The plastic stent has been used for EBD for more than three decades. Several studies have suggested that stents of 10- and 12-Fr diameters are superior in terms of patency compared to those with a 7-Fr diameter [4,5]. The average duration of patency is 3 to 4 months [6]. Therefore, a plastic stent diameter of at least 10 Fr is recommended for effective drainage for a particular period of time [5].

Plastic stents have advantages that include low cost, simple insertion, and easy removal. Plastic stents have been used to achieve biliary drainage in patients with a malignant hilar obstruction. However, the use of plastic stents for hilar cholangiocarcinoma has revealed poor results in some series [7-9]. Liu et al. [10] studied the usefulness of endoscopic insertion of plastic stents in 55 patients with hilar cholangiocarcinoma, and endoscopic stenting was attempted in 49 patients. Technical success rate, clinical success rate, and early complication rate were 73%, 41%, and 25%, respectively. The median patency of the first stent was just 1 week (range, 0 to 8). Stent patency was also relatively short (49 days) in another study [11].

The main disadvantage of plastic stents is their ease of clogging. Stent occlusion frequently results in cholangitis with rates as high as 20% to 40% for unilateral or bilateral plastic stents. Clogging is caused by a relatively narrow lumen (10 to 12 Fr) and is proportional to stent length [7]. The mechanism of stent clogging is multi-factorial and is associated mainly with bacterial contamination of undrained bile ducts [8,9]. Another disadvantage of plastic stents is the technical limitation of inserting a large diameter plastic stent. It is technically difficult to place more than one 10-Fr stent at the initial setting. Moreover, plastic stents in the hilum are prone to distal migration.

Due to the plastic stent clogging problem, stent exchange is required in one third of patients [12]. It is necessary to exchange plastic stents regularly every 3 to 4 months or when obstructive symptoms appear [13]. Many trials to prevent stent clogging, including increasing stent diameter, prophylactic administration of antibiotics, administration of drugs such as ursodeoxycholic acid that alter bile composition, and using newly modified stents have been tried, but most results have been disappointing [14-18]. Only increased stent diameter has prolonged stent patency [19,20]. Although no randomized controlled studies have compared plastic and metal stents, plastic stent use in patients with hilar cholangiocarcinoma has decreased because of its relatively short patency, difficulty of multiple stenting, and obstructive risk of side branches.

Self-expanding metal stents (SEMS) were developed to overcome the limitations of the plastic stent. The main advantage of SEMS is their large internal diameter (generally 10 mm; i.e., 30 Fr), and they provide patency duration of about 10 months, reducing the necessity for reintervention (Table 1) [8,10,11,13,21-29]. SEMS also allows drainage of secondary branches through the open side mesh of the stent. In a study comparing plastic and metal stents in 20 patients with hilar cholangiocarcinoma (Bismuth II to IV), SEMS showed higher patency rates, a lower rate of cholangitis (33.3% vs. 9.1%), and required fewer reinterventions (2.4 ± 2.6 vs. 0.4 ± 0.5). Median patency was not described accurately, but stent malfunction was observed in 50% of patients in the plastic group and 18.2% in the metal group after long-term follow-up (> 30 days) [13].

Raju et al. [22] compared plastic and metal stents in 100 patients with hilar cholangiocarcinoma. Metal and plastic stents were placed in 52 and 48 patients, respectively. Successful drainage was achieved in 46 (95.8%) in the SEMS group and 49 (94.2%) in the plastic group. Median patency was 5.56 months in the SEMS group and 1.86 months in the plastic group. Median survival was 9.08 and 8.22 months in the SEMS and plastic groups, respectively. More reintervention was needed in the plastic group [22].

Moss et al. [5] reported that metal stents are associated with a significantly lower relative risk (0.44) of stent occlusion than that of plastic stents. The median increase in the cost effectiveness ratio of metal stents is US $1,820 per endoscopic retrograde cholangiopancreatography (ERCP) prevented [6]. Another study reported that even in countries where the cost of the ERCP procedure are lower than those of SEMS, the average total cost is lower for the SEMS group compared with that in the plastic group [30].

Although their initial cost is higher, SEMS have been increasingly used based on the above results. Compared with plastic stents, a larger diameter SEMS results in improved stent patency and fewer repeat procedures [9,31,32]. In addition, the open-mesh design allows drainage of secondary branches through the side of the stent and less likelihood for obstructive segmental cholangitis. Insertion of an uncovered SEMS is preferred when unilateral drainage is planned due to prolonged patency and prevention of branch duct obstruction. Although few data are available on hilar cholangiocarcinoma, current evidence favors placement of SEMS as they provide superior palliation in comparison to plastic stents in terms of early and late complications [23,33].

The main problem with uncovered SEMS is tumor in growth into the stent lumen, which results in stent obstruction. Polyurethane or silicone covered SEMS have been developed to prevent tumor in-growth. The results of some studies support this hypothesis, but contradictory results have also been reported [34-37]. Another study performed with 65 patients with hilar cholangiocarcinoma compared 19 plastic stents, 98 bare metal stents, and nine covered metal stents. The patencies of the plastic stent, bare metal stent, and covered metal stent were 49, 165, and 104 days, respectively. The median survival time and median event-free survival time were also longer in the bare metal stent group than those in the plastic stent group. Although that study was limited by the heterogeneity of the patient groups and types of stents, the patency of the covered metal stent seemed not to be superior to that of the bare metal stent. A covered stent is not usually used in patients with hilar cholangiocarcinoma due to unintentional obstruction of contralateral ducts or side branch ducts and lack of improved stent patency. Thus, uncovered SEMS is the treatment of choice in patients with hilar cholangiocarcinoma.

Unilateral stenting is simple and similar to the conventional endoscopic stenting technique. An unresolved issue concerns which duct should be selected and drained when placing a unilateral stent. A developing approach to manage malignant hilar strictures is the use of magnetic resonance cholangiopancreatography (MRCP) or computed tomography (CT)-guided unilateral endoscopic drainage [42]. This technique involves obtaining an MRCP and/or CT image before the endoscopic intervention. MRCP/CT images are useful to determine the main hepatic duct that drains the largest number of viable segments, which avoids draining atrophied lobes [41,42,49]. The postprocedural cholangitis risk can be reduced by avoiding drainage of atrophied lobes. However, it is very difficult to select the intended duct endoscopically compared to the percutaneous route. Other authors have suggested no difference whether the right or left hepatic duct is drained, and that the most easily accessible duct should be used [45]. However, the right hepatic duct is frequently and more easily accessed anatomically, regardless of obstruction severity. A mildly obstructed duct that drains well and has less necessity for drainage is frequently selected and drained, instead of an obstructed duct or a duct causing cholangitis.

Bilateral drainage is technically difficult and challenging. Sphincterotomy is always performed with the use of prophylactic antibiotics. Care must be taken during the initial positioning of every guidewire to ensure that the left-sided guidewire is inserted first if possible because left stenting is more difficult to achieve than right stenting. When it is very difficult to initially place the guidewire into the left hepatic duct, "vice versa" is an alternative. "Stent by stent" and "stent in stent" are the two techniques available for endoscopic bilateral stenting. The "stent-by-stent" technique is a bilateral stenting procedure with parallel SEMS to drain both intrahepatic ducts. The technique is performed as follows: 1) initial insertion of bilateral semi-rigid hydrophilic guidewires with flexible tips, 2) ensuring that the guidewires are not entangled, 3) preliminary bilateral dilation of the stenosis using a hydrostatic balloon, and 4) initial insertion of the left-sided stent followed by insertion of the right-sided stent (Fig. 1). Several studies have reported endoscopic bilateral biliary drainage using this technique (Table 2) [13,24,25,29,36,54-56].

The "stent-by-stent" technique has some problems and difficulties that include potential entanglement of the two guidewires, fragility and dislodgement of the stents, and difficulty positioning both stents exactly. As an alternative, the "stent-in-stent" technique was developed to place bilateral SEMS in a Y configuration, in which the second stent traverses through the open-mesh wall of the first stent to enter the contralateral bile duct [57]. The transverse stent with the Y configuration, called a Y stent (Niti-S Biliary Y stent, Taewoong, Seoul, Korea), was developed as a hybrid of the spiral and Z components (Fig. 2).

The mesh of the central part of the Y stent becomes bigger by omitting the Z component (forming a central area of relatively wide-open mesh), resulting in an open-weave portion that is 10 mm long in the middle of the stent (the present form is composed of a longer central wide portion, whole large cell type, or open mesh design). The Y stent is 6 to 8 cm in length and 10 mm in diameter, with gold markers indicating the central mesh. The first Y stent has two major disadvantages: the tendency for easy tumor in-growth in the wider-central mesh portion compared to that of the smaller mesh portions on both ends, and poor stent expansion due to weak radial force. Several improved SEMS versions have been developed to overcome the disadvantages of the first Y stent (M-Hilar stent, Standard Sci-Tech, Seoul, Korea and LCD type Y stent, Taewoong, respectively). Several types of second stents (Niti-S, Niti-D, LCD type stent, Taewoong; M-Hilar stent, Standard Sci-Tech; Zilver stent, Cook Endoscopy, Winston-Salem, NC, USA), introduced via the central mesh of the Y stent, have a length of 6 to 8 cm and a diameter of 10 mm. The stent placement technique is as follows. After the strictures are negotiated with the guidewire into the left hepatic duct, the Y stent is deployed in the left hepatic duct. The guidewire left in place across the Y stent is carefully withdrawn with an ERCP catheter, without pulling it back completely, and is then inserted under fluoroscopic guidance into the right hepatic duct through the central open mesh of the Y stent. Another uncovered second metal stent is then introduced over the guidewire through the central open mesh. This stent is deployed in the right hepatic duct (Fig. 3).

When difficulties are encountered when inserting the guidewire into the intended hepatic duct, modified techniques involve using a more hydrophilic Terumo guidewire, a stiffer guidewire, a retrieval balloon catheter, a pull-type endoscopic sphincterotomy knife, or a triple-lumen catheter (Haber RAMP catheter, Cook Endoscopy) [24,29,42,54,58,59].

Passing a tapered-tip catheter or a stent through the stricture can be unsuccessful, even after obtaining guidewire access to the desired hepatic duct. In these cases, a 7-Fr Soehendra stent extractor (Cook Endoscopy), a 6 to 8 mm hydrostatic balloon (Hurricane CRE wire-guided esophageal/pyloric balloon dilation catheter, Boston Scientific, Cork, Ireland), or ENBD or guidewire left in place for several days can be used [60]. Few studies have compared the "stent-by-stent" and "stent-in-stent" bilateral stenting techniques. One retrospectively reviewed the results of these two techniques, but no statistical analysis was performed [61]. There is still no consensus as to which method is more useful to achieve bilateral stenting. However, the "stent-in-stent" method has become more widely used with recent improvements in technique and devices (Table 3) [27,43,57,58,62-66].