Gastric outlet obstruction (GOO) is a clinical syndrome secondary to luminal obstruction at the level of the distal stomach, pylorus, or duodenum. It can be caused by benign or malignant etiologies. Typical symptoms include postprandial vomiting, early satiety, abdominal pain, bloating, inability to tolerate oral intake, and weight loss, impacting patients’ quality of life.^1,2

Although benign conditions such as peptic strictures can cause GOO, most cases are caused by malignant conditions comprising the pyloric-duodenal region, such as gastric or pancreato-biliary neoplasms. The classical treatment for such patients has been surgical gastrojejunostomy (SGJ) by the open or laparoscopic approach. However, the drawbacks of SGJ include its high morbidity rate and delays in resuming oral intake and chemotherapy.³ A noteworthy point is that many patients are poor surgical candidates due to advanced underlying malignant disease, poor nutritional status, and a short life expectancy.^4–6

Endoscopic treatment with the placement of self-expandable metallic stents (SEMS) has emerged as a minimally invasive alternative to surgery,⁷ with reported technical and clinical success of 90%–100% and 70%–100%, respectively.^8,9 A meta-analysis including three randomized controlled trials comparing SGJ to endoscopic treatment with SEMS showed that SEMS had more favorable short-term outcomes, including a shorter time to restart oral feeding and shorter hospital stay.¹⁰ However, many patients may experience symptom recurrence due to stent obstruction from food contents or tissue ingrowth; the median patency is 70 days, and repeated interventions are required to maintain stent patency. Therefore, it has been advocated that SGJ may be a more suitable approach for patients with a better prognosis, while SEMS should be reserved for patients with poor clinical status and short-term life expectancy.¹¹

More recently, with the advent of different types of bi-flanged lumen-apposing metal stents (LAMS), endoscopic ultrasound-guided gastroenterostomy (EUS-GE) has been developed for the management of GOO, and various techniques have been developed to further improve its safety and efficacy.^12,13 In this review, we describe the technical aspects, current status, adverse events, and future perspectives of EUS-GE.

In 2002, Fritscher-Ravens et al¹⁴ developed a handmade suture mechanism (plication by tag) guided by echoendoscopy that made it possible to fix two organs together to assist in stent placement or less invasive methods to form anastomoses. However, the technique was not adopted due to the complexity of the procedure. After the development of a fully covered metal stent with bilateral lumen-apposing anchors (LAMS), Binmoeller and Shah¹⁵ introduced EUS-GE for the first time using LAMS in a porcine model study in 2012, thus enabling the expansion of the technique.

The basic principle of EUS-GE is to create an anastomosis between the stomach and enteric wall, similar to the surgical approach. EUS-GE is expected to lead to longer stent patency than conventional SEMS, owing to its shorter length and fully covered design.²

Symptomatic patients with GOO are candidates for EUS-GE regardless of the etiology of the stenosis. Benign conditions include peptic strictures due to gastric or duodenal ulcers, chronic pancreatitis, surgical anastomotic stricture, superior mesenteric syndrome, and other conditions. Malignant conditions include primary or metastatic tumors invading the pyloric-duodenal segment, such as gastric cancer, pancreatic cancer, biliary cancer, and lymph node metastasis.

EUS-GE procedures can be performed by puncturing the third or fourth portion of the duodenum (EUS-gastroduodenostomy) or by puncturing the jejunum (EUS-gastrojejunostomy). Preprocedural planning with computed tomography scan in both transverse and coronal views is helpful in deciding on the puncture site (Fig. 1). It is important to locate a bowel loop adjacent to the stomach to allow a safe puncture site for stent deployment. The presence of a large volume of ascites is an absolute contraindication to EUS-GE since it interferes with the adherence of the bowel loop to the stomach.¹⁶

Figure 1. Computed tomography-scan showing the proximity of the stomach (yellow arrow) to the distal duodenum (blue arrow), showing a good place to perform a safe puncture.

Patients with GOO symptoms often have significant gastric stasis, which may lead to serious adverse events during sedation. Therefore, it is important to maintain these patients on a low-residue liquid diet or no food intake for several days prior to the procedure to guarantee gastric emptying in association with enteral nutrition through a tube placed distally to the site of obstruction. The use of double-lumen tubes that make possible both gastric drainage and enteral nutrition is another common strategy.

The currently available standard LAMS include the AXIOS^TM stent (Boston Scientific Co., Marlborough, MA, USA), the HANARO^TM stent (M.I. Tech, Pyeongtaek, Korea), and the SPAXUS^TM and NAGI^TM stents (Taewoong Medical Co., Gimpo, Korea). Initially, stent implantation required placement of a guidewire and dilation of the needle tract using an electrocautery dilation catheter or a balloon. More recently, a LAMS attached at the tip to an electrocautery device (EC-LAMS) was developed. This system allows puncture, tract dilation, and stent placement in a single device, in a simpler, faster, and safer way. Even the guidewire is not needed as it can sometimes push the small intestine adjacent to the stomach away, resulting in poor positioning. Currently available EC-LAMS are HOT AXIOS^TM (Boston Scientific Co.) and HOT Spaxus^TM (Taewoong Medical Co.). For EUS-GE, both 15- and 20-mm LAMS are used; however, the 20-mm stent should preferably be chosen. A retrospective study showed a higher clinical success rate (100% vs 88%) and shorter hospital stay (4 days vs 5 days) by using the largest dimension stent.¹⁷ Table 1 summarizes the currently available LAMS.

Table 1 . Types of Lumen-Apposing Metal Stents.

Stent	Flange diameter (mm)	Length (mm)	Lumen diameter (mm)	Electrocautery enhanced option
AXIOSTM stent (Boston Scientific Co., Marlborough, MA, USA)	21, 24, 29	10, 15	10, 15, 20	Yes
HANAROTM stent (M.I. Tech, Pyeongtaek, Korea)	22–28	10, 30	10, 12, 14, 16	No
SPAXUSTM stent (Taewoong Medical Co., Gimpo, Korea)	23, 25, 31	20	8, 10, 16	Yes
NAGITM stent (Taewoong Medical Co., Gimpo, Korea)	20	10, 20, 30	10, 12, 14, 16	No

Different techniques have been described to perform EUS-GE. Despite the differences, the objective of all of these techniques is the release, through an echoendoscope, of a fully covered stent with wide flanges, which promotes the anchoring between the stomach and the intestine (or duodenum), avoiding risks of migration, leakage and tissue ingrowth. Fig. 2 summarizes the different EUS-GE techniques.

Figure 2. Schematic figures of endoscopic ultrasound (EUS)-guided gastroenteroanastomosis techniques. (A) Direct puncture by EUS and lumen filled with the aid of a stone extraction balloon catheter; (B) balloon-assisted; (C) ultra-slim endoscope assisted; (D) double balloon-occluded assisted.

Direct EUS-guided gastroenteroanastomosis

Step 1: Initially, the duodenum/gut distal to the point of obstruction is filled with 250 to 500 mL of saline solution combined with contrast and methylene blue (water should be avoided due to the risk of hyponatremia). This can be done by using an EUS 22-gauge needle as a test puncture or by using endoscopy accessories to fill up the intestinal lumen ahead of the obstruction (Fig. 2A). Intravenous administration of antimotility agents, such as butyl-scopolamine or glucagon, can be considered to decrease intestinal contractions.

Step 2: After endoscopic and fluoroscopic confirmation of the fluid-filled distended bowel, a single-step puncture is performed if an EC-LAMS stent is available (the “free-hand” technique). The LAMS is no longer inserted over the wire, because it can push the jejunum away. However, if an electrocautery stent is not available, puncture with a 19-gauge needle should be performed, followed by an enterogram with injection of contrast medium through the needle (or aspiration of the blue liquid from the needle), confirming that the target intestine can be correctly punctured. Then, the passage of 0.025- or 0.035-inch stiff guidewire, followed by tract dilatation (electrocautery dilator or a through-the-scope balloon) and deployment of the LAMS over the guidewire.^2,18,19

With the advent of EC-LAMS, this technique became faster and simpler, given the lack of extra accessories and the reduction of the number of steps; therefore, this technique is preferred by many authors. In a study comparing the direct technique versus the balloon-assisted technique, no differences were found regarding technical success, clinical success, adverse events, need for repeated interventions, or survival. However, the mean procedure time was significantly shorter in the direct technique than in the balloon-assisted group (35.7 min vs 89.9 min).²⁰ The main disadvantage of this technique is that the initially administered fluid may have migrated downstream away from the desired loop at the time EUS is performed.

Device-assisted EUS-guided gastroenteroanastomosis

Device-assisted EUS-GE uses a device, such as a balloon catheter, a thin nasobiliary tube, or an ultra-slim endoscope, to help to identify the target bowel loop.^2,19

Balloon-assisted technique

Step 1: A 0.025- or 0.035-inch guidewire is first advanced across the obstruction under endoscopic or fluoroscopic guidance.

Step 2: The endoscope is withdrawn and a dilation balloon or a stone extraction balloon catheter is advanced over the wire and then inflated with saline or dye.

Step 3: Stent placement can then proceed as described above, using the balloon as a target for EUS-guided needle puncture (Fig. 2B). In addition, the balloon helps to stabilize the bowel loop distally to the obstruction.

Double balloon-occluded-assisted technique

Fig. 3 summarizes the main steps of the double balloon-occluded (EPASS) technique.

Figure 3. Double balloon-occluded procedure. (A) A double-balloon-occlusion catheter is introduced inside the overtube. (B) The two balloons are inflated with 200 mL of dye to fix the bowel loop. The segment between the balloons is filled with a solution of saline and methylene blue. (C) Deployment of the distal flange of the lumen-apposing metal stent (LAMS). (D) Full deployment of the LAMS. (E) The distended bowel loop is located by echoendoscopy and then punctured with the electrocautery LAMS. (F) Endoscopic ultrasound-guided deployment of the stent. (G) Endoscopic view of the fully deployed LAMS. The double-balloon catheter is kept in place until confirmation of technical success.

Step 1: An endoscope (or enteroscope) attached to an overtube is introduced to the point of obstruction, followed by passage of the guidewire.

Step 2: The endoscope is removed, keeping the overtube next to the point of obstruction, and the double balloon-occlusion catheter is introduced over the guidewire inside the overtube. The two balloons (20 cm apart) are inflated with 200 mL of dye to fix a segment of duodenum (or jejunum), which is filled with a solution of saline and methylene blue.

Step 3: Puncture is performed between the two balloons following the technique already described.

Filling the bowel loop is essential for proper puncture, and even if a large amount of water is injected, the position of the small bowel can move and collapse due to the flow of water by peristalsis, even with the use of antimotility drugs. Thus, in theory, the use of a balloon, and especially a double-balloon enteric tube, could maintain bowel distention due to additional fluid injection, leading to greater technical safety of the procedure. However, there are no evidence-based differences in the rates of adverse events in the literature.

Nasobiliary drain-assisted technique

Step 1: A 0.025- or 0.035-inch guidewire is first advanced across the obstruction under endoscopic and/or fluoroscopic guidance.

Step 2: The endoscope is withdrawn and a nasobiliary drain is advanced only under fluoroscopic guidance. Alternatively, placement of the nasobiliary drain can be performed with the assistance of a therapeutic endoscope or duodenoscope, which allows insertion of the nasobiliary drain through the scope channel.

Step 3: After the infusion of liquid, the puncture is performed as explained above.

Ultra-slim endoscope device-assisted technique

Step 1: An ultra-slim endoscope is introduced at the point of obstruction and, when possible, passed through it. The fluid is administered directly through the endoscope to distend the target intestinal loop.

Step 2: The echoendoscope is then inserted and advanced into the stomach alongside the ultra-slim endoscope, and the distended loop puncture is performed.

Although this technique allows direct visualization of the puncture through the endoscope, it is not widely used due to the need for two separate endoscopic processes, in addition to the challenge of maneuvering a therapeutic echoendoscope alongside the ultra-slim endoscope (Fig. 2C).

The main limitation of these assisted techniques relates to the difficulties in orally advancing the devices through an often tight and angled duodenal/jejunal stenosis. In cases of complete GOO, pre-dilation using a balloon dilation catheter before inserting over-the-wire devices may be helpful.

Other techniques

Rendezvous technique

In this technique, after the ultrasound-guided puncture with the passage of the guidewire into the duodenum or small bowel, the guidewire is retrieved from the mouth by using a dilating balloon or a stone extraction balloon with a polypectomy snare or basket catheter. This maneuver, although technically difficult, allows traction of the guidewire, offering greater security in the passage of the stent.¹⁹

Retrograde deployment technique

After using the rendezvous technique for traction of the guidewire through the mouth, the echoendoscope is removed, a therapeutic endoscope is advanced over the guidewire, traversing the obstruction to the punctured site, and then the LAMS is deployed retrogradely from the duodenum (or jejunum) to the stomach. Although this technique has a lower risk of misdeployment, because the stomach is less likely to be pushed away, it may not be possible to pass the therapeutic endoscope through the obstruction, posing a risk of perforation.^2,19

In a systematic review and meta-analysis published in 2020 with 12 studies (various techniques) including 285 patients, technical success was achieved in 92% and clinical success in 90% of patients. Recurrence of symptoms or unplanned reintervention was needed in 9% of the patients, and adverse events, which include misdeployment of the stent, bleeding, pneumoperitoneum, peritonitis, abdominal pain, and gastrocolic fistula, were reported in 12% of the patients.²¹

Even for benign GOO, EUS-GE is a promising technique. In international retrospective series involving 5 tertiary centers and 26 patients, technical success was achieved in 96.2 % and clinical success in 84.0 %. The rate of unplanned reinterventions was 4.8 %, and adverse events were observed in 11.5% of cases (2 misdeployed stents and 1 gastric leak following elective stent removal).²²

Some authors place the direct EUS as the technique of choice due to greater technical simplicity, without the need for any device to assist in the procedure. In a study that compared the direct technique versus the balloon-assisted technique, the rates of technical success (94.2% vs 90.9%; P = 0.63), clinical success (92.3% vs 90.9%; P = 1.00), adverse events, reinterventions and survival, as well as the length of hospital stay, were similar, but with a shorter procedure time (35.7 ± 32.1 min vs 89.9 ± 33.3 min; P < 0.001) in the direct technique.²⁰

EUS-GE versus duodenal stent

The endoscopic placement of SEMS has been proven as a safe and effective option to relieve the symptoms of patients with malignant GOO.²³ This endoscopic procedure involves the deployment of an enteral stent crossing the malignant obstruction, typically located at the duodenum and/or at the proximal jejunum; therefore, it has been associated with significant rates of reintervention due to food impaction, stent migration, tumor ingrowth, and biliary obstruction.²³ What differentiates EUS-GE from the use of a duodenal stent (DS) is that its aim is to create a bypass from the stomach to the small intestine distally to the obstruction, ultimately decreasing the risk of tumor ingrowth and biliary obstruction.

A retrospective study published in 2017 by Chen et al²⁴ compared EUS-GE with DS in patients with malignant GOO, and showed similar technical and clinical success rates: 86.7% for EUS-GE vs 94.2% for DS (P = 0.2), and 83.3% for EUS-GE vs 67.3% for DS (P = 0.12), respectively. However, symptom recurrence and need for reintervention were significantly lower among patients in the EUS-GE group (4.0% vs 28.6%; P = 0.015). Ge et al,²⁵ in a retrospective study, showed not only a higher rate of reintervention with DS than with EUS-GE (32.0% vs 8.3%; P = 0.021), but also a higher clinical success rate in the EUS-GE group (95.8% vs 76.3%; P = 0.042) in patients with malignant GOO.²⁵ Two recent systematic reviews and meta-analyses of retrospective studies confirmed the lower rate of reintervention and the higher rate of clinical success with EUS-GE than with DS.^26,27

EUS-GE versus surgical gastrojejunostomy

Traditionally, SGJ has been the gold standard for patients with malignant GOO, while endoscopic therapies, like DS, were indicated for patients who were not good candidates for surgery.

A multicenter retrospective study of patients with benign and malignant GOO showed that EUS-GE, compared with laparoscopic gastrojejunostomy, had similar technical (88% vs 100%; P = 0.11) and clinical success (84% vs 90%; P = 0.11), but with fewer adverse events (12% vs 41%; P = 0.03).²⁸ The systematic review and meta-analysis by Boghossian et al²⁷ of four retrospective studies with patients with malignant GOO reported shorter hospital stays with EUS-GE than with SGJ, although a lower technical success was reported with EUS-GE (91.4% vs 100%; P < 0.01).

It is important to re-emphasize that EUS-GE is still a new and complex endoscopic procedure that requires advanced EUS skills. The learning curve for EUS-GE is still unknown, and techniques remain unstandardized in the literature, which may contribute to a lower technical success rate when compared to SGJ.

Collective experience shows excellent efficacy with low adverse events, even lower than with other techniques for treating GOO, with systematic reviews demonstrating complication rates from 11% to 12%.^20,29 The adverse events observed were peritonitis, bleeding, abdominal pain, gastric leakage, gastrocolonic fistula, and food impaction.

The most common technical complication is misplacement, which is primarily responsible for cases of technical failure and complications. Several reports in the literature described maneuvers to rescue misplaced stents. One option is to just remove the LAMS and close the gastric defects with clips, with no salvage attempt. However, many reports of successful salvage have been described in the literature. Repositioning the misplaced LAMS with grasping forceps is the simplest option and is usually the first technique to be attempted.

Some authors have described the successful salvage of misplaced LAMS by entering the peritoneal cavity through the gastric orifice and either rescuing the bowel segment or placing a guidewire with subsequent placement of a second LAMS or a fully covered metal stent.^30,31 The use of an echoendoscope in natural orifice transluminal endoscopic surgery (NOTES) to rescue the anastomosis has already been successfully described.³² In our previous experience with a misplaced LAMS during an EPASS procedure, we could not locate the punctured bowel loop by NOTES. Therefore, we kept the double-balloon catheter inflated, and a second EUS-guided LAMS puncture was successfully performed. The previous orifices were then closed with endoscopic clips.³³

In a multicenter study including 26 patients, stent misplacement occurred in seven cases (27%). In three cases of proximal flange misplacement, the tract was successfully bridged with a telescoping fully covered SEMS, and out of four cases of distal flange misplacement, two were solved by NOTES with the placement of a bridging LAMS, while two had the LAMS pulled back into the stomach and then one was closed with an over-the-scope clip and the other had an enteral SEMS deployed without any attempt to close the gastric puncture.^29,34 More recently, a multicenter retrospective study including 467 patients described stent misdeployment in 10% of them. In 70% of the cases, this adverse event occurred in the first 13 procedures. In most cases (60%) of stent misdeployment, the distal flange of the SEMS was deployed into the peritoneum, with no damage to the small bowel. The authors classified this situation as type I stent misdeployment. The treatment consisted of stent removal and endoscopic closure of the gastric defect, followed by another attempt of EUS-GE.³⁵

From our experience, tips to minimize adverse events related to LAMS misplacement are as follow:³³

1. The procedure should always be performed under general anesthesia.

2. An attempt should be made to obtain maximal distension of the bowel loop with fluid and the use of antispasmodics to create an easy and still target for puncture.

3. The puncture along with the use of cautery, should be done with a steady speed, neither too slow to push the bowel loop away nor too fast to transfix it.

4. Procedural stability is best achieved with the use of an EPASS catheter.

5. The double-balloon catheter should be kept inflated throughout the procedure until confirmation of correct LAMS placement to make sure the jejunal segment is kept in place.

6. In case of LAMS misplacement, the double-balloon catheter should be kept inflated, followed by endoscopic removal of the misplaced LAMS. A second EUS attempt at LAMS placement may be considered.

7. From our experience, the eventual bowel defects caused by LAMS puncture were only small orifices, easily treated by endoscopic clipping.

The available evidence on EUS-GE strongly suggests that this intervention is superior to duodenal stenting and similar to laparoscopic GE for the relief of GOO. However, correctly conducted randomized clinical trials are needed to clearly define the role of EUS-GE in the treatment of GOO. Another Achilles heel of EUS-GE is the lack of standardization of the technique, which should also be addressed by clinical trials. Finally, considering that EUS-GE may be offered to patients with longer life expectancy (e.g., GOO caused by a benign condition), it will be necessary to establish whether the stent should be removed and the anastomosis dilated, as opposed to periodic stent exchange.

EUS-GE is emerging as a promising technique with good short-term and long-term results, and with a low incidence of adverse events. While the optimal technique of the procedure has not been standardized, it should be reserved for referral centers and performed by endoscopists experienced in advanced therapeutic procedures. Larger prospective randomized trials comparing EUS-GE to surgery and endoscopic stents are necessary to better understand its role in the treatment of GOO. Improvements in LAMS design are desired, especially concerning luminal stent diameter, in order to maximize anastomosis performance and optimize clinical success.

None.

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