Introduction

Endoscopic biliary stent placement is widely accepted as palliation for malignant biliary obstruction or as a treatment of benign biliary stricture. Many decades have passed since self-expandable metallic stents (SEMS) were introduced, and although various biliary stent designs are now commercially available, no single ideal stent has yet been developed. An ideal stent should be patent until death, or surgery, in patients with resectable malignant biliary obstruction. Fewer complications, maneuverability, cost-effectiveness, and removability are also important factors. Alternatively, should we aim to develop a novel method for biliary drainage other than biliary stenting via endoscopic retrograde cholangiopancreatography (ERCP)? This article reviews the current status of biliary stenting for malignant biliary obstructions.

Resectable malignant biliary obstruction: plastic, metal, or no stent?

Preoperative drainage is considered essential in patients with resectable periampullary malignancy, although biliary drainage itself can increase postsurgical complications such as infection because of bacterial colonization or inflammation induced by preoperative drainage. Given the short time available, plastic stents (PS) are the standard method for preoperative drainage. Recently, van der Gaag and colleagues conducted a randomized controlled trial (RCT) of preoperative biliary drainage in patients with obstructive jaundice caused by pancreatic cancer. Data suggested no benefit as a result of preoperative drainage in terms of complications or mortality. Overall serious complication rates were 39% and 74% ( P <.001) and surgery-related complications were 37% and 47% in the early surgery and biliary drainage groups, respectively ( P = .14). These investigators concluded that routine preoperative drainage increases complication rates and is not recommended. However, 46% of the biliary drainage group experienced complications, which appeared too high, as pointed out in the editorial in this issue. Thus, more trials using either PS or SEMS are needed to confirm these investigators’ conclusion and to alter clinical practice. Because the incidence of complications with currently available covered SEMS (CMS) is high, the role of CMS in the preoperative setting is limited.

Neoadjuvant chemotherapy or chemoradiation therapy is a recent issue in borderline resectable pancreatic cancer. Most borderline resectable pancreatic cancer arises in the head of the pancreas and involves or abuts the major vessels. Longer stent patency is needed because the preoperative period in patients receiving neoadjuvant therapy is longer, and stent occlusion or stent-related complications cause cessation of neoadjuvant therapy. The occlusion rate of PS in patients with pancreatic cancer who underwent neoadjuvant chemoradiation therapy is as high as 55%. Therefore, SEMS are an option for biliary drainage in this setting. Wasan and colleagues compared SEMS and PS as preoperative biliary drainage in resectable pancreatic cancer. Although no stent-related postoperative complications occurred in either group, cholangitis or cholestasis was noted in 15% in the SEMS group and 93% in the PS group. In another retrospective study, use of SEMS did not increase surgery-related complications (34%) or morbidity (0%); stent-related complication rates were 7% in SEMS and 45% in PS ( P <.0001). These results suggest that SEMS offer longer stent patency without increasing complications and result in better cost-effectiveness in patients with resectable pancreatic cancer who require neoadjuvant treatment. However, these were retrospective studies at a single institution. Thus, the safety and efficacy, including cost-effectiveness, of SEMS in patients with resectable pancreatic cancer should be confirmed in RCTs.

Resectable malignant biliary obstruction: plastic, metal, or no stent?

Preoperative drainage is considered essential in patients with resectable periampullary malignancy, although biliary drainage itself can increase postsurgical complications such as infection because of bacterial colonization or inflammation induced by preoperative drainage. Given the short time available, plastic stents (PS) are the standard method for preoperative drainage. Recently, van der Gaag and colleagues conducted a randomized controlled trial (RCT) of preoperative biliary drainage in patients with obstructive jaundice caused by pancreatic cancer. Data suggested no benefit as a result of preoperative drainage in terms of complications or mortality. Overall serious complication rates were 39% and 74% ( P <.001) and surgery-related complications were 37% and 47% in the early surgery and biliary drainage groups, respectively ( P = .14). These investigators concluded that routine preoperative drainage increases complication rates and is not recommended. However, 46% of the biliary drainage group experienced complications, which appeared too high, as pointed out in the editorial in this issue. Thus, more trials using either PS or SEMS are needed to confirm these investigators’ conclusion and to alter clinical practice. Because the incidence of complications with currently available covered SEMS (CMS) is high, the role of CMS in the preoperative setting is limited.

Neoadjuvant chemotherapy or chemoradiation therapy is a recent issue in borderline resectable pancreatic cancer. Most borderline resectable pancreatic cancer arises in the head of the pancreas and involves or abuts the major vessels. Longer stent patency is needed because the preoperative period in patients receiving neoadjuvant therapy is longer, and stent occlusion or stent-related complications cause cessation of neoadjuvant therapy. The occlusion rate of PS in patients with pancreatic cancer who underwent neoadjuvant chemoradiation therapy is as high as 55%. Therefore, SEMS are an option for biliary drainage in this setting. Wasan and colleagues compared SEMS and PS as preoperative biliary drainage in resectable pancreatic cancer. Although no stent-related postoperative complications occurred in either group, cholangitis or cholestasis was noted in 15% in the SEMS group and 93% in the PS group. In another retrospective study, use of SEMS did not increase surgery-related complications (34%) or morbidity (0%); stent-related complication rates were 7% in SEMS and 45% in PS ( P <.0001). These results suggest that SEMS offer longer stent patency without increasing complications and result in better cost-effectiveness in patients with resectable pancreatic cancer who require neoadjuvant treatment. However, these were retrospective studies at a single institution. Thus, the safety and efficacy, including cost-effectiveness, of SEMS in patients with resectable pancreatic cancer should be confirmed in RCTs.

Unresectable malignant distal biliary obstruction: covered or uncovered?

The ideal stent for unresectable malignant biliary obstruction should be easy to deploy, patent until death without complications, and provide antitumor effects. The emergence of uncoated metal stents (UMS) prolonged stent patency significantly because of their larger diameter. Historically, the cost-effectiveness of UMS has been debated. In patients with a poor prognosis (<3–6 months or liver metastasis), PS are the reasonable choice in terms of cost-effectiveness. However, the prognosis of unresectable pancreatobiliary malignancy is improved by anticancer treatment, which necessitates longer stent patency. As a result, a role for SEMS in unresectable distal malignant biliary obstruction is now frequently reported.

The disadvantage of UMS is stent occlusion by tumor ingrowth or benign epithelial hyperplasia through the mesh. CMS were developed to overcome these disadvantages. The covering membrane of CMS, which does not allow embedding of the stent into the bile duct wall, prevents tumor ingrowth and epithelial hyperplasia. CMS were considered to be ideal when they first came into use, but they have their own problems, which are discussed later.

So far, 5 RCTs have been published, including 2 that used a percutaneous approach. We reported an RCT using diamond stents and showed the superiority of hand-crafted polyurethane-covered diamond stents over uncovered diamond stents. In this study, tumor ingrowth was not observed in CMS and the stent occlusion rates were 14% and 38% in CMS and UMS, respectively. As a result, stent patency was significantly longer in CMS, but a trend toward higher complication rates of pancreatitis and cholecystitis was noted. Subsequently, 2 RCTs of a percutaneous approach confirmed a longer stent patency in pancreatic cancer and extrahepatic bile duct cancer. However, no significant differences were observed in 2 recently published RCTs. CMS did not prevent tumor ingrowth in these studies, and stent migration was observed exclusively in 12% and 3% of patients.

Meta-analysis of these 5 RCTs confirmed longer stent patency of CMS, but stent migration (relative risk, 8.11), tumor overgrowth (relative risk, 2.02), and sludge formation (relative risk, 2.89) were significantly higher with CMS. CMS were believed to increase cholecystitis or pancreatitis by blocking the orifice of the cystic duct or pancreatic duct with their covering membrane, but no significant increase in these complications was detected in this meta-analysis.

These studies clarified 2 points: what CMS can achieve and what problems CMS present. First, CMS can show significantly longer stent patency if the cover on CMS achieves its original goal of preventing tumor ingrowth. Second, the problems of stent migration or sludge formation should be resolved with CMS. Apparently, current commercially available CMS do not fulfill these requirements but we are still searching for an ideal one because the concept of CMS has been proved to be legitimate.

What is missing from the covering membrane of current CMS?

CMS were originally developed to prolong stent patency by preventing tumor ingrowth via the stent mesh. Therefore, the durability of the covering membrane is one of the most important factors of CMS. Various materials are used as the covering membrane, such as polyurethane, silicone, or expanded polytetrafluoroethylene (ePTFE). Hydrolysis from long-term bile exposure is the Achilles’ heel of these covering materials. Durability comparing these materials was reported in a phantom model with bile flow and silicone, which is used in commercially available covered Wallstents (Boston Scientific Corp., Natic, MA, USA) or covered WallFlex stents (Boston Scientific Corp., Natic, MA, USA), as being more resistant to hydrolysis than other materials.

Stent clogging from bacterial biofilm formation was extensively studied in PS. Medical treatment to prevent clogging was attempted using antibiotics or ursodeoxycholic acid, but these medications did not improve stent patency in clinical trials. CMS present similar problems because they are similar to an extremely large-pore PS. The covering membrane of ePTFE is prone to bacterial biofilm formation after bile exposure because of its rough and uneven surface. Extensibility is another factor that can affect the mechanical characteristics, but no data are available about the influence of extensibility on different covering membranes.

The ideal covering membrane must be durable to long-term bile exposure and resistant to bacterial biofilms. More ex vivo data on this topic and their relation to clinical outcomes are awaited.

Mechanical properties: are all SEMS the same?

Mechanical properties of SEMS are defined by multiple factors such as stent design, type of wire, and covering materials. As a result of combinations of these variables, we proposed radial force (RF) and axial force (AF) as major mechanical properties that affect clinical outcomes. RF is well known as an expanding force. AF is a straightening or recovery force when SEMS are bent, which is a new concept of stent properties. The measurement of RF and AF in various SEMS has been reported.

RF affects stent patency in that dilation of biliary stricture and maintenance of luminal patency depend on the expanding force of SEMS. Two factors in RF exist in terms of time course. Immediate stent expansion at the time of stent deployment affects short-term outcomes, and chronic resistant force against tumor compression affects long-term outcomes. Insufficient expansion of SEMS immediately after deployment causes immediate stent occlusion by sludge or food impaction. In general, the chronic resistant RF is higher than the immediate stent expanding force because SEMS are made of a type of shape-memory alloy. This characteristic means that SEMS sometimes expand partially immediately after deployment and then gradually expand to their full extent, even although the RF is high. To prevent immediate SEMS occlusion, balloon dilation after SEMS placement is effective in cases of partial expansion.

AF is considered to define conformability of SEMS in the bile duct and may have a greater relationship with clinical outcomes than RF. After deployment in the bile duct, SEMS are fixed at the stricture by the tumor and AF causes compression to the bile duct at both stent ends. As AF increases, so does the compression of the bile duct or cystic duct/pancreatic duct orifice. Clinically, this situation may cause kinking of the bile duct, cholecystitis, or pancreatitis. In addition, less conformability of SEMS in the bile duct leads to stent migration.

We have reported these correlations between mechanical properties and clinical outcomes. In general, AF affects clinical outcomes such as stent migration and pancreatitis more than RF. These results are useful in the development of new SEMS.

What can we do to prevent SEMS complications?

Prevention of stent-related complications is important for the patient’s quality of life and prolongation of survival. Stent-related complications necessitate reinterventions, impair the quality of life, and cause suspension of anticancer treatments. Stent migration, pancreatitis, and cholecystitis are problems with SEMS.

Stent migration is almost exclusively observed in CMS, although this characteristic provides removability. The removability of CMS expands their indications into the benign biliary stricture or bile leak. It also makes reintervention easy at the time of stent dysfunction. CMS have also been considered to pose a risk factor of cholecystitis and pancreatitis, but recent studies including a meta-analysis revealed no differences between CMS and UMS.

Several factors can cause stent migration: poor conformability of SEMS, lack of anchoring, tumor regression from chemotherapy, or stent excretion because of impacted sludge or food. The last cause should be considered as stent occlusion, but distinguishing between these 2 conditions clinically is difficult. Low AF and anchoring systems are the keys to preventing stent migration. We compared 2 CMS with different AFs (ComVi stent [Taewoong Medical, Seoul, Korea] and covered Wallstent) and the migration rate was significantly lower in the ComVi stent (2.1% vs 17.0%, respectively). The ComVi stent has a characteristic structure with a covering membrane sandwiched between 2 uncovered metallic stents, which achieves a low AF. Anchoring structures such as fins, flares, and flaps can also prevent stent migration. Park and colleagues conducted a multicenter RCT of 2 CMS with flaps and flared ends in benign biliary stricture, and the anchoring flap was superior at preventing stent migration. Both stents were removable without complications. Recently, we conducted a multicenter prospective study of covered WallFlex stents. Newly developed covered WallFlex stents are made of nitinol wires with low AF and have flared ends. As a result, stent migration rates were significantly lower in partially covered WallFlex stents (8%) than in partially covered Wallstents (17%). Stent removal was successful in 96%. Thus, we should develop CMS with effective anchoring systems and maintain removability.

Pancreatitis after SEMS placement is also an unsolved problem. Theoretically, compressing the orifice of the pancreatic duct causes pancreatitis. Cote and colleagues reported a higher incidence of pancreatitis in SEMS compared with PS, but CMS or transpapillary stenting were not risk factors in a SEMS-specific analysis. Ternasky and colleagues suggested deflection force on the pancreatic orifice by PS as a cause of postbiliary stent pancreatitis. Kawakubo and colleagues conducted a multicenter retrospective study in the placement of 370 SEMS for malignant distal biliary obstruction, and the rate of pancreatitis was 6%. The rate of pancreatitis was 5.7% with CMS and 7.4% with UMS. The significant risk factors of pancreatitis were nonpancreatic cancer and SEMS with high AF. These results are in line with Ternasky and colleagues’ suggestion because high AF causes deflection and compression of the pancreatic duct. Thus, CMS do not increase the rate of pancreatitis but CMS with high AF are prone to both migration and pancreatitis.

Cholecystitis is another complication after SEMS placement, which impairs the patient’s quality of life. Previously, covering the orifice of the cystic duct was believed to cause cholecystitis after SEMS placement, but 2 retrospective studies showed no significant differences in the rate of cholecystitis between UMS and CMS. These 2 studies both showed that tumor involvement at the orifice of the cystic duct is a significant risk factor for cholecystitis. We speculate that an expansion of SEMS causes the occlusion of the orifice of the cystic duct, which is narrowed and fixed by the tumor as opposed to patent and elastic in patients without tumor involvement. Intraductal ultrasonography (IDUS) is more accurate than cholangiography in the diagnosis of tumor involvement and we routinely perform IDUS before SEMS placement to evaluate the risk of cholecystitis after SEMS placement. The risk factor for cholecystitis has been clarified but data are lacking on the prevention of cholecystitis after SEMS placement. Our preliminary analysis suggested that SEMS with high AF were more prone to cholecystitis. From our uni- and multivariate analysis of risk factors of cholecystitis, Tumor involvement of OCD (OR 5.34; 95% CI 2.19–13.0; P = .0001) and SEMS with high-AF (OR 5.18; 95% CI 1.69–22.6; P = .027) were significant risk factors (7% of cholecystitis among 356 cases placed at the OCD, presented at DDW 2012 by Isayama and colleagues).