Postoperative pancreatic fluid collections (POPFCs) are complications that frequently occur after pancreatic surgery.^1,2 If a pancreatic fistula persists, then it can lead to an intraperitoneal abscess, sepsis, or the formation and rupture of a pseudoaneurysm, which can be fatal. Traditional treatment for POPFCs has comprised surgical or percutaneous drainage. Recently, endoscopic procedures such as endoscopic transpapillary pancreatic duct drainage and endoscopic ultrasound-guided drainage (EUSD) have been suggested as treatment options.^3,4 Because of concerns regarding difficult intubation and postprocedural pancreatitis with transpapillary pancreatic drainage,⁴ we performed EUSD of POPFCs. However, there is no consensus regarding whether plastic stents/tubes or metal stents should be used for EUSD. Furthermore, when plastic stents/tubes are used, whether a nasocystic plastic tube (NPT), double-pigtail plastic stent (DPS), or a combination of both of these should be used has not been determined. This study aimed to analyze the outcomes of EUSD of POPFCs at our hospital and examine the safety and effectiveness of plastic stents/tubes.

Study design

This retrospective review of collected consecutive data was approved by the ethics committee of Saku Central Hospital Advanced Care Center (institutional review board number: R201704-11).

Patients

This study included patients who underwent EUSD of POPFCs at our hospital between October 2018 and January 2024. The indications for EUSD were computed tomography (CT) detection of the fluid collection cavity accompanied by increased inflammatory markers and symptoms such as fever, abdominal pain, nausea, and vomiting. EUSD was also indicated if the fluid collection cavity observed with CT enlarged in the absence of these symptoms.

Procedures

EUSD was performed using a therapeutic linear array echoendoscope (UCT-260; Olympus Corp.) and an ultrasound system (ME2 Premia Plus; Olympus Corp.). A 19-G needle (EZ Shot 3 Plus; Olympus Corp.) was used to perform the puncture. After puncturing the fluid collection cavity, a 0.025-inch guidewire (VisiGlide2; Olympus Corp.) was placed. After repeated passes over the guidewire at the puncture site using the puncture needle, the site was dilated with a 6-mm dilation balloon (REN; Kaneka Medical). An uneven double-lumen cannula (Piolax Medical Devices Inc.) was used to place an additional 0.025-inch guidewire (EndoSelector; Boston Scientific Corp.). A 7-Fr DPS (AdvanixJ [Boston Scientific Corp.] or Through & Pass Double Pit [Gadelius Medical K.K.]) was placed. Thereafter a 7-Fr NPT (pigtail-type QuickPlaceV ENBD; Olympus Corp.) was placed. In principle, we attempted to place one NPT and one DPS; however, we adjusted our stent choice based on the procedural difficulty and cognitive function of the patient. The NPT was removed after confirming the improvement of symptoms and inflammatory response as well as the reduction or disappearance of the fluid collection cavity through a CT evaluation or contrast examination. Complete removal of the remaining DPS was performed after confirming the disappearance of the fluid collection cavity using CT. Antibiotics were routinely administered intravenously to all patients before and after EUSD. Successful drainage case of one NPT and one DPS deployment is shown in Fig. 1.

Figure 1. Stent selection for the initial EUSD procedure and treatment course. EUSD, endoscopic ultrasound-guided drainage; NPT, nasocystic plastic tube; DPS, double-pigtail plastic stent; LAMS, lumen-apposing metal stent.

Evaluated parameters and definitions

The technical success rate, clinical success rate, type and number of stents, total number of additional interventions, stent retention duration, adverse event rate, and recurrence rate were evaluated. Technical success was defined as successful stent placement in the fluid collection cavity. Clinical success was defined as symptom improvement and resolution of the fluid collection cavity without conversion to techniques other than EUSD. The total number of additional interventions was defined as the number of required endoscopic procedures excluding initial EUSD and final stent removal. Adverse events were defined as bleeding, infection, and stent migration; these were categorized as early (within 14 days of EUSD) or late (≥ 15 days after EUSD). Recurrence was defined as the reappearance of a fluid collection cavity at the previously treated site after complete stent removal.

Baseline characteristics

Among the patients with POPFCs, 27 who underwent EUSD were included in this study. The baseline characteristics are shown in Table 1. The median age was 70 years (range, 16–83 years), and 20 (74.1%) patients were male. The most common surgical procedure was distal pancreatectomy, which was performed for 23 (85.2%) patients. The most common underlying disease was pancreatic cancer, which was observed in 13 (48.1%) patients. The median time from surgery to drainage was 14 days (range, 5–160 days). The time from surgery to drainage for 21 (77.8%) patients was within 30 days, and that for six (22.2%) patients was 31 days or more. The median size of the fluid collection cavity was 7.0 cm (range, 3.4–16.1 cm). The fluid collection cavity was unilocular in 26 (96.3%) patients and multilocular in one (3.7%) patient. The median follow-up period from the initial EUSD procedure to the final CT evaluation was 526 days (range, 36–1,703 days).

Table 1 . Baseline Characteristics.

Characteristic	Value
Age (yr)	70 [16–83]
Sex, male	20 (74.1)
Surgical procedure
Distal pancreatectomy	23 (85.2)
Pancreaticoduodenectomy	2 (7.4)
Laparoscopic and endoscopic cooperative surgery	1 (3.7)
Distal gastrectomy	1 (3.7)
Pathological diagnosis
Pancreatic ductal adenocarcinoma	14 (51.9)
Intraductal papillary mucinous neoplasm	3 (11.1)
Neuroendocrine neoplasm	2 (7.4)
Autoimmune pancreatitis	2 (7.4)
Mucinous cystic neoplasm	1 (3.7)
Solid-pseudopapillary neoplasm	1 (3.7)
Epidermoid cyst	1 (3.7)
Ampullary carcinoma	1 (3.7)
Gastric carcinoma	1 (3.7)
Duodenal heterotopic pancreas	1 (3.7)
Time from surgery to drainage (day)	14 [5–160]
Maximum diameter of POPFCs (cm)	7.0 [3.4–16.1]
Locularity of POPFCs
Unilocular	26 (96.3)
Multilocular	1 (3.7)
Follow-up period (day)	526 [36–1,703]

Values are presented as median [range] or number (%).

POPFCs, postoperative pancreatic fluid collections.

Clinical outcomes

Clinical outcomes are shown in Table 2. The technical success and clinical success rates were both 100%. The type and number of stents used during the initial drainage procedure were as follows: one NPT and one DPS for 19 (70.4%) patients; two DPS for four (14.8%) patients; one NPT for three (11.1%) patients; and one lumen-apposing metal stent (LAMS) for one (3.7%) patient. Six (22.2%) patients required one additional intervention and two (7.4%) patients required two additional interventions.

Table 2 . Clinical Outcomes.

Outcomes	Value
Technical success	27 (100)
Clinical success	27 (100)
Drainage procedure
One NPT and one DPS	19 (70.4)
Two DPS	4 (14.8)
One NPT	3 (11.1)
LAMS	1 (3.7)
Total number of additional interventions
One	6 (22.2)
Two	2 (7.4)
Time to NPT removal (day)	7 [3–23]
Time to DPS removal (day)	68 [14–223]
Adverse events	4 (14.8)
Early
Bleeding	1 (3.7)
Infections	0 (0.0)
Stent migration	0 (0.0)
Late
Bleeding	0 (0.0)
Infections	0 (0.0)
Stent migration	3 (11.1)
Recurrence	1 (3.7)

Values are presented as number (%) or median [range].

NPT, nasocystic plastic tube; DPS, double-pigtail plastic stent; LAMS, lumenapposing metal stent.

The treatment course after the initial drainage procedure is shown in Fig. 2. One NPT and one DPS were initially placed in 19 patients; of these patients, five required additional interventions. Two patients exhibited insufficient treatment effects with the initial combination of one NPT and one DPS and required additional intervention; the stent was converted to a LAMS for one patient and balloon dilation of the fistula was performed for the other patient. Both patients experienced subsequent improvement. Re-enlargement of the fluid collection cavity was observed in three patients with one DPS after removal of the NPT. In one of these patients, puncture and aspiration of the remaining fluid collection cavity (separate from the space where the stents had been placed) were performed. In another patient, an additional DPS was inserted. In the third patient, a new NPT was placed; after confirming reduction of the fluid collection cavity, it was replaced with two DPS. All three patients exhibited improvement after additional treatment. No patients required necrosectomy. The other seven patients (who initially received two DPS, one NPT, or one LAMS) did not require additional intervention after stent removal.

Figure 2. Successful EUSD of a POPFC with one NPT and one DPS. (A) CT image of POPFC (yellow arrow: POPFC). (B) EUS image of POPFC (yellow arrow: POPFC). (C) Fluoroscopic image of one NPT and one DPS. (D) CT image of the decreased POPFC, resulting in the decision to remove the NPT. (E) Endoscopic image of the removed DPS. (F) CT image after complete stent removal showing no recurrence. EUSD, endoscopic ultrasound-guided drainage; POPFC, postoperative pancreatic fluid collection; NPT, nasocystic plastic tube; DPS, double-pigtail plastic stent; CT, computed tomography.

Adverse events occurred in four (14.8%) patients. Bleeding was an early adverse event in one patient with pancreatic body cancer infiltrating the celiac axis to the common hepatic artery who underwent distal pancreatectomy with left gastric artery reconstruction as conversion surgery. On postoperative day 126, a CT evaluation revealed a fluid collection cavity at the pancreatic transection margin, and EUSD was performed on postoperative day 133. One NPT and one DPS were placed. The day after EUSD, hemorrhagic shock developed. Angiography revealed extravasation near the common hepatic artery transection site. Hemostasis was achieved by performing coil embolization. DPS migration occurred as a late adverse event in three patients. Although the DPS had completely deviated toward the gastrointestinal tract, an increased fluid collection cavity and infection were not observed, and no additional interventions were required.

Recurrence was observed in one patient. Re-enlargement of the fluid collection cavity was observed with CT 29 days after complete removal of two DPS; therefore, a second EUSD procedure was performed. At 285 days after the second EUSD procedure, recurrence did not occur and the two DPS remained in place.

At our hospital, the technical success and clinical success rates of EUSD using mainly using plastic stents/tubes were 100%, and the adverse event rate was 14.8%. A recent systematic review that included cases treated with plastic stents/tubes, fully covered self-expandable metal stents (FCSEMS), and LAMS did not perform analyses based on the type or number of stents; however, it reported an overall technical success rate of 94%, clinical success rate of 87%, and adverse event rate of 14%.³ The outcomes of EUSD using mainly plastic stents/tubes at our hospital were comparable to those previously reported. The ability to achieve these outcomes without necrosectomy, which is a more invasive procedure, is a positive aspect of this technique.

During this study, 77.8% of patients underwent EUSD within 30 days after surgery, indicating that the majority of fluid collections were treated early. Bacterial peritonitis associated with EUSD was not observed, suggesting that early treatment of fluid collections (within 30 days after surgery) with EUSD is feasible and safe; therefore, it could prevent complications that may occur if drainage is delayed. A systematic review of the timing of EUSD for POPFCs reported no significant differences in technical success rates, clinical success rates, or adverse event rates with early drainage (within 30 days) and late drainage (≥ 31 days thereafter).⁵ Fluid collection drainage after acute pancreatitis is recommended after 4 weeks, after the collection has become encapsulated, to avoid the risk of nonlocalized bacterial peritonitis associated with EUSD.⁶ However, POPFCs are often accompanied by adhesions caused by surgical trauma, even at an early stage, suggesting that the risk of nonlocalized bacterial peritonitis associated with early drainage may not be high.⁵ Therefore, EUSD of POPFCs during the early postoperative period is considered acceptable. However, whether earlier intervention is correlated with better outcomes is unknown because of insufficient data. Therefore, the ability to draw firm conclusions about the optimal timing of fluid collection drainage is limited.

The types of stents used for EUSD include plastic stents/tubes, FCSEMS, and LAMS. FCSEMS and LAMS are expected to provide effective drainage and facilitate necrosectomy because of their ability to maintain dilation of the fistula. However, necrosectomy is often unnecessary for POPFCs, and studies have reported no significant differences in clinical success rates when DPS were used.⁷ There are various placement configurations for plastic stents/tubes, including one DPS or multiple DPS, one NPT only, and one NPT and one DPS.^8–10 All of these methods have resulted in a high technical success rate of 100%, high clinical success rate of more than 93%, and low adverse event rate of 7.7% or less.^8–10 However, comparative studies of these different approaches have not been conducted. POPFCs have many major blood vessels near the pancreatic resection margin and are often small. Plastic stents/tubes are preferable for avoiding vascular injury caused by electrocautery needles of LAMS, or the ends of FCSEMS/LAMS. To reduce fluid viscosity by lavage of the fluid collection cavity with an NPT, we chose the approach comprising one NPT and one DPS as the first-line option.

Of the 19 patients who received one NPT and one DPS at the time of the initial drainage procedure, 14 experienced good outcomes and did not require additional treatment. Among the five patients who required additional interventions, two experienced inadequate drainage with the initial combination of one NPT and one DPS, and three experienced inadequate drainage with one DPS after removal of the NPT. The fluid collection cavity had a lobular shape in the two patients who experienced inadequate drainage with the initial combination of one NPT and one DPS. Because dilation of the fistula was effective in both patients, the placement of three or more plastic stents/tubes or FCSEMS/LAMS to create a larger space between the plastic stents/tubes and gastric wall was considered necessary for such lobular fluid collection cavities. On the other hand, for the three patients who experienced inadequate drainage with one DPS after removal of the NPT, except for one patient with a multilocular fluid collection cavity, plastic stent obstruction and blockage of the space between the gastric wall and plastic stent were considered the causes of these complications. To avoid these complications, the use of two DPS after removal of the NPT is preferable.

Although the NPT is useful for reducing fluid viscosity, its negative impact on patient comfort and daily functioning is a significant disadvantage. Although earlier conversion to DPS when the fluid viscosity is decreased is an alternative way to reduce the duration of NPT placement, the initial use of two or more DPS or FCSEMS/LAMS is clearly superior in terms of quality of life. A stepwise strategy of attempting initially three DPS and switching to FCSEMS or LAMS if drainage is inadequate or if infection occurs is the best option in terms of cost.

In this study, bleeding was observed as an early adverse event in one patient. We suspected that the site where the common hepatic artery was ligated during surgery had been weakened by the pancreatic fistula, thus leading to bleeding caused by contact with the stent. If the CT examination indicates that the distal end of the stent is positioned close to the cut vessel margin after EUSD, then bleeding could be avoided by correcting the stent position. Previous studies have reported bleeding after EUSD, including bleeding from the fluid collection site, puncture tract, or arterial injury; however, they did not mention the type of stent used or the cause of bleeding.¹¹ Contrast-enhanced CT should be performed before EUSD to identify pseudoaneurysms and vessels travelling in the fluid collection cavity, and color Doppler EUS should be performed to avoid vessels along the puncture route to reduce the risk of bleeding. Additionally, emergency management, such as interventions involving radiology or surgery, is required for bleeding events. DPS migration was observed as a late adverse event in three patients. However, they did not require additional interventions; therefore, DPS migration was not considered a serious complication. Additional data regarding the timing of DPS removal should be collected.

This study had some limitations. First, this retrospective study was conducted at a single center and included a small number of patients. Multicenter trials and larger studies are necessary to prove that plastic stents/tubes are superior to other options for managing POPFCs. Second, although we initially aimed to use one NPT and one DPS, the actual stent choices varied. However, this approach allowed stent selection flexibility (NPT, DPS, LAMS), which is advantageous when managing patients with varying anatomical complexities, cognitive functions, or procedural difficulties. Selecting stents based on patient-specific factors added a personalized aspect to treatment. Third, we could not compare one NPT and one DPS with the other stents of two or more DPS, FCSEMS, or LAMS. Our data only suggested that EUSD with plastic stents/tubes is effective; however, whether it is optimal is unclear.

In conclusion, the placement of plastic stents/tubes is safe and effective when performing EUSD of POPFCs.

None.

The data that support the findings of this study are available from the corresponding author on reasonable request.

No potential conflict of interest relevant to this article was reported.