Article Text

Download PDFPDF

Oesophageal free papers 012–025

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.


Z.H. Khan1, E. Simpson2, A.T. Cole1, I. Macdonald2, D. Pye2, A. Austin1, J.G. Freeman1. 1Department of Gastroenterology, Derby City Hospital, Derby, UK; 2University Department of Physiology and Medical Physics, Queens Medical Centre, Nottingham, UK

Aim: To investigate the potential for using thalidomide as an anti-cachectic agent in patients with advanced oesophageal cancer by studying its effect on body composition and weight.

Methods: 11 patients with non-obstructing and in-operable oesophageal cancer were included in the study.

Study protocol: Patients were established on an isocaloric diet over a 10-day period. Body weight, body composition studies with DEXA scanning, REE (resting energy expenditure) and serum levels of insulin, thyroxine, catecholamines and cortisol were measured at the entry and then after two weeks on diet alone. Patients were then started on thalidomide for 2 weeks and the measurements were repeated. Quality of life (QOL) was similarly measured as a secondary end point.

Results: Ten patients completed the study protocol. The average caloric intake remained the same throughout the study period in all these patients. 9/10 (95% CI 0.60, 0.98) lost weight on diet alone. The mean gain on thalidomide in the following two weeks was 1.29 kg (median 1.25kg). A similar trend was shown in lean body mass. There were missing data for one patient, so nine were analysed. 8/9 (95% CI 0.57, 0.98) initially lost mass on diet alone. The mean gain on thalidomide in the following two weeks was 1.75 kg (median 1.33 kg). The mean change in REE was 1.75 (95% CI –0.42, 3.91) on thalidomide. Amongst hormonal assay, changes in catecholamines approached statistical significance. The mean change in catecholamines on thalidomide was −0.71 (95% CI −1.60, 0.02).

Conclusions: In this sequential study of patients with progressive inoperable cancer, thalidomide treatment appeared to reverse loss of weight and lean body mass over the two week trial period. However to establish its role as an anti-cachectic treatment a full placebo-controlled trial is warranted.


B. Thjodleifsson1, A. Morocutti2, K.D. Bardhan3. 1University Hospital, Reykjavik, Iceland; 2Eisai Ltd, London, UK; 3Rotherham General Hospital, Rotherham, UK

Background: Many studies have found proton-pump inhibitors to be effective and safe in preventing relapse of gastro-oesophageal reflux disease (GORD) over a period of several months to a year. There is, however, little evidence from randomised trials about their long-term safety and efficacy.

Objectives: To compare the efficacy and safety of rabeprazole and omeprazole in the prevention of relapse in patients with healed gastro-oesophageal reflux disease during 5 years of treatment.

Methods: Patients were eligible for the study if they had previously been diagnosed with GORD, which had healed as shown by endoscopy. Patients received randomised, double-blind treatment with rabeprazole (10 mg or 20 mg) or omeprazole (20 mg) once daily for up to 5 years. The main outcome measure was endoscopically confirmed GORD relapse (Hetzel–Dent score = 2). Endoscopy was done after 13, 26, and 52 weeks, and yearly thereafter, or if symptoms suggested GORD relapse.

Results: 243 patients entered the study, of whom 123 completed all 5 years of treatment. Relapse rates were 9/78 (11.5%) in the 20 mg rabeprazole group, 8/82 (9.9%) in the 10 mg rabeprazole group, and 11/83 (13.3%) in the 20mg omeprazole group. The differences in relapse rates were not statistically significant. All three treatments were safe and well tolerated.

Conclusions: Rabeprazole at a daily dose of 10 mg is as effective as rabeprazole 20 mg or omeprazole 20 mg in preventing relapse of GORD over 5 years of treatment.


Y.H. Oo, D. Foy, H. Allen, P.J. Winwood. Department of Gastroenterology, Royal Bournemouth Hospital, Bournemouth, UK

Background: Oesophageal disease is a well-recognized cause of non-cardiac chest pain (NCCP). The role of Oesophageal Manometry (OM) and pH studies remain unclear, particularly in changing outcome.

Aim: To assess whether Oesophageal Manometry and pH studies affect the management and outcome of NCCP patients in a district hospital.

Methods: Retrospective study of patients with NCCP with repeated admissions to hospital (Negative ETT, normal Coronary Angiogram or normal Thallium scan) who were further investigated with OM and pH studies between November 1998 and May 2001 (2.5 years/60 patients). Diffuse Oesophageal Spasm (DOS), Nutcracker oesophagus and Achalasia, as defined by Spechler and Castell (Gut 2001;49:145–51), were the only motility disorders recognized as causes of NCCP in this study.

Results: All patients had normal endoscopy or barium swallows. 17 (28%) patients had significant reflux disease, 14 (23%) had DOS and 6 (10%) had nutcracker oesophagus (of whom 50% also had reflux). Normal studies were found in 25%. 5 patients had non-specific oesophageal dysmotility and 2 patients had hypomotility. All patients with significant reflux disease were treated with PPI and 3 patients had anti-reflux surgery. 90% of patients with nutcracker Oesophagus and DOS were treated with Nitrates or calcium blockers with/without PPI. 37% of patients had reflux symptoms and predictive values for significant reflux were 64% (positive), and 92% (negative). 22% of patients had dysphagia. Predictive values for significant dysmotility were 69% (positive) and 72% (negative). Management was changed in 67% (40 patients) who had OM and pH studies. The nature of the diagnosis was carefully explained in all patients with positive studies. Only one (1.6%) has been readmitted and one (1.6%) had further cardiac investigations (mean follow-up 1.5years).

Conclusions: A positive diagnosis of oesophageal dysmotility or reflux changed the management, reduced readmission rates and the need for further cardiac investigations. The presence or absence of GI symptoms has a high predictive value for OM and pH abnormalities in NCCP.


D. Kufeji, A. Anggiansah1, W.J. Owen1, E.H. Dykes. Department of Paediatric Surgery University Hospital Lewisham, London; 1Oesophageal Laboratory, St Thomas' Hospital, London, UK

Introduction: Sick neonates often require ventilation for prolonged periods of time. Gastro-oesophageal reflux (GOR) is very common in newborn infants, particularly those who are preterm. This can lead to significant morbidity and in extreme cases the neonate can only be successfully weaned off the ventilator after anti-reflux surgery.

Aim: To evaluate oesophageal motor function and acid clearance mechanisms in ventilated neonates.

Methods: Combined pressure and pH monitoring was undertaken in 10 neonates requiring assisted ventilation using Dentsleeve micromanometric assembly and a paediatric (1.5mm diameter) antimony pH sensor. Study repeated when baby was off the ventilator.

Results: Mean gestational age = 33 weeks and mean birth weight 1510 grams (range 28–36 weeks). Mean duration of recording = 58 minutes. LOS pressure = 20 mmHg off ventilation and 40.6 mmHg during positive pressure ventilation. A total of 683 pressure wave sequences were recorded. There were 4 major patterns normal peristalsis (69.8%, of which 16.5% low amplitude), reverse peristalsis 3.6%, synchronous activity 3.2%, non transmitted activity 21.7%. Eleven waveforms (1.6%) could not be adequately categorised. Reflux episodes (pH drop ≥ 0.5 for 10seconds) = 50 with a mean reflux duration of 22 seconds. An average of 2 normal swallows were required to return pH to pre reflux levels.

Conclusion: Ventilated neonates seem to have high oesophageal and LOS pressures that may protect them against reflux. However they exhibit a large proportion of ineffective oesophageal motor activity. During periods of reflux the oesophagus was cleared efficiently by peristaltic oesophageal contractions.


016 intragastric ph in ambulant subjects and its relations to physiological and pathological reflux

R.P. Arasaradnam, L.F. Smith, S.A. Riley. Dept of Gastroenterology & Oesophageal Studies, Central Sheffield Teaching Hospitals, Northern General Hospital, Sheffield, UK

Background and Aims: Episodes of gastro-oesophageal reflux (GOR) are usually associated with a loss of lower oesophageal sphincter (LOS) pressure. However, on many occasions barrier pressure is lost yet reflux does not occur. This suggests that other factors also influence the occurrence of reflux. The aim of this study was to measure pH at the gastric cardia in ambulant subjects and determine its relations to physiological and pathological reflux.

Methods: 17 asymptomatic volunteers (9 males, aged 21–33 years) and 17 patients (11 males, aged 33–53 years) with non-erosive reflux disease were studied. Standard station pull-through manometry was performed to locate the LOS. Under ambulant conditions, pH was measured at 5cm above and at 2 and 10cm below the LOS.

Results: As expected, oesophageal acid exposure (% time pH< 4) was greater in patients than volunteers (pre-prandial 8.5 v 0.9, p<0.0002 ; prandial 4.0 v 1.1, p<0.04; 0 to 60 min post-prandial 11.7 v 1.0, p<0.002; and while supine 13.7 v 2.3, p<0.001). Gastric cardia acid exposure (pH at 2 cm below the LOS) showed marked variability but was again greater in patients than volunteers (table). Transient buffering of cardia pH was seen in patients during ingestion of meals but rapidly returned to pre-prandial values. Gastric body acid exposure (pH 10 cm below the LOS) was consistently high and similar in patients and volunteers. Significant buffering was not seen.

Abstract 016

Conclusions: Under ambulatory conditions, the gastric cardia is variably exposed to acid. Transient buffering is seen following meal ingestion. Acid exposure is greater in patients with reflux disease and this is likely to influence the occurrence of reflux when barrier pressure is lost.


J. Fletcher, A. Wirz, J. Young, K.E.L. McColl. Dept of Medicine and Therapeutics, Gardiner Institute, Western Infirmary, Glasgow G11 6NT, UK

Background: The gastric cardia and distal oesophagus are common sites of upper GI disease and deserve further study. We have shown that after a meal there exists a pocket of highly acidic gastric juice in the proximal stomach that fails to be buffered by food. The location of this acid in relation to the cardia and distal oesophagus was unclear.

Aims: To establish the relationship between the unbuffered proximal acid pocket and the squamo-columnar junction ( Z-line).

Methods: Ten healthy subjects were studied using a dual channel pH electrode with 1cm distance markings. The squamo-columnar junctions (Z-line) was marked by attaching metal clips at endoscopy. The pH electrodes were withdrawn by 1cm increments from the stomach into the oesophagus. The minimum pH at each electrode position, the distance from the nostril to the pH step-up and from the nostril to metal clips (Z-line) shown on X-ray were measured in each subject under fasting conditions and after a meal of fish and chips.

Results: The pull through studies revealed a pocket of acid in the region of the gastro-oesophageal junction which escaped the buffering effect of meals, remaining highly acidic (pH 1.6) compared to the body of the stomach (pH 4.4) (p < 0.01). This pocket of acid (defined as < pH 2) extended over 2cm (range 1–4cm). The pH step-up distance moved after the meal (46.0cm fasting vs 44.4cm postprandial p < 0.05). In contrast the distance to the Z line did not (46.3cm fasting vs 46.2cm postprandial). The fasting pH step up corresponded to the Z-line and therefore the acid pocket extended from the cardia across the Z-line and 1.8cm into the distal oesophagus.

Conclusions: This study shows that after a meal unbuffered gastric acid traverses the Z-line and extends from the cardia to the distal oesophagus. This observation is likely to be relevant to the high prevalence of mucosal pathology recognised to occur at, just above and just below the squamo-columnar junction.


K. Ragunath1, N. Krasner1, V.S. Raman1, M.T. Haqqani2, W.Y. Cheung3. 1Dept of Gastroenterology and 2Pathology, University Hospital Aintree, Liverpool; 3School of Postgraduate Studies, University of Wales, Swansea, UK

Background: The value of methylene blue directed biopsies (MBDB) to detect specialised intestinal metaplasia (SIM) and dysplasia in Barrett's oesophagus remains unclear.

Aim: To compare the accuracy of MBDB technique against random biopsy (RB) to detect intestinal metaplasia and dysplasia in patients with Barrett's oesophagus.

Methods: A prospective randomised cross over trial was undertaken comparing MBDB and RB in patients with = 3cm Barrett's oesophagus without macroscopic evidence of dysplasia or cancer. Biopsies were taken from the stained and unstained mucosa in focal staining Barrett's segment and random four quadrantic in the case of diffuse and heterogeneous staining Barrett's segment. RB was done using standard endoscopic biopsy forceps from the four quadrants at 2 cm intervals. Dysplasia was defined as: indeterminate dysplasia (ID), low grade dysplasia (LGD), high grade dysplasia (HGD) and carcinoma (Ca). The histopathologist was blinded (unaware of which samples were methylene blue stained).

Results: Fifty-seven patients were recruited, of whom 44 were male. The mean age was 60 years range (31–85). The mean length of Barrett's was 5.4 cm, range (3–12). Using MBDB technique 651 biopsies were obtained (mean 11.42, range 5–23). SIM was present in 491 biopsies (75.42%). Dysplasia and carcinoma were diagnosed in 26 patients: ID1, LGD 21, HGD 2, Ca 2. Using RB technique 618 biopsies were obtained. SIM was present in 421 biopsies; mean 7.39 biopsies (68.12%). Dysplasia and carcinoma were diagnosed in 23 patients: ID 3, LGD 16, HGD 2, and Ca 2.

Conclusion: The diagnostic accuracy of MBDB technique was similar to RB technique in identifying HGD and Ca. However, there was a trend towards increased detection of SIM and LGD by MBDB technique. MBDB did not reduce the number of biopsies taken. Further studies involving larger number of patients are needed to detect a significant difference between the two techniques. Until then there is no role for MBDB in the routine use for Barrett's surveillance.


S.H. Doak, G.J.S. Jenkins, A.P. Griffiths1, E.M. Parry, J.M. Parry, J.N. Baxter2. Human Molecular Pathology Group, School of Biological Sciences, University of Wales Swansea, Singleton Park, Swansea SA2 8PP; 1Department of Pathology and 2Department of Surgery, Morriston Hospital, Swansea SA6 6NL, UK

Introduction: Barrett's oesophagus is a pre-malignant condition characterised by the conversion of the normal squamous cell oesophageal epithelium to a mucosa comprised of columnar cells as a result of chronic gastro-oesophageal reflux. This lesion progresses in a step-wise fashion through histologically identifiable stages and ultimately develops into oesophageal adenocarcinoma in approximately 10% of patients. To determine when specific genetic alterations arise during this neoplastic progression FISH was employed.

Methods: Gastroscope cytology brushes were used to exfoliate epithelial cells from patients at each stage of progression (Barrett's metaplasia to oesophageal adenocarcinoma). Interphase cell preparations were generated and subsequently analysed by application of fluorescently labelled centromeric probes for chromosomes 4, 8, 9, 20 & Y and locus specific probes for the p53, p16 & Rb genes.

Results: Increased copy numbers of chromosomes 4 & 8 occurred in 13/15 & 10/15 Barrett's metaplastic samples respectively, thus representing the most prominent and earliest alteration arising during neoplastic progression. Loss of the p16 tumour suppressor gene also arises during metaplasia (4/15) and was found to precede chromosome 9 amplifications, but in contrast, p53 loss is a later change first appearing in HGD. Increasing loss of chromosome Y occurs with progression.

Discussion: Aneuploidy is an early occurrence during the progression of Barrett's oesophagus with copy number increases of chromosomes 4 & 8 present in the majority of metaplastic samples. HGD appears to be the stage at which most aberrations accumulate, thus this genetic instability may possibly account for the high proportion of these patients that progress to cancer.


T.K.L. Wong1, L.B. Lovat1, P. Sirieix2, R.C. Fitzgerald2. 1National Medical Laser Centre, Department of Surgery, Royal Free and University College School of Medicine, London; 2Cancer Cell Unit, Hutchison/MRC Research Centre, Cambridge CB2 2XZ, UK

Background: Photodynamic therapy (PDT) is an emerging endoscopic treatment for patients with dysplasia in Barrett's oesophagus. Application of PDT to Barrett's oesophagus ideally leads to regeneration of non-dysplastic, stable squamous mucosa. A limitation of this technique is the persistence of Barrett's epithelium, including buried glands, which may still have dysplastic potential. Since the cellular microenvironment is crucial to epithelial repair it might be possible to manipulate this to promote squamous epithelial re-growth.

Aims: To investigate (a) differences in early repair (restitution) of an oesophageal cell monolayer following mechanical or PDT injury; (b) whether restitution can be altered by adding growth factors/cytokines.

Methods: Cell lines; Squamous (OE21), Barrett's (OE33) and co-cultures were injured mechanically or with PDT (5-aminolevulinic acid and blue light) using a novel applicator. Wounds were measured over 24 hours and immunofluorescence for cytokeratins identified squamous versus columnar cells. Transforming Growth Factor beta (TGF-β1), Hepatocyte Growth Factor (HGF), Interleukin 8 (IL-8) and Keratinocyte Growth Factor (KGF) were added individually to assess their effect on restitution compared with serum free media.

Results: In co-culture, squamous cells (OE-21) underwent greater restitution than columnar cells (OE-33). In both mechanical wound and PDT assays of co-cultures, TGF-β1 increased cell repair by restitution compared with controls (p<0.05). This effect was not seen in individually cultured cell lines. KGF and HGF stimulated restitution of squamous and co-culture cells after mechanical injury and also inhibited columnar cells significantly (p<0.05). IL-8 had no effect on cell restitution.

Conclusions: Restitution, in the first 24 hours after PDT and mechanical injury in vitro, can be influenced by growth factors. It may be possible to manipulate the microenvironment to favour squamous re-epithelialisation after PDT.


D.M. Aldulaimi, M. Cox, C.U. Nwokolo, D.E. Loft. University Hospitals Coventry and Warwickshire NHS Trust, UK

Aim: To establish whether Barrett's surveillance is worthwhile in terms of incident cancers and whether outcomes are favourable.

Method: A prospective non-randomised single centre Barrett's surveillance program commencing 1/1/1992 through 1/4/2001 (100 months). Oesophagectomy recommended for high grade dysplasia or carcinoma.

Results: Of 23,725 endoscopies, 506 patients were diagnosed as Barrett's oesophagus and 24 (5%) had carcinoma at diagnosis (prevalence cancers). 126 patients had at least one surveillance endoscopy. 248 surveillance endoscopies were performed spanning 338 patient years. 13 surveillance (incidence) cancers were detected. The surveillance cancers were all detected after at least one year of surveillance and no patient had dysphagia at diagnosis. In the prevalence cancer group 12 of the 24 patients underwent oesophagectomy. Lymph nodes showed evidence of metastases in 10 of the 12 resections. In the surveillance group 10 patients underwent oesophagectomy. All had carcinoma in the resection specimen. Lymph nodes showed evidence of metastases in 1of the 10 resections. 3 patients in the surveillance cancer group did not have an oesophagectomy. 1 of these patients died. 1 patient in the prevalence cancer group (4% of group; 8% of those operated) and 7 patients in the surveillance cancer group (54% of group; 70% of those operated) remain disease-free more than 2 years post-oesophagectomy. Assuming the 7 patients in the surveillance cancer group are cured and that the cost of endoscopy is £120, the cost per cancer cured is £4250. One curable cancer was detected per 48 patient years of surveillance (338/7).

Conclusion: 5% of Barrett's patients undergoing endoscopy have prevalent cancers. If surveillance is performed, 4% per year (13/338 %) develop cancer and 2 % per year are cured of their cancers. Most surveillance cancers are operable and of those undergoing surgery 70% are cured. Barrett's surveillance is cost-effective compared with other cancer screening or surveillance initiatives.


J.A. Fallowfield, P.J. Winwood, N.A. Davies, M. Lesna. Royal Bournemouth Hospital, UK

Background: The incidence of adenocarcinoma of the oesophago-gastric junction is rising. Barrett's oesophagus (BO) is considered a premalignant condition for this cancer. The effectiveness of endoscopic cancer surveillance programmes is unproven and controversial.

Aims: To measure the incidence and outcome of adenocarcinoma in a BO surveillance population over a 3 year period and to evaluate the effectiveness of endoscopic screening in a DGH.

Methods: All patients with BO attending Royal Bournemouth Hospital Endoscopy Unit between 1998–2001 were included. Cases were identified from the pathology computer database.

Results: We identified 299 patients with known BO in a biannual surveillance programme, with a mean age of 65 years. In the 3 year study period there were 34 BO-associated adenocarcinomas detected. 7 (19%) were identified as a result of a surveillance endoscopy. There were no interval cancers in the surveillance group. 27 (81%) were diagnosed de novo at index endoscopy. The mean age of patients with BO-adenocarcinomas was 69 years; 29/34 (85%) were male. Cancer incidence per patient year of follow-up was 1 : 79. All of the 7 BO-adenocarcinomas detected during endoscopy were early stage (<T2, N0) and had oesophagectomy (5/7) or endoscopic mucosal resection (2/7). All have survived to date (range 9–28 months). De novo BO-adenocarcinomas were generally more advanced at presentation. 17/27 were suitable only for palliative therapy; 16/17 have died.

Conclusions: New oesophageal cancers were found during surveillance endoscopy at a higher rate compared with most published studies. The reason for the high detection rate in this study may be due to the advanced age of this surveillance population. Nevertheless, most adenocarcinomas occurred in patients without a previous diagnosis of BO. There was a bias towards early stage cancers in patients with BO under surveillance. The outcome in these patients has been favourable compared with BO-related cancers diagnosed de novo at index endoscopy. Our experiences support endoscopic surveillance in selected patients with BO.


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

Background: Photodynamic therapy (PDT) is a minimally-invasive alternative to oesophagectomy for high-grade dysplasia (HGD) or intramucosal adenocarcinoma (T1m AdCa) arising in Barrett's columnar lined oesophagus. Initial reports using 5-amino laevulinic acid (ALA) suggest that HGD can be eradicated in 80% of patients. Our aim is to identify parameters associated with successful eradication of disease.

Methods: 15 previously untreated patients (13 M, 2 F) with HGD (11) or T1m AdCa (4) were treated over a 3-year period. Ethical approval was obtained. Patients were photosensitised with ALA 60mg/kg (light activation 635 nm) and received 30 PDT sessions: median of 2 per patient (range 1–3). ALA light doses were 500–1000J/cm diffuser fibre and treatments took approximately 40 minutes per 4cm length of columnar mucosa. Two delivery devices were tested (16 and 25mm diameter). 3 patients were treated additionally with the photosensitiser meso-tetrahydroxyphenylchlorin (mTHPC, 0.15mg/kg).

Results: Disease (HGD 5, AdCa 3) was eradicated in 8 patients (53%): 6 patients (40%) with ALA alone, 2 with additional mTHPC. Median follow-up is 10 months (range 2–29) with no deaths and no oesophageal strictures. One patient developed new HGD 2 years after successful treatment and is being re-treated. Benign glands “buried” under neo-squamous epithelium were seen in 7/15 cases. Failure of treatment was associated with the length of Barrett's segment (median in responders 6cm (4–8cm), non-responders 8cm (6–13cm), p=0.02) and with the presence of multi-focal disease (success in only 2/8 versus 6/7 for unifocal disease, p=0.04). Age, sex, presence of hiatus hernia, and size of delivery device did not appear to influence outcome.

Discussion: Using current treatment parameters, PDT with ALA for dysplasia and T1m carcinoma in Barrett's oesophagus is effective in less than one half of patients. A long Barrett's segment and multi-focal disease are associated with a poor outcome. Techniques to achieve a deeper effect (such as adding an iron chelator to ALA or using a different photosensitiser) may give better results.


Q. Arfin1, Z. Mahmood1, B.P. McMahon3, P. Byrne1, J.V. Reynolds1, E. Murphy1, V. Trimble1, D.G. Weir2. 1St James's Hospital; 2Trinity College; 3AMNCH, Dublin, Ireland

A method has been developed whereby sutures can be placed via an endoscope just below the oesophago-gastric junction (OGJ) whose purpose is to improve the function of the OGJ and thereby prevent oesophageal reflux. The aim of this work is to assess the safety and efficiency of the BARD Endocinch© for the treatment of GORD. 20 patients with symptoms of GORD were recruited; all were followed for 6 months and 14 for 1 year. The inclusion criteria included a dependence on proton pump inhibitor (PPI) drugs to control their reflux symptoms, and a documented oesophageal acid reflux. Exclusion criteria were age less then 18 years, pregnancy, dysphagia, obesity (BMI) > 40, previous upper intestinal surgery and an hiatus hernia > 2 cms. Pre-procedure assessment included symptom scoring, oesophageal endoscopy, manometry and 24 hour oesophageal pH, and completed quality of life (QOL) questionnaire. Post procedure symptomatology, QOL and adverse events were assessed at 1, 3, 6 and 12 months. Repeat endoscopy, manometry and 24 hour pH were performed at 3 months. Mean age was 37 (22 – 58 yrs.). All received conscious sedation (Midazolam and Pethidine). The median duration of the procedure was 50 minutes. The mean heartburn symptom score (heartburn frequency x severity) was 19 pre-procedure and 3 at six months (p = 0.0004). Moderate to severe regurgitation symptoms improved from 76% to 12% at 6/12 (p<0.004). Overall the mean pH Demeester score reduced from 39.7 to 30.8 (p = 0.03), upright events 13 ± 6.2 to 9.4 ± 6.0 (p = 0.008), numbered reflux episodes 168 ± 68 to 117.4 ± 67.7 (p = 0.015). 12/20 patients had a normal pH profile at 3 months follow up. Use of PPI reduced from 100% to 38%. No significant adverse events occurred. QOL assessment showed significant improvement in all modalities (p = 0.001). Endocinch is a safe and effective method of managing GORD. Three months post procedure demonstrated improved symptoms, reduced acid reflux and reduced requirement for PPI drugs. The degree of improvement in symptoms and PPI requirement remained constant for up to 1 year post procedure.


P.P. Tekkis1, P. McCulloch1, I.S. Benjamin1, J. Poloniecki2. RISK & ASCOT Group of Hospitals; Guy's Kings & St Thomas' School of Medicine; Public Health Sciences, St George's Hospital, London, UK

Introduction: Predictive models are increasingly being applied to evaluate surgical performance in upper GI surgery. We describe the use and limitations of hierarchical models in making quantitative comparisons between teaching and non-teaching institutions.

Methods: A longitudinal study of 981 patients undergoing major oesophagogastric resections from 31 UK hospitals from 1995 to 2000. Primary outcome was in-hospital mortality and risk-adjusted mortality. A two-level random effect logistic regression model was developed using age, pre-operative POSSUM and staging as “level 1” (patient) risk factors and teaching unit status as “level 2” risk factors.

Results: Mortality in the study was 11.3%. On univariate analysis crude operative mortality was significantly different between units (range 0% to 26.8%, p=0.001) and between teaching (8.8%, n=374) and non-teaching (13.3%, n=607) hospitals (p=0.032). Following risk-adjustment for patient related covariates, the teaching hospital status was not an independent predictor of outcome in the hierarchical model (Odds ratio 0.87, CI=0.63–1.21) despite a significant variation in inter-hospital operative mortality.

Conclusions: Although the divergence in performance may relate to bias in data collection, the study suggests that the `institution or surgeon effect' plays a determining role in the quality of healthcare provision in Upper GI surgery.


  • Funding: The Royal College of Surgeons of England.