Biliary drainage in palliative and curative intent European patients with hilar cholangiocarcinoma and malignant hilar obstruction: a retrospective single center analysis

Abstract

Background and aims

Relief of cholestasis in hilar cholangiocarcinoma is commonly undertaken in both curative and palliative treatment plans. There are numerous open questions with regard to the ideal biliary drainage strategy – including what constitutes clinical success (CS). In the existing data, curative patients and patients from the Western world are underrepresented.

Patients and methods

We performed a retrospective analysis of patients with complex malignant hilar obstruction (Bismuth-Corlette II and higher) due to cholangiocarcinoma who underwent biliary drainage at a German referral center between 2010 and 2020. We aimed to define CS and complication rates and directly compare outcomes in curative and palliative patients.

Results

56 curative and 72 palliative patients underwent biliary drainage. In patients with curative intent, CS was achieved significantly more often regardless of what definition of CS was applied (e.g., total serum bilirubin (TSB) < 2 mg/dl: 66.1% vs. 27.8%, p = < 0.001, > 75% reduction of TSB: 57.1% vs. 29.2%, p = 0.003). This observation held true only when subgroups with the same Bismuth-Corlette stage were compared. Moreover, palliative patients experienced a significantly greater percentage of adverse events (33.3% vs. 12.5%, p = 0.01). Curative intent treatment and TSB at presentation were predictive factors of CS regardless of what definition of CS was applied. The observed CS rates are comparable to published studies involving curative patients, but inferior to reported CS rates in palliative series mostly from Asia.

Conclusions

Biliary drainage in complex malignant hilar obstruction due to cholangiocarcinoma is more likely to be successful and less likely to cause adverse events in curative patients compared to palliative patients.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12876-024-03429-y.

Introduction

Among biliary tract cancers (BTC) tumors arising in the hilar region of the biliary tree (perihilar cholangiocarcinoma; pCCA) also known as Klatskin tumors are relatively common [1, 2]. pCCA of Bismuth-Corlette (BC) stage II or higher commonly presents with jaundice due to malignant hilar obstruction (MHO). While most patients present with advanced disease and will receive palliative treatment, some are amenable to a curative treatment, i.e., a treatment plan that includes a surgical procedure aimed at R0 resection of the tumor.

Relief of cholestasis, i.e., restoration of bile flow, as evidenced by a reduction in total serum bilirubin (TSB) to near normal levels, is usually attempted both in the curative and in the palliative treatment approach. With the curative approach, there is some evidence that a lower TSB prior to major hepatic resection is associated with reduced postoperative mortality [3]. With the palliative approach, persistent cholestasis with high TSB may limit the range and dose of anti-neoplastic therapies that can be administered. This consideration also applies for potentially curative patients since many will receive adjuvant treatment after resection and at our center bridging systemic therapy is used in some complex cases prior to resection [4, 5].

Moreover, cholestasis may cause pruritus and puts patients at risk of cholangitis. Finally, jaundice is perceived as a highly visual stigma of disease and hence should be addressed for reasons of patient comfort.

Currently, guidelines advise an interventional drainage therapy for malignant jaundice in both, curative and palliative patients [6, 7]. Biliary drainage in pCCA causing MHO is challenging since the liver lobes are functionally separated by the tumor and require separate drainage. Moreover, variations in biliary anatomy commonly require individualized approaches. Endoscopic biliary drainage (EBD) by means of endoscopic retrograde cholangiography (ERC), percutaneous transhepatic drainage (PTBD) or a combination of both is most commonly used. In Asian countries nasobiliary drainage catheters are used more commonly than in Western countries and there is some evidence suggesting they offer a favorable safety profile when compared with EBD [8]. A metanalysis of endoscopic treatment modalities for malignant biliary obstruction, but not exclusively MHO, suggest EBD and PTBD offer comparable results in terms of technical success with acceptable complication rates [9].

Endoscopic ultrasound-directed interventions provide a good option to drain the left lobe but do not usually reach the right liver lobe in patients with MHO.

Despite the large body of studies addressing biliary drainage in MHO, there is no clear consensus on the optimal modality of drainage (EBD vs. PTBD), type of stent (plastic vs. metal), or the required extent of drainage (unilateral vs. bilateral or limited vs. complete).

Among studies addressing biliary drainage in MHO, there is some overrepresentation of palliative patients and patients from Asia (Supplement Table S1). Moreover, studies are almost exclusively retrospective and there is wide variation in the definition of what constitutes clinical success in biliary drainage therapy. The aims of our study were threefold: first, since data from Western countries is limited and there is no comparison of curative and palliative cases, we retrospectively evaluated drainage modalities and outcomes in both curative and palliative patients observed at our center. Second, we evaluated the type, number and associated complications of interventions performed to achieve biliary drainage. Thirdly, we aimed to test the hypothesis that treatment plan (i.e., curative vs. palliative intent), TSB at presentation, bilateral or complete drainage and Bismuth-Corlette stage are predictors of successful biliary drainage in patients with pCCA and MHO.

Materials and methods

Study design, patients and outcome

In this monocentric study, we conducted a retrospective analysis of all patients treated due to a perihilar cholangiocarcinoma at the Center of Visceral Medicine (Asklepios Hospital Hamburg Barmbek) between 2010 and 2020.

The decision on treatment plans (curative versus palliative) was discussed by the hospital’s interdisciplinary tumor board. Treatment decisions are based on the current national and European guidelines [10, 11]. In addition, individual patient characteristics are considered to find the best tailored therapy approach. At our center, most BC 2 and 3 patients that are deemed fit for surgery are directed towards a curative concept. BC 4 patients are directed towards a curative concept if they are clearly fit for surgery and resection seems technically feasible. In BC 4 cases this will often require pre-operative induction of hepatic hypertrophy by portal or hepatic vein embolization. Curative patients, are surgically resected as soon as this is feasible. In many cases resection is followed by adjuvant systemic therapy. In cases where hepatic hypertrophy is necessary to make resection feasible, this is initiated first. If a significant delay until surgery is anticipated, we do use bridging systemic therapy in selected cases. In palliative patients’ biliary drainage is usually followed by combination chemotherapy (i.e. cisplatin + gemcitabine) or – more recently – immune-chemotherapy (i.e. cisplatin + gemcitabine + durvalumab) according to current guidelines and studies.

To be included in this analysis, patients had to be over 18 years of age with proven perihilar cholangiocarcinoma (i.e., histopathologically proven or clinically diagnosed in the cholangiogram by tumor board consensus). We excluded patients who received no interventional treatment to achieve biliary drainage, those who received an endoscopic ultrasound-guided biliary drainage procedure, and those with incomplete clinical data, missing laboratory parameters, anatomically uncertain perihilar localization and Bismuth-Corlette stage I. Baseline characteristics (i.e., sex and age), history of liver disease, laboratory data and peri-interventional data (i.e., number of interventions, localization of biliary drain, endoscopic technique and complications) were assessed from the electronic patient files. At our center, we aim to achieve resolution of jaundice and ideally normal or near normal TSB in both curative and palliative patients as long as their functional status is largely preserved, i.e., patients categorized as Eastern Cooperative of Oncology Group [ECOG] status 0–2 and some ECOG 3 patients. The drainage strategy at our center is similar for curative and palliative patients: we perform EBD first, changing to PTBD if sufficient drainage is not achieved. Throughout the study period when performing EBD, we aimed to place bilateral plastic stents in order to avoid implanting materials that cannot be exchanged during surgery or later endoscopic revisions. This strategy is in line with the German guideline published in 2022 [11] while the European guideline of 2017 recommend the use of uncovered metal stents in palliative patients [7]. Given the European recommendation and accumulating data favouring metal stents in palliative pCCA, we have begun to use bilateral uncovered metal stents in palliative patients with poor functional status and/or poor prognosis; these patients were not included in the analysis. We collected data on all EBD and PTBD interventions performed until near normal TSB was achieved or the patient’s drainage situation was considered not further improvable by two experienced interventional endoscopists. TSB measured before the first intervention performed at our center was defined as pre-therapeutic TSB; the lowest TSB measured within 100 days after the last intervention included in the analysis, but before any other therapeutic measures that might potentially change the drainage situation (e.g., surgery) was defined as the post-therapeutic TSB. To exclude bias due to different time points at which the final post-intervention TSB was evaluated, we also analyzed clinical success in 3 different subgroups according to observation period, i.e. final post-intervention TSB within 10 days, 11 to 20 days and over 20 days after the final drainage procedure.

Definitions

We used four alternative definitions of clinical success (CS) that have previously been used in studies addressing biliary drainage in pCCA with MHO:

• TSB < 2 mg/dl.

• TSB < 5 mg/dl.

• > 50% reduction of pretherapeutic TSB.

• > 75% reduction of pretherapeutic TSB.

Complications were all adverse events that occurred related to interventional treatment. Cholangitis was diagnosed in accordance with the Tokyo classification for cholangitis and findings from endoscopy [12]. Clinically significant bleeding was defined as any intervention-related bleeding leading to transfusion of blood products or termination of the intervention. Post-endoscopic retrograde cholangiography (ERC)-pancreatitis was defined as abdominal pain plus elevated serum lipase above three times the upper limit of normal. Perforation was diagnosed either through direct intraprocedural visualization of the defect or abdominal pain plus free air on imaging. The classification of adverse events was performed by the novel classification for adverse events in gastrointestinal endoscopy (AGREE) [13].

Statistical analyses

Study data was collected in Microsoft Excel and analysed with IBM SPSS Statistics 25 and R Core Team 2022. Categorical data are expressed as count (percentage) and continuous data are expressed as medians (quartiles). Depending on the particular data set, Mann-Whitney test, Wilcoxon rank sum test, or Pearson’s chi-squared test with Yates’ continuity correction was used to test for statistical significance. Binary logistic regression was used to identify predictors of clinical success. The results are presented as odds ratios (OR) and confidence intervals (CI) of 95%.

Ethics board approval and consent to participate

The study protocol was submitted to the ethics board of the Hamburg chamber of physicians and was approved without a formal vote because of the retrospective nature of the study. The need for informed consent was waived by the ethics board of the Hamburg chamber of physicians (reference numbers 2022-300198-WF, WF-078/20). The study was registered in the German Clinical Trials Registry (DRKS) on 7th of August 2023 (DRKS00032372).

Results

Baseline characteristics

We identified 128 patients with a diagnosis of MHO due to pCCA treated at the Center of Visceral Medicine at Asklepios Hospital Hamburg-Barmbek between 2010 and 2020 (Table 1). Of these, 77 (60.2%) were histologically confirmed. Median age was 72 (interquartile range 63–78) years and a non-significant male predominance (57%) was observed. All patients were discussed in our interdisciplinary tumor board: 56 (43.7%) patients were deemed suitable for a curative approach and for 72 (56.3%) patients, palliative therapy was recommended. Patients planned for curative therapy were significantly younger and had lower Bismuth-Corlette stage.

Table 1.

Baseline characteristics baseline-characteristics of patients presenting for perihilar Cholangiocarcinoma. Mann-Whitney test or Pearson’s Chi-squared test with Yates’ continuity correction was used to test for statistical significance

		Overall (n = 128)		Curative (n = 56)		Palliative (n = 72)		p
Age
Median (Q1, Q3)		72 (63, 79)		68 (61, 75)		76 (68, 82)		< 0.001*
Mean (SD)		71 (12)		67 (11)		73 (12)
Sex
Male % (n)		57 (73)		57.1 (32)		56.9 (41)		1
Female % (n)		43 (55)		42.9 (24)		43.1 (31)
Bismuth Classification, % (n)
II		10.2 (13)		14.3 (8)		6.9 (5)		0.29
IIIA		28.1 (36)		39.3 (22)		19.5 (14)		0.02*
IIIB		27.3 (35)		33.9 (19)		22.2 (16)		0.2
IV		34.4 (44)		12.5 (7)		51.4 (37)		< 0.001*
Pretherapeutic TSB (mg/dl)
Median (Q1, Q3)		10 (5, 16)		9.6 (3.1, 16.5)		10.6 (7, 16.5)		0.17
Mean (SD)		11.1 (7.2)		10.2 (7.3)		11.7 (7.1)

Interventional data

The median pre-therapeutic TSB was 10 (4.93, 16.5) mg/dl in the overall population and did not differ between palliative and curative patients (9.6 mg/dl vs. 10.6 mg/dl, p = 0.168). To achieve optimal biliary drainage, a total of 335 interventions, i.e., 2.6 interventions per patient on average, were performed (Table 2). These included 260 (77.6%) patients who underwent EBD and 75 (22.4%) who underwent PTBD. 47 patients underwent at least one PTBD procedure. Stent exchange due to occlusion was performed in 4 cases in the palliative subgroup but was not necessary in curative patients.

Table 2.

Interventional data Pearson’s Chi-squared test with Yates’ continuity correction was used to test for statistical significance

		Curative(n = 56)		Palliative (n = 72)		p
		n	%	n	%
Number of EBD procedures
0		0	0	1	1.4	1
1		23	41.1	19	26.4	0.16
2		19	33.9	29	40.3	0.58
>=3		14	25	23	31.9	0.51
Number of PTBD procedures
0		35	62.5	46	63.9	1
1		13	23.2	14	19.5	0.76
2		7	12.5	6	8.3	0.63
>=3		1	1.8	6	8.3	0.22
Unilateral drainage		25	44.6	38	52.8	0.46
Bilateral drainage		31	55.4	34	47.2	0.46
Complete drainage		37	66.1	30	41.7	0.01*

Bilateral drainage, i.e., drainage of both the right and the left lobe, was performed in 55.4% of curative and 47.2% of palliative patients; this difference was not statistically different (p = 0.46). Since variations in biliary anatomy are common, bilateral drainage does not necessarily equate with complete drainage of all liver segments. For the purpose of our retrospective study, complete drainage was defined as the absence of undrained liver segments in a post-therapeutic ultrasound or – if no post-therapeutic ultrasound report was available – bilateral drainage. Complete drainage was achieved in 52% of the overall population and significantly more often in curative patients (66.1% vs. 41.7%, p = 0.01).

Outcomes: clinical success

In our study, an average 79% reduction of the initial TSB was achieved, with a significantly greater reduction in curative compared to palliative patients (Fig. 1). Patients in the curative subgroup reached CS significantly more frequently than did those in the palliative subgroup, irrespective of what definition of CS was applied (Table 3): A TSB under 2 mg/dl was attained in 66.1% of curative patients and 27.8% of palliative patients (p < 0.001), as was a TSB under 5 mg/dl (94.6% vs. 59.7%, p < 0.001). Likewise, a decrease of more than 75% of initial TSB was achieved more frequently in curative patients (57.1% vs. 29.2%, p = 0.003) as was a 50% or more decrease (76.8% vs. 51.4%, p = 0.006). When considering only patients with Bismuth-Corlette stage III, CS was more frequently achieved in curative patients, irrespective of CS definition. When comparing patients with Bismuth-Corlette stage IV disease, TSB < 2 mg/dl was achieved significantly more often in curative patients (85.7% vs. 24.3%, p = 0.007). For TSB reduction under 5 mg/dl or reduction under 50%, we observed a sub significant trend towards higher efficacy in curative Patients. The percentage of patients who showed a decrease of more than 75% of the initial TSB did not differ significantly between therapy groups in patients with Bismuth-Corlette stage IV (Supplement S2). Next, we evaluated subgroups of patients according to how soon after the final drainage procedure the minimal TSB used to determine CS was measured. When observing TSB-reduction within 10 days after intervention, TSB < 2 mg/dl and TSB < 5 mg/dl was achieved significantly more often in curative patients (56.3% vs. 14% p = 0.003 and 100% vs. 46.5%) p < 0.001). For all other clinical success definitions and in all other subgroups (i.e. determination of final TSB determining CS within 10 days, 11 until 20 days and more than 20 days) we observed numerically higher CS rate in curative patients, but in these smaller subgroups the differences fell short of statistical significance (Supplement Tbl. S3).

Fig. 1.

Total Serum Bilirubin pre- and postinterventional in curative and palliative patients. Total Serum Bilirubin (TSB) before biliary drainage (pre) and lowest value within 100 days after intervention (post) in curative patients compared to those with a palliative therapy intention. Within group differences were tested for statistical significance using Wilcoxon signed ranks test; the difference in delta-TSB between the groups was tested using two-sided Wilcoxon-Mann Whitney test

Table 3.

Outcome Mann-Whitney test or Pearson’s Chi-squared test with Yates’ continuity correction was used to test for statistical significance

			Curative (n = 56)	Palliative (n = 72)	p
Pre-therapeutic TSB
	Median (Q1, Q3)		9.6 (3.1,16.5)	10.6 (6, 16.5)	0.17
	Mean (SD)		10.2 (7.3)	11.7 (7.1)
Post-therapeutic TSB
	Median (Q1, Q3)		1.45 (0.83, 2.8)	3.5 (1.7, 8.3)	< 0.001*
	Mean (SD)		1.9 (1.4)	5.5 (5.1)
Delta TSB
	Median (Q1, Q3)		7.8 (1.9, 12.4)	5.2 (1.8, 9.4)	0.11
	Mean (SD)		8.3 (6.8)	6.2 (7.2)
Percent reduction
	Median (Q1, Q3)		79 (62.8, 88.9)	50.4 (25.6, 78.7)	< 0.001*
	Mean (SD)		71.1 (23.2)	38.1 (70.7)
Clinical Success
< 2 mg/dl % (n)			66.1 (37)	27.8 (20)	< 0.001*
< 5 mg/dl % (n)			94.6 (53)	59.7 (43)	< 0.001*
> 50% TSB reduction % (n)			76.8 (43)	51.4 (37)	0.006*
> 75% TSB reduction % (n)			57.1 (32)	29.2 (21)	0.003*

Outcomes: adverse events

A total of 24.2% of patients experienced complications, with a significantly greater rate in the palliative compared to the curative subgroup (33.3% vs. 12.5%, p = 0.01) (Table 4). The most frequent adverse events were infectious complications, post-ERC pancreatitis, perforation and clinically relevant bleeding. The majority of complications were Class II or IIIa, only two complications had a complication rated Grade IV or higher. Both occurred in the palliative subgroup (Supplement Tbl. S4).

Table 4.

Complications Pearson’s Chi-squared test with Yates’ continuity correction was used to test for statistical significance

	All (n = 128)		Palliative (n = 72)		Curative (n = 56)		p
	n	%	n	%	n	%
Any complication	31	24.2	24	33.3	7	12.5	0.01*
Pancreatitis	6	4.7	4	5.6	2	3.6	0.92
Perforation	4	3.1	3	4.2	1	1.8	0.8
Pancreatitis + perforation	1	0.8	1	1.4	0	0	1
Infectious	12	9.4	11	15.3	1	1.8	0.02*
Cholangitis	8	6.3	8	11.1	0	0	0.03*
Pyogenic liver abscess	3	2.3	2	2.8	1	1.8	1
Sepsis	1	0.8	1	1.4	0	0	1
Clinically relevant bleeding	3	2.3	2	2.8	1	1.8	1
Bilioma	2	1.6	1	1.4	1	1.8	1
Cardiac arrest	1	0.8	1	1.4	0	0	1
Hemothorax	1	0.8	0	0	1	1.8	0.9
Stent induced cholecystitis	1	0.8	1	1.4	0	0	1

Predictors of CS

To evaluate the hypothesized predictors of clinical success, we conducted binary logistic regression models for all 4 CS cut-off values (Table 5). A curative therapeutic approach was a predictor of CS irrespective of the definition used; the highest OR was observed for TSB reduction under 5 mg/dl. TSB at presentation was also a significant predictor of CS irrespective of the definition used. BC stage IV was identified as a predictor of not achieving TSB reduction under 5 mg/dl and achieving a 75% reduction of TSB at presentation. Bilateral drainage was not a significant predictor of CS in this analysis.

Table 5.

Predictors for clinical success clinical important variables were included in the binary logistic model to test for predictive factors

CS-Grade	Corrected R2	Predictor	OR	95% CI		p
				Lower	Upper
< 2	0.26	Curative Therapy	5.3	2.21	12.88	< 0.001*
		TSB at presentation	0.92	0.87	1	0.007*
		Bilateral drainage	0.9	0.4	2.04	0.8
		BC < IV	0.9	0.35	2.28	0.81
< 5	0.43	Curative Therapy	7.28	1.82	29.1	0.005*
		TSB at presentation	0.88	0.81	0.95	0.001*
		Bilateral drainage	2.74	0.96	7.9	0.06
		BC < IV	3.44	1.18	10.04	0.02*
> 50%	0.36	Curative Therapy	4.12	1.54	11.04	0.005*
		TSB at presentation	1.19	1.1	1.29	< 0.001*
		Bilateral drainage	1.63	0.68	3.99	0.27
		BC < IV	2.4	0.91	6.34	0.78
> 75%	0.34	Curative Therapy	3.15	1.23	7.84	0.01*
		TSB at presentation	1.14	1.07	1.21	< 0.001*
		Bilateral drainage	2.22	0.95	5.2	0.07
		BC < IV	3.21	1.18	8.74	0.02*

Discussion

In pCCA patients with MHO, biliary drainage is a critical part of the treatment of patients planned for curative or palliative therapy. While several studies have addressed this topic, there is no clear consensus on the optimal strategy for biliary drainage with regard to several important questions, such as first line drainage procedure, type of stent, extent of drainage and target TSB. Most studies address palliative patients only and patients from the Western world are somewhat underrepresented in the existing literature. This is the first direct comparison of biliary drainage in palliative and curative patients with MHO. We observed that – despite comparable drainage strategies and also when controlling for Bismuth-Corlette stage – interventional biliary drainage in curative patients was robustly associated with better clinical success rates and fewer adverse events. In a separate subgroups analysis where patients were grouped according to how soon after the final drainage procedure the minimal TSB determining CS was measured, we consistently observed a trend towards higher CS rates in curative patients, although most comparisons between the subgroups did not reach statistical significance. The lack of statistical power due to the small size of subgroups is a likely explanation for this.

Current European guidelines advise against routine preoperative biliary drainage in curative patients with pCCA and MHO, but favor of a decision “by a multidisciplinary team based on patient characteristics and institutional experience”. They do not recommend any specific TSB goal [7]. Likewise, a statement from a consensus conference by a global group of surgeons merely noted that some centers favor preoperative biliary drainage aiming for an TSB of 2–3 mg/dl while other centers do not require routine biliary drainage; consensus existed that biliary drainage is indicated under specific circumstances such as cholangitis, planned preoperative anti-neoplastic therapy, preoperative portal vein embolization, hyperbilirubinaemia-induced malnutrition, hepatic insufficiency or renal insufficiency [3, 7]. In a European prospective multicenter study, Farges et al. observed that low preoperative TSB was associated with a favorable outcome (90-day mortality) of right-sided but not left-sided hepatectomy [14].

For palliative patients, the European guidelines recommend drainage of more than 50% of liver volume using uncovered self-expanding metal stents [7]. A large retrospective series from China supports this approach since it revealed a superior CS (TSB below 2 mg/dl in this case) of 99% compared to unilateral stenting of bilateral plastic stents [15]. In contrast, the more recent German guideline recommends multiple plastic stents arguing that metal stents that cannot be removed from the bile ducts make reinterventions more technically difficult [11]. This problem becomes more relevant as palliative CCA patients live longer as novel targeted and immunological treatment options become available. The rationale for treating cholestasis in palliative CCA is (1) relief of jaundice and pruritus, (2) protection from or treatment of cholangitis and (3) enabling anti-neoplastic treatment since gemcitabine, that is part of the current first line regime for CCA, is thought to have increased toxicity in patients with elevated bilirubin [16, 17]. Thus, biliary drainage to reduce TSB is an important goal in both curative and palliative patients as it allows for subsequent surgical or medical cancer-directed therapies.

In the absence of conclusive evidence and strong guideline recommendations at our center we pursue a biliary drainage strategy that is comparable in both palliative and curative patients aiming for complete drainage in most cases and utilizing EBD first, followed by PTBD if EBD does not achieve complete drainage. Our analysis indicates that despite the similar approach and even if the extent of hilar tumor growth, i.e., the Bismuth-Corlette stage, biliary drainage is more likely to be successful and adverse events are rarer in curative compared to palliative patients. This information is relevant to manage patient and provider expectations and guide planning of biliary interventions.

Our investigation suggests that TSB decrease to near normal values, i.e., below 2 mg/dl, is challenging in a real world setting; in our patient cohort this was achieved only in 66.1% and 27.8% in curative and palliative patients, respectively. This is in contrast to a recent series from China in which unilateral and bilateral plastic stent placement in palliative patients with MHO resulted in TSB below 2 mg/dl in 65.4% and 71.4%, respectively [15]. This is despite a high percentage of Bismuth-Corlette stage IV individuals in that cohort. However, the patients evaluated were younger and a broad range of tumor entities causing MHO such as hepatocellular carcinoma, gallbladder cancer and metastases of distant primary tumors were included. Similarly, Liang et al. reported CS rates (in this case TSB decrease > 75%) of 50.6% for unilateral and 69.6% for bilateral drainage (EBD or PTBD) in Chinese palliative patients with MHO compared to 29.2% in our cohort [18]. Again, patients were younger and tumor entities other than CCA were included. In a large older palliative cohort from Europe a functional success rate (TSB reduction of 75% or more, intention to treat) of 84.8% was reported (EBD using plastic stents) [19]. However, this group included a majority of Bismuth-Corlette stage I lesions. These patients were excluded from our analysis because in these cases the bile ducts of both liver lobes are still functionally connected and drainage is thus considerably easier. Our results are in line with results of Sanchan et al., who only included palliative patients at a large volume center in Thailand with complex MHO due to pCCA [20].

Overall, differences in included tumor entities and stages, drainage strategy (e.g. EBD vs. PTBD; plastic vs. metal stents; unilateral vs. bilateral drainage), age and comorbidities, and end points (definitions of clinical success; per protocol vs. intention to treat) make comparisons across studies very difficult. Moreover, it is unclear if patient cohorts from different global regions are directly comparable. For instance, patients in Asian cohorts tend to be younger and the percentage with underlying parasitic disease as a key risk factor for CCA is much greater.

Several studies indicate that for EBD in palliative patients metal stents are probably superior to plastic stents in terms of short term clinical success and duration of stent patency [15, 19–23]. However, there is concern that reinterventions will be more difficult with metal stents in situ that cannot be removed [11]. This may be of high relevance in a setting where most palliative patients receive state-of-the-art anti-neoplastic therapy. For this reason, over the course of the study period, we have favored an approach where EBD with plastic stents was performed first and, if this was unsuccessful, PTBD was used. Meanwhile we employ bilateral metal stenting in patients with very limited prognosis due to poor functional status and/or no good medical treatment options.

The situation may be different for curative patients, but data in this group is very limited. Coelen et al. performed a randomized controlled trial comparing EBD and PTBD in patients with potentially resectable pCCA [24]. The trial was stopped prematurely after inclusion of 54 patients due to high all-cause mortality in the PTBD group, although excess mortality was not clearly related to complications of the drainage procedure. Clinical success (TSB reduction by 20% or more and sonographic relief of cholestasis in the future liver remnant) was seen in 63% and 78% in the EBD and PTB group with a high rate of crossover form EBD to PTB. Thus, this study is very comparable to our study both in terms of drainage strategy and clinical success.

To our knowledge, this our analysis presented here is the first one directly comparing success and adverse events of biliary drainage between palliative and curative patients with complex MHO due to CCA. We found a curative intent treatment to be associated with successful biliary drainage in terms of cholestasis relief and a lower rate of adverse events compared to palliative patients. This association held true even though a very similar drainage strategy was applied, a similar number of interventions was performed and was still evident when controlling for Bismuth-Corlette stage. As this study is retrospective, we cannot provide a definitive explanation for these differences, but the most likely explanation is that palliative patients were older, likely had poorer functional status and more advanced tumor stage. Even when comparing subgroups with the same Bismuth-Corlette stage the palliative group may still have had more advanced tumor growth. These same factors may also account to an extent for a higher rate of complications and severe complications in the palliative group. The reported complication rate in palliative patients matches several studies [15, 21, 25–27]. To our surprise, curative patients in our study faced fewer complications than presented in the literature [24, 28]. General health and frailty are typical parameters discussed in tumor boards and therefore, patients in the curative group are likely to be healthier. Thus, patients in the palliative group are likely to be more frail – a known risk factor for adverse events in endoscopy [29]. Another factor contributing to the higher rate of adverse events may be that palliative remain with their stents in situ for a longer period of time and are more likely to suffer cholangitis eventually, while most curative patients eventually undergo surgical resection.

Our work has several limitations: first, the number of patients in our cohort is limited and this makes generalization of the findings more difficult. However, perihilar cholangiocarcinoma is a rare entity, especially in the West. The fact that we exclusively investigated this entity and did not include MHO due to other malignancies can be viewed as a strength of our study. Still, a larger sample size, preferably obtained by a prospective multicenter trial would provide more robust and reliable data.

Second and very importantly, the study is retrospective and there are numerous known and unknown differences between the groups. One issue following from the retrospective design is that the time points when the minimal post therapeutic TSB determining CS was measured, are heterogenous. Our analysis of patient subgroups with similar time points are an attempt to mitigate this limitation. In a prospective trial, one would clearly define the time points of TSB measurements as well as the overall observation period. Further, patients in a palliative setting face a poor prognosis with a short overall survival time. Therefore, drainage management could be less aggressive compared to patients in a curative setting.

Third, there are other relevant outcomes beyond TSB reduction in patients with pCCA such as complications or quality of life. Quality of life is very difficult to determine reliably from hospital records and is hence not addressed in this paper. Likewise, biliary events such as cholangitis and stent occlusion are difficult to determine reliably in a retrospective manner. Moreover, it is conceivable that curative and palliative patients were seen at different intervals and in different clinical settings which might bias observed complication rates. We thus opted to focus more of TSB reduction than on other outcomes. Furthermore, some important factors such as functional status and coexisting liver disease could not reliably be extracted for all patients from available records. Only a negligible number of patients received a therapy with metal stents for the reasons outlined above. Finally, while we could analyze unilateral vs. bilateral drainage there are frequent variations in biliary anatomy so that ultrasound confirmed complete drainage may be the more relevant parameter, but this information was not available in all cases.

Conclusion

In conclusion this retrospective analysis of European patients with complex MHO due to CCA provides a number of insights: (1) compared to curative patients, palliative patients have poorer clinical responses when biliary drainage is performed. (2) They are also more likely to suffer adverse events and these tend to be more severe. These findings provide potentially important information when planning biliary drainage in palliative patients and may help to manage both patient and provider expectations. (3) Studies on biliary drainage in MHO are extremely heterogeneous. Of note, we found success rates comparable to published series in curative patients, while recent series carried out on mostly Asian palliative patients report better clinical success rates even when trying to control for the drainage technique applied.

We think our investigation gives a relevant insight in the challenges clinicians face when preparing patients with MHO due to perihilar cholangiocarcinoma for surgery and/or systemic therapy. Even though new treatment options are emerging (i.e. endoscopic ultrasound-directed interventions, coated stents or metal stents) and may allow a decrease of TSB to near normal in even more patients, our study show encouraging outcomes attained with the currently available tools especially in the important subgroup of patients where potentially curative treatment is possible. Nonetheless, more data from different global regions and ideally more standardized study designs are needed.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

pCCA

Perihilar cholangiocarcinoma

EBD

Endoscopic biliary drainage

TSB

Total serum bilirubin

PTBD

Percutaneous transhepatic biliary drainage

BC

Bismuth-Corlette

MHBO

Malignant hilar biliary obstruction

CS

Clinical success

AGREE

Adverse events in GI endoscopy

OR

Odds ratio

CI

Confidence interval

ERC

Endoscopic retrograde cholangiography

ESGE

European Society of Gastrointestinal Endoscopy

Author contributions

Conceptualization: J.D., T.H., K.W., K.O., A.S.; Data curation: J.D., L.B., T.S, T.H., T.R.; Formal Analysis: J.D.; Investigation: T.B., T.H. Supervision: T.H.; Visualization: J.D., T.H.; Writing–original draft: J.D.; Writing–review & editing: all authors.

Funding

None.

Data availability

The datasets used and analyzed during the current study are available from the corresponding author on reasonable request due to data safety concerns by the hospital administration.

Declarations

Ethics approval and consent to participate

The study protocol was submitted to the ethics board of the Hamburg chamber of physicians and was approved without a formal vote because of the retrospective nature of the study. Further, the need for informed consent was waived by the ethics board of the Hamburg chamber of physicians (reference numbers 2022-300198-WF, WF-078/20).

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.

External funding sources specific to this paper

None.

Human ethics and consent to participate

Not applicable.

Clinical trial number

The study was registered in the German Clinical Trials Registry (DRKS) on 7th of August 2023 (Clinical trial number DRKS00032372).