Factors Impacting Treatment Outcomes in the Endoscopic Management of Walled-off Pancreatic Necrosis

Ji Young Bang; C Mel Wilcox; Jessica Trevino; Jayapal Ramesh; Shajan Peter; Muhammad Hasan; Robert H. Hawes; Shyam Varadarajulu


J Gastroenterol Hepatol. 2013;28(11):1725-1732. 

In This Article


Despite several improvements, necrotizing pancreatitis remains a challenging disease entity to manage. In this study, we have shown that creating multiple gateways in the gut facilitates better drainage of the necrotic debris and thereby improves short-term treatment success. In addition, not removing the transmural stents creates a conduit for secretions from the disconnected gland to drain through within the lumen of the gut, thereby decreasing PFC recurrence in the long-term.

MTGT requires the use of EUS to create multiple gateways for drainage, as described in a prior study.[17] This is difficult with standard endoscopy because the technique is dependent on the presence of a luminal compression that usually disappears after the initial puncture. In contrast, when EUS is used, the partially decompressed cavity can still be accessed for creating additional gateways for drainage. With this technique, irrigation can be performed with an 18-gauge nasogastric tube through one channel while others act as conduits for rapid drainage of the necrotic debris (Fig. 3). This technique, by facilitating better drainage, minimizes the risk of development of a superimposed infection and thereby precludes the need for a rescue endoscopic necrosectomy or surgery. This is in line with the findings of a recent randomized trial that compared surgery and the step-up approach where 35% of patients treated by percutaneous drainage alone did not require a subsequent intervention.[18] We believe that aggressive irrigation in conjunction with effective drainage precludes the need for percutaneous catheter placements or a more invasive intervention in a significant proportion of patients. In this study, the mean number of interventions for treatment success per patient was 1.3, which is significantly less than the six reported for endoscopic necrosectomy in a multicenter trial.[12] We also found that the need for more than one endoscopic intervention was significantly associated with treatment failure. This is probably because of the large amount of necrotic debris and superimposed infection caused by the index instrumentation in these patients. In our opinion, performing the index procedure using the MTGT facilitates rapid drainage of necrotic debris and minimizes the chance of procedure-related infection.

Figure 3.

(a) Computed tomography of the abdomen (axial image) showing a large walled-off pancreatic necrosis (> 20 cm) containing a large amount of necrotic debris. (b) Computed tomography of the abdomen (coronal image) showing marked resolution of the walled-off pancreatic necrosis in the same patient following drainage using the multiple transluminal gateway technique.

Although the rate of initial treatment success following transmural drainage of WOPN appears superior, PFC recurrence in the long term is higher when the upstream gland is disconnected.[13,14] This is because the viable upstream portion of the disconnected gland continues to secrete pancreatic juice, but without an internal conduit for drainage (Fig. 4). Although distal pancreatectomy and/or reconnection of the pancreatic duct to a Roux-en-Y limb can be effective treatment options, poor functional status, left-sided portal hypertension, ongoing inflammation, and adhesions from serial debridement oftentimes complicate this approach.[19] Prolonged stent placement has been prescribed as a minimally invasive treatment strategy in these patients, as it facilitates drainage of pancreatic secretions and thereby relieves the pressure in the main pancreatic duct upstream of the rupture site.[15,20] However, as the WOPN resolves, it causes adherence of the walls of the cavity, leading to migration of stents into the gut lumen in some patients. Despite spontaneous stent migration, this strategy appears to be effective, as scheduled removal of the stent before complete collapse of the cavity can lead to PFC recurrence, particularly when a communication exists between the cavity and the pancreatic duct. Therefore, the duration of stenting may be more important than whether the stents are still present or retrieved after an adequate stent placement period.

Figure 4.

(a) Pancreatogram of a patient with a walled-off pancreatic necrosis in whom double pigtail stents were placed. (b) Magnetic retrograde cholangiopancreatography of a patient following removal of the transmural stent showing opacification of the disconnected pancreatic duct and associated peripancreatic fluid collection.

Although placing permanent transmural stents appears to be effective, it does have some practical limitations. The migrated double pigtail stents can cause bowel obstruction, as was evident in two of our patients. This is particularly of concern when performing drainage using the MTGT approach, where we sometimes place more than 12 stents in an individual patient. As a consequence, our current practice is to remove all but two transmural stents to facilitate drainage of secretions from the upstream duct. We have not recently encountered any cases of bowel obstruction after adopting this approach; nonetheless, more data are needed to make definitive recommendations.

Despite the technical advances detailed in this report, nearly one-third of patients with treatment success were on long-term narcotics. It is unclear if these patients were drug-dependent or had parenchymal disease-induced pain or other neuromodulatory effects contributing to their continued need for narcotics. Two of 53 patients with treatment success eventually underwent pancreatectomy because of worsening symptoms. A EUS examination prior to surgery in these two patients revealed that the number of (sonographic) chronic pancreatitis criteria had increased from five to seven in one patient and from four to seven in the other.[21] Endoscopic options were limited in these two patients, as they did not have any ductal dilation. This suggests that despite advances in endoscopic therapy, only surgery can provide a definitive palliation in some patients.

There are several limitations to this study. Firstly, all procedures were performed at specialized medical centers with excellent support services that may not be available at smaller institutions. Secondly, although the data collection was prospective, the retrospective nature of the study design limits our ability to investigate the effectiveness of other variables on treatment outcomes. Thirdly, the MTGT and the use of permanent transmural stents need to be validated at other centers and by other investigators. Fourthly, selection bias remains a concern. However, we believe that this may not be a major limitation, because at our institution, sicker patients are generally referred for minimally invasive endoscopy rather than being subjected to surgery as the first-line treatment approach.

In summary, what lessons can we learn from this report? One, it is clear that the MTGT facilitates better irrigation and drainage of the necrotic cavity, thereby yielding superior short-term treatment outcomes. Two, the placement of permanent transmural stents minimizes the chances of PFC recurrence, although the possibility of stent migration remains a concern. Finally, close collaboration between specialties is paramount to optimize treatment outcomes.