Endoscopic retrograde pancreatography (ERP) is considered the first-line, standard treatment for treating main pancreatic duct (MPD) obstruction, stricture or disruption [1]. Although ERP is widely performed, it is an especially challenging procedure when the gastroduodenal anatomy has been surgically altered and the papilla is not easily accessible, or when advancement of the guidewire though the papilla or anastomosis is not possible [2]. These challenges are mainly faced in cases of failed cannulation due to high-grade stricture, inability to identify the pancreaticojejunostomy, and inability to reach the pancreaticojejunostomy. Endoscopic ultrasound-guided pancreatic duct intervention (EUS-PDI) is an emerging endoscopic approach allowing access and intervention to the MPD for patients with failed ERP or with surgically altered anatomy. The development of EUS-PDI plays an important role as an alternative to surgical intervention. On the other hand, this approach remains technically demanding with a high risk for complications, and so these procedures should only be performed by highly experienced endoscopists [3,4]. Proper patient selection is of utmost importance, and indication and relative contraindications must be carefully assessed (Table 1). There are multiple studies with small numbers of cases that have described this technique [5-17], however no large prospective, well-controlled studies have been performed. In this review, we describe an overview of the current endoscopic approaches, basic technical tips, and outcomes using these procedures.

EUS-PDI can be divided into two main approaches that are performed to achieve endoscopic ultrasound (EUS)-guided pancreatic duct (PD) drainage: EUS-guided antegrade drainage and EUS-guided rendezvous technique. When the papilla or anastomosis in surgically altered anatomy is not accessible, EUS-guided rendezvous approach is not applicable.

First, the MPD is visualized and carefully assessed with a linear echoendoscope. A therapeutic channel echoendoscope is preferred to allow for broader usage of accessories and insertion of larger caliber stents. Under combined fluoroscopic and EUS guidance, access into the MPD through the stomach or duodenum is achieved using a 19-gauge or 22-gauge fine aspiration needle [12]. Then, under fluoroscopic guidance, a pancreatogram is performed and a guidewire can be passed into the MPD. An 0.035-inch or 0.025-inch guidewire can be passed through a 19-gauge needle while 22-gauge needles require 0.018-inch guidewires. From this point onward, the approach between antegrade drainage and the rendezvous technique differs. The rendezvous technique is performed after the guidewire is successfully advanced across the papilla or anastomosis and coiled in the small intestine. The echoendoscope is then removed leaving the guidewire in place. Depending on the anatomy, a standard therapeutic duodenoscope, colonoscope, or balloon-assisted enteroscope is then advanced to the papilla or the anastomosis, where the PD can be accessed with the guidance of the EUS placed wire to perform retrograde interventions.

For antegrade PD drainage, the echoendoscope is used throughout the procedure for placement of a stent into the MPD via the stomach or the duodenum. Once guidewire access is achieved into the MPD, dilation of the transmural tract is performed. This can be done using the sheath on the fine needle aspiration (FNA) needle, tapered catheters, or cautery-assisted devices such as a needle knife or small caliber ringed catheter. Once catheter access to the MPD is achieved, the tract can be dilated using hydrostatic balloons prior to stent placement.

Currently there are no standard strategies for performing EUS-PDI, however here are our tips to consider when performing this procedure [19,27-29].

EUS-PDI is highly effective at achieving successful PD drainage but is associated with significant rates of complications. This procedure is technically demanding. Taking publication bias into account, actual success rates are likely lower than are reported. A limitation of this technique is that it has only been described in retrospective studies with small sample sizes. Although there are several studies reporting outcomes using EUS-PDI, overall the data are quite limited.

Fujii-Lau and Levy recently performed a systematic review of studies that focused only on EUS-guided PD access while excluding case reports [27]. They identified 222 patients who underwent EUS-PDI and demonstrated a 77% rate of technical success with a clinical success rate of 70% using either the antegrade or rendezvous technique. Complications developed in 19% of the patients, and included abdominal pain (7.7%), pancreatitis (3.1%), bleeding (1.8%), perforation (0.9%), peripancreatic abscess (0.9%), stripping of the guidewire coating (0.9%), and one patient each who developed fever, pneumoperitoneum, pseudocyst, pseudocyst with an aneurysm, and perigastric fluid collection (0.5%).

Recently, an international, multicenter, retrospective study on the safety and efficacy of EUS-PDI after failed ERP was published [33]. 80 patients who underwent EUS-PDI at 4 academic centers in 3 countries were analyzed. Technical success was achieved in 89% and clinical success in 81% of patients. The success rate in this study was higher than previously reported, which is likely due to increased operator experience and improvements in endoscopic equipment. The transpapillary or trans-anastomotic approaches to stent placement via rendezvous wire access seemed to be the more successful technique, with a trend toward an increased likelihood of complete symptom resolution after adjusting for sex, diagnosis, anatomy, prior failed ERP, and technical success, but that was not statistically significant. Immediate adverse events (AEs) (<24 hours) occurred in 20% of patients, with 15% experiencing major complications (6 patients with post-ERCP pancreatitis, 4 who developed pancreatic fluid collections, one with a MPD leak, and one with an intestinal perforation. Delayed AEs (>24 hours) occurred in 11% of patients (all of whom also had immediate AEs—2 pancreatitis, 1 MPD leak, and 4 abscesses treated with antibiotics). The method of approach (antegrade vs. rendezvous) was not a predictor of immediate or delayed AEs, however this could have been due to the small sample size [33].

While EUS-PDI has been shown to be effective, it appears to be limited by its high rates of complications. However there have been no comparative studies between EUS-PDI and standard ERP. A recent international, multicenter, retrospective study was performed to compare these 2 modalities in terms of technical success, clinical success, and AE rates in patients with post-Whipple anatomy. 66 patients underwent 75 procedures (40 EUS-PDI and 35 ERP). Technical success of EUS-PDI was 92.5% compared with 20% in the ERP group (odds ratio [OR], 49.3; p<0.001). Clinical success was achieved in 87.5% of EUS-PDI procedures compared with 23.1% in the ERP group (OR, 23.3; p<0.001). However, AEs occurred more commonly in the EUS-PDI group (35% vs. 2.9%, p<0.001). Procedure time and length of stay were not significantly different between the 2 groups. AEs included abdominal pain requiring hospitalization, intraabdominal abscesses and jejunal ulceration secondary to pancreatic stent placement [34]. Although there were no severe AEs with EUS-PDI in this study, the overall complication rate of 30% is very high. Even with the low technical success rates of 20%, ERP should remain a first-line treatment, even in patients with surgically altered anatomy, based on its superior safety profile. This is especially true considering the low case volume of EUS-PDI being performed, even at experienced, expert centers. It is also very important to emphasize that although this procedure may eventually become standard, improved accessories are needed. Finally, based on its technical difficulty and overall rarity, it is challenging to develop true expertise.

Potential contributing factors of treatment failure include small PD diameter, fibrotic pancreatic parenchyma, short length for guidewire insertion, lack of dedicated devices, lack of technical standard, and failure to navigate the guidewire through the site of obstruction, across the papilla or PJA [3,28]. Other factors that make this procedure challenging include difficulty maintaining the position of the echoendoscope along the axis of the MPD and the difficult angle at which the PD is accessed. Despite the theoretical concerns of puncturing a small caliber PD, it is only stent placement that appears to pose a challenge. In a multicenter, retrospective study, the method of approach (antegrade vs. retrograde) was not a predictor of technical success after controlling for prior failed ERP, altered anatomy, and diagnosis [33].

The lack of long-term clinical outcomes needs to be emphasized. It is difficult to determine the need for re-intervention and to predict long-term clinical outcomes after initial successful intervention. Will et al. reported that 29% of patients having EUS-PDI ultimately required surgical intervention during a follow-up period of 4 weeks to 3 years [10]. Stent dysfunction, including stent migration and occlusion, requiring multiple endoscopic interventions have been reported in up to half of the patients in several case series [9,13]. The use of double pigtail plastic stents with scheduled stent replacement may reduce the rate of stent dysfunction, especially in cases requiring prolonged stent duration.

Although the technical and clinical success rates of EUSPDI are improving, it remains a challenging procedure with a high risk of complication. More studies are needed to evaluate the safety, efficacy and long-term outcomes while we await improved and innovative devices. As opposed to biliary access where percutaneous drainage is an alternative for failed ERCP, treatment options after failed ERP are limited. Considering the major limitations in alternative treatment options after failed ERP, EUS-PDI has the potential to become standard-of-care by avoiding more invasive and involved surgical interventions.