Table 4UPPER GASTROINTESTINAL CANCER: signs and symptoms

AuthorSettingDescriptionNo.InclusionExclusionResultsGold StdQuality
Adachi et al 1993JapanRetrospective case series. it aimed to identify the most effective approach for detecting superficial oesophageal carcinoma was investigated through clinical histories using hospital charts.46 patients with superficial oesophageal cancer

49 patients with advanced oesophageal cancer.
Patients with superficial and advanced oesophageal cancerPatients who underwent preoperative therapy such as hyperthermia, chemotherapy and radiationSymptoms were more frequent and the size of lesions was larger with increasing depth of invasion. A piercing sensation was present mostly in superficial oesophageal carcinoma, while pain or dysphagia was present both in advanced oesophageal cancer and submucosal carcinoma.Pathology
Ahlgren 1996Pancreatic cancer risk factors were reviewed------Direct evidence linking specific dietary carcinogens to pancreatic cancer in humans was difficult to establish.
Bakkevold et al 1992Norway Primary careRetrospective case series. Data on signs and symptoms were reported from case history and information provided prospectively on sensitivities of diagnostic investigations.
The aim was to compare the symptoms and signs, delays in diagnosis, and the efficacy of diagnostic methods of pancreatic cancer at Norwegian hospitals
472Patients with verified carcinoma of the pancreas or the papilla of VaterPatients with endocrine tumour, cholangiocarcinoma, metastatic pancreatic tumour, cystadenocarcinoma, and unverifired primary pancreatic tumour.Jaundice without pain was present in 18%. Nonspecified symptoms occurred in 49%, the commonest being dyspepsia (12%), diarrhoea/steatorrhoea (12%) and nausea in 5%.
Jaundiced patients had less advanced tumours at staging (p=0.0000), but abdominal pain and/or weight loss predicted advanced disease (p=0.0001 and 0.004 respectively.
Histology or cytology
Crean et al 1982UKProspective Cohort study. A database was created to enable the development of a diagnostic decision system for dyspepsia by recording associated symptoms and clinical features.1000 patientsIndicants of gastric cancer were listed as age > 55, history < 1 year, daily pain, dark vomitus, early repletion, weight loss and interscapular radiation.(Endoscopic or radiological observations were made)
Crean et al 1994UKProspective cohort study. The study of dyspepsia was carried out in a primary referral hospital to elicit relevant clinical information on the principal diagnoses.1540 patientsPatients that met the definition of dyspepsia and were seen at a primary referral hospital between 1974 and 1987.--Of patients attending a pimary referral hospital, the commonest principal diagnosis were duodenal ulcer (26%), functional dyspepsia (22%) and irritable bowel syndrome (15%), alcohol related dyspepsia (4%) was as common as gastric carcinoma or symptomatic gall stones.Biopsy specimens were taken depending on the findings.
Delaney et al, 2000The aim was to determine the cost effectiveness of initial endoscopy compared with usual management in patients with dyspepsia over age 50 years presenting to their primary care physician.

patients were recruited and randomly assigned to initial endoscopy or usual management.
Primary outcomes were effect of treatment on dyspepsia symptoms and cost-effectiveness.
Secondary outcomes were quality of life and patient satisfaction.
422 patients213 (84%) patients had an endoscopy compared with 75 (41%) controls. Initial endoscopy resulted in a significant improvement in symptom score (p=0·03), and quality of life pain dimension (p=0·03), and a 48% reduction in the use of proton pump inhibitors (p=0·005).
The ICER was £1728 (UK£) per patient symptom-free at 12 months. The ICER was very sensitive to the cost of endoscopy, and could be reduced to £165 if the unit cost of this procedure fell from £246 to £100
Department of Health Guidelines, 2000The guideline was based on a report of the evidence prepared by a multi-disciplinary working group. It included a cost- effectiveness decision analysis of fast track referral for patients at risk of upper gastrointestinal malignancy.50The incidence of stomach cancer is decreasing, whereas the incidence of oesophageal cancer is increasing. Tumours at the junction between the stomach and oesophagus are increasing particularly rapidly.
Dysphagia is a relatively uncommon symptom in a community/general practice setting. Patients with difficulty swallowing food should always be referred for further investigation.
Dyspepsia is an extremely common problem in a community/general practice setting. The index of suspicion of cancer is very considerably raised if dyspepsia is combined with an ‘alarm’ symptom (weight loss, vomiting, anaemia). In patients aged over 55 years, recent onset of dyspepsia and/or continuous symptoms is associated with increased risk of cancer.
Duggan 1999UKThis paper evaluated the current cost of upper gastrointestinal disease in the UK, the base IGPCG algorithm and the 5 major alternative scenarios.The original IGPCG algorithm was the least costly option of all those considered, with additional H. pylori testing for all patients with suspected ulcer being the second least expensive option. Routine endoscopy for all patients or for all patients aged more than 45 years were the most expensive scenarios and would require a 16- or 13-fold increase, respectively, in the provision of endoscopy services in the UK.
The use of routine endoscopy for all patients aged more than 45 years who were presenting with upper gastrointestinal symptoms for the first time was a mid-priced option, but would still require a 5-fold increase in the provision of endoscopy services. The modelling process highlighted the fact that early stratification of patients into diagnostic and treatment groups, on the basis of history and symptom cluster, is a less costly approach than that of early routine endoscopy or H. pylori testing. If H. pylori testing is to be used routinely, then the least costly approach is to select those patients who have symptoms that are more indicative of ulcer disease.
Fielding et al 1980UKRetrospective case series. The signs and symptoms associated with early gastric cancer were reported.90Patients diagnosed with early gastric cancer.
Data obtained from cancer registry
--Number (n=90) and percentage of patients experiencing symptoms was given for epigastric pain 26 (28.9%), vomiting 21 (23.3%), abdominal pain 17 (18.9%), weight loss 17 (18.9%), anorexia 13 (14.4%), indigestion 11 (12.2%), haematemesis 6 (6.6%), dysphagia 1 (1.1%).Histology
Gillen et al 1999UKRetrospective case series. Aimed to establish whether endoscopy was justified in uncomplicated dyspepsia in patients aged less than 55.169 patients.Patients years diagnosed with gastroesophageal cancer. Patients were identified between 1989 and 1993 from the West of Scotland Cancer Registry.Prevalence of symptoms for gastric and for oesophageal cancer are listed as follows: weight loss 61.8% and 63.0% respectively; persistent vomiting 35.6% and 35.6%, dysphagia 23.7% and 84.9%, anaemia 22.4% and 5.5%, hem’sis melena 18.4% and 2.7% and palpable mass 9.2% and 0%.
Gold and Goldin 1998USAA review of epidemiological and risk factors for pancreatic cancer.
Heading et al 1999UKA systematic review was undertaken of all studies on the population prevalence of upper gastrointestinal symptoms.10 studiesStudies that had been published up to December 1997, if sample size and response rate were reported, if vague terms such as dyspepsia or indigestion were defined, abdominal pain or discomfort enquired about, and patients with a history or evidence of organic disease had not been excluded.Follow-up studies on groups of patients previously studiedThe reported prevalence of upper abdominal symptoms (mostly upper abdominal pain or discomfort) ranged from approximately 8% to 54% while the prevalence of heartburn and/or regurgitation ranged from 10% to 48% for heartburn, from 9% to 45% for regurgitation and 21% to 59% for both/either. Variations were attributed to varying definitions used.No meta analysis was undertaken and the studies were difficult to compare because the definition of signs and symptoms was not consistent.
Irving et al 2002UKRetrospective case series to assess the impact of DoH cancer referral guidelines (2000) to reduce delay from presentation to referral.90 patientsPatients with treated at a oesophago-gastric cancer unit at a hospital between 1 November 1999 and 30 December 2001.--Sixty-five patients were diagnosed with oesophageal cancer and 25 with gastric cancer. Dysphagia was the most common presenting symptom and it was experienced by 58 patients in the study (64%). It was much more prevalent in patients with oesophageal rather than gastric malignancies (77% versus 32%).HistologyThe study did not give details about the main focus of the study being on monitoring the speed with which patients with cancer were detected, referred and diagnosed.
Klamer et al 1982USARetrospective case series aimed to investigate epidemiologic factors, presenting symptoms, diagnostic methods, site and extent of cancer, treatment approaches and survival data associated with pancreatic cancer through examining patients’ charts.33The charts of all patients treated for cancer at Mount Sinai Medical Center between 1971 and 1978Patients with cancers arising from periampullary and islet cell tissueThe most common complaint leading to hospitalisation was abdominal pain, which occurred in 23 (70%), followed by jaundice in 19 (57%), anorexia in 15 (45%), weakness in ten (30%), and nausea in eight (24%). Six patients (18%) complained of pruritis or diarrhea.HistologyThe case series was based on a small sample size.
Lowenfels and Maisonneuve 2002USA and ItalyA review of epidemiologic factors in pancreatic cancer.Number of studies not mentioned----The proven risk factors were identified as being smoking, age and pancreatitis. Other potential risk factors were listed as being diabetes, peptic ulcer disease, gallstones, infections, salmonella, helicobacter pylori, obesity, diet, occupation, inherited and gene-environment factors. The relationship between smoking and pancreatic cancer has been studied extensively in case-control and cohort studies.
Age was discussed as being the strongest risk factor.
Pancreatic cancer is extremely unusual in patients younger than age 30 and is rare before age 50. The mean age of onset was about 65. Underlying benign disease is known to increase the eventual risk of malignancy. Examples include Barrett’s oesophagitis, and oesophageal cancer, gastritis and gastric cancer. Hereditary pancreatitis is a rare autosomal dominant disorder with a penetrance of about 80%. The clinical phenotype consists of involvement of siblings and multiple generations, early age of onset (generally <21 years old) and a course that resembles more common types of chronic pancreatitis.
NICE 2004Guidelines on the management of adults with dyspepsiaStill out for consultation
Numans et al 2001NetherlandsThis was a multicentre case series study of the diagnostic features of gastro-oesophageal malignancy. The usefulness of identified alarm symptoms in requesting gastroscopy was evaluated.861 patientsPatients who were investigated with first time gastroscopy between October 1986 and October 1988--Malignancy was found in 21 patients (2.4%). Five patients had oesophageal cancer.
Positive answers for the symptoms, weight loss (p<0.01) and dysphagia (p<0.01) together with negative answers on pain during the night (p<0.01) and heartburn, predicted malignancy in the study population with an area
Ojala et al 1982finlandRetrospective case series. Symptoms associated with carcinoma before medical attention was sought and prior to diagnosis was investigated.225 patientsAny inclusion and exclusion criteria applied were not mentioned--Patients with carcinoma of oesophagus or gastric cardia presented with dysphagia 93%; weight loss 46%; vomiting 33%; gastric pain 25%; thoracic pain 21%; anorexia 7%; haematemesis or melaena 6%; belching, hiccups or dyspepsia 4%; pharyngeal pain 4%; sensation of a lump 3%; anaemia 3%; cough, hoarseness 2% & others 9%.Histology
Shaheen and Ransohoff, 2002USAThe evidence linking Gastroesophageal reflux disease (GORD) and Barrett’s oesophagus to oesophageal carcinoma was examined. A MEDLINE search was performed to identify all English language reports about GORD, adenocarcinoma, and Barrett oesophagus from 1968 through 2001.Not mentionedStudies were of randomised controlled trials if available, case control data if trials were unavailable, and cohort-studies if case-control data were unavailable. Pertinent bibliographies were also reviewed to find reports not otherwise identified.--Cohort studies demonstrated that symptoms of GORD occurred monthly in almost 50% of US adults and weekly in almost 20%.
Three large case-control studies demonstrated a positive association between reflux symptoms and risk of adenocarcinoma of the oesophagus, with more prolonged and severe symptoms accentuating this risk. However, because of the low incidence of adenocarcinoma of the oesophagus and the ubiquity of reflux symptoms, the risk of cancer in any given individual with reflux symptoms was low.
Most studies on individuals with Barrett’s oesophagus reported a risk ratio of cancer that was 40 to 125 times higher than that of the general population. Estimates of the absolute risk of oesophageal adenocarcinoma varied widely from 0% to almost 3% per patient year. Recent larger studies and a meta analysis of these data suggested that a reasonable estimate was approximately 0.5% per-patient year.
Talley et al, 1998USAThis systematic review centred on evaluating the optimal management of patients with dyspepsia. A MEDLINE and Current Contents search was performed up to April 1997 using the MeSH term ‘dyspepsia’.36Studies reporting cancer rates using the findings on esophagogastro-duodenoscopy in patients with dyspepsia and in the general population.--Endoscopy was reported as consistently providing superior diagnostic accuracy in comparison with radiography.

Many of the studies indicated that dyspepsia was s symptom of cancer in approximately 2% of patients.
(The test performed was esophagogastro-duodenoscopy)
Tredaniel et al 1997A review and meta-analysis of undertaken to provide a quantitative estimate of the association between gastric cancer risk and tobacco smoking.40 studies.----All the cohort studies showed a significantly increased risk of gastric cancer of the order of 1.5 –2.5 for cigarette smokers. Evidence from case-control studies was less consistent. The results suggested a risk of stomach cancer among smokers of the order of 1.5–1.6 as compared to non-smokers.
Wilson et al 2000CanadaRetrospective case series. The incidence of signs and symptoms associated with oesophageal or gastric cancer were observed and reported. The objectives were to determine the symptoms experienced by patients with pancreatic cancer and the response by health professionals in providing supportive care in a large, tertiary centre99Patients diagnosed with pancreatic cancer--The most common symptoms were dysphagia (93%), weight loss (46%), vomiting (33%), gastric cancer (25%), thoracic pain (21%), anorexia (7%) and GI bleeding (9%). Bleeding and anaemia were found in the lower oesophagus tumours & gastric cardia. Infections, backache or pain in the lower abdomen occurred in 9% of patients.HistologyThere was no statistical evaluation of the results

From: Appendix A, Evidence Tables

Cover of Referral Guidelines for Suspected Cancer in Adults and Children
Referral Guidelines for Suspected Cancer in Adults and Children [Internet].
NICE Clinical Guidelines, No. 27.
Clinical Governance Research and Development Unit (CGRDU), Department of Health Sciences, University of Leicester.
Copyright © 2005, National Collaborating Centre for Primary Care.

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.