Biliary strictures remain the most challenging aspect of adult right lobe living donor liver transplantation (RLDLT). Between 04/2000 and 10/2005, 130 consecutive RLDLTs were performed in our center and followed prospectively. Median follow-up was 23 months (range 3–67) and 1-year graft and patient survival was 85% and 87%, respectively. Overall incidence of biliary leaks (n = 19) or strictures (n = 22) was 32% (41/128) in 33 patients (26%). A duct-to-duct (D-D) or Roux-en-Y (R-Y) anastomosis were performed equally (n = 64 each) with no difference in stricture rate (p = 0.31). The use of ductoplasty increased the number of grafts with a single duct for anastomosis and reduced the biliary complication rate compared to grafts ≥2 ducts (17% vs. 46%; p = 0.02). Independent risk factors for strictures included older donor age and previous history of a bile leak. All strictures were managed nonsurgically initially but four patients ultimately required conversion from D-D to R-Y. Ninety-six percent (123/128) of patients are currently free of any biliary complications. D-D anastomosis is safe after RLDLT and provides access for future endoscopic therapy in cases of leak or stricture. When presented with multiple bile ducts, ductoplasty should be considered to reduce the potential chance of stricture.
Biliary reconstruction remains a technical challenge in right lobe living donor liver transplantation (RLDLT). The reported incidence of biliary complications after RLDLT ranges from 15% to 60% (1–10). There are numerous conflicting reports in the literature concerning the method of performing the donor hepatectomy and biliary reconstruction to reduce stricture rates and to optimize patient and graft survival. Biliary tract problems with right lobe grafts are not necessarily surprising: anatomical variations are common, there is a high incidence of multiple bile ducts, and the blood supply may be poor (11). Biliary complications are associated with increased hospital stay and cost, decreased quality of life, and often repeated therapeutic interventions (1). We report our experience with biliary complications, with a specific focus on biliary strictures in a cohort of 130 consecutive RLDLTs and the strategies for treatment. We have attempted to identify risk factors for strictures, operative techniques in the donor hepatectomy to improve biliary complication rates, and changes in practice over time at a single institution in the Western hemisphere.
Between April 2000 and November 2005, 130 adults underwent RLDLT at Toronto General Hospital, University Health Network, University of Toronto. Two recipients died during the first postoperative week after surgery; therefore, 128 patients were included in this analysis. Research Ethics Board approval was obtained from University Health Network.
The donor evaluation process and our preferences for operative technique for right hepatectomy have been described in detail previously (12–16). The biliary anatomy was evaluated preoperatively with magnetic resonance cholangiography (MRC) (15). Prior to parenchymal transection, an intraoperative cholangiogram was performed with fluoroscopy to verify biliary anatomy and to determine whether there were biliary analomies present that would preclude resection. Biliary drainage of the left lobe by a duct terminating in the right anterior or posterior bile duct was considered a contraindication to donation. Three patients (2%) underwent only laparotomy and were excluded after this drainage pattern was found on intraoperative cholangiogram. A preoperative MRC had not been performed in two of these patients. During the hepatectomy, we avoided direct dissection of the bile duct and surrounding tissue in an attempt to maintain a good blood supply to the hepatic duct. The right bile duct(s) and the hilar plate were divided sharply after 80% of the parenchymal transection had been completed. In the latter half of our experience (years 2003–2005), a repeat cholangiogram was frequently performed immediately before dividing the duct, to help guide the level and angle of duct division (Figure 1). After the right lobe graft had been removed from the donor, it was flushed with one liter of chilled University of Wisconsin solution. During the flush, the number of bile ducts for anastomosis was recorded. A final cholangiogram was performed at the conclusion of the procedure to ensure the integrity of the left ductal system.
The recipient hepatectomy was performed with preservation of the native inferior vena cava (IVC). Venovenous bypass was not used. The number and type of biliary anastomosis performed (i.e. a duct-duct (D-D) or roux-en-Y hepaticojejunostomy (R-Y)), was influenced by the number and quality of the donor ducts and recipient factors (e.g. sclerosing cholangitis). All anastomoses were performed in an interrupted fashion. A D-D anastomosis was usually performed in the presence of a single duct if the native hepatic was not diseased. When possible, a ductoplasty was performed in grafts of multiple ducts to reduce the number of anastomoses (Figure 2); the walls of the bile ducts were approximated by absorbable stitches (6-0 PDS; Ethicon, Somerville, NJ) in a double-barrel fashion. External stents or T-tubes were not used routinely.
We analyzed donor and recipient demographics, operative findings and techniques, postoperative course and long term follow-up from a prospective database. The estimated graft-to-recipient weight ratio (GRWR) was calculated by dividing the right lobe volume by the recipient's weight; a ratio of at least 0.8% was considered acceptable. Model for End-Stage Liver Disease (MELD) score reported here do not include tumor exception points. Liver tests (serum bilirubin (μmol/L), alkaline phosphatase) were performed at 1 year and at the time of final analysis (January, 2006) to determine whether the presence of a biliary stricture had any effect on graft function. To determine if the incidence of biliary strictures changed over time, the first 65 cases were compared with last 65 cases.
Biliary complications included leaks and strictures. Bile leaks were labeled as originating from the liver cut surface or from an anastomosis. The presence of a bile leak was identified by imaging (CT), persistent bilious drainage from a Jackson-Pratt drain placed near the biliary anastomosis or at time of reexploration. A stricture was defined as a symptomatic (e.g. cholangitis, hyperbilirubinemia) narrowing in the biliary tree identified either by MRCP, percutaneous transhepatic cholangiogram (PTC) or endoscopic retrograde cholangiogram (ERCP). An early stricture was any stricture presenting within 6 months of RLDLT.
Variables were analyzed using the chi-square test for categorical values and the Student's t-test for continuous variables. Freedom from strictures and graft survival were estimated by Kaplan-Meier method, and the log rank test was used to compare variables (SPSS for Windows 14.0, Chicago, IL). Variables that appeared to be significantly associated with strictures (p < 0.1) were entered into a backwards binary logistic regression model to test for significant effects while adjusting for multiple factors simultaneously. Significance was defined as a p < 0.05.
Recipient selection and demographics
Recipient characteristics and demographics are summarized in Table 1. No patient was lost to follow up. Median follow-up was 23 months (range 3–67). The median recipient age was 53 years (range 19–69 years); 62% were male; and median MELD score at the time of RLDLT was 14. The most common causes of liver failure was hepatitis C (HCV; 31%), alcohol (15%), primary sclerosing cholangitis (PSC; 12%), primary biliary cirrhosis (PBC; 12%), nonalcoholic steatotic hepatitis (NASH; 8%), hepatitis B (6%) and autoimmune hepatitis (6%). One year patient and graft survival rates were 87% and 85%, respectively.
Table 1. Right-lobe living donor recipient demographics
N = 128 (%)1
MELD = model for end-stage liver disease; NASH = nonalcoholic steatotic hepatitis.
1Due to rounding, may not equal 100%.
MELD at listing (range)
MELD at transplant (range)
Indication for transplant
Primary sclerosing cholangitis
Primary biliary cirrhosis
Alpha-1 Antitrypsin deficiency
Donor age (range)
The median donor age was 39 years (range 18–61 years); 28 donors (22%) were older than 50 years. No donor had any known preexisting biliary condition such as PSC or PBC at time of donation.
Graft anatomy and characteristics
Table 2 highlights the graft characteristics. The median GRWR was 1.2 (range 0.6–2.7) and 58% of grafts included the middle hepatic vein (MHV). Forty-nine percent (n = 63) of grafts had a single duct for anastomosis, 44% (n = 56) had two ducts and 7% (n = 9) had three ducts. After ductoplasty in 24 grafts, 87 grafts (68%) had a single duct for anastomosis. Forty-one grafts had two or more ducts for anastomosis. D-D and R-Y biliary reconstruction in the recipient was performed equally (n = 64 each). A D-D anastomosis was performed in 58 of the 87 cases (67%) in which one duct was present in the right lobe graft and in six grafts with two or more ducts (7%).
Table 2. Graft anatomy and type of reconstruction
N = 128 (%)
GRWR = graft: recipient weight ratio; R hepatectomy = R hepatic lobectomy; MHV = middle hepatic vein.
1Data not available for all cases.
R hepatectomy with MHV
Number of bile ducts
Ducts to anastomose
Type of reconstruction
Rate of biliary complications
The overall incidence of biliary complications was 32% (41/128). Biliary leaks developed in 19 patients (15%), while bile duct strictures occurred in 22 patients (17%). Eight patients (6%) developed both biliary complications (Table 3). After ductoplasty, 17% (4/24) developed a subsequent complication (stricture = 2, leak = 2). Of the 41 patients with two or more ducts for engraftment who did not undergo ductoplasty, 46% (19/41) developed a biliary complication (stricture = 12, leak = 7; p = 0.02).
Table 3. Biliary complications and treatment after right lobe living donor liver transplantation
N = 128 (%)
Bile leak = defined leak from biliary anastomosis or cut surface of transected liver requiring therapeutic intervention; Early stricture = within six months; Late stricture = after six months; ERCP = endoscopic retrograde cholangiogram; PTC = percutaneous transhepatic cholangiogram; D-D = duct-to-duct biliary anastomosis; Roux-en-Y: hepaticojejunostomy employing roux-en-Y configuration.
Management of bile leaks
Percutaneous drain & ERCP stent
Initial management of strictures
Long-term management of strictures
Conversion D-D—> Roux-en-y
The majority of bile leaks were anastomotic in origin (n = 15; 79%); the others arose from the cut surface of liver (n = 4; 21%). The anastomotic leaks were diagnosed in part by persistent bilious output from the Jackson-Pratt drain near the biliary anastomosis (n = 7), at the time of reoperation (n = 5) or by CT (n = 3). This was in contrast to cut surface biliary leaks which were diagnosed at time of reoperation in all cases (n = 4). Biliary leaks were managed by reoperation in the majority of cases (n = 13/19; 68%); percutaneous drains with or without ERCP guided biliary stents was used in six cases (32%).
Nine of the 22 biliary strictures (41%; range 3–21 months) presented during the first 6 months and 13 were late. All strictures were managed nonsurgically initially (observation n = 6, PTC n = 11, ERCP n = 5) but four patients ultimately required conversion from D-D to R-Y. The median number of ERCP's and PTC's performed per patient with strictures was 3 (range 1–5) and 4 (range 1–10), respectively. There was no difference in outcome if the stricture presented early or late, and the management was the same in both instances.
Risk factors for biliary strictures and outcomes after RLDLT
Strictures developed more frequently in patients with bile leaks (p = 0.002), grafts with two or more ductal orifices for anastomosis (p = 0.05) and donors >50 years of age (p = 0.003) (Figure 3A–C). The type of biliary reconstruction (D-D vs. R-Y) was not a significant risk factor for future strictures; strictures developed more frequently after R-Y (14/64; 22%) than D-D (8/64; 12%) (Figure 3D).
In the multivariate analysis, donor age >50 years (p = 0.003; HR 5.6; 95% CI 1.8–17.2) and previous history of a bile leak (p = 0.002; HR 6.5; 95% CI 1.9–21.8) were significant independent predictors of biliary stricture (Table 4). Each 10 year increase in donor age was associated with a higher incidence of strictures only after age 40 but the most significant effect was evident in the fifth decade. We confirmed in a separate logistic regression analysis that older donor age was a risk factor for biliary stricture. Other factors including type of biliary reconstruction, recipient age, recipient gender, donor relationship, cold ischemic time, warm ischemic time, GRWR, cause of liver disease, MELD score and era (first 65 cases compared to last 65 cases) were not significant.
Table 4. Analysis of risk factors for biliary strictures
HR (95% CI)
MELD = model for end-stage liver disease; R-Y: Roux-en-Y hepatico-enteric anastomosis; D-D: duct-to-duct anastomosis.
Male sex (recipient)
Type of biliary anastomosis (R-Y vs. D-D)
Donor age >50 years
>1 Bile duct for anastomosis
The 6-month and 2-year stricture rate were 11% and 20%, respectively (Figure 4). No patient presented with a de novo biliary stricture beyond 21 months after RLDLT. With a median follow-up of 23 months, 96% (107/112) of patients who are still alive are currently free of any biliary complications. All five patients who remain symptomatic had an initial R-Y anastomosis. One graft (1%) was lost due to biliary problems in a patient with hepatitis B that is now relisted for deceased-donor transplantation. There was no difference in either the serum bilirubin (12.6 μmol/L vs. 15.3 μmol/L; p = 0.41) or alkaline phosphatase values at 1-year (199 vs. 184 IU/L; p = 0.45) and at the time of final analysis (p = 0.31 and p = 0.23, respectively) among patients in whom biliary complications did or did not develop.
This report highlights the pattern and management of biliary strictures after RLDLT over a 5-year span (2000–2005) from a single Western institution. Surgeons performing RLDLT must be able to individualize the type of reconstruction based on donor bile duct anatomy and quality of the recipient bile duct. When presented with multiple bile ducts, ductoplasty should be considered to reduce the risk of stricture and leak. Our findings confirm that D-D anastomosis is an attractive alternative to R-Y after RLDLT and provides access for future endoscopic therapy in cases of leak or stricture. The rate of biliary strictures is significantly higher in the recipients with a history of biliary leaks and in those that received grafts from donors older than 50 years of age. Although biliary strictures after RLDLT are common, in most recipients they can usually be managed nonoperatively and do not affect short-term graft survival; currently, 96% of patients are free of any biliary symptoms or complications. Longer follow-up will be necessary to determine the long-term outcome of biliary reconstruction after RLDLT.
The anatomy of the right bile duct can pose a surgical challenge. A single right hepatic duct is found in only 50% of cases. Even when there is a single duct, it is often short and quickly divides into anterior and posterior divisions. In this situation, we have found that a cholangiogram immediately before duct transection can be particularly helpful in guiding both the level and angle of transection. This practice increases the likelihood of obtaining a single duct while minimizing the risk to the donor left bile duct (4). Careful planning in the donor right hepatectomy, such as preoperative MRCP to identify high-risk anatomy (13,15), is also essential to minimize biliary complications.
We believe that several principles should be followed to increase the likelihood of obtaining a healthy well-vascularized right bile duct(s). First, dissection of the hilar plate should be minimized to avoid disruption of the microcirculation around the right hepatic duct and artery (8,17). We usually delay duct division until after 80% of the hepatic parenchyma has been divided as we find that it improves visualization of tissue planes at the biliary confluence, minimizing the need for direct dissection of the right duct. Dividing the parenchyma along the lateral aspect of the base of segment 4 rather than through the gall bladder fossa also helps to avoid over dissection of the right duct. Second, the duct should be divided sharply and perpendicular to its long-axis to minimize the risk of ‘skeletonizing’ the postero-medial aspect of the right hepatic (or the right posterior duct when there are two ducts) (Figure 4). In the latter half of our experience, we have often performed another cholangiogram with a metallic clip placed beside the proposed site of duct incision to increase accuracy. Third, it is important to preserve the vascular plexus around the common bile duct in the recipient, which is derived from the gastroduodenal artery and the right hepatic artery (1).
The A2ALL consortium reported a decreased incidence of bile leaks in the first 90 days from 38% to 24% when centers performed more than 21 RLDLT (18). We were unable to demonstrate a significant ‘learning curve’ in bile duct strictures after RLDLT despite changes in our operative technique and practice over time. Other recent large series from experienced centers in Kyoto, Hong Kong and New York have also been unable to demonstrate a learning curve with biliary strictures after RLDLT (5,9,17).
Biliary reconstructions with R-Y have traditionally been considered to be better than D-D and more durable because of more reliable blood perfusion and ability to obtain a tension-free anastomosis. Some reports suggest a R-Y reconstruction is associated with a lower stricture rate in RLDLT (8,17). We believe that a D-D anastomosis provides several advantages over R-Y: it is faster to perform; more physiologic; eliminates the need for intestinal manipulation and potential risk of ascending cholangitis; and facilitates investigation and management of postoperative biliary complications through ERCP (1,17,19). In addition, R-Y can be more technically challenging than D-D when there is mucosal edema resulting from portal hypertension and hypervolemia.
As pointed out by Fan et al., simply joining two ducts together may narrow the A-P diameter and increase the risk of stricture (8). Therefore, it is advisable to make an incision along the adjacent wall of each duct before suturing them together. The net effect is a much larger orifice, which facilitates the reconstruction, and likely reduces complications. A ductoplasty is not always feasible when the ducts are widely displaced or diverge at sharp angles away from each other (8), because of high wall tension or poor blood supply. Ductoplasty increased the number of grafts that required a single anastomosis by 17%. Only 4 of these 24 patients (17%) went on to suffer a biliary complication. This low incidence of morbidity contrasts with other studies that report a high leak rate after ductoplasty (4,5,17,19). Furthermore, the risk of biliary complications was significantly less after ductoplasty as compared to grafts with two or more ducts for anastomosis (17% vs. 46%).
Although strictures developed in 17% of patients after RLDLT, almost all patients currently maintain excellent graft function free of biliary symptoms. We are cognizant that absence of symptoms (e.g. recurrent fever and cholangitis) does not preclude the presence of functionally significant stricture. So far, we have not found any difference in long-term allograft function in these with or without biliary complications with respect to serial serum bilirubin or alkaline phosphatase levels, rejection or recurrent hepatitis C. The association of recurrent hepatitis C and biliary strictures is poorly defined; in our series, two of five patients (40%) with hepatitis C who developed biliary strictures also developed recurrent hepatitis C.
Conservative management of recurrent cholangitis in recipients with R-Y drainage using antibiotics and PTC drainage is prudent. In some patients, segmental strictures can result in lobar atrophy with resolution of clinical symptoms. All five patients who are still symptomatic underwent initial R-Y and were ineligible for further endoscopic therapy (Figure 5). Initial management of D-D anastomotic stricture should be with ERCP stent. Surgical correction of a biliary stricture after D-D anastomosis with R-Y reconstruction may be performed in patients with recurrent symptoms of cholangitis after ERCP stent removal or those patients who were unable to be initially stented. Currently, only one graft has been lost due to biliary complications. This graft was reconstructed with an initial R-Y anastomosis and the patient is currently relisted for deceased donor liver transplantation.
Risk factors for biliary stricture in our series included bile leak, two or more ducts for anastomosis, and older donor age greater than 50 years. In the multivariate analysis, only donor age greater than 50 years and previous bile leak were significant independent risk factors for future strictures. It would seem intuitive that the presence of bile leaks and multiple ducts for anastomosis would predispose a patient to future stricture; however, the impact of donor age was surprising. Olthoff and colleagues from the A2ALL consortium reported an increased risk of graft loss with donor age but this did not reach statistical significance (18). The average age of donors in their study was only 37 years. Our living donor population was slightly older with 22% of donors older than 50 years. The reason for the high incidence of strictures involving older donor grafts is unclear but may reflect differences in the quality of hilar tissue or microcirculation. Further investigation of a larger number of recipients is required to clarify this issue.
Biliary complications are the most common complication confronting centers performing RLDLT. These results confirm that although common, they can usually be managed nonoperatively. Despite this, biliary strictures require multiple interventions, increase hospital costs and affect the overall quality of life after RLDLT. Longer follow-up will be necessary to determine the long-term outcome of biliary reconstruction and if conservative management is appropriate and effective in most cases.