Article Text


Endoscopic surveillance in Barrett’s oesophagus
  1. J Ryan1
  1. 1Brighton Gastroenterology Associates, 2 Church St, Brighton, Victoria 3186, Australia; jeremy{at}
  1. H Barr2
  1. 2Gloucestershire Royal Hospital, Great Western Road, Gloucester GL1 3NN, UK; prof.barr{at}

Statistics from

I read with interest the debate on endoscopic surveillance in Barrett’s oesophagus (Gut 2002;51:313–14, 314–15). My reading of the literature supports the view of Dr Playford—there is insufficient evidence to justify surveillance endoscopy in this condition.

I am always interested in the uses and misuses of statistical data to support a personal view, and to that end I have some questions that should be honestly answered by those advocating screening: firstly, where is the evidence, prospectively collected, that shows that Barrett’s oesophagus is a consequence of acid reflux disease? The quoted references do not support this allegation. Secondly, I believe that it is deliberately obfuscatory to liken Barrett’s oesophagus to a colonic polyp in terms of malignant potential—abundant evidence supports the role of screening in the latter common condition. Lastly, this issue of absolute risk should indeed be addressed. Colon cancer is common—at least 20 times more common than the oft quoted epidemic of oesophageal adenocarcinoma—which may not be associated with Barrett’s oesophagus.

We need to remember that despite an increase in the reported incidence of oesophageal adenocarcinoma in recent decades, limited endoscopic resources may be better devoted to reducing the disease burden in a condition where reliable evidence supports surveillance.

Author’s reply

Dr Jeremy Ryan’s contribution to the debate is most welcome. He will not be surprised that I find his and Professor Playford’s arguments incorrect, but accept that they both pose insightful and valuable questions, which are correct.

Sufficient evidence

Firstly, without entering a philosophical dialectic, I must address his view of sufficient evidence. It is important to be mindful that “life is the art of drawing sufficient conclusions from insufficient premises” Samuel Butler. There are now further data to support the case for surveillance. A population based study of a cohort of patients with adenocarcinoma of the oesophagus and gastric cardia has concluded that surveillance detected Barrett’s oesophagus related cancers were associated with low stage disease and improved survival with no patient dying directly of cancer.1 A major problem of this study is that most patients were excluded because they did not have a diagnosis of Barrett’s oesophagus made six months prior to the diagnosis of cancer.2 The major challenge is finding Barrett’s oesophagus as it is an adaptive phenotype. It is becoming clear that many patients with Barrett’s oesophagus are asymptomatic, not complaining of reflux symptoms.3

Misuse of data supporting a personal view

Dr Ryan is absolutely correct that I do have a strong personal view, hopefully displayed in the debate. In the opening argument I alluded, perhaps obscurely, to the postmodern Nieztzschian philosophy of there being “no facts merely interpretations”. This is an approach I personally reject and in doing so, results correctly but perhaps harshly to a charge of lack of equipoise and misuse of data. I hope Dr Ryan accepts this explanation in mitigation. I am happy to inform him that there is a small amount of prospective evidence that reflux disease leads to Barrett’s oesophagus. McDougall and colleagues4 conducted a careful natural history study of patients with reflux oesophagitis. To my mind the most striking feature was that 11% of patients with oesophagitis developed Barrett’s mucosa after 3.5–4 years. These clinical data support substantial pathophysiological and experimental data.5,6

Barrett’s oesophagus and the colonic polyp

Again, I find that Dr Ryan is correct in part; stating that the burden of colon cancer is much greater than that of gastro-oesophageal cancer. My purpose in the debate was to highlight consequential rather than absolute risk. Most patients will survive a diagnosis of symptomatic colon cancer; very few will survive a diagnosis of symptomatic oesophageal cancer. Most Barrett’s patients are asymptomatic,3 and very few patients with a diagnosis of gastro-oesophageal cancer have a prior diagnosis of Barrett’s oesophagus.1 Those patients fortunate enough to be detected must be surveyed, as the consequence for the patient of ignoring their Barrett’s oesophagus is to inform them to return when they notice “alarm” symptoms of dysphagia. This latter strategy I strongly contend is wrong as the patient is unlikely to survive


View Abstract

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.