Clinical Success Impacts Overall Survival in Patients Undergoing Palliation for Malignant Gastric Outlet Obstruction Using Enteral Stents or Endoscopic Ultrasound Guided Gastroenterostomy

ABSTRACT

Background

EUS‐guided gastroenterostomy (EUS‐GE) and enteral stents (ES) are minimally invasive options for managing malignant gastric outlet obstruction (mGOO). We aimed to compare EUS‐GE with ES with respect to clinical success assessed objectively and overall survival.

Methods

Retrospective review of patients > 18 years who underwent either ES or EUS‐GE for palliation of mGOO between January 2021 and January 2024 was done. Primary outcomes were clinical success (defined as improvement of oral intake and tolerance of higher consistency diet compared to baseline by 2 points using GOOSS) and overall survival. Secondary outcomes were adverse events, reintervention rates, and reintervention‐free survival.

Results

One‐hundred and seventy‐two patients were included (112 ES group and 60 EUS‐GE). The most common background malignancy was pancreatic cancer (30.2%), followed by gall‐bladder cancer (26.2%). Technical success rates were similar (EUS‐GE 95.08% and ES 99.09%; p = 0.123). Clinical success was achieved more often in those who underwent EUS‐GE (94.7% vs. 77.5%, p = 0.004, OR 5.233). Overall survival (ES 112 days vs. EUS‐GE 177 days, p = 0.787) was not statistically different in the two groups. Although reintervention was needed in ES more often (18.9% vs. 3.3%, p = 0.005, OR 5.392), reintervention‐free survival (ES 93 days vs. EUS‐GE 177 days, p = 0.581) did not differ. Overall survival was lower in those who did not achieve clinical success (clinical success—no, 49 days; yes, 134 days, p = 0.013).

Conclusion

EUS‐GE could achieve better outcomes than ES in terms of the nature of oral intake feasible with lesser need for reintervention, with patients achieving clinical success surviving longer.

1. Introduction

Upper GI tract malignancies are frequently complicated by gastric outlet obstruction (GOO) which often leads to nausea, vomiting, and poor oral intake and can preclude the oral delivery of medications [1]. The primary approach for managing GOO is palliation of symptoms and reestablishing oral intake by restoring gastrointestinal continuity. Palliative care in terminal patients with GOO should ideally be as least invasive as possible, provide long‐term symptom relief, cause the least morbidity, and improve quality of life. In the past years, before the advent of therapeutic endoscopy, the treatment of choice for these patients was represented by the surgical bypass [2], but this intervention was burdened by a significant morbidity (25%–35%) and perioperative mortality of about 2% [3]. Endoscopic placement of a self‐expanding metal stent (SEMS) across the site of obstruction was first described in 1992 and became the first‐line therapy for malignant gastric outlet obstruction [4]. Multiple uncontrolled case‐series studies have demonstrated SEMSs to be safe and effective, with technical success of 90%–100% and clinical success of 67%–100% [5]. The advent of therapeutic endoscopic ultrasonography (EUS) with the development of lumen apposing metal stent (LAMS) with an electrocautery‐enhanced tip has revolutionized the practice of advanced endoscopy [6]. EUS‐guided gastroenterostomy (EUS‐GE) was first described by Binmoeller and Shah [7] in an animal model and subsequently came into practice with the first case reported by Khasab et al. [8]. The major advantage of EUS‐GE was the placement of the stent proximal to the tumor reducing the rate of tumor ingrowth. In addition, the placement of the stent proximally may be associated with better emptying of gastric contents with lesser need for propulsion [9]. However, EUS‐GE is technically demanding and requires a high level of skill and expertise in interventional EUS [10]. Moreover, severe adverse events requiring surgery can occur in cases of LAMS maldeployment [11]. On the other hand, ES is safe and effective (particularly for short‐term outcomes), widely available, and technically affordable and does not require advanced technical skills in interventional EUS [12]. Therefore, when dealing with mGOO, one must determine which procedure to employ. Till date, only a few comparative studies of EUS‐GE versus ES have been published, and most studies have defined clinical success by the ability to continue oral intake rather than considering the nature and consistency of oral intake that is possible. In addition, no large comparative studies are available from the Indian subcontinent. Survival is an outcome that is previously underreported in studies comparing EUS‐GE and ES. The study was undertaken to compare outcomes of EUS‐GE with ES in a large cohort of patients from a tertiary care hospital in western India in terms of clinical success and overall survival.

2. Methods

2.1. Study Population and Setting

This is a retrospective observational cohort study of patients who underwent either ES or EUS‐GE for palliation of gastric outlet obstruction between January 1, 2021, and January 31, 2024. All patients who were > 18 years of age and underwent palliation of mGOO with either EUS‐GE or ES were included. Patients with surgically altered anatomy and those who underwent surgical GE were excluded. Information was retrieved from prospectively collected data from the Endoscopy unit and Nutrition Clinic at a tertiary care oncology institute with the help of the electronic medical records and patient record sheets after Institutional Ethics Committee Approval (IEC 4483/2024).

2.2. Study Interventions

All procedures were performed under conscious sedation or general anesthesia. Patients were optimized prior to the procedure with intravenous fluids. For EUS‐GE, the antegrade free‐hand technique was used (Figure 1) [13]. First, a 0.025‐in. guidewire and a 7 Fr/10 Fr nasobiliary drainage tube were inserted beyond the site of GOO into the distal duodenum and proximal jejunum. The proximal small bowel loops' anatomical features were revealed, and the proper placement of the catheter was verified with the injection of contrast. The gastroscope was exchanged with a linear echoendoscope (GF‐UCT180; Olympus Medical Systems, Tokyo, Japan). The oro‐enteric catheter was used to insufflate the bowel with water, which was stained with methylene blue. Under EUS and fluoroscopy guidance, an electrocautery‐enhanced LAMS (Hot‐Axios 20X10 mm; Boston Scientific, Marlborough, United States) was used to puncture the distended proximal jejunal or distal duodenal loop. After deploying the distal flange into the target small bowel lumen (typically jejunum, but occasionally distal duodenum when anatomical considerations necessitate it) under endosonographic guidance, the proximal flange was released into the echoendoscope channel after pulling the distal flange and applying traction on the bowel toward the stomach. After the LAMS was released completely, the position was confirmed by backflow of methylene blue‐stained fluid from the proximal small bowel. In addition, contrast was injected to reconfirm the position followed by visualization of jejunal mucosa on trans‐illumination. Postprocedure patients were admitted for observation. Oral feeds were started in the evening after the procedure. Enteral stent placement was performed with either a side‐viewing duodenoscope (TJF 170, Olympus Medical systems, Tokyo, Japan) or a forward‐viewing therapeutic scope (GIF 2T 180, Olympus Medical systems, Tokyo, Japan). After assessment of the lesion and site of GOO, an endoscopic retrograde cholangiopancreatography catheter (Tandem XL, Boston Scientific, Marlborough, United States) and a conventional biliary guidewire (Visiglide, 0.025 in., Olympus Medical Systems, Tokyo, Japan) were used to negotiate the obstruction. Under fluoroscopic guidance, a water‐soluble contrast material was injected to determine the location and extent of the obstruction. Using both endoscopic and fluoroscopic guidance, the stent delivery system (Wallflex Duodenal 22 X 90 mm/22 X 120 mm; Boston‐Scientific, Marlborough, United States) was deployed across the stricture (Figure 2). The position was confirmed endoscopically and fluoroscopically. Comparative assessment of preprocedure and postprocedure Gastric Outlet Obstruction Symptom Score (GOOSS) was done on follow‐up at 2 weeks. In addition, patients were reviewed on follow‐up at 3 and 6 months as feasible. Reappearance of symptoms and need for reintervention, 6‐month stent patency, and overall survival were noted.

FIGURE 1.

Step‐by‐step approach to EUS‐GE: (A) narrowing at D1‐D2 junction; (B) 10 Fr nasobiliary tube placed as oroenteric catheter; (C) puncture taken into proximal jejunal loop distended with saline; (D) distal flange released under endosonographic vision; (E) proximal flange released using intrachannel release with jejunal loop seen across; and (F) fluoroscopic vision of deployed LAMS.

FIGURE 2.

Step‐by‐step approach to enteral stent placement: (A) narrowing at the pylorus due to carcinoma of the stomach and (B) stent placement across narrowing at the pylorus.

2.3. Study Variables

Demographic variables like age, sex, comorbid illnesses, primary malignancy, site of GOO (prepapillary—Type 1; papillary—Type 2; and postpapillary—Type 3), presence of ascites, preprocedure GOO scores, and technical details of endoscopic procedures were collected. Details of postprocedure oral intake, follow‐up at 2 weeks, 3 months, and 6 months were noted. In addition, the need for reintervention and the modality of reintervention were noted. Reintervention‐free and overall survival were also noted from the time of the index intervention.

2.4. Study Outcomes

Primary outcomes of the study were clinical success and overall survival in patients undergoing EUS‐GE and ES for palliation of mGOO. Technical success was defined as the successful creation of a gastrojejunostomy with the LAMS evidence of blue‐stained fluid flowing into the stomach or SEMS for ES covering the full length of the stenosis. Clinical success was defined as an improvement in oral enteral intake and increased tolerance of a higher consistency diet when compared to baseline by 2 points using the GOOSS. Patients who did not have a technically successful procedure were not included in the analysis of clinical success rates.

Secondary outcome measures were the comparative rate of adverse events, reintervention rates, and reintervention free survival. Adverse events were defined in accordance with the American Society of Gastrointestinal Endoscopy lexicon for reporting endoscopic adverse events [14]. Reintervention was defined as need for endoscopic or surgical intervention due to development of recurrent symptoms of GOO after initial relief after intervention.

2.5. Statistical Analysis Plan

Continuous variables were represented using mean with standard deviation or median with interquartile range, and comparison was made using independent t test or Mann–Whitney U test. Categorical variables were represented using percentage. Factors associated with clinical success and reintervention were assessed using binary logistic regression analysis using Backward likelihood ratio method. Variables included in the multivariate analysis for clinical success were age, sex, site of GOO (Types 1–3), presence of ascites, and type of intervention (EUS‐GE vs. ES). Survival analysis was done using Kaplan–Meier analysis with log rank test for comparison. Statistical analysis was done using SPSS Version 26.0 (IBM, Armonk, New York, United States).

3. Results

3.1. Patient Characteristics

Of the 180 patients who underwent intervention during the study period, 172 patients were included in the analysis. A total of 112 in the ES group (63 males, mean age: 53.62 ± 12.16 years) and 60 (30 males, mean age: 54.51 ± 12.75 years) in the EUS‐GE group were included. The most common etiology of GOO was pancreatic cancer (n = 52, 30.2%), carcinoma of the gall bladder (n = 45, 26.2%), gastric cancer (n = 40, 23.3%), and others (n = 35, 20.3%), and the stenosis was most frequently Type I (n = 91, 52.90%). Ascites was seen in 14 patients (8 in ES group and 6 in EUS‐GE group). The baseline characteristics of the study population are summarized in Table 1.

TABLE 1.

Baseline characteristics.

	EUS‐GE (n = 60)	Enteral stent (n = 112)	p value
Mean age (years)	54.52 ± 12.75	53.63 ± 12.16	0.658
Gender (M:F)	32:28	63:49	0.714
Baseline ECOG	ECOG 1—39	ECOG 1—74	0.244
	ECOG 2—18	ECOG 2—28
	ECOG 3—3	ECOG 3—10
Baseline GOOSS	Score 0—35	Score 0—43	0.012
	Score 1—25	Score 1—69
Primary malignancy	Pancreas—21	Pancreas—26	0.394
	Gall bladder—15	Gall bladder—30
	Stomach—10	Stomach—30
	Periampullary—09	Periampullary—13
	Others—10	Others—13
Site of obstruction	Type I—35	Type I—59	0.226
	Type II—20	Type II—49
	Type III – 4	Type III—4
Ascites	6 (10%)	8 (7.1%)	0.514

Note: The bold text represents significant values.

3.2. Outcomes

Technical success in patients undergoing EUS‐GE was achieved in 57/60 (95.08%) and 111/112 (99.09%) in ES patients with no significant difference between the two groups (p = 0.123). Out of the three patients in the EUS‐GE group who had technical failure, two had maldeployment into the peritoneal cavity (one due to device failure), and in both cases, the stents were retrieved. Both cases underwent surgical gastrojejunostomy. Another case had a colonic maldeployment of LAMS. In this case, the LAMS was removed at 7 days, and redo EUS‐GJ was done (not considered in the analysis). In the ES group, one patient had an impassable duodenal stenosis across which the guide‐wire could not be negotiated and was sent for surgical gastrojejunostomy.

Clinical success was achieved more often in those who underwent EUS‐GE as compared to ES (94.7% vs. 77.5%, p = 0.004, OR 5.233) (Figure 3). Type of intervention was the only predictor associated with clinical success on multivariate analysis (OR 5.50; 95% CI 1.53–19.71). No correlation was seen between the site of gastric outlet obstruction (p = 0.193), gender (p = 0.870), presence of ascites (p = 0.409), and baseline GOOSS (p = 0.465) with clinical success. These failures were managed by optimizing nutrition using dietary modification [13], use of feeding tubes [6], or additional EUS‐GE after ES or surgical gastroenterostomy [3] subsequently if no intake was feasible. Overall survival was not statistically significant between those who underwent ES and EUS‐GE (ES 112 days [95% CI 81.56–142.43], EUS‐GE 177 days [95% CI 93.55–260.44], p = 0.787) (Figure 4A). On subgroup analysis, in those with Types 1 and 2 stenosis, there was no difference in overall survival. In those with Type 3 stenosis, survival was better with EUS‐GE than ES (p = 0.030). However, overall survival was lower in those who did not achieve clinical success after intervention (clinical success—no, 49 days [95% CI 23.07–74.92]; yes, 134 days [95% CI 96.23–171.76], p = 0.013) (Figure 5).

FIGURE 3.

Sankey diagram showing the various outcomes of the study.

FIGURE 4.

Kaplan Meier Graphs of (A) overall survival and (B) reintervention‐free survival comparing EUS‐GE and ES.

FIGURE 5.

Kaplan–Meier graphs of overall survival in those that achieved clinical success.

Adverse events were comparable in both groups (5 in EUS‐GE vs. 3 in ES, p = 0.12). In the EUS‐GE group, two cases had severe adverse events in the form of maldeployment and perforation with the need for surgery. In the ES arm, one patient developed severe hemorrhage immediately postprocedure that required embolization of the gastroduodenal artery. Reintervention was needed after initial technical success in 21 patients who underwent ES and 2 patients who underwent EUS‐GE (p = 0.005; OR 4.4, 95% CI 1.1–16.7). Reintervention‐free survival was not statistically different in both groups (ES 93 days [95% CI 55.65–130.34], EUS‐GE 177 days [95% CI 96.89–257.10], p = 0.581) (Figure 4B). Table 2 summarizes the outcomes of the study.

TABLE 2.

Outcome measures.

	EUS‐GE	Enteral Stent	p value
Technical success	57/60 (95%)	111/112 (99.09%)	0.123
Primary outcomes
Clinical success (at 2 weeks)	55/57 (94.7%)	86/111 (77.5%)	0.004
Overall survival	177 days [95% CI 93.55–260.44]	112 days [95% CI 81.56–142.43]	0.787
Secondary outcomes
Adverse events	5/60	3/112	0.263
Reintervention needed	2 (3.3%)	21 (18.75%)	0.005
Reintervention free survival	177 days [95% CI 96.89–257.10]	93 days [95% CI 55.65–130.34]	0.581

Note: The bold text represents significant values.

4. Discussion

This is one of the largest comparative studies, comparing clinical outcomes and survival of EUS‐GE and ES. Our study showed that clinical success, when defined objectively by an increase in GOOSS by 2 points, was better with EUS‐GE. Although overall survival was not different between EUS‐GE and ES, those who achieved clinical success after intervention survived longer. In addition, the rate of reintervention was much lower in those who underwent EUS‐GE. There was no difference in hospital stay and adverse events related to the procedure. Among predictors of clinical success and reintervention, patients undergoing EUS‐GE were the only factor that was significantly associated with better outcomes.

Malignant gastric outlet obstruction is a common complication of various types of gastrointestinal cancers, including those of the stomach, pancreas, and duodenum [15]. Patients typically present with symptoms such as abdominal pain, postprandial epigastric fullness, and recurrent vomiting, which can lead to severe malnutrition, dehydration, and a poor quality of life [16]. Traditionally, the only therapeutic strategy for these patients was surgical gastroenterostomy, but this approach carries a relatively high risk of mortality and morbidity, as well as poor outcomes with persistent symptoms, delayed symptom relief, and prolonged hospital stays [17].

In recent decades, the availability of endoscopic interventions has provided an alternative approach for the palliation of malignant gastric outlet obstruction. Endoscopic placement of self‐expandable metallic stents has been used for the palliative treatment of patients with malignant obstruction of the gastrointestinal tract for over a decade [18]. Placement of these stents has been associated with higher clinical success rates, less morbidity, less time to start oral intake, lower rates of delayed gastric emptying, and shorter hospital stays when compared to palliative surgery [19]. EUS‐GE has emerged as a novel, minimally invasive technique for creating a gastroenterostomy bypass around the obstruction, without the need for surgical intervention [8]. One potential advantage of EUS‐GE is that it may be less prone to complications such as stent occlusion or migration, which can occur with enteral stenting. On the other hand, enteral stenting may be a more straightforward and potentially faster procedure [20].

Several studies have examined the comparative outcomes of these two endoscopic approaches for the palliation of malignant gastric outlet obstruction. Studies have reported high technical and clinical success rates for both ES and EUS‐GE in the palliation of malignant gastric outlet obstruction [21]. A retrospective analysis included 100 consecutive patients at a single center; 78 of them had enteral stents and 22 had EUS‐GE. The study also found that enteral stents had greater reintervention rates than EUS‐GE. Technical success did not differ; however, early clinical success rates were higher for EUS‐GE. Furthermore, there was a trend toward higher AEs following enteral stent implantation that was not statistically significant [22]. These results were in line with those of Chen et al. [23], who found that there was a statistically significant decrease in GOO recurrence and the requirement for reintervention. The most recent meta‐analysis evaluating the three procedures found that, in comparison to SEMS, EUS‐GE was linked to a reduced rate of tumor ingrowth, fewer severe adverse events, decreased stent blockage, a higher clinical success rate, and a lower requirement for reintervention [24]. In a recent multicenter randomized trial, Teoh et al. [25] showed that when compared to duodenal stenting, EUS‐GE can increase stent patency, decrease the frequency of reintervention, and improve patient‐reported eating habits. In our study, clinical success when defined using a 2‐point increase in GOOSS was much lower in ES. This may be due to poorer gastric emptying in these patients as compared to those who undergo EUS‐GE [26]. Clinical failure had to be managed using dietary modification, tube feeding, or additional intervention in the form of EUS‐GE or surgical gastroenterostomy. The clinical failure may impact oncologic therapies, with the possibility of delay or change in treatment plans due to suboptimal nutrition. However, considering the retrospective nature of our study, we could not study this aspect.

The main obstacle preventing EUS‐GE from being used widely at the moment is the procedure's difficulty, which calls for skilled EUS and fluoroscopic personnel with in‐depth understanding of the necessary accessories and patient anatomy. Bowel is free to move, unlike peripancreatic cysts, which are typically fixed. Compared to peripancreatic cysts, puncturing a freely flowing bowel with a LAMS delivery catheter is frequently challenging and risky. Maldeployment is the leading cause of technical failure and occurs at a significant rate, ranging from 6% to 27%, even in skilled hands [27]. Stent maldeployment was found in 9.85% of instances in a recent large multicenter retrospective analysis involving 464 patients [12]. Even though EUS‐GE has a technical and clinical success rate of over 90%, adverse events (AEs) occur in about one seventh of cases, with maldeployment being the most frequent. It is comforting to know that the combined incidence of major adverse events and mortality is still low [28, 29]. In our study, maldeployment was seen in three cases, with one undergoing repeat EUS‐GE and the other two undergoing surgical gastrojejunostomy.

Survival after enteral stent placement was shown to be between 49 and 183 days in a previous pooled analysis of 13 prospective studies [19]. In another series assessing long‐term outcomes after EUS‐guided gastroenterostomy bypass, the mean survival was 193 days [30]. However, comparative analysis of survival has not been done previously. In addition, reintervention‐free survival has also not been studied. Although our study showed no difference in overall and reintervention‐free survival in both groups, overall survival was much longer in those who achieved clinical success and had better oral intake. The significantly better overall survival observed with EUS‐GE in patients with Type 3 GOO (p = 0.030) warrants further discussion. We hypothesize that this difference is likely due to chance considering the small sample size in this subgroup. There were no significant reasons to explain this difference.

Our study has strengths in terms of large numbers with objective comparison of clinical success, providing more robust evidence to support EUS‐GE even in the short term. In addition, our study has studied reintervention‐free survival and overall survival, which has been a less addressed issue previously. The limitations include the retrospective nature of the analysis with a nonrandomized study design leading to potential bias. We have not performed propensity score matching. However, the baseline characteristics in our cohort were almost similar in both groups, thereby eliminating the need for the same. Most procedures were performed by or under the supervision of a single operator, further highlighting the issue about expertise for EUS intervention. Also, one of the driving factors for the use of ES is the lower direct cost of the stent and intervention in noninsurance‐based healthcare setups. Cost considerations in the short and long term could not be assessed in our study.

In conclusion, EUS‐GE is a safe and effective technique with better outcomes as compared to ES in terms of clinical success and reintervention rate. Although overall survival and reintervention‐free survival were not statistically different between the groups, patients achieving clinical success, regardless of the intervention type, demonstrated longer overall survival. EUS‐GE can be the primary modality for palliation of GOO in patients with malignant obstruction in expert hands, particularly given its superior clinical success and lower reintervention rates.

Conflicts of Interest

The authors declare no conflicts of interest.

Mahajan, A. , Sundaram, S. , Puri, R. , Shravan, G.  H. , Mane, K. , Kale, A. , Patkar, S. , Goel, M. , Shrikhande, S. , Chaudhari, V. , Patil, P. , and Mehta, S. (2025) Clinical Success Impacts Overall Survival in Patients Undergoing Palliation for Malignant Gastric Outlet Obstruction Using Enteral Stents or Endoscopic Ultrasound Guided Gastroenterostomy. Journal of Gastroenterology and Hepatology, 40: 2552–2559. 10.1111/jgh.70066.