Article Text

Download PDFPDF

Plenary posters 168–197

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.


I.D.R. Arnott, S. Ghosh.Gastrointestinal Unit, University Department of Medical Sciences, Western General Hospital, Edinburgh, UK

Introduction: Infliximab is a new treatment for refractory and fistulating Crohn's disease (CD). Clinical trials and audit data have proven efficacy on disease activity and health related quality of life although there is little data regarding cost effectiveness. Prior to the introduction of novel biological treatments hospitalisation and surgery were the major costs in CD treatment but the expense of Infliximab may change this.

Aim: We assessed whether Infliximab reduced hospitalisation or frequency of out patient clinic review within 6 months of infusion.

Methods: We analysed 30 well-characterised CD patients who had received a single infusion of Infliximab (5mg/kg) for refractory active CD. Clinical details and initial response rates have been published. Patients were followed prospectively and out patient visits, number of hospital admissions and total number of inpatient days were collected for the 6 months prior to and following Infliximab. Data was also compiled from the hospital electronic record of all patient episodes. Only patients that were cared for exclusively at our institution were included.

Results: Data for the 30 patients is displayed in the table. There are no significant differences in hospital attendance or admission rates. There remains no difference in the clinic visits and admission days pre and post Infliximab if patients are stratified as to whether they had a response or not and if they are on immunosuppressive or not.

Abstract 168

Conclusions: In the present study, Infliximab dose not reduce clinic visits or hospitalisation within 6 months of infusion. Although reductions may be seen with multi-dose regimes dramatic reductions in hospital episodes are not to be expected with a single dose alone. Infliximab remains a highly effective drug for active CD but it is likely that this will have implications when drug funding is allocated.


J.H. Li, X.S. Yang, S. Cullen, K.L. Lin, A. Armuzzi, T. Ahmad, M.J. Neville, R. Walton, D.P. Jewell.Gastroenterology Unit, Gibson Laboratories, Radcliffe Infirmary, University of Oxford, Oxford OX2 6HE, UK

Ulcerative colitis (UC) and Crohn's disease (CD), both forms of inflammatory bowel disease (IBD) are complex traits. Intercellular adhesion molecule-1 (ICAM1) is expressed on vascular endothelium and plays a key role in the transendothelial migration of neutrophils and T-cell activation. The region harboring the ICAM1 gene on 19p13 is linked to CD in a Canadian genome wide scan, and a growing body of evidence indicates that ICAM1 could play a role in IBD development. Our previous work has replicated the Canadian linkage of 19p13 to CD. ICAM1 is known to contain at least two polymorphic sites, situated in codons 241 (R/G 241) and 469 (K/E 469). A North American study has shown an association between IBD and the R241 polymorphism (Yang etc 1995). We have examined potential associations of ICAM1 polymorphisms in 132 UC and 67 CD and ethnically matched 131 controls. CD patients include subgroups of 26 ileal disease and 31 ileo colonic disease, 26 fistulating disease and 35 non-fistulating disease, 53 stenosing disease and 8 non-stenosing disease. UC patients include 37 patients who have undergone colectomy, 22 with mild total colitis, and 37 with proctitis. Both patients and controls were genotyped by PCR- SSP for ICAM1 polymorphisms at codon R/G241 and codon E/K469. There were no differences between the groups in the frequency of R/G241. The control frequency of ICAM1 exon 6 K469 was 38.2%. In CD overall, it was 76.9% (p<0.001). For CD patients with ileal disease, the frequency was 70.9% and 43.3% for ileo-colonic disease. For patients with fistulating disease, the K469 frequency was 69.2%, and 84.3% for non-fistulating disease. For patients with stenosing disease, the K469 frequency was 77.4%. For UC overall, the frequency of ICAM1 exon 6 K469 was 69.7% (p<0.001). For patients with severe disease requiring colectomy, with mild total colitis, and with proctitis, the frequencies of this polymorphism were 66.22 %, 75% and 81.1% respectively. This study suggests that the alteration in the amino acid sequences of E469 to K469 of the ICAM1 molecule may influence IBD.


T. Ahmad1, K. Mulcahy-Hawes1, M. Bunce2, A. Armuzzi1, K. Welsh3, S. Marshall3, D. Jewell1.1Departments of Gastroenterology,2Transplant Immunology, University of Oxford;3National Heart and Lung Institute, Imperial College, London, UK

Background: The clinical, endoscopic and histological features of intestinal BD are similar to those of IBD. BD is rare in Northern Europe where IBD is common, whilst the prevalence is high on the `Silk route' where IBD is rarely reported. This inverse relationship may reflect geographical differences in diagnostic practice. Diagnosis of BD, requires the presence of ROU. We have reported that Caucasian BD is associated with HLA-B*51 and B*57. We hypothesise that these markers might also be associated with ROU in IBD and therefore molecularly define an overlap group.

Aims: To determine the prevalence and genetic associations of ROU in patients with IBD.

Methods: History of ROU reported in questionnaires sent to 244 CD and 330 UC patients. Linkage disequilibrium mapping was carried out across 340 polymorphisms, broken down into 24 discrete gene haplotypic blocks. Genetic comparisons were made between IBD and healthy controls.

Results: 33.2% of UC and 38.9% of CD patients reported ROU (historical prevalence in general population 10%). In UC patients with ROU, associations with alleles on two extended HLA haplotypes were observed. Peak relative risk (RR) was at DRB1*0103 on the first (16.5% ROU+ vs 3.8% ROU-; P=0.0003; RR=5.0) and B*57 on the second (12.1% ROU+ vs 3.8% ROU-; P=0.009; RR=3.5). In CD patients with ROU a negative association was observed with B*51 (3.4% ROU+ vs 13.8% ROU-; P=0.01) However 11 CD patients who carried B*51 or B*57 reported ROU. 8 of these fulfilled the diagnostic criteria for BD.

Conclusions: 1. Prevalence of ROU in IBD is 3x greater than in the background population. 2. ROU in UC patients is associated with B*57, a BD susceptibility allele. 3. 8/11 CD patients with ROU who possess B*51 or B*57 fulfil the criteria for diagnosis of BD. 4. In our IBD clinic 5% of patients clinically and molecularly resemble BD.


D.M. Aldulaimi, J. Barclay, K. Getliffe, A.G. Morris, E. Karteris, E. Hillhouse, C.U. Nwokolo.University Hospitals Coventry and Warwickshire NHS Trust and Dept of Biological Sciences, Warwick University, UK

Background: Telomerase knockout mice develop ulceration and atrophy of the bowel. In humans colonic telomerase activity is decreased in ulcerative colitis but it is unknown whether this is restricted to the colon. The object of this study was to assess lymphocyte telomerase enzyme activity in patients with inflammatory bowel disease and to determine the role of the mRNA that encodes its catalytic sub-unit (hTERT) in the regulation of enzyme activity.

Methods: Blood was sampled from 47 patients with ulcerative colitis (UC), 37 with Crohn's disease (CD) and 37 controls. Lymphocytes were cultured for 72 hours with phytohaemagglutinin. Telomerase activity was measured in stimulated and unstimulated lymphocytes using the Telomerase Repeat Amplification Protocol (TRAP) assay. Lymphocyte hTERT mRNA was quantified in about 40% of samples (18 UC, 14 CD and 14 controls) by real-time PCR.

Results: Expressed as median (95 CI) in arbitrary units (Stimulated lymphocytes only). Telomerase enzyme activity in controls was 5.96 (3.3 – 9.2) and was decreased significantly in UC 3.1 (3.1–4.1) p < 0.05 and non-significantly in CD 2.27 (1.2–6). There was no difference in hTERT mRNA concentration between the three groups. Telomerase activity and hTERT mRNA were generally undetectable in unstimulated lymphocytes.

Conclusion: Lymphocyte telomerase activity is decreased in ulcerative colitis. This suggests that previously reported colonic telomerase deficiency in UC extends to non-colonic tissue and could represent a global defect. Factors other than hTERT mRNA expression may contribute to the regulation of telomerase activity in stimulated lymphocytes


S. Fradley, P.D. Lewis, P. Griffiths, J.M. Parry (introduced by J. Baxter).Human Molecular Pathology Group, University of Wales Swansea, Singleton Park, Swansea SA2 8PP, UK

The human mitochondrial genome (mtDNA) contains a short hypervariable, non-coding control region known as the D-Loop. A number of recent investigations have revealed that a proportion of colorectal adenocarcinomas harbour mtDNA mutations, not present in normal surrounding mucosa and are classified as tumour-specific. Tumour specific mtDNA mutations have also been observed in many other tumours including lung and bladder and it has been suggested that these mutations may serve as diagnostic markers for cancer. The D-Loop, used commonly in population studies and the most mutable region within mtDNA, may be rapidly and cost-effectively scanned for mutations. Previous studies involving the search for tumour-specific mtDNA mutations in colorectal cancer have relied on small sample sizes or have failed to reveal mutations within the D-Loop. Using PCR-SSCP and DNA sequencing we have undertaken a comprehensive survey of the D-Loop in adenocarcinoma and normal mucosa of twenty patients to (i) demonstrate the usefulness of the D-Loop in providing tumour-specific markers for colorectal cancer, (ii) reveal the types and distribution of mutations within the D-Loop in colorectal tumours, (iii) estimate the frequency of mutation within this region in adenocarcinomas, (iv) establish the levels of heteroplasmy in colorectal tumours. Of the colorectal adenocarcinomas, 25% showed tumour specific mutations which where not present in the normal mucosa. Sequencing revealed the mutations to be a 1-bp C:G deletion and a 1-bp C:G insertion at nucleotide position 309, two C:G/T:A transitions at nucleotide positions 61 and 52 and one T:A/A:T transversion at nucleotide position 55. This study is currently ongoing and we aim to increase the sample size by 100% and analyse adenomas and hyperplastic polyps. Our results will allow predictions to be made concerning the causative factors of mtDNA mutations in colorectal cancer and give insight into whether these mutations precede neoplasia or are a consequence of tumorigenesis.


R.P. Arasaradnam, N. Bhala, A.J. Shorthouse, I.J. Adam, A.J. Lobo.Dept of Gastroenterology & Colorectal Surgery, Central Sheffield Teaching Hospitals, Royal Hallamshire Hospital, Sheffield, UK

Background: Colorectal Cancer (CRC) is the commonest cancer in the United Kingdom among non-smokers. In accordance with the Government NHS Cancer Plan, the concept of a two week wait (tww) was introduced to ensure that all patients with a suspicion of CRC were seen by a specialist within two weeks of referral.

Aim: The aim of this study is to examine the effect of this implementation on tumour staging for patients with colorectal cancer.

Methods: TNM staging and demographics from 140 consecutive patients diagnosed with CRC and treated at a single regional unit were collected for one year before the introduction of the tww and for a year afterwards, to September 2001. Of the 107 elective patients, 90 were operated on. 33 patients presenting as emergencies were excluded from subsequent analysis.

Results: 24/40 (60%) patients were node negative in the pre tww group compared with 32/46 (70%) in the post tww group (κ2=0.21; ns). Similarly, before the introduction of the tww there were 9/40 (23%) patients regarded as being of good prognosis (T1N0, T1N1,, or T2N0) compared with 13/46 (28%) in the post tww (κ2=0.29, ns). The total number at each T stage pre tww (n=40) was T1=3; T2=8; T3=18; T4=11 compared with post tww (n=45) T1 =6, T2=7, T3=22 and T4=10 (κ2=0.55, ns). Irrespective of whether they were referred pre or post the tww, of all patients seen within 14 days (n=47) [median 9 & range 0–14 days], 25/47 (53%) were node negative. This compared with 38/53 (72%) of those seen greater than 14 days (n=60) [median 32 & range 15–216 days] after referral (κ2=0.01, ns). Of those seen within 14 days, 8/47 (17%) were in the good prognostic category compared with 14/53 (26%) seen after (κ2=0.19, ns). Median follow-up (days) was 513.5 (range 38–696) & 315.5 (range 12–454) for pre and post tww respectively. Overall survival (including emergency admissions) in the pre tww was 79.31% (n=59) while in the post tww (n=81) was 86.42% (p < 0.4749). Survival for elective admissions pre tww (n=47) was 91.30% compared to 91.67% (n=60) in the post tww (p=ns).

Conclusions: The two week wait has made no impact on TNM staging for CRC at initial presentation.


C.L. Brooks, E.A. Sanders, S. Hadfield, S. Green, W.M. Rosenberg.

Introduction: Hepatitis C virus (HCV) establishes a chronic infection in up to 85% of those exposed. The resultant immune mediated hepatitis leads to progressive fibrosis and cirrhosis in a significant proportion of patients. The combination of interferon-α (IFN-α) and ribavirin achieves a sustained virological response in around 40% of those treated. The addition of a Polyethylene Glycol (PEG) moiety to the IFN-α molecule significantly changes the pharmacokinetics and improves efficacy. HCV specific CD4+ T cell responses are weak or undetectable in patients with chronic hepatitis C (CHC), whilst they are brisk and multi-specific following spontaneous resolution of acute infection. We aimed to characterise CD4+ HCV specific responses in CHC patients at the start of treatment with PEG-IFN-α and ribavirin, and then repeat them serially throughout the course of treatment, correlating with viral response.

Methods: Peripheral blood mononuclear cells were isolated by density gradient centrifugation from 8 patients prior to starting anti-viral treatment with PEG-IFN-α and Ribavirin. These cells were set up in culture with either recombinant HCV antigens or appropriate positive and negative controls. Cellular proliferation was assessed by incorporation of tritiated thymidine. Cellular markers of activation and cytokine secretion were assayed by fluorescence activated cell scanning. The assays were repeated at regular intervals throughout treatment and follow up. At each time point, viral load was measured by a PCR based quantitation method.

Results: Throughout treatment, HCV specific lymphocyte proliferation increased in magnitude in 75% of patients, apparent from a median of 18 weeks. The greatest increase was seen in patients who achieved sustained virological response, in which stronger responses were maintained. There were also changes in the CD4+ cytokine profile induced by core and helicase recombinant proteins which correlated with virological response.

Discussion: There has been one study examining serial HCV specific CD4+ responses whilst on standard combination treatment [Cramp et al. 2000]. Our study has similar results, but the onset of increased magnitude of responses is later than seen on standard treatment, possibly reflecting the different pharmacokinetics. This study supports the hypothesis that successful anti-viral treatment allows tolerance to HCV to be broken.


M. Court, M.A. Aboshkiwa, M.F. Dixon, P.A. Robinson, J.E. Crabtree. Molecular Medicine Unit, St James's Hospital, Leeds LS9 7TF, UK

Introduction: Chronic H. pylori infection in Mongolian gerbils (MG) has been demonstrated to result in gastric cancer. The aims of this study were to assess the ability of the mouse adapted SS1 strain to colonise MGs and to induce gastric pathology, gastric epithelial cell proliferation and apoptosis.

Methods: MGs were orally challenged three times with H. pylori SS1 strain. Infected animals (n= 28) plus controls (n =23) were sacrificed following intra-peritoneal injection with bromodeoxyuridine at 4, 12 and 36 weeks post-infection (p.i.). Gastric epithelial cell proliferation and apoptosis were determined immunohistologically and by TUNEL assay. Infection was confirmed histologically and by culture. Strains were identified as SS1 by RAPD-PCR and sequence analysis of glmM.

Results: 27/28 of the inoculated MGs were H. pylori SS1 positive. At 4 weeks p.i. gastritis was antral predominant. Corpus gastritis and atrophy were present in 1/4 MGs at 12 weeks and 6/15 at 36 weeks. Gastric epithelial cell proliferation was significantly increased (p <0.05) in the antrum of infected MGs at 4, 12 and 36 weeks p.i. At 36 weeks p.i. MGs with corpus gastritis had significantly increased corpus epithelial cell proliferation compared to uninfected controls (p <0.005) and infected animals with no corpus gastritis (p= 0.06). H. pylori infection was associated with increased apoptosis in the glandular but not the superficial gastric epithelium. In the antrum epithelial apoptosis at 12 (p < 0.05) and 36 (p< 0.005) weeks p.i. was increased compared to uninfected controls. At 36 weeks p.i. a significant increase (p <0.005) in apoptosis in the corpus glandular epithelium was evident which was restricted to the infected MGs which had developed corpus gastritis.

Conclusions: The SS1 H. pylori strain will chronically infect Mongolian gerbils resulting in pangastritis by 36 weeks post infection. H. pylori infection is associated with increased gastric epithelial cell proliferation and apoptosis of the glandular epithelium in the antrum. Progression to corpus gastritis results in similar changes.

This study was funded by Yorkshire Cancer Research.


J.I.W. Jones, D.A. Kendall, P.J. Millns, C.J. Hawkey.

Introduction: Many patients suffer from acid sensitive dyspepsia yet the gastric mucosa is normally anaesthetic to luminal acid. We have previously reported that the pain caused by exposure of the gastric mucosa to the neural irritant capsaicin is pH dependent. We hypothesise that acid enhances the response of gastric mucosal nerves to capsaicin in a similar way to that has been observed for somatic neurons.

Aims: To determine the response of gastric mucosal nerves to capsaicin at physiological and non-physiological pH.

Methods: To study the effects of capsaicin on the cell bodies of gastric mucosal nerves, we injected a neuronal tracer, Texas Red, into the gastric mucosa of 4 Wistar rats 2–4 weeks before removal of their dorsal root ganglia (DRG). Cultured DRG cells were placed in a perfusion chamber mounted on a fluorescence microscope where those of gastric origin were identified by excitation of the Texas Red within them. The cells were loaded with the calcium sensitive ionophore, FURA 2-AM to detect the rise in calcium concentration accompanying cell activation and perfused with a HEPES based buffers at pH7.4 or pH7 containing capsaicin at a concentration of 10–9 to 10–6 molar to establish a dose response curve. Non Texas red containing cells were used as controls since the vast majority of these are of somatic origin.

Results: The percentage of gastric cells responding to capsaicin was less than half that of non-gastric cells but increased at lower pH (see fig 1).

Conclusion: The gastric mucosa has a lower percentage of capsaicin sensitive cells than somatic tissue but like somatic neurons responses are enhanced at reduced pH.


L.A. Houghton1, W. Atkinson1, P. Whitaker2, P.J. Whorwell1, M. Rimmer3.1Dept of Medicine, University Hospital of South Manchester;2Dept of Chemical Pathology, University of Leicester;3Glaxo Wellcome, Stockley Park, UK

Recent pilot data suggests that platelet depleted plasma 5-HT concentrations are undetectable under fasting conditions in both patients with irritable bowel syndrome (IBS) and healthy volunteers1. However, the number of subjects studied was small (n=5 and 6, respectively) and no data was provided as to the detection limits of the methods used. We have measured fasting platelet-depleted plasma 5-HT concentrations plus platelet 5-HT concentrations in 21 female patients with diarrhoea predominant IBS (aged 19–50 yrs) and 19 healthy female volunteers (20–46 yrs). 5-HT concentration was measured by reverse-phase high performance liquid chromatography with fluorimetric detection. α-Thromboglobulin, which is a marker of platelet activation and/or leakage (and thus a marker for adequate blood collecting technique), was also measured by ELISA method.

Results: Under fasting conditions, platelet 5-HT concentration was significantly higher in the female patients with diarrhoea predominant IBS (443.96ng/109 platelets, adjusted geometric mean) than healthy female controls (342.86ng/109 platelets; ratio IBS:healthy controls (95%CI), 1.30 (1.07,1.56); p=0.008). Platelet depleted plasma 5-HT concentration however, was similar in patients (4.25ng/ml) and healthy controls (4.01ng/ml; ratio IBS:healthy controls, 1.06(0.82,1.36, p=0.65). α-Thromboglobulin concentrations were undetectable in any of the samples measured.

Conclusions: Female patients with diarrhoea predominant IBS have larger platelet stores of 5-HT than healthy women, suggesting that they may have increased exposure to 5-HT in their systemic circulation. This supports the observations of Bearcroft et al1 that meal ingestion is associated with a greater increase in plasma 5-HT concentration in patients with IBS compared with healthy controls.

1Bearcroft et al, Gut 1988;42:42–46.


N. Andrews, G. Yu Lu, O. Gerasimenko, J. Rhodes.Departments of Medicine and Physiology, University of Liverpool, L69 3BX, UK

Introduction: Galectins are a family (12 to date) of β-galactoside-binding lectins. It is very unclear what are the functionally important natural ligands for the galectins but galectin-3 expression is implicated in tumour invasion and metastasis. We have therefore investigated the possible ligand(s) for recombinant galectin-3 in HT-29 colon cancer cells.

Methods and Results: HT-29 cells were grown to 70% confluence, harvested, and a cytoplasmic extract prepared. The extract was then separated by SDS-4% polyacrylamide gel electrophoresis (PAGE), electroblotted onto a nitrocellulose membrane and probed using recombinant human galectin-3 protein followed by anti-galectin-3 antibody overlay. A major ligand was then identified with a molecular weight of 420kDa. Since this corresponds to the mucin MUC-1, which is known to express the TF antigen (galactoseβ 1,3 N-acetylgalactosamineα), a known ligand for galectin-3, MUC-1 immunoprecipitate was prepared and probed as before. This confirmed binding of MUC-1 by recombinant galectin-3 protein (fig). Lane a shows three proteins identified by galectin-3 protein in HT-29 cytosolic extract, lane b shows two proteins identified by galectin-3 protein from MUC-1 immunoprecipitate.

The two protein bands observed for MUC-1 are due to the two different alleles of MUC-1. HT-29 cells were separated into invasive and non-invasive cell types by their ability to migrate through a 0.5mm Matrigel. The technique employed was a Membrane Invasion Culture System (MICS). Immuno-confocal microscopy of invasive HT-29 cells showed co-localisation of MUC-1 and galectin-3 whereas MUC-1 expression was weak in the non-invasive cells.

Conclusions: The transmembrane TF expressing mucin MUC-1 is a natural ligand for galectin-3. Its increased expression is correlated with a more invasive phenotype in colon cancer cells.


W.J.H. Griffiths, W.S. Sly1, T.M. Cox.University of Cambridge Dept of Medicine, Addenbrooke's Hospital, Cambridge;1St Louis University School of Medicine, MO, USA

Transferrin receptor 2 (TfR2) is a recently-identified homologue of the ubiquitous transferrin receptor (TfR) and is expressed in the liver and small intestine. Mutations in the HFE gene predispose to hereditary haemochromatosis and a role for HFE in the signalling of body iron status within small-intestinal crypt cells has been proposed. TfR2 mutations however account for rare forms of non-HFE related haemochromatosis suggesting a key role for this receptor sub-class in the control of intestinal iron absorption.

To investigate cellular interactions of HFE and TfR2 a panel of rabbit and avian polyclonal antisera was generated to specific peptide sequences of the human and mouse proteins. Antibodies were first characterized by Western immunoblotting. Using laser confocal microscopy in mouse and human duodenal sections, strong staining of TfR2 was observed in the crypts, where colocalisation occurred with HFE; no staining of HFE and TfR2 was observed in villus enterocytes. In contrast TfR expression, examined using a commercial murine antibody, was ubiquitous but did not colocalise with HFE. The localisation of HFE and TfR2 was further examined in situ in human Caco-2 cells, which have a small intestinal phenotype. Using confocal microscopy TfR2 stained abundantly with a vesicular pattern in undifferentiated Caco-2 cells. No colocalisation with TfR was observed by dual-label fluoresence studies confirming the specificity of the TfR2 antibody. HFE colocalised with TfR2 in an endosomal compartment following addition of iron-saturated transferrin to the culture medium.

Identification of TfR2 in small-intestinal crypt cells and the known effect of disabling mutations in the cognate gene would support a key regulatory role for this receptor in intestinal iron absorption. Colocalisation of endogenous TfR2 and HFE suggests functional coupling in an endosomal transport pathway for crypt cell iron signalling.


D.H.L. Ng, A.N.J. May, S.A. Wootton, A.A. Jackson, M.A. Stroud.Institute of Human Nutrition, University of Southampton, Southampton General Hospital, Southampton, UK

Background/Aims: Ileostomy patients, especially those with additional small bowel disease or resection, may have poor absorption or excess losses of calcium (Ca) and Vitamin D (Vit D) and may also be current or past users of steroid. Furthermore, they may also have magnesium (Mg) malabsorption and excessive Mg losses which could exacerbate bone demineralization (60% of total body Mg is in bone). Ileostomy patients are therefore at risk of osteoporosis but the extent of this risk has not been documented. The aims of this study were to examine bone mineral density (BMD) in ileostomy patients and its relationship with markers of Ca, Mg and Vit D status.

Methods: BMD of lumbar spine (LS) and right femoral neck (FN) were determined using DEXA in 57 unselected ileostomy patients (26–85 yr; 24F, 33M) including 13 (7F, 6M) who had had additional small bowel resection. Both plasma and 24-hour urinary excretion measures of Ca and Mg were made in all subjects along with circulating Vit D levels.

Results: 20 subjects (35%) had low BMDs (Z-score <-1.0) at LS or FN compared to an age matched reference population. However, 28 subjects (49%) had osteopenia (-1.0<T-score<-2.5) and 7 (12%) had osteoporosis (T-score<-2.5) by WHO definition. More patients with small bowel resection had Z-scores <-1.0 compared to those with colectomy alone (62% vs. 27%; p<0.05) and the mean LS BMD was also lower in this group (-0.592 vs. 0.221; p<0.05). Only 3 subjects (5%) had low plasma Mg (<0.7mmol/l) but 34 (60%) had low 24-hour urinary Mg (<3.3 mmol) suggestive of depleted total Mg stores. The mean BMDs were lower in these subjects compared to those with normal Mg excretion (LS Z-scores -0.302 vs 0.582, p<0.05; FN Z-scores -0.273 vs. 0.495, p<0.05) whereas abnormalities in plasma Ca and Vit D and urinary Ca excretion were fewer and had no apparent relationships to BMD.

Conclusions: Our results suggest that 1. patients with colectomy alone are not at increased risk of low BMD but this risk may be increased by additional small bowel resection and 2. many ileostomy patients have depleted Mg stores which may adversely affect BMD. Further studies on the relationship between Mg status and bone density are needed.


C.W.Y. Lai, R. Barlow, M. Barnes, A.B. Hawthorne.Departments of Gastroenterology, Radiology and Nutrition Support Team, University Hospital of Wales, Heath Park, Cardiff, UK

The success rate of unguided nasojejunal feeding tube (NJT) insertion is low, thus often requiring endoscopic or radiological assistance. The spiral end of the Bengmark NJT (spiral NJT) is supposed to aid post pyloric placement, but no comparative trial has been performed.

Methods: Patients requiring nasojejunal feeding were randomised to have either Medicina (straight) or Bengmark (spiral) NJT after stratification into those with or without normal gastric emptying. NJTs were placed at bedside in a standard fashion without radiological guidance by the same person (CWYL) for pre- and / or postoperative feeding. Bolus IV metoclopromide (10mg) was given prior to insertion in the abnormal gastric emptying group. Abdominal X rays were obtained at 4 and 24 hours. Successful placement was defined on X ray as passage through the duodenum from right to left, past the left border of the vertebrae.

Results: 47 patients were randomised of which 17 (11 straight, 6 spiral) could not tolerate the NJT. Of the 30 remaining patients, 16 had normal gastric emptying. Patients having straight or spiral NJT were well matched. Successful placement at 24 hours was achieved significantly more frequently with the spiral NJT (see table), mainly because of a higher success rate with the spiral NJT in the normal gastric emptying group. There was no difference in the duration of NJT in situ, complications, post-insertion hospital stay or final outcome.

Abstract 181

Conclusion: Bengmark spiral NJT should be used for bedside unguided post pyloric feeding.

(We are grateful to Nutricia for supplying the Bengmark tubes.)


S. Cherian, A. Cherukuri, P. Singh.Staffordshire General Hospital, Stafford, UK

Introduction: Previous studies suggest endoscopic balloon dilatation in gastric outlet obstruction from PUD does not achieve long-term remission and most patients eventually require surgery. However, these studies have not addressed the issue of altering the natural history of the underlying PUD.

Methods: We examined medical notes of 18 patients with PUD related gastric outlet obstruction treated by a single consultant gastroenterologist. In all patients, an attempt was made to establish and treat the aetiology of PUD. Where no cause was found or its removal not possible or where disease relapsed, long-term maintenance antisecretory therapy was given.

Results: Of the 18 patients, one presented with aspiration pneumonia and another with stroke and both succumbed to their illness. Of 16 available for follow up, 6 were men and 3 were smokers. Their median age was 69 years (range 43–94). Fourteen patients were treated with TTS balloon dilatation and drug therapy and 2 with drug therapy alone. The median number of dilatations was 2 (range 0–5). There were no complications from dilatation. The causes of PUD were as detailed in table 1. Nine of the 10 HP positive patients received eradication therapy. Eradication was confirmed in 5. NSAIDs were discontinued. Four patients stayed on aspirin for medical reasons. Remission was achieved in all 16 patients with a median follow-up of 30 months (range 5–54) including 2 who died from unrelated illness after being in remission for nearly 2 years. Three patients became asymptomatic without need for maintenance therapy (2 after successful HP eradication, and 1 after withdrawal of NSAID). The remaining 13 required long term maintenance therapy for the reasons detailed in table 2.

Abstract 182, Table 1

Abstract 182, Table 2

Conclusions: PUD related gastric outlet obstruction can be kept in long-term remission by using a structured approach combining dilatation and removal of the cause of PUD and/or maintenance antisecretory therapy. TTS balloon dilatation is a simple, effective, and safe procedure.


A.M. Thompson, R. Stuart for the Scottish Audit of Gastric and Oesophageal Cancer.

The population based Scottish Audit of Gastric and Oesophageal Cancer (SAGOC) accrued data over a 2 year period on 3,293 patients of whom 948, predominantly with oesophageal or junctional cancer, received endoscopic palliative treatment (EPT).

Stent placement alone (506 patients) or LASER treatment alone (117 patients) were popular, but combination approaches supported by radiotherapy (188 patients) or chemotherapy (134 patients) were also administered. There was significant variation in delivery of EPT by healthboard of residence (chi square test, P<0.001), but not by deprivation quintile.

Complications were recorded in 221/948 patients (23%) and were associated with multiple treatments (P<0.001). Oesophageal perforation was uncommon and occurred in 23 patients post stent and 3 patients after LASER.

Stent alone was used for the relief of grade 3 or 4 dysphagia, stent and radiotherapy for grades 2,3,4 and LASER for grades 1,2,3. The majority of patients (>65%) had normal physical activity or only strenuous activity restricted before undergoing EPT. Stents were deemed by the consultant looking after the patient to have been used appropriately for 95% of patients and LASER for 83% of patients.

Survival for all the patients receiving EPT was 40% at 6 months, 17% at 12 months 10% at 18 months and 6 % at 24 months, suggesting benefits from EPT intervention in patients with advanced disease.

While there may be differences between the symptoms and staging of patients receiving different types of EPT, there is evidence for regional variations in the approaches used. However, appropriate use of intervention may provide good palliation for several months.


S.J. Panter1, M.G. Bramble1.1James Cook University Hospital, Middlesbrough and CIHR, University of Durham, UK

Background: It is recommended that all gastric ulcers are biopsied to exclude malignancy with follow up endoscopies performed at 6–8 week intervals until healing is seen.

Aims: This prospective study aimed to identify whether these recommendations were being followed – especially at emergency endoscopy.

Methods: All patients over a 20-month period diagnosed with a gastric ulcer were identified and those with a definite ulcer (with a mucosal breach documented as >5mm) diagnosed at endoscopy were included in the study. A record of the macroscopic judgement of the ulcers as being benign, suspicious or malignant as stated by the endoscopist was made. This was correlated with the histology results.

Results: 250 patients were reportedly diagnosed with gastric ulcers. Of these 191 met the inclusion criteria. The male: female ratio was106: 85 and the mean age at diagnosis was 67 years (range 23–98). Of these 11 ulcers were diagnosed operatively, 79 were diagnosed as an emergency “bleeder” and 99 were diagnosed routinely (29% open access; 71% routine list via OP clinic). Of the “bleeders” only 55% had biopsies taken at the initial endoscopy, with the mean number of biopsies per ulcer being 2.7. This was despite only 30.4% actually requiring injection at the time of the endoscopy. Of those ulcers diagnosed routinely, biopsies were taken in 94.4%, with the mean number of biopsies being 3.5. Of the “bleeders” 6 patients required laparotomy for further bleeding of whom 1 had had a single biopsy taken at endoscopy. Overall 126 ulcers were thought to be benign, 28 to be suspicious and 17 frankly malignant. Of those ulcers diagnosed at an emergency endoscopy these figures were 60, 11 and 3 respectively. The predictive value of the macroscopic judgement can be estimated as the proportion of correct macroscopic diagnoses. The overall predictive value (PVpos) of a macroscopic judgement of definite or suspicious of malignancy was 0.52 (24/46) and the overall predictive value (PVpos) of a macroscopic judgement of benign was 0.96 (121/126). For the “bleeders” the PVpos were 0.43 and 0.97 respectively.

Conclusion: These results show that biopsies are often not taken at the time of an emergency endoscopy. However in this study only 2% of “benign-looking” bleeding gastric ulcers were ultimately diagnosed as malignant. Therefore the priority for benign looking bleeding gastric ulcers remains to establish haemostasis.


J.C. Brooker, S.G. Shah, B.P. Saunders.St Mark's Hospital, Harrow, UK

Background: Based on current knowledge, population screening and the removal of colonic adenomas would be an effective means of reducing the incidence of colorectal cancer. However, a sound understanding of the distribution of adenomas and risks of polypectomy is a prerequisite to implementation, and will assist in the selection of the optimal screening modality and the safest and most effective methods for polypectomy. We therefore studied polyp epidemiology and polypectomy techniques and complications in patients from a single endoscopy department.

Methods: Records of 938 patients (506 males; mean age 58.8 years [sd 14.3]; indications: symptoms 535, neoplasia surveillance 387, polyposis 55, IBD 37, not recorded 9) who had undergone 1023 consecutive colonoscopies with polypectomy during a 22-month period, were examined retrospectively. Complications were identified using a postal questionnaire.

Results: 2806 polypectomies were performed, 37.9% by hot-biopsy, 30.6% by snare and 23.3% by Argon Plasma Coagulation (APC). 44.7% were benign adenomas and 47.8% of these (excluding polyposis) were located proximal to the splenic flexure. However only 27% of advanced adenomas (size >1cm or villous histology) were proximal. Polypectomy failed to yield a specimen for analysis in 19.3% of snares and 5.2% of hot biopsies (p<0.0001). 67.6% patients replied to the complications questionnaire. There were no deaths, but 4 significant complications were identified, including one perforation after hot-biopsy (0.16%) and three episodes of major haemorrhage (0.47%). Intra-procedural bleeding requiring endoscopic therapy was significantly associated with aspirin ingestion (p=0.02), but post-procedure bleeding was not (p=0.3). No complications were reported after polypectomy using APC.

Conclusions: These data support previous observations of a proximal shift in the distribution of colorectal adenomas, although most advanced neoplasia were still found in the left colon within reach of the flexible sigmoidoscope. The incidence of polypectomy complications was lower than previously reported, which may reflect improved technique. Hot-biopsy and APC both appear to be safe for the removal of small polyps, but snare techniques, although safe, frequently fail to yield tissue for histology. No increase in secondary haemorrhage was seen in patients taking aspirin.


J.C. Brooker, S. Thomas-Gibson, S.G. Shah, C.J. Thapar, B.P. Saunders.St Mark's Hospital, Harrow, UK

Background: Colonoscopy withdrawal technique has been shown to be important in the detection of adenomas. We aimed to survey current practice in a single UK endoscopy department, as well as assessing the effect of video recording on examination quality.

Method: With ethics committee approval and patient consent, consecutive routine colonoscopy extubations were video recorded using a remote, closed-circuit TV system. Endoscopists were informed that recording would take place when a “recording-light”, conspicuously placed in the endoscopy room, was illuminated. However recording took place continuously. Endoscopy lists were randomly assigned to “open” or “blind” recording. The video footage was reviewed by a blinded panel of 3 experienced colonoscopists. Extubations were scored by consensus for seven segments of the colon using 5 parameters (looking behind folds, cleaning pools, adequacy of distension, time spent inspecting and quality of bowel preparation) using 50mm visual analogue scales. Patients with IBD, previous colonic resection, current colorectal-cancer or mean preparation score <30, were excluded from the analysis.

Results: 93 procedures by 16 endoscopists were included, 50 for colonic symptoms and 43 for neoplasia surveillance or a family history of cancer. 52.7% extubations attained a mean score of less than 30 (an arbitrary minimum score for acceptable quality). There was no difference in extubation scores or time with blinded compared with open video recording (see table). There was also no difference between groups in performance of retroflexion in the rectum (12% overall) or the detection of adenomas (p=0.24).

Abstract 186

Conclusions: Over 50% of extubations were sub-optimal, and awareness of video recording made no difference to quality. We suggest that both better training in withdrawal technique and allotting more time per procedure are required to improve examination quality.


N. Zakaria, J. Wendon, R. Arya, J. Devlin.King's College Hospital, London, UK

Introduction: The aetiology of severe pregnancy-related disorders remains poorly characterised. One hypothesis is that the clinical disorders recognised - Acute Fatty Liver of Pregnancy [AFLP], HELLP syndrome and Veno-occlusive disease - represent microangiopathic disorders. Thrombophilic disorders have previously been implicated in serious hypertensive complications of pregnancy where microangiopathy is evident. Accordingly, we tested the hypothesis that a higher incidence of genetic thrombophilia would be present in this population.

Methods: Twenty-eight patients with a history of peri-partum liver failure were tested (AFLP 22, HELLP syndrome 3, and Veno-occlusive disease 3. Median age 30 years (range 21–36). Twenty-eight patients were Caucasian and five Afro-Caribbean. All presented in the third trimester. Only one patint had a known pre-existing pro-thrombotic disorder. All patients had thrombophilia screens performed post-partum once fully recovered from their illness (factor V leiden [FVL] and prothrombin G20210A gene mutations, anticardiolipin antibody and lupus anti-coagulant, anti-thrombin III and protein C and S levels).

Results: A pro-thrombotic disorder was present in 12/28 (43%) patients. FVL heterozygosity was present in 5/23 Caucasian patients (17.5%). PT G20210A gene heterozygosity was present in 2/28 (7%). Anti-cardiolipin antibody was detected in 4/28 (14%). Lupus anticoagulant was detected in one patient with AFLP. In the patients where a diagnosis of AFLP was made; genetic thrombophilia was present in 10/23 (42%). Protein C and S deficiency was not detected in patient.

Conclusion: The above results lend some support to the hypothesis that microvascular thrombosis may play a pathogenic role in a subgroup of patients with peri-partum liver failure. However further predisposing factors for these serious disorders require to be elucidated.


M.B. Lewis, A. Stroud, M. Davies, P.D. Howdle.St James's University Hospital, Beckett Street, Leeds, UK

Introduction: Patients with liver failure who are not clinically encephalopathic can show evidence of neuropsychological impairment. Inadequate treatment of previous hepatic encephalopathy is often blamed. Repeated episodes of hepatic encephalopathy (HE) are thought to underlie the aetiology of the acquired hepatocerebral degeneration syndrome (AHCD) which consists of cognitive impairment associated with, tremor, ataxia, dyskinesias and various motor abnormalities.

Method: Patients listed for liver transplantation with no past history of hepatic encephalopathy were invited to participate in the study. Standard demographic data and patient characteristics were recorded. A battery of neuropsychological tests was used to assess cognitive function and a structured neurological examination was performed. Twenty-five healthy volunteers were used as controls. The data were compared using the unpaired t-test or the Mann-Whitney U test as appropriate.

Results: Twenty patients were included. The mean age was 53.25 years (sd 9.35). The median duration of illness was 60 months (iqr 6–156) and the median Childs-Pugh score was 7 (iqr 5–8). There were no significant differences between baseline characteristics of patients and controls. Highly significant global impairment of cognitive function was detected, with memory and visuo-spatial problems being prominent. No abnormalities were detected with the commonly used trailmaking test. Twelve patients had an abnormal neurological examination displaying many of the physical signs associated with AHCD.

Discussion: Cognitive and neurological dysfunction occurs in patients with end stage liver failure in the absence of previous HE. It is unlikely that the progressive neurological and cognitive decline that can occur in these patients is due to repeated episodes of HE. The impairment that occurs is severe, and affects many aspects of cognitive function. This has implications for the pre-operative counselling of these patients and for obtaining informed consent. Current methods of monitoring cognitive impairment clinically, such as trailmaking tests, may be inadequate.


A.K.P. Lim1,2, N. Patel1,2, G. Hamilton1,2, M.J.K. Blomley1, D.O. Cosgrove1, G.R. Foster2, R.D. Goldin3, H.C. Thomas2, S.D. Taylor-Robinson1,2.Departments of1Imaging,2Medicine, and3Histopathology, Imperial College School of Medicine, Hammersmith Hospital, Du Cane Road, London;2,3St Mary's Hospital, Praed Street, London, UK

Purpose: Non-invasive assessment of the severity of diffuse liver disease is problematic and biopsy is often needed. We evaluated hepatic vein arrival times (AT) of an ultrasound microbubble agent and carotid-delay times (CDT) in grading diffuse liver disease in patients with chronic hepatitis C (HCV) infection using histology as the gold standard.

Materials/Method: 51 untreated and 10 interferon-treated patients with biopsy-proven HCV liver disease were studied. Time-intensity curves of hepatic vein spectral Doppler signals and audio intensity from the carotid artery were analysed after an intravenous microbubble bolus. CDT was calculated as the difference between carotid and hepatic vein arrival times. Grading of fibrosis (F) and inflammatory activity (NI) was carried out using the modified HAI (Ishak) scoring system. Patients were divided into mild hepatitis, (F≤2/6, NI≤3/18); moderate/severe hepatitis, (F≥3/6, NI≥4/18) and cirrhosis, (F=6).

Results: There was a monotonic decrease in the mean ATs ± 1 s.d. and CDTs ± 1 s.d. for mild, moderate/severe hepatitis and cirrhosis: 50.5s ± 24.2, 33.6s ± 26.1, 14.8s ± 4.5 and 36.8s ± 29, 18.7s ± 22.6 and 5.8s ± 4.9 respectively (Kruskal Wallis ANOVA, p<0.001). An AT<24s and CDT<10s was 100% sensitive for cirrhosis but only 69% and 72% specific respectively, as subjects with fibrosis sometimes showed early AT and CDT. 10 interferon-treated patients showed earlier arrival times than comparative untreated subjects.

Conclusion: AT and CDT measurements hold promise in characterising liver disease in patients with HCV and is a highly sensitive marker of cirrhosis. Treatment with interferon appears to prolong AT. This is important for disease monitoring and may be useful in assessing the efficacy of treatment regimes non-invasively and possibly replace repeat liver biopsy in some situations.


K. Dabos, H.R. Whalen, P.N. Newsome, J.A. Parkinson, N.C. Henderson, I.H. Sadler, P.C. Hayes, J.N. Plevris.

A porcine model of paracetamol induced acute liver failure has been recently developed in our laboratory. This model is suitable for pathophysiological studies in acute liver failure. As a first step we investigated the gluconeogenic capacity of the porcine liver in that particular model.

Materials and Methods: Thirty five kilogram large white pigs were maintained under general anaesthesia with isoflurane and nitrous oxide. Three pigs acting as controls received no paracetamol while five other pigs received paracetamol by intravenous infusion for 12 hours keeping blood levels between 200 and 300 mg/L. Blood glucose was maintained within normal limits by continuous intravenous dextrose infusion. Using 1H NMR spectroscopy we measured concentrations of lactate, pyruvate, threonine, glycine and alanine at 5hourly intervals until the experiments were terminated. Experiments lasted for 28 hours and any surviving animals were then euthanased.

Results: In control pigs there were no significant differences in the concentrations of those substrates at any time point sampled. Animals who received paracetamol showed significant increases in the concentrations of lactate, pyruvate and the amino acids. Increase of lactate became significant at 15 hours and at 25 hours compared to t=0 an average increase of 405% was seen (p<0.003). Increase of pyruvate became significant at 20 hours and at 25 hours compared to t=0 an average increase of 150% was seen (p<0.018). Increase of threonine became significant at 20 hours and at 25 hours compared to t=0 an average increase of 82% was seen (p<0.048). Increase of glycine became significant at 5 hours and at 25 hours compared to t=0 an average increase of 390% was seen (p<0.005). Finally, increase of alanine became significant at 10 hours and at 25 hours compared to t=0 an average increase of 410% was seen (p<0.002).

Conclusion: In this model all gluconeogenic substrates studied are significantly increased whereas the end product of the pathway , glucose, is significantly decreased. This confirmed that in the model the gluconeogenic capacity of the liver is lost and further studies in humans are required to assess the observed phenomenon.


I. Mohammed1, L. Cherkas3, S.A. Riley2, T.D. Spector3, N.J. Trudgill1.1Sandwell General Hospital, West Midlands;2Northern General Hospital, Sheffield;3Twin Research Unit, St Thomas' Hospital, London

Background: A number of family pedigrees detail multiple members with gastro-oesophageal reflux disease (GORD). Aggregation of GORD symptoms within families of patients with documented GORD has also been demonstrated. This raises the possibility of a significant genetic contribution to the aetiology of GORD. We have therefore studied GORD symptoms in monozygotic (MZ) (100% of genes shared) and dizygotic (DZ) (approximately 50% of genes shared) twins to assess the contribution of genetic factors to GORD.

Methods: 4480 unselected twin pairs from a national volunteer twin register were asked to complete a previously validated questionnaire. GORD was defined as symptoms of heartburn or acid regurgitation at least weekly during the past year.

Results: 5032 respondents (56% response rate), including 1940 evaluable twin pairs. 922 MZ pairs (86 male, 836 female, median age 53(range 19–81)years) and 1018 DZ pairs (71 male, 947 female, age 54(20–82)years). The prevalence of GORD among the twins was 709/3880 (18%). Both pairwise and casewise concordance rates were significantly higher for MZ twins (see table). Heritability estimates suggest 50% (95%CI 39–61%) of the phenotypic variance in GORD is due to additive genetic factors (see table).

Abstract 191

Conclusion: This study strongly suggests a substantial genetic contribution to the aetiology of GORD.


S.H.C. Anderson, A. Anggiansah, M. Aratsu, R. Anggiansah, W.J. Owen.Oesophageal Laboratory, Guy's and St Thomas' Hospital, London, UK

Background: Achalasia characteristically presents with dysphagia and regurgitation. Food fermentation in the dilated oesophagus, or oesophageal distension, have been reported to cause heartburn, which may be misdiagnosed as gastroesophageal reflux disease (GORD).

Methods: We studied the medical notes of all patients diagnosed with achalasia in our laboratory over the past 10 years, and documented the onset and pattern of heartburn and other symptoms. Where available, the lower oesophageal sphincter pressure (LOSP) and pH studies were compared.

Results: 306 patients had a manometric diagnosis of achalasia. 81 were excluded having had a previous dilatation or surgery. Of the remaining 225, 10 (4%) were classified as vigorous achalasia and 5 had prior diffuse oesophageal spasm. The mean duration of symptoms was 3.5 years (range 1 month to 30 years). 110 (49%) experienced heartburn, 220 (98%) had dysphagia, 131 (58%) had regurgitation, 126 (56%) had weight loss and 87 (39%) had non-heartburn chest pain. In 54 patients (24%) the heartburn preceded the dysphagia and persisted, while in 25 patients (11%) the heartburn stopped with the onset of dysphagia. Heartburn developed after the onset of dysphagia in 46 patients (20%) and in 17 (8%) the symptoms began together. The mean LOSP was 20.2±9mmHg. There was no significant difference in LOSP between these different symptom groups. 106 patients (47%) were taking anti-dyspepsia medication, 58% of these being acid suppression therapy. A 24-hour pH study was performed in 58 patients. This showed GORD in 6 (10%) and was normal in 50. Two patients showed a pH drift to around pH4 but above pH3, indicating food fermentation in the dilated oesophagus. There was no correlation between the pH study and the pattern of symptoms or the LOSP.

Conclusion: There is a long delay in reaching a diagnosis of achalasia – up to 30 years in this study. The symptoms may not be characteristic, and weight loss, chest pain and heartburn are frequent symptoms. GORD is uncommon despite heartburn being present in half of patients. Reflux symptoms in these patients are therefore unreliable and should be investigated with a 24-hour pH study.


D.A. FreshwaterA, D. Al-Dulaimi, P. O'Hare, H. Randeva, C.U. Nwokolo.University Hospitals of Coventry & Warwickshire, Clifford Bridge Road, Coventry CV2 2DX, UK

Background: Decreasing H. pylori prevalence has been linked with increasing incidence of oesophageal adenocarcinoma. Increasing obesity leading to gastro-oesophageal reflux, Barrett's and oesophageal adenocarcinoma is one possible route of association. H. pylori gastritis is associated with increased gastric mucosal leptin and successful cure leads to a fall in gastric mucosal leptin. However, for leptin produced by the stomach to influence appetite and cause significant obesity a systemic effect should be detectable.

Methods: Ten H. pylori positive healthy subjects were studied before and after cure of H. pylori. Their mean age was 36.8 years and mean Body Mass Index was 25.80. After an overnight fast they were admitted to a research ward at 8 am. Blood was sampled hourly for six hours via an indwelling IV cannula. Plasma was separated and stored at -20 oC. Leptin was measured in duplicate by using the Linco Research Human Leptin Radioimmunoassay. 6 hour integrated leptin was calculated using the trapezoid rule. Paired comparison was made by the Wilcoxon rank sum test.

Results: Before H. pylori cure, median 6 hour integrated plasma leptin was 23.66 mcg/ (7.33 – 128.22), and after cure it was 26.75 mcg/ (7.43–72.81) p = 0.375.

Conclusions: Plasma leptin is unchanged following cure of H. pylori in healthy subjects. Decreased gastric mucosal leptin associated with cure of H. pylori does not affect plasma leptin. This study weakens the hypothesis that a fall in gastric leptin production may be responsible for increased appetite and obesity following H. pylori eradication.


R. Harvey, A. Lane, L. Murray, J. Donovan, M. Egger, I. Harvey, S. Nair.Frenchay Hospital, Bristol; Department of Social Medicine, University of Bristol, UK

Helicobacter pylori eradication therapy has been reported in some studies to result in increased heartburn and acid reflux. However, other studies show either no effect on these symptoms, or even a beneficial effect. We have investigated these contradictory findings, as part of the community-based Bristol Helicobacter Project.

Methods: 10,537 people aged 20–59 years gave informed consent to take part in the Bristol Helicobacter Project, a community-based prospective randomised controlled trial of the effects of H. pylori eradication. 1634 participants had a positive 13C-urea breath test, and were treated with either H. pylori eradication therapy (ranitidine bismuth citrate 400mg and clarithromycin 500mg twice daily for two weeks) or placebo. The prevalence, frequency and severity of heartburn and acid reflux were measured at baseline and two years after randomisation, using a validated questionnaire.

Results: There was an overall small benefit of active treatment, with 3.1% less heartburn and 2.5% less reflux at 2 years when compared with placebo. However, this small net benefit concealed complex differential effects. Active treatment had a more marked benefit over placebo in participants with mild or no initial symptoms - at 2 years, 6.8% fewer had heartburn and 4.3% fewer had reflux. Those with initially moderate symptoms showed little net benefit, and those with troublesome symptoms at randomisation were more likely to get worse after active treatment. Subjects with the most severe heartburn two years after treatment were almost twice as likely to have been treated with active therapy.

Conclusions: (1)There is a small net reduction in heartburn and acid reflux after H. pylori eradication therapy, but a significant subgroup of patients get worse, particularly if they have troublesome GORD symptoms initially. (2) The contradictory results of previous studies may reflect differences in the selection of patients.


S.A. Khan1, P.L. Carmichael2, S.D. Taylor-Robinson1, N. Habib3, H.C Thomas1.1Liver Unit, St Mary's Campus;2Department of Biological Chemistry, South Kensington Campus;3Department of Surgery, Hammersmith Campus, Faculty of Medicine, Imperial College, London, UK

Background and Hypothesis: Reported mortality from cholangiocarcinoma (CCA) has risen steeply in the UK and other industrialised countries over the past 20–30 years, the cause of which has not been adequately explained. DNA adducts are covalently modified bases resulting from carcinogen binding at the nucleotide level. Adduct formation is pro-mutagenic and clearly demonstrates exposure to a DNA damaging agent. It is a key step in toxin-induced carcinogenesis. We hypothesise that the increase in CCA mortality is caused by a temporally-associated rise in genotoxic environmental agent(s), causing cholangiocyte DNA damage.

Aims: To investigate and compare tumour and tumour-adjacent CCA tissue, and non-cancer control bile duct tissue, for the presence of DNA adducts as a biomarker of genotoxin exposure.

Methods: DNA from 28 CCA tissues, and in 24 cases adjacent non-tumour tissue samples from the same patients; and from bile ducts from 7 non-cancer patients (undergoing laparoscopic cholecystectomy for gallstones) were investigated for the presence of DNA adducts using the nuclease P1 method of 32P-postlabelling. Relative adduct labelling values (RAL, adducts/108 nucleotides) quantified.

Results: No difference was found in RALs between DNA from CCA tissue (mean 14, range 1–48) and tumour-adjacent tissue DNA (mean 14, range 1–52). RALs were significantly higher in tissue from CCA patients than from non-cancer patients (mean 6, range 1–31, p=0.04, Mann-Whitney test). Different adduct patterns were also seen CCA compared to non-cancer patients.

Conclusion: Quantitative and qualitative differences in adducts between cancer and non-cancer patients support the hypothesis that genotoxins play a role in the development of CCA.


F.W. Shek, A.B. Oyetade, P.A. Johnson, F.M.J. Walker, A,C. Bateman, C.D. Johnson, J.P. Iredale, D.R. Fine.Liver and Pancreatic Finrosis Group, University of Southampton, UK

Pancreatic Stellate Cells (PSCs) are central to pancreatic fibrosis. Our group have previously shown that recovery from liver fibrosis can occur and it is associated with apoptosis of Hepatic Stellate Cells. Nerve Growth Factor (NGF) stimulated apoptosis by activating the Low Affinity Nerve Growth Factor receptor (p75). We therefore studied the distribtion of NGF and its receptors in sections of human chronic pancreatitis (CP). We further studied the effects of NGF stimulaton of apoptosis in PSCs and the expression of p75 receptor.

With immunostaining techniques we examined the distrbution of NGF and p75 in sections of CP and normal pancreas. Passaged rat PSCs were used in all the experiments. PSCs were treated with NGF (1–100ng/ml) and apoptosis was assessedby nuclear morphology after cell staining with acridine orange. The proliferation rate of PSCs were determined by 3H-thymidine incoporation. Expression of p75 was assessed by Western blotting.

Immunostaining of the chronic pancreatitis tissues demonstrated that p75 was positively expressed in the fibrotic bands localised in the distribution of PSCs whereas NGF was expressed in the atrophic acinar parenchyma.

Exogenous NGF(10, 100ng/ml) significantly increased apoptosis under serum-free conditions by 42±11% and 30±8% as of control. The rate of proliferation of PSCs when treated with the same concentration of NGF(10, 100ng/ml) was also reduced by 57±0.8% and 73±2% respectively as compared to control. with western blotting we demonstratd that p75 was espressed in PSCs.

We therefore conclude that the expression of p75 on PSCs and the response to NGF stimulation by apoptosis may suggest a role in recovery of pancreatic fibrosis.


M.A. Karajeh, M.G. Ashton, M.J. Grundman.Gastroenterology Department, Chesterfield and North Derbyshire Hospital, Calow, Chesterfield S44 5BL, UK

Background: Endoscopic sphincterotomy and stone extraction is an established treatment for symptomatic common bile duct (CBD) stones. Surgery is recommended when endoscopic extraction of stones fails. For elderly and/or debilitated patients who are at high surgical risk, long-term biliary stenting may have a role as a definitive therapy.

Methods: A retrospective analysis was conducted of all patients who had biliary stents for retained CBD stones over a 7-year period (January 1993 to December 1999). 30 patients were identified (21 women, 9 men; median age 84 years, range 49–95 years). 7/30 (23%) patients had previous cholecystectomies. Follow-up data were obtained by referral to their case notes and contacting their general practitioners.

Results: Successful biliary drainage was achieved in all patients. The stent was considered to be a temporary measure in 1 patient while awaiting surgery and definitive in 29. Only 1 patient had subsequent surgery (elective cholecystectomy and choledocholithotomy). Early complications occurred in 2 patients (6.7%): both subsequently died. Late complications occurred in 5/30 (16.7%): cholangitis 4, recurrent jaundice 1. All of these had repeat endoscopic retrograde cholangiopancreatography with successful stone extraction in 2 and re-stenting in 3. During follow-up there were 5 unrelated deaths. The remaining 17/30 (57%) were well and asymptomatic at a median follow-up period of 20 months (range 10–63 months).

Conclusion: Endoscopic biliary stenting for irretrievable CBD stones is an effective method of establishing bile duct drainage as definitive treatment for patients at high surgical risk.