Interventions for Dysphagia in Oesophageal Cancer
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
To compare and summarise the efficacy of the different interventions used in palliation of dysphagia in patients with primary oesophageal cancer.
Background
The incidence of adenocarcinoma of the oesophagus and the gastro‐oesophageal junction is increasing (Cheng 1992), (Devesa 1998). Most oesophageal cancers are diagnosed at an advanced stage and in patients with co‐morbidity. This precludes curative treatment and makes palliative treatment a more realistic option in most of these patients (Weigel 2002). Dysphagia is the predominant symptom in more than 70% of patients with oesophageal cancer (Brierley 1998). Several management options have been developed in recent years to palliate malignant dysphagia. These include mechanical measures like endoluminal stenting or surgery, Anti neoplastic methods like external beam radiation, brachytherapy, chemotherapy, chemoradiotherapy, laser treatment, photodynamic therapy or ablation using injection of alcohol or chemotherapeutic agents. The optimum management of dysphagia caused by advanced primary oesophageal cancer is not established although continued progress is being made in recent years to achieve this goal (Bown 1991).
Radiotherapy has been shown to improve dysphagia in 50‐85% of patients with advanced primary oesophageal cancer. However, the response rate is slower and inferior compared to combined chemoradiotherapy or stent insertion (Albertsson 1989), (Herskovic 1992), (Cwikiel 1996). With the advances in therapeutic endoscopic techniques, interest has grown in the use of Self Expanding Metal Stents (SEMS), Thermal Laser therapy, Photodynamic therapy (PDT) and Argon Plasma Coagulation (APC). Randomised trials have shown metallic stents to be effective and safe with lesser procedural mortality and morbidity compared to plastic tubes (Roseveare 1998), (Siersema 1998) (De Palma 1996). However, the use of SEMS is not without problems. Stent migration, incomplete expansion of the stent, tumour ingrowth and overgrowth may require further intervention for recurrent dysphagia. Insertion of stents beyond the Gastro‐oesophageal junction has been observed to result in acid reflux (Adam 1997), (Sabharwal 2003), (Canto 2002).
Studies have shown laser treatment improves dysphagia in 80‐96% of patients. However, perforation has been reported to occur in up to 5% of patients (Carter 1992), (Maciel 1996). The disadvantages of laser treatment include the requirement of repeated sessions and its reduced efficacy in long tumours. Randomised studies comparing PDT with laser treatment have shown better overall palliation and lesser perforations with the former although photosensitivity has been reported in a significant proportion of patients (Heier 1995), (Lightdale 1995). A recent randomized study comparing metallic stents with laser treatment found comparable relief of dysphagia improvement with both forms of treatment. The median survival was longer in the laser group (125 vs 68 days) but the cost and median length of hospital stay were higher with laser treatment.
Current guidelines indicate a complementary use of the available modalities is required in palliating dysphagia in oesophageal cancer (Allum 2002). A search for previous systematic reviews resulted in one published review (Wong 2000). The systematic review included randomized studies comparing stents, radiotherapy, brachytherapy, chemotherapy and chemoradiotherapy with other modalities. Other interventions such as laser treatment and PDT were not studied. The published review did not include any meta analysis and the authors did not mention any details about whether the studies were considered for suitability to perform meta analysis. This review will aim to address these issues and include the more recent evidence available.
Objectives
To compare and summarise the efficacy of the different interventions used in palliation of dysphagia in patients with primary oesophageal cancer.
Methods
Criteria for considering studies for this review
Types of studies
Randomised studies comparing different interventions mentioned below will be included in this review. Blinding will not be possible with many of the interventions included in the review and hence both blinded and unblinded trials will be included. Both published and unpublished studies, full articles and abstracts will be included.
Types of participants
Patients with inoperable or unresectable primary oesophageal cancer who undergo palliative treatment. Patients with primary squamous or adenocarcinoma of the oesophagus or the Gastrooesophageal junction will be included. Where available, these patient sub groups will be considered for analysis and comparison separately.
Patients with extrinsic compression of the oesophagus from other tumours and patients with recurrence of dysphagia or recurrence of tumour after previous surgery will not be considered in this review.
Types of interventions
To be included in this review, the tested intervention should fall into one of the following categories (a to j)
a)External Beam Radiotherapy
b)Intra‐luminal brachytherapy
c)Chemoradiotherapy
d)Chemotherapy
e)Plastic stent insertion
f)Self expanding metal stent insertion
g)Thermal ablative therapy: Laser therapy, Argon plasma coagulation, BICAP probe electrocoagulation
h)Photodynamic therapy
i)Chemical ablative therapy: Alcohol injection, Chemotherapeutic agent injection
j)Oesophageal bypass surgery
The comparison should include one of the interventions mentioned above or oesophageal dilatation alone. Combination of interventions is acceptable if one of the interventions is included in both the arms of the study. Studies comparing different types of self expanding metal stents will be considered in the review to evaluate and compare the outcomes among the different brands or types of stents. This will include comparisons between:
1)Covered and Uncovered stents
2)Cuffed and Uncuffed stents to prevent Gastro‐oesophageal reflux
3)Different commercially available brands of stents
Types of outcome measures
The primary outcome of interest will be improvement in dysphagia. Several dysphagia scales have been used to assess improvement in dysphagia grade, across the studies. (Mellow 1985), (O'Rourke 1988), (Bown 1987). Recent studies have used these scales with modifications. If dichotomous outcomes could be extracted from studies, we will compare 1 point and 2 point improvement in dysphagia grade for each intervention between the studies. If continuous data is available then the primary outcome will be expressed and compared using mean and standard deviation. Standardised Mean Difference will be used for studies using different scales.
Secondary outcomes will include the following:
1) overall survival
2) time period from intervention to improvement or relief of dysphagia
3) recurrence of dysphagia
4) time period from intervention to recurrence of dysphagia
5) requirement for further interventions for recurrence of dysphagia
6) procedure related mortality,
7) major and minor adverse effects of each intervention
8) quality of life.
Trials including interventions with a curative intent and trials that looked at dysphagia improvement as a secondary outcome will be excluded. This is due to the potential for underestimation of improvement in dysphagia when this is not the primary outcome.
Cost‐effectiveness will not be addressed in this review.
Search methods for identification of studies
See: Collaborative Review Group search strategy.
The principal electronic search will be undertaken according to the Cochrane Upper Gastrointestinal and Pancreatic Diseases Review Group module (published in the Cochrane Library).
Cochrane Controlled Trials Register (CCTR)
Relevant studies will be retrieved from the CCTR using the broad search terms used in the title of the paper and the interventions mentioned above.
Complementary Published Electronic Database Searches
Electronic searches of MEDLINE (1966‐2004), EMBASE (1988‐2004) and CancerLIT (1985‐2004) will be undertaken using a combination of subject headings and text words related to the title words. Standard methodological filters will be applied to identify randomised controlled trials.
Correspondence
Experts in the field registered with the Cochrane Collaboration Upper Gastrointestinal and Pancreatic Disease (CC UGPD) review group will be contacted for leads on unpublished studies.
Abstracts
We will hand search Digestive Disease Week (DDW) and United European Gastroenterology Week (UEGW) abstract books between 1994 and 2004. Authors of trial reports published only as abstracts will be contacted and asked to contribute full datasets or completed papers.
We will hand search the reference list of identified articles for further relevant trials.
Data collection and analysis
Selection of studies
Two reviewers, (AS and SW), will independently assess the articles identified by the search for eligibility. A third reviewer will adjudicate in the event of discrepancies and a consensus view will be taken. The reasons for exclusion will be documented. Trials satisfying the inclusion criteria will be included in the review. The quality of each study will be assessed taking into account the following criteria stated in Cochrane Reviewers' Handbook:
Was the allocation truly random?
Was the concealment of treatment allocation adequate?
Were the baseline characteristics reported for the most important prognostic factors?
Were the number of withdrawals, dropouts and losses to follow‐up in each group completely described?
Was analysis performed by intention‐to‐treat?
Data Extraction
Two reviewers, (AS and SW), will independently extract the data using data extraction sheets designed a priori.
The following features will be recorded:
Setting: single centre vs multicentre
Method of randomisation : true vs pseudo, stated vs not stated
Adequacy of allocation concealment: stated vs not stated
Inclusion and exclusion criteria used
Baseline comparability between treatment groups
Dysphagia grade used: 4 point grade vs 5 point grade
Dysphagia improvement as an outcome: primary vs secondary
Location of cancer: upper, mid, lower oesophagus or gastro‐oesophageal junction
Type of cancer:squamous carcinoma vs adenocarcinoma
Length of cancer: stated vs not stated
Proportion of inoperable patients vs unresectable cancer, locally advanced vs metastatic cancer
Proportion of patients that had received previous chemo‐radiotherapy with curative intent.
Adverse effects of the intervention : individual adverse effects vs subclassification as major and minor effects
Data Analysis
All trials included in the systematic review will be entered into Review Manager 4.2 (RevMan). An intention to treat approach will be used in all analyses. Meta‐analysis will be performed only if sufficient trials with similar comparisons and outcome measures are found.
Primary outcome
The primary outcome will be improvement in dysphagia grades. We expect both dichotomous and continuous data in the trials assessing dysphagia improvement. Dichotomous data (1 point and 2 or more point improvement in dysphagia grade) will be expressed as odds ratio (OR) with 95% confidence intervals. If only continuous data is reported, these will be expressed as mean improvement in dysphagia grade with standard deviation following each intervention and will be compared between the groups at specific follow up periods Standardised mean difference will be used to compare the dysphagia improvement between studies using different grades of dysphagia. If dichotomous outcomes could be extracted, 1 point and 2 point improvement of dysphagia grade will be compared between the studies.
Secondary outcomes
Dichotomous data for secondary outcome measures will be expressed as odds ratios or relative risks with 95% confidence intervals. Continuous data for each outcome will be expressed as a means and compared between the intervention groups. Where quality of life indices are available, we intend to document the method used, difference observed between the compared interventions and if appropriate, compare these using group means.
Heterogeneity
Heterogeneity will be assessed using the Chi squared test along with visual inspection of the graph. A significance level less than 0.10 will be interpreted as evidence of heterogeneity. We will look for an explanation for heterogeneity and report this in the review. Sensitivity analysis will be performed using the potential sources of heterogeneity to test the robustness of the overall results. Where no significant heterogeneity is observed between study results, the fixed effects model will be used. Should variation between studies be observed, then the random effects model will be used The potential reasons for heterogeneity hypothesized a priori include:
1. Study quality
2. Study setting (multi centre vs single centre)
3. Dysphagia grade used (4 point vs 5 point grades)
4. Dysphagia improvement as primary outcome vs secondary outcome
5. Location of cancer: upper, mid, lower oesophagus or gastro‐oesophageal junction
6. Type of cancer: squamous carcinoma vs adenocarcinoma
7. Length of cancer
8. Radiotherapy dose fractionation
9. Different types of chemotherapy
10. Different types of stents used
11. Duration of laser and PDT treatment
12. Proportion of inoperable patients vs unresectable cancer, locally advanced vs metastatic cancer
13. Unplanned additional treatment modalities occurring in the intervention groups
