Article Text


Endoscopy Posters 352–391

Statistics from


M.J. Maida, J. Conibear, M.J. Dew. Gastroenterology Department, Prince Philip Hospital, Bryngwynmawr, Llanelli SA14 8QF, UK

An audit carried out last year examining the degree of informed consent in competent adults found that although the majority of patients understood what the procedure involved and had been interviewed by a health care professional, 34% had poor understanding of the benefits of the test and 75% had little or no idea of any associated risks1. This audit determines whether modifications made to information to include risk sent to patients in the light of these findings led to an improvement in the level of understanding.

118 patients were interviewed. As last year, 97% were aware which investigation was to be performed and the region to be examined. 69%(63% last year) had received information prior to the procedure, with 80% acknowledging receipt of printed information. 99% of patients had spoken to a health care professional, a proportion similar to last year. Only 60% (88% last year) were deemed to have a good understanding of what was to occur during the test and 66% (90% last year) were aware of what was to happen after the procedure. 70% (34% last year) had poor understanding of the benefits of the test. 48%(25% last year) were aware of the risks.

Conclusion: Changing the literature produced an improvement in understanding of associated risks, however, only 30% of patients were aware of the benefits of the test, and less than half were clear on what would happen during and after the investigation. This may reflect inter-observer variation, although the question does arise whether a certain proportion of patients have the cognitive ability to grasp the information presented to them—it may be pertinent to explore this issue further. Asking patients to complete a checklist is one way of ensuring that they have read material sent to them and been given the opportunity to ask questions.

1. Williams L. et al: Gut 2001;48(Supp 1): A94


F. Parente, S. Bargiggia, G. Bianchi Porro. Department of Gastroenterology, L.Sacco University Hospital, Milan, Italy

Background: A regional initiative, called the “Three Day Rule” (TDR), has been recently introduced in our country in order to facilitate the earlier diagnosis of malignancy. It requires patients with suspected severe diseases have a diagnostic procedure done within three working days of referral by General Practitioners (GPs).

Aim: To assess prospectively the effectiveness and compliance with TDR for upper digestive malignancies.

Methods: We studied prospectively patients referred for gastroscopy under the TDR initiative with contemporaneous open access referrals over a 12 month period at a single large teaching hospital of Western Milan. We compared prevalence of malignancies and other serious non-neoplastic diseases as well as the waiting times in the two groups. The appropriateness of the indications for each referral was also reviewed by a gastroenterologist blinded to the outcome of the test.

Results: 142 patients referred for gastroscopy under the TDR scheme and 767 routine referrals were studied. Significantly more oesophageal/gastric cancers (6% v 1%) and serious benign GI lesions (grade II–III oesophagitis or peptic ulcer) were diagnosed in TDR patients in comparison with routine referrals (p<0.05, table). The rate of inappropriate referral was significantly lower in the TDR than in the open access group (39% v 22%) (p<0.01). The estimated cost of the TDR scheme (in extra list examinations alone) was 10 780 Euros, which means about 1198 Euros per diagnosis of cancer but only 229.5 Euros per “useful” diagnosis (including PUD and oesophagitis).

Abstract 353

Conclusions: Significantly more upper GI cancers and serious benign diseases can be found within a short period to comply with the TDR scheme. However, some GPs appear to over read alarm symptoms, and this may lead to some inappropriate referrals. Better awareness of appropriate urgent referral criteria is needed in order to ensure that the best possible use is made of the resources available for this initiative.


D. Flynn1, P. van Schaik2, A. van Wersch2, P. Cann1, A. Douglass1.1Endoscopy Centre, James Cook University Hospital, Middlesbrough, UK;2University of Teesside (Social Sciences), Middlesbrough, UK

Background: Factors that influence GPs' referral decisions for GI complaints have received little attention in the literature. A greater understanding of this process is essential for developing guidelines and improving services, yet retaining aspects of the valuable primary care screening function. This has become ever more important since the implementation of “Two week rule” referrals and the Government's Direct Booking Project Initiative.

Aims & Objectives: To design a scale to measure (and rank order) the influence of medical and non-medical factors upon GP referral decisions for patients presenting with common upper and lower GI symptoms.

Methods: 21 GPs in a single health district were given an 80 item questionnaire separated into potentially potent medical (signs, symptoms, and clinical history) and non-medical factors (patient and GP characteristics—identified from the literature). GPs indicated on nine point Likert scales the relative weighting they assigned to the perceived importance of each factor when making referral decisions (very likely refer, no influence on decision, very likely not refer ..... +4, 0, -4).

Results: The strongest referral drivers (in rank order) were: a positive routine examination, the presence of dysphagia, symptoms consistent with GI referral guidelines, positive faecal occult blood test, and fear of malpractice. Contrary to expectations, patient age, patient anxiety, weight loss, and GPs' experience of GI symptoms had a negligible perceived impact upon referral decisions. The strongest referral inhibitors (in rank order) were: intermittent symptoms, symptoms inconsistent with GI referral guidelines, and the presence of a plausible explanation for presenting symptoms.

Conclusions: The scale was well received as a simple and quick method for assessing the relative importance GPs assign to factors involved in referral decisions. This tool is helping us to work with primary care colleagues to develop our direct access booking system.


H.L. Spencer, A.J. Lobo. Royal Hallamshire Hospital, Glossop Rd, Sheffield, S10 2JF, UK

Introduction: Gastroscopy is a fundamental investigation for the gastroenterologist. At present, approximately 1% of the UK population undergoes this procedure each year. The BSG and JAG have been involved in drawing up training guidelines to ensure technical and diagnostic competence. Historically however, the ethos has been “see one, do one, teach one”. Consistent reporting of endoscopic findings may therefore be lacking.

Methods: An audit was carried out of gastroscopies performed under the open access system or on the joint medical gastroscopy list from January 2000 to September 2001. Using the “Endoscribe” database, demographic details, use of sedation, indication for endoscopy and final endoscopic diagnosis were all retrieved and analysed by individual endoscopist (two GP clinical assistants, five specialist registrars, and one consultant).

Results: 2127 gastroscopies were analysed (female 54%, sedated 27%). The frequency of reporting the common diagnoses was very variable and the differences between endoscopists were highly significant when analysed using χ2 (see table).

Abstract 355

Conclusions: There are major discrepancies in the reporting of endoscopic examinations. We may be failing to accurately and consistently report our findings. This has major implications for the patients' perception of their illness, the GPs on-going management and consequently the nation's health costs.

Much more emphasis must be placed during training and on the correct interpretation of gastroscopies.


N.I. Church1, H.J. Dallal1, J. Masson2, A. Fraser2, N.A.G. Mowat2, D.A. Johnston3, G. Fullarton4, E. Radin5, M. Turner5, R. Prescott6, J. Plevris7 and K.R. Palmer1.1Western General Hospital, Edinburgh, UK;2Aberdeen Royal Infirmary, Aberdeen, UK;3Ninewells Hospital, Dundee, UK;4Gartnavel General Hospital, Glasgow, UK;5Scottish National Blood Transfusion Service, UK;6University of Edinburgh Medical School, Edinburgh, UK;7Royal Infirmary, Edinburgh, UK

Introduction: The heater probe is a widely used thermal modality, which is used in many centres to effect ulcer haemostasis. It has been stated that the large (3.2 mm) probe is more effective than the small (2.4 mm) probe, but this statement is not evidence based.

Method: As part of a large clinical trial, 247 patients with major peptic ulcer bleeding were treated with a heater probe in combination with endoscopic injection. Choice of heater probe was influenced by the availability of a large probe and also by the availability of an endoscope with a 3.7mm working channel. In 216 patients the small probe was used (group A); in 31 patients (group B) the large probe was applied.

Results: The amount of energy used in both groups was similar (120J and 160J respectively). Injection volume was 3.5 ml in both groups. The groups were well matched for age, shock, comorbidity, and endoscopic stigmata.

Failure of permanent haemostasis followed treatment in 17% of group A and 23% of group B patients; urgent surgery was necessary in 11% of group A and 19% of group B. Mortality at 30 days (9% and 7%) was similar in both groups. Adverse events occurred in 5% of group A and 7% of group B patients.

Conclusion: The efficacy and safety of both heater probes, used in association with endoscopic injection is similar.


D. Levine, R. Teague, S.Turner, J. Wong, J. Freeman. Royal Cornwall Hospitals Trust, Torbay Hospital and Bristol Royal Infirmary, UK

Although there has been substantial progress in the ability to teach gastrointestinal endoscopy there are still few valid and reliable tools for assessment of these skills. In part this reflects the lack of agreed standards for comparison. There is now agreement that training should result in colonoscopists being able to reach the caecum in 90% of patients but demonstrating this degree of competence presents problems.

Assessment should, in addition to showing compliance with a standard, provide feedback to stimulate improvement and help to evaluate training programs. Cumulative sum (Cusum) analysis is an established method of quality control in laboratories and has recently proved useful both in anaesthetics and cardiac surgery. Its use in assessment of colonoscopy has been previously described but it has never been widely adopted.

We have re-examined the value of this technique for assessing completeness of colonoscopy and describe a simple practical method for making it available to both trainees and experienced colonoscopists.

The basis of the analysis is the assignment, to pass/fail events, of appropriately different scores that are summated and charted to show when there is sustained achievement of the required standard. In the case of 90% success for complete colonoscopy being required, each success is assigned a score of minus 0.1 and each failure a score of plus 0.9. With sequential attempts the graph of this function will rise for as long as the success rate is below 90%, flatten out when the standard is attained and fall if the standard is consistently exceeded.

We present such a curve for one trainee's first 140 cases. After 20 incomplete examinations intensive structured training was given. This was followed by increasing success until, by about 70 cases, the standard was being achieved consistently.

We have applied the equation to an Excel program into which the colonoscopist enters each examination coded for “pass” or “fail” and is then able to produce a graph after suitable numbers of examinations.

We give each trainee the programme on a floppy disk with which they maintain and present the progress of training and which gives a graphic, motivating record of achievement over time. Programmes for more rigorous standards can be used by more experienced colonoscopists.


K. Burn1, R. Bicker2, G.D. Bell3, S. Dogramadzi4, C. Allen4.1School of Computing, Engineering and Technology, University of Sunderland, UK;2Mechanical, Materials and Manufacturing Engineering, Newcastle University, UK;3Medical Sciences Faculty, Sunderland University, UK;4Department of Electrical and Electronic Engineering, Newcastle University, UK

Introduction: There is increasing interest in the use of both variable stiffness colonoscopes and thinner “floppier” flexible sigmoidoscopes. We have previously argued (Gut 2001;49:154) that some form of simple beam displacement methodology to determine flexural rigidity (EI) has the advantage of being relatively easy, reproducible, and inexpensive to perform. The disadvantage is that in our experience it takes between one and two hours to accurately determine EI every 10 cm along the shaft of a typical colonoscope. The aim of this study was to develop a method of rapidly and accurately measuring EI.

Methods: A unique, computer based system has been developed that enables EI to be continuously measured along the entire shaft of an endoscope in <60 seconds. The user (1) places the shaft between three low-friction rollers/pulleys (2) applies a known load using a standard weight and (3) pulls or pushes the endoscope through the rollers. The system's software records data from optical encoders mounted on two pulleys, one of which measures the position of the load on the shaft relative to the instruments distal end, and the other the load deflection itself.

Results and Discussion: The new system accurately logged EI data in a fraction of the time (<60 seconds) that it had previously taken using our manual system. It recorded data at 1mm (compared to 100mm) intervals and had the additional advantage over the manual system that the EI data could be saved to computer in raw and processed forms. These could then be displayed in either numeric or graphical format using standard Windows applications such as Microsoft Excel. A systematic comparison of the EI of a number of commercially available endoscopes is now in progress.


C. Meaden, D.A.F. Lynch. Blackburn Royal Infirmary, Blackburn, BB2 3LR, UK

Intravenous benzodiazepine/opiate (IV) is normally used for sedation and analgesia in lower GI endoscopy. Disadvantages of this approach include significant cardiorespiratory complications and prolonged recovery time. The aim was to compare N2O/O2 (Gas) with IV in patients undergoing flexible sigmoidoscopy (FS), or colonoscopy (COL).

Consecutive patients (42 FS, 72 COL) were offered sedation with gas or IV. The endoscopist determined the amount of sedation to be given, and recorded the procedure duration and completeness. Patients recorded degree of discomfort and dissatisfaction using a standardised scoring system. An endoscopy nurse recorded recovery time. FS-gas was compared with FS-IV, likewise for COL. Data were analysed on an intention to treat basis. Median values (range), 95% confidence interval for difference of the means (CI) and p values are shown.

One patient refused gas for COL. Three patients undergoing COL required IV as well as gas. Duration, in minutes, of FS (CI 1.2 [-0.5–2.9]) and COL (CI 2.9 [-0.5–6.3]) and completion rate (100% FS both groups, and 92% and 88% COL for gas and IV, respectively) was not significantly different between the groups. Patient discomfort using gas is mild for FS and COL, but not as good as when using IV. For both FS and COL, recovery time is significantly shorter using gas. N2O/O2 is effective sedation and analgesia in lower GI endoscopy in the majority of patients.

Abstract 359


M.E.Vance, S.G. Shah, N. Suzuki, C.J. Thapar, C.B. Williams, B.P. Saunders. Wolfson Unit for Endoscopy, St Mark's Hospital, London, UK

Background: It has already been shown that nurses are effective in providing safe and accurate gastroscopy and flexible sigmoidoscopy. In this unit the nurse endoscopist has performed over 3000 diagnostic and therapeutic flexible sigmoidoscopies and assisted in 50 colonoscopies. A nurse colonoscopy training programme, based on JAG 2001 guidelines, has been developed to assess whether nurse endoscopists can provide a safe and effective colonoscopy service. A review of the first 160 procedures, performed by the nurse endoscopist is presented.

Aims: To review the first 160 colonoscopies performed by the nurse endoscopist.

Methods: A nurse endoscopist training programme was developed. The first 100 cases were performed under the direct supervision of an expert (BPS/SGS). An assessment of nurse endoscopist performance was evaluated after the first 100 procedures (BPS). The following 60 cases were performed without supervision. Routine polypectomy was performed by the nurse endoscopist on polyps <10mm. Details of referrals, examinations, and complications were recorded.

Results: 160 cases were performed (67 male, 93 female) in which the mean age was 56 years (16–94). Indications included pr bleeding (30%), assessment of IBD (15%), altered bowel habit (15%), pain (11%), diarrhoea (9%), anaemia (8%), cancer follow up (5%), polyp follow up (4%), FH cancer (2%) other (1%). The overall caecal intubation rate was 94% (146/160) (assistance given in 8% of cases due to looping/fixed sigmoid). Video documentation of the caecum and procedure extubation were recorded in 100% of cases. Median sedation administered was pethidine 25mg, midazolam 1.25mg, and buscopan 20mg. Overall findings were normal 70 (48%), IBD 25(17%), adenomas 18 (12%), diverticulosis 13 (9%), cancer 9 (6%), metaplastic polyps 7(5%), and others 4 (3%). Polypectomy was performed by the nurse endoscopist in 21 (14%) cases with no complications.

Conclusion: A nurse endoscopist with an experienced background in flexible sigmoidoscopy can, with specialised training, safely progress to perform colonoscopy for diagnostic referrals.


S.G. Shah, S. Moss, H.J. Pearson, E. Kewka, P.K. Jalal, B.P. Saunders. Wolfson Unit, St. Mark's Hospital, Harrow, London, Diana Princess of Wales Hospital, Grimsby, N.E. Lincolnshire, UK

Background: Precise localisation of colorectal lesions preoperatively directs appropriate surgical management and avoids confusion at subsequent surgery. Colonoscopy can be notoriously inaccurate and therefore other methods must be used to localise lesions. Magnetic endoscope imaging (MEI), a real-time, non x ray technique for imaging of the colonoscope, may assist in determining the location of lesions found at colonoscopy.

Methods: A prospective study was performed to determine the accuracy of MEI for anatomical localisation of the colonoscope tip. The MEI system was used to identify one of four predetermined locations within the colon. Once identified, two endoscopic marking clips were attached to the colonic mucosa and 400–500ml of Urografin™ radio-contrast medium injected to produce an air-contrast “enema”. The clips were subsequently localised by plain abdominal x rays, assessed by a single experienced radiologist, blinded to the colonoscopic findings.

Results: Twenty-nine consecutive patients were enrolled. The overall accuracy of MEI when compared to the air-contrast “enema” was 90% (26 of 29 cases). There were three slight errors of localisation (clips localised to proximal rather than mid-descending colon (1), or to either side of the mid-transverse (1) and hepatic flexure (1)) but which were not considered to be of surgical or clinical importance.

Conclusions: MEI is a reliable and accurate method for determining the anatomical position of the endoscope tip during colonoscopy. When commercially available, we believe the use of MEI will avoid the need for unnecessary barium enema for preoperative localisation of lesions, prior to definitive surgery.


J.A.H. Binnie, E.A.B. Cameron, W. Phillips, N. Carroll, S.J. Middleton. Department of Gastrointestinal Motility and Sciences, Addenbrooke's Hospital, Cambridge, UK

Background: Malignant dysphagia due to oesophageal cancer results in much patient morbidity. This is in addition to the late diagnosis of this cancer and the poor overall prognosis. This study is a review of our experience of metal oesophageal stent insertion for oesophageal cancer to identify factors associated with poor outcome.

Patients and Methods: Observational study of the protocol led management of 100 consecutive patients undergoing stent insertion for oesophageal cancer from 1 May 1996 to 31 December 2000.

Results: Patients can expect a mean improvement of 4.2 (SD 2.6) points in their dysphagia score, a median hospital stay of five days (IQR 2–10), and a median life expectancy post stent insertion of 86.5 days (IQR 30–168) There is an increased complication rate in older patients p=0.005). Improvement in dysphagia score appeared to be worse in those who experienced complications (p=0.09). Overall five day mortality in this series was 3% and perforation rate was 1%.

Conclusion: Palliation of malignant oesophageal strictures with expandable metal stents in the majority of cases provides significant improvement in dysphagia with a short hospital stay and is associated with a low incidence of serious complications. Complications are largely confined to elderly patients with co-morbid conditions. Optimisation of patients prior to stent insertion is vital if complications are to be minimised.


A. Fritscher-Ravens, D. Muckerjee, S. Mosse, P. Swain Address required

Background: Plastic stents are commonly used to provide drainage of obstructed biliary and pancreatic ducts for benign and malignant disease. They are simple flexible tubular structures, pre-shaped by heat to fit the obstructed duct, with flaps or pig tail curves to prevent displacement. They are delivered over guide wire, guide catheter, or both. A pushing catheter is used to push the stent into position. The pushing catheter can push but cannot pull the stent backwards towards the endoscope. One difficulty with stent insertion is that if the stent is inadvertently pushed too far through the ampulla it can be difficult or impossible to retrieve it so many endoscopists deploy longer stents than necessary to prevent inadvertent loss through the ampulla. But longer stents block more quickly and can erode into the duodenal wall.

Aim: To design, develop, and test a stent which can be pushed and pulled backwards over a guidewire or catheter.

Methods and results: An S-shaped curve was formed in the tip of a variety of plastic stents and a perfectly fitting s shape was also cut into the tip of the pushing catheter so that the stent and pushing catheter were locked together when placed over a guide wire or catheter but which would separate as soon as the wire or catheter was withdrawn. A variety of stents of 5, 7, 10, and 12 French size were created and various locking shapes were tested including dovetails and interdigitating bars. These stents were formed from nylon, polytetrafluoroethylene, included flaps and were either straight or heat formed into curves. Some locking mechanisms were formed with a metal tips to the stent and pusher, some were made with plastic. A friction fit lock was also tested. This mechanism was also used to deliver experimental anastomotic stents, which required the means to pull back the stent once in place to exert tension against another structure to form an anastomosis. They were first tested using duodenoscopes in models of the bile duct and pancreatic duct. Subsequently they were tested in post mortem dissections with the duodenoscope held in a jig. Finally these stents were delivered to the pancreatic and bile duct in survival studies in pigs. This type of stent could be delivered significantly more quickly and with fewer misplacements than comparable stents without the lock. The locks always released instantly.

Conclusion: A simple releasable lock on the tip of a plastic stent allowed better control of stent delivery. The stent can be pulled back into an optimum position.


C.J.Larkin, A. Workman, R.E. Wright, T.C.K. Tham. The Ulster Hospital, Dundonald, Belfast, N Ireland, UK

ERCPs are employed in the diagnosis and treatment of biliary and pancreatic disorders. Modern fluoroscopy systems usually allow acquisition of images for hard copy using a digital acquisition system (DAS) instead of a screen-film system (SFS). It is thought that the DAS is associated with a lower radiation dose.

The aim of the study was to prospectively compare the radiation doses to patients from digital imaging and the screen-film system.

DAP reading (Gy-cm2) which is a convenient measure of radiation exposure, screening time, and number of films were recorded.

Data was recorded on 33 patients (10 diagnostic) using the digital imaging and 20 (8 diagnostic) using the screen-film system. Average DAP for the DAS was 16.8 Gy-cm2 for diagnostic, and 63 Gy-cm2 for therapeutic ERCP. Average DAP for the SFS was 13.5 Gy-cm2 for diagnostic, and 66.8 Gy-cm2 for therapeutic ERCP. There was no significant difference in DAP between the two systems. Average screening time for the DAS was 2.7 mins for diagnostic and 7.4 mins for therapeutic ERCPs. Average screening time for the SFS was 2.3 mins for diagnostic and 10.5 mins for therapeutic ERCPs. The average number of films taken using the SFS was 2.8 for diagnostic and 3.7 for therapeutic ERCPs.

In conclusion, there was no significant difference between the radiation dose in the two systems of image acquisition. The number of images taken with the DAS was however higher, a possible explanation of which is the ease with which images can be taken since films do not need loaded each time.


R. Chaudhary, D. Parham, S. Catnach, I. Barrison. Hemel Hempstead Hospital, HP2 4AD, UK

Introduction: Open access endoscopy and the two week cancer initiative have increased pressures on endoscopy units. In our unit all requests for GI endoscopy are prioritised by consultants and allocated a routine, soon, or urgent slot. The purpose of this study was two fold, firstly to assess the relationship between the urgency of the endoscopy, as designated by the gastroenterologist, and the waiting time. Secondly, to determine the relationship of the assessment of the urgency of the procedure with the findings.

In a nine month period, 2584 gastroscopies and 798 flexible sigmoidoscopies were analysed using a computerised record system, enabling the urgency of the procedure to be recorded along with a coded diagnosis. Significant pathology was defined as grade 2 or more severe oesophagitis, Barrett's oesophagitis, oesophageal or gastric cancer, or peptic ulcer. For flexible sigmoidoscopy this was defined as neoplastic polyps, colorectal cancer, or inflammatory bowel disease.

Results: For upper GI endoscopy 28% were routine, 51% soon, and 22% urgent, the average waiting times were 154 days, 59 days, and 19 days respectively. The proportions of cases in each category and the waiting times were not significantly different for flexible sigmoidoscopy. Significant upper GI pathology was found in 33% of routine cases (6 cancers), 39% of soon cases (30 cancers), and 57% of urgent cases (35 cancers). Carcinoma of the stomach or oesophagus was found in eight times as many urgent cases as in routine cases (p<0.01). In flexible sigmoidoscopy, significant pathology was found in 24% of routine cases (l cancer), 44% of soon cases (10 cancers), and 54% of urgent cases (8 cancers). Nine times as many carcinomas were found in urgent cases as in routine cases (p< 0.01).

Conclusion: This study confirms that alarm symptoms are more predictive of the finding of serious pathology in patients with lower bowel symptoms and although alarm symptoms often reliably predict serious upper GI disease in over half of patients considered urgent, a significant proportion of patients with upper GI cancer will have no alarm symptoms.


J.K. Loehry, T.R. Smith, S.K. Vyas. Dept of Gastroenterology, Salisbury District Hospital, Salisbury, Wiltshire, UK

Introduction: Gastrointestinal endoscopy units UK wide continue to strive to meet the two week standard for diagnosis of suspected upper GI cancers. In many cases this has put additional strain on already overstretched departments with negligible improvement in outcome and potential counterproductivity.

Methods: We performed a retrospective audit of patients referred for urgent open access gastroscopy over a six month period. Referrals were made on a standard proforma or letter marked “urgent” and mentioning the word “cancer”. These were coordinated via our Cancer Office with fax and online facilities. We collected clinical and demographic data and analysed both the decision to refer to actual referral time (ideally 24 hours) and the receipt of referral to appointment time.

Results: Data were collected on 79 patients (46 female and 33 male), mean age 69 yrs, from April-September 2001. All but three were faxed proforma referrals. One patient DNA'd. 33% had a normal endoscopy. Only three cancers were detected (4%), all within 14 days of referral. These were all unresectable OG junctional adenocarcinomas.

Conclusion: Whilst the majority of patients were gastroscoped within two weeks of referral, a significant number fell short of the standard. This was due to in part to a delay in GP referral time. Despite this, it remains to be seen whether overall achievement of the two week standard translates into reduced mortality from upper GI cancer.

Abstract 366


A. Douglass1, D. Flynn1, P. van Schaik2, A. van Wersch2, P. Cann11Endoscopy Centre, James Cook University Hospital, Middlesbrough, UK;2University of Teesside (Social Sciences), Middlesbrough, UK.

Introduction: A previous retrospective pilot study identified key factors that influenced gastrointestinal (GI) referral decisions in two local practices.

Aim & Objectives: (1): to further assess the weighting of drivers and inhibitors, identified from the pilot, for GPs' referral decisions with patients presenting with common GI complaints; dyspepsia (D), change in bowel habit (CHB), and rectal bleeding (RB). (2): to examine the prioritisation and pathway of referrals after the initial decision to refer.

Methods: A fractional factorial design was used to construct case studies, of changing variables, to examine the impact of clinical factors, test results, and socio-economic status upon referral decisions. 21 GPs were interviewed about the factors involved in their decisions to refer patients with GI symptoms in the variable case studies.

Results: RB cases were more often referred (61%) than otherwise identical CBH (44%) and D (41%) cases. Males were more often referred than female patients for each main symptom complex. Increased duration of symptoms resulted in increased numbers of referrals for RB and CBH cases, but not for D. Older patients were more likely to be referred for each symptom complex; however, the impact was maximal for patients with a short history of lower GI symptoms. Abnormal blood tests prompted referral, but normal blood tests discouraged referrals for each symptom complex. Socioeconomic status had no impact on referral. Most of the “two week rule” referrals were for RB (27%); followed by CBH (15%) and D (10%). Almost equal numbers of RB and CBH referrals were assigned as urgent; the majority of D referrals were classified as routine. CBH and RB cases were evenly distributed between named consultant and open access. D cases were sent mainly to an open access route (83%)

Conclusions: Key drivers and inhibitors for GP referral have been examined in detail and will inform our development of a direct access booking system. The impact of gender was unexpected. GPs are not taking full advantage of our fast track open access endoscopy services.


A. Douglass, P.A. Cann. Endoscopy Centre, James Cook University Hospital, Middlesbrough, UK

Introduction: A pooled Medical/Surgical approach to GI services would enable referrals to be distributed according to Consultant specialist interest, allow an even spread of workload, avoid problems arising from leave, give direction to clinic, or procedure as first contact and facilitate efficient prioritisation. Do we know how much we all get of what?

Aims & Objectives: To examine the volume, distribution, and sub group clusters of GI referrals from GPs to a busy GI unit. To explore options for redistribution/prioritisation of these referrals.

Methods: The unit serves a 300 000 population and provides open access (OA) gastroscopy and colonoscopy services. It employs three GI physicians, three colorectal surgeons, and one upper GI surgeon. All GI (Medical, Surgical, and Open Access) referrals from GPs to our unit for a two month period were reviewed and analysed. General Surgical (non GI) referrals excluded. Main classes: dysphagia, upper GI (i.e. dyspepsia, GORD, or upper abdominal pain), and lower GI (i.e. isolated rectal bleeding (RB), change in bowel habit +/- rectal bleeding (Ch.BH +/- RB).

Results: Expressed as total number for main classes per sub-service and (% over 45 years of age). 1303 referrals were received in the two months. further examples, there were 12 patients with unexplained iron deficiency anaemia without any GI symptoms.111 referrals were labelled urgent and 26 as “Two week rule” referrals.

Discussion: This and the more detailed analysis of our total GI referral load from primary care, is allowing us to engineer a restructuring of the number of “slots” we must make available on the “pooled effort direct booking system” for each clinical sub-group category. It also informs our designation of “appropriateness” of these slots for relevant clinician, nature of first contact, and speed of first contact.

Abstract 368


R. Badreldin, P. Barrett, L. Hodgson, W.R. Ellis, A.F. Macklon, J.Y. Yiannakou. University Hospital of North Durham, North Road, Durham, DH1 5TW, UK

Background: The Olympus GIF Q240Z endoscope has a ×100 magnification allowing a dissecting microscope view of the mucosa. If it can be shown that the endoscope can accurately identify villous atrophy, there would be a number of potential benefits, including: (i) screening and selection of routine endoscopy patients for intestinal biopsy, and (ii) use in the assessment of patients with known Coeliac disease (CD) who have started a gluten free diet.

Methods: An endoscopic scoring system for villous appearance (Z score) was devised after studying patients with known CD, and comparing endoscopic photographs with histological appearances. Z1 = normal villi; Z2 = stunted villi; Z 3 = “ridges and pits”; Z4 = flat. Following this, 17 consecutive patients with known CD were studied over an eight month period. Z score was determined by the endoscopist. The pathology analysed by one individual (PB), blind of the endoscopic findings. Cases were classified as: Group A = minimal villous changes but expanded lamina propria (Marsh 1 and 2). Group B = mild to moderate villous atrophy (Marsh 3a). Group C = marked villous atrophy (Marsh 3b/c).

Results: There were six patients in group A, three characterised endoscopically as Z1, and three as Z2 (mean Z score 1.5). Four cases in group B, three were Z2, and one Z3 (mean 2.3). Seven cases were in group C, two were Z3 and five Z4 (mean 3.7). The sensitivity of “Zoom” endoscopy in identifying villous atrophy is 100%.

Conclusions: Although this is a small preliminary study, results suggest that the Zoom endoscope may well be effective at identifying villous atrophy and assessing its severity. We found that patients responded positively to seeing visual evidence of villous recovery.


K. Matthewson, S. White, M. Atkins, S. Clark, A. Torrance, P. Bottrill, R. Farrell. Royal Victoria Infirmary, Newcastle upon Tyne NE1 4LP, UK

PEJ insertion was attempted in 17 patients (10M, 7F, median 74 years) with an indication for enteral feeding. Four were unable to tolerate standard percutaneous endoscopic gastrostomy (PEG) feeding due to intractable vomiting or feed regurgitation and 13 were not suitable for standard PEG insertion. Of these, three had abnormal gastric anatomy preventing PEG insertion, four had previous gastric surgery, three had diabetic gastropathy, one had oesophago-gastric cancer, one had intractable vomiting with naso-gastric feeding, and one had meconium ileus equivalent due to cystic fibrosis requiring N-acetyl cysteine perfusion.

Patients were sedated with midazolam (mean 4.9mg) with the addition of pethidine (mean 50mg ) in six patients and hyoscine (mean 28mg) in five patients. Prophylactic co-amoxiclav was used in eight patients and other combinations of antibiotics in eight. A videoenteroscope was used in 11 and videogastroscope in six patients. The scope was advanced and withdrawn until good transillumination with single finger indentation of the jejunum was achieved, following which local anaesthetic was administered into the abdominal wall and the jejunum punctured with the trochar from a standard 16 French Gauge Merck Corflo PEG kit. The guide wire passed through the trochar was retrieved with biopsy forceps and the endoscope withdrawn. The feeding tube was then connected and pulled through the abdominal wall into position. If transillumination was not achieved within 30 minutes the procedure was abandoned.

The procedure was successful in 13 of the 17 patients (76%). The four failures were due to inability to transilluminate. Two of these had standard PEG insertion, one with a jejunal extension, and one had a surgical jejunostomy. One patient with advanced oesophagogastric cancer died three days post procedure from aspiration pneumonia. One patients developed refeeding syndrome from which they recovered. Seven patients died from their underlying disease a median of two months post procedure and five are alive a median of one year post procedure. One patient with successful PEJ insertion tolerated bolus feeding whilst the remainder had continuous infusions using a feeding pump. When conventional PEG feeding is not possible or poorly tolerated, PEJ insertion is a feasible option.


S. Thomas-Gibson, C.J. Thapar, S.G. Shah, J.C. Brooker, C.B. Williams, B.P. Saunders. St. Mark's Hospital and Northwick Park Hospital, Middlesex, UK

Background: Colonoscopy is the gold standard procedure for examining the colon. The procedure has been under great scrutiny in recent months.

Aims: To prospectively audit all aspects of colonoscopy performed at this unit including long term follow up.

Methods: Details of referral, examination, endoscopist, complications, and follow up were recorded prospectively. Patients completed 100 point visual analogue scales for pain and satisfaction following their procedure.

Results: 505 patients (246 male) underwent colonoscopy (by 27 different endoscopists), median age 57 years (13–92). 468 (93%) were outpatients. 64% patients were symptomatic and 36% patients were having surveillance or follow up colonoscopy. 19% of cases were stated as being “very urgent” or “urgent”. Median waiting times were four weeks (0–13) in urgent cases and 10 weeks (0–66) for routine cases. The overall caecal intubation rate was 93% (72–100%), with surgeons, physicians, and experienced trainees performing similarly well (89%, 92%, 94%) and experts 98%.

Photodocumentation of the caecum was performed in 85% of cases and verified by two experienced endoscopists in 88%. In only one case was an inexperienced trainee (<100 procedures) left unsupervised. Pain scores estimated by the endoscopist were well matched with those given by the patient 29 (0–100) and 26 (0–100). Median satisfaction score was 96 (0–100). Pain perception was less if CO2 insufflation was used median score 24.5 compared to 27.5 for air. Immediate complications were recorded in 16 (0.03%), five over-sedation, six vasovagals, three polypectomy haemorrhages, two mucosal injuries (neither requiring treatment). 69% patients returned their six month follow up questionnaire. They reported a median of one day off work or of being unwell at home. 32 patients (0.01%) reported minor bleeding, none requiring treatment. Three patients died within six months of follow up but none were related to their colonoscopy.

Conclusions: Completion rates were adequate for all endoscopists studied and a good level of supervision is available for trainees. There was a high level of satisfaction with the procedure and very few immediate or long term complications.


G.D. Bell1, J. Butler2, S. Dogramadzi3, J. Hancock4, D. Nylander4, J. Painter4, K. Burn5, R. Bicker6.1Medical Sciences Faculty, Sunderland University, UK;2Advanced Surgical Concepts, Bray, Eire;3Department of Electrical and Electronic Engineering, Newcastle University, Newcastle, UK;4Department of Medicine, Sunderland Royal Hospital, UK;5School of Computing, Engineering and Technology, University of Sunderland, UK;6Department of Mechanical, Materials and Manufacturing Engineering, Newcastle University, UK

Introduction: Our group have extensive experience in the use of different stiffening sigmoid overtubes in combination with both adult and paediatric colonoscopes (Gut 1998;42:(suppl 1):A18 and Gut 2000;46(suppl II):A33). All the overtubes were relatively expensive and had the additional disadvantage of not having an air-tight seal to prevent leakage of inflated air and fluid faecal material.

Aims and Methods: To assess in vitro a recently patented inexpensive disposable corrugated overtube developed by Advanced Surgical Concepts Ltd, Bray, Eire (ASCOT). We used the combination of a colonoscopy simulator and an Olympus variable flexion colonoscope (CF-240AL) set on its floppiest mode. Magnetic endoscope imaging was used to assess whether the ASCOT adequately splinted the left side of the colon

Results and Discussion: The ASCOT satisfactorily splinted the left side of the colon when repeated simulated colonoscopies were performed by three experienced endoscopists. The valve at its tip remained airtight despite numerous passages of the endoscope along the colon simulator. The ASCOT was less stiff ( p<0.001) than the Williams split overtube and its corrugated surface reduced friction between the colonoscope and its inner surface. The results are sufficiently encouraging to consider formal clinical assessment in lightly sedated patients undergoing colonoscopy with either adult or paediatric colonoscopes.


P.G.Thatcher, R.P.H. Thompson. St Thomas's Hospital, Lambeth Road, London, UK

Introduction: Barrett's oesophagus was first described in 1950 and is said to have an increased cancer risk of between 20 and 125 fold. This has led to endoscopic surveillance programmes.

Aims: To assess the number of patients with Barrett's oesophagus found on endoscopy between November 1999 and December 2000, histological characteristics and whether follow up had found any carcinomas during that period.

Methodology: Data were collected from the endoscribe reporting system using the terms “Barrett's” and “oesophagus”. Histology was found via the RRS computer reporting system and analysed in relation to the OGD reports.

Results: 280 patients were found (181 female, 99 male) to have an endoscopic diagnosis of Barrett's out of a total of approximately 7000 procedures. Histological confirmation was made in 75% of cases. 172 had no dysplasia, 29 mild dysplasia, one moderate, and two severe dysplasia. Eight had carcinomas, none were found as part of any surveillance programme. The three moderate/severe dysplasias had endoscopies every three months, but none progressed to carcinoma in the one year period.

Discussion: The discrepancy in endoscopic and histological diagnosis was probably due to the number of biopsies taken which varied between two and eight (not in keeping with the world congress of Gastroenterology recommendations). However, this audit did highlight that no carcinomas were found as part of any surveillance program and that the dysplastic patients didn't progress to carcinoma in this period. This further supports recent papers, which have challenged the effectiveness of endoscopic surveillance.


D.G. Power, A.N. Keeling, M.F. Byrne, E.W. Kay, F.E. Murray. Departments of Gastroenterology, Pathology and Clinical Pharmacology, Beaumont Hospital and RCSI, Dublin 9, Ireland

Aim: The aim of this study is to compare TI biopsies (BXs) taken by endoscopists with the histology of these BXs. Thus we can see if TI views at colonoscopy and subsequent BXs taken, actually are proven TI.

Methods: We analysed all the colonoscopy reports in one calendar year which reported taking TI BXs. Only colonoscopies performed by Consultant Gastroenterologists and Specialist Gastroenterology Registrars were used. The Endoscribe programme, set up in the eEndoscopy suite of our 630 bed Dublin teaching Hospital, was the means by which we accessed the reports. We then cross referenced each individual TI BX taken at colonoscopy with its histopathology report.

Results: Our study period was 2 Oct 2000—2 Oct 2001. A total of 111 TI BXs were sent for histology, with 106 being actual TI histologically and thus five proved to be colonic mucosa. These results showed a mismatch rate of 4.5% between endoscopic TI and histological TI. We also examined the indications for procedure, these included: altered bowel habit, unexplained abdominal pain, weight loss, blood per rectum, anaemia (iron and B12), polyp/carcinoma surveillance, radiological abnormalities, inflammatory bowel disease, previous ileal TB, pseudomembranous colitis, suspicion of Maltoma. We concluded that the most frequent indication for colonoscopy was investigation of altered bowel habit: a total of 49% of all procedures performed. The actual TI histological findings were varied and included: normal villous architecture (88%), apthous ulceration (6%), and others (6%). When TI BXs did not correlate with appropriate TI histology, we found normal colonic mucosa +/- mild architectural atypia.

Conclusion: Our gastroenterology service performed a total of 111 TI BXs at colonoscopy over the study period. 106 of these were actually TI. The remaining five were colonic mucosa, giving a 4.5% mismatch. This small retrospective study was interesting in that it highlighted the fact that an endoscopists view of TI does not always correlate with actual TI histologically. No reason was reported for the five erroneous BXs. It appears that experienced endoscopists have a lower endoscopic to histological mismatch.


S. Pathmakanthan, K. Smith, G. Thompson, H.G. Amer, M.T. Donnelly, S.A. Riley. Department of Gastroenterology, Northern General Hospital, Sheffield, UK

Background: Colonoscopy is the investigation of choice for examining the colon but as the demand for colonoscopy increases and JAG begin to specify quality targets, there have been shortfalls in practise confirmed by a recent national audit. We have developed the role of nurse colonoscopist and now present a prospective comparison of doctor and nurse performed colonoscopy.

Methods: Patients attending for routine diagnostic colonoscopy were invited to participate. Examinations were performed using standard videoendoscopies and were video recorded for independent review. Endoscopists were instructed to use midazolam, pethidine, and buscopan according to their usual practice. Endoscopists and nurse assistants graded the patients' pain and tolerance on visual analogue scales (VAS) and a validated questionnaire and VAS were given to the patient following the procedure.

Results: 84 patients (38 female, aged 29 to 83 years ) were examined by the doctor and 83 (48 female, aged 23 to 87years) by the nurse. Intention to treat caecal/ileal intubation rates were 98/83 % and 94/72 % respectively. Failure to reach the caecum was due to obstructing cancer (1) and diverticular disease (1) in the doctor group, and patient discomfort (3), poor preparation (1), and obstructing Crohn's stricture in the nurse group. Pain and tolerance scores were similar but the nurse used more midazolam and pethidine. The pathological spectrum was also similar but with more diverticular disease in the doctor examined group. No significant complications occurred.

Conclusions: Appropriately trained nurses perform safe and effective colonoscopy examinations with caecal and ileal intubation rates exceeding JAG recommendations. Expansion of the nurse colonoscopist role may help to meet the increasing demand for diagnostic colonoscopy.


A.A. Saeed1, M.T. Hendrickse2, P. Isaacs2, C.J. Shorrock2.1Queen Elizabeth Hospital, Gateshead, UK; 2Blackpool Victoria Hospital, Blackpool, UK

Background: Self expanding metal stents offer an attractive treatment modality for palliation in patients with large bowel obstruction due to cancer. We present our experience with endoscopically inserted stents, with and without the use of fluoroscopy.

Methods: 21 metal stents were placed in 18 patients, to relieve symptoms of obstruction due to malignancy. All patients had clinical or x ray evidence of obstruction or obstructive symptoms, such as distension and pain. None of the patients had complete obstruction. In 10 procedures, fluoroscopic imaging was not used and the stents were deployed under endoscopic control.

Results: Metal stents were successfully placed in 21 out of 22 procedures. Relief of symptoms was achieved in 14 out of 18 patients (78%). Of these three patients were alive after a median follow up of 4.3 months. 11 of them died after a median FU of 2.1 months. One patient died due to perforation. (Procedure related mortality 6%). The only other complications were incontinence and stent migration (33%). There was no difference in the success rate of the procedures done with or without fluoroscopic guidance.

Conclusions: When it is possible to traverse the malignant stricture endoscopically, palliative colorectal stenting can be safely performed without fluoroscopic imaging.


S.G. Nugent, A.F. Stone, M.J. Benson. Department of Gastroenterology, St. Helier Hospital, Carshalton, Surrey, UK

Background: Gastric outflow obstruction due to locally advanced malignant disease causes distressing symptoms. Historically, patients have been treated with surgical bypass, with a substantial morbidity. Self expanding metal enteral stents are now available enabling endoscopic palliation. We describe our experience with this technique over a 23 month period from December 1999 to October 2001 in a district general hospital setting.

Methods: All patients presenting with symptoms of mechanical gastric outflow obstruction due to inoperable malignant disease, were included. At endoscopy, a guide wire was manipulated across the stricture, under flouroscopic control. Placement of Wallstent enteral stent (Boston Scientific) was achieved using wire guided, “through the scope” technique and a large channel (4.2mm) operating gastroscope (Olympus GIF 2T240). Pre-dilatation of the stricture was not performed routinely.

Results: 15 patients, median age 78 years, (range 60–92 years) had total of 15 enteral stents. Outflow obstruction was due to carcinoma of antrum (n=3), pyloric canal (n=3), duodenum (n=1), bile duct (n=3), and pancreas (n=5). One patient had three stents placed in series to “bridge”a long stricture involving the gastric antrum and duodenal cap. One patient required a second stent four days later, due to proximal stent migration. One patient with a pyloric canal tumour was found to have a second distal duodenal stricture after placement of an enteral stent across the pylorus and was treated by surgical bypass. In three patients, although a wire was placed successfully across the stricture, it was not possible to position a stent. In each case, “hyperacute angulation” within stricture, prevented manipulation of the stent/delivery system assembly across the stricture. There were no procedure related complications.

Conclusion: Endoscopic placement of SEM enteral stents provide good palliation for malignant gastric outflow obstruction and offer a cost effective alternative to surgical bypass. We propose that simple modifications to the design of the stent/delivery system assembly, for example decreasing longitudinal rigidity and the addition of a flexible “leading edge” to delivery system, may improve success rates in difficult strictures.


K. Ragunath1, M. Waters2, C.J. Wright3.1Dept. of Gastroenterology, University Hospital of Wales, Cardiff, UK; 2Matrix Biology and Tissue Repair Research Unit, Cardiff Dental School, University of Wales College of Medicine, Cardiff, UK; 3Centre for Complex Fluids Processing, Department of Chemical and Biological Process Engineering, University of Wales, Swansea, UK

Background: The adhesion of protein and bacteria leads to the formation of biofilm in biliary stents, thus resulting in stent blockage. Stent internal surface roughness plays an important role in biofilm formation. Atomic force microscopy (AFM) is a novel technology that is uniquely placed to study the stent surface properties and any proposed modifications prior to clinical studies.

Aim: To use AFM technology to image a range of commonly used plastic stents and to quantify the surface roughness at a scale relevant to the adhesion of biological materials.

Methods: AFM (Dimension 3100, Nanoscope3) was used to image four different commercially available unused 10 Fr plastic stents at the 50μm2, 10μm2, and 1μm2 scale. Images of the internal stent surface were taken at the centre of the stent. Surface roughness (Ra) was then measured in order to compare stent surfaces quantitatively.

Results: The table lists the surface roughness of four different stents, which shows a wide variation in internal surface roughness. Roughness is an important indicator as to how surfaces will foul with biological materials. This data is now available to compare with clinical studies of the stents.

Abstract 378 The surface roughness of 4 biliary stents. Ra value is the average of 3 measurements from 10μm2 AFM images; standard deviation of the data is shown in the brackets

Conclusion: AFM images the native stent surface within air or liquid with a higher resolution and allows quantification of surface roughness. AFM is a useful tool for the study of stent materials and proposed surface chemistry modifications prior to clinical trials.


A. Bassi, R. Sturgess, K. Bodger. University Hospital Aintree, Liverpool, UK

Background: To improve access to investigation for patients with suspected cancer, the NHS Executive has introduced a two week out patient waiting time standard. Despite provision of an OAE service at our hospital (>2000 procedures/year; target population 330,000; waiting time <6 weeks), achieving this benchmark for patients with suspected UGI cancer (UGIC) demanded a new initiative. The RAUGICS was set up in parallel to OAE to allow GPs to request “fast track” direct access endoscopy (and subsequent clinic review) for “high risk” subjects.

Aims: (a) to evaluate the impact of introducing a RAUGICS on the total number of direct referrals (DRs) for endoscopy; (b) to compare the profile of endoscopic diagnoses for RAUGICS with that of OAE before (OAEB) and after (OAEA) introducing the new service; (c) to assess the resource implications of RAUGICS in terms of direct costs (endoscopy plus clinic) per cancer diagnosed.

Methods: Information was obtained from “Endoscribe” for OAE and prospectively for RAUGICS. Two six month periods were compared (OAEB: 01.01.00 to 30.06.00; OAEA & RAUGICS: 01.01.01 to 30.06.01). Major diagnoses and disease stage of cancers were verified by case record review.

Results: After launch of RAUGICS: (a) total DRs increased by 33% (953 to 1264) with 51% of DRs (645/1264) designated as requiring the “fast track” service (rapid endoscopy, then clinic review); (b) the prevalence of UGI cancer amongst DRs overall was unchanged (1.57% v 1.6%), but most cancers (20/21) were diagnosed in the RAUGICS group (3.1% yield); (c) Cost-per-cancer diagnosed within the new service was ∼ £7740; (d) Stage of cancers diagnosed by direct access endoscopy was unchanged (OAEB versus RAUGICS).

Conclusions: Referral criteria were effective in channelling patients with UGI cancer into the RAUGICS arm of the direct access service, but demand for “rapid access” was high (half of all DRs). The low cancer rate (3.1%) suggests either poor specificity of referral criteria or a high level of “inappropriate” referrals.

Abstract 379


E. Quinton, W.P. Goddard, R.G. Long, K. Teahon. Department of Gastroenterology, Nottingham City Hospital, Hucknall Road, Nottingham NG5 1PB, UK

Introduction: There is a significant body of literature assessing open access outpatient gastroscopy (OGD). Little is written about the use of organisation of the service OGD for inpatients except in the situation of gastrointestinal bleeding. In late 1998 we switched our inpatient gastroscopy service from a “Consultation first” service to an “Electronic Open Access Service”. This study evaluates the impact of thee change in service organisation.

Methods: (1) The number of and reason for requests for OGD since 1998 was retrieved from the electronic database. (2) “Reason for request” categories where >100 requests had been made in the last year were identified and the hospital records of the patients in these categories were selected randomly for review to allow review of 25% of the records. (3) On review of the records the demographic details, reason for admission, reason for referral for and outcome of OGD, eventual diagnosis, and routine laboratory parameters were recorded. (4) The pre OGD admission record of each patient was summarised and reviewed anonymously and independently by three consultant gastroenterologists to determine appropriateness of the request for OGD.

Results: During the period from 1998 there was a 19% increase in emergency medical and surgical admissions and a 27% increase in OGD requests but this varied from a 13% increase in referrals for melaena to a 56% and 80% increase in referrals for nausea/vomiting and abdominal pain respectively. When the appropriateness of referrals was assessed 67% of all requests were considered appropriate but the value varied (abdominal pain 41%, haematemesis 74%, melaena 83%, anaemia 84%, and nausea/vomiting 57%). The eventual diagnosis was achieved at OGD in all patients in 52% (abdominal pain 29%, haematemesis 79%, melaena 65%, anaemia 38%, nausea/vomiting 38%).

Conclusion: An electronic open access service is appropriate for inpatients with GI bleeding and anaemia although more selective referral questions may improve the appropriateness of referrals. Pre-procedure consultation is more appropriate for most other categories of inpatient referrals.


D. Flynn1, P. van Schaik2, A. van Wersch2, P. Cann1, A. Douglass1.1 Endoscopy Centre, James Cook University Hospital, Middlesbrough, UK; 2University of Teesside (Social Sciences), Middlesbrough, UK.

Introduction: A web based, direct booking system for GP referrals to a GI clinic or procedure is under development for our unit. Guidelines for referral of clinical sub groups will be incorporated as algorithms within the software of the computerised decision-support booking system. This will enable the most appropriate slot for the patient to be identified and booked directly in real time. “Appropriateness” will relate to relevant clinician, nature (clinic or procedure,) and speed of first contact and provide patients with some choice of appointments.

Aims & Objectives: To assess GPs' attitudes to the principles, design, and their eventual usage of such a system.

Methods: 21 GPs were given a description of the proposed system and asked to rate each of eight potential benefits on a 6 point scale. GPs were also given the opportunity to discuss the potential impact of the system on their clinical practice and patients.

Results: The five most highly scored potential benefits were (i) faster and more reliable communication between primary and secondary care (mean 5.22), (ii) appropriateness of the first assessment received by patients (mean 4.53), (iii) facilitating adherence to local clinical guidelines (mean 4.53), (iv) facilitating the explanation of referral decisions, tests and procedures to patients (mean 4.42), and (v) provision of information on evidence-based reports for GI symptoms (mean 4.37). 52% of GPs indicated that they would use the system in real time only if it generated no more than two additional minutes on consultation times. 21% of GPs were concerned that the system could increase administrative workload, and create unrealistic expectations of open access services for patients.

Conclusions: The majority of GPs recognised potential benefits of the proposed system, but the impact on consultation times is a core design issue that will determine acceptance and usage of such a system. Other members of the primary care team may help to minimise the need for the GP to spend “real time” online.


M.M. Walker, R.P. Negus, A. Rice, V. Loh, J. Teare, M.R. Thursz. Faculty of Medicine, Imperial College of Science, Technology and Medicine, St Mary's Campus, London W2 1PG, UK

The topography of gastritis predicts a likely outcome in Helicobacter pylori infection. Histological markers of cancer risk include severe atrophy and intestinal metaplasia. The updated Sydney system for reporting gastritis recommends five biopsy sites, which is time consuming. This study aims to establish if all five gastric sites are necessary for an accurate appraisal of gastric pathology in routine practice.

Patients and Methods: 100 patients, 45 men, median age 57, range 21–88 years, attending for upper gastrointestinal endoscopy had five gastric sites routinely biopsied according to the updated Sydney protocol. Biopsies were placed in order on a cellulose acetate strip, processed, and reported by two pathologists according to the Sydney System. Results were analysed to assess concordance of diagnosis by site and to determine which of the five sites were optimal to assess gastric pathology.

Results: The majority of presenting symptoms for endoscopy were dyspepsia or abdominal pain (47 patients). At biopsy for gastritis, including grade of gastritis, there was excellent concordance between body sites (100%), and good concordance between antral sites (94%). No further information was gained from separate analysis of the biopsy of the incisura. Intestinal metaplasia (IM) was found with greatest prevalence in A2 (inferior antrum) -17% all biopsies and B2 (greater curve, body) - 5%. In contrast, atrophy was found with greatest prevalence (21%) at A1 (superior antrum) but also as in IM, 10% of B2 biopsies showed atrophy. The overall prevalence of IM in this study was 19% and atrophy, 21% from analysis of biopsies of all sites.

Discussion: This study shows that when histological assessment of gastritis is required, in the absence of a visible lesion, two biopsies, one from the antrum and one from the body are adequate for a clinical diagnosis of grade and type of gastritis. However if only two sites are biopsied, IM (a histological marker of cancer risk) will be missed in 7% of cases relative to the full Sydney protocol.


M.C. Gallagher, F.R. Vicary, V.S. Wong. Department of Gastroenterology the Whittington Hospital London N1, UK

Background: Acute upper gastrointestinal haemorrhage (UGIH) constitutes a medical emergency with an incidence of 100 per 100 000 and mortality rate of 14%1. The Whittington Hospital departmental guidelines state that patients should undergo endoscopy the day after admission unless there are criteria for “emergency endoscopy” (out of the hours 9 am–5 pm): suspected varices, large bleed/continued bleeding, or history of aortic graft. Two consultant physicians, (and their SPRs) and two consultant surgeons provide 24 hour endoscopy cover.

Aims: To perform a six month prospective audit of the management of suspected cases of acute UGIH at the Whittington Hospital.

Methods: All suspected cases of acute UGIH were included. Data were collected on a standard proforma and case notes reviewed after discharge. Outcome measure included the timing of endoscopy, use of “out of hours” endoscopy services, endoscopic diagnosis, and mortality rates. Rockall scores were calculated for all patients who required endoscopy.

Results: 95 cases of suspected UGIH were identified. 76 patients (80%) presented via A&E and 19 (20%) were existing inpatients. 90 patients underwent endoscopy. In 92% of cases, endoscopy was performed within 24 hours of admission/inpatient bleeding episode. Of these, 20 patients achieved criteria for emergency endoscopy. 17 were endoscoped out of hours and three patients presented within hours. The average Rockall score for patients endoscoped within hours was 3.5 (range 1–10) and out of hours, 6.5 (range 1–10). Endoscopic diagnoses were as follows: peptic ulcer 37%, varices 15%, normal 12%, erosions 9%, oesophagitis 9%, Mallory-Weis tear 2%, other 16%. Endoscopic therapy was required in 29 cases (30%). The overall Mortality Rate was 10.5%.

Conclusion: The proportion of patients in this series endoscoped within 24 hours (92%) compares favourably with published data (50 and 56% in the BSG National Audits) 1, 2. 17/95 patients underwent out of hours endoscopy including seven with variceal bleeding. The appropriate provision of a 24-hour bleeding service is possible in a small unit with the collaboration of medical and surgical Endoscopists.

1. Rockall et al: BMJ 1995;311:222–6 2. Rockall et al: Gut 1997;41:606–11


T.A. Raven, J. Whelan, I. Fretwell, M. Aston, P. Whitney. Endoscopy Department, Chesterfield Royal Hospital, Derbyshire, UK

Background: Patients attending endoscopy for a flexible sigmiodoscopy (FS) examination were being prepared using picolax except one doctor who used phosphate enemas on arrival. Recent research by W. Atkin et al (2000)1 has shown that an phosphate enema administered on the day of the examination is as effective as picolax. To see if this could be successful in our endoscopy unit an audit was conducted.

Aims: To improve the way in which patients are prepared for (FS) without compromising the quality of bowel preparation.

Methods: 53 Patients were selected using two (FS) lists and had the option of taking the enema at home instead of picolax for bowel preparation. A questionnaire was prepared for patients to complete while they were waiting for the (FS). A note was made of patients unable/unwilling to administer the enema. Nursing and medical information was collected to assess the acceptability and quality of bowel preparation.

Results: 98% of patients administered the enema successfully at home. 74% of the (FS) tests were completed to 60cm. 100% of patients using the enema would use it again, 12% telephoned the department for advice, 38% experienced mild side effects, and 27% needed help from someone at home. The bowel preparation was excellent to good in 75% of cases.

Conclusion: Patients found the enema easy to use and acceptable as a method of bowel preparation. Although only a small number of cases were audited the results are very similar to those reported in the BMJ.

1. Atkin et al: BMJ 2000;320:1504–9


J. Hancock1, S. Dogramadzi2, G.D. Bell1, C. Allen2, K. Burn3, I. Fletcher3, R. Bicker4.1Medical Sciences Faculty, Sunderland University, UK; 2Department of Electrical and Electronic Engineering, Newcastle University, UK; 3School of Computing, Engineering and Technology, University of Sunderland, UK; 4Department of Mechanical, Materials and Manufacturing Engineering, Newcastle University, UK

Introduction and Aims: Torque steering is now the preferred method of teaching intubation of the tortuous sigmoid colon. One of the criticisms of many teaching simulators is that the “feedback” forces are not realistic. One of us (CA) had had extensive previous experience with the use of pressure sensor gloves to estimate the mechanical forces/ mechanical work done by female assembly plant workers1. We decided to see if (a) we could satisfactorily modify the equipment to measure similar forces being applied by an endoscopist's right and left hand during endoscopic procedures (real or simulated) and (b) if so whether these differed depending on the diameter and stiffness of the endoscope employed.

Aims and Methods: A single experienced endoscopist (JH) carried out a total of 10 endoscopic procedures (three OGDs, one flexible sigmoidoscopy, and six colonoscopies) while wearing the sensor gloves. Each glove containing 20 calibrated piezo electric sensors which sent pressure data to a PC fifty times a second. The data were represented both graphically and numerically using specially written software.

Results and Discussion: The equipment accurately and reproducibly measured the, at times, quite considerable forces generated by the right hand during torque steering. Right hand torque forces were predictably much higher during lower GI endoscopy than during OGD. Despite the small numbers of patients studied to date, it is becoming clear that during colonoscopy/flexible sigmoidoscopy greater right hand torque forces are required with thin endoscopes compared with thicker ones. We are now looking at the use of Neural Network/Artificial Intelligence methodology to help analyse the data in greater depth. We are confident we can apply the technique to looking at differences between endoscopists and endoscopes during both real and simulated examinations. The research could have implications in the design of instruments and teaching simulators as well as live “hands on” assessment of trainees during endoscopic procedures.

1. Allan CR et al: Mechatronics 1997;7:141–57


R. Zinicola1, M.D. Rutter1, G. Falasco1, V. Cennamo2, S. Contini2, B.P. Saunders1.

1Wolfson Unit for Endoscopy, St. Mark's Hospital, UK; 2Endoscopia Digestiva, Ospedale Bellaria, Bologna, Italy

Background: Management of haemorrhagic radiation proctitis (HRP) remains controversial. Recently both endoscopically delivered argon plasma coagulation (APC) & local rectal application of 4% formalin (LRAF) have been reported as effective treatments. However the exact role of these therapies is not clearly defined. We evaluated the efficacy of APC, and developed a new endoscopic score to guide therapy.

Methods: 12 patients with significant rectal bleeding due to HRP were retrospectively reviewed. Patients were classified using a new endoscopic grading of HRP from endoscopic videoprints, assessing confluence and distribution of telangiestasias, percentage of surface area involved, and presence of fresh blood. All patients were treated with APC initially.

Results: Utilising the new endoscopic grading, five patients were categorised as grade I (mild) HRP, four patients grade II (moderate) HRP, three patients grade III (severe) HRP. In 10 patients (83.3%), bleeding improved significantly following APC therapy. All patients with grade I & II were treated successfully by APC (median two sessions, range one to four). In two grade III patients APC failed, but subsequent formalin application was successful.

Conclusions: Our endoscopic score may help guide appropriate treatment for HRP. APC appears safe, efficacious, and should usually be considered first line therapy, particularly in grade I & II HRP. However with extensive (grade III) HRP, topical formalin application may be more effective.


M.C. Follows, B.J. Rembacken, D.M. Chalmers, A.T.R. Axon. Centre for Digestive Diseases, the General Infirmary at Leeds, Great George St, Leeds, LS1 3EX, UK

Aims: To correlate the histological and clinical findings in patients with a macroscopically normal, total colonoscopy.

Method: A search was made of all colonoscopy reports performed during 2000. Reports from patients with a macroscopically normal, total colonoscopy in which a set of serial biopsies (caecum, ascending, transverse, descending, and sigmoid colon and rectum) had been taken were identified. The histology result for each patient was found and the indication for the examination noted in each case. Cases were excluded if any endoscopic abnormality was noted or if a full set of biopsies had not been taken.

Results: See table.

Abstract 387

Conclusions: Five cases of IBD (inflammatory bowel diease/microscopic colitis) were identified in 282 patients. In all diarrhoea was a presenting symptom (i.e. 4% of patients with diarrhoea had IBD). If diarrhoea was not an indication no patients were found to have IBD1 on serial biopsies. We recommend serial biopsies should be taken in patients with a normal colonoscopy but only if the predominant symptom is diarrhoea.


S. Cherian, A. Cherukuri, P. Singh. Staffordshire General Hospital, Stafford ST16 3SA, UK

Introduction: Ileoscopy is not routinely attempted because of its perceived technical difficulty. We believe routine ileoscopy is useful in quality assurance and it provides additional diagnostic yield.

Methods: We examined colonoscopy data from September 1995 to October 2001 of a single gastroenterological firm. For documentation of completeness of examination, visualization of ileocaecal valve or ileal intubation were the only criteria used. A registrar endoscopist performed 80 colonoscopies independently. For analysis of intubation rates, these 80 and a further 56 procedures in patients with prior colonic resection were excluded. During the last year of the audit, data were prospectively collected on procedure times (PT).

Results: There were 1602 colonoscopies. The median age was 60 years (range 8–95). The male to female ratio was 4.5:5.5. The diagnostic yield from 73 ileoscopies and 67 sets of biopsies in 66 patients with colonic IBD was: 63 normal; seven Crohn's ileitis; and three backwash ileitis. Of those without colonic IBD, 904 had ileoscopy and 414 had biopsies. Ileal Crohn's was identified in 11, non-specific ileitis in three, discrete ileal ulcers in four (three related to NSAID use), and non-specific inflammation in ileum and colon in three. In addition to the above 21, nine additional positive histologies were as follows: five non-specific ileitis, three villous atrophy (two also had duodenal villous atrophy), and one amyloidosis). The main indications for obtaining histology in the absence of an endoscopically visible lesion were diarrhoea and/or anaemia. Overall ileoscopy rate (IR) was 67.5%. IR rose progressively to plateau for the past three years at approximately 85%. Twenty-two colonoscopies were excluded from analysis of PT in 259 patients (previous colonic resection=13; impassable stricture=7; incomplete colonoscopy=1; abandoned due to respiratory depression=1). The median PT and interquartile ranges in minutes were as follows: anus to caecum=7 (5–10); total time=17 (14–21.5); caecum to ileum=2 (1–4); and time in ileum=1 (0.5–2). The median length of ileum examined was 10 cm (10–15).

Conclusions: Ileoscopy is the gold standard in the documentation of completeness of colonoscopy. In skilled hands, it is easy, adds only three minutes to the procedure time, and contributes significantly to quality assurance and diagnostic yield.


R.P. Arasaradnam, P. Hancock, T. Woodward, G. James. Department of Gastroenterology, Doncaster Royal Infirmary, Doncaster, UK

Background: Laser therapy is useful not only in extensive oesophageal tumour and stent overgrowth but also, for inoperable recto-sigmoid tumours in reducing the risk of obstruction or bleeding.

Aim: To determine retrospectively outcome and complication rate of procedures carried out over a seven year period.

Subjects & Method: A total of 95 patients received laser therapy from 1994 until 2001. 31 of whom had lower, while 64 had upper gastrointestinal laser therapy

Results: Out of the 31 lower gastrointestinal procedures 15 (50%) were for rectal carcinoma, six (19%) recto-sigmoid carcinoma, five (16%) sigmoid carcinoma, three (9%) tubulovillous adenoma, one (3%) sigmoid adenoma and one (3%) colonic angiodysplasia. 19 (61%) were male; mean age of 78 (range 74–97). 13 (42%) had cardiovascular co-morbidities, three (9%) cerebrovascular accidents (CVA), 5 (16%) chronic obstructive airway disease (COAD), and three (9%) had other primary malignancies. Median number of procedures was 4 ± 1.2 over a duration of 4 ± 1.9 months. Nine (29%) remained alive (one colonic angiodyplasia, three tubulovillous adenoma, two rectal carcinoma (later had surgery), one sigmoid adenoma and one recto-sigmoid carcinoma). Complication rate was 0%

For the 64 upper gastrointestinal laser procedures, 39 (61%) had oesophageal carcinoma, 15 (23%) gastric carcinoma, six (8%) stent overgrowth, one (2%) oesophago-gastric carcinoma, one (2%) gastric arteriovenous ectasia (GAVE), one (2%) Hereditary haemorrhagic telangiectasia (HHT), and one (2%) gastric angiodysplasia. 38 (59%) were male; mean age of 75 (range 44–88). 27 (42%) had cardiovascular co-morbidities, eight (13%) CVA, 14 (22%) COAD, and five (8%) had other primary malignancies. Median number of procedures was 1.25 ± 0.9 over a 2.5 ± 0.5 month period. 12 (18%) remained alive (three gastric carcinoma, one gastric angiodysplasia, one HHT, one GAVE: eventually required Argon plasma coagulation and six oesophageal carcinoma: one eventually requiring a stent). Complication rate was 0.05% (two with aspiration pneumonia within two weeks of the procedure and one bled after 12 hours of treatment).

Conclusions: Laser therapy is safe and an effective method of palliation for inoperable malignancy when performed in a District General Hospital setting by an experienced endoscopist.


M. Allison, T. Tang, P. Roberts, R. Dickinson. Department of Gastroenterology, Hinchingbrooke Hospital, Hinchingbrooke Park, Huntingdon, PE29 6NT, UK

Background: The deployment of enteral stents under endoscopic and fluoroscopic guidance has emerged as an effective alternative to palliative surgery for malignant gastric outflow and duodenal obstruction.

Methods: Twenty-one consecutive patients with inoperable malignant (gastric (28%), pancreatic (67%), and metastatic deposit (5%)) upper GI strictures were prospectively studied between April 1999 to October 2001, on an “intention to treat” basis. Twenty-three procedures were performed using the Wallstent enteral prosthesis, implanted per-orally under fluoroscopic guidance. All patients had severe nausea, recurrent vomiting and their obstructions were deemed inoperable. Twelve (57%) were male and nine (43%) were female; mean age 71.6 years. Malignant obstruction occurred at the pylorus (19%), antrum (5%), first and second parts of the duodenum (66%), third part of duodenum (5%), and anastomotic sites (5%).

Results: Stent implantation was technically successful in eighteen (85%) patients. On two occasions a wire could not be passed and in a third, a previously placed plastic biliary stent could not be changed preventing enteral stent deployment. Nausea, vomiting, and dysphagia improved in all cases that had a stent successfully placed. There were no short term complications. Two patients (11%), who were successfully stented, complained of abdominal pain post insertion. Two patients were readmitted with vomiting at six and nine months respectively. The first was due to tumour overgrowth and the second to distortion caused by the stent. Both were successfully treated, by inserting further co-axial stents to bridge the stenoses. Twelve patients (67%) died during the follow up period from causes unrelated to the stent insertion, with a median survival time of three months. Five patients are still alive, the longest being seven months. One patient was lost to follow up.

Conclusion: In our experience, the endoscopically placed enteral stent is a safe and efficacious procedure that provides an effective means of palliation to patients with an inoperable malignant upper GI obstruction.


A. D. Gilliam, N. Davidson, K. Mylankal, D. Ziegler, E.P. Perry, J.D. Harrison. Department of Surgery, Scarborough Hospital, Woodlands Drive, Scarborough, YO12 6QL, UK

Aim: Laparoscopic cholecystectomy is the treatment of choice for symptomatic gallstone disease, although little data exists as to the long term ability of the procedure to alleviate the symptoms of patients.

Method: A prospective study of 212 patients undergoing laparoscopic cholecystectomy was performed. A detailed postal symptom questionnaire was designed and piloted. Patients completed these preoperatively and at three and 12 months postoperatively.

Results: 169 patients were female. The median age of patients was 57 years (range 21–81). Complete data was available on 68% (145 patients). 34 patients (23%) continued to be symptomatic. Only one patient (0.06%) reported their abdominal pain/discomfort had become worse. Of the symptomatic patients with continuing abdominal pain, 15 (44%) claimed their pain was identical to that pre-operatively. 55% (26) of the symptomatic patients consulted their GPs during the one year follow up period.

Conclusion: A significant number of patients are dissatisfied one year after laparoscopic cholecystectomy. All patients should be pre-operatively counselled about the risk of persistence of some pain symptoms after laparoscopic cholecystectomy. There should be a high index of suspicion for upper GI conditions and more patients should undergo diagnostic endoscopy before subjecting them to surgery.

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.