Endoscopic Ultrasound-Guided Gastroenterostomy for the Palliation of Gastric Outlet Obstruction (GOO): A Systematic Review and Meta-analysis of the Different Techniques

Introduction: Gastric outlet obstruction (GOO) is usually associated with a poor prognosis and a significant decrease in a patient’s quality of life. Endoscopic ultrasound-guided gastroenterostomy (EUS-GE) using lumen-apposing metal stents (LAMS) has emerged as a safe and effective palliation procedure for GOO in patients that are unfit for surgery. Without an exclusive gold-standard technique for EUS-GE, we aimed to compare the currently available ones in this systematic review and meta-analysis, the first on this subspecialty. Methods: A comprehensive search from multiple electronic databases was performed. The search had a particular emphasis on the techniques used in performing EUS-GE. We identified all the studies in which EUS-GE was performed as palliation for GOO from its inception to the current date. The outcomes analyzed were the following: technical and clinical success, total and severe adverse events (AEs), procedure duration, and length of hospital stay (LOHS). Results: Twenty studies involving 863 patients were the basis of this statistical analysis. Patients underwent the following techniques: direct gastroenterostomy (DGE) (n=718), balloon-assisted gastroenterostomy (BAGE) (n=27), and endoscopic ultrasound (EUS)-guided double-balloon-occluded gastrojejunostomy bypass (n=118). In comparison to balloon-assisted techniques, DGE had a lower rate of AEs, -0.121 (95% CI -0.191 to -0.051 p=0.001); and LOHS for the DGE group, -2.684 (95% CI -1.031 to -4.337 p=0.001). The other analyzed outcomes presented no statistically significant differences. On a sub-analysis, BAGE showed a lower rate of AEs than EUS-guided double-balloon-occluded gastrojejunostomy bypass, -0.196 (95% CI -0.061 to -0.331 p=0.004). Conclusions: EUS-GE is a safe and effective procedure for palliating GOO. When correctly administered, any of the analyzed techniques may be used to palliate GOO with similar technical and clinical outcomes. DGE had significantly lower rates of AEs and LOHS, which can be inferred as a safer procedure. These results should be interpreted cautiously due to the limited few studies that are available and accessible. Therefore, further well-designed, randomized clinical studies on the topic are warranted to compare the different techniques from more sources.


Introduction And Background
Gastric outlet obstruction (GOO) is a potential complication caused by various malignant and benign diseases of the upper gastrointestinal tract, which results in poor emptying of stomach content. When related to malignancy, it is usually associated with a poor prognosis, a decrease in quality of life, and an increase in morbidity. The primary etiologies of malignant GOO include gastric/duodenal cancer, cholangiocarcinoma, lymphomas, and metastasis. The patients' usual signs and symptoms are nausea, vomiting, weight loss, abdominal pain, inability to eat, and ascites [1][2].
The palliation approach results are variable and depend on the patient's clinical status [3][4]. Surgical gastrojejunostomy (SGJ), performed either as an open surgery or laparoscopy, is still preferable in patients with longer life expectancies due to the low reintervention rates. However, it is associated with considerable procedural-related morbidity [3][4][5][6]. The endoscopic approach to GOO has emerged with enteral, uncovered, self-expandable metal stents (SEMS). The use of SEMS rose due to their effectiveness and safety profile, especially in those patients with a medium-to-low life expectancy (<3 months). However, tumor ingrowth and the loss of patency of the SEMS happen due to its uncovered nature. Therefore, they are usually associated with the recurrence of symptoms, especially in long-term use (>3 months) [3][4][5][7][8].
The evolution of endoscopic ultrasound (EUS) has allowed endoscopic ultrasound-guided gastroenterostomy (EUS-GE) that uses lumen-apposing metal stents (LAMS) to emerge as a potential minimally invasive approach. It employs a cautery-enhanced LAMS to bypass the obstruction by creating an anastomosis between the stomach and the jejunum, distal to the obstruction [9][10]. Based on the initial studies done [11][12], it has been proven as a long-lasting luminal patency solution, with minimal risk of tumor ingrowth, surgical risk reduction, shorter procedural time, briefer hospital stays, and fewer adverse events (AEs) [11][12].
As a newly developed procedure, some variations have been reported for EUS-GE to achieve jejunal access. The three primary techniques described include direct gastroenterostomy (DGE), balloon-assisted gastroenterostomy (BAGE), and EUS-guided double-balloon-occluded gastrojejunostomy (EPASS) [9,13]. There is still no absolute and standard technique and only one retrospective comparative study [14] that measured the differences in the approaches. To the best of our knowledge, this is the first systematic review and meta-analysis that evaluate the efficacy and safety of the different techniques for EUS-GE.

Protocol and Registration
The study protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) under CRD42021272943, approved by the Ethics Committee of Hospital das Clińicas, Faculty of Medicine at The University of Saõ Paulo. This study was performed per the recommendations from the Cochrane Handbook of Systematic Reviews of Interventions and the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines [15].

Eligibility Criteria
All study designs were eligible to be included in this systematic review. Relevant published abstracts and full-text manuscripts describing EUS-GE techniques, regardless of either year of publication or language, were included. All included studies had to provide technical and clinical success and AE rates. Only the most recent study was included when articles concerning sample duplication were identified. The lead author attempted contact to acquire additional data whenever necessary. Studies with missing data and failed contact attempts were excluded.

Literature Search
From the study's inception through October 20, 2022, searches were performed in the following databases based on a standardized protocol. The MEDLINE search strategy was "(gastroenterostomy OR gastroenterostomies OR gastrojejunostomy OR gastrojejunostomies OR Billroth) AND (endoscopy OR endoscopic OR ultrasound OR EUS OR ultrasonography)." An equivalent strategy was performed for EMBASE, Cochrane, Lilacs, and Reference Citation Analysis. Two researchers independently conducted the eligibility screening. Duplicates were excluded, and potential eligible studies were selected for further evaluation. Any disagreements were resolved by consulting a third reviewer.

Definition of Techniques (Based on the Description of the Included Studies)
Direct gastroenterostomy (DGE): An endoscopy is performed to fill the duodenum or jejunal loop distal to the obstruction with a mixture of saline, contrast media, and methylene blue. Distal loop distention can be achieved by placing a nasobiliary tube or another catheter that can traverse the obstruction site. The gastric puncture is done either with an ultrasound-guided jejunal loop puncture with a 19-gauge needle followed by an over-the-wire placement of the non-cautery-enhanced LAMS or directly with a cautery-enhanced LAMS (also described as a "freehand" technique). We considered DGE in all the studies that had no balloon assistance for the jejunal/duodenal loop puncture.
Balloon-assisted gastroenterostomy (BAGE): An endoscopy is performed to place a guidewire and position a dilating balloon through the obstruction in the duodenal or jejunal loop. The balloon is filled with contrast and methylene blue and is punctured with a 19-gauge needle to confirm the correct location. A guidewire can be advanced through the needle, and then a cautery-enhanced LAMS is deployed over the wire.
Endoscopic ultrasonography-guided balloon occluded gastroenterostomy bypass (EPASS): An endoscopy is performed to place a guidewire or a procedure where an enteroscope with an overtube is placed through the obstruction. After removing the enteroscope or gastroscope, a double-balloon catheter is inserted over the wire or through the overtube. Both balloons are inflated. The fixed segment is filled with contrast and methylene blue. The puncture is performed either directly with cautery-enhanced LAMS or with a 19-gauge needle to confirm the correct location, followed by the guidewire through the needle and the placement of cautery-enhanced LAMS.

Data Items and Outcomes Definition
The selected studies included in the review and meta-analysis had the information extracted based on characteristics of study participants (age, sex, follow-up, primary disease); intervention performed (DGE, BAGE, or EPASS), and outcomes (technical and clinical success, total adverse events (TAEs); and severe adverse events (SAEs), procedure duration, and length of hospital stay (LOHS).
Technical success was based on the previously published literature. It was defined as "the ability to perform and complete the index procedure" (puncture of the distal bowel, release of the distal flange downstream from the obstruction, and the proximal flange upstream from the obstruction). Clinical success was variable among the studies, although the authors defined as "consuming at least a complete liquid diet without vomiting." Other relevant outcomes were AEs related to the procedure graded according to the lexicon classification for endoscopic AEs set by the 2010 American Society for Gastrointestinal Endoscopy (ASGE) [16]. If the AEs were presented in another classification, they were converted into the previously stated classification by consensus between the researchers and the disputes settled by a third researcher. Procedure duration and LOHS were also extracted.
The primary analysis was to compare the DGE and the balloon-assisted techniques (BTGE). A subgroup analysis was performed to compare BAGE and EPASS techniques.

Risk of Bias and Quality of Evidence
The risk of bias was assessed by the Joanna Briggs Institute (JBI) critical appraisal tools, a device for bias evaluation in case series [17], and by Cochrane's Risk of Bias in Non-randomized Studies of Interventions (ROBINS-I) for the comparative studies [18]. The quality of the evidence was assessed using the objective criteria of the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) for each outcome using the GRADEpro, a guideline development tool software [19].

Statistical Analysis
Comprehensive meta-analysis (CMA) V3 was used to create the tabular and graphical displays and perform the statistical analysis. Absolute values, means, and standard deviations were used in the data analysis. If a study provided medians and measures of variance, mathematical formulae were used to estimate means and standard deviations, thus promoting data standardization, according to Wan et al. [20].
The risk difference (RD) was calculated using the Mantel-Haenszel formula for dichotomous variables, with a corresponding confidence interval (CI) of 95%. The difference of means (MD) for continuous variables was calculated with inverse variance and a CI of 95%. All calculated p-values were two-sided, and p-values <0.05 were considered statistically significant. Heterogeneity (inconsistency) was assessed and quantified according to the Higgins method (I2). If the heterogeneity (I2) value was greater than 50%, it was considered high, and a random-effects model was chosen to evaluate this data. A fixed-effects model was preferred for the heterogeneity values lower than 50% [21]. A funnel plot was created and visually inspected for asymmetry and quantitative accuracy for publication bias analysis using Egger's regression testing [22].

Risk of Bias and Quality of the Evidence
Joanna Briggs Institute Critical Appraisal Tools assessed the risk of bias for the case series ( Table 2) and ROBINS-I for comparative studies ( Table 3). The quality of evidence for each outcome is described in Tables 4-5.

Discussion
To our knowledge, no other systematic review with meta-analysis comparing EUS-GE techniques is on record. However, previous reviews have provided comparisons of EUS-GE with surgical gastroenterostomy (SGJ) and enteral stenting (ES).
Boghossian et al. [41] recently published a systematic review with meta-analysis comparing EUS-GE with ES and SGJ. SGJ is a well-established treatment and is still the benchmark for patients with longer life expectancies (>3 months). It demands a good clinical status but has a low reintervention rate of 10% and high technical and clinical success (100% and 90%, respectively). ES is a good option for patients with a short life expectancies (<3 months); it has a high technical success of 98%, an early diet acceptance, and requires only a short LOHS (1.4 days shorter than EUS-GE) [41]. Compared to EUS-GE, ES has a higher reintervention rate (28% vs. 6%) and severe AE rate (31% vs. 11%), with similar technical success. SGJ has higher technical success than EUS-GE (100% vs. 91%), with similar clinical success (90% vs. 86%) and AE rate (11% vs. 10%) but has an extended LOHS (5 days longer than EUS-GE). Each patient must have an individualized approach, but both these reviews demonstrate that EUS-GE has the potential to become the exclusive standard for most cases [41].
Out of all the outcomes analyzed in our systematic review and meta-analysis, the TAEs and LOHS showed a statistically significant reduction in DGE compared to BTGE. Previous studies [12,42] have presented the hypothesis that the balloon-assisted techniques were safer and had lower AEs. DGE could have a possible disadvantage considering that an unassisted procedure (without a balloon catheter) may have an increased risk of an inadvertent puncture of a distal bowel loop or colon (as their differentiation under fluoroscopy may be tricky).
That was proven to be nonidentical from our study, where the incidence of AEs was significantly higher in the BTGE group than in the DGE group, with 21.4% and 9.3%, respectively. With the usage of a catheter to trespass the obstruction and the instillation of saline fluid combined with methylene blue to dilate the small bowel loop, the "freehand" puncture (DGE) is confirmed to be a suitable method. Using a guidewire through the puncture needle can push the loop away, increasing the risk of AEs, as some of the included studies [14,30] have suggested. The rates for SAEs were 3.1% for DGE and 8.2% for BTGE without a statistically significant difference. Caution should be taken when interpreting the lack of statistical significance in severe AEs, as the rate for DGE is 62% lower. The low number of patients and studies in the BTGE group may interfere with the lack of significance of these findings.
The hospitalization period alone could not allow us to conclude that any technique is superior to the others since different hospitals and centers have different discharge protocols. With lower AE rates in DGE, their association allows us to consider that this technique is safer for these patients, directly affecting their quality of life.
Although DGE and BTGE had similar technical success, the former has the advantage since it is less laborious and requires fewer materials to perform a functional gastroenterostomy compared to the balloonassisted techniques. Using fewer materials may decrease the total procedure cost and the learning curve. However, we could not evaluate this due to the scarcity of data available in the included studies. In general, there are fewer studies with a lower number of total procedures about the balloon-assisted techniques in which they are exclusively performed. Their main advantages are the visualization of the balloon(s) and creating a "safe window" of the fixated small bowel, which may facilitate correctly puncturing the desired jejunal loop. One of the disadvantages for the BTGE group is the unavailability of essential materials, such as the double-balloon catheter used in EPASS. This catheter has been developed specifically for this procedure, but it is available mainly in Asia, limiting its adoption in Europe and America. All techniques included in this study can be performed using a guidewire through the 19-gauge puncture before the insertion of the LAMS. Nevertheless, this technique, in contrast to the "freehand" technique, has been associated with a higher rate of stent misdeployments, decreased technical success rates, and increased AEs [42][43] due to the unintended pushing of the jejunal loop before LAMS deployment.
The clinical success was subjectively assessed and mainly from previous patient report charts and records, as most of the included studies are retrospective. The oral intake that varied from liquid to a complete diet was reported as clinically successful. Therefore, a vast difference in the quality of life is portrayed within the same group. This outcome needs to be assessed with more objective parameters in future studies.  [12] in the technique could justify the better results they have achieved. The lack of statistical significance in this outcome may have been influenced by the low number of studies and patients for the BTGE techniques, as a 25% lower procedure duration could be clinically and, possibly, statistically relevant if a more significant pool of patients should be included.
To confirm their technical and clinical equivalence, we opted to compare BAGE and EPASS, the balloonassisted techniques. Our analysis showed no statistical difference between the two in all the outcomes, except the rate of AEs. Combining these two techniques into just one category to compare with the more well-known DGE approach resulted in a systematic review with a larger patient pool.
This subgroup analysis should also be interpreted with caution. BAGE had an 8.9% rate of AEs and 4.2% of SAEs, while EPASS had 28.5% and 9.3%, respectively. Although the rate of AEs had a statistically significant difference, the low number of studies and the small pool of patients could have influenced the results. A larger pool of patients could deliver better data on these analyses.
Despite this being the first systematic review and meta-analysis evaluating the outcomes of the different EUS-GE techniques, this study is not exempt from limitations. Caution should be practiced when interpreting these results, as one significant limitation is the quality of the studies included. It consists mainly of retrospective case series. All studies were eligible to be included because of the very limited comparative studies. We opted to exclude other techniques, such as rendezvous and retrograde, as they only appeared in small case series. We also excluded studies that did not explicitly describe the technique used or did not separate the results from one another.
The EUS-GE being a novel procedure without a gold standard explains the various techniques developed and the studies with a few included patients. In addition, the procedures are concentrated in large referral centers, which interferes with the generalizability of these findings, compelling the authors to produce multi-technique and multi-center studies to attain a considerable sample size. The report of AEs in the included studies from this review had different classifications; thus, the conversion to the ASGE lexicon classification for endoscopic AEs may represent a source of bias in our study. Despite these limitations, our review, the only meta-analysis on this theme, shows strengths in demonstrating a summary of the efficacy and safety of the main EUS-GE techniques.
The different novel EUS-GE techniques demand a steep and laborious learning curve. The novelty may still interfere with the results in comparison to other modalities of therapeutic procedures for GOO, such as SGJ.
Mastering any technique may improve results, with lower AEs and faster procedures [13,25,42]. All the methods evaluated in this study are comparable and can be performed without additional harm to the patient. Further well-designed randomized clinical studies are warranted to compare the different techniques.

Conclusions
In summary, this systematic review and meta-analysis demonstrated that EUS-GE is a safe and effective treatment for the palliation of GOO. With the correct execution, any of the analyzed techniques may be used to palliate GOO with similar technical and clinical outcomes. Although DGE presented a statistically significant lower rate of AEs and LOHS, which can be inferred as a safer procedure, the best approach should be individualized, considering personal and local expertise and availability of material and devices.
relevant to this study. Diogo Turiani Hourneaux De Moura declare(s) personal fees from BariaTek. Advisory Board Member (Consulting fees). This was not relevant to this study. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.