The working party of the BSG has recently produced a document updating recommended guidelines for the management of Barrett's oesophagus (BO).1
In this article, the main changes in recommendations are highlighted and their value in the context of the numerous other guidelines and manuscripts that are already available are discussed.
The two key “new” recommendations are
- BO is defined as an endoscopically apparent area above the oesophagogastric junction that is suggestive of Barrett's which is supported by the finding of columnar lined oesophagus on histology. The presence of areas of intestinal metaplasia (IM), although often present, is not a requirement for diagnosis.
- For patients with BO but without dysplasia, the recommended surveillance protocols are two yearly, four quadrant biopsies every 2 cm, but jumbo biopsies are not required.
Additional recommendations include the advice that endoscopic screening of patients suffering from heartburn in order to detect BO is not recommended and that, in patients with non‐dysplastic BO, ablation should be performed only in the context of prospective randomised studies.
The new recommended definition of what constitutes BO requires a combination of macroscopic and microscopic identification. In the latest definition, in order to have Barrett's mucosa you have to be able to see it with an endoscope. This therefore excludes “ultra‐short Barrett's” and also does not require a
3 cm length (neither of which is likely to generate much controversy). In contrast, histological confirmation still requires the presence of a columnar lined oesophagus but does not
require areas of IM to be found. This therefore resembles older style guideline definitions and is out of synchrony with several other current guidelines from other countries.2
While it is agreed that adenocarcinoma probably usually originates from a segment containing IM, the rationale behind this decision is that sampling errors at the initial endoscopy may miss an area(s) of IM. Based on a publication,3
the guidelines support this decision by their conclusion that “If a sufficient number of biopsies are taken over an adequate period of time, IM can usually be demonstrated (in the majority of these patients)”. There is logic in this decision as, using the American guidelines, if a patient has an endoscopy and only areas of CLO without IM are found, by definition, the patient would not have BO and may not be entered into a programme purely due to sampling errors. Nevertheless, this decision may well be a major issue in future meta‐analyses and in the acceptance of the generality of clinical research findings obtained from opposite sides of the Atlantic as well as within Europe.
The other major area that is likely to produce heated debate is the advice that, in patients without dysplasia, the appropriate surveillance interval for UK patients is every two years. The advised intervals for surveying patients with BO without dysplasia in various guidelines have shown marked temporal and regional variations. The current recommendation is based on a new Markovian analysis by a member(s) of the guidelines group.1
Unfortunately, details of this analysis are only briefly stated in the document and it has not been published (or peer reviewed) elsewhere. As only minimal data on the assumptions used are provided, this causes real issues with regards to the validity of this decision which will have a major impact on costs and clinical services. In addition to concerns over the validity of this model, there is also the problem that many gastroenterologists have recently spent a large amount of time identifying patients who were “unnecessarily” undergoing annual surveillance and ensuring that they understood the rationale for prolonging to three yearly intervals (including setting up systems and follow up appointments to allow this occur). Is the new data really strong enough to go back and alter these patient pathways?
The current guidelines are a substantial document and provide an excellent overview of BO. For people with a particular interest in BO this will be a key resource. However, for the general gastroenterologist, it is a somewhat heavy read in order to obtain the key messages highlighted above. As stated in the guidelines, what we really need are new data to answer some of the fundamental questions and this is borne out by the fact that of the 22 principle recommendations, 17 are based only on grade C (professional opinions) evidence. With regard to this lack of robust data, two developments are worthy of mention. A trial of the use of aspirin prophylaxis to reduce the risk of cancer progression in BO (ASPECT) is currently underway and is likely to provide important answers in a few years' time. ASPECT is also comparing the effect of “normal” (20 mg) and high dose (80 mg) esomeprazole on cancer development. While this second component is of interest, it is severely limited by the failure to include either a “no proton pump inhibitor” or a true “symptom control (as required, PRN)” arm which is what is really required to answer the question “Does acid suppression affect the natural history of BO?”.
Because of the protocol being used, the ASPECT study is unlikely to answer many questions regarding the value of surveillance endoscopy. It is therefore timely that the Health Technology Assessment (HTA) group (part of the UK Government Healthcare Commission), who have been considering for several years what questions they would like to examine in the area of BO, are in the process of potentially funding a study on the value of endoscopic surveillance of BO. The cost of performing a full examination of this question however will be expensive (the ASPECT study has cost several million) but it is likely that HTA funding will be limited to around £300
000. As it is generally agreed that this study will require several thousand patients and run for 5–10 years, there is a real danger of it being under resourced. Part of this expenditure is that, if patients are to be truly randomised, patients who are entered into the “no surveillance” arm cannot be left to “fend for themselves” and are likely to need regular clinic appointments (which may well not have occurred if they were not in the study) with their associated costs. In the era of full economic costing, approaching centres and asking them to take part without funding support is therefore likely to prove difficult. In addition, while patients on the edge of suitability for surveillance (due to age or coexisting morbidity) are likely to be reasonably happy to take part, the willingness of younger people is likely to be lower and appropriate stratification must be taken into account to allow an answer to the regularly asked question “But what would you do if a 30 year old was found to have BO?”. What nobody wants is a trial that will take several years to perform, have low take up, or be dismissed as irrelevant if it is underpowered or confounded. Nevertheless, despite these caveats, it is refreshing to see that there appears to be a realisation that new solid data must be obtained if the old cyclical arguments between advocates and sceptics of the value of routine surveillance of BO is to be resolved. Watch this space (lumen)…