Reply to three letters: Dubravcsik et al, Murion et al and Fan et al
The horse must come before the cart …

We very much appreciate the interest generated by our recent publication in Endoscopy entitled “Urgent ERCP with pancreatic stent placement or replacement for salvage of post-ERCP pancreatitis” [1]. In the paper, we presented the second ever published case series involving placement of a pancreatic stent shortly after onset of pancreatitis, with the aim of ameliorating the course of predicted severe post-endoscopic retrograde cholangiopancreatography (ERCP) pancreatitis (PEP). Patients included were those without a pancreatic stent, and those patients in whom a pancreatic stent had migrated out prematurely. Dramatic resolution of predicted severe pancreatitis occurred in all patients. Three letters were sent to the Editor regarding this paper, and we have been asked to respond to all three collectively.

Murino et al. were concerned that the selection of patients for salvage represented only 14 of the 57 patients in the cohort, and wondered whether this introduced selection bias; they would have liked to have seen a comparison of outcomes with those not offered salvage therapy. Our response to this question is as follows. As clearly stated in the abstract and methods section, the 14 patients were selected for early repeat ERCP because they had predictors of severe pancreatitis, as defined by extreme pain and dramatic elevations of pancreatic enzymes combined with systemic inflammatory response syndrome (SIRS) and a bedside index for severity in acute pancreatitis (BISAP) score of ≥ 3. Collectively, SIRS and BISAP scores are now considered to be the best early predictors of severity of acute pancreatitis [2]. The other patients, who were not offered salvage pancreatic stenting, did not have such predictors of severe pancreatitis, and therefore would represent a very different comparison group with limited validity as their progress and condition were not a cause for concern. In terms of the definition of pancreatitis – pain and elevated amylase/lipase at 24 hours – we agree that patients undergoing salvage ERCP before that time did not technically meet the definition of PEP. However, waiting 24 hours would likely have been too late. The concept of this intervention, as with placing prophylactic pancreatic stents in the first place, or giving rectal nonsteroidal anti-inflammatory drugs (NSAIDS), is to interrupt the inciting process that leads to the pancreatitis inflammatory process before it has had time to evolve. Finally, Murino et al. raised concerns that the data should be considered carefully until a well-performed randomized trial can be conducted. We agree. In the final sentences of the paper, we stated that “Confirmation of safety and efficacy of salvage stenting would require a prospective controlled trial in which patients with predicted moderate to severe early PEP were randomized to early intervention with pancreatic stent placement or replacement compared with conservative management. Until then, the current data should be considered as hypothesis generating rather than a recommendation for practice.”

Ma et al. raised a similar suggestion for a randomized controlled trial. They suggest that we should have randomized the patients to intervention or standard conservative management. We would like to point out that prior to our paper, the only prior published report consisted of a similar paper by Madácsy et al., which involved only six patients [3]. We presented our paper as a hypothesis-generating study. If both series were combined, the total number of patients reported to undergo salvage ERCP is a mere 20. To conduct a valid randomized trial, we would first require treatment effect estimates compared with control, which is essential to generate meaningful sample size calculations. The treatment effect can only be estimated from previous treatment and data. In other words, the horse has to come before the cart!

Finally, Dubravscik et al., the very authors responsible for the pioneer report of “salvage” ERCP, raised yet more valid points, and also emphasized the need for a randomized controlled trial. We agree with their perspective, as stated in our final sentence in the discussion mentioned above. We do, however, disagree with one point. Very early outward stent migration was clearly temporally associated with moderately delayed onset PEP in our trial, as evidenced by normal serum lipase and amylase at 2 hours in many of the patients with protective pancreatic stents who within the next day developed severe pain and hyperenzymemia, correlating with documented early passage of the pancreatic stent. Delayed pancreatitis in such patients is not consistent with the natural course of acute recurrent pancreatitis, as only 3 of 14 patients undergoing salvage ERCP had such an indication for the original ERCP. The duration required for effective prophylactic pancreatic stenting has never been investigated, especially in patients suffering thermal injury to the pancreatic sphincter from interventions such as ampullectomy and pancreatic sphincterotomy. We suggest that such patients require a more prolonged duration of prophylactic stenting to allow thermal-induced edema to subside, and that the finding of more delayed onset of PEP in patients with a pancreatic stent, and especially in those with migrated stent, further reflects the importance of maintaining pancreatic drainage for at least several days. We also agree with the authors concern that failed placement of a pancreatic stent will induce more damage than no intervention, as shown by our group 10 years ago [4], and reinforced by a recent paper from the University of Michigan/Indiana University [5]. Thus, salvage ERCP is only likely to be of potential benefit in centers with extensive experience and success with pancreatic stent placement.

To summarize, we agree with all three authors writing letters to the Editor that a randomized controlled trial is in order. Challenges will be: 1) accurate estimation of treatment effect; 2) adequate sample size calculations, which will likely be very large; 3) difficulty recruiting adequate numbers of patients at expert centers because of the rarity of predicted severe PEP; and 4) funding for such a trial. We encourage all those working in the global ERCP community to consider undertaking such a trial. In the meantime, it seems reasonable to consider salvage stenting only for patients with predicted severe PEP as determined by SIRS and BISAP score, and only at expert centers with a proven very high rate of success at pancreatic stent placement.

• References

• 1 Kerdsirichairat T, Attam R, Arain M et al. Urgent ERCP with pancreatic stent placement or replacement for salvage of post-ERCP pancreatitis. Endoscopy 2014; 46: 1085-1091
• 2 Working Group IAP/APA. IAP/APA evidence-based guidelines for the management of acute pancreatitis. Pancreatology 2013; 13: e1-15
• 3 Madácsy L, Kurucsai G, Joó I et al. Rescue ERCP and insertion of a small-caliber pancreatic stent to prevent the evolution of severe post-ERCP pancreatitis: a case-controlled series. Surg Endosc 2009; 23: 1887-1893
• 4 Freeman ML, Overby CS, Qi DF. Pancreatic stent insertion: consequences of failure, and results of a modified technique to maximize success. Gastrointest Endosc 2004; 59: 8-14
• 5 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 2014; 81: 150-155