Introduction
The management of peripancreatic fluid collections (PFCs), particularly walled-off necrosis (WON), has evolved from open surgical to minimally invasive procedures, which include laparoscopic, percutaneous, or endoscopic approaches [1]. The endoscopic step-up approach is associated with a reduced risk of new-onset multiorgan failure, shorter hospital stay, reduced overall cost, and significantly reduced risks of external fistulas, with no significant difference in mortality [1]
[2]
[3]
[4]. Endoscopic ultrasound (EUS)-guided drainage of WON, with either plastic stents or metal stents, is the mainstay of WON management [5]
[6]; however, there is continuing debate and limited high quality evidence comparing the use of dedicated large caliber metal stents over multiple plastic stents during EUS-guided WON drainage.
Lumen-apposing metal stents (LAMSs) provide several advantages over plastic stents in the drainage of WON, which include their larger diameter that helps with the more efficient drainage of necrotic debris, permits direct endoscopic necrosectomy (DEN), and acts to tamponade any entry-site bleeding [4]
[5]. Bang et al., in a single-center randomized controlled trial (RCT), compared LAMSs with multiple plastic stents for drainage of WON. They observed no significant difference in treatment success, clinical adverse events (AEs), readmission rates, length of stay, or treatment cost, except for a shorter procedure duration with LAMSs [5]. Additional reinterventions within 72 hours were required for persistent symptoms in a quarter of subjects in the LAMS arm and >50% of patients in the plastic stent arm [5].
Karstensen et al., in their RCT, showed noninferiority of plastic stents in terms of technical and clinical success, AEs, and length of hospital stay [7]; however, a weekly dilation of the cystogastric tract was performed as per protocol, and was not necessarily based on outcome. This may have led to more effective drainage but also to multiple reinterventions.
In a retrospective study from our center, the endoscopic step-up approach showed that three-quarters of patients with WON improved with an index procedure using a biflanged metal stent (BFMS), without the need for further intervention [8]. These observations suggest that the use of wide caliber metal stents is associated with fewer reinterventions. The aim of this study was to evaluate the efficacy of BFMSs and plastic stents for WON drainage without the need for further reintervention, as part of the step-up approach.
Methods
Study design
This study was a prospective, single-center, randomized comparative trial conducted at the two units of the Asian Institute of Gastroenterology, India between September 2020 and February 2022. The study was approved by the Institutional Ethics Committee.
All consecutive patients admitted with symptomatic WON (jaundice, pain, anorexia, vomiting, visible lump, sepsis, organ failure) confirmed on cross-sectional imaging (either computed tomography or magnetic resonance imaging) and adhering to the inclusion criteria were recruited. WON was defined as mature encapsulated liquefied necrosis located within the pancreas or peripancreatic space, according to the revised Atlanta classification of acute pancreatitis ([Fig. 1]) [9]. Written informed consent was obtained from each patient before their enrollment in the study, according to the inclusion and exclusion criteria ([Table 1]).
Table 1 Inclusion and exclusion criteria for patients with walled-off necrosis (WON) undergoing step-up endoscopic drainage in this trial.
|
|
Criteria
|
|
EUS, endoscopic ultrasound; INR, international normalized ratio.
|
|
Inclusion
|
Age >18 years
|
|
Diagnosis of WON based on imaging (debris between 10% and 50%)
|
|
Documented history of acute pancreatitis
|
|
Symptomatic WON
|
|
Feasibility of EUS-guided drainage
|
|
Exclusion
|
Pregnancy or lactating mother
|
|
Severe coagulopathy (INR >1.5 and/or platelet count <50 × 109/L) that is not correctable
|
|
Fluid collection other than WON
|
|
Not willing to participate
|
Fig. 1 Endoscopic ultrasound image showing an area of walled-off necrosis.
Randomization and masking
All patients included in the study were allocated to receive either BFMSs (Group A) or plastic stents (Group B) in a ratio of 1:1 using block randomization, with each block consisting of four patients. A computer-generated randomization sequence was provided by the study coordinator and placed in sequentially numbered sealed opaque envelopes that were opened by the study investigators before the procedure. There was no blinding owing to the different nature of the two arms of the study.
Procedural technique
All EUS-guided drainage procedures were performed by experienced endoscopists after the patient had received intravenous antibiotics (preprocedure) and with the patient in the left lateral position under monitored anesthesia-care sedation that was supervised by an anesthetist.
Biflanged metal stents (BFMSs)
The biflanged metal stent used (length 20 mm, diameter 16 mm and 26 mm at the flares; NAGI; Taewoong Medical, Gyeonggi-do, South Korea) is a dedicated self-expandable fully covered metal stent (SEMS) for the drainage of PFCs ([Fig. 2]
a). The standard placement procedure was followed for the drainage, as described previously [8].
Fig. 2 Endoscopic images showing: a a biflanged metal stent; b plastic stents; c necrotic tissue requiring reintervention in a patient with a metal stent in situ.
Plastic stents
The standard steps of placement of plastic stents were followed. The WON was initially punctured using a 19G fine-needle aspiration needle and a guidewire was coiled within the WON under fluoroscopic guidance, which was followed by tract dilation up to 15 mm with a radial expansion balloon, passage of two guidewires, and finally placement of two 7-Fr double-pigtail plastic stents (length 4 or 5 cm; Zimmon biliary stent; Cook Medical, Indiana, USA) under endoscopic and fluoroscopic guidance ([Fig. 2]
b). The study methodology is shown in [Fig. 3] and Table 1s, see online-only Supplementary material.
Fig. 3 Consort diagram showing participant flow through the study and main outcomes. BFMS, biflanged metal stent; ITT, intention to treat; SIRS, systemic inflammatory response syndrome.
Reinterventions
All the patients were monitored as inpatients for 48 to 72 hours after the procedure. Those who improved symptomatically (hemodynamically stable off inotropes/absence of fever for 48 hours/resolution of systemic inflammatory response syndrome (SIRS) and organ failure/no requirement for analgesia in the past 24 hours) were discharged on a 7-day course of oral antibiotics. For patients who had persistent or new-onset symptoms, the endoscopic step-up approach was followed, which included one or more of the following: tract dilation with placement of a nasocystic drain, additional plastic stent placement into the same tract if in Group B, DEN ([Fig. 2]
c), percutaneous drainage, or surgery. The decision to intervene or to crossover to the other arm was taken by the treating physician.
Treatment protocol
Where there had been symptomatic improvement and reduction in the size of the WON, as
assessed by cross-sectional imaging at the end of 48–72 hours, the patient was discharged
and followed up within 4 weeks, or earlier if symptoms recurred and/or new symptoms
appeared. If there was complete symptomatic relief and the WON had reduced to <50% of the
original size and was <5 cm in its largest diameter at 4 weeks, both BFMSs and plastic
stents were removed. Magnetic resonance cholangiopancreatography was performed prior to
stent removal. In the plastic stent arm, if there was a persistent collection or evidence of
disconnected pancreatic duct syndrome (DPDS), plastic stents were not removed. In the BFMS
group, the metal stent was replaced by a plastic stent if there was a residual collection,
or evidence of DPDS.
Follow-up
All patients were assessed daily while in hospital, then at 1 and 4 weeks. At each visit, information was collected regarding signs and symptoms, and AEs. Patients were then followed up at 3 months via telephone and at 6 months via a hospital visit to assess their medium-term outcomes. At 6 months, cross-sectional imaging was done to assess for recurrence of the WON.
Definitions
Technical success was defined as the correct deployment of the stent with endoscopic visualization of WON fluid drainage. Clinical success was defined as the complete resolution of symptoms (hemodynamically stable off inotropes/absence of fever for 48 hours/resolution of SIRS and organ failure/no requirement for analgesia in the past 24 hours) and significant reduction in the size of the WON (<50% of original size and <5 cm in largest diameter at 4-week follow-up). Procedure duration was defined as the time taken from EUS scope insertion to the deployment of the stent with visualization of WON fluid drainage.
Adverse events were defined as any undesirable event during the study period whether related or unrelated to the EUS-guided drainage [10]
[11]. Stent migration was classified as clinically insignificant (spontaneous external migration of the stent after complete resolution of WON) or clinically significant, either internal (migration of the stent into the PFC cavity, with or without persistent WON) or external (migration of the stent into the stomach with persistent WON). Recurrence was defined as the occurrence of a new collection at 6 months after the achievement of initial clinical success.
Outcomes
The primary outcome was reintervention-free clinical success at 4 weeks. The secondary outcomes were: overall clinical success, AEs, number of reinterventions, and duration of hospital stay. For the medium-term outcomes, clinical success, recurrence, evidence of DPDS, chronic pancreatitis, or new-onset diabetes mellitus were noted at 6 months.
Statistical methods and sample size calculation
The sample size calculation was based on the hypothesis of detecting a statistically significant difference in the percentage intervention-free clinical success at 4 weeks in the BFMS and plastic stent groups. Based on previously published data, intervention-free clinical success in the BFMS group was expected to be 75% and in the plastic stent group was expected to be 45% [5]
[8]. Therefore 41 patients were required in each group to reject the null hypothesis with an 80% power and 5% type 1 error. In each group, an additional five patients were recruited to account for lose to follow-up and early mortality. Finally, a total of 92 patients were enrolled for the study.
Continuous variables are described as mean (SD) or median and interquartile range; categorical variables are described as numbers and percentages. Categorical variables were compared using the chi-squared test or Fisher’s exact test wherever appropriate. Continuous variables were compared using an independent t test or Mann–Whitney U test wherever appropriate. A P value of <0.05 was considered to indicate a statistically significant difference. All statistical analyses were performed using SPSS version 26 (IBM Corp., Armonk, New York, USA).
Results
A total of 165 patients were screened ([Fig. 3]) between September 2020 and February 2022, of whom 92 patients were finally included for randomization, with 46 patients allocated to each study arm. Alcohol and biliary pancreatitis were present in 38.0% (35/92) and 25.0% (23/92) of patients, respectively; while idiopathic and other etiologies were present in 37.0% (34/92). SIRS was present in 78.3% (72/92) of patients on admission. There were no significant differences in baseline characteristics between the two study arms ([Table 2]).
Table 2 Baseline characteristics of the study population allocated to receive either a biflanged metal stent (BFMS) or a plastic stent.
|
|
BFMS (n = 46)
|
Plastic stent (n = 46)
|
|
ASA, American Society of Anesthesiologists; IQR, interquartile range; SIRS, systemic inflammatory response syndrome; WON, walled-off necrosis.
|
|
Age, mean (SD), years
|
34.9 (12.4)
|
36.8 (11.1)
|
|
Sex, male, n (%)
|
38 (82.6%)
|
38 (82.6%)
|
|
Etiology of acute pancreatitis, %
|
|
|
14
|
21
|
|
|
16
|
7
|
|
|
16
|
18
|
|
Total leukocyte count, mean (SD), mm3
|
9133 (5407)
|
8459 (3948)
|
|
C-reactive protein, mean (SD), mg/L
|
43.6 (60.7)
|
40.03 (43.9)
|
|
SIRS, n (%)
|
37 (80.4)
|
35 (70.1)
|
|
Intensive care admission, n (%)
|
38 (82.6)
|
34 (73.9)
|
|
ASA score, n (%)
|
|
|
0 (0)
|
0 (0)
|
|
|
23 (50)
|
26 (56.5)
|
|
|
21 (45.7)
|
19 (41.3)
|
|
|
2 (4.4)
|
1 (2.2)
|
|
Indication, n (%)
|
|
|
43 (93.5)
|
42 (91.3)
|
|
|
31 (67.4)
|
33 (71.7)
|
|
WON location, n (%)
|
|
|
12 (26.1)
|
11 (23.9)
|
|
|
34 (65.2)
|
35 (76.1)
|
|
Debris, median (IQR), %
|
25 (20–30)
|
24 (15–30)
|
|
WON size, mean (SD), cm
|
14.2 (4.7)
|
15.0 (15.8)
|
Primary outcome
Clinical success at 1 month without any reintervention was observed in 67.4% of patients (31/46) in the BFMS arm and 43.5% (20/46) in the plastic stent arm (P = 0.02) on intention to treat (ITT) analysis ([Table 1]).
There were two patients in the plastic stent group, who crossed over to the BFMS group after randomization owing to an intraprocedural bleed in one case and persistent sepsis requiring DEN in another. There was one death in each study arm. On per-protocol analysis, the reintervention-free clinical success rate at 1 month was 68.1% (32/47) in the BFMS arm and 44.2% (19/43) in the plastic stent arm (P = 0.02).
Reinterventions
In the plastic stent arm, 24 patients underwent a total of 48 reinterventions ([Fig. 4]). The indications included persistent SIRS at 48 hours after the procedure and/or persistent symptoms. The reinterventions included: placement of a nasocystic tube (n = 17), removal of necrotic tissue cystogastrostomy site (n = 13), removal of plastic stents, followed by DEN and then replacement of plastic stents (n = 9), placement of percutaneous drainage tube (n = 7), and crossover to the BFMS group (n = 2).
Fig. 4 Sankey diagram showing the study outcomes. BFMS, biflanged metal stent.
In the BFMS arm, 13 patients required a total of 25 reinterventions, which included: removal of necrotic tissue to unclog the BFMS lumen (n = 10), placement of a nasocystic tube (n = 8), DEN (n = 5), and placement of a percutaneous drain (n = 2).
Secondary outcomes
Overall clinical success at 1 month was similar in the two groups. On ITT analysis, 41 patients (89.1%) in both arms achieved clinical success (P > 0.99). On per-protocol analysis, 43 patients (91.5%) in the BMFS arm and 39 (90.7%) in the plastic stent arm achieved overall clinical success (P = 0.90). The mean (SD) procedure time of the initial procedure was 19.6 (7.5) minutes in the BFMS group and 21.4 (9.0) minutes in the plastic stent group (P = 0.29). The mean (SD) total days of hospitalization were significantly lower in the BFMS group (7.0 [3.4] days) compared with the plastic stent group (9.1 [5.5] days; P = 0.04).
Adverse events
Intraprocedural bleeding at the puncture site was seen in three patients in the plastic stent arm. In two patients, the bleeding was self-limiting and did not require further intervention (not considered as AEs). In one patient, significant brisk bleeding was observed and the operator decided to crossover to a BFMS, with the patient requiring a blood transfusion (moderate severity, AGREE grade IIIa) [10]
[11]. No AEs led to prolonged hospital stay. There were however two deaths, one in each arm, both of which were related to persistent septic shock requiring multiple reinterventions.
Stent migration
Stent migration was seen in three patients (6.5%) in the BFMS arm. In one patient, there was partial internal migration of the BFMS, which was retrieved successfully by endoscopy. In two patients, there was complete external migration into the duodenum without symptomatic luminal obstruction and the stents were removed endoscopically. In the plastic stent arm, two patients’ stents (4.4%) migrated externally without any symptoms.
Medium-term follow-up
Three patients in the BFMS arm and six patients in the plastic stent arm were lost to
follow-up at 6 months. At 6 months, overall clinical success was seen in 95.3% of patients
(41/43) in the BFMS group and 94.6% (35/37) in the plastic stent group (P > 0.99). Recurrences of the fluid collection were similar in both groups
(2/43 [4.7%] in group A vs. 2/37 [5.4%] in group B; P > 0.99).
In the BFMS group, one of the recurrent PFCs was associated with active alcohol abuse with
acute pancreatitis and a second patient had DPDS. The two recurrences in the plastic stent
group were related to fresh episodes of recurrent acute pancreatitis. At 6-month follow-up,
there were similar incidences of DPDS, chronic pancreatitis, and new-onset diabetes mellitus
in both groups (all P>0.05) ([Table 3]).
Table 3 Comparison of clinical outcomes of the study participants in the biflanged metal stent (BFMS) and plastic stent groups.
|
|
BFMS
|
Plastic stents
|
P value
|
Relative risk (95%CI)
|
|
DPDS, disconnected pancreatic duct syndrome; IQR, interquartile range; ITT, intention to treat.
|
|
Reintervention-free clinical success at 1 month
|
|
|
31/46 (67.4%)
|
20/46 (43.5%)
|
0.02
|
0.6 (0.4–0.9)
|
|
|
32/47 (68.1%)
|
19/43 (44.2%)
|
0.02
|
0.6 (0.3–0.9)
|
|
Overall clinical success at 1 month
|
|
|
41/46 (89.1%)
|
41/46 (89.1%)
|
1
|
(0.3–3.2)
|
|
|
43/47 (91.5%)
|
39/43 (90.7%)
|
0.90
|
0.9 (0.2–3.4)
|
|
Number of reinterventions, median (IQR)
|
0 (0–1)
|
1 (0–2)
|
0.03
|
|
|
|
10
|
17
|
|
|
|
• Removal of necrotic tissue at cystogastrostomy site
|
8
|
13
|
|
|
|
|
5
|
9
|
|
|
|
|
NA
|
2
|
|
|
|
|
2
|
7
|
|
|
|
Procedure duration, mean (SD), minutes
|
19.58 (7.5)
|
21.44 (9.0)
|
0.29
|
|
|
Hospital stay, mean (SD), days
|
7.04 (3.4)
|
9.09 (5.5)
|
0.04
|
|
|
Adverse events
|
|
|
0.56
|
0.5 (0.0–5.6)
|
|
|
0
|
1 (2.2%)
|
|
|
|
|
1 (2.2%)
|
1 (2.2%)
|
|
|
|
Stent migration, n (%)
|
3 (6.6%)
|
2 (5.4%)
|
0.65
|
1.5 (0.2–9.6)
|
|
Medium-term outcomes (6 months), n (%)
|
N = 43
|
N = 37
|
|
|
|
|
41 (95.3%)
|
35 (94.6%)
|
>0.99
|
0.9 (0.1–5.8)
|
|
|
2 (4.7%)
|
2 (5.4%)
|
>0.99
|
1.0 (0.9–1.1)
|
|
|
19 (44.2%)
|
22 (59.5%)
|
0.17
|
1.3 (0.8–2.1)
|
|
|
7 (16.3%)
|
6 (16.2%)
|
0.99
|
1.0 (0.8–1.2)
|
|
|
1 (2.3%)
|
2 (5.4%)
|
0.59
|
1.0 (0.9–1.1)
|
Discussion
The current randomized study compares the outcomes of WON drainage using BFMSs or plastic stents. The observed overall clinical success at 1 month was similar in both groups; however, the reintervention-free clinical success rate at 4 weeks (primary outcome) was significantly higher in the BFMS arm compared with the plastic stent arm. For the secondary outcomes, BFMSs had significant benefits in terms of the number of reinterventions and duration of hospital stay. The primary procedural duration in the BFMS arm was slightly lower, although this was not statistically significant, probably because of the cold stent system. Medium-term follow-up at 6 months showed similar rates of clinical success, recurrent PFCs, evidence of chronic pancreatitis, and new-onset diabetes mellitus. This study emphasizes that a greater number of reinterventions are required in the plastic stent arm than in the BFMS arm to achieve similar clinical success.
The only two published RCTs comparing plastic and metal stents for WON have shown similar efficacy [5]
[7]. In contrast, several non-randomized studies have shown superiority of LAMSs [6]
[12]
[13]
[14]
[15]. Siddiqui et al. showed, in a large retrospective study, that the mean number of interventions required for WON resolution was significantly lower in the LAMS group compared with the plastic stent group [14]. They also showed, on multivariate analysis, that plastic stents were the sole negative predictor for successful resolution of WON (odds ratio [OR] 0.18, 95%CI 0.06–0.53) after adjusting for age, sex, and WON size. Bapaye et al., in their retrospective study comparing BFMSs and multiple plastic stents, observed higher clinical success, fewer DEN sessions and AEs, and shorter hospital stay in the BFMS group [13]. Kartensen et al., in their recent study comparing LAMSs and plastic stents for large WON, showed that the number of necrosectomy sessions required to achieve clinical success was similar in both groups; however, the treatment algorithm included a scheduled weekly dilation of the drainage tract in the plastic stent arm to prevent its spontaneous closure, a regime rarely followed in clinical practice [7]. This approach added reinterventions in the plastic stent group after the index drainage to achieve clinical success. Bang et al., in their RCT, observed that a higher proportion of patients achieved reintervention-free clinical success in the LAMS group (75%) compared with the plastic stent group (45%) [5].
Several published meta-analyses on the same subject have reported contradictory results in terms of clinical success and AEs; however, none evaluated the need for reinterventions to achieve clinical success [15]
[16]
[17]. A recent meta-analysis of RCTs showed that there is no significant difference in the clinical outcomes, including need for necrosectomy and number of reinterventions, between LAMSs and plastic stents [18]. In the current study, we observed that overall clinical success was similar in both groups; however, reintervention-free clinical success was higher in the BFMS group. Also, the number of reinterventions in the BFMS group were fewer compared with the plastic stent group.
The “step-up approach” is the accepted preferred management for any symptomatic WON [1]
[2]
[19]. Therefore, as expected, high clinical success was achieved at 1 month using this approach in both arms of our study. To achieve this outcome however, the plastic stent group had a greater number of reinterventions, which led to prolonged hospital stays. While the initial cost of a plastic stent may appear lower than that of a BFMS, it is important to consider all potential implications. Factors such as the need for more frequent reinterventions or follow-up procedures could contribute to higher overall treatment costs for plastic stents over time. It is worth exploring the full spectrum of costs and benefits associated with both options to make an informed decision that aligns with the best interests of the patient. This may also imply that sick patients and those with multiple co-morbidities and worse clinical parameters may benefit more from a metal stent than plastic stents, as they will have to undergo fewer procedures.
The overall AEs in the current study were comparable in the two groups. Significant
LAMS-related AEs are observed if the stent remains in situ for more than 3 weeks, and these
can be minimized by its timely removal by 3 weeks, after documenting WON resolution [5]. BFMSs and LAMSs are slightly different, as the former, owing to minimal
lumen-apposing capability, can spontaneously migrate externally on resolution of the WON. In
the current study, one patient in the plastic stent arm crossed over to a BFMS because of
intraprocedural puncture site bleeding. There were similar rates of stent migration (4%–6% at
1 month) in both groups. At 6-month follow-up, there were similar rates of clinical success,
recurrence of fluid collection, DPDS, chronic pancreatitis, and new-onset diabetes mellitus in
both groups; as was also reported in a previous RCT [5]. Similarly, in our study, the medium-term outcomes were not significantly different.
Although there was a high prevalence of DPDS in the current study, recurrent PFC was seen in
only one patient at 6 months. A recent study showed that replacement of metal stents with
plastic stents in patients with DPDS did not reduce PFC recurrence [20].
This is the first RCT comparing BFMSs (rather than LAMSs) with plastic stents and demonstrates similarly high overall success in both groups, although emphasizing that increased numbers of interventions and prolonged hospital stay were observed in those who had plastic stents placed at their index drainage. LAMSs have perpendicular flanges that secure tissue layers, maintaining adequate apposition of the two lumens. BFMSs have flared ends that are wider and smoother to prevent migration. Both types of stent come in variable diameters and allow for easy necrosectomy access; clinical success is similar with both metal stents [21]. Therefore, metal stents should be considered as first-line therapy in WON drainage; however, BFMSs should be removed within 4 weeks to avoid delayed stent-related AEs.
There are a few limitations of the current study, apart from being a single-center study. Cost analysis was not performed owing to non-standardized hospital care costs and the study was non-blinded. LAMSs are costlier than BFMSs, which makes the latter a more appealing option in resource-limited settings. We excluded WON that had >50% debris, which has been included in a few previously published studies [5]
[22]. As placement of a nasocystic tube was done in a number of patients, an alternate method to predict its requirement and its early placement during the primary procedure would likely reduce the requirement for another reintervention in the plastic stent arm.
In conclusion, this study reports that, in EUS-guided drainage of WON, BFMSs provide higher reintervention-free clinical success, reduce the overall number of reinterventions, and shorten the hospital stay compared with multiple plastic stents. These findings need to be confirmed in multicenter studies.
