Covered self-expanding metal stents for transmural drainage of infected pancreatic fluid collections: don't sell the skin till you have caught the bear or at least spotted the animal

We read with interest the article by Fabbri et al. entitled “Endoscopic ultrasound-guided transmural drainage of infected pancreatic fluid collections with placement of covered self-expanding metal stents: a case series” [1]. The authors state that from their experience in 18 patients endoscopic ultrasound-guided transmural drainage (EUS-GTD) of (infected) pancreatic fluid collections using covered self-expanding metal stents (CSEMSs) is easier, faster, and safer, and likely to be more cost-effective.

“A wise man,” wrote Hume, “proportions his belief to the evidence.” Although the concept of using CSEMSs to drain pancreatic fluid collections as presented in this case series is appealing, the real questions to be asked are: (i) how firm is the evidence, and (ii) should we all be using this approach from now on?

Our first comment challenging the paper by Fabbri et al. [1] refers to the methodology used to uphold their proposed conclusion. Case series can be considered of value in describing a clinical condition or to generate hypotheses for further study. However, because of potential selection, observation, performance, and publication bias, case series can almost never be relied upon to draw final conclusions. Even more crucial, lack of comparison, which reveals associations by exposing differences, can make case series appear to show an association between an intervention and an outcome when, in fact, there is none. So stating that there is an increased efficacy in terms of applicability, complications, resolution, and use of resources by using CSEMSs for EUS-GTD without a proper comparison with the supposed gold standard of using multiple 10-Fr plastic stents gives their study too much credit.

A second concern is the excessive devaluation of the methodology considered to be the current gold standard. First of all, the authors mention that “for nonfluid collections, the insertion of only one 10-Fr plastic stent will often result in inadequate drainage.” Yet, they use the presence of organized pancreatic necrosis (with minimal fluid component) as an exclusion criterion for the insertion of a CSEMS in their series, which makes this specific counterargument superfluous.

Next, they highlight that superinfection occurs quite frequently because of obstruction of plastic stents. This was indeed a problem in the past but has been largely overcome by inserting multiple stents, a practice which has been widely adopted since the analysis of Cahen et al. in 2005 [2]. Moreover, the overall reported technical success, clinical success, recurrence, and complication rates with EUS-GTD using plastic stents amount to 93.8 %, 90.6 %, 12.5 %, and 10.7 %, respectively [3]. There is not such an impressive difference between these figures and those of the currently reported series using CSEMSs, with an intention-to-treat success rate of 90.9 %, clinical success of 88.8 %, recurrence of 5 %, and complication rate of 16 %.

Third, the emphasis on the fact that the larger diameter access cystenterostomy provided by the CSEMSs contributed to the favourable outcome remains speculative without a direct comparison to multiple plastic stents. Furthermore, there was no uniform choice of type and size of CSEMSs because two different types and lengths were used.

Fourth, the argument that the use of CSEMSs would overcome the problem of limited endoscopic view when endoscopically “drowning” while evacuating pus or fluid from a pancreatic fluid collection seems somewhat oversimplified because the manner of stent insertion is not especially different nor does it preclude the need for vigorous monitoring of the correct positioning for the cystenterostomy. It is preferable that the latter is done under direct visual control. This is usually possible if one is capable of waiting and applies suction for long enough, but in more difficult cases it can be done under radiographic control using stents with radiopaque markers, which are commercially available for both plastic and metal stents.

A third consideration is cost. The authors state that their approach with CSEMSs is likely to be cost-effective owing to the savings made by reduction in the time to resolution, number of re-interventions, length of hospital stay, and patient discomfort. Again this premise without comparative data or at least without mention of a comparative cost analysis with a historical control group is given too much weight. Additionally, if you consider that, at least in Belgium, a CSEMS costs about 1000 euros whereas two plastic pigtails cost 100 euros and that the study of Arvanitakis [4] showed that leaving plastic stents in (without any obligatory re-intervention) was effective and did not cause discomfort (except in one patient), we question the cost-effectiveness of using CSEMSs.

A fourth and final remark concerns the prophylactic use of antifungals for 5 days when performing the EUS-GTD. In none of the current guidelines [5] [6] [7] [8] is there a recommendation for this approach. Therefore, this particular approach may also represent an interventional bias in this study.

In conclusion, the study of Fabbri et al. [1] is challenging and addresses the burning question of whether CSEMSs for EUS-GTD of pancreatic fluid collections may improve procedural efficacy and patient-related comfort, as is also being investigated in benign biliary and pancreatic strictures.

The truth is, of course, in the eye of the beholder but from this preliminary experience the only appropriate scientific conclusion one can make is that using CSEMSs for EUS-GTD is feasible and apparently safe, but nothing more and nothing less. The questions of superiority and cost-effectiveness in comparison with the current classical approach remain and will only be answered by a randomized controlled trial. The approach of using CSEMSs for EUS-GTD of pancreatic fluid collections therefore remains at present a personal rather than an evidence-based choice in endoscopic practice.

• References

• 1 Fabbri C, Luigiano C, Cennamo V et al. Endoscopic ultrasound-guided transmural drainage of infected pancreatic fluid collections with placement of covered self-expanding metal stents: a case series. Endoscopy 2012; 44: 429-433
• 2 Cahen D, Rauws E, Fockens P et al. Endoscopic drainage of pancreatic pseudocysts: long-term outcome and procedural factors associated with safe and successful treatment. Endoscopy 2005; 37: 977-983
• 3 Hookey L, Debroux S, Delhaye M et al. Endoscopic drainage of pancreatic fluid collections in 116 patients: a comparison of aetiologies, drainage techniques, and outcomes. Gastrointest Endosc 2006; 63: 635-643
• 4 Arvanitakis M, Delhaye M, Bali MA et al. Pancreatic-fluid collections: a randomized controlled trial regarding stent removal after endoscopic transmural drainage. Gastrointest Endosc 2007; 65: 609-619
• 5 Pezzilli R, Zerbi A, Di Carlo V et al. Practical guidelines for acute pancreatitis. Pancreatology 2010; 10: 523-535
• 6 UK Working Party on Acute Pancreatitis. UK guidelines for the management of acute pancreatitis. Gut 2005; 54: 1-9
• 7 Wu XN. Guidelines for treatment of severe acute pancreatitis. Hepatobiliary Pancreat Dis Int 2002; 1: 446-451
• 8 French Consensus Conference on Acute Pancreatitis: conclusions and recommendations. Paris, France, 25–26 January 2001. Eur J Gastroenterol Hepatol 2001; 13: 1-13