Technical Problems: Biliary

Epidemiology

BC rates were reported to be as high as 50% in early publications . During the last decades a constant decrease has been documented. Currently a range from 5% to 25% is reported.

A recent systematic review conducted by Akamatsu et al included 61 retrospective or prospective studies comprising a total of 14,359 liver transplantations. The overall incidence of BS was 13%, 12% among deceased donor liver transplantation (DDLT) patients and 19% among LDLT recipients. The overall incidence of biliary leakage was 8.2%, 7.8% in the DDLT group and 9.5% in LDLT recipients.

Sixty-six percent of BCs have been documented to occur in the first 3 months after OLT. However, particularly BS may occur up to several years after transplantation. As shown in Table 76-1 , there is a great variety of potential complications; however, BS and anastomotic leaks are the most common complications. BSs may be at the site of the anastomosis (AS) or nonanastomotic (NAS). The majority of BS (80%) are located at the anastomosis and occur in 9% to 12% of patients after DDLT. It was shown that the cumulative AS risk increases with time, being 6.6% at 1 year post OLT,10.6% at 5 years, and 12.3% at 10 years, respectively. The currently published rates of BLs, which comprise anastomotic leaks and T-tube–related leaks, range from 5% to 10%. Biliary filling defects (i.e., casts and stones) appear in approximately 5% of cases. Because DCD is a growing source of liver grafts in some countries due to a growing organ shortage, the epidemiology of BCs may alter profoundly. DCD may be performed under controlled (Maastricht categories 3 and 4) or uncontrolled conditions (categories 1 and 2). There is growing evidence that DCD is associated with distinctly higher rates of vascular complications and BCs, especially NAS, which even lead to decreased survival rates. One center analysis demonstrated BS rates as high as 37% at 3 years compared with 12% using grafts from DBD donors ( P < .0001). Several other studies confirmed higher NAS rates and lower survival rates as compared to matched DDLTs after DBD. This may even be aggravated with uncontrolled DCD. However, use of normothermic extracorporeal membrane oxygenation by two Spanish groups has dramatically improved outcomes after uncontrolled DCD transplantation.

The two most common forms of biliary reconstruction are choledochocholedochostomy (CC; duct-to-duct reconstruction) and choledochojejunostomy (CJ), usually with a Roux-en-Y loop. There is no consensus on the optimal biliary reconstruction, and considerable variability exists between transplant centers. Apart from the personal preference of the surgeon, the choice is influenced by multiple factors, such as the underlying liver disease, the size of donor and recipient bile ducts, and prior transplant or other biliary surgery. More than three quarters of DDLTs in adults are performed as CC, with only one tenth of pediatric cases. CC is preferred because it is technically easier to perform and incorporates the well-vascularized recipient common bile duct. It preserves the sphincter of Oddi, which serves as a natural barrier to reflux of enteric contents into the biliary tree, thus theoretically decreasing the risk for ascending cholangitis. It may be performed end-to-end or side-to-side and either with or without a T tube. According to current data, T-tube placement is applied in approximately 18% of duct-to-duct reconstructions. The incidence of BSs in a recent review was 10% with a T tube and 13% without a T tube, and the incidence of leakage was 5% with a T tube and 6% without a T tube.

End-to-end CC is the preferred procedure for biliary reconstruction at the majority of centers. However, a recent prospective, randomized trial of end-to-end versus side-to side CC revealed no significant difference in BCs. Some centers prefer the routine use of a T tube because it partially bridges the biliary anastomosis and allows for early postoperative assessment of bile quality and cholangiographic anatomy. Moreover, it may reduce the incidence of anastomotic stricture formation. Internal stents have been used in some patients but were reported to be associated with a high rate of serious complications, including obstruction, migration, and erosion with hemobilia. CC without the use of a T tube has been advocated by several groups to avoid postoperative BLs related to removal of the T tube. The current tendency to abandon T tubes in DDLT originates from two systematic reviews and metaanalyses that basically shared the same source of data. Both reported superior results for BS with use of a T tube, but no differences in terms of BL or overall BCs. However, several single-center prospective trials showed data favoring T-tube application. In a prospective, randomized trial, Weiss et al compared side-to side CC with T tube and without T tube and noted significant differences in BC incidence favoring the use of a T tube ( P < .0005). No complications were observed in this study after T-tube removal. Lopez-Andujar et al also analyzed the severeness of complications prospectively. The authors reported a significantly lower incidence of AS using a T tube, whereas the overall BC rate was similar. However, there were less severe complications in the T-tube group: 67% of complications in the T-tube group were grade 1 or 2 of the Clavien classification, whereas 95% of complications in the group without a T tube were grade 3a or 3b ( P < .0001). In conclusion, the latter data may challenge the current of opinion of abandoning T tubes in DDLT.

CJ is usually indicated for patients with diseased extrahepatic bile ducts (primary sclerosing cholangitis, cholangiocarcinoma, biliary atresia); prior biliary surgery; size disparities between donor and recipient ducts; reduced-size, split, and living donor grafts ( Fig. 76-1 ); retransplantation; or when the blood supply to the distal donor bile duct is suboptimal. CJ is the most common anastomosis used in pediatric liver transplant patients because of the small size or absence of bile ducts in this patient population. Problems and potential complications unique to Roux-en-Y CJ include increased surgical time, stricture or leakage of the CJ, bleeding at the jejunojejunostomy, bowel ischemia, Roux limb torsion, intestinal perforation, delayed mixing of bile with intestinal contents, altered cyclosporine absorption, and cholangitis induced by reflux of intestinal pathogens. In addition, creation of CJ restricts endoscopic evaluation of the biliary system, occasionally requiring percutaneous transhepatic cholangiography (PTC). Because magnetic resonance cholangiography (MRC) has emerged as a sensitive, noninvasive diagnostic tool, CJ-related diagnostic problems are likely to decrease.

The epidemiology and spectrum of BCs have also changed over the past decade because of the establishment of split-liver transplantation and reduced-size liver transplantation. The incidence of BCs in reduced-size liver transplantation was reported to be as high as 24%, however, with approximately 50% of complications related to cut-surface leakages. The initial experience with ex situ split-liver transplantation for infants and children resulted in an average biliary complication rate of 24% to 27%. Meanwhile, complication rates for ex situ split OLTs have been markedly reduced. In situ split OLT was reported to result in lower complication rates of the biliary system of about 0% to 15%, reflecting the better anatomical orientation during the in situ splitting process. In pediatric recipients a biliary complication rate of 8.7% to 15% was reported. A further modification has been seen with the advances in adult-to-adult LDLT. The incidence of BLs and BSs was reported to be as high as 30% in the early phase and may be attributed to initial difficulties in defining the dissection plane around the right hepatic duct and complex reconstructions including several orifices ( Fig. 76-2 ; see Fig. 76-1 ). A similar development has taken place in pediatric LDLT.

According to a recent survey of liver transplantation from living adult donors in the United States, the overall biliary complication rate was approximately 22%. Although the growing experience may result in improvement of complication rates over time, the recent review by Akamatsu et al reported even higher complication rates, with the BS rate being 19% and the BL rate being 9% after LDLT.

Pathophysiology and Risk Factors

A variety of factors have been identified as contributing to BCs. Several analyses compared risk factors for BC after deceased donor OLT before and after the Model for End-Stage Liver Disease (MELD) era. Duct-to-duct anastomosis (odds ratio [OR], 2.22), occurrence of a previous bile leak (OR, 2.24), donor age (OR, 1.01), and transplantation in the post-MELD era (OR, 2.30) were identified through a multivariate logistic regression analysis as factors associated with an increased rate of AS. The latter seemed to be associated with a higher number of DCD grafts. The only significant determinant for the development of a BL was the use of a T tube (OR, 3.38) in this study.

In an analysis by Welling et al, which excluded DCD grafts and patients with NAS, transcystic tubes (OR, 0.30) and organ preservation with histidine-tryptophan-ketoglutarate (HTK) solution (OR, 0.40) were revealed to be protective against BSs. A previous bile leak, however, was the only significant factor associated with an AS (OR, 3.63).

Although the exact mechanisms of bile duct injury unrelated to technical reasons sometimes remain elusive, the damage is believed to result from several contributing factors, including ischemic and immunological injury, preservation injury due to prolonged ischemia, and infection. A recent systematic review detected no significant differences in BCs between the use of HTK solution and University of Wisconsin (UW) solution. The same was found for UW versus Celsior solution.

Busquets et al reported the usefulness of postreperfusion biopsies in predicting bile duct complications. They correlated the extent of neutrophilic infiltrates indicative of preservation injury with BCs. Several other associated conditions were identified, including ABO incompatibility, cytomegalovirus (CMV) infection, age crossmatch, chronic ductopenic rejection, and also patients with a pretransplantation diagnosis of primary sclerosing cholangitis. The role of CMV infection has been particularly discussed. A higher BC rate was demonstrated among recipients with CMV antigenemia. Interestingly, 75% of patients with BC showed prior or concomitant CMV antigenemia, and in two patients CMV positivity in bile duct cells was proven after surgical reconstruction. Graft steatosis over 25% has also been associated with BCs.

Both the bile duct and the vascular endothelium were shown to be vulnerable to the damaging effects of humoral and cellular immune mechanisms. Both mechanisms can result in bile duct cell death and stricturing, either directly or because of a compromised vascular supply, and ultimately lead to ischemic cholangitis.

Patients with acute hepatic artery thrombosis and stenosis after transplant are at risk for ischemic AS and NAS unless the vascular flow is immediately reconstituted or adequate collateral blood flow exists. However, most BCs occur in the absence of arterial complications. In addition, low-flow phenomena in the hepatic artery unrelated to the anastomosis may occur in the case of a preexisting splenic artery steal syndrome or stenosis of the celiac trunk. Both phenomena should be subject to examination during the evaluation process of a potential recipient. Splenic artery steal syndrome may be treated sufficiently by interventional radiological methods reducing blood flow through the splenic artery and thus ameliorating portal hyperperfusion. Intermittent celiac trunk stenosis may be caused by the arcuate ligament impairing patency of the trunk during inspiration; a persistent stenosis may be congenital or related to arteriosclerosis. In the latter case an arterial jump graft is advised, and intermittent stenosis can be avoided by incision of the arcuate ligament. Further technical reasons may be related to the delicate vascular supply of the bile duct. Specifically, excess dissection of periductal tissue during organ procurement, which impairs the vascular supply of the donor’s bile duct, and the use of electrocautery for biliary duct bleeding control in both donor and recipient may contribute. In addition, excess tension on the ductal anastomosis and active bleeding from the cut ends of the bile duct before anastomosis were identified as risk factors for BCs.

The phenomenon of ITBLs was subject to several studies. Heidenhain et al identified donor age ( P = .028) and cold ischemic time ( P = .002) as significant risk factors for the development of ITBL. UW solution accounted for higher ITBL rates than HTK solution ( P = .036). The same applied to organs harvested externally versus those procured locally ( P < .001). Pressure perfusion via the hepatic artery significantly reduced the risk for ITBL ( P = .001), which was initially reported by Moench et al. The only further recipient factor that showed a significant influence was Child-Turcotte-Pugh score status C ( P = .021) in the pre-MELD era. Immunological factors were stated to have no significant impact on ITBL.

In contrast, Iacob et al demonstrated that specific chemokine receptor polymorphisms of the recipient are associated with development of post-OLT BSs, and altered cytokine profiles may contribute to enhanced fibrotic tissue remodeling and BS formation. In a univariate analysis, older donor and recipient age, partial OLT, high peak aspartate aminotransaminase levels, and CC chemokine receptor 5 delta32 loss-of-function mutation (CCR5Δ32) were associated with ITBL, whereas OLT for acute liver failure, ABO-compatible nonidentical OLT, presence of donor-specific anti-human leucocyte antigen (HLA) class II antibodies, and fractalkine receptor (CX3CR1)-249II allele were associated with AS. In the multivariate analysis, CCR5Δ32 was an independent risk factor for ITBL, whereas OLT for acute liver failure, ABO-compatible nonidentical OLT, and CX3CR1-249II allele remained predictive for AS. Serum levels of interferon-γ, interleukin-6, and interleukin-10 were significantly increased in patients with BSs. The authors advocated screening for anti-HLA antibodies as potentially useful for early identification of at-risk patients who could benefit from closer surveillance and tailored immunosuppressive regimens. Of note, other studies did not find an association between CCR-5Δ32 mutation and the development of ITBL.

Clinical Presentation

Patients with BCs may present with a variety of complaints. Complaints such as right upper quadrant abdominal pain, anorexia, abdominal distention, singultus, paralytic ileus, and right shoulder pain are typically associated with biliary tract disorders but are not specific or could be absent. Pain, a leading concern, can be totally absent in liver transplant patients due to hepatic denervation. Fever may accompany biliary leakage or cholangitis but is usually indicative for infections of various kinds. Jaundice, acholic stools, no bile drainage through the T tube, bilious ascites, or respiratory complaints caused by pleural effusion or elevation of the right diaphragm are usually late symptoms and signs. In all of these cases, a comprehensive examination is indicated. However, patients with BCs may remain asymptomatic for an extended period of time because abdominal symptoms may be masked by corticosteroid use. Bile leaks should also be considered in asymptomatic patients with unexplained elevations in serum bilirubin level, unexplained fluctuations in cyclosporine levels, bilious ascites, or intraperitoneal fluid collections on imaging studies.

Diagnostic Workup