Introduction

Biliary complications (BC) are the Achilles heel of living donor liver transplantation (LDLT), with the incidence ranging from 5 to 40%. Bile leaks (BL) account for 0–22% and biliary anastomotic strictures (BAS) occur in 4–25% of recipients, with BAS representing the majority (> 80%) [1]. Although there has been a steady decline in the BC rates following LDLT especially in some high-volume Asian centres, the rate of BAS continues to be high at around 20%–25% [2].

This high incidence of complications is attributed to various causes, the most important of which are the presence of small- and short-calibre donor duct stumps in the graft, multiple graft bile duct orifices, biliary anatomy variations in the graft, intimal damage to the duct during dissection, scar formation as a healing process in small calibre ducts or issues with blood supply to the graft duct due to compromise in the arterial supply during dissection. Flaws in surgical technique, and hepatic arterial complications in the recipient could be added factors. Non-surgical factors may also contribute, including arterial hypoperfusion of the liver graft secondary to portal hypertension, ischaemia reperfusion injury and immunological reactions [3].

The morbidity associated with biliary complications after LDLT is high, impairing the quality of life due to frequent hospitalisations. If not dealt with in a timely manner, these could progressively lead to cholestasis, bridging fibrosis, secondary biliary cirrhosis and ultimately graft failure, and rarely recipient mortality [3]. Hence, even today, BC’s are an unsolved dilemma even for experienced LDLT teams.

This review will try to summarise pertinent issues, novel insights and finally propose basic principles to be adhered to when dealing with the gamut of possible biliary complications after LDLT.

What is relatively standard, and what are the current dilemmas

In most adult LDLT’s, a duct-to-duct biliary anastomoses (DDA) is performed. Exceptions include recipients with primary sclerosing cholangitis, secondary biliary cirrhosis, those with a diseased native bile duct, and in patients undergoing a re-transplantation. Small children less than 10 kg or 2–3 years, or those undergoing LDLT for biliary atresia or other biliary pathology usually undergo a primary hepaticojejunostomy (HJ).

Endoscopic treatment with endoscopic retrograde cholangiography (ERC) and stenting is considered the first-line treatment in case of BL or BS, and percutaneous transhepatic cholangiography (PTC) with percutaneous transhepatic biliary drainage (PTBD) is usually an alternative or rescue therapy alone or in combination with ERCP (rendezvous technique), when first-line ERC fails. In patients with a pre-existing HJ, PTC is usually the first option. Other options that are rarely used in high-grade strictures (HGS) or difficult to traverse biliary strictures are combined procedures like hepatico-gastrostomy, magnetic recanalization, or re-do surgery, the latter being rare nowadays, given the expertise of interventional gastroenterologists and radiologists. Surgery continues to have a role in difficult biliary strictures not amenable to non-operative management, or in early bile leaks which are possibly best managed by early reoperation.

Pertinent dilemmas for which answers are still sought include: prediction of difficulty of endoscopic success in biliary strictures, optimal duration of indwelling stents and their planned removal, management options in HGS, and the long-term outcome of patients requiring endoscopic or percutaneous and other management strategies for BC’s. We will try and address these dilemmas in the light of studies published in the last half a decade or so.

Prevention is better than cure

Before coming to management of BCs, it is important to emphasise that good outcomes with respect to biliary anastomoses rely on what happens at the time of LDLT, a proper graft harvest, careful benching and then a meticulous biliary anastomoses. Given the prolonged morbidity, associated graft dysfunction and possibility of graft loss secondary to BC, all possible efforts should be made to prevent biliary complications in the first place. Apart from meticulous biliary reconstruction, comprehensive preoperative and intraoperative donor biliary anatomy assessment, and proper technique of donor hepatic duct isolation could be the key measures to reduce biliary complications. Preserving the blood supply of biliary system on both donor and recipient side, accurate approximation of mucosa, and tension-free anastomosis, with or without external stent may be key factors to prevent biliary complication in LDLT [1].

Modalities for managing biliary complications

Endoscopic balloon dilatation and stent placement across the stricture is indeed the first-line modality for majority of BAS after DDA, successful passage of a guidewire across the stricture being the critical step. Early BLs are also managed in this manner after prior drainage of the biloma though percutaneous route, unless reoperation is proposed. The technical success rate of ERC for BAS is 75%–80%. A majority of ERC failures are successfully treated by PTBD followed by stent internalisation, the so called rendezvous technique. A success rate of more than 90% is possible in patients with BAS using a combination of the two techniques provided the team has an experienced endoscopist and interventional radiologist.

ERC for post-LDLT strictures is technically more challenging compared to post-DDLT (deceased donor liver transplantation) strictures or ERC for benign biliary strictures due to other causes. Some of the reasons for this are the angulation of the biliary anastomoses with liver regeneration especially causing difficult access to the right posterior system in right-lobe LDLT, short stumps in the donor duct, and multiple bile duct orifices in the graft which are more common in LDLT given the not so rare biliary anatomy variations [4]. Post-ERC pancreatitis is a dreaded complication of ERC.

PTBD is useful in BAS that cannot be traversed by an endoscopic approach, provided there is sufficient dilatation of intrahepatic biliary radicles. PTC may also be helpful in some cases of completely disconnected bile duct anastomosis. Possible complications with the percutaneous approach include intraabdominal or biliary bleeding, hepatic subcapsular hematoma, BL from the surface, and very rarely, hepatic arterial pseudoaneurysm, arterioportal fistula, portal vein thrombosis [3].

The rendezvous technique is an approach using both the endoscopic and percutaneous techniques. The rendezvous technique combines double-balloon endoscopy with percutaneous approach, and is also the method of choice for treating BAS after hepaticojejunostomy in LDLT [5].

Predicting difficult to cannulate strictures: do classifications help?

Based on previous studies, there is general agreement that the clinical outcome and prognosis for the various types of BS, especially following right-lobe LDLT (majority of adult LDLTs in Asia), are significantly related to the safety and success rate of endoscopic treatment, which in turn depends on a categorisation of grade of stenosis based on bile ductography [6,7,8]. Further, the type of stricture may have an impact both on the number of stent exchanges required, as well as the possible need for an additional percutaneous interventional radiology approach to deal with difficult strictures.

There have been recent attempts at classifying post-LDLT biliary strictures based on cholangiography and cholangioscopy findings, to predict success with endotherapy.

One such classification considered the degree of inflammation as a key factor and proposed a “CUBES” classification to predict need for more sessions of endotherapy, and a longer duration for remodelling. They also demonstrated how cholangioscopy could help improving cannulation rates in difficult to traverse, or HGS [9].

Parlak et al. [10], in their proposed classification, considered characteristics of the anastomoses (number, location, and whether posterior segmental bile ducts cross the anterior ones), angle between the proximal and distal sites of the strictures, contrast enhancement pattern, and the morphology of the recipient’s proximal end of CBD in their proposed classification. They went a step ahead to say that analysis of their data could offer a clue in decision-making regarding the choice of anastomosis during LDLT in the first place.

The classification proposed by Park et al. [11] analysed anatomical types of biliary strictures, including intrahepatic biliary stricture (IHBS), after right-lobe LDLT based on data obtained during the first cholangiography, in an attempt to evaluate their prognosis. The strictures were classified as Type 1 (anastomosis), 2 (second-order branch [a, one; b, two or more; c, extended to the third-order branch]) and 3 (whole graft [a, multifocal strictures; b, diffuse necrosis]). All non-anastomotic strictures were considered IHBS, and the incidence of these was surprisingly high at 38.4% in this series. This was probably the first study to recognise the importance of, and classify IHBS in the LDLT setting. In this right liver LDLT cohort, multifocal (Type 3a) and diffuse necrosis (Type 3b) IHBS were associated with a persistent need for interventions and a reduced intervention‐free period when compared with Type 2 IHBS and Type 1 AS. They proposed that since Type 3 IHBS is associated with a higher risk of graft failure than other types, it requires more careful attention, and early re-transplantation might be considered in case of deterioration.

Difficult to traverse, or high-grade strictures (HGS): what next?

Few studies speak about long-term outcomes in patients with difficult BAS where the guide wire does not pass through the stricture site during endoscopy, and requires a PTBD with rendezvous technique.

Kim et al. [12] found that the stenting period of the rendezvous technique was longer in severe BAS than in cannulated ERC cases. They further observed that especially in the left liver, the position of the duct anastomosis changed to the right posteroinferior of the patient. Thus, the donor duct and the recipient duct are angulated, kinking worsens, and the stenting period becomes longer.

Rao et al. [13] looked at recipients with HGS, defined as those with the exclusive passage of 0.018-inch guidewire. In these patients, 5 Fr Soehendra stent retrievers were used to dilate the HGS over guidewire, using torsional movements. Technical success defined as successful placement of a biliary stent across the stricture was achieved in all the patients in this group, with no post-procedure complications or cholangitis.

Zhang et al. [14] reported on a guidewire puncture technique under cholangioscopy as a safe and feasible technique for complete stricture of biliary anastomosis after LT.

Kim MS et al. [15] reported on the increased likelihood of failure of ERC in post-DDA strictures in right-lobe LDLT, if the stricture involves the right anterior duct, if there is a bile duct anatomical variation and if a large angle is present between the recipient and graft bile duct (R-G angle). When the R-G angle was greater than 47.5°, the likelihood of ERC failure increased. They proposed the preferential use of PTBD instead of ERC as the first modality in such situations.

A “crane neck deformity” was described earlier in right-lobe LDLTs where the graft has 2 ducts by Kim ES et al. [7]. Often the anterior sector undergoes atrophy due to a compromise in venous outflow and the duct of the dominant posterior sector when involved in the stricture process leads to a duct dilation with the acute kink at the anastomotic site. PTBD as the first option, or even surgery as a salvage may be indicated in this group of patients.

Historically, one of the overwhelming questions in the management of post-LDLT strictures is the ideal duration of stenting. As reported by different authors, to achieve final success, an average of 2.2–6.3 ERCs have been required for an average period of 4.1–14.5 months [5].

The role of magnetic compression anastomosis (MCA) as a technique for treatment of disconnected bile duct after LDLT using the recently introduced through-the-scope magnet with a diameter of 2.4 mm and a length of 5 mm has been published in some studies recently. Ersoz et al. [16] reported successful biliary recanalization in six patients 13–42 days after the magnet placement procedure. The length of the biliary obstructed segment is one of the critical factors for the achievement of successful biliary recanalization, and more data are needed that demonstrate the efficacy of small magnets in the treatment of completely obstructed biliary segments longer than 20 mm. Although in a mixed population, the overall clinical success rate of MCA for bilio-biliary strictures was reported to be 87.5%, and the recurrence rate was 7.1% in a review by Jang et al. [17]. Kubo et al. [18] also reported a case of MCA for complete dehiscence of hepaticojejunostomy in a patient after LDLT.

Good success with cannulation, but what about the long term?

With growing expertise in the endoscopic management of early bile leaks and strictures post-LDLT, one of the key issues is the long-term outcomes in these LDLT recipients who need intervention for biliary complications. There is also a perceivable difference between technical success of cannulation and drainage of biliary strictures and resolution rate. The former is a temporary success, whereas resolution is a permanent solution to the biliary issue, where the patient is stent free after several endoscopic interventions [19].

Morbidity of biliary complications and impact on quality of life is of course of essence, but it is well known that unresolved BCs can have a significant impact on graft, and even patient survival. Bile leaks predispose to sepsis and graft dysfunction early after liver transplant, and especially in a partial liver graft as in LDLT, this could be responsible even for patient mortality.

A study looking at graft survival post-right-lobe LDLT in patients requiring endoscopic management of biliary complications concluded that those with persistent and unresolved biliary complications after endoscopic management and requiring prolonged biliary stenting had inferior graft survival compared to those without biliary complications and those with complications resolved by endoscopic procedures [20]. Other studies [21, 22] found contrary results. The discrepancies may be explained by the longer follow-up duration in the former study, as persistent stricture-related graft loss usually occurs beyond a year or more after LDLT.

A study by Guirguis et al. [23] found that biliary complications were a risk factor for the occurrence of chronic graft rejection and graft failure.

In addition or non-resolution, recurrence of BAS after initial successful management and indwelling stent removal is a serious issue as regards ERC or PTC for BAS and IHBS.

Park et al. [24] found that older donor age and non-B, non-C liver cirrhosis recipients had a higher incidence of stricture recurrence after stent removal. The median duration of indwelling stent in their series was 13 months and the median recurrence-free duration after final endoscopic success was also similar. They proposed that risk factors may need to be considered before stent removal. The authors did not find a significant association between stricture morphology and recurrence.

Nishikawa et al. [25] in their study found that after first resolution, the median interval to stricture recurrence was 7.6 months (range: 0.3–100.7). Anastomotic re-stenosis occurred in about half of the patients (8/18 cases) by 3 months from the first resolution, and about 80% (14/18 cases) by 3 years. New inside-stents were placed in 18 recurrent patients, and sustained resolution was achieved in 17 patients.

In a large series of LDLT recipients where percutaneous transhepatic treatment was used for biliary strictures after DDA, Lee et al. reported cumulative primary patency rates of 85.4%, 79.8%, 75.8%, and 68.4%, at 1, 2, 3, and 5 years after drainage catheter removal, respectively. Multiple drainage catheters were frequently placed in these patients because of multiple complex strictures. In their series, factors responsible for recurrence after drainage catheter removal were isolation of the sectoral duct and presence of concurrent bile leaks [26].

Principles of managing biliary complication post-LDLT: our take

The key issues in managing biliary complications using the various modalities available, endoscopy, interventional radiology and surgery, are related to the timing of intervention, choice of intervention and feasibility.

Post-LDLT, early BL noted within the first 2 weeks should be ideally managed with reoperation, in most cases culminating in a re-do biliary anastomoses using a Roux-en-Y hepaticojejunostomy. Exceptions include caudate or hilar plate leaks, or cut surface leaks occurring secondary to a slipped ligature or clip from a pedicle which could be managed by suturing the pedicle. A methylene blue test helps in identification of some non-apparent leaks at the time of reoperation. Conservative management with a controlled biliary fistula (drainage using a percutaneously placed drain) of a significant bile leak (defined on the basis of drain fluid bilirubin and amount of leak) without any future definitive management should be avoided when possible, given the risk of septic complications and future development of biliary strictures affecting patient recovery and outcome.

BL presenting beyond 2 weeks and up to 1–2 months usually are usually associated with a stricture as well. They usually present in the form of a biloma with signs of infection. Image-guided percutaneous or endoscopic ultrasound-guided drainage of the biloma, antibiotic cover, followed by ERC and stenting is usually indicated in these cases. It is ideal to give a 48–72-h period of drainage and antibiotic cover, before going ahead with ERC. Performing a MRCP prior to the procedure serves as a roadmap to the interventional endoscopist especially in case of a BAS. In the rare instance of a complete disruption/discontinuity of the biliary system, an initial PTBD may be indicated. This is indeed difficult in a non-dilated system due to the leak, and will require an expert interventional radiologist to succeed in external drainage. This can be later followed up with a rendezvous endoscopic stent placement.

Biliary strictures usually present after 4 weeks, and frequently beyond 2–3 months after LDLT. The first modality to approach them is an ERC; however, a combination of ERC and PTBD may be required to achieve success in by-passing difficult strictures not amenable to ERC alone, as noted above. Endoscopists and interventional radiologists prefer to have a MRCP roadmap, and also ensure patent vasculature to the graft before going ahead with their procedures. Presence of a concurrent biloma may be an important factor that predicts poor long-term survival in LDLT patients with biliary strictures [27].

Non-anastomotic stricture (NAS) are uncommon after LDLT and difficult to treat with ERC or PTBD and often require re-transplant as the only effective treatment option when the patient presents late with secondary effects of biliary complications on the graft. With a larger number of ABO incompatible liver transplants being performed now, rare cases of diffuse IHBS are encountered, they are usually associated with bad therapeutic outcomes and need a re-transplantation [28].

Some HGS are difficult to manage with ERC and PTBD alone. Expertise in MCA is limited. Some salvage procedures like hepatico-gastrostomy using the combined skills of the endoscopist and intervention radiologist may be used. However, surgery may be the only option in a select few of these patients in whom non-surgical interventions fail. Although difficult, there are reports of successful management with intraoperative delineation of intra-hepatic bile ducts using a hepatotomy (parenchymal transection) in some of these cases. Rarely, patients with long-standing biliary complications and secondary biliary cirrhosis may have re-transplantation as the only option.

Multiple graft bile duct orifices are known to be associated with higher rate of BCs. At our centre, multiple ducts were present in 810 (53%) amongst 1536 adult-to-adult right-lobe LDLT from 2011 to 2017. Multiple anastomoses (≥ 2) were performed in 374 (46.2%). At a median follow-up of 36 months, biliary complications rate was 16.9%, higher than overall series (13.5%; P = 0.03). However, the 5-year overall survival (OS) in this group with multiple ducts was similar to overall cohort (90.4 vs. 89%, P = 0.293). [1].

Finally, when dealing with biliary complications, and especially strictures, it is necessary to bear in mind the role of good vascular supply to the ductal system, especially patent hepatic arterial blood flow. Compromised vascularity to the bile ducts is a harbinger of future biliary complications. In recent years, we have also moved towards early conversion of duct-to-duct to bilioenteric anastomoses during surgical revascularisation of hepatic artery thrombosis (HAT) post-LDLT. A better vascularity in the bilioenteric anastomosis could help prevent the sequelae of HAT on the biliary system.