GI Lesions Research Articles

Mo1484 The Utility of Endoscopic Ultrasound (EUS)-Fine Needle Aspiration (FNA) for the Diagnosis of Pancreatic Cystic Neuroendocrine Tumors (PCNETs): a Decade of Single Center Experience

performance of EUS guided core biopsy. The yield of the core biopsy as compared to FNA with onsite cytopathologists is not known. Methods: All patients referred for EUS guided biopsy were considered for EUS guided core biopsy followed by EUS guided FNA between February 2011 and November 2012 at the Houston Veterans’ Affairs Medical Center. Patients were excluded if they had cystic lesions, small lesions ( 1cm) or if interposing vascular structures prevented the intervention. Study patients were evaluated with two core biopsies using 22 gauge needles followed by FNA guided by a bedside cytopathologist. Results of the core biopsy were reported by a second pathologist at a later date. The yield (percent accurate diagnosis) was compared between the two groups. The agreement between the 2 approached was examined using the Kappa (K) statistic. The final diagnosis was defined using a conglomerate of surgical pathology, follow-up studies, and clinical course. Results: Twenty six patients were included in the final analysis, of which 96% were male and 61% were Caucasian. 69.2% had pancreatic lesions and 30.8% had other upper GI lesions including lymph nodes and subepithelial lesions. The mean number of passes was 3.2 for FNA, and the mean procedure time was 39.4 minutes. The final diagnosis was malignant in 92.3% and benign in 7.7%. Sensitivity and specificity were 83% and 100% respectively for FNA, and 71% and 100% respectively for core biopsy. Diagnostic accuracy was 84.6% for FNA and 73.1% for core biopsy. The two approaches yield was in agreement in 88.4% with a kappa statistic of 0.66 (95% CI: 0.33 0.99) indicating substantial agreement. Conclusions: The diagnostic yield of two passes with a 22 gauge core biopsy needle is comparable to the current gold standard of FNA with a bedside cytopathologist. Performance of two core biopsies could represent a time efficient and widely available alternative to FNA with a bedside cytopathologist.

Chronic Constipation Caused by Neurofibromatous Proliferation in A Case of Von Recklinghausen’s Disease – A Case Report

Neurofibromatosis (NF) is an autosomal dominant genetic disorder with a myriad of clinical manifestations . A gastrointestinal involvement which is present in 10 -25% of patients, is usually a systemic manifestation of generalised NF. We are describing a case of NF 1 with chronic constipation, in whom colonoscopy revealed a thickening of the colon wall with narrowing. A mucosal biopsy showed neural hyperplasia .This case emphasizes the value of minimally invasive endoscopic biopsies of GI lesions in NF1, where despite a limited sampling, correlation with clinical and endoscopic features may help in reaching a diagnosis of a neurofibromatous proliferation. We have also discussed the differential diagnosis of gastrointestinal lesions in NF with neural hyperplasia.

Prevalence and Predictive Factors for Gastrointestinal Pathology in Young Men Evaluated for Iron Deficiency Anemia

The prevalence of gastrointestinal lesions in young men with iron deficiency anemia (IDA) is unknown, and there are no evidence-based recommendations for the evaluation of the gastrointestinal tract in this population. The purpose of this study was to assess the prevalence of significant GI lesions among young males with IDA, and to shed light on potential predictors of their presence. Clinical, endoscopic, and histological data was retrospectively collected from medical records of 347 young males with IDA. Clinically significant GI lesions were diagnosed in 62 %. Upper GI lesions were found in 35 %. Peptic disease was the most common finding, diagnosed in 30 %. Celiac disease was diagnosed in 4 %. Lower GI tract lesions were diagnosed in 34 %. The most common findings were hemorrhoids (17 %) and inflammatory bowel disease (16 %). Malignant lesions were not detected. GI lesions were encountered more frequently when respective symptoms were obtained. Multivariate analysis showed that the presence of GI symptoms and the use of proton pump inhibitors were associated with an increased likelihood of significant GI lesions. GI lesions are common among young men with IDA. GI evaluation is mandatory in symptomatic men and in asymptomatic men when the anemia is resistant to iron therapy. Symptoms may dictate the order of evaluation.

PMO-194 Polypoid lesions in the UGI tract in patients with portal hypertension; EUS before you biopsy!

IntroductionThe universal use of upper gastrointestinal (UGI) endoscopy in patients with portal hypertension in combination with increasing number of patients with liver disease has resulted in the detection of indeterminate...

Immunochemical FOBT was accurate for detecting colorectal cancer but less so for other GI lesions

Source Citation Chiang TH, Lee YC, Tu CH, Chiu HM, Wu MS. Performance of the immunochemical fecal occult blood test in predicting lesions in the lower gastrointestinal tract. CMAJ. 2011;183:1474-81. 21810951

P205 A new rapid test for fecal calprotectin (FC) predicts mucosal healing in ulcerative colitis (UC)

tertiary IBD center were reviewed. Inclusion criteria: (1) Small bowel CD assessed by conventional techniques; (2) Detailed clinical history; (3) No stenosis/obstructions; (4) Small bowelCD assessed by entero-CT/-MRI, small intestinal contrast ultrasonography, barium follow through and/or ileocolonoscopy (IC) within 6mths from SBCE; (5) SBCE images reviewed by one single gastroenterologist. SBCE performed using the Given Pillcam SB (Given, Israel) after 2L PEG. Findings considered: (a) CD extent before vs after SBCE; (b) upper GI lesions (i.e. above the distal ileum proximal to the valve or to the anastomosis); (c) SBCE transit times; (d) impact rate. Results: A total of 40 CD pts (20M, median age 34 yrs, range 18 70) fulfilled the inclusion criteria. Previous ileo-colonic resection was shown by 29/40 (72.5%) pts (median interval 10mths, range 3 300). CD localization involved the ileum in 37/40 (92%) CD (recurrence in 26/37) and also the upper GI in 4/37 (10%), while 3 pts showed no recurrence. SBCE showed marked interindividual variations in terms of gastric (median 29min, range 3 360) and small bowel transit times (valve: median 350min, range 240 433 or anastomosis in resected CD: median 231 min, range 61 466). SBCE did not visualize the entire small bowel in 9/40 (22.5%) pts (5/11 CD not resected, 4/29 resected pts). Upper GI lesions were detected by SBCE in 23/40 (57.5%) pts, known before SBCE in only 4/40 (10%) pts. In one pt, SBCE visualized an ulcerated jejunal stenosis requiring surgery not detected by SBFT and IC. In 1/40 (2.5%) pts, SBCE impact requiring surgery within a “cul de sac” of an ileo-ileal anastomosis was observed. Conclusions: Upper GI lesions may be detected by SBCE in a high proportion of CD pts, although their clinical relevance needs to be determined. Before SBCE, CD pts need a careful selection due to possible incomplete small bowel visualization and to the impact risk occurring also in the absence of stenosis.

Prognostic value of metabolic tumor volume on PET / CT in primary gastrointestinal diffuse large B cell lymphoma

Primary gastrointestinal (PGI) diffuse large B cell lymphoma (DLBCL) is a relatively common disease. Recent studies indicate that measurement of maximum standardized uptake value (SUV(max)) on pretreatment for (18)F-fluorodeoxyglucose PET is an important prognostic factor in PGI DLBCL. However, there is still an association between initial tumor burden and prognosis. Thus, in the present study, we investigated whether tumor volume by PET could have a potential prognostic value to predict the outcome. From 2006 to 2009, 165 Stage I E/II E PGI DLBCL patients were enrolled in the study. One hundred and five patients received cyclophosphamide, doxorubicin, vincristine, prednisone, and rituximab (R-CHOP) only, whereas 60 patients underwent surgery plus R-CHOP. Metabolic tumor volume (MTV) was defined initial tumor burden as target GI lesion above SUV, 2.5 by PET as a contouring border. Over a median follow-up period of 36.6 months, receiver operating characteristic (ROC) analysis indicated that the best cut-off values for MTV and SUV(max) were 160.1 cm(3) and 12.0, respectively. The estimated area under the ROC curve was higher for MTV than SUV(max). Thus, MTV was a better predictor for survival than SUV(max). In patients with a low MTV (<160.1 cm(3)), there were no significant differences in survival between patients undergoing R-CHOP alone and surgery plus R-CHOP (P = 0.347 for progression-free survival [PFS]; P = 0.148 for overall survival [OS]). Conversely, in patients with a high MTV (>160.1 cm(3)), survival was longer in those who underwent surgery plus R-CHOP than in those treated with R-CHOP alone (P < 0.001 for PFS; P < 0.001 for OS). Multivariate analysis revealed that high MTV is an independent factor for predicting survival. Even in the era of rituximab, treatment of PGI DLBCL is not easy in patients with a high MTV.

Endoscopic submucosal dissection with a flexible Maryland dissector: randomized comparison of mesna and saline solution for submucosal injection (with videos)

Endoscopic submucosal dissection (ESD) is increasingly used for en bloc removal of GI lesions. Current ESD techniques have limitations including long procedure times, technical difficulty, and complications. To compare mesna with saline solution for ESD. Blinded, randomized, controlled, porcine study in live animals. Animal laboratory. Twelve gastric lesions were marked by using electrocautery. After submucosal injection, a circumferential mucosal incision was created, and ESD was performed by using a flexible Maryland dissector. Half of the ESDs were performed with submucosal injection of mesna. Primary outcome was the time to dissect the submucosal plane. Secondary outcomes were total ESD time, specimen size, and procedure related complications. The average (± SD) time for dissecting the submucosal plane was 15 minutes (range 10-22 ± 4.8 min) in the group with submucosal mesna injection and 16 minutes (range 8-29 ± 8.3 min) in the control group (P = 1.0). Complete en bloc resection including all of the electrocautery markings was achieved in all cases. Injection of mesna did not provide any benefit over saline solution in terms of overall ESD time (24 ± 7.3 min vs 28 ± 11 min; P = .42). There were no perforations. Four hemorrhages requiring intervention were encountered during the procedures in the control group, compared with no bleeding in the mesna group (P = .09). Animal model, limited sample size. Submucosal mesna injection did not affect procedure times but was associated with a trend toward a lower incidence of intraprocedural bleeding.

Optical biopsies by confocal endomicroscopy prevent additive endoscopic biopsies before endoscopic submucosal dissection in gastric epithelial neoplasias: a prospective, comparative study

Confocal laser endomicroscopy (CLE) allows real-time in vivo histologic evaluation of GI lesions. To our knowledge, there is no reported prospective study comparing endoscopic and optical biopsy using CLE before endoscopic submucosal dissection (ESD) with post-ESD histopathology. We compared endoscopic and optical biopsy before ESD and thereby assessed the ability of CLE to effectively diagnose and differentiate gastric epithelial neoplasia. Single tertiary-care center, prospective comparative study. Soonchunhyang University Hospital, between September 2009 and April 2010. This study involved 31 patients with 35 gastric epithelial neoplasias, previously diagnosed by endoscopic biopsy, who were scheduled for ESD. Target lesions were imaged in vivo by using CLE. The overall accuracy of endoscopic and CLE diagnosis was compared with post-ESD histopathology. In histopathology after ESD, 11 of 35 lesions (31.5%) were adenomas and 24 (68.5%) were adenocarcinomas. The overall accuracy of CLE diagnosis of gastric adenomas and adenocarcinomas was significantly higher at 94.2% (95% confidence interval [CI], 81.3-98.4), versus 85.7% (95% CI, 70.6-93.7) for endoscopic biopsy (P = .031). The overall accuracy of CLE diagnosis of differentiated and undifferentiated adenocarcinomas also was higher (95.4%; 95% CI, 78.2-99.1) than that of endoscopic biopsy (84.2%; 95% CI, 62.4-94.4) but did not differ significantly (P = .146). Single tertiary-care center experience and small patient number. This study demonstrates the high accuracy of diagnosis of gastric epithelial neoplasia by using CLE. The use of CLE could possibly thus reduce the number of unnecessary biopsies and mistaken diagnoses before ESD.

Endoluminal full-thickness resection of GI lesions: A new device and technique

Full-thickness resection techniques are of growing interest in the field of endoscopic removal of tumors or their precursors in the digestive tract. A new dedicated full-thickness resection device has been developed based on the combination of the OTSC clip, an enlarged resection cap and an integrated snare. The device prototype allows combined resection of all gastrointestinal organ wall layers in one maneuver, up to a size of 3 x 3 cm. The device has been pre-clinically tested to demonstrate feasibility of full-thickness resection in the colon. Two clinical cases have been performed successfully to date.

Resection of lesions in the upper gastrointestinal tract: Results of one year of endoscopic mucosal resection and endoscopic submucosal dissection in a non-academic centre

Introduction: Endoscopic Resection of early neoplasias and their precursor lesions in the upper gastrointestinal tract has become standard in many endoscopy units. We report on the results, complications and outcome of patients referred to a non-academic centre for endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD) of upper GI lesions within one year.

Characteristics of Anemia in Elderly: A Hospital Based Study in South India

Anemia is a common concern in older people and can have significant morbidity and mortality. Because anemia is a sign, not a diagnosis, an evaluation is almost always warranted to identify the underlying cause. The purpose of this study was to study the clinical profile of elderly patients with anemia and to study characteristics of hematological types of anemia in such patients as well as the closest possible etiological profile. Hundred patients above the age of 60years were included in the study. Clinical profile with laboratory studies of Hemoglobin and diagnostic tests to fix the etiology. Majority of patients had normocytic blood picture. Renal failure was the most common underlying chronic disease. Significant number of patients were on non steroidal anti-inflammatory drugs which could contribute to the anaemia. 14% of the patients had an underlying malignancy. 73.3% of the patients in the microcytic group had an underlying GI lesion on endoscopy. Identifying anemia as an important aspect of a comprehensive geriatric assessment is absolutely essential further to clinical detection. Confirming the type of anemia is critical to direct the investigation for profiling the etiology since it is well known that the treatment of anemia goes a long way in improving the overall outcome and quality of life.

Prevalence of upper gastro intestinal endoscopy findings in endoscopy referred patients at Mosul city, a ten years retrospective study

Objectives: To analyze the findings of upper GI endoscopic examination carried out in order to assess the prevalence of upper GI disorders in the examined patients.Design and setting: A retrospective clinical case series study done over a 10 years period from the 1st of October 1998 to the 1st of October 2008, at the Endoscopy Unit of Al Salam Teaching General Hospital during which a total of three thousand seven hundreds and nineteen upper GI endoscopic examinations were performed and analyzed.Results: The mean age of our patients was 39.8 years and male to female ratio was 1.29:1 (males 56% and females 44%). Out of the total number of patients (3719), a sum of 1322 (35.5%) patients were typed as normal on endoscopic assessment while 2397 (64.5%) patient were found to retain abnormal findings with the duodenal lesions being the commonest 64%, followed by esophageal one 19%, then gastric lesions 17%. Duodenal ulcer was the most frequently recovered pathology [689 (29%) patients] followed by duodenitis [483 (20%) patients], then by gastroesophageal reflux disease [363 (15%) patients] that is followed by gastritis [312 (13%) patients].Conclusions: Upper GI lesions are frequent in Mosul population. The upper GI endoscopy is a valuable mean in the assessment of upper GI symptomatology. Keywords: GI, gastrointestinal.

"High" rate of upper GI lesions found in low-dose aspirin users

"High" rate of upper GI lesions found in low-dose aspirin users

Endoscopic submucosal dissection by using a flexible Maryland dissector: a randomized, controlled, porcine study (with videos)

Endoscopic submucosal dissection (ESD) is a minimally invasive method for en bloc removal of GI lesions. Current ESD methods and devices have limitations, including long procedure times, technical difficulty, safety, and availability. The aim of this study was to evaluate a blunt submucosal dissection technique and compare it with ESD by using the insulated-tip (IT) knife. Randomized, controlled, animal study. Animal facility laboratory. Twelve gastric lesions were marked by using electrocautery. After submucosal injection, a circumferential mucosal incision was created. ESD was then performed by using either the IT knife 2 (n = 6) or the flexible endoscopic Maryland dissector (n = 6). Median time for IT knife ESD versus median time for Maryland dissector ESD. Median time (IQR) for IT knife ESD was 43 minutes (range 36-50 minutes). The median time (IQR) for Maryland dissector ESD was 32 minutes (range 22-41 minutes; P = .09). The resection specimens obtained with the Maryland dissector tended to be larger, with a median (IQR) of 20.2 cm(2) (range 16.4-23 cm(2)) when compared with specimens resected with the IT knife, which yielded a median (IQR) resection area of 15.9 cm(2) (14.8-18.7 cm(2); P = .08). Complete en-bloc resection including all of the electrocautery markings was achieved in all cases. There were no perforations. There were two minor hemorrhages in the IT knife group and 3 in the Maryland dissector group. Nonsurvival setting, small sample size. The flexible Maryland dissector was demonstrated to be efficient, safe, and feasible for facilitating gastric ESD in a live animal model.

PTU-012 A novel submucosal injection solution: data on safety and efficacy from a large prospective EMR series

Introduction Submucosal injection is an integral part of EMR. The ideal solution will produce a long lasting high elevation, and an avascular field with delineation of tissue planes. This requires a combination of various drugs and solutions. However, no such mix is commercially available. This has resulted in a wide range of mix being used in this country, with no data on the safety, efficacy and long-term results of these solutions. We have developed an ideal mix of gelofusine (colloid) to provide a long lasting lift with adrenaline to reduce the vascularity of the field and indigo carmine to adequately delineate tissue planes. We aim to illustrate the feasibility, safety and efficacy and long-term effects of our novel submucosal lifting solution. Methods The lifting solution was: 500 mls gelofusine, 2 mls 1:1000 adrenaline and 5 mls indigocarmine 0.4%. We reviewed all EMRs since 2005. Records were assessed to collect data on efficacy of the solution, immediate and late mucosal and systemic effects. Data were collected on type and size of lesions, procedure time and volume of solution used. Results A total of 546 lesions were removed. 411 polypectomies and 135 upper GI EMRs. Volume used was between 5-250 mls, dependent on lesion size and procedure time (range 20 min–180 min). Distribution of upper GI lesions: 99 oesophageal EMRs, 20 gastric EMRs and 16 duodenal EMRs. Endoscopic follow-up data were available on 107 cases, with a mean interval of 3 months (range: 2–6 months). There was no evidence of immediate or late complications, including tissue injury, necrosis or stenosis related to the solution. For the lower GI lesions: 411 polyps removed. There were 143 polyps >2 cm in size. Left sided lesions=121 (median 45 mm range 20–150), right sided lesions=22 (median 31 mm, range 20–50). Procedure time 20 min–180 min. There were follow-up data on 117 colonic EMRs, with a mean interval of 4 months (range 0.5–60 months). There was no evidence of any late complications, including tissue injury necrosis or stenosis at the previous polypectomy sites. The endoscopist found the solution extremely effective and provided good lift after one injection in all of the cases examined. There was a short lived 5–20% rise in pulse rate. There were no immediate cardiac complications preventing completion of the procedures. No anaphylaxis or allergic reaction. There was no local or systemic sepsis related to the solution. No obvious renal or hepatic effects reported. Conclusion Largest reported series on submucosal injection solution in the world; Have clearly demonstrated the feasibility and efficacy of this novel injection mix; Shows immediate and long-term safety of this solution; Shows some systemic effects like tachycardia so caution is advised in cardiac patients.

Blue rubber bleb nevus syndrome

Blue Rubber Bleb Nevus Syndrome (BRBNS) is a rare condition characterized by multiple venous malformations involving the skin and internal organs. The gastrointestinal tract is always involved and intestinal haemorrhage is the most frequent clinical manifestation associated with iron deficiency anaemia. We describe a 10-year-old girl who, since birth, presented numerous venous malformations all over her body and a lymphangioma in the right leg. At the age of 5 years, she also had a severe episode of gastric bleeding requiring a blood transfusion. From this episode, she is suffering from chronic anaemia and this is the reason for admission into our hospital. The endoscopic examination of the gastrointestinal tract revealed multiple giant venous malformations in the oesophagus, stomach, duodenum and in all visible sections of the colon. Endoscopy is the gold standard technique for the diagnosis of BRBNS with GI lesions and also allows immediate therapeutic measures such as argon plasma coagulation, laser photocoagulation, sclerotherapy or band ligation. In addition, pharmacological treatments based on corticosteroids, interferon alfa, vincristine or octreotide have been described for BRBNS. Blue Rubber Bleb Nevus Syndrome is a congenital cutaneous and gastrointestinal haemangiomatosis. Its morbidity and mortality depends on involvement of visceral organs and particularly on GI bleeding. The treatment is based on pharmacological or surgical therapy. Overall, the most important step is the follow-up to the presence and the evolution of GI lesions and the possible bleeding.

Prevalence of gastric squamous ulceration in horses with abdominal pain

Prevalence of gastric ulcerations differs widely according to breed and circumstances of management. Further study of the biological variables involved is required in order to identify more exactly the reasons for the reported range. The objectives of this present study, which do not appear to have been addressed previously in the literature, were 1) the prevalence of gastric ulceration in horses with abdominal pain. 2) difference in prevalence in horses responding to medical therapy and those requiring surgical intervention. 3) whether gastric ulceration is associated with any particular gastrointestinal tract lesion. Horses were included in the study if gastroscopy was performed within 24 h of presentation. The presence and grade of gastric ulceration was recorded together with the medical records. Data were analysed categorically using a Fisher's exact test or Chi-squared test. One hundred horses met the selection criteria. Forty-nine percent (49/100) of horses had gastric ulceration, 63% (63/100) responded to medical therapy and 37% (37/100) to surgical intervention, and prevalence was higher in the former (59%) than the latter (32%). Horses with duodenitis-proximal jejunitis (DPJ) had a trend towards higher prevalence of gastric ulceration compared to those with other GI lesions. Sixty-eight percent (13/19) of horses diagnosed with DPJ, 32% (8/25) with a large colon impaction and 14% (1/7) with large colon volvulus had gastric ulceration. The presence of gastric ulceration in all horses with abdominal pain was moderate. Horses responding to medical therapy had a higher prevalence of gastric ulceration compared to horses requiring surgery and there was a trend towards higher prevalence in cases of DPJ compared to other GI lesions. There was not a statistically significant difference in gastric ulceration detected between specific lesions, including large colon impactions and large colon volvulus. The clinical relevance of ulceration is still unclear and further studies are required to differentiate between incidental and clinically important gastric ulceration.

Endoluminal resection for sessile neoplasia in the GI tract is associated with a low recurrence rate and a high 5-year survival rate

Endoluminal resection (ELR) is an alternative to surgery for the removal of superficial neoplastic lesions of the GI tract. We previously reported a complete resection (CR) rate of 89% when using ELR techniques. However, the long-term efficacy of ELR for the eradication of sessile lesions, as well as patient survival after ELR, is not known. To assess the long-term recurrence rate of GI lesions that were completely resected by ELR and to determine the long-term (5-year) survival rate of patients who had ELR for GI lesions that contained high-grade dysplasia (HGD) or adenocarcinoma (AdCa). From a larger cohort of 92 consecutive patients who had undergone ELR of sessile lesions from 1997 to 2000, long-term follow-up was obtained. Patient survival rates were considered in terms of disease-specific mortality. Endoscopic follow-up was possible in 44 patients (48%) and 46 lesions (46%). At the time of the initial ELR, the median age was 68.5 years. The median duration of follow-up was 3.8 years after ELR (range 7 months to 8.1 years). Resected lesions were located in the esophagus (25%), stomach (11%), duodenum (25%), colon (27%), and rectum (11%). Post-ELR histopathology consisted of adenomas (46%), HGD (16%), AdCa (11%), lymphoma (2%), leiomyoma (2%), carcinoid (2%), lesions indeterminate for dysplasia (2%), and nondysplastic lesions (18%). CR was achieved in 42 patients and 44 lesions (96%) after initial ELR procedures (range 1-3). Thirty-seven of 39 neoplastic lesions (95%) in this cohort achieved initial CR. Two lesions (found in 2 patients) that did not achieve initial CR were excluded from analysis. Four of 37 neoplastic lesions (10.8%) that underwent successful ELR had local recurrence of neoplasia (median time to recurrence was 1.9 years). Two of the 4 neoplastic recurrences were successfully eradicated by subsequent endoscopic techniques (95% overall eradication rate). Long-term endoscopic follow-up was possible in 12 of 20 patients (60%) with HGD or AdCa who achieved initial CR. These 12 patients had HGD (n = 7) or AdCa (n = 5) and were observed for an average of 4.9 years (range 1.4-7.9 years) after an ELR. Two of these patients died during follow-up, but only 1 death (8%) was lesion related. When comparing post-ELR pathology, HGD and AdCa trended toward an increased risk of recurrence compared with other neoplastic lesions (hazard ratio 4.75 [95% CI, 0.49-46.35], P = .18). A retrospective study with 52% of patients lost to long-term endoscopic follow-up, which limited the sample size and the number of events. ELR effectively eradicates sessile neoplastic lesions from the GI tract, with a low recurrence rate (10.8%) at long-term follow-up. Post-ELR surveillance is important, because local recurrences may be amenable to endoscopic eradication. HGD and cancers may carry an increased risk for a local recurrence, with a 5-year disease-specific survival of 92%. This study suggests that ELR is associated with a low recurrence rate for neoplastic lesions and appreciable long-term survival in patients with highly dysplastic lesions. Additional, long-term follow-up studies are necessary to investigate the role of ELR in the treatment of highly dysplastic lesions of the GI tract.

Diagnostic Yield of Capsule Endoscopy in Elderly Patients with Obscure GI Bleeding and Predictive Factors of Rebleeding: A Prospective Multicenter Study of the ANGH Group

Introduction: Capsule endoscopy (CE) is widely used to investigate obscure GI bleeding (OGIB) from small bowel origin. Aims: The aim of this prospective multicenter study was 1) to determine the findings and the diagnostic yield of CE in overt and occult OGIB in elderly patients and 2) to identify predictive factors associated with rebleeding. Methods: patients older than 70 years who had a previous negative endoscopic evaluation (at least one upper endoscopy and one colonoscopy) for OGIB and a rebleeding were eligible for inclusion. Rebleeding was defined as a drop in hemoglobin level from the high one of 2 g/dL or more, or by a new overt bleeding. A multivariate analysis was performed to identify predictors of lesions and to determine factors associated with rebleeding. Six French hospitals participated in this prospective multicenter study from January 2006 to April 2007. Results: 76 patients (42 women and 34 men, 76.6 ± 9 yrs) were included; 47 had an occult and 29 an overt OGIB. 73% of patients had one or more severe comorbidity. 40% of patients were taking one or more drug-induced GI lesion [NSAIDs (n = 2); aspirin (n = 12); vitamin K antagonist (n = 11), clopidrogel (n = 10)]. Visualization of the entire small bowel was achieved in 90%. No complication related to CE occurred. A lesion was identified in 49 (63%) cases, angiodysplasias were observed in 73% of cases. The yield (define by OGIB related to lesion) of CE was 58%, resulting in a potential specific treatment in 82%. The yield of CE for occult OGIB was 55% vs. 65% for overt OGIB (p = 0.7). Predictive factors of small bowel lesions were (OR [IC], p): male sex (6.7 [1.5-30]; 0,014) and cardiac disease (8.2 [1.7-39], 0.008). A specific treatment was possible in 40 (82%) of cases. 48 (63%) patients were followed up during a median period of 461 days (96-636 days). A rebleeding was observed in 11 (23%) cases. A trend towards but not significant, because of small sample size, to rebleeding was observed in case of 1) overt OGIB (36% vs. 18%), 2) identified lesion (27% vs. 14%) and 3) untreated lesion (42% vs. 18%). Conclusion: the diagnostic yield of CE is high in elderly patients with OGIB allowing an appropriate treatment in most of cases. Predictors of GI lesion identified by CE were male sex and a past history of cardiac disease. A rebleeding was observed in 11 (23%) patients, but this small sample size do not allowed to clearly identified factors of relapse.