Cohort characterization

A PTBD was attempted in a total of 56 LT adults between February 2013 and January 2023. Demographic and clinical characteristics at the time of initial PTBD insertion (= baseline) are demonstrated in Table 1. Median (Interquartile Range (IQR)) age was 49 (36 – 58) years with 36 (64%) male patients. The main reason for LT was primary sclerosing cholangitis (PSC) (41%), followed by hepatitis B or C infection (9%) and autoimmune hepatitis (7%). 14 patients had a variety of rare causes for LT, e.g. Budd-Chiari-syndrome or secondary sclerosing cholangitis (SSC), and were categorized as others. Almost all patients received a full organ transplant (86%) from a deceased donor (98%). A BDA was established in 47 patients (86%). Primary PTBD without previous ERC attempt was performed in 23 (41%) patients. A NAS was the most common indication for PTBD (55%), followed by a AS (48%). Biliary abscesses, cholelithiasis, PSC relapse and anastomotic leakage were concurring further diagnoses in addition to NAS and AS.

Table 1 Demographic and clinical characteristics at first PTBD insertion

Demographic and clinical characteristics of the whole patient cohort (n = 56) who received a PTBD at baseline (PTBD insertion, intention-to-treat). Multiple mentions are possible as indication for LT and laboratory indicators at baseline are shown for the whole patient cohort (n = 56). Values are presented as median (25% to 75% IQR) or if categorical as numbers and percentages.

Procedural characteristics and safety of PTBD

Initial PTBD insertion was successful in 55 patients (98%) (Table 2). PTBD was attempted to be directed with the distal tip into the small intestine in all patients. In 28 (50%) of patients the drainage could be placed into the small intestine at the time of initial PTBD insertion (internal). In the remaining cases an external drainage was placed initially and was either redirected to the small intestine throughout the subsequent interventions (external-internal) (12.5%) or kept as an exclusive external drainage (37.5%). Overall, an internal drainage was achieved either initially or through subsequent PTBD exchanges in 62.5% of patients (Table 2). Of note, in four patients two simultaneous PTBDs were established. In three of these patients a PTBD was placed in both liver lobes simultaneously. For simplification in patients with two drainages, only the first PTBD placement, which was exclusively external in all patients (three with additional external-internal and one with additional internal through following procedures), was taken into account. External-internal and external PTBDs were categorized as initially external PTBDs for further analysis unless otherwise specified. The median size of the first PTBD was 8 (8–9) French. Biliary dilatations in addition to push dilatation by PTBD exchanges, were performed by bougienage in 48 patients (86%) and subsequently, strictures were resolved in 63% of patients.

Table 2 Procedural and safety characteristics

Procedure related complications were defined as complications directly during or after PTBD admission and were noted in a total of 13 patients (23%). The most common procedural related complication was bleeding in six (11%), followed by cholangiosepsis in three (5%) and cholangitis in two (4%) patients (Table 2).

Procedural characteristics and procedural-related complications (defined as directly related to the PTBD procedure) for the PTBD insertion are shown. Of note, in four patients two simultaneous PTBDs were established. In three of these patients a PTBD was placed in both liver lobes simultaneously. For simplification in patients with two drainages only the first PTBD is taken into account (in all four patients extern). External-internal and external PTBDs were categorized as initially external PTBDs for further analysis unless otherwise specified. Values are presented as median (25% to 75% IQR) or if categorical as numbers and percentages.

Median follow-up time was 39.8 (15.0–67.2) months. PTBD could be removed in 31 patients (55%) overall after a median of 4 (1–9) PTBD changes and a median time of 12 (4–34) months (Table 2, Supp. Fig. 1A). Interestingly, PTBDs placed in patients with AS could be removed numerically more often than in those placed in NAS (Supp. Fig. 1B). PTBDs needed to be reinserted in 13 patients (23%), respectively (Table 2, Supp. Fig. 1C) without any differences between AS and NAS biliary strictures (Supp. Fig. 1D).

Longitudinal changes in laboratory indicators of cholestasis, graft function and inflammation following PTBD

Only patients with a complete biochemical follow-up over twelve months were taken into account in order to perform statistical tests accordingly (n = 37) (Supp. Table 1). Cholestasis parameters, such as alkaline phosphatase (AP), gamma glutamyl transferase (GGT), bilirubin as well as the C-reactive protein (CRP) were elevated at baseline representing cholestasis and cholangitis. Cholestasis was reduced significantly six and twelve months after PTBD admission (e.g. GGT: baseline: 322 (157–767) U/l vs 12 months: 172 (73–340) U/l, p = 0.004) (Fig. 2A–D). The concentration of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) representing liver injury were decreased following PTBD (Fig. 2C, Supp. Fig. 2A). Albumin serum concentration and cholinesterase (CHE), representing liver synthesis performance, increased significantly (CHE: baseline: 4.89 (3.06–6.91) kU/l vs 12 months: 6.31 (4.02–7.16) kU/l, p = 0.004) (Fig. 2 G,H). The commonly used surrogate for liver graft dysfunction, the MELD score, was significantly reduced six and twelve months after PTBD admission (baseline: 11 (9–16) vs 12 months: 10 (9–13), p = 0.027) (Fig. 2F). Of note, bilirubin (Fig. 2D) and creatinine as parameters of the MELD score (Fig. 2E) decreased, while INR remained unchanged (Supp. Fig. 2B).

Fig. 2

Changes in laboratory indicators of cholestasis, graft function and inflammation. Laboratory indicators of cholestasis (A–D), graft function (G–H) and inflammation (I) are shown as violin plots for baseline (before PTBD insertion) and follow-up (6 and 12 months following PTBD insertion) for all patients with a complete follow-up (n = 37). Values are presented as median (25% to 75% IQR). One-way ANOVA or Friedmann test were used accordingly. p-values < 0.05 were considered significant. ALT alanine aminotransferase, AP alkaline phosphatase, CHE cholinesterase, CRP C-reactive protein, GGT gamma glutamyl transferase, IQR interquartile range, MELD model of end stage liver disease, PTBD percutaneous transhepatic biliary drainage

CRP representing inflammation was also significantly reduced during the follow-up (baseline: 18 (6–52) mg/l vs 12 months: 8 (5–23) mg/l, p = 0.036), however white blood cell count remained unchanged (Fig. 2I, Supp. Fig. 2C).

Laboratory indicators of cholestasis, inflammation and graft function are shown for baseline (PTBD insertion) and follow-up (6 and 12 months) for all patients with a complete follow-up (n = 37). Values are presented as median (25 % to 75 % IQR). p-values < 0.05 are considered significant.

Biliary complications, need for surgical biliary revision and liver re-transplantation following PTBD

Prior to PTBD insertion biliary complications, such as cholangitis (76.4%) and cholangiosepsis (26.8%) were common and a surgical biliary revision was needed in 12 (21.4%) patients (Table 1).

Following PTBD insertion, biliary complications occurred in a total of 23 (41.1%) patients after a median of 7.3 (IQR 1.2–22.6) months during a one-year follow-up. Patients receiving a PTBD for AS were significantly less likely to develop biliary complications than patients with a NAS (cumulative incidence: AS 22.6% vs NAS 67.5%) (Fig. 3A, B). A surgical re-intervention or re-LT was performed in a total of seven patients (12.5%) during 12-month follow-up. Here, the nature of biliary stricture did not affect if a surgical re-intervention or liver re-LT needed to be performed (Fig. 3A, B). In a multivariate competing risk regression model (Fig. 3C), AS stenosis was the strongest predictor of subsequent biliary complications (adjusted sub-Hazard Ratio (SHR): 0.26, 95% CI: 0.09–0.78, p = 0.016). Patient age increased the risk for cholangitis (SHR 1.03, 95% CI: 1.00–1.05, p = 0.042), whereas a higher initial MELD score was associated with a numerically lower risk for cholangitis (SHR 0.91, 95% CI: 0.83–0.99, p = 0.026). Neither patient gender and body mass index (BMI) nor baseline AP and CRP concentration had an influence on the incidence of consecutive biliary complications following PTBD (Fig. 3C).

Fig. 3

Incidence of biliary complications, surgical biliary revision and liver re-transplantation during 12 months follow up after PTBD. A cumulative incidence function for biliary complications and need for surgical re-intervention or re-LT (competing risk) stratified by biliary stricture type (anastomotic- vs. non-anastomotic stricture) (A) and a multistate comparison of the cumulative incidence of biliary complications and need for surgical re-intervention or re-LT (B) are shown for a one year follow up demonstrating a lower incidence of biliary complications in patients with an anastomotic stricture. (C) Multivariable competing risk regression with subhazard ratios (SHR) and 95% CIs is shown as a forest plot with corresponding table. p-values < 0.05 were considered significant. AP alkaline phosphatase, AS anastomotic stricture, BMI body mass index, CRP C-reactive protein, IQR interquartile range, LT liver transplantation, MELD model of end stage liver disease, NAS non-anastomotic stricture, PTBD percutaneous transhepatic biliary drainage, SHR subhazard ratio

Long-term analysis within a follow-up period of 60 months since first PTBD insertion, confirmed the main finding of AS being the major predictor of reduced biliary complications following PTBD (Supp. Fig. 3A–C).

Short- and long-term survival in patients receiving PTBD following LT

12-month overall survival was 80% (95% CI: 70–91%) (Fig. 4A). Different potential predictors of 12 months survival were analyzed by multivariate cox-proportional hazard regression (Fig. 4B, C). Although AS was numerically associated with a lower SHR for 12-months mortality, this was not statistically significant. Of all investigated parameters, only higher CRP at baseline was associated with increased mortality at 12 months following PTBD (SHR 1.01, 95% CI: 1.00–1.02, p = 0.027). Long-term survival up to 60 months was 57% (Supp. Fig.4A) and was numerically higher in patients with AS compared to those with NAS (70% vs 46%, p = 0.11) (Supp. Fig. 4B). Again, higher baseline CRP concentration and also lower BMI was associated with higher mortality at 60 months (Supp. Fig. 4C).

Fig. 4

12-month survival following PTBD. Survival after initial PTBD insertion is shown as Kaplan–Meier graphs for all patients (A) or patients stratified by biliary stricture type (B) for a one-year follow-up. (C) Multivariate survival cox regression with hazard ratios (HR) and 95% CIs is shown as a forest plot with corresponding table. p-values < 0.05 are considered significant. p-values < 0.05 were considered significant. AP alkaline phosphatase, AS anastomotic stricture, BMI body mass index, CRP C-reactive protein, HR hazard ratio, IQR interquartile range, LT liver transplantation, MELD model of end stage liver disease, NAS non-anastomotic stricture, PTBD percutaneous transhepatic biliary drainage

Interestingly, patients, in which an internal PTBD drainage into the small intestine was achieved at first attempt had a significantly better 12 months survival than patients, in which internalization could either never or only on consecutive PTBD attempts be achieved (Fig. 5, p = 0.018) 25 (93%) patients with an initially internal placed PTBD survived after 12 months, whereas ‘only’ 19 (68%) patients with an initially external PTBD survived during the same follow-up. However, long term follow-up to 60 months did no longer show significant survival differences dependent on PTBD placement strategy (Supp. Fig. 5).

Fig. 5

Influence of PTBD internalization on 12-month survival. Survival after initial PTBD insertion is shown as Kaplan–Meier graphs for patients stratified by ability to internalize PTBD on first attempt (Internal) vs. on subsequent attempts (External-Internal) vs. never (External) for a one-year follow-up. PTBD Percutaneous transhepatic biliary drainage