Outcomes and complications of biliary drainage for malignant biliary obstruction: National prospective study

Abstract

Background and study aims

National data suggest that biliary drainage for malignant obstruction is associated with high complication rates and early mortality. This study examined factors associated with poor outcomes.

Patients and methods

RICOCHET was a national, prospective audit of patients with pancreatic cancer or malignant biliary obstruction between April and August 2018. This analysis reviewed outcomes including complications within 7 days and 30-day mortality following biliary drainage and associated factors.

Results

Biliary drainage was attempted in 773 patients, of which, 78.7% were successful at first attempt; but if unsuccessful, only 37% of subsequent attempts succeeded. Complications occurred following 11% of endoscopic retrograde cholangiopancreatographies (ERCPs) (including pancreatitis, 5%) and 12% of percutaneous transhepatic biliary drainages (PTBDs) (including cholangitis, 8%). Complications were associated with: potentially resectable cancer (odds ratio [OR] 1.93, 95% confidence interval [CI] 1.23–3.03); more than one biliary drainage attempt (OR 1.69, 95% CI 1.04–2.74); cholangiocarcinoma (OR 2.20, 95% CI 1.20–4.05), or radiological cancer diagnosis (OR 2.02, 95% CI 1.13–3.60). Thirty-day mortality rates following ERCP and PTBD were 21.4% and 21.4%, respectively, in unresectable cancer and 6% and 6.3%, respectively, in potentially resectable cancer. Increased 30-day mortality in patients with unresectable disease was associated with a performance status of 2 or more (HR 3.14 (1.65–5.97)). Thirty-day mortality was significantly higher in patients with unresectable cancer if a multidisciplinary team meeting had not reviewed and advised drainage prior to the procedure 50% vs 20.4% (P = 0.028).

Conclusions

Careful multidisciplinary consideration of risks and potential benefits should be undertaken prior to attempting malignant biliary drainage due to the high risk of complications and early mortality.

 

MissingLinkText

Keywords

Cholangiopancreatography Endoscopic Retrograde - Percutaneous Transhepatic Biliary Drainage - Cholangiopancreatography - Endoscopic Retrograde/adverse effects - Percutaneous Transhepatic Biliary Drainage/complications - Mortality

Introduction

Malignant biliary obstruction is a common complication of periampullary cancer. Biliary drainage is typically achieved either by endoscopic retrograde cholangiopancreatography (ERCP) or percutaneous transhepatic biliary drainage (PTBD). Endoscopic ultrasound-guided biliary drainage (EUS-BD) is presently limited largely to a few expert centers.

Both ERCP and PTBD are associated with significant risks [1] [2] [3]. Reported outcome data for ERCP commonly make no distinction between malignant or benign indications, but two retrospective analyses of national data demonstrate high 30-day mortality in patients undergoing ERCP for malignant biliary obstruction [4] [5].

Poor outcomes after biliary drainage can impact subsequent cancer pathways and chances of receiving optimal treatment. In potentially resectable disease, the focus should be on expediting safe surgery because evidence favors early surgery without attempting biliary drainage and increasing time to surgery is associated with lower resection rates [6] [7]. In unresectable disease, the focus should be on enabling chemotherapy or assisting best supportive care through relieving symptoms of biliary obstruction [8]. Among patients close to the end of life, the decision to undertake biliary drainage for symptomatic benefit needs to be carefully balanced against the discomfort and potential complications. This observational study aimed to describe the current care of patients undergoing biliary drainage for malignant biliary obstruction. Key outcomes included rates of, and factors associated with, failure to achieve biliary drainage on the first attempt, complications, and mortality within a national prospective cohort study of patients.

 

MissingLinkText

Patients and methods

The “Receipt of Curative resection Or palliative Care for Hepatopancreatic biliary Tumours” (RICOCHET) study was a United Kingdom (UK) prospective, observational audit of all patients presenting with pancreatic cancer or malignant biliary obstruction [9]. The RICOCHET Study Group is a UK trainee-led research collaboration formed in association with the West Midlands Research Collaborative (WMRC). For this analysis, only those patients undergoing attempted biliary drainage were included. Data collection took place over 16 weeks, starting in April 2018. Patients were followed up for 90 days from initial hospital contact. The Research Electronic Data Capture (REDCap) platform [10] [11] was used for data collection. Study data were collected and managed using REDCap tools hosted at the University of Birmingham. All UK healthcare providers involved in management of pancreatic cancer patients were invited to participate.

Inclusion criteria

Eligibility criteria for inclusion in the RICOCHET study were age over 16, suspected malignant pancreatic or peri-ampullary lesion, or suspected malignant biliary obstruction caused by a primary malignancy of the liver hilum or extrahepatic biliary tree (e.g. extrahepatic cholangiocarcinoma). Exclusion criteria were gallbladder or intrahepatic malignancy (e.g. hepatocellular carcinoma) or subsequent diagnosis of benign disease.

Patients were included in this analysis if they had biliary obstruction and underwent an attempt at drainage by ERCP, PTBD, or EUS-BD.

 

MissingLinkText

Definitions

Diagnosis of cancer was based on the decision of the local multidisciplinary team (MDT). Where a pathologic diagnosis was available, that was used. However, because many patients, particularly those with unresectable cancer, do not have a pathologic diagnosis, a diagnosis based upon clinical and radiologic features of suspected cancer was used in the absence of histology. Patients were included in the potentially resectable cancer cohort if they had at least one MDT outcome that indicated a potentially resectable disease. Providers were divided into specialist or non-specialist centers. Specialist providers performed pancreatic cancer resection (following UK national centralization of cancer services in 2001) [12]. Non-specialist providers were associated with a specialist provider and may have had range of multidisciplinary services to enable diagnosis, staging, interventions such as biliary drainage and treatments such as chemotherapy, but did not provide surgery. Biliary drainage was considered successful based on operator opinion.

 

MissingLinkText

Complications

Complications within 7 days were prospectively defined and included in a “data dictionary”. Post-procedure bleeding required either transfusion or intervention. Perforation and cholangitis were clinical diagnoses supported by imaging where appropriate. Cholangitis was assessed as a post-procedure complication. Patients who had cholangitis prior to biliary decompression were not included in the post-procedure cholangitis group because their cholangitis preceded the procedure. Pancreatitis required post-procedure abdominal pain along with an amylase three times above the local normal level [13], and acute kidney injury was a > 1.5-fold increase in baseline serum creatinine i.e. Kidney Disease Improving Global Outcomes stage 1 AKI or above [14].

 

MissingLinkText

Ethics approval

As determined by the national UK decision-making tool of the NHS Health Research Authority and the Medical Research Council [15], RICOCHET was a service evaluation and audit, and thus, did not require research ethics committee approval. Each participating hospital gained local approval prior to data collection, having registered RICOCHET as a service evaluation or audit.

 

MissingLinkText

Outcome measures

The primary outcome measure of this analysis was 30-day mortality following attempted biliary drainage by either PTBD or ERCP.

Secondary outcome measures were success of biliary drainage and incidence of recognized complications including bleeding, pancreatitis, perforation, cholangitis, acute kidney injury, and unplanned readmission.

 

MissingLinkText

Statistical analysis

For assessment of statistically significant differences in proportions of ERCP vs. PTBD, two-sample tests of proportions was used. Multivariable logistic regression models were constructed for success at first drainage attempt and complications post-drainage. Both success and presence of complications were binary outcomes, and therefore, a logistic model to identify associations was appropriate. Goodness of fit was confirmed using a Hosmer-Lemeshow test. Cox regression models were constructed for mortality within the study period and receipt of palliative treatment to patients on a non-curative pathway. Both outcomes were considered suitable for a time-to-event analysis, and therefore, a Cox model was employed. Included variables were identified by considering both results of univariable analyses and those with a compelling clinical rationale for all four models.

None of the regression models include an individual patient more than once. In the multivariable logistic regression model, post-procedure complications from any procedure that a single patient might undergo are included.

Due to data on EUS-BD only being available from a single provider, these data are not included in any of the models. Statistical analysis was undertaken in Stata-15. P < 0.05 was considered to be statistically significant. Missing data are reported in the demographics [Table 1].

 

MissingLinkText

 

MissingLinkText

Results

Study subjects

The RICOCHET audit collected data from over 2500 patients across 93 NHS hospitals (26 specialist and 67 non-specialist). A total of 773 patients underwent biliary drainage at a provider included in the RICOCHET audit and were included in this analysis. The consort diagram of case selection is shown in [Fig. 1]. Of the patients, 80.2% had a single drainage procedure, 16.7% had two procedures, and 3.1% had three or more procedures. Mean pre-drainage bilirubin was 248 µmol/L (SD 143). Of the patients, 28.8% had a final radiological cancer diagnosis only, 35.1% had a histological diagnosis of pancreatic ductal adenocarcinoma (PDAC), 14.7% had a diagnosis of cholangiocarcinoma, and 9.2% had a diagnosis of ampullary/duodenal cancer. Of the patients, 667 (86.3%) had biliary obstruction below the liver hilum, whereas only 97 (12.5%) had hilar obstruction. Of the patients, 42.2% were considered potentially resectable. Full demographic details of the cohort undergoing biliary drainage are shown in [Table 1].

Fig. 1 CONSORT diagram for the study.

Table 1 Demographic details of patients undergoing drainage of malignant biliary obstruction.

		Number	%
*Includes pancreatic neuroendocrine tumors or premalignant intraductal papillary mucinous neoplasms. ERCP, endoscopic retrograde cholangiopancreatography; EUS-BD, endoscopic ultrasound-guided biliary drainage; PDAC, pancreatic ductal adenocarcinoma; PTBD, percutaneous transhepatic biliary drainage.
Patients		773
Number of drainage procedures	1	620	80.2
	2	129	16.7
	3	21	2.7
	4	1	0.1
	5	2	0.3
Age (tertile)	< 69	260	33.6
	69–77	261	33.8
	> 77	252	32.6
Sex	Male	413	53.4
	Female	360	46.6
Charlson comorbidity score	< 5	196	25.4
	5–7	453	58.6
	> 7	124	16.0
Performance status	0	343	44.4
	1	217	28.1
	2	141	18.2
	> 2	61	7.9
	Not available	11	1.4
Potentially resectable	Yes	327	42.2
	No	446	57.8
Provider type	Non-specialist	396	51.2
	Specialist	377	48.7
Cancer diagnosis (by histology unless otherwise stated)	Radiological diagnosis only	223	28.8
	PDAC	272	35.1
	Cholangiocarcinoma	114	14.7
	Ampullary/duodenal carcinoma	71	9.2
	Other*	93	12.0
Initial biliary drainage procedure and location obstruction	ERCP for hilar obstruction	48	6.2
	PTBD for hilar obstruction	49	6.3
	ERCP for below hilum obstruction	588	76.1
	PTBD for below hilum obstruction	79	10.2
	EUS-BD	9	1.2
Multidisciplinary team meeting prior first drainage attempt	Yes	359	46.4
	No	414	53.6

 

MissingLinkText

Biliary drainage procedural details including factors associated with successful drainage

A total of 955 biliary drainage procedures were attempted in 773 patients: 719 ERCPs, 225 PTBDs, and 11 EUS-BDs. Outcomes of biliary drainage procedures are shown in [Table 2]. ERCP was usually the first drainage procedure (82.3%) rather than PTBD (16.6%); however, the proportion of first successful procedures for ERCP was lower compared with PTBD (76.6% vs. 87.5%; (-10.9%, 95% CI (-17.5 to -4.3)%, p=0.006). The rate of successful drainage on any subsequent attempt was only 37.4% per procedure. Drainage was eventually achieved in 46.3% of patients who had an unsuccessful first attempt. There was no statistical significance in the difference in proportion of first-time successful ERCP procedures between lesions below the hilum vs. hilar (77% vs.70.8%; 6.2%, 95% CI –7.1% to 19.5%). However, there was a statistically significant difference in proportion of first-time successful PTBD procedures for below the hilum vs. hilar lesions (94.9% vs 75.5%; 19.4%, 95% CI 6.4% to 32.4%; P = 0.0013). At ERCP the type of stent was recorded in 626 of 719 procedures (87.1%). A metal stent was placed in 351 (56.1%), a plastic stent in 141 (22.5%), and no stent in 134 (21.4%).

Table 2 Outcomes of biliary drainage procedures.

	ERCP		PTBD		Total	P value*	Difference with 95% CI*
	Number	%	Number	%
CI, confidence interval; ERCP, endoscopic retrograde cholangiopancreatography; PTBD, percutaneous transhepatic biliary drainage. *Two-sample test of proportions used. †Percentages based on responses for 626 ERCPs. ‡Frequency in either procedure is less than 5, and therefore, too low to allow for a meaningful comparison.
Drainage procedures	719	–	225	–	944	–	–
Outpatient procedure	146	20.3	7	3.1	153	< 0.001	17.2 (13.5 to 20.9)
Procedures in potentially resectable patients	318	44.2	80	35.6	398	0.023	8.6 (1.4 to 1.6)
First procedure	636	82.3	128	16.6	764	< 0.001	65.7 (61.9 to 69.5)
Success at first procedure	487	76.6	112	87.5	599	0.006	–10.9 (–17.5 to –4.3)
Success at first procedure (hilar)	34	70.8	37	75.5	71	0.601	–4.7 (–22.3 to 12.9)
Success at first procedure (below hilum)	453	77.0	75	94.9	528	< 0.001	–17.9 (–23.8 to –12.0)
Metal stent	351	56.1†	–	–	–	–	–
Plastic stent	141	22.5†	–	–	–	–	–
No stent inserted	134	21.4†	–	–	–	–	–
Any complication within 7 days	79	11.0	28	12.4	107	0.563	–1.4 (–6.3 to 3.5)
Any complication potentially resectable cancer	49	15.4	10	12.5	59	0.514	2.9 (–5.4 to 11.2)
Any complication unresectable cancer	30	7.5	18	12.4	48	0.074	–4.9 (–10.9 to 1.0)
Bleeding	7	1.0	3	1.3	10	0.703	‡
Perforation	1	0.1	1	0.4	2	0.343	‡
Cholangitis	21	2.9	18	8.0	39	< 0.001	–5.1 ( –8.9 to –1.3)
Pancreatitis	36	5.0	2	0.9	38	0.006	‡
Acute kidney injury	14	1.9	4	1.8	18	0.923	‡
7-day readmission (for outpatients)	24	16.4	1	14.3	25	0.883	‡
Death within 30 days (all)	107	14.9	36	16.0	143	0.688	–1.1 (–6.6 to 4.4)
Death within 30 days (unresectable)	86	21.4	31	21.4	117	1.0	0 (–7.8 to 7.8)
Death within 30 days (potentially resectable)	19	6.0	5	6.3	24	0.920	–0.3 (–6.2% to 5.6)

Results of logistic regression analysis of factors associated with success at first drainage attempt are shown in Supplementary Table 1. Patients with obstruction below the hilum (OR 1.90, 95% CI 1.09–3.30) were more likely to have a successful first drainage attempt, as were those undergoing PTBD as their first procedure (OR 3.17, 95% CI 1.70–5.96). Procedures attempted as an inpatient were less likely to be successful (OR 0.44, 95% CI 0.27–0.71).

 

MissingLinkText

Complications following attempted biliary drainage

Overall complications following ERCP and PTBD were 11.0% and 12.4%, respectively ([Table 2]). The most common complications were pancreatitis after ERCP (5%) and cholangitis after PTBD (8%).

Results of logistic regression analysis of factors associated with any complication following ERCP or PTBD are shown in [Table 3]. Cholangiocarcinoma (OR 2.20, 95% CI 1.20–4.05) and a radiological cancer diagnosis only (OR 2.02, 95% CI 1.13–3.60) both were associated with more complications compared with PDAC. More than one attempt at biliary drainage was associated with more complications (OR 1.69, 95% CI 1.04–2.74). Patients with potentially resectable cancer were associated with more complications (OR 1.93, 95% CI 1.23–3.03).

Table 3 Logistic regression analysis of patient level factors associated with complications following any attempted biliary drainage.

		OR	95% CI	P value
CI, confidence interval; OR, odds ratio; PDAC, pancreatic ductal adenocarcinoma.
Age (tertile)	< 69	Ref cat
	69–77	1.53	0.91–2.59	0.106
	> 77	1.31	0.72–2.37	0.280
Sex	Female	0.88	0.57–1.35	0.559
Performance status	0	Ref cat
	1	0.52	0.30–0.89	0.018
	2+	0.54	0.30–0.98	0.043
Cancer diagnosis	PDAC	Ref cat
	Radiological only	2.02	1.13–3.60	0.017
	Cholangiocarcinoma	2.20	1.20–4.05	0.011
	Duodenal/ampullary	0.80	0.33–1.95	0.629
	Other	1.50	0.72–3.15	0.278
Other factors	> 1 drainage procedure	1.69	1.04–2.74	0.033
	Potentially resectable	1.93	1.23–3.03	0.004

 

MissingLinkText

Thirty-day mortality following attempted biliary drainage

Thirty-day mortality after ERCP was 14.9%: 21.4% for unresectable cancer and 6.0% for potentially resectable cancer. Similarly, 30-day mortality after PTBD was 16.0%: 21.4% for unresectable cancer and 6.3% for potentially resectable cancer.

Results of Cox regression analyses of factors associated with mortality following biliary drainage in unresectable cancer are shown in [Table 4]. Female sex (HR 0.40 (95% CI 0.25–0.64)), success at first drainage attempt (HR 0.57, 95% CI 0.33–0.99) and a specialist provider (HR 0.52, 95% CI 0.32–0.85) were associated with lower mortality. A Performance Status of one (HR 2.17, 95% CI 1.09–4.32) and two or more (HR 3.14, 95% CI 1.65–5.97) were associated with increased mortality.

Table 4 Cox regression analysis of factors associated with 30-day mortality in patients with unresectable cancer.

		Hazard ratio	95% CI	P value
CI, confidence interval; PTBD, percutaneous transhepatic biliary drainage.
Age (quintile)	< 69	Ref cat
	69–77	0.90	0.47–1.71	0.747
	> 77	1.04	0.55–1.95	0.911
Sex	Female	0.40	0.25–0.64	< 0.001
Performance status	0	Ref cat
	1	2.17	1.09–4.32	0.027
	2+	3.14	1.65–5.97	< 0.001
Provider	Non-specialist	Ref cat
	Specialist unit	0.52	0.32–0.85	0.009
Other factors	Cancer below hilum	2.06	0.92–4.61	0.080
	Successful first drainage attempt	0.57	0.33–0.99	0.049
	PTBD as first drainage procedure	1.73	0.96–3.13	0.069

Among patients undergoing biliary drainage for unresectable disease, 30-day mortality following any drainage procedure was significantly higher if a MDT meeting had not resulted in a recommendation for drainage prior to the first attempt (50.0% [5/10] vs 20.4% [38/186]; P = 0.028).

 

MissingLinkText

Palliative chemotherapy following biliary drainage

Among patients undergoing biliary drainage with an unresectable cancer, 20.1% received palliative chemotherapy. Of those patients considered to be potentially resectable, 8.6% received palliative chemotherapy.

Cox regression analysis of associations with receiving chemotherapy within 3 months of first presentation are shown in [Table 5]. Chemotherapy was less common if performance status was > 1 (HR 0.24, 95% CI 0.11–0.53) or the cancer diagnosis was based on imaging only (HR 0.28, 95% CI 0.14–0.54). Patients at specialist providers were associated with an increased chance of receiving chemotherapy (HR 2.38, 95% CI 1.38–4.09).

Table 5 Cox regression analysis of factors associated with receipt of palliative chemotherapy within 90 days in patients with unresectable cancer.

		Hazard ratio	95% CI	P value
CI, confidence interval; PDAC, pancreatic ductal adenocarcinoma; PTBD, percutaneous transhepatic biliary drainage. *Other includes pancreatic neuroendocrine tumors or premalignant intraductal papillary mucinous neoplasms.
Age tertile	< 69	Ref cat
	69–77	0.73	0.38–1.40	0.342
	> 77	0.60	0.29–1.25	0.174
Sex	Female	1.67	0.96–2.89	0.067
Performance status	0	Ref cat
	1	0.73	0.38–1.38	0.328
	2+	0.24	0.11–0.53	< 0.001
Provider	Non-specialist	Ref cat
	Specialist	2.38	1.38–4.09	0.002
Cancer diagnosis	PDAC	Ref cat
	Radiological only	0.28	0.14–0.54	< 0.001
	Cholangiocarcinoma	0.63	0.27–1.49	0.293
	Duodenal/ampullary	0.39	0.11–1.31	0.128
	*Other	0.13	0.04–0.41	< 0.001
Success at first drainage procedure		1.50	0.77–2.92	0.238
Cancer below hilum		1.70	0.62–4.65	0.298
Complication of any procedure		0.54	0.19–1.52	0.242
PTBD as first drainage procedure		0.48	0.20–1.16	0.104

 

MissingLinkText

 

MissingLinkText

Discussion

This is the largest prospective national study of biliary drainage for malignant biliary obstruction. The key findings are as follows. There is a high incidence of complications, readmission, and 30-day mortality compared with previously reported ERCP outcome data [1] [16]. Overall success rates for biliary drainage were high. If the first drainage attempt was unsuccessful, the success rate for subsequent procedures dropped markedly. Post-procedure 30-day mortality and complication rates were notable for those with potentially curable cancer.

Several large-scale studies reported rates of complications following ERCP for all indications including both malignant and benign disease. A systematic review of prospective studies included 16,855 patients and reported a 6.85% complication rate, with pancreatitis in 3.47%, bleeding in 1.34%, and sepsis in 1.44% [1]. A large, single-institution review of 11,497 ERCPs reported complications in 4% of patients, including pancreatitis in 2.6% and sepsis in 0.3% [16]. In the current patient cohort with malignant disease, 11.0% had complications following ERCP, with 5.0% developing pancreatitis and 2.9% developing sepsis (cholangitis). Our prospective predefined data collection likely provides an accurate accrual of complications in a real-world clinical situation across a variety of providers rather than the data from expert centers included in previous publications.

The most frequently reported complication of ERCP is pancreatitis, with an incidence of 3% to 5% [17]. Varying definitions of pancreatitis make it difficult to compare studies, with some using the original consensus definition by Cotton et al and the revised Atlanta international consensus (RAC) [13] [18]. This study used an amylase level three times greater than the upper level of normal and abdominal pain as criteria for post-ERCP pancreatitis, as per Atlanta criteria [19], and is reflective of clinical practice. The Incidence of pancreatitis was higher than previously reported and is likely to have a significant clinical impact, particularly for those on a potentially curative pathway because it can delay surgery. It will also affect patient baseline physical status, potentially delaying postoperative recovery and adjuvant therapy. Pancreatitis may also delay or prevent patients receiving palliative chemotherapy.

Complications of PTBD are recognized, and unlike for ERCP, there are more data on complications specific to management of malignant biliary obstruction. Advances in endoscopic stenting techniques have reduced need for percutaneous intervention. However, there is still a significant role for PTBD in palliative management of malignant biliary obstruction following failed ERCP drainage and for endoscopically inaccessible strictures. A retrospective review of 16,822 patients undergoing PTBD for unresectable malignant biliary obstruction in the UK between 2001 and 2014 reported cholangitis in 3.9%, bleeding in 2.1%, and 30-day mortality of 23.1% [3]. Mortality was similar in this study with 21.4% 30-day mortality in unresectable disease (6.3% in resectable disease), but the rate of cholangitis was substantially higher at 8.0%. This likely represents under-reporting of cholangitis in the retrospective analysis. In the present study, success at first PTBD attempt was found to be associated with fewer complications. This is biologically plausible, because the need for repeated attempts to achieve drainage will likely relate to the complexity of the obstruction. Of the patients, 12.3% had hilar obstruction and successful decompression in the first procedure was achieved in 77% of patients. Success rates for ERCP and PTBD were 70.8% and 75.5%, respectively. This reflects heterogeneity in practice in the UK for hilar biliary compression. Lesions below the liver hilum had higher first-time procedure success rates. The number of attempts at biliary decompression were recorded, but the number of stents required to achieve successful biliary decompression in complex hilar strictures or internalization of stents following PTBD is beyond the scope of this study. For hilar decompression attempts, nearly an equal split of patients had ERCP ERCP (48.5%) or PTBD (49.5%) as their first procedure. For obstruction below the hilum, ERCP was first attempted in the majority of cases (87.2%), with a success rate of 77%. However, regression analysis underscored that lesions below the liver hilum had a higher first-time procedure success rate, reflecting the potential complexity of drainage of hilar strictures.

Aims of treatment are important to bear in mind when selecting patients. Post-ERCP complications were more common in those with potentially resectable cancer. Potential explanations for this finding include endoscopists accepting additional risks or persisting for longer to achieve drainage in patients with potentially curable cancer, with the consequence of an increased complication rate. Furthermore, patients with potentially curable cancer may have closer follow-up, and therefore, complications are more easily identified by the data collection team. In patients with unresectable disease, age was not significantly associated with mortality, however, performance status was, implying that the functional state of the patient rather than their age is the most important factor for procedural patient selection.

Significantly higher mortality was seen when a drainage procedure was undertaken despite an MDT meeting advising against this. Furthermore, success at first drainage attempt was associated with only half the 30-day mortality rate of subsequent attempts. This suggests that MDT discussion is critical in management of patients and should carefully consider the risks of biliary drainage procedures in the context of expected disease outcomes, previous attempts at drainage, and patient functional status. In the UK, all patients with suspected hepatobiliary cancer are discussed at MDT meetings, which tend to occur once a week and may not be in the same trust as that which performed the imaging. Therefore, some patients may undergo biliary decompression during an emergency episode prior to discussion at an MDT for logistical reasons.

The end goal of biliary drainage is not just technical success in relieving obstruction. In unresectable disease, it is to enable better quality of life and access to palliative chemotherapy, if appropriate. For resectable disease, it is to enable swift, definitive resection and adjuvant therapy. For many patients with unresectable cancer, a decision to avoid biliary drainage may be appropriate if they are not significantly symptomatic and will not be suitable for palliative chemotherapy. This requires a sensitive approach with frank discussion of risks versus benefits. A performance status of 2 or more was associated with a lower likelihood of palliative chemotherapy and increased 30-day mortality, highlighting the importance of careful assessment and decision making. Furthermore, a 30-day mortality risk prediction score at presentation can support a discussion between clinicians and patients that results in an informed decision about further management. Such a score has been developed in esophageal cancer [20]. For some patients with potentially resectable cancer, avoiding preoperative biliary drainage is desirable because level 1 evidence demonstrates reduced complications and a shorter time to treatment [21]. A 2017 systematic review and meta-analysis of preoperative biliary drainage favored no drainage where possible and the DROP trial randomizing to early surgery or drainage and delayed surgery between 4 to 6 weeks found a significantly higher 120-day rate of serious complications, which was attributed to a very high rate of complications associated with biliary drainage, as was found in the current study [12] [13].

It is unclear why patients managed by a specialist provider were more likely to receive palliative chemotherapy. It is plausible that fewer barriers to oncology services leads to improved access for this patient group. It is also possible that some patients considered unresectable by a specialist provider would have been considered potentially resectable by a non-specialist provider. Consequently, unresectable patients at a specialist provider may be better candidates for palliative chemotherapy, compared with unresectable patients at non-specialist providers. 42.2% were potentially resectable while 57.8% were unresectable. A histological diagnosis was not available in 28.8% of patients undergoing biliary decompression. Of this cohort, 32.6% were older than age 77 years, 58.6% had a Charlson comorbidity score of 5 to 7, and 16.0% had a score > 7. Therefore, a significant number of patients with unresectable disease would likely not be fit for chemotherapy, hence a histological diagnosis would be unlikely to change management. Of unresectable patients, only 20.1% received chemotherapy. In some patients, sampling may not be histologically diagnostic, with patients declining or being otherwise unfit for further invasive tests.

There are several limitations to the present study. Despite all UK providers being eligible to take part, the study only included data for patients newly diagnosed with malignant biliary obstruction from 93 providers in total over the course of just 16 weeks. Each year, approximately 3700 ERCPs are performed in England for cancer [5]. This excludes patients who subsequently undergo surgery. Using these data to validate RICOCHET, we would expect approximately 300 patients based on 12 weeks of recruitment and one-third of providers being involved. Although this audit collected data from over 2500 cases of malignant biliary compression over 16 weeks, only 773 cases had procedures for biliary decompression. The actual number of ERCPs with a histological diagnosis of PDAC, ampullary carcinoma, and cholangiocarcinoma that did not subsequently undergo curative surgery was 289, which suggests that the audit did not miss a significant number of cases. The higher rates of complications and lower success rates of biliary drainage than those previously reported are likely to reflect the unselected, multicenter, “real-world” setting of this study with heterogenous approaches to biliary drainage. Follow-up was 90 days from first hospital contact and not biliary decompression; therefore, stent changes are likely to have occurred outside the follow-up window. Further management after unsuccessful drainage or internalization of the drainage following PTBD was beyond the scope of this project. The two main biliary drainage procedures performed in the UK at the time of the study were ERCP and PTBD. EUS-BD is another option, but our study included only a small number of these procedures and all were performed by a single provider (Supplementary Table 2). In 2024, of 15 acute providers in the West Midlands region of the UK, only four providers offered EUS-BD. This limits the generalizability of this part of the data to the UK and was the reason for excluding it from the models constructed.

Furthermore, the audit found 30-day mortality for resectable and unresectable disease to be 19 (6%) and 86 (21.4%), respectively, but the cause of death went beyond the scope of the study because most patients will have died in the community rather than the hospital, and therefore, it was not possible to collect data on the precise cause of death.

 

MissingLinkText

Conclusions

In summary, the data presented in this paper provide a prospective multicenter assessment of the risk profile of biliary drainage for malignant biliary obstruction. They will enable clinicians to have more informed discussions with their patients around the most appropriate management of malignant biliary obstruction in both resectable and unresectable cancer settings.

 

MissingLinkText

 

MissingLinkText

Conflict of Interest

Keith Roberts has received honoraria and research funding from Mylan Pharmaceuticals. There are no other competing interests to declare.

Acknowledgement

We are grateful to the Birmingham Centre for Observational Non-Randomised studies at the University of Birmingham for the use of their servers for secure online data collection. The article is published under corporate authorship with all collaborators contributing to the project. We would like to thank the RICOCHET Study Group on behalf of the West Midlands Research Collaborative. Due to space limitations, the full list of collaborators is provided in the supplementary material of a related publication: Pancreatic enzyme replacement therapy in patients with pancreatic cancer: A national prospective study. Pancreatology. 2021 May 25;S1424-3903(21)00469-5. doi: 10.1016/j.pan.2021.05.299.

• References

• 1 Andriulli A, Loperfido S, Napolitano G. et al. Incidence rates of post-ERCP complications: a systematic survey of prospective studies. Am J Gastroenterol 2007; 102: 1781-8
• 2 Kochar B, Akshintala VS, Afghani E. et al. Incidence, severity, and mortality of post-ERCP pancreatitis: a systematic review by using randomized, controlled trials. Gastrointest Endosc 2015; 81: 143-149.e9
• 3 Rees J, Mytton J, Evison F. et al. The outcomes of biliary drainage by percutaneous transhepatic cholangiography for the palliation of malignant biliary obstruction in England between 2001 and 2014: a retrospective cohort study. BMJ Open 2020; 10: e033576
• 4 Bodger K, Bowering K, Sarkar S. et al. All-cause mortality after first ERCP in England: clinically guided analysis of hospital episode statistics with linkage to registry of death. Gastrointest Endosc 2011; 74: 825-833
• 5 Harvey PR, Baldwin S, Mytton J. et al. Higher volume providers are associated with improved outcomes following ERCP for the palliation of malignant biliary obstruction. EClinicalMedicine 2020; 18: 100212
• 6 Müller PC, Hodson J, Kuemmerli C. et al. Effect of time to surgery in resectable pancreatic cancer: a systematic review and meta-analysis. Langenbecks Arch Surg 2020; 405: 293-302
• 7 Pande R, Hodson J, Marudanayagam R. et al. Survival advantage of upfront surgery for pancreatic head cancer without preoperative biliary drainage. Front Oncol 2020; 10: 526514
• 8 National Institute for Health and Care Excellence. NICE guideline [NG85] Pancreatic cancer in adults: diagnosis and management. https://www.nice.org.uk/guidance/ng85
• 9 RICHOCET Study group, West Midlands Research Collaborative. Receipt of Curative Resection or Palliative Care for Hepatopancreaticobiliary Tumours (RICOCHET): Protocol for a nationwide collaborative observational study. JMIR Res Protoc 2019; 8
• 10 Harris PA, Taylor R, Thielke R. et al. Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform 2009; 42: 377-381
• 11 Harris PA, Taylor R, Minor BL. et al. The REDCap consortium: Building an international community of software platform partners. J Biomed Inform 2019; 95
• 12 Department of Health. The NHS Cancer plan; a plan for investment, a plan for reform. London, UK: Department of health; 2000.
• 13 Banks PA, Bollen TL, Dervenis C. et al. Classification of acute pancreatitis--2012: revision of the Atlanta classification and definitions by international consensus. Gut 2013; 62: 102-111
• 14 Khwaja A. KDIGO clinical practice guidelines for acute kidney injury. Nephron Clin Pract 2012; 120: c179-c184
• 15 NHS Health Research Authority. Is my study research?. https://www.hra-decisiontools.org.uk/ethics/
• 16 Cotton PB, Garrow DA, Gallagher J. et al. Risk factors for complications after ERCP: a multivariate analysis of 11,497 procedures over 12 years. Gastrointest Endosc 2009; 70: 80-88
• 17 Johnson KD, Perisetti A, Tharian B. et al. Endoscopic retrograde cholangiopancreatography-related complications and their management strategies: A "scoping" literature review. Dig Dis Sci 2020; 65: 361-375
• 18 Cotton PB, Lehman G, Vennes J. et al. Endoscopic sphincterotomy complications and their management: an attempt at consensus. Gastrointest Endosc 1991; 37: 383-393
• 19 Banks PA, Bollen TL, Dervenis C. et al. Classification of acute pancreatitis—2012: revision of the Atlanta classification and definitions by international consensus. Gut 2013; 62: 102-111
• 20 Driver RJ, Handforth C, Radhakrishna G. et al. The Glasgow Prognostic Score at the time of palliative esophageal stent insertion is a predictive factor of 30-day mortality and overall survival. J Clin Gastroenterol 2018; 52: 223-228
• 21 Van der Gaag NA, Rauws EAJ, van Eijck CHJ. et al. Preoperative biliary drainage for cancer of the head of the pancreas. N Engl J Med 2010; 362: 129-137

Publication History

Received: 02 January 2025

Accepted: 05 March 2025

Article published online:
23 July 2025