Guidelines on appropriate indications for upper gastrointestinal endoscopyBMJ 1995; 310 doi: https://doi.org/10.1136/bmj.310.6983.853 (Published 01 April 1995) Cite this as: BMJ 1995;310:853
- A T R Axon, consultant physiciana,
- G D Bell, consultant gastroenterologistb,
- R H Jones, Wolfson professorc,
- M A Quine, research fellowd,
- R F McCloy, senior lecturere
- a Centre for Digestive Diseases, General Infirmary, Leeds LS1 3EX
- b Ipswich Hospital NHS Trust, Ipswich IP4 5PD
- c Department of General Practice, United Medical and Dental Schools of Guy's and St Thomas's Hospitals, London SE11 6SP
- d Royal College of Surgeons, London WC2A 3PN
- e University Department of Surgery, Manchester Royal Infirmary, Manchester M13 9WL
- Correspondence to: Dr Axon.
- Accepted 12 December 1994
Upper gastrointestinal endoscopy is a valuable diagnostic tool, but for an endoscopy service to be effective it is essential that it is not overloaded with inappropriately referred patients. A joint working party in Britain has considered the available literature on indications for endoscopy, assessed standard practice through a questionnaire, and audited randomly selected cases using an independent panel of experts and an American database system. They used these data to produce guidelines on the appropriate and inappropriate indications for referral for endoscopy, although they emphasise that under certain circumstances there may be reasons to deviate from the advice given. The need for endoscopy is most difficult to judge in patients with dyspepsia, and this aspect is discussed in detail. Early endoscopy will often prove more cost effective than delaying until the indications are clearer.
Most patients referred for endoscopy complain of symptoms that come under the general heading of dyspepsia.1 2 3 4 5 Gastrointestinal symptoms are responsible for about 10% of the work of general practitioners,6 with upper abdominal symptoms (principally dyspepsia) accounting for about half. The severity and frequency of symptoms alone does not predict the likelihood of consultation, however; concerns about the meaning of symptoms, particularly in terms of malignancy and heart disease, play an important part in determining whether a patient consults. This is reflected in the finding that dyspeptic patients with negative findings on endoscopy have greatly reduced consultation and prescribing rates after the procedure.7
The diseases most commonly sought by endoscopy are reflux oesophagitis (and its complications), oesophageal varices, oesophageal cancer, gastric ulcer, gastric cancer, duodenal ulcer, and coeliac disease. The main purpose of the investigation is to identify the cause of symptoms in order to start suitable treatment. A secondary reason is to exclude organic upper gastrointestinal disease so that further investigations can be undertaken to identify the cause of symptoms or alternatively the patient and clinician can be reassured that no serious disease is present.
For endoscopy services to be used to their best advantage only those patients for whom the procedure is appropriate should be referred. A joint committee of the Royal College of Physicians of London, Royal College of Surgeons of England, Royal College of Anaesthetists, Association of Surgeons, the British Society of Gastroenterology, and the Thoracic Society of Great Britain has been auditing gastrointestinal endoscopy in England. On the recommendations of that committee a working party was set up to determine the appropriate indications for endoscopy. This report details the findings of the working party.
The working party did a systematic manual search and a Medline search of literature published on the appropriateness and inappropriateness of and indications for endoscopy since 1981. The epidemiology of upper gastrointestinal disease in the United Kingdom was specifically considered because the prevalence and types of dyspeptic disease differ widely throughout the world. In this report dyspepsia is defined as upper abdominal or retrosternal pain, discomfort, heart burn, nausea, vomiting, or other symptom considered to originate from the proximal alimentary tract.3 Heart burn is defined as an intermittent burning pain felt in the retrosternal area often but not always associated with a feeling of gastro-oesophageal regurgitation and typically relieved by taking antacids.
We also assessed current practice in Britain in two ways. We sent 1297 questionnaires to a random selection of doctors comprising 350 general physicians, 400 surgeons, 477 gastroenterologists, and 70 general practitioners.8 Questions were asked in the form of a brief clinical scenario to which the doctors had to reply whether or not they would usually refer for endoscopy.
An audit panel of seven judges (physicians and surgeons with an interest in gastroenterology) assessed 390 cases randomly selected by the audit coordinator from four endoscopy units throughout East Anglia (three in district general hospitals and one in a teaching hospital). The cases represented one month's work for these units. The cases were also assessed independently for appropriateness by using an American database last revised in January 1989.8 9 The methods and results of the questionnaire and assessment have been published8 and form the basis of this report.
Appropriateness was defined as follows.10 For a procedure to be appropriate its expected benefit should be greater than its expected negative consequences by a sufficiently wide margin to make the procedure worth doing. Both benefits and negative consequences are defined in the broadest terms. Therefore, in addition to reduced mortality and morbidity, benefits included reduced anxiety, pain, and time lost from work. Similarly, negative consequences included the risk of a false positive result (if the procedure under review is a diagnostic study) and the pain and anxiety caused by the procedure as well as its direct mortality and morbidity.
of questionnaire and assessments
Table I shows the results of the questionnaire. The response rate was 66%. The proportions of doctors who would request endoscopy in the various clinical situations varied from 4.5% to 99%. Although there was a wide range of opinion as to which patients are appropriate for endoscopy, agreement was good for the most and least serious clinical situations.
No difference was found in the referral pattern among the four endoscopy units from which the 390 cases were selected for further assessment. The audit panel considered 321 (82%) to have been appropriate referrals and 45 (12%) inappropriate; in 24 (6%) the indication was equivocal. No serious disease would have been missed by the panel. However, the American software system considered 120 (30%) of the referrals to have been inappropriate. With hindsight the computer system would have missed one gastric cancer, three duodenal ulcers, and two gastric ulcers (serious disease being missed in 5% of the patients identified as inappropriately referred). The commonest cause of inappropriateness found by the computer was that the patient had not had adequate formal antiulcer treatment. British doctors seem less likely than their American counterparts to defer endoscopy until a antiulcer treatment has been tried. The audit panel was more likely to label a case inappropriate when a clear history of irritable bowel disease or a chronic functional disorder was described.
Circumstances appropriate for endoscopy
From the published data, standard practice, and the results of the questionnaire, the audit panel, and American database analysis the working party identified specific indications or symptoms that usually indicate the need for endoscopy (box 1). These indications account for a relatively small proportion of patients referred for endoscopy. Most present with non-specific upper abdominal symptoms, and it is subdivision of this group that is most likely to identify those who are inappropriately referred for endoscopy. Those circumstances in which endoscopy would be appropriate for patients with dyspepsia are described below and in box 2.
Box 1—Symptoms or indications for which endoscopy is usually appropriate
Unexplained iron deficiency anaemia
Positive occult blood test result
Dysphagia (food sticking)
Odynophagia (painful swallowing)
Severe upper abdominal pain
Moderate, long standing upper abdominal pain
Unexplained weight loss
Severe heart burn
“Suspicious” barium meal result
Check gastric ulcer for healing
Suspected coeliac disease
Box 2—Appropriate and inappropriate indications for endoscopy in patients with dyspepsia
Any patient over the age of 45 with recent onset of dyspeptic symptoms or change in dyspeptic symptoms
Patients under the age 45 with dyspepsia who are positive for Helicobacter pylori on non-invasive testing or who have other risk factors such as treatment with non-steroidal anti-inflammatory drugs
Patients in whom continuous long term treatment with
H2 receptor antagonists, acid pump inhibitors, or prokinetic drugs is planned
Patients under the age of 45 with severe and persistent symptoms that do not respond to treatment
Patients of any age who have typical symptoms of irritable bowel syndrome rather than dyspepsia
Patients with mild or moderate reflux symptoms which respond to simple measures such as change in lifestyle, antacids, and alginates
Patients known to have duodenal ulcer who are responding to treatment
Patients who have had a single episode of dyspepsia and are now asymptomatic and not receiving treatment
Patients over the age of 45 with a recent onset of dyspeptic symptoms or change in dyspeptic symptoms—These patients should have endoscopy at an early stage so as not to miss gastric cancer.11 This disease is the fourth commonest lethal malignancy in England and Wales, accounting for over 10000 deaths a year. Surgery has most benefit in patients with early gastric cancer—that is, where the lesion is confined to the mucosa and submucosa but not invading the muscularis propria.12 14 Such patients present with mild symptoms that are often indistinguishable from benign disease.15 In the United Kingdom cancer is rare under the age of 45. The American computer system would prefer all patients with dyspepsia to receive a trial of treatment first. This argument was rejected by the British audit panel on the basis that the symptoms of cancer may respond to acid suppression thus delaying diagnosis. Indeed, in the 390 cases selected for analysis one patient with gastric cancer was regarded as inappropriate for endoscopy by the American computer system but would have had cancer identified using the criteria of British audit panel. One reason for the difference of approach between the two countries may be that gastric cancer is less common in the United States than in Britain.
Patients under the age of 45 with dyspepsia who are positive for Helicobacter pylori on non-invasive testing or who have other risk factors such as treatment with non-steroidal anti-inflammatory drugs—The main reason for endoscopy in this group is to diagnose gastroduodenal ulcer. Infection with H pylori is responsible for over 95% of duodenal ulcers and most gastric ulcers, and eradication of the bacteria leads to long term cure.16 Infection can be detected by a simple blood test, which is available in most hospitals. Five recent studies in Britain have assessed the predictive value of serological testing for H pylori,17 18 19 20 21 and all have confirmed that the likelihood of finding serious organic disease (duodenal ulcer, gastric ulcer, and gastric cancer) is small in those who are serologically negative and who are not taking non-steroidal anti-inflammatory drugs.
Patients in whom continuous long term treatment with H2 receptor antagonists, acid pump inhibitors, or prokinetic drugs is planned—Table II shows the cost of long term treatment with five drugs commonly prescribed for non-specific dyspeptic symptoms or gastro-oesophageal reflux disease. Long term treatment is expensive and in many cases unnecessary. Firm evidence that no serious disease is present may avoid the need for long term treatment. Conversely, it is desirable to know whether a patient has ulcerative oesophagitis, early stricturing, or a Barrett's oesophagus because these findings may influence management.
Patients under the age of 45 with no H pylori infection but severe and persistent symptoms that do not respond to treatment—Uncommon gastroduodenal conditions may affect this age group, including gastric cancer, gastric lymphoma, leiomyoma, erosive gastritis, and H pylori negative peptic ulcer. This was reflected in the fact that 97% of respondents to our questionnaire thought that endoscopy was appropriate for patients with resistant dyspepsia.
Inappropriate indications for endoscopy
Box 2 also lists the clinical situations in which endoscopy is inappropriate. The commonest reason the panel found for endoscopy being inappropriate was a clear history of irritable bowel syndrome without features of ulcer disease (30%). Typical symptoms of irritable bowel syndrome are abdominal distension, relief of pain with bowel movement, loose stools associated with pain, frequent stools with onset of pain, mucus passed per rectum, feeling of incomplete evacuation.22
Mild or moderate reflux symptoms may respond to simple measures such as change in lifestyle, antacids, and alginates. The changes in lifestyle to be considered include weight reduction, stopping smoking, avoidance of foods that precipitate symptoms, and if possible stopping any drugs that are associated with dyspepsia—for example, corticosteroids, non-steroidal anti-inflammatory drugs, and potassium chloride. Very heavy meals and excessive alcohol should also be avoided, and if symptoms are present at night the patient should be encouraged to eat earlier in the evening.
Patients who have had one episode of dyspepsia but are subsequently free of symptoms and taking no drugs were not considered appropriate for endoscopy by the audit panel of judges. In addition this was the second commonest reason for non-approval in the computer analysis.
Cancer screening or surveillance
Use of endoscopy to screen for cancer is a controversial area. Screening of patients with pernicious anaemia and surveillance of the stomach after surgery have been extensively analysed and regular endoscopy is probably not merited.26 27 28 Screening is common for Barret's oesophagus and in patients with familial adenomatous polyposis (to detect gastroduodenal cancer), but no authoritative analysis exists to support this.
Resource constraints have in the past tended to result in endoscopy being deferred as long as possible. However, up to 70% of all patients with persistent dyspeptic symptoms have either a barium meal examination or endoscopy at some stage,27 and it may be clinically and financially more appropriate to investigate earlier rather than later. This view is supported by clinical trials from Australia and Denmark in which empirical treatment followed by endoscopy (when necessary) was compared with early endoscopy followed by appropriate treatment.28 29 At six months no difference was apparent in the clinical and financial outcomes in the Australian study while after one year the Danish paper found a cost advantage for early endoscopy. A computer model study from Sweden also found that early endoscopy had the financial edge over deferred investigation.30 There may also be clinical advantage in the early detection of premalignant lesions or early gastric cancer in patients whose dyspepsia is investigated promptly.11
We believe that didactic recommendations cannot be made in a subject as non-specific as dyspepsia. Decision making must take account of the patient—for example, some patients with intermittent dyspepsia responding to simple treatment may be very worried that their problem has a serious underlying cause. If a doctor is unable to dispel a patient's concern by careful explanation and reassurance, endoscopy may be warranted.
Endoscopy must stand up to rigorous audit. The view that any upper abdominal symptom should be investigated in the first instance by endoscopy should be resisted. The procedure is expensive, somewhat unpleasant for the patient, and carries a small risk of complications. The cost and risk must be carefully weighed against the potential advantage. Nevertheless, the management of dyspepsia is rapidly changing with newer tests, newer concepts concerning the pathogenesis of dyspepsia, and more effective and powerful drugs. These recommendations will require modification in the light of changing circumstances.
The members of the joint committee were H Brendan Devlin (chairman), J D Hardcastle, Royal College of Surgeons of England; A T R Axon, G D Bell, P D Fairclough, R J Leicester, R F McCloy, British Society of Gastroenterology; J E Charlton, Royal College of Anaesthetists; A Hopkins, Royal College of Physicians of London; H R Matthews, Thoracic Society of Great Britain; D Watkin, Association of Surgeons of Great Britain and Ireland; R H Jones, Department of General Practice, United Medical and Dental Schools of Guy's and St Thomas's Hospitals, London; M A Quite, audit research fellow of Royal College of Surgeons.
We thank Miss C McCourt and Miss C Brizzolara, administrative assistants for the British Society of Gastroenterology and the Royal College of Surgeons, for their help in preparing this paper and Mrs A Scales for typing the manuscript.