Initial steps in the management of acute pancreatitis consist of aggressive supportive care including IV hydration with lactated ringer’s, pain management, and efficient steps to identify the cause of pancreatitis . Once a collection has been characterized, the complications of the collection may need to be addressed separately (see Introduction). If after the management of the acute pancreatitis and these complications , no or minimal symptoms remain, conservative management of the collection with watchful waiting can be attempted. Data on pancreatic collections show that in patients that can be managed without intervention, spontaneous resolution can occur in 30–60 % of patients [17–19]. More specific data on the different types of acute fluid collections are sparse, although based on a few studies, we know that fluid collections associated with non-necrotizing pancreatitis resolve faster, often within 2 weeks in about 70 % of patients [20] compared to spontaneous resolution in about 30 % by month 6 in patients with necrotizing pancreatitis [17].

In the case detailed above, a pancreatic duct disruption has occurred, which is likely resulting in the development of an acute peripancreatic fluid collection. Duct disruption regardless of etiology (acute pancreatitis, pancreatic surgery, or trauma) can be treated using a similar endoscopic approach. Initial conservative therapy for ductal disruptions can include nasojejunal feeding, somatostatin analogues, and pancreatic enzyme replacement. Nasojejunal feeding is associated with significantly higher spontaneous closure rate for post-surgical pancreatic duct fistulae, presumably by reducing pancreatic stimulation [21]. Somatostatin analogues are routinely used perioperatively for pancreatic surgeries [22, 23] although there are only a few studies describing its use for pseudocysts and the practice remains controversial [24].

If conservative management has not resulted in any improvement, either clinically or on imaging (as with our patient), ERCP with transpapillary pancreatic duct stent placement should be attempted [3, 5, 25, 26]. One scenario where transpapillary stenting is of unclear utility is the disconnected duct syndrome [5, 27]. This involves full transection of the pancreatic duct, and the proximal (tail) portion of the pancreas freely secretes pancreatic juices. On pancreatogram, there is either blowout of the duct or a complete cutoff with no duct opacified upstream from this area (Fig. 12.3). Given the full transection, a bridging stent cannot be placed and management focuses on creating a cystgastrostomy to allow a fistulous connection back to the lumen of the stomach or bowel. If this approach fails, combined percutaneous and endoscopic rendezvous procedures may succeed [27], and surgical options include pancreaticojejunostomy and distal pancreatectomy. Percutaneous drainage is not an appealing option as the rate of persistent external pancreatic fistula from the drains is high [27].

Timing for ERCP can vary significantly based on the clinical scenario, from immediately on recognition of the ductal leak to >4 weeks after the episode of acute pancreatitis . On the basis of the literature, we generally try to wait 4–6 weeks for patients with pancreatitis-related duct disruption, although patient’s discomfort or more acute symptoms may lead us to intervene sooner. ERCP is performed in a standard fashion, and the site of disruption is identified during pancreatography. In addition, the location of any stricture or stones should be noted. Most often, a 5Fr or 7Fr stent is utilized, although the exact size will depend on the clinical scenario and has not been shown to be related to successful closure of the PD leak. Any strictures or stones that may have contributed to the leak should be traversed and managed. In addition, the stent should ideally bridge the disruption to maximize chances of success (see next section). The stent should be left in place at least 6 weeks, although time to closure can vary significantly with studies reporting a median closure time as high as 4 months [3, 26]. Prophylactic antibiotics should be administered with all these procedures since ERCP introduces non-sterile contrast into an otherwise sterile fluid collection.

It should be noted that depending on the clinical scenario (including the degree of symptoms from the fluid collection and maturity of wall around the pseudocyst), many investigators have reported performing a cystgastrostomy in the same endoscopic session as the ERCP, particularly for large encapsulated fluid collections. Hookey at al [28] in a retrospective chart review and prospective follow-up, reported on the resolution of pancreatic fluid collections in 15 patients with transpapillary stent alone, 60 patients with transmural drainage alone, and a combined approach in 41 patients. Overall clinical success (resolution of symptoms and collection) was 88 % and was not dependent on the drainage technique performed. However, patient characteristics differed significantly among the different groups. For example, the 15 patients with transpapillary stenting alone typically had a smaller fluid collection (<7 cm) with evidence of pancreatic duct obstruction and communication between the pancreatic duct and the pseudocyst. A combined approach was used for larger cysts communicating with the pancreatic duct or when the transpapillary approach was unable to bridge the leak. These data contrast with a small study from Singh’s group in India, who employed a pure transpapillary approach for larger cysts (> 7 cm) in the tail of the pancreas and found resolution in only 33 % of patients, while the other 67 % were complicated by infection requiring further percutaneous drainage. Overall, transpapillary stenting alone is not recommended for large mature fluid collections where transmural drainage or combined transmural drainage and transpapillary stenting likely has the best outcome. Immature fluid collections cannot be drained transmurally, and would require transpapillary stenting if drainage is necessary.

ERCP with transpapillary stenting is fairly effective for the treatment of pancreatic duct leaks [3, 26, 29, 30]. Review of literature shows resolution of pancreatic duct leaks in 58–100 % of patients, although the etiology of the pancreatic duct leak in these studies included cases from chronic pancreatitis, trauma, and post-surgical leaks in addition to acute pancreatitis . In addition, the technique varied substantially and may explain the variability of the outcomes. Success was positively associated with findings of a partial duct disruption, location of the disruption in the body of the pancreas, maintaining the stents for at least 6 weeks, and placement of stent that bridges the duct leak. Data from our institution [3] showed that a stent that bridged a disruption was associated with successful resolution of a leak in 92 % of patients. Patients in whom the stent was placed only across the papilla or up to the disruption, by comparison, were associated with only a 44–50 % success rate (Fig. 12.4), stressing the importance of bridging a disruption.

Mortality of this procedure is rare and complications (7–9 %) are mainly associated with performance of the ERCP. More specific complications include fevers and infection following stent placement, stent occlusion, or transpapillary stenting alone of large pseudocysts. Recurrence in the setting of stent failure or stent dislodgement can be treated with repeat ERCP and restenting for another 6–8 weeks. With unsuccessful procedures occurring from either failure to place the pancreatic duct stent or failure to resolve the leak, a trial of draining the fluid collection alone can be attempted (see below, Endoscopic Ultrasound and Transmural Drainage of a Pseudocyst). About 4 % ultimately require surgery; surgical options include a pancreaticojejunostomy and a distal pancreatic resection for ductal disruption in the body and tail. Overall success of the surgical approach is high at 90–92 %, although with a mortality of 6–9 % [31].

Later sequelae of a pancreatic duct leak can include pancreatic ascites, pancreaticopleural fistula, and pseudocyst . The patient in our case has developed a pseudocyst with a well-defined wall. Management of a pseudocyst can include watchful waiting for asymptomatic lesions ; for symptomatic collections, percutaneous, surgical, or endoscopic strategies are available, and the choice should be made in a multidisciplinary fashion with the expertise of the particular institution in mind. Endoscopic drainage of a pseudocyst was first described by Gerald Rogers et al. in 1975 [32] with subsequent improvement by creating a fistula into the stomach by Richard Korazek in 1985 [33]. The methodology has since undergone several modifications.

Prior to starting a cystgastrostomy , consultation with a radiologist is often necessary, and at the minimum, careful review of abdominal CT and/or MRI should be done to ensure there is a mature wall around the fluid collection (Fig. 12.1b) and direct apposition of the stomach wall and the pseudocyst [34]. Of note, the radiology imaging should be a recent study obtained within 1–2 weeks of the planned procedure. A mature wall often develops in 4–6 weeks, although we have seen mature walls as early as 3 weeks [2]. An appropriate wall for a cystgastrostomy (and necrosectomy) is considered to be > 3 mm. While some centers consider the upper limit of wall thickness to be 1 cm since thicker walls are thought to increase the risk of complications [34], we do not limit our interventions to capsules less than 1 cm and have had technical success with much thicker walls.

Patients are given antibiotics prior to the procedure (usually intravenous ciprofloxacin and metronidazole). We use general anesthesia rather than conscious sedation to reduce the risk of aspiration when draining large amounts of fluid into the stomach and to allow maximal sedation if complications arise that require a longer duration of the procedure. We use CO₂ insufflation as it is rapidly absorbed by the body at the end of the procedure. The initial step is identification of the location of the pseudocyst and finding an appropriate puncture site. While traditionally direct endoscopic visualization of a bulge into the gastric or duodenal lumen was used as a marker, currently most centers use EUS guidance [35–39]. EUS is important for better localization of the fluid collections when there is no large extra-luminal compression, identification, and avoidance of blood vessels at the access site, and confirmation of the lack of necrotic tissue. The best data supporting the use of EUS was a randomized controlled trial by Varadarajulu et al. [35] comparing EGD to EUS-guided cystgastrostomy of pseudocysts, where the use of ultrasound was associated with greater success (100 % versus 33 %, P < 0.001) and trend toward decreased complications. EGD-related complications included one death from massive hemorrhage of gastric varices at the puncture site not visualized during the EGD.

Once a site has been identified using a therapeutic linear echoendoscope, we use a 19-gauge needle to puncture into the pseudocyst (Fig. 12.6). We try to puncture at an angle that ideally allows the endoscope to be kept straight in the body of the patient, allowing forces to be transmitted directly to the wall of the pseudocyst. Fluid can be aspirated for bacteriological and culture studies. In addition, contrast is instilled into the pseudocyst to identify borders of the pseudocyst and to maintain distention. We over-inject contrast to expand the pseudocyst by 5–10 mm to keep the pseudocyst distended and obviate the need for electrocautery when trying to traverse the wall of the pseudocyst (In press data). A 0.035-inch guidewire is then coiled numerous times into the cavity, and the access site is dilated using hydrostatic balloons. We start with a 6-mm biliary-dilating balloon (typically the Hurricane Biliary Balloon Dilation Catheter, Boston Scientific, Marlborough, MA) and serially dilate to 15 mm. Alternative strategies for tract creation include brief use of cautery with a needle-knife sphincterotome or cystotome. These strategies, however, likely pose increased risk of bleeding due to inadvertent cutting of a gastric vessel. Double pigtail stents are then placed through the tract to allow adequate drainage. We generally use three 10Fr double pigtails for drainage, with length depending on the depth necessary to adequately drain the pseudocyst.

On follow-up (Fig. 12.7), we reimage our patients in 4–6 weeks with a CT scan or MRI to evaluate for continued contraction or resolution of the pseudocyst and the presence of the stents. If the pseudocyst has resolved and the stents remain in place, we perform an EGD to remove the stents, although a small randomized controlled study has shown that leaving stents in place even after resolution of pseudocyst may decrease recurrence of the collection [40]. In this small study, patients either had their stents removed at a median of 2 months following cystgastrostomy or 12 months later with 38 % (5/13 patients) recurring in the former group compared to no recurrences with long-term stent patients. If the stents have fallen out by themselves, no further procedure is needed. If the pseudocyst is resolving, more time is allowed for drainage and involution. If the stents have fallen out and the patient remains symptomatic with an ongoing pseudocyst, a further cystgastrostomy may be needed to adequately drain the pseudocyst. Finally, if there is no change or increase in size of the pseudocyst, we reevaluate for pancreatic duct disruption and possible stent blockage. Surgical referral may also be appropriate as discussed below.

When performed in the setting of an experienced multidisciplinary team, the initial success rate in creation of the cystgastrostomy is high, up to 94–95 %, with a pseudocyst resolution rate of 90–100 % [41–43]. Complications can occur in 0–20 % of patients and include bleeding sometimes requiring surgery usually due to the use of electrocautery without EUS guidance to enter the pseudocyst , inadequate drainage, and pseudocyst infection. There is minimal mortality (< 1 %) with at least one death related to electrocautery used to access the pseudocyst without EUS. Recurrences can range from 0 to 16 %, and at least one recent paper shows resolution of the pseudocyst in all patients and no recurrence over at least 24-month follow-up [41]. Of note, in this paper, patients undergoing endoscopic cystgastrostomy had stents routinely removed 2 months following the procedure if imaging demonstrated resolution of the pseudocyst. Success is significantly lower if there is necrotic debris in the collection, underscoring the importance of preprocedural evaluation with imaging and EUS and ensuring the collection is a pseudocyst. Hookey et al. [28] showed that success of cystgastrostomy with or without transpapillary stenting of a pseudocyst was over 88 %, compared to only 25 % if necrotic debris is present.