Skip navigation

Right in the gut: Barrett's oesophagitis and colon cancer

Bandolier has found three interesting papers on evidence in gastroenterology. One is to do with the relationship between Barrett's oesophagitis and colon cancer, and the others tell us about risks associated with endoscopy and agreement beween clinicians. One is a systematic review and the other is an audit, but they illustrate how collecting evidence informs us.

Barrett's oesophagitis and colon cancer

Barrett's oesophagitis is a complication of long-standing gastro-oesophageal reflux disease (GORD). Prolonged and excessive exposure of the lining of the oesophagus leads to changes in the types of cell which make up the epithelium. Instead of the usual flat, squamous, cells lining the oesophagus, columnar cells take over. This can extend from just a few centimetres from the gastro-oesophageal junction to the full length of the oesophagus.

Development of columnar epithelium is associated with an increased risk of oesophageal cancer (about 1 in 50 to 170 patient years). But a recent systematic review [1] demonstrates that patients with Barrett's oesophagitis also have an increased risk of colon cancer.

Systematic review

The authors, from South Carolina, did a search to identify papers looking at colon cancer, polyps, neoplasms or adenomas in Barrett's oesophagus. They extracted data on patients with the disorder, and on controls, and on the numbers of patients found to have colon cancer or adenomas on colonoscopy.

They found five uncontrolled studies, in which the prevalence of colon cancer in patients with Barrett's oesophagus was 4.6% (8/174). The prevalence of colon adenomas was 27% (36/134).

They found nine papers with control groups. In size they varied from 17 to 175 patients and found rates of colon cancer among patients with Barrett's oesophagus of 0 to 14%. Overall 52 of 685 had colon cancer, an average rate of 7.6%.

The prevalence of benign plus malignant colon neoplasms ranged between 18 and 47%. Overall 176 of 510 had colon cancer, an average rate of 35%.

The authors also created a comparison cohort of 513 patients aged less than 80 years described in studies of colonoscopic screening for colorectal cancer. Of these 513, 8 (1.6%) had colon cancer and 169 (33%) had colon neoplasms.

So the conclusion is that while colon neoplasms occur at about the same rate in people who have and who do not have Barrett's oesophagitis, the risk of developing colon cancer is about five times greater in patients with Barrett's oesophagitis. Nearly 8% of them will have colon cancer. Is this prevalence high enough to consider screening?

Endoscopists agree with each other

Yes they do. A series of studies assessing the agreement obtained between experienced endocopists examining slides and videos, and between experienced and training endoscopists looking at slides, agreement was acceptable to excellent for identification of important features like mucosal breaks and complications of reflux disease [2].

Values for kappa ( Bandolier 37 ) were in the range 0.5 to 0.9.

Endoscopy safety

Risks are usually associated with therapy - some rare but severe adverse effect, for instance. But a prospective audit of endoscopy done in the UK gives us an insight into the risks of diagnosis. Not just having a diagnosis which could change your life, but the risk of the test itself.

A prospective audit of 13,036 diagnostic and 1,116 therapeutic upper gastrointestinal endoscopies were audited in the North West region and East Anglia [3]. There were 104 deaths within 30 days of the procedure. Some died as a result of perforation, but in many cases the relationship with endoscopy was debatable. For instance, deaths due to pneumonia, infarct and cerebrovascular accident often occurred in sick, elderly patients.

The conservative estimate was that in the 13,036 patients undergoing diagnostic endoscopy there were seven endoscopy-related deaths, or 1 in 2,000. The authors thought that this might be an under-estimate.

This is a thorough audit, well reported and with a host of details. It draws attention to ways in which things can be made as safe as possible, and to the appropriate guidelines.


  1. CW Howden, CA Hornung. A systematic review of the association between Barrett's esophagus and colon neoplasms. American Journal of Gastroenterology. 1995 90:1814-9.
  2. D Armstrong et al. The endoscopic asessment of esophagitis: a progress report on observer agreement. Gastroenterology 1996 111:85-92.
  3. MA Quine, GD Bell, RF McCloy, JE Charlton, HB Devlin, A Hopkins. Prospective audit of upper gastrointestinal endoscopy in two regions of England: safety, staffing, and sedation methods. Gut 1995 36:462-7.

previous or next story in this issue