Article Text


Endoscopy I
Low Yield of Colonoscopy for Melaena Following a Normal Upper Gastrointestinal Endoscopy: A 5-year Analysis
  1. S H Yung *,
  2. J Coughlan,
  3. S M Everett,
  4. A C Ford
  1. Leeds Gastroenterology Institute, Leeds General Infirmary, Leeds, UK


Introduction Patients with melaena with a normal upper gastrointestinal (GI) endoscopy often undergo colonoscopy, due to concerns they may have a caecal or right-sided colonic carcinoma. We investigated yield of colonoscopy in this group of patients, and the proportion with an entirely normal upper GI endoscopy, without any clinically relevant findings.

Methods This was a retrospective analysis of data collected prospectively within a single centre between January 2005 and September 2010. All patients undergoing colonoscopy for melaena were identified. Only data from patients who had also undergone an upper GI endoscopy, with melaena as the primary indication, in the prior 3 months were included. Gender, age, comorbidity (American Society of Anaesthesiologists (ASA) score), quality of bowel preparation, extent of colonoscopy, colonoscopic findings, and histopathology (where relevant) were recorded. A normal upper GI endoscopy, without clinically relevant findings, was defined as one where none of severe erosive oesophagitis, non-bleeding varices, non-bleeding gastric or duodenal ulcer, multiple non-bleeding gastric or duodenal erosions, gastro-oesophageal malignancy, or non-bleeding angiodysplasia were reported.

Results In total, 166 patients had colonoscopy for melaena with an upper GI endoscopy for the same indication in the previous 3 months. Of these, 139 (83.7%) had complete colonoscopy, and contributed data. Mean patient age was 66.6 years (range 22–92 years), and 87 (62.6%) were male. Median ASA score was 2. Quality of bowel preparation was adequate, or better, in 107 (77.0%) patients, poor in 17 (12.2%), and unreported in the remainder. A potential lower GI cause for melaena was found in 11 (7.9%) patients. Six had angiodysplasia, 4 histologically-confirmed inflammatory bowel disease or ulceration, and 1 (0.7%) a caecal cancer. Three other patients had visible blood or melaena, with no definite lower GI cause identified. Excluding these 3 from the analysis, 102 (75.0%) of 136 had an entirely normal upper GI endoscopy, without clinically relevant findings. 9 of 11 (81.8%) patients with and 93/125 (74.4%) without a lower GI cause of melaena at colonoscopy had an entirely normal upper GI endoscopy (p = 0.82). There were no significant differences in patient age (p = 0.39), gender (p = 0.89), or ASA status (p = 0.32) between those with and without an identifiable lower GI cause for their melaena.

Conclusion Up to 25% of patients who proceeded to colonoscopy had clinically relevant findings at upper GI endoscopy that may have accounted for their presentation with melaena. However, an entirely normal upper GI endoscopy did not predict findings at colonoscopy. Less than 1 in 10 patients with melaena who had colonoscopy had a lower GI cause detectable, and less than 1% had colon cancer.

Competing interests None.

  • colonoscopy
  • colorectal cancer
  • melaena

Statistics from

Request permissions

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.