Biliary strictures following liver transplantation: Past, present and preventive strategies



Biliary complications are still the major source of morbidity for liver transplant recipients. The reported incidence of biliary strictures is 5%-15% after deceased donor liver transplantation and 28%-32% after right-lobe live donor surgery. Presentation is usually within the first year, but the incidence is known to increase with longer follow-up. The anastomotic variant is due to technical factors, whereas the nonanastomotic form is due to immunological and ischemic events, which later may lead to graft loss. Endoscopic management of anastomotic strictures achieves a success rate of 70%-100%; it drops to 50%-75% for nonanastomotic strictures with a higher recurrence rate. Results of endoscopic maneuvers are disappointing for biliary strictures after live donor liver transplantation, and the success rate is 60%-75% for anastomotic strictures and 25%-33% for the nonanastomotic variant. Preventive strategies in the cadaveric donor include the standardization of the type of anastomosis and maintenance of a vascularized ductal stump. In right-lobe live donor livers, donor liver duct harvesting also involves a major risk. The concept of high hilar intrahepatic Glissonian dissection, dissecting the artery and the duct as one unit, use of microsurgical techniques for smaller ducts, use of ductoplasty, and flexibility in the performance of double ductal anastomosis are the critical components of the preventive strategies in the recipient. In the case of live donors, judicious use of intraoperative cholangiograms, minimal dissection of the hilar plate, and perpendicular transection of the duct constitute the underlying principals for obtaining a vascularized duct. Liver Transpl 14:759–769, 2008. © 2008 AASLD.


Since the early experiences of the 1960s, liver transplant surgery has evolved over the decades and is now the standard of care in patients with end-stage liver disease. Although there has been consistent improvement in the overall survival rates for transplant recipients, the Achilles' heel is still biliary tract complications.1, 2


AS, anastomotic strictures; DDLT, deceased donor liver transplantation; ERCP, endoscopic retrograde cholangiopancreatography; HAT, hepatic artery thrombosis; LDLT, live donor liver transplantation; NAS, nonanastomotic strictures; OLT, orthotopic liver transplantation; PTC, percutaneous transhepatic cholangiography.

Despite great improvements in the surgical techniques and standardization of the method of biliary reconstruction, the biliary tract is still the most common site for postoperative complications. The importance of this condition lies in the fact that it can be a serious source of morbidity and sometimes mortality.3, 4

These complications not only affect graft survival but also have a major impact on the quality of life for a liver allograft recipient, as they entail frequent readmission, reoperation, hospital stays, and escalating costs and add to the emotional trauma that patients suffer.

The conventional management of these conditions in the past was mainly surgical; however, with the rapid progress in the field of endoscopic surgery, the majority of these complications are now amenable to endoscopic interventions, with the role of surgery mainly confined to treatment failures as a second-line backup option.



The incidence of biliary tract complication after orthotopic liver transplantation (OLT) varies from 11% to 25%2, 5–19 and includes leaks, strictures, casts, sludge, stones, and sphincter of Oddi dysfunction. These high rates may signify an inherently sensitive nature of the biliary epithelium to ischemic damage in comparison with hepatocytes and vascular endothelium.20 Biliary strictures and bile leaks are the most common biliary complications following liver transplantation, although a decreasing trend has been noted in recent years.13, 19, 21, 22 The reported incidence of biliary strictures is 5%-15%4, 5, 19, 21, 23 after deceased donor liver transplantation (DDLT) and 28%-32% after right-lobe live donor liver transplantation (LDLT).24–28


Biliary strictures after liver transplantation have been classified as anastomotic strictures (AS) and nonanastomotic strictures (NAS).29, 30 The 2 types of strictures cannot be compared as they have inherent differences in their pathology, time to presentation, treatment, and response to treatment.

In the early 1990s, NAS were classified as NAS caused by hepatic artery thrombosis (HAT)31 and NAS caused by a non-HAT etiology, which were labeled as ischemic-type biliary lesions.31–34 The underlying pathologic process appears to be fibrosis following an injury to the biliary epithelium

NAS are heterogeneous entities, and on the basis of the etiology, Moench et al.35 proposed a classification dividing them into NAS secondary to macroangiopathy (HAT) and NAS secondary to microangiopathy (preservation injury, prolonged cold and warm ischemia times, donation after cardiac death, and prolonged use of vasopressors in the donor) and immunogenicity (chronic rejection, ABO incompatibility, autoimmune hepatitis, and primary sclerosing cholangitis; Figs. 1 and 2).

Figure 1.

Short and localized anastomotic stricture.

Figure 2.

Endoscopic retrograde cholangiopancreatography showing a long nonanastomotic stricture extending proximally from the site of the anastomosis.

NAS are usually multiple, longer in length, and located in the intrahepatic ducts and/or in the donor duct proximal to the site of anastomosis. Their incidence has been reported to vary widely from 1% to 19%,19 with most modern series suggesting it to be between 5% and 15%.13, 23, 30, 36 AS, on the other hand, are isolated, are localized to the site of the anastomosis, and are short in length. Their reported incidence in the modern literature is 4%-9%.5, 9, 15, 19, 37–43

Risk Factors

Deceased Donor Liver Transplantation (DDLT)

The risk factors for NAS include HAT, chronic ductopenic rejection,12, 44, 45 ABO incompatibility,46 primary sclerosing cholangitis as the primary pathology causing recurrent or ischemic strictures,47–49 donation after cardiac death donors,50 use of prolonged vasopressors in the donors,48, 51 older age of the donor,35, 42 preservative injury,52, 53 and prolonged cold and warm ischemia times.30, 31 In contrast to AS, NAS are known to cause graft loss.30

The underlying basis for AS includes ischemia or fibrosis following a suboptimal surgical technique or a bile leak in the postoperative period.11, 54, 55

Although the strictures can present at any time after the surgery, the mean interval at the time of presentation is 5-8 months after OLT,11, 19, 21, 23, 56 and the majority present within 1 year, but recent studies suggest that their prevalence continue to increase with the time after transplantation.55 NAS present earlier than AS, with the mean time to stricture development being 3.3-5.9 months.23, 30 Buis et al.57 further reported that NAS secondary to ischemic causes presented within 1 year of the transplant, whereas the occurrence after 1 year was more often related to immunological causes as the risk factors.

Living Donor Liver Transplantation (LDLT)

LDLT presents a unique problem as it includes not only the usual factors that are in play for DDLT but additional factors that are unique to liver donor allografts. The presence of bile leaks has emerged as one of the most important factors in addition to technical factors in the causation of these stricture55, 58, 59; the underlying process may be related to the local effect of the bile, which induces inflammation and subsequent fibrosis, or it may be a surrogate marker of poor vascularity in cases in which the bile leaks do not originate from cystic ducts. Hwang et al.60 demonstrated the small size of the duct (<4 mm) when used for duct-to-duct anastomosis as a risk factor for the development of strictures; however, the results have not been corroborated by other studies.61, 62 The presence of multiple bile duct anastomoses has not been conclusively associated with bile duct strictures as some studies have found either no relation or a statistically insignificant relationship.58, 62, 63 However, given the fact that these multiple duct anastomoses have been consistently found to have an association with bile leaks,63 which in turn are uniformly associated with biliary strictures, we believe that these entities are interlinked, and a large number of cases may be required to prove the statistical significance.

A recent study found an elderly donor and the presence of bile leaks in the postoperative period to be the only risk factors statistically significant in multivariate analysis.59 The mechanism of the development of strictures in elderly donors is not clear but may be related to the presence of degenerative age-related changes in the microcirculation of the biliary tract.


The presentation is usually suspected in the presence of cholestatic liver tests, although some patients may present with pruritus or cholangitis and abdominal pain;39, 64, 65 the biopsy findings rarely may suggest the presence of an obstructive pathology.66

A recent report of 15 patients highlighted the use of serum bilirubin > 1.5 mg/dL as a better indirect marker of biliary stasis in LDLT than alkaline phosphatase, which is overly sensitive.67


The first step in establishing the diagnosis is ultrasound of the abdomen, which in addition to displaying the vascular patency carries a high positive predictive value in the presence of a dilated duct. However, its sensitivity is too low at 38%-66% to be of any meaningful value in the decision-making process39, 68–70 as the absence of dilatation has been noted to be an unreliable indicator of adequate biliary drainage.71–73 In addition, the size of the duct has not been found to be a reliable indicator in following up these patients or in accessing the response to the treatment. Furthermore, there is a significant lack of correlation between the ducal dilatation on the ultrasound and the cholangiographic and clinical picture. It is unclear why the donor bile ducts do not respond to the distal obstruction by displaying the same degree of proportional dilatation as the nontransplant liver.67 Whether it is the denervation, ischemic-reperfusion injury, or preservation injury that results in subclinical damage to the overly sensitive biliary epithelium is still not clear. It is possible that the presence of variable degrees of fibrosis subsequent to the injury sustained at the time of the perioperative period results in less pliable ducts.70

The next logical step is the use of magnetic resonance cholangiopancreatography, which has a sensitivity and specificity close to 90% in establishing the diagnosis of biliary strictures.74–76 The chief disadvantage is the lack of its therapeutic ability. It can be used as the second step after ultrasound in patients for whom the use of diagnostic endoscopic retrograde cholangiopancreatography (ERCP) or percutaneous transhepatic cholangiography (PTC) carries a higher operational risk.

Scintigraphy of the biliary tree using 99-technetium–labeled iminodiacetic acid identifies strictures with 75% sensitivity and 100% specificity,77, 78 but the absence of any therapeutic benefit offsets its noninvasive advantage.

Cholangiography is considered by all to be the gold standard not only in establishing the diagnosis but also in allowing therapeutic intervention in the same setting.79, 80

ERCP has an advantage over PTC and is the first modality of choice as it is not only more physiological but also less invasive. An inherent limitation of ERCP is the problem of access in patients when the mode of reconstruction is either hepaticojejunostomy or choledochojejunostomy. However, newer approaches using the double balloon enteroscopic technique have been made it possible to reach this difficult area.19

The ERCP findings in NAS consist of mucosal abnormalities, narrowing, and prestenotic dilatation, whereas the findings in AS include narrowing at the anastomosis or bifurcation.


Historically, the management of post-OLT biliary strictures consisted of surgical reconstruction in the form of Roux-en-Y hepaticojejunostomy. However, the last decade has seen such tremendous growth in the evolution of endoscopic techniques that they are now considered the treatment of choice for biliary strictures.39, 42, 64, 65

Percutaneous therapy, although it has a success rate of 40%-85%, is still considered the second-line option because of the invasive nature of the procedure and the associated complications of hemorrhage, bile leaks, and significant morbidity. Surgery is now confined to patients for whom endoscopic modalities have failed and is considered to be the last-line back-up option.

DDLT: Anastomotic Stricture (AS)

The conventional method of endoscopic treatment consists of identification of the mouth of the stricture followed by cannulation by the guidewire, balloon dilatation of the stricture, and subsequent placement of plastic stents. The stents are generally replaced by larger stents every 3 months41, 64, 65 to prevent the complication of clogging, cholangitis, or stone formation. Dual or multiple stents, by providing greater dilatation, have shown better results than single stents.23, 65 The treatment is usually completed in 1 year with an average of 3-4 stent exchange sessions.12, 41, 43, 56, 64

It has been conclusively seen that the use of stents in addition to balloon dilatation achieves higher success rates and less recurrence in comparison with balloon dilatation alone.23, 64, 65 The overall success rate of endoscopic treatment for AS biliary strictures is in the range of 70%-100%.11, 23, 38, 39, 41, 43, 56, 64, 81

Factors reported but not widely corroborated for poor outcome and higher recurrence rates following the ERCP treatment of AS include delayed presentation post-transplant (>6 months) and the presence of tight strictures.11, 19, 55 The risk factors for the recurrence of strictures after stent removal include late initial presentation, presence of bile leaks, and use of T tubes.81

When AS are treated appropriately, the long-term results in terms of patient and graft survival are equivalent to those for matched controls without AS.12, 39, 41, 56

The management for AS complicating a hepaticojejunostomy includes the use of percutaneous transhepatic drainage that achieves success rates of 50%-75%.82–85 The role of surgical revision is confined to endoscopic failures, and the long-term results are good, with no effect on patient or graft survival.7, 19, 54, 86–88

DDLT: Non-Anastomotic Stricture (NAS)

NAS are difficult to treat, and any generalized treatment recommendations are difficult to make. NAS secondary to early HAT usually require urgent revascularization or retransplantation, whereas late HAT-causing NAS can be salvaged by endoscopic means. As a group, they require more endoscopic procedures, balloon dilatation, stents, and a longer duration of treatment.23, 41 Success depends on the severity of the strictures, the number of strictures, and the location of the strictures. Intrahepatic strictures not only are difficult to dilate but also respond poorly to endoscopic treatment.

NAS are more resistant to endoscopic treatment, the reported success rate being 50%-75%.12, 23, 41, 43, 56 An important result of inadequate treatment of these strictures is the development of repeated episodes of cholangitis and biliary cirrhosis or atrophy of the involved lobe. Ultimately, up to 50% of the patients with NAS either die in need of a liver or require retransplantation secondary to graft loss. The mortality reported in the literature varies considerably.30, 34, 46, 89

ERCP with balloon dilatation and stenting now has an established role in the multidisciplinary approach to the management of biliary strictures following OLT.

An analysis of the technical reasons for the failure of therapeutic ERCP shows that the presence of a stricture that is too tight to allow access to the central bile duct system is the commonest cause of ERCP failure in AS, NAS, and biliary strictures after LDLT.55, 57, 62 A stricture that is too tight may preclude successful cannulation by the guidewire as the procedure is done blindly in conventional ERCP with repeated attempts to force the guidewire through the stricture.

Once a stricture is cannulated with a guidewire, the reported success for this procedure is about 80%-90%.11, 39, 41–43, 55, 56, 64 These figures highlight the importance of cannulation of the stricture as an important and critical event in deciding the success or failure of the endoscopic treatment.

We recently used the Spyglass direct visualization system as the guidance system for the passage of the guidewire through a tight stricture where conventional ERCP failed (Fig. 3). The system, by providing direct visual access to the biliary tract for guidance, can be an important assisting device in this subset of patients. Not only did it facilitate the cannulation of the guidewire through a very tight stricture, but at the same time the procedure was done fairly quickly without us resorting to frequent hit and miss attempts to push the guidewire through the stricture.

Figure 3.

Spyglass view of the stricture.

Endoscopy in LDLT Biliary Strictures: AS

Although the therapeutic value of ERCP and PTC in the management of biliary strictures in DDLT is well established, their role in biliary strictures in LDLT is still under evaluation. The results of endoscopy intervention for biliary strictures in DDLT cannot be extrapolated to right-lobe LDLT as there are inherent differences between the two. The presence of multiple ductal anastomosis, smaller size, peripheral location, and increased risk of devascularization make these AS unique. So far, only 6 published trials have evaluated the efficacy of endoscopic approaches in AS secondary to LDLT.21, 27, 59, 62, 90, 91 Endoscopic maneuvers achieve a success rate of 60%-75% for AS, which is significantly less than that for AS in DDLT, for which the success rate is 80%-90%.21, 27, 59, 62, 90, 91 The commonest reason for the failure of endoscopic treatment was the inability to cannulate the stricture.61, 62 As in DDLT, the combination of balloon dilatation and stenting was more effective than either modality alone.62 Placement of the plastic stent can be a difficult proposition in these patients with complex and peripheral anastomosis.91

A subset of AS in LDLT, as defined by the Kyoto group, includes the crane neck deformity, in which the biliary anastomosis is located at a point that is far below the highest portion of the recipient duct. The occurrence of this event is probably related to compensatory hypertrophy of the donor lobe, which results in sharp angulation of the bile duct and hence a difficult endoscopic approach.91 These strictures have been reported to be resistant to endoscopic treatment.61

Endoscopy in LDLT Biliary Strictures: NAS

The results of endoscopic approaches have been particularly disappointing in the context of NAS in LDLT; the average success rate varies from 25% to 33%,61, 91, 92 which is way below the 60% success rate seen with NAS in DDLT.

Preventive Strategies in DDLT Biliary Strictures

The debate over the type of anastomosis has roughly been resolved in favor of duct-to-duct anastomosis. In the absence of a prospective trial comparing the two, data from retrospective studies suggest similar success rates for the two,4, 5, 7, 13, 19, 21, 22, 93, 94 with a slightly higher incidence of bacterial colonization, bleeding, leaks, and their consequences in the hepaticojejunostomy group.10, 29

Currently, duct-to-duct anastomosis is used in 80%-90% of adult liver transplants.30, 43, 56, 88, 95 The benefits include preservation of the sphincter of Oddi, less frequent colonization of the biliary tract, less operative time, fewer anastomoses, and the availability of endoscopic treatment as a back-up option if strictures develop.96 Roux-en-Y hepaticojejunostomy is used only under specific conditions, including gross disparity between the sizes of the ducts and diseased or unavailable ducts, and for revision surgery.29, 97–100

T tubes were widely used in the past because of suggested but unproven benefits101; however, the results of 4 prospective randomized studies failed to show any benefit from the use of T tubes, and at the same time, a higher incidence of complications (bacterial colonization of the biliary tract and bile leaks) and higher costs were noted in the T tube group.72, 102–104 The use of an internal stent, although theoretically eliminating the possible complications associated with T tubes and at the same time offering the potential benefits of a splint, has not shown uniformly reproducible results. Both increased incidence and decreased incidence of complications have been reported with the use of internal stents in comparison with T tubes.93, 105, 106 Their role as adjunct devices in biliary reconstruction is yet to be fully determined.

Technical factors that are usually blamed for strictures include the placement of too many sutures, uneven distribution and tension of the sutures, and too few sutures resulting in inadequate mucosal apposition leading to leaks.

A recent report highlighted the use of spatulation of the donor and recipient bile ducts and the use of continuous sutures as methods to increase the luminal diameter of the anastomosis and provide even distribution of the tension along the suture walls. The AS incidence rate was 2.5% versus 12.5% without spatulation.107 Although attractive, the concept needs further substantiation.

Preventive Strategies in Right-Lobe LDLT Biliary Strictures

The mode of reconstruction of the biliary enteric anastomosis in LDLT has traditionally been Roux-en-Y hepaticojejunostomy, which is used more out of caution than because of evidence. The proposed benefits of the procedure include a vascularized intestinal segment as one of the components of the anastomosis and its ability to counteract the effect of a short or diseased bile duct available for the anastomosis.108, 109 The disadvantages include the creation of an additional anastomosis, opening of the gut, risk of subsequent cholangitis as the protective effects of the sphincter are lost, and difficult endoscopic access if a postoperative stricture develops.24, 109

Growing experience in LDLT and improved surgical skills have resulted in more and more programs using the option of duct-to-duct anastomosis, and it has become the standard of care as it maintains the physiologic bilioenteric continuity.110, 111

Progression in preventive strategies is particularly impressive in right-lobe LDLT because of the high incidence of biliary strictures.24, 36 Various authors have used different strategies with the underlying principle of obtaining a single healthy vascularized right duct. The importance of an adequately vascularized duct cannot be overemphasized.112 The bile duct and its branches are supplied by a rich network of small vessels that form a plexus over the bile duct.113, 114 The feeding arteries to the plexus include the right and left hepatic arteries from the top and the gastroduodenal artery from the bottom; the preferential flow is upward. It makes sense to leave the donor stump as short as possible and the recipient stump as long as possible to maintain adequate blood supply to the ends of the stumps.62, 115 The concept of the high hilar dissection technique in the recipient hepatectomy is a step in this direction.116 However, too short a donor stump also makes the anastomosis technically demanding, and maintaining a fine balance is critical.

Underlying the problem of obtaining a healthy single right duct is the frequent anatomical variation associated with the right ductal system. A single right duct is identified in only 50% of donors, and even in those cases in which a single duct is identified, it frequently subdivides into anterior and posterior parts fairly quickly; this leaves very little margin for error (Figs. 4 and 5).59

Figure 4.

Intraoperative cholangiogram showing the delineation of the ductal anatomy.

Figure 5.

Intraoperative cholangiogram showing the right posterior duct clearly joining the left duct.

Surgical persistence in harvesting a single right duct is also fraught with danger as it frequently shifts the line of the transection, which is too close to the bifurcation and can jeopardize the left duct of the donor. Careful preoperative planning and maintaining flexibility are the critical components of effective prevention. The value of preoperative magnetic resonance cholangiopancreatography and intraoperative cholangiogram in delineating the anatomy and identifying any aberrations cannot be overemphasized.117–119

A recent report has further highlighted the value of a repeat cholangiogram at the completion of the parenchymal transection and immediately before the ductal transection. According to the authors, such a maneuver not only guides doctors in deciding the precise site and angle of the transection but also offers the advantage of improved visualization without the need for direct dissection of the right duct.59

Further surgical techniques that have been hypothesized to reduce the incidence of biliary strictures include minimal dissection of the hilar plate to prevent devascularization of the right duct and perpendicular division of the right duct to prevent ischemia to the lateral edges of the duct (Figs. 6 and 7).59, 108, 109

Figure 6.

Image of a live donor liver surgery showing the isolation of the right hepatic duct without dissection of the hilar plate.

Figure 7.

Intraoperative cholangiogram depicting the level and axis of the transection of the right duct.

The presence of more than one available duct for the anastomosis increase the complexity of the operation and the leak rates, but the effects on the incidence of biliary strictures have not been conclusively demonstrated.62, 63 Ductoplasty and the use of combined ducts for a single anastomosis, although increasing the luminal diameter of the anastomosis, have so far shown conflicting results as far as the leak rates are concerned (Fig. 8).24, 25, 59, 109, 117

Figure 8.

Tight biliary stricture that developed even when the wide lumen of the ductoplasty was part of the hepaticojejunostomy.

In the absence of a sufficient length of the ducts or when the ducts are separated too much or placed at sharp angles, the use of ductoplasty is fraught with danger as it invariably puts tension on the anastomosis and reduces the vascularity.108

Preventive steps in the recipient hepatectomy include high hilar dissection using the intrahepatic Glissonian approach and dividing the ducts at a point that is distal to the second-order branches of the bile duct in order to preserve a long length of the vascularized recipient duct.116 Dissection between the bile duct and the hepatic artery is kept minimal to prevent the disruption of the ductal vascularity.120 Use of synthetic monofilament nonabsorbable sutures has recently become preferred because of their feasibility and supposedly less induction of the tissue reaction.121, 122

Interrupted sutures have shown higher leak rates, whereas continuous sutures are associated with a higher stricture rate; using the benefits of both, the technique includes the continuous posterior wall and interrupted anterior wall.109

The use of microsurgical techniques has shown a clear benefit in hepatic artery anastomosis by reducing the incidence of HAT123, 124; an extrapolation of its use in biliary reconstruction (ducts < 2 mm) has recently been described, and it has shown a reduction in the incidence of biliary strictures.125


Strictures continue to be the commonest biliary complication after liver transplantation, although their incidence has been decreasing steadily over the years. The utilization of various preventive strategies, establishing an early diagnosis, defining when endoscopic failure has occurred, what exactly endoscopic failure is, and how soon doctors should resort to surgery, the effective use of ERCP and PTC, and the appreciation of the adjuvant endoscopic modalities are various strategies that may be examined as we progress further.