, to 7 when the risk score was 3+.
*Individual patient risk scores were determined by assigning 1.0 point for each major inclusion criterion and 0.5 points for each minor inclusion criterion, as outlined in the Methods section.
Discussion:
In this large, multicenter, comparative effectiveness trial, we did not demonstrate that a strategy of indomethacin alone is non-inferior to a strategy of indomethacin plus prophylactic pancreatic stent placement for preventing post-ERCP pancreatitis in high-risk cases. In fact, the intention-to-treat analysis showed that indomethacin alone was less effective than the combination and there were numerically fewer patients in the combination group who experienced severe post-ERCP pancreatitis and pancreatitis-related death.
The relative benefit of indomethacin plus stent was generally consistent across subgroups of patients in the trial, although cases at highest risk for pancreatitis appeared to derive the most benefit, such as when there was a prior history of post-ERCP pancreatitis, when a pancreatic sphincterotomy was performed, and among patients with higher post-ERCP pancreatitis risk scores. However, even among patients who experienced a difficult cannulation – generally considered to be at the lower end of the high-risk spectrum but to whom the findings of this trial are most likely to apply – noninferiority was not observed.
This trial compared a strategy of indomethacin alone versus a strategy of indomethacin plus prophylactic stent. A comparison of strategies is particularly informative for this clinical question because stent placement is not always technically successful, and a failed attempt can increase pancreatitis risk above baseline (15–16). Therefore, clinicians are faced with the question of whether to attempt prophylactic stent placement despite the possibility of failure and its consequences. In this trial, a favorable effect of prophylactic stenting was observed despite failure to place the stent in approximately 20% of patients who were assigned to receive one. Thus, even when accounting for the possibility of failure, a strategy of indomethacin plus a prophylactic stent is justified in high-risk cases.
This ~20% failure rate is higher than reported in prior trials (4) and may reflect the wide variability in experience levels, technical proficiency, and enthusiasm for prophylactic stenting among the >100 participating endoscopists, contributing to the generalizability of the findings. The observed benefit despite a relatively high failure rate suggests that the effectiveness of pancreatic stent placement could be augmented by interventions at the training and post-graduate levels to enhance the adoption of, and technical proficiency with stent placement. It is important to consider, though, that overambitious efforts to place a stent may result in net harm because the pancreatic trauma incurred in such cases may not be counteracted by the effect of the stent. However, our per-protocol analysis – which would generally be expected to exaggerate the benefit of stent placement by excluding those with failed attempts – showed similar findings to the primary analysis, highlighting the complex relationship between prophylactic stents and risk reduction and indicating that, in this trial, unsuccessful stent placement did not appear to be associated with an increased risk of pancreatitis. Additional research will be critical to elucidate the optimal procedural approach – including time and effort expended – for stent placement and the optimal combination of prophylactic interventions, including aggressive intravenous fluid hydration.
In this trial, we used a 5% non-inferiority margin on the basis of clinical and statistical reasoning. The investigators initially considered a 10% margin to achieve a more practical sample size. However, the hypothesis of this trial was that indomethacin would obviate the need for prophylactic stent placement, a major deviation from guideline-recommendations and common clinical practice at the time the trial was designed. Therefore, we judged that in order to influence clinical practice away from this well-established and routinely practiced intervention, the increased risk of post-ERCP pancreatitis associated with withholding a prophylactic stent would have to be minimal, and that a larger margin would not be in the best interest of the field.
In the last decade, however, there has been growing uncertainty about the role of prophylactic pancreatic stent placement in clinical practice and its use has decreased substantially (25–27). The findings of this trial indicate that this widespread abandonment of prophylactic stent placement is not justified. While the decision of whether the effect size observed in this study is worthwhile in specific cases remains a value judgement, the ~30% relative risk reduction in pancreatitis with numerically fewer cases of severe PEP and death, and the more pronounced benefit in higher-risk patients, indicate that prophylactic stent placement continues to play an important role in clinical practice.
The validity of our findings is augmented by addressing two important limitations in the design and execution of prior trials evaluating prophylactic stent placement – their open-label design and conduct at a limited number of expert centers. Because the development of post-ERCP pancreatitis can be affected by unequal co-interventions between study groups and because its definition is somewhat subjective (related to the interpretation of pain and the decision to hospitalize patients), we implemented rigorous study operations that aimed to ensure blinding at the level of the participant, anyone caring for the patient after ERCP, and outcomes adjudicator. Additionally, the study was conducted across many medical centers, with >100 participating endoscopists. Other than the randomization interventions (indomethacin and stent), procedural approach was not standardized, aiming to maximize generalizability by emulating routine clinical practice to the greatest extent possible. Indeed, in this trial the post-ERCP pancreatitis rate among patients receiving indomethacin alone was higher than that we observed in our prior RCT (6). This difference in PEP rates is likely explained by the design of this RCT in which there was blinded, independent adjudication of outcomes, and a much larger pool of endoscopists with varying levels of expertise and experience – which is more likely to produce an accurate estimate of the rate of PEP in this patient population in real-world practice. It also should be noted that this trial included a blinded sample size re-estimation which preserved the desired power of the study when it was identified that the assumed crossover and post-ERCP pancreatitis rates were underestimated. In fact, the DSMB recommended increasing the sample size based on this re-estimation and after closely monitoring the power of the trial made the final recommendation to stop enrollment when 1,950 participants were randomized.
Despite aiming to mitigate bias to the greatest extent possible, the results of this trial should be interpreted in the context of important limitations related to blinding and generalizability. First, despite extensive efforts, potential unblinding events did occur (most commonly through radiographic tests obtained within 48 hours after ERCP) and could have influenced care. Additionally, since clinicians who are aware of study group assignment could resume care of their patients after 48 hours, there was a decay in blinding rigor for the secondary outcome. Similarly, adjudicators were provided radiographic information when assessing the severity of post-ERCP pancreatitis, which may have influenced their decision-making. Additionally, endoscopists performing the ERCP may have made procedural decisions with knowledge of study group assignment, and it is possible that procedural approach could have been altered according to whether or not a stent would be placed. Lastly, all participating sites were tertiary referral centers and the findings therefore may not be applicable to community endoscopists.
In summary, a strategy of indomethacin alone was not as effective as a strategy of indomethacin plus prophylactic pancreatic stent placement. The relative benefit of stent placement appeared more prominent among those at highest risk for post-ERCP pancreatitis.
Supplementary Material
Research in context.
Evidence before this study
Acute pancreatitis is the most common and potentially devastating complication of endoscopic retrograde cholangiopancreatography (ERCP). Despite important advances in the prevention of this complication, post-ERCP pancreatitis still occurs in up to 15% of high-risk cases and mortality associated with this condition appears to be rising. We searched PubMed, Ovid, and Cochrane Library electronic databases and ClinicalTrials.gov (to identify ongoing trials) between Jan 1, 1990, and August 30, 2023, using the search terms “post-ERCP complications”, “post-ERCP pancreatitis”, “post-ERCP pancreatitis prevention”, “prophylactic pancreatic stent”, “NSAIDs”, and “indomethacin” with no restrictions on study type or language. Both prophylactic pancreatic stent placement and the administration of rectal indomethacin have independently been shown in clinical trials to reduce the incidence and severity of post-ERCP pancreatitis, and their combination is recommended for patients at elevated risk for this complication. While rectal indomethacin is widely available, safe, and easy to administer, prophylactic pancreatic stent placement is technically complex, time consuming, costly, and potentially harmful in certain situations. On this basis, and because hypothesis generating studies have suggested that the administration of indomethacin may obviate the need for prophylactic stenting, the use of stents has decreased substantially in clinical practice, perhaps contributing to the rising mortality associated with this condition. A large-scale, methodologically rigorous, comparative effectiveness trial had not been performed but is necessary to provide clarity on the role of prophylactic stent placement in clinical practice.
Added value of this study
In this 20-center, randomized, non-inferiority trial, patients at high risk for post-ERCP pancreatitis were assigned to receive indomethacin alone or the combination of indomethacin plus a prophylactic pancreatic stent. We found that post-ERCP pancreatitis occurred more commonly in the indomethacin alone group compared to the indomethacin plus stent group, and this difference was statistically significant. Thus, non-inferiority was not declared and in fact, indomethacin alone was found to be less effective than the combination. The relative benefit of stent placement was generally consistent across study subgroups but appeared more prominent among patients at highest risk for pancreatitis. The validity of these findings is augmented by addressing two important limitations in the design of prior trials evaluating prophylactic stent placement – their open-label design and conduct at a limited number of expert centers. Because the development of post-ERCP pancreatitis can be affected by unequal co-interventions between study groups and because its definition is somewhat subjective (related to the interpretation of pain and the decision to hospitalize patients), we implemented rigorous study operations that aimed to ensure blinding at the level of the participant, clinician caring for the patient after ERCP, and outcomes adjudicator. Additionally, the study was conducted across a large number of medical centers, with >100 participating endoscopists, maximizing the generalizability of the findings.
Implications of all the available evidence
The findings of this trial support the use of prophylactic pancreatic stent placement in addition to rectal indomethacin for preventing pancreatitis in high-risk patients, in accordance with existing international clinical practice guidelines. These findings should reverse the widespread abandonment of prophylactic pancreatic stenting which has been observed recently in the absence of an evidence-base to support it. In this trial, stents appeared protective in the large subgroup of patients who experienced a difficult cannulation during ERCP – a common situation in which risk is increased to some extent – but those at highest risk for the complication appeared to derive the most benefit, underscoring the importance of this intervention in the most vulnerable patients especially.
Acknowledgements:
This trial was supported by grant U01DK104833 from the National Institutes of Health (NIH). The findings are those of the authors and do not necessarily represent the official views of the NIH. We would like to express our deepest gratitude to the patients who participated in this trial, to the members of the Data and Safety Monitoring Board (Supplementary Appendix), and to Ms. Rebecca Torrance and Dr. Frank Hamilton from the NIH.
Declaration of interests:
The following authors have received honoraria, consulting fees, or research support from companies that manufacture prophylactic pancreatic stents or indomethacin within the last 2 years: S. Edmundowicz, V. Kushnir, N. Forbes, F. Willingham, S. Varadarajulu. J. Bang, A. Rastogi, M. Khashab, A. Wang, S. Gordon, J. Buxbaum, R. Keswani, R. Shah, A. Aadam, Y. Chen, A. Barkun, A. Ross, G. Papachristou, G. Cote. The other authors have no conflicts of interest to disclose pertaining to this manuscript.
Footnotes
ClinicalTrials.gov identifier: NCT02476279
Data Sharing Statement:
All individual de-identified participant data collected during the conduct of this trial and a corresponding data dictionary will be shared as part of a public use dataset in accordance with US National Institutes of Health policies. The study protocol and statistical analysis plan will also be shared. The data will become available in perpetuity 1 year after publication of this manuscript to any interested party through the NIH website.
References:
- 1).Akshintala VS, Kanthasamy K, Bhullar FA, et al. Incidence, Severity and Mortality of Post ERCP Pancreatitis: An Updated Systematic Review and Meta-Analysis of 145 Randomized Controlled Trials. Gastrointest Endosc 2023:S0016-5107(23)00363-2. doi: 10.1016/j.gie.2023.03.023 [DOI] [PubMed] [Google Scholar]
- 2).Mutneja HR, Vohra I, Go A, et al. Temporal trends and mortality of post-ERCP pancreatitis in the United States: a nationwide analysis. Endoscopy 2021;53:357–366. [DOI] [PubMed] [Google Scholar]
- 3).Tarnasky PR, Palesch YY, Cunningham JT, et al. Pancreatic stenting prevents pancreatitis after biliary sphincterotomy in patients with sphincter of Oddi dysfunction. Gastroenterology 1998;115:1518–24. [DOI] [PubMed] [Google Scholar]
- 4).Mazaki T, Mado K, Masuda H, Shiono M. Prophylactic pancreatic stent placement and post-ERCP pancreatitis: an updated meta-analysis. J Gastroenterol 2014; 49:343–55. [DOI] [PubMed] [Google Scholar]
- 5).Akshintala VS, Sperna Weiland CJ, Bhullar FA, et al. Non-steroidal anti-inflammatory drugs, intravenous fluids, pancreatic stents, or their combinations for the prevention of post-endoscopic retrograde cholangiopancreatography pancreatitis: a systematic review and network meta-analysis. Lancet Gastroenterol Hepatol 2021;6:733–742. [DOI] [PubMed] [Google Scholar]
- 6).Elmunzer BJ, Scheiman JM, Lehman GA, et al. A randomized trial of rectal indomethacin to prevent post-ERCP pancreatitis. N Engl J Med 2012;366:1414–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7).Gross V, Leser HG, Heinisch A, et al. Inflammatory mediators and cytokines—new aspects of pathophysiology and assessment of severity of acute pancreatitis? Hepatogastroenterology 1993;40:522–30. [PubMed] [Google Scholar]
- 8).Makela A, Kuusi T, Schroeder T. Inhibition of serum phospholipase A2 in acute pancreatitis by pharamacologic agents in vitro. Scand J Clin Lab Invest 1997;57:401–8. [DOI] [PubMed] [Google Scholar]
- 9).Dumonceau JM, Kapral C, Aabakken L, et al. ERCP-related adverse events: European Society of Gastrointestinal Endoscopy (ESGE) Guideline. Endoscopy 2020;52:127–149. [DOI] [PubMed] [Google Scholar]
- 10).Mine T, Morizane T, Kawaguchi Y, et al. Clinical practice guideline for post-ERCP pancreatitis. J Gastroenterol 2017;52:1013–1022. [DOI] [PubMed] [Google Scholar]
- 11).Buxbaum JL, Freeman M, Amateau SK, et al. American Society for Gastrointestinal Endoscopy guideline on post-ERCP pancreatitis prevention strategies: summary and recommendations. Gastrointest Endosc 2023;97:153–162. [DOI] [PubMed] [Google Scholar]
- 12).Das A, Singh P, Sivak MV Jr, et al. Pancreatic-stent placement for prevention of post-ERCP pancreatitis: a cost-effectiveness analysis. Gastrointest Endosc 2007; 65: 960–968. [DOI] [PubMed] [Google Scholar]
- 13).Zolotarevsky E, Fehmi SM, Anderson MA et al. Prophylactic 5-Fr pancreatic duct stents are superior to 3-Fr stents: a randomized controlled trial. Endoscopy 2011;43:325–330. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14).Elmunzer BJ, Higgins PD, Saini SD, et al. Does rectal indomethacin eliminate the need for prophylactic pancreatic stent placement in patients undergoing high-risk ERCP? Post hoc efficacy and cost-benefit analyses using prospective clinical trial data. Am J Gastroenterol 2013;108:410–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15).Bakman YG, Safdar K, Freeman ML. Significant clinical implications of prophylactic pancreatic stent placement in previously normal pancreatic ducts. Endoscopy 2009;41:1095–8. [DOI] [PubMed] [Google Scholar]
- 16).Freeman ML, Overby C, Qi D. Pancreatic stent insertion: consequences of failure and results of a modified technique to maximize success. Gastrointest Endosc 2004; 5:8–14. [DOI] [PubMed] [Google Scholar]
- 17).Choksi NS, Fogel EL, Cote GA, et al. The risk of post-ERCP pancreatitis and the protective effect of rectal indomethacin in cases of attempted but unsuccessful prophylactic pancreatic stent placement. Gastrointest Endosc 2015;81:150–5. [DOI] [PubMed] [Google Scholar]
- 18).Tarnasky PR, Palesch YY, Cunningham JT, et al. Pancreatic stenting prevents pancreatitis after biliary sphincterotomy in patients with sphincter of Oddi dysfunction. Gastroenterology 1998;115:1518–24. [DOI] [PubMed] [Google Scholar]
- 19).Chahal P, Tarn asky PR, Petersen BT, et al. Short 5Fr vs long 3Fr pancreatic stents in patients at risk for post-endoscopic retrograde cholangiopancreatography pancreatitis. Clin Gastroenterol Hepatol 2009;7:834–9. [DOI] [PubMed] [Google Scholar]
- 20).Cha SW, Leung WD, Lehman GA, et al. Does leaving a main pancreatic duct stent in place reduce the incidence of precut biliary sphincterotomy-associated pancreatitis? A randomized, prospective study. Gastrointest Endosc 2013;77:209–16. [DOI] [PubMed] [Google Scholar]
- 21).Akbar A, Abu Dayyeh BK, Baron TH, et al. Rectal nonsteroidal anti-inflammatory drugs are superior to pancreatic duct stents in preventing pancreatitis after endoscopic retrograde cholangiopancreatography: a network meta-analysis. Clin Gastroenterol Hepatol 2013; 11:778–83. [DOI] [PubMed] [Google Scholar]
- 22).Luo H, Zhao L, Leung J, et al. Routine pre-procedural rectal indometacin versus selective post-procedural rectal indometacin to prevent pancreatitis in patients undergoing endoscopic retrograde cholangiopancreatography: a multicentre, single-blinded, randomised controlled trial. Lancet 2016;387:2293–2301. [DOI] [PubMed] [Google Scholar]
- 23).Cotton PB, Lehman G, Vennes J, et al. Endoscopic sphincterotomy complications and their management: an attempt at consensus. Gastrointest Endosc 1991;37:383–93. [DOI] [PubMed] [Google Scholar]
- 24).Kent DM, Rothwell PM, Ioannidis JP, Altman DG, Hayward RA. Assessing and reporting heterogeneity in treatment effects in clinical trials: a proposal. Trials 2010;11:85–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25).Smith ZL, Elmunzer BJ, Cooper GS, Chak A. Real-World Practice Patterns in the Era of Rectal Indomethacin for Prophylaxis Against Post-ERCP Pancreatitis in a High-Risk Cohort. Am J Gastroenterol 2020;115:934–940. [DOI] [PubMed] [Google Scholar]
- 26).Issak A, Elangovan A, Ferguson RD, Waghray N, Sandhu DS. Underutilization of prophylactic rectal indomethacin and pancreatic duct stent for prevention of post-ERCP Pancreatitis. Endosc Int Open 2021;9:E979–E985. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27).Ashat M, Kandula S, Cote GA, et al. Utilization Pattern of Prophylactic Measures for Prevention of Post-ERCP Pancreatitis: A National Survey Study. Gastrointest Endosc. 2023. Feb 2:S0016-5107(23)00088-3. doi: 10.1016/j.gie.2023.01.049. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All individual de-identified participant data collected during the conduct of this trial and a corresponding data dictionary will be shared as part of a public use dataset in accordance with US National Institutes of Health policies. The study protocol and statistical analysis plan will also be shared. The data will become available in perpetuity 1 year after publication of this manuscript to any interested party through the NIH website.