Key points

Introduction

It is estimated that in 2014, a total of 46,420 cases of pancreatic cancer were diagnosed in the United States, as well as 10,650 cases of gallbladder and other biliary tumors with a trend toward a higher incidence of intrahepatic compared with extrahepatic cholangiocarcinomas. The median survival of all patients with biliary tract cancers is 4.8 months with 1-year and 5-year survival rates of 31% and 10%, respectively.

Given this poor prognosis, in a significant proportion of these patients biliary drainage is required because a palliative approach is indicated.

This review covers stenting of the biliary system in different clinical scenarios: patients with either palliative or resectable malignant biliary obstruction located either proximally or in the distal biliary tree. After a succinct discussion on drainage strategy, the authors focus on comparative data available to address technical and periprocedural considerations. This article does not cover issues pertaining to the evaluation, diagnosis, or treatment other than stenting or treatments incorporated within stents even if they are performed endoscopically (eg, photodynamic therapy or radiofrequency ablation).

Introduction

It is estimated that in 2014, a total of 46,420 cases of pancreatic cancer were diagnosed in the United States, as well as 10,650 cases of gallbladder and other biliary tumors with a trend toward a higher incidence of intrahepatic compared with extrahepatic cholangiocarcinomas. The median survival of all patients with biliary tract cancers is 4.8 months with 1-year and 5-year survival rates of 31% and 10%, respectively.

Given this poor prognosis, in a significant proportion of these patients biliary drainage is required because a palliative approach is indicated.

This review covers stenting of the biliary system in different clinical scenarios: patients with either palliative or resectable malignant biliary obstruction located either proximally or in the distal biliary tree. After a succinct discussion on drainage strategy, the authors focus on comparative data available to address technical and periprocedural considerations. This article does not cover issues pertaining to the evaluation, diagnosis, or treatment other than stenting or treatments incorporated within stents even if they are performed endoscopically (eg, photodynamic therapy or radiofrequency ablation).

Who should undergo biliary drainage?

Any biliary drainage is associated with possible complications, the risk of which has to be weighed against benefits in any given clinical situation ( Table 1 ). Meta-analyses have suggested more frequent negative outcomes in patients undergoing routine preoperative biliary drainage before surgery for potentially resectable distal malignant biliary obstruction. Therefore, biliary drainage is usually performed only in surgical patients who are candidates for neoadjuvant therapies, in patients with acute cholangitis, or in patients with intense pruritus or in whom surgery will be delayed.

When drainage is attempted for patients scheduled to receive neoadjuvant therapies, a plastic or short intrapancreatic covered SEMS is preferred, with recent data favoring SEMS insertion in this clinical situation because of premature plastic stent clogging in the face of a paucity of adequately controlled data for this clinical situation. In the case of hilar cholangiocarcinoma, a meta-analysis of observational studies has noted that in those with resectable tumors, the use of preoperative biliary drainage was also associated with greater overall postoperative complication and infectious rates ; the investigators did not differentiate between percutaneous and endoscopic approaches. Endoscopic palliative biliary drainage has also been associated with an improved quality of life as demonstrated in a randomized controlled trial (RCT) and in a real-life setting.

Biliary drainage for unresectable distal biliary lesions

In the case of unresectable pancreatic and peripancreatic tumors in which palliation is the goal, a meta-analysis of RCTs comparing surgical drainage procedures with an endoscopic approach has demonstrated a lower rate of recurrent biliary obstruction (relative risk [RR], 0.14; 95% confidence interval [CI], 0.03–0.63) with surgery and no difference in major complications or mortality, but endoscopy was associated with a shorter hospital stay. Of note, 4 of the 5 studies in this meta-analysis used plastic stents as definitive endoscopic drainage method, limiting the contemporary interpretation of these results in the era of metal stents. Issues related to actual patient resectability in an unselected group, coupled with the availability of timely surgery and the recent encouraging responses to preoperative neoadjuvant therapies for locally advanced tumors that become resectable, are perhaps the principal reasons why an endoscopic approach is currently the preferred method of drainage in the palliative setting in spite of these data. A more recent factor is the increasing effectiveness of triple chemotherapy for pancreatic cancer, which has introduced the possibility of downstaging of the tumor in several patients initially deemed unresectable and palliative.

Biliary drainage of hilar lesions: mapping and minimizing risk of endoscopic drainage

The optimal strategy involved in the biliary drainage of patients who present with hilar lesions is particularly complex owing to 3 main factors: the possible contamination of bile in sequestered biliary segments, the potential for resection with curative intent, and the need to optimize the preservation of liver function.

A new classification system for the prediction of survival in patients with perihilar cholangiocarcinoma, defined as tumors extending from the secondary branches of the right and left hepatic ducts to just above the site of cystic duct origin, has been found to exhibit better performance in predicting survival than the conventional TNM system. This categorization is useful when planning the best method of intervention.

To reduce the incidence of cholangitis, which is a serious complication attributable to undrained ducts after contrast injection, it has been advocated to plan in advance which duct to drain and to avoid unnecessary and possibly damaging use of contrast during endoscopic retrograde cholangiopancreatography (ERCP) as discussed further on.

Preprocedural planning and assessment of the site and extent of biliary obstruction, as well as the presence of lobar atrophy, should precede the insertion of any stent or percutaneous drain. This procedure includes imaging with a view to resectability (including appropriate identification of the hilar vascular anatomy and possible vascular invasion) and early surgical consultation. Preprocedural imaging also assesses the size of the liver segments and the possibility of preoperative segmental portal vein embolization to promote preoperative growth of residual liver volume when radical liver surgery is being considered. When a lobe is atrophied, suggesting vascular tumoral hilar involvement, there is little if any rationale to support stenting of that lobe. Draining 50% of the liver volume is associated with better drainage effectiveness and survival compared with draining less than 50% of the liver volume. Indeed, a study by Vienne and colleagues demonstrated that draining more than 50% of the liver volume produced the best results in terms of drainage effectiveness (odds ratio [OR], 4.5; 95% CI, 1.07–6.46) as well as patient survival (OR, 0.56; 95% CI, 0.32–0.82). Furthermore, intubating an atrophic sector resulting in less than 30% of the total liver volume was unhelpful and increased the risk of cholangitis (OR, 3.04; 95% CI, 1.24–7.48). A similar finding was demonstrated in a study by Chang and colleagues. The choice of liver segment or segments to be drained must also take into account any embolization and anticipated regional postembolization hypertrophy. Patient-related factors that have been associated with failure of resolution of jaundice include a high baseline bilirubin level, diffuse liver metastases, and an international normalized ratio greater than 1.5. Technical considerations in the biliary drainage of hilar lesions are described in later sections.

Prophylactic antibiotics

A meta-analysis that included 7 RCTs with 1389 cases did not find a reduction in routine post-ERCP cholangitis among unselected cases (RR, 0.58; 95% CI, 0.22–1.55). In a large population-based registry that included 31,188 ERCPs for a variety of indications, the postprocedure adverse event rate was 11.6% in patients who received prophylactic antibiotics compared with 14.2% in those who did not receive them (OR, 0.74; 95% CI, 0.69–0.79). This effect was also present in the subgroup of patients with obstructive jaundice, in whom a resultant reduction in septic complications was observed (OR, 0.76; 95% CI, 0.58–0.97). In contrast, a recent meta-analysis of 9 RCTs and 1573 patients showed a generalized reduction in cholangitis (RR, 0.54; 95% CI, 0.33–0.91), septicemia (RR, 0.35; 95% CI, 0.11–1.11), and bacteremia (RR, 0.50; 95% CI, 0.33–0.78), but no reduction in complications among patients undergoing a successful initial ERCP drainage (RR, 0.98; 95% CI, 0.35–2.69). The investigators of this Cochrane review concluded that further research was required to determine whether antibiotics should be given during or after an ERCP if it becomes apparent that biliary obstruction cannot be relieved during that procedure.

The authors thus suggest that in cases in which technical success is achieved by inserting the stent or stents in the desired segment or segments of the biliary tree (and a clinical success is anticipated with all opacified segments adequately drained), there is likely no need for antibiotic administration; in contradistinction, antibiotic administration is prudent in patients in whom there is failure or suspicion of failure to drain the targeted biliary system.

Available technologies for biliary drainage

Plastic Stents

Plastic stents are made from different materials such as polyethylene, Teflon, or polyurethane. Polyethylene stents have the advantage of being softer than Teflon stents and possibly change their shape, to a certain extent, in the bile duct ( Fig. 1 ). These stents also come in different designs, some with a slightly curved shape to adapt to the common bile duct (CBD) contour and to decrease migration. S-shaped stents are used for draining the left biliary system. Stents with single or double pigtail ends are also available for anchoring these stents. Tannenbaum stents with multiple side flaps, to prevent stent migration, but without side holes are also available, and they may adopt a double-layer design. Plastic stents with an antireflux valve have also been developed.

Available technologies for biliary drainage. EUS, Endoscopic ultrasound.

The size of plastic stents is limited by the size of the accessory channel of the scope, with 14F stents being the largest insertable, but these are difficult to deliver. Among plastic stents, those with 10F diameter have the longest demonstrated patency. However, the average plastic stent patency does not exceed 3 to 4 months, with as mentioned earlier, more recent data suggesting shorter actual patency times in real life ( Fig. 2 ).

A 64-year-old woman with metastatic adenocarcinoma of unknown origin and poor functional status developed biliary obstruction due to liver metastasis. ( A ) There is external compression of the common hepatic duct with dilatation of the right biliary system; selective cannulation of the duct is achieved. ( B ) A plastic biliary stent is inserted with satisfactory position.

Self-Expandable Metal Stents

SEMSs have been developed to overcome the limitations of plastic stents (early obstruction due to small caliber), but these stents can still occlude ( Box 1 ). Covered SEMSs are available for distal malignant biliary obstruction; their use in hilar tumors is not indicated because they would block the contralateral hepatic duct as well as intrahepatic side branches and potentially cause cholangitis. Two meta-analyses comparing covered with uncovered SEMSs have been published. Each analysis drew different conclusions but differed as to the included studies. The meta-analysis by Saleem and colleagues included 5 RCTs involving 781 patients. The use of covered SEMSs resulted in a longer median stent patency time with a weighted mean difference of 60.6 days (95% CI, 26.0–95.2); however, there were also greater associated rates of stent migration (RR, 8.11; 95% CI, 1.47–44.76), tumor overgrowth (RR, 2.02; 95% CI, 1.08–3.78), and sludge formation (RR, 2.89; 95% CI, 1.27–6.55). In contradistinction, the meta-analysis by Almadi and colleagues included 4 abstract publications in addition to the 5 fully published RCTs, totaling 1061 patients. The weighted mean difference in stent patency was 67.9 days, favoring covered SEMSs over uncovered SEMSs (95% CI, 60.3–75.5), but this conclusion was based on only 2 trials. There existed no difference in patency rates when comparing both stents at 6 months (OR, 1.82; 95% CI, 0.62–5.25) or 12 months (OR, 1.25; 95% CI, 0.65–2.39). These analyses are limited by marked variability in the adopted definitions of stent patency, patient selection, presence of metastasis, type of tumors, as well as the route of insertion across studies ( Figs. 3 and 4 ).

Box 1

• 
  

  Tumor ingrowth/mucosal hyperplasia

  
  
• 
  

  Tumor overgrowth

  
  
• 
  

  Sludge with/without stones

  
  
• 
  

  Hemobilia

  
  
• 
  

  Food impaction

  
  
• 
  

  Bile duct kinking

  
  
• 
  

  Others

Causes of stent occlusion

Adapted from the Isayama H, Hamada T, Yasuda I, et al. TOKYO criteria 2014 for transpapillary biliary stenting. Dig Endosc 2015;27:259–64.

A 58-year-old woman had poorly differentiated mucin-secreting adenocarcinoma of the gallbladder. ( A ) Initial ERCP was performed when the patient had cholangitis, and no clear diagnosis was reached. A long stricture is seen involving the common hepatic duct and extending into the right and left hepatic ducts. ( B ) A plastic biliary stent was inserted. ( C ) The patient developed cholangitis, and an internal/external percutaneous biliary drain was inserted through interventional radiology. ( D ) She subsequently developed signs of gastric outlet obstruction and did not want percutaneous drains, so an ERCP was performed and the duodenum was dilated to facilitate insertion of the duodenoscope. ( E ) Both the right and left biliary systems were cannulated with a wire, and an uncovered SEMS was inserted in to the left system. ( F ) Then, a second uncovered SEMS was inserted into the right biliary system. ( G ) Next, a duodenal stent was inserted. ( H ) Final image after removal of the percutaneous drain.

A 50-year-old man was diagnosed with pancreatic adenocarcinoma by fine-needle aspiration via endoscopic ultrasound that was deemed to be unresectable. ( A ) ERCP shows a distal common bile duct stricture. ( B ) A single uncovered SEMS was inserted. ( C ) Final image after deployment of the stent.

Furthermore, there exist differences across commercially available SEMS types in axial and radial forces, extent of covering (whether partial or complete), smoothness of the inner surface, as well as the presence of antimigration systems (flared ends, flanges, anchoring fins, flaps, etc). Two more recent RCTs have since assessed partially or completely covered SEMS with antimigration systems. Both demonstrated increased duration of stent patency for the covered SEMS with disparate conclusions about migration rates.

In a small study, the use of SEMSs with an antireflux system was studied in 13 patients with distal malignant biliary obstruction that had already occluded because of food debris. Although this technology was effective in this small sample and resulted in a longer patency time compared with the initial conventional SEMS use, the novel stent design carried a high migration rate (31%).

An RCT noted a survival advantage for the use of percutaneously placed covered Viabil SEMS (Conmed) compared with uncovered Wallstents (Boston Scientific) (243.5 vs 180.5 days, respectively; P value = .04) in palliating extrahepatic cholangiocarcinomas ; another RCT by the same group comparing covered with uncovered SEMS for patients with pancreatic head cancers did not show any difference in clinical outcomes.

The use of sphincterotomy

In a meta-analysis that included 3 RCTs, the use of sphincterotomy before stent insertion was associated with a lower rate of pancreatitis (OR, 0.34; 95% CI, 0.12–0.93) but a higher rate of bleeding (OR, 9.70; 95% CI, 1.21–77.75). Of note, a different type of stent was used in each of these studies including plastic, uncovered, and covered SEMSs. Furthermore, most of the cases of pancreatitis originated from the study by Zhou and colleagues that had an unusually high rate of post-ERCP pancreatitis (31.7%) in the nonsphincterotomy group. A recent RCT found no added benefit in performing a sphincterotomy in cases with nonresectable pancreatic cancer. Whether these considerations apply to perihilar tumor strictures when more than 1 stent may be needed, and in which the geometric forces exerted onto the pancreatic opening, is unclear.

The Otaru consensus has recently stated that it is unnecessary to perform a sphincterotomy in patients with pancreatic cancer requiring biliary stenting, although there were reservations in generalizing this statement to patients with all causes of distal malignant biliary obstruction.

Insertion of stents above versus across the sphincter of Oddi

It has been postulated that stent insertion where the distal tip remains above the papilla could prevent the reflux of bacteria and undigested materials into the biliary system as well as the stent and thus contribute to prolonged patency. In an RCT of patients with malignant biliary obstruction, there was no difference according to where the lower extremity of a plastic stent was placed with regard to its impact on stent patency duration, but the study might have been underpowered. A retrospective cohort study suggested that the most significant risk factor for the development of cholangitis was a transpapillary position of the SEMS but that the corresponding CI on multivariable modeling was extremely wide, whereas a second study found that inserting the SEMS across the sphincter was associated with developing pancreatitis (4.1% vs 25.0%, P value <.01) but that there was no effect on cumulative stent patency or patient survival. An international RCT of 84 patients with unresectable malignant biliary obstruction randomized subjects to either insertion of a fully covered SEMS above the sphincter of Oddi without a sphincterotomy versus insertion of the same SEMS across the sphincter of Oddi after a sphincterotomy. Placement of the SEMS above the sphincter of Oddi did not prolong SEMS patency and did not reduce the incidence of cholangitis without occlusion, whereas placement across the sphincter of Oddi resulted in more external migration. SEMS occlusion was also more frequent in those with pancreatic cancer if the SEMS was inserted above the sphincter of Oddi.

Plastic stents versus self-expandable metal stents

A meta-analysis of 7 trials (724 patients) suggested that SEMS were associated with greater patency rates when compared with plastic stents for malignant biliary obstruction with an RR of stent occlusion of 0.44 (95% CI, 0.3–0.63). In a meta-analysis comparing SEMSs to plastic stents in hilar malignant strictures, SEMSs were associated with a higher successful drainage rate (OR, 0.26; 95% CI, 0.16–0.42), lower early complication rate (OR, 2.92; 95% CI, 1.65–5.17), longer stent patency (hazard ratio [HR], 0.43; 95% CI, 0.30–0.61), and a patient survival advantage (HR, 0.73; 95% CI, 0.56–0.96). This result was emphasized in the Otaru consensus.

It has been suggested that it is cost effective to insert plastic stents if the patient’s survival is expected to be less than 4 months, whereas insertion of an SEMS is favored if anticipated survival is greater or if the cost of the SEMS is less than 50% of the cost of the ERCP. SEMS were recently found to be more cost effective in settings not only in which ERCP costs are high but also in which these are lower than the cost of SEMS. Most recently, a cost analysis in an RCT determined an average cost of US $6541 when an SEMS was inserted initially compared with US $19,054 when a plastic stent was first used. Furthermore, the chance of experiencing no occlusion over the 12 months was 65% for the SEMS group, whereas it was only 13.85% for the plastic stent group. A second very recent cost-effectiveness analysis of an RCT from the Netherlands confirmed no significant difference in the total cost per patient between plastic stent and SEMS in patients with short survival (€6555 vs €5719, respectively; P = .4) or metastatic disease (€6593 vs €6179, respectively; P = .69).