Bile duct anastomotic stricture after adult-to-adult right lobe living donor liver transplantation

Authors


Abstract

Duct-to-duct anastomosis (DDA) and hepaticojejunostomy (HJ) are options for biliary reconstruction in patients undergoing adult-to-adult right lobe living donor liver transplantation (ARLDLT), after which biliary anastomotic stricture (BAS) is common as a complication. The risk factors for BAS are not clearly defined. We aimed to determine the rate of post-ARLDLT BAS in our center and its associated factors. In 265 ARLDLT recipients, 55 (20.8%) developed postoperative BAS. The diagnosis was based on clinical, biochemical, histological, and radiological results. The BAS rates were 21.4% (43/201) for recipients undergoing DDA during transplantation, 18.9% (10/53) for recipients undergoing HJ, and 18.2% (2/11) for recipients undergoing both procedures. BAS and non-BAS patients had comparable demographics. The number of graft bile duct openings (P = 0.516) and the size of the graft's smallest bile duct (5 versus 5 mm, P = 0.4) were not significantly different between BAS and non-BAS patients. Univariate analysis showed that the factors associated with postoperative BAS were the recipient warm ischemia time (55 versus 51 minutes, P = 0.026), graft cold ischemia time (120 versus 108 minutes, P = 0.046), stent use (21.8% versus 7.1%, P = 0.001), postoperative acute cellular rejection (29.1% versus 11.0%, P = 0.001), and University of Wisconsin solution use (21.8% versus 7.1%, P = 0.001). Multivariate analysis showed that the cold ischemia time (odds ratio = 1.012, 95% confidence interval = 1.002-1.023, P = 0.014) and acute rejection (odds ratio = 3.180, 95% confidence interval = 1.606-6.853, P = 0.002) were significant factors. The graft survival rates of BAS and non-BAS patients were comparable. One patient required retransplantation for secondary biliary cirrhosis. In conclusion, BAS remains common after ARLDLT regardless of DDA or HJ. The graft cold ischemia time and postoperative acute cellular rejection are significantly associated with postoperative BAS. Liver Transpl 17:47–52, 2011. © 2011 AASLD.

Adult-to-adult right lobe living donor liver transplantation (ARLDLT) is one of the most complicated and technically demanding surgical procedures. It is associated with high morbidity and reoperation rates.1 In recent years, various advances in techniques and management have been introduced. Now a graft survival rate greater than 90% can be achieved even for high-risk recipients.2 In the past, hepaticojejunostomy (HJ) was the standard technique for bile duct reconstruction in ARLDLT, but now duct-to-duct anastomosis (DDA) is gaining popularity because of its various potential advantages, including a shorter operation time, fewer septic complications, more physiological enteric functions, and easier endoscopic access to the biliary tract.3 Nonetheless, despite these advantages, DDA may be associated with higher rates of biliary anastomotic stricture (BAS).4 BAS is a common complication after ARLDLT. It affects the long-term outcome of patients undergoing ARLDLT and the quality of life of recipients and sometimes causes graft loss or even patient death. The aim of this study was to determine the rate of BAS after ARLDLT in a single high-volume center and the factors associated with this complication.

Abbreviations:

ARLDLT, adult-to-adult right lobe living donor liver transplantation; BAS, biliary anastomotic stricture; DDA, duct-to-duct anastomosis; ERC, endoscopic retrograde cholangiography; ESLV, estimated standard liver volume; HJ, hepaticojejunostomy; LDLT, living donor liver transplantation; MELD, Model for End-Stage Liver Disease; PDS, polydioxanone suture; PTC, percutaneous transhepatic cholangiography.

PATIENTS AND METHODS

Prospectively collected data for 315 patients who underwent ARLDLT at the Department of Surgery of Queen Mary Hospital (Hong Kong, China) between July 1994 and August 2008 were reviewed. The first 50 cases were excluded because of the learning curve effect and because they had been reported in a previous study.5 We adopted the standard perioperative management protocol and surgical techniques for donor and recipient operations, which have been previously described.5, 6 HJ was adopted as the standard method for bile duct reconstruction before 2002, and DDA was the preferred method thereafter (from case 63 onward) unless there was a contraindication for safe anastomosis (eg, an unhealthy recipient bile duct or an anatomic variant in the donor bile duct).6 After March 2000, a T tube or internal stent was not placed routinely. Each liver graft was prepared at the back-table. Hepatic venous outflow reconstruction was performed as previously described,7, 8 and the hepatic duct was left untouched at all times.9 University of Wisconsin solution was used for graft preservation before August 2002; thereafter, histidine tryptophan ketoglutarate solution was used instead. The recipient warm ischemia time was defined as the time between the first venous anastomosis and portal reperfusion in the recipient. The graft cold ischemia time was the time between the cross-clamping of donor vessels and portal reperfusion in the recipient.

DDA and HJ Techniques

At the end of graft implantation when all vascular anastomoses were completed and hemostasis was achieved, DDA was performed. Bleeding points around the hepatic duct of the graft were plicated with fine sutures. For ductal openings adjacent to each other and less than 5 mm apart, ductoplasty was performed. Otherwise, a decision had to be made between 2 options: DDA to one duct and HJ to another or HJ for both ducts. The graft bile duct width and the recipient bile duct width were measured. The recipient bile duct might be partially closed with sutures to avoid a size discrepancy. End-to-end anastomosis was performed with 6/0 polydioxanone suture (PDS) continuous sutures for the posterior wall. A short segment of a cannula (a 3.5-Fr Argyle catheter) was temporarily inserted across the anastomosis until the anterior wall anastomosis was completed with multiple interrupted 6/0 PDS sutures. The cannula was then removed, and the sutures were tied. For HJ, a Roux jejunal loop of sufficient length was prepared with staples. Side-to-side jejunojejunostomy was performed with a single-layer 4/0 Prolene continuous suture, and it was fashioned at least 40 cm from the HJ anastomosis. HJ was performed through a retrocolic and preferably retrogastric route with 6/0 PDS continuous sutures for the posterior wall and multiple interrupted 6/0 PDS sutures for the anterior wall. After 2001, an abdominal drain was not inserted routinely unless the surgeon decided otherwise.

Immunosuppression Regimen

Before January 2001, the immunosuppression regimen consisted of tacrolimus and a corticosteroid. An intraoperative intravenous injection of 1 g of hydrocortisone was given. After transplantation, a tapering dose of methylprednisolone was administered (the initial dosage of 200 mg/day was tapered to 40 mg/day by day 7). The steroid treatment was then maintained with oral prednisolone at a dosage of 20 mg/day, which was gradually tapered off at 6 months after transplantation. Oral tacrolimus was given at a dosage of 0.15 mg/kg of body weight/day within 12 hours after transplantation. The dose was titrated to achieve a serum trough level of 10 to 15 ng/mL. Since January 2001, 20 mg of basiliximab (Simulect, Novartis, Basel, Switzerland) was given intravenously within 6 hours of graft reperfusion and on postoperative day 4. Hydrocortisone was injected intraoperatively (1 g) and on postoperative day 1 (500 mg). Oral tacrolimus at a dosage of 0.15 mg/kg of body weight/day was administered within 12 hours after transplantation, and the dose was titrated to achieve a lower trough level of 5 to 10 ng/mL. Mycophenolate mofetil (CellCept, Roche, Basel, Switzerland) was started within 48 hours after transplantation at a dosage of 1.5 g/day and was gradually tapered off at 3 months. A maintenance steroid was not given routinely.

For the treatment of acute rejection, a pulse steroid (1 g of methylprednisolone) was given with an increasing tacrolimus trough level. If no response was seen after 3 pulses, biopsy was repeated to exclude other possible causes of deranged liver function. For refractory rejection, muromonab-CD3 (Orthoclone OKT3, Janssen-Cilag) was given as the last resort.

Follow-Up and Diagnosis of BAS

All patients were followed up weekly for the first 2 months after discharge and then less frequently. Liver function was checked at each follow-up session, and clinical symptoms and signs of BAS were recorded. BAS was confirmed if the diameter of the anastomosis was ≤50% of the recipient bile duct. If liver function derangement or clinical symptoms developed (eg, itchiness, jaundice, and cholangitis), Doppler ultrasonography of the liver and liver biopsy, in addition to radioisotope scanning [99mtechnetium (Tc) diethyl-iminodiacetic (EHIDA) scan], were performed. According to the results of these investigations, a direct cholangiogram by percutaneous transhepatic cholangiography (PTC) for patients who had undergone HJ or by endoscopic retrograde cholangiography (ERC) for patients who had undergone DDA might have been required. Patients with confirmed BAS underwent endoscopic or percutaneous transhepatic biliary drainage and dilatation or reoperation (whichever was deemed appropriate). The diagnosis of acute cellular rejection was based on histological examination.

Statistical Analysis

The comparison of categorical variables was performed with the chi-square test or Fisher's exact test when it was appropriate. Nonparametric continuous variables were compared with the Mann-Whitney U test and are presented as medians and ranges. Significant variables were put into a multiple logistic regression analysis to determine independent predictive factors for BAS. Survival was estimated with the Kaplan-Meier method and was compared with the log-rank test. P values less than 0.05 were regarded as statistically significant, and all P values were 2-tailed.

RESULTS

Fifty-five of the 265 patients (20.8%) developed postoperative BAS. BAS patients and non-BAS patients had comparable demographics (Table 1). The median follow-up period was 55 months. The donor age, blood group compatibility, and severity of fatty change in the graft were not significantly different between the 2 groups of patients. The number of graft bile duct openings, the sizes of the graft bile ducts, and the biliary reconstruction method (DDA, HJ, or both) were not significant factors for BAS (Table 2). Patients who underwent DDA during transplantation had a BAS rate of 21.4% (43/201), those who underwent HJ had a BAS rate of 18.9% (10/53), and those who underwent both had a BAS rate of 18.2% (2/11). Hepatic artery thrombosis occurred in 5 non-BAS patients but not in the BAS group (Table 3). According to univariate analysis, the recipient warm ischemia time (55 versus 51 minutes, P = 0.026), graft cold ischemia time (120 versus 108 minutes, P = 0.046), stent use (21.8% versus 7.1%, P = 0.001), postoperative acute cellular rejection (29.1% versus 11.0%, P = 0.001), and use of University of Wisconsin solution (38.2% versus 20%, P = 0.005) were associated with a significantly higher chance of BAS. These 5 factors were put into a multiple logistic regression analysis. The graft cold ischemia time and postoperative acute cellular rejection were found to be significant factors associated with BAS (Table 4). Treatments for BAS are shown in Table 5. Essentially, patients who had undergone DDA during transplantation underwent ERC and dilatation, whereas those who had undergone HJ underwent PTC and dilatation. If dilatation still failed after 5 sessions, the mode of bile duct reconstruction was reviewed, and a mode switch was considered. Nine patients required a switch from DDA to HJ, and 4 patients required HJ revision because the stricture failed to dilate with repeated PTC and balloon dilatation. Thirty-eight patients underwent ERC and balloon dilatation with or without stenting. Three patients were given oral prophylactic antibiotics because their liver function was preserved and there was no evidence of cholangitis. One patient developed secondary biliary cirrhosis, for which retransplantation was performed. BAS patients had 1-, 3-, and 5-year graft survival rates of 97.3%, 90%, and 84.5%, respectively, whereas their counterparts had rates of 91.4%, 85.3%, and 82.5%, respectively (P = 0.574; Fig. 1).

Figure 1.

Graft survival in patients with or without BAS.

Table 1. Demographics of Patients With or Without BAS
FactorBAS Group (n = 55)Non-BAS Group (n = 210)P Value
Patient age (years), median (range)48 (17-61)49 (17-68)0.3
Patient sex (male:female), n47:8158:520.1
MELD score, median (range)28 (7-50)24 (6-59)0.2
Donor age (years), median (range)37 (18-57)34 (18-58)0.3
Donor sex (male:female), n19:3678:1320.7
Genetically related to donor, n (%)28 (50.9)132 (62.9)0.1
Blood group (identical:compatible), n37:18161:490.2
Graft weight/recipient ESLV (%), median (range)47.0 (41.7-52.9)49.9 (42.1-55.9)0.3
Fatty change in graft, n (%)  0.7
 No32 (58.2)119 (56.7) 
 Mild (<10%)20 (36.4)83 (39.5) 
 Moderate (10%-30%)3 (5.4)6 (2.9) 
 Severe (>30%)0 (0)2 (1.0) 
Cirrhosis, n
 Idiopathic20 
 Hepatitis B virus2394 
 Hepatitis C virus113 
 Alcohol12 
 Autoimmune01 
 Primary biliary06 
 Biliary atresia01 
Fulminant hepatic failure, n
 Idiopathic16 
 Hepatitis E virus10 
 Drug33 
 Hepatitis B virus02 
 Autoimmune01 
Cirrhosis with acute deterioration, n
 Idiopathic01 
 Hepatitis B virus1041 
 Hepatitis C virus01 
 Alcohol11 
 Wilson's disease03 
Chronic active hepatitis with acute flare: hepatitis B virus, n1228 
Chronic active hepatitis, n01 
Graft failure, n
 Hepatic artery thrombosis02 
 Recurrent hepatitis B01 
Caroli disease, n01 
Familial amyloid polyneuropathy, n01 
Simultaneous hepatocellular carcinoma, n1170 
Table 2. Operative Findings for Patients With or Without BAS
FactorBAS Group (n = 55)Non-BAS Group (n = 210)P Value
University of Wisconsin solution, n (%)21 (38.2)42 (20)0.005
Number of graft bile duct openings, n (%)  0.5
 136 (65.5)147 (70.0) 
 >119 (34.5)63 (30.0) 
Size of graft's smallest bile duct (mm), median (range)5 (4-6)5 (4-7)0.4
Ductoplasty, n (%)12 (21.8)25 (11.9)0.1
Recipient warm ischemia time (minutes), median (range)55 (47-65)51 (43-61)0.026
Graft cold ischemia time (minutes), median (range)120 (99-135)108 (96-125)0.046
Time between portal and arterial reperfusion (minutes), median (range)84 (75-95)83 (68.3-103.5)0.5
Type of bile duct reconstruction, n  0.9
 DDA43158 
 HJ1043 
 Both29 
T tube or internal stent, n (%)12 (21.8)15 (7.1)0.001
Operative time (minutes), median (range)670 (607-788)668 (572-772)0.3
Table 3. Postoperative Outcomes for Patients With or Without BAS
OutcomeBAS Group (n = 55)Non-BAS Group (n = 210)P Value
Acute cellular rejection, n (%)16 (29.1)23 (11.0)0.001
Cytomegalovirus antigenemia, n (%)4 (7.3)18 (8.6)0.1
Complications, n (%)
 Bile leakage4 (7.3)6 (2.9)0.1
 Hepatic artery stenosis/thrombosis0 (0)5 (2.4)0.2
 Hepatic vein stenosis/thrombosis1 (1.8)1 (0.5)0.3
 Portal vein stenosis/thrombosis9 (16.4)12 (5.7)0.4
Table 4. Multivariate Analysis of Factors Associated With BAS
FactorOdds Ratio (95% Confidence Interval)P Value
Postoperative acute cellular rejection3.180 (1.606-6.853)0.002
Graft cold ischemia time1.012 (1.002-1.023)0.014
Stent 0.1
University of Wisconsin solution 0.2
Recipient warm ischemia time 0.7
Table 5. Treatments for BAS
TreatmentN
  1. NOTE: Patients were not restricted to only 1 treatment.

ERC and dilatation with or without stenting38
Conversion to HJ from DDA9
Percutaneous transhepatic biliary drainage and dilatation7
Revision of HJ4
Conservative treatment3
Retransplantation1

DISCUSSION

BAS, a technical Achilles' heel of orthotopic liver transplantation,10, 11 is a common cause of morbidities after transplantation; sometimes, it even causes death. It has been reported to be related to many conditions, including a prolonged cold ischemia time, hepatic artery thrombosis, blood group incompatibility, and cytomegalovirus infection, as well as the use of reduced-size grafts, the use of University of Wisconsin solution, and the method of bile duct reconstruction.12 Compared to recipients of whole liver grafts, living donor liver transplantation (LDLT) patients have a consistently higher rate of BAS.12-15 This is considered to be related to the anastomosis blood supply and the presence of multiple and small-caliber donor ducts. With technical refinements and limited intraoperative manipulation of the donor bile duct, the incidence of BAS has decreased dramatically, as previously reported by our center.9

A previous study suggested that DDA was associated with a significantly higher BAS rate,4 but we did not find this association in the present study. In the study, DDA and HJ yielded comparable BAS rates; this agreed with the findings of a study by another high-volume center.16 Instead, the graft cold ischemia time, despite our every effort to minimize it, was still a factor significantly associated with BAS. We performed donor and recipient operations simultaneously, and the 2 teams of surgeons were always in good communication. This enabled us to keep the cold ischemia time at its shortest.

We found that in addition to the graft cold ischemia time, postoperative acute cellular rejection was also a factor significantly associated with BAS. There are reports showing that nonanastomotic strictures are associated with chronic ductopenic rejection.17 The present study is the first to show the importance of acute cellular rejection in the formation of BAS. This could be due to the rejection itself, the consequences of the treatment of acute rejection, or sampling bias.

The exact pathogenesis of rejection causing BAS is not yet known. A possible explanation is that more patients are subjected to liver biopsy for deranged liver function, so there are more incidental findings of acute cellular rejection together with bile duct obstruction [in the BAS group, 36 patients (65.5%) were biopsied; in the non-BAS group, 77 patients (36.7%) were biopsied].

We adopted an aggressive approach to managing BAS. For patients who had undergone DDA, vigorous ERC and balloon dilatation with or without stenting were performed by our own team members. For those who had undergone HJ, percutaneous transhepatic biliary drainage and balloon dilatation were performed by interventional radiologists. An internal biliary stent was inserted during ERC if the stricture was tight or a residual stricture was seen. The interval between 2 dilatations was no more than 6 weeks. In general, if the stricture could not be dilated satisfactorily after 5 sessions of dilatation, reoperation was considered. In the procedure, meticulous care was taken to avoid damage to the hepatic artery near the biliary anastomosis.9, 18 As a result, the 2 groups (BAS and non-BAS) had similar graft survival rates, and only 1 patient required retransplantation because of secondary biliary cirrhosis.

As for graft preservation, University of Wisconsin solution was used before August 2002. In August 2002, we switched to histidine tryptophan ketoglutarate solution because the former might be associated with BAS. As reported in the literature,19 University of Wisconsin solution might be associated with a high stricture formation rate after deceased donor liver transplantation.

One drawback of this study was bias in diagnosing BAS. It was easier to subject patients to ERC than PTC because ERC had a lower associated risk, and it was more difficult to arrange interventional radiologists for patients with tests for deranged liver function. The lack of an independent party to assess the occurrence of BAS was also a source of bias. A recent large retrospective cohort study by Freise et al.20 showed that there might be fewer bile duct strictures after HJ in LDLT patients, but no definite conclusion could be drawn because of the study's retrospective nature. In fact, to date, there is only 1 randomized study comparing DDA and HJ.21 However, in the comparison of LDLT and deceased donor liver transplantation by Freise et al., patients undergoing the former procedure were found to have significantly more biliary complications than patients undergoing the latter procedure.

In summary, BAS remains common after ARLDLT regardless of DDA or HJ, and the graft cold ischemia time and postoperative acute cellular rejection are 2 significantly associated factors. The former factor has been minimized in centers practicing LDLT. Now, preventing the latter should be the goal of all liver transplant surgeons.

Ancillary