Predictive factors and outcomes of endoscopic necrosectomy in patients with acute pancreatitis and walled-off necrosis

Abstract

BACKGROUND

Endoscopic ultrasound-guided drainage using lumen-apposing metal stents (LAMS) has emerged as the first-line approach for managing walled-off necrosis (WON). However, certain patients require escalation to direct endoscopic necrosectomy, for which the predictive factors have not been completely defined.

AIM

To determine the predictors of direct endoscopic necrosectomy following LAMS placement in patients with WON and to assess the clinical outcomes and safety.

METHODS

A retrospective analysis of prospectively collected data from patients with acute pancreatitis who were admitted to the Govind Ballabh Pant Institute of Postgraduate Medical Education in Delhi, India, between January 2020 and October 2023 was conducted. Patients with acute pancreatitis and symptomatic WON who underwent LAMS placement were included in the study. Patients aged < 18 years with asymptomatic WON, pseudocysts, postsurgical collections, or a history of percutaneous drainage were excluded. Data were collected using a predesigned form. Clinical details, treatments, interventions, and outcome data were recorded.

RESULTS

A total of 104 patients with symptomatic pancreatic WON who underwent LAMS placement were included in this study. Of these, 36 required endoscopic necrosectomy. Univariate analysis revealed that fever [odds ratio (OR) = 4.47, 95% confidence interval (CI): 1.85-10.79, P = 0.00], systemic inflammatory response syndrome (OR = 5.85, 95%CI: 2.03-16.83, P = 0.001), pancreatic necrosis > 30% (OR = 14.6, 95%CI: 1.87-113.86, P = 0.001), WON in the pancreatic head (OR = 4.246, 95%CI: 1.80-10.0, P = 0.001), and collection size (OR = 1.18, 95%CI: 1.04-1.34, P = 0.009) were the predictors of endoscopic necrosectomy. Subsequently, multivariate analysis indicated that the extent of necrosis was an independent predictor of the requirement for necrosectomy (OR = 1.085, 95%CI: 1.026-1.148, P < 0.004). Clinical success was higher in the non-necrosectomy group than in the necrosectomy group (88.2% vs 69.4%).

CONCLUSION

Early identification of these predictive variables can guide treatment planning for WON and facilitate early necrosectomy, thereby improving the clinical outcomes.

Key Words: Acute pancreatitis; Walled-off necrosis; Necrosectomy; Endoscopy; Lumen-apposing metal stent

Core Tip: Walled-off necrosis (WON) is a well-known complication of acute pancreatitis. Although many patients respond to endoscopic drainage alone, some require direct endoscopic necrosectomy for optimal outcomes. This retrospective study identified the clinical, biochemical, and radiological predictors of necrosectomy in patients undergoing placement of lumen-apposing metal stents. Of the examined factors, fever, anemia, hypoalbuminemia, systemic inflammatory response syndrome, pancreatic necrosis > 30%, WON size, and WON location were predictors of necrosectomy. Early identification of these predictors can guide clinicians in planning timely interventions, thereby improving outcomes and reducing morbidity associated with delayed or unnecessary necrosectomy.

INTRODUCTION

Acute pancreatitis (AP) is one of the most common gastrointestinal causes of hospitalization worldwide[1]. While most patients (80%) experience a mild and self-limiting course, 10%-20% of patients with AP develop severe disease characterized by necrosis of the pancreatic parenchyma or extrapancreatic adipose tissue[2]. These patients exhibit a complex, prolonged clinical course with organ failure and local and systemic complications, leading to high morbidity and associated mortality of up to 20%-30%[3].

Pancreatic and peripancreatic collections are well-recognized complications of AP. According to the revised Atlanta classification, these collections are classified into peripancreatic fluid collections, acute necrotizing collections, pseudocysts, and walled-off necrosis (WON), depending on the amount of debris and the time elapsed since the onset of pancreatitis[4]. WON is characterized by a mature encapsulated collection of pancreatic and/or peripancreatic necrotic tissue that occurs 4 weeks after disease onset. Approximately 50% of patients with WON are asymptomatic and may experience spontaneous resolution within 6 months of onset[5]. WON drainage is indicated in patients with superadded infections, persistent abdominal pain, anorexia, weight loss, gastrointestinal obstruction, or biliary obstruction[6].

In recent years, there has been a paradigm shift in the management of WON from surgical drainage to noninvasive percutaneous or endoscopic routes. Endoscopic ultrasound (EUS)-guided drainage has been simplified by the development of dedicated stents, specifically lumen-apposing metal stents (LAMS)[7]. LAMS are fully covered, “dumbbell”-shaped braided nitinol metal stents with broad antimigratory flanges, providing effective lumen-to-lumen apposition and a large caliber (up to 20 mm), which facilitates the drainage of solid debris[8,9]. The incorporation of a cautery tip into the delivery system (“hot” LAMS) further simplifies the drainage technique. Depending on the collection size and/or extent, WON may require either a single access or multiple transgastric or transduodenal accesses for adequate drainage. Several studies have established the efficacy and safety of endoscopic drainage with LAMS, resulting in improved clinical success rates, shortened procedure times, reduced morbidity and need for reintervention, and decreased recurrence rates compared with surgical or percutaneous drainage.

Approximately 60%-80% of patients with WON undergoing endoscopic treatment respond well to transmural drainage alone[6]. Nonetheless, a subset of patients may require additional direct endoscopic necrosectomy (DEN) because of persistent symptoms, infection, or inadequate drainage[10,11]. Identifying the predictors of necrosectomy remains challenging in clinical practice. Several imaging and clinical parameters have been assessed as potential predictors of necrosectomy. Factors such as the extent of necrosis (> 30%-50%), paracolic extension, non-homogeneous or loculated collections, and signs of infection have been implicated[10,12-14]. Furthermore, the timing of DEN - whether performed immediately at the time of stent placement or delayed - has emerged as a variable that may potentially affect the outcomes[8,15]. Despite the growing body of literature, consensus is lacking, and studies often vary in design, population, and endpoints. To address these shortcomings, this study examined the clinical, biochemical, and radiological predictors of necrosectomy in patients who underwent LAMS placement for WON.

MATERIALS AND METHODS

Study design and setting

This single-center retrospective observational study was conducted at the Department of Department of Gastroenterology, Govind Ballabh Pant Institute of Postgraduate Medical Education and Research, New Delhi, between January 2020 and October 2023. All adult patients (≥ 18 years of age) diagnosed with WON based on the revised Atlanta classification and who underwent EUS-guided drainage with LAMS placement were included[4,7]. EUS-guided cystogastrostomy or cystoduodenostomy was performed in patients with symptomatic WON, including those with infection, anorexia, weight loss, or gastrointestinal or biliary tract obstruction. Patients with asymptomatic WON, pancreatic pseudocysts, prior surgical or percutaneous catheter drainage, or incomplete clinical or imaging follow-up were excluded from the study. Ethical clearance was obtained from the Institutional Ethics Committee, No. F.1/IEC/MAMC/85/04/2021/No. 470.

Procedure details

EUS-guided cystogastrostomy: An experienced therapeutic endoscopist performed all procedures using a linear echoendoscope (GF-UCT180; Olympus, Tokyo, Japan) under intravenous anesthesia. The collection was identified, and a cautery-enhanced bi-flange stent (EASE^®, Scientific Healthcare, Faridabad, India) with a diameter of 16 mm and a length of 3 cm was deployed using the transgastric or transduodenal approach (Figure 1). Patients were monitored for symptom resolution and underwent repeat imaging (computed tomography or ultrasound) if symptoms persisted or at 4 weeks before stent removal. Antibiotics were administered either empirically or based on the results of microbial culture.

Figure 1 Magnetic resonance imaging and endoscopic ultrasound guided drainage of lumen-apposing metallic stents. A: Magnetic resonance T2-weighted imaging coronal; B: Axial views showing walled-off necrosis of the pancreas with significant debris; C: Endoscopic ultrasound (EUS) reveals a large walled-off necrosis with echogenic contents and debris (orange arrow); D: EUS-guided lumen-apposing metallic stents (LAMS) are being placed, with distal flanges (yellow arrow) deployed; E: Both distal flanges (yellow arrowhead) inside the cavity and proximal flanges (orange arrowhead) in the gastric cavity are deployed seen in EUS; F: Endoscopic image post-deployment of LAMS and pus seen oozing out from out the LAMS.

DEN: DEN was performed only for those who required it based on their clinical condition, such as persistent sepsis, systemic inflammatory response syndrome (SIRS), or incomplete resolution of WON. A forward-viewing endoscope (GF-UCT180; Olympus, Tokyo, Japan) was introduced into the cyst cavity through the indwelling LAMS, and the position of the metallic stent was identified. Any obstruction in the stent was cleared, and the cyst cavity was assessed for the presence of pus, debris, and the cyst wall. Snares, forceps, or baskets were used for debridement, and betadine irrigation was performed at the end of the procedure (Figure 2). The procedure was repeated only if necessary, based on the persistence of symptoms.

Figure 2 The procedure of direct endoscopic necrosectomy. A: Endoscopic image shows blocked lumen-apposing metallic stents with pus seen oozing through it; B: Blockage in the lumen-apposing metallic stents is removed, and the scope is negotiated into the cavity, which shows a large cavity with necrotic debris; C: Necrotic debris is removed with Roth net forceps; D: Post-necrosectomy with minimal debris and healthy wall cavity seen.

If imaging confirmed the resolution of WON, LAMS was removed 4 weeks after insertion[16]. If patients had persistent symptoms and incomplete WON resolution, the stents were removed, and 10 Fr × 7 cm double-pigtail plastic stents were placed[17].

Study objectives

Primary objective: To identify the clinical and radiological predictors of the need for endoscopic necrosectomy after LAMS placement.

Secondary objectives: To assess technical success, clinical success, and adverse events. To evaluate recurrence, reintervention, and mortality rates over 6 months.

Study definitions

Technical success: Accurate deployment of LAMS into the WON cavity under EUS guidance[18].

Clinical success: Resolution of symptoms, with ≥ 50% reduction in WON size on imaging within 4 weeks[19].

Necrosectomy: Endoscopic removal of necrotic debris through the LAMS tract.

Clinical failure: Reintervention need, recurrence, or death within 6 months of drainage.

Adverse events: Any complications following the procedure, such as infection, bleeding, pseudoaneurysm, stent migration, or perforation.

Statistical analysis

Data were analyzed using the SPSS software (version 26.0; IBM Corp., Armonk, NY, United States). All patient data were anonymized prior to analysis. Continuous variables were presented as mean ± SD or median with interquartile range and were compared using the independent samples t-test or Mann-Whitney U test, as appropriate. Categorical variables were expressed as frequencies and percentages and compared using χ² or Fisher’s exact test.

Variables identified in the univariate analysis were subjected to receiver operating characteristic (ROC) curve analysis to determine their predictive accuracy of the need for necrosectomy. Variables with P < 0.1 in the univariate analysis were included in the multivariate logistic regression analysis to identify independent predictors of necrosectomy. The predicted probabilities from the final multivariate model were subsequently used to generate an ROC curve and calculate the area under the curve (AUC), assessing the model’s discriminatory performance. Statistical significance was set at P < 0.05.

RESULTS

During the study period, 108 patients with symptomatic WON underwent EUS-guided LAMS placement. EUS-guided LAMS were successfully placed in 104 patients with WON. In two patients, the cyst cavity could not be punctured, and in two others, the LAMS was misdeployed inside the collection (technical success: 96.2%). Of the 104 patients, 36 (34.6%) required subsequent endoscopic necrosectomy, and the remaining 68 (65.4%) were managed successfully with LAMS alone (Figure 3). The two groups were comparable in terms of age and sex. The mean age of the study group was 34.93 ± 11.81 years. The most common etiology of AP in both groups was biliary (41 patients, 39.4%), followed by alcohol (38 patients, 36.5%). Furthermore, the most common indication for LAMS insertion in both groups was abdominal pain; however, fever as an indication was more prevalent in the necrosectomy group [odds ratio (OR) = 4.47, 95% confidence interval (CI): 1.85-10.79, P = 0.001]. Systemic complications, such as acute kidney injury, acute lung injury, and intensive care unit admission, were comparable between the groups; nonetheless, SIRS was more frequent in the necrosectomy group than in the LAMS-alone group (OR = 5.85, 95%CI: 2.03-16.83, P = 0.001) (Table 1). Furthermore, anemia (OR = 1.18, 95%CI: 1.04-1.34, P = 0.001) and hypoalbuminemia (OR = 0.330, 95%CI: 0.163-0.667, P = 0.001) were predictors of necrosectomy in the univariate analysis.

Figure 3 Study flow chart. ANP: Acute necrotising pancreatitis; WON: Walled off Necrosis; PCD: Percutaneous drainage; LAMS: Lumen apposing metallic stent.

Table 1 Comparison of baseline and outcome parameters of the patients in the study group, n (%).

Variable	Total (n = 104)	Necrosectomy group (n = 36)	Non-necrosectomy group (n = 68)	P value
Demographic parameters
Age (year), mean ± SD	34.93 ± 11.81	38.08 ± 10.82	33.26 ± 12.03	0.047
Gender				0.258
Male	79	25	54
Female	25	11	14
Smoking	20 (19.2)	7 (19.4)	13 (19.1)	0.968
Alcohol	37 (35.5)	12 (33.3)	25 (36.7)	0.728
Comorbidity
HTN	3 (2.8)	1 (2.7)	2 (2.9)	0.717
Diabetes mellites	6 (5.7)	3 (8.3)	3 (4.4)	1.000
Etiology
Alcohol	38 (36.5)	12 (33.3)	26 (38.2)	0.788
Biliary	41 (39.4)	16 (44.4)	25 (36.7)	0.564
Traumatic	1 (0.09)	0 (0)	1 (1.4)	0.453
Idiopathic	24 (23.0)	8 (22.2)	16 (23.5)	0.788
SIRS	66 (63.4)	31 (86.1)	35 (51.5)	0.001
Acute kidney injury	7 (6.7)	4 (11.1)	3 (4.4)	0.195
Acute lung injury	10 (9.6)	5 (13.8)	5 (7.3)	0.280
ICU admission	32 (30.7)	15 (41.6)	17 (25)	0.080
Indication of metallic stent insertion
Pain	98 (94.2)	34 (94.4)	64 (94.1)	0.946
Fever	51 (49.0)	26 (72.2)	25 (36.7)	0.001
Jaundice	8 (7.6)	4 (11.1)	4 (5.8)	0.341
Vomiting	10 (9.6)	2 (5.5)	8 (11.7)	0.307
Lump	5 (4.8)	0 (0)	5 (7.3)	0.095
Time from AP to LAMS (days)	49.04 ± 24.22	44.64 ± 23.89	51.38 ± 24.25	0.178
Hematological and Biochemical parameters, mean ± SD
Haemoglobin (gm/dL)	10.1 ± 1.86	9.23 ± 1.66	10.55 ± 1.81	0.001
Total leucocyte count/mm³	14341 ± 8054	17372 ± 7742	12908 ± 7805	0.006
Platelets (lac/mm3)	3.24 ± 1.37	3.30 ± 1.38	3.2074 ± 1.38	0.742
Urea (mg/dL)	26.4 ± 22.5	30.58 ± 32.72	24.19 ± 14.52	0.171
Creatinine (mg/dL)	0.8 ± 0.5	0.8 ± 0.6	0.7 ± 0.4	0.635
Total Bilirubin (mg/dL)	0.86 ± 1.027	1.03 ± 0.751	1.51 ± 0.631	0.186
AST (IU/L)	33.08 ± 23.8	39.17 ± 35.1	29.8 ± 14.06	0.580
ALT (IU/L)	24.5 ± 23.32	26.8 ± 31	23.3 ± 18	0.466
ALP (IU/L)	151.4 ± 94.8	186.5 ± 122	133 ± 70	0.006
Total protein (gm/dL)	6.72 ± 0.9	6.6 ± 0.8	6.7 ± 1.0	0.572
Albumin (gm/dL)	3.2 ± 0.68	2.9 ± 0.58	3.3 ± 0.6	0.001
Computed tomography imaging characteristics
Size of WON long axis	12.17 ± 3.76	13.48 ± 3.819	11.38 ± 3.37	0.005
WON location
Head	53 (50.9)	23 (63.8)	20 (29.4)	0.001
Body	99 (95.1)	35 (97.2)	64 (94.1)	0.481
Tail	76 (73.0)	30 (83.3)	46 (67.6)	0.086
Entire	26 (25)	17 (47.2)	9 (13.2)	0.001
Amount of necrosis, mean ± SD	53.98 ± 22.11	71.25 ± 12.5	48.09 ± 21.09	0.001
< 30%	21 (20.1)	1 (2.7)	20 (29.4)	0.001
> 30%	83 (79.8)	35 (97.2)	48 (70.5)
CTSI, mean ± SD	8.02 ± 1.35	8.56 ± 1.229	7.74 ± 1.334	0.003
Extension of WON	77 (74.0)	24 (66.6)	53 (77.9)	0.284
Pelvic	2 (1.9)	1 (2.7)	1 (1.4)	1.000
Paracolic	25 (24.0)	11 (30.5)	14 (20.5)	0.675
Additional drainage
PCD	13 (12.5)	6 (16.6)	7 (10.2)	0.350
DPT1	15 (14.4)	10 (27.7)	5 (7.3)	0.005
ENCD through LAMS	6 (5.7)	6 (16.6)	0 (0)	0.001
Outcome parameters
Clinical outcome: Success	85 (81.7)	25 (69.4)	60 (88.2)	0.022
Hospital stay duration	16.87 ± 9.64	21.56 ± 11.34	12.68 ± 7.95	0.001
LAMS duration	28.00 ± 10.22	31.21 ± 10.42	26.58 ± 9.80	0.033
WON recurrence in 6 months	13 (12.5)	8 (25)	5 (7.3)	0.032
Adverse events
Bleeding	6 (5.7)	2 (5.5)	4 (5.8)	0.356
Stent migration	11 (10.5)	7 (19.4)	4 (5.8)	0.136
Colonic fistula	1 (0.9)	0 (0)	1 (1.4)	0.561
6-month mortality	4 (3.8)	2 (5.5)	2 (2.9)	0.239

¹Double pig tail was placed in the lumen-apposing metal stent (LAMS) alone group while removal of LAMS, whereas in the Direct endoscopic necrosectomy group it was placed either after necrosectomy or after removal of LAMS.

HTN: Hypertension; SIRS: Systemic inflammatory response syndrome; ICU: Intensive care unit; AP: Acute pancreatitis; LAMS: Lumen-apposing metal stent; AST: Aspartate aminotransferase; ALT: Alanine aminotransferase; ALP: Alkaline phosphatase; WON: Walled-off necrosis; CTSI: Computed tomography severity index; PCD: Percutaneous catheter drainage; DPT: Double pig tail; ENCD: Endoscopic Naso cystic drainage.

There was no difference between the two groups in terms of the timing of LAMS insertion relative to the onset of AP. Imaging features, such as the amount of necrosis (OR = 1.082, 95%CI: 1.045-1.120, P = 0.001), pancreatic necrosis > 30% (OR = 14.6, 95%CI: 1.87-113.86, P = 0.001), number of WON (OR = 1.2, 95%CI: 0.504-2.857, P = 0.009), WON located in the pancreatic head (OR = 5.865, 95%CI: 2.248-15.307, P = 0.001), and WON replacing the whole pancreas (OR = 4.246, 95%CI: 1.80-10.0, P = 0.001), computed tomography severity index (CTSI) score > 6 (OR = 1.629, 95%CI: 1.165-2.274, P = 0.03), and WON size in the long axis (OR = 1.18, 95%CI: 1.04-1.34, P = 0.009) were associated with the need for necrosectomy compared with LAMS alone (Table 1). The extension of the WON to the paracolic gutter or pelvic extension and percutaneous catheter drainage placement were similar between the two groups; nevertheless, DPT placement was more frequent in the necrosectomy group (Table 1).

In the multivariate analysis, the amount of necrosis was the only independent predictor of necrosectomy [OR = 1.085, 95%CI: 1.026-1.148), P < 0.004]. The presence of fever and SIRS increased the risk of necrosectomy by > 1.5 times; however, these factors did not reach statistical significance (Table 2). In our study, the timing of DEN was not standardized and was determined by the presence of symptoms and signs. The timing of DEN was not significantly associated with clinical success (P = 0.944), WON recurrence at 6 months (P = 0.942), or mortality at 6 months (P = 0.471). The patients were stratified into two groups based on the timing of DEN administration: Within 7 days (n = 21) and after 7 days (n = 15). Further analysis was then conducted. No statistically significant differences were observed in any of the evaluated outcomes between the two groups (Table 3).

Table 2 Multivariate logistic regression analysis of predictors for necrosectomy.

Variable	OR	95%CI lower	95%CI upper	P value
CTSI score	0.729	0.416	1.281	0.272
Amount of necrosis	1.085	1.026	1.148	0.004
< 30% or > 30% of necrosis	0.381	0.013	10.781	0.571
WON location - head	1.294	0.234	7.158	0.768
WON location - entire	1.125	0.168	7.512	0.903
Size of WON	1.095	0.897	1.337	0.371
SIRS	1.584	0.304	8.24	0.585
Hemoglobin	0.815	0.567	1.171	0.269
TLC	1.0	1.0	1.0	0.914
Albumin	0.438	0.142	1.348	0.15
Fever	1.937	0.526	7.127	0.32

CI: Confidence interval; OR: Odds ratio; CTSI: Computed tomography severity index;.

WON: Walled-off necrosis; LA: Largest axis; SIRS: Systemic inflammatory response syndrome; TLC: Total leukocyte count.

Table 3 Comparison of outcomes based on timing of first direct endoscopic necrosectomy, n (%).

Outcome parameter	First DEN ≤ 7 days (n = 21)	First DEN > 7 days (n = 15)	P value
Mean number of DEN sessions	1.67 ± 0.98	1.33 ± 0.49	0.244
Clinical success	16 (76.1)	12 (80)	0.806
Hospital stay duration (days), mean ± SD	22.5 ± 12.7	20.3 ± 9.3	0.288
WON recurrence within 6 months	4 (19)	4 (26.7)	1.000
6-month mortality	1 (4.8)	1 (6.7)	1.000

DEN: Direct endoscopic necrosectomy; WON: Walled-off necrosis.

ROC curve analysis

ROC curve analysis was performed to assess the predictive value of the variables identified using univariate analysis for DEN requirement after LAMS placement. Of these variables, the amount of necrosis exhibited excellent predictive accuracy, with an AUC of 0.819. Hypoalbuminemia (AUC = 0.712), anemia (AUC = 0.706), total leukocyte count (AUC = 0.718), and fever (AUC = 0.677) demonstrated good discrimination ability. Factors such as the size of the WON on the long axis (AUC = 0.681), SIRS (AUC = 0.673), and WON involving the entire pancreas (AUC = 0.670) showed fair predictive value. Conversely, necrosis > 30% (AUC = 0.633) and the CTSI score (AUC = 0.659) displayed poor discriminatory ability. The location of WON in the pancreatic head (AUC = 0.500) had no predictive significance. ROC analysis of a multivariate logistic regression model incorporating all variables identified in the univariate analysis showed good discriminatory ability in predicting the need for necrosectomy, with an AUC of 0.892 (95%CI: 0.821-0.962), which was superior to that of the individual parameters (Figure 4).

Figure 4 Receiver operating characteristic curves for the multivariate predictive model and selected univariate predictors of necrosectomy. Hb: Hemoglobin; TLC: Total leucocyte count.

Clinical outcomes and complications

The overall clinical success rate was 81.7%. Patients requiring necrosectomy exhibited a slightly lower clinical success rate (69.4%) than those treated with LAMS alone (88.2%), and this difference was statistically significant (P = 0.022). Moreover, the mean hospital and LAMS durations were longer in the DEN group than in the LAMS group (Table 1). WON recurrence within 6 months was significantly higher in the necrosectomy group than in the non-necrosectomy group (25% vs 7.3%, P = 0.0032). Procedure-related complications included bleeding, pseudoaneurysm formation, and stent migration. The most common postprocedural complication in the study group was stent migration, followed by bleeding, which was comparable between the two groups. The incidence of pseudoaneurysms was similar between the two groups. At 6 months, four patients died, with two in each group. In the necrosectomy group, one patient developed persistent sepsis and multiorgan failure, and the other died owing to pseudoaneurysm-related hemorrhage. In the LAMS-only group, both patients died from the worsening of sepsis.

DISCUSSION

In this single-center retrospective analysis of patients who underwent EUS-guided drainage for WON using LAMS, 36 (34.6%) patients required subsequent necrosectomy. The findings indicated that the amount of pancreatic necrosis independently increased the need for necrosectomy by 8.5% (OR = 1.085, 95%CI: 1.026-1.148, P < 0.004). Clinical factors, such as the presence of fever and SIRS, increased the risk of necrosectomy by 4.47 times (OR = 4.47, 95%CI: 1.85-10.79) and 5.85 times (OR = 5.85, 95%CI: 2.03-16.83, P = 0.001), respectively, in univariate analysis. Laboratory parameters, including anemia (AUC = 0.706), leukocytosis (AUC = 0.718), and hypoalbuminemia (AUC = 0.712), had good predictive value for the need for necrosectomy. Furthermore, imaging factors, such as the CTSI score, number, location, and size of WON, were identified as predictors of necrosectomy. The combined model demonstrated superior discriminative ability (AUC = 0.892) compared with the individual parameters in the ROC analysis, suggesting that a complex interplay of multiple factors influences the need for DEN. The presence of early predictors could potentially inform the decision to implement immediate interventions in high-risk patients.

However, the timing of necrosectomy remains a matter of controversy. Several investigations on the management of WON advocate delayed drainage, and a similar protocol has been followed in necrosectomy, which is effective, safe, and associated with fewer necrosectomy sessions compared with other approaches. Delaying DEN allows tract maturation and alleviates the risk of complications. However, recent multicenter studies have examined the safety, feasibility, and benefits of immediate and early necrosectomy[20]. For instance, Yan et al[10] and Bang et al[21] showed that upfront endoscopic necrosectomy considerably reduced the number of reinterventions, improved the clinical outcomes, and shortened the length of hospitalization compared with the step-up approach, without increasing the adverse events. However, a recent multicenter study from Japan did not demonstrate the impact of necrosectomy timing on clinical success, which is consistent with the findings of our study[22]. The European Society of Gastrointestinal Endoscopy advocates a tailored approach regarding the timing of DEN based on clinical signs, symptoms (including persistent SIRS and organ failure), and response to passive drainage[23]. Given the conflicting results on the timing of DEN administration and clinical outcomes, further multicenter RCTs are needed to establish the optimal timing.

In addition to the timing of necrosectomy, several studies have attempted to identify the predictors of necrosectomy. Factors such as WON size, the percentage and extent of necrosis, persistent organ failure, and multiple organ failure are associated with worse outcomes and the need for more aggressive interventions, as reported in the literature. One of the most critical predictors identified in our study and in the literature is the presence and extent of necrosis[24]. González-González et al[12] identified necrosis as the only factor predicting the need for necrosectomy, a finding similar to that of the present study. They found that necrosis increased the need for necrosectomy by 18% (OR = 1.18, 95%CI: 1.01-1.39), whereas in our study, necrosis was associated with an 8% increase (OR = 1.085, 95%CI: 1.026-1.148). Cosgrove et al[24] and Chandrasekhara et al[25] have also shown that an extensive collection and the presence of necrosis increase the need for necrosectomy. Studies have shown that necrosis exceeding 30%-50% significantly reduces the likelihood of clinical success with drainage alone and correlates with an increased need for DEN[13,14].

Several studies have reported that hypoalbuminemia is a predictor of persistent organ failure and infected pancreatic necrosis[14,26]. This condition is independently associated with both inflammation and compromised nutritional status in AP, which predisposes patients to extended hospital stays and increased mortality. In our study too, albumin level was identified as a predictor of necrosectomy. Similarly, paracolic gutter extension of the necrotic collection, a marker of an extensive disease burden, has been shown to compromise drainage efficacy and predict the need for repeated debridement[25]. This anatomic extension is a complex, loculated cavity that cannot be adequately accessed via the standard transluminal route, which limits the efficacy of single-site drainage and necessitates multiple sessions of DEN or alternative access routes, such as retroperitoneal or percutaneous drainage. González-González et al[12] observed that patients with paracolic extension had a significantly higher need for reintervention and a lower clinical success rate with LAMS alone, necessitating adjunctive DEN in > 60% of the cases. Other studies have also reported that the paracolic extension is an independent predictor of clinical failure with transmural drainage alone, supporting its role in preprocedural planning and risk stratification[10,24]. Yan et al[10] showed that the paracolic extension was associated with an OR of 0.08 for successful resolution without DEN, highlighting its significance. However, in this study, this feature was not found to be significantly associated with the need for necrosectomy.

EUS-guided drainage using LAMS followed by selective DEN is a safe and effective strategy. Reported adverse events associated with necrosectomy include stent migration, infection, bleeding, and the development of pseudoaneurysms. Of these, stent migration is a known complication of LAMS, with reported rates ranging from 5% to 15%. Migration was slightly higher in our study, consistent with early placement in large necrotic cavities, which might have destabilized the stent. Walter et al[27] documented a 6% migration rate using LAMS, highlighting its reduced migration compared with plastic stents owing to its broad flanges and dumbbell shape. Considering other studies, Shah et al[8] and Siddiqui et al[7] detected migration rates of 3.5%-8%, with migration most commonly noted in collections > 12 cm or with paracolic extension. The bleeding rate in our cohort was 5.5%, which aligns with previous reports of 2%-8%, depending on the complexity of necrosis and the use of necrosectomy. Bang et al[21] observed a bleeding rate of 3% in 70 patients who had undergone necrosectomy, which was mostly controlled endoscopically. Although rare, pseudoaneurysm formation is a dreaded complication with a potentially fatal course, and its incidence increases in patients having large WON with multiple necrosectomy sessions. González-González et al[12] reported a pseudoaneurysm rate of 6.6%, which is comparable to the rate in our study (8.8%).

Our study has several limitations, including its retrospective design and single-center data collection. Moreover, in this study, necrosectomy was performed based on symptoms and was not protocolized, which could lead to potential variability in the timing and technique of DEN. Numerous studies have demonstrated that protocolized necrosectomy leads to fewer reinterventions and complications[19]. With the introduction of new devices and an expanding therapeutic armamentarium, the endoscopic management of WON is evolving; however, these were not used in our study, and can be viewed as a limitation. Cavity irrigation with hydrogen peroxide has been shown to improve clinical success. Maharshi et al[28] conducted a randomized controlled trial comparing nasocystic irrigation with H₂O₂ alone vs the placement of biflanged metal stents in WON. This study found that adverse events, technical success, and clinical success were comparable, although the H₂O₂ group experienced longer procedure times and hospital stays. A multicenter study comparing DEN with and without hydrogen peroxide for WON showed that necrosectomy with H₂O₂ was linked to a higher clinical success rate (OR = 3.30, P = 0.033) and a quicker resolution (OR = 2.27, P < 0.001), without a significant increase in adverse events[29]. Similarly, an extensive pooled analysis of 186 patients reported a clinical success rate of 91.6% and low complication rates, supporting the safety and efficacy of this technique[30]. Newer devices, such as the Endo Rotor (Interscope Inc., United States), Over-the-Scope Grasper tools (Ovesco Endoscopy AG, Germany), flexible flanged LAMS, and waterjet-assisted necrosectomy systems, have emerged as promising adjuncts in the management of WON, aiming to improve clinical success rates and reduce procedural failure[6,31,32].

Despite these limitations, the study’s strength lies in its comprehensive analysis of the variables and their alignment with global trends. Future research should focus on developing predictive models that integrate radiological, clinical, and procedural variables to facilitate individualized therapies. Furthermore, large-scale, multicenter, randomized controlled trials are essential for standardizing the protocol for endoscopic interventions, identifying accurate pre-endoscopy predictors of reinterventions and complications, and exploring novel techniques and devices that enhance efficiency and improve both short-term and long-term outcomes of the procedure.

CONCLUSION

WON management remains challenging and warrants a multidisciplinary approach to improve patient outcomes. Advances in EUS-guided interventions over the past decade have positioned them at the forefront of managing WON. Endoscopic necrosectomy techniques and devices continue to evolve, and dedicated devices are expected to enhance efficacy and safety while improving the comfort of both patients and endoscopists. However, our understanding of the timing and necessity of necrosectomy must be improved. This study reinforces the predictive ability of various clinical, laboratory, and imaging factors - such as the presence of fever, SIRS, leukocytosis, hypoalbuminemia, necrosis, and head of the pancreas WON - in identifying the need for necrosectomy in patients undergoing LAMS drainage for WON. Early recognition of these variables can help guide treatment planning, predict resource utilization, and augment clinical outcomes.