Article Text

Download PDFPDF

Oesophagus posters 442–464

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.



D.L. Morris, A. Ainsworth, C. Leckenby, S.M. Greenfield, P.B. McIntyre.

Background: The incidence of oesophageal cancer is increasing yet survival is poor and perioperative mortality remains high. Recent guidelines for referral and management issued by the department of health aim at earlier detection and improved effective appropriate management-but are they helpful?

Aims and Methods: A 2-year retrospective audit of oesophageal cancer referrals April 1999 to May 2001 in a district general hospital of catchment population ∼250,000. Comparison with guidelines for urgent referral (April 2000) and management (March 2001) of upper GI cancers. Patients identified with oesophageal cancer from PAS system and endoscopy database, and notes reviewed using proforma.

Results: 44 patients with oesophageal cancer identified. Mean annual incidence of 8.8/100,000/year. Median age 75 years. Histology found 57% adenocarcinoma, 23% squamous, 20% other type. Symptom analysis showed 70% presented with dysphagia, but significantly fewer with heartburn (20%) or reflux (9%) than referral guidelines. Some 37/44 (84%) fulfilled the symptom guidelines for urgent referral. Those who did not fulfil criteria presented with anaemia, GI bleeding and epigastric pain, without other alarm symptoms. Endoscopy was used in diagnosis in 43/44 patients. CT scanning was used for staging in 35/44 (79%), the remainder were deemed too frail for investigation. 2 patients had preoperative PET scans. EUS is unavailable. Surgical opinion was sought in 20/44 (45%) of which 9 (20%) had resections at a tertiary centre. 2 received neo-adjuvent chemotherapy. 1 patient was found to have inoperable disease at laparotomy. Some 20/44 (45%) were referred for oncological opinion. 3 (7%) received palliative chemotherapy and 8 (18%) radiotherapy at a designated centre. Endoscopic treatment was required in 26/44 (60%), 8 (18%) having metal stents. Outcome 6 months after the study period found only 10/44 (22%) alive, the median survival was 4 months from referral to death.

Conclusions: Symptom guidelines used alone would have missed 16% cancers. Management guidelines suggest a specialist team to identify those liable to benefit from treatment, but the overall outcome is likely to remain poor as comorbidity prevented treatment in many patients.


J.P. Cotton, M. Lopez, S. McLeod, J.A. Todd, D.A. Johnston, J.F. Dillon. Department of Digestive Diseases and Clinical Nutrition, Ninewells Hospital, Dundee DD1 9SY, UK

Introduction: GORD and Barrett's oesophagus have thought to be associated with greater deprivation, although recent data suggest that this may be changing. We aimed to examine the association between deprivation, GORD and it's complications..

Methods: A cohort of patients with GORD and Newly diagnosed Barrett's oesophagus were constructed from a database of patients with reflux symptoms. Deprivation index was obtained by utilising the post code area and sector to calculate the Carstairs Deprivation score (1 least deprived, 6 most deprived). Incidences and putative risk factors were examined.

Results: 658 patients were recruited with symptomatic GORD; when stratified for Carstairs Deprivation Category. In deprivation Category 1 (least deprived) 20% (14) had Barrett's oesophagus and 79% (275) had GORD, in deprivation category 6 (most deprived) 10% (22) had Barrett's oesophagus and 90% had GORD (Chi Squared for trend P=0.01). Symptom score, Acid suppression therapy exposure, Body Mass Index (BMI), Smoking (pack years), and Alcohol consumption (units per week) were further stratified for diagnosis and deprivation category. There was no difference in symptom score according to deprivation category or diagnosis (P>0.05). BMI, Smoking or Alcohol Consumption was equal for diagnosis and deprivation category. Patients with Barrett's oesophagus but not GORD in deprivation category one had a greater exposure and duration of therapy to proton pump inhibitors (PPI) compared with deprivation category 6 (Chi Squared for trend P=0.01) There was no difference in for Histamine 2 Receptor antagonists (H2RA).

Conclusions: Barrett's oesophagus seems to be associated with less deprivation and more proton pump inhibition, compared with GORD which is associated with more deprivation and less acid suppression therapy. This unusual finding may represent differing aetiologies for GORD and Barrett's oesophagus.


A. Chandra, A. Anggiansah, R. Moazzez, M.H. Arastu, W.J. Owen. Department of Surgery, Guy's & St Thomas' Hospital, London, UK

Introduction: Intraluminal Electrical Impedance (IEI) is a measure of resistance to current flow between two electrodes. The presence of a bolus in the oesophagus can be monitored with IEI. This preliminary study looks at ambulatory IEI (out-patient basis) in the investigation of reflux episodes (RE), comparing this to pH monitoring and manometry.

Patients and Methods: A preliminary sample of 10 patients with symptoms of gastro-oesophageal reflux disease prospectively underwent manometry. This was prior to an ambulatory study (for up to 24-hours). The combined catheter consisted of 4 pressure transducers placed at 0, 5, 10 and 23 cm with 2 impedance electrode pairs at 0 and 5 cm proximal to the pH sensor. The pH sensor was sited 5 cm proximal to the lower oesophageal sphincter. A pH-RE occurred where the distal oesophageal pH was less than 4, (>2 secs). A manometric RE (Man-RE) occurred where there was a distinct simultaneous increase in intra-oesophageal pressure (>10 mmHg) in 2/3 distal pressure sensors, with no evidence of peristalsis or upper oesophageal sphincter activity. An IEI-RE occurred where IEI values dropped by 50% from the baseline (>2 secs).

Results: Overall pH-RE had an associated IEI-RE on 310/538 (57.6%) occasions. IEI-RE were definitively acid (associated pH-RE) in 310/505 (61%) and non-acid in 105/505(21%). Gas RE were detectable using IEI. See table.

Abstract 445

Conclusions: Ambulant IEI has a good correlation with pH-monitored reflux events. IEI is a useful adjunct to pH monitoring in studying reflux, particularly in detecting gas and non-acid reflux events. Manometry is insensitive in detecting pH-RE, but provides information regarding gas, reflux and swallow events. This preliminary study highlights both the potential and the need for further validation and modification of IEI technique as an ambulatory technique.


B. Smith, J. Bagshaw, S.A. Riley. Department of Gastroenterology, Northern General Hospital, Sheffield, UK

Background: Many patients with oesophageal carcinoma present with advanced disease and relief of dysphagia is often the principal goal of therapy. SEMS have become a popular method of palliation but some have expressed concern that quality of life may be less good following SEMS insertion than following ablative methods of palliation. We have therefore undertaken a retrospective audit to identify factors that might predict a less favourable outcome.

Methods: Hospital records were reviewed in 100 consecutive patients in whom an oesophageal SEMS had been placed from June 1998 onwards. Pre-placement clinical characteristics were reviewed in relation to post-placement symptoms and survival.

Results: 69 men and 31 women underwent SEMS placement, ages ranged from 40 to 96 years and 61 patients were above 70 years. Tumours were predominantly distal; 67 in the lower third or at the gastro-oesophageal junction. 65 were adenocarcinomas. 88 Ultraflex and 12 Flamingo stents were placed, 50 straddled the gastro-oesophageal junction (GOJ). One patient died immediately following stent placement. A further 5 died within the first week and 66 within 6 months. Stent migration occurred in 2, tumour overgrowth in 4 and, despite dietary advice, food bolus obstruction in 23. 74 patients were able to take a normal “stent” diet, 13 a blended diet, 7 liquids only and 4 took little because of their poor general condition. Pain was reported in 19 before and 53 following SEMS placement. Regurgitation and vomiting were reported by 37 patients following SEMS. Relief of dysphagia, post stent pain and survival appeared independent of stent position. Vomiting, however, was seen more often when the stent straddled the GOJ (48%) than when in the lower oesophagus but not across the junction (36%) and least when in the mid oesophagus (16%).

Conclusions: SEMS provide reasonable relief of dysphagia in most patients with malignant oesophageal disease. However, post SEMS pain is common and often severe and vomiting is frequent when the stent is placed in the lower oesophagus particularly when it straddles the GOJ.


S.J. Dwerryhouse, A.D. Hollowood, C.P Armstrong. Department of Surgery, Frenchay Hospital, Bristol, UK

Introduction: Barrett's oesophagus can be a difficult subject to understand even for those with a special interest in it. For a newly diagnosed patient, getting to grips with the causes, risks and management of Barrett's oesophagus can be daunting. It is, therefore, important that patients have access to all the relevant information they need to help them understand the implications of their condition.

Aims: The aim of this study was to assess the adequacy of information on the controversies surrounding Barrett's oesophagus available on the internet.

Methods: Using the search term “Barrett's oesophagus”, an internet search was carried out using 3 commonly used search engines (HotBot, Yahoo and Altavista). The top 50 sites identified by each search were assessed for their relevance to patients and their discussion of cancer risk, surveillance and treatment.

Results: Only 98 of the 150 sites visited related specifically to Barrett's oesophagus and of those designed for patients, 60% were primarily concerned with oesophageal cancer. 33% (33/98) of sites mentioned surveillance of Barrett's and 38% (37/98) discussed treatment options for Barrett's oesophagus.

Conclusions: The information about Barrett's oesophagus on the internet provides an inaccurate perspective on the controversies related to this condition. In particular, the information is potentially misleading with regard to cancer risk, value of surveillance and treatment. As specialists with an interest in Barrett's Oesophagus, we should take a lead in ensuring accurate, balanced information is available to our patients from all sources including the internet.


D. Durai, E.D. Srivastava, M.C. Allison. Royal Gwent Hospital, Newport, Wales, UK

Background: Forthcoming BSG guidelines favour endoscopic surveillance of Barrett's oesophagus, which could lead to a major and increasing hospital workload and prove inconvenient for patients.

Aims: (i) to examine numbers of new patients entering our surveillance programme each year; (ii) to find how many Barrett's patients should and should not have entered the surveillance programme; and (iii) to review all Barrett's associated high-grade dysplasia and adenocarcinoma seen during 1995–2000 and examine the impact of surveillance.

Criteria for surveillance: Patients with Barrett's segment of 5cm or more, aged up to 70 years and without major co-morbidities, who could potentially withstand surgery, are offered annual endoscopic surveillance.

Results: There were 374 patients with endoscopic diagnosis of Barrett's oesophagus aged between 27 and 97 years. At January 1995 there were only 7 patients under surveillance. Between 1995 and 2000 99 further patients entered the programme. During the 6 years 24 Barrett's associated adenocarcinomas were diagnosed, of which 22 were symptomatic and found on index endoscopy and 3 had been detected in the surveillance group (one of which was an interval cancer). One 47-year-old man had potentially curative surgery of an asymptomatic surveillance cancer in 1999 and is well, but the other two were unfit for surgery. One 78-year-old woman with high-grade dysplasia was referred for photodynamic therapy. During the six years 29 other patients were lost to the surveillance programme because they had died or were discharged due to age/comorbidity or lost to follow-up. This still left 74 patients under surveillance, but as a result of this audit 21 further patients could be rejected from the programme: 19 had Barrett's <5cm and 2 were aged over 70 years.

Conclusions: Endoscopic surveillance presents an increasing burden year by year as more incident cases are diagnosed. The yield of surveillance in relation to our commonly diagnosed Barrett's associated cancers is disappointing. Regular audits of this kind are needed to ensure that surveillance is targeted towards those patients most likely to benefit.


N. Johnston1, P.E. Ross1, P.W. Dettmar2, M. Panetti3, J.A. Koufman3.

Background: Gastro-oesophageal reflux disease (GORD) is a common condition that has been extensively studied. It is now recognised that patients with laryngeal disease and voice disorders may also suffer from reflux of gastric contents into the upper aero-digestive tract (laryngopharyngeal reflux - LPR). Cellular defence mechanisms are important in protecting the mucosa from the damaging effects of gastric refluxate. It has been suggested that carbonic anhydrase (CA) enzymes may be important in this regard, generating HCO3- that could provide an important buffering mechanism in the oesophageal mucosa.

Aims: To investigate the pattern of expression of CA iso-enzymes in oesophageal and laryngeal mucosal biopsy specimens from patients with reflux disease.

Methods: The localisation and expression of CA iso-enzymes were determined in oesophageal and laryngeal mucosal biopsy specimens using standard immunofluorescent (IF) staining techniques combined with Western blot analysis.

Results: Oesophageal samples taken from patients with GORD demonstrate an increased expression of CA I & III in inflamed squamous epithelium, together with evidence that the enzymes were more widely expressed throughout the epithelium. Further increases in the level of expression of both CA iso-enzymes were detected in Barrett's mucosa and adenocarcinoma although in Barrett's mucosa IF studies revealed that the distribution of the immunoreactive enzyme was patchy. In contrast, laryngeal squamous epithelium did not demonstrate any change in expression of CA I in the presence of LPR but there was a notable decrease in CA III immunoreactivity.

Conclusion: The findings in patients with GORD suggest that expression of CA enzymes is modified and may be an important protective mechanism as it may increase the buffering capacity within the cells. The laryngeal mucosa does not show the same pattern of response, furthermore loss of expression of the CA III iso-enzyme may increase the sensitivity of this epithelium to damage associated with reflux disease.


K. Lauritsen1, on behalf of the Metropole Study Group, O. Junghard2, S. Eklund2 (introduced by P. Richardson). 1Odense Hospital, Odense, Denmark; 2AstraZeneca R&D Mölndal, Mölndal, Sweden

Aim: This study was conducted to compare the standard maintenance dose of esomeprazole 20mg once daily (od) with the maintenance dose of lansoprazole 15mg od for the prevention of recurrence of reflux oesophagitis (RO).

Methods: 1391 patients with endoscopically verified RO (LA classification) were enrolled in this randomised, double-blind, parallel-group, 14 country multi-centre trial. During the initial, healing phase of the study, all patients received 4–8 weeks' open treatment with esomeprazole 40mg. 1236 healed (identified by endoscopy at 4 and 8 weeks) and symptom-free (i.e. no heartburn or acid regurgitation during the last 7 days of treatment, investigator assessment) patients were randomised to 6 months maintenance treatment with esomeprazole 20mg od or lansoprazole 15mg od. Time to relapse (relapse of RO and/or discontinuation due to symptom recurrence) was analysed using a log-rank test.

Results: Esomeprazole maintained a significantly higher proportion of patients in remission than lansoprazole over the 6-month course of treatment (p<0.0001, ITT analysis-log rank test). After 6 months' treatment, 83% of esomeprazole recipients were in remission compared with 74% of lansoprazole recipients (Life Table estimates). Higher remission rates were consistently achieved with esomeprazole irrespective of baseline LA grade assessed at the beginning of the healing phase. Significantly more patients were free from heartburn in the esomeprazole group compared to the lansoprazole group at 1, 3 and 6 months (p<0.05, Chi-square test). Significant differences at 6 months between esomeprazole 20 mg od and lansoprazole 15 mg od were also observed for control of epigastric pain and acid regurgitation (p<0.05 and p<0.001, respectively, Chi-square test).

Conclusion: Esomeprazole 20 mg od is more effective than lansoprazole 15 mg od for maintaining healed RO and controlling accompanying GORD symptoms.


S.E. Jackson, J. Weinman, A. Chandra, A. Anggiansah, W.J. Owen. Department of Surgery, Guy's & St Thomas' Hospital, London, UK

Introduction: In a number of patients with symptoms of gastro-oesophageal reflux disease (GORD), investigations find no organic cause and the literature suggests that psychological factors may play a role. The aim of the study was to prospectively evaluate psychological profiles of symptomatic patients attending for investigation.

Patients and Methods: Of the 79 patients approached, 71 agreed to complete the questionnaire. This contained scales such as the revised Illness Perception Questionnaire (IPQ), which included a measure of perceived illness coherence, and the Positive and Negative Affect Schedule (PANAS). All patients underwent manometry and 24-hour ambulatory pH monitoring. GORD was defined where total percentage reflux time was greater than 5.78% (reflux occurred where the distal oesophageal pH was less than four). Analysis was performed on the questionnaires results, diagnosis after investigation, manometry and pH result. Spearman's rank correlation was used to examine correlations in the whole sample. The Mann-Whitney test for non-parametric non-dependent data was used to examine differences between groups.

Results: Illness coherence as measured by the IPQ was inversely correlated with total percentage reflux time (R= -0.33, p=0.007). There was a significant difference (p=0.008) in the illness coherence mean score between GORD (n=37, mean=8.26) and non-GORD (n=34, mean=10.23) groups. Mean values of positive affect as measured by the PANAS differed significantly (p=0.044) between subjects defined manometrically as having abnormal motility (n=46, mean=36.9) and those with normal motility (n=20, mean=33.5)

Conclusions: Patients who have a low positive affect tend to have normal motility when studied with manometry. A greater score of illness coherence implies a reduced understanding of and mystification by the illness. Those who had a greater illness coherence score did not have GORD. The presence of psychological factors as determined by this study seems to distinguish a subset of patients with no organic disease.


R. Moazzez1, A. Anggiansah2, D. Bartlett1, A. Chandra2, W.J. Owen2. 2Department of Surgery and 1Division of Conservative Dentistry, GKT

Introduction: Dental erosion and gastro-oesophageal reflux disease (GORD) are reported to be associated. However, the role of saliva in both conditions is unclear. This study aimed to investigate the relationship between saliva, dental erosion and reflux symptoms.

Patients and Methods: 104 patients attending the oesophageal laboratory complaining of symptoms of GORD were studied. A detailed history was obtained regarding the patients' diet, their oesophageal (heartburn, regurgitation, dysphagia and retrosternal pain) and extra-oesophageal (hoarseness, globus and chronic cough) symptoms. Tooth wear was assessed by grading each tooth surface using 4 pathological scores 2–5 with increasing severity. Score 2 represents dentine exposure whilst scores 4 and 5 represent severe wear either involving the pulp or needing treatment. Stimulated salivary flow rate and buffering capacity of these patients were assessed using a standard protocol. The patients subsequently had standard manometry and 24-hour pH tests.

Results: See table. The severity of tooth wear varied between patients. The mean (sd) tooth wear score of 3 for all sites was 4.3% (13.4%) and for scores 4 and 5 grouped together was 2.5% (12.5%). There were no associations between other symptoms of reflux or presence of GORD to tooth wear and no relationship between salivary parameters and tooth wear.

Abstract 452

Conclusions: Regurgitation was associated with increased tooth wear. Hoarseness was associated with decreased salivary flow rate. In this group of patients saliva was not associated with tooth wear or GORD.


J. Dearden, G. Hellawell, J. Pilling, K. Besherdas, N. van Someren. Chase Farm Hospital, The Ridgeway, Enfield, Middlesex, UK

Introduction: Balloon tamponade using Sengstaken-Blakemore (SB) tubes for oesophageal varices has been in use for almost 50 years. Despite the development of endoscopic techniques, SB tubes still have an important role in the management of variceal bleeding. Standard teaching recommends the use of a cooled SB tube that increases tube stiffness and aids insertion. We surveyed current clinical practice of SB tube use in our region and also assessed whether cooling SB tubes alters the stiffness of the tubes.

Methods: A telephone questionnaire was conducted of gastroenterology registrars and ITU departments in the North Thames region. The current clinical practice and the basis for this practice were determined in each case. The stiffness of SB tubes was measured at 0°C and 23°C by calculating the slope of the plot of load (kg) vs. strain (tube stretch/initial tube length). The time for tube warming from 0°C to 23°C when in stationary air and when in contact with skin was also measured.

Results: Fifty registrars were contacted and twenty ITU departments were surveyed. All ITU departments involved the gastroenterologists in the management of acute variceal bleeds. Eight registrars had never placed an SB tube. The majority of the remainder (95%) used a cooled SB tube. All of the registrars based this practice upon their clinical teaching, and 75% of these registrars thought cooling aided the insertion of the tube. There was no difference in the stiffness of the tubes at 0°C and 23°C. The time for SB tube warming from 0°C to 23°C was 120 seconds in stationary air, and 20 seconds when in contact with skin.

Conclusion: The current clinical practice of trainees for SB tube insertion is to cool the tubes in the belief that this `standard' practice aids tube insertion. We found no change in SB tube stiffness even after cooling to temperatures that would not be achieved during routine insertion. Furthermore, the rapid rise in tube temperature means that tubes approach room temperature by the time they reach the bedside. In the present era of evidence-based medicine the current dogma that SB tubes should be cooled must be discarded.


C.J. Kelty1, R. Ackroyd1, N.J. Brown1, S.B. Brown3, T.J. Stephenson2, M.W.R. Reed1 (introduced by A.G. Johnson). 1Academic Unit of Surgical Oncology, and 2Histopathology, Royal Hallamshire Hospital, Sheffield, UK; 3Centre for Photobiology and Photodynamic Therapy, University of Leeds, Leeds, UK

Introduction: Adenocarcinoma of the oesophagus is increasing in incidence more rapidly than any other cancer in the Western World. The major risk factor is Barrett's oesophagus (BO), an acquired condition where the normal squamous lining is replaced by columnar epithelium showing intestinal metaplasia. This metaplastic change confers a lifetime risk for developing adenocarcinoma of 10–15%. ALA-PDT is effective in the treatment of BO, but an optimum dosage regimen has yet to be established. Therefore the aim of this study is to determine the optimum dose and timing of ALA administration in PDT for BO.

Materials and Methods: Twenty-five patients with biopsy proven BO (median length 4cm, range 2–15cm) were randomised into 5 groups (n=5) and received: 30 or 60mg/kg ALA at 4 hours, 30 or 60mg/kg ALA at 6 hours, or 30mg/kg at 6 hours and 4 hours before light activation. All patients underwent laser endoscopy under sedation using a balloon applicator and 635nm light at 68mW/cm2, with a total fluence of 85J/cm2. Endoscopy with quadrantic biopsies was repeated 4 weeks later.

Results: All patients showed a macroscopic reduction in the length of BO, with biopsy proven squamous re-epithelialisation. This was greatest in the fractionated and 30mg/kg groups (median 60%, range 25–100%). There was no significant difference between groups (median 60%, range 20–100% across all groups). Side effects were minimal.

Conclusions: Low dose ALA-induced PDT with red light appears to be an effective protocol for safe and effective ablation of BO. It seems appropriate to use a lower dose to reduce cost, improve safety and minimise potential side effects.


M. van Blankenstein1, C.P.J. Caygill2, B.J. Johnston3. 1Gastroenterology Dept, Rotterdam U.H.; 2UK National Barrett's Oesophagus Registry, Royal Free Hospital, London NW3 2PF; 3Lady Sobell Gastrointestinal Unit, Wexham Park Hospital, Slough SL2 4HL, UK

Background: In the absence of population studies there are no reliable data on the prevalence of BO in the general population. However, the prevalence of BO found in a large endoscoped population should provide information on the pattern of distribution by age and sex.

Objective: To establish the prevalence of BO by age and sex over a 15 year period in a district general hospital.

Methods: Prevalence was calculated from all histologically proven cases of BO between the ages of 20–89, identified between 1982–96 and the number of first endoscopies, all stratified by age and sex.

Results: 491 cases of BO (316 in males, 175 in females) were identified in 21,899 endoscopies (10,939 in males, 10,960 in females). Prevalence rose incrementally from age 20–29, from 0.16% in males and 0% in females to a maximum at age 70–79 of 4.89% in males and 3.75% in females, prevalence declined in both sexes at age 80–89 to 3.21% and 2.44% respectively (see table). Binary logistic regression shows that the prevalence of BO in men was double that of females, O.R. 2.01, (95% C.I. 1.67–2.43). Fitted curves showed a ten year shift in prevalence between males and females.

Abstract 455

Discussion: This study shows that the prevalence of BO rose steeply with age in both sexes. However, in females this rise was far slower between the ages of 20–59 than in males, reflected in a 10 year delay in the onset of BO in females. This delay probably accounted for the 2:1 male:female ratio in the prevalence of BO These results suggest that pre-menopausal females are to some degree protected against the development of BO.

Conclusion: The prevalence of BO in a population is strongly influenced by its age and sex composition.


K.D. Bardhan, C. Royston, P.J. Willemse, S.J. Haggie, R.B. Jones, J.C. Cooper, L. Harvey, T. Marshall, D.N. Slater. Rotherham General Hospitals NHS Trust, Moorgate Road, Rotherham S60 2UD, UK

Introduction and Aim: We present our 20 year experience of Barrett's oesophagus identified from the open access endoscopy programme 1977–1996 and followed up in our DGH until 31.12.00. Patients (pts), methods Barrett's was diagnosed by histology or visually if ≥5cm length. Pts were treated with H2RA or PPI and followed by endoscopy and biopsy or by clinical means alone (including telephone survey) if elderly or unwell.

Results: see table. 1. Incidence 4.3% of reflux pts have Barrett's and the prevalence of both is rising. 2. Demography Barrett's pts are a decade older than reflux pts (mean age at diagnosis 62y vs 52y) and both have a slight male preponderance (58%, 55%). 3. Complications Presentation with haemorrhage and/or anaemia is more common in Barrett's (20% vs 5%) and oesophageal stricture is seen more often this group (11% vs 2%). 4. Mortality During a mean follow-up of 7.5 years (range 0–21y) 123/368 (33%) died, mean age 78y, 24 from oesophageal adenocarcinoma (OAC), 18 from other tumours and the remainder mainly from cardio-respiratory causes. 5. Risk of malignancy 5 presented with, and 5 developed OAC within a year (and are excluded as incident cancers). 20 of the remaining 358 (5.6%) developed OAC, mean follow-up 6.7y, range 1.1–14y i.e. 1 tumour per 136 patient years, at a mean age of 72y. 12 occurred during endoscopic follow-up (4 were cured after resection of whom 1 died 8 years later of an unrelated cause) and the other 8 were patients whose general condition precluded serial endoscopy.

Abstract 456

Discussion and Conclusion: 1. Barrett's affects an older reflux population and is associated with more complications. 2. One in twenty develop an OAC, generally lethal, but five more die from other causes. 3. As the prevalence of Barrett's continues to rise, the population risk of OAC is likely to increase.


E. Pohler, P.E. Ross, J.F. Dillon.

Ninewells Hospital and Medical School, University of Dundee, Dundee DD1 4HN, UK

Background: Oesophageal cancers have a poor prognosis; treatment of these cancers by chemotherapy, radiotherapy and surgery improves this in only a small minority. Mutations in the p53 gene are common in oesophageal cancer and are a poor prognostic indicator. The introduction of a wild type copy of p53 may, therefore, provide a novel treatment for oesophageal cancers. In this study we aim to determine the feasibility of introduction of wild type p53 into human oesophageal tissue.

Methods: Human normal oesophageal pinch biopsies, obtained at endoscopy were transfected with a human wild type p53 using liposomes as a vector; control biopsies were transfected with a control plasmid lacking the p53 gene. After transfection the biopsies were set up in short term organ culture. Samples were subsequently analyzed by semi-quantitative reverse transcriptase polymerase chain reaction (RT-PCR) and/or Western blotting to assess the expression of p53 and also its downstream transcriptional target p21WAF1.

Results: RT-PCR demonstrated that 11.1% of biopsies showed an accumulation of p53 mRNA following transfection with the p53 gene but not with the control construct. In other biopsies (44.4%), p53 transcripts were detected both in the control and in the p53-transfected samples. However, by comparison to the level of expression of β-actin, used as an internal standard, p53 expression levels were elevated in these latter samples. Similarly, an elevation in the levels of p21WAF1 mRNA was shown to occur in biopsies that had been transfected with the p53 gene as compared to the control. An increase in p53 expression was not detected by Western blotting, although p21WAF1 protein was readily detectable in biopsies transfected with the p53 gene, at levels significantly higher than those seen in the controls.

Conclusions: We have demonstrated that it is possible to introduce wild type p53 into human oesophageal tissue at sufficient dose so that this gene is expressed and its product causes activation of p21WAF1, a gene downstream in the pathway to cell cycle arrest. This has therapeutic potential.


B. Thjodleifsson1, N.M. Miller2, J.E. Salter2, K.D. Bardhan3. 1University Hospital, Reykjavik, Iceland; 2Eisai Ltd, London, UK; 3Rotherham General Hospital, Rotherham, UK

Background: Treatment with proton-pump inhibitors increases serum gastrin levels, but there is little evidence from prospective randomised trials about the effects of long-term treatment.

Objectives: The primary objective was to assess efficacy in preventing GORD relapse. The secondary objective reported here was to assess the effect of 5 years' treatment with rabeprazole or omeprazole on serum gastrin concentrations.

Methods: 243 patients were randomised to double-blind treatment with rabeprazole (10 mg or 20 mg) or omeprazole (20 mg) once daily for up to 5 years. Serum gastrin concentrations were measured during the study; treatment effects were investigated in an ANOVA model of the log-transformed area under the gastrin concentration–time curve (AUC).

Results: Mean serum gastrin concentrations are shown in the graph. The data had a highly skewed distribution, particularly in the omeprazole group. The differences among treatments in the AUC were not statistically significant.

Conclusions: Increases in serum gastrin concentrations during long-term rabeprazole or omeprazole treatment were modest in most patients, but were substantial in some patients taking omeprazole.


J. Darragh1, P.E. Ross1, N.M. Kernohan1, P.W. Dettmar2. 1Ninewells Hospital and Medical School, University of Dundee, UK; 2Reckitt Benckiser Healthcare Ltd, Hull, UK

Pre-malignant Barrett's oesophagus and reflux oesophagitis are associated with the reflux of stomach contents into the lower oesophagus. Although the role of acid and pepsin as damaging agents within this refluxate has been well established, attention is now becoming focused on the potential effect of bile acids. This study aims to determine the bile acid composition of refluxate in patients with GORD, so that an in vitro model can be set up to determine the effect bile acids may have, in inducing cellular injury to oesophageal epithelium.

Methods: Gastric juice samples were obtained from 170 patients with gastro-oesophageal reflux disease (112 with oesophagitis and 58 Barrett's oesophagus). The bile acid composition was determined by gas chromatographic analysis.

Results: Conjugated bile acids were detected in 95% of samples with concentrations ranging from 1.2 umol/l to 6.4 mmol/l, however only 11 samples contained concentrations exceeding 1 mmol/l. Mean concentrations of the primary bile acids, cholate and chenodeoxycholate were calculated to be 119 μmol/l and 112.1 μmol/l respectively, however the mean concentration of the secondary bile acid deoxycholic acid was 63.2 μmol/l. In comparison a much lower mean concentration of lithocholic acid, a particularly toxic bile acid, was found (mean concentration = 17 μmol/l). Sixty four percent of patients had unconjugated bile acids present in their gastric juice with only 18 patients having concentrations in excess of 20 μmol/l.

Conclusions: Bile acids are normally present in the gastric juice of patients with both Barrett's oesophagus and oesophagitis. For both groups conjugated bile acids are the predominant species, with primary bile acids being in excess of secondary. The presence of toxic bile acids within gastric juice suggests that they may have a role in the pathogenesis of GORD. Furthermore the present data have indicated appropriate concentrations of individual bile acids for investigating the effect of bile acids on apoptotic and stress responses in established oesophageal cells lines.

This research was funded in part by Reckitt Benckiser Healthcare (UK) Ltd, Hull, UK.


M.J. Gibbons, R.G.P. Watson, B.T. Johnston. Department of Medicine, Royal Victoria Hospital, Belfast, Northern Ireland

Gastro-oesophageal reflux disease (GORD) is a cause of significant morbidity in scleroderma (SSc). Proton pump inhibitors (PPI) are frequently used yet information is limited regarding their effect on oesophageal acid exposure in SSc.

The aim of this study was to examine the nature of pathological oesophageal acid exposure in SSc and, in addition, to evaluate PPI requirements.

SSc patients with reflux symptoms and who fulfilled ACR diagnostic criteria were recruited. Subjects underwent 24-hour ambulatory oesophageal pH monitoring. The following pH parameters were recorded: percentage of study time at pH<4 for the total, erect and supine periods, the number of long reflux events (>5 minutes) and the duration of the longest reflux event. Those subjects with exposure times above previously determined normal values were commenced on PPI therapy. Treatment was sequentially increased, with repeat pH monitoring after each increment, until a treatment endpoint of normalisation of acid exposure was reached.

Twenty-two subjects were recruited. Fourteen (64%) had an abnormal oesophageal pH profile. The median % of the total study time at pH<4 was 11.3 (normal <4.9), for the erect period 11.9% (normal <6.2) and for the supine period 4.1% (normal <1.2). The median number of long reflux events was 7 with a median longest event of 23 minutes. The longest reflux event occurred in the erect period in 6 (43%) and in the supine period in 8. Seven subjects (50%) had acid exposure normalised after 1 dose-study sequence on once daily PPI therapy. Six (43%) required 2 dose-study sequences and normalised on twice daily PPI therapy. One patient required 3 dose-study sequences and normalised on high dose twice daily therapy.

The majority of SSc patients with reflux symptoms have pathological acid exposure. Abnormal exposure is common in both the erect and supine positions. Modest PPI doses were required in most patients to normalise acid exposure. Refractory acid reflux was not seen.


R. Shah, A. Lim, S. Jones. Epsom General Hospital, Epsom, Surrey, UK

Expanding metal oesophageal stents are widely used for palliation of oesophageal cancer. These stents are designed to be inserted by endoscopy with fluoroscopic guidance. This need for X-ray can be limiting for endoscopists in centres with limited or no fluoroscopic services. Even with fluoroscopy, external marking of the position of the tumour can be inaccurate and internal marking of the tumour can be time consuming.

We have used a modified deployment technique, guided by direct endoscopic visualization without the need for fluoroscopy. Microvasive (Boston Scientific) stents were used. The proximal position of the stent is marked with a white proprietary marker paint (Tippex). A guide wire is inserted through the endoscope which is then removed. The stent is inserted over the wire, the endoscope is re-inserted and the stent is then deployed by direct visualization of the proximal end of the stent and the proximal end of the stricture.

We inserted 47 stents in 43 patients with obstructive malignant oesophageal strictures; 27 male, mean age 79 years (range 44–97). Histologically there were 22 adenocarcinomas, 7 squamous carcinomas, 7 undifferentiated cancers and 1 severe dysplasia. Covered and uncovered stents of 10, 15 and 17cm were used.

In 4 patients with almost complete stenosis, the stenosis was first dilated but the stents were inserted by the method described above. There were no cases of malposition and there were no immediate complications of stent insertion.

In our experience, stent insertion by direct endoscopic visualization was technically simple and successful. We found it simpler and more accurate than techniques that involve internal or external marking of the tumour followed by fluoroscopy. This technique is of particular use in centres with limited or no fluoroscopic services.


M.D. Gough, R.A. Ackroyd, N. Bird, A.W. Majeed (introduced by A.G. Johnson). Academic Surgical Unit, University of Sheffield, K Floor, Royal Hallamshire Hospital, Glossop Rd, Sheffield S10 2JF, UK

Aims: To develop a database for patients with Barrett's Oesophagus, which would aid clinical follow up, monitoring of disease progression and facilitate research studies.

Methods: We have developed a Barrett's Oesophagus Database for Sheffield patients (BODS). This has been produced in Microsoft Access™ and employs a drop down menu system to facilitate data entry and allows for the recording of patient demographics, relevant drug and past medical history, endoscopy summaries, 24 hour pH and manometry studies, histology, treatment, and follow up.

Results: Over 300 patients have been entered into BODS, we have found it to be user friendly and versatile e.g., the operator can choose an endoscopic surveillance interval for each individual patient, complete reports of a patients record with endoscopic findings and histology results can be easily produced. The query tool will automatically calculate future endoscopy intervals and dates thereby reducing clerical errors.

Conclusions: BODS is proving to be an effective tool in both the clinical and research settings and a rival to other systems such as Endoscribe™. In the future, it could be used in other hospitals and modifications could allow flexibility to include other conditions.

A multimedia presentation will demonstrate data entry, analysis, and report production.


N.F. Jamieson, A.J. MacRobert, S.G. Bown, L.B. Lovat. National Medical Laser Centre, Department of Surgery, Royal Free & University College School of Medicine, London, UK

Background: Oesophagectomy carries a significant risk of major morbidity or mortality. Photodynamic therapy (PDT) shows potential as a minimally-invasive treatment to ablate high-grade dysplasia in Barrett's oesophagus. The challenge is to completely but selectively remove mucosa, as damage to underlying tissue may result in oesophageal stricturing. Meso-tetrahydroxyphenylchlorin (mTHPC - QuantaNova) is a potent photosensitiser used for oesophageal PDT. With the standard regime, mTHPC may induce oesophageal muscle injury. Variation of treatment parameters may allow enhancement of mucosal selectivity.

Methods:Fluorescence studies – Wistar rats given mTHPC (0.1mg/kg) intravenously were killed at intervals. Frozen sections of colon were examined by fluorescence microscopy and the level of fluorescence quantified in the mucosa and muscle layers. PDT studies - Intravenous mTHPC (0.1, 0.05 or 0.025mg/kg) was administered. After an interval (drug-light interval) of 4, 24 hours or 4 days, normal rat colon was treated at laparotomy using a diffuser fibre and 652nm light. The light dose was 3, 9 or 27J at a power of 10, 30 or 100mW. After three days the extent of mucosal and muscle necrosis was graded histologically and a selectivity ratio derived (mucosal score/muscle score).

Results: Fluorescence in colonic mucosa peaked at 24 hours while muscle fluorescence rose until 5 days. Mucosal selectivity was greatest at 24 hours suggesting this was the optimum time for PDT. Histologically, after PDT some mucosal selectivity was seen in all animals. This was greater at 24 hours than other times (p<0.04). PDT at low power caused significantly more mucosal and muscle necrosis than at high power but selectivity ratio was not significantly better (p=0.09). Selectivity was not improved by lowering drug dose (p=0.07).

Conclusions: Mucosal selectivity was best achieved in this model using a 24-hour drug-light interval. Further work is needed to determine the optimum parameters for clinical use.


U. Dave, M.M. Walker, H. Ebrahim, E. Townsend1, M. Burke1, M.R. Thursz. Faculty of Medicine and Histopathology, Imperial College, St Mary's Campus; 1Harefield Hospital, UK

Barrett's oesophagus is a metaplastic change in the epithelium that is strongly associated with oesophageal cancer. Laminin, a component of the epithelial basement membrane, plays an important role in the regulation of cellular differentiation. There is limited information on the distribution of laminin chains in the upper gastrointestinal tract and furthermore the antibodies used in earlier studies have been ascribed incorrect specificity. The distribution of laminin chains in Barrett's oesophagus has not been previously described.

Method: The distribution of laminin isoforms was documented in the normal upper gastrointestinal tract and in Barrett's oesophagus from endoscopic biopsies and surgical oesophagectomy specimens. Immunohistochemistry was performed on frozen sections using the following primary antibodies (antigen indicated in brackets): BM2 (α3 chain), 545 (β1), S5F11 (β2), 6F12 (β3), B8B11 (γ2), MAb1920 (γ1), MAb1922 (α2), MAb1924 (α5), MAb2041 (β1), μαb19562 (γ2) MAb EB7 (α1).

Results: The α1 laminin chain is expressed in submucosal glands and duct basement membranes in squamous oesophagus and Brunner's glands in the duodenum but is not expressed in Barrett's oesophagus or elsewhere in the upper GI tract. The α2 chain is expressed in the cytoplasm of the basal cells in surface epithelium and basement membrane of the glandular epithelium. The β2 laminin chain is present in a continuous fashion in squamous epithelial basement membrane but in Barrett's epithelium, cardia, gastric body and duodenum it is discontinuous and not expressed at the surface and only expressed in the glandular basement membrane. Constituents of laminin 5 (α3, β3, γ2 chains) are expressed in a continuous fashion in squamous epithelial basement membrane but in cardia, gastric body, duodenum and Barrett's oesophagus are expressed in surface epithelial basement membrane with a sharp decline in glandular and gastric pit basement membranes.

Discussion: The site specific distribution of α1and β2 laminin chains may have an important role in Barrett's metaplasia in the oesophagus.