Interventions to relieve biliary and duodenal obstruction

National Guideline Alliance (UK)

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Guideline Alliance (UK). Pancreatic cancer in adults: diagnosis and management. London: National Institute for Health and Care Excellence (NICE); 2018 Feb. (NICE Guideline, No. 85.)

Pancreatic cancer in adults: diagnosis and management.

Show details

Contents

< Prev Next >

9Interventions to relieve biliary and duodenal obstruction

9.1. Biliary obstruction

Review question: What is the optimal treatment of biliary obstruction in adults with newly diagnosed or recurrent pancreatic cancer?

9.1.1. Introduction

Biliary obstruction causing obstructive jaundice is the most visible manifestation of pancreatic cancer in the head of pancreas. Although it is not present in all patients, the main symptom associated with obstructive jaundice is itching, which can be severe and debilitating. Other symptoms that may be caused or exacerbated by biliary obstruction include early satiety and nausea. The visible signs of biliary obstruction, which may most concern the individual, include yellow sclera and skin. Biliary obstruction leads to malabsorption of the fat soluble vitamins, resulting in a vitamin K deficiency if obstruction is prolonged, and consequent derangement of blood clotting.

In patients with resectable tumours, standard practice has been to relieve the obstruction via insertion of a stent, and normalise blood tests as far as possible prior to surgery; due to concern that operating on patients with significant biliary obstruction would increase operative morbidity and possibly mortality. As the jaundice worsens quickly, the delay between presentation and the date for surgery (which at best is only a few weeks but usually longer), can be associated with a significant worsening of jaundice.

In addition to whether or not jaundice needs to be relieved prior to surgery, another important issue is the timing of any drainage, relative to imaging for staging. This is because the process of placing a biliary stent (usually when endoscopic retrograde cholangiopancreatography [ERCP] is performed) has been associated with pancreatitis, which may make staging of the tumour more difficult. In addition, whilst plastic stents (which have a small diameter lumen) are cheap and have been used for drainage in the last few years, considerably more expensive self-expanding mesh metal stents (SEMS) (which have a larger diameter and therefore considerably better flow and longevity) have become widely available. Moreover, it is thought that SEMS cause less morbidity than plastic stents. Thus, in individuals with resectable tumours, it remains to be established whether or not drainage is required before surgery, whether SEMS are better than plastic stents, and - if it is indicated – when is the optimal time for drainage.

With regards to treatment of biliary obstruction in individuals with borderline resectable tumours, the issues are similar to those for individuals with resectable tumours (although they are perhaps clearer because the patient will not be considered for immediate surgery). The case for pre-operative drainage is stronger based on a patient’s symptoms and any jaundice will need to be relieved prior to neoadjuvant chemotherapy. However, which stent should be used for drainage and when drainage should occur are still open questions.

With regards to biliary obstruction in individuals with unresectable tumours, it is still unclear whether a plastic or metal stent should be used. One important issue is endoscopic management (ERCP and stenting), which is the most commonly-performed intervention, as it is perceived to be less invasive than alternative methods.

Guidance is needed on the optimal treatment of biliary obstruction in people with pancreatic cancer.

9.1.1.1. Review protocol summary

The review protocol summary used for this question can be found in Table 99. Full details of the review protocol can be found in Appendix C.

Table 99

Clinical review protocol summary for the review of optimal treatment of biliary obstruction.

9.1.2. Description of clinical evidence

Twenty-two RCTs were included in the review. Several of the studies included individuals that did not have pancreatic cancer. Generally, the Committee decided to only include studies that had at least 66% pancreatic cancer patients, though the quality of evidence for relevant outcomes was downgraded one level for indirectness.

Further information about the search strategy can be found in Appendix D. See study selection flow chart in Appendix E, forest plots in Appendix H, GRADE tables in Appendix I, study evidence tables in Appendix F and list of excluded studies in Appendix G.

9.1.2.1. Plastic stent versus self-expanding metal stents (SEMS) in adults with pancreatic cancer and biliary obstruction

Eight studies (n=815) compared the use of plastic stents with SEMS (Gardner et al., 2016; Isayama et al., 2011; Kaassis et al., 2003; Moses et al., 2013; Schmidt et al., 2015; Söderlund & Linder, 2006; Travis & Nicholson, 1997; Walter et al., 2015). Seven of these studies used ERCP to aid insertion of a stent, whilst only 1 used percutaneous transhepatic cholangiography (PTC) (Travis & Nicholson, 1997). Seven of the studies were in adults with either unresectable pancreatic cancer or unresectable malignant biliary obstruction (Isayama et al., 2011; Kaassis et al., 2003; Moses et al., 2013; Schmidt et al., 2015; Söderlund & Linder, 2006; Travis & Nicholson, 1997; Walter et al., 2015). One study included resectable and borderline resectable adult pancreatic cancer patients in addition to those whose tumours were unresectable (Gardner et al., 2016). A variety of plastic stents (e.g. polyethylene or polyurethane) and SEMS (e.g. covered, partially covered, or uncovered) were used.

9.1.2.2. Covered self-expanding metal stent versus uncovered self-expanding metal stent in adults with pancreatic cancer and biliary obstruction

Five studies (n=708) compared a covered SEMS with an uncovered SEMS (Gardner et al., 2016; Kitano et al., 2013; Krokidis et al., 2011; Kullman et al., 2010; Ung et al., 2013). The majority of the studies were in adults with unresectable pancreatic cancer.

9.1.2.3. Partially-covered self-expanding metal stent versus uncovered self-expanding metal stent in adults with pancreatic cancer and biliary obstruction

Two studies (n=243) compared a partially-covered SEMS with an uncovered SEMS (Telford et al., 2010; Walter et al., 2015) in adults with unresectable tumours.

9.1.2.4. Paclitaxel-eluting self-expanding metal stent versus covered self-expanding metal stent in adults with an unresectable distal malignant biliary obstruction

One study (n=52) compared a paclitaxel-eluting SEMS with a covered SEMS in adults with unresectable distal malignant biliary obstruction (Song et al., 2011). Although this study only included 51% pancreatic cancer patients, it was decided to include it and downgrade the quality of evidence two levels for indirectness for the relevant outcomes.

9.1.2.5. Preoperative endoscopic biliary drainage (PEBD) then surgery versus surgery in adults with suspected pancreatic cancer

One study (n=196) compared endoscopic preoperative biliary drainage using a plastic stent followed by surgery with surgery only in adults with obstructive jaundice due to suspected pancreatic head cancer (Eshuis et al., 2010). The study included resectable and unresectable tumour patients.

9.1.2.6. Endoscopic sphincterotomy then stent versus stent in adults with unresectable pancreatic cancer

Three studies (n=446) compared endoscopic sphincterotomy (ES) followed by the insertion of a stent with a stent only (Artifon et al. 2008; Giorgio & Luca, 2004; Hayashi et al., 2015) in adults with unresectable tumours. The majority of these studies used a partially-covered or covered SEMS.

9.1.2.7. Endoscopic sphincterotomy then stent versus surgical bypass in adults with unresectable pancreatic cancer

One study (n=30) compared endoscopic sphincterotomy (ES) followed by the insertion of a covered SEMS with surgical bypass only (Artifon et al., 2006) in adults with unresectable pancreatic cancer.

9.1.2.8. Endoscopic ultrasound-guided choledochoduodenostomy (EUS-CD) and stent versus percutaneous transhepatic biliary drainage (PTBD) in adults with an unresectable malignant biliary obstruction where either ERCP or EUS-guided transpapillary rendezvous has failed

One study (n=25) compared endoscopic ultrasound-guided choledochoduodenostomy (EUS-CD) and insertion of a partially-covered SEMS with percutaneous transhepatic biliary drainage (PTBD) (Artifon et al., 2012) in adults with an unresectable tumour where either ERCP or EUS-guided transpapillary rendezvous has failed. Although data regarding the number of individuals with pancreatic cancer in this study was not available, it was decided to include it but downgrade the relevant outcomes by two levels for indirectness.

9.1.2.9. Endoscopic ultrasound-guided choledochoduodenostomy (EUS-CD) and stent versus surgical bypass in adults with an unresectable malignant biliary obstruction where ERCP has failed

One study (n=32) compared endoscopic ultrasound-guided choledochoduodenostomy (EUS-CD) and insertion of a partially-covered SEMS with surgical bypass/drainage only (Artifon et al., 2015) in adults with unresectable tumour where ERCP has failed. Although data regarding the number of individuals with pancreatic cancer in this study was not available, it was decided to include it but downgrade the relevant outcomes by two levels for indirectness.

9.1.3. Summary of included studies

A summary of the studies that were included in this review are presented in Table 100.

Table 100

Summary of included studies.

9.1.4. Clinical evidence profiles

The clinical evidence profiles for this review question are presented in Table 101 to Table 109.

Table 101

Summary clinical evidence profile for plastic stent versus self-expanding metal stent in adults with pancreatic cancer and biliary obstruction.

Table 102

Summary clinical evidence profile for covered SEMS versus uncovered SEMS in adults with pancreatic cancer and biliary obstruction.

Table 103

Summary clinical evidence profile for partially covered SEMS versus uncovered SEMS in adults with pancreatic cancer and biliary obstruction.

Table 104

Summary clinical evidence profile for paclitaxel-eluting SEMS versus covered SEMS in adults with an unresectable distal malignant biliary obstruction.

Table 105

Summary clinical evidence profile for preoperative endoscopic biliary drainage then surgery versus surgery in adults with suspected pancreatic cancer.

Table 106

Summary clinical evidence profile for endoscopic sphincterotomy then stent versus stent in adults with unresectable pancreatic cancer.

Table 107

Summary clinical evidence profile for endoscopic sphincterotomy then stent versus surgical bypass in adults with unresectable pancreatic cancer.

Table 108

Summary clinical evidence profile for endoscopic ultrasound-guided choledochoduodenostomy and stent versus percutaneous transhepatic biliary drainage in adults with an unresectable malignant biliary obstruction where either ERCP or EUS-guided transpapillary (more...)

Table 109

Summary clinical evidence profile for endoscopic ultrasound-guided choledochoduodenostomy and stent versus surgical bypass in adults with an unresectable malignant biliary obstruction where ERCP has failed.

9.1.5. Economic Evidence

9.1.5.1. Systematic literature review

References to all included studies and evidence tables for all economic evaluations included in the systematic literature review of the economic evidence are presented in Appendix L. Economic evidence profiles of these studies are presented in Appendix K.

Two studies (Arguedas et al. 2002; Morris et al. 2014) were included in the current review of published economic evidence for this topic. Both economic evaluations considered different interventions in different patient groups and therefore meaningful comparisons between the studies could not be drawn. A bespoke economic model was also built to help inform recommendations for part of this topic.

Morris et al. (2014) compared preoperative biliary drainage (PBD) to direct surgery in patients with potentially resectable pancreatic or periampullary cancer and obstructive jaundice from a UK NHS and PSS perspective. The study was deemed to only have minor methodological limitations.

The effectiveness side of the model is nearly entirely based on 1 Cochrane Review of six RCTs comparing PBD to direct surgery. The utility values for the model were taken from patient responses to the EQ-5D questionnaire, scored using the UK population weightings, completed by people with hepatic colorectal metastases. Although this was not the patient group considered by the economic evaluation the study did report that the trends closely matched those reported in disease specific quality of life measures for the relevant patient group. However, the results of the model were not sensitive to this input and it noted that alternative plausible values were unlikely to change the preferred option. Cost inputs for the model were all sourced from NHS reference costs.

The model concluded that sending patients directly to surgery led to a cost saving of £2,552 per patient. It also led to a small increase in health of 0.006 QALYS. This result was robust to all sensitivity analyses performed with probabilistic sensitivity analysis showing a strategy of PBD prior to surgery being the preferred option in less than 10% of iterations when a £20,000 willingness to pay per QALY is assumed.

The economic evaluation did not explicitly consider the issues of capacity (i.e. operating theatres and surgeons being available when needed) although it was unclear if there would be additional costs to having to reorganise services or not. However, unless the increases in cost per patient were significant it would be unlikely to change the conclusions.

Arguedas et al. (2002) compared plastic stenting to metal stenting in patients with pancreatic cancer and obstructive jaundice presenting for palliative biliary stenting. The study took a US Societal Perspective and was deemed to have very serious methodological limitations. The study estimated that initial stenting with metal stents would lead to a cost saving of US$433 and a health increase of 0.033 QALYs. This result was robust to all parameters apart from length of survival. Given the age of the study, the US societal perspective, methodological issues and that a contemporary bespoke economic model had been built to answer an almost identical decision problem from a UK NHS and PSS perspective, for the purposes of this guideline it was difficult to give much weight to the conclusions of this economic evaluation.

9.1.5.2. Economic modelling

As this topic was deemed a high economic priority and the previous economic evidence did not fully answer the decision problem, a bespoke economic model was developed. The rationale for economic modelling, methods, results and discussion are reported in full in Chapter 12. This section provides an overview of the methods and results for the bespoke economic model.

9.1.5.3. Overview of methods

A decision-analytical model in the form of a Markov model was developed to evaluate the relative cost effectiveness of different strategies for stenting in people with unresectable or metastatic pancreatic cancer and obstructive jaundice. Three different strategies were considered by the model: a strategy of initial stenting with a plastic stent followed by stenting with a self-expanding metal stent (SEMS) upon dysfunction and initial stenting with SEMS followed by replacement/repositioning upon dysfunction (SEMS/SEMS) compared to a base case strategy of initial plastic stenting replaced with plastic stents upon dysfunction. The model did not consider different types of SEMS (covered, uncovered, partially covered) because it was determined there would not be significant cost differences by type and that the decision of the best type to use would be made wholly on clinical and not economic considerations if the strategies involving SEMS were cost effective. The main outcome of the economic model was incremental cost per QALY compared to the base case strategy. A NHS and PSS perspective was taken. The model had a time horizon of two years which was deemed sufficient to capture the lifetime of the vast majority of the cohort.

Clinical data were derived entirely from studies identified in the accompanying systematic review of clinical evidence. All costs were derived from NHS reference costs. The cost of initial stent insertion were taken from NHS reference costs. This figure would include all pre-operative imaging, the unit costs of the stents, the insertion of the stent and any peri-operative treatment and hospital stay.

NHS Reference costs gave a difference in total insertion costs between insertion of metal stents and plastic stents of £760, slightly less than the difference in unit cost of the different stents as reported in the NHS Supply Catalogues. Where the insertion of the stent is a secondary or later insertion the costs are assumed to be equal to those above apart from where a person is receiving a secondary SEMS stenting having previously received SEMS stenting (i.e. the SEMS/SEMS strategy). In this case the cost is assumed equal to that of receiving a plastic stent. This is because, unlike plastic stents, SEMS can be reused on migration or occlusion and thus the stent costs are not incurred again.

When occlusion or migration is suspected a patient would receive a diagnostic endoscopic procedure to investigate and confirm the suspicion and to rule out any other causes of the associated symptoms. Following this, patients would receive their secondary or later stenting.

During the base case analysis hospital days were not costed. Hospital days were not costed as the reference costs for stent placement allow for some days in hospital. It was likely that costing this difference could lead to double counting of this cost. Days in hospital above those in the perioperative period were costed in line with excess bed days for the procedure. In the base case analysis adverse events were not assigned a cost as it was assumed that these adverse events would often be treated as part of surgical treatment follow-up.

Quality of life weights were taken from 1 Dutch study (Walter et al. 2017), in an identical patient group, using the EQ-5D questionnaire, administered alongside an RCT. The EQ-5D questionnaire scored using Dutch population values showed no difference in quality of life between the SEMS and plastic stent groups. Therefore, the base case analysis was a de facto cost minimisation study. It was hypothesised that the EQ-5D questionnaire was not sensitive enough to pick up quality of life changes between the groups, therefore a secondary analysis was run using the values from the EQ-5D Visual Analogue Scale (VAS) to measure differences in quality of life between the different strategies.

All health and cost outcomes were discounted at a rate of 3.5% per annum.

9.1.5.4. Results of the economic model

In the base case analysis where overall survival and quality of life were assumed equal across the different strategies SEMS/SEMS was the least costly strategy with a cost saving, over the lifetime of one person of over £1500 when compared to the plastic/plastic strategy(Table 110). When scoring from the EQ-5D VAS was included in the secondary model the SEMS/SEMS strategy also lead to the largest amount of QALYs with an additional 0.024 QALYS compared to a plastic/plastic strategy. It was also cost saving and health improving compared to the plastic/SEMS strategy making it dominant compared to all other strategies considered in the base case analysis.

Table 110

Deterministic Base Case Results.

This result was only sensitive to overall survival with plastic stenting followed by plastic stenting becoming the least costly for survival less than 24 days. The robustness of the result is supported by the probabilistic sensitivity analysis. The initial stenting with SEMS strategy is cost saving compared to plastic stenting followed by plastic stenting in 98% of iterations.

9.1.5.5. Conclusions

A strategy of SEMS replaced with SEMS upon dysfunction was the preferred option in the base case results for both deterministic and base case results - being cost saving compared to the other two strategies. When quality of life data from the EQ-5D VAS was used this strategy was also health improving.

These conclusions were robust to both one way deterministic sensitivity analyses and probabilistic sensitivity analysis. SEMS/SEMS was the preferred option in nearly all deterministic sensitivity analysis. The robustness of these results are further highlighted by the probabilistic sensitivity analysis where a SEMS/SEMS strategy is cost saving in greater than 98% of iterations.

The results of this economic model were based on evidence from the clinical evidence review which was derived entirely from RCT evidence. The costings for the model were taken from UK NHS sources and quality of life from a European EQ-5D questionnaire administered alongside an RCT. The results, conclusions and sensitivities are almost identical to the 1 economic evaluation identified by the review of the previous economic evidence review (Arguedas et al. 2002).

9.1.6. Evidence statements

9.1.6.1. Plastic stent versus self-expanding metal stent in adults with pancreatic cancer and biliary obstruction

Relief of obstruction

Very low quality evidence from 3 RCTS (n=229) showed that, when used as either a primary or secondary stent, there is a clinically important difference favouring SEMS on time to dysfunction in adults with unresectable pancreatic cancer compared to plastic stents: HR 2.59 (95% CI 1.67-4.0).

Very low quality evidence from 2 RCTS (n=224) showed that when used as a primary stent, there is a clinically important difference favouring covered or partially-covered SEMS on time to dysfunction in adults with unresectable pancreatic cancer compared to plastic stents: HR 2.26 (95% CI 1.45-3.53).
Very low quality evidence from 1 RCT (n=117) showed that when used as a primary stent, there is a clinically important difference favouring uncovered SEMS on time to dysfunction in adults with unresectable pancreatic cancer compared to plastic stents: HR 3.0 (95% CI 1.45-6.2).
Very low quality evidence from 1 RCT (n=33) showed that when used as a secondary stent, there is a clinically important difference favouring partially-covered SEMS plastic stents on time to dysfunction in adults with unresectable pancreatic cancer compared to plastic stents: HR 6.69 (95% CI 1.39-32.07).
Very low quality evidence from 1 RCT (n=31) showed that when used as a secondary stent, there is a clinically important difference favouring uncovered SEMS on time to dysfunction in adults with unresectable pancreatic cancer compared to plastic stents: HR 9.97 (95% CI 3.46-28.74).

Low quality evidence from 6 RCTs (n=471) showed that there is a clinically important difference favouring SEMS on the number of adults with pancreatic cancer who experience stent occlusion compared to plastic stents: RR 2.25 (95% CI 1.67-3.02).

Very low quality evidence from 4 RCTs (n=258) showed that there is a clinically important difference favouring covered, partially-covered or uncovered SEMS on the number of adults with pancreatic cancer who experience stent occlusion compared to plastic stents: RR 2.2 (95% CI 1.45-3.35).
Very low quality evidence from 2 RCTs (n=213) showed that there is a clinically important difference favouring covered SEMS on the number of adults with pancreatic cancer who experience stent occlusion compared to plastic stents: RR 2.3 (95% CI 1.51-3.49).
Low quality evidence from 5 RCTs (n=417) showed that there is a clinically important difference favouring SEMS on the number of adults with unresectable pancreatic cancer who experience stent occlusion compared to plastic stents: RR 2.36 (95% CI 1.7-3.28).
Low quality evidence from 1 RCT (n=54) showed that there is no clinically important difference between SEMS and plastic stents on the number of adults with resectable, borderline resectable, or locally advanced pancreatic cancer who experience stent occlusion: RR 1.73 (95% CI 0.89-3.34).

Very low quality evidence from 1 RCT (n=113) showed that there is no clinically important difference between plastic stents and SEMS on the number of adults with pancreatic cancer who experience stent migration: RR 0.19 (95% CI 0.02-1.57).

Very low quality evidence from 1 RCT (n=117) showed that there is a clinically important difference favouring partially-covered or uncovered SEMS on the number of adults with pancreatic cancer who experience either stent occlusion or stent migration compared to plastic stents: RR 2.42 (95% CI 1.44-4.06).

Relief of symptoms

No evidence was identified to inform this outcome.

Treatment-related mortality

Very low quality evidence from 1 RCT (n=100) showed no clinically important difference between plastic stents and SEMS on treatment-related mortality in adults with unresectable pancreatic cancer: RR 2.88 (95% CI 0.12-69.16).

Treatment-related morbidity

Low quality evidence from 1 RCT (n=34) showed that there is no clinically important difference between wing-shaped plastic stents and SEMS on the number of adults with unresectable biliary obstruction caused by pancreatic cancer whose serum bilirubin levels decrease by 30% or more after their insertion: RR 0.94 (95% CI 0.79-1.1).

Low quality evidence from 1 RCT (n=98) showed that there is no clinically important difference between plastic stents and SEMS on the rate of change in total serum bilirubin (SMD 0.23 [95% CI −0.62-0.17]) after their insertion in adults with unresectable pancreatic cancer.

Treatment-related complications

Very low quality evidence from 7 RCTs (n=720) showed that there is no clinically important difference between plastic stents and SEMS on the number of adults with pancreatic cancer who experience pancreatitis after their insertion: RR 0.81 (95% CI 0.32-2.04).

Very low quality evidence from 4 RCTs (n=473) showed that there is no clinically important difference between plastic stents and covered, partially covered or uncovered SEMS on the number of adults with pancreatic cancer who experience pancreatitis after their insertion: RR 1.02 (95% CI 0.36-2.92).
Very low quality evidence from 2 RCTs (n=213) showed that there is no clinically important difference between plastic stents and covered, partially covered or uncovered SEMS on the number of adults with pancreatic cancer who experience pancreatitis after their insertion: RR 0.32 (95% CI 0.03-3.01).
Very low quality evidence from 5 RCTs (n=632) showed that there is no clinically important difference between plastic stents and SEMS on the number of adults with unresectable pancreatic cancer who experience pancreatitis after their insertion: RR 1.52 (95% CI 0.51-4.59).
Very low quality evidence from 1 RCT (n=54) showed that there is no clinically important difference between plastic stents and SEMS on the number of adults with resectable, borderline resectable or locally advanced pancreatic cancer who experience pancreatitis after their insertion: RR 0.12 (95% CI 0.01-2.01).

Low quality evidence from 4 RCTs (n=334) showed that there is a clinically important difference favouring SEMS on the number of adults with unresectable pancreatic cancer who experience cholangitis after their insertion compared to the insertion of plastic stents: RR 3.1 (95% CI 1.28-7.48).

Very low quality evidence from 2 RCTs (n=152) showed that there is a clinically important difference favouring covered, partially-covered or uncovered SEMS on the number of adults with unresectable pancreatic cancer who experience cholangitis after their insertion compared to the insertion of plastic stents: RR 1.71 (95% CI 0.5-5.89).
Very low quality evidence from 1 RCT (n=100) showed that there is no clinically important difference between plastic stents and covered SEMS on the number of adults with unresectable pancreatic cancer who experience cholangitis after their insertion: RR 4.81 (95% CI 0.24-97.68).
Very low quality evidence from 1 RCT (n=82) showed that there is a clinically important difference favouring partially-covered SEMS on the number of adults with unresectable pancreatic cancer who experience cholangitis after their insertion compared to the insertion plastic stents: RR 5.0 (95% CI 1.17-21.43).

Very low quality evidence from 4 RCTs (n=448) showed that there is no clinically important difference between plastic stents and SEMS on the number of adults with unresectable pancreatic cancer who experience cholecystitis after their insertion: RR 0.47 (95% CI 0.15-1.53).

Very low quality evidence from 2 RCTs (n=253) showed that there is no clinically important difference between plastic stents and covered, partially-covered or uncovered SEMS on the number of adults with unresectable pancreatic cancer who experience cholecystitis after their insertion: RR 2.56 (95% CI 0.33-20.1).
Very low quality evidence from 1 RCT (n=82) showed that there is no clinically important difference between plastic stents and partially-covered SEMS on the number of adults with unresectable pancreatic cancer who experience cholecystitis after their insertion: RR 0.2 (95% CI 0.01-4.04).
Very low quality evidence from 1 RCT (n=113) showed that there is no clinically important difference plastic stents and covered SEMS on the number of adults with unresectable pancreatic cancer who experience cholecystitis after their insertion: RR 0.11 (95% CI 0.01-1.91).

Very low quality evidence from 1 RCT (n=118) showed that there is no clinically important difference between plastic stents and covered SEMS on the number of adults with unresectable pancreatic cancer who experience post-endoscopic sphincterotomy haemorrhage after their insertion: RR 3.0 (95% CI 0.12-72.18).

Very low quality evidence from 2 RCTs (n=197) showed that there is no clinically important difference between plastic stents and SEMS on the number of days adults with unresectable pancreatic cancer are hospitalised after their insertion: SMD 0.49 (95% CI 0.21-0.77).

Overall survival

Very low quality evidence from 3 RCTS (n=247) showed no significant difference between plastic stents and SEMS on overall survival in adults with unresectable pancreatic cancer: HR 1 (95% CI 0.75-1.31).

Time to definitive treatment

No evidence was identified to inform this outcome.

Health-related quality of life

No evidence was identified to inform this outcome.

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.1.6.2. Covered self-expanding metal stent versus uncovered self-expanding metal stent in adults with pancreatic cancer and biliary obstruction

Narrative summary for overall survival

The 5 included RCTs did not report data for overall survival in a way that allowed a meta-analysis. Overall the studies were at high risk of bias due to selective (e.g. incomplete) reporting of outcomes, other sources of bias (such as significant differences at baseline), and insufficient information about the randomisation method or allocation concealment. None of the studies reported the hazard ratios and associated 95% confidence intervals. Unlike the other studies – all of which used ‘standard’ covered SEMSs (e.g. with a silicone membrane) - Krokidis 2011 used an SEMS with an expanded polytetrafluoroethylene/fluorinated-ethylene-propylene covering. Median overall survival of a covered SEMS ranged from 116 days to 285 days (1 study reported a mean of 71 days), whilst for an uncovered SEMS it ranged from 155 to 222 days. One study (Gardner et al., 2016) reported a mean overall survival of 71 (range 7-196) days for covered SEMS and 242 (range 122-453) days for an uncovered SEMS. One study (Krokidis et al., 2011) reported a near significant difference (p=0.06) on overall survival favouring a covered SEMS over an uncovered SEMS, three studies (Kitano et al., 2013, Kullman et al., 2010, Ung et al., 2013) reported no difference between them, and 1 study did not provide a p-value. However, all of the participants in this study were receiving neoadjuvant therapy.

Narrative summary for relief of obstruction (cumulative stent patency)

The 5 included RCTs did not report data for cumulative stent patency (time to obstruction) in a way that allowed a meta-analysis. Overall the studies were at high risk of bias due to selective (e.g. incomplete) reporting of outcomes, other sources of bias (such as significant differences at baseline), and insufficient information about the randomisation method or allocation concealment. None of the studies reported the hazard ratios and associated 95% confidence intervals. Unlike the other studies – all of which used ‘standard’ covered SEMSs (e.g. with a silicone membrane) - Krokidis 2011 used an SEMS with an expanded polytetrafluoroethylene/fluorinated-ethylene-propylene covering; all of the participants in this study were also receiving neoadjuvant therapy. Median stent patency for a covered SEMS ranged from 153 to 583 days, whilst for an uncovered SEMS it ranged from 127 to 314 days. One study (Gardner et al., 2016) reported a mean stent patency of 220 days (range 21-341) for a covered SEMS and 74 days (range 45-90) for an uncovered SEMS. Two studies (Kitano et al., 2013, Krokidis et al., 2011) reported a significant difference on stent patency favouring a covered SEMS over an uncovered SEMS, two studies (Kullman et al., 2010, Ung et al., 2013), reported no significant difference between them, whilst 1 study (Gardner et al., 2016) did not provide a p-value.

Relief of obstruction

Very low quality evidence from 5 RCTs (n=701) showed that there is no clinically important difference between covered and uncovered SEMS on the number of people experiencing stent dysfunction: RR 0.81 (95% CI 0.61-1.05).

Very low quality evidence from 3 RCTs (n=600) showed that there is a clinically important difference favouring uncovered SEMS on the number of stent dysfunctions caused by sludge formation compared to covered SEMS in adults with pancreatic cancer and biliary obstruction: RR 2.43 (95% CI 1.22-4.85).
Very low quality evidence from 2 RCTs (n=520) showed that there is no clinically important difference between covered and uncovered SEMS on the number of stent dysfunctions caused by stent migration in adults with pancreatic cancer and biliary obstruction: RR 13 (95% CI 0.74-229.23).
Very low quality evidence from 3 RCTs (n=600) showed that there is a clinically important difference favouring covered SEMS on the number of stent dysfunctions caused by tumour ingrowth compared to uncovered SEMS in adults with pancreatic cancer and biliary obstruction: RR 0.36 (95% CI 0.2-0.64).
Very low quality evidence from 3 RCTs (n=600) showed that there may be a clinically important difference favouring uncovered SEMS on the number of stent dysfunctions caused by tumour overgrowth compared to covered SEMS in adults with pancreatic cancer and biliary obstruction, although there is some uncertainty: RR 1.88 (95% CI 0.97-3.66).

Relief of symptoms

No evidence was identified to inform this outcome.

Treatment-related mortality

No evidence was identified to inform this outcome.

Treatment-related morbidity

No evidence was identified to inform this outcome.

Treatment-related complications

Very low quality evidence from 4 RCTs (n=668) showed that there is no clinically important difference between covered and uncovered SEMS on the number of adults with pancreatic cancer and biliary obstruction who experience adverse events: RR 0.89 (95% CI 0.52-1.51).

Very low quality evidence from 1 RCT (n=400) showed that there is no clinically important difference between covered and uncovered SEMS on the number of adults with pancreatic cancer and biliary obstruction who experience cholangitis (RR 0.67 [95% CI 0.28-1.6]) and retroperitoneal perforation (RR 1.0 [95% CI 0.06-15.88]).
Very low quality evidence from 2 RCTs (n=520) showed that there is no clinically important difference between covered and uncovered SEMS on the number of adults with pancreatic cancer and biliary obstruction who experience cholecystitis: RR 0.75 (95% CI 0.17-3.31).
Very low quality evidence from 2 RCTs (n=480) showed that there is no clinically important difference between covered and uncovered SEMS on the number of adults with pancreatic cancer and biliary obstruction who experience haemorrhage: RR 0.71 (95% CI 0.14-3.52).
Very low quality evidence from 3 RCTs (n=588) showed that there is no clinically important difference between covered and uncovered SEMS on the number of adults with pancreatic cancer and biliary obstruction who experience pancreatitis: RR 1.2 (95% CI 0.37-3.89).
Very low quality evidence from 1 RCT (n=80) showed that there is no clinically important difference between covered and uncovered SEMS on the number of adults with pancreatic cancer and biliary obstruction who experience peritoneal irritation: RR 1.5 (95% CI 0.26-8.5).
Very low quality evidence from 1 RCT (n=68) showed that there is no clinically important difference between covered and uncovered SEMS on the number of adults with pancreatic cancer and biliary obstruction who experience sepsis: RR 3.0 (95% CI 0.13-71.15).

Time to definitive treatment

No evidence was identified to inform this outcome.

Health-related quality of life

No evidence was identified to inform this outcome.

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.1.6.3. Partially-covered self-expanding metal stent versus uncovered self-expanding metal stent in adults with pancreatic cancer and biliary obstruction

Narrative summary for overall survival and relief of obstruction (cumulative stent patency)

The 2 included RCTs did not report data for overall survival and cumulative stent patency (time to obstruction) in a way that allowed a meta-analysis. Overall the 2 studies were at high/unclear risk of bias due to selective reporting of outcomes. None of the studies reported the hazard ratios and associated 95% confidence intervals. Only 1 study (Telford et al., 2010) reported median overall survival by group, which was not significant (227 days for a partially covered SEMS and 239 days for an uncovered SEMS). Median stent patency ranged from 285 to 357 days for a partially covered SEMS compared to 268 to 711 days for an uncovered SEMS. One study (Telford et al., 2010) reported no significant difference between partially covered and uncovered SEMS, whilst 1 study (Walter et al, 2015) did not provide a p-value.

Relief of obstruction

Very low quality evidence from 2 RCTs (n=243) showed that there is no clinically important difference between partially-covered and uncovered SEMS on the number of adults with pancreatic cancer and biliary obstruction who experience stent dysfunction from any cause: RR 1.35 (95% CI 0.81-2.23)

Very low quality evidence from 1 RCT (n=129) showed that there may be a clinically important difference favouring uncovered SEMS on the number of stent dysfunctions caused by stent migration compared to a partially-covered SEMS in adults with pancreatic cancer and biliary obstruction: RR 15.28 (95% CI 0.9-259.23).

Relief of symptoms

No evidence was identified to inform this outcome.

Treatment-related mortality

No evidence was identified to inform this outcome.

Treatment-related morbidity

No evidence was identified to inform this outcome.

Treatment-related complications

Very low quality evidence from 1 RCT (n=129) showed that there may be a clinically important difference favouring uncovered SEMS on the number of adverse events compared to a partially-covered SEMS in adults with pancreatic cancer and biliary obstruction, although there is some uncertainty: RR 1.4 (95% CI 1.0-1.96).

Very low quality evidence from 2 RCTs (n=-275) showed that there is no clinically important difference between partially-covered and uncovered SEMS on the number of adults with pancreatic cancer and biliary obstruction who experience pancreatitis (RR 0.97 [95% CI 0.14-6.58]) or other adverse events (RR 1.14 [95% CI 0.66-1.99]).
Very low quality evidence from 2 RCTs (n=-237) showed that there is no clinically important difference between partially-covered and uncovered SEMS on the number of adults with pancreatic cancer and biliary obstruction who experience cholecystitis: RR 0.98 (95% CI 0.21-4.59).

Time to definitive treatment

No evidence was identified to inform this outcome.

Health-related quality of life

No evidence was identified to inform this outcome.

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.1.6.4. Paclitaxel-eluting self-expanding metal stent versus covered self-expanding metal stent in adults with an unresectable distal malignant biliary obstruction

Relief of obstruction

Very low quality evidence from 1 RCT (n=52) showed that there is no clinically important difference between paclitaxel-eluting and covered SEMS on time to stent dysfunction in adults with an unresectable distal malignant biliary obstruction: HR 0.53 (95% CI 0.16-1.78).

Very low quality evidence from 1 RCT (n=25) showed that there is no clinically important difference between paclitaxel-eluting and covered SEMS on increasing time to stent dysfunction in adults with an unresectable distal malignant biliary obstruction caused by pancreatic cancer: HR 0.52 (95% CI 0.1-3.09).

Relief of symptoms

No evidence was identified to inform this outcome.

Treatment-related mortality

No evidence was identified to inform this outcome.

Treatment-related morbidity

No evidence was identified to inform this outcome.

Treatment-related complications

Very low quality evidence from 1 RCT (n=52) showed that there is no clinically important difference between paclitaxel-eluting and covered SEMS on the number of adults with an unresectable distal malignant biliary obstruction who experience cholangitis symptoms (RR 7.28 [95% CI 0.4-133.89]) and pancreatitis (RR 1.04 [95% CI 0.07-15.73]) within 30 days of surgery.

Overall survival

Very low quality evidence from 1 RCT (n=52) showed that there is no clinically important difference between paclitaxel-eluting and covered SEMS on overall survival in adults with an unresectable distal malignant biliary obstruction: HR 1.19 (95% CI 0.65-2.18).

Very low quality evidence from 1 RCT (n=25) showed that there is no clinically important difference between paclitaxel-eluting and covered SEMS on overall survival in adults with an unresectable distal malignant biliary obstruction caused by pancreatic cancer: HR 0.85 (95% CI 0.35-2.06).

Time to definitive treatment

No evidence was identified to inform this outcome.

Health-related quality of life

No evidence was identified to inform this outcome.

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.1.6.5. Preoperative endoscopic biliary drainage (PEBD) then surgery versus surgery in adults with suspected pancreatic cancer

Relief of obstruction

No evidence was identified to inform this outcome.

Relief of symptoms

No evidence was identified to inform this outcome.

Treatment-related mortality

Very low quality evidence from 1 RCT (n=196) showed that there is no clinically important difference between PEBD followed by surgery and surgery only in adults with pancreatic cancer on mortality at 120 days (RR 1.15 [95% CI 0.57-2.33]) nor on treatment-related mortality (RR 2.07 [95% CI 0.66-6.51]).

Very low quality evidence from 1 RCT (n=185) showed that there is no clinically important difference between PEBD followed by surgery and surgery only in adults with pancreatic cancer on mortality at 2 years: RR 0.96 (95% CI 0.84-1.09).

Treatment-related morbidity

No evidence was identified to inform this outcome.

Treatment-related complications

Very low quality evidence from 1 RCT (n=196) showed that there is a clinically important difference favouring surgery on the total number of adults with pancreatic cancer who experience protocol-specific complications (RR 1.87 [95% CI 1.42-2.46]), surgery-related complications (RR 1.26 [95% CI 0.91 to 1.76]), pre-surgery cholangitis (RR 12.44 [95% CI 3.04 to 50.89]), and the number that are hospitalised due to protocol-specific complications (RR 2.85 [95% CI 1.53-5.2]) compared to PEBD followed by surgery.

Very low quality evidence from 1 RCT (n=196) showed that there is a clinically important difference favouring surgery only on the rate of serious complications within 120 days of randomisation compared to PEBD followed by surgery: HR 1.86 (95% CI 1.41-2.45).

Very low quality evidence from 1 RCT (n=196) showed that there may be a clinically important difference favouring surgery only on the number of adults with pancreatic cancer who experience pre-surgery pancreatitis compared to PEBD followed by surgery, although there may be some uncertainty: RR 13.83 [95% CI 0.8 to 238.96].

Very low quality evidence from 1 RCT (n=196) showed that there is no clinically important difference between PEBD followed by surgery and surgery only on the number of adults with pancreatic cancer who experience pre-surgery post-ERCP haemorrhage (RR 4.61 [95% CI 0.22-94.83]), pre-surgery perforation (RR 4.61 [95% CI 0.22 to 94.83]), surgery-related haemorrhage (RR 0.46 [95% CI 0.09-2.46]), surgery-related cholangitis (RR 0.92 (95% CI 0.19 to 4.45) and surgery-related pneumonia (RR 1.66 [95% CI 0.58 to 4.77]).

Overall survival

Very low quality evidence from 1 RCT (n=113) showed that there is no clinically important difference between PEBD followed by curative surgery and curative surgery only in adults with resectable or borderline resectable pancreatic cancer after undergoing resection on overall survival at 2 years: HR 0.98 (95% CI 0.72-1.34).
Very low quality evidence from 1 RCT (n=67) showed that there is no clinically important difference between PEBD followed by palliative surgery and palliative surgery only in adults with unresectable pancreatic cancer on overall survival at 2 years: HR 1.02 (95% CI 0.63-1.67).

Time to definitive treatment

Very low quality evidence from 1 RCT (n=196) showed that there is a clinically important difference favouring surgery only on the delay to surgery in adults with pancreatic cancer compared to PEBD followed by surgery: MD 4.0 (95% CI 3.58-4.42).

Health-related quality of life

No evidence was identified to inform this outcome.

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.1.6.6. Endoscopic sphincterotomy then stent versus stent in adults with unresectable pancreatic cancer

Relief of obstruction

Very low quality evidence from 3 RCTs (n=454) showed that there is no clinically important difference between endoscopic sphincterotomy followed by a stent and stent only on decreasing the number of adults with unresectable pancreatic cancer who experience stent occlusion (RR 0.91 [95% CI 0.55-1.52]), stent migration (RR 1.84 [95% CI 0.75 to 4.54]).

Relief of symptoms

No evidence was identified to inform this outcome

Treatment-related mortality

No evidence was identified to inform this outcome.

Treatment-related morbidity

Moderate quality evidence from 1 RCT (n=200) showed that there is no clinically important difference between endoscopic sphincterotomy followed by a stent and stent only on the number of adults with unresectable pancreatic cancer that die due to disease progression: RR 0.86 (95% CI 0.72-1.02).

Treatment-related complications

Very low quality evidence from 2 RCTs (n=376) showed that there is no clinically important difference between endoscopic sphincterotomy followed by a stent and stent only on the number of adults with unresectable pancreatic cancer who experience early complications within 30 days (RR 1.24 [95% CI 0.61 to 2.5]) and early stent-related pancreatitis (95% CI RR 1.11 [0.49 to 2.54]).

Low quality evidence from 1 RCT (n=200) showed that there is no clinically important difference between endoscopic sphincterotomy followed by a stent and stent only on the number of adults with unresectable pancreatic cancer who experience early stent-related complications within 30 days (RR 1.0 [95% CI 0.52 to 1.93]) and late stent-related complications after 30 days (RR 1.2 [95% CI 0.38 to 3.81]).

Very low quality evidence from 3 RCTs (n=450) showed that there is no clinically important difference between endoscopic sphincterotomy followed by a stent and stent only on the number of adults with unresectable pancreatic cancer who experience pancreatitis within 30 days: RR 1.11 (95% CI 0.49 to 2.54).

Low quality evidence from 1 RCT (n=194) showed that there is no clinically important difference between endoscopic sphincterotomy followed by a stent and stent only on the number of adults with unresectable pancreatic cancer who experience perforation within 30 days: RR 0.34 (95% CI 0.01-8.25).

Low quality evidence from 1 RCT (n=184) showed that there is no clinically important difference between endoscopic sphincterotomy followed by a stent and stent only on the number of adults with unresectable pancreatic cancer who experience cholecystitis within 30 days and after 30 days: RR 0.26 (95% CI 0.03-2.24) for both outcomes.

Very low quality evidence from 1 RCT (n=182) showed that there is no clinically important difference between endoscopic sphincterotomy followed by a stent and stent only on the number of adults with unresectable pancreatic cancer who experience cholangitis after 30 days: RR 1.04 (95% CI 0.55 to 1.98).

Overall survival

No evidence was identified to inform this outcome.

Time to definitive treatment

No evidence was identified to inform this outcome.

Health-related quality of life

No evidence was identified to inform this outcome.

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.1.6.7. Endoscopic sphincterotomy then stent versus surgical bypass in adults with unresectable pancreatic cancer

Relief of obstruction

Low to very low quality evidence from 1 RCT (n=30) showed that there is no clinically important difference between endoscopic sphincterotomy followed by a covered stent and surgical bypass on relief of biliary obstruction in adults with unresectable pancreatic cancer: RR 1.0 (95% CI 0.88-1.13).

Relief of symptoms

No evidence was identified to inform this outcome.

Treatment-related mortality

No evidence was identified to inform this outcome.

Treatment-related morbidity

Low to very low quality evidence from 1 RCT (n=30) showed there is no clinically important difference between endoscopic sphincterotomy followed by a covered stent and surgical bypass on the number of people whose bilirubin level is less than 2.5 mg/dL on day 30 (RR 1 [95% CI 0.51 to 1.95]) nor on serum bilirubin levels at day 30 (MD −0.3 [95% CI −1.06-0.46]) in adults with unresectable pancreatic cancer.

Treatment-related complications

Very low quality evidence from 1 RCT (n=30) showed that there is no clinically important difference between endoscopic sphincterotomy followed by a covered stent and surgical bypass on treatment-related hospitalisation (RR 1.5 [95% CI 0.71-3.16]), stent-related complications (RR 9 [95% CI 0.53-153.79]), treatment-related early complications (RR 0.6 [95% CI 0.17-2.07]), treatment-related late complications (RR 0.75 [95% CI 0.2- 2.79]), post operative complications (RR 0.71 [95% CI 0.29-1.75]), pneumonia (RR 0.2 [95% CI 0.01-3.85]), post-ERCP pancreatitis (RR 3 [95% CI 0.13-68.26]) in adults with unresectable pancreatic cancer.

Overall survival

No evidence was identified to inform this outcome.

Time to definitive treatment

No evidence was identified to inform this outcome.

Health-related quality of life

Low quality evidence from 1 RCT (n=30) showed that there is a clinically important difference favouring endoscopic sphincterotomy followed by a covered stent on SF-36 overall quality of life scores at 30 days (SMD 0.78 [0.04-1.52]) and 60 days (SMD 0.75 [0.01-1.49]) in adults with unresectable pancreatic cancer, compared to surgical bypass.

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.1.6.8. Endoscopic ultrasound-guided choledochoduodenostomy (EUS-CD) and stent versus percutaneous transhepatic biliary drainage (PTBD) in adults with an unresectable malignant biliary obstruction where either ERCP or EUS-guided transpapillary rendezvous has failed

Relief of obstruction

No evidence was identified to inform this outcome.

Relief of symptoms

No evidence was identified to inform this outcome.

Treatment-related mortality

No evidence was identified to inform this outcome.

Treatment-related morbidity

Very low quality evidence from 1 RCT (n=25) showed that there is no clinically important difference in the effect of EUS-CD compared to PTBD on total serum bilirubin at 7 days (SMD −0.53 [95% CI −1.33-0.27]) and 30 days (SMD 0.42 [95% CI −0.37-1.22]) in adults with unresectable malignant biliary obstruction where ERCP has failed.

Very low quality evidence from 1 RCT (n=25) showed that EUS-CD has a clinically significant benefit of lowering gamma glutamyl transferase levels at 7 days in adults with unresectable malignant biliary obstruction where ERCP has failed compared to PTBD: SMD −0.87 (95% CI −1.69-−0.05).

Very low quality evidence from 1 RCT (n=25) showed that EUS-CD may have a clinically significant benefit in lowering alkaline phosphatase levels at 7 days in adults with unresectable malignant biliary obstruction where ERCP has failed compared to PTBD, although there is some uncertainty: SMD −0.73 (95% CI −1.54-0.08).

Treatment-related complications

Very low quality evidence from 1 RCT (n=25) showed that there is no clinically important difference in the effect of EUS-CD compared to PTBD on the number of adults with unresectable malignant biliary obstruction where ERCP has failed who experience treatment-related complications: RR 0.62 (95% CI 0.12-3.07).

Overall survival

No evidence was identified to inform this outcome.

Time to definitive treatment

No evidence was identified to inform this outcome.

Health-related quality of life

Very low quality evidence from 1 RCT (n=25) showed that there is no clinically important difference in the effect of EUS-CD compared to PTBD on SF-36 quality of life scores at 7 days (SMD −0.29 [95% CI −1.08-0.5]) and 30 days (SMD −0.31 [95% CI −1.1-0.48]) in adults with unresectable malignant biliary obstruction where ERCP has failed.

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.1.6.9. Endoscopic ultrasound-guided choledochoduodenostomy (EUS-CD) and stent versus surgical bypass in adults with an unresectable malignant biliary obstruction where ERCP has failed

Relief of obstruction

No evidence was identified to inform this outcome.

Relief of symptoms

No evidence was identified to inform this outcome.

Treatment-related mortality

No evidence was identified to inform this outcome.

Treatment-related morbidity

Very low quality evidence from 1 RCT (n=29) showed that there is no clinically important difference in the effect of EUS-CD after 7 days on the number of adults with unresectable malignant biliary obstruction (and where ERCP has failed) whose total serum bilirubin levels are reduced 50% or more compared to those who have surgical bypass: RR 0.77 (95% CI 0.54-1.09).

Very low quality evidence from 1 RCT (n=29) showed that EUS-CD may have a clinically significant effect on decreasing total serum bilirubin at 7 days compared to surgical bypass in adults with unresectable malignant biliary obstruction where ERCP has failed compared to those who have surgical bypass, although there is some uncertainty: MD 1.71 (95% CI −0.24-3.66).

Very low quality evidence from 1 RCT (n=29) showed that there is no clinically important difference in the effect of EUS-CD at 7 days on decreasing gamma glutamyl transferase (MD 116.46 [95% CI 34.63 to 198.29]) nor alkaline phosphatase (MD 64.54 [95% CI 16.34 to 112.74]), compared to surgical bypass in adults with unresectable malignant biliary obstruction where ERCP has failed.

Very low quality evidence from 1 RCT (n=29) showed that there is no clinically important difference in the effect of EUS-CD at 30 days on decreasing total serum bilirubin (MD 0.26 [95% CI −0.37-0.89]), gamma glutamyl transferase (MD 53.83 [95% CI −20.42-128.08], nor alkaline phosphatase (MD 11.39 [95% CI −22.16-44.94]), compared to surgical bypass in adults with unresectable malignant biliary obstruction where ERCP has failed.

Very low quality evidence from 1 RCT (n=25) showed that there is no clinically important difference in the effect of EUS-CD at 60 days on decreasing total serum bilirubin (MD 0.06 [95% CI −0.31-0.43]), gamma glutamyl transferase (MD 0.22 [95% CI −16.88-17.32]), nor alkaline phosphatase (MD 4.79 [95% CI −7.11-16.69]) compared to surgical bypass in adults with unresectable malignant biliary obstruction where ERCP has failed.

Very low quality evidence from 1 RCT (n=13) showed that there is no clinically important difference in the effect of EUS-CD at 90 days on decreasing total serum bilirubin (MD 0.01 [95% CI −0.58-0.6]), gamma glutamyl transferase (MD 14.43 [95% CI −2.3-31.16]) nor alkaline phosphatase (MD 5.4 [95% CI −4.87-15.67]), compared to surgical bypass in adults with unresectable malignant biliary obstruction where ERCP has failed.

Treatment-related complications

Very low quality evidence from 1 RCT (n=29) showed that there is no clinically important difference in the effect of EUS-CD on the number of treatment-related complications compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: RR 1.61 (95% CI 0.31-8.24).

Overall survival

Very low quality evidence from 1 RCT (n=29) showed that there is no clinically important difference in the effect of EUS-CD on overall survival, compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: HR 0.64 (95% CI 0.23-1.8).

Time to definitive treatment

No evidence was identified to inform this outcome.

Health-related quality of life

Very low quality evidence from 1 RCT (n=29) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 functional capacity at 7 days (MD 6.3 [95% CI −5.12-17.72]) and 30 days (MD 10.7 [95% CI 0.93-20.47]), compared to surgical bypass in adults with unresectable malignant biliary obstruction where ERCP has failed.

Very low quality evidence from 1 RCT (n=29) showed that there is no clinically important difference in the effect of EUS-CD on SF-36 physical health scores at 7 days (MD 1.5 [95% CI −11.76-14.76]) and 30 days (MD −4.9 [95% CI −18.55-8.75]) compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed.

Very low quality evidence from 1 RCT (n=29) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 pain scores at 7 days (MD −3.7 [95% CI −17.22-9.82]) and 30 days (MD 2.7 [95% CI −9.6-15.0]) compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed.

Very low quality evidence from 1 RCT (n=29) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 general health scores at 7 days (MD - 3.4 [95% CI −10.15-3.35]) and 30 days (MD −4.1 [95% CI −11.85-3.65]) compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed.

Very low quality evidence from 1 RCT (n=29) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 vitality scores at 7 days (MD 2.7 [95% CI −5.64-11.04]) and 30 days (MD 7.6 [95% CI −2.43-17.63]) compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed.

Very low quality evidence from 1 RCT (n=29) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 social role functioning scores at 7 days (MD −0.3 [95% CI −9.69-9.09]) and 30 days (MD 0.3 [95% CI −7.56-8.16]) compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed.

Very low quality evidence from 1 RCT (n=29) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 emotional role functioning scores at 7 days (MD 2.5 [95% CI −11.19-16.19]) and 30 days (MD 0.9 [95% CI −15.69-17.49]) compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed.

Very low quality evidence from 1 RCT (n=29) showed that there is a clinically important difference in the effect of EUS-CD on improving SF-36 mental health scores at 7 days (MD 9.1 [95% CI 1.49-16.71]) and 30 days (MD 12.9 [95% CI 4.63-21.17]) compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed.

Very low quality evidence from 1 RCT (n=26) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 functional capacity scores at 60 days compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: MD 9.9 (95% CI 1.04-18.76).

Very low quality evidence from 1 RCT (n=13) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 functional capacity scores at 90 days compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: MD −1.8 (95% CI −9.86-6.26).

Very low quality evidence from 1 RCT (n=26) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 physical health scores at 60 days compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: MD 6.8 (95% CI −5.67-19.27).

Very low quality evidence from 1 RCT (n=13) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 physical health scores at 90 days compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: MD −10.1 (95% CI −33.62-13.42).

Very low quality evidence from 1 RCT (n=26) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 pain scores at 60 days compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: MD −4.4 (95% CI −17.51-8.71).

Very low quality evidence from 1 RCT (n=26) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 general health scores at 60 days compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: MD −3.3 (95% CI −10.58-3.98).

Very low quality evidence from 1 RCT (n=13) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 general health scores at 90 days compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: MD 4.5 (95% CI −7.44-16.44).

Very low quality evidence from 1 RCT (n=26) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 vitality scores at 60 days compared to surgical bypass in adults with unresectable malignant biliary obstruction where ERCP has failed: MD 2.14 (95% CI −8.61-12.81).

Very low quality evidence from 1 RCT (n=13) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 vitality scores at 90 days compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: MD 14.6 (95% CI −3.2-32.4).

Very low quality evidence from 1 RCT (n=26) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 social role functioning scores at 60 days compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: MD −1.1 (95% CI −12.32-10.12).

Very low quality evidence from 1 RCT (n=13) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 social role functioning scores at 90 days compared to surgical bypass in adults with unresectable malignant biliary obstruction where ERCP has failed: MD 1.5 (95% CI −9.73-12.73).

Very low quality evidence from 1 RCT (n=26) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 emotional role functioning scores at 60 days compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: MD 9.5 (95% CI −11.05-30.05).

Very low quality evidence from 1 RCT (n=13) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 emotional role functioning scores at 90 days compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: MD 8.7 (95% CI −15.33-32.73).

Very low quality evidence from 1 RCT (n=26) showed that there may be a clinically important difference in the effect of EUS-CD on improving SF-36 mental health scores at 60 days compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed, although there is some uncertainty: MD 8.9 (95% CI 0.92-18.72).

Very low quality evidence from 1 RCT (n=14) showed that there is no clinically important difference in the effect of EUS-CD on improving SF-36 mental health scores at 90 days compared to surgical bypass, in adults with unresectable malignant biliary obstruction where ERCP has failed: MD 1.9 (95% CI −9.98-13.78).

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.1.7. Recommendations

33.

Offer resectional surgery rather than preoperative biliary drainage to people who:

have resectable pancreatic cancer and obstructive jaundice and
are well enough for the procedure and
are not enrolled in a clinical trial that requires preoperative biliary drainage.

34.

During attempted resection for pancreatic cancer, consider surgical biliary bypass if the cancer is found to be unresectable.

35.

If biliary drainage is needed in a person who has resectable pancreatic cancer and obstructive jaundice and is not yet fit enough for resectional surgery, offer endoscopically placed self-expanding metal stents.

36.

For people with suspected pancreatic cancer who may need their stent removed later on, consider endoscopically placed self-expanding fully covered metal stents.

37.

Offer endoscopically placed self-expanding metal stents rather than surgical biliary bypass to people with unresectable pancreatic cancer.

9.1.8. Evidence to recommendations

9.1.8.1. Relative value placed on the outcomes considered

The committee considered relief of obstruction, relief of symptoms, treatment-related mortality, treatment related morbidity, treatment-related complications, overall survival, time to definitive treatment, health-related quality of life, patient experience and PROMS to be the critical outcomes for this question.

Patient experience and PROMS were not reported for any comparisons of interest. Relief of obstruction, relief of symptoms, treatment-related mortality and morbidity, time to definitive treatment and quality of life were only reported by a few studies. Treatment related complications and overall survival were reported by the majority of studies. The majority of studies also reported the outcome of stent dysfunction which the committee agreed was a useful outcome to consider.

9.1.8.2. Quality of evidence

The quality of the outcomes for the comparisons identified by this review were as follows:

Plastic stent versus self-expanding metal stent (SEMS) – ranged from very low to low
Covered SEMS versus uncovered SEMS – very low
Partially-covered SEMS versus uncovered SEMS - very low
Paclitaxel-eluting SEMS versus covered SEMS - very low
Preoperative endoscopic biliary drainage then surgery versus surgery – very low
Endoscopic sphincterotomy then stent versus stent – ranged from very low to moderate
Endoscopic sphincterotomy then stent versus surgical bypass – ranged from very low to low
Endoscopic ultrasound-guided choledochoduodenostomy (EUS-CD) versus percutaneous transhepatic biliary drainage – very low
Endoscopic ultrasound-guided choledochoduodenostomy (EUS-CD) versus surgical bypass – very low

The committee noted that several of the studies included people who did not have pancreatic cancer. They agreed to focus on those studies which included at least 66% pancreatic cancer as they considered this proportion would be high enough for the data to be representative of the population under consideration by this guideline.

The committee decided to include three studies that either had less than 66% pancreatic cancer patients or did not report the composition of the samples because the studied interventions were deemed to be sufficiently novel to merit consideration. Each of these studies was the only study that contributed data to the relevant three comparisons: paclitaxel-eluting SEMS versus covered SEMS; EUS-CD versus percutaneous transhepatic biliary drainage; and EUS-CD versus surgical bypass. In relation to the two studies on EUS-guided biliary drainage, it was unclear how many patients had pancreatic cancer and the sample sizes were very small so it was difficult to draw conclusions from these data. Given this, and the fact that this is a relatively new technique, the committee agreed not to make any recommendations about this intervention.

9.1.8.3. Consideration of clinical benefits and harms

The committee noted, based on the evidence, that preoperative biliary drainage was associated with an increased delay to surgery, more complications, more serious complications within 120 days, more hospitalisations and more people experiencing pre-surgery pancreatitis compared to surgery alone. Given this evidence, and the results of the published economic analysis showing that going straight to surgery was both cost saving and health improving, the committee made a strong recommendation to offer surgery to people with resectable pancreatic cancer. Based on their clinical knowledge, the committee also noted that there are ongoing clinical trials which require the insertion of a biliary stent to meet the inclusion criteria of the trial protocol. They were conscious that they did not want these recommendations to restrict entry into such clinical trials and therefore agreed to add a caveat that surgery should be offered, when outside of a clinical trial of preoperative biliary drainage.

The committee noted, based on the evidence, that the time to dysfunction was shorter with plastic stents compared with SEMS and that there was a decrease in stent occlusion and stent migration with SEMS. Moreover, whilst there was no difference in the number of people experiencing pancreatitis or cholecystitis with the different types of stent, the number of people experiencing cholangitis was lower after the insertion of an SEMS. Given this evidence, and the results of the bespoke economic model showing that SEMS was the most cost effective intervention, the committee made a strong recommendation for the use of SEMS, rather than plastic stents, in people with pancreatic cancer and biliary obstruction. They agreed, based on their knowledge and experience, that stent placement should be done endoscopically as this is safer than percutaneous insertion.

The committee noted, based on their experience, that sometimes a stent has to be inserted to relieve the biliary obstruction before it is known whether pancreatic cancer is the cause of this obstruction. In those people where pancreatic cancer does not turn out to be the cause of the obstruction, the stent is likely to need removal. The committee noted that the evidence comparing covered and partially covered SEMS with uncovered SEMS had not identified any clinically significant differences in effects between the two. However, they agreed based on their knowledge, that fully covered metal stents should be considered where it is possible that stent removal may be required, because it can be very difficult to remove uncovered or partially covered metal stents. The committee also acknowledged the importance of fitness for reseactional surgery of people who have resectable pancreatic cancer and obstructive jaundice in need of biliary drainage. Based on the evidence on the effectivenss of SEMS committee therefore made a strong recommendation to offer endoscopically placed self-expanding metal stents to people who have resectable pancreatic cancer and obstructive jaundice and are not fit enough for resectable surgery.

The committee noted that there would be a group of people who had biliary obstruction but whose pancreatic cancer was unresectable and recommendations were needed for this group too. Based on the evidence, the committee agreed that endoscopic stenting was associated with improvements in quality of life compared to surgical bypass. They, therefore, made a strong recommendation for endoscopic stenting in people with unresectable pancreatic cancer as stent placement would avoid a major operation in someone who was likely to be quite poorly. Based on their knowledge and experience the committee also agreed to recommend that surgical biliary bypass should be considered for people whose pancreatic cancer was deemed unresectable during an attempted resection. This would mean the person would not need to have a potential additional procedure in future to insert a stent.

Given that the data for the other comparisons of interest had not demonstrated any difference between interventions, the committee agreed not to make any further recommendations.

The committee considered that the potential benefits of the recommendations made would be earlier treatment of biliary obstruction, improved symptom control, a reduction in the complications associated with stent insertion (as metal stents are less likely to occlude or migrate than plastic stents) and avoidance of unnecessary repeat stenting procedures (as metal stents are less likely to become dysfunctional). The committee noted that the potential harms could be duodenal perforation, bleeding and post procedure pancreatitis from stenting or biliary leaking and anastomotic leakage from surgical bypass. However, without these interventions the person would die so they considered that the harms were balanced by the potential benefits.

9.1.8.4. Consideration of economic benefits and harms

The literature search for previous economic evaluations identified 2 relevant economic evaluations (Morris [2015] and Arguedas [2002]). Both economic evaluations considered different interventions in different patient groups and therefore meaningful comparisons between the studies could not be drawn. A bespoke economic model was also built to help inform recommendations.

Morris (2015) compared preoperative biliary drainage (PBD) to direct surgery in patients with potentially resectable pancreatic or periampullary cancer and obstructive jaundice from a UK NHS and PSS perspective. The study was deemed to only have minor methodological limitations.

The effectiveness side of the model was nearly entirely based on 1 Cochrane Review of 6 RCTs comparing PBD to direct surgery. The utility values for the model were taken from patient responses to the EQ-5D questionnaire, scored using the UK population weightings and completed by people with hepatic colorectal metastases. As this was not the patient group considered by the model the committee found it difficult to say whether quality of life would be similar between these groups. The study did report that the trends closely matched those reported in disease specific quality of life measures for pancreatic cancer. However, the results of the model were not sensitive to this input and it was unlikely to change the preferred option. Costs inputs for the model were all sourced from NHS reference costs.

The model concluded that sending patients directly to surgery led to a cost saving of £2,552 per patient. It led to a small increase in health of 0.006 QALYS. This result was robust to all sensitivity analyses performed. Probabilistic sensitivity analysis showing a strategy of PBD prior to surgery being the preferred option in less than 10% of iterations when a £20,000 per QALY willingness to pay is assumed.

The committee were broadly in agreement with the inputs and findings of the economic analysis although raised concerns that issues of capacity (for example, operating theatres and surgeons being available when needed) had not been considered by the model. The committee agreed that this could be dealt with through reorganisation of surgical set-ups with no, or very limited, additional costs as there would be no increase in total number of operations. Whilst this reorganisation could be done in multiple ways, where costs were incurred they were likely to be in employing a coordinator for facilitating immediate access. Even with this wage cost, including on-costs, the conclusions of the economic evaluation were unlikely to be changed. The committee were, therefore, able to make a strong recommendation for sending patients with resectable pancreatic cancer and obstructive jaundice directly to surgery.

Arguedas (2002) compared plastic stenting to metal stenting in patients with pancreatic cancer and obstructive jaundice presenting for palliative biliary stenting. The study took a US Societal Perspective and was deemed to have very serious methodological limitations. The study estimated that initial stenting with metal stents would lead to a cost saving of US$433 and a health increase of 0.033 QALYs. This result was robust to all parameters apart from length of survival. Given the age of the study, the US societal perspective, methodological issues, and that a contemporary bespoke economic model had been built to answer an almost identical decision problem from a UK NHS and PSS perspective, the committee did not use this study in informing their recommendations.

The bespoke economic model considered 3 possible strategies for biliary stenting in patients with unresectable or metastatic pancreatic cancer and obstructive jaundice. The model compared a strategy of initial stenting with a plastic stent followed by stenting with a self-expanding metal stent (SEMS) upon dysfunction and initial stenting with SEMS replaced/repositioned upon dysfunction with a base case strategy of initial plastic stenting replaced with plastic stents upon dysfunction. The study took a UK NHS and PSS perspective and considered a 2 year time horizon which was adequate to represent the lifetime of over 99% of the patient group.

Clinical inputs and baseline values were largely taken from the accompanying clinical evidence review and cost inputs were exclusively taken from NHS reference costs. The utility values in the base-case were taken from a patient group, identical to that considered in the economic model, using the EQ-5D questionnaire and scored using Dutch population values. The questionnaire was completed alongside an RCT identified in the clinical evidence review. The hazard ratio for overall survival between plastic and metal stents in the clinical evidence review was equal to 1 (no difference) and there was no difference in deterioration in EQ-5D reported in the identified study. Therefore, the base case for the model assumed no difference on these parameters between the three strategies and the base case analysis became a de-facto cost minimisation. The committee, however, considered, based on their clinical experience, that quality of life, through reduced adverse events and lower need for repeat surgery would improve and therefore a secondary analysis was performed using the values reported in the same study but using the visual analogue scale. This measure reported that quality of life deteriorated at a lower rate with SEMS compared to plastic stents although this was not statistically significant.

In the base case a strategy of initial metal stenting followed by subsequent metal stenting was the least costly with a saving of over £1,500 per patient. When QoL was also considered it led to a small increase in QoL of 0.024 QALYs per patient. This result was only sensitive to overall survival with plastic stenting followed by plastic stenting becoming the least costly when survival was less than 24 days. The robustness of the result is supported by the probabilistic sensitivity analysis. The initial stenting with SEMS strategy is cost saving compared to plastic stenting followed by plastic stenting in 98% of iterations. The conclusions were broadly identical to that of Arguedas (2002), with metal stents being cost saving and results only being sensitive to survival. Although, given the differences between the studies described above, there is little validity to any comparison.

The committee, therefore, made a strong recommendation supporting the use of SEMS in this patient group. The economic model attempted to look at the type of SEMS used (covered, partially covered, uncovered) but results disaggregated by SEMS type were reported inconsistently and it was difficult to consider them as separate analyses. The 3 types of stents though have almost identical costs and the decision of which type to use was based on clinical and not economic considerations.

9.1.9. References

Artifon ELA, Aparicio D, Paione JB et al. (2012) Biliary Drainage in Patients With Unresectable, Malignant Obstruction Where ERCP Fails Endoscopic Ultrasonography-Guided Choledochoduodenostomy Versus Percutaneous Drainage. Journal of Clinical Gastroenterology 46: 768–774 [PubMed: 22810111]
Artifon ELA, Loureiro JF, Baron TH et al. (2015) Surgery or EUS-guided choledochoduodenostomy for malignant distal biliary obstruction after ERCP failure. Endoscopic Ultrasound 4: 235–43 [PMC free article: PMC4568637] [PubMed: 26374583]
Artifon EL, Sakai P, Cunha JE et al. (2006) Surgery or endoscopy for palliation of biliary obstruction due to metastatic pancreatic cancer. The American Journal of Gastroenterology 101: 2031–7 [PubMed: 16968509]
Artifon EL, Sakai P, Ishioka S et al. (2008) Endoscopic sphincterotomy before deployment of covered metal stent is associated with greater complication rate: a prospective randomized control trial. Journal of Clinical Gastroenterology 42: 815–9 [PubMed: 18285718]
Eshuis WJ, van der Gaag NA, Rauws EA et al. (2010) Therapeutic delay and survival after surgery for cancer of the pancreatic head with or without preoperative biliary drainage. Annals of Surgery 252(5): 840–849 [PubMed: 21037440]
Gardner TB, Spangler CC, Byanova KL et al. (2016) Cost-effectiveness and clinical efficacy of biliary stents in patients undergoing neoadjuvant therapy for pancreatic adenocarcinoma in a randomized controlled trial. Gastrointestinal Endoscopy 84(23): 460–466 [PMC free article: PMC4988865] [PubMed: 26972022]
Giorgio PD and Luca LD (2004) Comparison of treatment outcomes between biliary plastic stent placements with and without endoscopic sphincterotomy for inoperable malignant common bile duct obstruction. World Journal of Gastroenterology 10(8): 1212–4 [PMC free article: PMC4656363] [PubMed: 15069728]
Hayashi T, Kawakami H, Osanai M et al. (2015) No benefit of endoscopic sphincterotomy before biliary placement of self-expandable metal stents for unresectable pancreatic cancer. Clinical Gastroenterology & Hepatology 13: 1151–8.e2 [PubMed: 25632802]
Isayama H, Yasuda I, Ryozawa S et al. (2011) Results of a Japanese multicenter, randomized trial of endoscopic stenting for non-resectable pancreatic head cancer (JM-test): Covered Wallstent versus DoubleLayer stent. Digestive Endoscopy 23, 310–5 [PubMed: 21951091]
Kaassis M, Boyer J, Dumas R et al. (2003) Plastic or metal stents for malignant stricture of the common bile duct? Results of a randomized prospective study. Gastrointestinal Endoscopy 57: 178–182 [PubMed: 12556780]
Kitano M, Yamashita Y, Tanaka K et al. (2013) Covered self-expandable metal stents with an anti-migration system improve patency duration without increased complications compared to uncovered stents for distal biliary obstruction caused by pancreatic carcinoma: a randomized multicenter trial. The American Journal of Gastroenterology 108(11): 1713–1722 [PubMed: 24042190]
Krokidis M, Fanelli F, Orgera G et al. (2011) Percutaneous palliation of pancreatic head cancer: randomized comparison of ePTFE/FEP–covered versus uncovered nitinol biliary stents. Cardiovascular and Interventional Radiology 34(2): 352–361 [PubMed: 20467870]
Kullman E, Frozanpor F, Söderlund C et al. (2010) Covered versus uncovered self-expandable nitinol stents in the palliative treatment of malignant distal biliary obstruction: results from a randomized, multicenter study. Gastrointestinal Endoscopy 72(5): 915–923 [PubMed: 21034892]
Moses PL, Alnaamani KM, Barkun AN et al. (2013) Randomized trial in malignant biliary obstruction: plastic vs partially covered metal stents. World Journal of Gastroenterology 19: 8638–46 [PMC free article: PMC3870509] [PubMed: 24379581]
Schmidt A, Riecken B, Rische S et al. (2015) Wing-shaped plastic stents vsself-expandable metal stents for palliative drainage of malignant distal biliary obstruction: A randomized multicenter study. Endoscopy 47(5): 430–436 [PubMed: 25590188]
Söderlund C and Linder S (2006) Covered metal versus plastic stents for malignant common bile duct stenosis: a prospective, randomized, controlled trial. Gastrointestinal Endoscopy 63: 986–95 [PubMed: 16733114]
Song TJ, Lee SS, Yun SC et al. (2011) Paclitaxel-eluting covered metal stents versus covered metal stents for distal malignant biliary obstruction: a prospective comparative pilot study. Gastrointestinal Endoscopy 73: 727–733 [PubMed: 21288514]
Telford JJ, Carr-Locke DL, Baron TH et al. (2010) A randomized trial comparing uncovered and partially covered self-expandable metal stents in the palliation of distal malignant biliary obstruction. Gastrointestinal Endoscopy 72(5): 907–914 [PubMed: 21034891]
Travis S and Nicholson T (1997) Palliation of unresectable pancreatic malignant biliary obstruction: Results of a randomized trial comparing percutaneously placed metal and plastic endoprostheses. Journal of Interventional Radiology 12: 17–21
Ung KA, Stotzer PO, Nilsson Å et al. (2013) Covered and uncovered self-expandable metallic Hanarostents are equally efficacious in the drainage of extrahepatic malignant strictures. Results of a double-blind randomized study. Scandinavian Journal of Gastroenterology, 48(4): 459–465. [PubMed: 23373541]
van der Gaag NA, Rauws EA, van Eijck CH et al. (2010) Preoperative biliary drainage for cancer of the head of the pancreas. New England Journal of Medicine 362: 129–37 [PubMed: 20071702]

9.2. Duodenal obstruction

Review question: What is the optimal treatment of duodenal obstruction?

9.2.1. Introduction

Tumour invasion into the duodenum can result in obstruction to the flow of ingested food and secretions from the stomach into the duodenum. Gastric outflow obstruction results in recurrent large volume vomiting, fullness, dehydration and malnutrition. Duodenal obstruction is usually associated with advanced and unresectable pancreatic tumours and occurs in up to 20% of patients with pancreatic cancer.

When duodenal obstruction occurs in association with an operable tumour the definitive management of the obstruction will occur with resection of the tumour. For the majority of patients with duodenal obstruction who have inoperable disease, the options are between palliative surgery (gastrojejunostomy) to bypass the obstruction or the endoscopic placement of a self-expanding metal stent (SEMS). Placement of a SEMS may be tolerated better by frail individuals and are thought to be associated with faster recovery and symptom improvement, however the improvement may not be as marked or as durable as that achieved with surgery.

A proportion of individuals who undergo surgery with curative intent will be found to have inoperable disease at the time of surgery and will therefore not have a resection. Some of these individuals will subsequently develop duodenal obstruction due to disease progression. Prophylactic gastrojejunostomy performed during the operation when curative surgery is deemed not to be feasible may prevent the later development of duodenal obstruction.

Guidance is needed on the optimal treatment of duodenal obstruction in people with pancreatic cancer.

9.2.1.1. Review protocol summary

The review protocol summary used for this question can be found in Table 111. Full details of the review protocol can be found in Appendix C.

Table 111

Clinical review protocol summary for the review of optimal treatment of duodenal obstruction.

9.2.2. Description of Clinical Evidence

Six studies –2 RCTs (Lillemoe et al. 1999; Van Heek et al. 2004) from a recent Cochrane review (Gurusamy et al. 2013), and an additional 4 RCTs (Okuwaki et al. 2016; Jeurnink et al. 2010; Mehta et al. 2006; Shyr et al. 1997) were included in the evidence review. All the studies were in adults. A summary of the included studies is presented in Table 112.

Two RCTs (n=157) from a Cochrane review (Gurusamy et al. 2013) that compared prophylactic gastrojejunostomy (GJJ) combined with hepaticojejunostomy with hepaticojejunostomy only in patients with unresectable pancreatic cancer were included (Lillemoe et al. 1999; Van Heek et al. 2004).

Two RCTs (n=66) were found that compared laparoscopic GJJ with duodenal stenting as a means of palliating malignant gastric outflow obstruction in patients with pancreatic cancer (Jeurnink et al. 2010; Mehta et al. 2006). The sample in Metha et al. (2006) had only 56% pancreatic cancer patients and was thus downgraded for indirectness.

One RCT (n=45) was found that compared three types of GJJ for duodenal obstruction in patients with unresectable periampullary cancer (Shyr et al, 1997). Although the sample had only 51% pancreatic cancer patients, the study was included and downgraded for indirectness. The three types of GJJ differed according to the site of jejunum for the GJJ and the partition of duodenum: Type 1 (GJJ proximal to the Jejunal limb: Ligament of Treitz), Type 2 (Pylorus) and Type 3 (GJJ proximal to Roux-limb Jejunum).

One RCT (n=34) was found that compared two types of duodenal stents (WallFlex™ duodenal stent [W-group] and Niti-S™ pyloric/duodenal D-type stent) with different axial forces for alleviating duodenal obstruction in patients with pancreatobiliary cancer (Okuwaki et al. 2016). The sample in this study was 74% pancreatic cancer patients.

9.2.3. Summary of included studies

A summary of the studies that were included in this review are presented in Table 112.

Table 112

Summary of included studies.

9.2.4. Clinical evidence profile

The clinical evidence profiles for this review question are presented in Table 113 to Table 118.

Table 113

Summary clinical evidence profile for prophylactic gastrojejunostomy (GJJ) and hepaticojejunostomy versus hepaticojejunostomy only in adults with unresectable pancreatic cancer and gastric outlet obstruction.

Table 114

Summary clinical evidence profile for GJJ versus duodenal stent placement in adults with pancreatic cancer and gastric outlet obstruction.

Table 115

Summary clinical evidence profile for Type I GJJ (proximal to the Jejunal limb: Ligament of Treitz) versus Type II GJJ (Pylorus) in adults with pancreatic cancer and gastric outlet obstruction.

Table 116

Summary clinical evidence profile for Type I GJJ (proximal to the Jejunal limb: Ligament of Treitz) versus Type III GJJ (proximal to Roux-limb Jejunum) in adults with pancreatic cancer and gastric outlet obstruction.

Table 117

Summary clinical evidence profile for Type II GJJ (Pylorus) versus Type III GJJ (proximal to Roux-limb Jejunum) in adults with pancreatic cancer and gastric outlet obstruction.

Table 118

Summary clinical evidence profile for duodenal stent-1 versus duodenal stent-2 in adults with pancreatic cancer and duodenal obstruction.

9.2.5. Economic evidence

A literature review of published cost effectiveness analyses did not identify any relevant studies for this topic. Although there were potential implications for resource use associated with making recommendations in this area, other topics in the guideline were agreed as a higher economic priority. Consequently, bespoke economic modelling was not done for this topic.

9.2.6. Evidence Statements

9.2.6.1. Prophylactic GJJ and hepaticojejunostomy versus hepaticojejunostomy only

Relief of obstruction

Low quality evidence from 2 RCTs (n=152) showed that there is a clinically important difference favouring prophylactic gastrojejunostomy combined with hepaticojejunostomy on relief of obstruction compared to hepaticojejunostomy only in adults with unresectable pancreatic cancer and gastric outlet obstruction: RR 0.11 (95% CI 0.03-0.4).

Change in symptoms

No evidence was identified to inform this outcome.

Nutritional status

No evidence was identified to inform this outcome.

Adverse events

Low quality evidence from 2 RCTs (n=152) showed no clinically important difference between prophylactic gastrojejunostomy combined with hepaticojejunostomy and hepaticojejunostomy only on peri-operative mortality (RR 2.43 [95% CI 0.1-57.57]), bile leak (RR 1.23 [95% CI 0.28-5.34]), gastroenteral leak (RR 0.81 [95% CI 0.05-12.33]), delayed gastric emptying (RR 2.71 [95% CI 0.52-14.08]), wound infection (RR 3.09 [95% CI 0.52-18.36]), and chest complications (RR 0.44 [95% CI 0.08-2.35]) in adults with unresectable pancreatic cancer and gastric outlet obstruction.

Very low quality evidence from 1 RCT (n=87) showed no clinically important difference between prophylactic gastrojejunostomy combined with hepaticojejunostomy and hepaticojejunostomy only on cholangitis in adults with unresectable pancreatic cancer and gastric outlet obstruction: RR 1.95 (95% CI 0.38-10.12).

Very low quality evidence from 1 RCT (n=65) showed no clinically important difference between prophylactic gastrojejunostomy combined with hepaticojejunostomy and hepaticojejunostomy only on cardiac complications in adults with unresectable pancreatic cancer and gastric outlet obstruction: RR 1.61 (95% CI 0.32-8.19).

Overall survival

Low quality evidence from 2 RCTs (n=152) showed no clinically important difference between prophylactic gastrojejunostomy combined with hepaticojejunostomy and hepaticojejunostomy only on overall survival in adults with unresectable pancreatic cancer and gastric outlet obstruction: HR 1.02 (95% CI 0.84-1.25).

Health-related quality of life

Low quality evidence from 1 RCT (n=65) reported no statistically significant difference between prophylactic gastrojejunostomy combined with hepaticojejunostomy and hepaticojejunostomy only on EORTC quality of life at any time point in adults with unresectable pancreatic cancer and gastric outlet obstruction (no data reported).

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.2.6.2. GJJ versus duodenal stent placement

Relief of obstruction

Low quality evidence from 1 RCT (n=39) reported a statistically significant difference favouring duodenal stent placement on the number of days with a Gastric Outlet Obstruction Scoring System score of 2 or more compared to gastrojejunostomy (median 72 days vs 50 days, p=0.05) in adults with pancreatic cancer and gastric outlet obstruction.

Very low quality evidence from 1 RCT (n=39) showed no clinically important difference between gastrojejunostomy and duodenal stent placement on either persistent obstructive symptoms (RR 1.17 [95% CI 0.27-1.72]) or recurrent obstructive symptoms (RR 0.23 [95% CI 0.03-1.82]) in adults with pancreatic cancer and gastric outlet obstruction.

Change in symptoms

No evidence was identified to inform this outcome.

Nutritional status

Low quality evidence from 1 RCT (n=39) reported a statistically significant difference favouring duodenal stent placement on the number of days to restore the ability to eat compared to gastrojejunostomy (median 8 days vs 5 days, p<0.01) in adults with pancreatic cancer and gastric outlet obstruction.

Adverse events

Very low quality evidence from 1 RCT (n=39) showed no clinically important difference between gastrojejunostomy and duodenal stent placement on either major complications (RR 0.13 [95% CI 0.01-2.24]) or minor complications (RR 1.46 [95% CI 0.46-4.63]) in adults with pancreatic cancer and gastric outlet obstruction.

Overall survival

Very low quality evidence from 1 RCT (n=27) showed no clinically important difference between gastrojejunostomy and duodenal stent placement on overall survival in adults with pancreatic cancer and gastric outlet obstruction: HR 0.81 (95% CI 0.27-2.44).

Health-related quality of life

Very low quality evidence from 1 RCT (n=25) showed no clinically important difference between gastrojejunostomy and duodenal stent placement on either the SF-36 physical health (MD −7.9 [95% CI −22.74 to 6.94]) or mental health (MD 0.7 [95% CI −18.29 to 19.69]) subscales in adults with pancreatic cancer and gastric outlet obstruction.

Patient experience

No evidence was identified to inform this outcome.

PROMS

Very low quality evidence from 1 RCT (n=25) showed no clinically important difference between gastrojejunostomy and duodenal stent placement on self-reported pain visual analogue scale in adults with pancreatic cancer and gastric outlet obstruction: MD 2.0 (95% CI −0.36 to 4.36).

9.2.6.3. Types of gastrojejunostomy

9.2.6.3.1. Type I GJJ (proximal to the Jejunal limb: Ligament of Treitz) versus Type II GJJ (Pylorus)

Relief of obstruction

No evidence was identified to inform this outcome.

Change in symptoms

Very low quality evidence from 1 RCT (n=30) showed that there may be a clinically important difference favouring Type I gastrojejunostomy (proximal to the Jejunal limb: Ligament of Treitz) on change in clinical symptoms as assessed by the Gastric Outlet Obstruction Scoring System compared to Type II gastrojejunostomy (Pylorus) in adults with pancreatic cancer and gastric outlet obstruction, although there is some uncertainty: RR 3.5 (95% CI 0.86-14.18).

Very low quality evidence showed no clinically important difference between Type I gastrojejunostomy (proximal to the Jejunal limb: Ligament of Treitz) and Type II gastrojejunostomy (Pylorus) on change in symptoms of anorexia (RR 3.0 [95% CI 0.13-68.26]), epigastric fullness (RR 2.0 [95% CI 0.2-19.78]), nausea (RR 3.0 [95% CI 0.13-68.26]) and vomiting (RR 7.0 [95% CI 0.39-124.83]) as assessed by the Gastric Outlet Obstruction Scoring System in adults with pancreatic cancer and gastric outlet obstruction.

Nutritional status

Very low quality evidence from 1 RCT (n=30) showed no clinically important difference between Type I gastrojejunostomy (proximal to the Jejunal limb: Ligament of Treitz) and Type II gastrojejunostomy (Pylorus) on either minutes to gastric emptying (MD 40.8 [95% CI - 67.85 to 149.45]) or the number of patients with delayed gastric emptying (RR 3.0 [95% CI 0.35-25.68]) in adults with pancreatic cancer and gastric outlet obstruction.

Adverse events

No evidence was identified to inform this outcome.

Overall survival

No evidence was identified to inform this outcome.

Health-related quality of life

No evidence was identified to inform this outcome.

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.2.6.3.2. Type I GJJ (proximal to the Jejunal limb: Ligament of Treitz) versus Type III GJJ (proximal to Roux-limb Jejunum)

Relief of obstruction

No evidence was identified to inform this outcome.

Change in symptoms

Very low quality evidence from 1 RCT (n=30) showed that there may be a clinically important difference favouring Type I gastrojejunostomy (proximal to the Jejunal limb: Ligament of Treitz) on change in clinical symptoms as assessed by the Gastric Outlet Obstruction Scoring System compared to Type III gastrojejunostomy (proximal to Roux-limb Jejunum) in adults with pancreatic cancer and gastric outlet obstruction, although there is some uncertainty: RR 3.5 (95% CI 0.86-14.18).

Very low quality evidence showed no clinically important difference between Type I gastrojejunostomy (proximal to the Jejunal limb: Ligament of Treitz) and Type III gastrojejunostomy (proximal to Roux-limb Jejunum) on change in symptoms of anorexia (RR 1.0 [95% CI 0.07-14.55]), epigastric fullness (RR 2.0 [95% CI 0.2-19.78]), nausea (RR 3.0 [95% CI 0.13-68.26]) and vomiting (RR 7.0 [95% CI 0.39-124.83]) as assessed by the Gastric Outlet Obstruction Scoring System in adults with pancreatic cancer and gastric outlet obstruction.

Nutritional status

Very low quality evidence from 1 RCT (n=30) showed no clinically important difference between Type I gastrojejunostomy (proximal to the Jejunal limb: Ligament of Treitz) and Type III gastrojejunostomy (proximal to Roux-limb Jejunum) on either minutes to gastric emptying (MD −86.4 [95% CI −192.05 to 19.25]) or the number of patients with delayed gastric emptying (RR 3.0 [95% CI 0.35-25.68]) in adults with pancreatic cancer and gastric outlet obstruction.

Adverse events

No evidence was identified to inform this outcome.

Overall survival

No evidence was identified to inform this outcome.

Health-related quality of life

No evidence was identified to inform this outcome.

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.2.6.3.3. Type II GJJ (Pylorus) versus Type III GJJ (proximal to Roux-limb Jejunum)

Relief of obstruction

No evidence was identified to inform this outcome.

Change in symptoms

Very low quality evidence from 1 RCT (n=30) showed no clinically important difference between Type II gastrojejunostomy (Pylorus) and Type III gastrojejunostomy (proximal to Roux-limb Jejunum) on change in clinical symptoms as assessed by the Gastric Outlet Obstruction Scoring System in adults with pancreatic cancer and gastric outlet obstruction: RR 0.5 (95% CI 0.05-4.94).

Very low quality evidence showed no clinically important difference between Type II gastrojejunostomy (Pylorus) and Type III gastrojejunostomy (proximal to Roux-limb Jejunum) on change in symptoms of epigastric fullness (RR 1.0 [95% CI 0.07-14.55]), and nausea (RR 0.33 [95% CI 0.01-7.58]) as assessed by the Gastric Outlet Obstruction Scoring System in adults with pancreatic cancer and gastric outlet obstruction. (There were also no events on symptoms of anorexia and vomiting.)

Nutritional status

Very low quality evidence from 1 RCT (n=30) showed that there is a clinically important difference favouring Type II gastrojejunostomy (Pylorus) on minutes to gastric emptying compared to Type III gastrojejunostomy (proximal to Roux-limb Jejunum) in adults with pancreatic cancer and gastric outlet obstruction: MD −127.2 (95% CI −232.85 to −21.55).

Very low quality evidence showed no clinically important difference between Type II gastrojejunostomy (Pylorus) and Type III gastrojejunostomy (proximal to Roux-limb Jejunum) on the number of patients with delayed gastric emptying in adults with pancreatic cancer and gastric outlet obstruction: RR 1.0 (95% CI 0.07-14.55).

Adverse events

No evidence was identified to inform this outcome.

Overall survival

No evidence was identified to inform this outcome.

Health-related quality of life

No evidence was identified to inform this outcome.

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.2.6.4. Duodenal stent-1 versus duodenal stent-2

Relief of obstruction

Very low quality evidence from 1 RCT (n=31) showed no clinically important difference between WallFlex™ duodenal stents and Niti-S™ pyloric/duodenal D-type stents on the number of people who had recurrence of obstruction as assessed by the Gastric Outlet Obstruction Scoring System at 2 weeks in adults with pancreatic cancer and duodenal obstruction: RR 1.21 (95% CI 0.37-4.0).

Change in symptoms

Low quality evidence from 1 RCT (n=31) showed no clinically important difference between WallFlex™ duodenal stents and Niti-S™ pyloric/duodenal D-type stents on mean change on the Gastric Outlet Obstruction Scoring System at 2 weeks in adults with pancreatic cancer and duodenal obstruction: SMD 0.37 (95% CI −0.34 to 1.09).

Nutritional status

Moderate quality evidence from 1 RCT (n=31) showed no clinically important difference between WallFlex™ duodenal stents and Niti-S™ pyloric/duodenal D-type stents on mean change on BMI at 4 weeks, in adults with pancreatic cancer and duodenal obstruction: MD −0.3 (95% CI −1.22 to 0.62).

Adverse events

Low to very low quality evidence from 1 RCT (n=31) showed no clinically important difference between WallFlex™ duodenal stents and Niti-S™ pyloric/duodenal D-type stents on either mean change in Nausea and Vomiting Scoring System score (SMD 0.28 [95% CI −0.43 to 0.99]) or the number of procedure-related adverse events (RR 1.21 [95% CI 0.37-4.0]) in adults with pancreatic cancer and duodenal obstruction.

Overall survival

Low quality evidence from 1 RCT (n=31) showed no clinically important difference between WallFlex™ duodenal stents and Niti-S™ pyloric/duodenal D-type stents on overall survival in adults with pancreatic cancer and duodenal obstruction: HR 0.52 (95% CI 0.26-1.08).

Health-related quality of life

Low quality evidence from 1 RCT (n=31) showed no clinically important difference at 2 weeks between WallFlex™ duodenal stents and Niti-S™ pyloric/duodenal D-type stents on either mean change in Karnofsky Performance Score (MD 5.2 [95% Ci −5.47 to 15.87]) or mean change in Performance Score (MD −0.1 [95% CI −0.69 to 0.49]) in adults with pancreatic cancer and duodenal obstruction

Patient experience

No evidence was identified to inform this outcome.

PROMS

No evidence was identified to inform this outcome.

9.2.7. Recommendations

38.: During attempted resection for head of pancreas cancer, consider prophylactic gastrojejunostomy if the cancer is found to be unresectable.
39.: If possible, relieve symptomatic duodenal obstruction caused by unresectable pancreatic cancer.
40.: When deciding between gastrojejunostomy and duodenal stent placement, consider gastrojejunostomy for people with a more favourable prognosis.

9.2.8. Evidence to recommendations

9.2.8.1. Relative value placed on the outcomes considered

Relief of obstruction, change in symptoms, nutritional status, adverse events, overall survival, health-related quality of life, patient reported outcome measures and patient experience were considered to be the critical outcomes for this question.

Adverse events, overall survival and health-related quality of life were reported for all comparisons of interest except for gastrojejunostomy with duodenal partition versus other gastrojejunostomy types. Change in symptoms and nutritional status were reported for all comparisons of interest except prophylactic gastrojejunostomy versus no prophylactic gastrojejunostomy.

Relief of obstruction was only reported for duodenal stent placement and the comparison of prophylactic gastrojejunostomy with no prophylactic gastrojejunostomy. Patient reported outcome measures was only reported for the comparison of gastrojejunostomy with duodenal stent placement. Patient experience was not reported for any of the comparisons of interest.

The committee noted that the data on patient reported outcome measures looked at a self-reported pain score. They agreed that it was not possible to determine whether the pain was generated by the procedure or by the tumour itself, and consequently did not use this outcome when making recommendations.

9.2.8.2. Quality of evidence

The quality of the evidence was assessed by GRADE and the Cochrane risk of bias checklist. The evidence was either very low or low quality for all outcomes across all comparisons of interest.

The committee noted that the study looking at gastrojejunostomy with duodenal partition versus other gastrojejunostomy types was conducted in China. They considered that it had limited relevance to the UK setting, particularly because it used a type of gastrojejunostomy which is not done in the UK. The committee, therefore, agreed not to use the results of this study when making their recommendations.

The committee agreed that the study comparing different types of stent for relieving duodenal obstruction was not useful when making recommendations. This study was conducted in Japan and so had limited relevance to the UK healthcare setting. In addition, the aim of the study was to compare the effectiveness of two different types of stent. Given that there are several other types of stent available, which the study did not investigate, the committee agreed it would be difficult to draw robust conclusions as to which specific stent should be used.

The committee noted that the studies comparing gastrojejunostomy with duodenal stent placement had excluded people who were unfit for surgery. This is not representative of the group of people who get duodenal obstruction.

No evidence was found on the effectiveness of venting gastrostomy or resectional surgery for treating duodenal obstruction. Consequently, the committee did not make any recommendations for clinical practice for these interventions. The committee agreed that conducting further research in this area would not be practical because it would not be feasible to randomise people to these interventions and, therefore, did not make any research recommendations either.

The committee were not able to make any recommendations for people with resectable pancreatic cancer who have duodenal obstruction as there was no evidence available on this population.

9.2.8.3. Consideration of clinical benefits and harms

Due to the low quality evidence the committee was not able to make any strong recommendations.

The committee agreed, based on their knowledge and experience, that it is very important to relieve duodenal obstruction in people with unresectable pancreatic cancer. However they also recognised that people with unresectable pancreatic cancer may have more extensive disease, or may be too unwell for intervention, and this may make it difficult to relieve the obstruction. They, therefore, agreed to recommend that the obstruction should be relieved if possible.

The committee noted that the available evidence was of low quality and only covered some of the interventions of interest which made it difficult to specify the most effective method to relieve the obstruction. The evidence indicated a trend that stent placement was more effective in the short term whilst gastrojejunostomy was more effective in the longer term. This accorded with the committee’s knowledge and experience that gastrojejunostomy is normally done only in people likely to have longer overall survival because of the morbidity associated with surgery. They, therefore, agreed to recommend both duodenal stents and gastrojejunostomy as options for people with duodenal obstruction with gastrojejunostomy being considered for people with a more favourable prognosis.

Based on the evidence, the committee noted that prophylactic gastrojejunostomy was associated with less gastric outlet obstruction and no difference in the proportion of people developing adverse events. The committee noted, based on their knowledge and experience, that duodenal obstruction is a recognised complication of pancreatic cancer. It is associated with significant co-morbidities and is known to have a detrimental effect on quality of life. They, therefore, agreed that, in people with large tumours who were felt to be at risk of duodenal obstruction who were otherwise fit and had a relatively good prognosis, the prophylactic use of gastrojejunostomy could be considered.

The committee agreed that the potential benefits of the recommendations made would be symptom relief by an appropriate technique and improved quality of life. The potential harms of the recommendations made would be potential complications of surgery or stent insertion. The committee agreed that the potential benefits for the person would outweigh the risk of harm.

9.2.8.4. Consideration of economic benefits and harms

The committee noted that no relevant published economic evaluations had been identified and no additional economic analysis had been undertaken in this area.

The committee noted that current practice is for people with duodenal obstruction to receive a stent or a gastrojejunostomy. Both of these interventions are still options in the recommendations. The committee considered that the costs of stent placement and gastrojejunostomy are broadly similar. The stent insertion procedure is more expensive than a gastrojejunostomy but the length of hospital stay is normally shorter for stent placement and, therefore, associated with less cost than the hospital stay for a gastrojejunostomy. Therefore, whilst it is possible that the balance between stent placement and gastrojejunostomy may alter, the committee agreed this was unlikely to have a significant resource impact. The committee also noted that the recommendation for prophylactic gastrojejunostomy was unlikely to cause significant resource impact because the procedure will be done at the same time as the resectional surgery.

9.2.9. References

Gurusamy KS, Kumar S, Davidson BR (2013) Prophylactic gastrojejunostomy for unresectable periampullary carcinoma. The Cochrane Database of Systematic Reviews 2 [PMC free article: PMC7173743] [PubMed: 23450583]
Jeurnink SM, Polinder S, Steyerberg EW et al. (2010) Cost comparison of gastrojejunostomy versus duodenal stent placement for malignant gastric outlet obstruction. Journal of Gastroenterology 45(5): 537–43 [PubMed: 20033227]
Mehta S, Hindmarsh A, Cheong E, et al. (2006) Prospective randomized trial of laparoscopic gastrojejunostomy versus duodenal stenting for malignant gastric outflow obstruction. Surgical Endoscopy and Other Interventional Techniques 20(2): 239–42 [PubMed: 16362479]
Okuwaki K, Kida M, Yamauchi H et al. (2016) Randomized controlled exploratory study comparing the usefulness of two types of metallic stents with different axial forces for the management of duodenal obstruction caused by pancreatobiliary cancer. Journal of Hepatobiliary-pancreatic Sciences 23(5): 289–97 [PubMed: 26946214]
Shyr YM, Su CH, King KL et al. (1997) Randomized trial of three types of gastrojejunostomy in unresectable periampullary cancer. Surgery 121(5): 506–12 [PubMed: 9142148]

9.2.9.1. Studies included in Gurusamy et al., 2013 (n=1)

Lillemoe KD, Cameron JL, Hardacre JM et al. (1999) Is prophylactic gastrojejunostomy indicated for unresectable periampullary cancer?: a prospective randomized trial. Annals of Surgery 230(3): 322 [PMC free article: PMC1420877] [PubMed: 10493479]

Bookshelf ID: NBK536655

Contents

< Prev Next >

PubReader
Print View
Cite this Page
National Guideline Alliance (UK). Pancreatic cancer in adults: diagnosis and management. London: National Institute for Health and Care Excellence (NICE); 2018 Feb. (NICE Guideline, No. 85.) 9, Interventions to relieve biliary and duodenal obstruction.
PDF version of this title (16M)

Other titles in this collection

National Institute for Health and Care Excellence: Guidelines

Related information

PMC
PubMed Central citations
PubMed
Links to PubMed

Recent Activity

Clear Turn Off Turn On

Interventions to relieve biliary and duodenal obstruction - Pancreatic cancer in...
Interventions to relieve biliary and duodenal obstruction - Pancreatic cancer in adults: diagnosis and management
NCTC_3000
NCTC_3000
NCTC_3000
BioProject
BioProject Links for Nucleotide (Select 1539380133) (1)
BioProject
Profile neighbors for GEO Profiles (Select 83810912) (200)
GEO Profiles
PREDICTED: Homo sapiens solute carrier family 6 member 2 (SLC6A2), transcript va...
PREDICTED: Homo sapiens solute carrier family 6 member 2 (SLC6A2), transcript variant X5, mRNA
gi|2217307152|ref|XM_047434512.1|
Nucleotide

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on