Self-expandable metallic stents (SEMS) were first used in the early 1990s. Since then, use of these prostheses had increased exponentially and is now considered standard in the endoscopic management of malignant gastrointestinal (GI) luminal strictures. The obvious advantage of these stents over conventional non-expandable prostheses is their ability to be delivered while “compressed” through the narrow working channel of an endoscope (3.7 mm for a standard therapeutic gastroscope and 4.2 mm for a standard duodenoscope). Once they exit the scope and reach the target area (stricture), they can expand to a predetermined diameter, which is significantly larger than that of plastic stents. SEMS of various lengths and diameters can be used to restore patency of the oesophageal, gastric, duodenal, colonic and biliary lumens and provide palliation of obstructive symptoms or serve as “bridge” to more definitive surgical treatment with the possible combination of neo-adjuvant chemotherapy.¹

Endoscopic application of SEMS for the therapy of benign lesions has emerged as a treatment option in recent years. A major obstacle in the use of uncovered SEMS (UC-SEMS) in non-malignant obstructive lesions had been the need to eventually remove the stent after a period of time. The development of fully-covered SEMS (FC-SEMS) which are removable has largely overcome this problem allowing for a significant expansion in the use of these stents in situations where removal of the prosthesis in the future may be contemplated.^2,3 Other emerging indications for the use of FC-SEMS are (but not limited to) post-sphincterotomy bleeding, esophageal variceal bleeding, sealing of a perforation or bile leak and lately treatment of symptomatic biliary and pancreatic duct strictures.^4,5 Temporary placement of FC-SEMS can also be used to treat benign non-obstructive conditions in the GI tract, such as oesophageal and colonic fistulas.⁶

In recent years, endoscopic ultrasound (EUS) has emerged as a mainstream therapeutic modality for drainage and fine-needle injection.⁷ EUS-guided drainage has been successfully used in the treatment of peripancreatic fluid collections (PFCs).⁸

EUS-guided biliary drainage (EUS-BD) has also emerged as a second line modality in cases where retrograde drainage (via endoscopic retrograde cholangiopancreatography [ERCP]) or antegrade drainage (percutaneous transhepatic biliary drainage [PTBD]) fails or is not an option (unfavourable anatomy, duodenal obstruction, etc.).⁹ The two primary approaches to EUS-BD are EUS-guided hepaticogastrostomy (EUS-HGS) and EUS-guided choledochoduodenostomy (EUS-CDS).¹⁰ Finally, EUS-guided transgastric or transduodenal drainage of the gallbladder, first described in 2007, is now emerging as a potential alternative to ultrasound (US)-guided cholecystostomy in patients with acute cholecystitis unsuitable for emergency surgery and failing conservative management.^11,12

EUS-guided transmural drainage offers the advantage of minimally invasive endotherapy and is ideally suited to treat collections located adjacent to the GI tract. The simultaneous advances made recently in the field of specific accessories and metallic endoprostheses with enhanced safety profiles are likely to lead to widespread acceptance of these novel techniques in the foreseeable future. Here we will attempt to review the literature on the current state of these applications and evaluate future prospects.

The understanding of the natural history, pathophysiology and precise definition of PFCs has improved significantly in recent years largely due to the use of high definition cross-sectional imaging and EUS. The Atlanta classification (modified version, 2012) is widely used as a “common language” to define different types of PFCs based on the natural history of the underlying process and the characteristics of the cyst itself (Table 1).¹³

PFCs are a frequent complication of acute pancreatitis and can be seen in the acute phase in over 40% of patients; however, the significant majority of those collections will be absorbed spontaneously and only about 7% of patients will develop a pseudocyst later in the course of the disease.¹⁴

PFCs are commonly the result of containment of necrotic pancreatic tissue or disruption of the pancreatic ductal system.¹⁵ Other rare causes of PFC formation are abdominal trauma and surgery.^16,17 Risk factors associated with pseudocyst formation are younger age and alcohol-related aetiology. Since the natural history of simple chronic PFCs is that of spontaneous resolution in over 80% of cases, endoscopic drainage is reserved for collections complicated with pain (usually associated with cyst expansion), infection (Fig. 1), gastric outlet obstruction (Fig. 2), biliary obstruction (Fig. 3), and fluid leakage.^18–21

Historically, PFCs needing drainage have been managed either with surgical or percutaneous intervention. Surgical cystogastrostomy can be either open or laparoscopic and involves the anastomosis of the cyst with the gastric lumen or in some instances with the small bowel (cystojejunostomy). It is highly efficacious (recurrence of a pseudocyst in less than 5%), but complication rate is in the order of 30%.²² Open surgery for drainage of walled-off pancreatic necrosis (WOPN) is associated with very high morbidity and mortality of 6% to 28%.²³

Radiologically guided percutaneous drainage of PFCs has been shown to be an effective and safe treatment.^24,25 Inevitably, the percutaneous approach (computed tomography or US-guided) was compared with the “gold standard” approach of surgical cystogastrostomy and early studies showed a marginal survival benefit with the percutaneous approach.²⁶ An indwelling catheter is required for the percutaneous technique; this may become a focus for infection and potentially end up in the formation of a pancreatico-cutaneous fistula, which are the 2 main long-term complications of the technique.^27–29

Recently, most of the tertiary care centres have adopted the endoscopic approach for PFC drainage.^20,30 EUS-guided transmural drainage has become standard treatment and compares favourably to surgical cystogastrostomy in terms of efficacy, morbidity/ mortality, the length of hospitalization and cost-effectiveness.^31,32 It is therefore nowadays the first-line treatment for drainage of mature pancreatic pseudocysts in close proximity to the gastric or duodenal wall. Clinical success rates of 70% to 87% have been reported in multiple studies, with complication rates of 11% to 34%.^30,33–35

Drainage of more complex collections (acute necrotic collections [ANC], WOPN) can also be achieved surgically, via the percutaneous route, as well as endoscopically. There is controversy in the approach to these complex types of PFCs containing variable amounts of necrotic debris and this is reflected in the most recent consensus guidance from the International Association of Pancreatology (IAP) and the American Pancreatic Association (APA).³⁶ Indications to intervene are the presence of infected collection with clinical deterioration (i.e., septic shock) and/or on-going multiple organ failures for several weeks after the onset of acute pancreatitis. In the absence of infection, indications for intervention (surgical, percutaneous, or endoscopic) are persistent gastric outlet, intestinal, or biliary obstruction due to mass effect of the necrotic collection, on-going abdominal pain, debilitating constitutional symptoms and disconnected duct syndrome (i.e., separation of the upstream and downstream pancreatic ductal systems in the context of severe necrotizing pancreatitis with persistent symptoms [mainly pain]). In terms of the appropriate timing of the intervention, the general consensus is that if the patient’s clinical condition allows for it, drainage should be postponed for at least 3 to 4 weeks after the onset of pancreatitis.^36,37

The traditional approach to infected WOPN is an open surgical debridement, together with percutaneous drainage and continuous lavage. However, open surgical necrosectomy is associated with a very high risk of postoperative complications (50%–75%) and mortality (25%–56%).³⁸ In contrast, endoscopic transmural drainage of WOPN is associated with good initial and long-term clinical success and acceptable morbidity (30%) and mortality rates (5%). The advantages of endoscopic management are related to its minimal invasiveness.³⁹ Recent studies have shown that while endoscopic EUS-guided drainage of pseudocysts is highly successful and safe, drainage of necrotic collections is technically more challenging and carries higher risks.^35,40 This suggests that while EUS-guided puncture of a complex PFC-containing solid necrotic material is still technically feasible, some concerns arise as to the effectiveness of draining of the cyst contents through the cyst gastrostomy or cyst- duodenostomy fistulous tract.⁴¹ Some of the inherent risks of the endoscopic approach include the risk of air embolism, endoscopically uncontrollable bleeding and inadequate drainage through multiple plastic stents, as well as early occlusion of the fistulous tract.⁴² The more general application of CO₂ endoscopic insufflations is, therefore, necessary to reduce the risk of fatal air embolism during these procedures. In order to ensure adequate drainage, the technique of direct endoscopic necrosectomy (DEN) has been developed. DEN was initially described by Baron et al¹⁵ in 1996 and involved the placement of a nasocystic drain without mechanical removal of solid debris. The success rate was 81% with an overall complication rate of 36% were reported in this study. Similar success and complications were reported in a retrospective analysis of DEN in 53 patients.⁴³

The concept of dilation of the fistulous tract to allow entry of the endoscope inside the necrotic cavity to facilitate mechanical debridement was introduced a decade ago by Seewald et al.⁴⁴ In this small retrospective study of 13 cases, only 3 patients required rescue surgery at a later stage. Similar success rates were reported subsequently in other small retrospective studies.^45–47 Subsequently, two large multicentre trials from Germany and USA confirmed a clinical success rate of over 80% and an acceptable rate of complication of 14% to 20% for endoscopic necrosectomy.^48,49

Biliary FC-SEMS provide a larger diameter fistulous tract for drainage of WOPN but will not permit the advancement of an endoscope into the collection. Our group published a series of 4 patients treated with this novel approach combined with irrigation with a high-flow water jet system.⁵⁰ Briefly, the infected pancreatic necrosis or abscess was punctured with a cystostome and after balloon dilatation of the tract, a biliary FC-SEMS was placed into the cavity. Following stent deployment, a nasobiliary pigtail catheter was placed into the cavity to ensure continuous irrigation. The stent was removed endoscopically after 5 to 7 days and DEN as well as debridement were performed without pneumatic dilation of the tract with the use of a high-flow water-jet system with a flush knife, a Dormia basket, and hot biopsy forceps (Fig. 1, 4). The transluminal endotherapy was repeated 2 to 5 times daily during the next 10 days. Supportive care included parenteral antibiotics and jejunal feeding. All patients improved dramatically with complete resolution of sepsis. No procedure-related complications were observed. Similar success was reported in other small series.^51,52

As mentioned earlier, the main advantage of FC-SEMS over the traditional double pigtail stents is the size of the lumen allowing better drainage of the cyst contents. This is particularly important in ANC and WOPN, where often continuous irrigation of the cyst cavity and/or repeat attempts of DEN may be needed. In theory, better drainage may reduce the number of additional procedures a patient may need. A larger lumen single SEMS also reduces the length of the procedure (deployment of only one stent, no need for pneumatic dilation of the fistulous tract). A couple of small-size studies of PFCs drained using FC-SEMS (originally designed for biliary deployment) reported excellent technical success rate and over 78% complete resolution of the cyst with acceptable complication profile.^53,54 One particular complication appeared to be spontaneous stent migration, which, if proximal, may be problematic and necessitate attempts at endoscopic retrieval.

In order to overcome this potential complication, different types of FC-SEMS with enhanced anti-migratory properties were developed in recent years. Examples of such stents include those with anti-migratory fins, as well as large-bore wide flare SEMS (which allow performance of DEN through the stent lumen).^55,56

Novel dedicated transmural stents have been recently developed, such as the NAGI stent (Taewoong Medical, Gimpo, Korea) and the AXIOS stent (Boston Scientific, Marlborough, MA, USA). They both have bilateral flanges at the end, which are designed to facilitate stent anchoring, thereby preventing migration and leakage. Furthermore, both special stents are FC to prevent tissue in-growth. The active length between the flanges ranges of 10 to 20 mm and the diameter of these stents are 10 to 16 mm, which enables direct endoscopic passage and repeated necrosectomy through the lumen of the stent.^57,58 Several other series confirmed the high success rate in complete cyst resolution and a very low rate of serious complications with the use of these lumen-apposing metallic stents (LAMS).^59–62

These and other studies collectively demonstrate that EUS-guided transmural drainage of PFCs (especially those containing solid necrotic material and often needing several necrosectomy sessions) is feasible and safe. However, even in the most experienced hands, serious and potentially life-threatening complications may occur and a multidisciplinary approach (i.e., endoscopists, interventional radiologists and hepatopancreatobiliary surgeons) is essential for a positive outcome. The use of easily deployable SEMS with a lumen sufficiently large to allow DEN through it may well be an important breakthrough, making the index procedure as well as subsequent endoscopic necrosectomies easier, quicker, and potentially safer.

Acute cholecystitis and gallbladder empyema are common surgical emergencies. Laparoscopic cholecystectomy is the standard treatment in patients deemed sufficiently fit for such a procedure. Significant proportion of patients with acute cholecystitis, however, are elderly with co-morbidities and hence high operative risk. In this cohort, minimally invasive methods such as percutaneous transhepatic gallbladder drainage have been established as viable alternatives with very high efficacy and acceptable risk profile.⁶³

EUS-guided transmural gallbladder drainage (EUS-GBD) can be performed as an effective alternative therapy. Recent studies proved that EUS-guided gallbladder transmural drainage is as effective as percutaneous drainage of the gallbladder for patients with acute cholecystitis who are unfit for emergency surgery.^11,64,65 However, potential complications of EUS-GBD are air or bile leakage into the peritoneal cavity and related peritonitis, since plastic stents cannot isolate the fistulous tract and prevent leakage alongside the prosthesis.

To overcome these complications, recently developed LAMS, can also be applied for transmural GBD. de la Serna-Higuera et al⁶⁶ reported a series of 13 patients with acute cholecystitis treated with EUS-guided cholecystogastrostomy or cholesytoduodenostomy using the Hot Axios LAMS system and reported technical success in 11 out of those patients (85%) without any significant complications. Hot AXIOS combines an electrocautery-enabled access introducer device with the lumen apposing fully covered AXIOS metal stent for a streamlined, exchange-free procedure. Recently, Walter et al⁶⁷ reported a prospective safety and feasibility study in a multicenter trial on the use of LAMS for EUS-GBD in acute cholecystitis. Technical success was achieved in 90% and successful LAMS removal was achieved in 15 of 30 patients after a mean follow-up of 91 days. Interestingly, no stent-related complications were observed in the subgroup of patients in whom the stents were left in situ (either because removal was technically difficult or because of patients’ choice). The authors concluded that EUS-GBD is an elegant, safe and minimally invasive procedure when performed by experienced endoscopists. Larger prospective trials are needed to address technical issues and answer remaining questions, such as the need for stent removal and the optimal time to do so.

ERCP is currently the gold standard for establishing biliary drainage in cases of benign or malignant obstruction of the biliary system.⁶⁸ PTBD is an established alternative to retrograde biliary drainage but is generally associated with more complications such as bleeding, cholangitis, bile leakage, and pneumothorax.⁶⁹ The overall frequency of such complications is in the order of 5% to 25%, with a frequency of procedure-related deaths of up to 5.6%.^69,70 PTBD is contraindicated in the presence of ascites and it may need several stages to be completed, often leaving the patient with temporary (or permanent) external drains (leading to fluid and electrolyte imbalance and cosmetic issues potentially affecting the patient’s quality of life).

EUS has been recently used as an alternative to ERCP and PTBD to establish biliary drainage predominately in the context of malignant biliary obstruction. Giovannini et al⁷¹ first reported a case of EUS-CDS with the placement of a 10 F plastic stent after wire access to the common bile duct (CBD) from the duodenal bulb. EUS-HGS, an alternative approach to EUS-CDS, involves transgastric access to the biliary system of the left lobe of the liver and placement of a bridging stent, thus draining that system into the gastric lumen.⁷²

The basic technique of EUS-CDS is as follows: the CBD is identified with the EUS probe in the duodenal bulb; under EUS guidance a puncture is made with a 19 G needle and a guidewire advanced into the biliary system in a retrograde fashion. The newly created fistula is subsequently dilated (usually with a balloon catheter) and a stent (usually FC-SEMS) is advanced from the duodenum into the CBD.⁷³ In EUS-HGS, the left intrahepatic bile ducts are identified with the linear echoendoscope (usually from the proximal body of the stomach). Any left lobe dilated duct close to the stomach (usually second order duct to segment 3) can be punctured with 19 G or 22 G needle. After secure wire access and tract dilation, a stent (usually FC-SEMS) is placed into the intrahepatic ducts.⁷⁴

Currently, it is not appropriate to use EUS-BD as the first-line method to establish biliary drainage before ERCP, as the technique has not been standardized and there are no dedicated accessories.⁷⁵ A large retrospective Spanish national survey of EUS-BD conducted in 19 hospitals (a mixture of secondary and tertiary referral centres) was reported is 2012.⁷⁶ A total of 125 patients were treated with EUS-BD. Interestingly, none of the endoscopists participating in the study had undertaken more than a total of twenty EUS-BD procedures. Success rates for EUS-HGS and EUS-CDS were 64.7%, 86.3%, respectively. Four perforations were caused by inward stent migration after EUS-HGS (4/34, 11.8%) and 5 patients (4.0%) died, 2 of them from bile peritonitis, 2 from perforation and 1 from intraperitoneal bleeding. These results seem to provide more realistic clinical data without publication bias.⁷⁷ There appears to be a steep learning curve in training for EUS-BD and the rate of complications even in the most experienced hands is still relatively high. Some experts recommend undertaking the first 20 procedures under a mentor’s supervision. The development of animal models or three-dimensional printing model for EUS-BD may be of value in helping trainees acquire the necessary skills early on in their training.^78,79

It is also possible that the technical aspects of EUS-BD will be simplified and standardised resulting in an easier, faster and safer procedure for patients. The introduction of new “purpose-built” safer accessories could play a major role towards developing this technique. Recently, a retrospective multicentre European study reported on the use of the Axios and Hot Axios systems (AXIOS and Hot AXIOS; Boston Scientific) in 57 consecutive patients after failed ERCP for biliary drainage. The reported 98% technical and over 96% clinical success rate, with only a 7% major complications compares favourably to previously published data using “conventional” FC-SEMS.⁸⁰ Similarly, a small randomised phase 1 trial of EUS-CDS with a novel 7 F stent introducer with tapered metal tip which acts as a dilator for one-step stent placement found the new system as efficacious as the “three step” process (puncture, tract dilation, and stent deployment), resulting in significant shortening of the duration of the procedure.⁸¹ The stent used with the novel introducer system was a modified hybrid stent with an uncovered funnel-shaped proximal end and covered mid-body and a distal end; in addition, four anti-migratory flaps were placed in the distal end of the prosthesis to prevent spontaneous migration into the biliary system.⁸¹

Minimal invasive EUS-guided endoscopic procedures have been developed as alternatives to the more conventional surgical and percutaneous approach for drainage of PFCs, the inflamed gallbladder as well as the obstructed biliary system when ERCP fails. Currently, most of these procedures are not considered “mainstream” and are not included in conventional algorithms for the management of these conditions. The complication rate for some of these techniques (such as EUS-BD and especially EUS-HGS) remains high even in experienced hands. Most of these techniques are not part of the training curriculum of interventional EUS and clearly this will have to be addressed. Few specialized pancreaticobiliary endoscopists in high volume centres with appropriate interventional radiological and surgical backup should carry out these invasive EUS-guided interventions. Finally, further technological research is currently underway to develop new stent designs, such as the LAMS and encouraging preliminary data to support the notion that with the development of “purpose-built” accessories, the procedure can be simplified and risks minimized. With that in mind, it would not be surprising if these minimally invasive EUS-based techniques become first line techniques in the foreseeable future, replacing PTBD and even some ERCP procedures.