Advances in endoscopic management of biliary tract complications after liver transplantation


Address reprint requests to Martin L. Freeman, M.D., Division of Gastroenterology, Hepatology, and Nutrition, University of Minnesota, 406 Harvard Street Southeast, MMC 36, Minneapolis, MN 55455. Telephone: 612-625-8999; FAX: 612-625-5620; E-mail:


Biliary tract complications after liver transplantation (LT) most commonly include biliary leaks, strictures, and stone disease. Living donor recipients and donation after cardiac death recipients are at an increased risk of developing biliary complications. Biliary leaks usually occur early after transplantation, whereas strictures and stone disease occur later. The diagnosis of biliary complications relies on a combination of clinical presentation, laboratory abnormalities, and imaging modalities. Biliary leaks are usually diagnosed on the basis of bilious output from a surgical drain, fluid collections on imaging, or a cholescintigraphy scan demonstrating a leak. Magnetic resonance cholangiopancreatography (MRCP) is noninvasive, does not require the administration of an intravenous contrast agent, and provides detailed imaging of the entire biliary system both above and below the anastomosis. The latter not only helps in the diagnosis of biliary strictures and stones before patients undergo invasive procedures such as endoscopic retrograde cholangiopancreatography (ERCP) but also allows treating physicians to plan the optimal treatment approach. MRCP has, therefore, replaced invasive therapeutic modalities such as ERCP as the modality of choice for the diagnosis of biliary strictures and stones. There have been significant advances in endoscopic accessories, including biliary catheters, wires, and stents, as well as endoscopic technologies such as overtube-assisted endoscopy over the last decade. These developments have resulted in almost all patients, including those with difficult strictures or altered surgical anatomies (eg, Roux-en-Y hepaticojejunostomy), being treated via an endoscopic approach with percutaneous transhepatic cholangiography, which is more invasive and associated with significant morbidity, with surgery being reserved for a small minority of patients. Advances in the diagnosis and endoscopic management of patients with biliary complications after LT are discussed in this review. Liver Transpl 19:482–498, 2013. © 2013 AASLD.


anastomotic stricture


computed tomography


donation after circulatory death


deceased donor liver transplantation


endoscopic retrograde cholangiography


endoscopic retrograde cholangiopancreatography


living donor liver transplantation


liver transplantation


magnetic resonance cholangiopancreatography


nonanastomotic stricture


percutaneous transhepatic cholangiography


split liver transplantation

Complications of the biliary tract affect approximately one-third of patients after liver transplantation (LT) and result in significant patient morbidity and increased patient mortality.[1] The spectrum of biliary complications includes biliary leaks, strictures, choledocholithiasis, cast formation, papillary stenosis, and other less common conditions.[2, 3] An anticipatory approach and a clear understanding of the risk factors for biliary complications following LT can result in the prompt diagnosis and management of these conditions.

Biliary complications are usually diagnosed on the basis of clinical suspicion, laboratory abnormalities, and imaging of the biliary tree. Advances in noninvasive imaging, including computed axial tomography and magnetic resonance imaging with magnetic resonance cholangiopancreatography (MRCP), have greatly facilitated the precise identification of biliary tract complications, and invasive cholangiography, including endoscopic retrograde cholangiopancreatography (ERCP) and percutaneous transhepatic cholangiography (PTC), is thus used in directed therapeutics rather than diagnostic procedures in most circumstances.

Modalities available for the treatment of biliary complications include therapeutic ERCP, percutaneous transhepatic biliary drainage, and surgery, the last being reserved for severe complications or refractory conditions not amenable to less invasive techniques. ERCP allows the endoscopic treatment of complications in the majority of patients and is, therefore, considered the first-line treatment modality in most circumstances. PTC plays an important role in conditions for which ERCP is not successful (eg, a disconnected duct or high-grade stenosis that cannot be traversed at the time of ERCP). Traditionally, endoscopic biliary access was considered impossible after Roux-en-Y hepaticojejunostomy, in a substantial number of patients and thus percutaneous biliary drainage was the only nonsurgical option for such patients. However, significant advances in deep enteroscopy technologies over the last decade, including single-balloon enteroscopy, double-balloon enteroscopy, and spiral overtube–assisted enteroscopy (eg, the Spirus overtube), have increasingly allowed successful endoscopic access to and treatment of previously unapproachable biliary anastomoses in long Roux limbs. There is an ever increasing array of ERCP accessory devices, including biliary cannulation catheters, guide wires, and stents, that enable the successful endoscopic management of biliary complications in the majority of patients, and consequently, fewer patients require more invasive, nonendoscopic approaches such as PTC and surgery. The management of biliary complications, with a focus on newer endoscopic access and management techniques, is discussed in this review.


A clear understanding of the different types of surgical reconstruction during LT is key to the appropriate management of biliary complications. In addition, a thorough understanding of the normal and variant anatomies of the liver segments and their individual and sectoral ducts is a vital and often neglected prerequisite to the successful treatment of problems after LT.

The most common type of LT is performed orthotopically with a full-sized liver from a deceased donor who has experienced brain death. As a result of the widening gap between the number of patients eligible for LT and the number of such livers available for transplantation, various alternatives are increasingly being used to maximize the availability of donor livers for transplantation. These include donation after circulatory death (DCD) transplantation, living donor liver transplantation (LDLT), and split liver transplantation (SLT). The 2 main reconstructions of the biliary system are the choledochocholedochostomy (duct-to-duct anastomosis) and the Roux-en-Y hepaticojejunostomy or choledochojejunostomy. Indications for Roux-en-Y anastomosis include primary sclerosing cholangitis (in which the recipient bile duct is resected as much as possible to minimize the risk of recurrent disease and cholangiocarcinoma in the remaining native duct) and conditions in which there is no suitable recipient bile duct for a duct-to-duct anastomosis (eg, biliary atresia).[4-7] It is important for treating physicians to have a clear understanding of the differences between these variants of anastomoses because complications associated with a choledochocholedochal anastomosis are approached with conventional ERCP, whereas those associated with a hepaticojejunostomy can typically be approached only with deep enteroscopy techniques or percutaneous biliary drainage because of the long Roux limb.

With the LDLT technique, the living donor's right or left lobe or left lateral segment is used for transplantation. The common bile duct of the recipient is anastomosed to the donor's left hepatic duct in left LDLT or to the junction of the donor's right anterior and posterior sectoral ducts in right LDLT. The small caliber of the intrahepatic ducts and variations in the ductal anatomy make ductal anastomoses more challenging in LDLT versus deceased donor liver transplantation (DDLT). The variability of the donor's right anterior and posterior sectoral ducts may necessitate the creation of 2 separate anastomoses to the recipient's common bile duct and cystic duct or, in some instances, to the recipient's common bile duct and a Roux limb (hepaticojejunostomy). In SLT, a deceased donor liver is divided into 2 organs to allow 2 recipients to receive a transplant instead of just 1 recipient. The anastomoses of the right or left lobe are similar to those of living related donors.

Because of the variations in surgical techniques and ductal anatomy, it is vital for endoscopists and other providers to be aware of the type of surgical anastomosis, and if it is possible, every effort should be made to personally review the operative report and cross-sectional imaging studies such as MRCP in order to determine the appropriate approach and treatment plan for the patient.


Patients with biliary tract–related problems after LT may be asymptomatic or may present with jaundice, abdominal pain, or cholangitis. Asymptomatic patients typically present with laboratory abnormalities, and a diagnosis is usually made on the basis of imaging. The first step is to differentiate nonobstructive etiologies of cholestasis such as rejection (acute or chronic), recurrence of primary disease, and medication-related cholestasis from obstructive etiologies.

Imaging modalities employed to evaluate the biliary tree in patients with suspected biliary complications after LT include transabdominal ultrasound (with Doppler), computed tomography (CT) scanning, MRCP, endoscopic ultrasound, PTC, and ERCP (the last using a conventional duodenoscope in patients with duct-to-duct anastomoses and an enteroscopic approach in patients with Roux-en-Y anastomoses). The hepatobiliary iminodiacetic acid scan is a valuable functional imaging test for the evaluation of suspected biliary leaks. Doppler ultrasound may be helpful in evaluating patients for hepatic artery thrombosis because this complication is associated with the pathogenesis of biliary strictures and biliary leaks. In the appropriate setting, angiography may be used to confirm and treat hepatic artery thrombosis or hepatic artery stenosis. CT scanning can be useful in evaluating nonbiliary causes such as abscesses as well as fluid collections associated with bile duct leaks. Conventional CT scanning has a low sensitivity and specificity for directly diagnosing bile duct strictures and leaks. Contrast-enhanced CT with an intravenous dye excreted via the biliary system (eg, iodipamide meglumine) can be used to evaluate the biliary system in detail or to diagnose bile duct leaks; however, the risk of potentially life-threatening allergic reactions is considered to be higher with contrast agents excreted via the biliary system versus conventional contrast agents, and patients are, therefore, usually premedicated. In the United States, there is limited use of this technology in the evaluation of the biliary tract after LT.[8]

MRCP allows the detailed evaluation of the intrahepatic and extrahepatic biliary tree, and the accuracy of MRCP has been found to be comparable to that of ERCP in the evaluation of post-LT biliary complications.[8-12] As mentioned earlier, MRCP is noninvasive and thus avoids the risk of complications associated with invasive contrast–based cholangiography techniques (ERCP and PTC), and it can be performed under circumstances that would be challenging for direct cholangiography (eg, patients with Roux-en-Y anastomoses). MRCP allows the evaluation of the ducts above and below any focal abnormality of the biliary tree (eg, a biliary stricture or leak), and unlike contrast cholangiography with ERCP or PTC, MRCP thus allows the complete imaging of multiple ducts and completely obstructed or disconnected segmental ducts (Fig. 1). MRCP thus provides an important roadmap of the entire biliary system that can be used to plan therapy in advance and during invasive therapeutic procedures to target specific ducts. MRCP is particularly useful in patients with complex hilar or intrahepatic strictures and an anatomy difficult for direct cholangiography. It is the preferred modality for the evaluation of patients with a low or indeterminate suspicion of biliary complications and when endoscopic retrograde cholangiography (ERC) or PTC is not clearly indicated. Conventional MRCP without contrast enhancement allows a static evaluation of the biliary system; however, it does not allow an evaluation for duct leaks. Gadoxetate disodium (Eovist), a gadolinium-based intravenous contrast medium that is excreted via the biliary system, has recently been approved by the Food and Drug Administration. This agent allows dynamic imaging of the biliary system and has been shown to be effective in diagnosing biliary duct leaks and in differentiating extrahepatic fluid collections due to biliary leaks from other etiologies such as abscesses and cysts.[8, 9] It should be noted, however, that the agent has not been specifically approved for this indication. Limitations of contrast-enhanced magnetic resonance imaging with MRCP include the need for intravenous access, the longer duration of the imaging study, and the risk of contrast-induced nephrogenic systemic sclerosis associated with gadolinium-based dyes in patients with impaired renal function. Generally, the limitations of MRCP include its inability to be performed in patients with pacemakers, aneurysm clips, or severe claustrophobia and its poor quality if the patient is unable to remain still. A complete review of the accuracy of MRCP in the diagnosis of the various biliary complications after LT is beyond the scope of this article. MRCP has, however, been found to be as accurate as ERCP in the diagnosis of such complications.[10-12] In a recent study,[13] the authors correlated MRCP findings with other modalities, including ERCP, PTC, and surgery, in 120 patients suspected to have biliary complications after LT. MRCP had a sensitivity, specificity, positive predictive value, and negative predictive value of 98%, 94%, 94%, and 98%, respectively. MRCP may suggest the presence of stenosis at the level of a duct-to-duct anastomosis without an assessment of the severity or significance, and it has a lower sensitivity for sludge and stones < 5 mm in comparison with larger stones.[14] Despite these caveats, MRCP remains the principal technique for the imaging of biliary tract complications after LT.

Figure 1.

MRCP showing a patient after DDLT with an AS (arrow).


Biliary strictures after LT are characterized as anastomotic strictures (ASs) or nonanastomotic strictures (NASs), which are also called ischemic strictures.[15-17] ASs are usually short segmental areas of stenosis involving the choledochojejunostomy or choledochocholedochostomy (Figs. 1 and 2). Anastomotic stenosis of some degree is usually present on imaging such as MRCP, but it is considered clinically significant only if there is evidence of cholestasis or cholangitis. The mean time from LT to the diagnosis of strictures is typically approximately 2 months. ASs occur after 4% to 9% of LT procedures, and they are thought to develop as a result of localized fibrosis due to the operative technique, a postoperative bile leak, or localized ischemia. ASs respond well to endoscopic therapy and are not associated with lower graft or patient survival. The number of endoscopic procedures needed to achieve resolution for ASs is typically lower in patients with early-onset strictures (occurring within 30 days of transplantation) versus patients with late-onset strictures (occurring more than 30 days after transplantation).[3, 18]

Figure 2.

PTC showing a very high-grade AS (arrow) 1 year after DDLT. This stricture could not be traversed initially by ERCP, and 4 combined percutaneous-endoscopic procedures were required to traverse, dilate, stent, and resolve it.

NASs or ischemic strictures have been defined as strictures developing more than 5 mm proximal to the anastomosis (Fig. 3A-F). NAS are thought to be the result of ischemic injury to the duct and have been associated with vascular compromise from hepatic artery thrombosis, a prolonged ischemia time (cold and warm), DCD transplantation, prolonged vasopressor support for the donor, ABO-type incompatibility, primary sclerosing cholangitis, autoimmune hepatitis or hepatitis C infection in the recipient, cytomegalovirus infection, and chronic rejection.[19-26] It can be difficult to separate an AS from a more proximal NAS on the basis of cholangiography alone. True NASs typically are more diffuse and involve the hilum, and they commonly also involve multiple separate obstructions at the level of sectoral or segmental branch ducts. The incidence of NASs has been reported to range from 5% to 15% after conventional DDLT, with a higher incidence (20%-33%) reported for patients with DCD transplants. The management of patients with NASs or ischemic strictures is technically more difficult because of the need for dilation and stenting of multiple hilar and intrahepatic stenoses, and the outcomes are less satisfactory.[27-30] In contrast to patients with ASs, patients with NASs often require multiple procedures and sometimes permanent indwelling stents, with stent exchange scheduled every few months because of a high recurrence rate. Biliary sludge may accumulate repeatedly, with a high incidence of cholangitis, a need for hospitalization, and a risk of secondary biliary cirrhosis, and graft survival is lower. Early ischemic strictures are associated with hepatic artery compromise due to, for example, thrombosis, and patients typically require urgent angiographic intervention to re-establish hepatic artery flow; in severe cases, patients may require retransplantation. Endoscopic therapy is of limited value unless the hepatic artery blood flow can be re-established. Ischemic strictures that develop indolently from either late hepatic artery occlusion or other etiologies are more likely to respond to endoscopic treatment; however, as discussed previously, they are more resistant to treatment and require multiple procedures.

Figure 3.

(A) ERCP with a balloon occlusion cholangiogram showing nonanastomotic hilar ischemic stricturing with diffuse irregularities throughout intrahepatic ducts. (B) Balloon dilation of the right posterior sectoral duct. (C) Balloon dilation of the right anterior sectoral duct. (D) Balloon dilation of the left hepatic duct. (E) Fluoroscopic view of the placement of 3 long and flexible 8.5- to 10-Fr fenestrated biliary stents with multiple side holes into the right anterior, posterior, and left ducts. (F) Endoscopic view of the multiple biliary stents.

Denervation of the ampulla may theoretically lead to the development of a hypertonic sphincter. This may result in papillary stenosis or sphincter of Oddi dysfunction; however, this entity is somewhat controversial and of questionable clinical significance after LT. Sphincter of Oddi dysfunction may cause pain, cholestasis, or pancreatitis without another apparent etiology.

Bile leaks (Figs. 4A,B and 5A-E), bile duct stones, sludge, blood clots, and casts (Fig. 6A,B) may occur after LT.[2, 3, 17, 18, 31, 32] The severity of these complications is variable and is often dependent on the concomitant presence of biliary stricture disease. A thorough discussion of the pathogenesis of these conditions is beyond the scope of this review. Briefly, any condition that can impede the flow of bile is associated with biliary stone disease. Thus, ASs and biliary obstructions lead to stone formation, and conditions associated with biliary stricture formation such as hepatic artery stenosis, ischemia, and rejection are, therefore, also associated with the pathogenesis of stones, sludge, or casts. In general, patients with persistent NASs seldom have a durable response to repeated endoscopic treatment. Such patients often present with recurrent intrahepatic biliary stones and sludge. Biliary cast syndrome refers to the presence of multiple hard pigmented stone casts within the bile ducts. The pathophysiology of biliary cast syndrome is not known; however, this condition is also associated with ischemia and strictures. Several mechanisms such as acute cellular rejection, ischemia, infection, and biliary obstruction have been proposed to play an etiological role in the development of biliary casts. The reported incidence of biliary cast syndrome is 2.5% to 18%.[15, 31] Casts are associated with increases in morbidity, graft failure, the need for retransplantation, and mortality.

Figure 4.

(A) ERCP showing an anastomotic bile leak and diffuse ischemic abnormalities of the intrahepatic ducts after transplantation. (B) Fluoroscopic view of 2 flexible 10-Fr fenestrated biliary stents (16 and 18 cm long) with multiple side holes placed deeply into the right anterior sectoral and left hepatic ducts.

Figure 5.

(A) ERCP showing a bile leak with an abscess fistulizing into the stomach (arrows) after right lobe LDLT. (B) ERCP showing a puncture of the leak-related abscess through the gastric wall fistula. A biliary stent had already been placed by ERCP into the right lobe donor duct and also through the leak into the abscess. (C) Balloon extraction of debris beside the second stent placed through the leak into the abscess. (D) Placement of a 10-Fr double pigtail stent from the stomach into the perihepatic abscess and into the bile duct. (E) After the removal of the stents, the anastomotic leak and the stricture resolved.

Figure 6.

(A) ERCP showing an elongated intraductal filling defect above the anastomosis suggestive of a biliary cast. (B) Endoscopic view of the extracted elongated cast.

The clinical presentation of stone-related complications may be similar to that in the nontransplant setting. Patients may be asymptomatic or may present with pain, jaundice with or without pain, fever, sepsis, cholangitis or pancreatitis, or vague, nonspecific symptoms. It is important to have a low threshold for the further evaluation of patients suspected to have these complications because the use of immunosuppressive agents may mask symptoms and lead to rapid deterioration without laboratory abnormalities typically associated with biliary obstruction or cholangitis.

Bile leaks occur in up to 25% of patients after LT.[2, 31] Bile leaks may result from an anastomotic leak, an anastomotic ischemic injury due to hepatic artery ischemia, the T-tube insertion site into a biliary anastomosis, or the cut surface of the liver after LDLT or SLT. Leaks may result in fluid collections and abscesses and may be associated with disconnected or strictured ducts (Fig. 5A-E). Early recognition of bile leaks is important. The presence of bile in the output of operative drains is an early indicator of a bile leak, which can often be confirmed with a radionucleotide scan. CT or MRCP may reveal a fluid collection outside the biliary tree. As mentioned earlier, contrast-enhanced MRCP and CT can potentially be used to confirm the presence of a leak and, furthermore, localize the level of the leak.

Similar to strictures, bile leaks can be classified as early (occurring within the first month after transplantation) or late (occurring more than 1 month after transplantation). Early leaks are typically associated with anastomotic leaks, ischemic injury, or leakage around a T-tube. Late leaks may occur at the time of removal of a T-tube or above a biliary stricture, particularly after liver biopsy in a high-pressure duct.


Endoscopic therapy with ERCP is the primary approach for the management of biliary complications after LT, with PTC reserved for ERCP failures and disconnected ducts that cannot be traversed by a retrograde approach.[27-50] Recent developments in deep enteroscopy techniques have led to successful endoscopic treatment in many patients with Roux-en-Y hepaticojejunostomy,[34] with PTC reserved for the salvage of failures or rendezvous access in combination with ERCP. The spectrum of therapies employed at the time of ERCP includes biliary sphincterotomy, balloon dilation of the sphincter, balloon dilation of strictures, basket and balloon extraction of stones, sludge, and casts (Fig. 6A,B), and the placement of 1 or more biliary stents. Advanced endoscopic techniques using direct cholangioscopy enable direct visualization and directed therapy within the bile duct.[51-56] Direct visualization may help to differentiate biliary casts from strictures and aid in the directed acquisition of tissue for sampling purposes. Cholangioscopy can also be used for the removal of stones and sludge and for the documentation of complete ductal clearance. It can also be used to perform intraductal lithotripsy for large or difficult stones, although these are rare after LT. Surgery should be performed only as a last resort if endoscopic or percutaneous therapy is not possible or is unsuccessful in the treatment of resistant strictures with resulting progression of liver disease.

Because patients may present with minimal symptoms or abnormalities of laboratory tests only, it is important for clinicians to have a high index of suspicion for the development of biliary complications in order to allow early recognition and early endoscopic intervention and thus minimize the risk of subsequent problems in patients with biliary complications. Endoscopic management is generally very effective and has a low incidence of procedure-related complications; this makes it the preferred approach whenever it is feasible.

Bile Duct Sludge, Blood Clots, Stones, and Casts

Bile duct stone removal is accomplished via biliary sphincterotomy and stone extraction with various techniques, usually at a single ERC session. In patients with severe coagulopathy or thrombocytopenia, sphincterotomy may be relatively contraindicated, and there may be a role for balloon dilation of the intact sphincter. Pancreatic stent placement is generally recommended in this setting to reduce the risk of post-ERCP pancreatitis.[32] The removal of large bile duct stones and casts can be achieved through the combination of biliary sphincterotomy with large balloon dilation of the biliary orifice. This technique is increasingly being employed with improved efficacy and minimization of complications in the nontransplant setting.[33] ERCP may be technically more challenging in the setting of proximally located stones. In such patients, direct cholangioscopy is increasingly playing a useful role in diagnosing stones, aiding in stone removal, and ensuring complete duct clearance, all of which can be of limited efficacy with standard ERCP techniques guided only by contrast cholangiography. Cholangioscopy with fiber-optic (eg, SpyGlass fiber-optic single-operator cholangioscopy, Boston Scientific, Natick, MA) or video choledochoscopes can be used to employ advanced intraductal techniques such as electrohydraulic lithotripsy or a Holmium laser to achieve stone removal. In patients with favorable ductal and ampullary anatomy, direct choledochoscopy can be performed by advancing an ultra-slim (pediatric) forward viewing video endoscope directly in to the bile duct. The potential advantages of this approach include the fact that these endoscopes are already present in most endoscopy units, have very high image resolution, and, because of their larger working channels, have an enhanced ability to perform therapeutic maneuvers such as stone removal under direct visualization via their working channels. Their limitations include less flexibility in comparison with dedicated cholangioscopes and an inability to evaluate small-caliber ducts because of the size of the scopes. In our experience, endoscopic approaches for cast extraction are similar to those used for stone disease and include adequate sphincterotomy followed by large balloon dilation of the orifice and cast removal with stone retrieval devices such as balloon catheters and baskets. Deep enteroscopy techniques have allowed endoscopic access to biliary-enteric anastomoses in patients with long Roux-en-Y limbs. Balloon-assisted enteroscopy has been used to successfully remove a biliary cast in a patient with a Roux-en-Y anastomosis with an endoscopic approach.[34]

Papillary Stenosis and Sphincter of Oddi Dysfunction

Patients with suspected papillary stenosis, sphincter of Oddi dysfunction, or recurrent acute pancreatitis typically do not have an obstructive etiology found with noninvasive imaging techniques and are further evaluated with ERCP, ideally with manometry. Patients are typically treated with biliary sphincterotomy. Pancreatic sphincterotomy may be indicated for patients with recurrent pancreatitis and for those with an elevated pancreatic sphincter basal pressure. Because the risk of pancreatitis after ERCP is relatively high in these patients, temporary prophylactic pancreatic stents are recommended in most circumstances.[32]

Bile Leaks

In most cases, bile leaks can be managed nonoperatively. ERCP is indicated for patients whose diagnosis is established (eg, the presence of bile in an abdominal fluid collection or the diagnosis of a leak by noninvasive imaging is suspected) but not confirmed by other imaging modalities alone.[35] Biliary leaks are typically treated with biliary sphincterotomy and biliary stent placement. If there is an associated biliary stricture, the stricture can be carefully dilated, and 1 stent or multiple stents can be placed beyond both the stricture and the leak (Fig. 4A,B). Biliary leaks are typically treated with transpapillary stents for 1 to 3 months in order to ensure adequate healing of the leaks. In the setting of an associated stricture, ERCP is typically repeated 4 weeks after the initial ERCP procedure to evaluate the stricture and leak and to consider the placement of additional stents across the stricture. Endoscopic therapy results in the resolution of biliary leaks in more than 85% of patients.[31, 35-38] As discussed later, covered, self-expanding metal stents have been used in the treatment of biliary leaks considered to be refractory to conventional treatment. Despite improvements in endoscopic techniques, stents, and deep enteroscopy techniques, endoscopic therapy may not be successful or feasible in certain situations. Large anastomotic leaks (eg, in the setting of hepatic artery compromise) may not heal with endoscopic therapy. Similarly, leaks from the perforation or compromise of a Roux-en-Y anastomosis may require surgery because of an inability to reach the anastomosis or for definitive treatment.

In addition to the bridging and closure of the leak, drainage of the fluid collection or abscess is often required. This is usually done via percutaneous routes by interventional radiology, but for retrogastric collections that might be difficult to reach, this can also be done in selected circumstances by endoscopic ultrasound–guided transgastric drainage, just as for pseudocysts and pancreatic abscesses (Fig. 5A-E).

Biliary Strictures

Endoscopic therapy for biliary strictures generally involves biliary sphincterotomy followed by balloon dilatation and endoscopic stent placement. The use of balloon dilation alone in the treatment of post-LT strictures has been described in some reports; however, there is a general consensus that dilation alone is less effective than balloon dilation with biliary stent placement.[2, 15, 36, 39-42] Balloon dilation is generally avoided for very early strictures and for strictures in the setting of an anastomotic leak in order to avoid disruption of the biliary anastomosis.

The usual approach for the treatment of ASs by ERCP involves the temporary placement of multiple large-bore (usually 10-Fr) plastic stents for a period of up to 1 year. The stents are exchanged every 3 months to minimize the risk of stent occlusion and cholangitis. Endoscopic stenting is possible in the majority of cases. The upstream duct is often narrowed in the setting of ischemic strictures (eg, because of hepatic artery occlusion). In such settings, the narrow lumen of the hilar and intrahepatic ducts may limit the caliber of the biliary stent or stents placed. A more aggressive strategy using the placement of the maximum possible number of stents has been described; it involves up to 3 to 4 stents up to 10 Fr in diameter, with this depending on the location of the stricture and the size of the upstream ducts (Fig. 3E,F). The stricture is then evaluated every 3 to 5 months with stent exchange and, if possible, the placement of additional stents until the complete morphological resolution of the stricture. With this approach, long-term stricture resolution has been reported in more than 90% of patients with ASs.[43] In a recent study,[44] a similar approach was used to place up to 5 to 8 stents after initial balloon dilation, and ERCP was not repeated for up to 1 year unless there was clinical suspicion of stent occlusion. The median number of ERCP procedures per patient was 3 over a median duration of 15 months. With this approach, complete resolution was achieved in 94% of patients at a median follow-up of 11 months. The incidence of cholangitis due to stent occlusion was low despite a longer duration between ERCP procedures, and this was attributed to adequate biliary drainage due to the space between the stents. One of the major advantages of endoscopic therapy over a percutaneous approach is the ability to place multiple indwelling large-caliber stents to remold the stricture. In contrast, the percutaneous approach for stent placement is limited by the size of the percutaneous catheter exiting the patient's skin. The placement of additional catheters leads to significant discomfort and increased risk of morbidity from complications such as bleeding, bile leaks, and skin infections.

The endoscopic treatment of NASs or ischemic strictures, which usually involve the hilum and intrahepatic ducts, is significantly more challenging than the treatment of ASs. The small caliber of upstream and intrahepatic ducts may limit the caliber and number of stents placed. Some retrospective studies suggest improved efficacy for NASs (91% versus 31%) and less cholangitis (12% versus 25%) with the use of balloon dilation alone without stenting versus balloon dilation with stenting; this is analogous to the treatment of primary sclerosing cholangitis. However, in our experience, these strictures are often highly stenotic and thus require stenting to maintain duct patency, so such findings may represent case selection bias due to the retrospective, nonrandomized design of the study.[41] Another potential challenge of hilar strictures with the placement of conventional Amsterdam-type polyethylene stents is that these stents are relatively rigid, do not have side holes, and thus have the potential to occlude secondary branch ducts. Outward migration is very common with such stents and may result in cholangitis and even duodenal or intestinal perforation. Such limitations may partially explain the poor results reported with the use of conventional plastic stents for NASs.

Our approach to hilar and intrahepatic strictures has been to combine balloon dilation with the placement of multiple specialized biliary stents. We place large-caliber (8.5- to 10-Fr), long (12- to 20-cm), highly flexible fenestrated stents with multiple side holes (Johlin pancreatic wedge stents, Cook Endoscopy) deep into the intrahepatic ducts (Figs. 3E,F and 4B). The upstream wedge end is used to anchor the stents in the intrahepatic ducts. The stents allow adequate bile drainage through multiple side holes and interstent space (if 2 or more are placed) and, because of their flexibility, tend to conform to the tortuous contours (especially of the left hepatic ducts) and not migrate.

Biliary complications after LDLT occur frequently and may affect both the donor and the recipient.[45-48] According to a report from the multicenter Adult-to-Adult Living Donor Liver Transplantation Cohort Study in the United States,[45] 32% of recipients had a bile leak and 17% had a biliary stricture within the first year after transplantation. In the donor population, the incidence of bile leaks was 9.2%, and the incidence of biliary strictures was 0.5%.[46] The ductal anastomoses in LDLT are technically more difficult because of the smaller caliber of the ducts, and this may predispose patients to biliary complications. In addition, ischemic complications may also play a role in the development of biliary strictures. Endoscopic management in LDLT patients may be more challenging because of the complexity of duct-to-duct reconstruction. Because of the small caliber of the donor duct, patients often require multiple sessions of endoscopic therapy with the placement of multiple and sometimes smaller caliber (7- or 8.5-Fr) stents. Treatment for up to a year or more is sometimes required to achieve sustained long-term resolution of the strictures after stent removal. In some cases, the strictures may be refractory to endoscopic therapy and may require long-term stent exchange or surgical revision of the duct-to-duct anastomosis to a hepaticojejunal anastomosis.[49, 50]

Overall, the reported success rate of endoscopic therapy for biliary strictures after orthotopic LT is highly variable and ranges from 31% to 100%.[2, 15, 27-31, 36, 39-42] The success rate for some NASs and LDLT-related strictures is at the lower end of this spectrum. Patients with persistent strictures that do not resolve despite multiple attempts at stenting may be managed with scheduled stent changes, PTC, or surgical revision of the anastomosis. Ultimately, some patients (especially those who have complex ischemic intrahepatic strictures) may require retransplantation.

Self-expanding metallic biliary stents have a larger diameter (30 Fr for a 10-mm stent) and a longer duration of patency, and they have been shown to be superior to plastic stents in the setting of a malignant biliary obstruction.[57] These stents generally cannot be removed and, over time, have a high likelihood of stent occlusion due to reactive hyperplasia. This is often further complicated by stone formation above the stents. The placement of uncovered metal stents in the setting of benign biliary diseases is associated with poor outcomes, and as a result, these stents are typically contraindicated in the setting of biliary complications after LT.[58-60]

In contrast, fully covered, self-expanding metal stents such as the Viabil stent (ConMed Corp., Utica, NY), the fully covered Wallflex stent (Boston Scientific), and the fully covered Niti-S ComVi stent (Taewoong Medical, Korea) and partially covered self-expanding metal stents such as the covered Wallflex or Wallstent stent (Boston Scientific) can almost always be removed endoscopically because the outer coating of the stents prevents tissue ingrowth into the stent mesh (Fig. 7A-D). Over the last few years, the use of covered, self-expanding metal stents has been reported for LT strictures and leaks considered refractory to conventional treatment.[61-64] One of the main limitations of fully covered metal stents is their tendency to migrate out of the bile duct. The migration rate was 37.5% (6 of 16 patients) in the largest series of patients treated with a covered self-expanding metal stent for a refractory biliary stricture (n = 11) or leak (n = 5) after LT.[62] At the 2-month follow-up, the reported success rate was 87.5% (14 patients), and this included all 6 patients with a stricture in whom the stent had migrated. Notably, before the placement of the covered, self-expanding metal stent, no more than 2 plastic stents had been placed in the bile duct for the treatment of the biliary stricture. The short time in which the strictures resolved (despite migration in 6 patients) and the low number of plastic stents used raise the possibility of these strictures being responsive to an aggressive approach with the placement of multiple plastic stents. Our experience with fully covered metal stents (primarily covered Wallflex stents from Boston Scientific) has been limited by a high migration rate. Stent migration can occur inward above the stricture or downstream with the passage of the stent into the small bowel. One approach that we have employed with some success is to perform an ample sphincterotomy and place the covered metallic stent completely within the bile duct with the distal end of the stent just above the major papilla. The stent can subsequently be removed with a dilating balloon or under direct cholangioscopic visualization. Another approach is to use a fully covered metallic stent with external anchors (Viabil, ConMed). When they are placed above the biliary bifurcation, covered metal biliary stents occlude secondary branch ducts. These stents, therefore, cannot be used for complications involving the hilum or intrahepatic or living related donor transplants. If there is an indication for a self-expanding metallic stent, it is our opinion that only fully or partially covered stents should be used. Further experience and data are needed to better evaluate the usefulness of metal stents for complex posttransplant biliary strictures.

Figure 7.

(A) ERCP showing a high-grade AS (arrows) after DDLT. (B) Placement of a fully intraductal, fully covered Wallflex stent (10 mm × 6 cm) through the stricture. A persistent waist can be seen at the level of the stricture. (C) Complete resolution of the stricture with no waist (arrows) after the removal of the stent. Therapy was completed after 1 ERCP procedure with the placement of the fully covered metallic stent for 6 weeks (panels A-C courtesy of Rajeev Attam, M.D.). (D) Endoscopic view of the removal of a fully covered, self-expanding, metallic Wallflex stent.

Advances in endoscopic techniques are resulting in less reliance on PTC for the management of biliary complications. PTC is considered the primary approach in patients with biliary strictures with endoscopically difficult access to the duct (eg, Roux-en-Y hepaticojejunal anastomoses) and in patients with disconnected bile ducts with a complete breakdown or obstruction, which makes endoscopic retrograde access difficult (Fig. 2). In the past, PTC was proposed to be more advantageous than an endoscopic approach for difficult bile duct strictures such as those separately involving right and left hepatic ducts.[3] The percutaneous approach is, however, associated with a significant risk of complications, including bile leaks, intrahepatic or intraperitoneal hemorrhaging, and infections. Beyond this, the percutaneous catheter is often a source of significant patient discomfort and inconvenience. In addition, a nondilated biliary tree may make percutaneous access technically challenging or impossible. In the current era, most difficult hilar obstruction cases can be treated endoscopically with advanced techniques, the use of steerable ERCP cannulas (eg, SwingTip cannula, Olympus EndoTherapy, Tokyo, Japan), multiple guide wires, and aggressive dilation and stenting of multiple strictures. Whenever possible, ERCP offers the advantage of performing the procedure internally without the need for external drains. In addition, the ability to place multiple biliary stents simultaneously allows greater dilation of strictures over a long period of time in comparison with percutaneously placed catheters; for the latter, the extent of dilation is limited by the size of the catheters, and this ultimately results in a shorter period of treatment for successfully treating strictures[65] (Fig. 3A-F). The wider availability of direct choledochoscopy (eg, the SpyGlass Direct Visualization System, Boston Scientific) has made it possible to directly visualize posttransplant biliary strictures for diagnostic evaluation and to direct the video-guided placement of guide wires across difficult-to-traverse strictures.[52-54, 66, 67]

Treatment of Biliary Complications After Roux-en-Y Hepaticojejunostomy and Roux-en-Y Gastric Bypass

The 2 principle types of surgically altered luminal anatomies encountered after LT are (1) Roux-en-Y hepaticojejunostomy (as discussed) and (2) Roux-en-Y gastric bypass, which is being encountered more and more frequently as nonalcoholic fatty liver disease becomes an increasingly common indication for LT. Roux-en-Y gastric bypass results in a common limb of up to 150 cm plus a biliopancreatic limb of up to 150 cm, and this creates a very formidable challenge for endoscopic access to the ampulla or biliary-enteric anastomosis.

Conventional ERCP with duodenoscopes is generally impossible in patients with these types of surgically altered anatomies. As a result, percutaneous transhepatic biliary drainage has been used as the primary approach for the management of posttransplant strictures and other biliary complications in patients with a Roux-en-Y hepaticojejunostomy or gastric bypass. Deep enteroscopy with a pediatric colonoscope can sometimes be used to successfully perform therapeutic ERC in these patients, but it does not allow the anastomosis or papilla to be reached with consistency.[68] However, the field of deep enteroscopic ERCP is rapidly evolving. The introduction of the double-balloon enteroscope (Fujinon Corp., Saitama, Japan) in 2003, followed by the single-balloon enteroscope (Olympus, Tokyo, Japan) and the spiral enteroscopy overtube (Endo-Ease, Spirus Medical, Stoughton, MA), has led to several endoscopic options that permit deep enteroscopy. The long Roux limb is traversed to reach the biliary orifice, and ERC is performed through the forward-viewing enteroscope (Figs. 8A,B and 9A-E). The double-balloon enteroscope systems require a dedicated processor and endoscope system, whereas the spiral enteroscopy overtube is compatible with both single-balloon and double-balloon enteroscopes. Balloon-assisted enteroscopy uses a balloon attached to an overtube to anchor the enteroscope and overtube as the enteroscope is advanced through the small bowel. Spiral enteroscopy makes use of a spiral overtube that is placed over the enteroscope. As the spiral overtube is rotated, the small bowel is pulled onto the overtube, and this advances the enteroscope through the small bowel. Several authors have reported the successful treatment of biliary strictures and other biliary complications with deep enteroscopy techniques in patients with Roux-en-Y biliary anastomoses.[34, 69-72] The largest reported experience of deep enteroscopy in the setting of LT is a series of 25 pediatric patients with hepaticojejunal anastomoses.[69] The anastomosis was reached in 68% (16/25) of the patients overall; however, the success rate for completed procedures improved from 64.0% (16/25) before 2008 to 93.1% (27/29) after 2008. Therapeutic intervention was achieved with 83.7% of the procedures (36/47). Enteroscopic ERC may not be possible in some cases because of an unfavorable surgical anatomy or adhesions preventing access to the Roux limb. Other limitations to the use of these devices include the limited maneuverability of the scope in the region of the biliary anastomosis, the unfavorable orientation of the anastomotic site (which makes cannulation and treatment across the anastomosis difficult), the lack of an elevator device, and the availability of a limited number of small-caliber ERCP accessories and endoprostheses that can be used through these endoscopes. One unique application of deep enteroscopy that is essential and cannot easily be accomplished by any other route is the retrieval of retained surgical stents after LT in patients with a Roux-en-Y gastric bypass. There is no other suitable route because the percutaneous extraction of stents is very traumatic and difficult, and this leaves laparotomy or perhaps laparoscopic enterotomy as the only remaining solution (Fig. 10A,B). Such procedures are not yet widely available outside specialized centers and, as demonstrated in the aforementioned study, require highly specialized expertise for successful completion. A more invasive but direct approach to the major papilla in patients with a Roux-en-Y gastric bypass is creation of a gastrostomy, via a surgical approach (laparoscopic or open) or percutaneously using EUS guidance, followed by ERCP through the gastrostomy port (73). In recent study of patients with a Roux-en-Y gastric bypass, biliary intervention was achieved in all patients using a surgical approach compared to only 58% of patients in whom deep enteroscopy was performed (74) with lower success in patients with a Roux limb greater than 150 cm. Our preferred approach for patients with a Roux-en-Y gastric bypass is perform ERCP through a gastrostomy.

Figure 8.

(A) Use of spiral overtube–assisted deep enteroscopy to reach a hepaticojejunostomy after right lobe LDLT in a patient whose right posterior duct was separately anastomosed to a long Roux-en-Y limb. (B) Cholangiogram performed through a spiral overtube–assisted deep enteroscope and showing normal right posterior ducts without stenosis of the anastomosis.

Figure 9.

(A) ERCP in a patient with a stenotic hepaticojejunostomy after whole DDLT showing cannulation and balloon dilation of a stenotic biliary-enteric anastomosis. (B) Enteroscopic view of the highly stenosed hepaticojejunostomy at the time of ERCP performed via spiral overtube–assisted deep enteroscopy up a Roux-en-Y limb (upper right). In the upper right, a pinhole orifice to the stenosed anastomosis can be seen. (C) Guide wire passed through the stenosed hepaticojejunostomy. (D) Balloon dilation of the stenotic hepaticojejunostomy. (E) Widely patulous hepaticojejunostomy after balloon dilation.

Figure 10.

(A) Spiral overtube–assisted deep enteroscopy up the biliary-pancreatic limb in a patient with a retained surgical biliary stent 1 year after LT and a prior Roux-en-Y gastric bypass for morbid obesity. (B) Endoscopic retrieval of the retained surgical biliary stent from the patient shown in Figure 9A.

At many centers, PTC still remains the most commonly used treatment for biliary complications such as strictures in patients with long-limb biliary-enteric anastomoses.[3, 31] Limitations to the use of PTC include discomfort to patients (especially patients with large transhepatic catheters) and a higher risk of complications such as bleeding and bile leaks (especially in patients with coagulopathy or ascites). Percutaneous transhepatic biliary drainage can sometimes be used with a rendezvous approach to internalize external drains in patients whose biliary anastomosis can be reached endoscopically after the initial percutaneous drainage has been established. Another novel and useful application of combined techniques by endoscopists and interventional radiologists includes direct percutaneous transhepatic cholangioscopy through a previously established percutaneous tract. The technique involves the use of small-caliber forward-viewing endoscopes after the dilation of a percutaneous tract and the placement of a peel-away sheath by interventional radiology. This technique can be performed through the left or right lobe and allows direct digital video cholangioscopy and removal of intrahepatic bile duct stones, dilation of strictures, and any other interventional maneuvers (Fig. 11A,B).

Figure 11.

(A) Direct percutaneous transhepatic cholangioscopy with an ultraslim forward-viewing endoscope advanced through the right lobe of the liver via the percutaneous tract after dilation and the placement of a peel-away sheath. (B) Endoscopic percutaneous cholangioscopy image showing bifurcation of the intrahepatic ducts.


Biliary tract complications are an important source of morbidity after LT. A team approach involving practitioners from multiple disciplines, including transplant hepatology and surgery, interventional endoscopy, and interventional radiology, results in an appropriate evaluation, the interpretation of diagnostic tests, and the formulation of management strategies. In most cases, biliary complications can be approached and successfully treated with a nonoperative approach, so operative reintervention or retransplantation is rarely required. Biliary tract complications can mostly be treated via endoscopic techniques, but they present a particular challenge and require advanced expertise. Therapy now is mostly focused on ERCP and includes sphincterotomy, dilation, stone extraction techniques, stricture dilation, the placement of multiple and often intrahepatic stents, and sometimes direct cholangioscopy when it is necessary to identify filling defects and ensure the clearance of complex stones. Endoscopists have an ever increasing array of ERCP accessory devices, including catheters, wires, and stents, at their disposal to successfully treat biliary complications. Recent developments in deep enteroscopy techniques now allow endoscopic therapy of complications after Roux-en-Y hepaticojejunostomy in many cases. Percutaneous transhepatic biliary drainage is largely reserved for severely disconnected or strictured ducts that cannot be traversed by ERCP and for Roux-en-Y anastomoses when enteroscopy is not feasible or available. Removeable fully covered metal biliary stents may be used to treat biliary strictures and leaks but their effectiveness needs to be defined further prior to widespread use especially given the higher cost of these stents.