Short-term external biliary drainage in living donor liver transplantation using duct-to-duct anastomosis: a single-center experience Authors: Ji Su Kim, Tae Yoon Lee, Young Chul Yoon Published: 2025 ABSTRACT AbstractPurposeIn living donor liver transplantation (LDLT), biliary complications (BC) represent the most frequent and challenging postoperative issues. Some centers have adopted external biliary drainage (EBD) to mitigate these complications. While the conventional practice typically involves maintaining the drainage tube for 3 to 12 months, our institution pioneered a short-term placement protocol lasting 6 weeks. This study presents a single-center experience showcasing the efficacy and outcomes of this short-term EBD in LDLT.MethodsBetween January 2013 and November 2022, a total of 123 patients underwent liver transplantation at The Catholic University of Korea, Incheon St. Mary’s Hospital. Of these, a retrospective cohort study was conducted on 59 patients who underwent duct-to-duct anastomosis with EBD, excluding 53 patients who underwent deceased donor liver transplantation and 11 patients with insufficient data (due to follow-up loss or absence of EBD). EBD (feeding tube, 500 mm, 5 French) was placed across the biliary anastomosis during the operation. EBD was naturally drained for the first 1 to 3 weeks and was removed after 6 weeks.ResultsOverall, BC was observed in 22 patients (37.3%), comprising 4 cases (6.8%) of bile leakage and 18 cases (30.5%) of biliary stricture. Notably, without reoperation, all BC cases were successfully managed through minimally invasive interventions, including endoscopic retrograde biliary drainage and percutaneous transhepatic biliary drainage. Importantly, no mortalities were attributed to BC, nor were there any severe complications linked to the early removal of EBD.ConclusionA single institutional experience showed the effectiveness and safety of short-term EBD.Keywords: Anastomotic leak, Bile duct strictures, Drainage, Liver transplantation, Living donors INTRODUCTION Liver transplantation (LT) stands as a transformative therapy, offering a lifeline to patients suffering from chronic end-stage liver disease and liver failure for whom conventional treatments are no longer effective [1,2,3]. Since Thomas Starzl performed the first successful liver transplant in 1963, the field has witnessed remarkable advancements over the past 5 decades [4]. In Asia, where the scarcity of deceased donor organs remains a significant challenge, living donor LT (LDLT) has emerged as a predominant solution, with its annual utilization increasing tenfold over the past 20 years [5,6]. Despite these advancements, biliary complications (BC) persist as the unresolved Achilles’ heel of LDLT, posing significant risks of severe morbidity, graft failure, and postoperative mortality [7,8,9,10]. There are various biliary reconstruction methods, such as duct-to-duct anastomosis (DDA), hepaticojejunostomy (HJ), or choledochoduodenostomy (CD). However, the optimal approach to minimize BC remains controversial [11,12]. Surgeons typically select the most appropriate reconstruction method based on factors such as bile duct condition, size, prior surgical history, and the presence of cholangitis. Techniques like HJ or CD necessitate intestinal relocation and establish a direct connection between the bile duct and intestine, which poses the risk of recurrent cholangitis due to intestinal content reflux. Conversely, DDA offers notable advantages, including its simplicity, shorter operative time, and easier accessibility for endoscopic interventions in the event of BC. Consequently, DDA is the preferred method in most LDLT cases [11,12,13]. A method of reducing BC is to place a stent at the anastomosis site during DDA. Stents used in DDA include internal biliary drainage (IBD) and external biliary drainage (EBD). In 2012, Tranchart et al. [14] reported removable intraductal stenting in DDA in LT. IBD was removed at a median of 6.2 months after LDLT. The patients who underwent DDA with IBD had recurrent cholangitis and needed additional endoscopy to remove IBD and experienced dislocation or loss of the stent [14]. In 2014, Park et al. [15] reported DDA with EBD in LDLT. EBD was removed at a mean of 6–12 months after LDLT. EBD has statistically less early biliary leakage (BL) than non-EBD (EBD vs. non-EBD; 2/208 vs. 7/145, P = 0.036) [15]. The patients with DDA with EBD complained of discomfort in maintaining the catheter, commonly maintained for a long period of more than 6 months, and experienced catheter-related complications [15,16]. Departing from the conventional practice of maintaining drainage tubes for 3 to 12 months, our institution implemented a short-term placement protocol, limiting the duration to 6 weeks. This study reported the single institutional experience of this short-term EBD in LDLT. METHODS Ethics statement This study was approved by the Institutional Review Board of The Catholic University of Korea, Incheon St. Mary’s Hospital (No. 2024-0495-0001). It was performed in accordance with the principles of the Declaration of Helsinki. Patients Between January 2013 and November 2022, 123 patients underwent LT at The Catholic University of Korea, Incheon St. Mary’s Hospital. Of these, a retrospective cohort study was conducted on 59 patients who underwent DDA with EBD, excluding 53 patients who underwent deceased donor LT (DDLT) and 11 patients with insufficient data (due to follow-up loss or absence of EBD). Surgical techniques A single surgeon (YCY) performed all the surgeries in this study. A total hepatectomy with caval preservation was performed, followed by an allograft implantation using the donor hepatic vein to the confluence of conjoined recipient hepatic veins. Then, the portal vein was reconstructed. Once reasonable hemostasis was achieved, the hepatic artery was reconstructed, and then an intraoperative Doppler ultrasound was performed to assess the patency and satisfactory flow through the anastomosis. As for the biliary reconstruction, the graft’s and recipient’s bile duct sizes were compared, and DDA was performed if no significant discrepancy was found [17]. During DDA, EBD (feeding tube, 500 mm, 5 French [Fr]; JMS Co., Ltd.) was placed across the biliary anastomosis (Fig. 1). Depending on the number and size of bile ducts, a 5–6 Fr feeding tube was appropriately used. For biliary anastomosis, continuous running sutures using polydioxanone (PDS) 5–0 or 6–0 were performed on the posterior side, and simple interrupted sutures were performed on the anterior side. The proximal tip of the tube was placed on the intrahepatic duct and passed through the anastomosis site. A precise pinhole incision was created on the recipient’s bile duct, approximately 1 cm above the duodenum, allowing for the distal tip of the tube to be carefully exteriorized through this opening (Fig. 2). The distal tip of the tube passes through the abdominal wall and is exposed to the outside. A bile bag was connected here for biliary drainage. A tube was anchored at the hole of the recipient side duct using PDS 6–0 and to the skin using black silk 3–0. External biliary drainage. (A) Feeding tube (500 mm, 5 French; JMS Co., Ltd.). (B) The tube was placed through the anastomosis and was pulled out through the common bile duct of the recipient side. The tube was anchored to the anastomosis site and common bile duct with absorbable sutures. Surgical view of the placement of external biliary drainage. The distal tip of the tube passes through the abdominal wall and is exposed to the outside. A tube was anchored to the skin. The white arrowheads indicate the tube. The yellow arrow indicates the opening in the recipient’s bile duct (BD), where the distal tip of the tube is exteriorized. This opening is approximately 1 cm above the duodenum. The yellow dotted line represents the anastomosis site. Management of external biliary drainage EBD was naturally drained for the first 1–3 weeks and was removed after 6 weeks, postoperatively. If the liver function test (LFT) was near normal in 1–3 weeks, the tube was clamped, and, if there were no problems, it was removed in the 6th week. If the LFT was abnormal in 1–3 weeks, natural drainage of the tube was maintained and removed in the 6th week. When extracting the tube, it should be gently withdrawn without resistance. If resistance is encountered, the tube should be retracted by approximately 1 cm and completely removed the following day. Immunosuppression regimen All patients received induction immunosuppression with basiliximab, administered in 2 doses: 1 at the time of transplantation and the second on postoperative day 4. Methylprednisolone (250 mg) was given intraoperatively, followed by a gradual taper over 1 month until complete discontinuation. Standard maintenance immunosuppression consisted of tacrolimus and mycophenolate mofetil during the first postoperative year unless contraindicated or poorly tolerated. In cases of tacrolimus intolerance, cyclosporine or everolimus was used as an alternative. For ABO-incompatible (ABOi) recipients, preoperative desensitization included rituximab (300 mg/m2) administered 2 weeks before transplantation and 3–4 sessions of plasmapheresis before surgery. Analyzed factors Preoperative patient characteristics including age, sex, body mass index, American Society of Anesthesiologists physical status classification, primary diagnosis, MELD (Model for End-stage Liver Disease) score, Child-Pugh score, preoperative continuous renal replacement therapy (CRRT), ABOi, donor age, donor sex and intraoperative factors, including graft type, graft-to-recipient weight ratio, bile duct opening, bile duct size, ischemic time, operation time, and RBC transfusion were collected. Postoperative outcomes, including hepatic artery complications, BC, acute rejection, cytomegalovirus (CMV) infection, postoperative hospital stay, and mortality-related BC, were reviewed. Acute rejection included the patients who received steroid pulse therapy or were confirmed by liver biopsy within 3 months after LT. BC was classified as BL and biliary stricture (BS). A BC was defined as BL or a stricture that required surgery, interventional radiology, or endoscopic treatment. BC was confirmed by endoscopic retrograde cholangiography (ERCP) and cholangiography using EBD. Statistical analysis Continuous variables were described as median (range), and categorical variables were expressed as frequency (%). The 2 groups were compared using the Mann-Whitney U-test for continuous variables and the Fisher test for categorical variables. Risk factors of the BLs and BSs were determined by univariate analyses comparing the recipients with and without complications after LDLTs. All statistical analyses were conducted using IBM SPSS Statistics ver. 26 (IBM Corp.). Statistical significance was considered when the P-value was less than 0.05. RESULTS Baseline characteristics of patients A retrospective cohort study was conducted on 59 patients who underwent an LDLT with EBD and DDA. The median age of recipients was 54 years, and 43 patients (72.9%) were male. The reasons for LDLT for the 24 recipients were HBV-associated liver cirrhosis (HBV-LC), alcoholic liver disease (n = 21), HBV-LC with alcoholic liver disease (n = 11), HCV-LC (n = 1), autoimmune hepatitis (n = 1), and others (n = 1). Four patients had preoperative CRRT, and 10 had ABOi (Table 1). Table 1Demographics of total patients who underwent living donor liver transplantationValues are presented as number only, median (interquartile range), or number (%).ASA PS, American Society of Anesthesiologists physical status; HBV-LC, HBV-associated liver cirrhosis; HCV-LC, HCV-associated liver cirrhosis; MELD, Model for End-stage Liver Disease; CRRT, continuous renal replacement therapy; ABOi, ABO incompatible. Demographics of total patients who underwent living donor liver transplantation Values are presented as number only, median (interquartile range), or number (%). ASA PS, American Society of Anesthesiologists physical status; HBV-LC, HBV-associated liver cirrhosis; HCV-LC, HCV-associated liver cirrhosis; MELD, Model for End-stage Liver Disease; CRRT, continuous renal replacement therapy; ABOi, ABO incompatible. Perioperative characteristics and postoperative outcomes The predominant graft type was the right lobe, utilized in 58 of 59 cases, with only 1 patient receiving a left lobe graft. The median ischemic time was 73 minutes, and the median operative duration was 480 minutes. Hepatic artery thrombosis occurred in a single patient and was successfully managed with thrombolytic therapy via hepatic arteriography. BC was observed in 22 patients, comprising BL in 4 cases and BS in 18 cases. Additionally, 10 patients experienced acute rejection, while 18 developed CMV infection. The median length of postoperative hospitalization was 22 days. Notably, no mortality cases were attributable to BCs (Table 2). Table 2Perioperative characteristics and postoperative outcomes of patients who underwent living donor liver transplantationValues are presented as number (%) or median (interquartile range). GRWR, graft-to-recipient weight ratio; HAT, hepatic artery thrombosis; HAS, hepatic artery stenosis; CMV, cytomegalovirus. Perioperative characteristics and postoperative outcomes of patients who underwent living donor liver transplantation Values are presented as number (%) or median (interquartile range). GRWR, graft-to-recipient weight ratio; HAT, hepatic artery thrombosis; HAS, hepatic artery stenosis; CMV, cytomegalovirus. Catheter-related complication Among the patients, 52 had a single bile duct opening and received 1 EBD, while 7 had 2 or 3 openings and required 2 EBDs. The mean duration until EBD removal was 43 days postoperatively. Catheter-related complications occurred in 5 cases, including catheter malfunction in 3 patients and wound-related woundrelated issues at the catheter site in 2 patients. Additionally, 2 patients experienced abdominal pain following catheter removal, which resolved with conservative management without the need for intervention (Table 3). Table 3Data on external biliary drainage in patients who underwent living donor liver transplantationValues are presented as number (%) or median (interquartile range). Data on external biliary drainage in patients who underwent living donor liver transplantation Values are presented as number (%) or median (interquartile range). Development and management of biliary complications Twenty-two of the 59 recipients (37.3%) had developed a BC (4 BLs and 18 BSs). The perioperative characteristics of the 22 recipients with a BC are described in Supplementary Table 1. The 4 BLs occurred at a median of 19 days (range, 11–30 days) after the LDLT. A total of 4 BLs occurred at the anastomosis site. Three BLs developed among the patients with ABOi (2 acute rejections). One BL developed in the patients with multiple bile ducts (3 openings, 2 EBDs). All BLs were successfully managed by percutaneous drainage or endoscopic biliary drainage. BSs occurred at a median of 150 (range, 88–294.5) after the LDLT. All BSs occurred after removal of EBD (Fig. 3, Supplementary Table 1). Cases of biliary leakage (BL). The figures below show the cases of BL. A total of 4 BLs occurred at the anastomosis site. (A) In this case, the donor graft had three bile duct (BD) openings. Three donor BD openings were connected to the recipient’s single BD. The 2 external biliary drainages were inserted in the 2 largest openings (BD1, BD2). However, BL occurred in the remaining BD (BD3) without EBD. (B–D) Three BLs developed among the patients with ABO-incompatible transplantation; 2 occurred in patients with acute rejection. The yellow arrows indicate the points of BL. Comparing analysis according to the biliary complications The BLs were highly developed in patients with ABOi LT (P < 0.001). Except for ABOi LT, there were no statistically significant differences for BLs and BSs (Table 4). Table 4Comparing analysis according to the biliary complicationValues are presented as median (interquartile range) or number (%).ASA PS, American Society of Anesthesiologists physical status; HCC, hepatocellular carcinoma; ABOi, ABO incompatible; MELD, Model for End-stage Liver Disease; GRWR, graft-to-recipient weight ratio; HAT, hepatic artery thrombosis; HAS, hepatic artery stenosis; CMV, cytomegalovirus. Comparing analysis according to the biliary complication Values are presented as median (interquartile range) or number (%). ASA PS, American Society of Anesthesiologists physical status; HCC, hepatocellular carcinoma; ABOi, ABO incompatible; MELD, Model for End-stage Liver Disease; GRWR, graft-to-recipient weight ratio; HAT, hepatic artery thrombosis; HAS, hepatic artery stenosis; CMV, cytomegalovirus. The BCs are highly developed in patients with acute rejection (95% confidence interval, 1.349–51.344; P = 0.022). There were no other statistically significant risk factors for BCs. Although there was no statistically significant difference in patients with BCs, BCs were highly developed in elderly grafts (P = 0.073) (Supplementary Tables 2, 3). All 22 cases of BCs were successfully managed with endoscopic retrograde biliary drainage (ERBD) or percutaneous transhepatic biliary drainage (PTBD) without the need for reoperation. Notably, there were no mortality cases among patients who underwent ERBD or PTBD (Supplementary Table 1). DISCUSSION In LDLT, BCs are among the most prevalent and challenging postoperative issues. The incidence of BCs is notably higher in LDLT compared to DDLT [18], and early-onset BCs in transplant recipients can be a critical complication with potentially fatal consequences [7,8,9,10]. Therefore, minimizing the occurrence of BCs in LDLT is of paramount importance. Several factors contribute to the development of BCs, including the biliary reconstruction technique, biliary splinting, the type of LT procedure, organ preservation strategies, acute or chronic rejection, CMV infection, blood group incompatibility, and hepatic artery thrombosis. Additionally, both donor and recipient characteristics—such as age, type of graft (split/partial), presence of steatosis, and whether the graft originated from a donation after brain death or donation after circulatory death—play a significant role in BC risk [19]. The biliary reconstruction technique is the most significant risk factor for BCs. In the past, HJ with a T-tube was considered the gold standard technique in most transplant centers. However, its efficacy in preventing BC is controversial [20,21], leading to a decline in the popularity of HJ with T-tube. DDA with IBD was considered the subsequent option for preventing BCs, as it is easier to perform. However, patients who underwent DDA with IBD developed recurrent cholangitis and required additional endoscopic procedures to remove the IBD, and experienced dislocation or loss of the stent [14]. There are also institutions that perform DDA without stent placement [16,22,23]. These institutions are also reporting studies comparing the method of stent placement with that before stent placement [16,23]. In 2023, Jang et al. [16] reported the initial outcome of LDLT with EBD for 9–12 months. The EBD group had no BL compared with the non-EBD group (EBD vs. non-EBD: 0/40 vs. 23/150, P = 0.01). In 2018, Hong et al. [23] reported the long-term outcomes of BC after DDA in LDLT. EBD was placed for 6 months and had no BL compared with the non-EBD group (EBD vs. non-EBD: 0/116 vs. 8/24, P < 0.001). These studies show that EBD effectively reduces BL compared to non-EBD. In right-lobe LDLT at Asan Medical Center in Seoul, Korea, after regular application of small EBD and thorough dissection to maintain good arterial blood supply to the cut portion of the donor and recipient ducts, bile duct leakage disappeared [24]. EBD with DDA has several advantages. First, EBD can prevent leaks by minimizing intraductal pressure at the anastomosis site. Second, EBD can help maintain the open lumen in the early postoperative period [22]. This can be especially important when dealing with very small (<2 mm) accessory ducts or spiral ducts [22]. Third, it can help assess graft function by quantifying bile production [22]. In addition, cholangiography through EBD can be used to determine whether leaks or strictures have occurred [22,24]. If there is a biliary leak, intra-abdominal infection may occur. Hepatic arterial pseudoaneurysm rupture due to this infection may lead to re-LT and, in severe cases, death. The great advantage of reducing bile leaks using EBD is that it can prevent these complications. Although the efficacy of EBD in preventing BCs is well recognized, there are some disadvantages related to the duration of EBD. Several studies have reported methods for maintaining EBD for extended periods, ranging from 3 to 12 months [15,16,23]. However, prolonged EBD use presents several disadvantages. First, there is the risk of infection due to contamination at the EBD insertion site. Infection at the insertion site may require early removal of the EBD and antibiotic treatment. Patients also need periodic disinfection and care of the EBD insertion site. Second, there is a risk of self-removal or changes in the drain position due to pulling of the EBD. If the EBD is not safely removed by a doctor in a timely manner and/or is accidentally removed, it may lead to bile leakage from the biliary anastomosis site. It may require additional interventions, such as ERCP and antibiotic treatment. Third, damage to the EBD may occur. If the EBD is left in place for a long time, accidents such as cutting of the drainage tube or damage to the connection with the drainage bag can happen due to patient carelessness. Such damage to the EBD requires early removal. As a result, patients experience significant discomfort due to the need for prolonged attention and management of the EBD. To minimize these disadvantages, this study focuses on finding the shortest possible period of EBD use that still effectively reduces BL. To determine the duration of EBD, fistula maturity, tube size, tube type, and thread had to be considered. In cases of percutaneous transhepatic gallbladder drainage, some studies reported that track maturation occurred within 20 days [25,26,27]. It is said that BL can be minimized even if the drain is removed after a minimal 3 weeks [27]. It is also essential to use an appropriately sized stent so that the hole where it was placed is naturally occluded. Many studies have used small EBDs of 4–6 Fr [15,16,23]. Additionally, the thickness and type of thread used to secure the stent are essential to minimize tension in the biliary anastomosis and mounting hole during the stent removal process. This study used PDS 6–0 for biliary reconstruction and stent fixation. The PDS retains 50% of its tensile strength after 4 weeks and 25% after 6 weeks [28]. Therefore, the stent can be removed in about 6–8 weeks without thread resistance. BL that occurs after removal of the drainage tube in transplant patients is thought to be due to decreased fibrogenesis under immunosuppression using steroids, resulting in less development of fibrotic fistulas along the drainage tube [29]. In 2014, Park et al. [15] reported the outcome of LDLT using DDA with EBD for 6 to 12 months. And BL after stent removal was reported 11 of 208 (5.3%). In 2002, Ishiko et al. [30] reported that the BL after stent removal was 1 of 31 (3%) in LDLT with EBD for 2 months. These study results show that BL occurs after stent removal, no matter how long the stent is kept in place. Therefore, the minimum period for natural fistula development while reducing initial bile leak is 3 weeks. Also, considering the time when the thread dissolves to minimize the resistance to the bile duct when removing the drainage tube, at least 6 weeks is appropriate. This study presents a single institution’s experience with less than 6 weeks of short-term EBD in LDLT with DDA. Catheter-related complications occurred in 5 of 59 patients (8.5%), while post-removal complications were observed in 2 of 59 (3.4%), all of which resolved with conservative management. BL occurred in 4 patients (6.8%) and was successfully treated with endoscopic or percutaneous drainage. BS developed in 18 patients (30.5%), with 14 of 18 (77.8%) occurring beyond 3 months posttransplant. Notably, no mortality was associated with BCs. As previously established, ABO incompatibility and acute rejection were statistically significant factors for BCs. Despite its valuable insights, this study is limited by its retrospective nature, small sample size, and lack of direct comparison between short-term and long-term EBD. Additionally, follow-up duration varied from 1 to 10 years (2013–2022), underscoring the need for standardized long-term outcome assessments. Because of the small number of cases, the multivariable analysis could not identify independent risk factors for BL and BS. Further multicenter, randomized controlled trials are essential to refine the optimal duration of EBD placement and its impact on BSs. Above all, comparative studies involving many patients should be conducted in the future to clearly demonstrate the efficacy of EBD compared to non-EBD. In conclusion, this is the first report on short-term EBD (6 weeks) in LDLT with DDA, demonstrating its feasibility and safety. While our findings support short-term EBD as an effective strategy, further research is warranted to establish standardized placement protocols and assess long-term outcomes. Footnotes Fund/Grant Support: None. Conflict of Interest: No potential conflict of interest relevant to this article was reported. Author Contribution: Conceptualization: YCY.Data curation, Formal analysis, Methodology, Project administration: All authors.Visualization: JSK.Writing — Original Draft: YCY, JSK.Writing — Review & Editing: YCY, JSK. Conceptualization: YCY. Conceptualization: YCY. Data curation, Formal analysis, Methodology, Project administration: All authors. Data curation, Formal analysis, Methodology, Project administration: All authors. Visualization: JSK. Visualization: JSK. Writing — Original Draft: YCY, JSK. Writing — Original Draft: YCY, JSK. Writing — Review & Editing: YCY, JSK. Writing — Review & Editing: YCY, JSK. SUPPLEMENTARY MATERIALS Supplementary Tables 1, 2, 3 can be found via https://doi.org/10.4174/astr.2025.109.6.358. Supplementary Table 1 Summary of the recipients with biliary complications (n = 22) Supplementary Table 2 Characteristics of biliary complications Supplementary Table 3 Multivariate analysis of biliary complications (Cox regression) References Associated Data Supplementary Materials Supplementary Table 1 Summary of the recipients with biliary complications (n = 22) Supplementary Table 2 Characteristics of biliary complications Supplementary Table 3 Multivariate analysis of biliary complications (Cox regression)