Safety and efficacy of trans-afferent loop endoscopic ultrasound-guided pancreaticojejunostomy for post pancreaticoduodenectomy anastomotic stricture using the forward-viewing echoendoscope: a retrospective study from Japan

Article information

Clin Endosc. 2024;.ce.2024.089

Publication date (electronic) : 2024 August 26

doi : https://doi.org/10.5946/ce.2024.089

Ahmed Sadek^1,2, Kazuo Hara^,2, Nozomi Okuno², Shin Haba², Takamichi Kuwahara², Toshitaka Fukui², Minako Urata², Takashi Kondo², Yoshitaro Yamamoto², Kenneth Tachi^2,3

¹Department of Internal Medicine, Faculty of Medicine, Helwan University, Cairo, Egypt

²Department of Gastroenterology, Aichi Cancer Center Hospital, Nagoya, Japan

³Department of Medicine and Therapeutics, University of Ghana Medical School, Korle Bu Teaching Hospital, Accra, Ghana

Correspondence: Kazuo Hara Department of Gastroenterology, Aichi Cancer Center Hospital, 1-1 kanokoden, Chikusa-Ku, Nagoya City 464-8681, Japan E-mail: khara@aichi-cc.jp

Received 2024 April 12; Revised 2024 May 17; Accepted 2024 May 31.

Abstract

Graphical abstract

INTRODUCTION

Anastomotic stricture is a known long-term complication of any gastrointestinal anastomosis, and endoscopic management is the current standard initial treatment for most gastrointestinal anastomotic strictures.

Despite the improvement in the technical aspects, anesthesia, and postoperative care of pancreaticoduodenectomy (PD) surgery, pancreaticojejunal anastomotic stricture (PJAS) remains a long-term complication of this surgery, potentially leading to recurrent obstructive pancreatitis.1

Since PJAS is a frequently encountered medical problem, with a reported frequency of 2% to 11% of patients undergoing PD,2 its management necessitates more research to ensure standardized treatment options. Moreover, it is imperative because broadening the indications for PD to include benign pancreatic disease besides pancreatic cancer has led to improvement in overall patient survival after surgery, consequently giving more time for complications to occur.

Endoscopic treatment is considered the first line for PJAS management. This treatment includes direct visualization of the anastomosis so that endoscopic retrograde pancreatography (ERP) can be performed with strictural dilation or an EUS-guided approach, including either EUS-guided transmural drainage involving EUS-guided pancreaticogastrostomy (EUS-PGS) and EUS-guided pancreaticojejunostomy (EUS-PJS), through transmural puncture of the pancreatic duct from the stomach and the jejunum, respectively or EUS-guided rendezvous technique.

EUS-PGS is a widely accepted treatment modality for PJAS. However, multiple endoscopic reinterventions are required for regular stent exchange to maintain a patent fistula and continuous drainage. Furthermore, a subsequent high chance of adverse events (AEs) exists because of the indicated long duration of stent placement, and puncturing the tract through the abdominal cavity may cause stent-related complications, such as migration and leakage. EUS-PJS involves puncturing the main pancreatic duct through the anastomotic site (avoiding the abdominal cavity), and this approach could be superior to EUS-PGS in terms of the risk of stent-related AEs and creation of a permanent fistula without requiring multiple reinterventions.

We described the outcomes of EUS-PJS in this study using a forward-viewing curvilinear echoendoscope (FV-CLS) in patients presenting with PJAS after PD. This study represents the first evaluation of EUS-PJS as an interventional treatment for PJAS complicating PD in many patients.

METHODS

We retrospectively analyzed 10 patients who underwent EUS-PJS for PJAS between February 2019 and October 2023. Only naïve patients with PJAS who did not undergo EUS-guided drainage before EUS-PJS were included in the study. Patients who underwent EUS-PGS and had a residual pancreaticogastrostomy fistula before EUS-PJS were excluded. An expert endosonographer performed all the procedures. Early AEs, primary procedural details, and long-term outcomes were assessed. Long-term outcomes were described for late stent-related AEs, frequency of stent exchange procedures during the follow-up period, and whether a stent-free state was achieved. An intention-to-treat analysis was performed to assess the technical success rate.

Procedure steps

The management policy for patients with PJAS is shown in Figure 1. First, a colonoscope or enteroscope was inserted deep into the afferent jejunal limb till the pancreaticojejunal anastomosis site was reached under fluoroscopic guidance using CO₂ insufflation. If an anastomosis was identified, ERP was attempted, followed by dilation of the anastomotic stricture using a balloon catheter. EUS-PJS is attempted when ERP fails because of either failure to cannulate the pancreatic duct due to severe stricture or failure to identify the anastomotic stricture (by white light and indigo carmine spraying) due to complete scarring. If no evidence of the anastomosis site could be identified, the scope was inserted until it passed the choledocojejunal anastomosis (as the pancreaticojejunal anastomosis is usually located at a deeper site in the afferent jejunal limb), guided by fluoroscopy, so that the tip of the scope anatomically faced the pancreas (Fig. 2A).

Fig. 1.

Flowchart of our pancreaticojejunostomy anastomotic stricture management policy. ERP, endoscopic retrograde pancreatography; EUS, endoscopic ultrasound; EUS-PGS, EUS-guided pancreaticogastrostomy; EUS-PJS, EUS-guided pancreaticojejunostomy.

Fig. 2.

Endoscopic ultrasound-guided pancreaticojejunostomy procedure steps. (A) A colonoscope inserted into the afferent jejunal limb till the pancreaticojejunal anastomosis. (B) After endoscopic retrograde pancreatography failure, a guidewire is left in the jejunum, and the scope is exchanged. (C) A forward-viewing echoendoscope is inserted over the guidewire until anastomosis is achieved. (D) The pancreatic duct is punctured using a fine needle aspiration needle, and a guidewire is inserted. (E) Tract dilation using Tornus ES drill dilator. (F) Finally, a fully covered self-expandable metal stent is deployed inside the pancreatic duct across the anastomosis.

A clip was inserted to mark the anastomotic site, and a guidewire was inserted as deep as possible. The scope was then withdrawn, leaving the guidewire inside the afferent jejunal limb. An FV-CLS (TGF-UCT260 J; Olympus Medical Systems) was inserted over the guidewire using an endoscopic retrograde cholangiopancreatography cannula under fluoroscopic guidance until the anastomosis site was reached. In some patients, the FV-CLS could be inserted directly into the afferent jejunal limb without using the “over the wire” technique to reduce the insertion time. The decision to insert the FV-CLS directly or over a guidewire (following a colonoscope) depends on each patient’s experience when they had previously undergone endoscopy after PD surgery for varying reasons (such as postoperative cholangitis). If no reported difficulties were reported in scope insertion during previous interventions following surgery, direct insertion of the echoendoscope was attempted. For patients who were entirely new to the procedure or those who reported difficulty in scope insertion in previous interventions, a colonoscope was first inserted.

The anastomotic site was carefully examined using EUS (Fig. 3). The main pancreatic duct (MPD) was dilated, making it clear to visualize, and the pancreatic parenchyma and muscle layer of the jejunum were identified. The MPD was then punctured at its end, appearing as the closest point to the endoscope in the EUS image (this point represents the opening of the MPD into the anastomosis line), which ensures puncture through the anastomosis, not the parenchyma.

Fig. 3.

Endoscopic ultrasound image showing the anastomosis line. MPD, main pancreatic duct.

When the MPD was visualized, a 19-gauge needle (EZ Shot 3 Plus; Olympus Medical Systems) preloaded with a 0.025 guidewire (M-Through, Asahi Intecc Corp. or VisiGlide 2, Olympus Medical Systems) was used to puncture the MPD through the anastomosis, and the guidewire was advanced as deeply as possible. The needle tract was primarily dilated using either an electrocautery dilator (6 Fr, Cysto-Gastro-Set; Endo-Flex GmbH) or a Tornus ES drill dilator (0.025-inch guidewire-compatible type; Asahi Intecc Corp.). A balloon dilator catheter (4 or 6 m, REN; KANEKA Medics) was used for further tract dilation whenever needed. A biliary catheter (Uneven Double Lumen Cannula; PIOLAX) was used to inject the contrast medium for pancreatography and aspiration of the pancreatic juice.

We inserted a pancreatic stent after performing the ERP. We used a fully covered self-expandable metal stent (FCSEMS), 6 mm/6 to 8 cm HANAROSTENT Benefit (5.9 F delivery system; Boston Scientific Co.) with side holes artificially made at the proximal 1 to 2 cm of the intrapancreatic portion of the stent (Fig. 4) to avoid blocking the drainage of the side branches of the MPD. When stones were suspected, multiple strictures were present, or the MPD diameter was narrow, a 7 Fr/7 cm plastic stent (Through & Pass, Gadelius or Flexima, Boston Scientific Co.) was used.

Fig. 4.

Making artificial openings in the fully covered self-expandable metal stent. First, we deployed 1 to 2 cm of the stent outside the patient, made multiple openings in the stent cover using a simple needle or the sharp tip of a fine needle biopsy needle, and finally recaptured the stent.

The procedure steps are illustrated in Figure 2 and Supplementary Video 1.

Follow-up policy

Transmural stent exchange was scheduled every 2 to 3 months or earlier if symptoms of pancreatic duct obstruction recurred. Dilation of the anastomotic stricture was attempted at every scheduled stent exchange using a balloon dilator whenever needed. The decision to keep the patient stent-free is made when the patient is asymptomatic, the anastomotic opening appears wide enough (Fig. 5), with no MPD stricture in the fluoroscopy image, the contrast freely outflow from the MPD into the jejunum after injection, and with the absence of a notch when balloon dilation is attempted. Finally, we removed the stent, and the patient remained stent-free with no further stent exchange. For patients in whom the FCSEMS was inserted, the stent was replaced with a plastic stent on the first scheduled stent exchange date. A computed tomography scan was regularly performed every 6 months for early AEs the day after EUS-PJS and late AEs.

Fig. 5.

Endoscopic view of a widely open pancreaticojejunal anastomosis after stent removal in the endoscopic ultrasound-guided pancreaticojejunostomy.

RESULTS

Ten patients, seven females (70%) and three males (30%) with a median age of 63.5 years (23–79 years), underwent EUS-PJS at the Aichi Cancer Center Hospital during the study period. The primary disorders included three patients with branch-duct intraductal papillary mucinous neoplasm, one with mixed duct-intraductal papillary mucinous neoplasm, one had intraductal papillary mucinous cancer, two had pancreatic adenocarcinoma, two with duodenal gastrointestinal stromal cell tumor and one had solid pseudopapillary neoplasm. All patients underwent PD with child reconstruction complicated by symptomatic anastomotic stricture.

EUS-PJS was indicated because of either failure to visualize the anastomotic opening in six patients (60.0%) or failure to cannulate the anastomosis due to severe scarring in four patients (40.0%). Table 1 summarizes patient characteristics.

Table 1.

Patient characteristics (n=10)

The technical and clinical success rates were 100 % in 10 patients. The procedure time ranged from 12 to 33 minutes (median, 23.5 minutes). However, compared to EUS-PGS, FV-CLS insertion in the afferent limb may prolong the procedure time; hence, we thought it worth considering the scope insertion time as well. With the previously described insertion steps, which included either direct insertion of the FV-CLS or the use of an over-the-wire insertion technique after inserting a colonoscope or enteroscope, the median time until complete insertion of the FV-CLS and identification of the puncture point was 41 minutes (range, 26–64 minutes).

The median MPD diameter was 3.4 mm (2.6–8.2 mm). A 19 G needle was used in all patients for pancreatic duct puncture, and a 0.025-inch guidewire was used after pancreatic duct puncture. An M-Through guidewire was used in eight patients, whereas VisiGlide 2 was used in two patients. Tract dilation was primarily achieved using the Tornus Es drill dilator in nine patients, whereas the electrocautery dilator Cysto-Gastro-Set was used in only one patient. Further tract dilation was performed using a 4- or 6-mm balloon dilator catheter in six patients. In four patients, no further dilator device was used, and only the tornus dilator was sufficient for tract dilation.

We used a 7 Fr/7 cm plastic stent in two patients, while an FCSEMS was applied in the remaining eight patients. We used 6 mm/8 cm and 6 mm/6 cm stents in one and seven patients, respectively. The procedural details are summarized in Table 2.

Table 2.

Details of endoscopic ultrasound-guided pancreaticojejunostomy procedure (n=10)

Early AEs occurred in two patients as fever lasting for one or two days treated with antibiotics. No severe AEs, such as pancreatitis, bleeding during or after the procedure, abscess, or pancreatic juice leakage were observed. Furthermore, no early stent-related AEs required stent revision due to stent migration or occlusion.

One patient was converted to EUS-PGS 3 months after the EUS-PJS procedure and a 7 Fr/7 cm double pigtail plastic stent was inserted from the stomach to the jejunum because scope insertion into the afferent jejunal limb was technically difficult in this patient despite being successfully performed in the scheduled stent exchange; therefore, conversion to EUS-PGS was decided to allow easier reinterventions afterward. After EUS-PGS, frequent stent exchange procedures were required, and a stent-free state was achieved 2 years after the initial procedure. The long-term outcomes of EUS-PJS in this patient were not relevant for inclusion in the study. Long-term results were also excluded for another patient because of the relatively short follow-up period after the main procedure (two months).

The long-term outcomes were analyzed in eight patients and showed that no stent-related AEs in the form of internal or external migration or stent occlusion occurred in a median follow-up duration of 9.5 months (range, 5–14 months). No symptoms of MPD obstruction recurred. All patients underwent scheduled stent exchange every two or three months. The median number of stent exchanges was two (range, 1–5). Further dilation of the pancreaticojejunal anastomosis was performed in five patients in the scheduled stent exchange procedure using balloon dilation due to insufficient anastomotic opening after stent removal. The median duration of stent placement was 8 months (ranging from 5–12 months). Three patients achieved a stent-free state at the last scheduled follow-up at 10, 11, and 12 months after the initial procedure. These patients were asymptomatic after stent removal for 1 to 3 months. The long-term outcomes of the eight patients are summarized in Table 3.

Table 3.

Long-term outcomes (8 patients)

DISCUSSION

EUS-guided pancreatic duct drainage (EUS-PDD) has emerged as an interventional treatment option for PJAS following PD surgery in cases of ERP failure caused by conventional endoscopic interventions.4

After PD, using a double-balloon enteroscope or colonoscope in long and tortuous jejunal loops is technically feasible. However, even after reaching the anastomosis site, it is sometimes difficult to identify the anastomosis in cases of complete scarring, or difficult cannulation can occur due to severe stricture, leading to ERP failure, which is not uncommon. The ERP success rates ranged from 12.5% to 28.6%.5-7

EUS-guided PGS is a commonly used alternative to surgery for pancreatic duct drainage in PJAS,8 and many reports have described the outcomes of this procedure. However, it remains a challenging procedure that carries significant AEs, such as pancreatitis, pancreatic juice leakage, bleeding, pancreatic fluid collection, stent migration, and the necessity of frequent stent exchange.

Matsunami et al.8 explained the reasons, such as the instability of the echoendoscope in the large cavity of the stomach, the pancreas nonadherent to the stomach wall anatomically, and respiratory movements, for why the EUS-PGS is challenging. All of these factors can cause difficulty in puncturing the pancreatic duct.

One of the undesirable situations encountered during the EUS-PDD procedure is failure to deploy the stent despite successful puncture and dilation, leading to technical failure. In EUS-PGS, failure to deploy the stent may carry the risk of pancreatic juice leakage due to the anatomical distance between the stomach and pancreas, which may require salvage surgery.8,9

To our knowledge, only four reports have addressed the EUS-PJS procedure for PJAS management post-PD.10-13 The procedures were all technically successful; however, different dilation devices were used, and FV-CLS was used in only three reports, each of which evaluated EUS-PJS in one patient.11-13 To our knowledge, our study is the first to describe the safety and outcomes of EUS-PJS in several patients.

In our study, we achieved technical success in of 10/10 (100%) patients, in agreement with some reports that achieved 100% success rates with EUS-PGS.8,14-16 Imoto et al.4 summarized the technical success rates of EUS-guided transmural pancreatic drainage procedures in more than 20 studies. The total number of patients was 401, and the overall technical success rate was 339/401 (85%; range, 63%–100%).

The new Tornus ES drill dilator (0.025-inch guidewire-compatible type; Asahi Intecc Corp.) was safely and effectively used in 9/10 patients as the primary dilator device. The safety and efficacy of this dilator have been previously reported in EUS-guided hepaticogastrostomy.17 The safe and easy application of this drill dilator is the reason for our preference to use it in this procedure. However, dilation of the puncture tract in EUS-PJS is safer than EUS-PGS because the adhesions between the pancreas and jejunum reduce the risk of complications, so other types of dilators could also be used.

Overall, no severe immediate AEs, such as acute pancreatitis, bleeding, pancreatic juice leakage, peripancreatic collection, abscess, or perforation, occurred during the short-term follow-up period (2 weeks). In contrast, the rate of immediate AEs (including mild and severe events) ranged from 17.4% to 26.7% with EUS-PGS.4,8,14,15,18 Imoto et al.4 reported that the rates of severe AEs associated with EUS-PDD (involving EUS-PGS and pancreaticoduodenostomy) were 4.5% for acute pancreatitis, 2.7% for bleeding, and 2.4% for pancreatic juice leakage.

The technical challenge in this intervention, compared with EUS-PGS, is reaching the anastomosis using an echoendoscope. Using an oblique-viewing convex scope is technically difficult and risky; therefore, FV-CLS should be used instead.

During the follow-up, we did not encounter any long-term stent-related AEs in the form of stent migration or occlusion, which are not commonly reported with EUS-PGS because of the distance between the stomach and pancreas, probably causing stent migration or fracture. In a study by Oh et al.15 involving 23 patients with PJAS, long-term stent-related complications were reported in five patients with FCSEMS in the form of asymptomatic stent fracture at the gastric end, stent occlusion, and asymptomatic stent migration.

In EUS-PJS, we assume that the anatomical proximity and adhesions between the afferent jejunal limb and the pancreas after PD may have some advantages, such as a low risk of pancreatic juice leakage even in cases of failure of stent deployment and stent stability with a low risk of migration. Furthermore, we believe that puncturing the MPD through the anastomosis line while avoiding puncture through the pancreatic parenchyma may reduce the risk of early closure of the pancreaticojejunal fistula following stent removal, thus increasing the chance of achieving a permanent fistula and an earlier stent-free state (Fig. 6). In the current study, three patients were kept stent-free within one year of follow-up, and the number of stent exchange procedures in these patients was 2, 4, and 5. However, another study on long-term outcomes in a larger number of patients and a longer follow-up period is required to evaluate the efficacy of this procedure in providing a permanent patent pancreaticojejunal fistula.

Fig. 6.

Puncture tract in endoscopic ultrasound-guided pancreaticogastrostomy (EUS-PGS) and EUS-guided pancreaticojejunostomy (EUS-PJS). (A) The puncture tract in EUS-PGS, where the fistula is created through the pancreatic parenchyma. (B) Puncture tract in EUS-PJS, where the fistula is created through the anastomosis and not the parenchyma, using a forward-viewing scope.

The retrospective design and single-center location, along with the limited number of subjects, pose limitations to our study. However, this study is valuable because it evaluated the safety and usefulness of EUS-PJS for pancreatic drainage in PJAS and highlighted the possibility of creating a permanent fistula in these patients, thereby reducing the number of reinterventions required to maintain drainage.

Thus, EUS-guided PJS appears to be a safe and effective treatment option for patients with PJAS. Additional prospective randomized controlled studies with a large number of participants are necessary.

Supplementary Material

Article information Continued