Guidelines on small bowel enteroscopy and capsule endoscopy in adults

Transcription

1 on small bowel enterosopy and apsule endosopy in adults R Sidhu, 1 D S Sanders, 1 A J Morris, 2 M E MAlindon 1 Guidelines 1 Department of Gastroenterology, Royal Hallamshire Hospital, Sheffield, UK; 2 Department of Gastroenterology, Glasgow Royal Infirmary, Glasgow, UK Correspondene to: Dr M E MAlindon, Department of Gastroenterology, P39, P Floor, Royal Hallamshire Hospital, Glossop Road, Sheffield S10 2JF, UK; mark. malindon@sth.nhs.uk Revised 23 July 2007 Aepted 31 July 2007 Contents 1.0 Introdution 2.0 Formulation of guidelines 2.1 Grading of reommendations 2.2 Sheduled review 3.0 Summary of reommendations 4.0 Types of small bowel enterosopy 4.1 Enterosopy using a olonosope 4.2 Sonde enterosopy 4.3 Push enterosopy 4.4 Intraoperative enterosopy 4.5 Double balloon (push and pull) enterosopy 5.0 Capsule endosopy 5.1. Tehnique 5.2 Indiations for apsule endosopy 5.3 Compliations of apsule endosopy 5.4 Pateny apsule 6.0 Servie provision and training Referenes 1.0 INTRODUCTION The small bowel has historially been a diffiult area to examine due to its anatomy, loation and relative tortuosity. Examination beyond the duodenojejunal flexure is of importane in a number of small bowel disorders. Before the advent of enterosopy or apsule endosopy, radiographi studies had been the main investigative modality of the small bowel. Barium follow-through and enterolysis permits indiret examination of the small bowel but has a low diagnosti yield partiularly in the ontext of obsure gastrointestinal bleeding. 1 3 Capsule endosopy and enterosopy are now the preferred methods to examine the small bowel in most situations. These guidelines are intended to provide an evidene based doument desribing endosopi investigation of small bowel disorders. 2.0 FORMULATION OF GUIDELINES These guidelines were ommissioned by the Clinial Servies and Standards Committee of the British Soiety of Gastroenterology (BSG) and have been produed by the small bowel and endosopy setions of the BSG. The guidelines have been produed to onform to the North of England evidene based guidelines development projet. 4 5 They have been drawn up from a Medline, Embase and Ovid literature searh using terms enterosopy, push enterosopy, intraoperative enterosopy, double balloon enterosopy and apsule endosopy. There have been 180 peer review studies, seven review artiles, 58 ase reports and letters, and one set of Amerian guidelines on enterosopy. 6 The literature searh for apsule endosopy inludes 100 peer review studies, 51 review artiles, 74 ase studies and letters, 21 editorials, four pooled analyses and two sets of guidelines: Amerian and European on apsule endosopy Grading of reommendations Grade A requires at least one randomised ontrolled trial as part of a body of literature of overall good quality and onsisteny addressing the speifi reommendation (evidene ategories Ia and Ib). Grade B requires the availability of linial studies without randomisation on the topi of onsideration (evidene ategories IIa, IIb and III). Grade C requires evidene from expert ommittee reports or opinions or linial experiene of respeted authorities, in the absene of diretly appliable linial studies of good quality (evidene ategory IV). 2.2 Sheduled review The ontent and evidene base for these guidelines should be reviewed within 5 years of publiation. We reommend that these guidelines are audited. 3.0 SUMMARY AND RECOMMENDATIONS If there is a high suspiion of bleeding from an upper GI soure, a seond look endosopy should be undertaken prior to CE to ensure no pathology has been missed. (grade B) Patients presenting with obsure gastrointestinal bleeding with a negative gastrosopy and olonosopy should undergo apsule endosopy if no ontraindiations exist. (grade B) All patients undergoing CE for any indiation should be appropriately ounselled on the risks of apsule retention. (grade C) Non-passage of a apsule may our in the presene of a normal radiologial ontrast study. (grade B) Those patients with pathology/bleeding sites identified on CE should subsequently undergo either a PE or DBE (oral/anal route) depending on loation/site of bleeding. (grade B) Push enterosopy should ideally be performed using a dediated push enterosope. (grade B) Endosopi therapy should be attempted to minimise further bleeding episodes. (grade B) In patients with a negative CE and persistent OGB, a seond look apsule endosopy may be onsidered. If this is negative they should be referred for DBE. (grade C) Intraoperative endosopy should be reserved for patients with persistent signifiant GI Gut 2008;57: doi: /gut

2 bleeding in whom the bleeding soure remains undiagnosed. (grade B) CE should be onsidered in patients with a high suspiion of small bowel Crohn s disease based on the linial history and inflammatory markers undeteted by onventional means. Patients with abdominal pain as a signifiant feature should have radiologial imaging to exlude a striture prior to CE. (grade C) CE should be onsidered in patients with refratory oelia disease to look for oelia assoiated ompliations. (grade C) 4.0 TYPES OF SMALL BOWEL ENTEROSCOPY 4.1 Enterosopy using a olonosope The small bowel may be examined using a standard adult or a paediatri olonosope without the purhase of a dediated small bowel endosope. The olonosope is advaned as far as possible with the aid of abdominal pressure and hange of position of the patient. Using this method, up to 60 m of small bowel beyond the ligament of Treitz an be examined. 6 In pratie the stiffness of the adult olonosope makes advanement diffiult and the flexibility of the paediatri olonosope auses frequent looping therefore this tehnique is of limited value. 4.2 Sonde enterosopy The sonde fibreopti enterosope, first desribed by Tada et al in 1977, has a working length of m, and is passed orally or nasally. 10 It is advaned into the duodenum with the aid of another orally passed endosope. 6 It is then propelled through the small bowel by peristalsis. The main disadvantages are the lak of tip defletion, biopsy hannel and length of time (from 4 6 h) taken for this examination whih makes patient tolerane poor. 6 The use of this method of examining the small bowel has largely been superseded by other modalities. 4.3 Push enterosopy Push enterosopy is urrently the most frequently used endosopi method for small bowel examination. Dediated push enterosopes are m in length with biopsy hannels that an aommodate a range of aessories for therapeuti intervention Tehnique The endosope is introdued orally and passed into the duodenum beyond the ampulla of Vater. After traversing the urve of the seond part of the duodenum, the enterosope is straightened to redue any loops formed in the stomah. The enterosope is then pushed to the maximum length of insertion. 15 It is performed as an outpatient proedure, under onsious sedation and takes between 15 and 45 min Use of an overtube Initial studies using an overtube (first desribed in ) showed an inrease in depth of insertion with its use A number of reported ompliations, whih inlude muosal stripping, 20 duodenal perforation, 17 pharyngeal tear, 14 panreatitis and Mallory Weis tear 21 have been reported with the use of the overtube and this may limit its appliation during enterosopy. Later studies with graded stiffness enterosopes have questioned the additional value of the overtube, hene many units no longer use it in routine pratie The depth of insertion during push enterosopy and the length of small bowel examined ( m) is variable. Two methods an be used to measure the maximum length of small bowel examined: metri measurement from pylorus on withdrawal after straightening, or fluorosopy whih helps to asertain absene of a gastri loop Indiations for push enterosopy Push enterosopy is indiated in the following linial situations: (a) Diagnosti Obsure gastrointestinal bleeding Malabsorption and unexplained diarrhoea Exploration of radiographi abnormalities of the proximal small bowel Investigation of small bowel tumours (b) Therapeuti Thermooagulation of bleeding lesions Plaement of jejunostomy tubes () Surveillane Polyposis syndromes (a) Diagnosti Obsure gastrointestinal bleeding In most patients who present with gastrointestinal haemorrhage, prompt investigation by way of linial assessment and endosopy of the upper or lower gastrointestinal trat provides a satisfatory diagnosis. The main indiation for push enterosopy is obsure gastrointestinal bleeding (OGB) when initial gastrosopy and olonosopy have failed to detet the soure of bleeding. This ours in approximately 5% of patients who present with gastrointestinal haemorrhage. The investigation and management of OGB provides a resoure intensive hallenge for liniians through repeated hospital admissions, investigations, transfusions and medial or surgial therapy. 34 OGB is sub-lassified as overt with the presene of melaena or haematohezia, or oult with anaemia and/or positive faeal oult blood testing. 35 The diagnosis is often delayed due to slow or intermittent haemorrhage that is not deteted during endosopy or angiography. In the elderly, multiple potential bleeding sites may be seen without a lear indiation of whih lesion is the soure of haemorrhage. 36 The diagnosti yield of OGB with push enterosopy is between 12 and 80% with the highest yield in patients with overt bleeding. Push enterosopy has proven value in the investigation of patients with suspeted GI haemorrhage when initial onventional endosopy is normal. (reommendation grade B) Twelve to sixty-four per ent of lesions loated with push enterosopy are within the reah of a standard endosope Lesions ommonly missed are Cameron s ulers (linear uleration in large hiatus hernia), varies, pepti uler disease 43 and gastri antral vasular etasia (GAVE) whih an be diagnosed as gastritis by the inexperiened endosopist Repeat gastrosopy is reommended if an upper GI soure is suspeted despite the initial negative gastrosopy (reommendation grade B) Malabsorption and unexplained diarrhoea Duodenal biopsy during upper GI endosopy is the aepted approah to obtain histology in patients with suspeted malabsorption. There is a modest role for push enterosopy in patients with malabsorption when the duodenal biopsies are abnormal but non-diagnosti or if these individuals are endomysial antibody positive but have had a previously normal 126 Gut 2008;57: doi: /gut

3 duodenal biopsy In patients with refratory oelia disease, in one small study, PE identified lymphoma in all four patients that were referred for investigation of refratory disease. 48 In a similar ohort of eight patients, PE diagnosed ulerative jejunitis in 50%. 56 Push enterosopy has also been shown to be useful in smaller studies in deteting rarer auses of diarrhoea suh as lymphangietasia and atypial infetions (ylospora, mirosporidia) 57 and sprue related strongyloidosis, 17 when duodenal biopsies have been normal. Push enterosopy to obtain jejunal biopsies should be onsidered in patients suspeted of malabsorption with positive anti-endomysial antibody and non-diagnosti duodenal biopsies. (reommendation grade C) Radiologial abnormalities The use of push enterosopy in the evaluation of abnormal radiographi studies has been shown to be helpful in onfirming small bowel pathology in 33 83% of ases However the endosopist has to be onfident that the area in question has been reahed, to ensure the validity of a negative endosopi examination. The enterosope should be advaned beyond the area as far as possible and fluorosopi verifiation an be helpful. 28 Push enterosopy is useful in investigation of proximal small bowel abnormalities deteted by radiology. (reommendation grade C) Small bowel tumours Small bowel tumours aount for 5 7% of patients presenting with OGB It is the most ommon ause in patients under years of age presenting with obsure GI bleeding. These patients may be asymptomati at early stages or present with abdominal pain, episodes of obstrution or weight loss. The most ommon loation for both epithelial and nonepithelial small bowel tumours is the jejunum while arinoids are more ommon in the ileum. 61 Diagnosti methods for small bowel tumours inlude enterolysis, omputed tomographi sanning, magneti resonane imaging, arteriography and enterosopy. In unseleted ase series the yield of small bowel tumours diagnosed during enterosopy is between 3.5 and 11% However, in some of these ases, there was already a suspiious lesion identified by small bowel imaging. Push enterosopy offers the important opportunity of taking biopsies when the neoplasti lesion has been identified. (reommendation grade C) However, this approah an only be taken for lesions within the reah of an enterosope. The adjuvant use of apsule endosopy may enhane the seletion of patients in whom proximal small bowel lesions ould be reahed and histology obtained. (b) Therapeuti Thermooagulation of bleeding lesions Angioetasia are the single most ommon ause of bleeding in patients above the age of 50 years and may aount for up to 80% of the diagnoses. 63 Angioetasia should be treated with thermooagulation to redue the inidenes of reurrent haemorrhage (reommendation grade B). Follow-up studies of patients with OGB and treatment initiated at enterosopy, demonstrated a redution in rebleeding episodes and transfusion requirement. Feeding jejunostomy Perutaneous endosopi jejunostomy (PEJ) plaement is a modifiation of the PEG method (perutaneous endosopi gastrostomy) desribed by Ponsky and Gauderer 65 to provide alimentation diretly into the small bowel. Indiations inlude prior gastri resetion or failure to loate the stomah due to abnormal anatomy and reurrent aspiration. 66 It an either be plaed diretly into the small bowel 66 or as a jejunal extension from a PEG The endosopist is responsible for assessing the need for topial anaesthesia and sedation. 69 The urrent BSG guidelines advise prophylati antibiotis for insertion of PEGs. 70 Intravenous antibiotis suh as efotaxime or o-amoxilav have been shown to be effetive in reduing peristomal infetion Further studies are needed to assess their role in PEJ plaements. With diret PEJ insertion, push enterosopy is used to get into the jejunum. The tip of the enterosope is manoeuvred to obtain lear transillumination through the abdominal wall before the stylet is introdued into the jejunal lumen. Small bowel peristalsis may ause loss of the transilluminated site. 66 For this reason, are needs to be taken to prevent the jejunum sliding and inadvertent punture of other abdominal organs ourring. Compliations that have been reported with PEJ inlude bleeding, aspiration and oloni perforation. 66 Available data suggests that aspiration still ours despite more distal plaement of feeding tubes. This is thought to be due to aspiration of the patient s own oropharyngeal seretions due to underlying neurologial defiit or reflux of the feed With jejunal extensions, ommonly faed problems inlude olusion and kinking of the tube, as well as malposition or migration into the stomah. 68 Separation of the inner jejunal tube from the outer PEG tube and aspiration may also our. Push enterosopy is the method of hoie for endosopially plaed feeding jejunostomy. (reommendation grade C) () Surveillane Polyposis syndromes Patients with Peutz Jeghers syndrome (PJS), a hereditary disorder haraterised by muoutaneous pigmentation and hamartomatous polyposis of the GI trat, are at risk of developing ompliations as a result of small bowel obstrution, intussuseption and bleeding. The aim of management in these patients is to identify and remove the larger polyps endosopially or surgially before they ause ompliations. 61 Push enterosopy allows exploration and polypetomy in the jejunum whilst intraoperative enterosopy provides a supplementary means of removing polyps in the ileum Patients with familial adenomatous polyposis (FAP) are at risk of developing extra-oloni polyps, partiularly in the duodenum and periampullary region. Surveillane using a sideviewing endosope is reommended after the age of 20 years by experiened endosopists, unless the patient has symptoms that warrant investigations earlier. 77 The Spigelman lassifiation is used for staging of duodenal polyposis and is based on arhitetural parameters, grade of dysplasia, number and size of polyps. 78 Push enterosopy is used for endosopi sreening in FAP patients to identify high risk individuals. The best sreening method for small bowel polyps in both onditions is yet to be established. (reommendation grade C) 4.4 Intraoperative enterosopy Intraoperative enterosopy (IOE) allows omplete examination of the small bowel, and is the urrent gold standard for diagnosing obsure GI bleeding. It is performed when the soure of bleeding remains undiagnosed by onventional investigations and the bleeding is massive, ontinuous or reurrent. 79 The reported tehniques of IOE vary in several important aspets: approah to intra-abdominal aess (laparotomy versus laparosopy), enterosope used and tehnique of insertion (perorally or via multiple enterotomies) The introdutory route is Gut 2008;57: doi: /gut

4 hosen aording to the loation of the presumed pathology. The proedure is done jointly by the endosopist and a surgeon. The surgeon telesopes segments of the small bowel over the enterosope to aid passage. The muosa is inspeted on insertion to avoid muosal trauma being misdiagnosed as vasular lesions. The surgeon is also able to identify muosal lesions externally by transillumination from the enterosope. 89 The air-trapping tehnique, whih isolates segments of small bowel by gentle olusion of the distal aspet, avoids exessive air insufflation and allows metiulous muosal examination of eah segment Sequential segmental isolation and inspetion is done in an antegrade fashion. Bleeding sites an be oversewn or segmental resetions an be performed The diagnosti rate of intraoperative enterosopy for muosal disease has been reported to range from 70 to 100% Reported ompliations inlude prolonged post-operative ileus, muosal or serosal tears, wound infetion and multi-organ failure IOE is able to identify treatable lesions with resolution of bleeding It should, however, be reserved for a selet group, partiularly with the availability of double balloon enterosopy whih may allow omplete small bowel visualisation and endosopi treatment. Intraoperative endosopy should be reserved for patients with massive, ontinuous or reurrent gastrointestinal haemorrhage when other less invasive methods have failed to detet the soure of bleeding. (reommendation grade B) 4.5 Double balloon (push and pull) enterosopy The DBE system (Fujinon, In., Japan) onsists of a high resolution video endosope with an outer diameter of 8.5 mm and a working length of 200 m, and a flexible overtube with a length of 145 m and an outer diameter of 12 mm Air from a pressure ontrolled pump system is used to inflate and deflate the latex balloons that are attahed to the distal end of the enterosope. The inflated balloon on the overtube is used to maintain a stable position while the enterosope is advaned. The overtube balloon is deflated whilst the enterosope balloon is inflated, and the overtube is advaned along the distal end of the enterosope. This is desribed as the push proedure. This is followed by the pull proedure where both the enterosope and the overtube are pulled bak under endosopi guidane, with both balloons inflated. This proedure is repeated multiple times to visualise the entire small bowel The double balloon method redues looping of the endosope to a minimum. The average time for eah approah (per-oral or per-anal) is 75 min. 98 DBE an be performed under both onsious sedation and general anaestheti, the former being the preferred hoie in most studies Few ompliations have been reported with DBE: post-proedure abdominal pain whih may our in up to 20% of patients, 103 panreatitis, bleeding and small bowel perforation whih is more ommon after polypetomy of large polyps (.3 m in size) Comparison of DBE with other small bowel imaging modalities Abnormal lesions seen by apsule endosopy (CE) that are beyond the reah of the push enterosope have previously been managed either onservatively or by undertaking IOE or surgery. 112 DBE allows visualisation of the majority of the small bowel (by ombination of the oral and anal approah or oral approah alone). 105 DBE also has features of a onventional endosope suh as rinsing, sution, biopsy and, importantly, allows therapeuti intervention. The insertion route is hosen aording to the loation of the suspeted lesion. 106 Total enterosopy may not be neessary in the majority of patients where the small bowel pathology or bleeding soure is found and treated A suessful endpoint would be resolution of bleeding In addition, total enterosopy may not be ahieved in all ases However, in ases where total enterosopy is required, it is reommended that DBE via both anal and oral approah are not performed on the same day. 115 This limits the inreased risk of patient disomfort due to the longer proedure time and air insufflation. Insufflation of arbon dioxide during olonosopy, flexible sigmoidosopy and endosopi retrograde holangio-panreatography has been shown to redue patient disomfort in a small number of studies There are no published studies to date omparing the use of arbon dioxide versus air insufflation for routine DBE. Carbon dioxide insufflation has the potential to be a useful alternative in DBE due to the longer proedure time. For total enterosopy, the most distal point should be marked or tattooed. Studies omparing DBE and PE have shown that antegrade DBE is superior to PE in length of insertion A higher suess rate for deep intubation of the small bowel and improved diagnosti yield has been desribed. The diagnosti yield from DBE is between 43 and 83% with a subsequent hange in management for 57 84% of patients Whilst DBE may be more labour intensive, another advantage is that it allows to and fro observation and ontrolled movement. CE allows loalisation of lesions prior to DBE CE not only allows an initial imaging study for small bowel pathology but findings on CE may affet the endosopist s hoie of route of insertion for DBE. The ability to onfirm pathology and allow therapeuti appliation, makes DBE omplementary to CE. DBE may be preferable to IOE in angioetasia, as repeat proedures may be needed to ablate new lesions that develop over time. 121 In ases where surgery may still be required, biopsy sampling and India ink marking with DBE provides useful information to the surgeons. 95 There have also been other therapeuti appliations of DBE in the reported literature: the insertion of stents 132 and the removal of them in patients with previous Roux-en-Y surgery, 133 DBE assisted hromoendosopy in patients with FAP 134 and endosopi ultrasound (EUS) of the small bowel. 135 The use of EUS with DBE may be helpful to evaluate the depth of small bowel lesions and assessing the suitability of lesions for endosopi muosal resetion. DBE has also been used to remove retained apsules, preventing the need for surgery DBE should be used omplementary to CE partiularly in the ontext of therapeuti intervention beyond the reah of PE. (reommendation grade B) 5.0 CAPSULE ENDOSCOPY The apsule endosope (CE) is a 26 by 11 mm apsule ontaining a battery-powered omplementary metal oxide silion imager (CMOS), a transmitter, antenna and four light emitting diodes. 139 The imager is ativated by removal of the apsule from its magneti holder and takes two images per seond through the transparent plasti dome of the apsule. The apsule is swallowed and is propelled through the intestine by peristalsis. Currently, CE is manufatured by three ompanies: Pillam SB, Given Imaging Ltd, Yoqneam, Israel; Olympus Endoapsule from Olympus, Japan; and OMOM apsule endosope from Jinshan Siene and Tehnlogy Group, Chongqing, China (not urrently available in the UK). Whilst 128 Gut 2008;57: doi: /gut

5 the Pillam uses CMOS imaging, the other two prototypes of CE use harge-oupled devie (CCD) tehnology Tehnique Patients are fasted for between 8 and 12 h prior to the proedure. As the apsule usually leaves the stomah within 30 min, the patient is allowed to drink after 2 h and eat after 4 h. Images taken by the apsule are transmitted via eight sensors, whih are seured to the abdominal wall, to a batterypowered data reorder worn on a belt. The equipment is removed after 8 h (the approximate battery life) by whih time the apsule has reahed the aeum in 85% of ases. 140 On ompletion of the proedure, the data from the reorder is downloaded onto a omputer workstation whih allows approximately images to be viewed as a video. The average reading time of the video images takes between 40 and 60 min depending on the experiene of the endosopist. The yield of CE an be affeted by two problems: firstly, the presene of dark intestinal ontents in the distal small bowel whih may impair visualisation of the muosa, and seondly the rate of gastri emptying and small bowel transit whih ould lead to the exhaustion of the apsule batteries before the apsule reahes the ileo-aeal valve. Inomplete examination ours in 10 25% of ases There have been a small number of studies and numerous abstrats addressing the use of bowel preparation (polyethylene glyol solution/oral sodium phosphate) to improve small bowel visualisation and the use of prokinetis (metolopramide/domperidone/tegaserod/erythromyin) to aelerate transit times thereby improving the proportion of ases where the olon is reahed The urrent literature broadly suggests that by taking this approah, better quality of small bowel leanliness is ahieved; however, the optimal type of preparation, dosage and time of administration remains to be determined. In one prospetive randomised study, the diagnosti yield was also found to be higher after bowel preparation. 146 It has been also reported that aeal visualisation rates are lower in patients having apsule endosopy during hospitalisation. 143 Two small studies also suggested redued intra-luminal bubbles and improved muosal visibility after the administration of simethione prior to CE. The available data at present are insuffiient to make a firm reommendation for preparation of the patient for CE. (reommendation grade C) 5.2 Indiations for apsule endosopy Obsure gastrointestinal bleeding Small bowel Crohn s disease Assessment of oelia disease Sreening and surveillane for polyps in familial polyposis syndromes Obsure overt and oult gastrointestinal bleeding Capsule endosopy (CE) now has an established role in patients with persistent obsure gastrointestinal bleeding (OGB) who have had a negative gastrosopy and olonosopy. Most studies using CE in patients with OGB have been in omparison to other modalities of investigation of the small bowel. Prospetive studies have onsistently revealed a superior diagnosti yield for apsule endosopy ompared to push enterosopy in patients with OGB A reent meta-analysis (of 14 studies on patients with OGB) reported yields of 63% for CE and 28% for PE. 167 The yield of CE has also been shown to be superior to barium follow through and CT enterolysis in the ontext of OGB The seond meta-analysis of 17 studies (526 patients) supports these findings: the rate differene (ie, the absolute pooled differene in the rate of positive findings) between apsule endosopy and other investigative modalities for OGB was 37% (95% CI, 29.6 to 44.1). 165 The rate of rebleeding in patients with OGB and negative CE is signifiantly lower ompared to those with a positive CE (48% versus 4.6% respetively). 169 In patients with a negative CE and essation of bleeding, a onservative approah may be adopted. 169 In the subgroup of patients with negative results on initial apsule endosopy and persistent bleeding, a seond look apsule endosopy may be onsidered, as small studies have shown an additional yield of 35 75% (reommendation grade C) When omparing more invasive forms of endosopy (DBE) with apsule endosopy, diagnosti rates are similar. Studies omparing DBE and apsule endosopy have shown diagnosti yields of between % (for DBE) and % (for CE) Complete small bowel examination was ahieved more frequently by apsule endosopy 99 (90.6% ompared to 62.5%, respetively; p,0.05). Historially, intra-operative endosopy has been onsidered the gold standard in patients with OGB and negative standard endosopi evaluation. When ompared to intraoperative endosopy, apsule endosopy had sensitivity, speifiity, positive and negative preditive values of 95%, 75%, 95% and 86%, respetively. 172 An algorithm for investigation of patients with OGB is suggested in fig (reommendation grade B) Crohn s disease The small bowel is ommonly affeted by Crohn s disease. Endosopially, however, the small bowel is relatively inaessible. In addition, ileal intubation is not always ahieved at olonosopy. Small bowel ontrast studies have variable suess rates in diagnosing ative Crohn s disease Whilst CT may be effetive in diagnosing small bowel thikening and ompliations of Crohn s disease, its auray in determining the presene of muosal disease is unknown. This diffiulty partly explains a mean delay of between 1 and 7 years from onset of symptoms to diagnosis. A number of studies have now addressed the question of how best to investigate patients in whom onventional tests have failed to onfirm a diagnosis of ative Crohn s disease. These inlude patients with symptoms of pain, diarrhoea, weight loss, or investigational findings inluding iron defiient anaemia and an aute phase response. 179 Whih ombination of these features aurately predits a diagnosis of Crohn s disease is not known, but a onsensus group has suggested that further investigation using CE might be onsidered in patients with two or more of these riteria. 179 (reommendation grade C) A number of studies performed have ompared apsule endosopy with olonosopy and ileosopy, small bowel follow through, CT enterolysis and MRI In addition to onfirming suspeted Crohn s disease and assessing disease extent, CE has also been used in the ontext of reurrene of disease post-operatively. 185 Capsule endosopy versus endosopy Evidene of Crohn s disease was found by apsule endosopy in 43 71% of patients typially suspeted of having Crohn s disease in whih olonosopy (and small bowel radiography) had previously been normal An analysis of four prospetive omparative studies (total of 115 patients) showed a diagnosti yield of 61% for CE ompared to 46% for ileoolonosopy in the detetion of small bowel Crohn s (p = 0.02; Gut 2008;57: doi: /gut

6 Figure 1 Proposed role of apsule endosopy and enterosopy in obsure gastrointestinal bleeding. 95% CI, 2 to 27). 179 CE was also able to identify the extent of disease proximal to the terminal ileum. CE has been found to have a greater diagnosti yield when ompared to PE in patients known to have established Crohn s disease perhaps refleting the greater extent of small bowel muosa visualised during CE. The use of CE for reognition of disease reurrene within 6 months of ileo-oloni resetion, had a reported sensitivity of between 62 and 76% ompared to 90% for ileo-olonosopy. 185 However, CE did identify lesions outside the reah of a ileoolonosope. This data does not neessarily represent that of routine linial pratie: apsules entered the olon in all ases (ompared to a reported inomplete examination in 10 25% of other series) and all patients had suessful ileo-olonosopy (ompared to an average UK rate of 57% for aeal intubation). 188 Ileo-olonosopy has a higher yield in the detetion of reurrent disease ompared to CE in patients post ileo-oloni resetion. (reommendation grade C) Capsule endosopy versus small bowel radiology In patients with suspeted new or reurrent Crohn s disease, CE was more likely to identify ative disease than small bowel barium imaging. Studies omparing CT enterolysis with CE also showed a higher yield of small bowel uleration for CE. The two studies omparing CE and MR enterolysis showed either omparable or better yield for CE An important observation from most radiologial versus CE studies is that radiologial examination was able to delineate the presene of stritures whih preluded the use of CE in a signifiant number of patients. A reent meta-analysis made a omparison of CE versus other modalities in established and suspeted Crohn s disease. 190 In the evaluation of reurrene, CE is superior to both barium studies and ileo-olonosopy in established non-strituring Crohn s disease. (reommendation grade B) However, despite a higher yield of CE in omparison to other modalities in the suspeted Crohn s group, the sub-analysis did not show a statistially signifiant differene in favour of CE in this group. 190 Larger studies are needed to better establish the role of CE in the diagnosis of suspeted Crohn s disease. (reommendation grade C) Capsule retention remains a risk in patients with Crohn s disease even in the presene of radiologial investigations that do not show signifiant stritures. This is disussed in more detail in setion 5.3. In the studies referred to, with predominantly Crohn s patients, retention ourred in 0 6.7% of ases and apsules passed either after medial treatment of Crohn s disease, endosopi removal 191 or surgery The risk is greater in patients with established Crohn s disease ompared to patients suspeted to have Crohn s disease. 195 CE should be onsidered in patients with a high suspiion of small bowel Crohn s disease undeteted by onventional means. These patients should have radiologial imaging to exlude stritures prior to CE. (reommendation grade C) 130 Gut 2008;57: doi: /gut

7 Figure 2 The use of apsule endosopy and double balloon enterosopy in the investigation of Crohn s disease. An algorithm for the investigation of patients suspeted of having Crohn s disease using CE is suggested in fig. 2. (reommendation grade C) Coelia disease There have been two reported roles for the use of CE in oelia disease. Firstly, typial muosal hanges of oelia disease has been reognised at CE inluding a mosai pattern, salloping, otopus leg appearane, loss of muosal folds and atrophy As a result there have been small studies using CE as virtual histology in onjuntion with positive oelia serology, as the muosal hanges seen on CE is omparable to the marosopi appearane at endosopy. The sensitivity, speifiity, positive and negative preditive values of CE for oelia disease has been reported as 70%, 100%, 100% and 77%, respetively. 196 At present, duodenal biopsy remains the gold standard and there is insuffiient evidene for CE for the routine diagnosis of oelia disease. (reommendation grade C) The seond group of patients who would benefit from CE are those with known oelia disease established on a gluten free diet but with ongoing symptoms or those who develop alarm symptoms. These patients often undergo extensive radiologial and sometimes surgial evaluation to look for possible ompliations of ulerative jejunitis and small bowel lymphoma A reported study showed a yield of 60% in detetion of oelia related ompliations inluding ulerated muosa, striture and malignany. 198 CE may be indiated in the diagnosis of ompliations of oelia disease. (reommendation grade C) Familial polyposis syndromes There is a small number of studies looking at the use of CE in surveillane of polyposis syndromes (familial adenomatous polyposis and Peutz Jegher s syndrome) CE is more aurate in detetion of polyps than small bowel follow through and it an also detet smaller polyps in omparison to MRI. 205 Given the limited number of studies, the routine use of CE in patients with polyposis syndromes is urrently not advoated. The effet of CE on the hange of management in this group of patients also needs further larifiation. (reommendation grade C) 5.3 Compliations of apsule endosopy The main risk of CE is apsule retention. CE is ontraindiated in patients with known stritures or swallowing disorders. Patients with extensive small bowel Crohn s (disussed in setion 5.2.2) hroni usage of non-steroidal anti-inflammatory drugs and abdominal radiation injury are at higher risk. Patients should be fully informed about the risk of retention before onsent for CE is undertaken. It should be highlighted that further intervention inluding surgery may be required if Gut 2008;57: doi: /gut

8 passage of the apsule is impeded by a striture. Capsule retention has been defined by the International Conferene on Capsule Endosopy (ICCE) working group, as the apsule remaining in the digestive trat for 2 weeks or more requiring direted medial, endosopi or surgial intervention. 195 A large study (937 patients) reported an inidene of 0.75% of patients worldwide who required surgial intervention to remove a retained apsule. 208 An alternative imaging modality should be onsidered prior to CE in patients with obstrutive symptoms. (reommendation grade B) The absene of stritures on a barium study however does not entirely prelude the apsule being safely passed, as retention is known to our despite a normal barium or enterolysis study. In ertain situations, however, CE may be used to diagnose an obstruting lesion not identified by other tehniques and the apsule removed at surgery. (reommendation grade C) A plain abdominal radiograph should be obtained to onfirm exretion of apsule if the video fails to show that it enters the olon. Patients should not undergo magneti resonane imaging after CE until they have safely passed the apsule. Oasionally the apsule may be retained in the stomah due to gastroparesis. In these ases, speifially designed apsule delivery systems are reommended for delivery of the apsule diretly into the small bowel (reommendation grade C) There is theoretial potential for interferene between the radiofrequeny of the apsule, data reorder and permanent paemakers (PPM) and implantable ardia defibrillators (ICD). The manufaturers of CE have listed them as a relative ontraindiation for use of CE. Small studies have tested the use of CE in patients with these devies and have shown it to be safe without adverse events or interferene of apsule images Larger studies are required to verify its safe use. Advie should also be obtained either from the manufaturers of the ardia devie or the ardiologists to ensure that the apsule does not affet funtion of the ardia devie (reommendation grade C) 5.4 Pateny apsule The M2A pateny apsule was designed to overome the potential hazard of apsule retention in high risk patients. This apsule is idential to the video apsule in size and shape. It is filled with latose and proteted by a plug with a speifially sized hole that allows the influx of intestinal fluid if impated in stenosed bowel, whih in turn dissolves the latose in a predetermined time of approximately 40 h. 217 The pateny apsule also has a transmitter whih allows it to be deteted by a hand-held sanner plaed lose to the anterior abdominal wall. Small studies have reommended its safe use in patients with known small bowel stritures whilst one study showed that it an preipitate symptomati intestinal olusion. 219 The olusion may have ourred beause the latose plug requires fluid to dissolve and the distal side of an obstruted striture may be relatively dry. More reently, the Agile pateny apsule (Given Imaging, Yoqneam, Israel) whih has dissolvable plugs at both ends has been devised to improve its use as a non-invasive tool in the assessment of funtional pateny of intestinal stritures. Larger studies are needed before the pateny apsule an be reommended for routine use in the high risk group. (reommendation grade C) 6.0 SERVICE PROVISION AND TRAINING The demand for CE has risen sine its introdution in the United Kingdom. This is refleted by the inrease in the number of entres whih offer this servie. In addition to developing a role in the investigation pathway of OGB and IBD, the use of CE is ost effetive by preventing unneessary yles of investigations in patients. The reading of apsule endosopy videos remains a time onsuming exerise for gastroenterologists. Few studies have ompared the inter-observer variability between an experiened gastroenterology or endosopy nurse against a physiian Other investigators have also made omparisons between physiians of different levels of experiene (endosopy fellows or juniors endosopists versus experiened physiians). 228 These studies have shown that trainees were able to interpret CE images and reah the orret diagnosis in all linially relevant ases. Speialist registrars and nurse speialists who have an interest in the small bowel may wish to take up this role. Inorporation of a setion on apsule endosopy into the generi urriulum would help to formalise the training in this field. Despite the expansion of the servie of apsule endosopy, double balloon enterosopy is likely to remain as a regional servie. A DBE users group has reently been established to help promote standards, uniformity of pratie and training aross the UK. Like apsule endosopy, formal training and perhaps, in addition, a basi skills ourse should be mandatory for all wishing to pratise DBE. Regular audit of the servie should be arried out at appropriate intervals. (reommendation grade C) Competing interests: Delared (the delaration an be viewed on the Gut website at REFERENCES 1. Costamagna G, Shah SK, Riioni ME, et al. A prospetive trial omparing small bowel radiographs and video apsule endosopy for suspeted small bowel disease. Gastroenterology 2002;123: Nolan DJ, Traill ZC. The urrent role of the barium examination of the small intestine. Clin Radiol 1997;52: Voderholzer WA, Ortner M, Rogalla P, et al. Diagnosti yield of wireless apsule enterosopy in omparison with omputed tomography enterolysis. Endosopy 2003;35: Grimshaw J, Eles M, Russell I. Developing linially valid pratie guidelines. J Eval Clin Prat 1995;1: Eles M, Clapp Z, Grimshaw J, et al. North of England evidene based guidelines development projet: methods of guideline development. BMJ 1996;312: Eisen GM, Dominitz JA, Faigel DO, et al. Enterosopy. Gastrointest Endos 2001;53: Mishkin DS, Chuttani R, Croffie J, et al. ASGE Tehnology Status Evaluation Report: wireless apsule endosopy. Gastrointest Endos 2006;63: Rey JF, Gay G, Kruse A, et al. European Soiety of Gastrointestinal Endosopy guideline for video apsule endosopy. Endosopy 2004;36: Rey JF, Ladas S, Alhassani A, et al. European Soiety of Gastrointestinal Endosopy (ESGE)Video apsule endosopy: Update to guidelines (May 2006). Endosopy 2006;38: Tada M, Akasaka Y, Misaki F, et al. Clinial evaluation of a sonde-type small intestinal fibersope. Endosopy 1977;9: MaKenzie JF. Push enterosopy. Gastrointest Endos Clin N Am 1999;9: Chak A, Koehler MK, Sundaram SN, et al. Diagnosti and therapeuti impat of push enterosopy: analysis of fators assoiated with positive findings. Gastrointest Endos 1998;47: Davies GR, Benson MJ, Gertner DJ, et al. Diagnosti and therapeuti push type enterosopy in linial use. Gut 1995;37: Chong J, Tagle M, Barkin JS, et al. Small bowel push-type fiberopti enterosopy for patients with oult gastrointestinal bleeding or suspeted small bowel pathology. Am J Gastroenterol 1994;89: Hayat M, Axon AT, O Mahony S. Diagnosti yield and effet on linial outomes of push enterosopy in suspeted small-bowel bleeding. Endosopy 2000;32: Shimizu S, Tada M, Kawai K. Development of a new insertion tehnique in pushtype enterosopy. Am J Gastroenterol 1987;82: Landi B, Tkoub M, Gaudri M, et al. Diagnosti yield of push-type enterosopy in relation to indiation. Gut 1998;42: Taylor AC, Chen RY, Desmond PV. Use of an overtube for enterosopy does it inrease depth of insertion? A prospetive study of enterosopy with and without an overtube. Endosopy 2001;33: Iida M, Yamamoto T, Yao T, et al. Jejunal endosopy using a long duodenofibersope. Gastrointest Endos 1986;32: Yang R, Laine L. Muosal stripping: a ompliation of push enterosopy. Gastrointest Endos 1995;41: Gut 2008;57: doi: /gut

9 21. Barkin JS, Lewis BS, Reiner DK, et al. Diagnosti and therapeuti jejunosopy with a new, longer enterosope. Gastrointest Endos 1992;38: Keizman D, Brill S, Umansky M, et al. Diagnosti yield of routine push enterosopy with a graded-stiffness enterosope without overtube. Gastrointest Endos 2003;57: Lin S, Branh MS, Shetzline M. The importane of indiation in the diagnosti value of push enterosopy. Endosopy 2003;35: Sharma BC, Bhasin DK, Makharia G, et al. Diagnosti value of push-type enterosopy: a report from India. Am J Gastroenterol 2000;95: Bouhnik Y, Bitoun A, Coffin B, et al. Two way push videoenterosopy in investigation of small bowel disease. Gut 1998;43: Chak A, Cooper GS, Canto MI, et al. Enterosopy for the initial evaluation of iron defiieny. Gastrointest Endos 1998;47: Barkin JS, Chong J, Reiner DK. First-generation video enterosope: fourthgeneration push-type small bowel enterosopy utilizing an overtube. Gastrointest Endos 1994;40: Pennazio M, Arrigoni A, Risio M, et al. Clinial evaluation of push-type enterosopy. Endosopy 1995;27: Cuillerier E, Landi B, Cellier C. Is push enterosopy useful in patients with malabsorption of unlear origin? Am J Gastroenterol 2001;96: Morris AJ, Mokhashi M, Straiton M, et al. Push enterosopy and heater probe therapy for small bowel bleeding. Gastrointest Endos 1996;44: Benz C, Jakobs R, Riemann JF. Does the insertion depth in push enterosopy depend on the working length of the enterosope? Endosopy 2002;34: Szold A, Katz LB, Lewis BS. Surgial approah to oult gastrointestinal bleeding. Am J Surg 1992;163:90 2; disussion Gralnek IM. Obsure-overt gastrointestinal bleeding. Gastroenterology 2005;128: Prakash C, Zukerman GR. Aute small bowel bleeding: a distint entity with signifiantly different eonomi impliations ompared with GI bleeding from other loations. Gastrointest Endos 2003;58: Zukerman GR, Prakash C, Askin MP, et al. AGA tehnial review on the evaluation and management of oult and obsure gastrointestinal bleeding. Gastroenterology 2000;118: Marshall JK, Lesi OA, Hunt RH. Obsure gastrointestinal bleeding: an approah to management. Can J Gastroenterol 2000;14: Zaman A, Katon RM. Push enterosopy for obsure gastrointestinal bleeding yields a high inidene of proximal lesions within reah of a standard endosope. Gastrointest Endos 1998;47: Vakil N, Huilgol V, Khan I. Effet of push enterosopy on transfusion requirements and quality of life in patients with unexplained gastrointestinal bleeding. Am J Gastroenterol 1997;92: Lepere C, Cuillerier E, Van Gossum A, et al. Preditive fators of positive findings in patients explored by push enterosopy for unexplained GI bleeding. Gastrointest Endos 2005;61: Desamps C, Shmit A, Van Gossum A. Missed upper gastrointestinal trat lesions may explain oult bleeding. Endosopy 1999;31: Romelaer C, Le Rhun M, Beaugerie L, et al. Push enterosopy for gastrointestinal bleeding: diagnosti yield and long-term follow-up. Gastroenterol Clin Biol 2004;28: Lara LF, Bloomfeld RS, Pineau BC. The rate of lesions found within reah of esophagogastroduodenosopy during push enterosopy depends on the type of obsure gastrointestinal bleeding. Endosopy 2005;37: Nguyen NQ, Rayner CK, Shoeman MN. Push enterosopy alters management in a majority of patients with obsure gastrointestinal bleeding. J Gastroenterol Hepatol 2005;20: Taylor AC, Buttigieg RJ, MDonald IG, et al. Prospetive assessment of the diagnosti and therapeuti impat of small-bowel push enterosopy. Endosopy 2003;35: Linder J, Cheruvattath R, Truss C, et al. Diagnosti yield and linial impliations of push enterosopy: results from a nonspeialized enter. J Clin Gastroenterol 2002;35: Chen RY, Taylor AC, Desmond PV. Push enterosopy: a single entre experiene and review of published series. ANZ J Surg 2002;72: Fouth PG, Sawyer R, Sanowski RA. Push-enterosopy for diagnosis of patients with gastrointestinal bleeding of obsure origin. Gastrointest Endos 1990;36: Parry SD, Welfare MR, Cobden I, et al. Push enterosopy in a UK distrit general hospital: experiene of 51 ases over 2 years. Eur J Gastroenterol Hepatol 2002;14: Sidhu R, MAlindon ME, Kapur K, et al. Push enterosopy in the era of apsule endosopy. J Clin Gastroenterol (in press). 50. Shmit A, Gay F, Adler M, et al. Diagnosti effiay of push-enterosopy and longterm follow-up of patients with small bowel angiodysplasias. Dig Dis Si 1996;41: Gomez Rodriguez BJ, Ortiz Moyano C, Romero Castro R, et al. Diagnosti yield of 335 push video-enterosopies. Rev Esp Enferm Dig 2006;98: Jabbari M, Cherry R, Lough JO, et al. Gastri antral vasular etasia: the watermelon stomah. Gastroenterology 1984;87: Sidhu R, Sanders DS, MAlindon ME. Does apsule endosopy reognise gastri antral vasular etasia (GAVE) more frequently than onventional endosopy? J Gastrointestin Liver Dis 2006;15: Horoldt BS, MAlindon ME, Stephenson TJ, et al. Making the diagnosis of oelia disease: is there a role for push enterosopy? Eur J Gastroenterol Hepatol 2004;16: Hopper AD, Horoldt BS, Stephenson TJ, et al. Variability of histologi lesions in relation to biopsy site in gluten-sensitive enteropathy what are the impliations for diagnosing adult elia disease? Am J Gastroenterol 2005;100:2128; author reply Cellier C, Cuillerier E, Patey-Mariaud de Serre N, et al. Push enterosopy in elia sprue and refratory sprue. Gastrointest Endos 1999;50: Rossini FP, Arrigoni A, Pennazio M. Clinial enterosopy. J Clin Gastroenterol 1996;22:231 5; disussion O Mahony S, Morris AJ, Straiton M, et al. Push enterosopy in the investigation of small-intestinal disease. QJM 1996;89: Thompson JN, Salem RR, Hemingway AP, et al. Speialist investigation of obsure gastrointestinal bleeding. Gut 1987;28: Lewis BS, Kornbluth A, Waye JD. Small bowel tumours: yield of enterosopy. Gut 1991;32: Rossini FP, Risio M, Pennazio M. Small bowel tumors and polyposis syndromes. Gastrointest Endos Clin N Am 1999;9: Fouth PG. Angiodysplasia of the gastrointestinal trat. Am J Gastroenterol 1993;88: Askin MP, Lewis BS. Push enterosopi auterization: long-term follow-up of 83 patients with bleeding small intestinal angiodysplasia. Gastrointest Endos 1996;43: Swain CP. The role of enterosopy in linial pratie. Gastrointest Endos Clin N Am 1999;9: Ponsky JL, Gauderer MW. Perutaneous endosopi gastrostomy: a nonoperative tehnique for feeding gastrostomy. Gastrointest Endos 1981;27: Shike M, Latkany L, Gerdes H, et al. Diret perutaneous endosopi jejunostomies for enteral feeding. Gastrointest Endos 1996;44: Kaplan DS, Murthy UK, Linsheer WG. Perutaneous endosopi jejunostomy: long-term follow-up of 23 patients. Gastrointest Endos 1989;35: DiSario JA, Fouth PG, Sanowski RA. Poor results with perutaneous endosopi jejunostomy. Gastrointest Endos 1990;36: BSG. Guidelines on Safety and Sedation for Endosopi Proedures. British Soiety of Gastroenterology, June 2006). 70. BSG. Antibioti prophylaxis in gastrointestinal endosopy. British Soiety of Gastroenterology, Jan June 2006). 71. Gossner L, Keymling J, Hahn EG, et al. Antibioti prophylaxis in perutaneous endosopi gastrostomy (PEG): a prospetive randomized linial trial. Endosopy 1999;31: Dormann AJ, Wigginghaus B, Risius H, et al. Antibioti prophylaxis in perutaneous endosopi gastrostomy (PEG) results from a prospetive randomized multienter trial. Z Gastroenterol 2000;38: Prelik G, Grune S, Leser HG, et al. Prospetive, randomised, double blind trial of prophylaxis with single dose of o-amoxilav before perutaneous endosopi gastrostomy. BMJ 1999;319: Kadakia SC, Sullivan HO, Starnes E. Perutaneous endosopi gastrostomy or jejunostomy and the inidene of aspiration in 79 patients. Am J Surg 1992;164: Keshtgar AS, Losty PD, Lloyd DA, et al. Reent developments in the management of Peutz Jeghers syndrome in hildhood. Eur J Pediatr Surg 1997;7: Seenath MM, Sott MJ, Morris AI, et al. Combined surgial and endosopi learane of small-bowel polyps in Peutz Jeghers syndrome. J R So Med 2003;96: Sokol RJ, Winter H. Working group on endosopi and diagnosti tehniques. World Congress of Paediatri Gastroenterology, Hepatology and Nutrition. Lipppinott Williams & Wilkins, In. 2000: Spigelman AD, Williams CB, Talbot IC, et al. Upper gastrointestinal aner in patients with familial adenomatous polyposis. Lanet 1989;2: Lau WY, Fan ST, Wong SH, et al. Preoperative and intraoperative loalisation of gastrointestinal bleeding of obsure origin. Gut 1987;28: Lewis MP, Khoo DE, Spener J. Value of laparotomy in the diagnosis of obsure gastrointestinal haemorrhage. Gut 1995;37: Desa LA, Ohri SK, Hutton KA, et al. Role of intraoperative enterosopy in obsure gastrointestinal bleeding of small bowel origin. Br J Surg 1991;78: Ress AM, Benai JC, Sarr MG. Effiay of intraoperative enterosopy in diagnosis and prevention of reurrent, oult gastrointestinal bleeding. Am J Surg 1992;163:94 8; disussion Kendrik ML, Buttar NS, Anderson MA, et al. Contribution of intraoperative enterosopy in the management of obsure gastrointestinal bleeding. J Gastrointest Surg 2001;5: Ingrosso M, Prete F, Pisani A, et al. Laparosopially assisted total enterosopy: a new approah to small intestinal diseases. Gastrointest Endos 1999;49: Lau WY, Wong SY, Ngan H, et al. Intra-operative loalization of bleeding small intestinal lesions. Br J Surg 1988;75: Greenberg GR, Phillips MJ, Tovee EB, et al. Fiberopti endosopy during laparotomy in the diagnosis of small intestinal bleeding. Gastroenterology 1976;71: Ozmen MM, Yilmaz U, Kale IT, et al. Intra-operative enterosopy for obsure gastrointestinal bleeding. Hepatogastroenterology 1999;46: Douard R, Wind P, Panis Y, et al. Intraoperative enterosopy for diagnosis and management of unexplained gastrointestinal bleeding. Am J Surg 2000;180: Gut 2008;57: doi: /gut