Objectives: Our aim was to evaluate the influence of the localization of right posterior bile duct anatomy relative to portal vein of the donors on posttransplant bile duct complications.
Materials and Methods: We retrospectively investigated 141 patients who had undergone living donor liver transplant using right hemiliver grafts. The patients were classified based on the pattern of the right posterior bile duct and divided into infraportal and supraportal types. Clinical donor and recipient risk factors and surgical outcomes were compared for their relationship with biliary complications using logistic regression analyses.
Results: The 2 groups were similar according to demographic and clinical features. The biliary complication rate was 23.7% (9/38) in the infraportal group and 47.4% (37/78) in the supraportal group (P = .014). An analysis of risk factors for the development of anastomotic bile leak using logistic regression showed that a supraportal right posterior bile duct anatomy was a statistically significant positive predictor, with odds ratio of 18.905 (P = .012; confidence interval, 1.922-185.967). The distance of the right posterior bile duct from confluence was significantly lower in patients with biliary complications than in those without (mean of 7.66 vs 0.40 mm; P = .044). According to receiver operating characteristic analyses, the cut-off point for the length of right bile duct to right posterior bile duct from the hepatic confluence was 9.5 mm regarding presence of complications.
Conclusions: Factors influencing bile duct anastomosis leakage were supraportal-type donor bile duct anatomy and length of the right main bile duct from biliary confluence. Hepatic arterial complications were similarly a risk factor for biliary strictures. Because of the multiple factors leading to complications in living donor liver transplant, it is challenging to group these patients by operative risk; however, establishing risk models may facilitate the prediction of complications.
Key words : Biliary leakage, Infraportal, Supraportal
There are insufficient numbers of organs from deceased donors for the growing numbers of recipients on wait lists, adding to longer wait times and increased wait list mortality. Despite efforts, the number of liver donations has not increased adequately to maintain transplant needs. This growing imbalance has created a need for liver transplants from living donors. Living donor liver transplant (LDLT) was first started as a left lobe or left lateral segment transplant in pediatric patients1 and then in adults.2,3 Acquiring a size-matched living liver donation for every recipient is a difficult and limiting task. Because the procedure and outcomes have been successful as a result of innovative surgical techniques, advanced clinical care, and improved immunosuppression, this method soon extended to living transplants in adult patients where the right lobe is used.4
Unfortunately, LDLT is a more complicated procedure than deceased donor liver transplant, mainly because of its technical complexity and different physiologic requirements resulting from regeneration of a partial graft. Despite these issues, current recipient and donor outcomes have become acceptable with the establishment of standardized surgical techniques, a better understanding of the anatomic variations, and a better understanding of the hemodynamic properties of partial graft regeneration.5
Despite all surgical and technological advances, recipient biliary complications continue to be an important posttransplant issue, including mortality in patients undergoing liver transplant.6,7 Complications, including those involving bile ducts, are more frequent in living donor transplant procedures using partial grafts than in transplants with deceased donations.6-8 The reported risk factors for biliary complications after anastomosis are rather confusing. Although biliary anatomic variations in donors and surgical techniques are widely acknowledged risk factors for recipient biliary complications,9,10 it has been reported that donor age, donor body mass index, and macrovacuolar steatosis of the graft may be independent risk factors for biliary complications.11,12 The relation between bile ducts and portal vein confluence may be an independent risk factor for bile duct complications.
In this study, we aimed to determine whether localization of the right posterior bile duct (RPBD) anatomy relative to the portal donor vein has an influence on outcomes by assessing preoperative magnetic resonance cholangiopancreatography (MRCP) and intraoperative cholangiography images and intraoperative findings. We also aimed to determine factors, including anatomic properties, affecting posttransplant bile duct complications.
Materials and Methods
At our center, 525 liver transplant procedures were performed between 8 December 1988 and 28 February 2016. Of these procedures, 384 were LDLT and 141 were deceased donor liver transplants. Left lobe LDLT in children was first performed in March 1990 and in adults in April 1990 at our center.3,13 A combined liver and kidney transplant from a living related donor was the first procedure of its kind to be done anywhere in the world, which was performed in May 1992.14 Although we have been performing right lobe LDLT since April 2003, only patients who underwent MRCP before liver transplant and had at least 1 year of follow-up were included in the study, that is, eligible patients seen between January 2005 and December 2014. We retrospectively evaluated all consecutive patients who had undergone right lobe LDLT (n = 141). Thirteen patients who died while in the hospital or by 1 year after transplant were excluded from the study. After the exclusion of 9 patients with inaccessible intraoperative cholangiography images and patients with Hwang15 type C (n = 14) and type D (n = 2) donors according to preoperative MRCP and intraoperative cholangiography images, a total of 116 patients were included in the study.
Recipient and donor age, sex, the number and type of bile ducts of the graft in the donors, cold and warm ischemia time, duration of operation, the number and type of bile anastomosis, the cause of cirrhosis in recipients, Model for End-Stage Liver Disease/Pediatric End-Stage Liver Disease scores, cytomegalovirus infection during posttransplant follow-up, acute rejection, and bile duct complications and their treatment were retrospectively assessed. Complications after right lobe LDLT were evaluated under 2 categories, namely, anastomosis leak and biliary stricture. Postoperative anastomosis leak was defined as a contrast leak identified at the anastomosis site by endoscopic retrograde cholangiopancreatography or percutaneous transhepatic cholangiography or diagnosed at revision surgery. Stenosis at the anastomosis line or in intrahepatic bile ducts beyond 30 days after transplant was accepted as strictures. Patients with liver enzyme elevation, fever, and abdominal pain were suspected to have biliary complications, which we confirmed with imaging studies. A biliary leak was defined as a contrast leak identified at the anastomosis by endoscopic retrograde cholangiopancreatography or percutaneous transhepatic cholangiography. A biliary stricture was defined as a narrowing of anastomotic sites that required intervention. Cytomegalovirus infection was defined as patients who received preemptive or therapeutic antiviral treatment. Acute rejection was confirmed by liver biopsy.
Bile duct reconstruction
Duct-to-duct anastomosis or Roux-en-Y hepaticojejunostomy techniques were used for bile duct reconstruction. The anastomosis type to be performed was decided by considering the number of bile ducts and the diameter of the graft, diameter of the main bile ducts of the recipient, diameter mismatch between the recipient and donor, the disease cause in the recipient, and whether the procedure was a retransplant.
A duct-to-duct single anastomosis was preferred in recipients who had a single bile duct of the graft, when there was no diameter mismatch between the recipient or donor bile canal and the cause was not primary sclerosing cholangitis. When bile ducts were multiple but adjacent to each other, a single anastomosis was preferred in duct-to-duct or hepaticojejunostomy fashion after they were united as a single opening. When there were multiple bile ducts that were widely spaced from each other, duct-to-duct or hepaticojejunostomy anastomoses were performed separately.
Recipients were checked twice per day for the first 6 days after transplant by Doppler ultrasonography for vascular structures, intrahepatic bile ducts, ascites, collections, and the state of the graft parenchyma. Doppler ultrasonography was performed 1 month after transplant and was repeated at 3 months. At day 7 after transplant, a detailed radiologic assessment was done by abdominal computed tomography plus hepatic computed tomography angiography to detect leaks and strictures in addition to ultrasonographic-evaluated parameters.
Treatment of biliary complications
An internal-external or an external drainage catheter was placed for anastomosis leaks. When a leak was nonanastomotic and caused biloma, a percutaneous drainage was placed. When a control cholangiography demonstrated that a leak had been eliminated, the catheter was removed, whereas in cases of leak persistence, either surgical anastomosis revision was performed or duct-to-duct anastomosis was switched to Roux-en-Y hepaticojejunostomy.
In the case of a stricture, endoscopic retrograde cholangiopancreatography-guided balloon dilatation was primarily applied to the bile anastomosis. When a stenosis of more than 50% persisted despite balloon dilatation, an internal stent was placed. In cases where endoscopic retrograde cholangiopancreatography failed to relieve stenosis, the interventional radiology department performed percutaneous transhepatic cholangiography and balloon dilatation. After balloon dilatation, depending on the stenosis grade, either an internal stent or internal-external drainage catheter was placed, and the balloon dilatation procedure was repeated as needed. In patients with insufficient dilatation in control cholangiographies, either surgical anastomosis revision was performed or duct-to-duct anastomosis was switched to Roux-en-Y hepaticojejunostomy.
Determination of donor bile duct anatomy
Bile ducts, preoperative MRCP and computed tomography images, and intraoperative cholangiography images of all donors were reviewed by the surgical team and the radiologist in a comparative manner. Thus, it was aimed to evaluate the bile duct anatomy in a consistent manner. Additionally, detailed operative findings were recorded.
The Hwang classification, which is widely used, was also preferred at our center for the classification of right lobe bile duct anatomy. The features of this classification have been described elsewhere.15
Bile duct anatomy features of donors were categorized as 2 different types based on the anatomy of the right posterior hepatic canal on craniocaudal plane and its localization relative to the portal vein. Patients were divided into 2 groups: infraportal and supraportal types. The presence of a right posterior bile canal that opens into the right hepatic duct and courses beneath the portal vein on cranio-caudal plane was denoted as infraportal type (Figure 1). The presence of a right posterior bile canal that opens into the right hepatic canal and courses above the portal vein on cranio-caudal plane was denoted as supraportal type (Figure 2).
The length of the right bile duct from the biliary confluence to RPBD was also measured and recorded. In cases with right posterior, right anterior, and left main bile canal branched off as a trifurcation, the patients were grouped by the supraportal or infraportal localization of the right posterior bile canal (Figure 3). In donors with such a type of bile duct anatomy, this length was accepted as zero.
Statistical analyses were done with Statistical Package for the Social Sciences (SPSS) version 11.0 (SPSS Inc., Chicago, IL, USA) software package. Study data were expressed by descriptive statistics (mean, standard deviation, median, range, frequency, and percentage). Before the statistical analyses were done, study variables were tested for some assumptions. The Kolmogorov-Smirnov test was used to test normality of study variables, and the Levene test was used to test homogeneity of variances. Normally distributed parameters were compared between the 2 groups using t test. Nonnormally distributed parameters were compared between groups using Mann-Whitney U test. Receiver operating characteristic curve analysis was used to determine cut-off points for measurements of distance to the confluence. Clinical donor and recipient risk factors and surgical outcomes were studied for their relationship with biliary complications using logistic regression analyses. Qualitative data were compared using Pearson chi-squared test and the Fisher exact test. The results are presented at a confidence interval of 95% and a significance level of P < .05.
Although we have performed right lobe LDLT since 2003, only patients who had undergone MRCP before liver transplant and had at least 1-year follow-up between January 2005 and December 2014 were included in the study. The demographics of the 116 patients included in the study by RPBD classification are shown in Table 1, with 38 patients having infraportal-type classification (infraportal group) and 78 patients having supraportal-type classification (supraportal group). The 2 groups did not differ significantly regarding demographic properties. In total, there were 79 men and 37 women with a median age of 44 years (range, 6-64 y). Donors included 74 men and 42 women with a median age of 33 years (range, 19-55 y). Indications for transplant included viral cirrhosis (hepatitis B virus; 21.6%), followed by hepatocellular carcinoma (21.4%), Wilson cirrhosis (10.3%), cryptogenic cirrhosis (8.6%), alcoholic cirrhosis (6%), fulminant hepatic failure (5.2%), and miscellaneous causes (26.9%).
Comparison of the groups regarding clinical and operative outcomes
Both groups were similar with respect to transplanted graft volume, cold ischemia time, warm ischemia time, operative time, follow-up time, bile anastomosis type (duct-to-duct or hepaticojejunostomy), and the distance of RPBD to biliary branching (Table 2). Although patients with type A bile duct anatomy (according to Hwang classification15) were mostly in the infraportal group, patients with type B bile duct anatomy were mostly in the supraportal group (P < .001). Similarly, the number of bile ducts requiring more than 1 reconstruction was significantly higher in the supraportal group. Although arterial complications were more common in the supraportal group, the difference did not reach statistical significance (infraportal vs supraportal group was 10.5% vs 21.8%; P = .199). The rate of overall bile duct complications was similar in both groups, although the rate of bile leak was higher in the supraportal group. Bile leak was seen in 7.9% of patients in the infraportal group and 30.8% of patients in the supraportal group (P = .005). Two patients from the supraportal group had cut surface bile leaks, with no cut surface bile leaks in the infraportal group (P = 1.000). Bile duct stenosis was at similar rates in both groups. Cytomegalovirus infection rate and acute rejection rate were similar in both groups.
An analysis of risk factors for the development of bile leak using logistic regression showed that only RPBD anatomy was a statistically significant positive predictor, with odds ratio of 18.905 (P = .012; 95% confidence interval, 1.922-185.967) (Table 3). Arterial complications and type of biliary anastomosis were found to be significant risk factors for biliary stenosis, with odds ratio of 10.416 and 31.164 (P = .010; 95% confidence interval, 1.768-61.384, and P = .015; 95% confidence interval, 1.970-492.948) (Table 4).
The length of the right bile duct from biliary confluence to RPBD was significantly lower in patients who had biliary complications versus those without (mean of 7.66 vs 0.40 mm; P = .044). Considering this condition, we aimed to find a cut-off point for the distance from the confluence showing presence of complications. Receiver operating characteristic analysis and diagnostic screening tests were used to determine a cut-off point for each group (Table 5). This cut-off point for distance from the confluence for presence of complications was found to be 9.5 mm. This cut-off point of 9.5 mm had a sensitivity of 84.8% and specificity of 40.0%. The area under curve was 61.0%, and the standard error was 5.2% (P = .045) (Figure 4).
Lack of adequate numbers of deceased donations to meet liver transplant needs has obligated surgeons to transplant from living donors, and recent technical advances and appropriate patient selection have transformed transplant from living donors into a highly successful and safe procedure.5 Anatomic variability of the sectional ramification is much higher in the right liver, and distance to the sectional ramification is also shorter than that in the left liver,16 and this could lead to biliary complications. After we categorized patients regarding the localization of RPBD relative to portal vein, we found that both groups (infraportal and supraportal) had similar demographic and clinical characteristics and operative outcomes in our study. Patients in whom right lobe bile duct confluence were supraportal had a significantly higher rate of biliary anastomosis leak (infraportal vs supraportal was 7.9% vs 30.8%; P = .005). On the other hand, unlike biliary anastomosis leaks, biliary strictures were not significantly different between the groups in our study. Also, distance of RPBD from the biliary confluence of < 9.5 mm was a risk factor for bile duct complications.
Biliary complications are still a serious consideration because of the potential for morbidity in LDLT. In a retrospective cohort from the adult-to-adult living donor liver transplant cohort consortium study, biliary complications were found at higher rates in LDLT patients than in deceased donor liver transplant patients (40% vs 25%; P = .0046).17 These data were derived retrospectively using databases from 8 large transplant centers. Therefore, some data could be missing and factors influencing biliary complications were not totally assessed. However, biliary complication rates were similar to those shown in our supraportal group, whereas our infraportal group had better biliary complication rates. In a 10-year prospective study comparing living donor and deceased donor transplant, higher biliary complication rates were shown in patients with deceased donations (27.1% vs 17.6%; P = .026).18 These contrasting results need further evaluation in more detailed prospective studies. Anatomic features of bile ducts according to dissection plane and portal vein could be factors influencing biliary complications. The main risk factors for complications involving bile ducts after liver transplant include surgical technique, transplant from a living donor, graft type, cold ischemia time, posttransplant rejection episodes, complications involving hepatic artery, and factors related to recipient and donor.19
In a recent study, Jeon and associates20 found that the rate of biliary complications significantly increased in patients with supraportal localized RPBDs (infraportal vs supraportal group of 0% vs 17.6%-52.6%). In that study, biliary strictures and anastomosis leaks were combined for analyses. A logistic regression analysis of a variety of parameters that we investigated in our study revealed that a supraportal localization of RPBD significantly increased bile anastomosis leak risk (odds ratio = 18.905; P = .012). Arterial complications were found to be significant risk factors for biliary stricture (odds ratio = 10.416; P = .010). Similarly, hepaticojejunostomy anastomosis type was a significant risk factor for biliary stricture development (odds ratio = 31.164; P = .014). A review article and a meta-analysis investigating biliary complications and reconstruction outcomes found similar outcomes for duct-to-duct anastomosis and hepaticojejunostomy anastomosis.21 They recommended that the choice of anastomosis type be based on anatomic and personal characteristics of a patient.
Vascular anatomy of the donor, localization of bile ducts relative to the portal vein, and the number and diameter of bile canals are important risk factors for biliary complications developing after transplant.12 However, it is very difficult to verify the real effects of technical factors in biliary complications of LDLT because there are many perturbed variables in the surgical field. In addition, the preservation of blood supply to anastomosis is also important to prevent late biliary stricture.22 Unnecessary excessive dissection around donor hepatic duct should be avoided. During recipient hepatectomy, connective tissue around the native common duct was preserved to maintain the ascending axial vascular circulation for successful biliary anastomosis.5 These anatomic variations frequently encountered in right lobe grafts also increase the chances of technical errors and thus the complication risk.
Intragroup variabilities, particularly in patients with bile duct variations, have not been adequately studied. Bile duct anastomosis leaks were studied after left hepatectomy operations for cholangiocarcinoma and found to have a significantly lower rate in the group with an infraportal RPBD relative to portal vein than the group with a supraportal RPBD.23 It is difficult to predict the preventive effects of infraportal RPBD on biliary complications. It is obvious that exposure of a RPBD with supraportal localization will require a more difficult dissection; however, the appropriate dissection of the main right canal is sufficient for right hepatectomy. Jeon and associates suggested that this anatomic variation may lead to bile duct injury due to creating a need for a dissection within liver parenchyma.20 They considered that this difficulty is a factor for increasing biliary complications when hemiliver grafts were used. Supraportal-type RPBD is close to the midplane of the liver. Because of this, dissection could result in skeletonization of the duct or even direct injury to the duct, perhaps leading to a higher rate of biliary complications in supraportal-type procedures. Therefore, in donors with supraportal RPBD, parenchymal dissection should be conducted so not to injure the Glissonian capsule and not to skeletonize the upper part of the right hepatic duct.20 The possibility of RPBD with supraportal localization to form tension in anastomosis may be a causative factor for more frequent anastomosis leaks.
Aberrant bile duct anatomy and presence of multiple bile canals have been implicated as apparent risk factors for the development of biliary complications.24 In cases with right posterior, right anterior, and left main bile canal branching of as a trifurcation, or the site of the opening of the right posterior bile canal into the main hepatic canal being too close to the confluence, branching of the right lobe as anterior and posterior takes place more proximally than with standard anatomy. Hence, there is an increased chance for multiple bile duct ostia being present in the right lobe after the split is completed.8 Here, what is more important is to cut the bile duct on a correct plane when being transected. A transection on an incorrect plane not only increases the number of bile ducts to be anastomosed but increases ischemia risk and the associated risk of anastomosis leak by causing the posteromedial wall of the right posterior bile canal to be skeletonized.25 The reverse is also true. When there are multiple bile ducts in the right lobe, it becomes technically difficult to determine the transection line accurately, which may lead to oblique or irregular dissection or the injury of the ostia of bile ducts. Therefore, it is of utmost importance to draw a transection line on a perpendicular plane to the right lobe bile ducts under the guidance of an intraoperative control cholangiography.26
Biliary stricture and obstruction occur in 3.7% to 25.3% of posttransplant patients.27 Stricture most commonly develops at the anastomosis site. The most important causative factor for strictures is technical errors, followed by hepatic artery problems; AB0-incompatible transplant procedures, cytomegalovirus infection, and sclerosing cholangitis being the primary disease may also be included as causative factors. Despite the presence of studies reporting the opposite,28 many studies have reported higher stricture rates in the duct-to-duct technique than with the hepaticojejunostomy technique.15,29 Our study revealed that anastomosis type had no effect on anastomosis leak, whereas it showed that hepaticojejunostomy is a risk factor for stricture development. The fewer events per variable (n < 10), the greater the opportunity for the estimates of the regression coefficients to be unreliable; the sample variance of the model coefficients and confidence intervals will also be less accurate.30 The validity of statistical inference may also be adversely affected by having a small number of events per variable. The chances of biliary stenosis development were higher in patients undergoing hepaticojejunostomy anastomosis according to this evaluation; however, the extremely low number of patients who were operated with hepaticojejunostomy and < 5 patients in the 2 cells of logistic regression analysis table of hepaticojejunostomy group have limited the reliability of this finding.
The retrospective design and relatively low number of patients are the main limitations of our study. The variability of anatomic variations of bile ducts and multiple factors influencing complications made homogenous and adequate grouping difficult. However, not only anatomic variations of bile ducts but also systemic, immunologic, and hepatic arterial factors affect complication rates, and they should be reliably established.
In conclusion, the factors influencing bile duct anastomosis leakage were supraportal-type donor bile duct anatomy and length of the right main bile duct from biliary confluence. Our study patients with supraportal-type right lobe bile duct confluence were under risk of biliary anastomosis leak. Hepatic arterial complications were similarly a risk factor for biliary strictures. Determining complication risk preoperatively may aid in determining patient, donor, and surgical strategy to prevent complications. Because of the multiple factors leading to complications in LDLT, it is challenging to group these patients by operative risk; however, establishing risk models may facilitate the prediction of complications.
DOI : 10.6002/ect.2016.0200
From the 1Department of General Surgery, Baþkent University; the
2Þanlýurfa Mehmet Akif Ýnan Eðitim ve Araþtýrma Hastanesi; and the
3Department of Radiology, Baþkent University, Ankara, Turkey
Acknowledgements: The authors declare that they have no sources of funding for this study, and they have no conflicts of interest to declare.
Corresponding author: Tugan Tezcaner, Department of General Surgery, Baskent University Hospital, 5. Sokak No: 48, Bahcelievler, Ankara 06490, Turkey
Phone/Fax: +90 312 215 2629
Figure 1. (A) Magnetic Resonance Cholangiopancreatography of Infraportal Type of Right Posterior Bile Duct; (B) Intraoperative Cholangiography of Infraportal Type of Right Posterior Bile Duct
Figure 2. (A) Magnetic Resonance Cholangiopancreatography of Supraportal Type of Right Posterior Bile Duct; (B) Intraoperative Cholangiography of Supraportal Type of Right Posterior Bile Duct
Figure 3. (A)Magnetic Resonance Cholangiopancreatography of Trifurcation of Right, Left, and Right Posterior Bile Duct; (B) Intraoperative Cholang iography of Trifurcation of Right, Left, and Right Posterior Bile Duct
Figure 4. Receiver Operating Characteristic Analyses and Diagnostic Screening Tests for Determining a Cut-Off Point for Complications Regarding Length of Right Bile Duct to Right Posterior Bile Duct From the Hepatic Confluence
Table 1. Demographic Features of Recipients and Donors Grouped According to Right Posterior Bile Duct
Table 2. Clinical Features and Surgical Outcomes Grouped According to Right Posterior Bile Duct
Table 3. Logistic Regression Analysis to Assess Risk Factors of Anastomosis Leakage
Table 4. Logistic Regression Analysis to Assess Risk Factors of Anastomotic Stenosis
Table 5. Cut-Off Value of the Length of Right Bile Duct to Right Posterior Bile Duct From the Hepatic Confluence for Biliary Complications and Receiver Operating Characteristic Curve Analyses