Article Text


Gastroduodenal free papers 026–032

Statistics from


J. Deeks, B. Delaney2, D. Forman1, P. Moayyedi2. ICRF/NHS Centre for Statistics in Medicine, Oxford, UGPD Cochrane Group; 1University of Leeds; 2Primary Care Sciences Building University of Birmingham, UK

Objectives: We have published a Cochrane systematic review on the efficacy of H. pylori eradication therapy in non-ulcer dyspepsia (NUD). We reported that this intervention had a statistically significant effect in curing dyspepsia symptoms. A US systematic review suggested there was no significant effect of H. pylori eradication therapy on NUD symptoms. We explored reasons for these discrepant results.

Results: We identified six differences in methodology. The US review included all dual, triple and quadruple H. pylori eradication therapies, searched until December 1999, did not contact authors, included abstracts, assumed drops outs were treatment failures and analysed results as odds ratio for cure. The Cochrane review included all therapies proven to successfully eradicate H. pylori, searched until May 2000, contacted authors, only included abstracts if further information was available, excluded drop-outs from the analysis and analysed results as relative risk of remaining dyspeptic. The influences these factors had on the conclusion of the review are outlined in table. Excluding trials published in 2000 had a major impact on the results, reducing the number of trials in the review and widening 95% confidence intervals. Use of the odds ratio increased heterogeneity and a random effects model yielded a non-significant overall effect. Other differences in methodology did not make a difference in this instance.

Abstract 026

Conclusions: The results of this review in a fast developing field depend on inclusion of all relevant articles. The ability to continually update Cochrane reviews ensures that they are the more appropriate format for publishing reviews in research areas that are fast evolving.


H.R. Pearce1, N.J. Brown1, K.D. Bardhan2, J.C. Atherton3, N. Kalia1. 1Surgical & Anaesthetic Sciences, Royal Hallamshire Hospital, University of Sheffield; 2District General Hospital, Rotherham; 3University of Nottingham, Nottingham, UK

Background/Aims: Helicobacter pylori (H. pylori) infection is associated with delayed healing of peptic ulcers which in turn is dependent upon angiogenesis or new blood vessel formation. Proliferation and migration of endothelial cells (ECs) are crucial stages of angiogenesis. This study aimed to determine whether H.pylori inhibited these two processes in vitro.

Methods: Extracts of three H. pylori strains were tested on human umbilical vein ECs: a cagA+ vacA s1/m1 (toxigenic) strain, its VacA-isogenic mutant (non-toxigenic) and a cagA- vacA s2/m2 (non-toxigenic) strain. Campylobacter jejuni and Eschericia coli were also tested. To determine proliferation, ECs were exposed to extracts for 24, 48, 72 and 96 hrs. An MTT proliferation assay quantified EC viability and Hoescht/Propidium Iodide staining identified apoptotic, necrotic and viable ECs. Migration in response to vascular endothelial growth factor, (VEGF) was assayed over 4.5 hrs using a microchemotaxis chamber following a 24 hr pre-incubation period. Relevant controls were performed in all cases.

Results: Control ECs significantly proliferated at, 72, and 96 hrs (P<0.01). No proliferation was observed with the 3 H.pylori strains or C.jejuni. ECs treated with E.coli showed similar proliferation to controls. No significant increase in apoptotic or necrotic cell number was observed. VEGF significantly increased control migration (P<0.01) which was not inhibited by any of the bacterial extracts.

Conclusion:H.pylori extracts inhibit EC proliferation in vitro by a cytostatic mechanism but do not inhibit EC migration. This anti-proliferative effect is also demonstrated by C.jejuni. Inhibition of EC proliferation may decrease angiogenesis, despite no effect on migration, at the ulcer site which may in turn explain the delay in ulcer healing associated with H.pylori infection.


D. Gillen, A. Wirz, J.E.S. Ardill, J. Crabtree, K.E.L. McColl. University of Glasgow, Scotland, UK

Introduction: The presence of a Cag+ve strain of H. pylori is protective against oesophagitis and GO junction cancer. Some have suggested that this effect may be due to acid hyposecretion in Cag+ves. However, we have previously reported that Cag+ve subjects have a higher degree of hypergastrinaemia than Cag-ves, yet a similar level of acid secretion basally and in response to gastrin stimulation. It remained unclear why the higher plasma gastrin was not leading to an increased acid secretion in Cag+ve infection.

Aims: To determine the effect of Cag status on gastric physiology.

Methods: 15 Cag+ve and 11 Cag-ve H. pylori positive healthy subjects and 27 H. pylori negative healthy subjects had their acid output and serum gastrin measured basally (BAO) and in response to infusion of Gastrin 17 at 7,20,60.180 and 800pmol/Kg/h. This allowed one to calculate their sensitivity to gastrin ie gastrin concentration achieving 50% maximal acid output (MAO).

Results: The Cag+ves had a reduced sensitivity to gastrin compared with both Cag-ves and H. pylori negatives. However, the Cag+ves also have a higher gastrin level resulting in a similar acid output to both Cag+ve and H. pylori −ves (see table).

Abstract 028

Discussion: The higher gastrin and lower sensitivity to gastrin in Cag+ves are likely to be explained respectively by more severe antral gastritis and more severe body gastritis.

Conclusion: Any protective effect of Cag+ve infection in reflux disease cannot be explained by effects on acid secretion but might be explained by effects of hypergastrinaemia.


G. Blackshaw, J. Barry, P. Edwards, G.V. Thomas, M.C. Allison, W. Lewis. Royal Gwent Hospital, Newport NP20 2UB, UK

Background: The recent audit of oesophagogastric cancer in Wales demonstrated that many surgeons continue to undertake small caseloads and revealed an open and close laparotomy rate of 23%. Wider use of laparoscopy was advocated strongly.1

Aims: The aim of this study was to examine the benefit of universal staging laparoscopy in the preoperative staging of gastric cancer and to determine the strength of agreement with the true histopathological stage.

Methods: One hundred consecutive patients [median age 71 years (35–86), 59 male] were studied prospectively. All patients underwent staging computed tomography (Siemens somatom +4) prior to laparoscopy. The strength of agreement between the perceived preoperative radiological stage, the laparoscopic stage and the histopathological stage was determined by means of the weighted Kappa statistic (Kw).

Results: See table.

Abstract 029

Conclusion: Laparoscopy improved the perceived preoperative stage from fair to moderate for T stages and there was a significant twofold improvement from fair to good for M stages. This resulted in an open and close laparotomy rate of 12% rather than the 33% (Chi2 12.65, P<0.0001) that would have resulted without laparoscopy.

1Pye JK, Crumplin MKH, Foster ME, Biffin A, Charles J. One-year survey of carcinoma of the oesophagus and stomach in Wales. Br J Surg 2001;88:278–85.


P. Moayyedi1, S. Duffett2, S. Mason3, J. Brown3, D. Forman4, A.T.R. Axon2. 1Gastroenterology Unit, City Hospital, Birmingham; 2Centre for Digestive Diseases, The General Infirmary at Leeds; 3Northern and Yorkshire CTRU; 4Centre for Cancer Research, University of Leeds, UK

Introduction: Dyspepsia and irritable bowel syndrome (IBS) are associated with reduced quality of life (QoL). The temporal relationship between these events is unclear. We evaluated this in a cohort study.

Methods: This cohort study was nested in a randomised controlled trial that evaluated the clinical benefit of H pylori screening and treatment in the community. Subjects between the ages of 40–49 years were randomly selected to attend their local general practice. H pylori status was assessed by 13C-urea breath test and infected individuals were randomised to eradication therapy or placebo and followed up for two years. QoL was assessed by the Psychological General Well Being Index (PGWBI), dyspepsia by the Leeds Dyspepsia Questionnaire and IBS by the presence of 3 or more Manning's criteria. Assessments were made at baseline and at two years. Reduced QoL was defined as a PGWBI of < 106 (the mean score at baseline).

Results: 32,929 subjects were invited, 8,407 attended and were eligible, 1,769 were H pylori positive and had complete follow-up. Subjects that had dyspepsia or IBS at baseline were excluded. 71/576 (12%) of subjects with a PGWBI ≥ 106 that did not have dyspepsia at baseline had dyspepsia at two years compared with 76/388 (20%) of subjects with PGWBI < 106 (relative risk [RR] = 0.62; 95% confidence interval [CI] = 0.47 to 0.85; p=0.003). 30/772 (4%) of subjects with a PGWBI ≥ 106 that did not have IBS at baseline had IBS at two years compared with 67/622 (11%) of subjects with PGWBI < 106 (RR = 0.36; 95% CI = 0.24 to 0.55; p<0.0001). The associations between reduced PGWBI and subsequent development of dyspepsia and IBS remained in logistic regression models controlling for age, gender, H pylori eradication, NSAID use, social class, smoking coffee and alcohol intake.

Conclusion: Reduced QoL is an important risk factor for the subsequent development of dyspepsia and IBS. Drugs that improve these disorders may not improve QoL as much as cross-sectional surveys suggest.


E. Henry, A. Wirz, C. Mowat, V. Fyfe, K.E.L. McColl. Dept of Medicine and Therapeutics, Gardiner Institute, Western Infirmary, Glasgow G11 6NT, UK

Background and Aims: Vitamin C is denatured in gastric juice of high pH being converted irreversibly to diketogulonic acid. We have examined the effect of the elevation of intragastric pH which occurs during proton pump inhibitor therapy on the bioavailability of dietary vitamin C.

Methods: 29 healthy volunteers (13 female, 15 H. pylori positive) had their fasting plasma vitamin C measured on 4 occasions before and again on four occasions during the last week of a one month course of omeprazole 40mg/day. Vitamin C was measured by HPLC and a mean value calculated for each patient for before treatment and during the fourth week of treatment. 24h intragastric pH was also monitored in each patient before and during the last week of treatment. Dietary intake of vitamin C was measured over the week pre-treatment and last week of treatment using daily food diaries and the Diet 5 dietary analysis programme.

Results: Prior to commencing omeprazole, the mean plasma vitamin C concentration (μg/ml) in the H. pylori −ve subject was 25.1 (range 16.1–33) and substantially lower at 17.4 (6.7–29) in the H. pylori +ve subject (p<0.001). Mean daily dietary intake of vitamin C (mg/day) was also markedly lower in the H. pylori +ves (44, 10–130) versus −ves (141, 23–282) (p<0.001) and in the former below the recommended minimum value of 60mg/day. The 4 week course of omeprazole lowered the mean plasma vitamin C concentration by 15% (p=0.005) and the fall was similar in the H. pylori +ve and −ve subjects. Dietary intake of Vitamin C (mg/day) was the same before (94.7) and during omeprazole treatment (95.3).

Conclusion: Proton pump inhibitor therapy lowers the bioavailability of dietary Vitamin C. This is likely to be of clinical significance in H. pylori +ve subjects who have a deficient dietary intake and low plasma vitamin C concentration pre-treatment. The further reduction in systemic vitamin C in H. pylori +ves during proton pump inhibitor therapy may contribute to their propensity to develop atrophic gastritis during such therapy.


C.L. Sheen1, J. Wang1, J.F. Dillon2, D.N. Bateman3, K. Simpson4, T.M. MacDonald1. 1Medicines Monitoring Unit and 2Dept of Gastroenterology, Ninewells Hospital & Medical School, Dundee; 3Scottish Poisons Information Bureau and 4Scottish Liver Transplant Unit, Royal Infirmary, Edinburgh, UK

Introduction: Ibuprofen, a frequently used analgesic, is available without prescription (over the counter, OTC). Upper gastrointestinal complications (UGIC) ranging from minor dyspeptic symptoms to life threatening events such as haemorrhage and perforation may occur. Risks of UGIC depend on factors such as age, previous history of GI and other comorbid diseases, and the dose of ibuprofen used. We have calculated the excess number of UGIC requiring hospitalisation that may be expected from the amount of ibuprofen sold for OTC use in 2000 in the United Kingdom (UK) for a low risk population.

Methods: The risk for UGIC was calculated from the population in Tayside, Scotland who had redeemed a prescription for ibuprofen (= 1200mg/day, equivalent to the maximum daily dose (MDD) available OTC) between Jan 1989 and Dec 1995, and were low risk for GI events. We linked exposure to hospitalisation for UGIC in these patients exposed and not exposed to ibuprofen. IMS Health (UK) supplied data on the total weight of ibuprofen sold in the UK in 2000. Assuming the UGIC risk in Tayside was the same as the UK, the excess number of UGIC for the estimated OTC use in 2000 was calculated.

Results: The risk of UGIC whilst exposed to OTC MDD ibuprofen was 1.62 events/thousand patient years (TPY) and unexposed was 0.85 events/TPY. Thus, the excess risk was 0.75 events/TPY. 46,000 kg of ibuprofen was sold OTC in 2000. Assuming all usage at the MDD, 81 UGIC would be attributable to OTC ibuprofen exposure. An equivalent of 1.3 events per million population.

Conclusion: There is a small estimated excess risk of serious GI events associated with ibuprofen at doses available OTC. Ibuprofen when used at recommended OTC dosages in a low risk population must be considered very safe.

View Abstract

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.