There have been a number of studies comparing these methods of anastomosis including some randomized trials, but most including the randomized trials have focused on the early results rather than comparisons of late outcomes such as strictures at the PJA vs PGA. Tomimaru et al. studied 42 patients 2 years after pancreatoduodenectomy, 28 who had had PGS and 14 who had had PJS [9]. They noted that pancreatic duct diameter tended to increase more and that pancreatic atrophy was more severe after PGS [9]. Schmidt et al. studied QOL after PGS and PJS at a mean time of 6.4 years after surgery in about 100 patients [10]. In the PG group, there was an increase in steatorrhea as well as intolerance to certain foods. There was no difference in need for enzyme replacement or in onset of diabetes, and global QOL was also not different in the two groups. Ishikawa et al. studied glucose tolerance in 51 patients over a 7-year period. The patients were about equally divided between PJS and PGS. The decline in glucose tolerance after PG was not associated with type of pancreatic anastomosis. Konishi performed a prospective randomized trial of PGS vs PJS and followed the patients for 2 years [11]. They found no difference in change of pancreatic duct diameter or glucose tolerance but the study population was made up of only 25 patients. These results address the problem of stricture only tangentially but they suggest that there probably is not an advantage of one type of anastomosis over the other in retention of pancreatic exocrine function. The data regarding pancreatic endocrine function are probably more reflective of remaining parenchyma than anastomotic stricture.

The traditional ERP approach involves accessing the pancreatojejunostomy anastomosis (PJA) retrograde through the afferent loop of the gastroenterostomy. There are several challenges involved in performing ERP through the afferent loop for treatment of PJA stenosis, which limit technical success.

First, successfully advancing the endoscope to the PJA is challenging. The afferent limb is often difficult to engage with the side-viewing duodenoscope. Also, the afferent limb may be of variable length (depending on surgeon preference and location of jejunal loop in relation to the transverse mesocolon), sometimes making it impossible to reach the PJA with a standard duodenoscope (124 cm long). In these situations, forward-viewing instruments are required, typically either a pediatric or adult colonoscope (168 cm long) or enteroscope (234 cm long) with or without a balloon overtube. Using a forward-viewing instrument poses several difficulties. First, these instruments lack an elevator, that is, a metal lever at the distal tip of the working channel that provides an extra degree of motion to instruments exiting the tip of the scope. Second, the longer working channel length of colonoscopes and enteroscopes compared to duodenoscopes limits the number and type of instruments, which can be utilized during the procedure. Third, pediatric colonoscopes and enteroscopes have smaller diameter working channels, which limits the caliber of stents that can be used.

The second major difficulty encountered is that a stenotic PJA can be difficult to visualize (Fig. 24.2 ). The PJA is usually 15–20 cm beyond the usually well-visualized choledochojejunostomy and can be located at the stump of the afferent limb or, more commonly, approximately 5 cm proximal to the stump. Visualization of the anastomosis also depends on whether it is an end-to-end or end-to-side anastomosis, the latter being usually more difficult to visualize. When the PJA cannot be visualized, there are ways to help localize it. One method is to administer intravenous secretin and observe for a gush or trickle of pancreatic juice. However, another challenge is transparency of pancreatic juice. Visualization of the juice can be enhanced by spraying the mucosa with a dye, such as methylene blue.

Once the PJA is identified, a variety of catheters and wires may be utilized to achieve deep cannulation of the pancreatic duct. Usually, due to the pinhole size of the PJA, the smallest available 3-4-5 F taper tip catheter loaded with an 0.021ʺ caliber wire is used. Once the PJA is carefully engaged with the catheter, contrast is injected and retrograde opacification of the pancreatic duct is observed fluoroscopically. The wire is then advanced deeply into the pancreatic duct. Another approach is to attempt passage of a wire through the anastomosis prior to injection of contrast. Following wire placement deep into the pancreatic duct (Fig. 24.3 ), passage or balloon dilation of the PJA is performed. Cautery is avoided to reduce the risk of perforation at the PJA. Following dilation, a plastic stent is placed (Fig. 24.4 ). There are multiple different stent types of varying lengths, diameters, and shape (straight versus pigtail). Generally, stents are removed in 6 weeks, and the need for repeat dilation or stenting is assessed at that time (Fig. 24.5 ). However, the optimal duration of stenting is not well established and not evidence based.