Long-term outcome of combined radiologic and surgical strategy for the management of biliary complications after pediatric liver transplantation

Objectives We aimed to analyze the risk factors for management failure of BC after pediatric liver transplantation (pLT) by retrospectively analyzing primary pLT performed between 1997 and 2018 (n = 620 patients). Results In all, 117/620 patients (19%) developed BC. The median (range) follow-up was 9 (1.4–21) years. Patient survival at 1, 5 and 10 years was 88.9%, 85.7%, 84.4% and liver graft survival was 82.4%, 77.4%, and 74.3% respectively. Graft not patient survival was impaired by BC (p = 0.01). Multivariate analysis identified the number of dilatation courses > 2 (p = 0.008), prolonged cold ischemia time (p = 0.004), anastomosed multiple biliary ducts (p = 0.019) and hepatic artery thrombosis (p = 0.01) as factors associated with impaired graft survival. The number of dilatation courses > 2 (p < 0.001) and intrahepatic vs anastomotic stricture (p = 0.014) were associated with management failure. Thus, repeated (> 2) radiologic dilatation courses are associated with impaired graft survival and management failure. Overall, graft but not patient survival was impaired by BC. Supplementary Information The online version contains supplementary material available at 10.1186/s13104-024-06735-6.

However, managing biliary strictures and leaks in pLT can be challenging, requiring repeated radiologic procedures, long-term biliary drainage and even repeated surgeries [22,23] with a significant impact on graft, patient survival and quality of life [6,9,24].The outcomes of repeated PTC-C in pLT are inconsistent, complicating the determination of factors associated with treatment failure and the establishment of long-term outcomes [18].
Therefore, the primary aim was to identify factors leading to the failure to manage these complications.The secondary aim was to define long-term outcomes of BC.

Patients
All primary pLT procedures performed at our center between 1997 and 2018 with a minimal follow-up of 1 year were retrospectively reviewed.Exclusion criteria were: previous pLT and multiple organ transplantation to exclude BC related to prolonged cold ischemia times.Bile leaks that did not have a percutaneous cholangiography or surgery and were managed conservatively were excluded.All BC identified by the need for either percutaneous cholangiography or surgery were included.Patient characteristics including age at pLT, gender, indication for pLT, type of donor, type of liver graft, cold ischemia time, number of bile ducts within the anastomosis, type of biliary anastomosis, hepatic artery thrombosis/stenosis and episodes of biopsy-proven acute rejection were recorded.Initial immunosuppression was based on cyclosporine or tacrolimus and steroids from 1997 to 2003 and on basiliximab plus tacrolimus since 2003.Of note, split and reduced liver grafts were all performed ex-situ.Details concerning the surgical procedure of biliary anastomosis, diagnostic criteria, treatment of BC and outcome are detailed in Additional file 1.

Statistical analysis
SAS version 9.4 (SAS Institute, Cary, NC) was used for statistical analysis.Data were quoted as median (range).Comparisons were performed.Categorical variables were compared with the Chi-square test or a Fisher exact test.
Continuous data were compared with a Mann-Whitney test.Survival curves and confidence intervals were estimated with the Kaplan-Meier method.To evaluate the hazard ratios for the number of courses of PTC-C, the extended Cox model for time-dependent variables was used.Hazard ratios were estimated with univariate Cox proportional hazards models.Wald test was performed to determine the variables in univariate models with a p-value of < 0.20.These variables were then used to construct the multivariate model, with backward elimination.All the results were considered significant at a p-value of < 0.05.

Patient characteristics
In the 620 pLT recipients, 117 (19%) developed a BC including 16 bile leaks (strictures were identified in all 16 patients).One hundred six patients had a primary PTC-C and 11 primary surgery.
The demographic data and clinical characteristics of the 117 patients with BC are summarized in Table 1.Median follow-up was 9 years (range 1.4-21 years) (Table 1).

Risk factors for management failure or graft loss
In the primary PTC-C group, factors associated with management failure were, in univariate analysis: hepatic artery thrombosis, the intrahepatic vs. anastomotic stricture, lithiasis at first PTC-C, and the number of PTC-C courses (Table 2).In multivariate analysis, failure of PTC-C was associated with the number of PTC-C courses (p < 0.001) and intrahepatic vs anastomotic stricture (p = 0.014) (Table 2).
In the primary PTC-C group, factors associated with graft failure or patient death were, in univariate analysis: prolonged cold ischemia time, hepatic artery thrombosis, lithiasis at first PTC-C, duration of the first biliary drainage longer than 3 weeks, and the number of PTC-C courses (Additional file 4).In multivariate analysis, prolonged cold ischemia time (p = 0.004), hepatic artery thrombosis (p = 0.010), the number of PTC-C courses (p = 0.008), and two-bile ducts anastomosis (p = 0.019) were associated with graft loss or patient death (Additional file 4).The drain duration of more or less than 3 weeks was not associated to PTC-C treatment failure in the multivariate analysis (p = 0.370) (Additional file 4).
Finally, the study found that the management failure rates in the primary PTC-C group after the 4th and 5th PTC-C course were 42% and 50%, respectively, which were higher than the overall 28% failure rate of the surgical management group, although the difference was not statistically significant (p = 0.435 and p = 0.399 respectively).

Long-term survival after BC management
In our pLT cohort, the overall survival (OS) rates at 1, 5, and 10 years were 82.4%, 77.4%, and 74.3%, respectively.Patients without BC had OS rates of 86.7%, 83.5%, and 82.2% at 1, 5, and 10 years, respectively, and graft survival (GS) rates of 79.5%, 75.9%, and 73.2%, respectively (Additional file 5).Patients with BC had OS rates of 98.3%, 95.2%, and 94% at 1, 5, and 10 years, respectively, and GS rates of 94.9%, 83.8%, and 78.9%, respectively (Additional file 5).However, when using the extended Cox model for time-dependent variables to manage the immortal time bias and compare patients with and without BC, patients with BC had a shorter GS than patients without BC [HR = 2.483(1.568-3.934),p < 0.001].The same Cox model did not show impaired patient OS for patients with BC compared to patients without BC [HR = 0.948(0.449-2.002)].

Discussion
BC are among the most frequent complications after pLT [3-5, 9, 12, 13, 25], their treatment consisting mainly of PTC-C.To our knowledge, this is the first study to report an association between PTC-C failure rates and repeated radiologic procedures.For graft failure, besides repeated radiologic procedures, long drainage periods were also a risk factor.The high failure rates of PTC-C with repeated radiologic procedures (i.e.> 2) found in our series might be explained by fibrosis with retraction and scarred biliary tissue following repeated radiologic instrumentations of the bile ducts, etiologic findings already published by others [26,27].Even if centers report evaluating a surgical revision after 1-2 failed PTC-C for BC occurring 12 months after pLT [18], our findings provide evidencebased support for the clinician's decision.Because failure rates increase significantly after 3rd dilatation course, it appears that those strictures are unlikely to respond to subsequent dilatation attempts.
Long drainage periods after dilatation for biliary strictures in pediatric LT seem to be associated with high success rates in the available literature (Additional file 6) but success rates in the treatment of BC are difficult to compare [28]: frequently definitions are lacking; when described, the success reported might be technical, clinical or associated with improved patency rates of the biliary anastomosis [22,29,30].The difference between our study and the existing literature could be attributed to the different management strategies: short drainage period and external drainage in our center versus longer drainage period and internal-external drainage in the other series [18].The Society of Pediatric Liver Transplantation analysis failed to show the optimal management type (PTC-C, ERCP) or timing between PTC-C procedures (more or less than 3 weeks) underlining the need for a prospective study [17].The success of short draining periods that are used in our experience supports the concept of biodegradable stents for the treatment of biliary strictures in pLT [31].
The prolonged graft ischemia time, hepatic artery thrombosis, the number of duct anastomoses, and the number of PTC-C courses were also significant, underlying the intrication of graft survival with the risk of repeated and probably unsuccessful percutaneous procedures in small bile duct structures [26,27,32].Several other studies investigating the risk factors for BC found donor, immunological and ischemic factors [3,11,17,33].Nevertheless, their impact on treatment failure was not further investigated in our series.The high hepatic artery thrombosis rates found are explained by the long period of the study starting in 1997.Furthermore, the thrombosis rate encompasses both intra and postoperative occurrences with a significant resolution observed within the first month, yielding an 8% thrombosis rate at one month.Multiple biliary anastomoses are known to increase the risk of BC in pLT [6,34]; furthermore, in our series, the multivariate analysis identified multiple bile ducts anastomosis to be associated with increased risk of graft failure after PTC-C management of BC.A trend without statistical significance was already reported [34,35].While the diameter of the bile duct used for reconstruction is indeed a critical factor in biliary anastomosis, the type of graft was not associated in our series with PTC-C treatment failure or graft failure.The ratio of patients having either successful or favourable outcomes at the end of follow-up was as high as 72% with 48% having PTC-C successfully dilated strictures and free of recurrent strictures throughout extended follow-up.Throughout the literature, there are different PTC-C protocols (Additional file 6) and results vary accordingly.The length of the PTC-C course used within our series is shorter than the average of the above-mentioned series, approaching the model of the "three-session protocol" of Oggero et al. [36].The authors report a 71.4% success rate after the first PTC-C [36].The rationale for this short period [7-10 days] between dilatations is based on the hypothesis of a first dilatation injury followed by a repair process occurring in the 45 days after tissue injury (i.e.first dilatation session) with a peak deposit of fibroblasts at day 30 [37].By keeping the drainage periods short, we hypothesized a reduced drainageinduced inflammation and thus a shortened healing time.
The place of surgical treatment in the management of BC after pediatric LT is not well defined: in the past, it was the only treatment available for BC [3,38], being progressively indicated only in the management of bile leaks and in case of failure of interventional radiology [35,39,40], even if surgery is still being used as first-line therapy in some centers [21].Surgical revision was successful in our series for saving strictures not amenable to radiologic treatment.Even if not statistically significant, failure rates after the 3rd and 4th PTC-C were higher than after the surgical treatment, pointing to the need to evaluate the interest for earlier evaluation for surgical management after the 2nd PTC-C.
In our series, 66% of patients managed surgically for BC required further intervention, either PTC-C or additional surgery.This is in contrast to another series where only 20% needed repeat surgery, with a retransplantation rate of 12% [21].The discrepancy may stem from the latter series exclusively using surgery for all biliary complications, tracked over a variable follow-up period averaging 92 months [21].While a second surgery could reduce hospital stays and be less painful than multiple PTC-Cs, the benefits and drawbacks of each approach warrant further comparative research to establish optimal management strategies.

Limitations
This retrospective series offers an overview of the longterm outcomes of BC over 21 years.While management bundles changed little during the study period,  Furthermore, the study was performed over a long study period.Some variables for data that could have been of interest (i.e.donor data, immunological data) were not available in the revised records.
To conclude, in the long-term, graft and not patient survival is impaired for patients with BC after pediatric LT.Repeated (i.e.> 2 PTC-C courses) radiologic procedures are associated with impaired graft survival and management failure.Thus, an alternative surgical treatment could be proposed beyond two dilatation courses in the management of BC after pediatric liver transplantation.

Fig. 1
Fig. 1 Flowchart of patients with biliary complications after pediatric liver transplantation.PTC-C percutaneous transhepatic cholangiography with balloon cholangioplasty, Re LT redo-liver transplantation

Table 1
Characteristics of the patients with biliary complications, LT: Liver transplantation

Table 2
Univariate and multivariate analysis of risk factors for primary percutaneous transhepatic cholangiography with balloon cholangioplasty (PTC-C) group treatment failure.LT: Liver transplantation

Table 2
(continued)Bold was used for statistically significant values a Analyzed as a time-dependent variable b Continuous variables medical tools and fine-tuning of clinical practices may have impacted the outcomes.Patient age, and thus weight and size, vary from neonates to nearly teenagers.