Review
Expert Review of Gastroenterology & Hepatology Downloaded from informahealthcare.com by Washington University Library on 01/13/15 For personal use only.
Biliary complications after liver transplantation Expert Rev. Gastroenterol. Hepatol. Early online, 1–20 (2014)
Bala´zs Nemes*1, Gyo¨rgy Ga´ma´n2 and Attila Doros2 1 Division of Transplantation, Institute of Surgery, Clinical Centre, University of Debrecen, Moricz Zs. krt. 22, Debrecen, H-4032, Hungary 2 Clinic of Transplantation and Surgery, Semmelweis University Budapest, Baross u. 23–25, Budapest, H-1087, Hungary *Author for correspondence: [email protected]
Biliary complications (BCs) remain one of the most outstanding factors influencing long-term results after orthotopic liver transplantation. The authors carried out a systematic overview of 1720 papers since 2008, and focused on 45 relevant ones. Among 14,411 transplanted patients the incidence of BCs was 23%. Biliary leakage occurred in 8.5%, biliary stricture in 14.7%, mortality rate was 1–3%. Risk factors: preoperative sodium level; p = 0.037, model of end-stage liver disease score >25; p = 0.048, primary sclerosing cholangitis; p = 0.001, malignancy; p = 0.026, donor age >60, macrovesicular graft steatosis; p = 0.001, duct-to-duct anastomosis; p = 0.004, long anhepatic phase; p = 0.04, cold ischemic time >12 h; p = 0.043, use of T-tube; p = 0.032, insufficient flush of bile ducts; p = 0.001, acute rejection; p = 0.003, cytomegalovirus infection; p = 0.004 and hepatic artery thrombosis; p = 0.001. The management was surgical in case of biliary leakage, and interventional radiology or endoscopic retrograde cholangiopancreatography in case of biliary stricture. Mapping of miRNA profile is a new field of research. Nemes–Doros score is a useful tool in the estimation of hepatic artery thrombosis. Management of BCs requires a multidisciplinary expert team. KEYWORDS: biliary complication • ERCP • hepatitis C • intervention • ITBL • liver transplantation • overview • radiology • surgery • vanishing bile duct
Biliary complications (BCs) remain one of the most outstanding factors influencing the long-term results after orthotopic liver transplantation (OLT). This is the main factor of late complications besides the recurrence of hepatitis C, and primary sclerosing cholangitis (PSC), and the occurrence of de novo diseases, like diabetes mellitus, and kidney impairment as a result of drug side effects. BCs can be classified according to their anatomic site, the time of occurrence and by the etiology factor leading to this complication. Since the beginning, when Thomas Starzl declared that biliary anastomosis is the Achilles’ heel of liver transplantation, this issue has never been resolved completely. The authors give an overview about the field of BCs without the attempt of exploring the entire issue and giving explanations to all questions. This overview is based on the previous paper of Akamatsu et al. [1], and focuses only on the changes since this last overview (since 2008). In some minor part, it is also based on the own data and publications. According to the operative technique, the biliary anastomosis will always be performed at the end of an exhausting operation, informahealthcare.com
10.1586/17474124.2015.967761
when the most challenging part of the procedure (the recipient hepatectomy), and also the implantation and revascularization of the liver graft has already been performed. For this reason, it seems to be logical and easy to state that early BCs are technical ones in majority, and might be partly as a consequence of lack of concentration. On the other hand, challenging situations might also be sources of BCs. These are the size mismatch between donor and recipient, the sick bile ducts of PSC patients, the scars after previous abdominal operations, the previously stented bile ducts with severe cholangitis. However, the origin of biliary problem gives a wider range from the well-known arterial thrombosis, through the acute rejection, cytomegalovirus (CMV) infection and the improper flush of bile ducts, to the miRNA profile of the excreted bile. Material & method
The authors revised the publications on PubMed between 2008 and 2013, regarding the reports of liver transplantation and BCs. A total of 1720 papers were revised as first approach, 1675 of them were excluded from
2014 Informa UK Ltd
ISSN 1747-4124
1
Review
´ ma ´ n & Doros Nemes, Ga
was preferred (67, 29 and 86.11%). Biliary leakage (BL) occurred in 8.5% in total. It was 7.7% after DDLT, while it was Total DDLT LRLT 12% after LRLT. The management of Papers included 45 31 14 BL was surgical (32%), endoscopic retrograde cholangiopancreatography (ERCP) [2–7,10,14,23,44,65,81–99] [59,61,63,100–110] Ref. no. (13.1%), percutaneous transhepatic drainTransplantations 14411 12445 1966 age (PTD) (10.6%) and retransplantation Reconstructions (reOLT) (0.6%), and unreported in the majority (40%) of the cases. Mortality Duct-to10068 8375 1693 rate was very low (2.8%) in total. Biliary duct stricture (BS) occurred in 14.7% in HJ/CJ/CD 1726 1467 259 total, 13.7% after DDLT and 19.6% Biliary leakage 764/ 570/7413 LTs 194/1615 LTs after LRLT. ERCP was used in 40.28 9028 LTs and 56.3%, respectively. Reoperationreanastomosis was done in 21 and 18%, Management ERCP 42/320 LTs 29/203 LT 13/117 LTs while reOLT in 8.3 and 2.4%. Mortality PTD 34 19 15 rate was 4% in total, 1.6% after DDLT Surgical 103 55 48 and 7.1% after LRLT. Unreported in 6% of the cases. reOLT 2 1 1 Possible risk factors for BC are summaMortality 9 2 7 rized in TABLE 2. Comparing the results of Biliary stricture 1658/ 1287/9414 LTs 371/1895 LTs Akamatsu et al., the mentioned risk factors 11309 LTs were similar in the 2008–2013 period. Risk factors on the recipient side identified Management ERCP 352/745 LTs 170/422 LTs 182/323LTs as univariate analysis results were: preoperPTD 127 81 46 ative sodium level (p = 0.037), high model Surgical 148 89 59 of end-stage liver disease score (MELD) (>25) (p = 0.048), while as multivariate reOLT 43 35 8 results were: PSC as indication versus Mortality 30 7 23 non-PSC (p = 0.001) and malignancy The authors revised the publications on PubMed between 2008 and 2013, regarding the reports of LT and (p = 0.026). Donor factors were identified BCs. Finally, 45 publications were studied according to the following reported points: cadaveric or livingrelated transplantation, the presence of BC, stricture or leakage, the management of BC. The number of as advanced age (>60), macrovesicular cases/number of analyzed liver transplants are shown in this table. steatosis (p = 0.001 univariate, and 0.05 BC: Biliary complication; CD: Choledochoduodenostomy; CJ: Choledochojejunostomy; DDLT: Deceased donor liver transplantation; ERCP: Endoscopic retrograde cholangiopancreatography; HJ: Hepaticojejunostomy; multivariate). The operative factors were LRLT: Living-related liver transplantation; LT: Liver transplantation; OLT: Orthotopic liver transplantation; DD anastomosis over biliodigestive (BD) PTD: Percutaneous transhepatic drainage. (p = 0.004) in a cohort of 1798 patients. However, a hepaticojejunostomy (HJ) the title or abstract. The further 45 publications have been anastomosis over DD was also reported (p = 0.01) in a total of studied according to the following points: cadaveric deceased 486 patients. A long anhepatic phase (p = 0.04 multivariate) and donor liver transplantation (DDLT) or living-related liver extended cold ischemic time (CIT) (>12 h; p = 0.043) further transplantation (LRLT), the presence of BC, stricture or leak- the use of a T-tube (p = 0.032 multivariate), and insufficient age or necrosis. Timing/onset and management of BC, further flush of bile ducts at harvesting/preservation (p = 0.001). As postrisk factors reported. Data were expressed in tables. Own expe- operative factors, the acute rejection (p = 0.003), the duration of rience was expressed by reviewing the file of liver transplanted intensive care treatment (p = 0.006), the CMV infection patients retrospectively between 1995 and 2013. Further evalu- (p = 0.004) and hepatic artery thrombosis (HAT) (p = 0.001) ation of literature was performed in certain topics. were identified. Our data showed that the BCs were within the range of the Results international results. Preoperative epidemiologic data as shown The results of our review are summarized in TABLES 1 & 2. Own in TABLE 3 illustrate no significant difference between the patients data are shown in TABLES 3–5. According to the data shown according to postoperative BCs. As a specialty, there are fewer in TABLE 1, a total of 45 papers were included, 31 about BC in complications of the percutaneous interventions to endoscopic, DDLT and 14 in LRLT patients. As summary of these literature patients are treated mainly by radiological interventions and data of 12,445 DDLTs and 1966 LRLTs were evaluated. Both surgery. The success rate of reinterventions is similar to the in DDLT and LRLT, a duct-to-duct (DD) biliary anastomosis reported ones in the literature (TABLE 4). The local risk factors
Expert Review of Gastroenterology & Hepatology Downloaded from informahealthcare.com by Washington University Library on 01/13/15 For personal use only.
Table 1. Deceased donor liver transplantation and living-related liver transplantation articles reviewed.
doi: 10.1586/17474124.2015.967761
Expert Rev. Gastroenterol. Hepatol.
Biliary complications after liver transplantation
Review
Table 2. Risk factors for biliary complications in the literature. Analyzed for stricture or leakage
Number of cases
Univariate analysis
Multivariate analysis
Advanced age
Leakage
256
0.003
ns
Preoperative bilirubin level
All
160
0.007
[111]
Preoperative sodium level
All
279
0.037
[99]
High MELD score (>25)
All
95
0.048
[4]
Stricture
486
Risk factor
Versus
Ref.
Expert Review of Gastroenterology & Hepatology Downloaded from informahealthcare.com by Washington University Library on 01/13/15 For personal use only.
Recipient
PSC indication
Non-PSC
0.001
[89]
0.061
0.026
[99]
141
0.08
0.05
[112]
Stricture
177
0.001
0.05
[91]
0.008
Malignancy Donor factors
Advanced age (>60) Female donor
[98]
Male
Macrovesicular steatosis
Operative factor [2]
DD
HJ
Stricture
1798
0.004
HJ
DD
Stricture
486
0.012
WIT longer than 50 min
Leakage
256
0.03
0.044
[98]
Anhepatic phase
All
279
0.053
0.04
[99]
Use of T-tube
Stricture
0.002
0.032
[93]
Improper flushing of bile ducts at preservation
Stricture
[89]
403
0.001
[14]
471
0.043
[85]
Stricture
256
0.014
Intraoperative Perioperative blood transfusion >20 U
All
271
0.048
[113]
Acute rejection
All
471
0.003
[85]
ICU stay more than 20 days
All
290
0.005
[113]
CMV infection
Stricture
486
Hepatic artery abnormity
All
279
CIT >12 h UW HTK
UW HTK
0.005
[98]
Postoperative
0.001
0.004
[89]
0.001
[99]
The authors revised the publications on PubMed between 2008 and 2013, regarding the reports of liver transplantation and biliary complications. They have been studied according to the main risk factors of biliary complications. CIT: Cold ischemic time (min); CMV: Cytomegalovirus; DD: Duct-to-duct; HJ: Hepaticojejunostomy; HTK: Histidine–tryptophan–ketoglutarate; INR: International normalized ratio; MELD: Model of end-stage liver disease; OLT: Orthotopic liver transplantation; PTD: Percutaneous transhepatic drainage; U: Unit; UW: University of Wisconsin; WIT: Warm ischemia time.
informahealthcare.com
doi: 10.1586/17474124.2015.967761
Review
´ ma ´ n & Doros Nemes, Ga
Table 3. Biliary complications and epidemiologic data in the Hungarian Liver Transplant Program.
Expert Review of Gastroenterology & Hepatology Downloaded from informahealthcare.com by Washington University Library on 01/13/15 For personal use only.
With biliary complication n = 158
Without biliary complication n = 449
p
Stenosis n = 103
Leakage n = 64
Necrosis n = 32
ITBL n=8
Donor age
37.8 ± 13.1
36.8 ± 13.9
37.3 ± 15.1
37 ± 9.2
38.3 ± 13.4
ns
Donor BMI
23.8 ± 3.3
23.5 ± 2.9
23.2 ± 3.4
24 ± 2.9
23.6 ± 3.5
ns
Donor gender (male)
57.5%
56.5%
36.4%
75%
49.5%
ns
Recipient age
43.3 ± 12.6
43.7 ± 13.1
44.6 ± 10.6
47.5 ± 10
44.6 ± 13.5
ns
Recipient BMI
25.4 ± 4.9
24.9 ± 3.6
24.2 ± 3.6
25.3 ± 3.4
25.4 ± 4.9
ns
Recipient gender (male)
64.1%
62.5%
50%
50%
55.2%
ns
CP-score
8.8 ± 1.7
8.8 ± 2
8.6 ± 2
9.3 ± 1.6
8.95 ± 2
ns
MELD-score
13.7 ± 4.8
14.2 ± 4.5
12.3 ± 4.3
12.5 ± 5
14.3 ± 6.2
ns
Indication HCV
32%
32.8%
43.8%
25%
39%
ns
Indication PBC
7.8%
3.1%
3.1%
0%
5.3%
ns
Indication PSC
11.7%
17.2%
18.8%
0%
9.7%
ns
Indication ALD
20.4%
14.1%
6.3%
25%
13.9%
ns
Preoperative epidemiologic data showed no significant difference between the patient groups according to postoperative biliary complications. ALD: Alcoholic liver disease; CP: Child–Pugh; HCV: Hepatitis C virus; ITBL: Ischemic-type biliary lesion; MELD: Model of end-stage liver disease; PBC: Primary biliary cirrhosis; PSC: Primary sclerosing cholangitis.
had been analysed too: strictures are consequence of hepatic artery stenosis (HAS), and acute rejection and duration of hepatectomy and leakages are caused by multiple factors, led by HAS or HAT, acute rejection and preoperative deviation of serum Na (sodium), while HAT (and HAS), duration and hepatectomy and intraoperative blood loss are the main risk factors of necrosis. These results are not surprising, and are in concordance with other studies (TABLE 5). Discussion General assessments
BCs remain the Achilles’ heel of liver transplantation. The overall incidence of BSs did not change since 2008 (13 and 14.6%), respectively, compared with the study by Akamatsu et al.). There is a slight increase in DDLT cases (from 12 to 13.6%), and completely unchanged in case of LRLT (19% before and after 2008). BL remained grossly stable (8.2 and 8.5%) including the DDLT patients (from 7.8 to 7.6%; before and after 2008) and in LRLT patients (9.8–12%; before and after 2008). We believe that conclusions taken from previous studies had been taken into practice (e.g., to abandon the use of T-tube, or better flush the bile ducts), however, this has not been transferred to better statistics yet, and also that technical and anatomical challenges in case of a LRLT remains a source of BCs. According to our findings, this is questionable whether interventional radiology (IR) and ERCP is the first choice for the management of all BCs. The choice of surgeon between the DD and HJ (BD) anastomosis is usually a matter of a certain situation: doi: 10.1586/17474124.2015.967761
PSC patients with a sick bile duct, high proportion of LRLT cases and small-for-size livers sometimes can only be managed by a HJ. Therefore, to give a conclusion that DD is superior over a HJ is not fair. The gold standard remains that anatomical situation must be repaired as it was, when possible. Mortality rate is generally low (4%), but it is outstandingly high (7.1%) when a BS develops in a LRLT case. In contrast to the study by Akamatsu et al., we report that surgical treatment had been preferred for BL, while endoscopy or IR for BSs. This is in concordance with the facts discussed later. A recommended algorithm is detailed in FIGURE 1. Sundaram et al. reported that the strongest risk factors for stricture development were donor age (odds ratio [OR]: 1.01), presence of a prior bile leak (OR: 2.24), use of choledochocholedochostomy/DD anastomosis (OR: 2.22) [2]. In the study by Gantxegi et al., data of 300 OLT were evaluated. Results show that arterial thrombosis or strictures were related to BCs in a range from 10 to 33.3% [3]. Benitez et al. after analyzing the results of 95 DDLTs have found that factors like arterial thrombosis, MELD score >25 and obese donor seem to increase the risk of BCs following OLT. Biliary reconstruction without a T-tube is not related to an increased risk of BL, although stricture of the anastomosis is more frequent in this group of patients. The authors have concluded that ERCP is the most common therapeutic procedure for the resolution of BCs [4]. BCs following an OLT continue to pose a serious surgical problem. The treatment of choice when dealing with BC is ERCP, which has more than a 70% success rate. If ERCP or PTD is not successful, or when biliary peritonitis is present, we Expert Rev. Gastroenterol. Hepatol.
Expert Review of Gastroenterology & Hepatology Downloaded from informahealthcare.com by Washington University Library on 01/13/15 For personal use only.
Biliary complications after liver transplantation
Review
perform a reanastomosis of the bile duct (FIGURE 2). The combined effort of the surgeon, gastroenterologist and radiologist is the most important factor for a successful treatment [5]. Perrakis et al. analyzed 245 liver transplants over a 12-year period. Their findings are similar to other centers, for example, over the past 10 years, the primary treatment regimen has moved from surgical repair to interventional endoscopy. Only when endoscopy fails, should one consider surgical repair. The treatment after HJ is still primarily surgical [6]. As already mentioned above, the mortality rate was low in total. Gastaca et al. have reported a study cohort consisting of 743 patients who had endto-end choledochocholedochostomy with a T-tube. Of these, 73 patients (9.8%) experienced BC. Anastomotic strictures occurred in 17 patients (2.3%) and non-anastomotic strictures in 2 (0.3%). Fifteen patients with anastomotic strictures were successfully treated by dilatation and stenting. Bile leakage was diagnosed in 39 patients (5.2%). Leakage occurred at the anastomosis in 15 patients (2%), and at the exit site of the T-tube in 24 patients (3.2%). Tube opening was the only treatment used in 30 patients with bile leakage (76.9%). Seven patients experienced leaks after elective T-tube removal (1%). Overall, repeated surgery to manage BCs was needed in nine patients (1.2%). The mortality rate originated from BCs was 0.13% [7].
Table 4. Biliary complications and their management in the Hungarian Liver Transplant Program.
Early & late BCs
Regarding the timing, BCs might occur immediately during and after the operation. Early complications are usually recognized as anastomotic type or technical ones, while late complications are defined as non-anastomotic BC. Early BCs
The leading clinical symptom is usually the insufficient anastomosis, resulting in a BL, fever and/or too tight suture causing a relative stenosis, with a consequent jaundice. It is questionable whether the use of an intraoperative biliary drain might be protective or even a cause for an early biliary stenosis. According to Akamatsu et al., upon the revision of 14,359 OLTs in PubMed studies, the use of a T-tube was only observed in 18% of OLTs. T-tube did not have any protective influence on the incidence of BSs or BL [1]. Bypassing the papilla of Vateri, the drain can cause an ascending cholangitis resulting in an early stricture. Therefore, many centers have abandoned the use of the biliary drain during OLT. Biliary epithelium is a very sensitive tissue, especially for hypoxia [8]. The too precise preparation during the organ harvesting can result in a damaged common bile duct without a surrounding connective tissue that give the proper vascularization for the duct. It may cause an early ‘tip necrosis’ that will result in either leakage or stenosis [9]. Also a technical aspect is the proper flush out of the bile ducts during back-table procedure. Bile salts are known to be toxic [10]. Buis et al. collected postoperative bile samples from 111 liver transplant recipients. They found that patients who developed non-anastomotic BSs were initially clinically indiscernible from patients who did not develop nonanastomotic BSs, the biliary bile salts and phospholipids informahealthcare.com
Study period 1995–2013 Number of transplantations
n = 597
DD
499/597 (83.6%)
HJ
38/597 (6.4%)
Biliary complication
158/597 (26.5%)
Stenosis
103/597 (17.3%)
Leakage
64/597 (10.7%)
Necrosis
32/597 (5.4%)
ITBL
8/597 (1.3%)
Reintervention (346 cases/150 LT)
150/597 (25%)
Surgical
99/346 (28.6%)
IR
247/346 (71.4%)
PTD
111/247 (45%)
Stenting
65/247 (26.3%)
Dilatation
37/247 (15%)
Biloma drainage
19/247 (7.7%)
ERCP
33/150 LT (5.5%)
PTD
91/150 LT (15%)
Own experience of the management of biliary complications was expressed by reviewing the file of liver transplanted patients retrospectively between 1995 and 2013. DD: Duct-to-duct; ERCP: Endoscopic retrograde cholangiopancreatography; HJ: Hepaticojejunostomy; IR: Interventional radiology; ITBL: Ischemic-type biliary lesion; LT: Liver transplantation; PTD: Percutaneous transhepatic drainage.
secretion, as well as biliary phospholipids/bile salt ratio in the first week after transplantation was significantly lower in the former group. This supports the concept that bile cytotoxicity is involved in the pathogenesis of non-anastomotic BSs. Therefore, an insufficient washout of them can result in tissue damage, and a consequent necrosis or stenosis. An intraoperative aspect is a well-known factor: ‘never make an anastomosis under tension’ [11]. If there is a long distance between the donor and recipient edge of the bile duct after the placement of the liver to the abdomen, it is better to change for a BD anastomosis, than to force a DD one. The discrepancy between the calibers of the donor versus recipient bile duct will usually not give a great technical problem. A proper suture technique can be the resolution. However, the quality of the bile duct is also very important on the recipient side as well. Certain diseases, like PSC can affect the main bile ducts. The intraoperative frozen section of ring of the recipient common bile duct will tell us whether there is an active PSC or not. To prevent a stenosis caused by a recurrent PSC, a proper step is the intraoperative histology. In summary, the early BCs are mainly technical ones, and will appear as an anastomotic complication like BL, a biloma or doi: 10.1586/17474124.2015.967761
Review
´ ma ´ n & Doros Nemes, Ga
Pirenne et al. [14] analyzed 403 OLTs at a single center and upon a univariate analysis acute rejection and the absence of flushStricture Leakage Necrosis ing the donor bile ducts were independent HAS 0.037 0.036 ns risk factors for BC. They did not find CMV as a possible factor in contrast to the HAT ns 0.078 800,000 or 800,000 IU/ml was related to a statistically highly significant BC risk (hazard ratio: 4.8 compared with viral load
Expert Review of Gastroenterology & Hepatology Downloaded from informahealthcare.com by Washington University Library on 01/13/15 For personal use only.
Biliary complications after liver transplantation Expert Rev. Gastroenterol. Hepatol. Early online, 1–20 (2014)
Bala´zs Nemes*1, Gyo¨rgy Ga´ma´n2 and Attila Doros2 1 Division of Transplantation, Institute of Surgery, Clinical Centre, University of Debrecen, Moricz Zs. krt. 22, Debrecen, H-4032, Hungary 2 Clinic of Transplantation and Surgery, Semmelweis University Budapest, Baross u. 23–25, Budapest, H-1087, Hungary *Author for correspondence: [email protected]
Biliary complications (BCs) remain one of the most outstanding factors influencing long-term results after orthotopic liver transplantation. The authors carried out a systematic overview of 1720 papers since 2008, and focused on 45 relevant ones. Among 14,411 transplanted patients the incidence of BCs was 23%. Biliary leakage occurred in 8.5%, biliary stricture in 14.7%, mortality rate was 1–3%. Risk factors: preoperative sodium level; p = 0.037, model of end-stage liver disease score >25; p = 0.048, primary sclerosing cholangitis; p = 0.001, malignancy; p = 0.026, donor age >60, macrovesicular graft steatosis; p = 0.001, duct-to-duct anastomosis; p = 0.004, long anhepatic phase; p = 0.04, cold ischemic time >12 h; p = 0.043, use of T-tube; p = 0.032, insufficient flush of bile ducts; p = 0.001, acute rejection; p = 0.003, cytomegalovirus infection; p = 0.004 and hepatic artery thrombosis; p = 0.001. The management was surgical in case of biliary leakage, and interventional radiology or endoscopic retrograde cholangiopancreatography in case of biliary stricture. Mapping of miRNA profile is a new field of research. Nemes–Doros score is a useful tool in the estimation of hepatic artery thrombosis. Management of BCs requires a multidisciplinary expert team. KEYWORDS: biliary complication • ERCP • hepatitis C • intervention • ITBL • liver transplantation • overview • radiology • surgery • vanishing bile duct
Biliary complications (BCs) remain one of the most outstanding factors influencing the long-term results after orthotopic liver transplantation (OLT). This is the main factor of late complications besides the recurrence of hepatitis C, and primary sclerosing cholangitis (PSC), and the occurrence of de novo diseases, like diabetes mellitus, and kidney impairment as a result of drug side effects. BCs can be classified according to their anatomic site, the time of occurrence and by the etiology factor leading to this complication. Since the beginning, when Thomas Starzl declared that biliary anastomosis is the Achilles’ heel of liver transplantation, this issue has never been resolved completely. The authors give an overview about the field of BCs without the attempt of exploring the entire issue and giving explanations to all questions. This overview is based on the previous paper of Akamatsu et al. [1], and focuses only on the changes since this last overview (since 2008). In some minor part, it is also based on the own data and publications. According to the operative technique, the biliary anastomosis will always be performed at the end of an exhausting operation, informahealthcare.com
10.1586/17474124.2015.967761
when the most challenging part of the procedure (the recipient hepatectomy), and also the implantation and revascularization of the liver graft has already been performed. For this reason, it seems to be logical and easy to state that early BCs are technical ones in majority, and might be partly as a consequence of lack of concentration. On the other hand, challenging situations might also be sources of BCs. These are the size mismatch between donor and recipient, the sick bile ducts of PSC patients, the scars after previous abdominal operations, the previously stented bile ducts with severe cholangitis. However, the origin of biliary problem gives a wider range from the well-known arterial thrombosis, through the acute rejection, cytomegalovirus (CMV) infection and the improper flush of bile ducts, to the miRNA profile of the excreted bile. Material & method
The authors revised the publications on PubMed between 2008 and 2013, regarding the reports of liver transplantation and BCs. A total of 1720 papers were revised as first approach, 1675 of them were excluded from
2014 Informa UK Ltd
ISSN 1747-4124
1
Review
´ ma ´ n & Doros Nemes, Ga
was preferred (67, 29 and 86.11%). Biliary leakage (BL) occurred in 8.5% in total. It was 7.7% after DDLT, while it was Total DDLT LRLT 12% after LRLT. The management of Papers included 45 31 14 BL was surgical (32%), endoscopic retrograde cholangiopancreatography (ERCP) [2–7,10,14,23,44,65,81–99] [59,61,63,100–110] Ref. no. (13.1%), percutaneous transhepatic drainTransplantations 14411 12445 1966 age (PTD) (10.6%) and retransplantation Reconstructions (reOLT) (0.6%), and unreported in the majority (40%) of the cases. Mortality Duct-to10068 8375 1693 rate was very low (2.8%) in total. Biliary duct stricture (BS) occurred in 14.7% in HJ/CJ/CD 1726 1467 259 total, 13.7% after DDLT and 19.6% Biliary leakage 764/ 570/7413 LTs 194/1615 LTs after LRLT. ERCP was used in 40.28 9028 LTs and 56.3%, respectively. Reoperationreanastomosis was done in 21 and 18%, Management ERCP 42/320 LTs 29/203 LT 13/117 LTs while reOLT in 8.3 and 2.4%. Mortality PTD 34 19 15 rate was 4% in total, 1.6% after DDLT Surgical 103 55 48 and 7.1% after LRLT. Unreported in 6% of the cases. reOLT 2 1 1 Possible risk factors for BC are summaMortality 9 2 7 rized in TABLE 2. Comparing the results of Biliary stricture 1658/ 1287/9414 LTs 371/1895 LTs Akamatsu et al., the mentioned risk factors 11309 LTs were similar in the 2008–2013 period. Risk factors on the recipient side identified Management ERCP 352/745 LTs 170/422 LTs 182/323LTs as univariate analysis results were: preoperPTD 127 81 46 ative sodium level (p = 0.037), high model Surgical 148 89 59 of end-stage liver disease score (MELD) (>25) (p = 0.048), while as multivariate reOLT 43 35 8 results were: PSC as indication versus Mortality 30 7 23 non-PSC (p = 0.001) and malignancy The authors revised the publications on PubMed between 2008 and 2013, regarding the reports of LT and (p = 0.026). Donor factors were identified BCs. Finally, 45 publications were studied according to the following reported points: cadaveric or livingrelated transplantation, the presence of BC, stricture or leakage, the management of BC. The number of as advanced age (>60), macrovesicular cases/number of analyzed liver transplants are shown in this table. steatosis (p = 0.001 univariate, and 0.05 BC: Biliary complication; CD: Choledochoduodenostomy; CJ: Choledochojejunostomy; DDLT: Deceased donor liver transplantation; ERCP: Endoscopic retrograde cholangiopancreatography; HJ: Hepaticojejunostomy; multivariate). The operative factors were LRLT: Living-related liver transplantation; LT: Liver transplantation; OLT: Orthotopic liver transplantation; DD anastomosis over biliodigestive (BD) PTD: Percutaneous transhepatic drainage. (p = 0.004) in a cohort of 1798 patients. However, a hepaticojejunostomy (HJ) the title or abstract. The further 45 publications have been anastomosis over DD was also reported (p = 0.01) in a total of studied according to the following points: cadaveric deceased 486 patients. A long anhepatic phase (p = 0.04 multivariate) and donor liver transplantation (DDLT) or living-related liver extended cold ischemic time (CIT) (>12 h; p = 0.043) further transplantation (LRLT), the presence of BC, stricture or leak- the use of a T-tube (p = 0.032 multivariate), and insufficient age or necrosis. Timing/onset and management of BC, further flush of bile ducts at harvesting/preservation (p = 0.001). As postrisk factors reported. Data were expressed in tables. Own expe- operative factors, the acute rejection (p = 0.003), the duration of rience was expressed by reviewing the file of liver transplanted intensive care treatment (p = 0.006), the CMV infection patients retrospectively between 1995 and 2013. Further evalu- (p = 0.004) and hepatic artery thrombosis (HAT) (p = 0.001) ation of literature was performed in certain topics. were identified. Our data showed that the BCs were within the range of the Results international results. Preoperative epidemiologic data as shown The results of our review are summarized in TABLES 1 & 2. Own in TABLE 3 illustrate no significant difference between the patients data are shown in TABLES 3–5. According to the data shown according to postoperative BCs. As a specialty, there are fewer in TABLE 1, a total of 45 papers were included, 31 about BC in complications of the percutaneous interventions to endoscopic, DDLT and 14 in LRLT patients. As summary of these literature patients are treated mainly by radiological interventions and data of 12,445 DDLTs and 1966 LRLTs were evaluated. Both surgery. The success rate of reinterventions is similar to the in DDLT and LRLT, a duct-to-duct (DD) biliary anastomosis reported ones in the literature (TABLE 4). The local risk factors
Expert Review of Gastroenterology & Hepatology Downloaded from informahealthcare.com by Washington University Library on 01/13/15 For personal use only.
Table 1. Deceased donor liver transplantation and living-related liver transplantation articles reviewed.
doi: 10.1586/17474124.2015.967761
Expert Rev. Gastroenterol. Hepatol.
Biliary complications after liver transplantation
Review
Table 2. Risk factors for biliary complications in the literature. Analyzed for stricture or leakage
Number of cases
Univariate analysis
Multivariate analysis
Advanced age
Leakage
256
0.003
ns
Preoperative bilirubin level
All
160
0.007
[111]
Preoperative sodium level
All
279
0.037
[99]
High MELD score (>25)
All
95
0.048
[4]
Stricture
486
Risk factor
Versus
Ref.
Expert Review of Gastroenterology & Hepatology Downloaded from informahealthcare.com by Washington University Library on 01/13/15 For personal use only.
Recipient
PSC indication
Non-PSC
0.001
[89]
0.061
0.026
[99]
141
0.08
0.05
[112]
Stricture
177
0.001
0.05
[91]
0.008
Malignancy Donor factors
Advanced age (>60) Female donor
[98]
Male
Macrovesicular steatosis
Operative factor [2]
DD
HJ
Stricture
1798
0.004
HJ
DD
Stricture
486
0.012
WIT longer than 50 min
Leakage
256
0.03
0.044
[98]
Anhepatic phase
All
279
0.053
0.04
[99]
Use of T-tube
Stricture
0.002
0.032
[93]
Improper flushing of bile ducts at preservation
Stricture
[89]
403
0.001
[14]
471
0.043
[85]
Stricture
256
0.014
Intraoperative Perioperative blood transfusion >20 U
All
271
0.048
[113]
Acute rejection
All
471
0.003
[85]
ICU stay more than 20 days
All
290
0.005
[113]
CMV infection
Stricture
486
Hepatic artery abnormity
All
279
CIT >12 h UW HTK
UW HTK
0.005
[98]
Postoperative
0.001
0.004
[89]
0.001
[99]
The authors revised the publications on PubMed between 2008 and 2013, regarding the reports of liver transplantation and biliary complications. They have been studied according to the main risk factors of biliary complications. CIT: Cold ischemic time (min); CMV: Cytomegalovirus; DD: Duct-to-duct; HJ: Hepaticojejunostomy; HTK: Histidine–tryptophan–ketoglutarate; INR: International normalized ratio; MELD: Model of end-stage liver disease; OLT: Orthotopic liver transplantation; PTD: Percutaneous transhepatic drainage; U: Unit; UW: University of Wisconsin; WIT: Warm ischemia time.
informahealthcare.com
doi: 10.1586/17474124.2015.967761
Review
´ ma ´ n & Doros Nemes, Ga
Table 3. Biliary complications and epidemiologic data in the Hungarian Liver Transplant Program.
Expert Review of Gastroenterology & Hepatology Downloaded from informahealthcare.com by Washington University Library on 01/13/15 For personal use only.
With biliary complication n = 158
Without biliary complication n = 449
p
Stenosis n = 103
Leakage n = 64
Necrosis n = 32
ITBL n=8
Donor age
37.8 ± 13.1
36.8 ± 13.9
37.3 ± 15.1
37 ± 9.2
38.3 ± 13.4
ns
Donor BMI
23.8 ± 3.3
23.5 ± 2.9
23.2 ± 3.4
24 ± 2.9
23.6 ± 3.5
ns
Donor gender (male)
57.5%
56.5%
36.4%
75%
49.5%
ns
Recipient age
43.3 ± 12.6
43.7 ± 13.1
44.6 ± 10.6
47.5 ± 10
44.6 ± 13.5
ns
Recipient BMI
25.4 ± 4.9
24.9 ± 3.6
24.2 ± 3.6
25.3 ± 3.4
25.4 ± 4.9
ns
Recipient gender (male)
64.1%
62.5%
50%
50%
55.2%
ns
CP-score
8.8 ± 1.7
8.8 ± 2
8.6 ± 2
9.3 ± 1.6
8.95 ± 2
ns
MELD-score
13.7 ± 4.8
14.2 ± 4.5
12.3 ± 4.3
12.5 ± 5
14.3 ± 6.2
ns
Indication HCV
32%
32.8%
43.8%
25%
39%
ns
Indication PBC
7.8%
3.1%
3.1%
0%
5.3%
ns
Indication PSC
11.7%
17.2%
18.8%
0%
9.7%
ns
Indication ALD
20.4%
14.1%
6.3%
25%
13.9%
ns
Preoperative epidemiologic data showed no significant difference between the patient groups according to postoperative biliary complications. ALD: Alcoholic liver disease; CP: Child–Pugh; HCV: Hepatitis C virus; ITBL: Ischemic-type biliary lesion; MELD: Model of end-stage liver disease; PBC: Primary biliary cirrhosis; PSC: Primary sclerosing cholangitis.
had been analysed too: strictures are consequence of hepatic artery stenosis (HAS), and acute rejection and duration of hepatectomy and leakages are caused by multiple factors, led by HAS or HAT, acute rejection and preoperative deviation of serum Na (sodium), while HAT (and HAS), duration and hepatectomy and intraoperative blood loss are the main risk factors of necrosis. These results are not surprising, and are in concordance with other studies (TABLE 5). Discussion General assessments
BCs remain the Achilles’ heel of liver transplantation. The overall incidence of BSs did not change since 2008 (13 and 14.6%), respectively, compared with the study by Akamatsu et al.). There is a slight increase in DDLT cases (from 12 to 13.6%), and completely unchanged in case of LRLT (19% before and after 2008). BL remained grossly stable (8.2 and 8.5%) including the DDLT patients (from 7.8 to 7.6%; before and after 2008) and in LRLT patients (9.8–12%; before and after 2008). We believe that conclusions taken from previous studies had been taken into practice (e.g., to abandon the use of T-tube, or better flush the bile ducts), however, this has not been transferred to better statistics yet, and also that technical and anatomical challenges in case of a LRLT remains a source of BCs. According to our findings, this is questionable whether interventional radiology (IR) and ERCP is the first choice for the management of all BCs. The choice of surgeon between the DD and HJ (BD) anastomosis is usually a matter of a certain situation: doi: 10.1586/17474124.2015.967761
PSC patients with a sick bile duct, high proportion of LRLT cases and small-for-size livers sometimes can only be managed by a HJ. Therefore, to give a conclusion that DD is superior over a HJ is not fair. The gold standard remains that anatomical situation must be repaired as it was, when possible. Mortality rate is generally low (4%), but it is outstandingly high (7.1%) when a BS develops in a LRLT case. In contrast to the study by Akamatsu et al., we report that surgical treatment had been preferred for BL, while endoscopy or IR for BSs. This is in concordance with the facts discussed later. A recommended algorithm is detailed in FIGURE 1. Sundaram et al. reported that the strongest risk factors for stricture development were donor age (odds ratio [OR]: 1.01), presence of a prior bile leak (OR: 2.24), use of choledochocholedochostomy/DD anastomosis (OR: 2.22) [2]. In the study by Gantxegi et al., data of 300 OLT were evaluated. Results show that arterial thrombosis or strictures were related to BCs in a range from 10 to 33.3% [3]. Benitez et al. after analyzing the results of 95 DDLTs have found that factors like arterial thrombosis, MELD score >25 and obese donor seem to increase the risk of BCs following OLT. Biliary reconstruction without a T-tube is not related to an increased risk of BL, although stricture of the anastomosis is more frequent in this group of patients. The authors have concluded that ERCP is the most common therapeutic procedure for the resolution of BCs [4]. BCs following an OLT continue to pose a serious surgical problem. The treatment of choice when dealing with BC is ERCP, which has more than a 70% success rate. If ERCP or PTD is not successful, or when biliary peritonitis is present, we Expert Rev. Gastroenterol. Hepatol.
Expert Review of Gastroenterology & Hepatology Downloaded from informahealthcare.com by Washington University Library on 01/13/15 For personal use only.
Biliary complications after liver transplantation
Review
perform a reanastomosis of the bile duct (FIGURE 2). The combined effort of the surgeon, gastroenterologist and radiologist is the most important factor for a successful treatment [5]. Perrakis et al. analyzed 245 liver transplants over a 12-year period. Their findings are similar to other centers, for example, over the past 10 years, the primary treatment regimen has moved from surgical repair to interventional endoscopy. Only when endoscopy fails, should one consider surgical repair. The treatment after HJ is still primarily surgical [6]. As already mentioned above, the mortality rate was low in total. Gastaca et al. have reported a study cohort consisting of 743 patients who had endto-end choledochocholedochostomy with a T-tube. Of these, 73 patients (9.8%) experienced BC. Anastomotic strictures occurred in 17 patients (2.3%) and non-anastomotic strictures in 2 (0.3%). Fifteen patients with anastomotic strictures were successfully treated by dilatation and stenting. Bile leakage was diagnosed in 39 patients (5.2%). Leakage occurred at the anastomosis in 15 patients (2%), and at the exit site of the T-tube in 24 patients (3.2%). Tube opening was the only treatment used in 30 patients with bile leakage (76.9%). Seven patients experienced leaks after elective T-tube removal (1%). Overall, repeated surgery to manage BCs was needed in nine patients (1.2%). The mortality rate originated from BCs was 0.13% [7].
Table 4. Biliary complications and their management in the Hungarian Liver Transplant Program.
Early & late BCs
Regarding the timing, BCs might occur immediately during and after the operation. Early complications are usually recognized as anastomotic type or technical ones, while late complications are defined as non-anastomotic BC. Early BCs
The leading clinical symptom is usually the insufficient anastomosis, resulting in a BL, fever and/or too tight suture causing a relative stenosis, with a consequent jaundice. It is questionable whether the use of an intraoperative biliary drain might be protective or even a cause for an early biliary stenosis. According to Akamatsu et al., upon the revision of 14,359 OLTs in PubMed studies, the use of a T-tube was only observed in 18% of OLTs. T-tube did not have any protective influence on the incidence of BSs or BL [1]. Bypassing the papilla of Vateri, the drain can cause an ascending cholangitis resulting in an early stricture. Therefore, many centers have abandoned the use of the biliary drain during OLT. Biliary epithelium is a very sensitive tissue, especially for hypoxia [8]. The too precise preparation during the organ harvesting can result in a damaged common bile duct without a surrounding connective tissue that give the proper vascularization for the duct. It may cause an early ‘tip necrosis’ that will result in either leakage or stenosis [9]. Also a technical aspect is the proper flush out of the bile ducts during back-table procedure. Bile salts are known to be toxic [10]. Buis et al. collected postoperative bile samples from 111 liver transplant recipients. They found that patients who developed non-anastomotic BSs were initially clinically indiscernible from patients who did not develop nonanastomotic BSs, the biliary bile salts and phospholipids informahealthcare.com
Study period 1995–2013 Number of transplantations
n = 597
DD
499/597 (83.6%)
HJ
38/597 (6.4%)
Biliary complication
158/597 (26.5%)
Stenosis
103/597 (17.3%)
Leakage
64/597 (10.7%)
Necrosis
32/597 (5.4%)
ITBL
8/597 (1.3%)
Reintervention (346 cases/150 LT)
150/597 (25%)
Surgical
99/346 (28.6%)
IR
247/346 (71.4%)
PTD
111/247 (45%)
Stenting
65/247 (26.3%)
Dilatation
37/247 (15%)
Biloma drainage
19/247 (7.7%)
ERCP
33/150 LT (5.5%)
PTD
91/150 LT (15%)
Own experience of the management of biliary complications was expressed by reviewing the file of liver transplanted patients retrospectively between 1995 and 2013. DD: Duct-to-duct; ERCP: Endoscopic retrograde cholangiopancreatography; HJ: Hepaticojejunostomy; IR: Interventional radiology; ITBL: Ischemic-type biliary lesion; LT: Liver transplantation; PTD: Percutaneous transhepatic drainage.
secretion, as well as biliary phospholipids/bile salt ratio in the first week after transplantation was significantly lower in the former group. This supports the concept that bile cytotoxicity is involved in the pathogenesis of non-anastomotic BSs. Therefore, an insufficient washout of them can result in tissue damage, and a consequent necrosis or stenosis. An intraoperative aspect is a well-known factor: ‘never make an anastomosis under tension’ [11]. If there is a long distance between the donor and recipient edge of the bile duct after the placement of the liver to the abdomen, it is better to change for a BD anastomosis, than to force a DD one. The discrepancy between the calibers of the donor versus recipient bile duct will usually not give a great technical problem. A proper suture technique can be the resolution. However, the quality of the bile duct is also very important on the recipient side as well. Certain diseases, like PSC can affect the main bile ducts. The intraoperative frozen section of ring of the recipient common bile duct will tell us whether there is an active PSC or not. To prevent a stenosis caused by a recurrent PSC, a proper step is the intraoperative histology. In summary, the early BCs are mainly technical ones, and will appear as an anastomotic complication like BL, a biloma or doi: 10.1586/17474124.2015.967761
Review
´ ma ´ n & Doros Nemes, Ga
Pirenne et al. [14] analyzed 403 OLTs at a single center and upon a univariate analysis acute rejection and the absence of flushStricture Leakage Necrosis ing the donor bile ducts were independent HAS 0.037 0.036 ns risk factors for BC. They did not find CMV as a possible factor in contrast to the HAT ns 0.078 800,000 or 800,000 IU/ml was related to a statistically highly significant BC risk (hazard ratio: 4.8 compared with viral load
